Gregory
Murphy’s The Big Book of Concepts is large not only in size and scope
but also in significance. Murphy
provides a much-needed overview of important developments in the psychological
research on concepts during the last two decades. An enormous number of experiments related to concepts have been
performed in areas such as category learning, word meaning, conceptual
development in infants, the basic level of categorization, typicality effects,
and exemplar effects. Murphy
brilliantly weaves together a staggering quantity of experimental results and
theoretical explanations into a unified and enlightening survey of this
exciting area of research.

Murphy’s
engaging and easy-to-read writing style makes his densely packed book very
enjoyable to read. An impenetrable
writing style would have made a book of this size intolerably tedious. Murphy’s uses of non-technical language
means that any educated person can understand and benefit from reading this
book.

Although
The Big Book of Concepts is quite accessible, it also has plenty to
offer the serious student or scholar interested in doing research on
concepts. In addition to being packed
with information, Murphy makes suggestions regarding promising avenues for
future research in every chapter. For
example, he notes that most experimental work on concepts has subjects go
through series or lists of items until they can categorize them correctly. The experimenter is always the one to decide
which items go in which category. In
ordinary life, however, subjects experience new objects and events on their own
and must create new categories in which to place them. This activity of ‘category construction’ has
not, in Murphy’s opinion, been adequately studied in the research on concepts. Another research suggestion of Murphy’s
concerns ‘concept use.’ Ordinary
subjects do not—as they are forced to do in psychological experiments—simply
memorize which categories several items fit into. Instead, they interact with these items and use the concepts they
acquire to solve a variety of practical problems. Research has shown that concept use affects categorization, but
the major theories of concepts have not taken this fact into account. Murphy’s insightful suggestions will be a
valuable resource to graduate students and other interested researchers looking
for projects to pursue.

Murphy
begins his book by describing the demise of the so-called ‘classical theory of
concepts,’ which traces its roots all the way back to Socrates. According to the classical theory, concepts
are represented in the mind as definitions—or, as philosophers like to say, as
sets of necessary and sufficient conditions for the application of those
concepts. Murphy provides a very
readable account of how the important discoveries of typicality effects made by
Eleanor Rosch and her colleagues helped to bring about the demise of the
classical theory.

Rosch
found that subjects agree to a surprising extent about whether an example of a
category counts as a good member of that category. Robins, for example, but not ostriches are always rated as
typical or exemplary members of the category ‘bird.’ Items are rated as typical if they have features that are common
to their category and do not have features that are common to other categories. Rosch also discovered that subjects make
judgments about category membership with greater ease when the given instances
are typical. When faced with sentences
containing atypical instances, subjects take longer to read those sentences and
are more likely to make errors on categorization tasks concerning them. The classical theory did not predict these
results and has difficulty handling other associated phenomena, such as
borderline cases of category membership and distinctions between typical and
atypical members of a category. The discoveries
of Rosch and her colleagues not only undermined the classical theory of
concepts; they also revolutionized and invigorated the psychological study of
concepts.

The
two dominant theories that have arisen in the wake of Rosch’s findings since
the 1970s are the prototype theory and the exemplar theory. The prototype theory maintains that concepts
are represented in long-term memory by their best or most typical
instances. For example, my concept
‘dog’ would be represented as an image of an average-sized dog with average
ears and hair length and an average body shape. On the prototype theory, categorization involves comparing an
observed item to stored prototypes and seeing which prototype it is most
similar to.

The
exemplar view maintains that subjects store information about many or all
instances of a category rather than merely a summary or generalized
representation of that category. So, my
concept of ‘dog’ would consist in the set of memories of particular dogs that I
have seen—everything from bulldogs to dachsunds to golden retrievers. Categorization, according to the exemplar
theory, involves comparing an item to all the stored exemplars of a particular
category and determining the number of exemplars to which the target is similar
and the degree of similarity it bears to those exemplars.

Murphy
provides a helpful summary of experiments that have compared the prototype and
exemplar theories in the attempt to demonstrate the superiority of one theory
over the other. He strives to be
balanced and fair in his assessment of these experiments, pointing out the
strengths and weaknesses of both theories. 
According to Murphy, the exemplar theory comes out ahead in most
experiments that directly compare the two theories. However, he suspects that this apparent fact may be an artifact
caused by the kinds of experiments that are normally chosen to distinguish the
two. These experiments typically 1) involve artificial categories that contain
only a few exemplars, 2) have a simple or weak category structure and 3) are rather
difficult for subjects to learn. By contrast, ordinary categories experienced
in common life 1) contain many instances, 2) have a robust and rich structure
and 3) are easier to learn than the contrived categories typically used in
psychological experiments. 
Consequently, it is not clear what can be inferred from these
experiments about the structure of ordinary concepts. Murphy also believes that the exemplar theory is less able to
explain many phenomena that are found in experiments that do not directly
compare the prototype and exemplar theories—e.g., discoveries about the basic
level of categorization, induction, and concept learning in infants.

Throughout
the book Murphy does a fine job of separating important data from theoretical
explanations of those data. For
example, there are prototype phenomena, and then there is the prototype theory
of concepts. Prototype phenomena are
simply data showing that the typicality of an item affects subjects’
performance on various learning, recognition and categorization tasks. In contrast, the prototype theory of
concepts is an account of how concepts are represented in the mind. Murphy also clearly distinguishes exemplar
effects and the exemplar theory of concepts. 
Exemplar effects are experimental findings showing that subjects
sometimes rely upon particular instances of a category that are stored in
memory while performing various cognitive tasks. Again, the exemplar theory, as a theory of conceptual structure,
is something else altogether.

Murphy
also discusses the effects of background knowledge on concept learning and
use. He notes that studies of concepts
have traditionally tried to separate as much as possible the tasks subjects
perform from any background knowledge those subjects may have. Subjects are typically asked to learn
artifical categories composed of certain geometric shapes, alphanumeric
strings, patches of color, and dot patterns. 
However, Murphy argues that it is also important to study how subjects
construct categories and make categorization judgments in cases where their
background knowledge may be relevant. 
Stimuli that subjects receive in real world settings are typically
richly structured and connected to categories and objects subjects know a lot
about. Exemplar and prototype theories
of concepts have typically neglected the role of knowledge in concept learning
and use. Knowledge-based theories (or
theory theories, as they are often called) are sometimes put forward as
promising successors to the exemplar and prototype theories because they can
handle knowledge effects. However,
those theories cannot handle the typicality and exemplar effects the way the
other two theories can. So, Murphy
thinks the abandonment of the prototype and exemplar theories in favor of
knowledge-based theories would be premature. 

Murphy
explains the rather clever methodologies that have been developed in recent
years to study the concepts of infants and the significant—and often
surprising—results they have yielded. 
He spends more time discussing developmental issues in the study of
concepts than is typical for a review of the concepts literature. The last two decades of developmental
research have overturned earlier conceptions of infants as lacking the same
sort of concepts adults have. Very young
infants reveal prototype effects, knowledge effects, and a preference for
basic-level categorization—just like adults. 
Whatever differences there are between concept-learning in young
children and concept-learning in adults, Murphy claims, can be chalked up to
differences in life experience, knowledge about specific domains, and
processing capacity.

Murphy
discusses many other topics, including the relationship between word meanings
and concepts, the use of category knowledge in induction, and conceptual
combination. If you can make it all the
way through this very big and splendid book on concepts, you will receive a
first-class introduction to the psychology of concepts. I highly recommend this book to students and
scholars who wish to know more about this exciting area of psychological
research.

© 2003 James R. Beebe

James Beebe, Ph.D., teaches philosophy
at Louisiana State University. His
research and teaching interests lie primarily in epistemology and the
philosophy of cognitive science.

Categories: Philosophical, Psychology