This book was awarded the 2002 Eleanor Macoby Book
Award by the American Psychological Association. This award is given to a book
that has had or promises to have a profound impact on developmental psychology.
Thus the book has already been judged by other professionals to be an important
one. Yet the book is clearly intended also for a general audience, and appears
to already have had some success at attracting such an audience. (For instance,
Bloom was interviewed about this book on the NPR talk show The Todd Mundt Show on June 3, 2002). It is written in a clear,
accessible style and doesn’t have any of the detailed reports of statistical
analyses that are de rigueur in
articles published in professional psychological journals. Data tables and
graphs are kept to a minimum, and where they are used they are usually very
easy to interpret.

Bloom says that each of the 9 main chapters is
“self-contained enough to be read on its own, but they do have a logical
progression, and each rests to some extent on evidence and arguments introduced
earlier.” (p. 22) I agree, though I would say that to appreciate the main points
Bloom is arguing for, it is necessary to read chaps. 2 – 6 and 8. The remaining
chapters do seem relatively independent, taking up interesting but somewhat
tangential issues, such as what is involved in distinguishing representation
from reality and how children learn to name pictures, how children learn number
words, and what the relation is between thought and language. (Personally I
found these chapters to be amongst the most interesting, so I am not suggesting
that they be skipped!)

The main purpose of Bloom’s book is to explode some
myths about early word learning. He begins his book with the following
scenario:

It looks simple. A 14-month-old toddles after the family dog, smacking
it whenever she gets close. The dog warily moves under the table. “Dog”, the
child’s mother tells her. “You’re chasing the dog. That’s the dog.” The child
stops, points a pudgy hand at the dog, and shrieks, “Daw!” The mother smiles:
“Yes, dog.”

Many people have the impression that word leaning is
the simplest part of language learning, and that children learn words by a
simple associative process in which caregivers name objects in the children’s
environment, thereby enabling the children to set up mappings between words and
their meanings. Bloom argues that early word learning is far from simple and in
fact relies on several sorts of capacities. These include “an understanding
that the world contains objects, events, and relations, kinds and individuals;
an appreciation that the nature of some categories does not reduce to their
superficial features; an ability to appreciate the referential intentions of
others, to understand what they are referring to when they communicate.” (p.
258) As Bloom goes on to observe, many philosophers and psychologists have
thought that these capacities are the product
of word learning, so his claim that they are prerequisites for early word learning stands the traditional
associationist picture on its head.

One of the central anti-associationist claims is
defended in chapter 3. It is that early word learning requires children to
recognize a speaker’s referential intentions. This means that even very young
infants must have some sort of “theory of mind”. Some of the main evidence in
favor of this claim comes from the very interesting work of Dare Baldwin and
her colleagues. In one experiment children were given an object to play with
while another was placed in a bucket. While the child’s attention was focused
on the object she had been given, the experimenter looked at the object in the
bucket and said “It’s a modi”. When later asked to find the modi, children as
young as 18 months were able to pick out the object that the adult had named.
In other words, they did not simply associate the novel name with the novel
object that they were paying attention to when the experimenter said “It’s a
modi”. Instead, children were able to infer the speaker’s referential
intentions (presumably by relying on cues such as the direction of the adult’s
gaze – although gaze following itself is something that children seem to engage
in only when they are interacting with entities that show signs of having
intentional states and of acting in a goal-directed manner).

This chapter assembles a lot of other evidence in
favor of the claim that a “theory of mind” (ToM) is essential to early word
learning, as well as addressing some of the main objections that might be
raised against this claim. The chapter also contains a section on autism and
language learning. It has been claimed that autism involves a deficit in the
“theory of mind module” (ToMM), the faculty of the mind that enables normal
people to understand the beliefs, desires and intentions of others. This raises
the question of how autistic individuals can learn word meanings if they have a
deficit in their ToM and yet a ToM is a prerequisite for word learning. Bloom
presents evidence that severely autistic individuals do indeed have problems
learning words, and that those “who have relatively preserved language skills
are the same individuals who tend to perform well on tasks designed to tap
their understanding of the thoughts of other people.” (p.80).

Many people have been convinced by the work of
Perner & Wimmer, Wellman, and others that a theory of mind is a late
acquisition, since children younger than three years of age do badly on the
so-called “false belief” task, which is supposed to be a test of a child’s
ability to ascribe representational mental states to others. In one version of
this task an object is hidden in one location in the presence of the child, an
experimenter and a puppet. The puppet then “leaves” the scene (often by simply
being hidden under the table). After the puppet leaves, the experimenter moves
the object to a second location. The child is then asked where she thinks the
puppet will look for the object when the puppet returns. Children under the age
of three generally have difficulties with this task, and say that the puppet
will look in the second location, where the object actually is, even though the
puppet was (supposedly) not present when the object was moved, and so should be
assumed not to know it has been moved. If a theory of mind is a late
acquisition, and yet children have begun to learn words before their second
birthdays, then it may seem that we are forced to the conclusion that word
learning cannot require a theory of mind. I believe that this is a mistake.
Whatever the “false belief” task is testing, it does not show that young
children do not have a theory of mind. Bloom does a convincing job in his book
of assembling evidence that children as young as 15 months have an awareness of
the intentions of others, and that they are able to harness their
“mind-reading” abilities in word learning.

I have focused on children’s theory of mind in this
review, but Bloom’s book explores a wealth of other issues. Besides the myth
that word learning is a simple associationist process, there are other myths
that Bloom tries to explode. For instance, children do not have to have adults
around to name objects for them in order to learn words. In some cultures
adults do not engage in the sort of object labeling practices that some Western
parents engage in. Yet children in those cultures still learn the meanings of
words. They are able to infer the meanings of words by observing the
conversational exchanges of others.

Another popular conception that Bloom challenges is
the idea that the process of word learning imposes some sort of conceptual
order on the child’s awareness of the world, which is initially a “blooming,
buzzing confusion”. On the contrary, Bloom assembles evidence that children
bring some already well-formed conceptual apparatus to bear on the task of word
learning. For example, they must already have concepts of ordinary middle-sized
objects and their identity and persistence conditions. They must have some
conception (perhaps arrived at by abstraction and generalization from their
concept of an object) of parts and collections of objects. They must have an
appreciation of the fact that what unites objects under a common natural or artifact
kind is often something unobservable. And as already mentioned at some length,
they must have some concept of other speakers as intentional agents. In short,
children bring both a theory of objects and a theory of mind to the task of
word learning.

An important component of word learning that I
haven’t yet mentioned is the role of syntax in word learning. Bloom discusses
this in chapters 5, 8 and 9. In these chapters he discusses how children learn
proper names, pronouns, substance names, verbs, adjectives, prepositions, and
number words. In all these cases children must be relying partly on syntactic
clues to figure out what sort of thing a word is referring to (whether it
refers to an individual or to a kind, whether to an object kind or a substance kind,
whether to a property or a relation, whether to a property of an object or a
property of a set of objects). For
instance, English-speaking children are sensitive to the different syntactic
environments in which mass and count nouns occur and can use this as a clue to
whether they are dealing with an object kind or a substance kind. (Children
learning languages that do not syntactically mark the mass/count distinction –
such as Chinese – must obviously be relying on non-syntactic means to figure
this out). However, Bloom stresses that such syntactic clues are only a part of
the story and that it is syntactic abilities together with a theory of objects
and a theory of mind that together explain early word learning. Moreover, he
maintains that some ur-words must be learned independently of any syntactic
bootstrapping.

Besides the popular misconceptions about word
learning that Bloom wishes to undermine, he also takes on certain claims that
have been made by other child language researchers. For instance, he raises
some skeptical worries about the claim that children go through a word spurt at
around 50 words. The idea of such a word spurt may be more terminological than
real. A child could satisfy the criteria for having experienced a word spurt on
some definitions (e.g., learning 10 or more new object words in a 3 week
period) even though she learned these words at a steady rate, so that nothing
spurt-like happened in the child’s experience. Bloom also challenges the idea
that there is any sort of discontinuity between early word learning and later
word learning, or that children are better than adults at word learning. There
is no “critical period” for word learning as is arguably the case for
productive syntax. In general Bloom rejects the idea that word learning is
under the control of special-purpose language mechanisms. According to Bloom,
children rely on capacities that are used also in other non-language domains,
such as the domains of social cognition and of spatial reasoning.

In this brief review I have not been able to do
justice to this richly argued book. I concur with those who awarded it the
Eleanor Macoby prize that this book is likely to have a profound impact on the
field of child language. It articulates a framework that challenges the traditional
associationist picture of word learning, and does an excellent job of
assembling the evidence that we already have in favor of a rival
anti-associationist view. It also does a good job of pointing out where there
are gaps in our understanding and where further research will be required.

©
2002 Anne Bezuidenhout

Anne Bezuidenhout
teaches at the University of South Carolina in the Philosophy Department and
the Linguistics Program. Her research interests lie mainly in the philosophy of
language, psycholinguistics and experimental pragmatics. Her work has appeared
in such journals as Philosophical Review,
Mind, Noûs, Mind & Language,
Journal of Pragmatics and Pragmatics & Cognition.

Categories: General, Philosophical