U.S. flag

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

  • Publications
  • Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

  • Advanced Search
  • Journal List

Logo of springeropen

Personality Psychology: Lexical Approaches, Assessment Methods, and Trait Concepts Reveal Only Half of the Story—Why it is Time for a Paradigm Shift

Comparative Differential and Personality Psychology, Freie Universität Berlin, Habelschwerdter Allee 45, 14195 Berlin, Germany

This article develops a comprehensive philosophy-of-science for personality psychology that goes far beyond the scope of the lexical approaches, assessment methods, and trait concepts that currently prevail. One of the field’s most important guiding scientific assumptions, the lexical hypothesis, is analysed from meta-theoretical viewpoints to reveal that it explicitly describes two sets of phenomena that must be clearly differentiated: 1) lexical repertoires and the representations that they encode and 2) the kinds of phenomena that are represented. Thus far, personality psychologists largely explored only the former, but have seriously neglected studying the latter. Meta-theoretical analyses of these different kinds of phenomena and their distinct natures, commonalities, differences, and interrelations reveal that personality psychology’s focus on lexical approaches, assessment methods, and trait concepts entails a) erroneous meta-theoretical assumptions about what the phenomena being studied actually are, and thus how they can be analysed and interpreted, b) that contemporary personality psychology is largely based on everyday psychological knowledge, and c) a fundamental circularity in the scientific explanations used in trait psychology. These findings seriously challenge the widespread assumptions about the causal and universal status of the phenomena described by prominent personality models. The current state of knowledge about the lexical hypothesis is reviewed, and implications for personality psychology are discussed. Ten desiderata for future research are outlined to overcome the current paradigmatic fixations that are substantially hampering intellectual innovation and progress in the field.

Science provides a special way of constructing knowledge about the world. Unlike nonscientific knowledge construction, science provides a way of thinking simultaneously about phenomena and the means of producing knowledge about them—this presupposes meta-theory and methodology (Althusser and Balibar 1970 ; Toomela 2011 ). Meta-theory refers to the philosophical assumptions about the theoretical nature of the phenomena to be studied and to the questions that are asked about them. Methodology refers to the ways (i.e., approaches) in which these questions can be answered and to the techniques (i.e., methods) that can therefore be used (Sprung and Sprung 1984 ).

All sciences have meta-theories. They determine which elements of real phenomena can be reduced to precisely those subsets of elements that are considered relevant to and defining of concrete scientific phenomena and the ways in which they can be reduced (Althusser and Balibar 1970 , p. 84, analysing the philosophy of science of Marx 1867 ; Køppe 2012 ; Weber 1949 ). Meta-theories determine what is considered data in a particular field (in a particular historical time; Kuhn 1962 ), and how the thus-defined data can be analysed and interpreted (Køppe 2012 ; Wagoner 2009 ). Hence, the first step of reducing real elements into facts is already a theoretical decision (Weber 1949 , p. 173). “Alles Faktische ist schon Theorie”—all facts are already theory (Goethe 1907 , p. 127). In other words, “it is the theory which decides what can be observed” (Einstein to Heisenberg in 1926, cited in Heisenberg 1989 , p. 10). Meta-theories and methodologies are the rules that govern the effective practices of sciences. Therefore, they should be explicated to enable researchers to constantly scrutinise all levels from epistemology and ontology up to the specific theories about the phenomena studied (Toomela 2011 ). This also includes rethinking the very reduction of real phenomena into scientific phenomena (Utz 2005 ).

Contemporary psychology largely follows meta-theory and methodology implicitly, sometimes even “blindly” (Toomela 2011 , p. 22)—this occurs in particular when research methods are decided first and research questions are adapted to the methods rather than vice versa (Omi 2012 ; Westen 1996 ). “The understanding that research methodology comprises an essential part of scientific theories about phenomena that are studied is not always brought into the center of theoretical reasoning” (Toomela 2009 , p. 45). Ignoring the meta-theoretical assumptions underlying particular methods can result in mismatches between methods and research questions that hamper the scientific understanding and the explanations of the phenomena being studied—and thus scientific progress (Loftus 1996 ; Toomela and Valsiner 2010 ; Weber 1949 ). More profoundly, a priori decisions on methods preclude posing the primary question of what the phenomena to be studied actually are (Toomela 2011 ).

These meta-theoretical and methodological challenges are characteristic of contemporary personality psychology and of taxonomic personality research in particular (Uher 2008a , b , 2011a , b ; Uher, Methodological approaches to personality taxonomies: The Behavioural Repertoire x Environmental Situations Approach—A non-lexical alternative, unpublished). The “discovery” of five major dimensions of individual differences is considered a milestone in modern Western psychology (De Raad 1998 ; Digman 1990 ; Goldberg 1990 ). This “break through” was made possible by the assumption that people encode in their everyday languages all those individual differences that they perceive as most salient in everyday encounters and that they consider to be socially relevant. This so-called lexical hypothesis, first articulated by Galton ( 1884 ), has provided a stringent rationale for using the lexica of human languages as finite sources of information to unravel a few major dimensions of individual differences (Allport and Odbert 1936 ; Cattell 1943 ). In English and some other languages, the reduction of the pertinent lexical repertoires to five major dimensions has received the most support in Western scientific communities (Goldberg 1993 ; John et al. 1988 ). The lexical hypothesis also suggests that enquiring about the everyday psychological ideas that people develop of themselves and of other individuals could be a suitable method for the scientific measurement of personality (Block 2010 ; Westen 1996 ). Assessments by laypeople have become the standard methods of investigation (Baumeister et al. 2007 ; Matthews et al. 2003 ) and the “primary source of data” in personality psychology (Schwarz 1999 , p. 93). The assumptions expressed in the lexical hypothesis might also contribute to the fact that the prevailing strategies of the scientific explanation of individual differences largely follow structures that are deeply rooted in everyday psychology.

In spite of its enormous importance as one of the most widely used theoretical assumptions to have guided personality psychology (Ashton and Lee 2005 ), the lexical hypothesis has remained untested (Toomela 2010a ; Westen 1996 ), and still today, its statements have been considered only partially. This article systematically explores the meta-theoretical assumptions that underlie this hypothesis. It starts by highlighting the explicit reference of the lexical hypothesis to two sets of phenomena that must be clearly differentiated; these are, on the one hand, people’s lexical repertoires and the representations that these repertoires encode, and, on the other hand, the kinds of phenomena that are perceivable in everyday life and that are being represented. Thus far, personality psychology has focused primarily on just one of these two sets—on lexically encoded representations—but has failed to systematically investigate the second set, namely, the kinds of phenomena that are being perceived and represented.

This article elaborates the distinct natures of the different kinds of phenomena to which the lexical hypothesis refers and explores their commonalities, differences, and interrelations from a philosophy-of-science perspective. This allows for scrutiny to be applied to the phenomena that are being lexically encoded and sheds new light on the fundamental questions of which kinds of phenomena can actually be captured by assessments—and which ones cannot. These analyses reveal that erroneous meta-theoretical assumptions underlie the established beliefs about what these phenomena actually are. They show that contemporary personality psychology is largely based on everyday psychological knowledge. Furthermore, the present meta-theoretical analyses allow us to scrutinise the prevailing psychological strategy of explaining individual differences—explicitly or implicitly—by assuming the existence of “traits”. The philosophy-of-science perspective identifies explanations based on trait concepts as fundamentally circular. These findings seriously challenge the established assumptions about the causal and universal status of the phenomena described by prominent personality models.

Various researchers have carefully crafted serious concerns about the established five factor models of personality (e.g., Block 1995 , 2001 , 2010 ; Eysenck 1992 ; McAdams 1992 ; Westen 1996 ), about assessment data and their interpretation (e.g., Brower 1949 ; Michell 1997 , 2003 ; Omi 2012 ; Rosenbaum and Valsiner 2011 ; Schwarz 2009 ; Trendler 2009 ; Valsiner 2012 ), and about the explanations provided by trait psychology (e.g., Allport 1961 ; Bock 2000 ; Cervone et al. 2001 ; Lamiell 2003 ; Mischel and Shoda 1994 , 1995 ). This article moves these critical objections coherently to meta-theoretical and methodological levels of consideration. It expands them substantially from the philosophy-of-science perspective and elaborates a fundamental critique of the lexical approaches, assessments by laypeople, and trait concepts that currently dominate personality psychology. It closes by outlining 10 desiderata for future research to overcome the limitations revealed in the analyses and to stimulate new directions in the field.

Lexical Encodings—Constructs and Representations

The lexical hypothesis states that people encode in their everyday languages all those differences between individuals that they perceive to be salient and that they consider to be socially relevant in their everyday lives. Encodings about differences within individuals over time are not explicitly mentioned, however (see below). Among the many phenomena that individuals can perceive, recurrent patterns are particularly meaningful because such patterns may allow predictions of future events while facing the uncertainty of the future. Individuals therefore seek to identify recurrent patterns in their experiences with their personal world (Kelly 1955 ). To describe, explain, and to predict the likely occurrences of events, people construct and represent their experiences in private ideas (Valsiner 2012 ). These ideas are called subjective or individual representations (Jovchelovitch 2007 ; Moscovici 1984 ). Representations referring to experiences with one’s own person and the—especially social—world are called personal constructs in personality psychology (Kelly 1955 ).

To communicate their perceptions and mental constructs of what they perceive and to negotiate socially shared meanings in order to cope with the world collectively, groups of individuals create and use social constructs. These constructs are called intersubjective or social representations (Jovchelovitch 2007 ) or folk concepts in personality psychology (Tellegen 1993 ). Constructs and representations referring to the perceptions of persons are called personality constructs . The lexical hypothesis states that, over time, socially shared constructs of self- and other-perception become encoded in the natural human languages. Pertinent lexical encodings reflect the body of everyday psychological ideas, beliefs, values, and practices that people have developed about individuals (Block 2010 ; Wagner et al. 1999 ; Westen 1996 ).

What is Constructed and Represented—Phenomena and Patterns

What is it that is constructed and represented as “personality”? Psychological definitions of personality—a few prominent ones picked out of many—provide the first insights into what scientific psychologists construct as personality. Some frequently cited definitions refer fairly generally and descriptively to personality as

those characteristics that account for a person’s consistent patterns of feeling, thinking, and behaving (Pervin and John 1997 , p. 4) or as
an individual’s characteristic patterns of thought, emotion, and behaviour, together with the psychological mechanisms–hidden or not–behind those patterns (Funder 2004 , p. 5).

These definitions include psychological and behavioural phenomena. Allport ( 1937 ) also included psychophysical systems and environmental adaptivity by defining personality as

the dynamic organization within the individual of those psychophysical systems that determine his unique adjustments to his environment (p. 48).

Eysenck ( 1947 ) furthermore specified genetic and environmental causality, biological systems, and ontogenetic development when he defined personality as

the sum-total of the actual or potential behaviour-patterns of the organism, as determined by heredity and environment; [that] originates and develops through the functional interaction of the four main sectors into which these behaviour-patterns are organized: the cognitive sector (intelligence), the conative sector (character), the affective sector (temperament), and the somatic sector (constitution) (p. 25).

McCrae and Costa ( 2008 ) extended the scope of consideration by conceptualising particular personality constructs as reflecting “universals” of human nature that are “invariant across human cultures” (McCrae and Costa 1997 , p. 510), and thus also potentially phylogenetic in origin (McCrae 2009 ).

These definitions are sufficient for highlighting fundamental issues. First, they refer descriptively to particular patterns (e.g., those described as “characteristic” and “consistent”) and to the different kinds of phenomena in which these patterns occur (e.g., physiological, psychological, and behavioural phenomena). Second, they include assumptions about the causation of the described patterns in the described phenomena by other kinds of phenomena internal to the individual (e.g., psychological phenomena) and/or external (e.g., environmental conditions) considering various explanatory perspectives (e.g., proximate, adaptive, ontogenetic, phylogenetic). Astonishingly, there is no mention of people’s lexical encodations and representations of individual differences although these are central to the lexical hypothesis.

From philosophy-of-science perspectives, the different kinds of phenomena studied with regard to personality must be clearly differentiated. This is essential because interrelations among them can be untangled only if they are explored each in their own right and if a priori assumptions about specific interrelations are avoided. This is rarely done in psychology. Behaviourists have focused too much on external conditions and behaviours, but have largely ignored psychological phenomena. Subsequent researchers have tried to overcome the limitations of behaviourism and have therefore focused strongly on psychological phenomena and on individuals’ representations of their world, but behaviour has faded into the background. The prevailing investigative strategies are preoccupied with causal processes and the conditions in which behavioural phenomena occur—especially psychological ones—with the result that clear differentiations of psychological and behavioural phenomena are missing.

The meta-theoretical definition of behavioural phenomena as “external activities or externalisations of living organisms that are functionally mediated by the environment (Millikan 1993 ) in the present” (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished) allows such differentiations. This definition highlights the intrinsic relatedness of behaviour to properties of the immediate external environment that are defined as environmental situations and emphasises that it is inherently bound to the present. Externality differentiates behaviour from thoughts, emotions, and other psychological phenomena, which are also bound to the present (Gillespie and Zittaun 2010 ; Toomela 2010a ; Valsiner 1998 , 2012 ) but are internal to the individual (Toomela 2008 ). This meta-theoretical definition generally defines all behavioural phenomena without specifying particular ones (e.g., goal-directed action or response) because many of these concepts include a priori assumptions about causally related internal processes (e.g., psychological phenomena) and external conditions in the environment (e.g., particular stimuli) that are separate kinds of phenomena. To scrutinise which kinds of phenomena people can perceive in everyday life and construct as personality, it is helpful to first specify the particular patterns referred to in definitions of personality.

Defining Patterns

For scientific definitions, the notions of “a person’s consistent patterns”, “individual characteristics”, and “individual uniqueness” are surprisingly vague. They do not specify what is meant to be “consistent” with what, nor do they indicate which patterns are considered “different, “unique”, and “characteristic” and why. Moreover, they fail to mention within-individual variability and structural complexity—both within and between individuals. This is remarkable because these patterns refer to phenomena (e.g., those of the psyche and of behaviour) that are heterogeneous, complex, dynamic, and thus highly fluctuating. In such phenomena, differences among individuals are necessarily apparent at any given time (Uher 2011a ) and determinations of meaningful consistency—both within and between individuals—are matters of mere convention. Meta-theoretical analyses help to carve out basic criteria for determining which patterns can be considered specific to particular individuals, and thus different from others, which ones cannot, and why. So what does individual-specificity mean?

Probabilistic Patterns

In the steadily fluctuating flow of events in highly dynamic phenomena, individuals necessarily show considerable within-individual variability over time. Consequently, they can be characterised only by probabilistic patterns. What types of probability 1 are able to describe these patterns? First, events of psychological and behavioural phenomena are not random, such as events in throws of dice. There are no equipossible elementary events in well-defined sample spaces that can be studied in repeated experiments as in Laplacian theory. Moreover, the events are not independent from one another—this applies to events of the same phenomenon (e.g., several smiles) as well as to events of different phenomena (e.g., a smile and a laugh). Rather, different events may co-occur and even depend upon one another, particularly if the phenomena are functionally similar or related. Their events need not be exclusively disjunctive (e.g., smiles and laughs). Thus, finite (limiting) frequencies, which are defined as empirical occurrences relative to all possible occurrences (von Mises 1928 ), are inadequate for describing individual probabilities in psychological and behavioural phenomena.

The concept of propensity probabilities seems to be better suited for this purpose. Propensities are considered properties of physical objects or situations that are assumed to cause empirical frequencies, even if only single-case observations are made (Popper 1959 ). They are therefore conceived also as properties of repeatable conditions that determine empirical frequencies in the long run (Gillies 2000 ). 2 But this very assumption precludes the idea that propensity probabilities can be applied to individuals because their defining property as living systems is the fact that they undergo continuous changes during ontogenetic development. This is true for the molecules in their bodies and the phenomena constructed as personality alike. In fact, personality is assumed to develop and change gradually over the course of the lifespan (e.g., Caspi and Roberts 2001 ; Cattell 1950 ; Soldz and Vaillant 1999 ). This implies that the individual’s empirical frequencies cannot converge (i.e., stabilise) in the long run as may be true for physical systems. By contrast, the probabilities themselves change over time—an idea typically not considered by theories of propensity probabilities (cf. Gillies 2000 ; Popper 1959 ).

Further differences are essential. In psychological and behavioural phenomena, there are no discrete events such as there are in coin tosses. Well-defined natural entities on which probability estimations could be based are lacking. Instead, it is the scientists who must decide which entities of real phenomena make up particular scientific phenomena. The ways in which they define entities as categories depend not only on the phenomena themselves, but also on the purposes of their reduction (Weber 1949 ). Moreover, the processual character and microgenetic development of many psychological and behavioural phenomena (Rosenthal 2004 ) entails that their occurrences can be construed as frequencies only in some cases. In other cases, they may be construed more accurately as durations of nonfixed and often highly variable length (or as latencies, which are durations with specified start times).

Consequently, in fluctuating phenomena such as those of the psyche and of behaviour, events can be considered only by their empirical occurrences (i.e., in terms of frequencies and durations) relative to particular time periods. This new probability type will therefore be called a time-relative probability . For example, in observations of free play in a kindergarten group, a time-relative probability of 10 min per hour of observation time during which a child plays by him/herself may be determined for a particular 5-year-old child. The time periods in which time-relative probabilities are determined must be specified precisely—both in terms of the occasions and spans of time in which the data were collected to determine the probabilities and in terms of historical times in the individual’s ontogeny. Both are essential to define individual-specificity. For example, the 10 min per hour probability could have been determined from 20 h of regular observation obtained across a time period of 2 weeks; the ontogenetic stage of the individual is specified by an age of 5 years, which can be categorised as early childhood.

This new probability type enables ratio-scaled quantifications that are essential for quantitative comparisons across time and situational contexts and between individuals, groups, populations, and species (Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality; An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, both unpublished). It also allows for the merging of concepts of averages across occasions, as studied in personality research, with concepts of ranges , especially maxima, as studied in intelligence and achievement research 3 (Ackerman 1994 ), and concepts of variability in terms of differences between consecutive occasions (i.e., fluctuations) as studied in physiological and behavioural research (De Weerth et al. 1999 ). For example, the child’s time-relative probability could be on average 10 min per hour, ranging from 5 to 20 min per hour on a daily basis; the magnitude of day-to-day variability (for comparable occasions across the days) as indicated by the coefficient of variation could be CV  = 0.5. This coefficient specifies the standard deviation standardised by the mean (because the standard deviation is sensitive to the sample mean) to allow for comparisons between different samples (for an overview of methods of analyses of variability, see van Geert and van Dijk 2002 ). Such comprehensive analyses of patterns are important because within-individual variability in psychological and behavioural phenomena is pronounced and bears theoretical and empirical importance. For example, within-individual variability can indicate ongoing processes and can reflect phenomena that are important for explaining changes and development (van Geert and van Dijk 2002 , p. 344). Moreover, within-individual variability often substantially exceeds between-individual variability (Shweder and Sullivan 1990 ; Uher, An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, unpublished).

Differential Patterns

Time-relative probabilities that characterise all individuals in the same way cannot be individual-specific. If all children in the observed group have the same time-relative probability of self-playing, then these probabilities cannot characterise any one of them individually. Time-relative probabilities can reflect individual uniqueness only if they deviate from those of other individuals of a particular reference group of interest (e.g., a social group, culture, or species)—that is, if they are differential . This is the case if the determined time-relative probabilities differ between children such that their individual averages vary interindividually, for example, between 2 min and 30 min per hour, their individual ranges vary, for example, between 15 min and 50 min, and their individual within-individual variabilities vary, for example, between CV  = 0.3 and 1.3.

The concept of time-relative probabilities presupposes that the number of occasions and periods of time considered for the probability estimations are comparable between individuals. This is important when considering the effects of aggregation on the reliability of probability estimations (cf. Spearman 1910 ) and when taking into account the fact that the individuals’ probabilities themselves may gradually change over time (i.e., develop). In the given example, this means that the time-relative probabilities of the other children should also be estimated based on 20 h of observation across a time span of 2 weeks. These children should also be of similar age in order to disentangle differences between individuals from differences between age groups. Furthermore, the situational contexts in which different individuals are studied should be comparable between them because the average time-relative probabilities of particular groups of individuals may generally shift across different situations (Uher, An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, unpublished.) Analyses of differentiality therefore require statistical standardisation of time-relative probabilities within the time periods and within the situational contexts in which they are determined (Uher 2011a ). Statistical standardisation removes the information on absolute time-relative probability scores, however, because the data are converted into relative, but still ratio-scaled data. Yet, the absolute scores can always be traced when needed for interpretation and for later comparisons with other samples or with future investigations of the same sample.

Temporal Patterns

Differential patterns in time-relative probabilities that change rapidly cannot characterise an individual’s uniqueness. To reflect individual-specificity, they must be stable across time periods that are longer than those in which the probabilities were first ascertained. As with the analyses of differential patterns, the analyses of temporal patterns presuppose that probability estimations are based on numbers of occasions and periods of time that are comparable between the individuals being studied and between the time periods being contrasted. In the above example, this means that the differential patterns in the children’s time-relative probabilities for self-play should be similar when estimated again some time later in a second period of 2 weeks. Temporal stability is an essential prerequisite for justifying interpretations of the obtained standardised aggregate scores as reflecting individual-specific patterns (Uher 2011a ).

Temporal stability can be analysed for differential patterns in the individuals’ averages, ranges, and variabilities. For example, it is quite possible that individuals with similar differential scores in their average probabilities will differ from one another in stable ways with regard to their ranges and patterns of within-individual variability. The magnitude of temporal stability considered meaningful to construct individual-specificity is necessarily a matter of convention that can and must be explicitly defined. It depends on the ontogenetic stages of the individuals under study as well as on their species-specific life expectancy (Uher 2009 ). It also depends on the phenomena being studied and the meta-theory used. For example, in biology, weak test-retest correlations of individual differences in behaviour across just a few days or weeks of r  = .20 are considered ecologically and evolutionarily meaningful (Sih and Bell 2008 ), whereas in human psychology, such weak correlations are considered indicators of unreliable measurements of individual-specific patterns (Uher 2011a ).

In conclusion, for fluctuating and dynamic phenomena such as those of the psyche and behaviour, individual-specificity refers to differential patterns in time-relative probabilities that are stable in ways that are considered to be meaningful . This concept fits well into Brunswik’s ( 1955 ) probabilistic theory for functional psychology. Brunswik ascertained that although the environment may be lawful in terms of physical principles, to the individual, “it presents itself as semi-erratic … therefore all functional psychology is inherently probabilistic” (Brunswik 1955 , p. 193). An empirical application of this meta-theoretical concept of individual-specificity to behaviour was demonstrated in a study on capuchin monkeys based on 141 contextualised behavioural variables (Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished).

Individual-specific patterns that have been identified in these ways can be subsequently studied across longer periods of time to explore patterns in their gradual change and ontogenetic development (Uher 2011a ). This presupposes again that the occasions and time spans as well as the historical times that are studied should be specified and should be comparable between the individuals under study.

The concept of time-relative probabilities allows for the close examination of individual cases in specified situations. By studying many individuals in this way, it also allows for the study of lawful patterns at the sample level. The former type of study is commonly referred to as ideographic; the latter one as nomothetic. But ideographic approaches in terms of single case studies in and of themselves need not be informative about particular individuals’ specificity because they fail to disentangle individual-specific from group-, population-, or species-specific patterns (Uher, An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, unpublished). Single case studies are informative about individual-specificity only if they are based on the meta-theoretical assumption that individuals generally differ from one another in a given population. This is commonly assumed for humans, but not necessarily for individuals of other species. Single case studies of, for example, cockroaches can also be based on the assumption that all individuals are alike. Thus, studying one individual is assumed to provide information about all individuals, but not necessarily about the peculiarities of a particular one. Ideographic approaches are therefore informative about individual-specificity only if they are conceptualised as ideographic-nomothetic approaches such as in the above-introduced concept of individual-specificity that is based on differential patterns in time-relative probabilities that are stable in meaningful ways.

Up to this point in the article, psychological and behavioural phenomena have been discussed only exemplarily to illustrate the meta-theoretical analyses of individual-specificity in fluctuating phenomena. I will now scrutinise the particular kinds of phenomena in which individual-specific patterns are constructed as personality. According to the lexical hypothesis, the lexically encoded representations of individual-specificity refer to phenomena that are (1) perceivable and salient and (2) important in everyday life, in particular in social encounters (John et al. 1988 ). Systematic explorations of the kinds of phenomena to which these criteria apply encounter two fundamental challenges.

Fundamental Challenges

With regard to phenomena that people perceive and construct as personality, one must consider that individual-specific patterns cannot be directly perceived at any given moment in any given individual because they refer to probabilistic, differential, and temporal patterns. This has far-reaching implications for their investigation in particular in phenomena as dynamic and fluctuating as those of the psyche and of behaviour (discussed in detail below; Michell 1997 , 1999 ; Rosenbaum and Valsiner 2011 ; Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished).

With regard to people’s constructs and representations of these phenomena, one must consider that the construing activity of the human mind leads people to overlook and to transcend gaps in perception (Brunswik 1952 ; Daston and Galison 2007 ; Valsiner 2012 ). These are challenges for scientific psychologist as well. In contrast to scientists in many other disciplines, personality psychologists study phenomena that are important in everyday life—including their own—in individuals of their own kind. They start researching these phenomena and such individuals only after having acquired a substantial everyday psychological knowledge base and a pertinent lexical repertoire in their pre-academic lives (Uher 2011a ). Their everyday psychological knowledge and the sociocultural context in which it is embedded may inevitably influence and “confuse” their thinking as researchers (James 1893 , p. 196; Weber 1949 , p. 54)—despite all scholarly efforts (Jovchelovitch 2007 ; Komatsu 2012 ; Valsiner 2012 ).

These challenges, which will be further explored below, require precise specification and clarification of the central statements of the lexical hypothesis. To identify individual-specificity, (1) people must be able to directly perceive the phenomena in question in individuals. Only then can they consider the occurrences of their events and compare these between individuals to identify differential patterns in time-relative probabilities; and only then can they consider temporal patterns therein—the three meta-theoretical criteria of individual-specificity in fluctuating and dynamic phenomena (Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished). This is also implied by the assumption that it is salient phenomena that become lexically encoded over time—salience presupposes direct perceivability. The notion of perceiving does not imply intention, purpose, or conscious awareness—in contrast to observing. (2) Directly perceivable phenomena can become socially relevant and important in everyday life only if these phenomena are also directly involved in the individuals’ interactions with and in their relations to their—in particular social—environment. If particular phenomena cannot directly interact with an individual’s environment but depend completely on mediation through other phenomena, then these former phenomena in and of themselves cannot become socially relevant and important in everyday life without the phenomena that mediate them . Thus, they are secondary. For example, by itself, a private thought cannot become effective in an individual’s life unless the individual eventually externalises it in language or acts upon it (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished). This issue will be discussed in detail below.

Scrutinising Phenomena

To which kinds of phenomena do these two criteria—(1) direct perceivability and (2) direct involvement in the individual’s environmental relations—apply? The lexical hypothesis specifies no particular kinds of phenomena, but the above-mentioned psychological personality definitions do.

Psycho-Physiological Phenomena

Some definitions specify psycho-physiological phenomena. Individuals cannot directly perceive any of these phenomena, however, either in themselves or in any other individual (criterion 1 failed). Moreover, psycho-physiological phenomena may be involved only indirectly in the individuals’ relations to their environment and depend on mediation by other phenomena, such as of physiology (e.g., metabolic) and of behaviour (criterion 2 partly failed; see Fig.  1 ).

An external file that holds a picture, illustration, etc.
Object name is 12124_2013_9230_Fig1_HTML.jpg

Behaviour—The essential bridge from the psyche to the environment. Note. The box in the dark frame on the left side of the figure and the two smaller boxes that mirror its shape and composition on the right side symbolise individuals. The arrows indicate the directions of the interactions that are possible between the different types of phenomena considered; in the smaller symbols, only some of them are depicted. Of the kinds of phenomena frequently mentioned in personality definitions, phenomena of behaviour and of outer appearance are the only ones that fulfil both criteria implied by the lexical hypothesis—i.e., (criterion 1) they are directly perceivable by individuals in themselves and in other individuals, and (criterion 2) they can be directly involved in the individuals’ interactions with and their relations to the environment. Behaviour is the essential interface between an individual’s psychological phenomena, which are entirely internal, and that individual’s (abiotic and biotic, especially social) environment. Those environmental details that functionally mediate the individual’s externalisations in the present (i.e., the individual’s behaviour) are called the environmental situation (indicated with a dotted line). In the graphic, the lower of the two individuals (indicated as small boxes) is part of the environmental situation of the individual on the left side (indicated as large box), whereas the upper one is not

Psychological Phenomena

Psychological phenomena, such as those described as emotions, thoughts, motives, beliefs, attitudes, and many others are essential elements of many psychological personality definitions. But all these phenomena are entirely internal (Toomela 2008 ). They cannot directly interact with any of the individual’s external systems (Bronfenbrenner 1979 ). Hence, by themselves , psychological phenomena cannot be directly involved in the individuals’ environmental relations (criterion 2 failed). And because they are entirely internal, people can perceive them only in themselves through introspection (Wundt 1904 ), but they are unable to perceive psychological events in any other individual (Locke 1689 ; Toomela 2008 , 2011 ; criterion 1 semi-fulfilled). The psychological events of others can be inferred only from events of external—and thus directly perceivable—phenomena that may mediate between their internal and external systems, such as from behavioural phenomena (see Fig.  1 ), including parts of language (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished).

Difficulties arise, however, because these mediating phenomena may not reflect psychological phenomena unequivocally or accurately. Similar behavioural phenomena can be related to different psychological phenomena and different behavioural phenomena to similar psychological phenomena (Cervone et al. 2001 ; Kagan 1994 , 1998 ; Lewin 1935 ; Mischel and Shoda 1994 ). Clear meta-theoretical differentiations between behavioural and psychological phenomena are therefore essential (see above). Moreover, differences in behavioural structures may be but are not necessarily associated with differences in psychological structures (Toomela 2011 ). Further complicating the matter is the fact that individuals must adapt their mediating behaviours to the (subjectively interpreted; Rotter 1954 ) realities of the physical and social environment—“acting is not the same as thinking” (Lahlou 2008 , p. 21, translated). This means that individuals can (and must) control the mediators to some extent depending on the given contexts. This substantially complicates the ability to make inferences to underlying psychological events. The reduced ability shown by babies and young children to control behaviour may result in stronger relations between the events of their psychological phenomena and those of their behavioural phenomena. Yet their inability to provide self-reports hinders our ability to “validate” any inferences that we make to their underlying psychological events.

But self-reports cannot be used for straightforward inferences either, not even in adults, given how people acquire their semiotic repertoire for describing psychological phenomena during childhood. Children learn the shared meaning of symbols only because adults infer psychological phenomena from the children’s behaviours and because they label these inferences with particular symbols that they had learned in the same way. But because they cannot perceive the children’s psychological phenomena directly, they may vary—and also err—in their interpretations. Not everyone is “good at understanding” children; that is, at making inferences to their psychological events with relatively high validity.

The horizon of one’s own realm of experiences plays an important role as well. Some individuals experience psychological phenomena that others obviously never experience or do not experience in the same way—as reflected in the very assumption of individual differences. How can individuals understand what these symbols that denote particular psychological phenomena refer to if they have never experienced such phenomena themselves? How can colour-blind people understand the actual visual experience of people who are able to perceive particular wavelengths of light (Ludlow et al. 2004 )? Locke ( 1689 ) ascertained that individual differences in the physical properties of sensory organs should produce different perceptions and thus different ideas in different individuals. But this could never be detected because nobody can pass into another individuals’ body to perceive the perceptions that those organs produce. It seems inevitable that—even within semiotic communities—individuals develop slightly different personal connotations for socially shared symbols denoting particular psychological phenomena (cf. Westen 1996 ). Consequently, it can never be ascertained whether particular behaviours or particular symbols of shared meaning refer, in fact, to exactly the same psychological phenomena in different individuals. Their presence and quality in others cannot be detected with absolute certainty because every individual can directly access only his or her own psychological phenomena and cannot access anyone else’s (Locke 1689 ).

Behavioural Phenomena

Behavioural phenomena, by contrast, are directly perceivable by individuals both in themselves and in other individuals (criterion 1 fulfilled); however, this perception need not be conscious (see notion of perceiving above). As external activities or externalisations that are functionally mediated by the present environment, behavioural phenomena are defined by being directly involved in the individuals’ interactions with and in their relations to their environment (criterion 2 fulfilled).

Consequently, among those kinds of phenomena that are frequently incorporated in contemporary personality definitions, behaviour is the only one that fulfils both criteria implied by the lexical hypothesis. For this reason, personality is also conceptualised as “dimensions of behavioural space” (Cattell 1950 , p. 221), “individual differences and profiles in behaviour and behavioural performances” (Pawlik 2006 , p. 16, translated), and as individual-specific behavioural phenotypes (Uher 2011a ).

Phenomena of Outer Appearance

Other phenomena of ecto-phenotype fulfil these criteria. Individuals can directly perceive individual-specific patterns in natural outer appearance , such as body size and build, physiognomy, hair, or smell both in themselves and in others (criterion 1 fulfilled). These phenomena can also be directly involved in the individuals’ relations to their—in particular social—environment (criterion 2 fulfilled). Artificial modifications that individuals often add to their natural outer appearance to stress their individuality and that are socioculturally considered “adornments” such as clothing, colour, hair style, jewellery, and fragrances, are also—and actually purposefully made to be—(1) directly perceivable by other individuals. They can also be (2) directly involved in the individuals’ environmental interactions, their social interactions in particular, for example, as means of social communication (“fine feathers make fine birds”) or by influencing how individuals are socially perceived in terms of their “social stimulus value” (Allport and Odbert 1936 ; Cattell 1950 ). Hence, ecto-phenotypical phenomena other than behaviour can also become socially relevant in everyday life.

In summary, the meta-theoretical assumptions of the hypothesis that humans have lexically encoded all individual differences in phenomena that are (1) directly perceivable and (2) socially relevant and important in everyday life, and thus directly involved in their environmental interactions, apply only to phenomena of behaviour and outer appearance. But they do not apply to physiological and psychological phenomena.

This highlights two central points in which contemporary personality psychology deviates from the meta-theoretical assumptions of the lexical hypothesis. First, outer appearance is commonly not considered in scientific personality concepts. This is surprising because—as to be expected from the lexical hypothesis—such individual differences are encoded in everyday language. When Allport and Odbert ( 1936 ) compiled lexical personality descriptors from Webster’s New International Dictionary of the English Language ( 1925 ) for first comprehensive lexical studies, they also found terms describing physical appearance. Norman ( 1967 ) counted 882 such terms in a later edition of that dictionary. According to the lexical hypothesis, “these characteristics should be included in an exhaustive specification of personality” (John et al. 1988 , p. 186), but descriptors of outer appearance are typically excluded from lexical personality research (Saucier and Goldberg 1998 ).

Interestingly, some early personality theorists explicitly considered outer appearance in their definitions of personality. Guilford ( 1959 , p. 7) conceived morphology as one of seven “modalities of traits representing different aspects of personality”. Eysenck ( 1947 , p. 25) conceived constitution (the somatic sector) as one of four main sectors of personality. Outer appearance is also central to constitutional theories of personality (regardless of their methodological insufficiencies), such as those focusing on body build (Kretschmer 1977 ; Sheldon and Stevens 1942 ), skull shape (phrenology; Bouts 1931 ), and physiognomy (physiognomics; Aristotle, 300 BC, 1963 ; Lavater 1775–1778 ). Outer appearance is also increasingly considered in research on person-perception at zero-acquaintance (Borkenau and Liebler 1992 ; Hartung and Renner 2011 ), yet in relation to and not as part of personality.

A second point in which contemporary psychology deviates from the meta-theoretical assumptions of the lexical hypothesis is that psycho-physiological and psychological phenomena are frequently incorporated in personality concepts although they do not fulfil the two criteria. Psycho-physiological phenomena (cf. Allport 1937 ) are not perceivable by individuals in themselves or in others nor are they well represented in the everyday lexica. Psychological phenomena are prominently conceived as (the) core elements of personality both in scientific psychology (Matthews et al. 2003 ) and in everyday psychology. In fact, the human lexica contain many terms that are (also) used to describe individual-specific patterns in psychological phenomena (e.g., anxious, playful). Granted the lexical hypothesis is valid, so how can this be?

Genesis of Personal Constructs of Self- and Other-Perception

According to the lens model of human perception (Brunswik 1952 , 1955 ), in a given moment, individuals can perceive from their complex environment only details although they have flexibility regarding which particular details they focus on. Brunswik ( 1952 , p. 23) therefore considered the human perceptual systems to be “imperfect machines” that allow for the creation of only imperfect mental representations. Representational accuracy depends on the ecological validity with which individuals can perceive their environment by shifting the narrow lens filters of their perceptual systems. The accuracy of representations of highly dynamic and fluctuating phenomena, such as behaviour, may be particularly impaired.

However, developing perceptual constructs involves more than merely representing sketchy inputs as modelled in artificial neuronal networks. These networks passively map any input they are receptive to. But perceptual constructs are not imprints that external phenomena breaking through the individuals’ perceptual lenses leave on their inner mental images as light breaking through an optical lens darkens photobase paper. Representations are not copies (Jovchelovitch 2007 ). Mental representations shape and sharpen the individuals’ perceptual lens filters, thus changing their perception (Fig.  1 ). Moreover, individuals can also perceive actively. They can focus their perceptual lenses purposefully on particular details that could be meaningful and that they therefore seek to perceive (Brunswik 1952 ). Particularly meaningful in complex environments are recurrent patterns because their recognition and mental representation may provide orientation and predictive control of future events (Kelly 1955 ). This may lead individuals to selectively perceive and actively construct more patterns than there actually are in the perceivable phenomena. Similarly, they may seek to primarily perceive those structures and phenomena that match the mental representations that they have already established (e.g., their category systems); whereas they may give less consideration to those that do not match (i.e., thus perceiving “through coloured glasses”). These processes of perception and mental representation are iterative and self-reinforcing.

Essential for individuals are perceptions and constructs of patterns in their experiences with their own person and their own relations to the environment (Kelly 1955 ). Because individuals can directly perceive psychological phenomena in themselves, they may recognise relations between patterns of their own psychological experiences, patterns of their own behaviours, and patterns of their environment. It may be for this reason that such relations are essential parts of personal constructs. However, phenomena of both the psyche and behaviour are fluctuating and highly dynamic as are many phenomena in the environment, the social environment in particular (Brunswik 1955 ). The details that individuals can perceive from these phenomena through shifting their perceptual lenses are necessarily inconsistent and fragmentary (Brunswik 1952 ). Construing coherent and meaningful patterns in these phenomena and in their interrelations must transcend these perceptual gaps.

The Mind-Environment Connection and the Essential Role of Behaviour

The most fundamental gap is that between the phenomena of the psyche and those of the environment. This gap is particularly puzzling and difficult to understand because it occurs in just one direction. Environmental phenomena may directly influence psychological phenomena through perception. But there is no direct connection in the other direction (Fig.  1 ). As entirely internal phenomena, psychological phenomena may be able to directly influence other internal phenomena. For example, they may influence physiological phenomena such as in psycho-neuro-endocrinological systems, but they cannot directly influence any phenomenon in the individual’s environmental systems (Bronfenbrenner 1979 ). No thought and no motivation can move an object in physical space. No emotion can directly influence another individual. This is only possible for phenomena that are externalised or external—such as the ecto-phenotypical phenomena of behaviour (including some parts of language; Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished) and of outer appearance. Thoughts, motives, emotions, abilities, and all other psychological phenomena can become effective in an individual’s environment and in that individual’s environmental relations only if these phenomena can somehow reach the external world through mediation of other phenomena at least at some point.

Behavioural phenomena are particularly suited for this mediation because their dynamics and flexibility may be somewhat comparable to those of the psyche. This enables timely mediation to the external world. This nearness-in-time is particularly important for the individual’s adaptation to and his or her interaction with dynamic and flexibly changing environments, especially in the social domain. Phenomena of outer appearance are comparably less flexible and more static, especially natural outer appearance. Artificial modifications may be more flexible—but they are ultimately results of behaviour. Psychological phenomena cannot directly influence outer appearance; even laughter and worry lines in the face are results of behaviour.

Consequently, behaviour is the essential interface between the individual’s internal (biological and psychological) and external (abiotic and biotic, in particular social) systems. It serves as the vital—and for psychological phenomena, the only —bridge between the individual’s internal and external worlds (Fig.  1 ).

No individuals demonstrate the essential nature of this vital bridge and the one-sidedness of the mind-environment connection more dramatically than those suffering from locked-in syndrome. Individuals in this condition are patients who are awake and conscious, but who have no means of producing speech, limb, or facial movements because their voluntary control over muscles is (sometimes completely) paralysed (Laureys et al. 2005 ). This loss of motor control burns the sole bridge between their psyche and the external world. They can still perceive their environment with (almost) all of their senses and continue to construct it in their minds (Laureys et al. 2005 ). But they cannot behave and thus transmit information from their internal worlds to the external world anymore. Experiencing the mind-environment connection in its full-blown one-sidedness may perhaps cause their greatest suffering. Locked-in syndrome can be recognised and differentiated from vegetative states only in those patients who are still able to perform minimal behaviours, be it sniffs or eye blinks. Such behaviours, as minuscule as they may be, are mediating channels that enable patients to externalise information from their inner psychological world. This mediation can be successful, however, only if the social environment interprets these externalisations correctly—and if it still tries at all to infer psychological phenomena in seemingly nonbehaving individuals.

The one-sidedness of this connection seems so puzzling and incompatible with the logic of the human mind and, at the same time, behaviour fulfils its mediating function so promptly and so smoothly, that individuals fill this gap (often not knowingly) with assumptions of direct mutual connectedness. When individuals recognise recurrent patterns in their own psyche and in their environment, they frequently construct direct mutual connections and causal associations—seldomly becoming aware of the impossibility of such. Personal constructs of self-perception therefore regularly contain assumptions about direct mutual connectedness and causal relations. These constructs constitute oversimplified representations that are parsimonious and often sufficiently viable in everyday life and that can provide some orientation and predictability in complex and uncertain environments (Kelly 1955 ). The iterative and self-reinforcing processes of perception and of mental construction contribute to the maintenance of these erroneous assumptions and make the one-sidedness of the mind-environment connection for individuals increasingly harder to recognise.

Preconditions for Identifying Individual-Specific Patterns

Predictability is particularly important in the social environment. Therefore, individuals seek to perceive and construct patterns that recur in many individuals—that is in the average individual—of their social environment. Of particular importance are the environmental interactions of the average individual—that is normative patterns of behaviour and outer appearance. Individuals also seek to recognise patterns that recur in particular individuals and their environmental interactions—that is in individual-specific patterns of behaviour and outer appearance. They construct these individual-specific patterns of environmental interactions as personality (see above; cf. lexical hypothesis; Allport 1937 ; Gray 1999 , p. 563).

But how can individual-specific patterns be constructed at all? In fluctuating phenomena, individual-specificity refers to differential patterns in time-relative probabilities that are meaningfully stable over time (see above). As such, they cannot be perceived at any given moment. “There is no present for all the elements and structures of conceptual systems at once” (Althusser and Balibar 1970 , p. 318). Whereas behaviour is defined by its functional reference to the present (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished), constructs of personality—because they refer to time-relative probabilistic patterns that are stable across time—also involve phenomena of the past that have already ceased to be. Therefore, the only way in which individuals can capture personality is to—literally—(re)construct someone’s personality based on individual-specific patterns that they can (re)cognise in past perceptions of past phenomena of behaviour and outer appearance.

However, the past and the future are not realities. They are constructions of the human mind. The ancient philosopher Augustine (354–430 AD, 1961 ) considered linear time perception an achievement of the human mind and its “enduring attention” (Hausheer 1937 ). Hence, constructs of personality, because they describe patterns over time, require conscious awareness and time perception. Only those phenomena that people can become consciously aware of are able to become the subject matter of their personal constructs. And only those constructs that they are consciously aware of can be socially shared and lexically encoded. However, people cannot consciously perceive all behavioural phenomena that are important and socially relevant in everyday life (Westen 1996 ). Pheromones (Grammer et al. 2005 ) and subliminal odours (Li et al. 2007 ) influence human behaviour, in particular mate selection (Bhutta 2007 ), but they are not consciously perceivable. Such phenomena are blind spots that likely evade personal and social constructions of personality (Kagan 1998 ; Uher, Methodological approaches to personality taxonomies: The Behavioural Repertoire x Environmental Situations Approach—A non-lexical alternative, unpublished).

Recognising temporal patterns requires time perception. Individuals who are unable to perceive time, and thus temporal patterns, cannot (re)cognise individual-specific patterns—and thus cannot (re)construct personality. Babies seem to have no sense of time at all; it develops only slowly during childhood (Fraisse 1964 ; Piaget 1969 ). This may hinder young children even after language acquisition from providing reliable reports about their own habitual behaviours or those of others. In adults, subjective time perception generally accelerates with age and differs among individuals (Carrasco et al. 2001 ; Joubert 1984 ). It is also flexible within individuals in that, sometimes, time seems to fly or to drag slowly. Individuals may therefore have different perceptions of temporal patterns of individual behaviour and also of (especially artificially modified) outer appearance—and thus vary in how they construct personality differences.

Of the sketchy details that individuals can perceive through their shifting perceptual lenses (Brunswik 1952 ), many may enter their minds just briefly. Those that are memorised are not stored as the “pure” perceptions that they were when they entered the person’s mind. Instead, perceptions are transformed into, organised with, and integrated with information that is already stored—in constructs and representations that thereby become continuously modified (Kelly 1955 ). “Experiences are cognitive resultants of past lives” in categorised form (Peirce 1902 , CP 2.84). Thinking is making categories that link the experiences of the present moment with those of the semiotically reconstructed past; the construction of the new is bound to and constrained by the reconstructions of the past (Valsiner 2012 ).

Consequently, individuals do not and cannot access large accumulations of “pure” perceptions of past events. They cannot base the recognition of individual-specific patterns in unbiased ways on large mental “data bases” of perceptions. Processing such enormous masses of information to identify probabilistic, differential, and temporal patterns would perhaps even exceed their cognitive capacities. All the more given that the identification of such patterns is substantially hindered by the fact that, in behavioural and psychological phenomena, within-individual variation often considerably exceeds between-individual variation (Shweder and Sullivan 1990 ; Mischel 1968 ; Uher, An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, unpublished). Instead, individuals can access only information that they have already integrated, transformed, and (re)constructed in iterative processes of perception and construction over time. “Recognition is not a simple accumulation of elements, but an integrated whole” (Komatsu 2012 online, p. 3).

For these reasons, the particular phenomena of behaviour and outer appearance and the particular situations in which they are perceived along with the particular individuals, the particular occasions and the particular spans of time that people actually consider when they mentally (re)construct individual-specificity (i.e., personality; Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished) all remain unknown—also to the construers themselves. These mental (re)constructions necessarily suffer from constraints, biases, and errors in perception, memory, and reasoning (Fahrenberg et al. 2007 ; Gigerenzer et al. 1999 ; Shweder and D'Andrade 1980 ).

Genesis of Socially Shared Constructs of Self- and Other-Perception

The effective ability to communicate with others substantially increases the information that individuals can obtain. Communication (verbal and nonverbal) may direct an individual to focus its perceptual lens on particular details that others have perceived and that they consider important, thereby promoting the socially shared perception of particular phenomena and thus their salience. Verbal communication enables individuals to exchange information also beyond what is available in the present (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished). This opens up further avenues to obtain information that is relevant for (re)cognising individual-specific patterns and that may, at the same time, modify the individual’s pertinent constructs and representations. Sharing information about individuals is obviously so important to human social life that gossiping about who-is-doing-what-with-whom comprises up to about two thirds of conversation time (Dunbar 1996 ) and occurs in all human cultures (Brown 1991 ). Central to this gossip is information about average individuals and about particular individuals in the social environment. Obtaining this information inevitably entails comparisons among individuals.

Possibilities for Identifying Individual-Specific Patterns

The rather static phenomena of outer appearance can be perceived in many individuals at the same time and can therefore be compared rather directly between individuals. By contrast, the dynamic and fluctuating nature of behavioural phenomena hinders their simultaneous perception in many individuals—and hence hinders direct comparisons between individuals. Only in particular situations, such as in footraces in which individuals perform exactly the same behaviour spatially and temporally in parallel to one another, do differences between individuals become directly perceivable. But such situations and uniform performances of the same behaviours in spatial and temporal parallelism are not typical of everyday life.

Psychological phenomena, in turn, can generally not be directly perceived in other individuals (Locke 1689 ; Toomela 2008 )—hence, they cannot be directly compared between individuals. This is a crucial point for psychology, especially for personality psychology. Ultimately, how could people identify differences between individuals in phenomena that everyone can directly perceive in just one single individual—in just oneself? The impossibility of perceiving psychological phenomena in others (Locke 1689 ) also entails that the validity of inferences from others’ externalisations, such as from their behaviour (nonverbal and verbal; cf. Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished), can never be exactly ascertained (cf. Toomela 2011 ). Any comparisons based on phenomena that potentially mediate the mind-environment connection are necessarily compromised. They are but social conventions (Valsiner 2012 ). Even if each individual developed his or her own private metric against which that person could quantify his or her own psychological phenomena—there would be no point of direct comparison by which to convert these metrics across individuals.

But people cannot directly compare their own psychological phenomena even within themselves across time. Psychological phenomena are “of maximum uniqueness—they occur each only once, at the minuscule border of the future and the past we construct as the ‘present’” (Valsiner 1998 , 2012 , p. ix). They are “actualities” (Gillespie and Zittaun 2010 , p. 72). Their perception is inevitably bound to the present, which is described as the “here and now” in the field of psychology. (For considerations of the present and its perception, see James 1893 , and Le Poidevin 2004 , 2011 ). Comparisons of perceptions of psychological phenomena within the individual would require comparisons with past ones—which, however, have already ceased to be. All that can remain are memorised perceptions (Le Poidevin 2011 ) that have already been integrated, organised, abstracted, and (re)constructed—the “cognitive resultants of past lives” (Peirce 1902 , CP 2.84; Valsiner 2012 ). Consequently, perceptions of ongoing psychological phenomena can be compared only with (re)constructions of past perceptions. Direct comparison between present and past events of psychological phenomena—and thus quantifications of them—are impossible. Attributions of quantitative properties to psychological phenomena therefore must be—and are—fundamentally challenged (e.g., Brower 1949 ; Michell 1997 ; 2003 ; Omi 2012 ; Rosenbaum and Valsiner 2011 ).

The actively construing minds of individuals again transcend these impossibilities—largely unnoticed as they transcend the one-sided gap of the mind-environment connection. Because each individual can directly perceive his or her own psychological phenomena, and constructs relations among them and to his or her own behavioural phenomena, social exchange may promote the assumption that other individuals have similar experiences and constructions as well. Likewise, social exchange may contribute to the awareness that others also experience perceptual gaps. Vice versa, realising the existence of perceptual gaps and the resultant lack of understanding may trigger social exchange and the formation of socially shared constructs (Moscovici 1961 ; Voelklein and Howarth 2005 ). These processes may explain why the entirely internal phenomena of the psyche are commonly incorporated into socially shared constructs of personality. These constructs are, in the strictest sense, socially shared constructs of self- perception, but not of other -perception. The latter is possible only for the phenomena of behaviour and outer appearance.

Through social exchange, individuals can also incorporate into their own personal or subjective constructs those that other individuals have subjectively constructed; thereby merging and shaping them in socially shared ways into intersubjective or social representations (Jovchelovitch 2007 ) or folk concepts (Tellegen 1993 ). Because everyone constructs the world differently (Kelly 1955 ), inferring psychological phenomena from the externalisations of other individuals and comparing the inferred phenomena across individuals and with those of one’s own psyche is necessarily a collective act. Compromises are required in order to achieve consensus on the particular inferences drawn and on the particular comparisons made. This may lead to different outcomes in different sociocultural communities as the diversity of, for example, parental ethnotheories shows (Pillai 2012 ). But social exchange and intersubjective compromise need not necessarily resolve contradictions between personal and socially shared constructs. Individuals and semiotic communities can simultaneously maintain competing constructs in terms of cognitive polyphasia (Howarth et al. 2004 ; Moscovici 1961 ).

Semiotic systems are elementary for social exchange—lexical symbols in particular. They are essential for making subjectively constructed meanings accessible to others and for intersubjectively exchanging them among construing minds. However, “language is not merely a reproducing instrument for voicing ideas but rather is itself the shaper of ideas” (Whorf 1958 , p. 5). The very communication tools that people use to build up and to express their representations already implicitly contain some structures (Lahlou 1996 ). Language mediates thoughts and enables abstract thinking (Neuman et al. 2012 ).

A crucial process through which language enables abstraction is hypostatic abstraction in which perceivable qualities (e.g., sweet) are converted into objects (e.g., sweetness; Peirce 1902 , CP 4.227). This reification makes perceivable qualities functionally independent of their embodied experience. As objects, these qualities can be associated with signs that are arbitrarily related. These reified qualities can become objects of reflection and contemplation and can be linked to various other modes of perception and to connotations (Neuman et al. 2012 ). Languages therefore contain not only concrete words that refer to directly perceivable phenomena; they also contain abstract words that refer to ideas and concepts that are distant from immediate perception (Vygotsky 1962 ). Because the concrete references of abstract words cannot be easily traced anymore, these words are loaded with meanings that likely vary across sociocultural and historical contexts (Neuman et al. 2012 ). Hence, all words have meaning; the meaning of every word is a construct (Vygotsky 1962 ).

Contextuality of Constructs of Self- and Other-Perception

Constructs reflect recurring patterns that have been perceived in the world (Kelly 1955 ). As such, they represent knowledge. Constructs of self- and other-perception represent knowledge about individuals—average and particular ones—and thus about comparisons between them (see above; Westen 1996 ).

Knowledge is always context-dependent (Jovchelovitch 2007 ). The knowledge that constructs of self- and other-perception reflect may depend on the context of particular individuals . It can therefore be ideographic and personalistic (Allport 1937 ; Kelly 1955 ; Lamiell 2003 ; Murray 1938 ; Stern 1935 ). It also depends on the individuals’ stages of development (Toomela 2010a ), on their knowledge and educational level (Vygotsky 1962 ), and on their abilities to reflect about themselves, their world (Omi 2012 ), and their life histories (McAdams 1985 ). Different individuals therefore associate standardised items on personality questionnaires with different “fields of meaning” that need not be identical to the interpretations that the researchers have of these items (Arro 2012 ; Rosenbaum and Valsiner 2011 , p. 47). This knowledge may also depend on the context of particular situations (Bandura 1986 ; Lazarus 1981 ; Mischel 1968 ; Rotter 1954 ) and on the states of mind that individuals may have in particular situations (Omi 2012 ), such as particular goals and motivations (Biesanz and Human 2010 ). It may also depend on the context of particular groups of individuals; therefore, reference group effects can be found (Heine et al. 2002 ).

The knowledge that constructs of self- and other-perception reflect also depends on the context of the particular semiotic system in which it is encoded. Bilingual people therefore describe themselves differently depending on the language they use—even in translated versions of the same questionnaire (Veltkamp et al. 2012 ). It is not possible to separate a semiotic system from the sociocultural context in which it was developed (Køppe 2012 ; Whorf 1958 ). The knowledge that constructs of person perception reflect therefore also depends on the context of the particular culture in which it was developed. Culture is “the inherent core of human psychological functions” (Valsiner 2009 , p. 5). The particular patterns and phenomena that particular groups of individuals perceive as salient, that they construct as socially relevant, and that they semiotically encode (cf. lexical hypothesis) is culture-dependent (Church and Katigbak 1988 , 1989 ). “Cultures select a limited range from among the spectrum of [patterns] to encode in their lexicon, and they may select differently. Languages differ not only in the precise … terms they include (as every translator knows), but more broadly in the aspects … their vocabularies emphasize” (McCrae and Costa 1997 , p. 510; Angleitner et al. 1990 ).

Finally, person-related knowledge also depends on the context of the historical era in which it is developed and encoded (Kagan 1998 ). The socio-cultural norms of interpretation and appraisal of particular patterns of behaviours and of outer appearances may change over time—and with it the pertinent everyday psychological knowledge. Over time, new lexical symbols (John et al. 1988 )—and new meanings (Toomela 2010a ; Vygotsky 1962 )—are created, whereas others become outdated and are no longer used.

The preceding analyses have shown that lexical approaches and assessment methods investigate everyday psychological knowledge, but not the phenomena that this knowledge is about. The philosophy-of-science perspective will now be used to scrutinise the prevailing strategy to explain individual differences—explicitly or implicitly—with assumptions of “traits”.

Strategies of Scientific Explanation Based on Trait Concepts

Over the last century, psychologists have tried to tackle the major task of systematically and comprehensively categorising individual differences in order to develop explanatory theories of personality. “For, we can look at a theory of personality as a specification of the most important individual differences and then as a model of how they come about” (Goldberg 1981 , p. 141). To do so, it was necessary to meet three challenges.

The first challenge was to devise suitable strategies for deciding which kinds of individual differences to select for empirical categorisations (Goldberg 1981 ; Uher 2008a , b , Uher, Methodological approaches to personality taxonomies: The Behavioural Repertoire x Environmental Situations Approach—A non-lexical alternative, unpublished)—a task that seemed “hopelessly complex” (Thurstone 1934 , p. 14). The lexical hypothesis suggested that categorising the collective knowledge about individual differences that people had encoded in their natural languages could be a suitable strategy for developing comprehensive taxonomic models of individual differences. The compilations of lexical descriptors in comprehensive lexica—the “storehouses of folk knowledge” (John et al. 1988 , p. 174)—that were already available for various languages made this a viable, though labour-intensive approach (Allport and Odbert 1936 ; Norman 1967 ).

The second challenge was to devise methods of scientific measurement of individual differences (Uher 2008a ). The lexical hypothesis suggested that enquiring about lay people’s constructs about individual differences could be a suitable scientific method (Block 2010 ; Westen 1996 ). The ease of collecting such data from many individuals made assessments by lay people the “preferred” (Matthews et al. 2003 , p. 5) “if not the standard” method (Rosenbaum and Valsiner 2011 , p. 50; italics added) on which personality psychology “relied heavily” (Baumeister et al. 2007 , p. 396).

The third challenge was to devise methods for the systematic reduction of the complex data sets obtained through assessments, be they based on lexical or other approaches (cf. Uher 2008a ; Uher, Methodological approaches to personality taxonomies: The Behavioural Repertoire x Environmental Situations Approach—A non-lexical alternative, unpublished). Many statistical methods have been developed to identify the structures that underlie empirical data to enable parsimonious descriptions of their complex manifest structures (e.g., factor analyses; Thurstone 1934 ; Cattell 1952 ).

Statistics describe structures that are latent to data . Assessment data reflect how much people believe (i.e., construct) that particular lexical descriptions apply to particular individuals. Lexical descriptions encode the socially shared knowledge about individuals. This knowledge comprises everyday psychological descriptions of patterns of behaviour and of outer appearance as well as potential causal explanations—be they accurate or not—in terms of inferred psychological phenomena in particular. Assessments rely on this knowledge. They also rely on the personal and socially shared constructions about how to quantify—the not directly perceivable—individual-specific patterns in these phenomena (see above). Consequently, the latent structures that can be identified by lexical or other assessment studies reflect structures that underlie the socioculturally shared construction, interpretation, appraisal, and explanation of consciously perceived individual differences (Borkenau and Ostendorf 2008 ; John 1990 ). These structures describe categorical systems of personality descriptions in natural-language terms (Goldberg 1982 ; John and Srivastava 1999 ) and thus comprise models of everyday beliefs about personality (Westen 1996 )—i.e., the systems of collective knowledge.

The prevailing psychological interpretation of the structures that are latent to assessment data, however, is that they reflect stable structures that are latent to the individual-specific phenomena that are being described and assessed. Hence, as an analogy to statistical methods in which true (i.e., stable) structures are modelled as latent “traits” that underlie complex empirical (observable) data and that can statistically explain these data, these structures are assumed to reflect stable structures in psychological, physiological, and other phenomena internal to the individual that causally influence and can thus explain that individual’s (perceivable) behaviour. These structures are called “traits” or “dispositions” in personality psychology (Allport 1937 ; Matthews et al. 2003 ).

The statistical approximation of true scores that are free of measurement error refers to precision in scientific measurement. Measuring personality “traits” that are stable rather than that fluctuate across occasions and across situations, as is the case for behaviour seems therefore intrinsically plausible and scientifically valuable. But this reasoning is based on a fundamental fallacy entailed by the human language.

Language enables abstraction because it allows us to make perceivable qualities independent from their immediate perception by reifying them into linguistic objects (i.e., hypostatic abstraction, Peirce 1902 , CP 4.227; see above; Neuman et al. 2012 ). Reifications are prone to the fallacy of misplaced concreteness in which abstractions are treated as concrete (Whitehead 1929 ). The idea of individual-specific patterns in behavioural and other phenomena requires high levels of abstraction. The linguistic reification of this complex and highly abstract idea through lexical encodings, such as by the terms “traits” or “dispositions”, is therefore inherently prone to this fallacy. It misleads people to treat “traits” and “dispositions” as concrete actual entities.

The reification of constructs of self- and other-perception as causal entities that underlie the phenomena that are being perceived and constructed are widespread in everyday psychology. For example, individuals who behave aggressively at some point in time are ascribed as having aggressive “traits” or “dispositions”; and the construction of them as having these “traits” is used to explain the fact that they behaved in this manner in a given moment. In everyday life, linguistic reification facilitates and accelerates the exchange of information, despite its inexactitude. It may therefore disseminate more quickly than more accurate yet cumbersome descriptions. Everyday psychology is therefore full of reifications and circular explanations (Laucken 1974 ).

The assumption of “traits” as heritable structures internal to the individual (Brody 1994 ) that “determine behaviour in a defined situation” (Cattell 1950 , p. 222) ascribes to these abstract ideas not only a concrete existence, but furthermore a causal status (Bock 2000 ). But assessments of “traits” do not and cannot capture causal entities inside the individuals assessed (Komatsu 2012 ). “Traits” are lexically encoded and socially shared constructs about recurring patterns in phenomena perceivable in the assessed individuals—i.e., in behaviours and in outer appearances. As such, “traits” are categorical summary statements about a person’s behaviour (Wiggins 1979 ). Using these summary statements to explain and predict behaviour means “in effect using a description of behaviour to explain it” (Mischel and Shoda 1994 , p. 157).

The circularity of psychological explanations that are based on trait concepts may not be as directly apparent as that of the explanations that are provided by everyday psychology. Scientific constructs of “personality traits” are the result of long and highly complex series of abstraction processes. The phenomena to be explained are first filtered through iterative and repeated processes of perception, personal and socially shared (re)construction, lexical encoding, and assessment by lay psychologists (see above). The outcomes of these processes are then further condensed by the application of complex statistical methods by scientific psychologists. The complexity of these processes of abstraction, of which the everyday psychological ones are still poorly understood (Rosenbaum and Valsiner 2011 ), may mask the circularity of these explanations—but cannot overcome it.

In a nutshell, psychological explanations based on trait concepts equate the statistical structures underlying the assessments of encoded constructions about perceivable phenomena with the phenomena that underlie these perceivable phenomena. This explanation is circular at a very high level of abstraction and complexity, but it is essentially circular. Figure  2 illustrates this schematically. Based on the erroneous assumption that D is a measure of C, which underlies A, it is concluded that D measures what underlies A and that it can thus causally explain A. But meta-theoretical analyses show that D can only be the latent structure of B, which are the constructs and representations that people develop of A. Explaining A with D is thus circular.

An external file that holds a picture, illustration, etc.
Object name is 12124_2013_9230_Fig2_HTML.jpg

The fundamental circularity of psychological explanations based on trait concepts. Note. Based on erroneous meta-theoretical assumptions about what phenomena are actually reflected by assessment data, D is considered to be a measure of C, which are the phenomena that underlie A. Therefore, it is concluded that D measures what underlies A and that D can thus causally explain A. But D can be the latent structure only of the phenomena B, which are the constructs and representations that people develop of A. Thus, explaining A with D is circular

Given that assessments reflect socially shared knowledge about recurring patterns in behaviour and outer appearance, it is not surprising—but rather to be expected—that “personality predicts behaviour” (Paunonen and Ashton 2001 ) and that, vice versa, individuals can use “thin slices of behaviour” (Borkenau et al. 2004 ), “minimal appearance cues”, such as “facial structures” (Kramer and Ward 2010 ) or shoes (Gillath et al. 2012 ) to activate this knowledge and use it to “accurately assess others’ personality” even at zero acquaintance (Shevlin et al. 2003 ). But such cues are not “signals” of entities internal to individuals (Kramer and Ward 2010 ) that people are able to decipher. These findings merely reflect the socially shared knowledge about how particular individual-specific patterns in behaviour and outer appearance typically go together in a particular semiotic community and how that community (causally) explains this. This knowledge may also be reflected implicitly in lexical descriptors that do not refer explicitly to outer appearance (which were excluded from lexical approaches). Likewise, it is also to be expected that personality assessments are stable across decades (Conley 1984 ; Soldz and Vaillant 1999 ). This stability is also a property of the socially shared knowledge they reflect.

At the bottom of psychological explanations based on trait concepts is the—anything but new—insight that “past behaviour is the best predictor of future behaviour”. It may be circular to predict behaviour with behaviour—and this is why assumptions of independent internal entities are so tempting. But it is not circular to predict future events from past events, in particular not in phenomena as dynamic and fluctuating and inherently bound to the present, such as those of behaviour. In complex dynamic environments, such as in the social world, identifying recurrent patterns in these phenomena can provide orientation and predictive control. But prediction is not yet an explanation. Past behaviour can be causally related to future behaviour, though only indirectly, such as through the mediation of environmental (in particular social) transactions or through the knowledge that people have developed about it. But this cannot explain how behavioural patterns emerge in the first place. The direct causes of behaviour, which are primarily sought in psychological, biological, and other internal phenomena as well as in external environmental conditions, cannot be captured through assessments and causal reifications of behavioural descriptions. They can be explored—in themselves and as possible causes of individual behaviour—only if they are studied in their own rights (see below).

Implications for Personality Psychology

The circularity of scientific explanation based on trait concepts derives from the fundamental challenges that the different kinds of phenomena entail that are described in the lexical hypothesis. Of the phenomena that people construct and represent, many are bound to the present, such as those of behaviour and the psyche, which makes direct perceptions of individual-specific patterns impossible. Their (re)cognition requires repeated perception, memorisation, and mental reconstruction (Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished). But the dynamics and fluctuations of these phenomena— and the individual’s narrow perceptual lens filters that shift across these phenomena’s fluctuating events—enable only fragmentary and inconsistent perceptions (Brunswik 1952 ), which thus hinders the (re)cognition of individual-specific patterns. Nonetheless, to enable orientation and sense of control, human minds transcends these gaps (Daston and Galison 2007 ; Valsiner 2012 ) and reify—promoted through language (Neuman et al. 2012 )—mental (re)constructions of the perceived into causal entities (Bock 2000 ; Lamiell 2003 ), thus blending a phenomenon’s description with its explanation (Mischel and Shoda 1994 ). This obscures the vital differentiation between the phenomena to be explained, the explananda , from those explaining them, the explanantia (Hempel and Oppenheim 1948 ; Popper 1934 )—and inevitably entails explanatory circularity.

These are challenges to everyday psychologists (Laucken 1974 ) and scientific psychologists alike. The latter are always both. “All scientists are victims of a linguistic structure that partially controls our thinking” (Howard 1994 , p. 400). Underlying this structure are the basic assumptions that their particular semiotic communities maintain about the “lived” relations between individuals and their world—i.e., their ideologies (Althusser and Balibar 1970 , p. 314). Ideology precedes science; it “constitutes the prehistory of a science” (p. 45). By breaking radically with the epistemological frame of reference of prescientific (i.e., ideological) notions and by constructing new patterns through scientific methodology—i.e., through thinking about phenomena and the means of producing knowledge [ German: Wissen schaffen ] about them—science [ German: Wissenschaft ] is produced. But alongside science, ideologies survive as essential elements of every social formation (Althusser and Balibar 1970 ). Sciences are always embedded in the sociocultural ideologies that operate through the meanings that semiotic communities attribute to their activities through common sense (Valsiner 2012 ).

This may explain why scientific psychologists—at least in Western societies, where basic assumptions of homo rationalis (Simon 1993 ) prevail—are so confident about the abilities of everyday psychologists to comprehensively identify, accurately describe, and validly quantify all individual differences that are socially relevant and important in everyday life—as this is prominently reflected in the lexical hypothesis and its fundamental significance as one of the most important—if not the most important—guiding theoretical assumptions in modern personality psychology (cf. Ashton and Lee 2005 ).

The Current State of Knowledge about the Lexical Hypothesis

Despite its enormous importance, the lexical hypothesis itself has not yet been made an object of scientific inquiry. Still today, almost 130 years after its first articulation (Galton 1884 ), it has remained but a hypothesis . To gain a more comprehensive understanding of this hypothesis and its current state, it is important to explore its possible origins.

The lexical hypothesis was not derived from a specific theory; rather, some specific theories were developed based on parts of its basic statements and on the results obtained from empirical applications of it (Goldberg 1981 ). Most likely, the lexical hypothesis was generated by abduction —a form of logical inference that Peirce ( 1901 , CP 7.218) introduced as “guessing” and considered “merely preparatory”. Abduction is an essential epistemological step in scientific reasoning. It begins with surprising facts that are in need of an explanation that is not available at the outset. Consideration of the surprising facts suggests a hypothesis that, if it was true, would explain the observed facts as a matter of course. The particular hypothesis is suggested by the resemblance between the facts and the consequences of the hypothesis if it is found to be true. “Hence, there is reason to suspect the generated hypothesis is true” (Peirce 1903 , CP 5.188-5.189). But its truth is still uncertain because abduction seeks to explain facts ascertained in the present by inferring possible causes of these facts, which necessarily lie in the past and thus have already ceased to be (Valsiner 2012 ).

For Galton ( 1884 ), it may not have been a single surprising fact that led him to abduce the lexical hypothesis, though the richness of lexical descriptors about individuals could indeed surprise someone who is unfamiliar with human languages. But rather, it may have been the reduction problem of a science that is concerned with the surprising complexity of the real phenomena of individual differences. If the lexical hypothesis was true, lexical descriptors of individual differences would be self-evident consequences. Thus, some of its statements could provide an explanation for the existence of the pertinent repertoires in everyday languages and—based on this—a possible solution to the reduction problem of personality psychology.

Once hypotheses are generated abductively, they are used as general premises to reach specific conclusions in scientific reasoning. Two approaches to argumentation are thereby possible: deductive and inductive. If the truth of the general premise is believed to definitely establish the truth of the conclusions, then the argument is deductive . If the truth of the general premise is not assumed to definitely establish the truth of the conclusions, but to provide good reasons to believe that they are true and unlikely to be false, then the argument is inductive . Thus, deduction seeks a logically certain conclusion that is derived from the premise, whereas induction seeks facts that support or falsify the premise. But neither deduction nor induction can leave the realm of analysis that the hypothetical premise defines and delimits. They cannot create new knowledge. This is possible only for abduction because it “seeks a theory” (Peirce 1901 , CP 7.218). Hence, abduction creates new knowledge—yet at the expense of a priori certainty about its truth. This uncertainty allows—and intrinsically calls—for their constant reconsideration and reconstruction (Valsiner 2012 ; cf. also Althusser and Balibar 1970 ; Weber 1949 ).

For the lexical hypothesis, this has not yet been done. Personality psychologists took this hypothesis as a true premise by which to reach the specific conclusion that the human lexical repertoires constitute finite sources of information for categorising individual differences (Allport and Odbert 1936 ; Cattell 1943 ; Goldberg 1993 ; John et al. 1988 ). But inductive investigations have not yet been done; personality psychology did not—and still does not—seek facts that may support or falsify this hypothesis. It is still unknown whether in fact all individual differences that are important in interpersonal life are lexically encoded as stated by the lexical hypothesis. For example, how well are individual differences in behaviours that are not consciously perceivable yet are known to influence social behaviour (see above) represented in everyday psychological knowledge? And how is this knowledge lexically encoded? Where do the differences in taxonomic models come from that have been derived from different languages? Do they reflect actual differences in the individual-specific behaviours and outer appearances between different language communities, just differences in what these communities consider salient and socially relevant or in how they interpret and appraise what they perceive, or all of these together (see below)?

The status of the truth of the lexical hypothesis is still as uncertain as it was at the time of its abduction almost 130 years ago (Galton 1884 ). Even worse, at this historical point in time, personality psychology is de facto unable to test this hypothesis empirically because its basic statements have been only partially considered so far. Psychologists have so far systematically categorised only people’s lexically encoded constructs and representations, but they have not even tried to achieve this for those phenomena that are constructed and represented in everyday life—that is, for individual-specific patterns in behaviour (Uher, Methodological approaches to personality taxonomies: The Behavioural Repertoire x Environmental Situations Approach—A non-lexical alternative, unpublished) and in outer appearance. Psychologists have revealed only half of the story of personality so far. However, comprehensive categorisation systems of individual-specific behaviours and appearances are needed to unravel how lexically derived models actually represent perceivable individual differences; the ways in which they may reflect different perceptions, interpretations, and appraisals; and how individual behaviours and outer appearances actually vary within and among different sociocultural and language communities (cf. Block 2010 ). This knowledge is also essential for systematic investigations of the actual causal mechanisms behind many fascinating findings that have been found by using lexically derived taxonomic models. Ten desiderata for future research are specified below.

Lack of an Elaborated Philosophy of Science for Personality Psychology

As scientists concerned with phenomena that are important in each individual’s everyday life, personality psychologists are particularly susceptible to the “psychologist’s fallacy” (James 1893 , p. 196)—the unwanted confusion between our own personal standpoints and the scientific objects we study. “It is true that in our sciences, personal value-judgments have tended to influence scientific arguments without being explicitly admitted” (Weber 1949 , p. 54). “Psychology is a science which is peculiarly liable to distortion by pseudo-scientific, political and religious intrusions” (Cattell 1950 , p. 11). Like all individuals, researchers have socioculturally guided minds (Valsiner 1998 ). Ideologies about the relations of individuals—average and particular ones—to their social world—i.e., about personality—are particularly pronounced. The influences of such ideologies on the scientific practices in personality research are therefore particularly profound and difficult to recognise and to confine.

Scientific ideas should be separated from the sociocultural ideologies into which they are embedded and that all individuals as members of their particular semiotic communities have in their minds (Valsiner 1998 ). Unfortunately, an explicit and elaborated philosophy-of-science framework that could help personality psychologists to continuously rethink and challenge their own ways of generating knowledge has not yet been established—despite Rychlak’s ( 1968 ) promising start half a century ago and apart from brief sections on ideographic and nomothetic approaches and discussions of the “images of man” of major schools of thought contained in many textbooks. But philosophy-of-science tools are essential for regularly rethinking and challenging the meta-theoretical structures that have became established in schools of thought and in the scientific practices of a discipline (Kuhn 1976 ).

Nature is vastly too complex to be explored even approximately at random. Something must tell the scientists where to look and what to look for, and that something, though it may not last beyond his generation, is the paradigm with which his education as a scientist has supplied him. Given that paradigm and the requisite confidence in it, the scientists largely ceases to be an explorer at all, or at least to be an explorer of the unknown. Instead, he struggles to articulate and concretize the known. (Kuhn 1976 , p. 61)

Paradigmatically fixed and no longer challenged assumptions close the doors to intellectual innovation (Valsiner 2012 ). In personality psychology, the paradigmatic establishment of trait concepts has hindered elaborations and explications of the meta-theoretical assumptions that underlie this particular concept and the school of thought in which it was developed, which has impeded the recognition of its fundamental explanatory circularity. But meta-theoretical assumptions are always effective in any scientific practice—even if they may be effective only implicitly. The assumption that the “description of personality must precede, not follow, personality theory” (Costa and McCrae 1992 , p. 861) overlooks the methodological matter of fact that all empirical facts are defined as facts only based on particular theoretical assumptions.

At the same time, the dogmatic and unquestioned status of assessments by laypeople as the standard methods of “scientific measurement” has hindered the explication of the meta-theoretical basis on which they rest. Selecting methods a priori rather than purposefully for given questions narrows researchers’ perceptions of real phenomena and blinds them to particular elements (Gillespie and Zittaun 2010 ; Omi 2012 ; Toomela 2011 ; Westen 1996 ). The tempting assumption that assessments constitute a “direct approach to find out what people think or feel” by using a “simple and straightforward task” (Rosenbaum and Valsiner 2011 , p. 50) that requires only “abbreviated introspection” (p. 47) has blinded many psychologists to the fact that the phenomena of the psyche and of behaviour are inherently bound to the present (Gillespie and Zittaun 2010 ; Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished; Valsiner 1998 , 2012 ) and cannot be captured by retrospective assessments at all (Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished).

Their efficiency with regard to data collection in large samples in conjunction with sophisticated statistical methods for data analysis has given assessment methods a scientific shape that has consolidated their status as the standard empirical method (Omi 2012 ; Sato et al. 2010 ; Toomela 2011 ; Westen 1996 ). But this a priori status has precluded an enquiry as to what the phenomena actually are that are being converted into data in assessments—and thus has precluded a rethinking of how these data can be analysed and interpreted at all (Gillespie and Zittaun 2010 ; Køppe 2012 ; Toomela 2011 ; Valsiner 2012 ). This has hindered many from recognising that assessments can reflect only knowledge structures, but not past events of psychological or behavioural phenomena. It has also hindered the recognition that, in the fluctuations of behaviour, individual-specific patterns cannot be directly perceived and thus cannot be directly quantified. The a priori status of these methods has also hindered the recognition that psychological phenomena may not—and actually cannot—have quantitative properties that could be “measured psychometrically” because they, too, are phenomena that are inherently bound to the present (Brower 1949 ; Michell 1997 , 2003 ; Omi 2012 ; Rosenbaum and Valsiner 2011 ; Schwarz 2009 ; Trendler 2009 ; Valsiner 2012 ) and that therefore, they cannot be compared with one another directly.

The lack of an elaborated meta-theory has also hindered many personality psychologists from considering the fact that analyses of individual assessments on the basis of their between- individual variation cannot yield results that are generalisable to the understanding and explanation of within- individual variation—much less of individual development or of the dynamics of internal mechanisms (Cervone et al. 2001 ; van Geert and van Dijk 2002 ; Omi 2012 ; Molenaar 2004a , b ). The particular numbers of dimensions of individual differences and their compositions are not natural entities to be “discovered” anyway. They depend on the meta-theoretical decisions made by scientists with regard to how real phenomena can be conceptually reduced to scientific phenomena as well as their decisions about how the thus-derived data can be reduced statistically (cf. Block 2010 ; Lamiell 2003 ). Statistics such as factor analyses are not “de facto theories” (Gigerenzer 1991 ) or “mechanical truth generators” (Meehl 1992 , p. 152). Rather, they are but “mathematical thinking tools” (Valsiner 2012 , p. 201; Loftus 1996 ).

The focus of personality psychology on trait concepts, lexical encodings, and assessment methods has hindered the ability of researchers to elaborate the meta-theoretical assumptions on which the lexical hypothesis is based. Personality psychologists have therefore failed to recognise that the hypothesis explicitly refers to different sets of phenomena of which they have systematically explored only lexically encoded representations so far, but have seriously neglected the phenomena that are represented. Assumptions that “personality and its assessment are intimately bound with natural language” (McCrae and Costa 1997 , p. 510) and that “the idea of personality traits may be as old as human language itself” (Matthews et al. 2003 , p. 3) clearly refer to personality as the lexically encoded structures of everyday psychological knowledge. But these structures cannot be understood and explained without researching the phenomena that they are about .

The Dialectical Interplay Between Encodings, Constructs, and the Phenomena Being Constructed

Individual-specific patterns of behaviour can also be found in babies and preverbal children (e.g., Kagan et al. 1988 ) as well as in many other species (Allport 1937 , p. 24). Among these species are those that are closely related to humans, such as nonhuman primates (Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ) unpublished), but also some that are related only remotely, such as octopuses (Sinn and Moltschaniwskyj 2005 ). Hence, individual-specific patterns in behaviour (and in outer appearance; Darwin 1859 ) are phylogenetically much older than the human semiotic systems that describe such patterns. Ancestral humans initially developed lexical and other symbols to refer to something that they had perceived, (re)cognised, and mentally and socially constructed—and that had thus already been there. Physiology, the psyche, and behaviour develop(ed) before language—both in ontogeny and in phylogeny.

It would be fascinating and insightful to study individual human ancestors who existed before humans had developed the faculty of language. Direct evidence of when humans started to develop socially shared constructs of individual-specific patterns of behaviour and outer appearance and pertinent semiotic symbols in phylogenetic history is lacking. But indirect evidence comes from Palaeolithic dog fossils that suggest that canine domestication had already begun at least some 30,000 years ago (Germonpré et al. 2009 ) and from an impressive 40-year breeding experiment in another canine species. In farm foxes, strong selective breeding solely for low fearfulness of and low aggressiveness to humans (i.e., tamability) over only 30–35 generations entailed a host of changes in genes, morphology, physiology, and behaviour in which the present day’s domesticated species differ markedly from their wild relatives. These findings suggest that individual-specific patterns in behaviour rather than size or reproductive capacity have been the key factors of artificial selections that humans have imposed on some species during domestication (Belyaev 1969 ; Trut 1999 ). Hence, humans must have had been able to mentally construct and socially represent individual-specific patterns in behaviour in some nonhuman species already tens of thousands of years ago. Very likely, pertinent representations and semiotic symbols referring to their conspecific individuals had already been developed before.

Individual members of many nonhuman species also develop representations, such as of their environment (Tolman 1948 ). In some species, in particular those with complex social systems, individuals also develop socially shared representations as reports on community formation and intercommunity warfare in chimpanzees suggest (Goodall 1986 ; Standford 1998 ). Individual members of some animal species develop representations that are socially shared even with humans, among them bonobos (Savage-Rumbaugh and Fields 2000 ), gorillas (Patterson and Linden 1981 ), dogs (Kaminski et al. 2004 ), and grey parrots (Pepperberg 2002 ). But none of these species have developed complex semiotic systems on their own so far. Without pertinent semiotic systems, however, the development and propagation of socially shared representations is substantially impeded.

Whether some nonhuman species are able to recognise and mentally construct individual-specific patterns of behaviour among their conspecifics is still unknown. Evidence of individualised dyadic relationships among nonkin (e.g., in Rhesus macaques; Weinstein and Capitanio 2008 , 2012 ) suggests some degrees of pertinent abilities at least in nonhuman primates. But more research on the abilities that are prerequisites for such abilities (e.g., perception of fluctuating phenomena, perception of time, see above) is needed. For example, many animal species perceive their environments in far greater detail than most humans are able to (Grandin and Johnson 2005 ), such as chimpanzees that have been shown to have eidetic working memories (Inoue and Matsuzawa 2007 ). Brunswik’s ( 1952 ) concept of narrow perceptual lens filters shifting flexibly across the environment may not apply well to them. On the one hand, these abilities may allow these nonhuman species to perceive individual behaviours more accurately than humans. But on the other hand, this massive amount of information and their more limited capacity for information processing and long-term memory may hinder their ability to recognise and construct individual-specific patterns in these perceptions (see above). Research on the pertinent abilities of humans who have eidetic perceptual or memory abilities (e.g., some people on the autistic spectrum) can provide important insight in this regard.

But perhaps nonhuman species’ lack of language or of a similarly efficient semiotic system to exchange information about individuals and to enable abstract thinking (Peirce 1902 , CP 4.227; Vygotsky 1962 ) does not allow them to construct individual-specificity in the highly abstract form that humans are able to. Individualised relationships may indicate that individual members of some species are able to construct that some specific others show particular behavioural regularities towards them, which may allow some predictions of their future behaviour. But they may perhaps not be able to construct that individuals in their social environment generally differ from one another in particular ways without referring to specific ones . Constructs of personality differences enable humans to quickly assess strangers (McAdams 1994 ) based on category systems that have proven to be socially significant within the community. Along with the ability to form impressions of others rapidly, which may have co-evolved dialectically, this allows people to gain some cognitive control of interactions with unknown individuals (cf. Goldberg 1981 ). Maybe the ability to construct personality differences is an essential cognitive tool that enables humans to deal with anonymous others, which seems to be a uniquely human ability (cf. Blaffer-Hrdy 2009 ) at least among mammalian species. This hypothesis shall be called the personality-constructs-promote-peaceful-anonymous-contacts hypothesis . It suggest that ancestral humans must have developed these abilities at least when they began dealing with unknown individuals outside their particular community on a more regular basis, such as for peaceful traffic, exchange, and trade, for which direct archaeological evidence exists.

Humans of the present day are born into a world full of complex semiotic systems and social representations. Individuals’ own perceptual constructs, those that are socially shared within their sociocultural community, and their lexical encodings therefore develop in tight dialectical interplays with one another and cannot be disentangled anymore (Lahlou 1996 , 2001 , 2008 ). They also develop in dialectical interplays with the individual-specific patterns in behaviours and in outer appearances to which they refer. Individuals like to belong and therefore adjust their behaviours in order to socially conform. At the same time, they like to set themselves apart as individuals within their communities and to stress their individual-specificity—both in behaviour and in an artificially modified outer appearance (though to different degrees in different sociocultural communities). The possibility for semiotic exchange about individual differences may therefore have promoted individual diversification in humans—as compared to other species that cannot directly communicate about their perceptual constructs at the level of their constructed meaning. Instead, they each have to develop these constructs individually based on their own perceptions and interpretations of what they perceive. Consequently, the development of semiotic systems could have been a driving force for individual diversification during human phylogeny. This fascinating hypothesis, which shall be called the language-promotes-individualisation hypothesis, can be investigated only if research methodology is substantially enlarged and diverse species are studied (see below).

For modern personality psychology, semiotic exchange entails additional important implications. Through semiotic exchange, scientific knowledge also becomes disseminated to the public. Many scientific concepts—some as old as Galens’s four temperamental types, others as scientifically contentious as Freud’s personality theory—have become established parts of everyday psychological knowledge. How scientific concepts disseminate in everyday psychology was studied in psychoanalytic concepts and their dissemination among the French public in the 1960s (Moscovici 1961 ). Everyday psychologists learn about scientific developments through the public media, education guidebooks, and popular science books. The interest in personality is particularly lively given its importance in each individual’s everyday life. Terms originating from science, such as melancholic, choleric, extraverted, neurotic, or psychotic, are widely used in everyday psychology—correctly or not—as are the ideas that childhood experiences and genetic inheritance influence adult individual behaviour. The body of everyday psychological knowledge is not free of scientific knowledge anymore as it may have been in prehistoric times.

This has far-reaching implications for scientific personality psychology. If lexical descriptions—in lexical approaches and in assessments—continue to be the primary source of information for the scientific study of personality, this not only triggers further explanatory circularity, but also leads to a growing convergence between everyday psychological and scientific constructs of personality. As a result, scientific concepts of personality increasingly appear to “explain” lexically encoded individual differences better—in particular because (psychology) students comprise important parts of empirical samples. Such findings suggest progress in science—when in fact scientific understanding of the described phenomena has not necessarily advanced. John B. Watson’s and Sigmund Freud’s theories of personality development had profound (and fairly different) impacts on educational practices and on everyday explanations of individual behaviour (Wiggins 2003 ). These practices and explanations in turn may have influenced the individuals’ behaviours themselves—though not necessarily in the ways postulated by these theories.

The strong focus of personality psychology on human language, be it through lexical research or assessment methods, has hindered these illuminating insights. The ancient philosopher Plato (428–348 BC) had already recognised that “one should leave the study of words behind to investigate the realities expressed through them” (Politeia, 438d2-439b8, cited in Bordt 1998 , p. 156; translated). This is essential because, when words are studied instead of real phenomena, criteria for judging empirical results cannot be established because there will be only many different names or descriptions for phenomena, but no way to decide which descriptions—if any at all—should be accepted (Toomela 2011 ). Personality psychology should finally explore the fascinating phenomena with which it concerns itself in their entirety and from much broader perspectives than previously done.

Ten Desiderata for Future Research

It is not important that we better understand which particular behaviours socially shared knowledge about personality allows us to predict or which minimal cues of whatever kind can activate this knowledge because these questions are intrinsically circular (see Fig.  2 ). Studying global structures of everyday psychological knowledge without setting them in relation to the particular contexts on which they depend is also not very important to do. The “discovery” of five major dimensions has provided merely a snapshot of the everyday psychological knowledge that particular semiotic communities had encoded about individuals in particular sociocultural environments and in particular historic times.

It is essentially important, however, to fully consider in their entirety the meta-theoretical assumptions that underlie the lexical hypothesis, to systematically test the hypothesis using inductive approaches, and to move personality psychology beyond this particular hypothesis to create new knowledge. Ten desiderata are thereby primary.

Desideratum 1: A Philosophy-of-Science Framework for Personality Psychology

A comprehensive philosophy-of-science framework that explicitly incorporates seven central issues ( a – g in the following) should be developed.

  • Meta-theoretical definitions of the scientific phenomena constructed as personality, i.e. , of individual-specific patterns in behaviour and outer appearance and of the everyday psychological knowledge about such patterns, including specifications of the principles and conditions for their demarcation from real phenomena and specifications of their ontology . For individual-specific patterns in behaviour, such a meta-theoretical framework has already been elaborated (see above, Uher 2011a ) and successfully applied in empirical studies on capuchin monkeys (Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished) and great apes (e.g., Uher et al. 2008 ). This framework comprises principles for defining behavioural phenomena (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished) and for defining entities (i.e., categories) among their qualitative elements (Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished). It explicitly considers both within- and between-individual variations in the contexts of situations, within-species populations (e.g., cultures), and species using ipsative-normative-populative-speciative approaches (Uher, An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, unpublished). To differentiate between the different sets of phenomena to which the lexical hypothesis refers, the framework relies on two meta-theoretical kinds of taxonomic personality constructs—one for the phenomena of encoded knowledge and one for the phenomena that this knowledge is about (i.e., scientific-statistical second-order and first-order constructs; Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished). Corresponding frameworks for individual-specific patterns of outer appearance and for the pertinent everyday psychological knowledge still have to be elaborated. The development of these frameworks can capitalise on the meta-theoretical elaborations of individual-specificity in the dynamic and fluctuating phenomena outlined above, but should also draw on developments made by other disciplines that study physical appearance (e.g., research on kin recognition and attractiveness) and language and knowledge systems (e.g., psycholinguistics, semiotics, computational linguistics).
  • Questions that can be asked about the thus-defined scientific phenomena . Tinbergen’s ( 1963 ) four key questions about the causation, ontogeny, adaptivity, and phylogeny, originally formulated for the species-level study of behaviour, constitute meaningful directions also for the individual-level study of behaviour (Uher 2008a , 2011a ) and outer appearance. They can also be applied in part to study the systems of everyday psychological knowledge about individuals, the systems of their semiotic encodings, and the sociocultural and historic variations in these systems.

To establish comprehensive taxonomies of individual-specific behaviours, a new manifest system approach that is grounded in the above-mentioned meta-theoretical framework (Desideratum 1a)—the Behavioural Repertoire x Environmental Situations Approach (Uher 2008a , b , 2011a , b ; Uher, Methodological approaches to personality taxonomies: The Behavioural Repertoire x Environmental Situations Approach—A non-lexical alternative, unpublished)—has been developed. This approach bases element selection on the behavioural-ecological systems of a population or species as far as these have already been scientifically reduced (i.e., are manifest). It formulates a systematic strategy to generate constructs and select their operationalisations based on the established scientific descriptions of the average individual’s behaviours and of the environmental situations in which these typically occur. In line with the philosophy of science outlined above, these constructs and the taxonomic constructs that can be empirically derived from them are descriptive; they are not ascribed an a priori causal status. As is true for any research, this approach necessarily relies on language. But in contrast to lexical approaches, it is not guided by the lexical encodings of lay people’s representations in which descriptions of behaviour are often blended with interpretations, appraisals, and explanations and loaded with implicit sociocultural meanings. This new approach has already been successfully applied in studies on capuchin monkeys (e.g., Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished) and great apes (Uher 2011b ; Uher et al. 2008 ) in which it yielded substantial empirical evidence for individual-specific patterns in behaviour not previously described by other approaches, including those derived from human everyday language. Corresponding methodological approaches for taxonomising individual-specificity in outer appearance should be developed by capitalising on the expertise of other disciplines (cf. Desideratum 1a).

In assessment-based investigations, the encoding of qualitative and quantitative properties of the selected elements into data largely relies on implicit decisions by lay people. The particular elements of the sets B, S, T, and I that people consider in their representations and assessments are thus ill defined, as are the elements of the sets of interpretations, appraisals, and explanations that these may comprise. Everyday psychological knowledge should therefore be analysed by using qualitative methods (e.g., Diriwächter et al. 2004 ; Rosenbaum and Valsiner 2011 ) and lexical analyses of texts, conversations, and reports obtained in everyday life (e.g., with methods of computational linguistics; Bolden and Moscarola 2000 ) that allow for the explication of the conversion and encoding decisions made.

This is an important issue for any kind of taxonomic research that is based on statistical methods. Assessment tools are developed by selecting variables that yield data with high internal consistency and that thus measure the same concept—i.e., redundancies (Block 2010 ). But in behaviour, redundancies may be rare and constrained by evolutionary processes (Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished). In fact, it is well known in psychology that the cross-situational consistency of individual behaviour is often only moderate (Mischel 1968 ). Even within a given situation, the internal consistency of behaviours that are assigned to the same construct in both everyday and scientific psychology is often low to zero, such as with regard to gaze aversion, long pauses in speech, hesitant speaking, and restricted gestures all of which are assigned to the construct of “shyness” (Asendorpf 1988 ).

In a nutshell, it must be considered that behavioural data have different properties than assessment data because they capture different kinds of phenomena. The distributional patterns of behavioural events may differ from those of the mental representations that people develop of them. Behavioural measurements may not fulfil the psychometric standards that have been established for assessments that use predefined scales. This need not indicate insufficient utility for scientific investigations, but instead may reflect real patterns that are important and should be researched (Fenson et al. 2000 ). The normal curve distributions of individual scores that are frequently assumed for the five major dimensions (Jang et al. 2002 ; Loehlin et al. 1998 ) might simply result from processes of mental information processing, from the limited response formats of questionnaires, and from the artificial selection of only those questionnaire items that yield such distributional patterns in the target population (Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished).

To meet these challenges, new meta-theoretical principles of data reduction have been proposed within the framework of the Behavioural Repertoire x Environmental Situations Approach (Uher, Methodological approaches to personality taxonomies: The Behavioural Repertoire x Environmental Situations Approach—A non-lexical alternative, unpublished; Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished). They allow researchers to employ a two-step reduction procedure. First, behavioural data are reduced based on the particular behaviours’ scientifically established functional importance—regardless of potentially low internal consistencies among behavioural measurements. Then the thus-derived composite measures of functionally defined constructs (rather than the raw behavioural data) are subjected to statistical reduction techniques. The first reduction step corresponds to the intuitive processes of mental (re)construction, but is, in contrast to them, made explicit and based on scientific knowledge. This step can therefore always be traced and reconsidered if needed.

Statistical principles of data analyses should also be reconsidered with regard to within-individual variability. Prevailing methods are based on the meta-theoretical assumption that variability reflects variance derived from measurement error and from random variation around a hypothetical true score (i.e., a “trait”). But variability can be an important phenomenon in itself. Dynamic system theories, for example, consider variability to be a driving force of development and a potential indicator of ongoing processes. Studying variation patterns can therefore offer important insights into how individuals change and develop (Thelen and Smith 1993 ). Analytical principles that are adapted to the developmental nature of behavioural phenomena—and thus of personality—and that allow for analyses of variability should be further developed (see van Geert and van Dijk 2002 ).

Finally, interpretations should carefully consider what phenomena the obtained data (can) actually reflect (Gillespie and Zittaun 2010 ; Køppe 2012 ; Toomela 2011 ; Valsiner 2012 ). For example, analyses of lay people’s assessments cannot reveal the relative contributions of genetic inheritance of and environmental influence on individual differences nor can they be insightful about their evolutionary genetics because assessments capture everyday psychological knowledge structures. This knowledge is transmitted nongenetically through behavioural and symbolic inheritance systems (Jablonka and Lamb 2005 ). The results reported in pertinent assessment studies can thus mean two things. If—despite fragmentary human perception, imperfect mental (re)construction, and socioculturally negotiated ascriptions of meaning— this knowledge accurately captures some real structures in the phenomena about which it has been developed, then they are promising. But if—through mental and social (re)construction—this knowledge reflects more structure than there actually is in the phenomena that are being constructed or if it reflects structures in phenomena other than the phenomena that this knowledge is assumed to be about, then these results may be misleading. Psychologists can find this out only if they study the phenomena that can be influenced directly and indirectly by genetic transmission (e.g., physiology, behaviour) rather than the pertinent symbols and representations that lay people develop of them.

  • Scientific language . A scientific language that transcends common sense and that encodes knowledge in specially established sign systems should be developed purposefully for personality psychology as has already been done for the (hard) natural sciences (cf. Valsiner 2012 ). For example, although the statistical derivation of taxonomic personality constructs is described explicitly and precisely in lexical research (e.g., with factor loadings), their meanings remain implicit because they are derived from abstract everyday psychological terms. These terms cannot be easily traced to concrete perceivable phenomena anymore and are loaded with implicit meanings that are specific to particular sociocultural communities. Instead, terminology for personality constructs that begins with specific terms and constructs that are close to the directly perceivable qualitative and quantitative properties of the defined study phenomena for encoding the data should be developed (cf. Desideratum 1d). Then, these specific constructs should be processed during the reduction and analysis of the data using formal-logical operations to generate in a stepwise fashion more abstract and complex constructs that can be clearly traced to the concrete references from which they were derived. For these traceable constructs, abstract, complex, and even new terms should be developed for scientific psychology. This will help to reduce the unintended confusion with everyday psychological terms that all personality researchers have in their minds as well. For example, thus-derived complex constructs of individual-specific patterns of behaviour can be labelled based on the bio-psycho-socio-ecological functions of the particular behaviours and environmental situations to which they refer (Uher 2008b ; Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished).
  • Structures that allow for the constant scrutinising of epistemology, ontology, and specific theories that have been developed about the phenomena studied . Meta-theoretical (Toomela 2011 ), cultural and ethical structures that facilitate discussions at the philosophy-of-science level and that help to delimit the paradigmatic and ideological practices that hinder intellectual innovation should be established in psychology communities and their publication media. Among all seven issues, this may be the most difficult to achieve.

Desideratum 2: Comprehensive Taxonomies of Individual-Specific Patterns in Behaviour and of Individual-Specific Patterns in Outer Appearance

The structures of individual-specific patterns in behaviour and in outer appearance are still largely unknown (cf. Uher 2008a , b ; Westen 1996 ). They should be researched and taxonomised using the philosophy-of-science framework outlined above (Desideratum 1) to the same systematic and comprehensive extent as has already been done for the knowledge structures of some language communities—for humans and for some other species as well. Three issues must be considered.

  • Behaviour is a phenomenon of the present and thus requires realtime recording. To capture the defined elements of behaviour in the flow of events, methods that help reduce the limitations that are entailed by human perceptual abilities and by the actions required for recording what is perceived should be used and further developed. For example, computerised life observations, video techniques, and software-assisted coding techniques enable the creation of detailed records of complex and quickly occurring events (e.g., Uher et al. 2008 ; Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished). Methods of first-person recording (Lahlou 2011a , b ) and of ambulatory monitoring (Fahrenberg and Myrtek 2001 ; Fahrenberg et al. 2007 ) enable computerised measurements of ongoing behaviours in daily life settings. Studies on outer appearances could employ photographical techniques.
  • Individual-specific patterns are temporal phenomena and are thus not phenomena of the present. Their identification requires repeated measurement occasions over time and posterior statistical analyses; they cannot be measured directly (Uher, Meta-theoretical foundations of objectivity versus subjectivity in quantifications of behaviour and personality, unpublished).
  • Phenomena of behaviour and of artificially modified outer appearance must be studied in the contexts in which they emerge. As phenomena that are functionally mediated by the present environment (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished), behaviour and artificially modified outer appearance must be studied together with the contexts in which they emerge, such as particular situations, groups, populations (e.g., nations, sociocultural, or semiotic communities), times (e.g., generations, historical eras), and species. This requires taxonomic approaches for categorising contexts. Such can be taken and further developed from approaches that are designed to study the behavioural contexts of average individuals, such as from behaviour setting theory (Barker 1968 ; Schoggen 1989 ), world installation theory (Lahlou 2008 ; 2011a ), cross-cultural psychology (Berry et al. 2011 ), and behavioural ecology. This also includes analyses of the similarities and differences of the obtained taxonomic structures of individual-specific behaviours and outer appearances within and across contexts to identify patterns that are specific to particular contexts or universal across multiple defined contexts (e.g., situations, sociocultural and semiotic communities, species; Uher 2008a , b , An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, unpublished).

Desideratum 3: Analyses of the Relations Between the Taxonomic Structures of Individual-Specific Behaviours and those of Individual-Specific Outer Appearances

The identified taxonomic structures of individual-specific patterns in behaviour and in outer appearances should then be analysed purposefully and systematically for interrelations between them. The background of defined contexts in which they have been studied (e.g., situations, groups, populations, times, species) must thereby be carefully considered.

Desideratum 4: Analyses of How People Perceive and Construct Structures of Individual-Specific Patterns of Behaviour and of Outer Appearance

Still little is known about what exactly people perceive in individual behaviour and in outer appearance in everyday life, and about what specific communities actually consider salient and relevant. “Although [the lexical hypothesis] is reasonably clear, the criteria of importance that have shaped the personality lexicons of everyday people are not well understood” (John et al. 1988 , p. 175). For example, within-individual variance often considerably exceeds between-individual variance in humans (Shweder and Sullivan 1990 ) and also nonhuman species (Uher, An integrative meta-theoretical framework for research on individual behaviour in context—situations, populations, species, unpublished). But still, between-individual variability seems to be more salient to human minds and more central to pertinent representations. What actually happens during the processes of the perception and mental (re)construction of the perceived phenomena that this is possible? The concept of time-relative probabilities allows researchers to determine the magnitudes of within-individual and between-individual variation (measured in set-theoretically objective ways, cf. Desideratum 1d) that particular people—both individuals and communities—perceive as salient in other humans and in some other species (Uher et al., Through the human personality glasses: Constructed personality taxonomy and typology in Crab-eating macaques ( Macaca fascicularis ), their crossmethod coherences and 24-month stabilities, unpublished). Researchers should also study how people construct, interpret, appraise, explain, and represent the perceived (see e.g., Laucken 1974 ), and in which ways this may be different in different contexts (e.g., cultures and semiotic systems). In a nutshell, the processes of perception and mental and social construction that create such knowledge and their lexical encodings—rather than just their outcomes—should be systematically studied (Gillespie and Zittaun 2010 ; Jovchelovitch 2007 ; Komatsu 2012 ; Pillai 2012 ; Rosenbaum and Valsiner 2011 ; Westen 1996 ). A portfolio of methods (see Desideratum 1d) should thereby be used, qualitative and microgenetic ones in particular (Wagoner 2009 ; Valsiner 1998 ).

Desideratum 5: Comprehensive Taxonomies of Everyday Psychological Knowledge About Individual-Specific Patterns of Behaviour and of Outer Appearance

The pertinent bodies of everyday psychological knowledge have so far been taxonomised primarily by beginning with the systems of their lexical encodings (John et al. 1988 ). An entirely alternative strategy that begins with the phenomena that are represented in this knowledge (i.e., from behaviours and outer appearances) should be pursued. To taxonomise behaviour-related everyday psychological knowledge, the Behavioural Repertoire x Environmental Situations Approach (cf. Desideratum 1a) can be applied. For such investigations, the elements selected from the scientifically described behavioural ecology of a population or species are described in the everyday language of the target community. It is thereby important to use specific terms and descriptions that are close to perceivable qualitative properties (e.g., in contextualised behaviour-descriptive verb sentences). These descriptions can then be used in empirical investigations of pertinent representations using assessment methods or qualitative methods (see Desideratum 1d). This approach has already been applied to study (using assessments) the representations that human observers develop about individual-specific behaviours in other species; specifically those that zoo keepers developed about great ape individuals (Uher and Asendorpf 2008 ) and those that students and researchers developed about individual crab-eating macaques (Uher et al., Through the human personality glasses: Constructed personality taxonomy and typology in Crab-eating macaques ( Macaca fascicularis ), their crossmethod coherences and 24-month stabilities, unpublished). Similar approaches should be used to taxonomise everyday psychological knowledge about individual-specific outer appearances. Ideally, taxonomising knowledge structures should be based on the empirically established comprehensive taxonomies of individual-specific behaviours and outer appearances and on the knowledge about their empirical interrelations (see Desiderata 2 and 3). But because these are difficult to establish given the enormous logistic efforts required, the aforementioned approach will provide a viable alternative until such taxonomies can be established empirically.

In addition to descriptions, researchers should systematically study patterns of interpretation, appraisal, and explanation and compare them among sociocultural and language communities. Such analyses can be done only with qualitative methods, e.g., using interviews or analyses of texts, conversations, and reports (Lahlou 2011a ). For example, Laucken ( 1974 ) systematically collected and catalogued all verbal manifestations of lay psychology explanations and predictions of human behaviour, such as in newspapers and everyday conversations that he came across over a period of 1.5 years. Then he analysed and categorised the structures and the conceptual elements underlying these verbal manifestations to reconstruct the “naïve behaviour theory” that people were using in everyday life in the population that he studied at the time of his investigation.

Desideratum 6: Analyses of the Relations Between the Taxonomic Structures of Everyday Psychological Knowledge and Those of Individual-Specific Behaviours and of Individual-Specific Outer Appearances

The taxonomies of everyday psychological knowledge should be set systematically in relation to the taxonomised structures of behaviours and outer appearances to explore convergences and divergences therein both within and between different cultural and semiotic communities. Such analyses will unravel how taxonomies about everyday psychological knowledge and those derived from their lexical encodings actually represent individual-specific patterns in behaviour and in outer appearance; the ways in which they may reflect different perceptions, interpretations, appraisals, and explanations; and how this may vary within and among different sociocultural and language communities. The above-mentioned studies on the representations of human observers of great apes and crab-eating macaques found systematic coherence between the human observers’ pertinent represented knowledge and contextualised measurements of individual-specific behaviours (likewise selected using the Behavioural Repertoire x Environmental Situations Approach). But they also found interesting deviations that provide insight into how humans may perceive and (re)construct such patterns in these species (Uher 2011b ).

Desideratum 7: Systematic Investigations of Causally Related Phenomena

The established descriptive-structural taxonomies of individual-specific behaviours and of outer appearances and the taxonomies of the related contexts (cf. Desideratum 2c) should be used to guide causal-explanatory investigations of these phenomena. This strategy has crucial advantages over approaches that begin with the causally related phenomena themselves because behaviours, outer appearances, and environments are exterospectively accessible and can therefore be systematically categorised and quantified in set-theoretically objective ways (see above). But this is not possible for psychological processes and for representations. Because they are not directly perceivable; approaches that begin with these phenomena are prone to conceptual biases on the part of the researchers. This is also in line with the methodological insight made by social-cognitive personality theorists that descriptive-structural investigations of behaviours in their situational contexts are essential for uncovering the psychological processes through which distinctive individual behaviours emerge and endure in specific situations (e.g., Wright and Mischel 1987 ; Wright and Zakriski 2003 ).

The phenomena that may be causally related to individual-specific behaviours and outer appearances, and to their pertinent representations should be conceptually integrated into the philosophy-of-science framework outlined above (for a meta-theoretical definition of the psyche, see, e.g., Toomela 2010b ). As with behavioural and psychological phenomena, it is important to study the potential causes of behaviours, outer appearances, and representations separately because they need not be the same for these different kinds of phenomena (cf. Toomela 2011 ). For example, the cognitive and social processes that create encoded knowledge structures (Witkowski and Brown 1978 ) can but do not necessarily causally influence behaviours and outer appearances as well. Given their order of emergence in ontogeny and phylogeny, they may not even be the primary causal influences of behaviour. Research on nonhuman species can be particularly insightful in this regard. Among causal phenomena internal to the individual, phenomena of physiology, (neuro)anatomy, and genes can be measured quantitatively because these are physical phenomena. But for investigations of psychological phenomena, three important issues must be considered.

  • Most psychological processes are unconscious (Freud 1915 ). Individuals are consciously aware of only small subsets of their psychological processes (Westen 1996 , 1999 ) and many of these processes cannot be easily verbalised (Brower 1949 ; Kelly 1955 ; Komatsu 2012 ; Valsiner 2012 ).
  • Psychological phenomena are entirely internal and accessible only through introspection . Psychological events cannot be perceived in other individuals ( Locke 1689 ; Toomela 2008 , 2011 ). They can only be inferred from phenomena that are external and that may mediate between an individual’s psychological events and his or her environment (see above, Fig.  1 ), such as behavioural phenomena and some parts of human language (Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished). In oneself, psychological events can be directly perceived—through introspection. But scientific introspection imposes challenges because reflection and attention inevitably introduce changes in the course of psychological events (Wundt 1904 ). Moreover, the results of introspection can be investigated only based on their externalisation through behaviour and language. Direct investigations of psychological phenomena are impossible.
  • Psychological phenomena are bound to the present and thus can be studied only in the present (Gillespie and Zittaun 2010 ; Uher, What is behaviour? And (when) is language behaviour? A meta-theoretical definition, unpublished; Valsiner 1998 , 2012 ). Disentangling constructs about psychological phenomena (e.g., anxiousness) from ongoing events of psychological phenomena (e.g., experiencing anxiety in a given moment) is difficult because both are internal and are externalised primarily in a verbal manner. For the phenomena of thought, this entails particular difficulties because constructs are the memorised resultants of past thoughts (Peirce 1902 , CP 2.84; Valsiner 2012 ). The lexical encodings used to externalise psychological events are constructs as well (Vygotsky 1962 ), including those that refer to ongoing events of psychological phenomena. Given that one can never ascertain whether words denoting particular psychological phenomena actually refer to exactly the same phenomena in different individuals ( Locke 1689 , see above), (re)constructions by investigators are inevitably involved in research on psychological phenomena.

The fact that all psychological phenomena are actualities and inherently bound to the present (Gillespie and Zittaun 2010 ; Valsiner 1998 , 2012 ) opens up some ways—within the limitations mentioned— in which ongoing psychological events can be distinguished from pertinent constructs because constructs inherently also involve events from the past. As for behavioural phenomena, this requires realtime investigations of ongoing psychological phenomena, such as through microgenetic methods (Wagoner 2009 ) or methods of experience sampling in real-life settings (Mehl and Connor 2012 ). But because psychological phenomena cannot be investigated directly, this requires that researchers differentiate their qualitative properties by explicitly categorising the phenomena through which they can be externalised (see Desideratum 1d). These categories may refer, for example, to particular information externalised in language (e.g., complains about pains, expressed worries) or to nonverbal behaviours. Humans usually make these assignments intuitively, for example, between behavioural and psychological assignments as is apparent in concepts of “emotional behaviour” or “helping” that already contain assumptions about underlying emotions or intentions. Because these assumptions may be erroneous, in particular if individual members of other sociocultural communities and of other species are concerned, the assignment of externalised phenomena to particular psychological phenomena should be made explicit to enable continuous scrutinising. For quantifications of individual-specific patterns, realtime records of events of externalisation categories should be repeatedly obtained and accumulated over time. Probabilistic, differential, and temporal patterns of their occurrences can then be quantified based on time-relative probabilities (see above). Compared to the ease of obtaining lay people’s assessments, this may appear to be a fairly cumbersome and rough way to investigate psychological phenomena. But given their peculiarities, this may be the only way in which the psychological phenomena themselves—rather than lay people’s knowledge about them—can be somewhat quantified in set-theoretically objective ways.

Desideratum 8: Investigations of Ontogenetic Development

The development of individual-specific behaviours, outer appearances, pertinent knowledge structures, and encodings should be studied systematically. Close attention should be paid to the fact that, in present-day humans, none of these patterns can be explored and understood independently of the others because they always develop in tight dialectic interplay. Methods of investigation and developmental theories must therefore be primarily dialectical rather than demonstrative as those currently prevailing in psychology (Rychlak 1968 ; Sameroff 2010 ; Valsiner 2012 ). Particularly illuminative are investigations of early ontogeny when the patterns of behaviour and outer appearances of human individuals can be studied before and while these individuals develop the abilities to mentally construct pertinent representations and acquire complex repertoires of socioculturally shared meanings and symbols. The influences of the latter should be explored in comparative investigations of a broad range of human cultures (Keller 2007 ) by explicitly considering cultures that are markedly different from the Western societies on which much of today’s personality psychology is focused. Adult individuals should be studied with regard to their own behaviours, outer appearances, and pertinent representations, but also with regard to the representations that they develop about young children who do not yet share the adults’ social representations comprehensively and whose behaviours are therefore still not so strongly influenced by the sociocultural appraisals they comprise.

Research on various nonhuman species will illuminate patterns of ontogenetic development of individual-specific behaviours (Suomi 2005 ) and also of individual-specific outer appearances that may be universal to some groups of species. For example, in various species such as horses and dogs, individuals with more gracile bones and limbs also seem to differ in their individual-specific behaviours from conspecifics that are more robust in their physical appearance (Grandin and Johnson 2005 ). Furthermore, the abilities that are prerequisite for the recognition and mental reconstruction of individual-specific patterns in behaviour should be studied in various species as well as in humans with special abilities (e.g., eidetic perception) or particular impairments (e.g., deficiencies in time perception). Differences in these abilities should be set in relation to evidence that may indicate the recognition and reconstruction of individual-specific patterns, such as individualised dyadic relationships. Research on nonhuman individuals that have acquired semiotic systems that are socially shared with humans can provide insights into the role that language may have played in the development of pertinent abilities during human evolution.

Desideratum 9: Alternative Conclusions from the Lexical Hypothesis

So far, psychologists have deduced just one conclusion from the lexical hypothesis (i.e., all pertinent terms in the lexica constitute the universe of elements for empirical studies). Alternative deductive conclusions should be derived systematically, such that person perceptions may also be encoded in idioms and proverbs. In contrast to single words, idioms and proverbs usually do not translate well; literal translations often change their meaning or render them meaningless. What do these different ways of encoding knowledge reveal about the different sociocultural and linguistic communities and their ways of representing their social world?

Furthermore, inductive conclusions should be derived and facts should be sought that may support or even falsify the hypothesis. For example, researchers should purposefully study how people construct and encode the phenomena that they perceive, even when these perceptions are not conscious (e.g., subliminal odours, pheromones). The fact that within-individual differences are not mentioned in the lexical hypothesis should be scrutinised. Do people perceive and mentally represent this at all? And do they notice that some individuals are more consistent within themselves and across situations, whereas others are not both in humans (cf. Caspi and Roberts 1999 ; Mischel et al. 2002 ) and in (at least some) nonhuman species as well (Uher et al., Contextualised behavioural measurements of personality differences obtained in behavioural tests and social observations in adult capuchin monkeys ( Cebus apella ), unpublished)?

Desideratum 10: Alternative Nonlexical Hypotheses

Fundamentally new hypotheses about individual-specificity should be developed, in particular about the theoretical nature of the phenomena in which it can be found, why it emerges in these phenomena at all, and how. These questions can be explored only if personality psychology dismisses its narrow research focus on Homo sapiens and the phenomena reflected in its language.

The shift in perspective from humans to the enormous diversity of today’s species opens a huge field of research that allows profound and illuminative insights …. What is unique about Homo sapiens compared to all other species in the phylogenetic tree? What personality [differences 4 ] may have contributed to Homo sapiens’ accelerated development in the most recent evolutionary past and to its unmatched success in conquering almost every habitat on earth? There is no better opportunity to understand the phylogenetic basis, adaptive significance and ecological relevance of personality and its role in speciation than studying the evolved diversity of species. (Uher 2008a , p. 427–428)

The much wider focus of cross-species comparative psychology may challenge some deep-rooted ideologies about human personality. But it opens doors to alternative hypotheses that trigger new approaches and insights into the phylogenetic history of these fascinating phenomena in humans and other species, such as the personality constructs-promote-peaceful-anonymous-contacts hypothesis , the language-promotes-individualisation hypothesis (see above), the central nervous system hypothesis, or the motility hypothesis (Uher 2009 ) to name just a few. All these new hypotheses—just as the lexical hypothesis is—are centred on the fascinating finding that individual members of the human species as well as of other species develop individual-specific styles of interaction with their environments that we construct as personality.

Those who say it can’t be done are usually interrupted by others doing it. (James Arthur Baldwin)


I am very grateful to Manfred Schmitt, Jochen Fahrenberg, Jaan Valsiner, Grete Arro, and an anonymous reviewer for thoughtful comments on earlier drafts of the manuscript. The views expressed herein are mine and should not be attributed to any of the persons who provided commentaries. Preparation of the manuscript was supported by a grant from the Deutsche Forschungsgemeinschaft DFG (grant number UH249/1-1).

received her PhD from Freie Universität Berlin in 2009 where she is heading the research group Comparative Differential and Personality Psychology that she founded in 2010. Her research is transdisciplinary concentrating on philosophy-of-science issues of psychological and behavioural research on individuals in humans and nonhuman species from species-comprehensive and culture- and species-comparative perspectives. For demonstrating her metatheoretical and methodological developments, she has been working empirically with primate species in particular human children, the great apes, capuchin monkeys, and various macaques, and with plant species using noninvasive experimental and observational studies. Additionally, she has been conducting assessment studies with adult observers of young children and with human observers of nonhuman individuals respectively. Her empirical research projects are rooted in international and national collaborations with scholars from different disciplines. She has been working at the Max Planck Institute for Evolutionary Anthropology in Leipzig (2003–2005), and has also been a visiting scholar at the Institute of Cognitive Sciences and Technologies, National Research Council of Italy (ISTC-CNR) in Rome (2011–2012).

1 For the present meta-theoretical analyses, theories considering subjectivistic probabilities of the occurrences of events (e.g., degrees of belief; Gillies 2000 ; Rychlak 1968 ) will not be considered here (for reasons described in detail below).

2 The concept of propensity probabilities reifies observed properties into causal entities of an unkown and undefined kind. As such, the explanatory strategy is inherently circular and faces the same challenges as those elaborated for trait psychology in this article.

3 Personality research is often distinguished from intelligence research in American psychology. In European traditions (e.g., Cattell, Eysenck, Stern, Pawlik), intelligence and achievement are considered inherent parts of an individual’s personality.

4 The attentive reader will notice that the article referred to here and some additional ones are written based on the assumptions of trait psychology. This documents the educational origins of the author and the meta-theoretical development of her work.

  • Ackerman PL. Intelligence, attention, and learning: Maximal and typical performance. In: Detterman DK, editor. Current topics in human intelligence : Volume 4. Theories of intelligence . Norwood: Ablex; 1994. pp. 1–27. [ Google Scholar ]
  • Allport GW. Personality: A psychological interpretation. New York: Macmillan; 1937. [ Google Scholar ]
  • Allport GW. Pattern and growth in personality. New York: Holt, Rinehart and Winston; 1961. [ Google Scholar ]
  • Allport GW, Odbert HG. Trait names: a psycholexial study. Psychological Monographs. 1936; 47 :1. doi: 10.1037/h0093399. [ CrossRef ] [ Google Scholar ]
  • Althusser L, Balibar E. Reading capital. London: New Left Books; 1970. [ Google Scholar ]
  • Angleitner A, Ostendorf F, John OP. Towards a taxonomy of personality descriptors in German: a psycho-lexical study. European Journal of Personality. 1990; 4 :89–118. doi: 10.1002/per.2410040204. [ CrossRef ] [ Google Scholar ]
  • Aristotle (1963[1936]). Physiognomics . In Aristotle. Minor works (pp. 81–137). trans. W. S. Hett. Cambridge, MA and London: Harvard University Press.
  • Arro, G. (2012). Peeking into personality test answers: Inter- and intraindividual variety in item interpretations. Integrative Psychological and Behavioral Science . doi:10.1007/s12124-012-9216-9. [ PubMed ]
  • Asendorpf, J. B. (1988). Individual response profiles in the behavioral assessment of personality. European Journal of Personality, 2 , 155–167.
  • Ashton MC, Lee K. A defence of the lexical approach to the study of personality structure. European Journal of Personality. 2005; 19 :5–24. doi: 10.1002/per.541. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Augustine, Saint (397–398 AD/1961). Confessions . Translated by Pinecoffin, R. S. (Ed.), Harmondsworth: Penguin.
  • Bandura A. Social foundations of thought and action. Englewood Cliffs: Prentice Hall; 1986. [ Google Scholar ]
  • Barker, R. G. (1968). Ecological psychology. Concepts and methods for studying the environment of human behavior . Stanford University Press
  • Baumeister RF, Vohs KD, Funder DC. Psychology as the science of self-reports and finger movements: whatever happened to actual behavior? Perspectives on Psychological Science. 2007; 2 :396–403. doi: 10.1111/j.1745-6916.2007.00051.x. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Belyaev DK. Domestication of animals. Science Journal (U.K.) 1969; 5 :47–52. [ Google Scholar ]
  • Berry JW, Poortinga YH, Breugelmans SM, Chasiotis A, Sam D. Cross-cultural psychology. Research and applications. Cambridge: Cambridge University Press; 2011. [ Google Scholar ]
  • Bhutta M. Sex and the nose: human pheromonal responses. Journal of the Royal Society of Medicine. 2007; 100 :268–274. doi: 10.1258/jrsm.100.6.268. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Biesanz JC, Human LJ. The cost of forming more accurate impressions: accuracy motivated perceivers see the personality of others more distinctively but less normatively. Psychological Science. 2010; 24 :589–594. doi: 10.1177/0956797610364121. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Blaffer-Hrdy S. Mothers and others: the evolutionary origins of mutual understanding. Cambridge: Belknap; 2009. [ Google Scholar ]
  • Block J. A contrarian view of the five-factor approach to personality description. Psychological Bulletin. 1995; 117 :187–229. doi: 10.1037/0033-2909.117.2.187. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Block J. Millenial contrarianism: the five factor approach to personality description 5 years later. Journal of Research in Personality. 2001; 35 :98–107. doi: 10.1006/jrpe.2000.2293. [ CrossRef ] [ Google Scholar ]
  • Block J. The five-factor framing of personality and beyond: some ruminations. Psychological Inquiry. 2010; 21 :2–25. doi: 10.1080/10478401003596626. [ CrossRef ] [ Google Scholar ]
  • Bock PK. Culture and personality revisited. American Behavioural Scientist. 2000; 44 :32–40. doi: 10.1177/00027640021956071. [ CrossRef ] [ Google Scholar ]
  • Bolden R, Moscarola J. Bridging the quantitative-qualitative divide: the lexical approach to textual data analysis. Social Science Computer Review. 2000; 18 :450–460. doi: 10.1177/089443930001800408. [ CrossRef ] [ Google Scholar ]
  • Bordt M. Platon Werke. Übersetzung und Kommentar , Band V 4. Lysis. Göttingen: Vandenhoeck & Ruprecht; 1998. [ Google Scholar ]
  • Borkenau P, Liebler A. Trait inferences: sources of validity at zero acquaintance. Journal of Personality and Social Psychology. 1992; 62 :645–657. doi: 10.1037/0022-3514.62.4.645. [ CrossRef ] [ Google Scholar ]
  • Borkenau P, Ostendorf F. NEO-Fünf-Faktoren-Inventar nach Costa & McCrae . 2. neu normierte und vollständig überarbeitete Auflage. Göttingen: Hogrefe; 2008. [ Google Scholar ]
  • Borkenau P, Mauer N, Riemann R, Spinath FM, Angleitner A. Thin slices of behavior as cues of personality and intelligence. Journal of Personality and Social Psychology. 2004; 86 :599–614. doi: 10.1037/0022-3514.86.4.599. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Bouts P. La psychognomie. Paris: Dervy-Livres; 1931/1986. [ Google Scholar ]
  • Brody N. Heritability of traits. Psychological Inquiry. 1994; 5 :117–119. doi: 10.1207/s15327965pli0502_3. [ CrossRef ] [ Google Scholar ]
  • Bronfenbrenner U. The ecology of human development: Experiments by nature and design. Cambridge: Harvard University Press; 1979. [ Google Scholar ]
  • Brower D. The problem of quantification of psychological science. Psychological Review. 1949; 56 :325–331. doi: 10.1037/h0061802. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Brown DE. Human universals. New York: McGraw-Hill; 1991. [ Google Scholar ]
  • Brunswik E. The conceptual framework of psychology. Chicago: The University of Chicago Press; 1952. [ Google Scholar ]
  • Brunswik E. Representative design and probabilistic theory in a functional psychology. Psychological Review. 1955; 62 :193–217. doi: 10.1037/h0047470. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Carrasco MC, Bernal MC, Redolat R. Time estimation and aging: a comparison between young and elderly adults. International Journal of Aging & Human Development. 2001; 52 :91–101. doi: 10.2190/7NFL-CGCP-G9E1-P0H1. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Caspi, A., & Roberts, B. W. (1999). Personality continuity and change across the life course. In Pervin, L. A. & John O. P. (Eds). Handbook of Personality: Theory and Research (pp. 300–326, 2 nd ed). New York, NY: Guilford Press.
  • Caspi A, Roberts BW. Personality development across the life course: the argument for change and continuity. Psychological Inquiry. 2001; 12 :49–66. doi: 10.1207/S15327965PLI1202_01. [ CrossRef ] [ Google Scholar ]
  • Cattell RB. The description of personality I. Foundations of trait measurement. Psychological Review. 1943; 50 :559–594. doi: 10.1037/h0057276. [ CrossRef ] [ Google Scholar ]
  • Cattell RB. An introduction to personality study. London: Hutchinson; 1950. [ Google Scholar ]
  • Cattell RB. Factor analysis. New York: Wiley; 1952. [ Google Scholar ]
  • Cervone D, Shadel WG, Jencius S. Social-cognitive theory of personality assessment. Personality and Social Psychology Review. 2001; 5 :33–51. doi: 10.1207/S15327957PSPR0501_3. [ CrossRef ] [ Google Scholar ]
  • Church, A. T., & Katigbak, M. S. (1988). The emic strategy in the identification and assessment of personality dimensions in a non-Western culture: Rationale, steps, and a Philippine illustration. Journal of Cross-Cultural Psychology, 19 , 140–163.
  • Church, A. T., & Katigbak, M. S. (1989). Internal, external, and self-report structure of personality: An investigation of cross-language and cross-cultural generalizability. Journal of Personality and Social Psychology, 57 , 857–872.
  • Conley JJ. Longitudinal consistency of adult personality: self-reported psychological characteristics across 45 years. Journal of Personality and Social Psychology. 1984; 47 :1325–1333. doi: 10.1037/0022-3514.47.6.1325. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Costa PT, Jr, McCrae R. Four ways five factors are basic. Personality and Individual Differences. 1992; 13 :653–665. doi: 10.1016/0191-8869(92)90236-I. [ CrossRef ] [ Google Scholar ]
  • Darwin CR. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. London: John Murray; 1859. [ PMC free article ] [ PubMed ] [ Google Scholar ]
  • Daston L, Galison P. Objectivity. New York: Zone Books; 2007. [ Google Scholar ]
  • De Raad, B. (1998). Five big, Big Five issues: Rationale, content, structure, status, and crosscultural assessment. European Psychologist, 3 , 113–124.
  • De Weerth C, van Geert P, Hoitink H. Intra-individual variability in infant behavior. Developmental Psychology. 1999; 35 :1102–1112. doi: 10.1037/0012-1649.35.4.1102. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Digman JM. Personality structure: emergence of the five-factor model. Annual Review of Psychology. 1990; 41 :417–440. doi: 10.1146/annurev.ps.41.020190.002221. [ CrossRef ] [ Google Scholar ]
  • Diriwächter, R., Valsiner, J. & Sauck, C. (2004). Microgenesis in Making Sense of Oneself: Constructive Recycling of Personality Inventory Items. Forum Qualitative Sozialforschung/Forum: Qualitative Social Research, 6 , Art. 11.
  • Dunbar R. Grooming, gossip, and the evolution of language. Cambridge: Harvard University Press; 1996. [ Google Scholar ]
  • Eysenck HJ. Dimensions of personality. London: Routledge and Kegan Paul; 1947. [ Google Scholar ]
  • Eysenck HJ. Four ways five factors are not basic. Personality and Individual Differences. 1992; 13 :667–673. doi: 10.1016/0191-8869(92)90237-J. [ CrossRef ] [ Google Scholar ]
  • Fahrenberg J, Myrtek M, editors. Progress in ambulatory assessment. Computerassisted psychological and psychophysiological methods in monitoring and field studies. Göttingen: Hogrefe & Huber; 2001. [ Google Scholar ]
  • Fahrenberg J, Myrtek M, Pawlik K, Perrez M. Ambulatory assessment—monitoring behavior in daily life settings. A behavioral-scientific challenge for psychology. European Journal of Personality Assessment. 2007; 23 :206–213. doi: 10.1027/1015-5759.23.4.206. [ CrossRef ] [ Google Scholar ]
  • Fenson L, Bates E, Dale P, Goodman J, Reznick JS, Thal D. Measuring variability in early child language: Don’t shoot the messenger. Comment on Feldman et al. Child Development. 2000; 71 :323–328. doi: 10.1111/1467-8624.00147. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Fraisse P. The psychology of time. London: Eyre & Spottiswoode; 1964. [ Google Scholar ]
  • Freud S. The unconscious. Standard Edition. 1915; 14 :159–214. [ Google Scholar ]
  • Funder DC. The personality puzzle. 3. New York: W.W. Norton & Co; 2004. [ Google Scholar ]
  • Galton F. Measurement of character. Fortnightly Review. 1884; 36 :179–185. [ Google Scholar ]
  • Germonpré M, Sablin MV, Stevens RE, Hedges REM, Hofreiter M, Stiller M, et al. Fossil dogs and wolves from Palaeolithic sites in Belgium, the Ukraine and Russia: Osteometry, ancient DNA and stable isotopes. Journal of Archaeological Science. 2009; 36 :473–490. doi: 10.1016/j.jas.2008.09.033. [ CrossRef ] [ Google Scholar ]
  • Gigerenzer G. From tools to theories: a heuristic of discovery in cognitive psychology. Psychological Review. 1991; 98 :254–267. doi: 10.1037/0033-295X.98.2.254. [ CrossRef ] [ Google Scholar ]
  • Gigerenzer G, Todd P, the ABC Group . Simple heuristics that make us smart. New York: Oxford University Press; 1999. [ Google Scholar ]
  • Gillath, O., Bahns, A. J., Ge, F., & Crandall, C. S. (online 2012). Shoes as a source of first impressions. Journal of Research in Personality .
  • Gillespie A, Zittaun T. Studying the moment of thought. In: Toomela A, Valsiner J, editors. Methodological thinking in psychology: 60 years gone astray? Charlotte: Information age publishing; 2010. pp. 69–88. [ Google Scholar ]
  • Gillies D. Varieties of propensity. The British Journal for the Philosophy of Science. 2000; 51 :807–835. doi: 10.1093/bjps/51.4.807. [ CrossRef ] [ Google Scholar ]
  • Goethe JW. In: Maximen und reflexionen. Schriften der goethe-gesellschaft, 21. Hecker M, editor. Weimar: Verlag der Goethe-Gesellschaft; 1907. [ Google Scholar ]
  • Goldberg LR. Language and individual differences: The search for universals in personality lexicons. In: Wheeler L, editor. Review of personality and social psychology: Vol. 2. Beverly Hills: Sage; 1981. pp. 141–165. [ Google Scholar ]
  • Goldberg LR. From Ace to zombie: Some explorations in the language of personality. In: Spielberger CD, Butcher JN, editors. Advances in personality assessment (Vol. 1) Hillsdale: Erlbaum; 1982. pp. 203–234. [ Google Scholar ]
  • Goldberg LR. An alternative “description of personality”: the big-five factor structure. Journal of Personality and Social Psychology. 1990; 59 :1216–1229. doi: 10.1037/0022-3514.59.6.1216. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Goldberg LR. The structure of phenotypic personality traits. American Psychologist. 1993; 48 :26–34. doi: 10.1037/0003-066X.48.1.26. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Goodall J. The chimpanzees of gombe: Patterns of behavior. Cambridge: The Belknap Press; 1986. [ Google Scholar ]
  • Grandin T, Johnson C. Animals in translation. New York: Scribner; 2005. [ Google Scholar ]
  • Grammer K, Fink B, Neave N. Human pheromones and sexual attraction. European Journal of Obstetrics, Gynecology, and Reproductive Biology. 2005; 111 :135–142. doi: 10.1016/j.ejogrb.2004.08.010. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Gray P. Psychology. 3. New York: Worth; 1999. [ Google Scholar ]
  • Guilford JP. Personality. New York: McGraw-Hill; 1959. [ Google Scholar ]
  • Hartung F-M, Renner B. Social curiosity and interpersonal perception: a judge x trait interaction. Personality and Social Psychology Bulletin. 2011; 37 :796–814. doi: 10.1177/0146167211400618. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Hausheer H. St. Augustine’s conception of time. Philosophical Review. 1937; 46 :503–512. doi: 10.2307/2180833. [ CrossRef ] [ Google Scholar ]
  • Heine SJ, Lehman DR, Peng K, Greenholtz J. What’s Wrong with cross-cultural comparisons of subjective likert scales?: the reference-group effect. Journal of Personality and Social Psychology. 2002; 82 :903–918. doi: 10.1037/0022-3514.82.6.903. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Heisenberg, W. (1989). Encounters with Einstein: And other essays on people, places, and particles . Princeton University Press.
  • Hempel, C. G., & Oppenheim, P. (1948). Studies in the logic of explanation. Philosophy of science, 15 , 135–75; reproduced in Hempel, Carl G. (1965). Aspects of scientific explanation . New York: Free Press.
  • Howard GS. Why do people say nasty things about self-reports? Journal of Organizational Behavior. 1994; 15 :399–404. doi: 10.1002/job.4030150505. [ CrossRef ] [ Google Scholar ]
  • Howarth C, Foster J, Dorrer N. Exploring the potential of the theory of social representations in community-based health research—and vice versa? Journal of Health Psychology. 2004; 9 :229–245. doi: 10.1177/1359105304040889. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Inoue S., & Matsuzawa, T. (2007). Working memory of numerals in chimpanzees. Current Biology, 17 , R1004–R1005. [ PubMed ]
  • Jablonka E, Lamb M. Evolution in four dimensions—Genetic, epigenetic, behavioral, and symbolic variation in the history of life. Cambridge: MIT Press; 2005. [ Google Scholar ]
  • James W. The principles of psychology. New York: H. Holt and Company; 1893. [ Google Scholar ]
  • Jang KL, Livesley WJ, Angleitner A, Riemann R, Vernon PA. Genetic and environmental influences on the covariance of facets defining the domains of the five factor model of personality. Personality and Individual Differences. 2002; 33 :83–101. doi: 10.1016/S0191-8869(01)00137-4. [ CrossRef ] [ Google Scholar ]
  • John OP. The “Big Five” factor taxonomy: Dimensions of personality in the natural language and in questionnaires. In: Pervin LA, editor. Handbook of personality: Theory and research. New York: Guilford; 1990. pp. 66–100. [ Google Scholar ]
  • John OP, Srivastava S. The big five trait taxonomy: History, measurement, and theoretical perspectives. In: Pervin LA, John OP, editors. Handbook of personality: Theory and research. 2. New York: Guilford; 1999. pp. 102–138. [ Google Scholar ]
  • John OP, Angleitner A, Ostendorf F. The lexical approach to personality: a historical. Review of trait taxonomic research. European Journal of Personality. 1988; 2 :171–203. doi: 10.1002/per.2410020302. [ CrossRef ] [ Google Scholar ]
  • Joubert CE. Structured time and subjective acceleration of time. Perceptual and Motor Skills. 1984; 59 :335–336. doi: 10.2466/pms.1984.59.1.335. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Jovchelovitch S. Knowledge in context: Representations, community and culture. London: Routledge; 2007. [ Google Scholar ]
  • Kagan J. Galen’s prophecy. New York: Basic Books; 1994. [ Google Scholar ]
  • Kagan J. Three seductive ideas. Cambridge: Harvard University Press; 1998. [ Google Scholar ]
  • Kagan J, Reznick JS, Snidman N. Biological bases of childhood shyness. Science. 1988; 240 :167–171. doi: 10.1126/science.3353713. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Kaminski J, Call J, Fischer J. Word learning in a domestic dog: evidence for ‘fast mapping’ Science. 2004; 304 :1682–1683. doi: 10.1126/science.1097859. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Keller H. Cultures of infancy. New York: Tailor and Francis; 2007. [ Google Scholar ]
  • Kelly GA. The psychology of personal constructs (Vol. 1 And 2) New York: Norton; 1955. [ Google Scholar ]
  • Komatsu K. Temporal reticence of the self: who can know my self? Integrative Psychological and Behavioral Science. 2012; 46 :357–372. doi: 10.1007/s12124-012-9199-6. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Køppe S. A moderate eclecticism: ontological and epistemological issues. Integrative Psychological and Behavioral Science. 2012; 46 :1–19. doi: 10.1007/s12124-011-9175-6. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Kramer RSS, Ward R. Internal facial features are signals of personality and health. Quarterly Journal of Experimental Psychology. 2010; 63 :2273–2287. doi: 10.1080/17470211003770912. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Kretschmer E. Körperbau und Charakter (1st ed. 1921) 26. Berlin: Springer; 1977. [ Google Scholar ]
  • Kuhn TS. The structure of scientific revolutions. 1. Chicago: University of Chicago Press; 1962. [ Google Scholar ]
  • Kuhn TS. The function of dogma in scientific research. In: Averill IR, editor. Patterns of psychological thought. Washington: Hemisphere Publishing Co.; 1976. pp. 47–66. [ Google Scholar ]
  • Lahlou S. Propagation of social representations. Journal for the Theory of Social Behaviour. 1996; 26 :157–175. doi: 10.1111/j.1468-5914.1996.tb00527.x. [ CrossRef ] [ Google Scholar ]
  • Lahlou S. Functional aspects of social representations. In: Deaux K, Philogene G, editors. Representations of the social: Bridging theoretical traditions. Oxford: Blackwell; 2001. pp. 131–146. [ Google Scholar ]
  • Lahlou, S. (2008). L’Installation du Monde: De la représentation à l’activité en situation . Aix-en-Provence, Université de Provence: Habilitation à Diriger des Recherches en Psychologie, 375.
  • Lahlou S. How can we capture the subject’s perspective?: an evidence-based approach for the social scientist. Social Science Information. 2011; 50 :607–655. doi: 10.1177/0539018411411033. [ CrossRef ] [ Google Scholar ]
  • Lahlou S. Socio-cognitive issues in human-centred design for the real world. In: Boy G, editor. The handbook of human–machine interaction. Aldershot: Ashgate; 2011. pp. 165–188. [ Google Scholar ]
  • Lamiell JT. Beyond individual and group differences: Human individuality, scientific psychology, and William Stern’s critical personalism. Thousand Oaks: Sage Publications; 2003. [ Google Scholar ]
  • Laucken U. Naive verhaltenstheorie. Stuttgart: Klett; 1974. [ Google Scholar ]
  • Laureys S, Pellas F, Van Eeckhout P, Ghorbel S, Schnakers C, Perrin F, et al. The locked-in syndrome: what is it like to be conscious but paralyzed and voiceless? Progress in Brain Research. 2005; 150 :495–511. doi: 10.1016/S0079-6123(05)50034-7. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Lavater JC. Physiognomische fragmente physiognomische fragmente, zur beförderung der menschenkenntniß und menschenliebe. (Bde. I-IV.) Leipzig und Winterthur: Weidmann und Steiner; 1775–1778. [ Google Scholar ]
  • Lazarus AA. The practice of multimodal therapy: Systematic, comprehensive, effective psychotherapy. New York: McGraw-Hill; 1981. [ Google Scholar ]
  • Le Poidevin R. A puzzle concerning time perception. Synthese. 2004; 142 :109–142. doi: 10.1023/B:SYNT.0000047710.71824.b1. [ CrossRef ] [ Google Scholar ]
  • Le Poidevin, R. (2011). The experience and perception of time. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Fall 2011 Edition). http://plato.stanford.edu/archives/fall2011/entries/time-experience/ .
  • Lehner PN. Handbook of ethological methods. Cambridge: Cambridge University Press; 1998. [ Google Scholar ]
  • Lewin K. A dynamic theory of personality. Selected papers. New York: McGraw-Hill; 1935. [ Google Scholar ]
  • Li W, Moallem I, Paller KA, Gottfried JA. Subliminal smells can guide social preferences. Psychological Science. 2007; 18 :1044–1049. doi: 10.1111/j.1467-9280.2007.02023.x. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Locke J. Essay concerning human understanding. Book II. (chapter 32) Oxford: Oxford University Press; 1689/1975. [ Google Scholar ]
  • Loehlin JC, McCrae RR, Costa PT, Jr, John OP. Heritabilities of common and measure-specific components of the big five personality factors. Journal of Research in Personality. 1998; 32 :431–453. doi: 10.1006/jrpe.1998.2225. [ CrossRef ] [ Google Scholar ]
  • Loftus GR. Psychology will be a much better science when we change the way we analyze data. Current Directions in Psychological Science. 1996; 5 :161–171. doi: 10.1111/1467-8721.ep11512376. [ CrossRef ] [ Google Scholar ]
  • Ludlow, P., Nagasawa, Y., & Stoljar, D. (Eds.) (2004). There’s something about Mary. Essays on phenomenal consciousness and Frank Jackson’s knowledge argument . MIT Press/A Bradford Book.
  • Marx K. Das kapital. Buch I. Hamburg: Verlag Otto Meissner; 1867. [ Google Scholar ]
  • Matthews G, Deary IJ, Whiteman MC. Personality traits. 2. Cambridge: Cambridge University Press; 2003. [ Google Scholar ]
  • McAdams DP. Power, intimacy, and the life story: Personological inquiries into identity. New York: Guilford; 1985. [ Google Scholar ]
  • McAdams DP. The five-factor model in personality: a critical appraisal. Journal of Personality. 1992; 60 :329–361. doi: 10.1111/j.1467-6494.1992.tb00976.x. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • McAdams DP. A psychology of the stranger. Psychological Inquiry. 1994; 5 :145–148. doi: 10.1207/s15327965pli0502_12. [ CrossRef ] [ Google Scholar ]
  • McCrae R. The physics and chemistry of personality. Theory and Psychology. 2009; 19 :670–687. doi: 10.1177/0959354309341928. [ CrossRef ] [ Google Scholar ]
  • McCrae RR, Costa PT., Jr Personality trait structure as a human universal. American Psychologist. 1997; 52 :509–516. doi: 10.1037/0003-066X.52.5.509. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • McCrae RR, Costa PT., Jr . The five-factor theory of personality. In: John OP, Robins RW, Pervin LA, editors. Handbook of personality: Theory and research. 3. New York: Guilford; 2008. pp. 159–180. [ Google Scholar ]
  • Meehl P. Factors and taxa, traits and types, differences of degree and differences of kind. Journal of Personality. 1992; 60 :117–174. doi: 10.1111/j.1467-6494.1992.tb00269.x. [ CrossRef ] [ Google Scholar ]
  • Mehl MR, Connor TS, editors. Handbook of research methods for studying daily life. New York: Guilford Press; 2012. [ Google Scholar ]
  • Michell J. Quantitative science and the definition of measurement in psychology. British Journal of Psychology. 1997; 88 :355–383. doi: 10.1111/j.2044-8295.1997.tb02641.x. [ CrossRef ] [ Google Scholar ]
  • Michell J. Measurement in psychology. Cambridge: Cambridge University Press; 1999. [ Google Scholar ]
  • Michell J. The quantitative imperative: positivism, naïve realism and the place in quantitative methods in psychology. Theory and Psychology. 2003; 13 :5–31. doi: 10.1177/0959354303013001758. [ CrossRef ] [ Google Scholar ]
  • Millikan R. White queen psychology and other essays for Alice. Bradford: MIT Press; 1993. [ Google Scholar ]
  • Mischel W. Personality and assessment. New York: Wiley; 1968. [ Google Scholar ]
  • Mischel W, Shoda Y. Personality psychology has two goals: must it be two fields? Psychological Inquiry. 1994; 5 :156–158. doi: 10.1207/s15327965pli0502_15. [ CrossRef ] [ Google Scholar ]
  • Mischel W, Shoda Y. A cognitive-affective system theory of personality: reconceptualizing situations, dispositions, dynamics, and invariance in personality structure. Psychological Review. 1995; 102 :246–268. doi: 10.1037/0033-295X.102.2.246. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Mischel W, Shoda Y, Mendoza-Denton R. Situation-behavior profiles as a locus of consistency in personality. Current Directions in Psychological Science. 2002; 11 :50–54. doi: 10.1111/1467-8721.00166. [ CrossRef ] [ Google Scholar ]
  • Molenaar PC. A manifesto on psychology as idiographic science: bringing the person back into scientific psychology, this time forever. Measurement: Interdisciplinary Research and Perspectives. 2004; 2 :201–218. doi: 10.1207/s15366359mea0204_1. [ CrossRef ] [ Google Scholar ]
  • Molenaar PC. Forum discussion of the manifesto’s aggregational act. Measurement: Interdisciplinary Research and Perspectives. 2004; 2 :248–254. doi: 10.1207/s15366359mea0204_3. [ CrossRef ] [ Google Scholar ]
  • Moscovici S. La psychanalyse, son image et son public. Paris: PUF; 1961. [ Google Scholar ]
  • Moscovici S. The phenomenon of social representations. In: Farr RM, Moscovici S, editors. Social representations. Cambridge: Cambridge University Press; 1984. pp. 3–70. [ Google Scholar ]
  • Murray HA. Explorations in personality. New York: Oxford University Press; 1938. [ Google Scholar ]
  • Neuman Y, Turney PD, Cohen Y. How language enables abstraction: a study in computational cultural psychology. Integrative Psychological and Behavioral Science. 2012; 46 :129–145. doi: 10.1007/s12124-011-9165-8. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Norman T. 2,800 personality trait descriptors: Normative operating characteristics for a university population. Ann Arbor: Department of Psychology, University of Michigan; 1967. [ Google Scholar ]
  • Omi Y. Tension between the theoretical thinking and the empirical method: is it an inevitable fate for psychology? Integrative Psychological and Behavioral Science. 2012; 46 :118–127. doi: 10.1007/s12124-011-9185-4. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Patterson F, Linden E. The education of Koko. New York: Holt, Rinehart and Winston; 1981. [ Google Scholar ]
  • Paunonen SV, Ashton MC. Big five factors and facets and the prediction of behavior. Journal of Personality and Social Psychology. 2001; 81 :524–539. doi: 10.1037/0022-3514.81.3.524. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Pawlik K. Objektive tests in der persönlichkeitsforschung. In: Ortner TM, Proyer RT, Kubinger KD, editors. Theorie und praxis objektiver persönlichkeitstests. Bern: Hans Huber; 2006. pp. S. 16–S. 23. [ Google Scholar ]
  • Peirce CS. Collected papers of Charles sanders Peirce (CP 7.218—1901, On the logic of drawing history from ancient documents especially from testimonies) Cambridge: Harvard University Press; 1901/1935. [ Google Scholar ]
  • Peirce, C. S. (1902/1958). The simplest mathematics (CP 4.227-323). In Collected papers of Charles Sanders Peirce, Vols. 1–6, C. Hartshorne & P. Weiss (eds.), vols. 7–8, A. W. Burks (ed.), Cambridge, MA: Harvard University Press. 1931–1935.
  • Peirce, C. S. (1903/1997). Lecture five: the normative sciences. In C. S. Peirce (Ed.), Pragmatism as a principle and method of right thinking (pp. 205–220). Edited by P. Turrisi. Albany, NY: Suny Press [Harvard Lectures on Pragmatism—CP 5.188-89].
  • Pepperberg IM. In search of King Solomon’s ring: cognitive and communicative studies of grey parrots ( Psittacus erithacus ) Brain, Behavior and Evolution. 2002; 59 :54–67. doi: 10.1159/000063733. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Pervin LA, John OP. Personality: Theory and research. 7. New York: Wiley; 1997. [ Google Scholar ]
  • Piaget J. The child’s perception of time. London: Routledge & Kagan Paul; 1969. [ Google Scholar ]
  • Pillai P. Cultural directions and origins of everyday decisions. Integrative Psychological and Behavioral Science. 2012; 46 :235–242. doi: 10.1007/s12124-012-9196-9. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Popper KR. The propensity interpretation of probability. The British Journal for the Philosophy of Science. 1959; 37 :25–42. doi: 10.1093/bjps/X.37.25. [ CrossRef ] [ Google Scholar ]
  • Popper KR. Logik der forschung. Wien: Springer; 1934. [ Google Scholar ]
  • Rosenbaum PJ, Valsiner J. The un-making of a method: from rating scales to the study of psychological processes. Theory and Psychology. 2011; 21 (1):47–65. doi: 10.1177/0959354309352913. [ CrossRef ] [ Google Scholar ]
  • Rosenthal, V. (2004). Microgenesis, immediate experience and visual processes in reading. In A. Carsetti (Ed.), Seeing, thinking and knowing—Meaning and self-organisation in visual cognition and thought , (pp. 221–243). Kluwer Academic Publishers.
  • Rotter JB. Social learning and clinical psychology. New York: Prentice-Hall; 1954. [ Google Scholar ]
  • Rychlak JF. A philosophy of science for personality theory. Boston: Houghton Mifflin; 1968. [ Google Scholar ]
  • Sameroff A. A unified theory of development: a dialectic integration of nature and nurture. Child Development. 2010; 81 :6–22. doi: 10.1111/j.1467-8624.2009.01378.x. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Sato T, Wakabayashi K, Nameda A, Yasuda Y, Watanabe Y. Understanding a personality as a whole. In: Toomela A, Valsiner J, editors. Methodological thinking in psychology: 60 years gone astray? Charlotte: InfoAge Publications; 2010. pp. 89–119. [ Google Scholar ]
  • Saucier G, Goldberg LR. What is beyond the big five? Journal of Personality. 1998; 66 :495–524. doi: 10.1111/1467-6494.00022. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Savage-Rumbaugh ES, Fields WM. Linguistic, cultural and cognitive capabilities of bonobos ( Pan paniscus ) Culture & Psychology. 2000; 6 :131–153. doi: 10.1177/1354067X0062003. [ CrossRef ] [ Google Scholar ]
  • Schoggen, P. (1989). Behavior settings. A revision and extension of Roger G. Barker’s ecological psychology . Stanford University Press.
  • Schwarz N. Self-reports: how the questions shape the answer. American Psychologist. 1999; 54 :93–105. doi: 10.1037/0003-066X.54.2.93. [ CrossRef ] [ Google Scholar ]
  • Schwarz, M. (2009). Is psychology based on a methodological error? Integrative Psychological and Behavioral Science, 43 , 185–213. [ PubMed ]
  • Sheldon, W. H. & Stevens, S. S. (1942). The varieties of temperament: a psychology of constitutional differences. (1st ed). Harper & Brother.
  • Shevlin M, Walker S, Davies MNO, Banyard P, Lewis CA. Can you judge a book by its cover? Evidence of self-stranger agreement on personality at zero acquaintance. Personality and Individual Differences. 2003; 35 :1373–1383. doi: 10.1016/S0191-8869(02)00356-2. [ CrossRef ] [ Google Scholar ]
  • Shweder RA, D’Andrade RG. The systematic distortion hypothesis. In: Shweder RA, editor. Fallible judgement in behavioral research: New directions for the methodology in social and behavioral science, Vol. 4. San Francisco: Jossey-Bass; 1980. pp. 37–58. [ Google Scholar ]
  • Shweder RA, Sullivan M. The semiotic subject of cultural psychology. In: Pervin L, editor. Handbook of personality: Theory and research. New York: Guilford Press; 1990. pp. 399–416. [ Google Scholar ]
  • Sih A, Bell AM. Insights from behavioral syndromes for behavioral ecology. Advances in the Study of Behavior. 2008; 38 :277–281. doi: 10.1016/S0065-3454(08)00005-3. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Simon, H. A. (1993). Homo rationalis: die Vernunft im menschlichen Leben . Frankfurt am Main: Campus. [engl. Simon, H. A. (1983). Reason in human affairs . Stanford, California: University Press.]
  • Sinn DL, Moltschaniwskyj NA. Personality traits in dumpling squid ( Euprymna tasmanica ): context-specific traits and their correlation with biological characteristics. Journal of Comparative Psychology. 2005; 119 :99–110. doi: 10.1037/0735-7036.119.1.99. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Soldz S, Vaillant GE. The big five personality traits and the life course: a 45-year longitudinal study. Journal of Research in Personality. 1999; 33 :208–232. doi: 10.1006/jrpe.1999.2243. [ CrossRef ] [ Google Scholar ]
  • Spearman C. Correlation calculated from faulty data. British Journal of Psychology. 1910; 3 :271–295. [ Google Scholar ]
  • Sprung L, Sprung H. Grundlagen der methodologie und methodik der psychologie. Berlin: Deutscher Verlag der Wissenschaften; 1984. [ Google Scholar ]
  • Standford CB. The social behavior of chimpanzees and bonobos: empirical evidnce and shifting assumptions. Current Anthropology. 1998; 39 :399–420. doi: 10.1086/204757. [ CrossRef ] [ Google Scholar ]
  • Stern, W. (1935). Allgemeine Psychologie auf personalistischer Grundlage . Haag: Nijhoff. English translation (1938). General Psychology From the Personalistic Standpoint . New York: Macmillan.
  • Suomi SJ. How gene-environment interactions shape the development ofimpulsive aggression in rhesus monkeys. In: Stoff DM, Sussman EJ, editors. Developmental psychobiology of aggression. New York: Cambridge University Press; 2005. pp. 252–268. [ Google Scholar ]
  • Tellegen A. Folk concepts and psychological concepts of personality and personality disorders. Psychological Inquiry. 1993; 4 :122–130. doi: 10.1207/s15327965pli0402_12. [ CrossRef ] [ Google Scholar ]
  • Thelen E, Smith LB. A dynamic systems approach to the development of cognition and action. Cambridge: MIT Press; 1993. [ Google Scholar ]
  • Thurstone LL. The vectors of mind. Psychological Review. 1934; 41 :1–32. doi: 10.1037/h0075959. [ CrossRef ] [ Google Scholar ]
  • Tinbergen, N. (1963). On aims and methods of ethology. Zeitschrift für Tierpsychologie, 20 , 410–433.
  • Tolman E. Cognitive maps in rats and men. Psychological Review. 1948; 55 :189–208. doi: 10.1037/h0061626. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Toomela A. Variables in psychology: a critique of quantitative psychology. Integrative Psychological and Behavioral Science. 2008; 42 :245–265. doi: 10.1007/s12124-008-9059-6. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Toomela A. How methodology became a toolbox—and how it escapes from that box. In: Valsiner J, Molenaar P, Lyra M, Chaudhary N, editors. Dynamic process methodology in the social and developmental sciences. New York: Springer; 2009. pp. 45–66. [ Google Scholar ]
  • Toomela A. In: Methodological thinking in psychology: 60 years gone astray? Toomela A, Valsiner J, editors. Charlotte: Information Age Publishers; 2010a. pp. 1–26. [ Google Scholar ]
  • Toomela A. In: Yearbook of idiographic science (Vol. 2) Salvatore S, Valsiner J, Gennaro A, Simon Travers JB, editors. Rome: Firera & Liuzzo Group; 2010b. [ Google Scholar ]
  • Toomela A. Travel into a fairy land: a critique of modern qualitative and mixed methods psychologies. Integrative Psychological and Behavioral Science. 2011; 45 :21–47. doi: 10.1007/s12124-010-9152-5. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Toomela A, Valsiner J, editors. Methodological thinking in psychology: 60 years gone astray? Charlotte: Information Age Publishers; 2010. [ Google Scholar ]
  • Trendler G. Measurement theory, psychology and the revolution that cannot happen. Theory and Psychology. 2009; 19 :579–599. doi: 10.1177/0959354309341926. [ CrossRef ] [ Google Scholar ]
  • Trut LN. Early canid domestication: the farm-fox experiment. American Scientist. 1999; 87 :160–169. [ Google Scholar ]
  • Uher J. Three methodological core issues of comparative personality research. European Journal of Personality. 2008; 22 :475–496. doi: 10.1002/per.688. [ CrossRef ] [ Google Scholar ]
  • Uher J. Comparative personality research: methodological approaches. European Journal of Personality. 2008; 22 :427–455. doi: 10.1002/per.680. [ CrossRef ] [ Google Scholar ]
  • Uher, J. (2009). Botanical personality? On the origins of consistent individual differences in behavior. In J. Uher Comparative research on personality in Animalia and Plantae . PhD-Dissertation. Berlin: Freie Universität.
  • Uher J. Individual behavioral phenotypes: an integrative meta-theoretical framework. Why ‘behavioral syndromes’ are not analogues of ‘personality’ Developmental Psychobiology. 2011; 53 :521–548. doi: 10.1002/dev.20544. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Uher J. Personality in nonhuman primates: What can we learn from human personality psychology? In: Weiss A, King JE, Murray L, editors. Personality and behavioral syndromes in nonhuman primates. New York: Springer; 2011. [ Google Scholar ]
  • Uher J, Asendorpf JB. Personality assessment in the great apes: comparing ecologically valid behavior measures, behavior ratings, and adjective ratings. Journal of Research in Personality. 2008; 42 :821–838. doi: 10.1016/j.jrp.2007.10.004. [ CrossRef ] [ Google Scholar ]
  • Uher J, Asendorpf JB, Call J. Personality in the behaviour of great apes: temporal stability, cross-situational consistency and coherence in response. Animal Behaviour. 2008; 75 :99–112. doi: 10.1016/j.anbehav.2007.04.018. [ CrossRef ] [ Google Scholar ]
  • Utz K. Philosophie des Zufalls: Ein Entwurf. Paderborn: Schöningh; 2005. [ Google Scholar ]
  • Valsiner J. The guided mind. Cambridge: Harvard University Press; 1998. [ Google Scholar ]
  • Valsiner J. Cultural psychology today: innovations and oversights. Culture & Psychology. 2009; 15 :5–39. doi: 10.1177/1354067X08101427. [ CrossRef ] [ Google Scholar ]
  • Valsiner J. A guided science: History of psychology in the mirror of its making. New Brunswick: Transaction Publishers; 2012. [ Google Scholar ]
  • Van Geert P, van Dijk M. Focus on variability: new tools to study intra-individual variability in developmental data. Infant Behavior & Development. 2002; 25 :340–374. doi: 10.1016/S0163-6383(02)00140-6. [ CrossRef ] [ Google Scholar ]
  • Veltkamp, G. M., Recio, G., Jacobs, A. M., & Conrad, M. (2012). Is personality modulated by language? Evidence from bilinguals’ NEO-FFI scores. International Journal of Bilingualism. Published online before print April 16, 2012, doi:10.1177/1367006912438894
  • Voelklein C, Howarth C. A review of controversies about social representations theory: a British debate. Culture & Psychology. 2005; 11 :431–454. doi: 10.1177/1354067X05058586. [ CrossRef ] [ Google Scholar ]
  • Von Mises R. Probability, statistics and truth. New York: Dover Publications; 1928/1981. [ Google Scholar ]
  • Vygotsky LS. Thought and language. Cambridge: MIT Press; 1962. [ Google Scholar ]
  • Wagner W, Farr R, Jovchelovitch S, Lorenzi-Cioldi F, Marková I, Duveen G, et al. Theory and method of social representations. London: LSE Research Online; 1999. [ Google Scholar ]
  • Wagoner B. The experimental methodology of constructive microgenesis. In: Valsiner J, Molenaar P, Chaudhary N, Lyra M, editors. Handbook of dynamic process methodology in the social and developmental sciences. New York: Springer; 2009. pp. 99–121. [ Google Scholar ]
  • Weber M. In: The methodology of the social sciences. Shils E, Finch H, editors. New York: Free Press; 1949. [ Google Scholar ]
  • Webster’s new international dictionary (1925). (2nd unabridged ed.) Springfield: Merriam.
  • Weinstein TAR, Capitanio JP. Individual differences in infant temperament predict social relationships of yearling rhesus monkeys ( Macaca mulatta ) Animal Behaviour. 2008; 76 :455–465. doi: 10.1016/j.anbehav.2008.01.024. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Weinstein TAR, Capitanio JP. Longitudinal stability of friendships in rhesus monkeys (Macaca mulatta): individual - and relationship - level effects. Journal of Comparative Psychology. 2012; 126 :97–108. doi: 10.1037/a0025607. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Westen D. A model and a method for uncovering the nomothetic from the idiographic: an alternative to the five-factor model. Journal of Research in Personality. 1996; 30 :400–413. doi: 10.1006/jrpe.1996.0028. [ CrossRef ] [ Google Scholar ]
  • Westen D. The scientific status of unconscious processes: is Freud really dead? Journal of the American Psychoanalytic Association. 1999; 47 :1061–1106. doi: 10.1177/000306519904700404. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Whitehead AN. Process and reality. New York: Harper; 1929. [ Google Scholar ]
  • Whorf BL. Language and stereotypes. In: Maccoby EE, Newcomb TM, Hartley EL, editors. Readings in social psychology. 3. New York: Holt; 1958. pp. 1–9. [ Google Scholar ]
  • Wiggins JS. A psychological taxonomy of trait-descriptive terms: the inter-personal domain. Journal of Personality and Social Psychology. 1979; 37 :395–412. doi: 10.1037/0022-3514.37.3.395. [ CrossRef ] [ Google Scholar ]
  • Wiggins JS. Paradigms of personality assessment. New York: Guilford; 2003. [ PubMed ] [ Google Scholar ]
  • Witkowski SR, Brown CH. Lexical universals. Annual Review of Anthropology. 1978; 7 (427):451. [ Google Scholar ]
  • Wright JC, Mischel W. A conditional approach to dispositional constructs: the local predictability of social behavior. Journal of Personality and Social Psychology. 1987; 55 :454–469. doi: 10.1037/0022-3514.55.3.454. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Wright JC, Zakriski AL. When syndromal similarity obscures functional dissimilarity: distinctive evoked environments of externalizing and mixed syndrome children. Journal of Consulting and Clinical Psychology. 2003; 71 :516–527. doi: 10.1037/0022-006X.71.3.516. [ PubMed ] [ CrossRef ] [ Google Scholar ]
  • Wundt WM. Principles of physiological psychology. London: Allen; 1904. [ Google Scholar ]

Library homepage

  • school Campus Bookshelves
  • menu_book Bookshelves
  • perm_media Learning Objects
  • login Login
  • how_to_reg Request Instructor Account
  • hub Instructor Commons
  • Download Page (PDF)
  • Download Full Book (PDF)
  • Periodic Table
  • Physics Constants
  • Scientific Calculator
  • Reference & Cite
  • Tools expand_more
  • Readability

selected template will load here

This action is not available.

Social Sci LibreTexts

3.2: Personality Traits

  • Last updated
  • Save as PDF
  • Page ID 10607

  • https://nobaproject.com/ via The Noba Project

University of Utah, University of Virginia, Michigan State University

Personality traits reflect people’s characteristic patterns of thoughts, feelings, and behaviors. Personality traits imply consistency and stability—someone who scores high on a specific trait like Extraversion is expected to be sociable in different situations and over time. Thus, trait psychology rests on the idea that people differ from one another in terms of where they stand on a set of basic trait dimensions that persist over time and across situations. The most widely used system of traits is called the Five-Factor Model. This system includes five broad traits that can be remembered with the acronym OCEAN: Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. Each of the major traits from the Big Five can be divided into facets to give a more fine-grained analysis of someone's personality. In addition, some trait theorists argue that there are other traits that cannot be completely captured by the Five-Factor Model. Critics of the trait concept argue that people do not act consistently from one situation to the next and that people are very influenced by situational forces. Thus, one major debate in the field concerns the relative power of people’s traits versus the situations in which they find themselves as predictors of their behavior.

learning objectives

  • List and describe the “Big Five” (“OCEAN”) personality traits that comprise the Five-Factor Model of personality.
  • Describe how the facet approach extends broad personality traits.
  • Explain a critique of the personality-trait concept.
  • Describe in what ways personality traits may be manifested in everyday behavior.
  • Describe each of the Big Five personality traits, and the low and high end of the dimension.
  • Give examples of each of the Big Five personality traits, including both a low and high example.
  • Describe how traits and social learning combine to predict your social activities.
  • Describe your theory of how personality traits get refined by social learning.


When we observe people around us, one of the first things that strikes us is how different people are from one another. Some people are very talkative while others are very quiet. Some are active whereas others are couch potatoes. Some worry a lot, others almost never seem anxious. Each time we use one of these words, words like “talkative,” “quiet,” “active,” or “anxious,” to describe those around us, we are talking about a person’s personality — the characteristic ways that people differ from one another. Personality psychologists try to describe and understand these differences.

A person sits on a chair almost completely hidden inside a long sweater.

Although there are many ways to think about the personalities that people have, Gordon Allport and other “personologists” claimed that we can best understand the differences between individuals by understanding their personality traits. Personality traits reflect basic dimensions on which people differ (Matthews, Deary, & Whiteman, 2003). According to trait psychologists, there are a limited number of these dimensions (dimensions like Extraversion, Conscientiousness, or Agreeableness), and each individual falls somewhere on each dimension, meaning that they could be low, medium, or high on any specific trait.

An important feature of personality traits is that they reflect continuous distributions rather than distinct personality types. This means that when personality psychologists talk about Introverts and Extraverts, they are not really talking about two distinct types of people who are completely and qualitatively different from one another. Instead, they are talking about people who score relatively low or relatively high along a continuous distribution. In fact, when personality psychologists measure traits like Extraversion , they typically find that most people score somewhere in the middle, with smaller numbers showing more extreme levels. The figure below shows the distribution of Extraversion scores from a survey of thousands of people. As you can see, most people report being moderately, but not extremely, extraverted, with fewer people reporting very high or very low scores.

This figure shows that most people score towards the middle of the extraversion scale, with fewer people who are highly extraverted or highly introverted.

There are three criteria that are characterize personality traits: (1) consistency, (2) stability, and (3) individual differences.

  • To have a personality trait, individuals must be somewhat consistent across situations in their behaviors related to the trait. For example, if they are talkative at home, they tend also to be talkative at work.
  • Individuals with a trait are also somewhat stable over time in behaviors related to the trait. If they are talkative, for example, at age 30, they will also tend to be talkative at age 40.
  • People differ from one another on behaviors related to the trait. Using speech is not a personality trait and neither is walking on two feet—virtually all individuals do these activities, and there are almost no individual differences. But people differ on how frequently they talk and how active they are, and thus personality traits such as Talkativeness and Activity Level do exist.

A challenge of the trait approach was to discover the major traits on which all people differ. Scientists for many decades generated hundreds of new traits, so that it was soon difficult to keep track and make sense of them. For instance, one psychologist might focus on individual differences in “friendliness,” whereas another might focus on the highly related concept of “sociability.” Scientists began seeking ways to reduce the number of traits in some systematic way and to discover the basic traits that describe most of the differences between people.

The way that Gordon Allport and his colleague Henry Odbert approached this was to search the dictionary for all descriptors of personality (Allport & Odbert, 1936). Their approach was guided by the lexical hypothesis , which states that all important personality characteristics should be reflected in the language that we use to describe other people. Therefore, if we want to understand the fundamental ways in which people differ from one another, we can turn to the words that people use to describe one another. So if we want to know what words people use to describe one another, where should we look? Allport and Odbert looked in the most obvious place—the dictionary. Specifically, they took all the personality descriptors that they could find in the dictionary (they started with almost 18,000 words but quickly reduced that list to a more manageable number) and then used statistical techniques to determine which words “went together.” In other words, if everyone who said that they were “friendly” also said that they were “sociable,” then this might mean that personality psychologists would only need a single trait to capture individual differences in these characteristics. Statistical techniques were used to determine whether a small number of dimensions might underlie all of the thousands of words we use to describe people.

The Five-Factor Model of Personality

Research that used the lexical approach showed that many of the personality descriptors found in the dictionary do indeed overlap. In other words, many of the words that we use to describe people are synonyms. Thus, if we want to know what a person is like, we do not necessarily need to ask how sociable they are, how friendly they are, and how gregarious they are. Instead, because sociable people tend to be friendly and gregarious, we can summarize this personality dimension with a single term. Someone who is sociable, friendly, and gregarious would typically be described as an “Extravert.” Once we know she is an extravert, we can assume that she is sociable, friendly, and gregarious.

Statistical methods (specifically, a technique called factor analysis ) helped to determine whether a small number of dimensions underlie the diversity of words that people like Allport and Odbert identified. The most widely accepted system to emerge from this approach was “The Big Five” or “ Five-Factor Model ” (Goldberg, 1990; McCrae & John, 1992; McCrae & Costa, 1987). The Big Five comprises five major traits shown in the Figure 3.2.2 below. A way to remember these five is with the acronym OCEAN (O is for Openness ; C is for Conscientiousness ; E is for Extraversion ; A is for Agreeableness ; N is for Neuroticism ). Figure 3.2.3 provides descriptions of people who would score high and low on each of these traits.

Openness: The tendency to appreciate new art, ideas, values, feelings, and behaviors. Conscientiousness: The tendency to be careful, on-time for appointments, to follow rules, and to be hardworking. Extraversion: The tendency to be talkative, sociable, and enjoy others; the tendency to have a dominant style. Agreeableness: The tendency to agree and go along with others rather than assert one's own opinions and choices. Neuroticism: The tendency to frequently experience negative emotions such as anger, worry, and sadness, as well as being interpersonally sensitive.

Scores on the Big Five traits are mostly independent. That means that a person’s standing on one trait tells very little about their standing on the other traits of the Big Five. For example, a person can be extremely high in Extraversion and be either high or low on Neuroticism. Similarly, a person can be low in Agreeableness and be either high or low in Conscientiousness. Thus, in the Five-Factor Model, you need five scores to describe most of an individual’s personality.

In the Appendix to this module, we present a short scale to assess the Five-Factor Model of personality (Donnellan, Oswald, Baird, & Lucas, 2006). You can take this test to see where you stand in terms of your Big Five scores. John Johnson has also created a helpful website that has personality scales that can be used and taken by the general public:


After seeing your scores, you can judge for yourself whether you think such tests are valid.

Traits are important and interesting because they describe stable patterns of behavior that persist for long periods of time (Caspi, Roberts, & Shiner, 2005). Importantly, these stable patterns can have broad-ranging consequences for many areas of our life (Roberts, Kuncel, Shiner, Caspi, & Goldberg, 2007). For instance, think about the factors that determine success in college. If you were asked to guess what factors predict good grades in college, you might guess something like intelligence. This guess would be correct, but we know much more about who is likely to do well. Specifically, personality researchers have also found the personality traits like Conscientiousness play an important role in college and beyond, probably because highly conscientious individuals study hard, get their work done on time, and are less distracted by nonessential activities that take time away from school work. In addition, highly conscientious people are often healthier than people low in conscientiousness because they are more likely to maintain healthy diets, to exercise, and to follow basic safety procedures like wearing seat belts or bicycle helmets. Over the long term, this consistent pattern of behaviors can add up to meaningful differences in health and longevity. Thus, personality traits are not just a useful way to describe people you know; they actually help psychologists predict how good a worker someone will be, how long he or she will live, and the types of jobs and activities the person will enjoy. Thus, there is growing interest in personality psychology among psychologists who work in applied settings, such as health psychology or organizational psychology.

Facets of Traits (Subtraits)

So how does it feel to be told that your entire personality can be summarized with scores on just five personality traits? Do you think these five scores capture the complexity of your own and others’ characteristic patterns of thoughts, feelings, and behaviors? Most people would probably say no, pointing to some exception in their behavior that goes against the general pattern that others might see. For instance, you may know people who are warm and friendly and find it easy to talk with strangers at a party yet are terrified if they have to perform in front of others or speak to large groups of people. The fact that there are different ways of being extraverted or conscientious shows that there is value in considering lower-level units of personality that are more specific than the Big Five traits. These more specific, lower-level units of personality are often called facets .

Facets of Openness: Fantasy prone; open to feelings; open to diverse behaviors; open to new and different ideas; open to various values and beliefs. Facets of Conscientiousness: Competent; orderly; dutiful; achievement oriented; self-disciplined; deliberate. Facets of Extraversion: Sociable; warm; assertive; active; excitement-seeking; positive emotionally. Facets of Agreeableness: Trusting; straightforward; altruistic; compliant; modest; tender-minded. Facets of Neuroticism: Anxious; angry; depressed; self-consciousness; impulsive; vulnerable.

To give you a sense of what these narrow units are like, Figure 3.2.4 shows facets for each of the Big Five traits. It is important to note that although personality researchers generally agree about the value of the Big Five traits as a way to summarize one’s personality, there is no widely accepted list of facets that should be studied. The list seen here, based on work by researchers Paul Costa and Jeff McCrae, thus reflects just one possible list among many. It should, however, give you an idea of some of the facets making up each of the Five-Factor Model.

Facets can be useful because they provide more specific descriptions of what a person is like. For instance, if we take our friend who loves parties but hates public speaking, we might say that this person scores high on the “gregariousness” and “warmth” facets of extraversion, while scoring lower on facets such as “assertiveness” or “excitement-seeking.” This precise profile of facet scores not only provides a better description, it might also allow us to better predict how this friend will do in a variety of different jobs (for example, jobs that require public speaking versus jobs that involve one-on-one interactions with customers; Paunonen & Ashton, 2001). Because different facets within a broad, global trait like extraversion tend to go together (those who are gregarious are often but not always assertive), the broad trait often provides a useful summary of what a person is like. But when we really want to know a person, facet scores add to our knowledge in important ways.

Other Traits Beyond the Five-Factor Model

Despite the popularity of the Five-Factor Model, it is certainly not the only model that exists. Some suggest that there are more than five major traits, or perhaps even fewer. For example, in one of the first comprehensive models to be proposed, Hans Eysenck suggested that Extraversion and Neuroticism are most important. Eysenck believed that by combining people’s standing on these two major traits, we could account for many of the differences in personality that we see in people (Eysenck, 1981). So for instance, a neurotic introvert would be shy and nervous, while a stable introvert might avoid social situations and prefer solitary activities, but he may do so with a calm, steady attitude and little anxiety or emotion. Interestingly, Eysenck attempted to link these two major dimensions to underlying differences in people’s biology. For instance, he suggested that introverts experienced too much sensory stimulation and arousal, which made them want to seek out quiet settings and less stimulating environments. More recently, Jeffrey Gray suggested that these two broad traits are related to fundamental reward and avoidance systems in the brain—extraverts might be motivated to seek reward and thus exhibit assertive, reward-seeking behavior, whereas people high in neuroticism might be motivated to avoid punishment and thus may experience anxiety as a result of their heightened awareness of the threats in the world around them (Gray, 1981. This model has since been updated; see Gray & McNaughton, 2000). These early theories have led to a burgeoning interest in identifying the physiological underpinnings of the individual differences that we observe.

Another revision of the Big Five is the HEXACO model of traits (Ashton & Lee, 2007). This model is similar to the Big Five, but it posits slightly different versions of some of the traits, and its proponents argue that one important class of individual differences was omitted from the Five-Factor Model. The HEXACO adds Honesty-Humility as a sixth dimension of personality. People high in this trait are sincere, fair, and modest, whereas those low in the trait are manipulative, narcissistic, and self-centered. Thus, trait theorists are agreed that personality traits are important in understanding behavior, but there are still debates on the exact number and composition of the traits that are most important.

There are other important traits that are not included in comprehensive models like the Big Five. Although the five factors capture much that is important about personality, researchers have suggested other traits that capture interesting aspects of our behavior. In Figure 5 below we present just a few, out of hundreds, of the other traits that have been studied by personologists.

This table lists personality traits other than those that are part of the Big 5. These include Machiavellianism, Need for Achievement, Need for Cognition, Authoritarianism, Narcissism, Self-Esteem, Optimism, and Alexithymia.

Not all of the above traits are currently popular with scientists, yet each of them has experienced popularity in the past. Although the Five-Factor Model has been the target of more rigorous research than some of the traits above, these additional personality characteristics give a good idea of the wide range of behaviors and attitudes that traits can cover.

The Person-Situation Debate and Alternatives to the Trait Perspective

College students in a classroom.

The ideas described in this module should probably seem familiar, if not obvious to you. When asked to think about what our friends, enemies, family members, and colleagues are like, some of the first things that come to mind are their personality characteristics. We might think about how warm and helpful our first teacher was, how irresponsible and careless our brother is, or how demanding and insulting our first boss was. Each of these descriptors reflects a personality trait, and most of us generally think that the descriptions that we use for individuals accurately reflect their “characteristic pattern of thoughts, feelings, and behaviors,” or in other words, their personality.

But what if this idea were wrong? What if our belief in personality traits were an illusion and people are not consistent from one situation to the next? This was a possibility that shook the foundation of personality psychology in the late 1960s when Walter Mischel published a book called Personality and Assessment (1968). In this book, Mischel suggested that if one looks closely at people’s behavior across many different situations, the consistency is really not that impressive. In other words, children who cheat on tests at school may steadfastly follow all rules when playing games and may never tell a lie to their parents. In other words, he suggested, there may not be any general trait of honesty that links these seemingly related behaviors. Furthermore, Mischel suggested that observers may believe that broad personality traits like honesty exist, when in fact, this belief is an illusion. The debate that followed the publication of Mischel’s book was called the person-situation debate because it pitted the power of personality against the power of situational factors as determinants of the behavior that people exhibit.

Because of the findings that Mischel emphasized, many psychologists focused on an alternative to the trait perspective. Instead of studying broad, context-free descriptions, like the trait terms we’ve described so far, Mischel thought that psychologists should focus on people’s distinctive reactions to specific situations. For instance, although there may not be a broad and general trait of honesty, some children may be especially likely to cheat on a test when the risk of being caught is low and the rewards for cheating are high. Others might be motivated by the sense of risk involved in cheating and may do so even when the rewards are not very high. Thus, the behavior itself results from the child’s unique evaluation of the risks and rewards present at that moment, along with her evaluation of her abilities and values. Because of this, the same child might act very differently in different situations. Thus, Mischel thought that specific behaviors were driven by the interaction between very specific, psychologically meaningful features of the situation in which people found themselves, the person’s unique way of perceiving that situation, and his or her abilities for dealing with it. Mischel and others argued that it was these social-cognitive processes that underlie people’s reactions to specific situations that provide some consistency when situational features are the same. If so, then studying these broad traits might be more fruitful than cataloging and measuring narrow, context-free traits like Extraversion or Neuroticism.

In the years after the publication of Mischel’s (1968) book, debates raged about whether personality truly exists, and if so, how it should be studied. And, as is often the case, it turns out that a more moderate middle ground than what the situationists proposed could be reached. It is certainly true, as Mischel pointed out, that a person’s behavior in one specific situation is not a good guide to how that person will behave in a very different specific situation. Someone who is extremely talkative at one specific party may sometimes be reticent to speak up during class and may even act like a wallflower at a different party. But this does not mean that personality does not exist, nor does it mean that people’s behavior is completely determined by situational factors. Indeed, research conducted after the person-situation debate shows that on average, the effect of the “situation” is about as large as that of personality traits. However, it is also true that if psychologists assess a broad range of behaviors across many different situations, there are general tendencies that emerge. Personality traits give an indication about how people will act on average, but frequently they are not so good at predicting how a person will act in a specific situation at a certain moment in time. Thus, to best capture broad traits, one must assess aggregate behaviors, averaged over time and across many different types of situations. Most modern personality researchers agree that there is a place for broad personality traits and for the narrower units such as those studied by Walter Mischel.

The Mini-IPIP Scale

(Donnellan, Oswald, Baird, & Lucas, 2006)

Instructions: Below are phrases describing people’s behaviors. Please use the rating scale below to describe how accurately each statement describes you. Describe yourself as you generally are now, not as you wish to be in the future. Describe yourself as you honestly see yourself, in relation to other people you know of the same sex as you are, and roughly your same age. Please read each statement carefully, and put a number from 1 to 5 next to it to describe how accurately the statement describes you.

1 = Very inaccurate

2 = Moderately inaccurate

3 = Neither inaccurate nor accurate

4 = Moderately accurate

5 = Very accurate

  • _______ Am the life of the party (E)
  • _______ Sympathize with others’ feelings (A)
  • _______ Get chores done right away (C)
  • _______ Have frequent mood swings (N)
  • _______ Have a vivid imagination (O)
  • _______Don’t talk a lot (E)
  • _______ Am not interested in other people’s problems (A)
  • _______ Often forget to put things back in their proper place (C)
  • _______ Am relaxed most of the time (N)
  • ______ Am not interested in abstract ideas (O)
  • ______ Talk to a lot of different people at parties (E)
  • ______ Feel others’ emotions (A)
  • ______ Like order (C)
  • ______ Get upset easily (N)
  • ______ Have difficulty understanding abstract ideas (O)
  • ______ Keep in the background (E)
  • ______ Am not really interested in others (A)
  • ______ Make a mess of things (C)
  • ______ Seldom feel blue (N)
  • ______ Do not have a good imagination (O)

Scoring: The first thing you must do is to reverse the items that are worded in the opposite direction. In order to do this, subtract the number you put for that item from 6. So if you put a 4, for instance, it will become a 2. Cross out the score you put when you took the scale, and put the new number in representing your score subtracted from the number 6.

Items to be reversed in this way: 6, 7, 8, 9, 10, 15, 16, 17, 18, 19, 20

Next, you need to add up the scores for each of the five OCEAN scales (including the reversed numbers where relevant). Each OCEAN score will be the sum of four items. Place the sum next to each scale below.

__________ Openness: Add items 5, 10, 15, 20

__________ Conscientiousness: Add items 3, 8, 13, 18

__________ Extraversion: Add items 1, 6, 11, 16

__________ Agreeableness: Add items 2, 7, 12, 17

__________ Neuroticism: Add items 4, 9,14, 19

Compare your scores to the norms below to see where you stand on each scale. If you are low on a trait, it means you are the opposite of the trait label. For example, low on Extraversion is Introversion, low on Openness is Conventional, and low on Agreeableness is Assertive.

19–20 Extremely High, 17–18 Very High, 14–16 High,

11–13 Neither high nor low; in the middle, 8–10 Low, 6–7 Very low, 4–5 Extremely low

Outside Resources

Discussion Questions

  • Consider different combinations of the Big Five, such as O (Low), C (High), E (Low), A (High), and N (Low). What would this person be like? Do you know anyone who is like this? Can you select politicians, movie stars, and other famous people and rate them on the Big Five?
  • How do you think learning and inherited personality traits get combined in adult personality?
  • Can you think of instances where people do not act consistently—where their personality traits are not good predictors of their behavior?
  • Has your personality changed over time, and in what ways?
  • Can you think of a personality trait not mentioned in this module that describes how people differ from one another?
  • When do extremes in personality traits become harmful, and when are they unusual but productive of good outcomes?
  • Allport, G. W., & Odbert, H. S. (1936). Trait names: A psycholexical study. Psychological Monographs, 47 , 211.
  • Ashton, M. C., & Lee, K. (2007). Empirical, theoretical, and practical advantages of the HEXACO model of personality structure. Personality and Social Psychological Review, 11 , 150–166.
  • Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality development: Stability and change. Annual Reviews of Psychology, 56 , 453–484.
  • Donnellan, M. B., Oswald, F. L., Baird, B. M., & Lucas, R. E. (2006). The mini-IPIP scales: Tiny-yet-effective measures of the Big Five factors of personality. Psychological Assessment, 18 , 192–203.
  • Eysenck, H. J. (1981). A model for personality .New York: Springer Verlag.
  • Goldberg, L. R. (1990). An alternative description of personality: The Big Five personality traits. Journal of Personality and Social Psychology, 59 , 1216–1229.
  • Gray, J. A. (1981). A critique of Eysenck’s theory of personality. In H. J. Eysenck (Ed.), A Model for Personality (pp. 246-276). New York: Springer Verlag.
  • Gray, J. A. & McNaughton, N. (2000). The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system (second edition) .Oxford: Oxford University Press.
  • Matthews, G., Deary, I. J., & Whiteman, M. C. (2003). Personality traits . Cambridge, UK: Cambridge University Press.
  • McCrae, R. R., & Costa, P. T. (1987). Validation of the five-factor model of personality across instruments and observers. Journal of Personality and Social Psychology, 52 , 81–90.
  • McCrae, R. R. & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60 , 175–215.
  • Mischel, W. (1968). Personality and assessment . New York: John Wiley.
  • Paunonen, S. V., & Ashton, M. S. (2001). Big five factors and facets and the prediction of behavior. Journal of Personality and Social Psychology, 81 , 524–539.
  • Roberts, B. W., Kuncel, N. R., Shiner, R., Caspi, A., & Golberg, L. R. (2007). The power of personality: The comparative validity of personality traits, socioeconomic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science, 2 , 313-345.


The Lexical Hypothesis


The Lexical Hypothesis, proposed by Gordon W. Allport and Henry S. Odbert in 1936, suggests that the most salient and important aspects of human personality are represented by the words found in natural language dictionaries.


According to the Lexical Hypothesis, individuals tend to describe and differentiate personality traits using words and phrases found in their linguistic systems. It assumes that if a characteristic or trait is significant enough to be recognized by a culture, it is likely to have a corresponding descriptor in the language of that culture.

Key Aspects:

The Lexical Hypothesis is built on the following key aspects:

  • Synonym Frequency: Personality traits that are important to individuals are more likely to be addressed by multiple synonyms or descriptors in a given language. For example, the trait of “honesty” may also be captured by words like “truthfulness,” “integrity,” or “sincerity.”
  • Importance and Visibility: Traits that hold greater significance and are more observable or noticeable tend to be represented by a larger number of words in a language. For instance, personality traits like “extraversion” or “conscientiousness” are typically described by numerous terms in dictionaries.
  • Cross-Cultural Consistency: The Lexical Hypothesis assumes that universally recognized personality traits exist across cultures and languages. While the specific words used to describe these traits may differ, the underlying concepts are expected to be consistent.


The Lexical Hypothesis has contributed significantly to personality psychology by guiding research on trait structure and taxonomy. It has aided in developing widely used trait models such as the Big Five personality traits (neuroticism, extraversion, openness, agreeableness, and conscientiousness). Researchers have extensively explored and validated lexical approaches to measure personality traits and identify their underlying factors.

The hypothesis has also informed various areas of applied psychology, including personnel selection, job fit assessments, clinical assessments, and relationship compatibility evaluations.


The Lexical Hypothesis emphasizes the importance of language in understanding and describing human personality. It posits that the traits individuals consider significant and observable are reflected in the words and phrases they use. By leveraging natural language, researchers have gained valuable insights into trait structure, allowing for a better understanding of human personality and its implications in various areas of life.

Psychology Dictionary


the theory that important natural characteristics and traits unique to individuals have become intrinsically embedded in our natural- language lexicon over time.

Avatar photo

Leave a Reply

Your email address will not be published. Required fields are marked *

Latest Posts

what is lexical hypothesis

Meeting the Milestones: A Guide to Piaget's Child Developmental Stages

what is lexical hypothesis

Counseling, Therapy, and Psychology: What Is The Difference?

what is lexical hypothesis

The Psychology of Metaphysical Belief Systems

what is lexical hypothesis

4 Key Considerations When Supporting a Loved One Through a Legal Battle for Justice 

How Exercise Can Boost Your Mental Health as You Age

Finding Balance: The Psychological Benefits of Staying Active

what is lexical hypothesis

The Psychology of Winning: Case Studies and Analysis from the World of Sports

what is lexical hypothesis

Transitioning to Digital Therapy: Navigating the Pros and Cons

what is lexical hypothesis

From Loss to Liberation: The Psychological Journey Of Seniors Receiving All-On-4 Dental Implants

what is lexical hypothesis

How to Create Family History Interview Questions?

what is lexical hypothesis

The Most Common Addiction in the United States

Road to recovery: tools and resources for mental health treatment success.

what is lexical hypothesis

Do Cat Allergy Shots for Humans Work?

Popular psychology terms, medical model, hypermnesia, affirmation, brainwashing, backup reinforcer, message-learning approach, affiliative behavior, behavioral modeling, approach motivation.

  • You are here
  • Everything Explained.Today
  • A-Z Contents
  • Lexical hypothesis

Lexical hypothesis explained

In personality psychology, the lexical hypothesis [1] (also known as the fundamental lexical hypothesis , [2] lexical approach , [3] or sedimentation hypothesis ) generally includes two postulates :

1. Those personality characteristics that are important to a group of people will eventually become a part of that group's language. [4]

and that therefore:

2. More important personality characteristics are more likely to be encoded into language as a single word. [5] [6] [7]

With origins during the late 19th century, use of the lexical hypothesis began to flourish in English and German psychology during the early 20th century. [8] The lexical hypothesis is a major basis of the study of the Big Five personality traits , [9] the HEXACO model of personality structure [10] and the 16PF Questionnaire and has been used to study the structure of personality traits in a number of cultural and linguistic settings. [11]

Early estimates

Sir Francis Galton was one of the first scientists to apply the lexical hypothesis to the study of personality, stating:

Despite Galton's early ventures into the lexical study of personality, more than two decades passed before English-language scholars continued his work. A 1910 study by George E. Partridge listed approximately 750 English adjectives used to describe mental states, [12] while a 1926 study of Webster's New International Dictionary by M. L. Perkins provided an estimate of 3,000 such terms. [13] These early explorations and estimates were not limited to the English-speaking world, with philosopher and psychologist Ludwig Klages stating in 1929 that the German language contains approximately 4,000 words to describe inner states. [14]

Psycholexical studies

Allport & odbert.

Nearly half a century after Galton first investigated the lexical hypothesis, Franziska Baumgarten published the first psycholexical classification of personality-descriptive terms. Using dictionaries and characterology publications, Baumgarten identified 1,093 separate terms in the German language used for the description of personality and mental states. [15] Although this number is similar in size to the German and English estimates offered by earlier researchers, Gordon Allport and Henry S. Odbert revealed this to be a severe underestimate in a 1936 study. Similar to the earlier work of M. L. Perkins, they used Webster's New International Dictionary as their source. From this list of approximately 400,000 words, Allport and Odbert identified 17,953 unique terms used to describe personality or behavior.

This is one of the most influential psycholexical studies in the history of trait psychology . Not only was it the longest, most exhaustive list of personality-descriptive words at the time, it was also one of the earliest attempts at classifying English-language terms with the use of psychological principles. Using their list of nearly 18,000 terms, Allport and Odbert separated these into four categories or "columns":

Column I : This group contains 4,504 terms that describe or are related to personality traits. Being the most important of the four columns to Allport and Odbert and future psychologists, its terms most closely relate to those used by modern personality psychologists (e.g., aggressive, introverted, sociable). Allport and Odbert suggested that this column represented a minimum rather than final list of trait terms. Because of this, they recommended that other researchers consult the remaining three columns in their studies.

Column II : In contrast with the more stable disposition s described by terms in Column I, this group includes terms describing present states, attitudes , emotion s, and moods (e.g., rejoicing, frantic). As a result of this emphasis of temporary states, present participles represent the majority of the 4,541 terms in Column II.

Column III : The largest of the four groups, Column III contains 5,226 words related to social evaluations of an individual person's character (e.g., worthy, insignificant). Unlike the previous two columns, this group does not refer to internal psychological attributes of a person. As such, Allport and Odbert acknowledged that Column III did not meet their definition of trait-related terms. Predating the person-situation debate by more than 30 years, [16] Allport and Odbert included this group to appease researchers of social psychology , sociology , and ethics.

Column IV : The last of Allport and Odbert's four columns contained 3,682 words. Termed the "miscellaneous column" by the authors, Column IV contains important personality-descriptive terms that did not seem appropriate for the other three columns. Allport and Odbert offered potential subgroups for terms describing behaviors (e.g., pampered, crazed), physical qualities associated with psychological traits (e.g., lean, roly-poly), and talents or abilities (e.g., gifted, prolific). However, they noted that these subdivisions were not necessarily accurate, as: (i) innumerable subgroups were possible, (ii) these subgroups would not incorporate all of the miscellaneous terms, and (iii) further editing might reveal that these terms could be used for the other three columns.

Allport and Odbert did not present these four columns as representing orthogonal concepts. Many of their nearly 18,000 terms could have been classified differently or put into multiple categories, particularly those in Columns I and II. Although the authors attempted to remedy this with the aid of three other editors, the average degree of agreement between these independent reviewers was approximately 47%. Noting that each outside reviewer seemed to have a preferred column, the authors decided to present the classifications performed by Odbert. Rather than try to rationalize this decision, Allport and Odbert presented the results of their study as somewhat arbitrary and unfinished.

Warren Norman

Throughout the 1940s, researchers such as Raymond Cattell [4] and Donald Fiske [17] used factor analysis to explore the more general structure of the trait terms in Allport and Odbert's Column I. Rather than rely on the factors obtained by these researchers, Warren Norman performed an independent analysis of Allport and Odbert's terms in 1963. [18] Despite finding a five-factor structure similar to Fiske's, Norman decided to use Allport and Odbert's original list to create a more precise and better-structured taxonomy of terms. [19] Using the 1961 edition of Webster's International Dictionary, Norman added relevant terms and removed those from Allport and Odbert's list that were no longer in use. This resulted in a source list of approximately 40,000 potential trait-descriptive terms. Using this list, Norman then removed terms that were deemed archaic or obsolete, solely evaluative, overly obscure, dialect-specific, loosely related to personality, and purely physical. By doing so, Norman reduced his original list to 2,797 unique trait-descriptive terms. Norman's work would eventually serve as the basis for Dean Peabody and Lewis Goldberg's explorations of the "Big Five" personality traits. [20] [21] [22]

Juri Apresjan and the Moscow Semantic School

During the 1970s, Juri Apresjan, a founder of the Moscow Semantic School, developed the systemic, or systematic, method of lexicography which utilizes the concept of the language picture of the world . This concept is also termed the naive picture of the world in order to stress the non-scientific description of the world which is found in natural language. [23] In his book "Systematic Lexicography", which was published in English in 2000, J.D.Apresjan puts forward the idea of building dictionaries on the basis of "reconstructing the so-called naive picture of the world, or the "world-view", underlying the partly universal and partly language specific pattern of conceptualizations inherent in any natural language". [24] In his opinion, the general world-view can be fragmented into different more local pictures of reality, such as naive geometry, naive physics, naive psychology, and so forth. In particular, one chapter of the book Apresjan allots to the description of lexicographic reconstruction of the language picture of the human being in the Russian language. [25] Later, Apresjan's work was the basis for Sergey Golubkov's further attempts to build "the language personality theory" [26] [27] [28] which would be different from other lexically-based personality theories (e.g. by Allport, Cattell, Eysenck, etc.) due to its meronomic (partonomic) nature versus the taxonomic nature of the previously mentioned personality theories. [29]

Psycholexical studies of values

In addition to research on personality, the psycholexical method has also been applied to the study of values in multiple languages, [30] [31] providing a contrast with theory-driven approaches such as Schwartz's Theory of Basic Human Values . [32] [33]

Similar concepts

Concepts similar to the lexical hypothesis are basic to ordinary language philosophy . [34] Similar to the use of the lexical hypothesis to understand personality, ordinary language philosophers propose that philosophical problems can be solved or better understood by an examination of everyday language. In his essay "A Plea for Excuses," J. L. Austin cited three main justifications for this method: words are tools, words are not only facts or objects, and commonly used words "embod[y] all the distinctions men have found worth drawing...we are using a sharpened awareness of words to sharpen our perception of, though not as the final arbiter of, the phenomena". [35]

Despite its widespread use for the study of personality, the lexical hypothesis has been challenged for a number of reasons. The following list describes some of the major critiques of the lexical hypothesis and personality models based on psycholexical studies. [7] [36]

  • The use of verbal descriptors as material for analysis brings a pro-social bias of language into the resulting models. [37] [38] [39] Experiments using the lexical hypothesis indeed demonstrated that the use of lexical material skews the resulting dimensionality according to a sociability bias of language and a negativity bias of emotionality, grouping all evaluations around these two dimensions. [37] This means that the two largest dimensions in the Big Five model of personality (i.e., Extraversion and Neuroticism) might be just an artifact of the lexical method that this model employed.
  • Many traits of psychological importance are too complex to be encoded into single terms or used in everyday language. [40] In fact, an entire text may be the only way to accurately capture and reflect some important personality characteristics. [41]
  • Laypeople use personality-descriptive terms in an ambiguous manner. [42] Similarly, many of the terms used in psycholexical studies are too ambiguous to be useful in a psychological context. [43]
  • The lexical hypothesis relies on terms that were not developed by experts. As such, any models developed with the lexical hypothesis represent lay perceptions rather than expert psychological knowledge.
  • Language accounts for a minority of communication and is inadequate to describe much of human experience. [44]
  • The mechanisms that resulted in the development of personality lexicons are poorly understood.
  • Personality-descriptive terms change over time and differ in meaning across dialects, languages, and cultures.
  • The methods used to test the lexical hypothesis are unscientific. [45]
  • Personality-descriptive language is too general to be represented by a single word class , [46] yet psycholexical studies of personality largely rely on adjective s. [47]
  • 16 Personality Factors
  • Linguistic relativity

External links

  • Language Personality Theory

Notes and References

  • Book: Personality Theory . Oxford University Press . Crowne, D. P. . 2007 . Don Mills, ON, Canada . 978-0-19-542218-4.
  • An alternative "description of personality": The Big-Five factor structure . Goldberg, L. R. . Journal of Personality and Social Psychology . December 1990 . 59 . 6 . 1216–1229 . 2283588 . 10.1037/0022-3514.59.6.1216. 9034636 .
  • Book: The Psychology of Personality: Second Edition . Wiley-Blackwell . Carducci, B. J. . 2009 . Malden, MA . 978-1-4051-3635-8.
  • Cattell . R.B. . The description of personality: basic traits resolved into clusters . Journal of Abnormal and Social Psychology . 1943 . 38 . 4 . 476–506 . 10.1037/h0054116.
  • The lexical approach to personality: A historical review of trait taxonomic research . John, O. P. . Angleitner, A. . Ostendorf, F. . European Journal of Personality . 1988 . 2 . 3 . 171–203 . 10.1002/per.2410020302 . 15299845 . https://web.archive.org/web/20141026125224/https://pub.uni-bielefeld.de/luur/download?func=downloadFile&recordOId=1779427&fileOId=2312707 . 26 October 2014.
  • Book: Miller . George A . The science of words . registration . 1996 . Scientific American Library . New York . 978-0-7167-5027-7.
  • A defence of the lexical approach to the study of personality structure . Ashton, M. C. . Lee, K. . European Journal of Personality . 2004 . 19 . 5–24 . 10.1002/per.541 . 145576560 . https://web.archive.org/web/20120317220226/http://www.psy.uwa.edu.au/davidm/203/2007/lexical%20studies%20Ashton%20and%20Lee%202005b.pdf . 17 March 2012.
  • Book: Personality: Determinants, Dynamics, and Potentials . Cambridge University Press . Caprara, G. V.. Cervone, D. . 2000 . New York . 978-0-521-58310-7.
  • Goldberg. Lewis. The structure of phenotypic personality traits.. American Psychologist. 1993. 48. 1. 26–34. 10.1037/0003-066x.48.1.26. 8427480. 20595956.
  • Ashton. Michael C.. Lee. Kibeom. Perugini. Marco. Szarota. Piotr. de Vries. Reinout E.. Di Blas. Lisa. Boies. Kathleen. De Raad. Boele. A Six-Factor Structure of Personality-Descriptive Adjectives: Solutions From Psycholexical Studies in Seven Languages.. Journal of Personality and Social Psychology. 86. 2. 2004. 356–366. 0022-3514. 10.1037/0022-3514.86.2.356. 14769090.
  • Book: Handbook of Personality: Theory and Research, Third Edition . The Guilford Press . John, O. P.. Robins, R. W.. Pervin, L. A. . Paradigm Shift to the Integrative Big-Five Trait Taxonomy: History, Measurement, and Conceptual Issues . 2008 . New York . 114–158 . 978-1-59385-836-0.
  • Book: An Outline of Individual Study . Sturgis & Walton . Partridge, G. E. . 1910 . New York . 106–111.
  • The teaching of ideals and the development of the traits of character and personality . Perkins, M. L. . Proceedings of the Oklahoma Academy of Science . 1926 . 6 . 2 . 344–347 . 27 March 2012 . https://web.archive.org/web/20160303235502/http://digital.library.okstate.edu/OAS/oas_pdf/v06/p344_347.pdf . 3 March 2016 .
  • Book: The Science of Character . George Allen & Unwin . Klages, L. . 1929 . London.
  • Book: Trait-names: A psycho-lexical study. . Psychological Review Company . Allport, G. W.. Odbert, H. S. . 1936 . Albany, NY.
  • The person-situation debate in historical and current perspective . Epstein, S.. O'Brien, E. J. . Psychological Bulletin . November 1985 . 98 . 3 . 513–537 . 10.1037/0033-2909.98.3.513 . 4080897.
  • Consistency of the factorial structures of personality ratings from different sources . Fiske, D. W. . Journal of Abnormal and Social Psychology . July 1949 . 44 . 3 . 329–344 . 10.1037/h0057198. 18146776 .
  • Toward an adequate taxonomy of personality attributes: Replicated factor structure in peer nomination personality ratings . Norman, W. T. . Journal of Abnormal and Social Psychology . June 1963 . 66 . 6 . 574–583 . 10.1037/h0040291 . 13938947.
  • Book: 2800 personality trait descriptors: Normative operating characteristics for a university population . University of Michigan, Dept. of Psychology . Norman, W. T. . 1967 . Ann Arbor, MI .
  • Book: Problems with Language Imprecision: New Directions for Methodology of Social and Behavioral Science . Jossey-Bass . Fiske, D. W. . Developing a taxonomy of trait-descriptive terms. . 1981 . San Francisco, CA . 43–65.
  • Some determinants of factor structures from personality-trait descriptors . Peabody, D.. Goldberg, L. R. . Journal of Personality and Social Psychology . September 1989 . 57 . 3 . 552–567 . 10.1037/0022-3514.57.3.552 . 2778639. 19459419.
  • An alternative "description of personality": The Big-Five factor structure . Goldberg, L. R. . Journal of Personality and Social Psychology . December 1990 . 59 . 6 . 1216–1229 . 10.1037/0022-3514.59.6.1216 . 2283588. 9034636 .
  • Book: Apresi͡an, I͡Uriĭ Derenikovich. Systematic Lexicography . 2000. Oxford University Press. 978-0-19-823780-8.
  • Web site: Apresjan J. (1992). Systemic Lexicography. In Euralex-92 Proceedings (part 1). p.4 . 2014-01-14.
  • Book: Apresi͡an, I͡Uriĭ Derenikovich. Systematic Lexicography. https://books.google.com/books?id=WWPf0c024E8C. 2000. Oxford University Press. 978-0-19-823780-8. 101–143. The Picture of Man as Reconstructed from Linguistic Data: An Attempt at a Systematic Description.
  • Web site: Golubkov's Language Personality Theory . webspace.ship.edu.
  • Golubkov S.V. . 2002 . The language personality theory: An integrative approach to personality on the basis of its language phenomenology . Social Behavior and Personality . 30 . 6. 571–578 . 10.2224/sbp.2002.30.6.571 .
  • Web site: PsychNews 5(1) . userpage.fu-berlin.de.
  • Golubkov S.V. . 2002 . The language personality theory: An integrative approach to personalityon the basis of its language phenomenology . Social Behavior and Personality . 30 . 6. 573 . 10.2224/sbp.2002.30.6.571 .
  • De Raad . Boele . Van Oudenhoven . Jan Pieter . A psycholexical study of virtues in the Dutch language, and relations between virtues and personality . European Journal of Personality . January 2011 . 25 . 1 . 43–52 . 10.1002/per.777. 227275239 .
  • De Raad . Boele . Morales-Vives . Fabia . Barelds . Dick P. H. . Van Oudenhoven . Jan Pieter . Renner . Walter . Timmerman . Marieke E. . Values in a Cross-Cultural Triangle . Journal of Cross-Cultural Psychology . 28 July 2016 . 47 . 8 . 1053–1075 . 10.1177/0022022116659698. 151792686 .
  • Schwartz . Shalom H. . Theory-Driven Versus Lexical Approaches to Value Structures . Journal of Cross-Cultural Psychology . 15 March 2017 . 48 . 3 . 439–443 . 10.1177/0022022117690452. 151514308 .
  • De Raad . Boele . Timmerman . Marieke E. . Morales-Vives . Fabia . Renner . Walter . Barelds . Dick P. H. . Pieter Van Oudenhoven . Jan . The Psycho-Lexical Approach in Exploring the Field of Values . Journal of Cross-Cultural Psychology . 15 March 2017 . 48 . 3 . 444–451 . 10.1177/0022022117692677. 28502995 . 5414899 . free .
  • Five big, Big Five issues: Rationale, content, structure, status, and crosscultural assessment . De Raad, B. . European Psychologist . June 1998 . 3 . 2 . 113–124 . 10.1027/1016-9040.3.2.113.
  • Book: Philosophical Papers . Oxford University Press . Austin, J. L. . A Plea for Excuses . 1970 . London . 175–204.
  • Book: Dumont, F.. A History of Personality Psychology: Theory, Science, and Research from Hellenism to the Twenty-first Century. Cambridge University Press. Trait theories and the psychology of individual differences.. 2010. New York. 149–182 . 978-0-521-11632-9.
  • Trofimova . I. N. . 2014 . Observer bias: an interaction of temperament traits with biases in the semantic perception of lexical material. . PLOS ONE . 9 . 1 . e85677 . 2014PLoSO...985677T . 10.1371/journal.pone.0085677 . 3903487 . 24475048 . free.
  • Trofimova. I.. Robbins. TW. Sulis. W.. Uher. J.. 2018. Taxonomies of psychological individual differences: biological perspectives on millennia-long challenges . Philosophical Transactions of the Royal Society, Biological Sciences. 373. 1744. 10.1098/rstb.2017.0152. 29483338 . 5832678 . free.
  • Trofimova, I . etal . 2022. What's next for the neurobiology of temperament, personality and psychopathology? . Current Opinions in Behavioral Sciences. 45. 101143. 10.1016/j.cobeha.2022.101143. 248817462.
  • A contrarian view of the five-factor approach to personality description. . Block, J. . Psychological Bulletin . March 1995 . 117 . 2 . 187–215 . 10.1037/0033-2909.117.2.187 . 7724687.
  • Openness to experience: Expanding the boundaries of Factor V . McCrae, R. R. . European Journal of Personality . November 1994 . 8 . 4. 251–272. 10.1002/per.2410080404. 144576220 .
  • A model and a method for uncovering the nomothetic from the idiographic: An alternative to the Five-Factor Model . Westen, D. . Journal of Research in Personality . September 1996. 30. 3. 400–413 . 10.1006/jrpe.1996.0028 . https://web.archive.org/web/20120507121610/http://www.psychsystems.net/Publications/1996/2.%20uncovering%20the%20nomothetic%20from%20the%20idiographic_Westen_jrnl%20of%20research%20in%20pers%201996.pdf . 7 May 2012. .
  • Book: Personality Description in Ordinary Language . Wiley. Bromley, D. B. . 1977 . London . 978-0-471-99443-5 . registration .
  • Book: Silent Messages . Wadsworth . Mehrabian, A. . 1971 . Belmont, CA . 978-0-534-00910-6 . registration .
  • The Big Five versus nobody? . Shadel, W. G.. Cervone, D. . American Psychologist . December 1993 . 48. 12. 1300–1302 . 10.1037/0003-066x.48.12.1300.
  • Personality-descriptive verbs . De Raad, B.. Mulder, E.. Kloosterman, K.. Hofstee, W. K. B. . European Journal of Personality . June 1988 . 2 . 2 . 81–96 . 10.1002/per.2410020204. 146758458.
  • Eysenck . Hans Jurgen . The structure of phenoytypic personality traits: Comment . The American Psychologist . 1993 . 48 . 12 . 1299–1300 . 10.1037/0003-066x.48.12.1299.b.

This article is licensed under the GNU Free Documentation License . It uses material from the Wikipedia article " Lexical hypothesis ".

Except where otherwise indicated, Everything.Explained.Today is © Copyright 2009-2024, A B Cryer, All Rights Reserved. Cookie policy .

  • Search Menu
  • Browse content in Arts and Humanities
  • Browse content in Archaeology
  • Anglo-Saxon and Medieval Archaeology
  • Archaeological Methodology and Techniques
  • Archaeology by Region
  • Archaeology of Religion
  • Archaeology of Trade and Exchange
  • Biblical Archaeology
  • Contemporary and Public Archaeology
  • Environmental Archaeology
  • Historical Archaeology
  • History and Theory of Archaeology
  • Industrial Archaeology
  • Landscape Archaeology
  • Mortuary Archaeology
  • Prehistoric Archaeology
  • Underwater Archaeology
  • Urban Archaeology
  • Zooarchaeology
  • Browse content in Architecture
  • Architectural Structure and Design
  • History of Architecture
  • Residential and Domestic Buildings
  • Theory of Architecture
  • Browse content in Art
  • Art Subjects and Themes
  • History of Art
  • Industrial and Commercial Art
  • Theory of Art
  • Biographical Studies
  • Byzantine Studies
  • Browse content in Classical Studies
  • Classical Literature
  • Classical Reception
  • Classical History
  • Classical Philosophy
  • Classical Mythology
  • Classical Art and Architecture
  • Classical Oratory and Rhetoric
  • Greek and Roman Papyrology
  • Greek and Roman Archaeology
  • Greek and Roman Epigraphy
  • Greek and Roman Law
  • Late Antiquity
  • Religion in the Ancient World
  • Digital Humanities
  • Browse content in History
  • Colonialism and Imperialism
  • Diplomatic History
  • Environmental History
  • Genealogy, Heraldry, Names, and Honours
  • Genocide and Ethnic Cleansing
  • Historical Geography
  • History by Period
  • History of Emotions
  • History of Agriculture
  • History of Education
  • History of Gender and Sexuality
  • Industrial History
  • Intellectual History
  • International History
  • Labour History
  • Legal and Constitutional History
  • Local and Family History
  • Maritime History
  • Military History
  • National Liberation and Post-Colonialism
  • Oral History
  • Political History
  • Public History
  • Regional and National History
  • Revolutions and Rebellions
  • Slavery and Abolition of Slavery
  • Social and Cultural History
  • Theory, Methods, and Historiography
  • Urban History
  • World History
  • Browse content in Language Teaching and Learning
  • Language Learning (Specific Skills)
  • Language Teaching Theory and Methods
  • Browse content in Linguistics
  • Applied Linguistics
  • Cognitive Linguistics
  • Computational Linguistics
  • Forensic Linguistics
  • Grammar, Syntax and Morphology
  • Historical and Diachronic Linguistics
  • History of English
  • Language Evolution
  • Language Reference
  • Language Variation
  • Language Families
  • Language Acquisition
  • Lexicography
  • Linguistic Anthropology
  • Linguistic Theories
  • Linguistic Typology
  • Phonetics and Phonology
  • Psycholinguistics
  • Sociolinguistics
  • Translation and Interpretation
  • Writing Systems
  • Browse content in Literature
  • Bibliography
  • Children's Literature Studies
  • Literary Studies (Romanticism)
  • Literary Studies (American)
  • Literary Studies (Modernism)
  • Literary Studies (Asian)
  • Literary Studies (European)
  • Literary Studies (Eco-criticism)
  • Literary Studies - World
  • Literary Studies (1500 to 1800)
  • Literary Studies (19th Century)
  • Literary Studies (20th Century onwards)
  • Literary Studies (African American Literature)
  • Literary Studies (British and Irish)
  • Literary Studies (Early and Medieval)
  • Literary Studies (Fiction, Novelists, and Prose Writers)
  • Literary Studies (Gender Studies)
  • Literary Studies (Graphic Novels)
  • Literary Studies (History of the Book)
  • Literary Studies (Plays and Playwrights)
  • Literary Studies (Poetry and Poets)
  • Literary Studies (Postcolonial Literature)
  • Literary Studies (Queer Studies)
  • Literary Studies (Science Fiction)
  • Literary Studies (Travel Literature)
  • Literary Studies (War Literature)
  • Literary Studies (Women's Writing)
  • Literary Theory and Cultural Studies
  • Mythology and Folklore
  • Shakespeare Studies and Criticism
  • Browse content in Media Studies
  • Browse content in Music
  • Applied Music
  • Dance and Music
  • Ethics in Music
  • Ethnomusicology
  • Gender and Sexuality in Music
  • Medicine and Music
  • Music Cultures
  • Music and Media
  • Music and Culture
  • Music and Religion
  • Music Education and Pedagogy
  • Music Theory and Analysis
  • Musical Scores, Lyrics, and Libretti
  • Musical Structures, Styles, and Techniques
  • Musicology and Music History
  • Performance Practice and Studies
  • Race and Ethnicity in Music
  • Sound Studies
  • Browse content in Performing Arts
  • Browse content in Philosophy
  • Aesthetics and Philosophy of Art
  • Epistemology
  • Feminist Philosophy
  • History of Western Philosophy
  • Metaphysics
  • Moral Philosophy
  • Non-Western Philosophy
  • Philosophy of Language
  • Philosophy of Mind
  • Philosophy of Perception
  • Philosophy of Action
  • Philosophy of Law
  • Philosophy of Religion
  • Philosophy of Science
  • Philosophy of Mathematics and Logic
  • Practical Ethics
  • Social and Political Philosophy
  • Browse content in Religion
  • Biblical Studies
  • Christianity
  • East Asian Religions
  • History of Religion
  • Judaism and Jewish Studies
  • Qumran Studies
  • Religion and Education
  • Religion and Health
  • Religion and Politics
  • Religion and Science
  • Religion and Law
  • Religion and Art, Literature, and Music
  • Religious Studies
  • Browse content in Society and Culture
  • Cookery, Food, and Drink
  • Cultural Studies
  • Customs and Traditions
  • Ethical Issues and Debates
  • Hobbies, Games, Arts and Crafts
  • Lifestyle, Home, and Garden
  • Natural world, Country Life, and Pets
  • Popular Beliefs and Controversial Knowledge
  • Sports and Outdoor Recreation
  • Technology and Society
  • Travel and Holiday
  • Visual Culture
  • Browse content in Law
  • Arbitration
  • Browse content in Company and Commercial Law
  • Commercial Law
  • Company Law
  • Browse content in Comparative Law
  • Systems of Law
  • Competition Law
  • Browse content in Constitutional and Administrative Law
  • Government Powers
  • Judicial Review
  • Local Government Law
  • Military and Defence Law
  • Parliamentary and Legislative Practice
  • Construction Law
  • Contract Law
  • Browse content in Criminal Law
  • Criminal Procedure
  • Criminal Evidence Law
  • Sentencing and Punishment
  • Employment and Labour Law
  • Environment and Energy Law
  • Browse content in Financial Law
  • Banking Law
  • Insolvency Law
  • History of Law
  • Human Rights and Immigration
  • Intellectual Property Law
  • Browse content in International Law
  • Private International Law and Conflict of Laws
  • Public International Law
  • IT and Communications Law
  • Jurisprudence and Philosophy of Law
  • Law and Society
  • Law and Politics
  • Browse content in Legal System and Practice
  • Courts and Procedure
  • Legal Skills and Practice
  • Primary Sources of Law
  • Regulation of Legal Profession
  • Medical and Healthcare Law
  • Browse content in Policing
  • Criminal Investigation and Detection
  • Police and Security Services
  • Police Procedure and Law
  • Police Regional Planning
  • Browse content in Property Law
  • Personal Property Law
  • Study and Revision
  • Terrorism and National Security Law
  • Browse content in Trusts Law
  • Wills and Probate or Succession
  • Browse content in Medicine and Health
  • Browse content in Allied Health Professions
  • Arts Therapies
  • Clinical Science
  • Dietetics and Nutrition
  • Occupational Therapy
  • Operating Department Practice
  • Physiotherapy
  • Radiography
  • Speech and Language Therapy
  • Browse content in Anaesthetics
  • General Anaesthesia
  • Neuroanaesthesia
  • Clinical Neuroscience
  • Browse content in Clinical Medicine
  • Acute Medicine
  • Cardiovascular Medicine
  • Clinical Genetics
  • Clinical Pharmacology and Therapeutics
  • Dermatology
  • Endocrinology and Diabetes
  • Gastroenterology
  • Genito-urinary Medicine
  • Geriatric Medicine
  • Infectious Diseases
  • Medical Toxicology
  • Medical Oncology
  • Pain Medicine
  • Palliative Medicine
  • Rehabilitation Medicine
  • Respiratory Medicine and Pulmonology
  • Rheumatology
  • Sleep Medicine
  • Sports and Exercise Medicine
  • Community Medical Services
  • Critical Care
  • Emergency Medicine
  • Forensic Medicine
  • Haematology
  • History of Medicine
  • Browse content in Medical Skills
  • Clinical Skills
  • Communication Skills
  • Nursing Skills
  • Surgical Skills
  • Medical Ethics
  • Browse content in Medical Dentistry
  • Oral and Maxillofacial Surgery
  • Paediatric Dentistry
  • Restorative Dentistry and Orthodontics
  • Surgical Dentistry
  • Medical Statistics and Methodology
  • Browse content in Neurology
  • Clinical Neurophysiology
  • Neuropathology
  • Nursing Studies
  • Browse content in Obstetrics and Gynaecology
  • Gynaecology
  • Occupational Medicine
  • Ophthalmology
  • Otolaryngology (ENT)
  • Browse content in Paediatrics
  • Neonatology
  • Browse content in Pathology
  • Chemical Pathology
  • Clinical Cytogenetics and Molecular Genetics
  • Histopathology
  • Medical Microbiology and Virology
  • Patient Education and Information
  • Browse content in Pharmacology
  • Psychopharmacology
  • Browse content in Popular Health
  • Caring for Others
  • Complementary and Alternative Medicine
  • Self-help and Personal Development
  • Browse content in Preclinical Medicine
  • Cell Biology
  • Molecular Biology and Genetics
  • Reproduction, Growth and Development
  • Primary Care
  • Professional Development in Medicine
  • Browse content in Psychiatry
  • Addiction Medicine
  • Child and Adolescent Psychiatry
  • Forensic Psychiatry
  • Learning Disabilities
  • Old Age Psychiatry
  • Psychotherapy
  • Browse content in Public Health and Epidemiology
  • Epidemiology
  • Public Health
  • Browse content in Radiology
  • Clinical Radiology
  • Interventional Radiology
  • Nuclear Medicine
  • Radiation Oncology
  • Reproductive Medicine
  • Browse content in Surgery
  • Cardiothoracic Surgery
  • Gastro-intestinal and Colorectal Surgery
  • General Surgery
  • Neurosurgery
  • Paediatric Surgery
  • Peri-operative Care
  • Plastic and Reconstructive Surgery
  • Surgical Oncology
  • Transplant Surgery
  • Trauma and Orthopaedic Surgery
  • Vascular Surgery
  • Browse content in Science and Mathematics
  • Browse content in Biological Sciences
  • Aquatic Biology
  • Biochemistry
  • Bioinformatics and Computational Biology
  • Developmental Biology
  • Ecology and Conservation
  • Evolutionary Biology
  • Genetics and Genomics
  • Microbiology
  • Molecular and Cell Biology
  • Natural History
  • Plant Sciences and Forestry
  • Research Methods in Life Sciences
  • Structural Biology
  • Systems Biology
  • Zoology and Animal Sciences
  • Browse content in Chemistry
  • Analytical Chemistry
  • Computational Chemistry
  • Crystallography
  • Environmental Chemistry
  • Industrial Chemistry
  • Inorganic Chemistry
  • Materials Chemistry
  • Medicinal Chemistry
  • Mineralogy and Gems
  • Organic Chemistry
  • Physical Chemistry
  • Polymer Chemistry
  • Study and Communication Skills in Chemistry
  • Theoretical Chemistry
  • Browse content in Computer Science
  • Artificial Intelligence
  • Computer Architecture and Logic Design
  • Game Studies
  • Human-Computer Interaction
  • Mathematical Theory of Computation
  • Programming Languages
  • Software Engineering
  • Systems Analysis and Design
  • Virtual Reality
  • Browse content in Computing
  • Business Applications
  • Computer Games
  • Computer Security
  • Computer Networking and Communications
  • Digital Lifestyle
  • Graphical and Digital Media Applications
  • Operating Systems
  • Browse content in Earth Sciences and Geography
  • Atmospheric Sciences
  • Environmental Geography
  • Geology and the Lithosphere
  • Maps and Map-making
  • Meteorology and Climatology
  • Oceanography and Hydrology
  • Palaeontology
  • Physical Geography and Topography
  • Regional Geography
  • Soil Science
  • Urban Geography
  • Browse content in Engineering and Technology
  • Agriculture and Farming
  • Biological Engineering
  • Civil Engineering, Surveying, and Building
  • Electronics and Communications Engineering
  • Energy Technology
  • Engineering (General)
  • Environmental Science, Engineering, and Technology
  • History of Engineering and Technology
  • Mechanical Engineering and Materials
  • Technology of Industrial Chemistry
  • Transport Technology and Trades
  • Browse content in Environmental Science
  • Applied Ecology (Environmental Science)
  • Conservation of the Environment (Environmental Science)
  • Environmental Sustainability
  • Environmentalist Thought and Ideology (Environmental Science)
  • Management of Land and Natural Resources (Environmental Science)
  • Natural Disasters (Environmental Science)
  • Nuclear Issues (Environmental Science)
  • Pollution and Threats to the Environment (Environmental Science)
  • Social Impact of Environmental Issues (Environmental Science)
  • History of Science and Technology
  • Browse content in Materials Science
  • Ceramics and Glasses
  • Composite Materials
  • Metals, Alloying, and Corrosion
  • Nanotechnology
  • Browse content in Mathematics
  • Applied Mathematics
  • Biomathematics and Statistics
  • History of Mathematics
  • Mathematical Education
  • Mathematical Finance
  • Mathematical Analysis
  • Numerical and Computational Mathematics
  • Probability and Statistics
  • Pure Mathematics
  • Browse content in Neuroscience
  • Cognition and Behavioural Neuroscience
  • Development of the Nervous System
  • Disorders of the Nervous System
  • History of Neuroscience
  • Invertebrate Neurobiology
  • Molecular and Cellular Systems
  • Neuroendocrinology and Autonomic Nervous System
  • Neuroscientific Techniques
  • Sensory and Motor Systems
  • Browse content in Physics
  • Astronomy and Astrophysics
  • Atomic, Molecular, and Optical Physics
  • Biological and Medical Physics
  • Classical Mechanics
  • Computational Physics
  • Condensed Matter Physics
  • Electromagnetism, Optics, and Acoustics
  • History of Physics
  • Mathematical and Statistical Physics
  • Measurement Science
  • Nuclear Physics
  • Particles and Fields
  • Plasma Physics
  • Quantum Physics
  • Relativity and Gravitation
  • Semiconductor and Mesoscopic Physics
  • Browse content in Psychology
  • Affective Sciences
  • Clinical Psychology
  • Cognitive Psychology
  • Cognitive Neuroscience
  • Criminal and Forensic Psychology
  • Developmental Psychology
  • Educational Psychology
  • Evolutionary Psychology
  • Health Psychology
  • History and Systems in Psychology
  • Music Psychology
  • Neuropsychology
  • Organizational Psychology
  • Psychological Assessment and Testing
  • Psychology of Human-Technology Interaction
  • Psychology Professional Development and Training
  • Research Methods in Psychology
  • Social Psychology
  • Browse content in Social Sciences
  • Browse content in Anthropology
  • Anthropology of Religion
  • Human Evolution
  • Medical Anthropology
  • Physical Anthropology
  • Regional Anthropology
  • Social and Cultural Anthropology
  • Theory and Practice of Anthropology
  • Browse content in Business and Management
  • Business Ethics
  • Business History
  • Business Strategy
  • Business and Technology
  • Business and Government
  • Business and the Environment
  • Comparative Management
  • Corporate Governance
  • Corporate Social Responsibility
  • Entrepreneurship
  • Health Management
  • Human Resource Management
  • Industrial and Employment Relations
  • Industry Studies
  • Information and Communication Technologies
  • International Business
  • Knowledge Management
  • Management and Management Techniques
  • Operations Management
  • Organizational Theory and Behaviour
  • Pensions and Pension Management
  • Public and Nonprofit Management
  • Strategic Management
  • Supply Chain Management
  • Browse content in Criminology and Criminal Justice
  • Criminal Justice
  • Criminology
  • Forms of Crime
  • International and Comparative Criminology
  • Youth Violence and Juvenile Justice
  • Development Studies
  • Browse content in Economics
  • Agricultural, Environmental, and Natural Resource Economics
  • Asian Economics
  • Behavioural Finance
  • Behavioural Economics and Neuroeconomics
  • Econometrics and Mathematical Economics
  • Economic History
  • Economic Methodology
  • Economic Systems
  • Economic Development and Growth
  • Financial Markets
  • Financial Institutions and Services
  • General Economics and Teaching
  • Health, Education, and Welfare
  • History of Economic Thought
  • International Economics
  • Labour and Demographic Economics
  • Law and Economics
  • Macroeconomics and Monetary Economics
  • Microeconomics
  • Public Economics
  • Urban, Rural, and Regional Economics
  • Welfare Economics
  • Browse content in Education
  • Adult Education and Continuous Learning
  • Care and Counselling of Students
  • Early Childhood and Elementary Education
  • Educational Equipment and Technology
  • Educational Strategies and Policy
  • Higher and Further Education
  • Organization and Management of Education
  • Philosophy and Theory of Education
  • Schools Studies
  • Secondary Education
  • Teaching of a Specific Subject
  • Teaching of Specific Groups and Special Educational Needs
  • Teaching Skills and Techniques
  • Browse content in Environment
  • Applied Ecology (Social Science)
  • Climate Change
  • Conservation of the Environment (Social Science)
  • Environmentalist Thought and Ideology (Social Science)
  • Natural Disasters (Environment)
  • Social Impact of Environmental Issues (Social Science)
  • Browse content in Human Geography
  • Cultural Geography
  • Economic Geography
  • Political Geography
  • Browse content in Interdisciplinary Studies
  • Communication Studies
  • Museums, Libraries, and Information Sciences
  • Browse content in Politics
  • African Politics
  • Asian Politics
  • Chinese Politics
  • Comparative Politics
  • Conflict Politics
  • Elections and Electoral Studies
  • Environmental Politics
  • European Union
  • Foreign Policy
  • Gender and Politics
  • Human Rights and Politics
  • Indian Politics
  • International Relations
  • International Organization (Politics)
  • International Political Economy
  • Irish Politics
  • Latin American Politics
  • Middle Eastern Politics
  • Political Behaviour
  • Political Economy
  • Political Institutions
  • Political Theory
  • Political Methodology
  • Political Communication
  • Political Philosophy
  • Political Sociology
  • Politics and Law
  • Public Policy
  • Public Administration
  • Quantitative Political Methodology
  • Regional Political Studies
  • Russian Politics
  • Security Studies
  • State and Local Government
  • UK Politics
  • US Politics
  • Browse content in Regional and Area Studies
  • African Studies
  • Asian Studies
  • East Asian Studies
  • Japanese Studies
  • Latin American Studies
  • Middle Eastern Studies
  • Native American Studies
  • Scottish Studies
  • Browse content in Research and Information
  • Research Methods
  • Browse content in Social Work
  • Addictions and Substance Misuse
  • Adoption and Fostering
  • Care of the Elderly
  • Child and Adolescent Social Work
  • Couple and Family Social Work
  • Developmental and Physical Disabilities Social Work
  • Direct Practice and Clinical Social Work
  • Emergency Services
  • Human Behaviour and the Social Environment
  • International and Global Issues in Social Work
  • Mental and Behavioural Health
  • Social Justice and Human Rights
  • Social Policy and Advocacy
  • Social Work and Crime and Justice
  • Social Work Macro Practice
  • Social Work Practice Settings
  • Social Work Research and Evidence-based Practice
  • Welfare and Benefit Systems
  • Browse content in Sociology
  • Childhood Studies
  • Community Development
  • Comparative and Historical Sociology
  • Economic Sociology
  • Gender and Sexuality
  • Gerontology and Ageing
  • Health, Illness, and Medicine
  • Marriage and the Family
  • Migration Studies
  • Occupations, Professions, and Work
  • Organizations
  • Population and Demography
  • Race and Ethnicity
  • Social Theory
  • Social Movements and Social Change
  • Social Research and Statistics
  • Social Stratification, Inequality, and Mobility
  • Sociology of Religion
  • Sociology of Education
  • Sport and Leisure
  • Urban and Rural Studies
  • Browse content in Warfare and Defence
  • Defence Strategy, Planning, and Research
  • Land Forces and Warfare
  • Military Administration
  • Military Life and Institutions
  • Naval Forces and Warfare
  • Other Warfare and Defence Issues
  • Peace Studies and Conflict Resolution
  • Weapons and Equipment

The Oxford Handbook of Reading

  • < Previous chapter
  • Next chapter >

The Oxford Handbook of Reading

10 Individual Differences Among Skilled Readers: The Role of Lexical Quality

Sally Andrews, School of Psychology, University of Sydney

  • Published: 05 December 2014
  • Cite Icon Cite
  • Permissions Icon Permissions

Theories of visual word recognition and reading have been based on averaged data from relatively small samples of skilled readers, reflecting an implicit assumption that all skilled readers read in the same way. This chapter reviews evidence of systematic individual differences in the early stages of lexical retrieval among samples of above-average readers that challenges this assumption. Individual differences in patterns of masked priming and parafoveal processing during sentence reading provide evidence of variability in the precision and coherence of lexical knowledge that are consistent with Perfetti’s (2007) construct of lexical quality. This evidence is compatible with neuroimaging evidence that literacy drives the development of specialized neural systems for processing written words. Understanding these dimensions of individual differences among expert readers has important implications for future refinements of theories of visual word recognition and reading.

Even a casual glance at the chapters in this volume reveals that we know a lot about how the average skilled reader reads words. Hundreds of papers reporting average data for samples of 20 to 30 university students in both single word and text reading tasks have confirmed a range of robust phenomena that generalize across samples, tasks, and languages. These findings provide the foundation for a rich landscape of theories of lexical retrieval and reading. Many of these models have been computationally implemented and systematically assessed against empirical benchmarks. Despite this degree of specificity and systematic evaluation, the validity of the different theoretical accounts remains a focus of intense debate. The central argument of this chapter is that investigations of individual differences among skilled readers can—and should—play an important role in future refinement of these competing models.

Most theories of skilled word recognition and reading have been validated against average data for samples of 20 to 30 university students. I am not questioning the value of this experimental research strategy, which underpins much of the empirical evidence and theoretical frameworks discussed in other chapters on skilled reading. Too early a focus on individual differences may have impeded extraction of these general principles of skilled reading behavior ( Andrews, 2012 ). However, continuing to rely on averaged data to test the validity of different theories may lead the field astray.

Many computational models of visual word recognition can accurately simulate the average effects of major variables such as word frequency, length, and regularity on people’s performance. But more fine-grained evaluation, such as comparisons of performance for different items, reveals that the models are very poor at predicting more detailed aspects of human performance. For example, the highly influential dual-route cascaded (DRC) ( Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001 ) and parallel distributed processing (PDP) ( Plaut, McClelland, Seidenberg, & Patterson, 1996 ) models account for less than 10% of variance in the average human naming latencies from a number of large-scale databases ( Balota & Spieler, 1998 ). Computational modelers’ solution to this problem has been to develop more refined models, often by combining the features of earlier models, to achieve a better fit with the human data (e.g., Perry, Ziegler, & Zorzi, 2007 ). However, the problem may not lie in the models but in the validity of the data used to evaluate them.

Relying on averaged data to evaluate theories reflects an implicit uniformity assumption ( Andrews, 2012 ): that all skilled readers have developed the same cognitive architecture and read in the same way. The evidence reviewed in later sections shows that averaged data can obscure systematic individual differences and potentially lead to misleading conclusions about underlying cognitive processes. Rather than modifying models to better fit average patterns of performance, it is time for experimental psycholinguists to consider whether and how individual differences modulate skilled lexical retrieval.

The more specific goal of this chapter is to review and evaluate the contribution of lexical quality to individual differences among skilled readers. This construct was first introduced by Perfetti (1985) in the context of his verbal efficiency theory and subsequently defined as the “extent to which a mental representation of a word specifies its form and meaning components in a way that is both precise and flexible” ( Perfetti, 2007 , p. 359). High-quality representations are said to be orthographically fully specified, phonologically redundant, and semantically more generalized and less context-bound. The orthographic, phonological, and semantic constituents of a word are also assumed to be more tightly bound. These strong connections lead to more reliable, synchronous activation of the various constituent codes that define a word’s identity, which in turn supports flexible use of this information to achieve the reader’s goals ( Perfetti, 2007 , p. 360).

What critically distinguishes the lexical quality hypothesis (LQH) from many other accounts of individual differences in reading is that it assigns a causal role to lexical knowledge . Rather than attributing reading difficulties to dysfunctions in phonological, semantic, or working memory mechanisms , “the LQH is about knowledge that has not been acquired or practiced to a high-enough level” ( Perfetti, 2007 , p. 380). Ineffective or inefficient processes are a consequence of low quality knowledge representations.

The utility of the construct of lexical quality depends on how clearly it is defined. Perfetti’s (2007) definition emphasizes the precision and flexibility of lexical knowledge. Precision is identified with the content of lexical knowledge: the specificity and completeness of the orthographic, phonological, grammatical, and semantic constituents of the lexical representation. As elaborated in what follows, orthographic precision is particularly critical because written forms are the gateway to lexical knowledge ( Dehaene, 2009 ). Flexibility arises from the interconnectedness or binding between the different constituents of lexical representations. Strong connections allow printed word forms to trigger synchronous, coherent activation of all components of the word’s identity required for comprehension ( Perfetti, 2007 ).

According to Perfetti’s (2007) definition, lexical quality is both word-specific and graded. The quality of lexical representations varies within individuals, as they gradually increase the size of their vocabulary and refine the specificity of the knowledge stored for existing words through reading experience. There are also differences between individuals in the extent to which they have established high-quality representations for most of the words in their written vocabulary. Such differences could arise from genetic differences in the capacity for orthographic learning ( Byrne et al., 2008 ) or environmental factors, such as methods of instruction, as well as from differences in reading strategy. Readers who rely heavily on context to identify words may devote little attention to the details of words’ internal structure and therefore be less likely to develop fully specified orthographic codes for all words ( Frith, 1986 ).

Most investigations of lexical quality have focused on children, reflecting the common assumption that, although individual differences in word recognition are a major source of variance during early reading development, they make little contribution to variability in successful reading comprehension at later grades (e.g., Eason & Cutting, 2009 ). However, the university students who provide the averaged data in published studies vary considerably in their level and profile of performance in tests of written language proficiency ( Andrews, 2012 ). Recognition of this variability, combined with difficulties in distinguishing between models, has increased interest in individual differences in word-level processing among skilled readers (e.g.,   Ashby, Rayner, & Clifton, 2005 ; Kuperman & Van Dyke, 2011 ; Yap, Balota, Sibley, & Ratcliff, 2012 ; Yap, Tse, & Balota, 2009 ). However, these studies have used a range of different measures to assess proficiency that vary in how effectively they capture the knowledge that Perfetti (2007) identified with lexical quality.

The chapter is structured around three broad questions that are central to evaluating the contribution of individual differences in lexical quality to skilled reading.

Why is lexical quality important for skilled reading?

What is lexical quality?

How does lexical quality influence skilled reading?

Why Is Lexical Quality Important for Skilled Reading?

Reading is a culturally engineered skill ( Carr, 1999 ) that was acquired too recently in the evolutionary history of our species to be part of our genetic blueprint ( Dehaene, 2009 ). The first writing systems were invented about 5,000 years ago, and the alphabet is less than 4,000 years old. Humans are therefore not evolutionarily prepared to learn to read as they appear to be for comprehending and producing spoken language ( Pinker, 1994 ). Reading is also acquired much later in a child’s development than spoken language. Gough and Hillinger (1980) argued that this is because learning to read is “an unnatural act” that requires children to crack the code that connects written and spoken language so that they can harness their well-established spoken-language knowledge and skills for the new task of reading. Achieving these feats requires children to gain new metalinguistic insights, such as phonological awareness, that were not necessary for spoken-language acquisition ( Ehri , this volume). These cognitive developments can be fostered by a supportive, literate environment. Learning to read is therefore subject to systematic influences of both the instructional and home environment ( Sénéchal , this volume).

However, at least in cultures with universal access to education, genetic factors are the strongest predictor of early reading acquisition ( Byrne et al., 2009 ). Behavioral genetic analyses of continuities and discontinuities in shared genetic and environmental variance, within a large-scale cross-national longitudinal study of twin pairs from before preschool until grade two, have demonstrated substantial shared genetic variance between decoding, spelling, and orthographic learning that is independent of the genetic variance associated with intelligence and significantly stronger than the variance due to environmental factors ( Byrne et al., 2008 ). Phonological processing measures, such as phonological awareness and rapid naming speed, showed significant heritability at both preschool and kindergarten time points, demonstrating a genetic basis for the well-established contribution of phonological processes to early reading acquisition. However, analyses tracking changes to grade two showed that genetic effects of phonological awareness were shared with measures of orthographic knowledge, such as word decoding and spelling ( Byrne et al., 2009 ). Byrne et al. (2008) argued that this common genetic variance is most parsimoniously attributed to a learning rate factor that influences the efficiency of learning associations between orthography and phonology.

Consistent with this evidence of genetic contributions to the development of orthographic knowledge, recent neuroimaging evidence from skilled readers has shown that the same highly localized brain regions respond selectively to written text across a wide range of languages. Meta-analyses of studies of readers of languages that vary in both script type and reading direction have identified a reproducible location of activation in the lateral occipito temporal sulcus. Dehaene (2009) argues that the uniformity of this visual word form area (VWFA) across individuals and languages presents a reading paradox: despite its evolutionary recency, reading involves “fixed cortical mechanisms that are exquisitely attuned to the recognition of written words . . . as though there was a cerebral organ for reading” (p. 4). Understanding how humans have developed a uniform solution to such a recent cultural invention is important not only for understanding reading itself but also for understanding how the brain adapts to a new cognitive skill.

The resolution to the reading paradox provided by Dehaene’s (2009)   neuronal recycling hypothesis is that the specialized visual word form system develops by partially recycling “a cortical territory evolved for object and face recognition” ( Dehaene & Cohen, 2011 , p. 254). This region is specialized for extracting features such as line orientation, junctions, and contours that are important for identifying objects. Such features are also nonaccidental properties of all writing systems because these cultural inventions have been shaped by the “learnability constraints” of the human brain ( Dehaene, 2009 ). The anatomical connectivity of this region is also ideally suited to serve as the “brain’s letterbox” in reading ( Dehaene, 2009 ): filtering visual word forms from the perceptual stream and funneling them to temporal and frontal areas involved in phonological and semantic processing. Direct evidence for the emergence of the visual word form system with reading experience is provided by Dehaene et al.’s (2010) comparison of neural activity in populations varying in levels of literacy. This study found that the acquisition of literacy is associated with increased activation of the VWFA area by words and reduced activation of this region by faces and objects.

Convergent evidence for the important role of early visual processing in skilled reading is provided by neuroimaging comparisons of successful and unsuccessful readers. Pugh et al.’s (2001) review of such evidence concluded that successful reading was associated with higher levels of activation in the ventral visual processing areas corresponding to the VWFA during the early stages of word identification, as well as with stronger activation in temporoparietal areas of the dorsal visual system during later processing. However, successful readers showed less activation than disabled readers in frontal regions, centered around Broca’s area, which are associated with phonological output processes such as fine-grained articulatory coding. Pugh et al. interpreted these data as indicating that the dorsal system plays an important role during the initial stages of reading as children learn to integrate the orthographic, phonological, and semantic features of words. Such integration enables development of the specialized “linguistically structured, memory-based word identification system” of the VWFA ( Pugh et al., 2001 , p. 245). Disabled readers rely on a different reading circuit because deficits in the temporoparietal processes required to integrate orthographic, phonological, and semantic features impede development of a well-structured VWFA system. They are therefore forced to rely on compensatory strategies indexed by increased activation of frontal areas associated with phonological processing. Pugh et al. suggest that this may reflect increased reliance on covert pronunciation to compensate for limitations in the automatic retrieval of the phonological codes required for comprehension.

The Cognitive Architecture of Skilled Reading

This selective overview of evidence about the neural circuitry underlying reading has important implications for the current discussion of individual differences in skilled reading because it demonstrates how the cognitive system of the developing reader is shaped by the process of learning to read. Dehaene et al. found changes in populations who acquired literacy in adulthood, emphasizing that “both childhood and adult education can profoundly refine cortical organization” (2010, p. 1359). This evidence of late plasticity provides a mechanism for the gradual refinements to the quality of lexical knowledge that is implied by the assumption that lexical quality is a word-specific property which “accumulates with age and experience” ( Perfetti, 2007 , p. 380). The VWFA’s selective response to the visual features that define letters and words also demonstrates how orthography-specific features come to be embedded in the “mosaic of category-specific regions in ventral cortex” ( Dehaene et al., 2010 , p. 362) to take advantage of areas with appropriate receptive fields and anatomical connectivity to the brain regions that code semantic information. The VWFA therefore provides a neural mechanism that is compatible with the construct of orthographic precision emphasized by the LQH.

Pugh et al. (2001) ’s systems-level analysis demonstrates how strengths and weaknesses in the processes required to integrate orthography with existing representations of spoken language can shape the functional architecture of the reading system. Successful readers develop a specialized automatic word identification system that supports fast, efficient access to all components of a word’s identity. This reduces the need for compensatory, phonologically mediated word identification strategies, like those employed by less proficient readers.

This interpretation is consistent with the well-established interactive compensatory framework of individual differences in reading ( Stanovich, 1992 ). This framework assumes that bottom-up decoding processes and top-down comprehension processes both contribute to word identification. The balance between them depends on a range of factors, including reading skill. In early applications of this framework, the central causal factor determining the balance between bottom-up and top-down processes was working memory. Automatic word decoding was assumed to benefit comprehension by reducing the drain on central processing resources so that they could be directed to comprehension. As reviewed earlier, Perfetti (2007) shifted the causal focus to the quality of lexical knowledge. High-quality representations support effective comprehension not simply because they support automatic retrieval but also because coherent, synchronous activation of the word’s constituents provides the semantic and grammatical information required for comprehension. Readers who fail to develop such knowledge need to co-opt other processes, such as articulatory recoding and contextual prediction, to support both word identification and comprehension. This leads to differences in the balance of labor between automatic word identification and top-down processes.

The neuroimaging evidence therefore provides an answer to the question of why the precision and coherence of lexical knowledge is important to skilled reading. Precise knowledge about orthographic word forms provides rapid, specialized access to the tightly interconnected lexical information necessary for effective comprehension.

Lexical Quality and Models of Skilled Reading

This characterization of skilled reading is also consistent with modular views of lexical retrieval in reading ( Forster, 1989 ), which assume that perceptual features are automatically mapped to lexical forms with minimal, if any, influence of context. This is the implicit or explicit assumption of many current theories of visual word recognition, which focus on retrieval of orthographic word forms without specifying how they interact with semantic knowledge (e.g., Coltheart et al., 2001 ; Davis, 2010 ; Perry et al., 2007 ).

Many models are based on the hierarchical interactive-activation (IA) architecture ( McClelland & Rumelhart, 1981 ), which assumes that activation of orthographic word forms is a prerequisite for accessing semantic knowledge (see Taft , this volume). There are two senses in which these models implicitly assume that all skilled readers have developed high-quality representations for all words in their vocabulary. At the micro level, all words have fully specified connections between letter and word units, paralleling the precision criterion for lexical quality. At a systems level, few models allow the possibility that readers vary in the time course or weighting of different sources of knowledge, reflecting differences in the strength of connections between lexical constituents that might influence the coherence of their activation.

The limitation of these models from the perspective of the LQH is that they fail to accommodate the graded nature of lexical quality both within and between individuals or to recognize that the refinement of lexical representations may be a gradual process that continues at least into adolescence and shapes the cognitive architecture of the reading system. I will consider how the models might be modified to accommodate individual differences in lexical quality among skilled readers in the concluding section. But it is first necessary to consider how lexical quality should be defined and measured.

What Is Lexical Quality?

As reviewed earlier, Perfetti (2007) identified lexical quality as precise, stable, word-specific knowledge that supports coherent activation of all components of a word’s identity. Although this definition incorporates both form and meaning, the neuroimaging evidence of a specialized system for processing the features of written words highlights the special status of orthographic knowledge in the development of literacy. This conclusion converges with theories of reading development that also emphasize the importance of precise orthographic representations in successful reading acquisition ( Ehri, 2005 , this volume).

Orthographic Precision and Lexical Quality

Phase theories of reading development (see Ehri, this volume) propose a sequence of transitions in children’s representations of written words. As children acquire phonological awareness, their initial prealphabetic codes, based on idiosyncratic visual features and context (e.g., identifying McDonalds by the golden arches of its initial letter), are gradually connected to developing knowledge of letter–sound relationships to form partial alphabetic codes. To progress to the next stage of full alphabetic codes, in which all of a word’s letters are connected to their pronunciations, requires children to be able to segment words into phonemes and map them to graphemes. Achieving this transition benefits from systematic phonics instruction ( Ehri, Nunes, Stahl, & Willows, 2001 ) and training that emphasizes articulatory features ( Boyer & Ehri, 2011 ). Full alphabetic codes are assumed to support accurate identification and spelling of most familiar words and relatively effective decoding of novel words.

However, fully automating word identification depends on a further transition to a consolidated alphabetic phase that is defined by unitization of larger orthographic chunks, such as syllables, rimes (e.g., ‹amp› in camp, damp, stamp ), and morphological units such as roots (e.g., ‹mit› from admit, submit ) and affixes (e.g., ‹pre›, ‹ed›). Word codes that are fully unitized can be read as whole units, reflected in reading speeds as fast as for single digits ( Ehri & Wilce, 1983 ). Consolidated representations also enhance acquisition of new words by facilitating the use of analogy strategies. Representing multiple levels of orthographic-phonological (O-P) correspondence facilitates automatic word identification by securely bonding words’ spellings to their pronunciations so that they are directly activated by print. It also reduces the number of links required to connect them—for example, interesting can be linked as four syllabic chunks rather than multiple grapheme-phoneme units ( Ehri, 2005 ).

Ehri’s description of the phases of lexical development aligns well with the LQH’s assumption that high-quality lexical knowledge is acquired gradually at the level of the individual word. Transition between phases is gradual rather than discrete, so that a reader’s vocabulary may include words at different phases of development. Ehri also emphasizes the importance of establishing strong connections between the different components of lexical identity. Like Perfetti (2007) , she argues that this amalgamation of consolidated alphabetic codes with their semantic and syntactic features supports synchronous coactivation of these constituents, which facilitates automatic word identification,

The distinction between full and consolidated alphabetic representations fleshes out the critical construct of orthographic precision. According to Ehri, full alphabetic representations specify all the letters of a word and link them to phonology, the properties that Perfetti (2007) identified with precision. However, they do not necessarily support direct, automatic identification, particularly for words that have many orthographically similar word neighbors that contain many of the same letters. Consolidated alphabetic codes that represent multiple grain size s of O-P correspondence ( Ziegler & Goswami, 2005 ) facilitate discrimination between similar words by reducing ambiguities in spelling-sound correspondence. In English, for example, many irregularities in O-P mapping are reduced when larger chunks are considered ( Kessler & Treiman, 2001 ). For example, the grapheme ‹ea› takes many different pronunciations (e.g., bead, dead, great ), but it is more likely to be pronounced /ɛ/ when followed by ‹d› than by any other letter (e.g., dead vs dean; head vs heal ). Context also reduces ambiguity in mapping from phonology to orthography. For example, the phoneme /ʊ/ can be spelled in different ways, but ‹oo› is more likely before /k/ (e.g., book, took, look ) while ‹u› is more common before /l/ or /ʃ/ (e.g., full, bull, push, bush ).

The averaged performance of skilled readers is sensitive to these statistical patterns in both spelling-sound mapping, reflected in the pronunciations assigned to nonword items (e.g., Andrews & Scarratt, 1998 ; Treiman, Kessler, & Bick, 2003 ) and sound-spelling mapping, assessed in spelling of both nonwords and words (e.g., Treiman, Kessler, & Bick, 2002 ). Investigations of children have shown that this sensitivity develops gradually and appears to take longer to manifest in spelling than in reading: Vowel pronunciations of nonwords were significantly modulated by the following consonants as early as grade three ( Treiman, Kessler, Zevin, Bick, & Davis, 2006 ), but corresponding effects on spelling performance only became evident between grades 5 and 7 ( Treiman & Kessler, 2006 ). These findings are consistent with evidence that larger orthographic units play an increasing role in more skilled readers (Ehri, this volume). Treiman et al. (2002) showed that spelling ability was correlated with contextual influences on vowel spelling, but there is little systematic evidence about individual differences in sensitivity to these factors among skilled readers.

Structured orthographic codes that represent larger orthographic chunks such as rimes would facilitate discrimination between words that are similar at the level of individual letters and graphemes (e.g., good and book would be less similar to goon and boot than if they were coded purely as grapheme-phoneme units). They also contribute to securing the representation of the order of letters within the word. A representation that only codes individual grapheme-phoneme units is vulnerable to being confused with another word that contains the same letters in a different order (e.g., calm/clam; eager/agree ). This vulnerability is reduced if the representation also codes rime (e.g., ‹am› is not contained in calm ) or syllable units (e.g., agree does not contain ‹er›). Unitized morphemic units provide another dimension that can differentiate between similar words (e.g., hunter and hunger share five of their six letters, but contain different root morphemes). Morphologically structured orthographic representations also support the early morphological parsing and lemma extraction proposed in Taft’s (this volume) AUSTRAL model.

The refinement of orthographic representations that results from integrating multiple grain sizes of mapping to phonology has been referred to as lexical restructuring ( Ziegler & Goswami, 2005 ) and lexical tuning ( Castles et al., 2007 ) to highlight its impact on the similarity relationships between orthographic forms. More finely tuned representations overlap less with the representations of orthographically similar words, making them less vulnerable to being activated by letter strings that contain similar components.

Thus, converging evidence from developing and skilled readers highlights the critical role of precise orthographic representations of words. However, there are major gaps between these bodies of evidence. Critically, there is little evidence about the later phases of reading development. Few developmental studies go past the early primary years, as if mastery of the alphabetic principle is sufficient to ensure success at reading. This view ignores the further refinements in orthographic learning required to achieve consolidated representations.

There are also differences in conceptual approach. Studies of reading development have attempted to identify the factors that predict orthographic learning but have paid little attention to underlying mechanisms ( Castles & Nation, 2008 ). By contrast, the skilled-reading literature has used averaged data to build very detailed models of orthographic processing (see Grainger, 2008 ) but virtually ignored individual differences.

However, because orthographic learning operates at the item level ( Share, 2004 ) and the acquisition of written vocabulary takes more than a decade ( Nation, 2005 ), it must continue at least into adolescence ( Andrews, 2012 ). Moreover, a substantial proportion of readers’ written language vocabulary is acquired through reading. Of the 60,000 to 90,000 unique words known by the average high school graduate, as few as 200 are explicitly taught at school and less than 20,000 are encountered outside the school context ( Landauer, Kireyev, & Panaccione, 2011 ). If rich, consolidated representations enhance effective application of analogy strategies to learn new words, as Ehri (this volume) argues, individual differences in the quality of lexical knowledge across readers will magnify over adolescence because readers who fail to develop consolidated representations for many words will be able to identify fewer words automatically and be less effective at developing stable representations of new words. This will encourage greater reliance on contextual prediction to identify words, which will, in turn, reduce the attention to orthography necessary to extract the larger orthographic chunks that contribute to consolidated representations.

Frith (1986) argued that this partial cue reading strategy of using coarse orthographic cues, supplemented by contextual prediction, is characteristic of good readers/poor spellers who demonstrate above average reading comprehension and knowledge of grapheme-phoneme correspondences but who perform well below their reading level on tests of spelling. A contextually driven strategy supports fast, efficient comprehension in many reading contexts but impedes extraction and unitization of the larger grain size units required for consolidated representations. Adoption of such a strategy is therefore likely to perpetuate and exaggerate individual differences in the precision of readers’ orthographic representations. Differences in orthographic precision are likely to remain a source of variability in adult populations of competent readers ( Ziegler & Goswami, 2005 ). The next section reviews evidence evaluating the validity of this hypothesis.

How Does Lexical Quality Influence Skilled Reading?

Most of the relatively sparse literature on individual differences among skilled readers has used measures of reading comprehension or vocabulary as predictor variables (e.g., Ashby et al., 2005 ; Perfetti & Hart, 2001 ; Yap et al., 2009 , 2012 ). While such studies have provided useful evidence of systematic variability within this population, these relatively gross measures cannot provide direct evidence that these differences are due to lexical quality.

The most systematic body of evidence about individual differences in word identification among skilled readers has relied on measures of vocabulary, because it is an easily administered and calibrated index of lexical knowledge. It has therefore been included in a number of recent megastudies of skilled word recognition, including the English Lexicon Project ( Balota et al., 2007 ), which collected measures of performance in a range of word identification tasks for large samples of words from participants from different universities (see Yap & Balota , this volume). To complement their analyses of the averaged data collected in this project, Yap, Balota, and colleagues have recently investigated how vocabulary modulates lexical decision and naming performance ( Yap et al., 2012 ) and semantic priming effects ( Yap et al., 2009 ). Vocabulary level is interpreted as an index of lexical integrity , which is identified with the coherence and stability that define high-quality lexical representations ( Perfetti, 2007 ).

Using sophisticated analysis of response time data, Yap et al. (2009) showed that a sample of participants with high average vocabulary showed an additive relationship between word frequency and semantic priming, while a sample of lower vocabulary participants showed larger priming effects for low-frequency words, particularly for slower responses. Similarly, Yap et al.’s (2012) analysis of naming and lexical decision data showed that vocabulary was most strongly associated with variability in slow responses and that higher vocabulary was associated with reduced effects of frequency and O-P consistency. These findings were interpreted as showing that higher lexical integrity is associated with more automatic, modular lexical retrieval processes.

Similar conclusions have been drawn from studies using the combination of vocabulary and reading comprehension to index individual differences among skilled readers. Using the more naturalistic method of assessing eye movements during sentence reading, Ashby et al. (2005) compared the effects of predictability and frequency on the eye movements of highly skilled and average readers classified according to a median split on a composite measure of reading comprehension and vocabulary. The highly skilled group showed similar patterns of reading behavior for predictable, neutral, and unpredictable sentences, suggesting that they relied on automatic, context-independent word identification processes. By contrast, the average readers were strongly influenced by sentence context. Ashby et al. concluded that even within this sample of generally skilled readers, lower levels of proficiency were associated with increased reliance on context to compensate for less efficient word identification. They suggested that this contextually based strategy might “interfere with feedback processes that help form the crisp, stable … representations needed to access that word in the future” (p. 1083). This interpretation attributes the automatic, context-independent word identification strategy adopted by highly skilled readers to the precision of their orthographic representations. However, the measures of reading comprehension and vocabulary used to assess proficiency do not provide a direct measure of this construct.

Reading comprehension is a multifaceted skill that depends on domain-general processes such as working memory and inferential processes, as well as knowledge and processes that are specific to reading. Tests of reading comprehension vary in how effectively they assess these different components ( Keenan, Olson, & Betjeman, 2008 ). Vocabulary size, breadth, and depth are also multiply determined and not specific to reading. Comprehension and vocabulary also tend to be at least moderately correlated and are therefore often combined to provide a more reliable index (e.g., Ashby et al., 2005 ). However, given the range of other cognitive processes that contribute to both measures, there is no guarantee that the variance they share is related to the quality or precision of lexical knowledge. Most notably, these measures fail to capture the critical construct of orthographic precision highlighted by both the neuroimaging and developmental evidence.

Isolating the Components of Lexical Quality

The goal of recent research in my laboratory has been to go beyond global measures of lexical quality to investigate specific contributions of the precision and coherence of lexical knowledge to skilled word identification and reading. To assess the contribution of word-specific orthographic knowledge, I have developed measures of spelling ability. Separate tests of spelling dictation and spelling recognition are combined to provide a broad index of spelling ability. These highly correlated measures ( r between .7 and .8) both have high test-retest reliability ( r > .9) and yield a wide range of scores in our university student samples ( Andrews & Hersch, 2010 ).

A central feature of our approach to assessing orthographic precision has been to combine measures of spelling ability with tests of reading comprehension and vocabulary in order to isolate variance specific to orthographic precision. As would be expected, these three measures are intercorrelated: In a recent, typical sample of 200 students from my laboratory, vocabulary was moderately correlated ( r = .54 and .55) with both reading comprehension and spelling ability, which were more weakly related ( r =. 37). Nevertheless, by using reasonably large samples of participants and a small focused test battery, we have successfully applied multivariate regression procedures ( Baayen, 2008 ) to distinguish variance common to all tests from variance unique to spelling. Using this approach, we have made substantial progress in demonstrating specific contributions of individual differences in orthographic precision to both word identification ( Andrews & Hersch, 2010 ; Andrews & Lo, 2012 ) and sentence processing ( Andrews & Bond, 2009 ; Hersch & Andrews, 2012 ; Veldre & Andrews, 2014 , in press ) and are now extending this work to distinguish individual differences in precision and lexical coherence ( Andrews & Lo, 2013 ; Andrews, Xia, & Lo, 2014 ).

It is important to emphasize that our participants are screened to exclude non-native English speakers to ensure that variability is not due to bilingualism. Our participants also tend to be of superior reading ability. According to the US four-year college norms for the Nelson-Denny Reading Test ( Brown, Fishco, & Hanna, 1993 ), our participants generally fall above the 50th percentile (e.g., medians of 80 and 86 for the sample of 200 referred to previously). We focus on this population of above-average readers because, as in other domains of expertise, understanding how lexical experts perform the task of reading provides insight into the optimal cognitive architecture for this complex skill ( Andrews, 2008 ).

Another important feature of our approach has been the choice of experimental paradigms. Detecting individual differences among above average readers requires methods that test the limits of the reading system to reveal constraints and limitations that are masked or compensated for under less demanding conditions. At the same time, it is important to avoid tasks that are difficult because they place heavy demands on central resources. For example, a common method of increasing the demands of word identification is to present stimuli in a degraded form. Comparisons of the effects of degradation have also been used to investigate individual differences ( Ziegler, Rey, & Jacobs, 1998 ). But high-quality representations facilitate both more efficient extraction of perceptual information and more effective use of educated guessing strategies based on partial information, which depend on central resources. Any interactions with individual difference observed in such tasks may reflect general differences in cognitive capacity rather than reading-specific effects.

The more general methodological problem in isolating processes associated with lexical quality is that more proficient readers tend to be better at all the processes required for effective reading. The LQH attributes causality to precise lexical representations but assumes that such representations also support effective higher-level processes. Many tasks conflate these sources of variability, making it difficult to disentangle the effects of precise lexical knowledge from the flexibility of integration and decision strategies that they afford. For example, meaning judgment tasks that require participants to judge the synonymy of word pairs ( Perfetti & Hart, 2001 ) or decide whether a word is related to the meaning of a sentence ( Gernsbacher & Faust, 1991 ) require a combination of lexical retrieval and decision processes. This makes it difficult to isolate the cause of any individual differences that are observed.

To avoid these problems, our recent research has relied on two major methodological approaches. To investigate the contributions of orthographic precision and lexical coherence to isolated word identification, we have used variants of the masked priming task. To assess how lexical quality contributes to sentence processing, we are investigating individual differences in eye movements during sentence reading. I focus primarily on masked priming data but also briefly describe some of our recent eye movement data.

Individual Differences in Masked Orthographic Priming

The masked priming paradigm developed by Forster and Davis (1984) is ideal for separating individual differences in early lexical retrieval from those due to decision processes. In this task participants respond to a clearly presented uppercase target stimulus, which is preceded by a brief (~50 ms) lowercase prime that is forward-masked by a symbol string (#####) and backward-masked by the target. The fact that responses are made to a clear, visible target eliminates the problems associated with using stimulus degradation to tax the system. Instead, the ability to extract information from the degraded, masked prime is assessed by investigating the impact of different types of primes on responses to the target. Because participants are usually unaware of the presence or identity of the prime, such influences are unlikely to reflect prediction or decision strategies. These higher order processes should affect overall speed and sensitivity to the impact of target attributes, while individual differences that affect early lexical retrieval should appear as interactions with priming manipulations.

Our initial investigations of the contribution of orthographic precision to lexical retrieval focused on masked neighbor priming ( Andrews & Hersch, 2010 ) because data from this paradigm had been used in both the developmental and skilled reading literature to track the tuning of lexical representations with reading experience. Using the typical definition of orthographic neighbors as stimuli that share all but one letter with the target word, Castles, Davis, Cavalot, and Forster (2007) found that third-grade children showed facilitation from masked nonword neighbor primes (e.g., pley-PLAY ), which disappeared by the time they were in fifth grade. Castles and colleagues interpreted this finding as suggesting that as children’s written word vocabulary increases, their orthographic representations are fine-tuned to reduce the overlap between similar words. Similar conclusions derive from differences in skilled readers’ masked priming for different items. Averaged data for samples of skilled readers typically show facilitatory masked neighbor priming for words with few neighbors (e.g., eble-ABLE ), but not for targets (e.g., tand-SAND ) that are similar to many words ( Forster, Davis, Schoknecht, & Carter, 1987 ). The lack of priming for words with many neighbors suggests that the representations of these words may have been restructured by incorporating larger grain size units ( Ziegler & Goswami, 2005 ) to enhance discrimination between similar words ( Forster & Taft, 1994 ).

These data are consistent with the LQH’s assumption that the acquisition of precise orthographic representations is a gradual process that varies between the words in an individual’s vocabulary. We hypothesized that if the extent and success of the tuning process varies among skilled readers, as the LQH implies, individual differences in masked neighbor priming should still be evident within samples of skilled readers ( Andrews & Hersch, 2010 ). Further, if spelling ability provides an index of the orthographic precision of word representations, it should be the best predictor of individual differences in neighbor priming. Our masked neighbor priming lexical decision data supported these predictions. The averaged data replicated Forster et al.’s (1987) finding that neighbor priming only occurred for targets with few neighbors. However, analyses including individual difference measures showed that the null priming for many neighbor targets in the averaged data obscured significant effects of spelling: good spellers showed inhibitory priming from neighbor primes, while poor spellers showed facilitation. We interpreted these findings as showing that spelling ability is selectively associated with inhibitory effects of lexical competition. However, the inhibitory priming shown by better spellers was specific to targets with many word neighbors. For targets with few neighbors, better spellers showed facilitation effects equivalent to those shown by poorer spellers.

This pattern is consistent with interactive-activation models, in which orthographic priming reflects the combination of facilitation due to sublexical overlap and inhibition due to lexical competition ( Perry, Lupker, & Davis, 2008 ). Neighbor primes benefit target processing at the sublexical level because they preactivate most of the target’s letters. However, if the prime becomes sufficiently strongly activated at the word level it will inhibit the representations of other similar words, potentially including the target word. Spelling ability appears to selectively predict the strength of the competition between words with many neighbors.

To provide further insight into the basis of these individual differences in orthographic precision, we subsequently compared the effects of neighbor primes with transposed-letter (TL) primes , which contain all the letters of the target word in a different order (e.g., clam-CALM ). Transposed-letter items have played an important role in theories of visual word recognition because they provide insight into the relative contribution of letter identity and letter order to word recognition ( Kinoshita , this volume). Averaged data for skilled readers has revealed that they are relatively poor at resolving letter order: TL nonwords (e.g., jugde ) are often misclassified as words, and masked TL nonword primes (e.g., jugde-JUDGE ) tend to facilitate performance more than neighbor primes (e.g., jurge-JUDGE ) (e.g., Grainger, 2008 ). In contrast, TL word primes (e.g., salt-SLAT ) often yield interference (e.g., Andrews, 1996 ). Such evidence has stimulated a substantial recent body of empirical research and computational modeling focused on “cracking the orthographic code” of skilled reading (see Grainger, 2008 ). However, this research has relied on averaged data. Our evidence that spelling ability modulates neighbor priming suggests that the averaged effects of TL priming may also obscure systematic individual differences.

To investigate these issues, we directly compared TL and neighbor primes for the same target words ( Andrews & Lo, 2012 ). To disentangle facilitatory effects of sublexical overlap from lexical competition, we compared word and nonword primes (e.g., plot, colt, CLOT; clib, cilp, CLIP ). The sample of 90 students in this study showed higher correlations between reading ability, spelling ability, and vocabulary ( r = .64–.75) than those tested by Andrews and Hersch (2010) . To separate the effects of general proficiency from effects specific to spelling, we therefore used principal components analysis to define two orthogonal individual difference measures: an index of overall proficiency that was essentially the average of the three composite scores and a second factor corresponding to the difference between spelling and reading/vocabulary that I will refer to as the spelling-meaning factor.

Analyses including these two independent individual difference measures showed that overall proficiency was associated with faster overall RT, confirming the expected association between higher proficiency and more efficient lexical classification, as well as stronger effects of prime lexicality. As summarized in Figure 10.1 , which presents average priming effects for TL and neighbor primes relative to unrelated primes, lower proficiency participants showed minimal, and equivalent, priming for both types of prime. 1 In contrast, high proficiency participants showed facilitatory priming from nonword neighbor primes (e.g., clib-CLIP ) but inhibitory priming from TL word primes (e.g., calm-CLAM ). Over and above these effects of overall proficiency, the second, spelling-meaning factor also significantly modulated priming. As depicted in Figure 10.2 , individuals with unexpectedly high spelling ability for their level of reading/vocabulary showed inhibitory priming, particularly for TL primes, that was not evident in those with poorer spelling ability than expected from their reading/vocabulary scores. The latter subgroup showed facilitatory priming, particularly from neighbor primes.

Average transposed-letter (TL) and neighbor (N) priming effects (in ms) separately for participants above the median (high overall proficiency) and below the median (low overall proficiency) on the overall proficiency factor. Positive priming scores indicate facilitation relative to the unrelated prime condition, while negative scores indicate inhibition.

Average transposed-letter (TL) and neighbor (N) priming effects (in ms) separately for participants above the median (spelling>meaning) and below the median (meaning>spelling) on the spelling-meaning factor. Positive priming scores indicate facilitation relative to the unrelated prime condition, while negative scores indicate inhibition.

These results complement and extend Andrews and Hersch’s (2010) findings. By using principal components analysis, we identified two independent dimensions of individual difference that both significantly interacted with priming. The composite measure of overall written language proficiency was associated with faster performance and strong sensitivity to the lexical status of the prime and its orthographic relationship to the target. These data suggest that the dimension tapped by this index of overall proficiency is associated with rapid, efficient processing of the briefly presented prime. When the prime is a nonword neighbor that orthographically overlaps with the target (e.g., clib-CLIP ), the shared sublexical constituents preactivate the target representation so that it is more quickly retrieved and classified when the target is presented. However, a TL nonword prime that contains all the letters of the target in a different order (e.g., cilp-CLIP ) creates perceptual ambiguity about letter position/order that reduces the extent of target preactivation. When the prime is a similar word, rapid prime processing allows the prime word’s representation to become sufficiently strongly activated to inhibit the representation of the target word, counteracting the facilitation derived from sublexical overlap. If the prime is a word neighbor of the target (e.g., plot-CLOT ), the net effect is no priming. But TL word pairs (e.g., colt-CLOT ) yield strong inhibitory priming that systematically increases with reading proficiency. This result suggests that lexical competition may be stronger between more orthographically similar words, perhaps particularly those that differ only in letter order ( Davis, 2010 ).

Andrews and Hersch (2010) found that lexical competition was selectively associated with spelling ability, but they did not directly evaluate the contribution of overall proficiency. Andrews and Lo’s (2012) evidence that this dimension was also associated with stronger sublexical facilitation suggests that it taps broader aspects of lexical quality than orthographic precision. By combining spelling, reading comprehension, and vocabulary, this measure captures both orthographic and semantic aspects of lexical knowledge. It may therefore index the strength of connections between different lexical constituents that both Perfetti (2007) and Ehri (this volume) identify as critical to the coherent, automatic activation of the multiple constituents of a word’s identity necessary for flexible, effective use of lexical knowledge.

Over and above the effects of overall proficiency, spelling also made an additional contribution to predicting priming reflected in stronger inhibition—particularly from TL primes—in individuals with unexpectedly high spelling ability for their level of reading and vocabulary. This independent source of variance in inhibitory priming was insensitive to prime lexicality, suggesting this factor may tap perceptual processes involved in resolving letter order that are specifically associated with orthographic precision. Many current models of orthographic coding attribute TL priming effects to the greater perceptual ambiguity of information about letter order than letter identity (e.g., Gomez, Ratcliff, & Perea, 2008 ; Norris & Kinoshita, 2012 ). The unique variance tapped by the spelling-meaning factor may reflect differences in the efficiency of resolving ambiguous information about letter order, which is a prerequisite for the rapid activation of prime word representations required to trigger inhibition of similar words. However, the strength of that lexical competition depends on broader properties of high-quality representations, such as tight constituent binding and coherence. Nonword primes yield diffuse activation across multiple letter and word representations that does not converge on a single lexical identity, as occurs for word primes. The more tightly interconnected the lexical constituents of the prime word, the more strongly and coherently they will be activated by the brief prime presentation, and the more they will inhibit representations of similar words. It is therefore the coherence and synchrony of the early activation associated with a stimulus, tapped by the overall proficiency factor, that distinguishes word from nonword primes and predicts effects of prime lexicality.

These results provided insights into the sources of variance differentiating skilled readers that go beyond the subgroup ( Andrews & Bond, 2009 ) and regression ( Hersch & Andrews, 2012 ) approaches that we have used previously. They provide evidence for independent contributions of orthographic precision, indexed by the spelling-meaning factor, and the broader aspects of lexical quality indexed by overall proficiency, which I will refer to as lexical coherence . In masked orthographic priming tasks, lexical coherence predicts both greater benefit from sublexical overlap and greater sensitivity to lexical competition. The second dimension of orthographic precision adds an additional, perhaps more perceptually based, component that is associated with stronger competition between similar words in individuals with the orthographic profile of higher spelling than reading/vocabulary ability.

The apparent paradox of identifying high lexical quality with competition processes that are detrimental to target identification is specific to the task demands of the masked priming paradigm. Under these conditions, rapid activation of the prime hurts performance because faster access to the prime leads to stronger inhibition of similar words, including the target. In normal reading, where readers do not receive contradictory perceptual input, rapid, precise lexical retrieval benefits reading ( Andrews, 2012 ). As discussed later, orthographic precision also facilitates parafoveal processing during text reading ( Veldre & Andrews, 2014 , in press ).

Individual Differences in Masked Morphological Priming

To provide further insight into the role of individual differences in orthographic precision and lexical coherence, we have begun to assess semantic processing of masked primes. The modular models of skilled lexical retrieval described earlier typically assume that processing follows a hierarchical, form-first sequence in which activation of orthographic forms precedes, and operates independently of, activation of meaning units at higher levels of the hierarchy ( Forster, 2004 ). These assumptions have been most explicitly elaborated in models of morphological priming, which assume that complex words are segmented into morphemes that serve as access units to semantic and conceptual information about whole words. These morphographic models vary in their assumptions about the precise representational levels involved and how they interact (see Taft, this volume), but they share a commitment to an early orthographic decomposition process that is insensitive to semantic similarity (e.g., Rastle & Davis, 2008 ). Other researchers are equally committed to the alternative view that semantic information is available very early in processing and influences morphological segmentation and decision processes (e.g., Feldman, O’Connor, & Moscoso del Prado Martin, 2009 ).

Efforts to distinguish these competing theoretical claims have relied on comparisons of masked morphological priming for semantically transparent morphologically related pairs (e.g., dreamer-DREAM ); pseudomorphemic pairs (e.g., corner-CORN ), which have the same apparent morphological structure as transparent pairs but are not semantically related; and orthographic control pairs (e.g., brothel-BROTH ) in which the prime word is not morphologically complex. An informal meta-analysis of masked priming studies comparing these conditions led Rastle and Davis (2008) to conclude that transparent and pseudomorphemic pairs showed equivalent priming relative to control pairs, as predicted by the orthographic decomposition account. However, a reanalysis of essentially the same data by proponents of the early semantic view concluded that priming was stronger for transparent than pseudomorphemic pairs ( Feldman et al., 2009 ). The studies evaluated by both groups all relied on averaged skilled reader data.

We reasoned that these contradictory conclusions may reflect individual differences in orthographic precision and lexical coherence paralleling those observed in masked orthographic priming tasks ( Andrews & Lo, 2013 ). Given Yap et al.’s (2009) evidence, reviewed earlier, that individual differences in vocabulary predict semantic priming, we used measures of spelling and vocabulary as predictors in order to tease apart individual differences in reliance on orthographic and semantic knowledge. Following the procedures in Andrews and Lo (2012) , principal components analysis was used to separate overall proficiency from a second, independent factor reflecting the difference between spelling and vocabulary.

Average morphological priming effects for the form, pseudomorphemic, and transparent conditions separately for participants above the median (spelling > meaning) and below the median (meaning > spelling) on the spelling-meaning factor. The positive priming scores indicate facilitation relative to the matched unrelated prime for each condition.

The results of Andrews and Lo (2013) showed that high overall proficiency predicted significantly faster responses to transparent stems but did not interact with any measures of priming. Significant interactions with morphological priming did, however, emerge for the second, spelling-meaning factor. As summarized in Figure 10.3 , individuals with higher spelling than vocabulary showed equivalent priming for transparent and pseudomorphemic pairs, which both produced stronger priming than orthographic controls. However, those with the opposite profile of higher vocabulary than spelling showed priming for transparent but not pseudomorphemic or control pairs. Thus the orthographic profile was associated with the pattern predicted by form-first models, while the reverse semantic profile predicted the pattern identified with early semantic influences.

Andrews and Lo (2013) emphasized that the orthographic and semantic profiles should not be seen as distinct subgroups but as evidence of an independent source of individual differences over and above differences due to overall proficiency. In the morphological priming task overall proficiency did not interact with priming, perhaps because the critical stimuli were equated on orthographic overlap. Rather, differences in the pattern of priming were predicted by the independent spelling-meaning dimension that appeared to tap the availability and use of orthographic relative to semantic knowledge in lexical classification.

Individual Differences in Masked Semantic Priming

To provide further insight into this dimension of individual difference among skilled readers and assess its generality across tasks, we are currently investigating masked semantic priming in the semantic categorization task ( Andrews et al., 2014 ). In contrast to the lexical decision task, which is relatively insensitive to semantic priming (e.g., Forster, 2004 ), semantic categorization judgments for broad categories like “animal” yield category congruence effects from masked primes, which implicate activation of semantic attributes of the prime ( Quinn & Kinoshita, 2008 ). These effects are indexed by faster categorization of target words preceded by primes from the same category as the target: both faster “yes” responses to exemplar targets preceded by animal than nonanimal primes (e.g., mole-EAGLE vs. boots-EAGLE ) and faster “no” responses to nonexemplars preceded by nonanimal than animal primes (e.g., boots-RIFLE vs. mole-RIFLE ). Forster (2004) found that such effects only occurred for narrow, finite categories like single-digit numbers or months, for which decisions can be based on searching the entire category. However, Quinn and Kinoshita (2008) reported significant category congruence priming on animal categorization responses, but only for category congruent primes high in semantic feature overlap with the target (e.g., hawk-EAGLE ; pistol-RIFLE ).

We have recently replicated this evidence that high overlap semantic primes significantly facilitate average semantic categorization judgments, but we found that the extent and breadth of priming was significantly modulated by both the overall proficiency and spelling-meaning factors. Higher proficiency participants showed equivalent congruence priming for both high and low overlap primes (e.g., pistol-RIFLE and boots-RIFLE ), while low proficiency participants only showed priming for high overlap primes. Significant interactions with the independent spelling-meaning factor showed that semantic priming was stronger in those with the semantic profile of higher vocabulary than spelling ability and entirely absent in individuals with above average spelling ability but below average vocabulary knowledge. These findings converge with those of Andrews and Lo (2013) in revealing individual differences in sensitivity to semantic attributes of masked primes. However, in contrast to our morphological priming data, in which semantic influences were predicted by the spelling-meaning dimension of variability, overall proficiency was a stronger predictor of priming in the semantic categorization task. The different contributions of the two dimensions of variability to predicting morphological and semantic priming may be due to task requirements.

The lexical decision task that we used to assess morphological priming appears to rely more on orthographic than semantic information ( Balota, Cortese, Sergent-Marshall, & Spieler, 2004 ). This may account for why variability in morphological priming was predicted by the more specific spelling-meaning factor rather than overall proficiency. By contrast, congruence priming indexes facilitation from primes that share semantic but not orthographic features with the target. Such priming depends on the tight connections between different lexical constituents that enable rapid activation of the prime’s semantic features, which facilitates rapid categorization of targets that share these features. This coherence dimension of lexical quality is indexed by overall proficiency, which was associated with stronger semantic priming, particularly from more distantly related category congruent primes. Nevertheless, paralleling Andrews and Lo (2013) , there was an additional, independent contribution from the second spelling-meaning dimension reflecting an association between orthographic precision and reduced semantic priming.

There are at least two ways of interpreting the contributions of the spelling-meaning dimension of individual differences to masked semantic priming. One possibility is that the absence of semantic priming associated with the orthographic profile is a direct consequence of orthographic precision: Greater sensitivity to the orthographic features of the prime prevents preactivation of words that are orthographically different from the prime, even when they share semantic features. From this perspective, the reduced semantic priming may be specific to the masked priming task, which presents contradictory orthographic information without any cues to allow the system to distinguish between prime and target events ( Norris & Kinoshita, 2012 ).

The second possibility is that the spelling-meaning factor indexes variability in the speed and efficiency of accessing semantic activation about words. This may be a lingering relic of the poor comprehender profile identified in developmental populations ( Nation & Snowling, 1998 ) who have age-appropriate decoding skills but weak vocabulary and working memory. When the latter deficits are not too severe, cognitively able individuals, perhaps particularly those who are genetically endowed for efficient learning of orthographic-phonological correspondences ( Byrne et al., 2008 ), achieve above average levels of reading skill but remain poorer at reading comprehension than expected from their word-level skills. From the perspective of the LQH, this profile reflects the precision required for rapid lexical retrieval without the synchronous, coherent activation of other components of word identity required for effective, flexible use of this knowledge.

According to this view, orthographic precision is an independent dimension of lexical quality that can coexist with deficits in word-level semantic knowledge that impede lexical coherence. Precision without coherence is not major disadvantage in the lexical decision task, where decisions depend on activation of the word’s orthographic form, but is reflected in stronger sensitivity to discrepancies in letter order and sublexical morpheme units and reduced benefit from semantic overlap between morphologically related words. However, in the semantic categorization task, the reduced speed of activation of semantic information is reflected in reduced benefit from semantic feature overlap.

Individual Differences in Parafoveal Processing

The masked priming studies that I have reviewed provide clear evidence of individual differences in word-level processes. To evaluate whether they also influence reading of connected text, we are currently using eye-movement methods to investigate the contribution of lexical quality to sentence reading. The most exciting and informative strand of that research, led by Aaron Veldre, is focused on individual differences in parafoveal processing (see Schotter & Rayner , this volume). We first demonstrated that reading and spelling ability both predict skilled readers’ perceptual span ( Veldre & Andrews, 2014 ): the size of window of text that readers use during reading ( Rayner, 2009 ). The combination of high reading and spelling ability was associated with a greater benefit from being provided with a larger window of text to the right, but not the left, of fixation, and a greater disruption from being given a very small window ( Veldre & Andrews, 2014 ). This result indicates that higher overall proficiency is associated with greater reliance on parafoveal information.

We are now using gaze-contingent boundary paradigms ( Cutter, Liversedge, & Drieghe , this volume) to investigate how individual differences modulate the nature of the information extracted from the parafovea. Investigations of individual differences in sensitivity to parafoveal previews of word and nonword neighbors converge with our masked priming data in showing that high proficiency readers who are better at spelling than reading show inhibitory parafoveal preview effects of word neighbors on early fixation measures ( Veldre & Andrews, in press ). This result suggests that, in the same way that orthographic precision facilitates rapid processing of masked primes leading to inhibition of the target word representation, precision also supports more rapid extraction of lexical information from the parafovea. When the parafoveal preview is a word neighbor of the target its identification inhibits orthographically similar words, leading to longer fixation durations on targets.

The eye movement measures also provide insight into the costs associated with lower quality lexical knowledge. Low scores on both factors were associated with facilitation from word previews on early fixation measures, paralleling the facilitatory masked priming effects shown by low proficiency readers. This result implies that these readers only extract sublexical features of parafoveal words. Low proficiency readers also showed inhibitory effects of word neighbor previews on later regressions to the target, suggesting that they failed to resolve the conflict between the target and the word neighbor preview, leading to disruptions in later integration processes.


The data reviewed in the preceding section provide converging evidence of systematic differences among skilled readers in the early stages of lexical retrieval tapped by the masked priming task. Varying the nature of the similarity relationships between the prime and the target and the decision requirements of the task revealed two independent dimensions of variance among skilled readers: an overall proficiency dimension, indexed by the shared variance between measures of reading comprehension, spelling and vocabulary, and a separate spelling-meaning factor indexed by the discrepancy between spelling ability and reading comprehension/vocabulary ( Andrews & Lo, 2012 , 2013 ; Andrews et al., 2014 ).

Both factors also significantly modulate skilled readers’ extraction of parafoveal information during sentence reading ( Veldre & Andrews, 2014 , in press ). The eye movement data provide insight into one of the mechanisms underlying the benefits of lexical quality for the reading process: Precise orthographic representations support rapid lexical retrieval, which in turn enables deeper processing of upcoming words (see “Future Directions”).

Implications for Models of Visual Word Recognition

This evidence of systematic individual differences among skilled readers has important implications for computational models of visual word recognition that have simulated the average skilled reader, implying that all skilled readers read in the same way. The models that most obviously lend themselves to accommodating individual differences are the various multiple-route models of both reading aloud (e.g., Coltheart et al., 2001 ) and morphological priming (e.g., Diependaele, Sandra, & Grainger, 2005 ), which allow for differences between items as a function of lexical attributes such as frequency, regularity, and morphological structure. Within such frameworks, higher lexical quality can be conceptualized as an extension of these item differences: Readers with high-quality lexical representations have more functionally high-frequency words that are identified as whole units rather than through computational, sublexical processes. But such accounts describe rather than explain the source of the individual differences ( Andrews & Lo, 2013 ).

An alternative, more perceptual approach to incorporating individual variability in lexical quality is to assume that readers vary in the efficiency with which they extract the visual features of written words and map them to representations in lexical memory. Many models of orthographic coding attribute TL priming effects to difficulties in resolving information about letter position ( Gomez et al., 2008 ; Norris & Kinoshita, 2012 ), and this is clearly a parameter that could be varied to accommodate individual differences. Systematic modeling is required to determine whether individual differences in perceptual efficiency are sufficient to explain the differential orthographic priming from word and nonword primes and the modulation of semantic influences on morphological and category congruence priming. These effects appear to implicate individual differences in the online activation of lexical and semantic knowledge that depend on more than perceptual efficiency. However, these could be consequences of faster perceptual analysis of the prime. Within interactive activation models, word detectors must reach a threshold level of activation before they begin to inhibit other words. If this is achieved faster because of more efficient perceptual processing, there is a greater probability that the prime word will both inhibit other similar words to yield inhibitory orthographic priming and activate its semantic features to support category congruence effects. This analysis suggests that, within a hierarchical interactive-activation architecture, the cascading effects of individual differences in the efficiency of low level perceptual processes could potentially contribute to stronger inhibitory effects of masked orthographic primes and stronger facilitation from masked semantic primes. However, it remains to be seen whether such patterns can be simulated by modulating parameters governing low level perceptual processes. If not, other modifications, such as the addition of noise or weaker connectivity in higher layers of the network, may be necessary to effectively simulate the consequences of inefficient perceptual processes for the precision of lexical knowledge.

Another critical issue raised by the present data concerns the role of lexical competition in word identification. Such competition is central to interactive activation frameworks’ account of lexical retrieval. Modulating the strength of lexical competition may simulate the association between overall proficiency and inhibitory neighbor priming ( Perry et al., 2008 ). The additional assumption that competition is stronger between more orthographically similar words, embodied in Davis’s (2010) spatial coding model, offers another potential source of individual differences in sensitivity to inhibition.

Fully understanding how individual differences in lexical quality emerge during reading development and shape the organization of lexical knowledge and the architecture of the reading system requires the development of computational accounts of how lexical knowledge is learned and refined, rather than further tweaking of models that stipulate the form and organization of lexical knowledge. Individual differences among skilled readers provide empirical constraints that should play an important role in future empirical and theoretical progress toward this goal.

Future Directions

Symbolic versus distributed lexical representations.

From one perspective, the gradual refinement of lexical representations required to develop high-quality lexical knowledge seems compatible with emergentist theories of cognition such as PDP accounts, which attribute the rich, complex structure of human language and thought to simple processes that dynamically modify the weightings of connections between neuron-like processing units. McClelland et al. (2010) contrast these approaches with theories that assume abstract units corresponding to psycholinguistic entities such as phonemes, words, and concepts, and identify cognitive processing with the manipulation of these symbols. Proponents of emergentist approaches argue that these symbolic approaches are misleading approximations that fail to capture the dynamic, flexible similarity relationships that characterize real-world knowledge. However, acknowledging the dynamic interactions underlying the development of symbolic knowledge does not undermine the possibility that the “advantage of imposing a specific structural form on knowledge [may] outweigh its disadvantages” ( McClelland et al., 2010 , p. 354) and contribute to achieving expertise in highly practiced skills like identifying written words.

How Does Lexical Quality Contribute to Sentence Comprehension?

Skilled reading clearly involves more than word identification. The LQH assumes that high-quality representations directly contribute to reading by supporting automatic retrieval of the lexical knowledge required for effective comprehension. Support for this view is provided by our recent eye-movement evidence, reviewed briefly earlier, showing that skilled readers make more use of parafoveal information than less skilled readers ( Veldre & Andrews, 2014 ; Veldre & Andrews, in press ). Higher quality lexical knowledge supports rapid foveal processing, which in turn allows more efficient extraction of parafoveal information that facilitates processing of subsequently fixated words and enhances the efficiency and effectiveness of oculomotor planning. We are conducting further investigations of individual differences in skilled readers’ eye movements to better understand the interplay between early lexical retrieval and the integrative processes required for comprehension.

Does Orthographic Precision Matter for All Languages?

Orthographic precision may not be equally important for all languages. According to Ziegler and Goswami’s (2005)   psycholinguistic grain size theory, differences in the consistency of the relationship between orthography and phonology determine the grain size of the O-P units that readers need to extract. The drive to develop consolidated representations may therefore be stronger in deep orthographies like English, in which inconsistent O-P correspondences “force the reading system into developing multiple grain size mappings” ( Ziegler & Goswami, 2005 , p. 18) to achieve the precision and redundancy required for automatic word identification. Consistent with the LQH, this pressure toward lexical restructuring is assumed to be word-specific: Words that have many similar neighbors will experience more pressure to reflect multiple O-P correspondences than those from less dense orthographic neighborhoods. The grain size account also assumes that lexical restructuring is an ongoing process that shapes the long-term organization and dynamics of the adult reading system.

Arguing that languages differ in the extent to which they drive the extraction and representation of larger orthographic units does not contradict claims for a universal theory of reading ( Frost, 2012 ). Effective processing of all writing systems depends on learning to efficiently map orthography to phonology. However, the specific strategies and representational structures required to achieve this depend on the covariations between the phonology, orthography, and semantics of the language. Our data show that individual differences among skilled readers provide insight into the lexical structures that emerge for readers of English. Comparisons of individual differences in skilled readers of different writing systems will further contribute to understanding how universal cognitive constraints interact with the characteristics of different languages to shape the cognitive architecture of the skilled reading system.

The analyses reported by Andrews and Lo (2012 , 2013 ) used continuous measures of both the overall proficiency factor and the spelling-meaning factor. To summarize the significant interactions of priming with these individual difference measures, Figures 10.1 , 10.2 , and 10.3 present average priming effects for participants above and below the median on the relevant factor.

Andrews, S. ( 1996 ). Lexical retrieval and selection processes: Effects of transposed-letter confusability.   Journal of Memory and Language, 35, 775–800.

Google Scholar

Andrews, S. ( 2008 ). Lexical expertise and reading skill. In B. H. Ross (Ed.), The psychology of learning and motivation: Advances in research and theory (Vol. 49, pp. 247–281). San Diego, CA: Elsevier.

Google Preview

Andrews, S. ( 2012 ). Individual differences in skilled visual word recognition and reading: The role of lexical quality. In J. Adelman (Ed.), Visual word recognition : Volume. 2. Meaning and context, individuals and development (pp. 151–172). Hove, England: Psychology Press.

Andrews, S. , & Bond, R. ( 2009 ). Lexical expertise and reading skill: Bottom-up and top-down processing of lexical ambiguity.   Reading and Writing: An Interdisciplinary Journal, 22, 687–711.

Andrews, S. , & Hersch, J. ( 2010 ). Lexical precision in skilled readers: Individual differences in masked neighbor priming.   Journal of Experimental Psychology: General, 139, 299–318.

Andrews, S. , & Lo, S. ( 2012 ). Not all skilled readers have cracked the code: Individual differences in masked form priming.   Journal of Experimental Psychology: Learning, Memory, and Cognition, 38, 152–163.

Andrews, S. , & Lo, S. ( 2013 ). Is morphological priming stronger for transparent than opaque words? It depends on individual differences in spelling and vocabulary.   Journal of Memory and Language, 68, 279–296.

Andrews, S., & Scarratt, D. R. ( 1998 ). Rule and analogy mechanisms in reading nonwords: Hough dou peapel rede gnew wirds?   Journal of Experimental Psychology: Human Perception and Performance, 24, 1052–1086.

Andrews, S. , Xia, V. , & Lo, S. (2014, November). Individual differences in masked semantic priming: The role of lexical quality. Paper presented at the meeting of the Psychonomic Society, Long Beach, CA.

Ashby, J. , Rayner, K., & Clifton, C. ( 2005 ). Eye movements of highly skilled and average readers: Differential effects of frequency and predictability.   Quarterly Journal of Experimental Psychology, 58A, 1065–1086.

Baayen, R. H. ( 2008 ). Analyzing linguistic data: A practical introduction to statistics using R . Cambridge, England: Cambridge University Press.

Balota, D. A. , Cortese, M. , Sergent-Marshall, S., & Spieler, D. ( 2004 ). Visual word recognition of single-syllable words.   Journal of Experimental Psychology: General, 133, 283–316.

Balota, D. A., & Spieler, D. ( 1998 ). The utility of item-level analyses in model evaluation.   Psychological Science, 9, 238–240.

Balota, D. A. , Yap, M. , Cortese, M. , Hutchinson, K. , Kessler, B. , Loftis, B. , . . . Treiman, R. ( 2007 ). The English Lexicon Project. Behavior Research Methods, 39, 445–459.

Boyer, N., & Ehri, L. ( 2011 ). Contribution of phonemic segmentation instruction with letters and articulation pictures to word reading and spelling.   Scientific Studies of Reading, 15, 440–470.

Brown, J. I. , Fishco, V. V. , & Hanna, G. ( 1993 ). Nelson-Denny reading test . Rolling Meadows, IL: Riverside.

Byrne, B. , Coventry, W. L. , Olson, R. K. , Samuelsson, S. , Corley, R. , Wilcutt, E. G. , . . . DeFries, J. ( 2009 ). Genetic and environmental influences on aspects of literacy and language in early childhood: Continuity and change from preschool to grade 2.   Journal of Neurolinguistics, 22, 219–236.

Byrne, B. , Olson, R. K. , Hulslander, J. , Wadsworth, S. , Defries, J. C., & Samuelsson, S. ( 2008 ). A behavior-genetic analysis of orthographic learning, spelling and decoding.   Journal of Research in Reading, 31, 8–21.

Carr, T. H. ( 1999 ). Trying to understand dyslexia: Mental chronometry, individual differences, cognitive neuroscience and the impact of instruction as converging sources of evidence. In R. M. Klein & P. A. McMullen (Eds.), Converging methods for understanding reading and dyslexia (pp. 459–491). Cambridge, MA: MIT Press.

Castles, A., Davis, C., Cavalot, P., & Forster, K. ( 2007 ). Tracking the acquisition of orthographic skills in developing readers.   Journal of Experimental Child Psychology, 97, 165–182.

Castles, A., & Nation, K. ( 2008 ). Learning to be a good orthographic reader.   Journal of Research in Reading, 31, 1–7.

Coltheart, M. , Rastle, K. , Perry, C. , Langdon, R. , & Ziegler, J. ( 2001 ). DRC: A dual route cascaded model of visual word recognition and reading aloud.   Psychological Review, 108, 204–256.

Davis, C. J. ( 2010 ). The spatial coding model of visual word identification.   Psychological Review, 117, 713–758.

Dehaene, S. ( 2009 ). Reading in the brain . New York, NY: Penguin Viking.

Dehaene, S., & Cohen, L. ( 2011 ). The unique role of the visual word form area in reading.   Trends in Cognitive Science, 15, 254–262.

Dehaene, S. , Pegado, F. , Braga, L. , Ventura, P. , Filho, G. , Jobert, A. , … Cohen, L. ( 2010 ). How learning to read changes the cortical networks for vision and language.   Science, 3, 1359–1364.

Diependaele, K. , Sandra, D., & Grainger, J. ( 2005 ). Masked cross-modal morphological priming: Unravelling morpho-orthographic and morpho-semantic influences in early word recognition.   Language and Cognitive Processes, 20, 75–114.

Eason, S. H., & Cutting, L. E. ( 2009 ). Examining sources of poor comprehension in older poor readers. In R. K. Wagner , C. Schatschneider , & C. Phythian-Sence (Eds.), Beyond decoding: The behavioral and biological foundations of reading comprehension (pp. 263–283). New York, NY: Guilford.

Ehri, L. ( 2005 ). Development of sight word reading: Phases and findings. In M. J. Snowling & C. Hulme (Eds.), The science of reading (pp. 135–154). Cornwall, England: Blackwell.

Ehri, L. , Nunes, S. , Stahl, S., & Willows, D. ( 2001 ). Systematic phonics instruction helps students learn to read.   Review of Educational Research, 71, 393–447.

Ehri, L., & Wilce, L. ( 1983 ). Development of word identification speed in skilled and less skilled beginning readers.   Journal of Educational Psychology, 75, 3–18.

Feldman, L. B. , O’Connor, P. A., & Moscoso del Prado Martín, F. ( 2009 ). Early morphological processing is morphosemantic and not simply morpho-orthographic: A violation of form-then-meaning accounts of word recognition.   Psychonomic Bulletin & Review, 16, 684–691.

Forster, K. I. ( 1989 ). Basic issues in lexical processing. In W. Marslen-Wilson (Ed.), Lexical representation and process (pp. 75–107). Cambridge, MA: MIT Press.

Forster, K. I. ( 2004 ). Category size effects revisited: Frequency and masked priming effects in semantic categorization.   Brain and Language, 90, 276–286.

Forster, K. I. , & Davis, C. ( 1984 ). Repetition priming and frequency attenuation in lexical access.   Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 680–698.

Forster, K. I. , Davis, C. , Schoknecht, C., & Carter, R. ( 1987 ). Masked priming with graphemically related forms: Repetition or partial activation?   Quarterly Journal of Experimental Psychology, 39, 211–251.

Forster, K. I. , & Taft, M. ( 1994 ). Bodies, antibodies and neighborhood-density effects in masked form priming.   Journal of Experimental Psychology: Learning, Memory, and Cognition, 20, 844–863.

Frith, U. ( 1986 ). A developmental framework for developmental dyslexia.   Annals of Dyslexia, 36, 69–81.

Frost, R. ( 2012 ). Towards a universal model of reading.   Behavioral and Brain Sciences, 35, 263–279.

Gernsbacher, M., & Faust, M. ( 1991 ). The mechanism of suppression: A component of general comprehension skill.   Journal of Experimental Psychology: Learning, Memory, and Cognition, 17, 245–262.

Gomez, P. , Ratcliff, R., & Perea, M. ( 2008 ). The overlap model: A model of letter position coding.   Psychological Review, 115, 577–600.

Gough, P., & Hillinger, M. ( 1980 ). Learning to read: An unnatural act.   Bulletin of the Orton Society, 30, 179–196.

Grainger, J. ( 2008 ). Cracking the orthographic code: An introduction.   Language and Cognitive Processes, 23, 1–35.

Hersch, J., & Andrews, S. ( 2012 ) Lexical expertise and reading skill: Bottom-up and top-down contributions to sentence processing.   Scientific Studies of Reading, 16, 240–262.

Landauer, T. , Kireyev, K., & Panaccione, C. ( 2011 ). Word maturity: A new metric for word knowledge.   Scientific Studies of Reading, 15, 92–108.

Keenan, J. M. , Olson, R. K., & Betjemann, D. ( 2008 ). Assessment and etiology of individual differences in reading comprehension. In R. K. Wagner , C. Schatschneider , & C. Phythian-Sence (Eds.), Beyond decoding: The behavioral and biological foundations of reading comprehension (pp. 227–245). New York, NY: Guilford.

Kessler, B. , & Treiman, R. ( 2001 ). Relationships between sounds and letters in English monosyllables.   Journal of Memory and Language, 44, 592–617.

Kuperman, V. , & Van Dyke, J. ( 2011 ). Effects of individual differences in verbal skills on eye-movement patterns during sentence reading.   Journal of Memory and Language, 65, 42–73.

McClelland, J. , Botvinick, M. , Noelle, D. , Plaut, D. C. , Rogers, T. , Seidenberg, M., & Smith, L. ( 2010 ). Letting structure emerge: Connectionist and dynamical systems approaches to cognition.   Trends in Cognitive Sciences, 14, 348–356.

McClelland, J. L. , & Rumelhart, D. E. ( 1981 ). An interactive activation model of context effects in letter perception: Part 1. An account of basic findings.   Psychological Review, 88, 375–407.

Nation, K. ( 2005 ). Children’s reading comprehension difficulties. In M. Snowling & C. Hulme (Eds.), The science of reading: A handbook (pp. 248–265). Oxford, England: Blackwell.

Nation, K., & Snowling, M. ( 1998 ). Semantic processing and the development of word recognition skills: Evidence from children with reading comprehension difficulties.   Journal of Memory and Language, 39, 85–101.

Norris, D., & Kinoshita, S. ( 2012 ). Reading through a noisy channel: Why there’s nothing special about the perception of orthography.   Psychological Review, 119, 517–545.

Perfetti, C. A. ( 1985 ). Reading ability . New York, NY: Oxford University Press.

Perfetti, C. A. ( 2007 ). Reading ability: Lexical quality to comprehension.   Scientific Studies of Reading, 11, 357–383.

Perfetti, C., & Hart, L. ( 2001 ). The lexical basis of comprehension skill. In D. Gorfein (Ed.), On the consequences of meaning selection: Perspectives on resolving lexical ambiguity (pp. 67–86). Washington, DC: American Psychological Association.

Perry, C. , Ziegler, J., & Zorzi, M. ( 2007 ). Nested modeling and strong inference testing in the development of computational theories: The CDP+ model of reading aloud.   Psychological Review, 27, 301–333.

Perry, J. R. , Lupker, S. J. , & Davis, C. J. ( 2008 ). An evaluation of the interactive-activation model using masked partial-word priming.   Language and Cognitive Processes, 23, 36–68.

Pinker, S. ( 1994 ). The language instinct . Cambridge, MA: Harvard University Press.

Plaut, D. , McClelland, J. L. , Seidenberg, M. S. , & Patterson, K. ( 1996 ). Understanding normal and impaired word reading: Computational principles in quasi-regular domains.   Psychological Review, 103, 56–115.

Pugh, K. R , Mencl, W. E. , Jenner, A. R. , Lee, J. R. , Katz, L. , Frost, S. J ., . . . Shaywitz, B. A. ( 2001 ). Neuroimaging studies of reading development and reading disability.   Learning Disabilities Research and Practice, 16, 240–249.

Quinn, W., & Kinoshita, S. ( 2008 ). Congruence effects in semantic categorization with masked primes with narrow and broad categories.   Journal of Memory and Language, 58, 286–306.

Rastle, K., & Davis, M. H. ( 2008 ). Morphological decomposition based on the analysis of orthography.   Language and Cognitive Processes, 23, 942–971.

Rayner, K. ( 2009 ). Eye movements and attention in reading, scene perception, and visual search.   Quarterly Journal of Experimental Psychology, 62, 1457–1506.

Share, D. L. ( 2004 ). Orthographic learning at a glance: On the time course and developmental onset of self-teaching.   Journal of Experimental Child Psychology, 87, 267–298.

Stanovich, K. E. ( 1992 ). Speculations on the causes and consequences of individual differences in early reading acquisition. In P. B. Gough , L. C. Ehri , & R. Treiman (Eds.), Reading acquisition (pp. 307–342). Hillsdale, NJ: Erlbaum.

Treiman, R., & Kessler, B. ( 2006 ). Spelling as statistical learning: Using consonantal context to spell vowels.   Journal of Educational Psychology, 98, 642–652.

Treiman, R. , Kessler, B., & Bick, S. ( 2002 ). Context sensitivity in the spelling of English vowels.   Journal of Memory and Language, 47, 448–468.

Treiman, R. , Kessler, B., & Bick, S. ( 2003 ). Influence of consonantal context on the pronunciation of vowels.   Cognition, 88, 49–78.

Treiman, R. , Kessler, B. , Zevin, J. , Bick, S., & Davis, M. ( 2006 ). Influence of consonantal context on the reading of vowels: Evidence from children.   Journal of Experimental Child Psychology, 93, 1–24.

Veldre, A. , & Andrews, S. ( 2014 ). Lexical quality and eye movements: Individual differences in the perceptual span of skilled adult readers.   Quarterly Journal of Experimental Psychology, 67, 703–727.

Veldre, A. , & Andrews, S. ( in press ). Parafoveal activation depends on skilled reading proficiency.   Journal of Experimental Psychology: Learning, Memory, and Cognition .

Yap, M. J. , Balota, D. A. , Sibley, D. E., & Ratcliff, R. ( 2012 ). Individual differences in visual word recognition: Insights from the English Lexicon Project.   Journal of Experimental Psychology: Human Perception and Performance, 38, 53–79.

Yap, M. J. , Tse, C.-S. & Balota, D. A. ( 2009 ). Individual differences in the joint effects of semantic priming and word frequency revealed by RT distributional analyses: The role of lexical integrity.   Journal of Memory and Language, 61, 303–325.

Ziegler, J., & Goswami, U. ( 2005 ). Reading acquisition, developmental dyslexia and skilled reading across language: A psycholinguistic grain size theory.   Psychological Review, 131, 3–29.

Ziegler, J. , Rey, A., & Jacobs, A. ( 1998 ) Simulating individual word identification thresholds and errors in the fragmentation task.   Memory & Cognition, 26, 490–501.

  • About Oxford Academic
  • Publish journals with us
  • University press partners
  • What we publish
  • New features  
  • Open access
  • Institutional account management
  • Rights and permissions
  • Get help with access
  • Accessibility
  • Advertising
  • Media enquiries
  • Oxford University Press
  • Oxford Languages
  • University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

  • Copyright © 2024 Oxford University Press
  • Cookie settings
  • Cookie policy
  • Privacy policy
  • Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Measuring Lexical Quality: The Role of Spelling Ability

  • Published: 13 April 2020
  • Volume 52 , pages 2257–2282, ( 2020 )

Cite this article

  • Sally Andrews 1 ,
  • Aaron Veldre 1 &
  • Indako E. Clarke 1  

6073 Accesses

24 Citations

2 Altmetric

Explore all metrics

The construct of ‘lexical quality’ (Perfetti Scientific Studies of Reading 11 , 357–383, 2007 ) is widely invoked in literature on word recognition and reading to refer to a systematic dimension of individual differences that predicts performance in a range of word identification and reading tasks in both developing readers and skilled adult populations. Many different approaches have been used to assess lexical quality, but few have captured the orthographic precision that is central to the construct. This paper describes, evaluates, and disseminates spelling dictation and spelling recognition tests that were developed to provide sensitive measures of the precision component of lexical quality in skilled college student readers – the population that has provided most of the benchmark data for models of word recognition and reading. Analyses are reported for 785 students who completed the spelling tests in conjunction with standardized measures of reading comprehension, vocabulary, and reading speed, of whom 107 also completed author recognition and phonemic decoding tests. Internal consistency analyses showed that both spelling tests were relatively unidimensional and displayed good internal consistency, although the recognition test contained too many easy items. Item-level analyses are included to provide the basis for further refinement of these instruments. The spelling tests were moderately correlated with the other measures of written language proficiency, but factor analyses revealed that they consistently defined a separate component, demonstrating that they tap a dimension of variability that is partially independent of variance in reading comprehension, speed, and vocabulary. These components appear to align with the precision and coherence dimensions of lexical quality.

Similar content being viewed by others

what is lexical hypothesis

Literature reviews as independent studies: guidelines for academic practice

Sascha Kraus, Matthias Breier, … João J. Ferreira

what is lexical hypothesis

Natural Language Processing

what is lexical hypothesis

Reconsidering the Evidence That Systematic Phonics Is More Effective Than Alternative Methods of Reading Instruction

Jeffrey S. Bowers

Avoid common mistakes on your manuscript.

This paper has both practical and theoretical goals. The major practical goal is to validate and disseminate two measures of spelling ability that were developed to discriminate among samples of skilled, native English-speaking university students. Although other measures of spelling ability are available in batteries such as the Wide Range Achievement Tests (Wilkinson & Robertson, 2017 ), they were not specifically developed for university student populations. This may, in part, account for why spelling ability has received little attention in the literature on individual differences among skilled readers relative to measures of word and nonword reading (e.g., Kuperman & Van Dyke, 2011 ), reading speed (e.g., Jackson & McClelland, 1975 ), reading comprehension (e.g., Ashby, Rayner, & Clifton, 2005 ), and vocabulary (e.g., Yap, Balota, Sibley, & Ratcliff, 2012 ). The lack of attention to spelling ability also reflects a common assumption that, at least for skilled readers, measures of word identification and spelling tap the same dimension of individual variability, typically conceptualized as word identification or decoding . While spelling is seen as a useful index of reading development (e.g., Treiman, 2017 ), at higher levels of skill, reading and spelling words are often assumed to be 'two sides of the same coin' (Ehri, 2000 ), suggesting that spelling is unlikely to account for unique variance. More generally, the contribution of individual differences in word identification to explaining variability in reading comprehension is typically assumed to reduce across reading development relative to measures of comprehension-related factors like vocabulary and listening comprehension (e.g., Braze et al., 2016 ).

Challenging these assumptions, a series of studies using the tests of spelling dictation and spelling recognition evaluated in this paper has demonstrated that spelling ability accounts for unique variance in a variety of measures of performance across a range of single-word and sentence-processing tasks within samples of monolingual, English-speaking university students (see Andrews, 2008 , 2012 , 2015 ; Andrews & Veldre, 2019 , for reviews). Most of this research has been conducted in Andrews' laboratory at the University of Sydney, Australia, where the tests were developed, but they are now beginning to be used by independent psycholinguistic research groups (Adelman et al., 2014 ; Emmorey, Midgley, Kohen, Sevcikova Sehyr, & Holcomb, 2017 ; Eskenazi, Swischuk, Folk, & Abraham, 2018 ; Meade, Grainger, Midgley, Emmorey, & Holcomb, 2018 ; Slattery & Yates, 2018 ; Tan & Yap, 2016 ). It is therefore timely to report analyses evaluating the tests' internal consistency and validity and to make them publicly available along with details of the typical administration procedures and norms so that they can be consistently applied by other researchers to investigate individual differences among skilled readers.

The Lexical Quality Hypothesis

The theoretical goal is to evaluate the validity and utility of combining measures of spelling ability with other measures of written language proficiency to assess individual differences in lexical quality among skilled readers. Perfetti ( 1985 ) coined the term 'lexical quality' to refer to a critical determinant of the efficiency and effectiveness of the procedures involved in retrieving linguistic codes during reading comprehension. He subsequently refined the definition to refer to qualities of skilled readers' lexical knowledge (Perfetti & Hart, 2001 ): high-quality representations are "orthographically fully specified", represent redundant word-specific, context-sensitive phonology, and are "semantically more generalized and less context-bound" (Perfetti, 2007 , p. 359). This focus on the causal role of lexical knowledge distinguishes the lexical quality hypothesis (LQH) from many other accounts of individual differences in reading. Rather than attributing reading difficulties to deficits in phonological, semantic, or working memory processes , “the LQH is about knowledge that has not been acquired or practiced to a high-enough level” (Perfetti, 2007 , p. 380) to achieve the properties required for efficient, effective retrieval and subsequent higher-level processing.

The utility of the construct of lexical quality depends on how clearly it is defined. Perfetti ( 2007 ) emphasized the precision and flexibility of lexical knowledge. Precision is identified with the content of lexical knowledge: the specificity and completeness of the orthographic, phonological, grammatical, and semantic constituents of the lexical representation. Orthographic precision is particularly critical, because written forms become the gateway to lexical knowledge (Dehaene, 2009 ). Flexibility arises from the interconnectedness or binding between the different constituents of lexical representations. The precision and redundancy of high-quality representations strengthens the binding between the orthographic, phonological, and semantic constituents that define a word’s identity. These strong connections allow printed word forms to trigger synchronous, coherent activation of all components of the word’s identity required for the higher-order meaning integration processes underpinning effective comprehension (Perfetti, 2007 ). This property of greater coherence between the constituents of a lexical complex may be at least partially independent of the precision of the orthographic representation of a word (Andrews, 2015 ).

According to Perfetti’s definition, lexical quality is a graded, word-specific attribute. The quality of lexical representations varies within individuals as they gradually increase the size of their vocabulary and refine the specificity, redundancy, and interconnectedness of the knowledge stored for existing words through reading experience. There are also differences between individuals in the extent to which they have established high-quality representations for most of the words in their written vocabulary. Such differences could arise from genetic differences in foundational skills such as phonological awareness or orthographic learning (Byrne et al., 2008 ), or from environmental factors such as reading experience and methods of instruction. They may also interact with differences in reading strategy: readers who rely heavily on context to identify words may devote little attention to the details of words’ internal structure and therefore be less likely to develop fully specified orthographic codes for all words (Frith, 1986 ).

Assessing lexical quality

Perfetti's ( 2007 ) construct of lexical quality is increasingly widely invoked in the literature on both developing and skilled readers to refer to a systematic dimension of individual differences that predicts performance in a variety of word identification and reading tasks (e.g., Breadmore & Deacon, 2019 ; Rossi, Martin-Chang, & Ouellette, 2019 ; Slattery & Yates, 2018 ), but a range of different measures have been used to assess it. The most widely used indices of lexical quality are measures of vocabulary and decoding skill. Although word-level measures of decoding continue to predict significant variance in reading comprehension, even in skilled readers (e.g., Landi & Perfetti, 2007 ), vocabulary has been found to account for more variance than decoding at later stages of reading development (e.g., Protopapas, Sideris, Mouzaki, & Simos, 2007 ). In Verhoeven's systematic studies of the contribution of lexical quality to the development of reading comprehension in late-primary-age Dutch children, tests of both the breadth and depth of vocabulary accounted for a substantial proportion of variance in reading comprehension in a large cross-sectional sample, after controlling for decoding measures of word and nonword reading, short-term memory, and nonverbal intelligence (Swart et al., 2017 ). Vocabulary and category fluency also predicted growth in reading comprehension in a longitudinal study over grades 4 to 6 (Nouwens, Groen, Kleeman, & Verhoeven, 2017 ). Measures of 'lexical richness' obtained from tests of synonym knowledge and verbal analogies have also been found to predict eye-movement measures of lexical processing and semantic integration in English-speaking adolescents' sentence reading (Luke, Henderson, & Ferreira, 2015 ).

Individual differences in vocabulary, decoding, and meaning retrieval have also been used to assess lexical quality in adult readers. Many of the investigations conducted by Perfetti's group have used scores derived from factor analyses of a broad battery of tests of reading-related skills administered to large samples of college-student readers to define "functionally distinct dimensions of variability" (Taylor & Perfetti, 2016 ). Early studies (e.g., Perfetti & Hart, 2002 ; Landi & Perfetti, 2007 ) identified separate 'meaning knowledge' (assessed by vocabulary and reading comprehension) and 'form knowledge' factors (defined by spelling, phonology, and decoding) and suggested that form knowledge split into separate orthographic and phonological factors among less-skilled readers (Perfetti & Hart, 2002 ), supporting the LQH's assumption that lexical constituents become more integrated at higher levels of reading skill. Taylor and Perfetti’s ( 2016 ) more recent research using an expanded battery including subjective reports of reading history (Lefly & Pennington, 2000 ) and a measure of print exposure based on author recognition (Stanovich & West, 1989 ) found that the strongest source of shared variance was a 'reading experience' factor defined by measures of reading speed, print exposure, book reading, and reading attitude, which was independent of a robust 'lexical knowledge' factor defined by decoding, word recognition, and spelling skill. Both factors predicted more efficient eye movements – more word skipping and fewer regressions – during reading of short passages, but the lexical factor was a stronger predictor of these reading behaviors for texts containing unfamiliar, recently acquired words. Convergent evidence that word identification predicts individual differences in eye movements in a broader community sample of adults derives from Kuperman and Van Dyke's ( 2011 ) finding that word identification and rapid naming were the only measures from a large battery of tests assessing decoding, working memory, and listening and reading comprehension that significantly predicted unique variance in eye movements during sentence reading. Better performance on these measures was associated with efficient oculomotor control that Kuperman and Van Dyke suggested as being "an affordance of the overall quality of … [the reader's] lexical representations" (p. 56).

The test-battery approach adopted in the studies described above is very resource-intensive, particularly for the large samples that are desirable for individual-differences research. Many researchers have therefore attempted to capture individual differences in lexical proficiency with a single or limited set of measures. Vocabulary has been systematically used as a coarse index of 'lexical integrity' in Yap and colleagues’ thorough investigations of how individual differences modulate both behavioral performance and the underlying processes revealed by mathematical modeling methods across a range of word identification tasks (e.g., Pexman & Yap, 2018 ; Yap et al., 2012 ; Yap, Tse, & Balota, 2009 ). Another relatively widely used single measure that aims to assess the reading experience assumed to underpin lexical quality is the Author Recognition Test (ART). Originally developed by Stanovich and West ( 1989 ), and subsequently refined and extended by Acheson, Wells, and MacDonald ( 2008 ), these quick, easily administered tests use the ability to discriminate between real and fabricated authors (or magazine titles; Acheson et al., 2008 ) as a surrogate measure of the extent of exposure to print (see Moore & Gordon, 2015 , for a review).

The accumulating evidence that a range of measures of lexical proficiency predict systematic variance in skilled readers' performance not only in single-word identification and priming tasks but also in eye-movement measures of sentence reading confirms that individual differences in word-level processes remain significant predictors of variance among adult readers. However, to provide evidence for a specific contribution of lexical quality requires specification of how the predictor variables map to the precision, redundancy, and coherence of lexical representations that define their quality. As highlighted by Braze, Tabor, Shankweiler, and Mencl ( 2007 ), vocabulary knowledge is central to general linguistic comprehension processes that are shared between spoken and written language processing. Correlations between vocabulary and reading performance may, therefore, tap general comprehension processes rather than the quality of reading-specific lexical representations. Consistent with this interpretation, factor analyses typically show that vocabulary loads on the same latent factor as listening comprehension (Braze et al., 2016 ). However, in two independent studies of large community samples of adult readers, vocabulary also accounted for a small, but significant, component of unique variance in reading comprehension when word decoding and listening comprehension were controlled for (Braze et al., 2007 ; Braze et al., 2016 ), suggesting a reading-specific contribution. Braze et al. ( 2016 ) suggested this may be because higher-quality representations "that incorporate subtle gradations of meaning, may integrate more flexibly into representations of discourse of narrative and … be more readily recognized in context" (p. 447). Nevertheless, in studies that only measure vocabulary, it is not clear whether observed effects reflect lexical quality specifically, or factors related to general comprehension – or some combination of the two. Similarly, ART measures of print exposure are typically at least moderately correlated with measures of vocabulary and word identification (e.g., Moore & Gordon, 2015 ). Moreover, as well as contributing to the refinement of lexical representations, the reading experience that these tests are presumed to assess will influence a range of reading processes (Falkauskas & Kuperman, 2015 ).

Orthographic precision and lexical quality

Precision is central to Perfetti's ( 2007 ) definition of lexical quality, but little attention has been paid to establishing the extent to which the measures of individual differences used in studies of skilled reading capture this attribute. Orthographic precision is particularly critical to lexical quality: successful word identification requires readers to extract the relevant features from the perceptual input and map them to existing lexical representations. Phonology shapes the orthographic units that need to be extracted, but neuroimaging evidence suggests that skilled readers develop a specialized visual system for mapping visual input to word-specific knowledge (Dehaene et al., 2010 ). Spelling ability provides a direct index of the orthographic precision of readers' lexical representations of known words that the LQH suggests may play a specific role in predicting effective reading, independently of other measures of lexical proficiency.

Consistent with this view, recent studies of skilled readers that have indexed lexical quality by combining the spelling dictation and recognition tests described in this paper with standardized measures of vocabulary and reading comprehension have demonstrated that spelling ability predicts unique variance both in masked priming studies of single-word identification tasks (Andrews & Hersch, 2010 ; Andrews & Lo, 2012 , 2013 ; Andrews, Lo, & Xia, 2017 ) and in behavioral (Andrews, 2008 ; Hersch & Andrews, 2012 ) and eye-movement indices of sentence reading (Drieghe, Veldre, Fitzsimmons, Ashby, & Andrews, 2019 ; Veldre & Andrews, 2014 , 2015a , 2015b ; Veldre & Andrews, 2016a , 2016b ; Veldre, Drieghe, & Andrews, 2017 ). Independent studies using one or both of the same spelling tests have provided converging evidence for the unique contribution of spelling ability to predicting adult readers’ lexical decision performance (Adelman et al., 2014 ), eye movements during sentence reading (Slattery & Yates, 2018 ), and electrophysiological indices of word processing (Meade et al., 2018 ). A second electrophysiological study found that the selective effect of spelling was more marked in pre-lingually deaf adults than hearing readers (Emmorey et al., 2017 ), suggesting that orthographic precision may play a particularly important role in reading development when phonological processes are compromised.

This accumulated evidence confirms that tests of spelling ability account for individual differences among skilled adult readers that are not captured by other measures of written language proficiency. However, further evidence about the internal consistency and validity of these tests is required to confirm and elaborate the relationship between spelling and lexical quality. To address these issues, the present paper reports analyses of data collated from over 800 individuals tested across nine independent samples of between 46 and 110 University of Sydney students who completed individual-differences tests in conjunction with their participation in eye-tracking studies of sentence reading.

The present research

The effects of spelling ability summarized above were derived from tests of spelling dictation and spelling recognition developed to discriminate among samples of skilled university student readers that have been briefly described in a number of previous papers (e.g., Andrews, 2008 , 2012 , 2015 ), and in greatest detail by Andrews and Hersch ( 2010 ). Spelling production tasks, like dictation, directly test the precision of orthographic knowledge, but some theories of spelling assume that spelling production relies on different information from that required for recognition of correct spellings. Readers may rely on partial orthographic information to identify words (e.g., Frith, 1980 ) and therefore be able to correctly recognize words for which they cannot produce a correct spelling. Performance on spelling recognition tasks is also influenced by the type of misspellings included. Katz and Frost ( 2001 ) found that participants were more likely to accept a repeated misspelling as being correct when it was phonologically plausible, even if it had been correctly rejected on its first presentation, and adopted laxer criteria for judging spelling acceptability when phonologically implausible spellings were included in the recognition list than when all misspellings were phonologically plausible. Such findings demonstrate the role of decision processes in spelling recognition tasks and suggest that participants can be induced to vary the relative weighting of orthographic and phonological information. More extreme differences between production and recognition tasks are suggested by cognitive neuropsychological evidence of dissociations between reading and spelling performance in brain-injured patients that has been interpreted as indicating separate input and output representations that can be independently accessed and damaged (e.g., Ellis, 1993 ). Such views predict that readers may be able to correctly recognize the spelling of words that they cannot accurately produce by relying on their more accurate input representation (Holmes & Babauta, 2005 ).

Thus, spelling recognition tasks may tap factors not captured by spelling dictation. To comprehensively assess spelling ability it is therefore important to include both measures. As reported by Andrews and Hersch ( 2010 ), the spelling dictation and recognition tests both demonstrate high test–retest reliability ( r  = 0.90 and 0.93, respectively), but they have not previously been analyzed for internal consistency. This is the goal of the first set of analyses reported here.

The further goal of the present research is to extend the evidence for the convergent and divergent validity of the spelling tests by assessing their relationships with other measures of written language proficiency across a large sample of participants. Before tackling this question, we addressed an issue relevant to assessing reading comprehension and vocabulary, two of the other major indices of lexical quality. Specifically, we investigated the implications of varying the time limits allowed for the reading comprehension and vocabulary subtests of the Nelson-Denny Reading Test (NDT; Brown, Fishco, & Hanna, 1993 ), a widely used standardized measure of adults’ reading proficiency. This test was developed for use with students from grade 9 of high school through to the fourth year of college/university. Our extensive experience administering the test to samples of university students has revealed that a substantial number of participants complete one or both subtests before the recommended time limit expires. Presumably because of the potential reduction in the tests’ discriminative power, combined with pragmatic constraints on testing time, applications of these tests in Perfetti’s studies of individual differences in lexical quality among skilled readers have typically reduced the time limits for each subtest to half the length required for standard administration (e.g., Perfetti & Hart, 2001 ; Taylor & Perfetti, 2016 ). The same reduced time limits have been used in approximately half of the studies collated for the present research, while the remainder used the standard time limits. This provided an opportunity to evaluate the impact of administration time on the distribution of NDT scores and their relationship to scores on other tests.


Data were collated for 813 students who participated in exchange for credit in introductory Psychology courses. The mean age of the participants was 19.59 years Footnote 1 (SD = 3.65 years), and approximately 71.3% of the sample was female. The recruitment criteria specified that participants spoke English, and began learning to read and write English by no later than age 6. Most participants (87.1%) reported that English was the first language they learned to speak, and English was the first language that virtually all (96.6%) the sample learned to read and write.

The individual-differences data were collected over a period of approximately 5.5 years (late 2011–early 2017) as part of a series of nine eye-movement experiments investigating the role of written language proficiency in sentence reading. Almost all participants completed the two spelling tests (798 participants). In addition, most participants also completed the vocabulary, reading comprehension, and reading rate subsections of the NDT (785 participants). The Author Recognition Test (ART; Moore & Gordon, 2015 ) and the phonemic decoding subtest of the Test of Word Reading Efficiency (TOWRE; Torgesen, Wagner & Rashotte, 1999 ) were completed by 107 participants. The individual-differences battery was administered either before or after the eye-movement experiment, individually or in small groups.

Spelling dictation test

This test consists of 20 words selected from a larger set administered to samples of Australian university students by Burt and Tate ( 2002 ) to cover a broad range of spelling accuracy. In their sample, the words were correctly spelled by between 35% and 92% of participants, and discriminated between above-average and below-average spellers. The experimenter read aloud each word and included it in a short sentence to resolve any ambiguity. Most participants handwrote the word on a response sheet, but one subsample ( n  = 62) completed a computerized version of the test (implemented in Qualtrics but administered in the laboratory) in which they typed the word into a response box. Handwritten responses were scored manually by the experimenter, and typed responses were automatically scored by Qualtrics. Administration of the spelling dictation test typically took between 5 and 7 minutes.

Spelling recognition test

This test consists of 88 items, half correctly spelled and half incorrect. Incorrect spellings were constructed to be phonologically plausible to increase the difficulty of the test and encourage reliance on orthographic knowledge (Katz & Frost, 2001 ). Participants were given unlimited time to select all incorrectly spelled items. They viewed all the items together in four columns and recorded their responses either by circling items on paper or by selecting items in a Qualtrics survey. Responses were manually scored by the experimenter or automatically scored in Qualtrics. Scores on the spelling recognition test are calculated out of a total possible score of 88, subtracting the number of correct spellings selected (i.e., false alarms) and the number of incorrectly spelled items that were not selected (i.e., misses). Participants typically took between 5 and 10 minutes to complete the spelling recognition test.

Nelson-Denny Test (NDT)

All participants completed Form H of this instrument, which includes two subtests. In the 80-item vocabulary test, participants are given a word and asked to select the best-matching word or phrase from five options. The separately timed comprehension subtest includes 38 items relating to seven short passages on a range of topics. It also provides an assessment of reading rate by instructing participants to mark their progress through the first passage after 1 minute has elapsed. A total of 361 of the participants were administered the vocabulary and comprehension sections with standard timing procedures (i.e., 15 minutes for vocabulary and 20 minutes for comprehension); the remaining 345 participants were allowed only half the usual time limit for each subtest. To allow a direct comparison of scores under the two administration procedures, a further sample of 107 participants (referred to as the Full+Half sample) completed the full-timed version of the NDT but marked where they were up to at the halfway point for each section so that both full- and half-timed scores could be computed. This sample also completed the two additional tests described below.

Author Recognition Test (ART)

The Full+Half sample of 107 participants completed this test, which required them to identify which of a list of 100 names they recognized to be authors. The test items were taken from Moore and Gordon ( 2015 ), who removed 15 poorly discriminating author names from the ART scale developed by Acheson et al. ( 2008 ), leaving 50 author names and 50 foils. All names were listed in alphabetical order by surname on a response sheet, and participants circled the names they recognized as authors. Participants were instructed not to guess, because they would be penalized for incorrect responses. Administration of the ART took between 3 and 5 minutes. Scores were computed by subtracting the number of false alarms to foil names from the number of correctly selected authors, the standard method of scoring the ART (Acheson et al., 2008 ). Moore and Gordon ( 2015 ) found that this scoring method yielded slightly higher correlations with measures of reading behavior than a measure of hit rate alone.

Phonemic decoding

The Full+Half subsample also completed the phonemic decoding subtest of the TOWRE (Torgesen et al., 1999 ), which consists of a list of 63 nonwords ranging from one to four syllables. After completing a practice list of eight items, participants were given 45 seconds to read aloud as many nonwords as they could. Responses were recorded by a microphone, and the number of correctly pronounced items was checked offline by a research assistant. Pronunciations were deemed correct if they applied plausible grapheme-phoneme correspondences for the complete nonword string. Partial scores were not applied.

Results and discussion

Descriptive statistics.

Table 1 presents the descriptive statistics for each measure – spelling dictation, spelling recognition, vocabulary, reading comprehension, reading rate, author recognition, and phonemic decoding efficiency – separately for the three subsamples administered full-timed and half-timed versions of the NDT.

As expected for a selected sample of predominantly native English-speaking university students, the average level of reading performance was relatively high: the means for the NDT vocabulary and comprehension tests administered with standard timing corresponded to the 74th and 75th percentiles, respectively, of the norms for US students in the first year of a 4-year college program, while the average reading rate corresponded to approximately the 60th percentile of that cohort. The mean and standard deviation of the ART scores were very similar to those obtained by Moore and Gordon ( 2015 ) for the 50-item test (M = 13.75; SD = 6.81), while the average phonemic decoding score was above the 90th percentile for 6th grade children – the oldest members of the Australian normative sample (Marinus, Kohnen, & McArthur, 2013 ). Despite the high average performance, the wide score ranges indicate that the samples showed considerable variability on all measures in the battery.

Unsurprisingly, the NDT vocabulary and comprehension scores were lower for the half-timed version, but vocabulary scores were relatively less affected by the reduced time limit than comprehension. The reasons that the two subtests differ in sensitivity to speed pressure and the implications for their discriminative power are explored below. The manipulation of administration time did not affect the NDT reading rate measure, because it is based on only the first passage of the comprehension test. The similarity between the mean reading speeds of the three subsamples provides reassurance that they are of relatively equivalent reading proficiency.

Internal consistency of spelling tests

To address our first aim of assessing the psychometric properties of the two novel tests of spelling ability, analyses of the internal consistency of each test were conducted.

Spelling dictation

Factor analysis of the 20 items from the spelling dictation test using maximum likelihood estimation to test for unidimensionality showed a clear break in the scree plot after the first factor (from an eigenvalue of 4.56 to 1.21), which accounted for 18.88% of the variance. Footnote 2 Factor loadings ( λ ) and extracted communalities ( h 2 ), which represent the proportion of the item’s variance accounted for by this factor, are shown in Table A1 with descriptive statistics for each item (see Appendix for full test and norms). All but three items ( warranty , asymmetry , diligent ) had factor loadings greater than 0.30, indicating adequate prediction of the item from the latent construct assessed by the test as a whole. Internal consistency indexed by Cronbach’s alpha of 0.814 is in the range classified as ‘good’, and the mean inter-item correlation of 0.18 falls at the lower end of the range taken to indicate test homogeneity (Clark & Watson, 1995 ). Further evidence of homogeneity is provided by the point biserial correlations ( r pb ; see Table A1 ), which index whether the item discriminates between individuals in the same way as the total test. All items had positive point biserial correlation coefficients > 0.30 (65% above 0.40), indicating that the items consistently predicted the total test score.

To provide further insight into the psychometric properties of the test, Rasch analysis of the 20 items was conducted using RUMM2030, yielding the Rasch item difficulty indices (in logits) shown in Table A1 for each item, and summarized in the person-item map in Fig. 1 . Rasch item difficulty indices estimate the level of person ability at which an item has a 50% chance of being correctly/incorrectly endorsed. Higher Rasch difficulty values indicate more difficult items (i.e., a greater level of ability required). The most difficult item ( conciliatory ), which yielded a Rasch index of 2.45, was correctly spelled by less than 20% of the sample, while over 88% correctly spelled the easiest item ( euphoric , Rasch index of −1.94). Rasch analysis also allows items and individuals to be measured on the common scale of Rasch logit units depicted in the person-item map (Fig. 1 ), which plots the participants by ability against the test items by difficulty. In general, the distribution of item difficulty aligns with the distribution of person ability; however, the range of the ability distribution is wider than that of the item difficulty distribution. This indicates that the test items may not effectively capture individual differences in spelling ability at the lowest and highest levels. Footnote 3

figure 1

Plot of participants (’PERSONS’) by performance on the Spelling Dictation Test against the 20 test items (’ITEMS’) by Rasch difficulty (in logit units). Better test performance and more difficult items appear at the top of the figure

Spelling recognition

Four of the 88 items ( appreciate , distinguish , exhibition , annual ) of the spelling recognition test were identified as correctly spelled by all participants and so were excluded from item-level analyses. The remaining 84 items were factor analyzed using maximum likelihood estimation to check for unidimensionality. Examination of the scree plot showed a clear break after the first factor, which accounted for 8.83% of the variance. Factor loadings ( λ ) and extracted communalities ( h 2 ) are shown in Table A3 , with descriptive statistics for each item (see Appendix for complete test and norms). Overall internal consistency was ‘good’ (Cronbach’s alpha = 0.859). However, the factor loadings for some items were marginally negative (e.g., sufficient , elementary , inhibition ), and many items had low communalities (e.g., attitude , consequence , parallel ), indicating inadequate prediction of the item from the latent construct of spelling recognition ability. The mean inter-item correlation was 0.06, indicating that the test is likely not homogeneous. The point biserial correlations ( r pb ) for many items were also negative or near-zero (e.g. senior , fulcrum , guitar ), indicating poor discrimination of spelling ability. These limitations arise, at least in part, because the test contains too many easy items. Although the percentage of participants correctly identifying each item as either a correct or incorrect spelling ranged from 26.6% to 100%, more than half of the items were correctly classified by over 90% of participants, limiting its sensitivity.

This problem is evident in the person-item map for the spelling recognition test presented in Fig. 2 . The Rasch item difficulty indices (in logits) reveal considerable variability among items: consequence was the easiest, with a Rasch score of −3.55, and vigilant , at 3.71, was the most difficult. Participants were also relatively normally distributed on the Rasch difficulty scale, but the distributions of ability and item difficulty were not aligned. Many items fell below the lowest level of spelling ability and therefore made little contribution to discrimination between people. It is noteworthy that the majority of the items in the easier half of the difficulty range are correct spellings (e.g., 37 of the 41 items with Rasch values below 0), suggesting a bias to classify items as correctly spelled, i.e., participants were much more likely to fail to identify an incorrect spelling than to falsely classify a correct spelling as incorrect.

figure 2

Plot of participants (’PERSONS’) by test performance on the Spelling Recognition Test against 84 test items (’ITEMS’) by Rasch difficulty (in logit units). Better test performance and more difficult items appear at the top of the figure. Items in italicized font are incorrect spellings

Thus, the results of item-level analyses for the spelling recognition test clearly indicate that there is room for psychometric improvement. This might be achieved by removing items with low communalities, negative factor loadings, and/or negative and near-zero point biserial correlations, but, as elaborated in the Discussion, attention will need to be paid to the distribution of correct and incorrect spelling and response requirements of the recognition task. The present set of items may also be more effective in discriminating between individuals in samples with lower levels of written language proficiency than those included in our studies. Even though a large proportion of the items yielded very high average accuracy, the Rasch item distribution for the easy items is graded, and only four items were correctly classified by all participants. Importantly, even in our restricted sample, scores on the recognition test were highly correlated with spelling dictation, and the factor analyses reported below show that they tapped the same dimension of variability between people.

Effects of administration time on vocabulary and comprehension performance

To investigate our second research question of how differences in administration time influence the estimates of vocabulary and reading comprehension obtained from the NDT, scores were collated separately for participants using the standard ‘full-timed’ procedures ( n  = 361), those tested using ‘half-timed’ procedures ( n  = 345), and the Full+Half NDT sample ( n  = 107) for whom both full-timed and half-timed measures of vocabulary and reading comprehension were obtained. As illustrated in Fig. 3 , the distribution of scores obtained in both tests was highly negatively skewed under the standard full-timed conditions (vocabulary: skew = −0.97; comprehension: skew = −1.54), reducing discrimination at the upper end of the score distribution, but substantially more normal with the half-timed limit conditions (vocabulary: skew = −0.11; comprehension: skew = 0.46). The half-timed procedure therefore increased discrimination among more proficient readers. Similar differences in the distribution of full-timed and half-timed scores were evident in the Full+Half NDT sample. To determine whether the differential discrimination of the half-timed procedure influenced the relationship of vocabulary and comprehension to other measures, the analyses of convergent and divergent validity reported below were conducted separately on the full-timed and half-timed samples.

figure 3

Frequency distributions of scores on the Vocabulary (upper panels) and Comprehension (lower panels) subtests of the Nelson-Denny Reading Test for participants tested under full-timed versus half-timed administration conditions


Table 2 shows the correlations among spelling and reading measures for the samples that completed the NDT under full- and half-timed administration. All measures were moderately to strongly positively correlated, although the relationship of reading rate with the other measures tended to be weaker than those between spelling, vocabulary, and comprehension. The two spelling tests were very highly correlated in both the full-timed and half-timed NDT subsamples, and they showed almost identical strong correlations between vocabulary and comprehension, and similar moderate relationships between these measures and the two spelling tests. The most substantial difference between the samples was a significantly higher correlation between NDT comprehension and reading rate in the half-timed ( r  = 0.51) than the full-timed sample ( r  = 0.30), z  = 3.018, p  = 0.003, suggesting that the relationship between reading speed and comprehension is stronger under time constraints. However, in general, the simple correlations indicate that the half-timed version of the NDT captures similar dimensions of individual differences among skilled readers as the standard full-timed version.

Further support for this conclusion derives from the similar patterns of correlations observed for the full- and half-timed measures in the Full+Half subsample (see Table 3 ). The data for this sample also showed that the additional ART and phonemic decoding scores were only weakly related to each other, but both were at least moderately correlated with all of the measures of spelling and reading proficiency.

Principal component analysis

To evaluate whether these positively correlated measures of linguistic processing can be reduced to a smaller number of independent dimensions, principal component analyses with promax rotation were conducted. Footnote 4 Each analysis was conducted for both the full-timed and half-timed NDT subsamples, excluding the Full+Half subsample to determine whether administration time changed the dimensional structure. Two components were extracted in each analysis.

The first set of analyses was conducted on the spelling dictation, spelling recognition, vocabulary, and comprehension scores that have been used in our previous published investigations of individual differences in masked priming and sentence reading. As summarized in Table 4 , for both full-timed and half-timed subsamples, the two moderately correlated components accounted for around 85% of variance, with the spelling tests forming Component 1, and NDT vocabulary and comprehension forming Component 2. A second set of analyses added reading rate (rightmost columns of Table 4 ) to evaluate whether it moderated the impact of administration time. Again, both full-timed and half-timed subsamples yielded two similar, moderately correlated components on which the two spelling tests formed Component 1. In the full-timed subsample, vocabulary, comprehension, and reading rate again formed Component 2. However, for the half-timed subsample, comprehension and reading rate loaded selectively on Component 2 but vocabulary loaded equally on both components. There was no evidence of a similar cross-component contribution of vocabulary when reading rate was not included as a predictor, suggesting that reading rate captures variance shared with vocabulary that particularly affects performance in half-timed conditions.

A third set of principal component analyses were conducted on the Full+Half subsample. Separate analyses were conducted using full-timed and half-timed measures of vocabulary and comprehension to evaluate the similarities and differences in the structure observed for samples administered with the independent full- and half-timed measures. These analyses also included ART and phonemic decoding scores to confirm and refine understanding of the dimensions of individual differences. Two components were extracted using a promax rotation. Component loadings are shown in Table 5 .

Paralleling the results for the independent samples, analyses including both the full- and half-timed scores yielded moderately correlated components that separated the spelling tests from the other measures: the two spelling tests fell on Component 2, and the NDT measures of vocabulary, comprehension, and reading rate on Component 1. Footnote 5 The two additional individual-differences measures collected for this sample provided useful evidence of convergent and divergent validity because they loaded on different components. The ART index of print exposure loaded on Component 1 with the three NDT measures, while the phonemic decoding score loaded on Component 2 with the two spelling tests. As well as providing converging evidence for the two components of reading proficiency identified in the principal component analyses, the differential loadings of these two measures also provide useful additional information about the nature of the dimensions they assess. The relationship between phonemic decoding and spelling ability is consistent with the view that orthographic precision depends on the amalgamation of orthography and phonology (Ehri, 2015 ; Perfetti, 2007 ), and that this dimension is partially independent of the higher-order linguistic knowledge and skills captured by tests of vocabulary, comprehension, and reading speed, and the broad reading experience tapped by the ART.

General discussion

The central aims of the present research were to assess the internal consistency and validity of two recently developed tests of spelling ability that have been used in several investigations of individual differences among skilled readers and to evaluate whether they assessed a dimension of variability that is not effectively captured by other widely used tests of written language proficiency.

The results confirmed previous evidence that the tests of spelling dictation and spelling recognition were highly correlated across the entire sample of close to 800 skilled readers, and in the three independent sub-samples of 107–361 participants ( r  = 0.75–0.82) defined by the different administration timing conditions of the NDT. There is therefore little evidence that dictation and recognition tests tap independent input and output representations. Rather, spelling appears to be a relatively unitary ability in skilled adult readers, but, as in other domains, production tests like dictation are more difficult than recognition tests. However, the failure to demonstrate independent dimensions of spelling ability may reflect psychometric limitations of the test of spelling recognition discussed below.

Factor analyses showed that the items in each test had good internal consistency (Cronbach’s alpha > 0.8) and assessed a relatively homogeneous dimension of individual variance. Rasch analysis revealed that the item and ability distributions for the spelling dictation test were relatively well aligned, although the items failed to tap either the upper or lower extreme of the ability distribution. The spelling recognition test was more poorly calibrated. The distributions of ability scores for both participants and items were relatively normal, but the test contained too many easy items. The analyses also revealed a strong bias to withhold ‘incorrect’ responses – yielding high accuracy for correct spellings at the expense of a high miss rate for incorrect spellings. This conservative strategy may reflect the response requirement of the present recognition task, i.e., to check incorrect spellings. The opposite instruction – to endorse correct spellings – may yield the reverse bias: participants may be cautious about endorsing spellings as correct and therefore show a high miss rate to avoid making false alarms to incorrect spellings. Such biases might be reduced by telling participants that 50% of the items are correctly spelled. However, a more effective strategy for developing a psychometrically sound spelling recognition test may be to use a forced-choice format in which two alternative spellings of a single item are simultaneously presented to avoid the inherent problems of yes/no procedures.

Both spelling tests could undoubtedly be psychometrically improved to yield better ability-item alignment by adding more difficult items, particularly to the recognition test. However, care would be required to ensure that increasing discrimination did not reduce the specificity of the tests. Difficulty in producing or recognizing a correctly spelled word may reflect low vocabulary or reading experience rather than a specific limitation in the precision of orthographic knowledge for known words. Consistent with this possibility, there were small, significant correlations between Rasch difficulty and log frequency estimates from both the SUBTLEX and HAL corpora for both correctly ( r  = –.24 and −.30, respectively) and incorrectly spelled words ( r  = –.40; –0.37). Adding further items that are sufficiently difficult to be discriminating may therefore risk confounding spelling ability with vocabulary. As discussed in more detail in the next section, the importance of distinguishing the effects of spelling and vocabulary depends on the research goals.

The relatively poor discriminative power of the spelling recognition test may also, in part, reflect the relatively elite population of participants our samples were drawn from: monolingual tertiary students from a metropolitan, research-intensive, high-entry university. The test may yield more effective discrimination between individuals in more diverse samples of adult readers. Nevertheless, even within our restricted samples, it demonstrated strong convergent validity with the spelling dictation test. The two measures were highly correlated, and they defined a separate principal component in the three independent samples of participants, regardless of differences in the administration time of the NDT vocabulary and comprehension tests. Thus, despite its poorer discrimination, specific variance in performance on the recognition test was clearly shared with the dictation test rather than the other measures of written language proficiency.

The inclusion of the additional measures of ART and phonemic decoding in the Full+Half sample contributed to establishing the convergent and divergent validity of the two components. Both measures showed at least moderate simple correlations with the tests defining both factors, despite their low correlation with each other. Confirming that they tapped different dimensions of individual difference, they loaded on different components in analyses using both full- and half-timed NDT measures. Their differential loadings are conceptually consistent with the lexical quality hypothesis. The broad index of reading experience tapped by the ART loaded with other measures of higher-level reading processes, while phonemic decoding ability selectively loaded with the two tests of spelling ability. This evidence that phonemic decoding is specifically related to spelling ability is consistent with the view that knowledge of orthographic-phonological correspondences is a critical foundation for the development of precise orthographic knowledge (Ehri, 2015 ); Nation, 2017 ). The foundational role of the alphabetic principle and phonological decoding in learning to read English has been well established in developmental populations (e.g., Byrne, 1998 ; Share, 1995 ) and has been argued to provide “the means for orthographic learning – the gradual accumulation of orthographic knowledge, via reading experience” (Nation, 2017 , p. 3). The present evidence that phonemic decoding is specifically related to spelling ability even in skilled adult readers provides strong support for such claims.

The consistency of the component structure regardless of whether vocabulary and comprehension were assessed in time-pressured conditions also suggests that the separable dimensions of individual differences exert a stronger influence than the method variance associated with strategic adjustments of speed-accuracy in response to different timing constraints. As summarized in Fig. 3 , reducing the time limits for these two tests had a dramatic influence on the distribution of scores for both subtests of the NDT. Such differences would be expected to influence the discriminative capacity of the tests and compromise their sensitivity to different sources of variability. However, the overall proportion of variance accounted for, and the component structure, was very consistent across variations in administration time. The only observed change in component structure was due to the inclusion of NDT reading rate in the half-timed sample. In this analysis, NDT vocabulary showed similar moderate loadings on both principal components, rather than loading selectively with the higher-level processing measures as it did in all other analyses. No such shift was evident in the analysis using half-timed scores for the Full+Half sample, which was tested under the same longer time limits as the independent full-timed sample but instructed participants to mark where they were up to at the half-timed limit. This sample was therefore not subject to the same time pressure as the half-timed sample. The contribution of vocabulary to the ‘precision’ component therefore appears to be associated with strategies that are enhanced under conditions of speed pressure.

Thus, with respect to the practical goals of this research, the analyses demonstrate that the two tests of spelling ability provide relatively internally consistent, converging measures of a source of individual differences among skilled readers that is partially independent of the more typically used measures of vocabulary, reading comprehension, and speed. Despite some psychometric limitations, these instruments have already yielded new empirical insights, briefly elaborated in the next section, that demonstrate their utility in research on individual differences in reading.

One practical issue that may warrant further investigation concerns the value of assessing both spelling dictation and spelling recognition. The high correlation between the two tests provides little support for the view that spelling production and recognition tap different representations or processes (e.g., Ellis, 1993 ). Nevertheless, the use of two converging measures that involve different encoding and response requirements may contribute to more effectively isolating variance specifically associated with the spelling ability. If a single test was selected, spelling dictation would certainly be favored on psychometric grounds. Assessing production of the complete orthographic form also provides a more conceptually valid index of precise orthographic knowledge and yields errors that can potentially be analyzed to diagnose the source of spelling problems. However, the informal experience we have gained by making the tests available to other research groups is that the spelling recognition test is preferred, presumably because of both its ease of administration and greater acceptability to participants: spelling dictation tests often evoke performance anxiety even in skilled readers. It may be possible to develop a more psychometrically sound spelling recognition test by using a forced-choice rather than yes/no format to reduce the biases of recognition tasks, and selecting distractor items that systematically manipulate different dimensions of similarity to the target (e.g., orthographic, phonological, morphological). Such an approach may allow construction of a recognition test that achieves the same level of discrimination and diagnostic capacity as spelling dictation. The item analyses reported here provide the foundation for such developments.

The role of spelling in assessing lexical quality

The consistent evidence that tests of spelling ability tap a partially independent dimension of variability in reading skill aligns with the precision component of Perfetti’s ( 2007 ) construct of lexical quality. It also confirms Andrews’ ( 2008 , 2012 , 2015 ) proposal that measures of spelling ability can be used to tap lexical precision and account for unique variance that is not captured by the measures of reading comprehension and vocabulary more typically used to assess individual differences among skilled readers.

Lexical quality is defined by both the precision of orthographic representations and the coherent, synchronous activation of their associated phonological and semantic codes (Andrews, 2015 ; Perfetti, 2007 ). The second component consistently identified in the principal component analyses was defined by measures of higher-level knowledge and processes: the semantic knowledge indexed by vocabulary, text-level measures of reading comprehension and speed, and the ART index of reading experience. It therefore appears to tap the processes associated with efficient, synchronous retrieval and use of the complex of lexical codes associated with a specific word form – processes related to the ‘coherence’ dimension of lexical quality (Perfetti, 2007 ). This is consistent with evidence that individuals with higher scores on the coherence than the precision component show stronger semantic influences of masked primes in both lexical decision (Andrews & Lo, 2013 ) and semantic categorization tasks (Andrews et al., 2017 ). The present evidence that NDT vocabulary loaded equally on both the precision and coherence factors for the half-timed sample only when reading rate was included as a predictor suggests that knowledge of word meanings combines with high orthographic precision to facilitate rapid, coherent retrieval of the complex of lexical codes that define a word, as well as contributing to higher-level word integration processes. It also implies that the contribution of the coherence between orthographic and semantic codes to early lexical retrieval may be more effectively tapped when vocabulary is assessed under speed pressure.

The multidimensional nature of lexical quality revealed by these analyses highlights the value of including multiple measures of lexical proficiency in studies designed to investigate individual differences among highly skilled readers. As confirmed by the present data, measures of written language proficiency are generally at least moderately inter-correlated, so a single reliable measure can potentially tap variability in the broad dimension of lexical proficiency. For example, Yap’s systematic investigations of the role of vocabulary in predicting individual differences across a range of tasks including single-word lexical decision (Yap et al., 2012 ) and semantic categorization tasks (Pexman & Yap, 2018 ) and studies of unmasked (Yap et al., 2009 ) and masked priming (Tan & Yap, 2016 ) have been interpreted as demonstrating the role of the “integrity” of readers’ lexical representations in facilitating automatic lexical retrieval. However, vocabulary is also correlated with measures of higher-order linguistic processing such as listening comprehension (Braze et al., 2016 ). Similarly, in the present analyses, vocabulary most consistently loaded with the text-based measures of reading comprehension and reading rate rather than with spelling ability. The consistency of the modulating effects of vocabulary across a range of word-identification tasks may reflect the fact that it contributes to both the precision and coherence components of lexical quality, at least when it is assessed under speeded testing conditions. Vocabulary tests may therefore be a useful index of the broad construct of lexical quality when researchers are limited to a single measure.

However, studies using a broader test battery that allows different facets of lexical quality to be separated provide additional insights into how lexical quality contributes to effective reading that are obscured by more global measures. Investigations of masked orthographic priming of lexical-decision performance have shown that spelling selectively predicts competition from similar words on both behavioral performance (Andrews & Hersch, 2010 ; Andrews & Lo, 2012 ) and the N400 component of the event-related potential waveform (Meade et al., 2018 ). An ‘orthographic profile’ of relatively higher spelling than vocabulary is also associated with form-based morphological decomposition indexed by equivalently strong masked priming for genuine (e.g., hunter-HUNT ) and pseudo-morphemically related ( corner-CORN) prime-target pairs (Andrews & Lo, 2013 ). In contrast, the opposite ‘semantic profile’ of higher vocabulary than spelling ability predicts strong masked priming for genuinely morphological pairs (Andrews & Lo, 2013 ) and stronger semantic congruence priming in a semantic categorization task (Andrews et al., 2017 ).

Isolating the different dimensions of lexical quality has also yielded new insights into the processes underlying skilled readers’ eye movements during sentence reading. Reading comprehension is typically the best predictor of the overall speed and efficiency of sentence reading, consistently predicting shorter fixation durations and sentence/passage reading times, but spelling ability selectively predicts measures related to eye-movement control such as saccade length (Veldre & Andrews, 2014 ) and word skipping (Slattery & Yates, 2018 ; Veldre & Andrews, 2016b ; Veldre et al., 2017 ). The combination of spelling and reading comprehension is also a stronger predictor of the extent and depth of parafoveal processing than either measure alone (Veldre & Andrews, 2015a , 2015b ). However, the two measures have counteracting effects on semantic processing of parafoveal previews: higher reading comprehension was associated with stronger preview benefit from a semantically related or contextually plausible word, whereas better spellers showed reduced semantic/plausibility preview benefit relative to poorer spellers (Veldre & Andrews, 2016a , 2016b ; see Andrews & Veldre, 2019 , for a review). These data converge with the masked morphological and semantic priming results in suggesting that high spelling ability supports rapid retrieval of a word’s orthographic form. In the masked orthographic priming paradigm, this yields competition from word primes that are orthographically similar to the target. This competition eliminates the benefits of sublexical overlap between the prime and target which is responsible for priming in poor spellers. In studies of parafoveal preview, the conflict between the orthographic form of the preview and the target eliminates the benefits of the preview’s contextual acceptability – which is the source of semantic preview benefit in poorer spellers. In both cases, the detrimental effects of high orthographic precision are due to the conflicting perceptual input introduced by the masked prime, or parafoveal preview. In normal reading, where such misleading information is not presented, rapid retrieval of either orthographic or semantic features of a briefly presented, or parafoveal, word will facilitate lexical retrieval and enhance the efficiency of reading.

This evidence of systematic individual differences among skilled readers reviewed above contributes to resolving some of the contradictory findings obtained in typical analyses of the averaged data for samples of skilled readers that have sustained ongoing debates about visual word recognition and reading (Andrews, 2012 ). The evidence that spelling ability selectively modulates lexical competition suggests that efforts to ‘crack the orthographic code’ (emphasis added) for reading (e.g., Grainger, 2008 ) may be misguided. Similarly, the differential effects of vocabulary and spelling on morphological and semantic priming suggest that there may not be a single answer to the debate between form-first versus cascaded activation accounts of semantic retrieval (e.g., Forster, 2013 ; Rodd, 2004 ). Individual differences also influence the interactions between foveal and parafoveal processing that provide a critical source of evidence for distinguishing between serial and parallel models of eye movement control (e.g., Engbert, Nuthmann, Richter, & Kliegl, 2005 ; Reichle, Pollatsek, Fisher, & Rayner, 1998 ; Snell, van Liepsig, Grainger, & Meeter, 2018 ).

It is important to emphasize that the different components of lexical quality are not entirely independent. Across each of our samples, the correlation between the two rotated components identified in our factor analyses was consistently around 0.5. In unrotated factor solutions, all the NDT and spelling tests loaded on the first common factor, which accounted for 54% and 56% of variance in the full- and half-timed samples, respectively; while the discrepancies between spelling and NDT measures emerged on the second component, which captured an additional 19–20% of variance. Thus, the primary dimension of individual differences tapped by our test battery is common to spelling, vocabulary, and reading comprehension – an index we have referred to as ‘overall proficiency’ (e.g., Andrews, 2015 ; Andrews & Veldre, 2019 ). However, when this shared variance is partialed out, a second, weaker dimension of variance is revealed that is defined by discrepancies between individuals’ level of spelling and vocabulary/comprehension ability.

The fact that spelling taps a common dimension of variance that is shared with vocabulary and comprehension is not surprising. Knowledge of the meanings and the spellings of words both depend critically on experience reading (and writing). Although spoken language experience clearly makes an important contribution to vocabulary acquisition, the majority of children’s vocabulary learning occurs through reading (Nagy, Herman, & Anderson, 1985 ), and print exposure is clearly essential for spelling in both children and adults (e.g., Mol & Bus, 2011 ) – particularly in the quasi-regular writing system of English. Footnote 6 Nation ( 2017 ) argues that repeated exposure to words across a diversity of contexts, episodes, and experiences supports the development of a rich interconnected database and leads to “local variation at the word level: a ‘lexical legacy’ that is measurable during word reading behavior” (p. 1). Vocabulary and spelling knowledge both depend on the capacity to extract invariant patterns from text and establish connections both between the different components of a lexical form (orthography, phonology, and semantics) and between related linguistic units. Individual differences in these capacities, and their consequences for the quality of the lexical database that is acquired through reading experience, may account for the shared variance in spelling, vocabulary, and reading. However, the evidence of a second dimension of individual variability that is tapped by the discrepancy between orthographic and semantic knowledge implicates an additional source of variation in either the knowledge extracted from reading experience, or the manner in which it is applied to word identification and reading. Further systematic investigations of orthographic learning in developing readers (e.g., Nation, 2017 ; Joesph, Wonnacot, Forbes, & Nation, 2014 ; Tamura, Castles, & Nation, 2017 ) and computational modeling of this development (e.g., Ziegler, Perry, & Zorzi, 2014 ) will contribute to determining the individual, instructional, and contextual factors responsible for these differences.


The spelling dictation test and spelling recognition test described here each take only approximately 5 minutes to administer but capture unique variance among the population of skilled, college-aged readers that has been shown to modulate effects in masked priming and sentence reading tasks. Both spelling tests were unidimensional, displayed good internal consistency, correlated with other measures of reading ability, and formed a distinct component across a large college-aged sample. The item-level analyses reported in the present paper provide a basis for further refinement of these instruments, particularly the spelling recognition test, to better discriminate among participants at the highest levels of proficiency, but the present versions may discriminate more effectively in less-skilled adult samples. The results also demonstrate that using shorter time limits for standard tests of vocabulary and comprehension enhances discrimination among skilled readers without substantially changing the relationships between different components of written language proficiency. These practical contributions to measuring lexical quality will support the development of a richer body of empirical evidence about how individual differences modulate skilled word recognition and reading. Incorporating such variation into models of these processes is widely acknowledged to be a critical step for future theoretical development in understanding these essential educational and vocational skills (e.g., Andrews & Reichle, 2019 ; Radach & Kennedy, 2013 ; Rayner, Abbott, & Plummer, 2015 ).

Demographic data were not collected from all participants. The summary statistics reported here are based on between 466 and 609 participants, but are representative of all samples tested.

Confirming that a single-factor solution captured the majority of predictable item variance, a second factor analysis that extracted two factors revealed that they accounted for 19.04% and 2.42% of variance, respectively, and were highly correlated (r = 0.678).

To evaluate whether these limitations in discrimination were due to categorical scoring of spellings as correct/incorrect, the dictation responses for a subset of 125 participants were rescored by calculating the Levenshtein distance between the correct spelling and the participant’s response using the vwr package (Keuleers, 2013 ) in R (R Core Team, 2019 ). On average, participants’ spellings were very similar to the correct spelling – the mean edit distance was less than one letter (M = 0.66, SD = 0.46). Levenshtein-based scores of spelling accuracy were also highly correlated with the categorical dictation scores by both subjects ( r  = −0.78) and items ( r  = −0.95), and with the Rasch difficulty index based on categorical scores ( r  = −0.97). The Levenshtein scores also showed similar, albeit slightly weaker, relationships as the categorical scores, with variability in NDT vocabulary (−0.42 vs. 0.51), comprehension (−0.36 vs. 0.46), and reading rate (−0.29 vs. 0.37), and spelling recognition (−0.58 vs. 0.80). Thus, at least for this set of items, the greater precision offered by Levenshtein distance measures does not appear to improve discrimination among skilled readers, presumably because their spelling errors are generally limited to specific ambiguous phoneme-grapheme correspondences.

All principal component analyses were also conducted with no rotation. Component structure prior to rotation also yielded two components in which the first centroid had high loadings from all variables, and the second differentiated the spelling tests from the other measures.

The same two-factor structure was observed in analyses including the smaller subset of predictors available for the full- and half-timed samples (see Appendix Tables A5 and A6 ). NDT vocabulary loaded selectively on the same component as NDT comprehension both when NDT reading rate was and was not included as a predictor.

An issue that is beyond the scope of this paper, but warrants further research, is whether the independent effects of spelling ability on reading behavior generalize to languages other than English. The construction of precise word-specific orthographic representations may be a specific response to the notoriously inconsistent and idiosyncratic spelling-sound correspondences of English which have been demonstrated to be more complex than any other alphabetic orthography (Share, 2008 ). This complexity may drive extraction of multi-letter units at a range of different ‘grain-sizes’ (Ziegler & Goswami, 2005) to capture systematic, higher-order consistencies in the mapping between orthography and phonology (e.g., Kessler & Treiman, 2001 ). If so, the independent contributions of spelling ability to predicting reading behavior may be limited to ‘orthographically deep’ scripts (Frost, Katz, & Bentin, 1987 ) but absent in transparent alphabetic script like Finnish and Spanish, for which word-specific orthographic representations may be unnecessary or redundant. Alternatively, the contribution of orthographic knowledge to reading behavior may reflect general cognitive principles of skill acquisition (Anderson, 1981 ). Across a range of domains of skilled human behavior, the transition from novice to expert performance is characterized by a shift from slow, deliberate, effortful, algorithmic processing to rapid, automatic performance mediated by direct-retrieval mechanisms (e.g., Logan, 1988 ). From this perspective, unitized orthographic representations may be a signature of expert word recognition and reading even in highly transparent scripts. However, spelling tests like those reported here are unlikely to provide a sensitive measure of orthographic knowledge in such languages, so other methods will be required in order to investigate this possibility.

Acheson, D.J., Wells, J.B. & MacDonald, M.C. (2008). New and updated tests of print exposure and reading abilities in college students. Behavior Research Methods, 40, 278–289.

PubMed Central   Google Scholar  

Adelman, J. S., Johnson, R. L., McCormick, S. F., McKague, M., Kinoshita, S., Bowers, J. S., . . . Davis, C. J. (2014). A behavioral database for masked form priming. Behavior Research Methods, 46 , 1052–1067.

Anderson, J.R. (1981). Cognitive skills and their acquisition. Hillsdale, NJ: Erlbaum.

Google Scholar  

Andrews, S. (2008). Lexical expertise and reading skill. In B. H. Ross (Ed.), The Psychology of Learning and Motivation: Advances in Research and Theory (Vol. 49, pp. 247–281). San Diego, CA: Elsevier.

Andrews, S. (2012). Individual differences in skilled visual word recognition and reading: The role of lexical quality. In J. S. Adelman (Ed.), Visual Word Recognition (Vol. 2, pp. 151–172). London: Psychology Press.

Andrews, S. (2015). Individual differences among skilled readers: The role of lexical quality. In A. Pollatsek & R. Treiman (Eds.), The Oxford Handbook of Reading (pp. 129–148): Oxford University Press.

Andrews, S. & Hersch, J. (2010). Lexical precision in skilled readers: Individual differences in masked neighbor priming. Journal of Experimental Psychology: General, 139 , 299–318.

Andrews, S. & Lo, S. (2012). Not all skilled readers have cracked the code: Individual differences in masked form priming. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38 , 152–163.

PubMed   Google Scholar  

Andrews, S. & Lo, S. (2013). Is morphological priming stronger for transparent than opaque words? It depends on individual differences in spelling and vocabulary. Journal of Memory and Language, 68 , 279–296.

Andrews, S., Lo, S., & Xia, V. (2017). Individual differences in automatic semantic priming. Journal of Experimental Psychology: Human Perception and Performance, 43 , 1025–1039.

Andrews, S., & Reichle, E.D. (2019). The cognitive architecture of reading: The organization of an acquired skill. In P. Hagoort (Ed)., Human Language: From genes and brains to behaviour . MIT Press.

Andrews, S., & Veldre, A. (2019). What is the most plausible account of the role of parafoveal processing in reading? Language and Linguistics Compass .

Ashby, J., Rayner, K., & Clifton, C. (2005). Eye movements of highly skilled and average readers: Differential effects of frequency and predictability. Quarterly Journal of Experimental Psychology, 58 , 1065–1086.

Braze, D., Katz, L., Magnuson, J.S., Mencl, W.E., Tabor, W., Van Dyke, J.A. et al., (2016). Vocabulary does not complicate the simple view of reading. Reading and Writing, 29 , 435–451.

Braze, D., Tabor, W., Shankweiler, D., & Mencl, W.E. (2007). Speaking up for vocabulary: Reading skill differences in young adults. Journal of Learning Disabilities, 40, 226–243.

PubMed   PubMed Central   Google Scholar  

Breadmore, H.L., & Deacon, S.H. (2019). Morphological processing before and during children’s spelling. Scientific Studies of Reading, 23, 178–191.

Brown, J. I., Fishco, V. V., & Hanna, G. (1993). Nelson-Denny Reading Test . Rolling Meadows, IL: Riverside Publishing Company.

Burt, J. S., & Tate, H. (2002). Does a reading lexicon provide orthographic representations for spelling? Journal of Memory and Language, 46 , 518–543.

Byrne, B. (1998). The foundation of literacy: the child’s acquisition of the alphabetic principle . Hove, UK: Psychology Press.

Byrne, B., Coventry, W. L., Olson, R. K., Hulslander, J., Wadsworth, S., DeFries, J. C., . . . Samuelsson, S. (2008). A behaviour-genetic analysis of orthographic learning, spelling and decoding. Journal of Research in Reading, 31 , 8–21.

Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7 , 309–319.

R Core Team. (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/

Dehaene, S. (2009). Reading in the brain . New York: Penguin Group.

Dehaene, S., Pegado, F., Braga, L., Ventura, P., Filho. G., Jobert, A., Dehaene-Lambertz, G., Kolinsky, R., Morais, M. & Cohen, L. (2010). How learning to read changes the cortical networks for vision and language. Science, 3, 1359–1364.

Drieghe, D., Veldre, A., Fitzsimmons, G., Ashby, J., & Andrews, S. (2019). The influence of number of syllables on word skipping during reading revisited. Psychonomic Bulletin & Review, 26 , 616–621.

Ehri, L. (2000). Learning to read and learning to spell: Two sides of a coin. Topics in Language Disorders, 20, 19–37.

Ehri, L. (2015). How children learn to read words. In A. Pollatsek & R. Treiman (Eds.), The Oxford handbook of reading (pp. 293–310), Oxford, England: Oxford University Press

Ellis, A. W. (1993). Reading, writing and dyslexia: A cognitive analysis . Hillsdale, NJ: Lawrence Erlbaum Associates.

Emmorey, K., Midgley, K.J., Kohen, C.B., Sevcikova Sehyr, Z. & Holcomb, P.J. (2017). The N170 ERP component differs in laterality, distribution and association with continuous reading measures for deaf and hearing readers. Neuropsychologia, 106, 298–309.

Engbert, R., Nuthmann, A., Richter, E., & Kliegl, R. (2005). SWIFT: A dynamical model of saccade generation during reading. Psychological Review , 112 , 777–813.

Eskenazi, M. A., Swischuk, N. K., Folk, J. R., & Abraham, A. N. (2018). Uninformative contexts support word learning for high-skill spellers. Journal of Experimental Psychology: Learning, Memory, and Cognition, 44 (12), 2019–2025.

Falkauskas, K. & Kuperman, V. (2015). When experience meets language statistics: Individual variability in processing English compound words. Journal of Experimental Psychology. Learning, Memory, and Cognition, 41, 1607–1627.

Forster, K.I. (2013). How many words can we read at once? More intervenor effects in masked priming. Journal of Memory and Language, 69, 563–573.

Frith, U. (1980). Unexpected spelling problems. In U. Frith (Ed.), Cognitive Processes in Spelling (pp. 495–515). London: Academic Press.

Frith, U. (1986). A developmental framework for developmental dyslexia. Annals of Dyslexia, 36 , 69–81.

Frost, R., Katz, L., & Bentin, S. (1987). Strategies for visual word recognition and orthographical depth: A multilingual comparison. Journal of Experimental Psychology: Human Perception and Performance , 13 , 104–115.

Grainger, J. (2008). Cracking the orthographic code: An introduction. Language & Cognitive Processes, 23, 1–35.

Holmes, V. M., & Babauta, M. L. (2005). Single or dual representations for reading and spelling? Reading and Writing, 18 , 257–280.

Jackson, M. D., & McClelland, J. L. (1975). Sensory and cognitive determinants of reading speed. Journal of Verbal Learning and Verbal Behavior, 14 , 565–574.

Joesph, H.S., Wonnacot, E., Forbes, P., & Nation, K. (2014). Becoming a written word: Eye movements reveal order of acquisition effects following incidental exposure to new words during silent reading. Cognition, 133, 238–248.

Katz, L. & Frost, S.J. (2001). Phonology constrains the internal orthographic representation. Reading and Writing, 14, 297–332.

Kessler, B., & Treiman, R. (2001). Relationships between sounds and letters in English monosyllables. Journal of Memory and Language, 44, 592–617.

Keuleers, E. (2013). vwr: Useful functions for visual word recognition research (R package version 0.3.0). https://CRAN.R-project.org/package=vwr

Kuperman, V. & Van Dyke, J.A. (2011). Effects of individual differences in verbal skills on eye-movement patterns during sentence reading. Journal of Memory and Language, 65, 42–73.

Landi, N., & Perfetti, C. A. (2007). An electrophysiological investigation of semantic and phonological processing in skilled and less-skilled comprehenders. Brain and Language, 102 , 30–45.

Lefly, D. L., & Pennington, B. F. (2000). Reliability and validity of the Adult Reading History Questionnaire. Journal of Learning Disabilities, 33 , 286–296.

Logan, G.D. (1988). Toward an instance theory of automatization. Psychological Review, 95, 492–527.

Luke, S. G., Henderson, J. M., & Ferreira, F. (2015). Children's eye-movements during reading reflect the quality of lexical representations: An individual differences approach. Journal of Experimental Psychology: Learning, Memory, and Cognition, 41 , 1675–1683.

Marinus, E., Kohnen, S., & McArthur, G. (2013). Australian comparison data for the Test of Word Reading Efficiency (TOWRE). Australian Journal of Learning Difficulties, 18 , 199–212.

Meade, G., Grainger, J., Midgley, K.J., Emmorey, K. & Holcomb, P.J. (2018). From sublexical facilitation to lexical competition: ERP effects of masked neighbor priming. Brain Research, 1685, 29–41.

Mol, S.E. & Bus, A.G. (2011). To read or not to read: a meta-analysis of print exposure from infancy to early adulthood. Psychological Bulletin, 137, 267–296.

Moore, M. & Gordon, P. (2015). Reading ability and print exposure: item response theory analysis of the author recognition test. Behavior Research Methods, 47, 1095–1109.

Nagy, W.E., Herman P.A., & Anderson, R.C. (1985). Learning words from context. Reading Research Quarterly, 20, 439–458.

Nation, K. (2017). Nurturing a lexical legacy: reading experience is critical for the development of reading skill. npj Science of Leaning, 2 :3.

Nouwens, S., Groen, M. A., Kleeman, T., & Verhoeven, L. (2017). The role of semantic retrieval in children's reading comprehension development in the upper primary grades. Journal of Research in Reading, 41 , 597–614.

Perfetti, C. & Hart, L. (2001). The lexical basis of comprehension skill. In D. Gorfein (Ed.), On the consequences of meaning selection: Perspectives on resolving lexical ambiguity (pp. 67–86). Washington DC: American Psychological Association.

Perfetti, C. A. (1985). Reading Ability . New York: Oxford University Press.

Perfetti, C. A. (2007). Reading ability: Lexical quality to comprehension. Scientific Studies of Reading, 11 , 357–383.

Perfetti, C. A., & Hart, L. (2002). The lexical quality hypothesis. In L. Verhoeven, C. Elbro & P. Reitsma (Eds.), Precursors of functional literacy (pp. 189–214). Amsterdam, The Netherlands: John Benjamins Publishing.

Pexman, P.M. & Yap, M.J. (2018). Individual differences in semantic processing: Insights from the Calgary Semantic Decision Project. Journal of Experimental Psychology. Learning, Memory, and Cognition , 44 , 1091–1112.

Protopapas, A., Sideridis, G. G., Mouzaki, A., & Simos, P. G. (2007). Development of lexical mediation in the relation between reading comprehension and word reading skills in Greek. Scientific Studies of Reading, 11 , 165–197.

Radach, R., & Kennedy, A. (2013). Eye movements in reading: Some theoretical context. Quarterly Journal of Experimental Psychology, 66 , 429–452.

Rayner, K., Abbott, M. J., & Plummer, P. (2015). Individual differences in perceptual processing and eye movements in reading. In P. Afflerbach (Ed.), Handbook of individual differences in reading: Reader, text, and context (pp. 348–363). New York, NY: Routledge.

Reichle, E. D., Pollatsek, A., Fisher, D. L., & Rayner, K. (1998). Toward a model of eye movement control in reading. Psychological Review,105, 125–157.

Rodd, J.M. (2004). When do leotards get their spots? Semantic activation of lexical neighbors in visual word recognition. Psychonomic Bulletin & Review, 11, 434–439.

Rossi, M., Martin-Chang, S., & Ouellette, G. (2019). Exploring the space between good and poor spelling: Orthographic quality and reading speed. Scientific Studies of Reading, 23, 192–201.

Share, D.L. (1995). Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cognition, 55, 151–218.

Share, D.L. (2008). On the anglocentricities of current reading research and practice: The perils of overreliance on an “outlier” orthography. Psychological Bulletin, 134 , 584–615.

Slattery, T.J., & Yates, M. (2018). Word skipping: Effects of word length, predictability, spelling and reading skill. Quarterly Journal of Experimental Psychology, 71 , 250–259.

Snell, J., van Liepsig, S., Grainger, J., & Meeter, M. (2018). OB1-reader: A model of word recognition and eye movements in text reading. Psychological Review, 125 , 969–984.

Stanovich, K. E., & West, R. F. (1989). Exposure to print and orthographic processing. Reading Research Quarterly, 24 , 402–433.

Swart, N. M., Muijselaar, M. M. L., Steenbeek-Planting, E. G., Droop, M., de Jong, P. F., & Verhoeven, L. (2017). Differential lexical predictors of reading comprehension in fourth graders. Reading and Writing, 30 , 489–507.

Tamura, N., Castles, A., & Nation, K. (2017). Orthographic learning, fast and slow: Lexical competition effects reveal the time course of word learning in developing readers. Cognition, 163, 93–102.

Tan, L. C., & Yap, M. J. (2016). Are individual differences in masked repetition and semantic priming reliable? Visual Cognition, 24 , 182–200.

Taylor, J. N., & Perfetti, C. A. (2016). Eye movements reveal readers’ lexical quality and reading experience. Reading and Writing, 29 , 1069–1103.

Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (1999). Test of word reading efficiency . Austin, TX: ProEd.

Treiman, R. (2017). Learning to spell: Phonology and beyond. Cognitive Neuropsychology, 34 , 83–93.

Veldre, A. & Andrews, S. (2014). Lexical quality and eye movements: Individual differences in the perceptual span of skilled adult readers. Quarterly Journal of Experimental Psychology, 67 , 703–727.

Veldre, A. & Andrews, S. (2015a). Parafoveal lexical activation depends on skilled reading proficiency. Journal of Experimental Psychology. Learning, Memory, and Cognition , 41 , 586–595.

Veldre, A. & Andrews, S. (2015b). Parafoveal lexical activation is modulated by the precision of skilled readers’ lexical representations. Journal of Experimental Psychology: Human Perception and Performance , 41, 219–232.

Veldre, A., & Andrews, S. (2016a). Semantic preview benefit in English: Individual differences in the extraction and use of parafoveal semantic information. Journal of Experimental Psychology: Learning, Memory, and Cognition, 42 , 837–854.

Veldre, A., & Andrews, S. (2016b). Is semantic preview benefit due to relatedness or plausibility? Journal of Experimental Psychology: Human Perception and Performance, 42 , 939–952.

Veldre, A., Drieghe, D., & Andrews, S. (2017). Spelling ability selectively predicts the magnitude of disruption in unspaced text reading. Journal of Experimental Psychology: Human Perception and Performance, 43 , 1612–1628.

Wilkinson, G.S., & Robertson, G.J. (2017). Wide Range Achievement Test (5 th edition). Pearson Education Inc.

Yap, M.J., Balota, D.A., Sibley, D.E. & Ratcliff, R. (2012). Individual differences in visual word recognition: Insights from the English Lexicon Project. Journal of Experimental Psychology: Human Perception and Performance, 38 , 53–79.

Yap, M. J., Tse, C. S., & Balota, D. A. (2009). Individual differences in the joint effects of semantic priming and word frequency revealed by RT distributional analyses: The role of lexical integrity. Journal of Memory and Language, 61 , 303–325.

Ziegler, J.C., Perry, C., & Zorzi, M. (2014). Modelling reading development through phonological decoding and self-teaching: implications for dyslexia. Philosophical Transactions of the Royal Society B, 369, 20120397 .

Download references

Author information

Authors and affiliations.

School of Psychology, The University of Sydney, Sydney, NSW, 2006, Australia

Sally Andrews, Aaron Veldre & Indako E. Clarke

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Sally Andrews .

Additional information

The data, test instruments, and norms for the Spelling Dictation Test and the Spelling Recognition Test are available at https://osf.io/t4x7r/ . This research was supported under the Australian Research Council’s Discovery Projects funding scheme (project numbers DP16203224, DP190100719).

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Spelling Dictation Test

Administration instructions: The experimenter reads the word aloud to the participant, followed by the sentence containing the word.


The ex-alcoholic found it very difficult to maintain complete abstinence from drinking.


She knew the woman as an acquaintance, but she was not a close friend.


The nurses had to blend the food to make it digestible for the patient.


She tried to adopt a conciliatory approach to avoid further conflict.


The biscuit with pistachio nuts was delicious.


The TV seemed a good buy because it had a 3 year warranty.


The old woman suffered rheumatic pain in all of her joints.


The music reached a crescendo towards the end of the symphony.


She found the asymmetry of the design very appealing.


By comparison with most Asian countries, Australia is very affluent.


Most of the students are lazy but this boy is very diligent in completing his work.


The noise from the next classroom was a constant aggravation to the teacher.


It is usually not appropriate to use colloquial language in an essay.


The student felt euphoric when she completed her last exam.


Children often dislike broccoli and other green vegetables.


The gymnast turned a somersault before landing back on the bar.


Many sportsmen achieve fame when they are young but then sink into oblivion.


The rhythmical beat of the drums was mesmerizing.


She had missed lunch so by dinner time she felt ravenous.


She tried to persuade him to her point of view.

Spelling Recognition Test

Administration instructions: The participant is given unlimited time to select all the incorrectly spelled items. Note: The item ‘behaviour’ should be replaced by ‘behavior’ if administering the test to speakers of US English.


Rights and permissions

Reprints and permissions

About this article

Andrews, S., Veldre, A. & Clarke, I.E. Measuring Lexical Quality: The Role of Spelling Ability. Behav Res 52 , 2257–2282 (2020). https://doi.org/10.3758/s13428-020-01387-3

Download citation

Published : 13 April 2020

Issue Date : December 2020

DOI : https://doi.org/10.3758/s13428-020-01387-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

  • Reading ability
  • spelling ability
  • individual differences
  • lexical quality hypothesis
  • Find a journal
  • Publish with us
  • Track your research

Personality Trait Theory

practical psychology logo

The words “personality” and “trait” go hand in hand. When you take online quizzes about your personality, you probably get traits as your answer. These quizzes illustrate our unique traits compared to others, yet occasionally, we discover commonalities in our responses with friends and neighbors.

But how does the trait perspective of psychology work? How have psychologists organized different traits? Let's find out. 

What Is The Trait Theory of Personality?

Traits describe stable, consistent, and meaningful differences among individuals. Using language, traits describe people's objective behaviors. They are displayed on a dimension or spectrum. Personalities are made up of these traits assigned to individuals to show how they differ from others.

What Are Traits?

Before we dive deeper into personality, let's talk about traits.

Meaningful Differences

Traits describe meaningful differences among individuals. For example, everyone eats food - a behavior we all participate in to survive. It is driven by biology. However, The way you eat your food could be a sign of different personality traits. Sometimes, the habits in which you eat your food can be signs of cultural differences. Some people are taught to eat with their hands; others are taught to keep their mouths closed or slurp their soup.

But if someone is in a room with people who have different eating habits, their personality may greatly influence how they approach the situation. Are they conscious of how people are eating around them? Do they care to fit in?

Already, you can see how genetics, culture, and personality all form a complicated spider web that can be hard to trace.

Stable and Consistent

Second, traits are stable and consistent. People display signs of personality traits across different situations throughout their lives. Again, culture, rules, and the context of a situation will greatly impact how someone behaves. But if someone is honest , this trait will heavily influence what it takes for the person to lie (or justify these actions later.) Social psychology really likes to look at instances when people break their normal personality traits. 

For example, if you were to take a personality test , you should answer the questions in the same manner for many years.

Dimensions of Spectrums

Third, traits are usually displayed as dimensions or spectrums with extremes at both ends. Introvert vs. extrovert is one of the most common sets of personality traits that we know and talk about. But I’m sure that you know not everyone identifies or displays the behaviors of an extreme introvert or an extreme extrovert. Some people are “ambiverts” and fall between these two extremes.

Fourth, traits rely on language. We can't call it a trait if we don’t have a word describing how a person “is” or how they act.

Lexical Hypothesis in Personality Psychology 

Lexical Hypothesis is a theory that says if there's a behavior so prominent throughout time, we create a word for that word. If we don't have a word that describes a trait, then it must not be very prevalent or useful to personality psychology.

Objective Behavior 

Last but certainly not least, traits are objective behavior. This is especially important to remember when describing yourself or another person’s personality traits. An introvert is not necessarily “good,” full stop. Introversion may benefit a person in certain situations, but you can’t write any trait off as “good” or “bad.” Culture, again, plays a part here. Competitive behavior may be advantageous in growing your startup, but it may be a disadvantage in developing meaningful relationships.

You may be raised in a culture that teaches you to be agreeable or amicable; someone worldwide may be raised to be independent and put themselves first. Both traits may seem more positive or negative depending on your goals, values, or beliefs.

Also, traits must be a behavior. For example, we can say that someone is 6 feet tall. That isn't a personality trait; it's a physical trait. It's not a behavior!

There are four people to know in the world of trait psychology. These psychologists have spent their lives organizing traits into a central group of terms or spectrums that can be applied to everyone. (Basically, they interpret what our answers are and what the responses mean.)

Let’s get to know them.

Gordon Allport 

Gordon Allport

Gordon Allport is a great trait theorist to start with. In the early part of the 20th century, he searched the dictionary and found over 4,500 words that could be considered personality traits. (Nowadays, we have about 18,000 trait-descriptive adjectives.) From those 4,500 words, he came up with three different types of traits.

Cardinal Traits

The first category consists of cardinal traits. These traits and behaviors rule how you approach the things you are passionate about. Punctual is a classic example of a cardinal trait; it is usually influenced by some desire to impress or be ready to get to work. If someone had to describe you in three words, these three words would most likely be Cardinal Traits. Some traits are named after people: Machiavellian , Freudian, Christ-like. 

Central Traits

The second category is central traits. These traits are found to a certain degree in every person. Honesty, agreeableness, or jealousy may all be considered central traits that may or may not come from our genetic makeup.

Secondary Traits 

Last is secondary traits. These traits may apply to different situations depending on the context of said situation. In general, you may be a respectful person. But if you dislike a certain authority figure or person, people may see a rude side to you. Another word for these are "attitudes" or "preferences". 

Looking at 4,500 words, you’re bound to find some repeats and synonyms. In the 60s, Cattell took the 4,500 trait words from Allport and narrowed them down to 171 traits. He wasn’t done yet. He used factor analysis to look for trends in these 171 words and narrowed them down to the most influential traits.

He came down to 16 using a process called Factor Analysis. Factor analysis can be used to look at enormous amounts of data to look for trends and to see which elements are the most influential or important.

Remember what we said about how traits can be on a spectrum? So were Cattell’s 16 personality traits. Each of these 16 words had a direct opposite. Most people fit somewhere in the middle of these two extremes. Here's a table of those 16, along with their dimensions:

As you can see, some of these personality traits are very similar. For example, "Private" under Introversion is similar to "Loner" under Independence. 

Eysenck had a specific job when he developed his theory. Around the same time Cattell was developing his theories on personality, Eysenck worked at a psychiatric hospital in London. His job was to make an initial assessment of the patients. Eysenck noticed certain trends; he found that soldiers, for example, seemed to answer questions in a similar way. Maybe, these answers revealed specific traits that led a person to become a soldier.

Eysenck called these traits first-personality traits.

What Eysenck is most known for, however, is the PEN Model. He condensed the most important personality traits into just three traits: psychoticism, extraversion, and neuroticism.

These seem like negative traits, but let's look at what they mean:

Psychoticism: When an individual engages in risky and irresponsible behavior. People with high psychoticism usually have a more aggressive temperament. 

Extraversion : When an individual engages in a lot of social activities. Also, an extravert is considered "under-aroused," and their cortical arousal can be measured with skin conductance. Skin conductance measures the skin's electrical conductivity, which increases with sweat gland activity. In extroverts, who are considered "under-aroused," this method helps assess their level of arousal during social activities.

Neuroticism : When an individual's mood and emotions fluctuate more than normal. Eysenck said these people experienced more flight-or-fight reactions than most people.

Again, these are all on spectrums. Eysenck theorized that we all displayed some level of these traits, but we just expressed them differently. Part of his theory comes from the belief that our personality traits come from our genetics.

I save The Big Five for last because we will go through this theory in more detail in another article. Psychology credits a small group of psychologists with the development of this theory. It is a “happy medium” between the three personality traits developed by Eysenck and the 16 developed by Cattell. 

These are the Big Five, the OCEAN Theory, or the Five Factor Model. Similar to the PEN Model, OCEAN is an acronym for five different traits that all humans display some degree of.

Five Factor Model

Take some time to think about your personality traits. You certainly have a lot of ways to assess the traits that you hold. Family and friends probably already have a few choice words to describe your personality.

The more you reflect on your personality and remember that these traits are objective, the more you can understand yourself. This self-awareness will help you find and approach opportunities that best fit your personality traits. If you know you are an introvert, you can use this knowledge to create a schedule or pursue opportunities that allow introverts to shine.

Related posts:

  • 17+ Cardinal Traits (Definition + Examples)
  • What is Personality Psychology?
  • Big 5 Factor Model of Personality (Free Test + OCEAN)
  • Biological Theory of Personality
  • 40+ Famous Psychologists (Images + Biographies)

Reference this article:

About The Author

Photo of author

Personality Psychology:

What is Personality?

Personality Theories

Trait Theory

Behavioral Theory

Humanistic Theory

Biological Theory

Psychoanalytic Theory

Personality Tests


Openness to Experience




The Dark Triad

Myers Briggs Type Indicator

Keirsey Temperament Sorter

Inkblot and TAT

House Person Tree Test

Defense Mechanisms

Manipulation Tactics

Type A and Type B Personalities

Personality Disorders


Physical Health


Sex Differences in Personality

Child Birth Order

Nature vs Nurture

what is lexical hypothesis

PracticalPie.com is a participant in the Amazon Associates Program. As an Amazon Associate we earn from qualifying purchases.

Follow Us On:

Youtube Facebook Instagram X/Twitter

Psychology Resources




Serial Killers

Psychology Tests

Personality Quiz

Memory Test

Depression test

Type A/B Personality Test

© PracticalPsychology. All rights reserved

Privacy Policy | Terms of Use

what is lexical hypothesis

  • May 31, 2022
  • 10 min read

Personality 101: The Trait Approach & the Lexical Hypothesis

Human personality is a well-known concept in both academic and non-academic circles. This concept has raised the most diverse conclusions in both circles: from well-established factorial solutions to classifications of people based on what the Sorting Hat from Hogwarts would estimate. For a long time now, research in psychology has gained plenty of knowledge about human personality and its implications in everyday life, but this knowledge is either unknown or misunderstood by the general population. This situation calls for efforts to close this gap between what is known based on science and what is assumed to be true based on our random and subjective experience. The following 101 series explores the most consensual contemporary conceptualization of personality, the creative steps scientists took to arrive at this conceptualization using the lexical hypothesis, the specificities of this contemporary conceptualization by looking at each one of the Big Five traits, and the implications of each trait for individuals and societies.

The Personality 101 series is divided into 5 chapters:

Personality 101: The Trait Approach and The Lexical Hypothesis

Personality 101: Conscientiousness

Personality 101: Agreeableness and Extraversion

Personality 101: Emotional Stability

Personality 101: Openness to Experience

Thirty-one -year-old Robert often feels restless. He has problems sitting at a desk for more than a few minutes, cannot get organized, loses his keys and wallet, and forgets about his plans for the evening. He fails to achieve up to his potential at work. During the conversation, his mind wanders and he interrupts others, blurting out what he is thinking without considering the consequences. He gets into arguments. His mood swings and periodic outbursts make life difficult for those around him. Now his marriage is in trouble (John, 2021) .

This is a typical description of someone’s personality in a clinical setting. It conveys a good idea of how Robert is: disorganized with things and time causing evident consequences in his work, disorganized in speech, restless and impulsive. With this small paragraph, a clinician might already have an idea of what Robert is and what the goals would be in clinical intervention. However, something is lacking here. Surely, a short paragraph cannot summarize the wholeness of a human being, all their intricacies and idiosyncrasies; besides, the paragraph lacks some mention of the positive characteristics, too. Despite the risk of losing information, the task of trying to summarize someone's personality is useful, it helps with making life decisions easier: decisions like the search for a job, choosing a career, a partner, an adequate therapeutic procedure, and even one’s friends and hobbies.

What is needed is to summarize the personality of someone with the least loss of information possible, i.e. a descriptive model or a taxonomy of personality. “One of the central goals of scientific taxonomies is the definition of overarching domains within which large numbers of specific instances can be understood in a simplified way” (John, 2021, p. 38 ) . In plain words, a taxonomy about personality would be useful because it will allow the interpretation of the massive amount of information contained in one person’s behaviour using a very small group of categories. In this article, the best taxonomy for personality is going to be introduced along with the methodology followed to design it.

A taxonomy of human personality has been searched for a long time. Even in Ancient Greece Theophrastus would ponder: “why is that, while all Greece lies under the same sky and all the Greeks are educated alike, it has befallen us to have characters variously constituted?” (Theophrastus, 1909, p.77) . The most famous of the ancient attempts is Hippocrates' taxonomy in which he believed that different proportions of four bodily fluids or humour would manifest in the way people think, feel, and behave. A predominance of blood constituted a sanguine or social character, phlegm constituted a phlegmatic or easygoing character, black bile constituted a melancholic or analytical character, and yellow bile constituted a choleric or extraverted character (Chiao, 2018) .

what is lexical hypothesis

Before going into more taxonomies of traits a note of recognition must be conceded to many other conceptualizations of personality that are not trait-based. Whereas the trait approach is one of the most frequently used nowadays, many authors in the history of psychology proposed models based on their own scientific and theoretical framework (Funder, 2012) . Freud’s framework, for instance, is based on the psychosexual development of the person, and he would argue that personality suffers many changes during childhood, also known as the stages of psychosexual development, but once adolescence arrives, personality becomes rather stable. Another key aspect of his theory is the division of personality into three components: the less conscious aspect, the id; the conscious experience of the person, the ego; and the social demands internalized in the individual, the super-ego (Funder, 2012) .

Later, other authors would propose different psychological processes as important components of human experience. Jung proposed the collective unconscious, the Anima and the Shadow as crucial mechanisms of the human psyche. More humanistic perspectives, like the ones championed by Abraham Maslow, Carl Rogers or Positive Psychology, focused on self-actualization processes, the acceptance of one's own experience, the hierarchies of motivations and needs, and the positive aspects of psychology like strengths and virtues (Funder, 2012) .

These are important contributions but some of them are elusive to scientific investigation, meaning that it is hard to apply the scientific method to answer the questions they pose. This is why, currently, they do not receive the same attention as the trait approach, and though there are some attempts to understand them, their significance is not as universal as the trait approach (Mastnak, 2021) .

The most basic tenet of the trait approach is, understandably, the trait. According to Allport (1931) , a trait has more than a nominal existence, is more than a generalized habit, is dynamic, it can be empirically or statistically found, is not unrelated to other traits, is not a moral quality, is not disproven if other behaviors appear in the behavioral repertoire of an individual, and it can be spotted in the individual and also in the population. Therefore, a trait is something real, can be found using statistical tools, and it is not disproven if other behaviors or emotions that go against this trait appear. This last feature of traits reveals an important nuance: a trait is the natural tendency of an individual, what the person would do almost spontaneously, the default mode of operating; this does not mean that, in a particular situation, an individual could not show behaviors, feelings, or thoughts that deviate from this natural tendency (Fleeson & Law, 2015) . Practically, this means that an extravert can sometimes act as an introvert, and vice versa.

what is lexical hypothesis

Once the concept of trait has been cleared, let’s look at some of the attempts to define a taxonomy of personality in the modern history of Psychology. Raymond Cattell, a British-American psychologist, proposed 16 factors, or overarching categories obtained by means of statistical procedures, that comprise traits like Warmth or being outgoing and supportive, Social Assertiveness or being uninhibited and bold in social situations, Introversion or being reserved and clear-headed, and Independence or being self-sufficient (Cattell & Mead, 2008) . His theory led to the development of the Sixteen Personality Factor Questionnaire (16PF) that is still being used in vocational and educational settings. Another well-known theory of traits is the Eysencks’ theory of personality, in which there were only two bipolar factors accounting for all the variation in human personality: extroversion-introversion, and emotional stability-instability (Furnham et al., 2008) .

The methodology used by these authors to obtain such taxonomies is mostly based on the following process. The author would first get enough information about the topic by reading or gathering the conclusions obtained after years of experience in therapy and consultation. Once they think they have a solid theoretical framework from which to talk about personality, they will enumerate a set of traits that could explain and summarize humans across time and places. Although this is an oversimplification of the process, it serves one purpose: to show that, although it may be helpful for the patients of the author, it is not replicable or its replicability could be easily questioned. Therefore, since human personality is a universal phenomenon, a taxonomy that could replicate itself across contexts and individuals is needed (Mischel, 1996) .

There is a different process that overcomes the limitations of the previous one. This is the so-called lexical hypothesis, proposed by Galton (1949) , which states that every important human phenomenon must be somehow represented in the lexicon of a language, and since most languages are easily translated to others, the universality of the phenomenon could also be guaranteed. Based on this proposition, what authors would normally do is gather all the words used in a particular language for the phenomenon of interest from a representative sample of words in that language (some examples include dictionaries but also the transcripts of contemporary famous TV shows and movies), then they would ask a group of experts to analyze this list and determine which words are better at capturing the phenomenon under study (Ashton & Lee, 2007; Oreg et al., 2020; Parrigon et al., 2017) . This implies discarding synonyms and uncommon words from the sample. Later, they would approach a representative sample of individuals to categorize the phenomenon of interest (e.g. personality) according to the words established in the previous step. This will allow, by using proper psychometric and statistical techniques (i.e., factorial and multivariate statistical analyses), both to filter the best words that will be used and to create a refined measure of the phenomenon.

what is lexical hypothesis

As of now, the scientific consensus is that the lexical hypothesis is the best solution found so far for the classification and understanding of personality (DeYoung et al., 2007) . In fact, one of the most famous and used taxonomies used at the moment, both in research, educational, clinical, and vocational settings, is the Big Five Taxonomy of personality, initially proposed by Costa & McCrae (1992) , and that have been further developed ever since. This taxonomy summarizes personality in five overarching traits: Conscientiousness, Agreeableness, Emotional Stability, Openness to Experience, and Extraversion. Each one of this will be further developed in the following posts of this 101 series.

One important note of warning is that, although this Big Five solution has proven to be very useful, the traits that conform it should not be understood as casual entities of human behaviors, instead they should be understood according to their real nature: a descriptive explanation of human personality. They do not explain casual relationships, they describe personality, they summarize it so that it can be easier to understand it (Fajkowska & Kreitler, 2018) . A second note is that these traits are not separable and completely distinguishable entities in and of themselves, they are correlated and there is some degree of overlap between some of them (Van der Linden et al., 2012) .

A final note, and the most important one, is that this taxonomy is based, as almost everything else in Psychology, on self-reports accounts. This is an important issue because, as Jung would say, “you are not what you say you’ll do, but what you do” (PsycholoGenie, 2014) . This fact poses a challenge, specially in personality research: are the tests really measuring psychological phenomena if they rely solely on self-report accounts? Are scientists not purposefully biasing their findings because self-reports suppose cheaper costs in research than observational or experimental reports (Galic et al., 2016; Olino & Klein, 2015) ? These questions are fueling some alternative research directions, like gathering behavioral data by means of wearables, cameras, and smartphones which, as of now, are both showing coincidences with the already extant research on the topic, and pushing its boundaries forward (Ihsan & Furnham, 2018) .

what is lexical hypothesis

Personality psychology is an important and flourishing branch of Psychology. Its aim is to better understand human behaviors, thoughts, and emotions so that its knowledge would allow people to make better informed decisions in their life. Throughout history, personality has raised many questions to philosophers, writers, thinkers, and scientists and there have been many attempts to understand it. Though the validity of some of them could be recognized from a phenomenological point of view, the scientific method is not yet capable of working with them. Now, the consensus obtained in science is that the best solution found so far is the lexical hypothesis, as manifested in the relevance and extended use of the Big Five Theory of personality. These five traits have been proven useful in the description of many aspects of the human experience and their research is still a burgeoning theme in Psychology, though they are not free of limitations and improvements. Contemporary technologies are being gradually incorporated in the study of personality and their conclusions are solidifying the field and incorporating new findings. Only time will show how far the study of personality will get, and it is almost breathtaking to think that it all started with a man pondering why there were so many differences in people that were born under the same Ancient Greek sky.

Bibliographical References

Allport, G. W. (1931). What is a trait of personality? The Journal of Abnormal and Social Psychology , 25 (4), 368–372. https://doi.org/10.1037/h0075406

Ashton, M. C., & Lee, K. (2007). Empirical, Theoretical, and Practical Advantages of the HEXACO Model of Personality Structure. Personality and Social Psychology Review , 11 (2), 150–166. https://doi.org/10.1177/1088868306294907

Cattell, H. E. P., & Mead, A. D. (2008). The Sixteen Personality Factor Questionnaire (16PF). In The SAGE handbook of personality theory and assessment, Vol 2: Personality measurement and testing (pp. 135–159). Sage Publications, Inc. https://doi.org/10.4135/9781849200479.n7

Chiao, E. (2018, October 4). New study reveals four major personality types . The Johns Hopkins News-Letter. https://www.jhunewsletter.com/article/2018/10/new-study-reveals-four-major-personality-types

Costa, P. T., & McCrae, R. R. (1992). Four ways five factors are basic. Personality and Individual Differences , 13 (6), 653–665. https://doi.org/10.1016/0191-8869(92)90236-I

DeYoung, C. G., Quilty, L. C., Peterson, J. B., & nueva, E. a sitio externo E. enlace se abrirá en una ventana. (2007). Between facets and domains: 10 aspects of the Big Five. Journal of Personality and Social Psychology , 93 (5), 880–896. https://doi.org/10.1037/0022-3514.93.5.880

Fajkowska, M., & Kreitler, S. (2018). Status of the Trait Concept in Contemporary Personality Psychology: Are the Old Questions Still the Burning Questions? Journal of Personality , 86 (1), 5–11. https://doi.org/10.1111/jopy.12335

Fleeson, W., & Law, M. K. (2015). Trait enactments as density distributions: The role of actors, situations, and observers in explaining stability and variability. Journal of Personality and Social Psychology , 109 (6), 1090–1104. https://doi.org/10.1037/a0039517

Funder, D. C. (2012). The Personality Puzzle (pp. xxvii, 466). WW Norton & Co.

Furnham, A., Eysenck, S. B. G., & Saklofske, D. H. (2008). The Eysenck personality measures: Fifty years of scale development. In The SAGE handbook of personality theory and assessment, Vol 2: Personality measurement and testing (pp. 199–218). Sage Publications, Inc. https://doi.org/10.4135/9781849200479.n10

Galic, Z., Bubić, A., & Parmac Kovacic, M. (2016). Alternatives to self-reports: Conditional reasoning problems and IAT-based tasks. In The Wiley Handbook of Personality Assessment (p. 215.-227.). https://doi.org/10.1002/9781119173489.ch16

Galton, F. (1949). The Measurement of Character. In Readings in general psychology (pp. 435–444). Prentice-Hall, Inc. https://doi.org/10.1037/11352-058

Ihsan, Z., & Furnham, A. (2018). The new technologies in personality assessment: A review. Consulting Psychology Journal: Practice and Research , 70 (2), 147–166. https://doi.org/10.1037/cpb0000106

John, O. P. (2021). History, measurement, and conceptual elaboration of the Big‑Five trait taxonomy: The paradigm matures. In Handbook of personality: Theory and research, 4th ed (pp. 35–82). The Guilford Press.

Mastnak, W. (2021). Psychoanalysis and Qualitative Factor Analysis: A comparative meta- theoretical perspective . https://doi.org/10.13140/RG.2.2.11022.89922

Mischel, W. (1996). Personality and Assessment . Psychology Press. https://doi.org/10.4324/9780203763643

Olino, T. M., & Klein, D. N. (2015). Psychometric Comparison of Self- and Informant-Reports of Personality. Assessment , 22 (6), 655–664. https://doi.org/10.1177/1073191114567942

Oreg, S., Edwards, J. A., & Rauthmann, J. F. (2020). The situation six: Uncovering six basic dimensions of psychological situations from the Hebrew language. Journal of Personality and Social Psychology , 118 (4), 835–863. https://doi.org/10.1037/pspp0000280

Parrigon, S., Woo, S. E., Tay, L., & Wang, T. (2017). CAPTION-ing the situation: A lexically-derived taxonomy of psychological situation characteristics. Journal of Personality and Social Psychology , 112 (4), 642–681. https://doi.org/10.1037/pspp0000111

PsycholoGenie. (2014, August 12). A Comprehensive Collection of 60 Famous Quotes By Carl Jung. Psychologenie . https://psychologenie.com/collection-of-famous-quotes-by-carl-jung

Theophrastus. (1909). The characters of Theophrastus (R. C. Jebb, Trans. & J. E. Sandys, Ed.). London: Macmillan.

Van der Linden, D., Tsaousis, I., & Petrides, K. V. (2012). Overlap between General Factors of Personality in the Big Five, Giant Three, and trait emotional intelligence. Personality and Individual Differences , 53 (3), 175–179. https://doi.org/10.1016/j.paid.2012.03.001

David Saeteros

Arcadia _ Logo.png

Arcadia, has many categories starting from Literature to Science. If you liked this article and would like to read more, you can subscribe from below or click the bar and discover unique more  experiences  in our articles in many categories

Let the posts come to you.

Thanks for submitting!

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • View all journals
  • My Account Login
  • Explore content
  • About the journal
  • Publish with us
  • Sign up for alerts
  • Review Article
  • Open access
  • Published: 27 January 2017

Nurturing a lexical legacy: reading experience is critical for the development of word reading skill

  • Kate Nation 1  

npj Science of Learning volume  2 , Article number:  3 ( 2017 ) Cite this article

8756 Accesses

63 Citations

88 Altmetric

Metrics details

  • Human behaviour
  • Learning and memory

The scientific study of reading has taught us much about the beginnings of reading in childhood, with clear evidence that the gateway to reading opens when children are able to decode, or ‘sound out’ written words. Similarly, there is a large evidence base charting the cognitive processes that characterise skilled word recognition in adults. Less understood is how children develop word reading expertise. Once basic reading skills are in place, what factors are critical for children to move from novice to expert? This paper outlines the role of reading experience in this transition. Encountering individual words in text provides opportunities for children to refine their knowledge about how spelling represents spoken language. Alongside this, however, reading experience provides much more than repeated exposure to individual words in isolation. According to the lexical legacy perspective, outlined in this paper, experiencing words in diverse and meaningful language environments is critical for the development of word reading skill. At its heart is the idea that reading provides exposure to words in many different contexts, episodes and experiences which, over time, sum to a rich and nuanced database about their lexical history within an individual’s experience. These rich and diverse encounters bring about local variation at the word level: a lexical legacy that is measurable during word reading behaviour, even in skilled adults.

Similar content being viewed by others

what is lexical hypothesis

Universal and specific reading mechanisms across different writing systems

Xingshan Li, Linjieqiong Huang, … Jukka Hyönä

what is lexical hypothesis

Rapid online assessment of reading ability

Jason D. Yeatman, Kenny An Tang, … Benjamin W. Domingue

what is lexical hypothesis

Nurturing the reading brain: home literacy practices are associated with children’s neural response to printed words through vocabulary skills

Cléa Girard, Thomas Bastelica, … Jérôme Prado

Most children cannot read before they go to school, but fast forward a few years and they are working their way through Harry Potter. How does this learning happen? The science of reading has taught us much about the genesis of reading. In alphabetic languages such as English, we know that an understanding of phonology—the sound system of spoken language—underpins the development of the alphabetic principle 1 —the insight that print represents meaning via sound. Armed with this insight, children discover the spelling-sound mappings that characterise their language and from this, they have a means to access language from print. Our scientific understanding of the end point of learning is also advanced. Cognitive psychology is abound with studies examining how adults process written words 2 and much is known about the neural systems that support word reading. 3 , 4 , 5 Despite a rich understanding of both beginning reading and its end state, how children move from one to the other is not well understood. The lexical legacy hypothesis, introduced in this paper, provides a new perspective on the transition from novice to expert.

The focus of this paper is with how people read words. While reading comprehension requires much more than the identification of individual words, comprehension can not happen without it. 6 , 7 Thus, understanding how word reading expertise develops is critical. Importantly, however, as will become clear once the lexical legacy hypothesis is described, we can not divorce the processes involved in word identification from the reality that words are not experienced in an isolated vacuum. Words occur in meaningful context, in both spoken and written language. The lexical legacy account argues that this is important. It sees skilled word reading as, in part, a consequence of experiencing words in diverse and meaningful language environments during reading experience. Reading experience provides the substrate that allows a person to build knowledge of an individual word, not just of its spelling and pronunciation, but knowledge of its meaning and how it connects to other words. This rich knowledge base underpins reading fluency and reading comprehension. But before elaborating further, we need to begin with what happens before then, for reading experience can only exert its influence once children are able to read words. So what needs to happen to get the system kick-started?

Beginning reading: from overt phonological decoding to orthographic processing

Compare your experience of reading words with that of a young child. For you, word reading is usually fast, accurate and largely without effort. For a novice, reading is characterised by phonological decoding, whereby letter strings are closely analysed and laboriously ‘sounded out’ to form words. This moment of introspection highlights the essence of what needs to develop: word reading becomes more automatic and less effortful. How is this achieved?

Let us first focus on beginning reading. There is clear consensus and abundant evidence (for review, see refs 1 , 8 , 9 ) that in alphabetic languages, phonological decoding is at the core of learning to read words. Put simply, learning how letters (or graphemes) relate to sounds (or phonemes) allows children to begin to learn the skills required to access the spoken form of a word from its written form. This takes time to develop and requires instruction and practice. Initially, decoding attempts may be only partially correct and certainly will be effortful; with practice in applying their knowledge of grapheme-phoneme relations to read words, children’s decoding skills improve and reading becomes more fluent. According to Ehri’s phase theory of reading development, learning to decode is a connection-forming process in which the spelling patterns of words become tightly bonded with their pronunciations. These unitised representations are retained in memory, supporting efficient visual word recognition and access to meaning. 8

Share’s self-teaching hypothesis also has phonological decoding at the foundation of learning to read—indeed, he describes it as indispensible and absolutely necessary: the sine qua non of reading acquisition. 9 Why is this the case? According to Share, although phonological decoding might initially be effortful and laborious, by forcing the translation from print to sound, it provides an opportunity to acquire word-specific orthographic information about the word, its spelling pattern and its pronunciation. This will then be available on future encounters with the word, lessening the reliance on overt and effortful phonological decoding. As well as word-specific knowledge, this process supplies children with a means to gradually accumulate knowledge about how their orthography (that is, their writing system) works. This might include knowledge of regularities and sub-regularities, orthographic conventions and exceptions to those conventions, statistics, which sum over time to provide each child with their own experience-based database of orthographic knowledge. A good deal of evidence supports the central aspects of the self-teaching hypothesis, including its critical foundation in phonological decoding and how it facilitates word reading development; 10 , 11 , 12 it is also supported by a computational implementation. 13

The self-teaching hypothesis describes how a small system can expand rapidly. It is, however, largely silent as to how orthographic expertise develops. The basics of word reading in hand, the task ahead is nevertheless enormous: with just 26 letters to represent the many thousands of written words we encounter, the amount of orthographic overlap between words is considerable. 14 Experiments with adults show how words interact and compete with each other during processing, pointing to a system that is tuned to be highly efficient at getting us from print to meaning quickly. 2 In terms of development, as overt and serial decoding declines, more automatic and parallel phonological activation from print emerges and indeed, this remains a stable feature of skilled word recognition. 15 , 16 , 17 Alongside this, development brings critical changes in orthographic processing, with evidence that coarse-grained orthographic coding increases with reading level, as the system becomes more adult like. 15 It is likely that children also develop increased sensitivity to morphological complexities and regularities, allowing them to capitalise on the relationships between a word’s morphological structure and its spelling. 18 , 19 , 20 Clearly, something develops as children move from novice to expert, and while this has its roots in phonological decoding, much more research on the development of orthographic expertise is needed. 10

Reading experience and the development of lexical quality

Perfetti’s lexical quality hypothesis 6 , 21 defines lexical quality as the extent to which a word's mental representation specifies its spelling, sound and meaning. High quality representations contain tightly bound orthographic, phonological and semantic constituents that together comprise a word’s identity. Higher quality representations are considered to be more fully-specified, more stable and less context-bound than those of lower quality. As a result, they support efficient word identification during reading, freeing cognitive resources for the ultimate purpose of reading: comprehension.

The lexical quality hypothesis argues that for all of us, there are words we know well and others we know less well. Greater expertise is associated with a higher mean lexical quality: on average, adults will have a higher mean lexical quality than young children. An attractive feature of the lexical quality hypothesis is that it unites ideas about knowledge with ideas about cognitive processing. Knowledge is represented by lexical quality and differences in lexical quality lead to differences in processing. In turn, effective processing provides opportunities to gain knowledge and this serves to further tune lexical quality to the benefit of future processing. In this way, lexical quality is both a cause and a consequence of individual and developmental differences in reading skill, although the mechanisms that bring about change in lexical quality are not yet detailed. 6

This brings us to what has to be the broad answer to the novice-expert question: experience. Reading is a skill and like other skills, practice is critical to gaining expertise. Once phonological decoding is in place, practice allows basic skills to be honed and reading experience provides the substrate from which lexical processes can be tuned to the specific orthography being learned. This fits with the finding that print exposure (estimates of how much an individual reads) is a powerful predictor not just of reading outcomes in children, 22 but of word reading processes in skilled adults too. 23 , 24

The lexical legacy hypothesis

Reading experience provides opportunities to refine knowledge about orthography-phonology mappings. Importantly though, it provides much more than repeated exposure to individual words: words are usually encountered in meaningful sentences, stories and texts. What is the relevance of this type of experience? The lexical legacy hypothesis sees it as critical to variations in lexical quality. At its heart is the idea that reading (and spoken language) provides many different contexts, episodes and experiences which, over time, sum to a rich and nuanced database about a word, its connections to other words and its lexical history within an individual’s experience. The hypothesis suggests that these rich and diverse encounters bring about local variation at the word level: a lexical legacy that is measurable during word reading behaviour.

This account is related in spirit to theories of word knowledge based on lexical co-occurrence and the principle that “you shall know a word by the company it keeps.” 25 Mathematical models of word knowledge based on latent semantic analysis demonstrate the utility and psychological validity of this statistical approach to meaning, in which words come to occupy a position in semantic space, based on encounters during the course of language experience. Put simply, a word’s position in semantic space at any one point in time, relative to other words, captures its meaning (for reviews see refs 26 , 27 ). The extension of this approach to the development of word reading (rather than meaning) is speculative, but several lines of evidence converge to suggest it is one worth exploring. To illustrate its utility, consider the frequency effect.

Frequency (how often a word appears in a language corpus) enjoys special status as a powerful item-level predictor of lexical processing. 28 , 29 It is represented in models of word recognition in various ways, consistent with the notion that seeing a word more frequently adjusts its recognition threshold so that it is more easily processed on subsequent encounters. Clearly, frequency is the product of cumulative experience. Some words are seen more often than others. Less clear, however, is whether the frequency effect arises solely from variations in repetition. How often a word appears in a corpus is correlated with many other factors, including the local semantic and syntactic contexts in which the word appears throughout the corpus. This type of linguistic co-occurrence predicts the frequency effect, 30 suggesting that frequency might serve as an umbrella for complex lexical experience. On this view, frequency is a powerful predictor of reading as it subsumes other features, co-captured by experience, 31 as well as being an indicator of repetition in and of itself.

Consistent with a word’s frequency representing more than just the number of times it is likely to have been seen, the number of unique documents a word appears in is more closely associated with lexical processing in adults than raw frequency. 32 And, it seems likely that it is the semantic diversity of those different contexts that matters most: words experienced in more varied semantic contexts enjoy a processing advantage in word reading and lexical decision, 31 , 33 , 34 , 35 , 36 relative to words of equivalent frequency that occur in more redundant contexts (lexical decision is a commonly used psycholinguistic task in which participants respond yes if a stimulus is a word and no if it is a pseudoword). One way to interpret this finding, supported by both behavioural and computational evidence, 33 , 36 is that change is needed to bring about learning. Simply repeating a word in isolation or across identical documents will not update a word’s lexical history; in contrast, differences in linguistic environment associated with a changing semantic context will cause the word’s lexical representation to be updated, and so enhance learning.

This interpretation of the frequency effect requires us to rethink what frequency captures and why it is an important item-level predictor of word reading. To return to the lexical legacy hypothesis, semantic diversity might be relevant beyond raw frequency as it captures the linguistic environment a word has been experienced in, with variations in this being reflected in lexical quality. Other item-level predictors might exert their influence for similar reasons. Consider classic semantic variables such as imageability, number of semantic features and number of senses. These variables are also associated with the ease of word recognition in skilled readers. 28 , 37 , 38 , 39 This might be the measurable legacy that follows from reading experience, where instances with words in meaningful text brings about differences in lexical quality.

Questions to ask about the lexical legacy hypothesis

The alphabetic principle underpins word reading development. Once the basics are in place, further development comes from reading experience. The lexical legacy hypothesis helps us to understand how this input allows distributional information to be massed over time, influencing reading skill. A number of features make this hypothesis attractive. It offers a means by which differences in lexical quality emerge. It helps us to understand the relationship between word reading skill and print exposure. It also forges direct links between lexical learning and lexical processing: how easily a word is processed, even by skilled adults, is a product of the learning opportunities afforded by an individual’s lexical experience. It is, however, speculative and under-specified. I end by setting out some questions that need to be addressed.

Most current data investigating the relationship between linguistic experience and lexical processing are limited by their correlational nature, as in the correlation between print exposure and reading development noted earlier, for example refs 22 – 24 . To move beyond this, studies that explicitly manipulate and control variables are needed. Encouragingly, training experiments with adults suggest that linguistic diversity supports learning. 33 , 36 , 40 Extending this work to children will shed light on whether diversity has relevance for how children learn written words. If the lexical legacy hypothesis is correct, experiencing words in more diverse linguistic contexts should bring about better learning of orthographic forms. Training studies that control for frequency of exposure to each new word while manipulating the number or nature of contexts or episodes each is experienced in will be particularly informative. For example, children could read some novel words embedded in a series of stories where the context varies from story to story. If diversity is more critical than repetition, experiencing words in different contexts should result in better learning than when the novel words are encountered the same number of times, but in non-diverse contexts. A related set of questions is what is meant by diverse contexts: is it variation in the number of contexts (for example, the number of different books a child sees a word in) or the temporal spacing of the encounters (for example, three times on one day vs. once per day over three days). Or, is it more about the nature of the linguistic context that characterises each encounter, and the similarity of those contexts to previous lexical experiences? Again, carefully designed training experiments have utility here, as do computational models.

Clearly, word knowledge is multifaceted. We may wish to know how well a child has learned the written form of the word; or, we may be more interested in how well meaning has been acquired; or, we may be interested in how readily word meaning is activated from written forms when children read words in text. Regardless of our question of focus, sensitive measures of learning will be needed to tap knowledge that is partial and incremental, as it builds over time with each exposure. 10 , 41 , 42 Comparing learning across different exposure conditions should then reveal what type of experience is optimal to bring about learning. Using this type of design, a recent training experiment found that children better learned the meaning of new verbs when they were experienced in episodes built around a common scenario. 43 Potentially, this type of contextual experience promoted semantic connections between words and in doing so, promoted learning of meaning.

Another question is whether it is reading experience that matters for providing diversity, rather than language experience more generally. While spoken language experience is of course relevant, there are reasons to propose that experience with text also plays a role. Once children are able to read, the majority of new vocabulary is learned via reading, not listening. 44 Even text written for young children is more lexically diverse than speech, 45 and there are differences in syntax too, 46 consistent with the idea that experience with text affords unique learning opportunities. And text is clearly needed for children to learn and refine their knowledge about how spelling patterns relate to spoken language: word reading skill demands efficient mappings between orthography and phonology, 47 as well as orthography and morphology, 18 , 19 tuned to the individual’s language system. 48 One way to investigate the impact of spoken vs. written language experience on reading development would be to extract lexical statistics such as frequency and semantic diversity from corpora that sample either children’s spoken or written language experience, respectively. If linguistic experience gained via text is important, the item-level association between reading behaviour and text-corpora statistics should be closer than its association with spoken-corpora statistics.

In closing, it is important to emphasise what the hypothesis is not saying. It is not the case that young children learn to read words by contextual guessing, as is clear from a large evidence base. 49 , 50 , 51 , 52 To the contrary, the foundation of learning to read in English is the alphabetic principle and from this, the development of high quality phonological decoding skill. 1 , 8 , 9 This provides the means for orthographic learning—the gradual accumulation of orthographic knowledge, via reading experience. 9 , 10 , 11 , 12 Building on this, the lexical legacy hypothesis situates expertise as the product of reading experience in a broader sense. How often words occur, how they are used, how they look and sound, what they come to mean and how they relate to other words all feed into a dynamic database of knowledge, continuously updated by experience. Reading behaviour, for an individual word averaged over people, or an individual person averaged over words, is the product of this rich experience at that point in time.

Byrne, B. The Foundations of Literacy: The Child’s Acquisition of the Alphabetic Principle . (Psychology Press, 1998).

Adelman, J. A. Visual Word Recognition, Vol. 1 (Psychology Press, 2012).

Taylor, J. S. H., Rastle, K. & Davis, M. H. Can cognitive models explain brain activation during word and pseudoword reading? A meta-analysis of 36 neuroimaging studies. Psychol. Bull. 139 , 766–791 (2013).

Article   CAS   PubMed   Google Scholar  

Dehaene, S. & Cohen, L. The unique role of the visual word form area in reading. Trends Cogn. Sci. 15 , 254–262 (2011).

Article   PubMed   Google Scholar  

Dehaene, S. et al. How learning to read changes the cortical networks for vision and language. Science 330 , 1359–1364 (2010).

Perfetti, C. Reading ability: Lexical quality to comprehension.   Sci. Stud. Read. 11 , 357–383 (2008).

Article   Google Scholar  

Perfetti, C. & Stafura, J. Word knowledge in a theory of reading comprehension. Sci. Stud. Read. 18 , 22–37 (2014).

Ehri, L. C. Learning to read words: Theory, findings, and issues. Sci. Stud. Read. 9 , 167–188 (2005).

Share, D. L. Phonological recoding and self-teaching - sine-qua-non of reading acquisition. Cognition. 55 , 151–218 (1995).

Nation, K. & Castles, A. Putting the learning in to orthographic leanring. (eds K. Cain, D. Compton and R. Parrila), John Benjamins (in the press).

Castles, A. & Nation, K. in From Inkmarks to Ideas: Current Issues in Lexical Processing (ed. S. Andrews) 151–179 (2006).

McClung, N. A., O’Donnell, C. R. & Cunningham, A. E. in Visual Word Recognition Vol. 1 (ed. J. A. Adelman) (Psychology Press, 2012).

Ziegler, J. C., Perry, C. & Zorzi, M. Modelling reading development through phonological decoding and self-teaching: implications for dyslexia. Philos. Trans. R. Soc. B-Biol. Sci. 369 . doi: 10.1098/rstb.2012.0397 (2014).

Davis, C. J. in Visual Word Recognition Vol. 1 (ed. J. A. Adelman) 185–206 (Psychology Press, 2012).

Ziegler, J. C., Bertrand, D., Lete, B. & Grainger, J. Orthographic and phonological contributions to reading development: tracking developmental trajectories using masked priming. Dev. Psychol. 50 , 1026–1036 (2014).

Grainger, J., Lete, B., Bertand, D., Dufau, S. & Ziegler, J. C. Evidence for multiple routes in learning to read. Cognition. 123 , 280–292 (2012).

Jared, D., Ashby, J., Agauas, S. J. & Levy, B. A. Phonological activation of word meanings in grade 5 readers. J. Exp. Psychol-Learn. 42 , 524–541 (2016).

Rastle, K. & Davis, M. H. Morphological decomposition based on the analysis of orthography. Lang. Cogn. Process. 23 , 942–971 (2008).

Beyersmann, E., Castles, A. & Coltheart, M. Morphological processing during visual word recognition in developing readers: evidence from masked priming. Q. J. Exp. Psychol. 65 , 1306–1326 (2012).

Beyersmann, E., Grainger, J., Casalis, S. & Ziegler, J. C. Effects of reading proficiency on embedded stem priming in primary school children. J. Exp. Child. Psychol. 139 , 115–126 (2015).

Perfetti, C. A. & Hart, L. in Precursors of Functional Literacy (eds L. Vehoeven et al. ) 189–213 (John Benjamins, 2002).

Mol, S. E. & Bus, A. G. To read or not to read: a meta-analysis of print exposure from infancy to early adulthood. Psychol. Bull. 137 , 267–296 (2011).

Chateau, D. & Jared, D. Exposure to print and word recognition processes. Mem. Cognit. 28 , 143–153 (2000).

Acheson, D. J., Wells, J. B. & MacDonald, M. C. New and updated tests of print exposure and reading abilities in college students. Behav. Res. Methods. 40 , 278–289 (2008).

Article   PubMed Central   Google Scholar  

Firth, J. R. Papers in Linguistics (Oxford University Press, 1957).

Landauer, T. K. & Dumais, S. T. A solution to Plato’s problem: the latent semantic analysis theory of acquisition, induction, and representation of knowledge. Psychol. Rev. 104 , 211–240 (1997).

Landauer, T. K. Latent semantic analysis: a theory of the psychology of language and mind. Discourse Process. 27 , 303–310 (1999).

Balota, D. A., Cortese, M. J., Sergent-Marshall, S. D., Spieler, D. H. & Yap, M. J. Visual word recognition of single-syllable words. J. Exp. Psychol. -Gen. 133 , 283–316 (2004).

Brysbaert, M. & New, B. Moving beyond Kucera and Francis: a critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English. Behav. Res. Methods 41 , 977–990 (2009).

Baayen, R. H. Demythologizing the word frequency effect: a discriminative learning perspective. Ment. Lex. 5 , 436–461 (2010).

McDonald, S. A. & Shillcock, R. C. Rethinking the word frequency effect: the neglected role of distributional information in lexical processing. Lang. Speech 44 , 295–323 (2001).

Adelman, J. S., Brown, G. D. A. & Quesada, J. F. Contextual diversity, not word frequency, determines word-naming and lexical decision times. Psychol. Sci. 17 , 814–823 (2006).

Jones, M. N., Johns, B. T. & Recchia, G. The role of semantic diversity in lexical organization. Can. J. Exp. Psychol. 66 , 115–124 (2012).

Hoffman, P. & Woollams, A. M. Opposing effects of semantic diversity in lexical and semantic relatedness decisions. J. Exp. Psychol.-Human 41 , 385–402 (2015).

Plummer, P., Perea, M. & Rayner, K. The influence of contextual diversity on eye movements in reading. J. Exp. Psychol. -Learn. 40 , 275–283 (2014).

Johns, B. T., Dye, M. & Jones, M. N. The influence of contextual diversity on word learning. Psychon. Bull. Rev. (in the press).

Taylor, J. S. H., Duff, F. J., Woollams, A. M., Monaghan, P. & Ricketts, J. How word meaning influences word reading. Curr. Dir. Psychol. Sci. 24 , 322–328 (2015).

Pexman, P. M. in Visual Word Recognition Vol. 2 (ed. J. S. Adeleman) 24–43 (Psychology Press, 2012).

Nation, K. Form-meaning links in the development of visual word recognition. Philos. Trans. R. Soc. B-Biol. Sci. 364 , 3665–3674 (2009).

Bolger, D. J., Balass, M., Landen, E. & Perfetti, C. A. Context variation and definitions in learning the meanings of words: an instance-based learning approach. Discourse Process 45 , 122–159 (2008).

Tamura, N., Castles, A. & Nation, K. Orthographic learning, fast and slow: lexical competition effects reveal the time course of word learning in developing readers. (Submitted).

Joseph, H. S. S. L., Wonnacott, E., Forbes, P. & Nation, K. Becoming a written word: Eye movements reveal order of acquisition effects following incidental exposure to new words during silent reading. Cognition 133 , 238–248 (2014).

Chilton, M. W. & Ehri, L. C. Vocabulary learning: sentence contexts linked by events in scenarios facilitate third graders’ memory for verb meanings. Read. Res. Quart. 50 , 439–458 (2015).

Google Scholar  

Nagy, W. E., Herman, P. A. & Anderson, R. C. Learning words from context. Read. Res. Quart. 20 , 233–253 (1985).

Montag, J. L., Jones, M. N. & Smith, L. B. The words children hear: picture books and the statistics for language learning. Psychol. Sci. 26 , 1489–1496 (2015).

Article   PubMed   PubMed Central   Google Scholar  

Montag, J. L. & MacDonald, M. C. Text exposure predicts spoken production of complex sentences in 8-and 12-year-old children and adults. J. Exp. Psychol. -Gen. 144 , 447–468 (2015).

Ziegler, J. C. & Goswami, U. Reading acquisition, developmental dyslexia, and skilled reading across languages: a psycholinguistic grain size theory. Psychol. Bull. 131 , 3–29 (2005).

Ziegler, J. C. & Goswami, U. Becoming literate in different languages: similar problems, different solutions. Dev. Sci. 9 , 429–436 (2006).

Landi, N., Perfetti, C. A., Bolger, D. J., Dunlap, S. & Foorman, B. R. The role of discourse context in developing word form representations: A paradoxical relation between reading and learning. J. Exp. Child. Psychol. 94 , 114–133 (2006).

Nation, K. & Snowling, M. Individual differences in contextual facilitation: evidence from dyslexia and poor reading comprehension. Child. Dev. 69 , 996–1011 (1998).

Stanovich, K. E., Nathan, R. G., West, R. F. & Valarossi, M. Childrens word recognition in context—spreading activation, expectancy, and modularity. Child. Dev. 56 , 1418–1428 (1985).

Stanovich, K. E., Cunningham, A. E. & Feeman, D. J. Relation between early reading acquisition and word decoding with and without context—a longitudinal-study of 1st-grade children. J. Educ. Psychol. 76 , 668–677 (1984).

Download references


I thank Niina Tamura for thoughtful discussion and comments on an earlier draft. This work was supported by grants from The Economic and Social Research Council (ES/M009998/1) and The Leverhulme Trust (RPG-2015-070).

Competing interests

The authors declare no competing interests.

Author information

Authors and affiliations.

University of Oxford, Oxford, UK

Kate Nation

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Kate Nation .

Rights and permissions

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Reprints and permissions

About this article

Cite this article.

Nation, K. Nurturing a lexical legacy: reading experience is critical for the development of word reading skill. npj Science Learn 2 , 3 (2017). https://doi.org/10.1038/s41539-017-0004-7

Download citation

Received : 09 June 2016

Revised : 04 December 2016

Accepted : 13 December 2016

Published : 27 January 2017

DOI : https://doi.org/10.1038/s41539-017-0004-7

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

The influence of oral vocabulary knowledge on individual differences in a computational model of reading.

  • Ya-Ning Chang

Scientific Reports (2023)

How phonological and orthographic decoding complicates the simple view of reading in Chinese: examining mediation through listening comprehension

  • Haomin Zhang

Cognitive Processing (2023)

Development of silent reading fluency and reading comprehension across grades 1 to 9: unidirectional or bidirectional effects between the two skills?

  • Maria Psyridou
  • Asko Tolvanen
  • Minna Torppa

Reading and Writing (2023)

Structural differences in the semantic networks of younger and older adults

  • Dirk U. Wulff
  • Thomas T. Hills

Scientific Reports (2022)

Sensitivity to meaningful regularities acquired through experience

  • Anastasia Ulicheva
  • Marco Marelli
  • Kathleen Rastle

Morphology (2021)

Quick links

  • Explore articles by subject
  • Guide to authors
  • Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

what is lexical hypothesis

Help | Advanced Search

Computer Science > Computation and Language

Title: deep lexical hypothesis: identifying personality structure in natural language.

Abstract: Recent advances in natural language processing (NLP) have produced general models that can perform complex tasks such as summarizing long passages and translating across languages. Here, we introduce a method to extract adjective similarities from language models as done with survey-based ratings in traditional psycholexical studies but using millions of times more text in a natural setting. The correlational structure produced through this method is highly similar to that of self- and other-ratings of 435 terms reported by Saucier and Goldberg (1996a). The first three unrotated factors produced using NLP are congruent with those in survey data, with coefficients of 0.89, 0.79, and 0.79. This structure is robust to many modeling decisions: adjective set, including those with 1,710 terms (Goldberg, 1982) and 18,000 terms (Allport & Odbert, 1936); the query used to extract correlations; and language model. Notably, Neuroticism and Openness are only weakly and inconsistently recovered. This is a new source of signal that is closer to the original (semantic) vision of the Lexical Hypothesis. The method can be applied where surveys cannot: in dozens of languages simultaneously, with tens of thousands of items, on historical text, and at extremely large scale for little cost. The code is made public to facilitate reproduction and fast iteration in new directions of research.

Submission history

Access paper:.

  • Other Formats

license icon

References & Citations

  • Google Scholar
  • Semantic Scholar

BibTeX formatted citation

BibSonomy logo

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

  • Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

  • Reviews / Why join our community?
  • For companies
  • Frequently asked questions

Lexical Activation Hypothesis

Literature on lexical activation hypothesis.

Here’s the entire UX literature on Lexical Activation Hypothesis by the Interaction Design Foundation, collated in one place:

Learn more about Lexical Activation Hypothesis

Take a deep dive into Lexical Activation Hypothesis with our course AI for Designers .

In an era where technology is rapidly reshaping the way we interact with the world, understanding the intricacies of AI is not just a skill, but a necessity for designers . The AI for Designers course delves into the heart of this game-changing field, empowering you to navigate the complexities of designing in the age of AI. Why is this knowledge vital? AI is not just a tool; it's a paradigm shift, revolutionizing the design landscape. As a designer, make sure that you not only keep pace with the ever-evolving tech landscape but also lead the way in creating user experiences that are intuitive, intelligent, and ethical.

AI for Designers is taught by Ioana Teleanu, a seasoned AI Product Designer and Design Educator who has established a community of over 250,000 UX enthusiasts through her social channel UX Goodies. She imparts her extensive expertise to this course from her experience at renowned companies like UiPath and ING Bank, and now works on pioneering AI projects at Miro.

In this course, you’ll explore how to work with AI in harmony and incorporate it into your design process to elevate your career to new heights. Welcome to a course that doesn’t just teach design; it shapes the future of design innovation.

In lesson 1, you’ll explore AI's significance, understand key terms like Machine Learning, Deep Learning, and Generative AI, discover AI's impact on design, and master the art of creating effective text prompts for design.

In lesson 2, you’ll learn how to enhance your design workflow using AI tools for UX research, including market analysis, persona interviews, and data processing. You’ll dive into problem-solving with AI, mastering problem definition and production ideation.

In lesson 3, you’ll discover how to incorporate AI tools for prototyping, wireframing, visual design, and UX writing into your design process. You’ll learn how AI can assist to evaluate your designs and automate tasks, and ensure your product is launch-ready.

In lesson 4, you’ll explore the designer's role in AI-driven solutions, how to address challenges, analyze concerns, and deliver ethical solutions for real-world design applications.

Throughout the course, you'll receive practical tips for real-life projects. In the Build Your Portfolio exercises, you’ll practise how to  integrate AI tools into your workflow and design for AI products, enabling you to create a compelling portfolio case study to attract potential employers or collaborators.

All open-source articles on Lexical Activation Hypothesis

The self-generation effect: how to create more memorable user interfaces.

what is lexical hypothesis

  • 3 years ago

Open Access—Link to us!

We believe in Open Access and the  democratization of knowledge . Unfortunately, world-class educational materials such as this page are normally hidden behind paywalls or in expensive textbooks.

If you want this to change , cite this page , link to us, or join us to help us democratize design knowledge !

Privacy Settings

Our digital services use necessary tracking technologies, including third-party cookies, for security, functionality, and to uphold user rights. Optional cookies offer enhanced features, and analytics.

Experience the full potential of our site that remembers your preferences and supports secure sign-in.

Governs the storage of data necessary for maintaining website security, user authentication, and fraud prevention mechanisms.

Enhanced Functionality

Saves your settings and preferences, like your location, for a more personalized experience.

Referral Program

We use cookies to enable our referral program, giving you and your friends discounts.

Error Reporting

We share user ID with Bugsnag and NewRelic to help us track errors and fix issues.

Optimize your experience by allowing us to monitor site usage. You’ll enjoy a smoother, more personalized journey without compromising your privacy.

Analytics Storage

Collects anonymous data on how you navigate and interact, helping us make informed improvements.

Differentiates real visitors from automated bots, ensuring accurate usage data and improving your website experience.

Lets us tailor your digital ads to match your interests, making them more relevant and useful to you.

Advertising Storage

Stores information for better-targeted advertising, enhancing your online ad experience.

Personalization Storage

Permits storing data to personalize content and ads across Google services based on user behavior, enhancing overall user experience.

Advertising Personalization

Allows for content and ad personalization across Google services based on user behavior. This consent enhances user experiences.

Enables personalizing ads based on user data and interactions, allowing for more relevant advertising experiences across Google services.

Receive more relevant advertisements by sharing your interests and behavior with our trusted advertising partners.

Enables better ad targeting and measurement on Meta platforms, making ads you see more relevant.

Allows for improved ad effectiveness and measurement through Meta’s Conversions API, ensuring privacy-compliant data sharing.

LinkedIn Insights

Tracks conversions, retargeting, and web analytics for LinkedIn ad campaigns, enhancing ad relevance and performance.

LinkedIn CAPI

Enhances LinkedIn advertising through server-side event tracking, offering more accurate measurement and personalization.

Google Ads Tag

Tracks ad performance and user engagement, helping deliver ads that are most useful to you.

Share the knowledge!

Share this content on:

or copy link

Cite according to academic standards

Simply copy and paste the text below into your bibliographic reference list, onto your blog, or anywhere else. You can also just hyperlink to this page.

New to UX Design? We’re Giving You a Free ebook!

The Basics of User Experience Design

Download our free ebook The Basics of User Experience Design to learn about core concepts of UX design.

In 9 chapters, we’ll cover: conducting user interviews, design thinking, interaction design, mobile UX design, usability, UX research, and many more!

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access


Research Article

Liking music with and without sadness: Testing the direct effect hypothesis of pleasurable negative emotion

Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Validation, Visualization, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliation Empirical Musicology Laboratory, School of the Arts and Media, UNSW Australia, Sydney, NSW, Australia

ORCID logo

  • Emery Schubert


  • Published: April 10, 2024
  • https://doi.org/10.1371/journal.pone.0299115
  • Reader Comments

Table 1

Negative emotion evoked in listeners of music can produce intense pleasure, but we do not fully understand why. The present study addressed the question by asking participants (n = 50) to self-select a piece of sadness-evoking music that was loved. The key part of the study asked participants to imagine that the felt sadness could be removed. Overall participants reported performing the task successfully. They also indicated that the removal of the sadness reduced their liking of the music, and 82% of participants reported that the evoked sadness also adds to the enjoyment of the music. The study provided evidence for a “Direct effect hypothesis”, which draws on the multicomponent model of emotion, where a component of the negative emotion is experienced as positive during music (and other aesthetic) experiences. Earlier evidence of a mediator, such as ‘being moved’, as the source of enjoyment was reinterpreted in light of the new findings. Instead, the present study applied a semantic overlap explanation, arguing that sadness primes emotions that share meaning with sadness, such as being-moved. The priming occurs if the overlap in meaning is sufficient. The degree of semantic overlap was defined empirically. The present study therefore suggests that mediator-based explanations need to be treated with caution both as a finding of the study, and because of analytic limitations in earlier research that are discussed in the paper.

Citation: Schubert E (2024) Liking music with and without sadness: Testing the direct effect hypothesis of pleasurable negative emotion. PLoS ONE 19(4): e0299115. https://doi.org/10.1371/journal.pone.0299115

Editor: Maja Vukadinovic, Novi Sad School of Business, SERBIA

Received: December 5, 2023; Accepted: February 5, 2024; Published: April 10, 2024

Copyright: © 2024 Emery Schubert. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: Data contain potentially identifying or sensitive participant information because open ended responses about personal experiences to music could have been reported. The decision to restrict data sharing was part of the approval given by the institutional ethics committee. The email contact for the institutional ethics advisory committee that granted approval for this design is [email protected] .

Funding: Initials of the authors who received each award: ES Grant numbers awarded to each author: FT120100053 (ES) The full name of each funder: Australian Research Council URL of each funder website: https://www.arc.gov.au/ Did the sponsors or funders play any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript?: No.

Competing interests: The authors have declared that no competing interests exist.


A considerable portion of the population (estimates ranging from around 25% to 50%) will report that music they love can also make them feel negative emotions such as sadness [ 1 – 6 ]. This finding has mystified researchers. How can a loved activity simultaneously produce a negative feeling, and yet lead the same individual to eagerly seek out the experience?

The Indirect effect hypothesis

Much theorising has been proposed to explain the conundrum as it applies to music listening and the contemplation of the arts in general. A dominating approach argues that the ‘sadness’ (the negative emotion that is the focus of the current investigation, and one that has received much attention) evoked by the music serves some non-negative purpose. The negative emotion is not in and of itself enjoyed. We will refer to such explanations as part of the ‘Indirect effect hypothesis’, meaning that a negative emotion such as sadness itself cannot or should not directly play a role in the generation of pleasure. The Indirect effect hypothesis is old, with written origins in Aristotle’s concept of catharsis from 4 th century BCE–where certain negative emotions in response to the arts act as a psychic cleanser, which removes bad or negative emotions from the soul [ 7 , 8 ]. The enduring concept of catharsis suggests an Indirect effect hypothesis because the negative emotion itself is not enjoyed directly. Rather, it is the cleansing, or the product of the cleansing that feels good. (Please note that in this article, the terms enjoyment, pleasure, feels-good, preferred, loved and liked are treated, more or less, as substitutable synonyms; see [ 9 ]) The negative impact of the emotion is thus compensated for by the positive effect on the soul or, in early 21 st century parlance, the mind.

A more recent version of the Indirect effect hypothesis is that sadness produces pleasure indirectly by triggering an intermediary step, sometimes referred to as a ‘mediator’. ‘Being moved’, for example, has been reported as the underlying reason for listening to otherwise sad music. Being moved can be seen as consisting of positive aspects, in addition to negative aspects [ 10 – 13 ]. It is the positive aspects of being moved that are responsible for the pleasure of the otherwise sadness-inducing music. Such explanations argue that the negative emotion occurs alongside a mediator, and so itself is not the direct cause of the positive aspects of the experience, thus eradicating the paradoxical aspect of the phenomenon.

A common technique to test the Indirect effect hypothesis is to ask participants to listen to a piece of music and rate the felt sadness and enjoyment experienced, in addition to rating the alleged mediator. If the enjoyment ratings are correlated with the mediator, and provided this correlation is overall stronger than is sadness with enjoyment, we have evidence, albeit correlational, that the mediator is the direct cause of the liking, not the sadness, supporting the Indirect effect hypothesis. To date, being moved has produced the strongest evidence of mediating sadness [ 3 , 14 – 16 ]. But other contenders that have been proposed, including beauty, wonder and nostalgia [for an overview, see 3 , 17 ].

Limitations of the Indirect effect hypothesis

An inherent weakness of Indirect effect hypothesis, and in particular the mediator-based explanation, is that it does not consider the phenomenal experience of the individual who claims that they both experience sadness, and that the sadness itself, for them, forms at least part of the pleasure [e.g., 6 ]. There are also limitations with research methods that are used to test the mediator explanation in the extant literature, as elucidated in the Method section.

Another limitation specifically concerns the mediator driven approach because it does not explain why the negative emotion would be present at all if it is the mediator that is driving the pleasure. If music is pleasurable because it is moving, and not because it evokes sadness, why would the listener not just seek the music that is moving but not sadness evoking? Is it because the mediator generates the negative (sad) emotion, as a by-product? But this would suggest that the occurrence of enjoyed negative emotion experiences such as sadness in response to music should be nothing more than an outlier, and be rarely reported as an enjoyed part of the experience (presumably well under the 25% of reports that are typical of published research, as indicated at the Introduction). Mediation theory therefore only explains why listeners claim to enjoy felt negative emotions to a limited extent. An alternative explanation is worth considering, and here the Direct effect hypothesis is proposed.

The Direct effect hypothesis

The Direct effect hypothesis argues that there is something intrinsic about felt negative emotion evoked by music that attracts the listener, without mandating a mediator or some factor outside the negative emotion itself. The presence of accompanying affects (such as being moved) are not excluded, but they are not essential. One line of research that supports this hypothesis is the link between individual differences and enjoyment of sad music. Such research does not exclude the Indirect effect account, but it does suggest that individual factors attract the listener to sadness in music, raising the possibility that there is something peculiar about some negative emotions that allow them to be enjoyed in their own right.

Strong contenders for the disposition of people who enjoy the sadness evoked by music are empathisers, fantasisers, ruminators, those who demonstrate an openness to experience, and those with a high propensity to fall into states of absorption [ 2 , 3 , 16 , 18 – 22 ]. Current thinking is that these personal characteristics, especially empathising, absorption and openness to experience, allow the individual to connect with fictional narratives while suspending disbelief, and so exhibit a good capacity to “make-believe” [ 23 , 24 ], a capacity which generalises to emotions in music listening [e.g., see 16 , 25 – 27 ]. This explanation also presents an alternative theoretical perspective to the above cited literature, because rather than presenting sadness as a mere by-product of mediation or as a means to some beneficial end, the sadness can be ‘enjoyed’ for its own sake (directly). It is not real-sadness, but a make-believe, or aesthetic, kind of sadness, still experienced as sadness, but with some real-life negative aspect of the sadness not triggered [ 28 ].

The Direct effect hypothesis has a theoretical foundation. Emotion researchers such as Frijda [ 29 ] and Scherer [ 30 ] have conceptualised emotion as consisting of multiple phases or components operating in synchrony. This view is both reflective of contemporary understandings of emotion, and defined networks in the brain. In one instantiation of a componential model, Sander, Grandjean and Scherer [ 31 ] proposed five components/networks of emotion building on Scherer’s model: ‘Expression’ (e.g., a facial expression that communicates the emotion), ‘Action Tendency’ (e.g., motivation to approach toward, or flee from the cause of the emotion), ‘Autonomic Reaction’ (e.g., changed heart rate), ‘Feeling’ (what the emotion feels-like, such as ‘I feel sadness’) and ‘Elicitation’ (the internally triggered cause of the emotion through interpretation of environmental situation, association and instinct) such as prolonged loneliness eliciting sadness.

In the case of the enjoyment of negative emotions Schubert [ 32 ] proposed that when contemplating aesthetic stimuli the Action tendency component of an emotion is experienced as positive (motivation to approach) while other components remain as they would for real-life, non-aesthetic experiences of such emotions. The individual is not compelled to act in a withdrawn or aversive manner to the stimulus or event under contemplation because the perceiver has an implicit awareness that it is presented in an aesthetic or make-believe context. This dissociated response occurs because the individual has an intrinsic understanding of the safe, make-believe context in which the causal stimulus/event is perceived [ 33 – 35 ].

Limitations of the direct effect hypothesis

The Direct effect hypothesis of enjoyment of negative emotion has arguably been difficult to test. If emotions happen to be correlated (such as sadness and being moved), researchers typically take this as an indication in favour of the Indirect effect hypothesis. But such interpretations do not exclude the possibility that the enjoyment directly stems from the sadness. While there is some evidence that those who enjoy negative emotion in music are indeed enjoying the negative emotion, there has been little systematic investigation of the experiential aspect of enjoyment of negative emotion in music. Other approaches to falsifying the Direct effect hypothesis are needed.

The approach taken in the present research is in the form of an ‘empirical thought experiment’, which has origins in so-called experimental philosophy [ 36 ]. Thought experiments, also referred to as mental simulation or ‘prefactual thinking’, rely on the participant’s capacity to imagine a situation and provide a response to that situation. The method can be particularly useful when a real-life stimulus-effect manipulation of interest is not possible or ethically compromising [e.g., 37 ]. It has been applied successfully to the empirical investigation of a range or research questions [ 38 ] and, of relevance here, to scenarios involving mental simulation of emotions [ 39 – 42 ].

Probing listeners to mentally simulate manipulating aspects of sadness induced by music is a simple approach to address both the Direct and Indirect effect hypotheses of enjoyment of experienced negative emotion in music. In brief, if a listener reports experiencing the sadness induced by a piece of music as pleasurable, the thought experiment to address the question of interest (to test if the sadness is the cause of the pleasure) is to ask the participant to imagine that the felt sadness, and only the felt sadness, can somehow be removed. If enjoyment is consequently diminished (as a result of the mentally simulated, excised sadness), the Direct effect hypothesis will be supported. Assurances would need to be set in place that the sadness was experienced (felt) and not just expressed by the music [ 43 ], and that the music was responsible for triggering the sadness, not some (extramusical) association (as discussed in the Method section).

The aim of this study was to investigate whether negative emotion in music, in this case sadness, can be both experienced and enjoyed. Two competing hypotheses were tested:

H1 –the Indirect effect hypothesis, which predicts that: Sadness removed from a liked piece of music will increase or not change enjoyment. This is because it is not the sadness that is enjoyed, but something external to the sadness, such as being moved or some other mediator.

H2 –the Direct effect hypothesis predicts that: Sadness removed from a liked piece of sadness will decrease enjoyment. This is because the sadness itself is somehow enjoyed, regardless of the impact of correlated variables (such as being moved, etc.).

Methodological and data analysis issues

This preamble to the method examines four key issues encountered in extant methods and data-analysis conventions stemming from controversy about use of experimenter- versus participant-selected stimuli. These issues are: Confounding extramusical association, Phenomenon of interest, Demand characteristics and Prospective mediators. This is followed by a discussion of problems that have emerged in experimenter-selected stimulus, and, as a result, a justification for the use of participant-selected music is then presented.

Confounding extramusical association.

There has been growing consensus that investigations of enjoyed sadness in music should be assessed through experimenter-selected music. Participant- or ‘self’-selected music has the disadvantage that the music can have personal or other non-musical associations, meaning that it is not the music that is directly responsible for triggering sadness, but previously formed, ‘extramusical’ associations with the music. Self-selected music could therefore lead to confounding extramusical associations that evoke sadness: the music acting as a mere go-between with the external cause of the sadness and the experience of sadness, and therefore potentially lead to false conclusion of negative emotion being caused by the music. Furthermore, self-selected music does not assure that findings would be generalisable to other participants who did not self-select the same piece. Self-selected music is inevitably music that is familiar. Personal meanings and associations with familiar music could well lead to idiosyncratic responses, peculiar to one or a small number of individuals [for a detailed discussion on limitations in use of familiar music, see 44 ].

Although one of the main drivers for using experimenter-selected music is to avoid confounding extramusical associations , it is possible that even for unfamiliar (experimenter-selected) music a participant will have an emotional response to music because it triggers an external factor, rather than emanating from the music itself [ 45 ]. For example, while Day and Thompson [ 46 ] found that familiar music is more successful at evoking visual imagery (and hence increasing the likelihood of extramusical emotional associations), they also observed the important role of fluency, where music that is complex (low in fluency) is more likely to trigger visual imagery than music that is less complex (high in fluency), regardless of familiarity. Furthermore, autobiographical memories have been reported to be triggered by unfamiliar music, although to a lesser extent than familiar music [ 47 , 48 , see also 49 ]. Thus experimenter-selected music can help to diminish the likelihood of data pollution through confounding extramusical associations , even if not eliminate it.

Phenomenon of interest.

Use of unfamiliar music that is rated by an independent panel, or some other means, as evoking sadness and being pleasurable has been proposed to remedy the problem of confounding extramusical association [e.g., 14 , 16 ]. However, this approach also has its shortcomings. Others deciding what music is likely to evoke sadness will not necessarily evoke sadness to a sufficient degree in a randomly sampled participant to address the phenomenon of interest (enjoyment of evoked negative emotion in music). It is well documented that familiar music can evoke stronger emotions than unfamiliar music, with self-selected music being a particularly effective way to elicit the strong emotions [e.g., 43 , 50 – 56 ]. Similarly, others deciding what music someone likes is riddled with problems. Music preference calls into play several factors such as familiarity [ 57 ], making the assumption of an absolute, objective rating of pleasure in response to a given piece of music problematic. This constitutes a considerable drawback of experimenter-selected design because additional precautions need to be taken to assure that participant experiences capture the phenomenon of interest (both strong liking and experiencing of sadness), as discussed below.

Demand characteristics.

Another problem with self-selected music is that it may attract demand characteristics bias. This bias can occur when the participant infers the research question [ 58 , 59 ]. For self-selected music the research objective can be inferred by the participant, in particular if they are asked to select music that they love that also evokes sadness. In this situation, the participant may guess that the study is concerned with enjoyment and experiencing sadness. If consciously or subconsciously they wish to please the experimenter, they may inflate their assessment of the amount of enjoyment the music generates or the amount of sadness it evokes or both. Furthermore, during participant recruiting, if mention is made that people are sought who experience sadness in response to loved music, it is self-evident that the participant pool will be biased, because only those who have the targeted experience are likely to participate, overlooking the opportunity to estimate how common the phenomenon is in a general population.

Prospective mediators.

Overall, the studies adopting experimenter-selected designs have used interval rating scale measurements of the variables of interest (enjoyment, sadness, and the prospective mediator variables, such as being moved). In addition, other variables are rated to help reduce the likelihood that the participant will successfully intuit the aim of the study, and to capture information about alternative, prospective mediators. Interval rating scales have the advantage of being convenient for correlation based data processing procedures, such as statistical mediation analysis [ 60 ].

Problems with experiment-selected designs.

Although research using experimenter-selected music designs have claimed to manage several methodological problems identified in self-selected music designs to address the current research question, as summarised above, experimenter-selected stimuli based approaches nevertheless have their own limitations (some overlapping with self-selected music approaches).

As mentioned above, experimenter-selected music is less likely to evoke strong emotions compared with self-selected music, and so it is possible that a person who is capable of experiencing intense sadness in response to loved music will not have that experience for music selected by the best-intentioned experimenter. Even with self-selected music, some studies have shown that only about one quarter to one third of participants report experiencing negative emotions such as sadness in response to music they love (see Introduction ). Schubert (6) used the self-selection approach while considerably circumventing the problem of demand characteristics. He asked participants to select a piece of music that they love, but not revealing the research interest in negative emotions. As it turned out, about one third (25/73) of the participants spontaneously reported experiencing negative emotions, with specific mention made of sadness in 12/72 (i.e., one sixth of) cases (p. 17). In that study it was not clear, however, whether the sadness emanated from the music itself, or through some confounding extramusical association . Nevertheless the method mitigated demand characteristics bias, and above all, it ensured that the piece selected was highly liked, something which experimenter-selected approaches rarely guarantee. Konečni [ 61 ] also argued that fully-fledged aesthetic experiences in response to music are rare even under regular listening circumstances. Therefore, the phenomenon of interest would occur in an even smaller proportion of cases in studies applying experimenter-selected music, even if the stimuli have been previously screened for sadness evocation and enjoyment by individuals other than the participant them/her/himself.

Another related limitation of studies using experimenter-selected pieces concerns the response format itself, which commonly employs an integer-based rating scale for each of the affective variables of interest. The problem is not the use of rating scales per se , but the tradition of publishing rating scale results. Studies typically report scale (i.e., item) mean (X) and standard deviation (SD) scores, and/or the correlation coefficient (usually the Pearson product moment coefficient, r) for pairs of variables. The chief problem with such reporting is they imply assumptions about the distribution of the responses. Providing these descriptive statistics, and in particular when the data are then applied to parametric statistical analysis procedures, infers that the distribution of the data are normal, have homogenous variance and are linear [ 62 , p. 311]. If these assumptions are taken at face value, it means that the density of responses diminish as data points are located further away from the mean, with the diminution per scale step being more rapid when the standard deviation is small. Consequently, when there is no explicit information provided about the nature of the distribution, the number of responses that meet the criterion for the phenomenon of interest could be relatively small, and risk not providing statistically sufficient power for meaningful analysis. A simple visual diagnosis can be made through scatterplots of felt sadness versus liking ratings. The decision needs to be made as to where the cut off mark is for sadness and liking scores above which count as satisfying the phenomenon of interest .

This weakness in extant research constitutes the most serious problem of the mediation-based explanation, which, to the author’s knowledge, has exclusively employed experimenter-selected stimuli and use of interval rating scales with X/SD/r reporting, assuming that any amount of sadness evoked by a piece of music should be proportionally implicated in its enjoyment. The assumption is incorrect because it asserts that a linear relationship is evidence of the phenomenon of interest . In fact, the phenomenon of interest is not concerned with enjoyed that accompanies low levels of sadness because when sadness levels are low, other reasons for enjoying the music are still perfectly viable. Evidence of this problem is reflected to some extent by the generally low correlations reported between sadness and liking scores, usually with a small effect size [r < .3, see 63 ]. When the correlation coefficient is small, no conclusion can be drawn about the phenomenon of interest because low correlation only reveals a lack of (non-zero) linearity, rather than information about the modality of the bivariate distribution. That is, a small correlation coefficient provides no information regarding the location of the mode of the distribution, or whether a desirable mode (also) exists in the high sadness, high liking region of the distribution.

In short, by not diagnosing the nature of the bivariate response distribution, the analytic approaches adopted for currently available experimenter-selected designs potentially exclude cases of high evoked sadness that accompany high liking, meaning that they have not captured the phenomenon of interest and so cannot make conclusions about it, or should do so with caution. One solution for future research employing ratings for all variables of interest while maintaining the advantages of the experimenter-selected stimuli approach is to recruit a sufficiently large random sample so that enough cases happen to fall in the desired range spontaneously. However, using self-selected music is more efficient because the phenomenon of interest is achieved by categorical self-selection.

Using self-selected stimuli–justification.

With the above arguments, the stimulus self-selection approach can be justified provided some modifications are made to the way the approach has been applied in the past. These are itemised here in six points. Based on the above overview, the main innovations to note are points 2, 3c, 3d and 4. Square bracketed text following each point indicates the main methodological issue(s) discussed above that are addressed by each of the proposed actions.

  • Correspondence used for recruiting participants is not to indicate that the study is concerned with experiencing sadness in music, its enjoyment, or both [as per recommendations by 58 , 59 ]. [Demand characteristics]
  • During the study, request that the participant selects music that is loved, not just liked, to ensure that the desired (high) liking category of music is attained [ 64 ]. [Phenomenon of interest]
  • that the music is highly liked,
  • the sadness is indeed felt,
  • the sadness emanates directly from the music, and not through extramusical association, and
  • the experienced sadness is implicated in the enjoyment of the music. [Confounding extramusical association; Phenomenon of interest]
  • A control condition is employed, for example where instead of requesting sadness-evoking music, music evoking another emotion that is not paradoxical is requested, such as a mediator proposed in previous research. An obvious choice is moving music (that is loved). [Demand characteristics; Phenomenon of interest]
  • A number of affect terms, including sadness and the control condition emotion should be added to a list of emotions rated in both test and control conditions to allow for comparison, and help identify prospective mediators. [Prospective mediators]
  • Since participants are explicitly asked to have potentially powerfully sad emotions evoked, towards the end of the study an additional stimulus is rated that requires evocation of a positive emotion. This satisfies potential ethical concerns where sadness experience could influence mood negatively, and allows the option of further comparisons with affects in the test condition that were prospective mediators. [Prospective mediators]


103 participants, recruited from an English speaking tertiary institution, consisting mostly of undergraduate music students, completed the study. They were randomly assigned to one of the two conditions. Fifty participants were randomly assigned to the Sadness condition and 53 to the Moving condition in a between-subjects design. The research received ethics approval from the UNSW Australia institutional review board Human Research Advisory Panel B: Arts, Architecture, Design and Law. Participants were recruited from June 4, 2021 until June 9, 2021. Consent to participate was provided at the opening of the online survey, with a checkbox selected if the participant agreed to participate. No minors participated in the study.

The Qualtrics survey platform ( https://www.qualtrics.com ) was used for human data collection. Self-selected music was identified through online links searched for and reported within the survey by the participant. The participant used an electronic device, such as a laptop, iPad or tablet. They were encouraged to wear earphones to listen to music, but this was not enforced. Affect terms consisted of a list of terms that are drawn from Schindler, Hosoya [ 65 ] and Schubert [ 66 ], as presented in the Procedures.

Prior to commencing the study, informed consent was requested verbally through the online interface, with all participants being asked to read an online participant information sheet, which included information about being free to withdraw from the study at any time. They were informed that their data would be treated confidentially, and were encouraged to ask questions if needed, and then to indicate if they wished to commence the study. Participants were randomly assigned into a Sadness (test) or Moving (control) condition. We describe the sadness condition here, but the moving condition is identical, except that ‘sad’ and ‘sadness’ is replaced with ‘moved’/’being moved’ and ‘movingness’ (respectively). Otherwise, where grammatically straight-forward ‘[CONDITION]’ is shown, which was replaced by ‘sadness’ or ‘moved’/’being moved’, depending on the assigned condition. After the tasks for the test or control condition were completed, all participants were invited to select another piece, but this time one that made them feel happy. Although this step of the study was completed by all participants, it will be referred to as the Happy ‘condition’ for convenience. The steps of the study are listed below. They followed one another in sequence, and the participant could not return to a step once they had answered the questions in that step and progressed.

  • Participants were asked to self-select a piece that they both loved and that evoked sadness. They were encouraged to think about this for a few minutes if necessary. For those who could not come up with a piece that met these criteria, some alternative pieces were proposed, from which they could select, or, have further opportunity to select another piece. Details of the piece were collected.
  • Enjoyment of the piece was rated: "How much do you like this piece?” (anchors: 0 = dislike it a lot; 100 = like it a lot)
  • Open-ended felt emotions requested: “Please indicate in as much detail as possible any emotions that you feel in response to this piece. Be sure to include [CONDITION], of course.” (Free text response.)
  • Affects felt . 26 felt affect terms were rated on a 3-point scale (A lot, A little, Not felt) on the extent to which each terms was felt. The wording of each terms was presented to the participant as—1: Being absorbed/completely immersed in the music; 2. Anger; 3. A sense of awe; 4. Feeling of beauty; 5. Calm; 6. Chills; 7. Compassion; 8. Empathy; 9. Euphoria; 10. Fear; 11. A feeling that is sublime; 12. Goosebumps; 13. Grief; 14. Happiness; 15. Joy; 16. Being moved; 17. Nostalgia; 18. Peacefulness; 19. Powerful feelings; 20. Release or relief (sometimes referred to as ’Catharsis’); 21. Sadness; 22. Tears/wanting to cry/feeling like crying/actually crying; 23. Tenderness; 24. Transcendence; 25. Tragedy; 26. Wonder.
  • Confirm felt and direct . Confirm that: Affect terms marked as present in the previous step (‘A lot’ or ‘A little’) were (a) felt and (b) that they were triggered directly by the music, not by thoughts, memories, images, etc. (Yes/No for each of (a) and (b)).
  • I would like the piece a LITTLE LESS;
  • It would make NO DIFFERENCE;
  • I would like the piece a LITTLE MORE;
  • I would like the piece a LOT MORE.
  • Affects that add to liking . The same 26 Affect terms listed in step iv were rated on a 3-point scale (Adds to the pleasure, Does not add to the pleasure, Don’t know/not relevant) to assess whether the “the felt emotions add to the liking, pleasure, attraction or enjoyment”.
  • Cooling down. The above procedure was repeated for a self-selected happy piece, but without any ratings of the 26 Affect terms requested (i.e. steps iv, v & vii excluded).
  • Background (age, gender, music background) data were collected after which the participant was thanked and farewelled.

Some researchers, such as [ 67 , 68 ], treat the concepts of affect and emotion as distinct. In the present study the distinction is partly made for the convenience of distinguishing between participant open-ended response in step iii (emotion) versus their selection from a predetermined list of terms in steps iv, v & vii (affect). The term ‘emotion’ rather than ‘affect’ was used in all of these instruction steps because the former term was considered better understood by participants, regardless of whether referred to as emotion or affect in this article.

Data validation

Participant profile by condition..

Inferential tests demonstrated that the Sadness and Moving groups were statistically identical in terms of gender, age and years of music lessons ( Table 1 ). Also comparable across the groups was the overall rating of liking, averaging over 90 on a 0–100 scale, with upper quartiles (Q3) demonstrating a ceiling effect in both conditions which supports the use of self-selected music for generating high levels of pleasure.


  • PPT PowerPoint slide
  • PNG larger image
  • TIFF original image


Check that the emotion was felt and evoked emotion was directly due to the music.

There was overall high confirmation that the emotions were felt (over 96% of participants) and over 90% of participants in both conditions confirmed that the sadness was triggered intrinsically by the music (not triggered by something outside the music). See Table 1 for breakdown by condition. Overall, participants from both conditions were successful at experiencing the target emotion (Sadness or Being moved) and confirmed that, as requested, the music was directly responsible for triggering the emotion, rather than due to some extramusical factor. All participants were retained for further analysis.

Most frequently reported music excerpts.

All participants selected a piece that met the music selection criteria. Although researcher-suggested pieces were prepared in case a participant could not identify a self-selected piece meeting the criteria, none of the participants requested the researcher-suggested option, and so the research-suggested options were never used in the study. A selection of the self-selected items is presented in Table 2 , showing composers/artists reported by at least three participants across the cohort, and listing the works reported at least twice across the cohort. Interesting similarities can be observed across conditions, with composers Beethoven, Chopin and Debussy, and artists Taylor Swift and Bon Iver appearing in the Moving and Sad conditions. Furthermore, for the Beethoven, two pieces were mentioned in both of these conditions: Für Elise and Moonlight Sonata (1st Movement). These selections reflect the shared tastes across the groups, and at the high proportion of musicians, in particular pianists, who participated (all of the more frequently selected Beethoven, Chopin and Debussy pieces were for piano). Table 1 reveals the overall high average years of music lessons reported across the cohort [ 69 ]. These selections also indicate the capacity for the same piece of music to evoke different emotions (being moving and sadness).



Emotion profile of sad music: Open-ended

After selection of a piece in their assigned condition, participants were asked to provide free descriptions of the emotions they felt in response to the selected piece (self-selected sad or self-selected moving music). The reported terms were pre-processed by identifying all reported emotion terms (participants could report more than one), correcting spelling mistakes, checking context and lemmatizing terms. This was followed by a frequency count of these terms for each condition. The target emotion was expected to be reported frequently in each condition.

Table 3 lists the emotion terms in descending order of frequency for each condition (including the Happy condition, where the same task was requested of participants in both conditions, but for a happy piece), with the most frequent words shown (down to a count of five). The selection of most frequent terms shown with an asterisk in the top rows of the table (above the horizontal cell divider) was determined by the ‘Power Fitted Elbow’ (PFE) technique that builds on word frequency distribution characteristics [ 70 – 73 ]. The expected target emotion (shown in italics font in the table) is reported most frequently in all conditions. Noteworthy is that sad was reported frequently in the Moved condition, while negative emotions were reported exclusively among the most frequently reported Sad condition emotions. Nostalgia was frequently reported in all conditions. In the Sad condition, the lemma Moved (not shown in the table) was mentioned 4 times, but was not reported frequently, according to the PFE criterion. Another interesting finding is that none of the frequently investigated mediator emotions (Being moved, in particular), appear in the most frequently reported items of the Sad condition list (sad, nostalgia, loss, melancholy and lonely). In contrast, the Moved condition did lead to frequent open-ended reporting of sadness.



Emotion profile of sad music: Felt Affect term ratings

After open-ended responses were reported, participants were asked to indicate the extent to which each of 26 affect terms were felt when listening to the music. Again, the target affect terms were expected to be rated highest. The ratings for each affect term within and between conditions were examined.

what is lexical hypothesis

Means for each affect term by condition are summarised in Fig 1 . Ratings of the same affect term between conditions were analysed using Bonferroni adjusted independent samples t-tests. Felt sadness was rated higher in the Sad condition, but (non-significantly) higher ratings were given to felt Power, Moved and Absorption ratings in the Sad condition. For the Moved condition the affect term Being moved was rated as the second highest scale (second to Absorption), and the rating was statistically the same as for the rating of Being Moved in the Sad condition. Other differences within and across the two conditions can be observed in Fig 1 . Differences for within conditions are not shown because of the large number that were significantly different at p = .05. The highest scoring (with mean rating in at least one condition > 1.5) affect terms were Absorption, Awe, Beauty, Moved, Power, and Sadness.



In these data, a relatively high rating of Being moved can be observed in the Sad condition, and it received a higher rating than the target emotion (Sadness) by M = .122, though non-significantly (p = 1.0), which could be taken to support the action of a mediator, being moved, as responsible for the pleasure generated by the music, despite the accompanying rating of sadness.

Affects that add to enjoyment

The above results indicate the presence of emotion during the enjoyable music experience. However this does not necessarily confirm that the emotion itself is implicated in the enjoyment of the music. The next step of the study addressed this with an explicit question about the contribution of each affect term to the enjoyment of the music. The 26 Affect terms were presented again this time to be classified as contributing, not contributing, or being irrelevant to the enjoyment of the music. Table 4 lists the counts across each of the three possibilities for each Affect term, by Condition. Chi-Square tests identified whether the Affect words add to enjoyment of the music by chance or not.



Significant Chi-square test statistics (at p = .05 with Bonferroni correction) ranged from 15.500 (Fear) to 83.400 (Absorption) for the Moving condition and 14.596 (Fear) to 82.383 (Being moved) for the Sad condition (at p = .05). Chi-Squared tests for Sad and Moving conditions pooled produced statistically significant results for all emotions at p = .05 with Bonferroni correction, ranging from χ 2 = 13.273 (Tragedy) to 158.606 (Being Moved), with second highest χ 2 = 156.85 (for Absorption) and third highest χ 2 = 84.061 (for Sadness).

Self-selected sad music was associated with good likelihood of reporting felt sadness as adding to the pleasure of the experience (83% of response in the Sad condition versus 71% in the Moving condition). The same applies for the affect term rating of Being moved in the Moving condition.

All emotions contributed to the enjoyment of the self-selected music, with the exception of Anger, Fear, Tragedy (both conditions for each, though Tragedy was approaching significance), Grief (Moving condition), Euphoria, Sublime, Happiness, Joy, Peacefulness and Wonder (Sad condition for each). Absorption and Being Moved made the most consistently positive contribution to enjoyment of music, with each being reported as contributing to enjoyment by 90% or more of participants regardless of condition ( Table 4 ).

Fewer nominally negative emotions add to enjoyment in the Moving condition, whereas fewer positive emotions add to enjoyment in the Sad condition. Sadness and crying are emotions with nominally negative connotations, but were reported as adding to the pleasure, regardless of the condition.

Additional emotions that add to liking

The 26 Affect terms might not have exhaustively covered all the emotions that could be experienced, or enjoyed. Therefore, a final question invited participants to list any other emotions that added to the enjoyment of the music.

Only one expression was reported by different participants more than once—Hopelessness (3 independent mentions, one in the Moving condition). 72 participants indicated that no additional emotions contributed to enjoyment (36 in the Moving condition and 36 in the Sad condition). A higher proportion of participants who did report additional emotions mentioned ones that could be interpreted as negative in the Sad condition compared to the Moving condition, but because of the heterogeneity of the responses, which included some words that were already among the 26 Affect terms, no strong conclusion can be drawn, except that the set of Affect terms was effective in identifying the feelings implicated in pleasurable musical experiences.

Hypothesis test–Sadness is liked because the music is sad

For the responses to the Sadness removed step, the following scoring was applied to responses: -2 for ‘I would like the piece a LOT LESS’, -1 for ‘I would like the piece a LITTLE LESS’, 1 for ‘I would like the piece a LITTLE MORE’, 2 for ‘I would like the piece a LOT MORE’, and 0 for NO DIFFERENCE. If the Direct effect hypothesis is supported, we would expect liking to reduce when sadness is removed from the experience. The Indirect effect hypothesis, on the other hand, predicts that removal of sadness would not change liking (change of 0) or increase liking. A single sample t-test supported the Direct effect hypothesis, with an overall reduction of .83 (SD = .916) in liking on the scale of -2 to +2 (t(46) = -.6.207, p < .001, Cohen’s-d = .916). For comparison, in the control condition, removal of movingness also led to a reduction in liking (M = -.77, SD = .807, t(51) = -.6872, p < .001, Cohen’s-d = .807). Taken together the data from this step of the study supports the Direct effect hypothesis.

Based on an overall interpretation of the data, the Direct effect hypothesis is supported. In the specific part of the study that tested the hypothesis, the Sadness removed step, participants reported overall significant reduction in pleasure if the felt sadness, and only the felt sadness evoked by the music, were excised. If sadness were not in itself enjoyed, we may have expected participants to attribute non-sad emotions to the enjoyment, or be unable to perform the task. As it turned out, we can confirm that 83% of participants could perform the task and verify that the sadness was specifically enjoyed, suggesting that the phenomenon of interest is empirically demonstrable. To further ascertain if this is a plausible interpretation, the results are interpreted through the alternative, Indirect effect hypothesis, lens by examining whether mediators still play a commensurate or dominant role in the effect.

Mediation explanation

In the results where affect terms were all rated, a term can be viewed as a mediator if its score or count is statistically equal to or higher than the score or count of the target emotion. Based on this criterion, several steps of the study could be interpreted as supporting the presence of a mediator. In the Open-ended felt emotions step Nostalgia, a prospective mediator of sadness-enjoyment, was spontaneously reported ( Table 3 ). However, Being moved was not, despite previous evidence that Being moved is the stronger candidate of the two [ 15 ]. Nostalgia appeared frequently in the Moved condition as well, but in the Moved condition no mediator was expected because the target emotion (being moved) itself already contained an implicitly positive component. Furthermore, Sadness was also frequently reported in the Moved condition, but, again, there is no reason that being moved would require a mediator. The Indirect effect hypothesis does not predict a mediator that is itself negatively valenced. Thus a mediator based explanation for these results is not straight forward.

In the Affects felt step a more credible impact of prospective mediators can be observed. In the Sad condition, Absorbed (rated highest, with M = 1.796), Being moved (rated higher than Sadness by M = .122, though non-significantly [NS], p = 1.0) and Powerful feelings (rated higher than Sadness by M = 0.020, NS p = 1.0) are all rated as high or higher than the target emotion (Sadness). In the Moved condition only Absorbed (M = 1.942) is rated higher than Being moved (by M = .135, NS p = .074). If we set aside the finding for the Moved condition, the mediator-based explanation is supported, triangulating extant evidence that two of these affects (absorbed and moved) are mediators of sadness.

So it is possible to find support for the Indirect-effect hypothesis, and the mediator-based explanation in particular. However, the findings refer to the presence of emotions. There is no assurance that any of the emotions identified are adding to the pleasure, with the exception of the target emotion, since that requirement was made explicit in the procedure.

The Affects that add to liking step addressed the matter. Being moved, Absorption, and Powerful feelings (but not Nostalgia) all had the same or higher counts than the target (Sadness) emotion, indicating that they add to enjoyment in the Sad condition ( Table 4 ). For example, the affect term Being moved was voted as ’adding to pleasure’ by 96% of participants in the Sad condition, compared to the affect term Sadness ’adding to pleasure’ according to 83% of participants. This supports the Indirect effect hypothesis ( Table 4 ).

Here we have the strongest evidence of mediators in explaining enjoyment of sadness, and this aligns with evidence from previous research [as discussed in the introduction, see 17 ]. But Absorption (adds to pleasure according to 92% of participants) also has a higher count than the target emotion (90%) in the Moved condition. Does that mean that Absorption also mediates Being moved? As pointed out above, that seems unlikely because Being moved already contains a positive aspect, and so should not need a mediator. Using the mediator-based explanation, Absorption adding to enjoyment votes should have (at least) been fewer than the votes for Being moved in the Moving condition (which was not the case). Furthermore, in the Sadness condition, the target emotion itself received statistically significant votes as adding to pleasure, meaning that the alleged mediators may not have served any essential purpose in contributing to the enjoyment. The mediation explanation is only able to partially explain the results. An alternative explanation is proposed by applying the concept of ‘semantic overlap’.

Semantic overlap explanation

Semantic overlap is a phenomenon concerned with the mental organisation of concepts and word meanings. Words with similar meanings (synonyms) are more linked with one another in a mental space than words with unrelated meanings. This is often characterised in network inspired models of the mind, foundationally proposed by Quillian and the notion of the semantic network [ 74 , 75 ]. Word meanings are organised in a complex yet systematic manner according to network principles, of particular interest here being through similarities in the meaning of words, where expressions that are more similar in meaning appear ‘closer together’ in the mental network. This means that when a word is triggered (e.g., heard or read), the semantically more closely related words are more primed (ready to be raised to conscious attention) in the mental network than less closely related words. Cognitive linguists by and large agree that words are pointers or approximate representations of concepts and experiences stored in memory [ 76 , 77 ]. The implication is that words can be mapped onto points in multidimensional semantic space, with distance between words reflecting (of interest here) degree of conceptual dissimilarity between the words. Considerable effort has been devoted to organising emotions by similarity [e.g., 78 – 83 ]. Semantic distance may therefore explain why Being moved frequently appears for sad evoking music (a frequently reported result), and the novel findings identified in the present study.

It is possible to estimate the relative semantic distance between the two words moving and sadness by looking up the terms in a published list of words with quantified point estimates of locations in theoretical semantic space. A large such database was developed by Mohammad [ 82 ], where estimates of location in semantic space of some 20,000 English words were produced. The semantic space in that research adopts a conventional representation of the space, particularly relevant for emotions, referred to as ‘VAD’ space. Emotions can be reasonably well expressed in terms of two dimensions, labelled valence (V) and arousal (A), where the former refers to the positive or negative aspect of the word’s meaning (e.g., happy and calm exhibit positive valence, while sad and angry negative) and the degree of activity associated with the word’s meaning (e.g., joyous and furious are high arousal, while calm and sad are low arousal). Some have argued that two dimensions are only partially sufficient for describing the meaning of an emotion [ 81 , 84 – 87 ], and a frequently proposed third dimension is dominance (D) (where words such as angry and energetic exhibit high dominance, while fear and innocuous are low in dominance), leading to the VAD (Valence, Arousal, Dominance) abbreviation for this three dimensional configuration [other examples: 85 , 88 , for a review, see 89 , 90 ]. Mohammad (82) provided numerical VAD scores for each term scaled to a score between 0 and 1 (negative to positive for valence, low to high for arousal and for dominance) based on human ratings. From these data it is possible to estimate the semantic distance between emotions.

Through calculations using the VAD word list published by Mohammad (82), Moved and Sadness have a semantic distance in VAD space of 0.607 units (numbers closer to 0 indicating greater similarity). With Sadness as the reference, positive emotions appearing in the Affect term list have distances that range from 0.852 for Calm to 1.243 for Joy (all greater than the distance between Sadness and Moving), while negative emotions have scores ranging from 0.469 for Grief (closest negative emotion to Sadness from the Affect terms presented) to 0.768 (Anger), which apart from Anger are all closer to Sadness than Moving is to Sadness. That is, Moving has more semantic overlap with Sadness than does Anger and the positive emotions Joy and Happiness, suggesting semantic overlap as a viable alternative to mediation as to why being moved appears in tandem with sadness. The VAD data also suggest that Moving is semantically more closely related to Sadness than Catharsis, since Catharsis has a distance of 0.633 from Sadness (slightly more distant than Moving). High ratings of Moving for a Sad-evoking context can therefore be explained by semantic overlap. Such an interpretation strengthens the case for supporting the Direct effect hypothesis, because being moved need not be treated as surrogate for sadness.

The Direct effect hypothesis proposes that pleasure is experienced by contextualised re-appraisal or ‘dissociation’ of the Action tendency component of an otherwise negative emotion. The consequent positive experience (enjoyment, pleasure, preference) provides another clue for the remaining Affect terms that were rated the same or higher than the target emotion in each condition. The mediation account fails to explain why Sadness was voted (by 71% of participant) as adding to enjoyment in the Moving condition. The mediator based explanation is also poor at explaining why Absorption was reported as adding to enjoyment, and for doing so in both conditions.

The semantic overlap approach can better explain these results, too. Affect terms such as Absorption and Powerful feelings are affects related to enjoyment when experiencing art. Consider the Absorption in Music scale developed by Sandstrom and Russo [ 91 ]. The 34 item scale contains several items related to the pleasure of being engaged with music in different ways [see also 2 , 18 , 92 , 93 ]. Powerful experiences are reported during special, personal experiences that occur during strong positive aesthetic experiences [ 94 – 96 , p. xiv]. That is, the task itself, of identifying a loved piece of music, also produces semantic overlap of these terms. Furthermore, in the Sad condition several positive emotions were reported more frequently as having no relevance to enjoyment, in comparison to the Moved condition: Euphoria (57% in the Sad condition versus 15% in the Moving condition), Happiness (43% vs 8%) and Joy (49% vs 12%). Mediation struggles to explain why purely positive affect terms are not voted as adding to enjoyment. Semantic overlap, on the other hand, suggests that the activation of sadness is more likely to be associated with other negative emotions, while being moved would be more associated with emotions of both positive and negative valence. In addition to the possibly misleading interpretations of enjoyed-sadness in music research employing a mediator-based approach to explaining the phenomenon, discussed in the Method section, semantic overlap offers an explanation of the results that is superior to the mediator-based explanation.


This study investigated whether the experience of sadness, evoked by music, can itself be highly enjoyable. A novel method was applied where participants were asked to imagine how enjoyment would be impacted should the felt sadness somehow be removed. The results demonstrated that sadness is directly implicated in the enjoyment of such music, providing support for the ‘Direct effect hypothesis’. This hypothesis states that when sad music is enjoyed, the sadness itself directly contributes to the enjoyment. A theoretical position has been presumed by the hypothesis–that the experience of sadness contains a component that can be dissociated from regular experience of the negative emotion when contemplating music or any aesthetic event. The presence of emotions such as being moved were explained by the concept of semantic overlap, where an emotion concept is not activated as a lexical singular, but rather as the meaning that the emotion encompasses, or that is spread to other related emotions, according to how similar they are (in this case to the concept of sadness). Being moved is sufficiently close in meaning to sadness to allow it to be activated during a sadness evoking music experience, regardless of the extent to which it is enjoyed, meaning that the presence of an emotion such as being moved does not necessarily explain (and is not needed to explain) why felt sadness can be enjoyed. Absorption is another affect that accompanied loved, sadness-inducing music. This, too, was explained by semantic overlap, with the positive component of the sadness activating other, reasonably nearby, positive affects, including Absorption. The state of absorption may also play a causal role in attraction to music [ 20 , 97 ], and so there could well be some feedback loop between absorption and other aspects of the experience, including evoked emotions. Suggestions were made for further research to test whether the semantic overlap account and the Direct effect hypothesis better characterise enjoyment of negative emotion in music than mediators (such as being moved and absorption) that themselves have a positive component, through which enjoyment is indirectly generated.

The results of the present study were enhanced by applying a modified version of research using self-selected stimuli that minimised demand characteristics, while ensuring that the phenomenon of interest was investigated. Methodologically, the study took the critical step of ensuring that the impact of particular affects on enjoyment of the music were investigated, not just their presence. Future research is likely to continue the more popular method of using experimenter-selected stimuli which are then rated along various affect terms. This paper made recommendations on how such research could be more successful at identifying the phenomenon of interest, and in so doing better address the debate on the enjoyment of felt sadness and other felt negative emotions in music.

  • View Article
  • Google Scholar
  • PubMed/NCBI
  • 23. Walton KL. Mimesis as make-believe: On the foundations of the representational arts. Cambridge, MA: Harvard University Press; 1990.
  • 29. Frijda NH. The emotions. Cambridge: Cambridge University Press; 1986.
  • 54. Van den Tol AJM, Ritchie TD. Emotion memory and music: A critical review and recommendations for future research. In: Rosaria SM, Alessandra R, editors. Music, memory and autobiography: An interdisciplinary approach. Naples, Italy: Liguori Publisher; 2015.
  • 60. Hayes AF. Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press; 2013.
  • 67. Quigley KS, Lindquist KA, Barrett LF. Inducing and measuring emotion and affect: Tips, tricks, and secrets. In: Reis HT, Judd CM, editors. Handbook of research methods in social and personality psychology: Cambridge University Press; 2014. p. 220–52.
  • 70. Hills T, Hurford AC, Stroup WM, Lesh R. Formalizing learning as a complex system: Scale invariant power law distributions in group and individual decision making. Foundations for the future in mathematics education. London: Routledge; 2007. p. 225–44.
  • 74. Quillian MR. Semantic memory: PhD Thesis, Carnegie Institute of Technology; 1966.
  • 82. Mohammad S, editor Obtaining reliable human ratings of valence, arousal, and dominance for 20,000 English words. Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long papers); 2018; Melbourne, Australia: Association for Computational Linguistics.
  • 87. Trkulja M, Janković D, editors. Towards Three-Dimensional Model of Affective Experience of Music. 12th International Conference on Music Perception and Cognition (ICMPC) and 8th Triennial Conference of the European Society for the Cognitive Sciences of Music (ESCOM); 2012; Theealoniki, Greece.
  • 94. Gabrielsson A. Strong experiences with music: Music is much more than just music. Oxford: Oxford University Press; 2011.
  • 95. Gabrielsson A, Whaley J, Sloboda J. Peak Experiences in Music. In: Hallam S, Cross I, Thaut MH, editors. The Oxford Handbook of Music Psychology. 2nd ed. Oxford: Oxford University Press; 2016. p. 745–58.
  • 96. Wordsworth W, Coleridge ST. Lyrical ballads, with other poems. In two volumes. London: printed for T. N. Longman and O. Rees, by Biggs and Co. Bristol; 1800.


  1. How to Write a Hypothesis

    what is lexical hypothesis

  2. How to Write a Hypothesis: The Ultimate Guide with Examples

    what is lexical hypothesis

  3. Research Hypothesis: Definition, Types, Examples and Quick Tips

    what is lexical hypothesis

  4. What is the Difference Between Lexical Analysis and Syntax Analysis

    what is lexical hypothesis

  5. How to Write a Strong Hypothesis in 6 Simple Steps

    what is lexical hypothesis

  6. Figure1: learning theory in Lexical Approach (Lewis, 2000)

    what is lexical hypothesis


  1. The Good Genes Hypothesis

  2. lexical Analysis|ATC

  3. Synonyms in Lexical Relations|by Muhammad Alamgir


  5. Lexical Analysis In Compiler|AUTOMATA THEORY AND COMPILER DESIGN|

  6. Linguistics relativity (Sapir-Whorf hypothesis)


  1. Lexical hypothesis

    The lexical hypothesis is a major basis of the study of the Big Five personality traits, the HEXACO model of personality structure and the 16PF Questionnaire and has been used to study the structure of personality traits in a number of cultural and linguistic settings. History ...

  2. Lexical Hypothesis

    The Lexical Hypothesis is a significant concept in the field of personality psychology. Broadly speaking, it proposes that the most relevant and universally acknowledged human personality traits are encoded in our language. These traits are believed to be so crucial to communication and social interaction that our ancestors developed specific ...

  3. APA Dictionary of Psychology

    lexical hypothesis. Share button. Updated on 04/19/2018. the supposition that any significant individual difference, such as a central personality trait, will be encoded into the natural-language lexicon; that is, there will be a term to describe it in any or all of the languages of the world.

  4. Personality Psychology: Lexical Approaches, Assessment Methods, and

    Lexical Encodings—Constructs and Representations. The lexical hypothesis states that people encode in their everyday languages all those differences between individuals that they perceive to be salient and that they consider to be socially relevant in their everyday lives. Encodings about differences within individuals over time are not explicitly mentioned, however (see below).

  5. 3.2: Personality Traits

    Lexical hypothesis The lexical hypothesis is the idea that the most important differences between people will be encoded in the language that we use to describe people. Therefore, if we want to know which personality traits are most important, we can look to the language that people use to describe themselves and others. ...

  6. Personality psychology: Lexical approaches, assessment methods, and

    This article develops a comprehensive philosophy-of-science for personality psychology that goes far beyond the scope of the lexical approaches, assessment methods, and trait concepts that currently prevail. One of the field's most important guiding scientific assumptions, the lexical hypothesis, is analysed from meta-theoretical viewpoints to reveal that it explicitly describes two sets of ...

  7. Lexical Hypothesis

    The Lexical Hypothesis emphasizes the importance of language in understanding and describing human personality. It posits that the traits individuals consider significant and observable are reflected in the words and phrases they use. By leveraging natural language, researchers have gained valuable insights into trait structure, allowing for a ...

  8. 10 The Lexical Foundation of the Big Five Factor Model

    The Lexical Hypothesis Before the two versions of the five factorial trait model started playing a significant role in the psychological literature, there is a rather clear timeline from Galton (1884) to Cattell ( 1943a , 1943b ) in which ideas for a glossary of personality were given form and in which the large list of personality trait terms ...


    lexical hypothesis By N., Sam M.S. the theory that important natural characteristics and traits unique to individuals have become intrinsically embedded in our natural- language lexicon over time.

  10. Personality Structure

    The lexical hypothesis has been particularly generative for defining the domain of variables that are relevant to mapping personality structure (see Saucier's chapter on the "Lexical Approach" in this volume). One early model of personality structure developed using an approximation to the lexical approach is Cattell's 16-factor model, which was developed by factor-analyzing 35 ...

  11. lexical hypothesis definition

    The lexical hypothesis is a concept in personality psychology and psychometrics that proposes the personality traits and differences that are the most important and relevant to people eventually become a part of their language. It goes further to suggest that the most important concepts in personality become single descriptive words in a language.

  12. Lexical hypothesis explained

    The lexical hypothesis [1] (also known as the fundamental lexical hypothesis, [2] lexical approach, [3] or sedimentation hypothesis) is a thesis, current primarily in early personality psychology, and subsumed subsequently by many later efforts concerning that topic. [4] Despite some variation of its definition and application, the hypothesis ...

  13. Individual Differences Among Skilled Readers: The Role of Lexical

    What critically distinguishes the lexical quality hypothesis (LQH) from many other accounts of individual differences in reading is that it assigns a causal role to lexical knowledge. Rather than attributing reading difficulties to dysfunctions in phonological, semantic, ...

  14. Measuring Lexical Quality: The Role of Spelling Ability

    The construct of 'lexical quality' (Perfetti Scientific Studies of Reading 11, 357-383, 2007) is widely invoked in literature on word recognition and reading to refer to a systematic dimension of individual differences that predicts performance in a range of word identification and reading tasks in both developing readers and skilled adult populations. Many different approaches have been ...

  15. Personality Trait Theory

    Lexical Hypothesis in Personality Psychology . Lexical Hypothesis is a theory that says if there's a behavior so prominent throughout time, we create a word for that word. If we don't have a word that describes a trait, then it must not be very prevalent or useful to personality psychology.

  16. Understanding the Lexical Hypothesis

    The lexical hypothesis posits that important aspects of human personality and behavior are encoded in language. It suggests that traits and characteristics w...

  17. Testing the lexical hypothesis: Are socially important traits more

    The findings suggest that the lexical hypothesis may not be strongly supported in some ways it is commonly understood but is supported in the manner most important to investigations of trait structure. Specifically, trait terms with greater relational impact tend to more strongly correlate with other terms in lexical sets and thus have a ...

  18. Personality 101: The Trait Approach & the Lexical Hypothesis

    This is the so-called lexical hypothesis, proposed by Galton (1949), which states that every important human phenomenon must be somehow represented in the lexicon of a language, and since most languages are easily translated to others, the universality of the phenomenon could also be guaranteed.

  19. Nurturing a lexical legacy: reading experience is critical for the

    The lexical legacy hypothesis, introduced in this paper, provides a new perspective on the transition from novice to expert. The focus of this paper is with how people read words.

  20. Deep Lexical Hypothesis: Identifying personality structure in natural

    Deep Lexical Hypothesis: Identifying personality structure in natural language. Recent advances in natural language processing (NLP) have produced general models that can perform complex tasks such as summarizing long passages and translating across languages. Here, we introduce a method to extract adjective similarities from language models as ...

  21. Lexical Quality Matters: Effects of Word Knowledge Instruction on the

    sion. Because the lexical quality hypothesis emphasizes the interface between these two components, we decided to turn some of its theoretical assumptions into principles for teaching. Specifically, we chose to empha-size word knowledge as the focal point of our interven-tion. Moreover, we wanted to enhance lexical knowledge

  22. Lexicalist hypothesis

    Lexicalist hypothesis. The lexicalist hypothesis is a hypothesis proposed by Noam Chomsky in which he claims that syntactic transformations only can operate on syntactic constituents. [ambiguous][jargon] [1] It says that the system of grammar that assembles words is separate and different from the system of grammar that assembles phrases out of ...

  23. What Is Lexical Analysis?

    Lexical analysis, or scanning, is a fundamental step in NLP. In programming languages, this process involves the lexical analyzer (lexer or scanner) reading the source code character by character to group these characters into tokens, the smallest units in the code that convey meaning.

  24. What is Lexical Activation Hypothesis?

    Learn more about Lexical Activation Hypothesis. Take a deep dive into Lexical Activation Hypothesis with our course AI for Designers . In an era where technology is rapidly reshaping the way we interact with the world, understanding the intricacies of AI is not just a skill, but a necessity for designers. The AI for Designers course delves into ...

  25. Liking music with and without sadness: Testing the direct effect

    Limitations of the Indirect effect hypothesis. An inherent weakness of Indirect effect hypothesis, and in particular the mediator-based explanation, is that it does not consider the phenomenal experience of the individual who claims that they both experience sadness, and that the sadness itself, for them, forms at least part of the pleasure [e.g., 6].