inventory are positively correlated (Kline, 1991). Spearman thought that the
prevalence of positive correlations reflected a physical property of the brain,
namely, a kind of mental energy that some brains happened to possess more of
than other brains (Spearman, 1927). He labeled this idea ‘‘g,’’ to stand for the
general factorthat underlies all intellectual activity. More recent but similar
interpretations of g are that g reflects the capacity to pay attention to informa-
tion (Hunt, 1980; Jensen, 1979), reflects nerve conduction velocity and rate of
neural decay (Jensen, 1993), or reflects the ability of neurons to change con-
nections (Larson & Saccuzzo, 1989).
Spearman, and many others since, have noted that subtests of the IQ inven-
tory that are similar to one another are even more positively correlated than are
dissimilar subtests. For example, two different subtests that measure spacial
reasoning will be more highly correlated than a subtest that measures spacial
reasoning and a subtest that measures vocabulary. This pattern of correlations,
analyzed by a statistical technique called factor analysis, is sometimes in-
terpreted as indicating that intelligence has a general (also known asfluid)
component that reflects some genetically determined biological aspect of the
cognitive system, and a series of specialized (also calledcrystallized)compo-
nents that reflect various learned skills (Kline, 1991).
There is other evidence for the unitary nature of intelligence. Correlations
among IQ tests are significant even when one IQ test is verbal and the other IQ
test is nonverbal. For example, the correlations between the Raven’s Matrices (a
nonverbal IQ test) and conventional IQ tests range from aboutþ.40 toþ.75
(Anastasi, 1988). That IQ scores predict performance in very different situa-
tions, such as school settings and job settings, also suggests that there is a uni-
tary aspect to intelligence.
What Underlies Unitary Intelligence? Contributions of Information Processing
Elsewhere I have criticized information processing models that postulate that
all problems are solved by the same, generic information processing system. A
similar sort of information processing perspective has been used as an account
for why intelligence seemingly has a unitary character.
A generic information processing approach to intellectual differences has all
intellectual tasks performed by a single information processing system. Indi-
vidual differences in intellectual performance reflect differences in the speed
and efficiency with which the various components of the system are executed.
I do wish to note that information processing cognitive psychologists need
not postulate a generic information processing model of individual differences.
Perhaps human cognition is composed of many different, relatively autono-
mous information processing systems. However, the idea of a generic informa-
tion processing system is implicit in most information processing approaches
to cognition (Lachman, Lachman, & Butterfield, 1979), and so it is the generic
form of it that I will critique here.
The information processing approach rose to prominence in the 1950s, 1960s,
and 1970s. An important claim of information processing is that any given
cognitiveprocesscanbebrokendownintoasetoffundamentalcomponents,
such as perceiving information, transforming information, storing information
in memory, and retrieving symbols from memory. Most information processing
784 R. Kim Guenther