Abstract

Since the beginning of the 20th century, intelligence has been conceptualized
as a qualitatively unique faculty (or faculties) with a relatively fixed
quantity that individuals possess and that can be tested by conventional
intelligence tests. Despite the logical errors of reification and circular
reasoning involved in this essentialistic conceptualization, this view
of intelligence has persisted until the present, with psychologists still
debating how many and what types of intelligence there are. This paper
argues that a concept of intelligence as anything more than a label for
various behaviors in their contexts is a myth and that a truly scientific
understanding of the behaviors said to reflect intelligence can come only
from a functional analysis of those behaviors in the contexts in which
they are observed. A functional approach can lead to more productive methods
for measuring and teaching intelligent behavior.
Copyright Psychological Record Winter 2003

Follow up:

Since the beginning of the 20th century, intelligence has been conceptualized
as a qualitatively unique faculty (or faculties) with a relatively fixed
quantity that individuals possess and that can be tested by conventional
intelligence tests. Despite the logical errors of reification and circular
reasoning involved in this essentialistic conceptualization, this view
of intelligence has persisted until the present, with psychologists still
debating how many and what types of intelligence there are. This paper
argues that a concept of intelligence as anything more than a label for
various behaviors in their contexts is a myth and that a truly scientific
understanding of the behaviors said to reflect intelligence can come only
from a functional analysis of those behaviors in the contexts in which
they are observed. A functional approach can lead to more productive methods
for measuring and teaching intelligent behavior.

Few topics have sparked such heated debate within the academic community
and society at large as that of intelligence and intelligence testing.
Some of the contentious issues in the debate include the very definition
of intelligence, the controversy concerning IQ and race, the ever present
nature-nurture problem (Weinberg, 1989), and even the question of whether
intelligence exists (Howe, 1990). The debate was reignited most recently
by the publication in 1994 of The Bell Curve: Intelligence and Class Structure
in American Life by Richard J. Herrnstein and Charles Murray. The sturm
and drang created by the publication of this book was, among other things,
the motivation behind the creation of a task force in 1995 by the Board
of Scientific Affairs of the American Psychological Association to prepare
an authoritative report on the current status of research on intelligence
and intelligence testing. The flurry of response generated by The Bell
Curve, both in the academic community and in the media, also prompted a
letter to the Wall Street Journal in December, 1994, in which 50 professors
"all experts in intelligence and allied fields" signed a statement, titled
"Mainstream Science on Intelligence."1 The purpose of this statement was
to respond to the public outcry over the suggestions and social implications
of The Bell Curve by outlining "conclusions regarded as mainstream among
researchers on intelligence, in particular, on the nature, origins, and
practical consequences of individual and group differences in intelligence"
("Mainstream Science on Intelligence," 1994).

One of the reasons for the persistent concern about intelligence is that
intelligence tests have been used to support nativistic theories in which
intelligence is viewed as a qualitatively unique faculty with a relatively
fixed quantity. Historically, proponents of nativistic theories have succeeded
in persuading those with political power that standardized tests reliably
measure intelligence; and these tests have been used to make important
decisions about vast numbers of individuals including immigrants, U.S.
soldiers during the first World War, normal school children, and the developmentally
disabled. Not surprisingly, there exists a substantial literature documenting
the history of the intelligence testing movement (e.g., Bolles, 1993; Fancher,
1985; Gould, 1981; Herrnstein & Murray, 1994; Kamin, 1974).

Background: General Intelligence

Although many people have contributed to the conception of intelligence
as a unitary, qualitatively unique trait, we may trace its origin in the
history of the intelligence testing movement to the British psychologist
and statistician Charles Spearman. When Spearman factor analyzed scores
on different tests of intelligence, including his own Galtonian-inspired
sensory acuity tests, as well as those designed by Binet and Simon, he
discovered that not only were the correlations positive, but they fell
into a roughly hierarchical pattern with the highest correlations on tests
that Spearman believed required higher level skills, such as abstract thinking.
For Spearman, these positive intercorrelations provided evidence of a common
underlying factor that tied them all together; he called this factor g
for general intelligence (Spearman, 1904, 1927). Crinella and Yu (2000)
explain:

People who are proficient at solving a given problem tend to be proficient
at solving others; those less capable of solving that problem tend to be
less capable of solving others. The psychometric representation of this
phenomenon is the general intelligence or g factor, obtained whenever scores
on a battery of diverse problem solving tests are factor analyzed. (p.
299)
Thus g factor, or simply g as it is called, is a numerical outcome-an algebraic
factor-resulting from a complex series of statistical manipulations, called
factor analysis. Although Spearman believed that different types of skills
required their own type of intelligence, which he called s for specific
factors, he still viewed g as the most important factor. Spearman further
believed that a person's general intelligence was inherited and, thus,
was somehow represented in the physiology of the brain (Fancher, 1985).

As Fancher (1985) notes, Spearman saw what he wanted to see in his data.
Despite problems with his original calculations, however, Spearman's conclusions
regarding the positive intercorrelations between individual tests and their
hierarchical arrangement have been generally accepted. In addition to problems
with Spearman's research design and statistical analyses, there are several
logical problems with his theory of general intelligence that would not
concern us as much today had the core of his theory not been adopted by
many who followed him. For one, as critics such as Gould (1981), among
others, have pointed out, Spearman committed the logical error of reification.
He took an abstract mathematical correlation and reified it as the general
intelligence that someone possesses. Although Spearman was probably not
the first person to commit this error regarding intelligence, his use of
mathematics and statistics lent the appearance of scientific credibility
to this practice. Once the error of reification is committed, it is easy
to commit another logical error, circular reasoning, in which the only
evidence for an explanation of some phenomenon is simply the phenomenon
itself. In Spearman's case, the only evidence for g, or general intelligence,
were the positive correlations, even though it was those positive correlations
he was trying to explain in the first place. Had no positive correlations
been obtained, Spearman would not have posited g. As some authors have
correctly noted, g is a "statistical extraction" (Gottfredson, 1998) or
"a construct . . . representing individual differences in performance on
multiple cognitive measures" (Crinella & Yu, 2000, p. 300); it is not a
thing being measured. What is measured is the behavior of large numbers
of people on various tests. The positive intercorrelations that result
from factor analysis of their test scores are themselves far removed from
the behavior of any individual in the test-taking situation or, for that
matter, in any other context.

If we do not accept prima facie the claims by Spearman and his intellectual
descendants that the positive intercorrelations obtained on intelligence
tests are a reflection of innate intelligence possessed by the individual
test takers, then the positive correlations need to be explained some other
way as do the performances of subjects on individual tests. As scientists
know, correlation does not mean causation, although we apparently need
to be reminded of this dictum fairly regularly (e.g., Gould, 1981; Howe,
1988a; Kamin, 1995; Layzer, 1972; Schlinger, 1996). In fact, variables
may correlate for a number of possible reasons, only one of which is that
they may be causally related. However, correlations based on different
intelligence tests in large populations cannot be explained in the traditional
sense of the term; they can only be accounted for statistically. The performance
of individuals on a test or while engaging in behavior we call "intelligent"
can be explained scientifically but this has not been the interest or strategy
of intelligence researchers.

Because g is a statistical construct, theorists can see in it what they
want and, consequently, the interpretation of g has changed over time.
It initially represented a factor resulting from factor analysis but it
has evolved semantically into general intelligence as a mental capacity
and, more recently, into general cognitive ability (Plomin, 1999). The
value, if any, of g is the same as that of any factor derived from factor
analysis, namely it is a useful tool for detecting covariation in a set
of complex data. Unfortunately because of the potential problems with external
validity, a factor could emerge that is just an artifact of the way intelligence
tests are designed; it might not represent anything fundamental about the
test takers.

The concept of the inheritance of intelligence has also undergone some
revision. Galton and Spearman believed that intelligence was inherited.
Modem behavior geneticists claim that the heritability of 10 scores lies
somewhere between .30 and .70 (Plomin, 1990), with some claiming an outright
.80 heritability quotient (Jensen, 1969). What must be remembered is that
behavior genetic research can only estimate to what extent the variance
of one measurement, such as intelligence test scores, is correlated with
the variance of another measurement, such as genetic differences, in a
population of individuals. The heritability of a trait, for example, intelligence
(as reflected by IQ), refers to the proportion of that variation that is
associated with genetic differences among the individuals. Although behavior
genetic researchers make strong claims for the validity of the (separated
twin and adoption) research that they have used to estimate the heritability
of intelligence, some of the core assumptions underlying such research
have been seriously called into question. For example, according to some
authors (e.g., Beckwith, 1999; Joseph, 1998), the equal environment assumption-namely
that the environments of identical and fraternal twins are equal or even
similar-is demonstrably false, whereas other critics (e.g., Layzer, 1972,
1999) seriously question the assumption that the variation of genetic and
nongenetic factors contribute additively and independently to intelligence
and, even if they do, that the heritability of IQ can be estimated from
the extant data.

All of these concerns raise the issue as to just exactly what is being
measured by intelligence tests. These and other problems occasioned by
Spearman's theory would not demand our attention at the present time had
the theory died with him. But Spearman's concept of g still plays a central
role in intelligence research and theory, especially among those psychologists
and social scientists who champion standardized tests and factor analysis
as methods to discover what intelligence is and how to measure it, and
who either conduct or support behavior genetic research (e.g., Carroll,
1993; Gottfredson, 1998; Jensen, 1998). Dissatisfaction with this strong
(or restricted) view of g by other intelligence theorists, however, has
led to different, but related concepts of intelligence which are alive
and well in contemporary psychology.

Recent Conceptions of Intelligence

Until fairly recently, Spearman's concept of g had been the accepted view
of human intelligence among intelligence theorists. In recent years, however,
other theories have been proffered in which intelligence is seen not as
a single general trait, but rather as a number of different traits or capacities,
a trend that has not gone uncriticized by hard-line g-factor theorists
(e.g., Gottfredson, 1998). For example, Sternberg's (1984) triarchic theory
includes three types of intelligence-analytical, creative, and practical-all
of which combine to make up what we call "intelligence." Gardner (1983)
has postulated no fewer than six intelligences, including linguistic and
musical intelligence, both of which are aural-auditory; logicalmathematical
and spatial intelligence, which are visual; and bodily kinesthetic and
personal intelligence. Although Gardner explicitly warns against the trap
of reification when discussing intelligence, he offers the following general
rule: "Intelligences should be thought of as entities at a certain level
of generality, broader than highly specific computational mechanisms (like
line detection) while narrower than the most general capacities like analysis,
synthesis, or a sense of self (if any of these can be shown to exist apart
from combinations of specific intelligences)" (p. 68, emphasis added).
This statement notwithstanding, Gardner states that his intelligences are
not "physically verifiable entities" but, rather, "potentially useful scientific
constructs" (p. 70).

One of the notable differences in both Sternberg's and Gardner's subdivision
of intelligence into multiple forms is their inclusion of intelligences
not specifically related to academic success. In fact, a relatively recent
trend in the intelligence literature has been to expand the concept to
include capabilities or skills explicitly not related to success in school
and, thus, not assessable by conventional tests. Toward that end, some
psychologists (e.g., Neisser, 1976; Sternberg, Wagner, Williams, & Horvath,
1995) have distinguished between academic and practical intelligence. Academic
intelligence enables one to perform well on school-related tasks and can
be assessed by conventional standardized intelligence tests. Practical
intelligence, known by the ordinary term, "common sense," enables one to
perform well in the real world but cannot be fairly assessed through standardized
intelligence tests.

Consider the following example of practical intelligence offered by Sternberg
et al. (1995). Apparently, garbage collectors in Tallahassee, Florida used
to have to manually lift each huge city-issued trash container, empty it
into the truck, and then carry the empty container back into the resident's
yard. One day a new man on the job came up with the idea, the authors called
it an insight, of wheeling each empty container into the yard of the next
resident and so on, rather than returning each container to the yard from
which it came. This meant that the garbage collectors only had to make
one trip to each yard instead of two. Instituting this practice cut the
work almost in half. In an attempt to understand such intelligent behavior,
Sternberg et al. (1995) asked the following question: "What kind of intelligence
enables a person to come up with this kind of strategy for reducing effort
by half, a strategy that had eluded well-educated observers such as the
present authors, other garbage collectors, and the managers who trained
them" (p. 912)? By assuming that a "kind of intelligence" must be responsible
for such an insight, however, Sternberg et al. commit the error of reification
and, by extension, circular reasoning. They simply observed a particular
behavior in its context, and because the actual proximate and ultimate
causes of the behavior were not easily identifiable, they inferred a "kind
of intelligence" as the cause of the behavior. In most, if not all, instances
the only evidence for the inferred intelligence is the very behavior said
to reflect it. If others do not exhibit the behavior of interest, then
theorists are likely to assume, as Sternberg et al. did, that they do not
possess that particular "kind of intelligence." Once again, the actual
behaviors said to reflect intelligence, in this example, the idea (i.e.,
the verbal statement) of wheeling trash containers into the yard of the
next resident, remain unexplained and scientists may be distracted from
studying the variables, for example, the particular experiences, which
determine such behavior.

More recently, psychologists have moved even farther away from the conventional
and narrow conception of intelligence as a "genetic given that cannot be
changed by life experience" (Goleman, 1995, p. xi) and toward a conception
that includes "emotional intelligence" (see Mayer & Geher, 1996; Mayer
& Salovey, 1993). For Mayer and his colleagues, emotional intelligence
is "defined as the capacity to reason with emotion in four areas: to perceive
emotion, to integrate it in thought, to understand it and to manage it"
(Mayer, 1999). Of course, such a definition is predicated on a definition
of emotions, which, according to Mayer, Caruso, and Salovey (2000), "are
internal events that coordinate many psychological subsystems including
physiological responses, cognitions, and conscious awareness (p. 267).
Notwithstanding the authors' assertion that emotional intelligence meets
the criteria required for scientific legitimacy, one of which is its ability
to be operationalized as a set of abilities, their definition of emotion
is inherently subjective - how can one observe and measure the internal
events?

Psychologists, such as Goleman (1995), who have popularized the concept
of emotional intelligence2 believe a challenging question needs to be answered:
"What can we change that will help our children fare better in life" (pp.
xi-xii)? For Goleman the answer lies in the abilities he calls "emotional
intelligence," which include, among other capabilities, "self-control,
zeal and persistence, and the ability to motivate oneself' (p. xii). For
Goleman, addressing emotional intelligence is not only a practical issue
but a moral one as well. Moreover, he suggests that parents and the culture
at large should better prepare children for life by placing more emphasis
on their emotional education relative to their intellectual education.
The behaviors involved when we speak of "self-control, zeal and persistence,
and the ability to motivate oneself' are no doubt important in some social
contexts, but by adding the term "intelligence" Goleman opens up a Pandora's
box of difficulties, including the abovementioned problems of reification
and circular reasoning. For example, it might be said that some people
have more emotional intelligence than others and that this causes them
to behave with more zeal and persistence, or to be more self-motivated.
Meanwhile, and most importantly, the behaviors said to reflect emotional
intelligence, remain unexplained. Although labeling behavior as "emotional
intelligence" may have some value, it does nothing to explain the behavior.
Except for the common sense appeal of terms such as "emotional intelligence,"
they bring with them numerous logical and scientific pitfalls.

Even this very cursory review of recent trends regarding human intelligence
compels us to ask just exactly what intelligence is and how many types
there are. Such questions have the appearance of being scientific, although
they are really semantic (or philosophical) questions that may distract
behavior scientists from the more important questions about the variables
responsible for the actual behavior we call "intelligent." There are obviously
many different ways of being smart. In fact, one implication of the position
offered in this article goes even further to suggest that the intelligence
of an individual be determined on a case-by-case or, more specifically,
a behavior-by-behavior basis. Thus, the question of whether intelligence
is one general, qualitatively distinct trait, or several specific categorical
types seems moot. Rather than debating how intelligence should be defined
or how many forms of intelligence there are, perhaps behavior scientists
should instead take a more Darwinian view and look at the specific behaviors
in their contexts that we label "intelligent," and then analyze them according
to their function in those contexts.

Essentialistic Versus Functional Approaches To Intelligence

The logical problems of reification and circular reasoning described above
are often found in essentialistic approaches to behavior. The concept of
"essentialism" [the term was coined by the philosopher Karl Popper (1957)]
has a long history in Western thinking. As Mayr (1982) points out, we may
trace the concept back to Plato who likened phenomena in the natural world
to geometric forms. For example, just as a triangle always has the same
unchanging form and is thus discontinuous, that is, different from any
other form, so are phenomena in the natural world reflections of a limited
number of fixed and changeless forms, or essences as they came to be called.
For example, any one cat, although different to varying degrees from other
cats, is an example of a fixed unchanging form that defines all cats. According
to an essentialistic view, genuine change can occur only through the sudden
and abrupt origin of new essences or forms, a view at odds with a selectionist
view of change. According to Mayr (1982), the philosophy of essentialism
dominated biological thinking for almost two millennia until selectionism,
first formally introduced by Charles Darwin, replaced it. More recently,
Palmer and Donahoe (1992) extended Mayr's distinction between essentialism
and selectionism to psychology as well. They argued that, with perhaps
the single exception of the selectionist program articulated by B. F. Skinner,
theorizing about human behavior, in particular by many cognitive psychologists,
has been dominated by essentialistic thinking.

A clue to essentialistic thinking can often be found in the practice of
formally defining terms before any functional analysis occurs. Palmer and
Donahoe (1992) state that a priori formal definitions of terms can be said
to have essential properties to the extent that they "precede any example
of the category." For example, they point out that Noam Chomsky's concepts
of abstract structures and language faculty are essentialistic in that
they are wholly abstract constructs which have been posited without reference
to physical mechanisms explained by scientific principles of selection.
Interestingly, Chomsky viewed such constructs in the same way that Spearman
viewed general intelligence-as inborn, fixed properties of human beings.

Many of the ways the concept of intelligence has been historically discussed
reflect essentialistic thinking. For example, simply asking the question,
What is intelligence? implicitly assumes that there is a quality, or essence,
of human nature with essential, immutable qualities. Asking this question
begins the process of reifying intelligence because it suggests that intelligence
is an entity possessed by individuals that determines their behavior. Once
intelligence as an essence or quality is assumed, the next logical step
is to provide a formal definition. The futility of this tactic was demonstrated
by Sternberg and Detterman (1986) who asked two dozen prominent theorists
to define intelligence and got two dozen different definitions. After a
formal definition is offered, researchers then typically construct hypotheses
regarding how the essence should manifest itself in behavior, or how it
can be tested or studied. All of this information may then be used to construct
a research program into the nature of the essence and the processes through
which it works to generate behavior. Researchers observe the actual behavior
usually only as a way of understanding the underlying essence which, itself,
can never be observed directly-a practice that has not gone without criticism
(e.g., Schlinger, 1993).

A functional, selectionist approach follows a different path. Rather than
constructing formal definitions a priori and then looking for instances
of them, scientists must discover the definition (see Palmer & Donahoe,
1992). This is done by experimentally analyzing behavior and looking for
order therein. Consider, for example, the behaviors involved in solving
a particular mathematical problem. A conventional, essentialistic approach
would assume that children possess mathematical abilities (i.e., intelligence)
to varying degrees. It follows, then, that although we may attempt to teach
all children how to solve the mathematical problem, some will learn to
solve it with greater ease than others. Those who learn to solve the problem
easily are said to possess more mathematical ability than those who do
not. Using a functional approach, researchers might first try to determine
through experimental analysis what constituent behaviors are involved in
solving such problems in general and in what order those behaviors must
occur. Teaching methods based on such an analysis can then be constructed
to increase the chances that as many children as possible learn the behaviors.
A functional approach cannot only provide better and more elegant explanations
of the behavior by pointing to the variables of which the behavior of interest
is a function, but it also can ensure through the application of the scientific
principles that a greater proportion of children will be able to perform
the behaviors under the appropriate circumstances whether it is solving
mathematical problems, thinking and reasoning, or writing poetry.

Some psychologists will object to an approach that discounts questions
about the nature or definition of intelligence and focuses only on the
behaviors said to reflect it. Therefore, it is not unreasonable to ask
whether there is really any advantage in positing the existence of intelligence
separate from behavior. The answer, I believe, is no. Returning to the
example above, if the only evidence for mathematical ability is the very
behaviors necessary to solve the mathematical problem and from which we
infer mathematical ability (or the lack thereof), then a more productive
approach is to analyze the variables responsible for those behaviors. By
focusing on mathematical ability, we might be distracted from understanding
the function of the very behaviors that lead us to talk about that ability
in the first place. Once scientists are able to produce the mathematical
behaviors by arranging the variables responsible for them, talking about
some construct called "intelligence" seems superfluous except perhaps,
as Howe (1990) suggests, as a broad summary term or as a descriptive term
in applied psychology to refer to how well someone might perform in a given
context.

In contrast to an essentialistic approach, a functional approach suggests
at least two different questions with respect to intelligence or any other
construct. First, what behaviors in what contexts cause us to use the term
intelligence, and second, what accounts for the behaviors? The second question
is about the actual behaviors we call intelligent, but the first question
is largely about the verbal behavior of psychologists and social theorists.
We may shed some light on the answer to this question by looking at the
derivation of the term intelligence. The word intelligence comes from the
Latin intellegere, meaning to perceive or understand, from the roots inter
meaning between or among, and legere, meaning to gather, pick or choose.
It is interesting to note that these roots do not refer to inferred essences
or qualities, but rather to behaviors, in this case, gathering, picking,
and choosing. Thus, intelligence, or, more aptly, intelligent behavior,
is what we observe when we say that an individual perceives (i.e., reacts
to) relationships between or differences among situations.

With the emphasis shifted from intelligence as a unique quality or qualities
to the actual behavior we call intelligent, it may be useful to revisit
the logical errors made by Spearman and his intellectual descendants regarding
the concept of intelligence with an eye toward correcting similar errors
we make in our everyday language. First, when we speak of intelligence,
we are not speaking of a thing or an entity that individuals possess and
that determines their behavior. Intelligence is simply a descriptive term
for an instance of behavior in a particular context or for a group of related
behaviors (e.g., those called verbal, spatial, mathematical, etc.). We
are fooled into believing that the term refers to a tangible entity because
other words in our language, such as cat or table, have tangible referents.
Intelligence, however, does not have a referent per se. Intelligence is
like other terms such as mind or personality in that the only objective
referents are the behaviors that occasion the terms. And such behaviors
and their contexts are as varied as the number of different types of cats
that cause us to say "cat," although they obviously have some common features
that enable one summary term to apply to them.3

This simple locution has significant and radical implications for the way
we talk about intelligence. Alfred Binet, the originator of the modern
intelligence test, seemed to know this intuitively. Rather than referring
to intelligence, and defining it a priori, which would have tempted him
to search for the referent, Binet used an adjectival form, intellectual
skill, where skill was synonymous with behavior. Once it is understood
that what we speak of as intelligence is more appropriately spoken of as
intelligent behavior, then we can begin to examine the variables of which
the behavior is a function. Once the functional determinants of intelligent
behavior are discovered, they can be altered, as Binet believed, through
teaching. This optimistic view is not just wishful thinking; scientifically
based teaching techniques have for years demonstrated how intelligent behavior
can be effectively taught (e.g., Johnson & Layng, 1992).

Another significant implication of this simple change in language is that
instead of the endless debate about whether there are different types of
intelligence, or the differences between academic, practical, and emotional
intelligence, behavior scientists can experimentally analyze the behavior
in which they are interested. Moreover, they can no longer say that intelligence
tests measure intelligence; rather, they measure the correctness or appropriateness
of certain behaviors in contrived contexts (Schlinger, 1992). Finally,
it is erroneous to say that some observed behavior occurred because the
person was intelligent or because they possessed intelligence. These types
of explanations simply take the name or label given to a behavior and convert
it into the explanation of that very same behavior; in other words, they
are circular explanations. Understanding behavior in the scientific sense
means being able to specify the historical and contemporary conditions
or variables necessary for its occurrence, and this is accomplished only
through systematic experimentation.

From General Intelligence to Specific Skills:
Toward a Functional Environmental Approach to Intelligent Behavior

Recall that even though Spearman argued for a general intelligence underlying
all skills, he did posit specific factors, or s, that influenced specific
skills or abilities. As I outlined previously, in the last two decades,
some psychologists have gone beyond the concept of a unitary general intelligence
and have suggested many different types of intelligence, some operating
autonomously, from three (Sternberg, 1984) to seven (Gardner, 1983). Despite
their apparent break from traditional intelligence theorists who accept
g as a unitary general capacity (neither Sternberg nor Gardner signed the
statement, "Mainstream Science on Intelligence," that appeared in the Wall
Street Journal after the publication of The Bell Curve), both Sternberg
and Gardner accept g in a restricted sense. What is important for the present
argument, however, is that psychologists are gradually coming to the realization
that in contrast to the approach that there is one general intelligence,
measurable by standardized tests, and reflecting considerable heritability,
it may be more scientifically productive to investigate individual skills
in an effort to determine their causes. As a group, behavioral psychologists,
like Alfred Binet before them, are already ahead of the game. Behaviorists
have known for some time that intelligence is simply a term that refers
loosely to numerous behaviors the performance of which cannot be separated
from their specific contexts. Cognitive psychologists, (e.g., Howe 1989,
1996) are moving toward a more behavioral view by concentrating on a person's
abilities or skills, not as reflective of any general ability or intelligence,
but rather as autonomous behaviors that may be learned independently (although
any two or more may share common features), and are under the control of
the stimuli that define their specific context.
If psychologists really want to understand the behaviors that we call intelligent
and that are often tested on intelligence tests, perhaps they should look
at more objective variables that do not reify intelligence or posit any
hypothetical constructs. A major step in that direction was taken by Hart
and Risley (1995) in their longitudinal study of 40 American children from
birth until 2 1/2 years of age. In this study, language interactions between
children and their families were observed in the home 1 hour a month for
2 1/2 years. Of several family characteristics, including child gender
and race, and whether both parents were employed, the only one that made
a difference in the amount of talk a child heard was socioeconomic status.
For example, 11 - to 18-month-old children of professional parents heard
an average of 642 utterances per hour of which 482 were addressed to the
children, whereas children of welfare parents heard an average of 394 utterances
of which only 197 were addressed to the children. A closer look at the
data revealed strong correlations between what the parents were actually
doing during interactions with their language-learning children and the
children's language development. Specifically, the researchers found strong
correlations between the children's cumulative vocabulary growth rate and
several derived variables of family experience, in particular, language
diversity (the sum of different nouns plus modifiers) and feedback tone
(approval plus repetition of children's utterances divided by approval
plus prohibitions). These correlations, unlike those with the families'
socioeconomic status, held up at age 3, and with performance on standardized
achievement and intelligence tests (i.e., IQ scores) when the children
were 9 years old and in the 3rd grade. In addition, negative correlations
were found between the richness of parent initiations, imperatives, and
prohibitions and their children's vocabulary growth and use. In particular,
children in lower SES families heard proportionally more imperatives and
prohibitions from parents and siblings and they initiated verbal interactions
less often.

This major study by Hart and Risley, unlike the research conducted by Spearman's
intellectual descendants that relies on factor analysis, correlates the
actual behaviors of parents and children in the home during the critical
years of language learning and intellectual development. The results support
the assertion that what the parents do at home is strongly related to the
child's intellectual behaviors both at home and later in life and offers
a more productive approach to understanding the genesis of some of the
behaviors we call "intelligent." Such data set the stage for behavior theorists
to interpret the interactions between parents and children according to
the principles of operant and social learning. As a result, the ultimate
causes of the children's behaviors and the importance of the first three
years of life may come into sharper focus.

Counterpoint

Many psychologists may contest the assertion that intelligence is still
reified in modern psychology and that it is only a label for behaviors
in certain contexts. Their argument is that intelligence (or any other
similar construct) is at the very least a hypothesis to be tested. If no
evidence can be found for intelligence other than behaviors in certain
contexts, then its value as a scientific concept is diminished. So what
other evidence for intelligence is there? Intelligence theorists assert
that general intelligence as measured on a variety of standardized tests
correlates with a number of other educational, occupational, economic,
and social variables (Jensen, 1994), as well as various CNS measures such
as brain size, speed of nerve conduction, and EEG pattern (Jensen & Sinha,
1993). In assessing the potential explanatory power of the concept of intelligence,
Howe (1988b) looked at 10 potential conditions, any one of which, if confirmed,
could warrant using the term "intelligence" in an explanatory fashion.
These conditions included the relationship between measured intelligence
to observable physiological correlates, variability in basic mental processing
mechanisms such as reaction time, the ability to learn and remember, and
the complexity of a person's cognitive functioning, among others. Howe
concluded that any observed correlations were either insupportable scientifically
or there were alternative and more parsimonious explanations. Howe (1988b)
explained:

Although the possibility that an indication of a person's measured intelligence
can be explanatory or informative ... cannot be ruled out absolutely, there
are no strong grounds for believing that identification of someone's measured
intelligence justifies any meaningful statement about the individual's
qualities, achievements, attributes, or even detailed predictions except
in narrowly circumscribed circumstances. (p. 358)

Of course, intelligence theorists and behavior geneticists interested in
intelligence have argued for decades that the factor analysis of scores
on standardized intelligence tests provides substantial evidence for the
existence of intelligence as a mental or cognitive capacity as well as
for its heritability. Their claims have always rested on the legitimacy
of their use of factor analysis. However, the use of factor analysis on
standardized intelligence test scores has been seriously called into question
by a number of authors, but none more convincing than the Harvard astrophysicist
David Layzer. In 1972, just 3 years after Arthur Jensen's article, "How
Much Can We Boost 10 and Scholastic Achievement?" appeared in the Harvard
Educational Review, and just 1 year after Richard Herrnstein's article,
"10," appeared in the Atlantic Monthly, Layzer, in his article "Science
and Superstition: A Physical Scientist Looks at the IQ Controversy" (1972,
1995) (which has been reprinted in two edited books since) provided a comprehensive
and scientific critique not only of the statistical approach taken by Spearman's
intellectual descendants, but also the underlying assumptions and rationale
for their research. Layzer concluded that all estimates of the heritability
of IQ are "unscientific and indeed meaningless" in part because (a) the
research violates the equal environment assumption mentioned previously,
which in population genetic terms means that the range of relevant environmental
variations is assumed to be small; (b) intelligence theorists mistakenly
assume that intelligence represents a metric character like height or weight,
and that IQ tests measure that character, and, relatedly, because (c) IQ
tests do not measure what they purport to. Layzer (1995) explains:

IQ does not measure an individual phenotypic character like height or weight;
it is a measure of the rank order or relative standing of test scores in
a given population. (p. 666)

Finally, it could be argued that "intelligence" as an unobservable hypothetical
construct that determines behavior is a hypothesis to be tested and not
to be rejected out of hand. This argument is based on the fact that in
other scientific fields hypothesized unobservable constructs, such as extrasolar
planets or genes, were subsequently confirmed after refinements in observational
technology. However, as I have pointed out elsewhere (Schlinger, 1998),
inferring unobserved constructs in other sciences is predicated on an already
established factual base of experimentally derived relationships. The field
of intelligence research has no such base and, in fact, is built solely
on a-correlational foundation which renders conclusions about an internal
construct that we may call "intelligence" premature at best and completely
misdirected at worst. Among other things, this means that any significant
correlations between 10 scores and variables such as brain size or function
do not necessarily confer additional meaning on g or intelligence. At best
they represent possible neurophysiological underpinnings of behaviors,
some of which might be assessed with intelligence tests. They do not, however,
represent causes in any scientific sense of the term.

Why Has The Myth of Intelligence Persisted?

Given the logical problems with the concept of intelligence as it has evolved
in modern psychology and problems with many of the research methods used
to study it, one might ask why most psychologists persist in mythologizing
it. One simple answer is that as psychology students are trained, the myth
is passed on through the literature, only a small sample of which I have
referenced in this article. The myth of intelligence as a distinct quality
possessed by individuals, however, is not unlike the myth of personality,
memory, or the myriad other distinct faculties human beings are said to
possess. Thus, conceptualizing intelligence in essentialistic ways is part
of a broader approach to human behavior that has persisted for centuries.
It may have had its roots in primitive animistic philosophies of human
behavior, but it has become the standard vocabulary of ordinary citizens.
Whereas the other sciences have developed technical vocabularies distinct
from the primitive ones used by early philosophers, the vocabulary of psychology,
with some exceptions, is still mired in early animistic-like philosophies
that place many of the causes of human behavior inside the individual in
terms of mind, will, and so forth. Howe (1990) explains:

Psychology is a difficult scientific discipline for the unusual reason
that we come to it already furnished with firm habits of thought that have
been acquired from daily exposure to folklore and the unscientific (and
sometimes illogical) "commonsense" psychological thinking that permeates
everyday life. (p. 490)

Another reason that some psychologists persist in believing in intelligence
is that they have a significant personal and professional investment in
it. For example, those of Spearman's intellectual descendants who continue
to assert the heritability of general intelligence do so, at least in part,
because of a strong bias in favor of heritability and against environmental
determinants of intelligence. Beckwith (1999) writes:

What the history of IQ research shows is that fundamental assumptions or
mind-sets that permeate the field reflect ab initio attitudes toward the
relative importance of environment in the development of human behaviors
and aptitudes. (p. 168)

And contrary to claims by some intelligence theorists that their conclusions
about general intelligence and its heritability have no ab initio implications
for social policy, ever since Galton first posited the heritability explanation
of individual differences and eugenics as a solution to a general decrease
in intelligence, the initial assumptions of those who conduct intelligence
research have repeatedly been used to promote social policy (Beckwith,
1999). Beckwith notes:

What is most striking about this state of affairs is that the simplifying
assumptions practically dictate the conclusions. The research field appears
to have been imbued with a social perspective from its outset to its presentation
to the public. (1999, p. 168)

Coda

Intelligence, then, as an essence or quality, is a myth. Intelligence is
first and foremost a word that psychologists and others use to refer to
various behaviors in varying contexts. It may be useful for social reasons
to distinguish between behaviors that are more relevant to academic, practical,
or emotional contexts, or to subdivide and categorize behaviors according
to other socially relevant criteria. However, we should remember that these
are not scientific distinctions, at least not yet. At best they are arbitrary
social distinctions (verbal discriminations), regardless of whether they
turn out to be useful or not. Even the recent report on intelligence issued
by the Task Force established by the American Psychological Association
concluded: "Because there are many ways to be intelligent, there are also
many conceptualizations of intelligence" (Neisser et al., 1996, p. 95).
Taken to its extreme, we could potentially postulate a different kind of
intelligence for each instance of intelligent behavior observed. Palmer
and Donahoe (1992) remind us, however, that although we are free to define
such categories as we please, they may not reflect distinctions in nature.
It is for scientists to determine the natural lines of fracture of the
particular phenomenon under study, and they will do so by finding order
in the subject matter through experimentation.

Psychologists who take a more functional approach will necessarily follow
a different path than traditional intelligence researchers have. A functional
route will have Alfred Binet, E. L. Thorndike, and B. F. Skinner as the
progenitors rather than Francis Galton and Charles Spearman and it will
emphasize behaviors that we call intelligent rather than statistical constructs.
It will consist of experimental (rather than correlational) methods for
determining the controlling variables of the behavior and will emphasize
teaching methods based on discovered functional units of behavior. Rather
than debating the unresolvable question of whether some statistical factor,
g, is heritable, those interested in intelligent behavior will recognize
that although nature and nurture interact to produce human behavior, the
hallmark of human beings is a flexible central nervous system which is
especially sensitive to environmental input early in life. Overall, this
more functional approach will lead to new ways of thinking about how the
environments of human beings can be structured to generate more intelligent
behavior.

1With some exceptions, the list of cosigners reads like a Who's Who of those
theorists (e.g., Thomas J. Bouchard, Jr., John B. Carroll, Raymond B. Cattell,
Hans Eysenck, Linda S. Gottfredsen, Seymour W. Itzkoff, Arthur R. Jensen,
Robert Plomin, J. Philippe Rushton, and Vincent Sarich) who have continued
Spearman's tradition of factor analyzing intelligence test scores to generate
a theory of general intelligence - g - and some of whom (e.g., Thomas J.
Bouchard, Robert Plomin) claim that behavior genetic research supports
the conclusion that g is highly heritable, and others of whom (e.g., Arthur
Jensen, J. Philippe Rushton, Seymour Itzkoff) have written highly emotionally
charged articles arguing that the research supports the conclusion that
group differences on intelligence tests reflect genetic differences.

2Mayer (1999) and Mayer, Caruso, and Salovey (2000) contend that the concept
of emotional intelligence popularized by others and by the media, stretches
the meaning of their term-usually as a list of personality characteristics,
which they call a "mixed conception," and overestimates the predictions
about important life outcomes or about people who are said to be highly
emotionally intelligent.

3Howe (1990) makes a similar argument about the error of reification in
which the assumption is made that just because the noun intelligence exists,
that it refers to some underlying thing or entity.

References

BECKWITH, J. (1999). Simplicity and complexity: Is IQ ready for genetics?
Cahiers de Psychologie Cognitive, 18 (2), 161-169.

BOLLES, R. C. (1993). The story of psychology: A thematic history. Pacific
Grove, CA: Brooks/Cole.

CARROLL, J. B. (1993). Human cognitive abilities: A survey of factor-analytic
studies. New York: Cambridge University Press.

CRINELLA, F. M., & YU, J. (2000). Brain mechanisms and intelligence. Psychometric
g and Executive Function. Intelligence, 27(4): 299-327. FANCHER, R. E.
(1985). The intelligence men: Makers of the IQ controversy. New York: Norton.

GARDNER, H. (1983). Frames of mind: The theory of multiple intelligences.
New York: Basic Books.

GOLEMAN, D. (1995). Emotional intelligence: Why it can matter more than
IQ. New York: Bantam.

GOTTFREDSON, L S. (1998). The general intelligence factor. Scientific American
Presents. [Orn-ine] Available: www.seam.corm/1998/1198intelligene/11 98gottfred.html
GOULD, S. J. (1981). The mismeasure of man. New York: Norton.

HART, B., & RISLEY, T. R. (1995). Meaningful differences in the everyday
experiences of young American children. Baltimore: Paul Brookes.

HERRNSTEIN, R. J., & MURRAY, C. (1994). The bell curve: Intelligence and
class structure in American life. New York: The Free Press.

HOWE, M. J. A. (1988a). The hazards of using correlational evidence as
a means of identifying the causes of individual ability differences: A
rejoinder to Sternberg and a reply to Miles. British Journal of Psychology,
79, 539-545.

HOWE, M. J. A. (1988b). Intelligence as an explanation. British Journal
of Psychology, 79, 349-360.

HOWE, M. J. A. (1989). Separate skills or general intelligence: The autonomy
of human abilities. British Journal of Educational Psychology, 59, 351-360.

HOWE, M. J. A. (1990, November). Does intelligence exist? The Psychologist,
490-493.

HOWE, M. J. A. (1996). Concepts of ability. In I. Dennis & P. Tapsfield
(Eds.), Human abilities: Their nature and measurement. New York: Erlbaum.

JENSEN, A. J. (1969). How much can we boost IQ and scholastic achievement?
Harvard Educational Review, 39, 1-123.

JENSEN, A. R. (1994, December 5). Paroxysms of denial. National Review,
46 (23) pp. 48-51.

JENSEN, A. R. (1998). The g factor. Westport, CN: Praeger.

JENSEN, A. R., & SINHA, S. N. (1993). Physical correlates of intelligence:
A review. In P. A. Vernon (Ed.), The biological basis of intelligence.
Norwood: Ablex.

JOHNSON, K. R., & LAYNG, T V. J. (1992). Breaking the structuralist
barrier: Literacy and numeracy with fluency. American Psychologist, 47, 1475-1490.

JOSEPH, J. (1998). The equal environment assumption of the classical twin
method: A critical analysis. The Journal of Mind and Behavior, 19, 325-358.

KAMIN, L. J. (1974). The science and politics of IQ. Potomac, MD: Lawrence
Erlbaum.

KAMIN, L. J. (1995, February). Behind the curve. Scientific American,
272, 99-103.

LAYZER, D. (1972). Science or superstition? (A physical scientist
looks at the IQ controversy). Cognition, 1, 265-299.

LAYZER, D. (1995). Science or superstition? In R. Jacoby & N. Glauberman
(Eds.), The bell curve debate: History, documents, opinions. New York:
Times Books.

LAYZER, D. (1999). Comment on "Misconceptions of Biometrical IQists" by
C. Capron et al. Cahiers de Psychologie Cognitive, 18 (2), 214-216.

MAINSTREAM SCIENCE ON INTELLIGENCE. (1994, December 13). Wall Street Journal.
[On-line] Available: www.mugu.com/cgi-bin/Upstream/Issues/bellcurve/support-bell-curve.htm

MAYER, J. D. (1999, September). Emotional intelligence: popular or scientific
psychology?. APA Monitor On Line. [On-line) Available: www.apa.org/monitor/sep99/sp.html

MAYER, J. D., CARUSO, D. R., & SALOVEY, P. (2000). Emotional intelligence
meets traditional standards for an intelligence. Intelligence, 27, 276-298.

MAYER, J. D., & GEHER, G. (1996). Emotional intelligence and the identification
of emotion. Intelligence, 22, 89-113.

MAYER, J. D., & SALOVEY, P. (1993). The intelligence of emotional intelligence.
Intelligence, 17, (4), 433-442.

MAYR, E. (1982). The growth of biological thought. Cambridge, MA: Harvard
University Press.

NEISSER, U. (1976). General, academic, and artificial intelligence. In
L. B. Rescind (Ed.), The nature of intelligence (pp. 135-144). Hillsdale,
NJ: Erlbaum.

NEISSER, U., BOODOO, G., BOUCHARD, T. J., BOYKIN, A. W., BRODY,
N., CECI, S. J., HALPERN, D. F, LOEHLIN, J. C., PERLOFF, R., STERNBERG, R.
J., & URBINA, S. (1996). Intelligence: Knowns and unknowns. American Psychologist,
51, 77-101.

PALMER, D. C., & DONAHOE, J. W. (1992). Essentialism and selectionism in
cognitive science and behavior analysis. American Psychologist, 47,1344-1358.

PLOMIN, R. (1990). The role of inheritance in behavior. Science, 248, 183-188.

PLOMIN, R. (1999, December 2). Genetics and general cognitive ability.
Nature, 402, C25-C29.

POPPER, K. (1957). The poverty of historicism. Boston: Beacon Press.

SCHLINGER, H. D. (1992). Intelligence: Real or artificial. The Analysis of Verbal
Behavior, 10, 125-133.

SCHLINGER, H. D. (1993). Learned expectancies are not adequate scientific
explanations. American Psychologist, 48, 1155-1156.

SCHLINGER, H. D. (1996). How the human got its spots: A critical analysis
of the just so stories of evolutionary psychology. Skeptic, 4, 68-76.

SCHLINGER, H. D. (1998). Of planets and cognitions: The use of deductive
inference in the natural sciences and psychology. The Skeptical Inquirer,
22, 49-51.

SPEARMAN, C. (1904). General intelligence, objectively determined and measured.
American Journal of Psychology, 15, 201-293.

SPEARMAN, C. (1927). The abilities of man. London: Macmillan.

STERNBERG, R. J. (1984). Toward a triarchic theory of human intelligence.
Behavior and Brain Sciences, 2, 269-315.

STERNBERG, R. J., & DETTERMAN, D. K. (Eds.). (1986). What is intelligence?
Contemporary viewpoints on its nature and definition. Norwood, NJ: Ablex.

STERNBERG, R. J., WAGNER, R. K., WILLIAMS, W. M., & HORVATH, J. A.
(1995). Testing common sense. American Psychologist, 50, 912-926.

WEINBERG, R. (1989). Intelligence and IQ. American Psychologist, 44, 98-104.

I am grateful to David Palmer, Julie Riggott, and two anonymous reviewers
for their helpful comments and suggestions. Correspondence concerning this
article should be addressed to Henry D. Schlinger, Department of Psychology,
California State University, Los Angeles, Los Angeles, CA 90032-8277. E-mail: hschlin@calstatela.edu

The Psychological Record was granted permission to reprint this article. The full citation is:
Schlinger, H. D. (2003). The myth of intelligence. The Psychological Record, 53, 15-32.