Cognitive and Language Development in Children - J. Oates, A. Grayson (Blackwell, 2004) WW
Cognitive and Language Development in Children - J. Oates, A. Grayson (Blackwell, 2004) WW
Cognitive and Language Development in Children - J. Oates, A. Grayson (Blackwell, 2004) WW
Development in Children
This publication forms part of an Open University course ED209 Child Development.
Details of this and other Open University courses can be obtained from the Student
Registration and Enquiry Service, The Open University, PO Box 197, Milton Keynes,
MK7 6BJ, United Kingdom: tel. +44 (0)870 333 4340, e-mail [email protected]
Alternatively, you may visit the Open University website at http://www.open.ac.uk where
you can learn more about the wide range of courses and packs offered at all levels by The
Open University.
To purchase a selection of Open University course materials visit www.ouw.co.uk, or
contact Open University Worldwide, Michael Young Building, Walton Hall, Milton Keynes
MK7 6AA, United Kingdom for a brochure. tel. +44 (0)1908 858785;
fax +44 (0)1908 858787; e-mail [email protected]
Development in Children
First published 2004 by Blackwell Publishing Ltd in association with The Open University
Reprinted 2006
MK7 6AA
www.blackwellpublishing.com
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
licence from the Copyright Licensing Agency Ltd. Details of such licences (for reprographic
reproduction) may be obtained from the Copyright Licensing Agency Ltd of 90 Tottenham
Open University course materials may also be made available in electronic formats for use
by students of the University. All rights, including copyright and related rights and database
rights, in electronic course materials and their contents are owned by or licensed to The
law.
In using electronic course materials and their contents you agree that your use will be
solely for the purposes of following an Open University course of study or otherwise as
Except as permitted above you undertake not to copy, store in any medium (including
or show in public such electronic materials in whole or in part without the prior written
consent of The Open University or in accordance with the Copyright, Designs and Patents
Act 1988.
A catalogue record for this title is available from the British Library.
Printed and bound in the United Kingdom by The Alden Group, Oxford.
1.2
Contents
Foreword
21
First words
61
MARGARET HARRIS
113
163
205
Understanding minds
231
EMMA FLYNN
259
303
Acknowledgements
333
Name index
336
Subject index
339
Foreword
We would very much like to acknowledge the valuable contributions that were
made by the many people who made this book possible. First, we would like to
thank our team of editors; Julia Brennan, Helen Lanz and Bridgette Jones. Their
careful and insightful reading of the chapters and the many changes that they
suggested have contributed immensely to the readability and clarity of the book.
We feel greatly privileged to have had the opportunity to work with all of our
consultant authors and we have welcomed their exibility in working with us to
produce what we hope will prove to be interesting and effective learning
material. We would also like to thank the critical readers Rachel George, Marianne
Puxley and Joanne Kelly for their encouragement and constructive criticism of the
drafts. Our academic colleagues Peter Barnes, Sharon Ding, Karen Littleton and
Clare Wood deserve thanks for their contributions to the book and for diligently
commenting on what must have seemed like an endless series of revisions. We
thank Iris Rowbotham for her supportive help with managing this project. Last,
but not least, we wish to thank Stephanie Withers for her patient and careful
secretarial help with typing and formatting, Jonathan Davies, the book designer,
Jon Owen, our illustrator, and Nikki Tolcher, our compositor.
John Oates
Andrew Grayson
Introduction: perspectives on
cognitive and language
development
John Oates and Andrew Grayson
Contents
1
10
understanding of grammar
2.5 Chapter 5: Executive functions in childhood
10
12
12
cognitive development
13
13
14
3.2 Constructivism
3.3 Social constructivism
14
17
References
19
11
11
Cognitive
Pertaining to the
gaining, storage
and use of
knowledge.
Two 3-year-old children, Tom and Amy, are in the living room, playing together
with some small toy gures and a toy horse and cart. The television is on in the
corner of the room and someone is clattering about in the kitchen cooking a
meal. Amy puts a gure into the cart and says, Off to market! Tom says, And
Daddy comes too, putting another gure into the cart. Amy then makes the horse
trot along, pulling the cart, while making clip-clop noises. Tom says, No, dont,
theyre not ready. Theyve lefted the children behind. Oh! Theyre crying and he
puts two child gures into the cart. Amy puts a small plastic brick in the cart and
says, Thats their food. Then Tom makes the horse trot along pulling the cart,
saying Off to market! and makes trotting sounds, as Amy had done earlier.
This is a simple, everyday scene, but one that contains a wealth of cognitive
activity. What knowledge have these two children gained and stored? How are
they using it? This scenario can be analysed at a number of levels and illustrates
many of the aspects of childrens cognitive development dealt with in this book.
First, we can note that these two children know a lot about people, and horses
and carts, and what they can do. This is shown in the way that they use this
knowledge to make the toy horse and cart do appropriate things like carrying
people to market. But this is an interesting sort of knowledge, since it is unlikely
that they have actually seen a horse and cart taking people to market. What they
may have done is to put together some pieces of knowledge gained from books
or other media, or from other children and adults, and used them in their play.
We can note also that Tom knows that the children will be upset at being left
alone. He appears to be able to understand the workings of other peoples minds
knowledge of a somewhat different kind.
In addition we can see that Tom and Amy both know how to organize their
behaviour into structured sequences; to have a plan and to put it into action. They
know how to work together on this, taking account of and adapting to each other.
This collaboration is demonstrated through, and dependent on, the language that
they both use in order to communicate with each other.
Furthermore, it seems likely that a certain amount of knowledge is being
gained in this brief episode. Perhaps Tom is acquiring something from observing
Amys way of playing, and maybe Amy becomes more aware of the emotional
consequences for children of being left behind.
But a most signicant aspect of this scene is the way that Tom and Amy are
knowingly using the toy objects to stand for equivalent objects in the real world
and the way in which they play with them as if they were real. This capacity to
symbolize, to use one thing to represent another, is a uniquely human capacity.
Amys and Toms interactions and play raise a number of important questions
relating to cognitive development:
.
How did they come to know about the characteristics of things such as
horses, carts, markets and people? At what point in their development were
they able to categorize Daddy and the children together in a way that
differentiated people from horses and vehicles?
At what stage did the two children learn the names of the objects that they
use? Did they have an understanding of the characteristics of these objects
before the names emerged, or were the names acquired before any real
understanding?
Are Amy and Toms brains pre-programmed to acquire the language that
they now use so easily and exibly, or have they had to learn it all from
others?
How have these children developed the ability to construct sentences out of
the words that they have acquired that enable them to communicate
apparently so effortlessly with each other? What about the sentence Theyve
lefted the children behind? Has Tom copied that from someone else, or did
he make it up himself?
What processes lie behind Amy and Tom being able to plan a trip to market,
and to organize their behaviour and language into sequences that make
their plans happen?
At what age did Amy and Tom become capable of understanding the
thoughts and feelings of other people? How signicant for their cognitive
development are their interactions with each other and with other people?
10
This rst chapter introduces a theme that runs through much of the book; that
as development proceeds, mental representations become progressively more
abstract and freed from the specic contexts and concrete experiences from
which they spring. In this chapter the mental representations in question are
infants representations of categories. Also, the chapter shows that even the very
early categories formed by infants can be exible and adaptable. The picture that
begins to emerge is that of development as an active, constructive process, and
that this can be seen from one of the earliest cognitive, meaning-making activities
of young infants namely, category formation.
11
12
13
14
3.2 Constructivism
The work of Jean Piaget (18961980) made a key contribution to the eld of
developmental psychology, especially in the area of cognitive development. It set
the scene for a much more scientic approach to understanding this area than
had ever been attempted before. Piagets theory was based rmly on the
observation of childrens behaviour. He found that there are many problemsolving tasks in which children behave quite differently from adults. Furthermore,
he believed that childrens mental abilities pass through a regular series of stages
as development progresses.
Piagets model of development (Piaget, 1955, 1959 and 1973) involves the child
becoming increasingly freed up from the constraints of their own perspective and
the concrete objects around them, as their mental operations become more
abstract. In Piagets theory, children are at rst egocentric, dominated by, or
centred on, their own perceptions, because they are still very much tied to the
concrete world and their actions within it. Similarly, Piaget held that the young
child is unable to comprehend points of view different from his or her own. A
classic example of this particular difculty is provided by childrens responses to
the three mountains task.
As used by Piaget, this task involves sitting a child beside a three-dimensional
model of three mountains (see Figure 1). A doll is placed at the edge of the
model, with a different view to that of the child, and the childs task is to show the
experimenter what the dolls view is. Piaget used various ways of doing this so
that the child would not have to describe it verbally. For example, the child has to
arrange three cardboard mountains (like the model ones), or choose one of a set
of drawings of the mountains model, drawn from different viewpoints, to indicate
what view of the mountains the doll has.
BOX 1
A
(a)
(b)
Figure 1 (a) Plan of the three mountains task; (b) view of the three mountains from
position A (Source: Cox, 1980).
The child was asked to represent the view of the mountains from his own position at A.
Then a wooden doll (about two to three cm high) was put at position C and the child
15
16
was asked to represent the view that could be seen from there. This procedure was
repeated for positions B and D. The child was then asked to move to position B (or C
or D) and to represent the view from there; in addition he was asked to represent the
view from A or other positions he had already occupied.
Piaget and Inhelder elicited three different modes of response. First, the child was asked
to reproduce the dolls view by arranging three pieces of cardboard shaped like the
mountains. Second, the child was asked to select the dolls view from a set of ten
pictures (each measuring 20 by 28 cm). In the third task, the child was asked to choose
one picture and then decide which position the doll must occupy to have that particular
view of the mountains.
Source: adapted from Roth, 1990, p. 99.
n
Presented in this way, the problem is usually difcult for a child of up to about 7
years old to solve correctly. A common response is for children to indicate that
the view that they themselves have is also the one that the doll has. Piaget saw
this as reecting the inability of younger children to decentre from their own
position. His explanation was that children in this pre-operational stage have not
developed the ability to recognize that there can be many different viewpoints on
a scene like the three mountains, their own being just one of these. Nor are they
able to hold in mind the idea that the viewpoint changes as the viewing position
changes. This process reaches its end-point in the nal stage of Piagets
developmental model, when mental operations become formal when wholly
abstract thinking can be achieved and the child becomes able to reason
hypothetically and systematically. Although this book does not cover the details
of Piagets work, the developmental journey that he describes, from concrete to
progressively more abstract ways of thinking, is an important theme in the
account of cognitive and language development that we give.
A classic Piagetian task, the pendulum test, which is widely used in physics
teaching in many schools, is a good illustration of these points. The childs task is
to nd out what factor(s) affect the rate at which a pendulum swings to and fro.
There are several factors that could affect the rate, for example the weight of the
pendulum, the distance through which it swings, the length of the string, the force
of the initial push, the thickness of the string and so on. A child whose reasoning
capacities are still developing may vary these factors without a systematic
approach, perhaps trying various ideas more than once as a result, whereas a
child who can reason more formally can work out that varying each factor
independently to isolate the important one(s) is the most effective approach.
Piagets theory describes a child who is progressively elaborating a more
abstract and general capacity to tackle problems in the world, but in an essentially
solitary, independent way. This view of development was used in the second half
of the twentieth century to support the pedagogic principles of discovery
learning, which hold that it is a rich learning environment, rather than direct
teaching, that is essential for cognitive development.
17
18
framework. So, for example, when Piagets three mountains task was changed so
that the question asked was whether a policeman could see a robber who was
trying to hide from them (see Figure 2), children aged around 4 years who failed
the classic Piagetian task showed that, in this new situation, they could decenter
and see things from the perspective of another person. A policeman looking for a
robber made greater sense to children of this age than the more abstract notion of
another characters view of a set of mountains. Several other experiments in the
same vein carried out around this time conrmed that childrens cognition shows
itself to be more advanced when the tasks that they tackle are situated in a
meaningful context.
BOX 2
B
D
Figure 2 Plan of the experimental layout in the hiding game (Source: adapted from
Donaldson, 1978).
Source: adapted from Roth, 1990, p. 115.
References
Chomsky, N. (1965) Aspects of the Theory of Syntax, Cambridge, MIT Press.
Interedition.
Piaget, J. (1959) The Language and Thought of the Child, translated by M. Gabain
and R. Gabain, London, Routledge and Kegan Paul (rst published 1923).
19
20
Piaget, J. (1973) The Childs Conception of the World, trans. by J. Tomlinson and A.
Tomlinson, London, Paladin.
Roth, I. (1990) Introduction to Psychology, London, Erlbaum Associates/The
Open University.
Vygotsky, L.S. (1962, 2nd edn) Thought and Language, Cambridge, MA, MIT
Press (rst published 1934).
Wood, D. (1988) How Children Think and Learn, Oxford, Blackwell.
Chapter 1
concepts
Contents
Learning outcomes
23
Introduction
1.1 Cognitive structure
23
24
27
28
28
30
33
35
37
category representations?
3.1 Use of cues
39
39
41
43
43
44
46
46
48
49
50
50
51
52
54
54
Conclusion
57
References
57
56
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 dene categories and concepts and discuss the differences between
them;
2 explain how category formation in infants can be studied experimentally;
3 describe and discuss the evidence that infants can form categories before
they can speak;
4 understand that infants may use certain cues to aid categorization;
5 explain what is meant by single-process and dual-process models of
categorization;
6 evaluate these models in relation to evidence from infants categorization of
human and non-human animals, and developmental trends;
7 consider possible links between categorizing and naming.
Introduction
In this rst chapter we look at how very young infants begin to make sense of the
world in which they nd themselves. We start by considering how important it is
in everyday life to be able to see the links between similar experiences and to
group them together. It is through this grouping that our mental world becomes
structured and that we form concepts.
When you, as an adult, see an unfamiliar thing, it is a basic reaction to try to t it
into a category, to recognize what sort of thing it is. And much of the time this is
not really a conscious process; it is only if we come across something that is really
unfamiliar, such as a strange organism in an aquarium, that we are aware of our
categorizing. We might for example see an odd thing in a tank that might be a
seaweed, might be a worm, or might be some other sort of living being. A lot of
categorizing is tied up with language; if there is someone with us at the aquarium,
we are very likely to say What is that? to our companion, when we see the
strange creature, to try to t it into a category that has a name.
These observations point to the idea that much of our making sense of the
world is to do with the very rich and extensive system of categories that we all
have. These have been acquired from all sorts of different sources, from books
and other media, from schooling and from talking with other people, as well as
our own direct experience.
But what about very young babies, who cannot ask what things are, who do
not have words for things? What sense do they make of the sights and sounds
that confront them in the rst weeks and months after their births? Is each
successive experience totally new, or do they begin to register the similarities
among repeated experiences of the same kinds of things?
23
24
t
Activity 1
Allow about
5 minutes
This activity will help you to think about the ways in which people form and use categories.
Look at the pictures in Figure 1 and see whether there are any fruits that you recognize.
Think about what types of information you used from the pictures and from your own past
experience to make this judgement. Next, decide which picture is the odd one out and again
think about what information you used to make that judgement.
Comment
Before we turn to the questions about the information you used to make these judgements,
let us rst consider what was going on in this activity. These are all pictures of tropical fruit.
They are all members of a category; in fact a subcategory, tropical fruit, nested in the broader
category fruit.
In the rst sentence of the activity, the word recognize was used. Recognize is made up of
two parts, cognize (a word rarely used by itself), which refers to forming knowledge and re,
which suggests a repetition. If you have seen all these fruits before, if you have re-cognized
them, you have made use of your previous experience and knowledge and probably found it
quite easy to separate one from another. But to achieve that category distinction you made
use of some features of the fruits to do so: What were they? Probably they were a mixture of
visible attributes like the shape of the fruit, its inside structure and the texture of the surface,
along with characteristics like taste and smell, which you might know about.
But if you are unfamiliar with some or all of these fruits, you might still have grouped two
together and separated one out using visual features only. You might have focused on the fact
that the inside of the durian fruit (Figure 1, right ) is different from the other two fruits that
have similar insides. Thus you were categorizing the longan (Figure 1, middle) and the
rambutan (Figure 1, left) as members of one category, and the durian as a member of another
category, (although they are all members of the category, tropical fruit).
25
26
Activity 1 should have helped to give you an awareness of the mental processes
behind re-cognition. To break this process down further, you could say that to
recognize something as a thing that has been experienced before, as a member
of a category or class, involves two basic cognitive processes. First, the storing of
some sort of memory of an experience and, second, the comparing of a later
experience with the memory trace and registering that there is a similarity. Human
cognition thus involves the building up of an internal, mental structure that has a
relation to corresponding structure in the outside world. These mental structures
lie at the core of our psychological functioning, since life without any of the
regularity and predictability that they provide would be chaotic and highly
inefcient. Life would hardly be possible if each new encounter with the world
involved working out anew, from scratch, appropriate ways of responding and
behaving. Instead, it is a central, almost a dening feature of mental and
behavioural function that objects and events in the world are encountered in
terms of their similarities and differences; that the elements of our world are seen
as falling into many classes and subclasses. The ability to recognize new things as
being members of classes that we have experienced before gives humans a real
advantage in responding to situations rapidly, appropriately and effectively.
Across the animal kingdom there are numerous ways that categorical forms of
mental structure can be genetically programmed, as, for example, in many
animals reactions to potential predator-like entities. This is one of the most
effective ways in which evolution provides animals with survival mechanisms. It
is obviously a great advantage to be able to rapidly identify something as
belonging to the class potential predator and to behave accordingly. For some
organisms, usually those that inhabit highly predictable niches, this sort of
cognition is all that is necessary. For example, a shrimp shoots backwards if
anything large looms nearby. However, other organisms need to be able to adapt
their behaviour to more varied environments, and for them it is also important to
be able to generate new classes and forms of corresponding behaviours. For
humans, this is the predominant way of classifying the world, through learning
about things and building new mental structures that aid recognition. Thus, a
central question for developmental psychologists is How is mental structure
formed and elaborated?
Clearly, once a child has acquired language, this provides them with a powerful
tool for learning about new classes. Indeed, language is in itself a vast repository
of classes. Nouns represent a multitude of objects and events, and verbs represent
a vast number of actions. The natural reaction to a childs question What is that?
is simply to give the class name of whatever it is that the child is asking about.
Chapter 2 focuses on the rst stages of language acquisition and considers how a
child comes to understand the referential function of language and that words
may represent invitations to form categories (Waxman and Markow, 1995). But
before infants have access to the use of language in this way, do they lack the
structures that enable them to represent objects? Clearly not, because very young
infants can quickly come to group objects based on their similarity, which is at
least an initial form of classication.
Finding out more about infants abilities to classify entities in the world, before
they start to use language as an aid, offers an opportunity to uncover the basic
cognitive processes that language builds on and to gain a better understanding of
human cognition and its development. New experimental methods, which get
around the problem that we cannot ask infants directly, have begun to uncover
some basic features of this foundational period of development.
27
28
involved in each? And are there changes in how infants categories develop
between the time that they are born and the emergence of language? Before we
can begin to answer these questions, however, it is necessary to investigate how
and to what extent infants are able to identify, form and store categories. This is
the subject of Sections 2 and 3.
most likely that the infants attention will be given more to the novel one in preference
to the now familiar one. This is called a novelty preference.
The familiarization/novelty-preference procedure has two stages, as you can see in
Figure 2. In Stage 1, Familiarization, the infants are shown a number of different
exemplars (members) of a single category one after another. In Figure 2, the category
chosen was cats and so the infants are shown four different cat pictures. In Stage 2,
Novelty-preference, the infants are given a preference test. In the test, a novel
member of the familiar category (in Figure 2, a cat) is shown alongside a novel member
of a novel category (in Figure 2, a dog). The results from the preference test are then
observed and interpreted. If infants look longer at the novel member of the novel
category, this is a good indicator that they have formed a representation in their
memories of the familiar category a category representation. So what these
experiments do is play off the novelty value of a previously unseen exemplar from
within a familiar category (of which infants have just seen several exemplars) against the
novelty value of a previously unseen exemplar from a novel category. If infants give more
attention to the latter, then we can assume that they are seeing the novel exemplar
from the familiar category as just another one of those, whereas the response to the
novel category exemplar is more like Aha! Heres a new sort of thing!
Stage 1.
Familiarization set of stimuli.
First, these are shown
to the infant, one after another.
Category
exemplar 1
Category
exemplar 2
Category
exemplar 3
Category
exemplar 4
Stage 2.
Test pair of stimuli.
Next, these are shown to the infant,
both at the same time, to see which
attracts more attention.If the infant
looks longer at the dog (the novel
exemplar of the novel category)
then we could infer that the infant
is applying a category something
like cat, based on the
familiarization set and possibly also
on previous knowledge.
Novel exemplar
of novel category
Novel exemplar
of familiar category
These experiments are usually carried out in a laboratory so that the conditions can be
well controlled. Stimuli are often shown to infants on computer screens and are
computer-controlled, so that the presentations of the images can be accurately timed.
Infants are sat at a xed distance from the stimulus display, close enough so that their
limited ability to focus does not affect their perception.
29
30
t
Activity 2
Allow about
5 minutes
This activity will raise your awareness of possible aws or limitations in the familiarization/noveltypreference method.
When analysing scientic studies, it is always important to look for alternative, perhaps
simpler, explanations of observations. With this in mind, think about the following question.
If infants generally look for longer at the dog in the experiment illustrated in Figure 2, are
there other possible explanations for this result, other than the one offered in Box 1? Try to
come up with two others, and make a note of your answers.
Comment
One possibility is that infants might generally prefer to look at the picture of the dog, perhaps
because it is more colourful or has a more interesting shape. A way of avoiding this sort of
problem is to take another group of infants who have not seen the familiarization stimuli, and
test them on a number of different dogs paired with different cats. If there is no prior
preference, we can be reasonably condent that any preferences found are due to the effect
of the set of images shown during the familiarization phase.
Another possibility is that young infants might not be as good as adults at distinguishing
between different cats; one might look much the same as another to the infant, but that does
not mean that they are seeing them both as members of a single category, i.e. cats. A way of
checking this out would be to show another group of infants the picture of one cat followed
by just a picture of another cat. If the infants attention declines as they looked at the rst
picture and then increases for the next, this would be good evidence that they can
discriminate between the two within-category exemplars. These are the kinds of controls used
in well-designed experiments on infant categorization that can be used to guard against
misleading results.
It is worth noting that the results from the sorts of studies that we will be considering in this
chapter come from analysing responses made by a number of infants. There is, typically, a lot
of variation in how infants behave. Some fail to engage properly with the task and hence do
not provide data. The ndings reported are based on signicant trends observed across whole
groups of infant participants.
novel category, e.g. a triangle. This method works on the following rationale: if
infants form a category representation that recognizes similarities among the
distorted versions of the shape, then we can expect the representation to be close
to the undistorted prototype shape (e.g. a square).
Comment
If infants had established the category square from the set of distorted exemplars
during familiarization, which of the two test stimuli (a prototype square or a
prototype triangle) do you think they would find more interesting? (Clue: remember
that infants generally prefer looking at novel, unfamiliar things.)
Research summary 1 gives details of one of these studies, so that you can see how
the data from such research are used to nd out more about infant categorization.
RESEARCH SUMMARY 1
Distortion
Good
Intermediate
Poor
Figure 3 Good, intermediate and poor prototypes and their distortions (Younger and
Gotlieb, 1988).
Next, the infants were shown a test pair that included a prototype (i.e. an undistorted
shape) of the distorted forms that they had just seen, alongside a prototype of another
form. This was carried out in order to test whether the infants were treating the
31
32
prototype of the distorted forms as more familiar than a different form. (Bear in mind
that they had not actually seen this prototype before; they had only seen a series of
distorted versions of it.) The percentage of time that infants spent looking at the novel
prototype was then observed and analysed as shown in Table 1.
Table 1 Preferential looking times for novel prototype at different distortion
levels of familiarization stimuli
Between-category comparison
percentage of time spent looking at novel prototype
Form
3-month-old
infants
5-month-old
infants
7-month-old
infants
Good
M
SD
71.41%**
9.12
61.83%**
59.13%*
15.54
14.63
Intermediate
M
47.05%
64.92%**
62.38%**
SD
13.84
23.23
20.14
53.87%
53.57%
59.66%*
SD
20.78
19.95
12.60
Poor
The rst thing to note about these results is that in all conditions, except for the
intermediate form for 3 month olds, infants were tending to look for longer at the
novel prototype, which belonged to a class from which they had not previously seen
distorted exemplars. This is indicated by the fact that all the percentage looking times
(except one) are greater than 50 per cent. This shows that the novel prototype was
attracting more attention than the prototype of the exemplars that they had seen
before. This means that they were behaving as if the prototype of the distorted forms
that they had seen was more familiar to them (even though they had not actually seen
this undistorted prototype before). However, this preference was only reliably
greater than chance variations (shown by the asterisks) for the good form for 3 month
olds, for the good and intermediate forms for the 5 month olds and for all forms for
the 7 month olds. Clearly, the older the infants, the better they were at extracting the
prototype as a category from the distorted exemplars.
In the experiment described in Research summary 1 the novel prototype from the
novel category was preferred by most infants, indicating that they had formed a
category representation as a result of looking at the familiarization set of distorted
exemplars. Even though they had never actually seen the prototype of the familiar
category, they were, nevertheless, treating it as familiar. Their representation of
the category was sufciently structured so as to include novel exemplars from the
same, familiar category and exclude novel exemplars from novel categories.
33
34
Figure 4 Cat and dog stimuli (Quinn et al., 1993; Eimas and Quinn, 1994).
In addition, infants of this age range who were familiarized with horses treated
novel horses as familiar, but showed greater interest in cats, giraffes, and zebras
(Eimas and Quinn, 1994). These ndings indicate that young infants can form
separate representations for cats and horses each of which excludes instances of
the other category, as well as excluding exemplars of other related animal species.
Furniture domain
Animal domain
Global/superordinate
Furniture
Animal
Basic/intermediate
Chair
Cat
Specic/subordinate
Garden chair
Siamese cat
35
36
Although, as we have seen, infants can group objects into categories, perhaps
they do not at rst form these sorts of nested sets of categories? It is possible that
at rst young infants simply recognize different classes of things. For example,
they might group cats into one category and dogs into another, but not recognize
that at a higher level, both of these categories fall within the category of domestic
pets or the broader category of animals.
Behl-Chadha (1996) wanted to nd out whether infants are able to form the
sorts of categories within categories outlined in Table 2. To do this, she extended
the ndings of studies looking at early categorization differentiation among
animal species and designed an experiment to discover whether infants could
form separate categories for particular types of furniture. Furniture was chosen as
a stimulus class because infants have a lot of exposure to furniture from birth
onwards.
In the experiment, infants of 3 and 4 months of age were rst familiarized with
twelve realistic photographs of chairs (including examples of the subordinate
categories: armchairs, desk chairs, kitchen chairs, rocking chairs and stuffed
chairs). Then, they were shown novel chairs, along with non-chair furniture
(either couches, beds or tables). The infants gave more attention to these nonchair pictures than to the novel chairs. This indicates that they had formed a
category equivalent to chair that did not include other types of furniture. When
they were familiarized with a set of photographs of couches, they treated novel
couches as familiar, but showed novel category preferences for chairs, beds and
tables by looking at these for longer, showing that in this case they had formed
category representations for couch. This is good evidence that 3 and 4 month
olds can form individuated representations for chairs and couches as separate
categories, separate also from the categories of beds and tables. In other words,
nested within the infants category furniture were further, basic or intermediate
categories for chairs, couches, beds and tables.
Behl-Chadha (1996) also found that infants could form more global (inclusive)
category representations for broader classes of stimuli. In one experiment, 3 and 4
month olds were familiarized with photographs of mammals (deer, domestic cats,
elephants, horses, rabbits, squirrels, tigers and zebras). Then, photographs of an
animal from a novel mammal category were paired with either (1) a novel example
of a familiar mammal category (2) a non-mammalian animal (bird or sh) or (3) an
item of furniture. The infants preferred birds, sh and furniture to instances from
the novel mammal categories, but they did not prefer members of novel mammal
categories to novel members of familiar mammal categories. These ndings
indicate that young infants can form a broad category representation for mammals
that includes examples from novel mammal categories that they had not seen
during familiarization, and excludes some non-mammals (birds and sh) as well as
furniture. However, what is even more interesting is that within the broad category
representation animal, they can use nesting categorization skills. In other words,
within the animal category, they can form basic category representations for
mammals.
Results from the same series of experiments backed up these ndings, showing
that 34 month olds can make the same kinds of nested distinctions the other
way round i.e. for furniture versus animals. The results showed that infants
could form a category representation for furniture that includes beds, chairs,
couches, cabinets, dressers and tables, but excludes mammals. This evidence
suggests that young infants can form global category representations for at least
some natural (mammal) and artefact (furniture) categories and this is strong
evidence that infants can, indeed, form nested category representations. These
experimental outcomes are highly signicant because they indicate quite clearly
that young infants have the cognitive abilities to group stimuli into categories that
are similar to many of the cognitive groupings that older children and adults use.
This is also strong evidence that infants are forming categories on the basis of
their experience, because evolution could hardly have provided us with innate
templates for types of furniture!
37
38
(a)
(b)
The results were that both groups of infants showed a preference for the novel spatial
category, consistent with the idea that they had formed category representations for
the above and below relations between the dot and the horizontal bar. More recent
research using the same method indicates that by 6 or 7 months of age, infants can form
category representations for between and outside (Quinn et al., 2003).
n
Given that infants can categorize objects in quite sophisticated ways, the nding
that they can also categorize the ways in which objects are spatially located shows
that they have the basic cognitive building blocks (primitives, as they are often
called) to form quite complex representations of the physical world. These
representations also provide infants with a foundation for learning to
communicate about different types of information: to learn the words dog, table
and under, and then say things like dog under table if a dog is below a table. So,
clearly, before they can speak, infants have the necessary mental abilities to
construct category representations that will be useful in later development to
support their use of language (see Chapter 2).
.
.
What is a cue?
In its everyday sense, a cue is some sort of signal that launches a sequence of
behaviour or some other response. For example, a green trafc light is a cue to
the behaviour of moving forward across a road junction. The word cue is used in
a similar way in psychology, to refer to something that indicates or triggers
something else, such as behaviour or recognition. A cue could be in any sensory
modality; it could be a visual signal, a sound, a touch or a taste. If you were
meeting someone at a railway station, for example, spotting their hat might serve
as a visual cue to trigger your recognition of them. Or, if you were waiting for a
particular programme on the radio, the signature tune would give you an auditory
cue that the programme is about to start. In categorization, then, we often use
specic cues as indicators of which category something belongs to. Specic cues
might include the values of attributes such as colour or shape.
One strategy that has been used to identify the cue (or cues) that infants use to
form a particular category representation is to show that they form the category
representation when a particular cue is present, but not when the cue is absent.
Such a strategy has been used to nd out, for example, how infants form distinct
category representations for dogs versus cats. Because the two species have
considerable perceptual overlap (both have facial features, a torso, four legs, fur
and tails) it is not immediately obvious what information infants might be using to
form and differentiate the categories.
t
Activity 3
Allow about
5 minutes
This activity will help you to identify the possible cues that might be used to differentiate the category
What visual features mark the difference between dogs and cats for you? In other words,
what would you look for to decide whether an animal that you saw for the rst time is a dog
or a cat? You might like to try visualizing (or drawing) a prototypical dog and cat, and see
what are the main features of your visual images and how the dog and cat differ.
39
40
Comment
For many people, this is not a very easy activity. Although they assert that they have little
difculty in distinguishing a cat from a dog, they have a lot more difculty in saying how they do
so. This activity serves to remind us that our categorization processes are not always fully
accessible to consciousness.
It is possible that infants are able to make categorical distinctions on the basis of
differences in single attributes, in the pattern of correlations among attributes or
in the overall combination of attributes. For example, in the case of the visible
features of birds, beaks and feathers are reliably correlated; they almost always
co-occur. Hence, this correlation between attributes might feasibly form a basis
for categorical distinctions. Or, in the case of the snake category, a long,
cylindrical body is a reasonable basis, although using this cue alone would
include worms, slugs and legless lizards in the category. One way of nding out
how infants select from the range of visible information is to vary systematically
the attributes shown by a category exemplar.
Whether infants use a subset of cues in order to categorize was the subject of a
series of studies by Quinn and Eimas (1996b), Spencer et al. (1997) and Quinn
et al. (2001a). These studies found that infants categorized animals when the
exemplars presented during familiarization and test trials showed only the head
region of the animal, but they failed to show the categorical distinction when only
the body region was displayed (Quinn and Eimas, 1996b; Quinn et al., 2001a).
Infants also formed category representations based on the head (and not the
body) when they were familiarized with whole cats or dogs and were then
preference-tested with a pair of hybrid stimuli: a novel cat head on a novel dog
body versus a novel dog head on a novel cat body (Spencer et al., 1997).
Examples of hybrid stimuli used in an experiment by Quinn and Eimas (1996b)
are shown in Figure 6.
So it seems that, at least in the case of dogs and cats, infants do not make use of all
the available cues for categorization, but they do tend to focus on a specic
subset. These results suggest that infants make the most use of perceptual cues in
the head region to distinguish cats from dogs and that cues from the rest of these
animals bodies are less important. From this set of experiments at least, we can
say that infants may select certain attributes as a basis for categorization.
n
There is evidence that salient part differences (e.g. legs versus wheels) can play a
role in signalling broader (global) category contrasts in older infants. Rakison and
Butterworth (1998) examined categorization of toy objects by 1422-month-old
infants, using the sequential touching procedure, and found that the infants
would categorically differentiate between animals and vehicles. However, in a
subsequent experiment, when the legs were removed from the animals and the
wheels from the vehicles, the infants no longer showed the category
41
42
differentiation. This supports the idea that older infants use specic, salient
attributes to make broad contrasts between global categories of objects, for
example, using legs versus wheels to distinguish animals from vehicles.
Dynamic point
light displays
A display
produced by
putting several
small light sources
(point lights) on
different parts of a
moving object.
When the object is
then lmed
moving in
darkness, all that
can be seen is a
series of lights
moving around.
This often gives a
strong impression
of the form of the
object.
Infants are good at selecting certain cues, out of all those that are
available, to categorize things.
Infants use visual cues from cats and dogs heads to categorize these
kinds of animals.
Older children can also categorize by using parts of objects, like wheels
or legs, to distinguish between the vehicle and animal categories.
It is not just static visual cues that infants use to categorize things: sounds
and movement may also be used.
As you have seen from the research ndings given in this chapter so far, the
evidence is now clear that infants do indeed have remarkable abilities to group
things into categories. Given this important nding, psychologists are interested
to know how these initial categories develop into the more mature structures that
we described in Section 1 as concepts. This has been a matter of some debate
among developmental psychologists (Mandler, 2000; Quinn and Johnson, 2000)
and two competing theories have been put forward. One view proposes that a
single process is operating, while the opposing view is that a second process
comes into play, working along with the perceptually-based process.
43
44
their summary of the single-process model, a representation like animal that may
begin by picking out relatively simple features from seeing and other sensory
modalities comes over time to have sufcient knowledge to permit specifying the
kind of thing something is through a single continuous and integrative process of
enrichment (Quinn and Eimas, 2000, p. 57). We can call this a perceptual learning
model. Computer simulations (connectionist models) of learning networks have
supported the validity of the single-process framework to some extent. It has also
been found that the same computer model could form category representations
based only on perceptual features or on arbitrary labels that classify exemplars
(Quinn and Johnson, 1997).
Perceptual
schema
A simple form of
category based
only on
supercial,
perceptual
features.
Image schema
A form of
category that
incorporates
multiple
characteristics of
its exemplars,
including less
immediately
obvious, more
abstract features
such as how the
exemplars
behave.
Procedural
knowledge
Knowing how to
do something, but
not necessarily
being able to
describe it.
Do you think the single-process model accounts for the way in which infants form
categories? Or do you think that at some point in development there might be a
change? Mandler has argued that, early in development, infants start to use a
different approach, based on more abstract attributes a view closer to the idea
that you came across in Section 1 that seeing is not the same as thinking
(Mandler, 1992, 2000). Mandler argues that category representations based on
perceptual features are at rst simply perceptual schemas that dene what a group
of things looks like, but they do not really dene the meaning of something.
Mandlers argument has been described as a dual-process model. She describes
the dual-process model as involving a second process that works separately from
the formation of perceptual schemas, that carries out more analysis of the
perceptual features and also links them with dynamic information such as motion,
sound and function. She proposes that this second process leads to the formation
of image schemas (Mandler, 1992, 2000). Image schemas are the forerunners of
mature concepts and can, for example, separate animals from non-animals by
what have been called conceptual primitives. These are more abstract attributes,
such as whether the members of the concept are self-starters (i.e. they move by
themselves without a visible external force acting on them). These attributes are
more abstract because they cannot be actually seen (or heard or felt) but instead
have to be inferred in some way. This is a very different sort of process from
straightforward perception. This dual-process model proposes that both
perceptual schemas and image schemas (which are different systems of
representation for perception and the formation of concepts respectively) operate
in parallel. This dual process then leads to the formation of true concepts that are
much more suffused with meaning than perceptual schemas.
The concept of representational re-description of Karmiloff-Smith (1986)
outlined in Box 3, describes a similar, related developmental process whereby
childrens knowledge moves from being implicit and procedural to being explicit
and potentially thought about and talked about. Karmiloff-Smiths theory holds
that a new ability is at rst something that children can just do, without being able
to reect on it, so it is at this point represented at a procedural level. Then, once it
has been practised enough and applied to a range of different problems, it
becomes redescribed and becomes an object of thought; a different form of
t
Activity 4
Allow about
10 minutes
This activity will help you to understand the differences and possible similarities between these two
Consider these two models of category formation. Do you think that they really are
alternative, competing explanations, or might they both have some validity? One way of
approaching this activity could be to list the features of each model in a table, to help you to
see points of similarity and divergence between the two models.
45
46
Comment
You might feel that perhaps the dual-process model is not so very different from the singleprocess model, in that a single-process model could incorporate functions and names, which
are clearly not just straightforward perceptual features. As well, the idea of function might
itself be seen as a form of redescription of perceptual schemas, in other words, based on
things that are visible. Similarly, we might question whether an attribute like has legs, which at
rst sight seems an obvious perceptual feature might also be tied up with the use of legs for
locomotion, a functional aspect: in other words, is the notion of legs also tied up with these
limbs being used for walking?
Experiments have not yet been able to determine conclusively whether a singleor dual-process model best describes how infants categories develop; the
question is still an open one. Although this specic issue has yet to be resolved,
researchers have claried some other aspects of infants categorization abilities.
One of these aspects is the way in which infants categories are organized into
larger structures of knowledge about the world, a topic which we turn to in the
next section.
.
.
Furniture domain
Animal domain
Global/superordinate
Furniture
Animal
Basic/intermediate
Chair
Cat
Specic/subordinate
Garden chair
Siamese cat
Object examination
Object examination is a method that has been used for studying category
representation in older infants from about 6 months of age upwards (Oakes et al.,
1991). This procedure, like the familiarization/novelty-preference method, depends on
infants gradually losing interest in successive new members of a single category and
then regaining interest in a member from a novel category.
For example, if a vehicle versus animal category contrast were being studied, infants
would be given toy vehicles (e.g. a tractor, a car, a lorry, a bus, a quad-bike and a
motorcycle) one after the other, each for a xed number of seconds (typically 2030).
Infants are allowed to look at and manipulate the toys. They are then given a toy animal,
for example, a horse (novel exemplar from a novel category), for the same number of
seconds. A measure of active examination (that combines the amounts of handling and
looking at each toy) is then used to see whether infants examine the novel exemplar
from the novel category signicantly more than the previous one or two vehicle toys. If
we know that the infants can discriminate among the instances from the familiar
category and that they do not have a pre-existing preference for the novel category
exemplar (remember the precautions that are taken to rule out these possibilities) then
one can conclude that categorization is taking place.
47
48
A perceptual learning account (i.e. single-process model) has been put forward to
describe this differentiation-based pattern of ndings on the early development of
object categorization (Quinn and Eimas, 1997, 2000; Quinn and Johnson, 1997,
2000; Quinn, 2002). At rst, infants seem to learn global categories like mammals
and furniture based on the presence or absence of salient features. Then they go
on to differentiate these into basic-level categories like cats and dogs based on
specic values of shared features. This observation is supported by the fact that
simple connectionist networks also learn category representations in a global-tobasic order when provided only with input features (e.g. leg length, body length)
that are measured directly from the stimuli shown to the infants (in other words,
they are basic perceptual attributes). This again lends some support to the singleprocess model of category development.
Figure 7 Example of spatial relations test stimuli (based on Quinn et al., 1996).
t
Activity 5
Allow about
5 minutes
This activity will help you to understand how the familiarization/novelty-preference method can be
used to address a new research question about infants capacities to categorize locations separately
from the objects found in those locations.
If infants are able mentally to separate an objects position (above/below) from the shape of
the object that is in that position, which of the test stimuli shown in Figure 7 will they look at
for longer? Make a note of your answers. (Remember that infants tend to prefer novelty:
think about which of the two test stimuli would seem more novel.)
Comment
In the familiarization phase, infants became familiar with a particular shape (e.g. a dot) and a
particular location (e.g. above). If they can separate out object and position information, then,
because they have been familiarized to above, they should see the right-hand test stimulus as
novel because it shows the spatial relation below and they will look at it for longer.
In contrast to the infants tested in the original above versus below experiment
described in Section 2, the infants tested in this new version of the categorization
task did not show a preference for the novel spatial category test stimulus, instead
they divided their attention equally across both test stimuli. This result suggests
that infants of this age do not form the abstract category representations for
above and below separately from the particular objects showing the relation.
A follow-up experiment (Quinn et al., 1996) found that older infants, aged 67
months, did prefer the novel spatial relation in the test phase. This result shows that
by this age they were now able to separate their more abstract categories of spatial
location from the changes in the identity of the objects. A similar concrete-toabstract developmental pathway has been reported for the spatial relation
between in the age range from 67 months to 910 months (Quinn et al., 2003).
These results, taken together, support the idea that category representations of
spatial relations may be initially tied up with the objects depicting the relations, but
later become independent of the objects.
49
50
.
.
If this is true, then it should be relatively easy to modify the way that young
infants categorize non-human animals. One study found that infants who were
familiarized with a series of cats, then showed a novelty preference for a dog over a
novel cat, suggesting that in this condition they were indeed using a cat category
that excluded dogs (Quinn et al., 1993). However, in a second condition, another
group of infants who were shown a series of dogs did not then show a novelty
preference for a cat over a novel dog. This suggests that the dog category formed in
this second condition was an inclusive category it was broad enough to
encompass cats as well. So in some conditions infants can form narrow categories
(e.g. a category of cats that excludes dogs) and in other conditions they can form
broad categories (e.g. a category of dogs that includes cats). This suggests that
infants do not have rigid categories for cats and dogs that they apply in all
conditions.
Two further experiments support the idea that these are exible categories and
that the dog versus cat asymmetry only arose because of the particular sets of stimuli
that were used (Quinn et al., 1993; French et al., 2001). In one of these experiments
(Quinn et al., 1993), the variability of the dog stimuli was reduced (the stimuli were
chosen to be more similar to each other). In this case, 34-month old infants did not
include cats in the dog category that they found. In another experiment with
infants of the same age, new sets of dog and cat pictures resulted in the cat category
this time including dogs (French et al., 2001). Thus, when familiarized with
exemplars that show a lot of variation in their features and overlap between
categories, infants tend to form broad, inclusive categories. But if they are
familiarized with very similar exemplars within a category and no overlap between
categories, their category representation tends to be much narrower. These
ndings, taken together, indicate that these non-human animal categories were
quite uid and determined more by the experimental condition than any prior
experience.
The studies of the categorization of cat and dog images by young infants makes
the important point that the infants seem to be forming their category
representations for non-human animals over the course of the familiarization
trials (i.e. bottom up), rather than tapping into pre-existing concepts that had
been formed prior to arriving at the laboratory (i.e. top down). If infants in the cat
versus dog categorization studies had simply been tapping into category
representations established before the experiments, then the experimental
changes should have had no effects. The fact that infant responsiveness did vary
across experiments suggests that the categories were being formed during the
familiarization experience and that the boundaries of categories can be relatively
easily affected by modifying the familiarization stimuli. Thus we can see that
infants can not only form complex category representations, but also that these
are open to change in the light of new experiences.
51
52
values of attributes (e.g. clothing, and types of clothing, upright stance and
different gaits, hair colours and different styles, speech, etc.).
A study by Quinn and Eimas (1998) familiarized 3- and 4-month-old infants
with photographs of twelve humans (male and female) in a variety of standing,
walking and running poses. Examples of the human stimuli used in the
experiment are shown in Figure 8. The infants were then preference-tested with a
novel human paired with a cat and a different novel human paired with a horse.
Most surprisingly, no novel category preferences for either the cats or horses were
observed. (There was no spontaneous preference among the infants for humans
over non-human animals, which might have mistakenly led us to believe that
infants categorized humans as different from horses and cats.) It seems unlikely
that infants cannot tell the difference between humans, and cats and horses, so
something rather odd seems to be going on.
53
54
Name
First
Single-category grouping
Second
Third
Two-category grouping
Object
permanence test
A test of whether
infants believe
that an object
continues to exist
after it has been
hidden from view.
Meansend test
A test that
involves carrying
out one (or more)
actions to achieve
some other result.
For example,
pulling a cloth on
which a toy is
placed in order to
get the toy.
Categorization
Vocabulary spurt
Object permanence
0.36
0.48
0.70*
Meansends
0.19
0.017
Categorization
0.78**
55
56
It is important to note that categorization was not strongly linked with the other
cognitive measures, nor were these other measures strongly linked with each
other. This helps to rule out the possibility that what was being observed was just
a general improvement in cognitive ability in the children, but that something
more specic, linking categorization and vocabulary, was going on.
When children are given a set of objects they spontaneously sort them;
this gives an indication as to how they categorize the objects.
The development of childrens spontaneous sorting seems usually to
follow a xed sequence.
Most children show a sudden spurt in the growth of their vocabulary
some time between the ages of 1 and 2 years.
There is evidence that this spurt is linked to a particular stage in the
development of categorizing, suggesting that there is some connection
between language development and categorization.
Conclusion
In this chapter we have shown that new experimental techniques have provided
very productive ways of gaining a window into infants minds and how they
operate. We have seen how infants are able to group things into categories in
remarkably exible and sensitive ways. They are selective in the cues that they
use to categorize different sorts of things and they can, very early in their lives,
build the beginnings of complex, multi-levelled meaning structures. Having the
abilities to form these rich structures from the earliest months of life gives infants a
powerful tool to help them in making sense of the world of people and things
that surround them and provides an essential foundation for almost all aspects of
human experience.
This chapter has also explored the developmental pathway of infants
categorization abilities and looked at how this ts in with language development.
We found that infants initial categories seem very much based on what things
look like and that more abstract, less immediately obvious features are only used
in categorization later on in the period of infancy. It is only towards the end of this
period, as infants begin to use their rst words, that their categorizing becomes
better at handling more abstract ideas separately from the visible objects to which
they relate. And when their language vocabulary suddenly begins to race ahead,
this appears to have a basis in a further development in categorization, which
may be the point at which categories make the transition to becoming true
concepts that are richly imbued with meanings. This is the exciting time when
children are learning to communicate effectively with language and enter more
fully into their social worlds. But that is a topic for the next chapter.
Acknowledgements
The preparation of this chapter was supported by NSF Grant BCS-0096300 and
NIH Grant HD-42451.
References
Arterberry, M. E. and Bornstein, M. H. (2001) Three-month-old infants
categorization of animals and vehicles based on static and dynamic attributes,
Journal of Experimental Child Psychology, vol. 80, pp. 33346.
Arterberry, M. E. and Bornstein, M. H. (2002) Infant perceptual and conceptual
categorization: the roles of static and dynamic stimulus attributes, Cognition, vol.
86, pp. 12.
Behl-Chadha, G. (1996) Basic-level and superordinate-like categorical
representations in early infancy, Cognition, vol. 60, pp. 10541.
Bomba, P. C. and Siqueland, E. R. (1983) The nature and structure of infant form
categories, Journal of Experimental Child Psychology, vol. 35, pp. 294328.
57
58
Deloache, J. S., Kolstad, V. and Anderson, K. (1991) Physical similarity and young
childrens understanding of scale models, Child Development, vol. 62,
pp. 11126.
Eimas, P. D. and Quinn, P. C. (1994) Studies on the formation of perceptually
based basic-level categories in young infants, Child Development, vol. 65,
pp. 90317.
Fantz, R. L. (1963) Pattern vision in newborn infants, Science, vol. 140, pp. 2967.
French, R. M., Mermillod, M., Quinn, P. C. and Mareschal, D. (2001) Reversing
category exclusivities in infant perceptual categorization: simulations and data, in
Stenning, K. and Moore, J. (eds) Proceedings of the 23rd Annual Conference of
the Cognitive Science Society, pp. 30712, Mahwah, NJ, Erlbaum.
Gopnik, A. and Meltzoff, A. N. (1987) The development of categorization in the
second year and its relation to other cognitive and linguistic developments, Child
Development, vol. 58, pp. 152331.
Gopnik, A. and Meltzoff, A. N. (1992) Categorization and naming: basic-level
sorting in eighteen-month-olds and its relation to language, Child Development,
vol. 63, pp. 1091103.
Karmiloff-Smith, A. (1986) From meta-processes to conscious access: evidence
from childrens metalinguistic and repair data, Cognition, vol. 23, pp. 95147.
Madole, K. L. and Oakes, L. M. (1999) Making sense of infant categorization:
stable processes and changing representations, Developmental Review, vol. 19,
pp. 26396.
Mandler, J. M. (1992) How to build a baby: II. Conceptual primitives,
Psychological Review, vol. 99, pp. 587604.
Mandler, J. M. (1997) Development of categorization: perceptual and conceptual
categories, in Bremner, G., Slater, A. and Butterworth, G. (eds) Infant
Development: recent advances, pp. 16390, Hove, Psychology Press.
Mandler, J. M. (2000) Perceptual and conceptual processes, Journal of Cognition
and Development, vol. 1, pp. 336.
Mandler, J. M. and McDonough, L. (1993) Concept formation in infancy,
Cognitive Development, vol. 8, pp. 291318.
Mandler, J. M. and McDonough, L. (1996) Drinking and driving dont mix:
inductive generalization in infancy, Cognition, vol. 59, pp. 30735.
Mandler, J. M., Bauer, P. J. and McDonough, L. (1991) Separating the sheep from
the goats: differentiating global categories, Cognitive Psychology, vol. 23,
pp. 26398.
Millikan, R. G. (1998) A common structure for concepts of individuals, stuffs, and
real kinds: more mama, more milk, and more mouse, Behavioural and Brain
Sciences, vol. 21, pp. 55100.
Oakes, L. M., Madole, K. L. and Cohen, L. B. (1991) Infants object examining:
habituation and categorization, Cognitive Development, vol. 6, pp. 37792.
Poulin-Dubois, D. and Graham, S. (1994) Infant categorization and early objectword meaning, in Vyt, A., Bloch, H. and Bornstein, M. H. (eds) Early Child
Development in the French Tradition: contributions from current research,
pp. 20725, Hillsdale, NJ, Lawrence Erlbaum Associates.
Quinn, P. C. (1987) The categorical representation of visual pattern information
by young infants, Cognition, vol. 27, pp. 14579.
Quinn, P. C. (1994) The categorization of above and below spatial relations by
young infants, Child Development, vol. 65, pp. 5869.
Quinn, P. C. (2002) Early categorization: a new synthesis, in Goswami, U. (ed.)
Blackwell Handbook of Childhood Cognitive Development, pp. 84101, Oxford,
Blackwell.
Quinn, P. C. and Eimas, P. D. (1996a) Perceptual organization and categorization
in young infants, Advances in Infancy Research, vol. 10, pp. 136.
Quinn, P. C. and Eimas, P. D. (1996b) Perceptual cues that permit categorical
differentiation of animal species by infants, Journal of Experimental Child
Psychology, vol. 63, pp. 189211.
Quinn, P. C. and Eimas, P. D. (1997) A re-examination of the perceptual-toconceptual shift in mental representations, Review of General Psychology, vol. 1,
pp. 27187.
Quinn, P. C. and Eimas, P. D. (1998) Evidence for a global categorical
representation of humans by young infants, Journal of Experimental Child
Psychology, vol. 69, pp. 15174.
Quinn, P. C. and Eimas, P. D. (2000) The emergence of category representations
during infancy: are separate perceptual and conceptual processes required?
Journal of Cognition and Development, vol. 1, pp. 5561.
Quinn, P. C. and Johnson, M. H. (1997) The emergence of perceptual category
representations in young infants: a connectionist analysis, Journal of
Experimental Child Psychology, vol. 66, pp. 23663.
Quinn, P. C. and Johnson, M. H. (2000) Global-before-basic object categorization
in connectionist networks and 2-month-old infants, Infancy, vol. 1, pp. 3146.
Quinn, P. C., Eimas, P. D. and Rosenkrantz, S. L. (1993) Evidence for
representations of perceptually similar natural categories by 3-month-old and 4month-old infants, Perception, vol. 22, pp. 46375.
Quinn, P. C., Eimas, P. D. and Tarr, M. J. (2001a) Perceptual categorization of cat
and dog silhouettes by 3- to 4-month-old infants, Journal of Experimental Child
Psychology, vol. 79, pp. 7894.
Quinn, P. C., Adams, A., Kennedy, E. et al. (2003) Development of an abstract
category representation for the spatial relation between in 6- to 10-month-old
infants, Developmental Psychology, vol. 39, pp. 15163.
Quinn, P. C., Cummins, M., Kase, J. et al. (1996) Development of categorical
representations for above and below spatial relations in 3- to 7-month-old
infants, Developmental Psychology, vol. 32, pp. 94250.
59
60
Quinn, P. C., Slater, A. M., Brown, E. et al. (2001b) Developmental change in form
categorization in early infancy, British Journal of Developmental Psychology,
vol. 19, pp. 20718.
Rakison, D. and Butterworth, G. (1998) Infants use of object parts in early
categorization, Developmental Psychology, vol. 34, pp. 4962.
Rakison, D. and Poulin-Dubois, D. (2001) Developmental origin of the animate
inanimate distinction, Psychological Bulletin, vol. 127, pp. 20928.
Spencer, J., Quinn, P. C., Johnson, M. H. et al. (1997) Heads you win, tails you
lose: evidence for young infants categorizing mammals by head and facial
attributes, Early Development and Parenting, vol. 6, pp. 11326.
Waxman, S. R. and Markow, D. B. (1995) Words as invitations to form categories:
evidence from 12- to 13-month-old infants, Cognitive Psychology, vol. 29,
pp. 257302.
Younger, B. A. and Gotlieb, S. (1988) Development of categorization skills:
changes in the nature or structure of infant form categories?, Developmental
Psychology, vol. 24, pp. 61119.
Younger, B. A. and Fearing, D. D. (1999) Parsing items into separate categories:
developmental change in infant categorization, Child Development, vol. 70,
pp. 291303.
Chapter 2
First words
Margaret Harris
Contents
Learning outcomes
63
Introduction
1.1 Understanding and producing words
63
63
Recognizing speech
2.1 Identifying speech sounds
2.2 Cues to word boundaries
65
65
68
73
73
78
81
81
82
83
vocabularies
84
85
85
87
91
91
93
Conclusion
94
References
Readings
Reading A: Is Dutch native English?
95
98
98
106
2 FIRST WORDS
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 describe the task that confronts the young child in learning rst words and
explain how this compares with the adult experience of acquiring
vocabulary in an unfamiliar language;
2 describe how infants can use cues to identify word boundaries in a stream
of continuous speech;
3 outline Bruners arguments about the importance of the social context for
early word learning;
4 discuss ways in which reliable data can be collected on childrens early
comprehension and production of words using a variety of methods;
5 describe the typical time course of the development of childrens
comprehension and production vocabulary while being aware of
individual differences in development;
6 describe the stages in babbling;
7 understand the range of meanings that is found in early words.
Introduction
In the previous chapter you learned that very young babies are able to construct
and represent categories that group together similar objects. In this chapter you
will discover how young children learn about the names for objects in these
categories as well as the words used to describe actions and personal names. I
will be focusing on the rst words that children say but I will also be considering
the fundamental processes and developments that lay the foundations for the
emergence of rst word use.
Parents usually regard the moment when children begin to say words as very
signicant but, by the time this happens, children already know a great deal both
about the way that words sound and about the way in which language is used in
familiar situations. As you have seen, they also know about object categories. All
of this knowledge is crucial as children search for the meaning of words.
63
64
t
Activity 1
Allow about
10 minutes
This activity asks you to reect on the strategies you use when learning new words and to compare
your experience of word learning to that of a young child at the beginning of language development.
Imagine that you are listening to someone talking and you hear a word that you do not
recognize. Note down how you could discover what the new word meant if you could not
look it up in a dictionary or ask someone else about the words meaning. What difculties
would you face?
Next, note down the difference between your experience of hearing an unfamiliar language
and what you imagine the experience of a young child to be. Remember, you already know a
lot about at least one language your own native tongue. Think about the advantage this
gives you when you hear a new language for the rst time. Then try to imagine what it would
be like if you did not know any words in any language.
Comment
At the very beginning of language development, young children do not know about the
meaning of any words so they are rather in the position that you would be in if you heard a
new language for the rst time. However, you have an advantage over the young child you
already know a lot about language from speaking your own native tongue. Young children do
not know any words in any language.
The chart below summarizes the different stages that young children must go
through in learning about rst words. You will see that there are more
requirements for the production of words than for comprehension. At the end of
the chapter you will look at this chart again so that you can see the various
aspects of early word learning that you have covered.
Nativism
A theoretical
position in which
structures of
thought are held
to be innate,
development
being relatively
unaffected by
experience.
WORD PRODUCTION
Speech stream
The ow of sound
produced when
people speak,
made up of
different
frequencies. If a
graphical plot of
this stream is
examined, it is
often unclear
where words
begin and end.
So far I have been talking about the skills that are involved when children learn
their rst words, and learning tends to imply a non-nativist view. You might ask
whether there are nativist theories of word learning. Several theorists, most
notably Noam Chomsky, have taken a nativist view of language development,
2 FIRST WORDS
Phonology
The set of sounds
that make up the
basic building
blocks of speech,
which vary from
one language to
another and even
within languages
to some extent.
arguing that the human capacity to learn language arises from an innate
mechanism.
Interestingly, nativist theories of language development have little, if anything,
to say about early word learning (although they have made some claims about
phonology that do not concern us here). Traditional nativist theories have been
mainly concerned with childrens mastery of morphology and syntax, that is, the
grammatical rules for combining words into phrases. You will be reading about
these aspects of language development in Chapters 3 and 4.
.
.
Most infants comprehend many more words than they can produce. It is
therefore essential to consider early word comprehension before early
word production.
Babies need to draw on more skills for word production than for word
comprehension.
Babies do not know the meaning of any words, so their experience of
word learning differs greatly from the adult experience of learning
another language.
Recognizing speech
65
66
RESEARCH SUMMARY
2 FIRST WORDS
Recognizing voices
These dramatic ndings about babies ability to recognize speech sounds before
birth can help to explain why they soon develop preferences for familiar speech
sounds after they are born. Babies prefer the human voice to other sounds and, a
few days after birth, according to an earlier study by DeCasper and Fifer (1980)
using the pressure-sensing dummy technique, they will increase their rate of
sucking in order to hear a recorded human voice but not to hear recorded music
or a rhythmical non-speech sound. Over the rst weeks of life, this general
preference for voices over other sounds becomes more specic, and by 4 weeks
of age, infants prefer their own mothers voice to other female voices (Mehler and
Dupoux, 1994).
Prosody
The music of
speech; aspects
such as
intonation, stress,
rhythm and rising
and falling
patterns.
As well as preferring familiar voices, babies also rapidly develop a preference for
familiar languages. The language that is familiar to a baby will typically be the one
they heard before they were born and it is also the one that they hear other
people speaking after they are born. Mehler et al. (1994) gave 4-day-old French
babies (who had heard French being spoken before and since birth) the
opportunity to hear both French and Russian a language that they had not heard
before. As you can predict from reading about the DeCasper and Spence
experiments, the babies preferred French.
How is it that such young babies can tell the difference between two languages
and between two voices? A study by Christophe and Morton (1998) goes some
way towards answering this question. Following on from the study by Mehler
et al. (1994) that compared babies ability to differentiate between French and
Russian, Christophe and Morton presented 2-month-old English babies with two
different language comparisons. The rst was between English and Japanese.
These two languages were compared because they have a very different
rhythmical (or prosodic) pattern. The other comparison was between English and
Dutch. These languages are more similar to each other in prosody.
Christophe and Morton predicted that, if babies use prosody to distinguish one
language from another, then they would be able to tell the difference between
English and Japanese but not the difference between English and Dutch.
t
Reading
At this point you should turn to the end of this chapter and read Reading A Is Dutch native
English? Linguistic analysis by 2-month-olds which is the paper written by Christophe and
Morton (1998). While reading, think about the method that was used to assess the babies
perception of the two languages. Make notes on the authors conclusions about the cues that
the babies were using to tell the difference between the two languages.
67
68
Cue
A specic feature
or marker that is
used to identify an
object or event, or
distinguish one
object or event
from another.
Figure 1 Sound spectrogram of the phrase How are you?. Reproduced from an original
recording provided by Mark Tatham (University of Essex).
2 FIRST WORDS
Syllable stress
There are several possible prosodic cues to word boundaries. One is syllable
stress. Many languages, such as Italian or Greek, have a very regular pattern of
stress within words. In English, however, stress is variable as you will see in the
next activity.
t
Activity 2
Allow about
10 minutes
This activity will help you to identify stress patterns in spoken English.
Say the following words aloud to yourself and underline the syllable that is stressed. For
example, telephone is stressed on the rst syllable since we normally say telephone rather
than telephone. If you are not sure where the stress falls, it may help to think of stress in
terms of emphasis. You could also try saying the word in different ways, varying the stress,
and see which one sounds natural.
telephone
caterpillar
pencil
belong
accident
cricket
under
apple
ambulance
basket
across
cushion
tortoise
zebra
tickle
arctic
football
panda
guitar
ointment
Comment
In typical, conversational English, around 90 per cent of content words (i.e. nouns, verbs,
adjectives, adverbs and personal names) have stress on the rst syllable. You should have
found that only belong, across and guitar were exceptions to this rule. So if babies could use the
presence of a stressed syllable as a guide to the beginning of a word, they would be correct
most of the time.
Stress patterns vary from language to language and they make up an important part of a
languages prosodic structure. As you read in the Christophe and Morton (1998) study, babies
can tell one language from another on the basis of prosody so it is perhaps not surprising to
nd that they are very sensitive to stress patterns.
Transitional probabilities
Although syllable stress is a good cue to the beginning of a new word, it is not
always reliable. However, there are other cues available for babies to identify
word boundaries which are not based on prosody. One further example of a cue
is the probability of certain syllables appearing together. This is known as
transitional probability.
69
70
Johnson a n d J u s c z y k s r s t s t u d y
In their rst study, Johnson and Jusczyk looked at the effect of transitional
probabilities on 8 month olds. In order to consider transitional probabilities, they
had to control the infants prior experience of syllable sequences. To do this,
Johnson and Jusczyk invented words by taking twelve syllables and combining
them into four sequences to make pakibu, tibodu, golatu and daropi.
There were two phases in the experiment. During the rst phase called the
familiarization phase the infants listened to the words repeated over and over
in random order for 3 minutes, with no pauses between them. Because there
were no pauses, what the infants heard was a long, uninterrupted string of
syllables. Furthermore, because the order of the words was randomized, the
sequences of syllables that resulted from two words being next to each other (for
example, bu-go-la, heard only on those occasions when pakibu was followed by
golatu) were heard less frequently than the sequences of syllables that formed the
words (pa-ki-bu).
Because the sequences of syllables that formed words occurred more often
than those that occurred as a result of one word following another, transitional
probabilities could be used to distinguish the words from other syllable
sequences. To get some idea of what the babies were listening to, try saying this
sequence out loud, making sure that you keep the interval between each syllable
the same: pa/ki/bu/ti/bo/du/go/la/tu/da/ro/pi/go/la/tu/pa/ki/bu.
The second phase of the experiment was the test phase. In the test phase,
having listened to the syllables during the familiarization phase, the babies were
presented with all the words that they had heard and also with part-words where
syllables from two words were recombined. For example the part-word, tudaro,
was formed from the last syllable of golatu and the rst two syllables of daropi.
2 FIRST WORDS
Part-word
8
Word
8 month olds
Johnson a n d J u s c z y k s s e c o n d s t u d y
Having shown that infants could detect and remember transitional probabilities in
their rst study, Johnson and Jusczyk went on to investigate the role of syllable
stress in word boundary detection. Using a similar experimental paradigm, with a
familiarization phase followed by a test phase, they manipulated stress so that
every time a part-word appeared during the familiarization phase it was stressed
on the rst syllable. There were three possible outcomes for the second test
phase.
.
.
.
The infants might still spend longer listening to the part-words, as in the rst
experiment. This would indicate that stress was not a strong cue; that the
infants had not made use of the stress to highlight part-words.
The infants might listen equally to words and part-words. This would imply
that both cues were equally strong.
The infants might spend longer listening to the words than the part-words,
that is, they would show the opposite response to the one observed in the
rst experiment. This would suggest that the stress cues were very strong
and had overwhelmed the transitional probabilities, that the stress on the
rst syllables had effectively labelled these as words too.
71
72
t
Activity 3
Allow about
10 minutes
This activity will help you to compare the results of Johnson and Jusczyks two studies and think about
the implications for babies detection of word boundaries.
Figure 3 shows the results that Johnson and Jusczyk obtained in their second experiment.
Compare these results with those from their rst experiment shown in Figure 2 which you
looked at earlier.
How have the infants responded to the
words and part-words in the second
experiment?
Comment
The addition of stress as a variable brought
about a signicant change in the infants
behaviour. During the second test phase they
now listened longer to words than to
part-words. This is the opposite of the
response found in the rst study. Remember
that infants listen more to unfamiliar stimuli so
this indicates that they were now perceiving
the part-words as more familiar than the
words.
9
Part-word
8
Word
7
Mean orientation time (s.)
6
5
4
3
2
1
0
8 month olds
Johnson and Jusczyk concluded that prosodic speech cues such as stress are more
important than transitional probabilities in the detection of word boundaries. In
fact, adults and older children can take account of transitional probabilities even
when these are in conict with stress. In everyday speech you often hear words
spoken with atypical stress patterns but you will have no difculty in recognizing
them. For example, in Activity 2 there are the examples of guitar and across.
2 FIRST WORDS
Bruners theory
One theory that provides a useful framework for thinking about early word
learning was put forward by Bruner (1975, 1993). His key insight was that young
children encounter language in highly familiar social contexts because people
generally talk to them about familiar events and objects. So they hear speech that
relates to events, objects and people that they are already familiar with and that
are quite likely to be present or happening along with the speech.
To illustrate Bruners theory, consider the following specic example of the
way that early word meanings arise from a familiar social routine. The example
comes from the diary records I kept of my daughter, Francesca, as she began to
master her rst words.
73
74
t
Activity 4
Allow about
30 minutes
This will help you to examine the way in which things that are said to young children often relate
closely to what they are doing, and how this impacts on language development.
The following is part of a transcript made in a study by myself and two colleagues
(Harris, et al., 1983). It records a conversation between a mother and her
10-month-old daughter.
When the transcript begins, the child (who is just learning to walk) is standing holding onto
the arm of a chair with one hand. The mother is sitting on the oor next to her child.
The rst column of the transcript shows the time (in minutes and seconds) at which a
particular utterance was made during the session. The second column records the
mothers utterances and activity, and the third column describes what the child was
doing and saying. If you look at the childs column you will see that she is not yet saying
any words but she is babbling (recorded as vocalizes on the transcript). Where an utterance
is shown on the same line as an activity, they happened at the same time.
2 FIRST WORDS
Look at the mothers utterances in the transcript (numbered 113) and work out what she
is talking about in each case. Then draw a table as shown in the illustration below the
transcript. In your table, write each of the mothers utterances in the appropriate
numbered row in the second column. Then, in the third column write a brief description
of the topic of each of the mothers utterances. Be sure to include the following information:
.
Is the mother responding to something that has just happened?
.
Is the mother describing something that the child is doing?
.
Is the mother talking about something that she is doing?
There are two example answers shown in Figure 4 to help you.
Transcript of a conversation between a Mother (M) and her 10-month-old child (C).
Utterances are shown in bold type
Time
Mother
00.01
00.24
00.38
00.50
00.59
75
76
Comment
In our study we found that the majority of a mothers utterances referred to objects or events
that her child was currently attending to. The extract that you have just analysed was typical of
what we found. All of the mothers utterances refer to something that has just happened, to
something that the child is attending to, or the utterance actively directs the childs attention
onto a particular object. The rst two utterances occur immediately after the child has
touched a teddy on the oor. Then the child picks up the teddy and puts it next to a toy emu
(which was a glove puppet). The mother immediately responds by saying, And that one as
well?. She then picks up the emu, puts her arm inside it and says, Thats a very strange looking
thing. As she says this, she makes the emu peck the child.
You will see that most of the mothers utterances follow on from something the child had
done immediately beforehand. However, as well as taking her lead from the child, the mother
also provides supporting actions of her own that serve to make her general meaning clear.
Notice that utterances 811 are all concerned with the childs attempts to walk towards her
mother, hence they are focused on the childs current actions rather than objects. The
conversation then returns to the topic of the toy animals.
Bruner himself does not give many specic examples of exactly how familiarity
with the social context might help young children to understand what adults are
saying to them. It is clear, however, that different kinds of words place different
demands on young children. Like the childs attempts at walking in the activity
above, Francescas understanding of Are you ready? was not focused on a
particular object since she was learning to associate a question (Are you ready?)
with an action (lifting her head in anticipation of being pulled up). However, in
many other cases, a young child has to learn that a particular personal name (such
as mummy) is associated with the presence of a particular person, or that a
2 FIRST WORDS
Cues to reference
The development of an understanding of associations of this kind between
sounds and personal names or objects is only possible if a child has some way
of deciding what an adult is talking about. There are two specic cues that seem
to be signicant. These are gaze direction and pointing. Gaze direction alone can
be difcult to interpret. However, when someone turns to look at a particular
object or person, they often turn their head. Together, gaze, head turning and
pointing can provide invaluable cues to reference, that is, what someone is
referring to when they use a particular word.
t
Activity 5
Cues to reference
Allow about
5 minutes
This activity will help you to evaluate the different cues to reference that people use.
Imagine that you are listening to an unfamiliar language and have begun to recognize the
sounds of a few words. Now you want to know they mean. Think about how much you could
tell from where a speaker is looking or from where they are pointing. Which is the better
cue? Can you think of any other cues that might be used? Make a brief note of your answer
before reading further.
Comment
It turns out that pointing provides a more accurate cue to reference than either gaze or head
movement alone. This is true for both infants and adults. Adults are better at working out
what another person is referring to when the other person points than when they merely turn
to look at a particular object, and infants can locate an object more accurately when someone
points at it rather than merely looking at it (Butterworth, 1998). Another cue that people
sometimes use is to touch or pick up an object that is being named. Parents sometimes pick
up and animate objects that they are naming for children.
Pointing
Baldwin (1995) looked at the signicance of pointing in the acquisition of new
vocabulary. She studied the amount of time that infants looked at a novel object
when an adult pointed to it and compared this with the time spent looking at an
object when there was no pointing. Infants, who were as young as 10 months of
age, spent signicantly longer looking at objects when an adult pointed. When an
adult named an object, as well as pointing to it, the amount of looking was even
greater, suggesting that the young child is most predisposed to look at objects that
are singled out both through pointing and through naming.
However, there appears to be an even closer relationship between pointing
and rst words than that illustrated by Baldwin. In a longitudinal study, Harris, et
al., (1995a) found that the age at which children rst pointed which was around
77
78
10 months was highly correlated with the age at which they rst showed signs
of understanding the names of objects. This close relationship between pointing
and the understanding of object names is well illustrated by another example
from Francesca. Again, the example comes from the detailed diary records that I
kept.
The rst object name that Francesca understood was nose. She began by
touching the nose on a toy koala bear when asked and then, the following
day, she was asked Wheres mummys nose? and Wheres daddys nose?.
Francesca reached out and touched her parents noses. This rst occurred
when she was just over 9 months old. The very same day she pointed at a
plant in the conservatory. This was the rst time that she pointed.
This close relationship in time between rst referring to objects in the world by
pointing at them and rst understanding an object name suggests that these are
closely interlinked processes that may well have a common origin. For example,
when sighted children read picture books with their care-givers, there is a shared
attention to individual objects and animals through pointing. This seems to be one
way in which infants acquire object names. However, since using pointing in this
way obviously depends on a childs vision, it is relevant to ask what happens if an
infant is blind and this activity is not accessible in the same way. Blind children have
been found to produce signicantly fewer words for discrete objects than sighted
infants (Norgate, 1997) and this supports the importance of pointing in acquiring
object names.
t
Reading
At this point you should turn to the end of this chapter and read Reading B, Developments in
early lexical comprehension: a comparison of parental report and controlled testing by
Harris and Chasin (1999). This paper considers the accuracy of parental reports of early
comprehension and compares data collected using the MacArthur checklist with data
derived from experimental testing of comprehension. As you read this paper, think about
the factors that might contribute to the accuracy of parental reports about vocabulary.
2 FIRST WORDS
Figure 5 shows the average number of words understood by boys and girls
between 8 and 16 months of age as reported by their parents who completed the
MacArthur checklist. You can see two clear patterns. Girls are generally ahead of
boys in the number of words they can understand. However, for both boys and
girls, the overall pattern of development is essentially similar. The total number of
words comprehended grows fairly slowly up to about 12 months of age. Then
there is a sudden increase in vocabulary size.
300
250
Female
200
Male
150
100
50
10
11
12
13
14
15
16
Age (months)
Figure 5 Average numbers of words understood by boys and girls between 8 and 16 months of
age (adapted from Fenson et al., 1994, p. 74).
79
80
Connectionist
model
A computer
simulation with
many nodes and
connections
between them. As
stimuli are
presented, the
activation levels
of the nodes
change, as do the
weightings of the
connections
between the
nodes. Such
models can be
trained to learn
about and
recognize
particular stimuli,
and are often seen
as useful models
for how brains
learn.
2 FIRST WORDS
4.1 Babbling
Phonetic
contrasts
Differences
between the
individual sounds
that make up
speech.
Very young infants can distinguish a wide range of phonetic contrasts and, over
the rst year of life, they gradually lose the ability to discriminate contrasts that are
not present in the language they hear around them (Werker and Tees, 1984).
Over the same time period, signicant developments are taking place in infants
abilities to produce the sounds that they can hear. The development of babbling
as this early sound-making is called is divided into a number of stages.
.
At rst, babies do not produce speech-like sounds but they do make a
number of different sounds. By 3 months, these sounds begin to be used
communicatively and special sounds are used when babies are interacting
with other people. This is known as cooing.
.
By about 4 months, babies begin to engage in vocal play when they
experiment with the loudness and pitch of their vocalizations and the
position of their tongue. This experimentation gradually allows infants to
produce adult-like vowels and some of the features of adult-like consonants.
.
Around 6 months, another important change occurs in babbling. At this age,
babies rst begin to produce recognizable syllables. These are made up of a
consonant sound and a vowel. Very early sounds include da and ba. Oller
(1980) describes this prototypical or basic stage as canonical babbling.
.
A little later, at around 8 months, babies begin to produce reduplicated
babbling in which the same sound is repeated as in da-da and ba-ba.
.
Around 10 months there is another change as babies become capable of
variegated babbling (Oller, 1980). At this nal stage, babies begin to follow
one sound with another, similar sound such as ba-da or da-de.
81
82
t
Activity 6
Allow about
15 minutes
This will help you to compare the physical characteristics of the adult vocal tract with those of the
infant vocal tract and reect on how such differences might impact on the production of speech
sounds.
Adult human
Infant human
Adult chimpanzee
Soft
palate
Soft
palate
Tongue
Soft
palate
Tongue
Tongue
Epiglottis
Epiglottis
Vocal
cords
Larynx
Vocal cords
Epiglottis
Vocal cords
Larynx
Larynx
Figure 7 The anatomy of the vocal tract of chimpanzees and infant and adult humans.
Look at the three pictures in Figure 7 which show the vocal tracts of an adult chimpanzee, an
infant human and an adult human. Look carefully at the size and position of the tongue and
the position of the larynx (voice box) in each picture. Next, draw and ll in a table like the
one below and use the information you have gathered to decide whether the infant is more
like an adult human or more like a chimpanzee in the structure of its vocal tract. In order to
compare the position of the tongue and larynx you may nd it helpful to draw a line from the
lips to the larynx and look at the angle that is created.
Chimpanzee
Size of tongue relative to
mouth
Position of tongue
relative to lips
Position of larynx relative
to lips
Infant human
Adult human
2 FIRST WORDS
Comment
The infant humans vocal tract is not simply a miniature version of the adult human tract and,
up to the age of 3 months, it actually resembles the vocal tract of a chimpanzee more closely
than that of an adult human. The infants larynx is positioned high up so that the epiglottis
nearly touches the soft palate at the back of the mouth. The babys tongue is large in relation
to the size of the mouth, nearly lling the oral cavity, while the pharynx is very short compared
to that of an adult, allowing little room for the back portion of the tongue to be manipulated.
The reason for the characteristic shape of the infant vocal tract at birth becomes clear
once you realize that the main purpose of the infants tongue in the rst weeks after
birth is to enable the strong piston-like movements that are essential for sucking.
Once the rst 4 months are over, and the baby has gained in weight, sucking
becomes less of a priority. At this point, the vocal tract gradually changes into a more
adult-like form so that infants can produce the complex range of movements that
will be required for speech. Changes in the anatomy of the vocal tract are
accompanied by neural maturation of the related motor areas in the brain. Together
these developments enable infants to develop increasing control over the ne motor
movements that are essential for producing the full range of speech sounds.
150
125
100
Female
75
50
Male
25
10
11
12
13
14
15
16
Age (months)
Figure 8 Average numbers of words produced by boys and girls between 8 and 16 months of
age (adapted from Fenson et al., 1994, p. 75).
83
84
are taken from the MacArthur study (Fenson et al., 1994). Although parents are
often unsure about childrens early understanding of words, they can usually
provide very reliable information about the new words that their children are
producing. When collecting such data it helps to provide parents with clear
instructions about what counts as a word. This is important because young
children often simplify and shorten long words (e.g. saying nana for banana or
mama for mummy), and hence such utterances, although not strictly words, need
to be counted as such.
The MacArthur data show that most children produce their rst word at around
10 months and gradually produce more words over the next few months. As with
comprehension, there is a sudden increase in the rate of learning new words. This
occurs around 13 months, and at this point girls are, on average, signicantly
ahead of boys in the number of words they are able to produce.
2 FIRST WORDS
.
.
.
So far you have read about the growth in childrens comprehension and
production of rst words. However, I have not yet said anything about what rst
words mean. Early word meaning has been a topic of considerable debate among
child language researchers and there has been a signicant shift in the prevailing
view.
85
86
You have just read about Jamess use of mummy and there which were both
context-bound. However, it is notable that James also used words in a
contextually exible way. He used teddy to refer to one particular teddy in a
variety of different contexts (for example, when he was sitting on a teddy and
when he was pointing to the teddys reection in a mirror). He also used the word
more when he was reaching into a toybox for some bricks, taking another drink
from his cup or holding out an empty bowl. Another child in our study, a girl
called Madeleine, rst used the word shoes in a range of situations including
looking at pictures of shoes in a book, pointing at her own shoes and also when
holding her dolls shoes.
In our study (Harris et al., 1988), we looked at the rst ten words produced by
four different children. Table 1 shows the number of context-bound and
contextually exible words that the children produced.
Table 1 The number of context-bound and contextually exible words produced
by four children. Words in bold type were initially used as names for objects
Type of word
James
Jacqui
Jenny
Madeleine
Context-bound
Mummy
Go
Quack
There
Buzz
Moo
Boo
Wee
Hello
Mummy
Here
No
Down
More
Go
Choo-choo
Bye-bye
There
There
Hello
Here
Bye-bye
Contextually exible
Teddy
Ball
More
Jacqui
Bee
Teddy
Doggy
Moo
Shoe
Car
Mummy
No
Teddy
Shoes
Brum
Woof
Baby
Yes
You will see from the table that all the children produced words of both types
although there was variation from child to child in the relative number of each
type. Jenny and Madeleine produced more contextually exible words, whereas
James and Jacqui both produced more context-bound words.
You will also see that several of the words in the table the ones in bold were
initially used as the names for objects. This is consistent with the view that
children do not suddenly develop a naming insight but in fact use object names
right from the beginning. Data from a study by Goldeld and Reznick (1990)
support this view. In a longitudinal study of 24 children, starting from the age of
just over 1 year, they found that half of the early words used by the children
before the vocabulary spurt were object names.
2 FIRST WORDS
However, this same study also found that a small number of children showed a
more gradual rate of lexical growth. These children seemed to be acquiring a
broader range of different word types, while most of the children, who were
acquiring mainly object words in their early language development, were more
likely to show the rapid increase in rate of learning. This suggests that there may
be different pathways in early language development, but a common feature for
many children is a period when learning names for things is a dominant strategy.
Issues relating to individual variability and their links to the vocabulary spurt will
be discussed further in Section 6.
87
88
Table 2 gives some other examples of the childrens rst words and their
mothers use of the same word.
You can see from these examples that hello was rather like Mummy in that the
child had a single use that corresponded exactly to the mothers use. As in the
case of Jamess use of Mummy, the corresponding mothers use was her most
frequent use of hello when talking to her child. The other two examples are more
complex because the rst use of these words is contextually exible. Teddy is an
object name that the child rst used only to refer to one particular teddy. (We
know that it was an object name rather than a personal name because the child
rapidly used it to refer to other teddies.) The mothers use was similar because she
also used teddy to refer to a particular big teddy. No was used by Jenny in a range
of situations where an anticipated action was not going to be carried out.
Interestingly Jenny used no both to refuse an action that was requested by her
mother and also in connection with her own action. Again you can see that the
mothers own use of this word covered a similar range.
Table 2 Examples of four childrens rst words and their mothers use
of the same word
Word
Hello
(Context-bound)
Teddy
(Contextually exible)
No
(Contextually exible)
2 FIRST WORDS
a precedent in the mothers own use dropped to 45 per cent. This compared to
the gure of 93 per cent that we had found for initial uses of the rst ten words.
A study by Hart (1991) also shows that the importance of direct experience
becomes much less important as children acquire more words. Hart compared
early vocabulary with later vocabulary to see whether there was a difference in
the frequency with which children heard the words that they acquire early on.
Hart found that there was a difference. Childrens rst words tended to be the
ones that their parents frequently used when talking to them on average these
words occurred 30 times in a monthly observation session. However, when the
children were 6 months older, the words that they were acquiring had typically
been used only twice in parental speech during an observation session.
Some effects of language experience do, however, appear to be of longer
duration. Children learning English have a predominance of object names in their
early vocabulary and relatively few verbs. In contrast, Gopnik and Choi (1995)
found that verbs appear earlier, and form a greater proportion of early vocabulary
in the speech of children acquiring Korean, while Tardif (1996) found that 21month-old children who were learning Mandarin Chinese had as many different
verbs as nouns in their vocabulary. These patterns reect the fact that Koreanand Mandarin-speaking mothers use relatively more verbs than English-speaking
mothers when they are talking to young children. It is thought that this is a result
of the difference in structure between Korean and Mandarin (both of which
emphasize verbs), and English (which emphasizes nouns).
89
90
t
Activity 6
Allow about
5 minutes
This will help you to think about possible reasons why mothers speech to their children with language
delay is different.
Read the four paragraphs above on language experience and language delay again. Note
down at least two contrasting explanations for the differences in the speech of mothers of
children with language delay, as compared with mothers of children with typical language
development.
Comment
There are at least two possible explanations for the relationship between the mothers speech
and the childrens language.
One possibility is that the differences in the mothers speech in the two groups arose because
of differences in the language ability between the groups. In other words, the two groups of
mothers might have been talking differently because the two groups of children responded in
different ways to what they were saying.
Another possible explanation is that these differences in maternal speech were responsible, at
least in part, for the differences in the childrens language ability.
There is good reason to prefer the second of the explanations offered above.
When we sampled the speech of the two sets of mothers, both groups of children,
who were then 16 months old, were producing similar speech. Evidence of
differences among the children did not appear until several months later. At 2
years of age, the slower developers were still producing single words but the
typical developers were producing utterances that were several words long. The
groups also differed in their vocabulary size, with the typical developers
producing signicantly more words than the slower developers. Because these
differences between the groups did not emerge until several months after the
mothers speech was sampled, it seems unlikely that the speech style of the
mothers was inuenced by the speech of their children, but rather that the
maternal speech style was inuencing the language development. These ndings
can therefore be seen as evidence that the close tying of maternal speech to the
current social context is an important factor in early language development.
However, the issue of the causes and effects of language delay is complex. For
example, it is possible that the 16-month measures of childrens speech did not
capture some aspects of the childrens speech that were differentially affecting the
mothers behaviour. It is also possible that the children with language delay were
less able to make use of their mothers speech to them. What we can be sure of is
that there is an interactive relationship between mothers speech and their
childrens language development.
2 FIRST WORDS
First words can have a variety of meanings. Some are context-bound and
are used only in a single context. Others are contextually exible which
means they are used in a less restricted way. Some of these contextually
exible words are object names or the names of people. Others are used
to describe events.
There is a close relationship between childrens use of rst words and
their experience of hearing words being used by adults in conversation
with them. Early words arise from situations where children hear the
same words being used in a consistent way to describe particular objects
and events.
As children acquire more words, they soon become less dependent on
their experiences. Later words have to be heard less often in order for
children to acquire them.
Children who do not have the opportunity to hear words being used
consistently in familiar situations can nd that the process of acquiring
rst words is more difcult.
Whereas it was once thought that childrens rst words were always
context-bound, there is now evidence to suggest that this is not always
the case.
91
300
+ 1 SD
250
Female
200
Male
150
_ 1 SD
100
50
10
11
12
13
14
15
16
Age (months)
Figure 9 Amount of variation in the numbers of words understood between the ages of 8 and
16 months by boys and girls (Fenson et al., 1994, pp. 745).
150
Number of words produced (average score)
92
+ 1 SD
125
100
Female
75
50
Male
25
_ 1 SD
8
10
11
12
13
14
15
16
Age (months)
Figure 10 Amount of variation in the numbers of words produced between the ages of 8 and
16 months by boys and girls (Fenson et al., 1994, pp. 745).
Although not all children will be producing words by 18 months of age, most will
comprehend a considerable number of words at this age. This suggests that the
rst thing to look for when a child is not yet talking is vocabulary comprehension.
If comprehension is poor relative to the norms, this could well be a sign of
language delay. Another indication that can be used in younger children is
babbling. The stages of babbling that you read about in Section 4 are fairly similar
2 FIRST WORDS
93
Children vary in the rate at which they acquire rst words and also in the
age at which they produce their rst word. There are smaller differences
in early comprehension.
Children also show variation in the content of their early vocabularies.
Some children, who adopt a referential style, have a large number of
object names in their rst 50 words. Other children, who have an
expressive style, have a relatively low number of object names.
Conclusion
At the beginning of this chapter, I outlined a number of essential steps that are
necessary for children to understand and produce their rst words. You have now
learned about each stage in this complex process and I want to end by asking you
to reect on the wide range of skills that come together when young children
begin to build up a vocabulary.
Look again at the chart of component skills that you saw in Section 1.
WORD PRODUCTION
94
When you consider all the skills that come together in learning rst words, you
can begin to appreciate what an extraordinary achievement it is when children
rst comprehend and produce words. And that, as you will see in Chapter 4, is
only the beginning of a learning process that continues for many years until
children become fully competent in their native language.
2 FIRST WORDS
References
95
96
2 FIRST WORDS
97
98
Readings
Reading A: Is Dutch native English? Linguistic analysis
by 2-month-olds
Anne Christophe and John Morton
Abstract
A variant of the non-nutritive habituation/dishabituation sucking method was
used to test 2-month-old English infants perception of languages. This method
tests for the spontaneous interest of the baby to a change in the stimulus. English
and Japanese were clearly discriminated. The difference between French and
Japanese was equally clearly not of interest to babies using this procedure, the
babies behaving as though both languages were classied simply as foreign. In
order to further specify babies representation of native and foreign language, we
used Dutch, which shares a number of suprasegmental features with English. The
results from our last 2 experiments indicate that a portion of our 612 week-old
babies consider Dutch as native, suggesting that we tapped in a transition period
where the babies are still rening the suprasegmental specication of their native
language.
One of the most important tasks for a new-born infant is to learn its native
language. The majority of babies grow up in a multi-lingual environment and
must learn some characteristics of their mother tongue as early as possible so as to
distinguish it from other languages. This is a particularly crucial ability, since
infants could not possibly learn the syntax of a language (that is, discover the
regularities shared by a number of sentences) if they worked on a database
containing sentences from several different languages (Mehler et al., 1994).
It has been shown that newborns can discriminate between their mother tongue
and a foreign language. Mehler, Jusczyk, Lambertz et al. (1988) found that 4-dayold French infants discriminate between French (their mother tongue) and
Russian stimuli, showing a preference for their native language (see also Moon et
al. 1993, for equivalent results with English and Spanish). In addition, newborns
are able to discriminate between utterances in two foreign and unfamiliar
languages (Mehler and Christophe, 1995; Nazzi, Bertoncini, and Mehler, 1998).
Most of these studies have been replicated successfully using speech which has
been low-pass ltered at 400 Hz. Under these conditions, prosodic features such
as intonation and rhythm are preserved, whereas most phonemic information is
missing. It is therefore probable that babies ability to discriminate between
languages is based on a representation of speech prosody. It is very likely that the
infants preference for their native language comes from their having learned its
prosody in utero. However, we still do not know the precise nature of the
prosodic representation that babies use to classify languages.
Babies of 2 months of age behave slightly differently from newborns. They still
discriminate between their native language and other languages but they fail to
show any recovery of interest when switched from one foreign language to
another. Thus, Mehler et al. (1988) showed that while 2-month-old American
2 FIRST WORDS
babies were able to discriminate between English (their mother tongue) and
Italian, they did not discriminate between French and Russian. A possible
interpretation of this counter-intuitive result is that, while newborns still attempt
to analyse in detail any speech sample they are exposed to, 2-month-old infants
have sufcient knowledge of their mother tongue to be able to lter out any
foreign language as being not relevant.
Hesketh, Christophe and Dehaene-Lambertz (1997) developed a variant of the
contingent sucking response method which has the advantage that it can be used
both with new-borns and with 2-month-old infants and can be used with
extended segments of speech. With this technique, 2-month-old English babies
distinguished clearly between English and Japanese. It is this technique which we
used to explore the infants abilities further.
Method
The method for the Hesketh et al. experiment will rst be briey described (see
Hesketh et al., 1997, for details). The other experiments to be reported used the
same technique apart from changes in language. The stimuli consisted of 80
sentences, half in English, half in Japanese, between 15 and 21 syllables long.
These were recorded by four female native English speakers and four female
native Japanese speakers respectively. Speakers were nave as to the aim of the
experiment and were instructed to read as naturally as possible. Ten sentences
from each speaker were selected and matched for syllabic length (17.8 syllables)
and duration (3.1 seconds). Each infant underwent two changes in stimulation,
one experimental (language) change, the other control (or speaker) change. The
key measure was the difference between these two changes.
Half the babies received the experimental change rst and the control change
second. In addition, the order of presentation of languages and of speakers was
counter-balanced across subjects. This yielded eight conditions. In each of the
three phases the baby heard sentences from two speakers with the idea of making
speaker change mundane.
Subjects were seated in a car seat placed in a sound-proofed chamber and offered
a standard (steam sterilised) pacier. One experimenter, out of view of the baby
and deaf to the stimuli, checked that the pacier stayed in the babys mouth
throughout the experiment. A second experimenter monitored the experiment on
the computer outside the chamber. The computer recorded the pressure of the
infants sucks via an analogue-digital card (NIDAQ), detected the sucking
responses and delivered the sentences through a ProAudio 16 sound board
according to the reinforcement schedule (see below). The computer also saved
both the moment and amplitude of each suck as well as the stimuli triggered by
the sucks. Hesketh et al. (1997) reported that the number of sentences triggered
was a cleaner measure than the number of sucking responses. Only this measure
will be reported here.
The experiment started with a short period without stimulation (about 30 secs) to
settle the infants. The rst phase of the experiment then began, during which
infants heard sentences in either English or Japanese contingently upon their
high-amplitude (HA) sucks. After a short shaping phase, three HA sucks were
required to trigger each sentence (such that there was less than one second
99
100
between two consecutive sucks). There was an ISI of at least 600 ms between
consecutive sentences. When reaching the end of an ISI period after presentation
of one sentence, the program looked back to see if HA sucks had occurred
recently: any sequence of three HA sucks such that the last one occurred within
the last 600 ms was used to instantly trigger a new sentence. This procedure
ensured uent presentation of sentences in case of sustained sucking activity.
Within each phase of the experiment, the order of presentation of the sentences
was quasi-random for each baby.
A switch in stimulation occurred after a predened habituation criterion had been
met. For two consecutive minutes the infants HA sucking rate had to be less than
80% of the maximum sucking rate from the beginning of the experiment. Each
phase of the experiment lasted at least 5 full minutes. Sixteen babies aged
between 612 weeks participated in the study, mean age 8 weeks 6 days. Subjects
were randomly assigned to one of the eight conditions prior to testing.
To assess the effect of the experimental manipulation, two kinds of analyses were
performed on the data: ANOVAs and non-parametric tests. For each baby we
counted the number of sentences triggered during the two minutes before and
after the experimental (language) switch. The difference between these two
values gives us a measure of dishabituation to the language shift. The equivalent
measure was computed for the control (speaker) switch. The difference between
these two dishabituation scores represents a discrimination index for each baby:
whenever this value is positive, the baby reacted more to the language change
than to the speaker change. These values are shown in Figure 1 (left hand
column).
A Wilcoxon signed ranks test showed that the median of the discrimination index
for the number of sentences triggered was signicantly above zero (Z = 3.4,
p < 0.001). In the ANOVAs, the dependent measure was the dishabituation scores
for the Experimental and Control switches. There was one within-subject factor
(Experimental vs Control switch) and two between-subject counterbalancing
factors, Order (experimental switch rst, versus control switch rst), and
Language (English rst vs Japanese rst). There was a main effect of the
Experimental factor (F (1,12) = 11.6, p < 0.01), no signicant effect of any of the
counterbalancing factors, and no interactions between the Experimental and
counterbalancing factors.
2 FIRST WORDS
Figure 1 Results of four language discrimination experiments using the modied non-nutritive sucking method
(Hesketh et al., in press) where sentences are presented contingently upon sucking responses and each baby is submitted
to two shifts of stimulation, one experimental (language change) and one control (speaker change). A discrimination
index is computed for each baby: it represents the difference between the increase in the number of sentences triggered
for the language change and the increase for the speaker (or control) change. Whenever this value is positive, the baby
showed more interest in the language change than in the speaker change. Increases in the number of sentences triggered
are computed using two minutes before and after each shift of stimulation.
101
102
The distribution of the discrimination index can be seen in the second column of
Figure 1. The infants gave no indication of being more interested in language
change than in speaker change (Wilcoxon, Z<1; ANOVA: F (1,12) <1). This result
is signicantly different from the results of the experiment with English and
Japanese. In an ANOVA contrasting the distribution of discrimination indices in
these experiments, F (1,28) = 8.32, (p < 0.01).
The lack of interest shown by 2-month-olds in the differences between foreign
languages is in line with previous work. Paired with Nazzi et al.s (1998)
conrmation that newborn infants can discriminate between sentences belonging
to two foreign languages (with the same experimental technique as here), this
result conrms the developmental trend already described. Our best
interpretation is that 2-month-old infants have enough knowledge of the
properties of their native language to be able to lter out foreign input as being
irrelevant to their language learning. In that case, both French and Japanese
would simply be classied as foreign and would not be analysed to a sufcient
depth to allow the differences to be detected.
The results of the rst two experiments immediately pose a new question: how
specied is the 2-month-olds representation of their mother-tongue? What do
they consider native, and what do they lter out as being foreign? To answer this
question, we picked a language which shares with English a number of prosodic
properties. Dutch, like English, has vowel reduction, complex syllabic structure,
and the same sort of word stress as English. These factors lead to both English and
Dutch as being heard as stress-timed (Cutler et al., 1997). In fact, Dutch and
English have already been shown to be rather similar to babies ears: Nazzi et al.
(1998) demonstrated that French newborns do not distinguish between Dutch
and English ltered sentences.
2 FIRST WORDS
Discussion
Using the modied contingent sucking response we have shown that English
2-month-olds discriminate English from Japanese but not French from Japanese.
Given that this habituationdishabituation technique measures infants interest in
changes in auditory stimulation, it allows us to evaluate their spontaneous
partitioning of perceptual space into categories. In the present case, the results
suggest that babies form two major categories, one for English, which could be
termed native or mother tongue, and one, with French and Japanese, of foreign
languages.
In the last 2 experiments of this paper, we studied English infants perception of
Dutch, a language that is prosodically very similar to English. We contrasted
Dutch to English and to Japanese. If English babies treat Dutch as native, they
should not be able to discriminate between English and Dutch, but should readily
distinguish Dutch from Japanese; in contrast, if they have already set up their
native category such that Dutch is excluded, they should distinguish between
Dutch and English but ignore the difference between Dutch and Japanese, both
of which would be in the category foreign. Both experiments gave marginally
signicant results, indicating that some English babies consider Dutch as native
but others do not. The former would distinguish Dutch from Japanese but not
from English; the latter group would distinguish Dutch from English but not from
Japanese. If we tested babies in both conditions, we predict that whenever one
condition works the other would not. What factors may account for this individual
variation? The most obvious candidate is age. At one month, all infants might
103
104
regard Dutch as native, whereas by four months they might all have excluded it.
Further, we would expect early exposure to different languages to affect the
speed of setting up a tight specication of native but it is unlikely to be a factor in
our experiments since we selected the babies to come from monolingual English
households.
Eventually we will need to distinguish between environments where second
languages are addressed to the infant from those where second languages are
present but not directly addressed. The second case might accelerate the
denition of native whereas the rst case, true bilingualism, might lead to
confusion. Recent experiments by Bosch and Sebastian (1997) showed that by
four months of age, bilingual Spanish/Catalan babies already behaved differently
from monolingual babies (either Spanish or Catalan). Monolingual babies
oriented faster to their mother tongue than to English. In contrast, bilinguals
orient to Spanish or to Catalan signicantly more slowly than to English. Is this
because of confusion? Apparently not, since in more recent and still unpublished
work, these authors showed that, although Spanish and Catalan are close, both
monolingual and bilingual 4-month-olds can discriminate between them. Of
course the gap between these 4-month-olds and our 2-month-olds is enormous
and it could be that at 2 months bilingual babies are confused. At any rate, it has
become clear that, from birth, infants work hard at learning what language is
native.
Acknowledgements
The work reported in this paper was assisted by a grant from the Human Frontiers
Science Programme, the Human Capital and Mobility Programme, and the
European Science Foundation. We especially want to thank Sarah Hesketh, Jon
Bartrip and Sarah Minister for their help in recruiting and testing subjects.
2 FIRST WORDS
Mehler, J., & Christophe, A. (1995). Maturation and learning of language in the
rst year of life. In M. S. Gazzaniga (Ed.), The Cognitive Neurosciences: A
handbook for the eld (pp. 943954). Cambridge, Mass.: MIT Press.
Mehler, J., Dehaene-Lambertz, G., Dupoux, E., & Nazzi, T. (1994). Coping with
linguistic diversity: The infants viewpoint. In J. L. Morgan & K. Demuth (Eds.),
Signal to Syntax: Bootstrapping from speech to grammar in early acquisition (pp.
101116). Mahwah, New Jersey: LEA.
Mehler, L., Jusczyk, P. W., Lambertz, G., Halsted, G., Bertoncini, J., & Amiel-Tison,
C. (1988). A precursor of language acquisition in young infants. Cognition, 29,
143178.
Moon, C., Cooper, R., & Fifer, W. (1993). Two-day-olds prefer their native
language. Infant Behavior and Development, 16, 495500.
Nazzi, T., Bertoncini, J., & Mehler, J. (1998). Language discrimination by
newborns: Towards an understanding of the role of rhythm. Journal of
Experimental Psychology: Human Perception and Performance, 24, 111.
Source: Christophe, A. and Morton, J. (1998) Is Dutch native English?
Linguistic analysis by 2-month-olds, Developmental Science, vol. 1,
pp. 21519.
105
106
Introduction
This paper compares data on early lexical comprehension derived from parental
report with that from systematic experimental testing. A major source of data on
the composition of early comprehension vocabularies comes from Fenson, Dale,
Reznick, Bates, Thal & Pethick (1994) who describe an extensive sample of
children whose vocabulary was assessed with the MacArthur Communicative
Development Inventories (Infant Scale). These data are cross-sectional rather than
longitudinal but they do provide important evidence about the rst 50 words that
children understand.
Pooling data across subjects, Fenson et al. found that the words understood by
the youngest children (aged 0;8) were the names of people or were related to
games or routines. At 50 words, the main category was nouns (comprising
household items, animal names, toys, clothing, food and drink, body parts,
furniture and rooms) which accounted for 48% of items. The other categories
were games and routines (20%), actions words (16%), personal names (10%) and
sounds (6%). By 1;4 when mean comprehension vocabularies were reported as
192 words 52% of words understood were nouns, 19% were verbs, 9% were
words stemming from games and routines and 3% were personal names. (For a
complete list see Fenson et al. (1994) table 16.)
The use of the MacArthur Communicative Development Inventory for the
assessment of early comprehension vocabulary has been questioned since it relies
exclusively on parental report. Although there is a long tradition of using parental
report for reliable assessment of production it is less clear that this can provide
equally reliable assessment of early comprehension since it is often difcult to
determine from observation alone whether a child understands a word or is,
2 FIRST WORDS
Method
Participants
Six children took part in the study, four boys (Ben, Andrew, Sebastian and
George) and two girls (Katherine and Katy). All the children were rst born and
English was the only language spoken in the home. At the time of the rst
observation the children were 0;6 (range 0;5.240;6.06). Observation continued
until the children were 2;0 although all children had attained a comprehension
vocabulary of at least 100 words by 1;6. Data for the early production and
comprehension of these children have been reported in Harris, Yeeles, Chasin &
Oakley (1995a) and Harris, Barlow-Brown & Chasin (1995b).
Procedure
Assessment of comprehension
Full details of the procedure used to assess comprehension can be found in Harris
et al. (1995a). Briey, three different sources of evidence were used in the rst
instance. These were parental diary records, home observation (supplemented by
videotaping) and a comprehension checklist which contained the most
commonly understood words organized into categories (e.g. toys, food and drink,
people, games, actions). The checklist was a modied version of the one used by
Benedict (1977).
107
108
Denition
Examples
Personal name
Object name
Context-bound
object word
Action word
Assessment of production
The development of production was monitored through diary records, maternal
interviews, home observations and video recordings. Details of the procedure are
set out in Harris et al. (1995a). Briey, a vocalization was counted as a word if it
was reported in the diary record and observed either during home observation or
in a videorecording. If there was no maternal report, three observations were
required before a vocalization was counted. Unlike comprehension, controlled
testing was only carried out if there was some ambiguity about the range of
contexts in which the child produced a word as, for example, where a diary entry
and an observation were not identical.
2 FIRST WORDS
Results
The proportion of words comprehended in each of the four categories was
calculated for vocabulary sizes of 20, 60 and 100 words. The proportion of words
in each category was found to change as the size of comprehension vocabulary
increased (see Table 2). At the 20-word level, the number of personal names and
object names was almost equal and together they accounted for two thirds of the
total. The remaining two categories context-bound object words and action
words each made up about one sixth of the total. At 60 words, the proportion of
object names and action words had both increased while there was a marked
decrease in the proportion of personal names. By the 100-word level the
proportion of personal names had decreased even further and there was a
corresponding rise in the proportion of object names. The proportion of action
words remained the same as at the 60-word level as did the proportion of contextbound object words which was identical at all three points.
Table 2 Mean percentage (and range) of words in each category in relation to size
of comprehension vocabulary
Vocabulary
size
Personal
names
Object names
Context-bound
object words
Action words
20
32.5 (2060)
35 (1560)
15 (030)
17.5 (1025)
60
18 (1223)
40 (3052)
15 (823)
27 (1537)
100
15 (1019)
43 (3056)
15 (1120)
27 (1337)
Gender
Comprehension
Production
Andrew
59
11
Ben
48
46
George
52
Sebastian
80
39
150
30
Katherine
224
65
Katy
111
25
210
95
Individual data for the six children reected the overall pattern. The number of
object names understood by each child generally increased with the size of their
comprehension vocabulary. However there was considerable individual variation
in the number of different categories of word understood at each stage. A
109
110
Discussion
Our data support the nding of earlier studies that the proportion of different
classes of words comprehended changes with vocabulary size (Benedict, 1977;
Bates, Bretherton & Snyder, 1988; Gunzi, 1993; Fenson et al., 1994). Personal
names gured prominently in early vocabulary but they made a relatively smaller
contribution as the total number of words understood by the children increased.
This decrease in the importance of personal names occurred as the proportion of
both object names and action words increased. The proportion of context-bound
object words remained stable throughout the period of development.
At 60 words, the mean proportion of object names was 40%. This is very
comparable to Benedicts (1977) data for 50 words but somewhat lower than the
48% reported by Fenson et al. (1994). The mean proportion of action words at the
same vocabulary size was 27% which was considerably lower than the
proportions reported by both Benedict and Gunzi for 50 words but very similar to
the total proportion reported by Fenson et al. for action words plus words related
to games and routines (both of which were classied as action words in the
present study). Our data suggest that some early words that appear to be object
names are, in fact, context-bound object words and that a parental checklist may,
therefore, over-represent the number of object names. There is also some
suggestion from our data that parents over-estimate the number of words that
children understand since even the most precocious child that we tested attained
a score that was under the 75th percentile on the Fenson et al. norms; and there
2 FIRST WORDS
111
112
Chapter 3
development
Contents
Learning outcomes
115
Introduction
1.1 Innateness and epigenesis
115
116
118
120
122
123
124
125
127
132
Language
4.1 An instinct for language?
4.2 The evidence from developmental cognitive
neuroscience
4.3 An instinct for language: review
135
135
cognitive development
5.3 The prefrontal cortex and the acquisition
of new skills
5.4 A note on relating brain structure and function
138
143
145
145
146
147
149
Conclusion
150
Further reading
150
References
150
Readings
Reading A: The fragmented mind from
154
154
158
Learning o u t c o m e s
After you have studied this chapter you should be able to:
3 discuss the case for and against innate specication of cortical function;
cognitive development;
5 outline some aspects of the relationship between brain development and
cognitive and language development.
6 In addition you should have:
. enhanced your understanding of principles of self-organization;
. further developed your understanding of plasticity;
. reinforced your understanding of nativism and constructivism;
. extended your knowledge of research methods.
Introduction
This chapter is about the development of the human brain in infancy, and about
how changes in the physical structure of the brain can be related to childrens
cognitive and language development. In the course of this chapter we will
address the following questions.
.
What is the basic course of pre- and postnatal brain development?
(Section 2)
.
How do different areas of the brain come to perform different functions?
(Sections 3 and 4)
.
Are different areas of the brain innately pre-specied to take on their
respective cognitive functions? (Sections 3 and 4)
.
How can cognitive function be related to the development of brain
structure? (Section 5)
.
How much can be learned about cognitive development from studying
brain development? (Sections 4 and 5)
.
How is it possible to study brain function? (All sections)
We start in Section 2 by taking a brief look at some basic features of brain
development in the child before birth and in the rst few months after birth. A key
idea in this section is that some brain development goes on postnatally,
maximizing the opportunities for brain and cognitive development to be
inuenced by both genes and environment, in interaction with each other.
115
116
Cerebral cortex
The layer of cells
on the outer
surface of the
forebrain; only
found in
mammals, and
particularly well
developed in
humans.
In Section 3 we examine the idea that different areas of the cerebral cortex (see
Figure 1, p. 118) specialize in performing different cognitive functions. One way
of thinking about this functional specialization within the human brain is in
terms of cognitive modules. Cognitive modules are hypothetical constructs that
help to map what we know about brain function (what the brain does) to brain
structure (its physical make-up). Section 3 considers the debate about how the
cognitive functions performed by such modules might develop. Do they unfold
from a predened genetic blueprint, or are they dependent upon the childs
interactions with the environment? We conclude that these cognitive functions
develop according to epigenetic principles, with genes and environment
inextricably linked. Section 4 considers this argument in detail in relation to
language development.
In Section 5 we look at how psychologists relate knowledge about childrens
cognitive development to knowledge about the developing structure of the
cortex. This section focuses on the prefrontal cortex (see Figure 1, p. 118) because
of its importance within the eld of developmental psychology. The prefrontal
cortex has received considerable attention from psychologists because of its
involvement in higher mental functions, such as the planning and initiation of
actions, and the inhibition of irrelevant behaviour (so called executive functions;
see Chapter 5).
The chapter as a whole is about the relatively new eld of developmental
cognitive neuroscience (Johnson, 1997). This interdisciplinary eld has emerged
partly as a result of new and improved ways of investigating brain structure and
function (such as advances in neuroimaging) which have led to signicant
progress in understanding how brains are constructed during development.
These methods will be described and illustrated throughout the chapter. Their
importance for the psychological study of child development is that they allow us
to investigate the relationship between the development of brain structure on the
one hand and cognitive and behavioural development on the other.
Epigenesis
Development
by means of
interaction
between genes
and their
environment.
Neuroimaging
The use of various
technologies for
the non-invasive
measurement of
brain activity,
aimed at
specifying the
functions of
different brain
regions. Also
referred to as
brain mapping.
t
Activity 1
Allow about
10 minutes
This activity is a gentle introduction to some of the complex issues you will encounter concerning
genetic and environmental inuences on child development.
Think of ways in which someone you know well is psychologically similar to their parents.
You might focus on personality traits, skills, interests, habits and so on. Write down a list of
two or three examples.
For each example, think about the following questions and make brief notes.
.
What factors have caused this person to be like this? Has this trait been developed
through experience, or has it been inherited from their parents?
.
Was it inevitable that this person would develop in this way, or could other choices or
experiences have produced different outcomes?
117
118
Comment
At the very least this activity should highlight some of the complexities (perhaps the
impossibility) of trying to separate genetic and environmental inuences on human
development. If children grow up to be like their parents in some ways, is this because they
learned to be like this from them (an external factor) or because they inherited these
similarities from them (an intrinsic factor)?
2
Plasticity
The ability of a
system (such as a
human brain) to
change and adapt
to circumstances.
CEREBRAL CORTE
Dorsolateral
region
Parietal
lobe
Prefrontal
cortex
Visual
cortex
Temporal
lobe
Figure 1
The sequence of events that goes towards building human brains is very similar to
that observed in other mammals, but the timescale over which these events occur
is signicantly more extended. This slower timescale has two major
consequences. First, there is a prolonged period of postnatal development during
which the later stages of brain development can be inuenced by interaction with
the outside environment. If the timescale of human brain development were
quicker, and babies were born with more mature brains, then much more of their
brain development would take place in the relatively limited environment of the
mothers womb.
Second, the more delayed the general time course of the development of a
species, the larger the relative volume of the later developing areas of the brain.
In humans, the slowed rate of development is associated with a relatively larger
volume of cerebral cortex, and an especially large prefrontal cortex (see Figure 1).
The prefrontal cortex is considered by most investigators to be critical for many
high-level cognitive abilities (Milner, 1982; Goldman-Rakic, 1987; Fuster, 1989)
which you can read about in Section 5, and in Chapter 5 of this book. It is
signicant that the region of brain that undergoes most postnatal development,
in interaction with the rich, external environment, is the region most closely
associated with high-level cognitive abilities such as the planning and execution
of complex sequences of behaviour.
t
Activity 2
Allow about
10 minutes
This activity explores what psychologists mean by the term environment, and emphasizes that
environmental inuences on child development operate before birth.
Imagine for a minute the prenatal environment of the child (the mothers womb). In what
ways do you think it is similar to the childs postnatal environment? In what ways is it
different? Write a list of characteristics under the headings as suggested below. See Figure 2
for an example to start you off.
Prenatal environment
Postnatal environment
Constant temperature
119
120
Comment
The environment in developmental psychology means everything outside the child. So,
postnatally, childrens environments include all aspects of their physical and social worlds.
Clearly this is considerably richer in many ways than the restrictive environment of the
mothers womb. The sample answer in Figure 2 indicates that the postnatal environment is
much more variable, even at a simple physiological level. In addition, in the mothers womb the
child cannot move easily, and has very limited access to information from the senses (which
are also at an early stage of development). However, it is important to realize that the
mothers womb is an environment, so even prenatal development cannot be put down
exclusively to the unfolding of a genetic blueprint. Look back at the work of DeCasper and
Spence (1986) and DeCasper et al. (1994) in Chapter 2 (Section 2) of this book to see
examples of how the environment impacts on the prenatal child.
circumstances are right, pass the signal down the neurons axon and on to other
neurons. The pattern of branching of dendrites is important, because it affects the
amount and type of signals the neuron receives. The points of communication
between neurons are called synapses (see Figure 3). Synapses begin to form in the
brain in the early weeks of gestation. The generation of synapses occurs at
different times in different areas of the cortex.
Synapses
connecting
other
neurons
Neuron
cell body
Axon
Myelin sheath
around axon
Synapse
(a)
Neuron
(b)
Figure 3 (a) A diagram of a neuron showing an axon (much shortened) with axon terminal
(synapse) and dendrites with synaptic connections to other neurons; (b) a more realistic picture of
the density of a dendritic tree in a human neuron (adapted from Stewart, 1991).
121
122
Motor cortex
The section of
the cortex that is
associated with
the control of
movement.
Human genome
The complete
set of genetic
information
contained in
human
chromosomes.
Visual cortex
The region at the
rear of the brain
that processes
visual
information.
(a) Neonate
Figure 4 Dendrites in the visual cortex of human infants. As the dendritic trees of neurons
extend, so the overall connectivity in the brain increases (adapted from Conel, 193963).
2.3 Plasticity
Plasticity is an inherent property of the developing brain. At birth the brain
appears to be highly adaptable or plastic. If one area is damaged, perhaps
through a localized brain injury, or a stroke, other brain regions can take over the
processing from the damaged region. In adult life, the brain is considerably less
plastic, and localized brain damage is much harder to overcome.
Brain growth involves a process of increasing specialization in the sense that
tissue and cells become more differentiated in their structure and functioning as
development progresses. Sometimes this increased specialization is referred to as
a restriction of fate because at the outset of its life a cell might take on any
number of forms and functions, but as development progresses, and as it assumes
its mature form and connectivity, its fate becomes increasingly specied.
Plasticity simply represents the state of not yet having achieved specialization at
some level. As an example, consider a piece of tissue from the cerebral cortex that
may not yet have developed its specialization for processing a certain category of
information when a neighbouring region is damaged. The same developmental
mechanisms that would have ensured specialization for one type of processing
may now bias the tissue towards the type of processing normally undertaken by
its damaged neighbour. Thus in many instances, atypical patterns of brain
specialization in developmental disorders may reect the action of normal
developmental processes following some earlier disturbance to the typical
developmental pathway. Identifying and understanding the mechanisms
underlying specialization, particularly in postnatal life, remains one of the major
challenges for developmental cognitive neuroscience.
123
124
Darwinian
selection
The process of
natural selection
identied by
Darwin and also
known as
survival of the
ttest. When
applied to a
population of
neural pathways
in the brain, this
concept suggests
that the strongest,
most active
pathways survive
at the expense of
weaker, more
infrequently used
ones.
Encapsulation
The process by
which a neural
pathway comes to
have a specialized
function that is
relatively
unaffected by the
activities of other
neural pathways.
As you will see in Section 3, the adult brain is characterized by cortical regions
that are dedicated to different types of information processing and different
cognitive functions. As neurons become increasingly differentiated, their fate is
restricted, and their function becomes more specialized. But how does this
happen? What processes are involved in the development of specialist cortical
regions?
There is no clear agreement on this matter, so we will briey look at one
possible view of these processes, called selectionism (Changeux, 1985).
Remember that cortical development sees a rise in the number and density of
synaptic connections, followed by a fall. The time course of this rise and fall
differs from region to region. According to selectionism the decrease in synaptic
connectivity that can be observed is a function of a kind of Darwinian selection.
Neural circuits that are used more frequently (perhaps in response to stimuli from
the environment) are preserved, and those that are activated less frequently are
weakened. As the selective loss of redundant pathways increases, so does the
separation of the surviving pathways. In the long term only neural circuits that are
actively employed remain intact, and these become increasingly encapsulated
(separated off from other pathways, and specialized in function). In essence,
selectionism holds that increased specialization of cortical structure comes about
through selective loss of connectivity. Note that stimulation from the environment
is one determinant of which pathways are employed and ultimately preserved, so
these processes can be thought of as a type of learning.
There is some evidence to support this account. For example, there is evidence
to suggest that information from different senses (touch, vision, hearing) may not
be clearly differentiated at a perceptual level in very young infants. For example,
Lewkowicz and Turkewitz (1981) showed that a newborn infants visual
preference for looking at a bright or dim visual stimulus was affected by prior
exposure to an auditory stimulus. Such confusion of different modes of sensory
input (cross-modal transfer) indicates a relatively unspecialized sensory cortex;
perhaps one with a proliferation of somewhat haphazard connections, which is
not yet able to transmit information across specialized neural pathways without
triggering activity in other pathways.
Similarly, Stetri (1987) found evidence of cross-modal transfer of tactile and
visual information at 3 months of age, but no such transfer at 5 months of age.
Meltzoff and Borton (1979) found that 1-month-old infants preferred to look at
the shape of a dummy which they had previously been given to suck. This crossmodal effect could not be detected in a sample of 3 month olds (Maurer, 1993).
These studies suggest that infants become increasingly able to differentiate
between information from the various senses. Perhaps the relevant pathways
have, in these older infants, become sufciently encapsulated to allow the
processing of information from one sensory input to be undertaken relatively
independently of other sensory pathways.
Synaesthesia
The occurrence
of multi-channel
sensory experiences. The way in
which some
people taste
colour, or hear
visual stimuli.
Summary
.
.
.
3
Neuropsychology
The branch of
psychology that
specializes in the
study of the brain
and behaviour.
of Section 2
125
126
This section will examine how such modules develop. Are they genetically
pre-specied, simply unfolding as the child matures? Or are the functions and
underlying structures of modules inuenced by interactions with the
environment?
t
Activity 3
Allow about
20 minutes
This activity encourages you to reect on the way in which psychologists reason about genetic and
environmental inuences on development. It also aims to get you thinking about how the different
functional areas of the brain might develop.
The fact that most typically developing children and adults end up with the same sort of
perceptual, cognitive and motor functions performed by the same regions of cortex seems
to suggest that the brain develops according to a genetic blueprint (in the same way that
other features of the human body develop in similar ways in different people on the basis of a
genetic blueprint). But such reasoning is not really sound.
To understand the problem with this reasoning it is necessary to consider the shared
environment in which people develop. At one level it would seem that peoples
environments are unique to each individual. But if you take a few steps back and look at the
bigger picture, you should realize that typically developing children and adults live in very
similar environments. Spend a few minutes thinking about the following questions and write
down your answers before reading further.
1
In what ways can one persons environment differ from the environment of another?
Comment
Peoples environments vary in all sorts of ways, and it should not be difcult to list just a few of
these. However, at a more global level there are also many ways in which it can be said that
peoples environments are alike. For example, typically developing people are surrounded by
very similar perceptual worlds, with similar physical constraints (owing to the laws of physics
and biochemistry) on what they perceive and what they do. They live in a world of shared
objects that behave in certain predictable ways. It is necessary to think at this more global level
to understand why similar cortical structures and functions do not necessarily imply that genes
are the only factor involved in brain development. Genetic determination may play a part, but
they may also be the result of a shared environment. It is more likely that the cortical
structures and functions that typically developing adults have in common are a result of
epigenetic processes which involve the interaction of genetic and environmental factors.
3
Some people experience atypical patterns of development; for example people with
motor disabilities such as cerebral palsy, or sensory disabilities such as hearing loss or
visual impairment. Can you think of ways in which people with these kinds of disabilities
might not share a similar environment with typically developing children and adults?
What implications might this have for our understanding of brain structure and function
in atypical development?
Comment
People with different kinds of sensory disabilities operate in a world in which they rely on
different kinds of sensory input. For example, people with an autism spectrum disorder may
live in rather distinctive sensory worlds (Williams, 1996). Children with motor disabilities may
not act in and on their environment in the same way as typically developing children. If the
epigenetic view is correct we might expect such factors to have an effect on brain structure
and function. In other words, if differences in brain structure and function are detected when
comparing typically and atypically developing people, we should not assume that these
differences are the cause of the disability; they might result from the disability.
Before we start the discussion of how modules develop, we need to make a very
important conceptual distinction one that was alluded to in the introduction to
this chapter. This distinction is between neural modules and cognitive modules.
Neural modules are real, structural units of the brain. There is plenty of evidence
for their existence. Cognitive modules, on the other hand, are hypothetical
constructs that provide a way of thinking about how the different functions of the
brain might be performed. There is considerable debate about whether cognitive
modules exist (that is, whether the brain works in a modular way), and if they
do, what sorts of functions they perform. Developmental cognitive neuroscience
can illuminate the extent to which functions of the brain (cognitive modules) map
onto structures of the brain (neural modules). Some of this work is described in
Section 5. It is important to realize that, unless otherwise stated, whenever the
word module is used in this chapter it is being used to refer to cognitive
modules. The debate that we shall look at in this section is between two theorists
Fodor and Karmiloff-Smith who share a view that the brain does function in a
modular fashion. Their disagreement is with regard to how modules develop.
One theorist holds that they are innately specied; the other holds that
modularization occurs as the brain develops.
Fodor (1983) is the key advocate of innately specied modularity. Arguing from
a nativist standpoint, he asserts that humans are born with the innate capacity to
develop information processing systems that allow them to make sense of the
world in which they have evolved. In this way he is affording the environment a
crucial role in the development of cognitive modules in one sense, but not in
another. To Fodor, the environment inuences the structure and function of
modules over the course of evolution in a phylogenetic sense (to do with the
long-term development of a species over many generations). In other words
human brains have become modular through adaptations to the environment,
during the course of evolution.
In contrast he does not believe that the environment plays a crucial role in
their development in an ontogenetic sense (that is when it comes down to
understanding the development of a given individual child). Fodors argument is
summarized in Box 1.
127
128
Comment
Phylogenesis is to do with the development of a species. Ontogenesis is to do with
the development of an individual member of a species. When discussing the role of
the environment in development, it is important to be clear about whether one is
talking phylogenetically (that is, about the development of a species through
evolution over many generations of individuals) or ontogenetically (about the
development over one lifespan of one or more members of the species from the
same evolutionary stage).
BOX 1
t
Reading
Now read Reading A The fragmented mind from What Infants Know by Mehler and Dupoux
(1994). This excerpt further illustrates the notion of cognitive modularity by drawing on
some clinical case studies. As you read the descriptions of the cases, consider how they
relate to Fodors model of the mind.
t
Reading
Now read Reading B Is the initial architecture of the infant mind modular? from Beyond
Modularity by Karmiloff-Smith (1992). This summarizes Karmiloff-Smiths objections to
Fodors nativist views, and sets out some features of her constructivist alternative. While
reading, focus on the notions of domain-specicity and domain-generality, and on her argument
that the brain becomes modular as a result of development.
What does the wider research evidence have to say on the matter of innate
modularity (Fodor) versus modularization (Karmiloff-Smith)? As ever the picture
is complex and incomplete. Some researchers (for example, Rakic, 1988) argue
that the structural differentiation of the cortex could be explained in terms of a
molecular and genetic specication (just as other organs in our body, such as the
heart, lungs and liver, are differentiated one from another, without particular
reference to interactions with the environment). If, once differentiated, the
various structures of the cortex always took on the same cognitive functions in the
developing childs mind, then this would support Fodors position.
However, the balance of evidence (OLeary, 1989; Elman et al., 1996; Katz
and Shatz, 1996; Johnson, 1997) suggests that for most regions of the cortex
differentiation of function results from an epigenetic system. That is,
differentiation results from an interaction between molecular and genetic factors
129
130
Neural activity
The electrochemical activity
of neurons as they
transmit
information from
one to another.
on the one hand and environmental factors on the other. This view holds that
neural activity, itself the product of both biological and environmental factors, is
the key to understanding differentiation of function in the cortex. Before birth,
this neural activity is probably largely a spontaneous and intrinsic process, driven
by genetic and molecular factors (though note our earlier reminder that the
mothers womb provides an environment in which the baby can, for example,
hear). After birth, however, neural activity is substantially inuenced by sensory
and motor experience as well as by intrinsic factors.
What evidence can be cited to support this view? Petersen et al. (1990) used
positron emission tomography (PET) (see Box 2) to study the responses of native
English speaking adults to written stimuli in the form of (a) English words, (b)
pseudowords that obeyed English spelling rules (for example oop and toglo),
(c) nonsense strings of letters (such as pxqlo) and (d) letter-like forms (false
fonts). They found that a specic region of the left visual cortex only responded
to the English words and to the pseudowords. It is implausible to suggest that
native English speakers are genetically programmed to develop an area in the
brain that will respond to the shapes of letters, but only when they are grouped
together to form English words or words that follow the rules of English spelling
(note that the activation was in visual areas, so the ndings are unlikely to be to
do with the pronounceability of the different types of letter strings). A much
better explanation of this nding would be that experience of a particular
language environment has inuenced the development of this specialist area of
cortical function. The fact that the location of this specialist area is shared by
native English speakers suggests that there are similar genetic and molecular
processes working in harmony with a similar language environment, to produce
a similar structural and functional outcome.
There are many more studies that provide evidence for the effect of experience
on cortical structure and function. One example is Nevilles (1991) work using
scalp recorded event-related potentials (ERPs) with congenitally deaf participants
(see Box 2). For these participants she found that regions of the temporal lobe
(see Figure 1, p. 118) which in typical development respond to auditory, or multimodal input (input from more than one of the senses) had become dominated by
responses to visual input. This suggests that the function of these regions had
been inuenced by the distinctive sensory experiences of the participants.
BOX 2
131
132
A further example, also from Neville and her colleagues (Mills et al., 1993) shows
how, with experience, certain types of processing are performed by progressively
smaller (more localized) regions of cortex. For instance, data from scalp-recorded
ERPs suggest that processing of known words and control stimuli is initially
spread over a relatively large area of cortex. This processing narrows to an area
over the left temporal lobe only when the childs vocabulary reaches about 200
words, irrespective of maturational age. We will look at further evidence relating
to whether modularization occurs as the brain develops when we discuss the
specic case of language in Section 4.
Comment
The area of the brain shown by Neville and colleagues (Mills et al., 1993) to process
some aspects of word recognition appears to change at a certain point in vocabulary
development, rather than at a certain age. How can we interpret these findings? If
this narrowing of function to the left temporal lobe happened for all children at a
similar age, it could be argued that this was due to an innate pre-programmed
specification of function that was simply unfolding with maturity. Indeed this could
still be the case if it happened at different ages, given that children mature at different
speeds. The fact that it happens at different ages for different children, and that it
relates to something so dependent on experience (vocabulary development)
suggests that the environment is playing a part and supports the modularization
position.
3.2 Self-organization
Section 3.1 provided some evidence to support the view that the environment
plays a part in the development of the structures and functions of the brain. If this
is the case, how might it happen? What processes might be involved in this gene
environment interaction that we are invoking? One answer lies with the concept
of self-organization.
Self-organization occurs when structure emerges in response to a systems
dynamic interactions with an environment. All stages of brain development
involve an element of self-organization (Keslo, 1995; Johnson, 1997). Initially the
neural system is relatively undifferentiated (randomly organized) but, as a result
of small adaptive changes, an order begins to emerge among the elements of the
system. In the brain, synaptic adjustment rules (such as the Hebb rule; see Box 3)
can lead to ordered connection patterns that in turn lead to structured behaviours.
In arguing that cognitive modules emerge as a result of development, KarmiloffSmith (1992) is proposing that the brain is a self-organizing system. The child
interacts with its environment, and as a result particular neural structures and
modules emerge. Self-organization is regarded by many researchers as a
fundamental characteristic of the brain (Changeux et al., 1984; von der Malsburg,
1995).
BOX 3
n
Self-organization occurs in systems with a large number of degrees of freedom;
that is, in complex systems where there are seemingly endless possibilities for
how things might turn out. In such systems it is difcult to see how all the
elements can be following a predetermined plan. For example, in the case of
brain development, the cerebral cortex alone contains some 100,000,000,000,000
(also written as 1014) synapses. It is unlikely that the genes can (in any direct way)
encode the full information necessary to generate this level of complexity
according to a predetermined plan (Elman et al., 1996). This is another argument
in favour of an epigenetic view of brain development, and in favour of KarmiloffSmiths argument that modularization happens as a product of development.
There are many well-studied examples of self-organization in the physical and
biological sciences. For example, in snowakes and in many other crystals, a
complex structure emerges from the apparently random addition of new material
(see Figure 6). Simple local rules of where new material can be added result in
intricate global patterns appearing in the nal crystal. Processes of selforganization can explain many complex patterns in nature, from the shape of
storm clouds to the spots on a leopard (Goldeld, 1995).
133
134
Figure 6 A snowake is a complicated, non-random shape that looks designed, yet it arises
from a very simple process; there is no design for a snowake built into water molecules. When an
ice crystal starts to form it initially takes a simple hexagonal shape. With the crystal suspended in
humid, cold air, water molecules join on more easily at the corners, starting to form a star. Alternate
melting and freezing then creates more corners where molecules join, leading to the crystal
becoming more and more complex.
Language
This section extends the argument from Section 3. It looks in more detail at
evidence relating to the question of whether specialist brain functions are
genetically predetermined, or the result of a more epigenetic system in which
genes and environment interact. We will focus on the case of language. Of all
human abilities, language has been regarded as the most biologically special
(Johnson, 1997, p. 137). It is in this domain that the strongest nativist case for
innateness of function has been made, notably by Chomsky (1965, for example).
Chomskys ideas have been further developed in the persuasive arguments of
Pinker (1994), which are the topic of this section of the chapter.
135
136
Pidgin
A make-shift
jargon ... choppy
strings of words
borrowed from
the language of
[others] ... highly
variable in order
and with little in
the way of
grammar (Pinker,
1994, p. 33).
Creole
A fully
grammatical
language
developed out of
a pidgin.
communities and of the non-native signers, are actually doing no more than every
child does they are reinventing language anew as they use it. He sees this as
compelling evidence for a language instinct:
... children actually reinvent [language] generation after generation not
because they are taught to, not because they are generally smart, not
because it is useful to them, but because they just cant help it.
(Pinker, 1994, p. 32; emphasis as in original)
The second element of Pinkers case is drawn from Chomskys own argument
from the poverty of the input (Pinker, 1994, p. 42). The core of this argument is
that children produce language that they can never have heard before, indicating
that they are not learning or imitating it, but generating it anew. Pinker gives the
example of how children are able to produce grammatically correct questions,
even though they will not have heard these questions asked before, and even
though the rule that they have to follow is comparatively complex, and one that
they are unlikely to have learned.
t
Activity 4
Allow about
10 minutes
This activity will help you to reect on Pinkers second piece of evidence for the innateness of
language the fact that children formulate sentences correctly, even when the rules they are using
are complex. Pinkers evidence is based on Chomskys classic illustrations of how children appear to
know the deep structure of their language.
1
Turn the following statement into a question: A unicorn is in the garden. Can you
devise a rule that you could follow to convert a statement like this into a question?
Comment
You should have come up with something like Is a unicorn in the garden? or Is there a
unicorn in the garden? The rule that you followed might have been take the rst is and
move it to the beginning of the sentence, and add a question mark. (Whether or not you
added there is not relevant to this activity.)
2
Now formulate a question from the following sentence: A unicorn that is eating a
ower is in the garden. Write your question down. Does your original rule still work?
Comment
Your original rule would produce the non-grammatical (in fact nonsensical) question Is a
unicorn that___ eating a ower is in the garden?. This is clearly not correct. This time you have
to take the second is and move it to get Is (there) a unicorn that is eating a ower in the
garden?.
137
138
that is eating the ower in the garden? (Pinker, 1994, pp. 412). According to the
nativists, children achieve the right grammatical form for the question because
they have an innate grasp of the deep structure of language and its units of
meaning.
The third element of Pinkers argument is related to the second. Pinker asserts
that many languages around the world use auxiliary verbs and that these
languages move the auxiliary to the beginning of the sentence to formulate a
question. (The is in the sentences in Activity 4 is an English auxiliary verb.) If you
think about it, there are limitless ways in which a question might be formulated.
Why not turn the sentence backwards, or exchange the rst and last words? So
why have all these languages hit upon the same way of doing this? It is as if
isolated inventors miraculously came up with identical standards for typewriter
keyboards or Morse code or trafc signals. (Pinker, 1994, p. 43). Pinkers answer
to this question is that it is not miraculous, and it is not a coincidence; rather, it
reects a commonality in the structure of the human brain.
The nal and fourth part of Pinkers argument brings us back to the specic
topic of this chapter; that is, to studies of brain and cognitive development. He
argues that there is an identiable seat [for language] in the brain, and perhaps
even a special set of genes that help wire it into place (Pinker, 1994, p. 45). We
will look in more detail at this claim, but now from the perspective of the wider
developmental cognitive neurosciences literature.
Hemisphere
The cerebral
cortex is divided
into two halves,
known as the left
and right
hemispheres.
One way of tackling the issue of innateness from the point of view of cognitive
neuroscience is to ask whether there are any areas of cortex that are critical for
language processing or acquisition. If there are such areas, without which
language simply cannot happen, then a working assumption would be that these
areas have an innate, language-specic processing capacity. If, on the other hand,
there are a variety of cortical regions that can support language acquisition, then
this would suggest that areas of language specialization emerge from more
general structures, connectivity and processes within the developing brain, in
concert with the rich language environment.
Sometimes this debate is framed in terms of the concept of equipotentiality.
This refers to the hypothesis that at birth the left and right hemispheres of the
brain both have equal potential for developing language. Evidence in favour of
equipotentiality would seriously undermine the nativist case for innate languagespecic areas of the cortex. Bear in mind that in typically developing adults, areas
of the left hemisphere are strongly associated with language processing, and
damage to those areas in adulthood can cause a variety of lasting language
difculties. There is no doubt that by adulthood there are specialist cortical
regions that are devoted to language processing (language modules). The
question is whether language processing must be carried out by these areas, or
whether early damage, or indeed atypical developmental pathways, (for example,
through deafness) can lead to other cortical regions performing the same
functions.
On balance, the evidence from neuroscientic studies points to the conclusion
that several cortical regions are capable of supporting the development of
language, but does not support the hypothesis of full equipotentiality. For
example, Neville et al. (1998) used the fMRI technique (see Box 2, Section 3.1) to
examine the brain regions that are involved in language processing in deaf and
hearing participants. When the deaf participants viewed American Sign Language
(ASL) sentences (see Figure 7b) they showed activation in language areas in the
left hemisphere, as did the hearing participants on reading English sentences (see
Figure 8a). However, the deaf participants (who were all native ASL signers) also
displayed a level of activation in the right hemisphere that was not observed for
the hearing participants (compare Figure 7b with Figure 8a).
Left hemisphere
Right hemisphere
Left hemisphere
Right hemisphere
t
Activity 5
Allow about
10 minutes
This activity encourages you to think about the implications of Neville et al.s pattern of ndings.
Look carefully at the ndings reported for Neville et al. (1998) (see Figures 7 and 8). You will
see that they found that both hearing and deaf participants in their study showed activation in
language areas of the left hemisphere when processing language stimuli in their native
language (ASL for deaf participants, written English for the hearing participants). In addition,
the level of right hemisphere activation was higher for the deaf participants than for the
hearing participants. What implications do these ndings have for the hypothesis of
139
140
Comment
Bearing in mind that these ndings are from just one study (and you should therefore exercise
caution in interpreting them) they indicate the following:
1
The language areas of the left hemisphere are prepared in some way to process languagerelated information, irrespective of the form of the language. Both signed and written
sentences were processed by the classic language areas of the left hemisphere.
Some areas of the right hemisphere had been captured for language processing in the deaf
participants, perhaps because of the role of visuo-spatial information in sign languages.
Both these ndings suggest some kind of bias towards language processing in areas of the left
hemisphere, but also that other areas in the right hemisphere are capable of supporting
language-related information processing. Note also the right hemisphere activation for deaf
participants when reading English sentences (Figure 8b).
Semantic
To do with
meaning.
Neville and her colleagues (Neville et al., 1992) also noted different patterns of
cortical response (as measured by ERPs) in hearing and deaf participants whilst
reading grammatically signicant words in sentences, but noted similar patterns
of response across the two groups when they were reading semantically
important words. So perhaps the question of how biologically special language is
may require different answers for different aspects of language processing and
acquisition. Neville et al.s ndings indicate that grammatical aspects of language
may be more sensitive to experience than semantic aspects, because hearing and
deaf participants (who have a different experience of language) responded in
different ways to the grammatical information. After summarizing a decade of
their research, Neville and Bavelier conclude that the evidence supports the
hypothesis that there are constraints on the organisation of the neural systems
that mediate formal language ... however, it is clear that the nature and timing of
sensory and language experience signicantly affect the development of the
language system of the brain (Neville and Bavelier, 2001, p. 283).
Further evidence to support the hypothesis that different areas of the cortex are
capable of supporting language processing comes from studies of children with
localized brain damage (focal lesions) that happened either before or during
birth. In a sample of children aged between 3 and 9 years, Reilly and colleagues
(Reilly et al., 1998) found that the group of participants with a focal lesion
performed worse on a series of language tasks than the group of typically
developing control participants (see Research summary 1). However, the children
with a focal lesion showed a pattern of catching up on these measures, then
lagging behind at the next stage of language development, then catching up
again. The crucial points here are: rst, functional recovery from a focal lesion
appears to be an ongoing process in childhood, and second, there is an
implication that functions affected by the original damage to a localized area of
the cortex were taken over in later development by undamaged areas of the
cortex.
RESEARCH SUMMARY 1
In study children were asked to tell a story from a 24-page picture book about a boy, a
frog and a dog. The researchers aimed to examine the effects of focal lesions which
happen before 6 months of age on childrens subsequent narrative language skills. The
study involved 62 children, of which thirteen had right-hemisphere damage (RHD),
eighteen had left-hemisphere damage (LHD) and 31 were typically developing
(neurologically intact) control participants. All the RHD and LHD children had focal
lesions in one hemisphere only, which had occurred before 6 months of age. The age
range of participants was 3 years and 5 months (3;5) to 9 years and 4 months (9;4), with
a mean age of just over 6 years. The participants were also subdivided into two age
levels: younger than 5 years (seven LHD, four RHD and eleven controls) and older than
5 years (eleven LHD, nine RHD and 20 controls). This gave the researchers an
experimental design that allowed them to test hypotheses relating to age (comparing
the younger group with the older group), neurological status (comparing the children
with focal lesions with the neurologically intact group), lateralization (comparing RHD
with LHD), and interactions among all these factors. This is known as a factorial design.
1.50
1.25
Control
LHD
RHD
Regression
lines
1.00
0.75
LHD
0.50
RHD
0.25
Control
0.00
2
10
Age in years
The researchers collected both quantitative and qualitative data. The quantitative data
involved systematic measurements of various features of the narrative account of the
story produced by each child.
The researchers found that the children with brain injury showed some delays in
grammatical and narrative skills when compared to their control group. But the data
also showed clear evidence of change and development on all fronts (Reilly et al., 1998,
p. 358). For example, Figure 9 plots a measure of grammatical errors (morphological
errors) on the vertical axis against age on the horizontal axis. Each child is represented
by a marker within the graph. The higher up the marker is, the more errors the child
made. The regression lines (or lines of best t) for the three groups, all of which slope
downwards from left to right, show that in all groups (LHD, RHD and control) older
141
142
children did, on average, better than younger children. Note how the lines converge
with age; at 4 years of age there is quite a gap between the performance of the control
group, on the one hand, and the LHD and RHD groups on the other. By 6 years of age
this gap in performance had narrowed, and by 9 years of age it had narrowed even
further.
The qualitative data were used to analyse the nature of any grammatical errors. From
this analysis the researchers were able to conclude that for the brain injured group the
raw data suggest ... a delay in morphological development, but there is no evidence here
for a qualitative deviation from the normal pattern (Reilly et al., 1998, p. 360). The
younger children in the LHD and RHD groups made signicantly more errors than their
age-matched controls, but the errors that they made were normal errors.
n
Like Reilly and colleagues (1998), Stiles and Thal (1993) also studied children with
focal lesions that had occurred either to the left or right hemisphere before 6
months of age. The children were examined longitudinally, and were found to
have language delays irrespective of which hemisphere had been damaged.
Surprisingly, though, the most signicant delays in word comprehension were
found among the right hemisphere lesion group. In adults one would expect to
see this kind of problem when an area of the left hemisphere becomes damaged.
This suggests that the regions responsible for language acquisition are not
necessarily the same as those used for language processing in the adult.
The evidence presented in this section does not in itself rule out the
proposition that the left hemisphere is innately pre-wired for language. It only
shows that other areas of the brain can do language, if necessary. However, there
is a proposition that ts the evidence a little more securely than this. This is that
the architecture and connectivity of the left temporal lobe (the region most clearly
related to language processing in adults) biases the system towards building
neural modules that specialize in language processing in these regions. These
regions may be slightly better suited to processing rapidly changing information
than other areas of the cortex, and so can develop more efcient computational
properties for processing language. However, the regions are not the only place
where language processing can nd a home. There is sufcient plasticity in the
developing childs cortex to enable other regions to take on these functions,
perhaps with some loss of efciency. While regions of the left temporal lobe may
be best suited to language processing, they are not critical since language can
develop in a close to normal way without [them] (Johnson, 1997, p. 141). Overall,
therefore, the evidence from developmental cognitive neuroscience supports a
compromise position somewhere between equipotentiality on the one hand, and
genetically predetermined language-specic cortical regions on the other.
143
144
t
Activity 6
Allow about
30 minutes
This activity encourages you to review the debate about the language instinct, and to weigh up the
evidence for and against the nativist approach.
Review the nativist arguments of Chomsky and Pinker in Section 4.1 in the light of the
evidence presented in Section 4.2, together with the evidence presented in Section 3. Does
the evidence support or refute their position? As usual, the answer is probably not clear cut.
To help you carry out your review, construct a list of points in favour of the nativist stance,
and points against it. Structure your list as shown in Figure 10, remembering to cite evidence
wherever possible (we have provided some examples to get you started).
Universality of language
.
.
In typically developing adults, areas in the left temporal lobe deal with
many language-specic functions. Damage to these areas in adulthood
can seriously disrupt key language functions.
It is likely that the left temporal lobe becomes the seat of language
because of slight advantages in its capacity for processing complex,
rapidly changing information, rather than because it has an innate
capacity to process language.
Broadly speaking, the neuroscientic evidence supports an epigenetic
view of the development of language-specic functions in the adult
brain.
In this section we leave behind arguments about the origin of cognitive modules,
and turn more directly to the question of how the development of brain structure
can be related to the development of cognition and behaviour in infants and
children. We focus on the case of the prefrontal cortex for the reasons outlined in
Section 5.1. Thereafter, Sections 5.2 and 5.3 provide two competing views of how
structural developments in this area of the brain can be related to advances in
cognitive ability.
Working
memory
The part of the
memory system
that deals with the
active
manipulation of
information (as
opposed to longterm memory
which is more to
do with storage
of information).
The region of the frontal lobe in front of (anterior to) the primary motor and
premotor cortex, the prefrontal cortex (see Figure 1), accounts for almost onethird of the total cortical surface in humans and is considered by most
investigators to be critical for many higher cognitive abilities (Milner, 1982;
Goldman-Rakic, 1987; Fuster, 1989). The types of cognitive processing that have
been associated with the prefrontal cortex concern the planning and execution of
sequences of action, the maintenance of information over short delays in working
memory, and the ability to inhibit a set of responses that are appropriate in one
context but not another. Collectively these cognitive processes are known as
executive function. This aspect of cognitive function will receive more detailed
treatment in Chapter 5 of this book.
The prefrontal cortex is the part of the brain that shows the most prolonged
period of postnatal development, with changes in synaptic density detectable
even into the teenage years (Huttenlocher, 1990). For this reason it has been the
part of the brain most frequently associated with cognitive development.
145
146
Phenylketonuria
An inborn
condition which
affects the infants
ability to process
a protein
(phenylalanine)
and which can
lead to severe
learning
difculties if
untreated by a
dietary
intervention.
Neurotransmitters
Chemicals that are
released by
neurons and
which affect the
transmission of
information
between neurons.
in EEG responses recorded from over the prefrontal cortex were correlated with
the ability to respond successfully over longer delays in delayed response object
permanence tasks; the more activation there was in the prefrontal cortex, the
longer the infants response could be delayed and still be successful.
A second source of evidence is work on cognitive development in children
with phenylketonuria (PKU). Even when treated, this inborn condition can cause
a reduction in the levels of a neurotransmitter, dopamine, in the prefrontal cortex.
Neurotransmitters are chemicals that enable neurons to communicate with each
other across their synapses, so a reduction in the level of a neurotransmitter will
affect the efciency of neural connections. In these cases, reductions in dopamine
levels in the DLPC result in impairments on tasks thought to involve parts of the
prefrontal cortex such as object permanence and object retrieval tasks. The
affected infants show no impairment on tasks thought to depend on other regions
of cortex.
147
148
RESEARCH SUMMARY 2
Milliseconds
There are 1000
milliseconds (ms)
in a second.
When an adult is looking at something, and a brief visual stimulus is displayed elsewhere
in their eld of view, their attention is drawn momentarily to that stimulus. If, within
about 150 milliseconds (ms) (nearly one seventh of a second), this stimulus is shown
again in the same location, together with a competing visual stimulus elsewhere in their
eld of view, the adult will look at it more readily than they will at the competing
stimulus. If there is a gap of around 700 ms before the second presentation, the adult
will be less likely to look at it than at the competing stimulus. The rst effect is called
facilitation of detection (because the rst showing of the stimulus helps detection of
the same stimulus the second time round). The second effect is called inhibition of
return (the rst showing seems to inhibit attention to the stimulus when it is shown for
a second time). Both effects have been well documented in studies of attention shifting.
By about 7 months of age, typically developing infants also display these patterns of
attention shifting (Johnson and Tucker, 1996).
Johnson et al. set out to discover what areas of the brain are involved in these patterns
of attention shifting. They studied seven, 7-month-old infants who had sustained
localized damage to an area of their brain during or shortly before birth. Three of these
infants had damage on the left-hand side of the prefrontal cortex, and four had damage
in other regions of the cortex.
The infants were seated in front of three computer screens, positioned side by side. On
the middle screen a visually interesting pattern was played which engaged their
attention. At a given point in time a stimulus was ashed onto either the right-hand or
left-hand screen. After a set interval (133, 200, 700 or 1200 ms) the same stimulus
would be ashed onto both side-screens. Each infant experienced 32 trials, eight at each
of the four different intervals. Video recordings of the infants looking behaviour were
made from a camera located above the middle screen, and a record was made of which
way the infant looked on each occasion that the double stimulus was displayed.
The researchers found that the infants without left side damage to the prefrontal
cortex showed similar patterns of attention shifting behaviour to the typically
developing infants from the Johnson and Tucker study (1996). The infants with left side
damage, on the other hand, showed no sign of the facilitation of detection effect for
the 133 ms interval. This suggests that the prefrontal cortex plays a role in visual
attention tasks in infancy. In adults only those with damage to the parietal lobe (see
Figure 1, p. 118) show such a pattern of results.
n
The ndings from the Johnson et al. (1998) study, and from Csibra et al. (1998)
are consistent with the view that the cortical regions that are crucial for a
particular ability change with the stage of acquisition of that ability. Perhaps there
is greater prefrontal activity on visual attention tasks in infancy than in adulthood
because infants are still at an early stage of developing visual attention skills. By
adulthood the skills have been learned and have moved to a different, more
specialized area of the cortex.
t
Activity 8
Allow about
15 minutes
This activity will help you to review the material presented in Section 5 of this chapter.
Summarize the evidence that supports each of the two perspectives presented in Section 5
on prefrontal cortex development in infancy. The two perspectives are:
1
Various aspects of the cognitive and behavioural development of infants can be related
to structural changes in the prefrontal cortex.
The prefrontal cortex is specically associated with the early stages of the development
of a new skill, and its involvement decreases with the development of that skill.
5.4
Section 5 has been about efforts to bridge the conceptual gap between brain
structure (how the different regions of the brain are physically organized) and
function (what these regions do). This is a major challenge for developmental
cognitive neuroscience. Much is known about the brain at a structural level. Much
is also known about behaviour and cognition. But there is still a great deal to
learn about how structures underpin behaviours and cognitive functions, and
about how behaviours and cognitive functions inuence the development of
structures. One tool for investigating the relationship between structure and
function in this respect is connectionist computational modelling (neural
networks). One aim of connectionist modelling is to see how the brain structures
that we know about could possibly give rise to human behaviour and cognition.
You will encounter more about connectionist modelling in Chapter 4.
149
150
Conclusion
Further reading
Elman, J. L., Bates, E. A., Johnson, M. H. et al. (1996) Rethinking Innateness: a
connectionist perspective on development, Cambridge, MA, MIT Press.
Johnson, M. H. (1997) Developmental Cognitive Neuroscience, Oxford, Blackwell.
Nelson, C. A. and Luciana, M. (2001) Handbook of Developmental Cognitive
Neuroscience, Cambridge, MA, MIT Press.
References
Bell, M. A. (1992) A not B task performance is related to frontal EEG asymmetry
regardless of locomotor experience, in Rovee-Collier, C. (ed.) Proceedings of the
VIIIth International Conference on Infant Studies 15 (Special ICIS Issue), p. 307,
Miami Beach, FL: Infant Behavior and Development.
Bell, M. A. and Fox, N. A. (1992) The relations between frontal brain electrical
activity and cognitive development during infancy, Child Development, vol. 63,
pp. 114263.
Casey, B. J., Trainor, R. J., Orendi, J. L. et al. (1997) A developmental functional
MRI study of prefrontal activation during performance of a go-no-go task,
Journal of Cognitive Neuroscience, vol. 9, pp. 83547.
Conel, J. L. (193963) The Postnatal Development of the Human Cerebral Cortex,
Vols IVI, Cambridge MA, Harvard University Press.
Changeux, J.-P. (1985) Neuronal Man: the biology of mind, New York, Pantheon.
Changeux, J.-P., Heidman, T. and Patte, P. (1984) Learning by selection, in
Marler, P. and Terrace, H. S. (eds) The Biology of Learning, pp. 11533, Berlin,
Springer-Verlag.
Chomsky, N. (1965) Aspects of the Theory of Syntax, Cambridge, MA, MIT Press.
Cowan, M. W., Fawcett, J. W., OLeary, D. M. and Staneld, B. B. (1984)
Regressive events in neurogenesis, Science, vol. 225, pp. 125865.
Csibra, G., Tucker, L. A. and Johnson, M. H. (1998) Neural correlates of saccade
planning in infants: a high-density ERP study, International Journal of
Psychophysiology, vol. 29, pp. 20115.
DeCasper, A. J. and Spence, M. J. (1986) Prenatal maternal speech inuences
newborns perception of speech sounds, Infant Behavior and Development,
vol. 9, pp. 13350.
DeCasper, A. J., Lecanuet, J-P., Busnel, M-C., Granier-Deferre, C. and Maugeais, R.
(1994) Fetal reactions to recurrent maternal speech, Infant Behavior and
Development, vol. 17, pp. 15964.
Diamond, A. (1985) Development of the ability to use recall to guide action, as
indicated by infants performance on AB, Child Development, vol. 56, pp. 86883.
Diamond, A. and Goldman-Rakic, P. S. (1986) Comparative development of
human infants and infant rhesus monkeys of cognitive functions that depend on
prefrontal cortex, Neuroscience Abstracts, vol. 12, p. 274.
Diamond, A. and Goldman-Rakic, P. S. (1989) Comparison of human infants and
infant rhesus monkeys on Piagets AB task: evidence for dependence on
dorsolateral prefrontal cortex, Experimental Brain Research, vol. 74, pp. 2440.
Elman, J. L., Bates, E. A., Johnson, M. H., Karmiloff-Smith, A., Parisi, D. and
Plunkett, K. (1996) Rethinking innateness: a connectionist perspective on
development, Cambridge, MA, MIT Press.
Fodor, J. A. (1983) The Modularity of Mind: an essay on faculty psychology,
Cambridge, MA, MIT Press.
Fox, N. A. and Bell, M. A. (1990) Electrophysiological indices of frontal lobe
development, in Diamond, A. (ed.) The Development and Neural Bases of Higher
Cognitive Functions, pp. 67798, New York, New York Academy of Sciences.
Fuster, J. M. (1989, 2nd edn) The Prefrontal Cortex, New York, Raven Press.
Goldeld, E. C. (1995) Emergent Forms: origins and early development of human
actions and perception, Oxford, Oxford University Press.
Goldman-Rakic, P. S. (1987) Development of cortical circuitry and cognitive
function, Child Development, vol. 58, pp. 60122.
Hebb, D. O. (1949) The Organisation of Behavior: a neuropsychological
approach, New York, Wiley.
Huttenlocher, P. R. (1990) Morphometric study of human cerebral cortex
development, Neuropsychologia, vol. 28, pp. 51727.
Johnson, M. H. (1997) Developmental Cognitive Neuroscience, Oxford, Blackwell.
Johnson, M. H. and Tucker, L. A. (1996) The development and temporal
dynamics of spatial orienting in children, Journal of Experimental Child
Psychology, vol. 63, pp. 17188.
151
152
Johnson, M. H., Tucker, L. A., Stiles, J. and Trauner, D. (1998) Visual attention in
infants with perinatal brain damage: evidence of the importance of anterior
lesions, Developmental Science, vol. 1, pp. 538.
Karmiloff-Smith, A. (1992) Beyond Modularity: a developmental perspective on
cognitive science, Cambridge, MA, MIT Press.
Katz, L. C. and Shatz, C. J. (1996) Synaptic activity and the construction of cortical
circuits, Science, vol. 274, p. 113338.
Keslo, S. (1995) Dynamic Pattern: the self-organization of brain and behavior,
Cambridge, MA, MIT Press.
Kwong, K. E., Belliveau, J. W., Chesler, D. A. et al. (1992) Dynamic magnetic
resonance imaging of human brain activity during primary sensory stimulation,
Proceedings of the National Academy of Sciences, vol. 89, pp. 56759.
Lewkowicz, D. J. and Turkewitz, G. (1981) Intersensory interaction in newborns:
modication of visual preferences following exposure to sound, Child
Development, vol. 52, pp. 82732.
Maurer, D. (1993) Neonatal synesthesia: implications for the processing of speech
and faces, in de Boysson-Bardies, B., de Schonen, S., Jusczyk, P. et al. (eds)
Developmental Neurocognition: speech and face processing in the rst year of life,
pp. 10924, the Netherlands, Kluwer.
Mehler, J. and Dupoux, E. (1994) What Infants Know, Oxford, Blackwell.
Meltzoff, A. N. and Borton, R. W. (1979) Intermodal matching by human
neonates, Nature, vol. 282, pp. 4034.
Mills, D. M., Coffey, S. A. and Neville, H. J. (1993) Language acquisition and
cerebral specialization in 20-month-old infants, Journal of Cognitive
Neuroscience, vol. 5, pp. 32642.
Milner, B. (1982) Some cognitive effects of frontal-lobe lesions in man,
Philosophical Transactions of the Royal Society of London, vol. 298, pp. 21126.
Nelson, C. A. and Bloom, F. E. (1997) Child development and neuroscience,
Child Development, vol. 68, pp. 97087.
Neville, H. J. (1991) Neurobiology of cognitive and language processing: effects
of early experience, in Gibson K. R. and Peterson A. C. (eds) Brain Maturation
and Cognitive Development: comparative and cross-cultural perspectives,
pp. 35580, New York, Aldine de Gruyter.
Neville, H. J. and Bavelier, D. (2001) Variability of developmental plasticity, in
McClelland J. L. and Siegler, R. S. (eds) Mechanisms of Cognitive Development:
behavioral and neural perspectives, pp. 27187, London, Lawrence Erlbaum
Associates.
Neville, H. J., Mills, D. and Lawson, D. (1992) Fractionating language: different
neural subsystems with different sensory periods, Cerebral Cortex, vol. 2,
pp. 24458.
Neville, H. J., Bavelier, D., Corina, D. et al. (1998) Cerebral organisation for
language in deaf and hearing subjects: biological constraints and effects of
153
154
Readings
Reading A: The fragmented mind from
What Infants Know
Jacques Mehler and Emmanuel Dupoux
Like vision, language also appears to us to be a whole. It even seems
indissociable from the other higher faculties. But we have known for more than
a century that certain cortical lesions localized particularly in the left hemisphere
can have specic effects on linguistic capacity. Some patients lose the ability to
understand language, while others lose the ability to produce sentences. Others,
more seriously affected, lose both abilities. Neuropsychologists have been trying
to link these symptoms with the affected areas of the cortex since the nineteenth
century. Thus it was believed that anterior lesions in particular caused problems
with production of speech, Brocas aphasia, and that lesions of the posterior parts
of the cortex tended to create difculties in language comprehension, Wernickes
aphasia. Today, thanks to a better description of these symptoms and
deciencies, this gross classication has proved to be of limited interest. On the
other hand, a more specic description has permitted us to show how traumas
can cause a loss that remains limited to a very specic component of the linguistic
system. Thus, some patients experience difculty identifying words, while others
have problems with grammar, reading, or writing. Without enumerating all the
different forms of aphasia, it is possible to say that clinical studies have facilitated
the relatively precise description of the various stages of linguistic data
processing. They seem to be specic and modular.
As Broca discovered, certain cerebral traumas can cause loss of language without
affecting other faculties. Aphasics suffering from a serious lesion in the left
hemisphere thus can still reason and calculate. However, one might think that
since both language and music are auditory, patients who have a hard time
understanding what is said to them would also have a problem recognizing a
piece of music. But this is not true, as several celebrated cases demonstrate.
At the age of seventy-seven, a well-known composer and organist, who was
blind, suffered a stroke. A cerebral lesion of the left hemisphere left him aphasic.
He could no longer speak words or utter sentences. And yet his musical abilities
were unaffected: he could still decipher scores by ear. But above all, and even
more astonishingly, while he could read music in braille, the accident had left him
almost unable to read braille texts. This discrepancy was all the more surprising
since the symbols representing notes and letters are identical in braille. The
nature of the system used was therefore not a factor. Musical ability would then
be a function of the right hemisphere, while linguistic competence would be
handled by the left one.
A, a musical score; B, the same score in braille; C, a transcription of the braille score with the letters
used. The patient can play the score presented in braille, but has great difculty reading E, which
represents ... le pe`re [the father]. He reads: Le ... Pa ... le ... ta ... le fre`re [the brother] ... la ... le par
... il y a [there is] pe. (Signoret et al., 1987).
The example of another patient, also a professional musician, bears out this
theory. It concerns a right-handed conductor who had led the La Fenice orchestra
in Venice and the La Scala orchestra in Milan for many years. At sixty-seven, a
stroke injuring his left cerebral hemisphere caused the loss of all linguistic
aptitude, and he was unable to speak for the next six years. On the other hand,
his musical capacities remained almost intact. He was incapable of naming notes,
but could sing or play them. He made such rapid progress that, despite his
inability to communicate with an orchestra except by gestures, he continued to be
a completely professional conductor capable of detecting the most subtle musical
errors. His interpretation of Verdis Nabucco was especially appreciated by critics.
The lesion resulting from his rst stroke therefore affected only his left
hemisphere, the seat of language. Six years later, a second vascular accident partly
destroyed his right hemisphere. Unfortunately, this time he did not recover and
his death prevented our testing his musical capacities. Other cases of lesions
suppressing the aptitude for music conrm, however, that musical ability has its
seat in the right hemisphere of the brain.
As indicated above, commissurotomy cases illustrate the dissociation that exists
between language and other mental functions, for in these patients we can
present data to one hemisphere only. For example, we can arrange for the image
of a word to reach the left hemisphere only. Subjects can read it and even point to
the corresponding object. But this is possible only if they use their right hand.
This hand is directly linked to the left hemisphere, in other words, to the one that
saw the word.
If, on the contrary, we present a word to the opposite hemisphere, subjects are
unable to read it, and even claim not to have seen anything. However, they can
locate the object, with their left hand. In these cases, the left hemisphere, which is
responsible for language, has not seen the word presented, and subjects are
therefore unable to formulate a verbal response. To summarize, if we analyze the
subjects behavior in terms of what they perceive and produce with their right
hemispheres, we conclude that they have a very diminished linguistic capacity
similar to that of certain aphasics. Yet if we analyze their behavior in terms of
what they control with their left hemisphere, we conclude that they are in full
possession of their linguistic capacities.
155
156
How is this possible? How can one both be and not be aphasic? In fact, this
discrepancy between mental faculties in the same person is only partial and
manifests itself only in very specic experimental situations. Under normal
conditions, both hemispheres function in a coherent fashion, for they have many
means of communicating with another. For example, eye movements permit the
duplication of visual data in both hemispheres, and information can also circulate
through the intermediary of motor activity. For instance, we have observed a
subject give an incorrect verbal response, raise his eyebrows, and then
immediately correct himself. No doubt the right hemisphere, which knew the
correct answer, prompted the raised eyebrows, which informed the left
hemisphere of its error and led him to change his response. It is also possible that
such phenomena sometimes occur in healthy individuals.
Be that as it may, these surprising observations clearly show that it is
counterproductive to see our mental faculties as an inseparable whole. In the light
of neuropsychological studies, our psychological mechanism seems rather to be
composed of independent and autonomous faculties like the perception of faces
and of language. And language itself is not all of one piece either. Very specic
afictions illustrate this. Some patients, for example, lose the use of conjunctions
but not of nouns and adjectives, or even the use of a whole semantic group like
fruits and vegetables, or trees. An apparently simple and homogeneous task
therefore calls upon many distinct and specialized submechanisms, which can
break down while the rest of the system continues to function normally.
Commonsense psychology, as we clearly see, cannot explain this type of
phenomenon. Nor is the perspicacity forged by years of experience of any use.
We therefore must assume that a part of our cognitive system is divided into
functional units, or modules, which are responsible for a given aptitude and
operate in an autonomous fashion oblivious to what is happening elsewhere in
the system. The more research advances, the less our intelligence seems to
resemble a whole with indistinct equipotential parts, rather like a bowl of jello,
and the more it seems to be subdivided into a great number of functions which
have a certain autonomy. Galls intuition has proven to be right. As we already
mentioned, it was taken up again in more modern form by Jerry Fodor.
According to Fodors formulation, we can compare a function, or module, to a
physical organ. Our body is made up organs which have their specic functions
and which, through their interaction, contribute to the functioning of the whole.
However, modules do not exchange uids or energy, but information. Thus each
module constitutes an abstract organ which we could describe, at least in a rough
approximation, in terms of concepts borrowed from computer science
(representation, structure of data, manipulation of symbols, etc.). In a sense, then,
a module is closer to a computer program than to a heart or liver. However, like
its physiological counterpart, it has a specic function: it only processes a
fragment of the information circulating through the whole psychological
mechanism and can only use rather restricted, predetermined channels of
communication. Conscientious, but rather limited, it is an expert in its eld.
Furthermore, just as in the case of physical organs, the growth and organization
of cognitive modules is guided by genetic program specic to the species.
And nally, as the example of language shows, a module is not distributed
throughout the entire brain, but involves a specic nervous structure, circuit, and/
or cortical area.
From this perspective, we can no longer imagine that our psyche is controlled by
a central intelligence. The data used by each module are limited and its eld of
operation is restricted. A so-called superior intelligence is no longer necessary. It
is the system as a whole that exhibits a behavior we can term intelligent, not its
parts taken separately, and not just one part in particular.
How many modules are there? How are they organized? According to one of
Fodors hypotheses, there are many specialized systems which are consolidated
into large modules, each specialized in the rapid and automatic processing of one
type of data (e.g., language). Each module delivers information to a central
processing unit which compares the different entries as well as all the other
knowledge available to the organism, making it possible to elaborate the longterm planning of actions.
It is quite obvious that we have not yet been able to explore this extremely
complex modular architecture in detail. We are still taking the rst baby steps
in a discipline which is centuries behind the natural or exact sciences. This
functionalist representation of the psychological mechanism is above all a
guideline: it enables us to formulate hypotheses and submit them to experimental
verication. The actual existence of a particular system or a particular exchange of
data is not a metaphysical postulate. The beauty and perfection of the theory
matter less than its ability to raise questions that experiments can then proceed to
answer.
157
158
159
160
161
Chapter 4
understanding of grammar
Contents
Learning outcomes
165
Introduction
165
166
167
2.2 Morphology
168
2.3 Syntax
2.4 Learning morpho-syntax
170
171
173
173
174
174
179
word endings
4.1 Dual route theory
182
182
183
185
developmental disorders
188
189
5.1 What, where, why and who: asking questions about syntax 192
194
Conclusion
197
Further reading
198
References
198
Reading
Reading A: An extract from The Articulate Mammal
201
201
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 describe the key stages in childrens grammatical development;
2 understand Chomskys theory of Universal Grammar;
3 show familiarity with methodological approaches to studying childrens
grammatical development;
4 offer a critical account of single and dual route explanations of grammatical
development;
5 compare and contrast nativist and empiricist accounts of grammatical
development.
Nativists
Theorists who
argue that
development is
primarily driven
by innate (inborn)
constraints.
Empiricists
Theorists who
argue that
development is
primarily driven
by factors in the
environment.
Introduction
If you have ever tried to learn a foreign language as a teenager or an adult, you
will know that one of the most complex aspects is understanding and
remembering its grammatical rules. Learning vocabulary is a relatively simple
affair, but knowing how to combine those words to form legitimate sentences that
are an accurate representation of what you want to communicate is a more
sophisticated skill. Moreover, the fact that languages can differ greatly from each
other in their grammatical rules further emphasizes the complexity of the task. So,
consider how much more difcult it would be if you were never explicitly taught
what these rules were, and were left to discover them for yourself. Think also
about how difcult this task would be if you were not already uent in a language
of your own. And yet, this is exactly the position that infants are in during the rst
few years of life.
In Chapter 2, First words, you read about how children come to acquire their
spoken vocabulary. In this chapter the focus is on how children achieve the
apparently insurmountable task of acquiring grammatical understanding. We will
consider the nature of spoken language and how it is constructed before looking
at methodological approaches to this area of research. We will then examine two
approaches to explaining how children acquire knowledge about the grammar of
their rst language. The two approaches reect and further develop the longstanding debate over the fundamental source of human knowledge, be it
linguistic or non-linguistic, between nativists and empiricists.
What you will also notice while reading this chapter is how childrens
grammatical development further illustrates an underlying theme of this book:
that cognitive development is characterized by a progression from a piecemeal
knowledge of the environment to an organized, systematic understanding that
imposes structure on what is experienced.
165
166
Before engaging with the question of how young children acquire grammatical
understanding, it is rst necessary to appreciate the complexity and the
signicance of spoken language as a structure. When people speak, they are
producing and combining sounds to communicate their ideas. The capacity to
combine sounds in different ways enables the speaker to communicate different
ideas. The capacity to combine sounds in novel ways enables the language user
to communicate ideas that have never been expressed before. So, the creative use
of language is entirely dependent on the ability to assemble simple building
blocks of sound into the complex structures we call sentences. Speakers of a
language appreciate the organization of their systems of linguistic combination.
They recognize when a sentence is spoken incorrectly, say by a child or a
foreigner. Speakers can even classify the type of mistake in an erroneous
utterance. For example, they may comment that it was not correctly pronounced,
that it contained an unknown word, or that it was grammatically incorrect. See
Box 1 for a reminder of some grammatical terms.
In learning to combine sounds in a way that is appropriate, children develop a
largely implicit understanding of phonology, morphology and syntax (see
Figure 1).
Spoken language
Phonology
The structure
of speech sounds
Grammar
Morphology
The structure
of words
Syntax
The structure
of sentences
BOX 1
sentence The child is riding the bike, the nouns are child and bike.
Auxiliary verb: a verb that precedes the main verb in a sentence in order to
complete the main verb. For example, the verb to be is the auxiliary verb in our
Subject: the person or thing that is performing the action in the sentence. For
example, in The child is riding a bike, it is the child who is doing the riding, and is
Object: the person or thing that is the recipient of the action in the sentence. In our
Active and passive sentences: The child is riding the bike is an active sentence, as
the verb is being directly executed by its subject. The passive version of the sentence is
The bike was ridden by the child. Note the change in wording: the object now appears
rst, and to accommodate this change we now have to use an auxiliary verb and the
word by has also been added (the presence of these factors indicates a passive
statement). The object of the sentence (the bike) now has prominence over the subject
(the child).
2.1 Phonology
Phonology and phonotactics are terms that are used to refer to the knowledge that
a speaker possesses about the sound patterns of his or her native language. For
example, in English, speakers readily distinguish sounds like [b] as in bat from
sounds like [p] as in pat, whereas Arabic speakers ignore this distinction.
t
Activity 1
Phonological distinctions
Allow about
5 minutes
This activity will help you appreciate the subtlety of the phonological distinctions that are made in
understanding language.
Say pig and big out loud, paying attention to the very beginning of each word, and what
your mouth and throat have to do to produce these sounds. In what way are they the same,
and what is the crucial difference?
Comment
For both words, the mouth, lips and tongue are in the same position. The main difference is
that the /p/ in pig is voiceless (i.e. the vocal chords do not move as the initial sound is made)
but for the /b/ in big voicing starts sooner. If you put your hand on your throat while saying
just the words you will notice this more easily. Other sounds, such as [t] and [d], and also [k]
and [g], differ in the same way. Although the difference between these two sounds is subtle,
the impact that it has on the meaning of an utterance is huge (e.g. big and pig simply differ in
whether the rst brief sound in the word is voiced or voiceless, but they have very different
meanings!).
167
168
Aspirated
These are
phonemes that
are accompanied
by a puff of air as
they are spoken.
Similarly, other languages such as Thai, recognize other subtle distinctions which
English speakers ignore; the aspirated [ph], as in pit, versus the unaspirated [p] as
in spit, are treated as the same sound by English speakers, but signal a difference
in meaning to speakers of Thai. To notice the difference between the aspirated
and unaspirated [p], try saying pit while holding your hand in front of your face;
you should feel a burst of air on your hand. Now do the same as you say spit.
You will notice that the burst of air is absent, or at least a great deal weaker.
Likewise, there are combinations of sounds that English speakers would regard
as foreign. For example, the /p/ sound cannot be followed by the /s/ sound at
the beginning of a word in English but it is permitted in French. So, in the English
word psychology, the letter p is silent, but in the French word psychologie the
/p/ and the /s/ sounds are both pronounced. Speakers who are not careful in
articulating these subtle cues, or listeners who do not pay attention to them, are
likely to be misunderstood or to misunderstand the communications of others.
Indeed, as you saw in Chapter 2, much of language development during the rst
year of life is concerned with mastering the mechanics of speech production, or
deciphering the sound patterns of the language that dominate the infants acoustic
environment.
2.2 Morphology
Grammatical
morpheme
A morpheme is a
unit of speech that
has meaning. A
grammatical
morpheme is a
unit of speech that
modies the
meaning of the
word to which it
is added, such as
ed when added
to the end of a
word like kick.
English-speaking children begin to use plurals early in their third year. Most
children will also have used some past tense verbs by about 30 months of age
(see Table 1, Section 3.3). In English, plural and past tense forms are marked
either by a process of adding an ending (dog ? dogs, walk ? walked) or through
some change of the word stem (man ? me n, see ? sa w). The process whereby
an ending is added is known as sufxation. This is the regular (typical) way of
forming plurals and past tenses in English. Other changes are called irregular
(though do note that some nouns, for example deer, and verbs, for example hit,
do not change their form at all in the plural or the past tense). Although a wider
range of nouns and verbs are inected through sufxation, irregular plurals and
past tenses are commonly used in everyday adult language. Indeed, irregular
plural and past tense forms are amongst the earliest acquired by young children.
Again, it can be inferred that children have a developing sensitivity to
inections from their early utterances. That is, initially children produce plural
and past tense verb forms perfectly. However, towards the end of their third year
they start to make errors like these:
My teacher holded the baby rabbits and we patted them.
I love cut-upped egg.
(Pinker, 1995, p. 109)
Instances where children incorrectly add /s/ to the end of a word to indicate
plurality (e.g. sheeps), or apply the add /ed/ rule for indicating that something
has happened in the past are known as over-regularization errors. These types of
error continue for several years. However, such errors are relatively rare, with
over 95 per cent of irregular past tense verb use continuing to be perfect (Pinker,
1995; see also Figure 2). The incidence of over-regularization then decreases as
children get older before gradually returning to 100 per cent performance in
adulthood. The reason behind this U-shaped development, as it is termed by
psychologists, is linked to childrens increasing sensitivity to morphological rules
(see Figure 2). That is, initial perfect production of past tense verbs is likely to be
due to children simply copying the speech that they hear. The period of overregularization suggests that children have inferred a rule about how to indicate
that things happened in the past add /ed/ to the end of the verb which they
then have to learn to apply appropriately. We will discuss different explanations
of how children learn to do this later in the chapter.
169
100
90
80
70
60
50
40
30
20
10
0
20
(b)
100
90
80
70
60
50
40
30
20
10
0
25
(a)
170
35
45
Age in months
55
120
220
Vocabulary size
320
Figure 2 Graphs showing the percentage of irregular verb errors (a) observed in the speech of
Adam (one of the children from Browns study, 1973), and (b) made by a computer model that is
learning how to produce irregular verbs (Plunkett and Marchman, 1993). Note that in both cases
the overall rate of correct responses is high, but both are characterized by a small dip in
performance over time and as vocabulary increases. This is the so-called U-shaped pattern of
development.
2.3 Syntax
Syntax is the term used to refer to the knowledge that a speaker possesses
regarding the manner in which words can be combined to form sentences. In
English, syntax provides the information needed to determine who did what to
whom. For example, English speakers have no difculty distinguishing the
meanings of the two sentences:
(1) John gave Mary the owers.
(2) Mary gave John the owers.
In the rst sentence, John does the giving whereas in the second sentence it is
Mary. The meanings of the words by themselves do not yield this information. It is
necessary to pay attention to their position in the sentence. It might be tempting
on the basis of this simple observation to conclude that syntax is synonymous
with word order. For example, we might infer that the noun that occurs rst is
always the subject of the sentence. However, it is relatively easy to show that
syntax is more complicated than this. In the sentence,
(3) Mary was given the owers by John.
the subject of the sentence is the last noun mentioned, yet it is not difcult to
understand that John did the giving. Figuring out who does what to whom
requires the listener to be sensitive to the cues that indicate the type of sentence
used. The language learner must decipher these cues in this case, whether it is a
passive (as in example (3) above) or an active sentence (as in example (1)). In
general, this means paying attention to the structure of the sentence.
One of the achievements of grammarians has been to discover the kinds of
structures listeners must recognize when understanding sentences. For example,
passive sentences are easily identied by the presence of the word by. However,
sometimes the structural cues are far more subtle. In Chapter 3, Activity 4, you
considered the example,
(4) A unicorn is in the garden.
and how easy it is to turn this into a question by moving the word is to the front
of the sentence. However, you also noted that this simple rule does not apply
when a sentence such as (5) is converted into a question.
(5) A unicorn that is eating a ower is in the garden.
Such examples demonstrate peoples sensitivity to more complex aspects of
sentence structure.
171
172
children can remember the word sequences or identify the inections on the
words and then work out how they relate to the events described by the speaker.
An alternative solution is that young children already know that languages can
be syntactically or morphologically oriented. That is, they are born with an
implicit understanding that languages can signal who did what to whom through
syntactic or morphological devices. In Chapter 3 you were introduced to the idea
of language development being underpinned by an innate predisposition to
acquire it; a position advocated by the work of Noam Chomsky, and further
developed by Steven Pinker. With respect to acquiring grammatical
understanding, Chomsky (1965) argued that humans have an innate knowledge
of potential language structures, which he refers to as Universal Grammar.
Universal Grammar contains a set of constraints on language processing that can
be switched on or off through exposure to spoken language. Languages which
exploit morphological cues switch on the morphological constraints, whilst
those which exploit syntactic cues switch on the syntactic constraints.
Chomsky argued that this kind of innate knowledge is necessary because
children are rarely presented with coherent, grammatically complete speech that
maps directly onto things happening in their immediate environment. For
example, speakers might say a variety of different things in the context of <John
chasing Mary> such as Look or Not him again, or the conversation may be
about something that happened in the past that the child knows nothing about.
How are children to know which are the important parts of language to focus on?
Chomskys answer to this question is that it is Universal Grammar that directs the
childs linguistic attention to the right kind of information.
t
Reading
At this point you should read Reading A, which is an extract from The Articulate Mammal by
Jean Aitchison. This reading offers a simple introduction to Chomskys notion of a Universal
Grammar which we will return to later in this chapter.
.
.
.
173
174
4.00
3.50
Eve
Adam
Sarah
3.00
2.50
2.00
1.50
16 18 20 22 24 26 28 30 32 34 36 38 40 42 44
Age in months
MLU
Forms
1.75
Telegraphic speech
II
2.25
III
2.75
IV
3.5
4.0
175
176
Expressive
language
The language that
a person actually
produces.
Systematic transcription
Systematic transcription is a process whereby periods of speech are recorded and then
transcribed. The recording almost invariably includes excerpts of speech from the
childs conversational partner, thereby providing a context in which to interpret the
childs utterances. The continuous collection of data (sometimes over several hours)
offers the investigator a representative sample of the childs language. Indeed, recording
in different situations allows the investigator to determine how speech varies from one
context to another. The cost of this enriched source of data is the time taken to collect
and transcribe the material. For example, a single hours talk can take many hours to
transcribe.
There are two ways that speech can be transcribed. The rst of these is phonetic
transcription which involves representing each utterance using a special phonetic
alphabetic. This differs from standard alphabets, which may use the same letter symbol
to represent more than one sound (e.g. in English the letter c can represent the sound
[k] or [s]), or more than one symbol to represent the same sound (e.g. in English the
letters f and ph both represent the sound [f ]). Moreover, each language only uses a
proportion of all the sounds that it is possible to produce. In contrast, the phonetic
alphabet has one symbol each for all the sounds that could possibly be produced by the
human voice and also distinguishes between sounds that a speaker of a single language
may not recognize (see Section 2.1). This makes it ideal for transcribing the talk of
young children whose early utterances can include foreign sounds.
Many researchers opt instead for the second way, that is producing orthographic
transcriptions. This form of transcription glosses over the peculiarities of the childs
pronunciation, and identies words from adult language that the investigator believes
that the child is attempting to produce. The addition of video-recording methods and
digital enhancements provides investigators with even richer sources of information to
interpret and analyse childrens utterances. For example, it is now possible to view on a
computer screen a digital video recording of a child simultaneously with a transcription
of what the child or his or her conversational partner is saying.
However, systematic transcription records do not necessarily offer an accurate
reection of what children know about language. After all, children may know a lot
more about language than they actually say. Also, given the naturalistic situation in
which the data are typically collected, it can be very difcult to assess what could be
contributing to a childs linguistic behaviour. Take, for example, the case of a child who
produces an especially complex utterance, potentially indicative of a sophisticated
knowledge of language. Was it, in fact, only produced because of something an adult has
just said or done, or because of the presence of other objects and events at the time?
Consequently, investigators have turned to more experimental methods to assess
childrens comprehension of linguistic forms, which offer the investigator more control
over the situation in which the childrens talk occurs.
n
Receptive
language
The language that
a person actually
understands, but
that may not
necessarily feature
in their own
language use.
Relatively little work has been done to determine the grammatical abilities
present in one year olds receptive language. In general, the ability to
comprehend language is better developed than the ability to produce language.
For example, children are unlikely to produce the word dog unless they are able
to understand it. Consequently, there is good reason to believe that the
developmental prole of speech presented in this section may underestimate the
grammatical abilities of the one year old.
Experimental procedures have been developed to enable investigators to
better assess young childrens potential to understand grammatical rules. One
such procedure is the inter-modal preferential looking task. This procedure was
rst developed by Spelke (1979) who used it to investigate infants perceptual
abilities. She showed infants two events (e.g. something being dropped or
hands clapping) but only played the sound associated with one of those events
(e.g. clapping). The children were found to spend more time looking at the event
that was consistent with the sound that they heard, their extended gaze
suggesting that they correctly associated the sound with the event. This technique
has since been adapted to analyse very young childrens understanding of
grammar and how this might facilitate the acquisition of new words (see Research
summary 1).
177
178
RESEARCH SUMMARY 1
Transitive verb
A verb that must
refer to an object
in a sentence. For
example the verb
to hit is transitive
it is not
legitimate just to
say I am hitting.
It is necessary to
say what is being
hit.
Intransitive verb
A verb that does
not require its
object to be
specied in a
sentence. For
example, to read
is intransitive as it
is possible to say
I am reading. It is
not necessary to
say what is being
read.
Figure 4 Experimental set-up of the preferential looking task (from Naigles and Kako,
1993, p. 1670).
The child was shown actions where one actor had an effect on something (causative
actions) and actions that had no effect on another thing (non-causative actions). So the
child saw, for example, footage of a duck making a rabbit bend over (a causative action)
while both animals were making arm movements (the non-causative actions). This
scene was paired with a nonsense verb (e.g. gorp) which was presented in one of two
sentences. One sentence, The duck is gorping the bunny, implied that the nonsense
word was a transitive verb (one that requires a corresponding object). The
other sentence, The duck and the bunny are gorping, implied that the nonsense word
was an intransitive verb (one that does not require an object).
Then the actions were shown separately. On one screen the duck was shown bending
the rabbit over; on the other screen the duck and the rabbit were shown just moving
their arms (see Table 2).
Naigles found that the children looked at the causative actions, such as the duck forcing
the rabbit to bend over, for longer than they did at the other screen when they heard
the phrase The duck is [word]-ing the bunny!. They looked at the non-causative
actions, such as the duck and rabbit making arm movements, for longer when they
heard the phrase The duck and the bunny are [word]-ing!. These results were taken as
evidence that children are able to use their sensitivity to grammatical forms to learn
about verb meanings.
Table 2 Structure of the stimulus video tapes for the verb gorp
Video tape 1
Audio
Video tape 2
Black screen
Black screen
Find gorping!
179
180
Age 2;3: Play checkers. Big drum ... A bunny-rabbit walk ...
Age 2;6: What that egg doing? No, I dont want to sit seat ...
Age 2;9: Where mommy keep her pocket book? Show you something funny ...
Age 3;0: I going wear that to wedding ... Those are not strong mens ... You dress
t
Activity 2
Allow about
5 minutes
This activity will help you to reect on the ways in which childrens grammatical competence changes
over time.
Look back at the examples of Adams speech. What do you notice about the way that Adam
forms his utterances over this period?
Comment
The most obvious thing to note is the change in the length of Adams utterances. His attempts
to combine words are not always grammatical, but they do increase in length over time. The
early utterances omit function words (the, a, on, etc.) as expected. He attempts to
formulate questions quite early on in his multi-word utterances although is is often omitted.
He also makes an interesting error mens, where he correctly inects the noun man, but still
adds the conventional /s/ ending to indicate that there is more than one. The later utterances
are more in line with adult speech (I want to have some espresso) although they are not
always error free (... the mailman can know where I are ...). The main source of error
throughout is from omitting words, rather than getting them wrong. You can see that simply
by inserting the missing words the sentences become grammatical: the actual order of the
words that are produced is correct, suggesting that Adam shows some grasp of syntax.
Even when they make errors, children still show sensitivity to correct forms, even
Child:
You readed some of this too ... She readed all the rest.
Parent:
Child:
Parent:
Child:
Parent:
Child:
Although children are continuously bombarded during their waking hours with a
stream of richly structured linguistic input, they seem to assimilate the structures
of their native language without explicit instruction or correction (Brown and
Hanlon, 1970). The ease with which all this seems to occur belies the underlying
complexity of the skill acquired. As is often the case with many complex abilities,
a lot can be learned from observing the order in which parts of the skill are
acquired, and tracking down the rare mistakes that individuals make in exercising
their skills. This is precisely what will be discussed in the next two sections. The
discussions in these sections will help you to evaluate the extent to which it can
be said that language acquisition is driven by innate cognitive mechanisms (like
Universal Grammar) or experience of a linguistic environment.
181
182
record, the rule system is blocked and the memorized (irregular) form is
produced. For example, in an attempt to inect the word go, the language user
will discover that the memory system contains went, which blocks the
application of the add /ed/ rule, preventing the error goed.
t
Activity 3
Allow about
5 minutes
This activity will enable you to assess how well you understand dual route explanations of inection.
Assume that a boys memory system contains only the following irregular past tense verb
forms: went, broke and held. What does the dual route theory predict the child will say when
he is required to produce the past tense of the following words: go, want and eat?
Comment
According to the dual route model, he would be expected to say went rather than goed
because he has a representation of this irregular form in his memory system. He should say
wanted because if there is no corresponding irregular representation in the memory system
the add /ed/ rule will be applied. However, the boy will say eated instead of ate, because
there is no stored representation of the irregular form, and so the add /ed/ rule will, once
again, be applied in this case.
According to this account, the development of the memory system is critical to the
proper operation of the inectional process. In order to identify an irregular
inection such as went or men, the irregular forms need to be stored and
retrievable from memory. Failure to retrieve exceptions from memory will result
in a failure of blocking and hence an over-regularization error. Memory failure is
most likely due to lack of knowledge of particular words. Hence, irregular words
that only rarely occur in speech are most susceptible to over-regularization. With
greater experience (as the child gets older) memories for rare types of inection
are consolidated and errors diminish. Dual route theory suggests that the
inection that occurs across the greatest number of words is established as the
default rule (Marcus et al., 1992); for example -ed is the most common type of
past tense inection in English, so it becomes the rule.
183
184
BOX 3
Connectionism
Connectionist networks are computer models loosely based on the principles of neural
information processing. They are sometimes referred to as neural networks. It is only
necessary to have a basic knowledge of how connectionist models work in order to
understand what they can tell us about cognitive processes (such as language
understanding and memory) and cognitive development.
A connectionist model consists of units, which can be thought of as roughly analogous
to neurons. A typical network is organized as follows. A set of units is designated as
input units. These can be thought of as equivalent to human sensory neurons which are
activated by a stimulus from the environment. Another set of units is designated as
output units. It might be helpful to think of these as the equivalent of motor neurons,
which produce a behaviour. In between the input and output units are various
congurations of hidden units. These can be thought of as equivalent to all the neurons
which are part of the black box of cognition; everything that goes on between input
(stimulus) and output (response).
A connectionist network needs to be able to accept an input (a stimulus), make various
computations, and produce an output (a response). The job of a connectionist network
when it is being used as a model of human information processing, is to accept a stimulus
and to learn to make the response that a person would make to that stimulus. For
example, if a connectionist network were being built to model a childs developmental
mastery of the English past tense it would be provided with the stimuli of English verb
stems. It would need to learn to respond to each verb with the appropriate past tense
form. Its behaviour would be observed as it learned (developed) and then compared to
the course of development that can be observed in children.
n
Plunkett and Juola (1999) trained a connectionist network on some 5,000 different
English inections. They found that, like children, their network went through an
initial period of producing nouns and verbs correctly, which was then followed
by a period of intermittent over-regularization. They also found that the network
learned plural inections before past tense inections. If you look back at
Table 1, you will see that this is consistent with the results of Browns (1973) study
of childrens grammatical development. Given the small scale of these networks
compared to a human brain (see Chapter 3, Section 2.1), the proposal that a single
system might memorize all of the inections in the language does not seem so
unlikely after all.
Single route theory differs from dual route theory in its account of why children
begin to make over-regularization errors. The explanation is this. As more and
more inections are stored in memory, competition for memory resources
increases. This competition between words results in interference effects, where
words that are very similar to each other are easily confused. For example, fan
and man sound very similar to each other but one has a regular plural inection
(fans) and the other has an irregular inection (me n). In the case of the
irregular word man, if there are lots of regular words stored that have a similar
phonological structure (e.g. fan, ban, van and span) then man may be
accidentally treated as if it too has a regular inection (mans).
This is exactly what Rumelhart and McClelland (1987) found when they trained
their connectionist network on regular and irregular verb inections. Early in
training, the network only had a few verbs to memorize and so it managed to
inect the verbs correctly. However, as its vocabulary grew, competition for
network resources intensied and over-regularization errors were found. With
further training, the irregular verbs became more robustly stored in the network
and were able to resist the interference effects of regular neighbours.
t
Activity 4
Computational modelling
Allow about
5 minutes
This activity will encourage you to reect on the advantages of computational modelling as a
technique in research into child development.
Why would you want to model a developmental process? What would be the aim of such a
model?
Comment
The aim of any model is to test specic predictions about the mechanisms that underlie the
developmental process in question. It is one thing to propose a model on paper, but quite
another thing to build a computational model that actually has to function. For one thing it
forces researchers to be explicit about how all the different structures and processes that they
are proposing work. If the nished model shows patterns of behaviour that are similar to those
shown by children, then perhaps the structures and processes built into the model are similar
to those going on in the developing child. Perhaps should be emphasized because even if two
systems (in this case a child and a computer model) produce the same behaviour, that does
not necessarily mean they are doing it in the same way. If you search your memory for a piece
of information, and a computer searches its memory for the same piece of information, the
output from you and the computer may be the same, but the search process may have been
completely different.
185
186
t
Activity 5
Allow about
10 minutes
This activity will give you another opportunity to test your understanding of single and dual route
accounts.
As discussed earlier, single route theories and dual route theories differ in the types of
mistakes that they predict children will make when learning about inections. Here are two
types of child-like inection errors [with the correct form of the verb in brackets]:
(a) I holded [held] the books for my teacher.
(b) The paperclips were lank [linked] together.
Can a dual route account explain both these types of error? Can a single route account
explain both these types of error?
Comment
Sentence (a) is an example of an over-generalization error, and hopefully you remembered
that both theories can explain this sort of mistake. Sentence (b) is an example of what
happens when a regular word takes an irregular inection. While very rare, these do occur in
childrens speech. Dual route theory predicts that regular words are not susceptible to
interference effects from irregular words in the way illustrated by sentence (b). So errors like
lank should not occur. However, single route theory predicts that interference effects will
impact on childrens production of both regular and irregular words. That is, just as regular
inections interfere with the production of irregular words (just as fold becomes folded in
the past tense, so hold wrongly becomes holded), so should irregular words interfere with
the production of regular inections (just as sink becomes sank, so link wrongly becomes
lank, instead of linked). Consequently, this kind of evidence favours single route theory.
The two theories also differ in their account of how children generalize inections
to new words. For example, given the novel word wug, children and adults
would most likely use the word wugs as its plural form. Dual route theory
explains this creative use of inections in terms of the operation of the rule, and
there being no entry in memory to block the application of the rule. Single route
theory explains generalization in terms of the similarity of the new word to
existing words in the language (mug, slug, hug, etc.).
There is controversy here as to what the language user does with words that are
not similar to any other word stored in memory. This situation is particularly
poignant in the case of children with relatively small vocabularies. Dual route
theory predicts that the word will simply be inected according to the default
rule, whereas single route theory predicts that an educated guess will be made on
the basis of a weighted similarity to all the other words in the individuals
vocabulary. Experimental evidence provided by Marchman (1997) has examined
just this question and favours the single route account (see Research summary 2).
RESEARCH SUMMARY 2
n
Comment
So far, we have been careful not to relate either the dual or single route theories to
either nativist or empiricist positions. This is because both theories are simply about
specifying the nature of the cognitive mechanism that might account for the patterns
of development observed in young children. This debate can be seen as separate
from discussions of whether the mechanism is believed to be innately preprogrammed or to develop in response to experience with the linguistic
environment. However, historically nativists have aligned themselves with the dual
route theories (note that Pinker is an advocate of dual route theory) and
empiricists (such as Marchman) have supported the single route position. The basic
distinction is that dual route theorists generally propose that the rule mechanism is
innate whereas single-route theorists contest this view. However, it is entirely
possible to present a view in which a dual route system might arise as the result of
learning.
187
188
Szagun found a much wider variety of errors in children learning German than that
reported by Marcus et al. (1995) with little evidence of the s plural dominating in the
way they suggested. The childrens development in using the different plural inections
was consistent with the occurrence of those inections in their parents speech. The
frequency of the childrens use of the different inections was the same as that of their
parents by the time they were 3 years and 8 months old. Finally, the children made
errors right from the beginning of their use of plural forms. These ndings are
consistent with a single route account of German inectional morphology.
Williams
Syndrome
A non-hereditary
syndrome caused
by a chromosomal
abnormality.
Brain and physical
development can
be affected, and
symptoms can
include poor coordination, some
muscle weakness,
possible heart
defects,
occasional kidney
damage and high
calcium levels.
Development is
generally delayed
in these children.
Specic
language
impairment
(SLI)
A specic and
severe difculty in
acquiring some
aspect of
language
development,
despite no
apparent general
cognitive or
neurological
impairment.
In this extract the regular verb to watch is not marked for past tense, although
the irregular verb to have is produced correctly. Gopnik and Crago noted that
the children were signicantly more likely to make mistakes on regular verbs than
on irregular ones. Moreover, there was evidence that when the teacher corrected
past tense errors of this kind, the children did appear to learn the correct form in
that case, but were unable to generalize the add /ed/ rule to other regular verbs.
This pattern of results reects a double dissociation between regular and
irregular inection. Double dissociation refers to evidence that two cognitive
processes appear to be unrelated to each other. This is because not only does one
group of people show ability at A but impairment B (a single dissociation), but
another group of people exists who show ability at B but impairment at A. The
189
190
Comment
There are other arguments, aside from the linguistic evidence, that favour the single
route model. Firstly, it is a parsimonious account; that is, it is a simple account,
having only one component. It also appears to offer a biologically plausible account of
how neural connections might operate in the human brain to produce the observed
patterns of inflection use. Finally, the idea of distributed representation, that
knowledge is not represented in one location in the brain but is the product of a
network of processing units (e.g. neurons) that is consequently resistant to minor
damage, is one that has appeal.
.
.
So far you have seen that there are two competing explanations of how children
acquire inectional morphology, and that historically they have been associated
with either nativist or empiricist accounts of language acquisition. To extend and
further illustrate the nature of the nativistempiricist discussion, we will now
introduce you to explanations of how children come to understand the
signicance of syntax for constructing meaningful sentences.
Telegraphic speech (childrens early use of highly simplied sentences) offers
the rst opportunity to examine childrens mastery of the syntax of their language.
On the surface, children seem to be doing a pretty good job even by their second
birthday. Words are produced more or less in the right order (Mummy gone
work). They are creative in their combinations (All gone Mummy) and they
know how to use these sequences in conversation to ask questions, make
statements and issue commands. All that seems to be missing are the function
words that are characteristic of well-formed adult sentences (Mummy has gone to
work).
191
192
Notice that both sentences (10) and (11) contain the necessary ingredients for the
correct formulation of questions. The child has simply failed to place the auxiliary
verb is in front of the subject (Daddy or Mummy). Notice that sentences (10)
and (11) stripped of their respective Wh-words constitute legitimate statements.
Perhaps the rst attempts to produce Wh-questions build upon the ability to make
statements? At some point, however, the child must discover the importance of
subject-auxiliary inversion for question formation. How is this achieved?
Negative evidence
Brown and Hanlon (1970) argued that parents do not seem to correct the grammatical
errors of their children in a systematic fashion. They are more likely to correct the
factual errors or meaning of their childrens language. Explicit attempts to tell the young
language user that they have made a mistake No you said that wrongly are very
rare indeed. Language researchers therefore assumed that any theory of language
acquisition needs to be able to deal with an environment that does not provide the
learner with negative evidence. This assumption has not gone unquestioned. For
example, Demetras et al. (1986) suggested that indirect or implicit feedback to the child
may occur when a grammatical mistake is made. Indirect feedback might take the form
193
194
n
How then are children to correct their grammar? Two solutions seem possible here.
First, the language learner knows that certain types of verbs allow dative
alternations and others do not. Second, the language learner only uses dative
alternations on verbs for which there is evidence that they occur and refrains from
using them elsewhere. However, both of these solutions beg the question of why
children make the over-generalization in the rst place and how are they able to
correct themselves.
Pinkers solution is that language learners know what verbs allow dative
alternations, by virtue of their knowledge of a set of rules linking the meaning of
the verb to the grammar associated with it. Subscribing to Chomskys view that
Universal Grammar is innate, Pinker (1989) proposes that dative overgeneralizations are a consequence of imprecise knowledge about word
meanings. For example, a child may think of report as meaning the same as tell.
As a consequence, the same rules of grammar might also apply to it. As the
childs understanding of word meanings develops and the differences between
the two verbs are rened, the meaning structures associated with report will
trigger a different set of rules in Universal Grammar which will prevent the
inappropriate dative alternations.
The important feature of this solution to the problem of correcting overgeneralization errors in syntax is that it assumes that the grammar is correct all
along. It is just the semantics (word meaning) that the child has mistaken. This
approach is entirely compatible with the view that Universal Grammar is innate.
account. If these categories exist at all in the mind of the child, they are emergent
properties of the learning process, rather than prerequisites for it.
So, according to this approach, how do children extend their knowledge to
new situations and new sets of words? It is not possible for children to memorize
all the utterances they hear for later reproduction. However, a possible
explanation for their creativity with language is analogy; children may initially
memorize some of the utterances they hear and then adapt them as prototypes for
the construction of new utterances. In fact, one of the earliest accounts of
childrens grammatical development (Braine, 1963, 1976) comes very close to
adopting this position (see Box 5 for a discussion of early diary methods of
studying language development).
Braine (1963) examined his son Martins early word combinations and found
that they exhibited clear ordering patterns. Some words only ever occurred in one
position in the utterance (initial or nal). He called these words Pivot words. Other
words were free ranging. He called these Open words. Of course, most words can
occur in a wide range of positions in a sentence. However, Martin may have
noticed that some words vary more in their distribution than others. For example,
nouns move around a lot in the sentence (subject versus object position) whilst
verbs tend to stay put. Did Martin observe his parents and other adults producing
speech where the position of the Open words in sentences varied more than Pivot
words and imitate these patterns to structure his own sentences?
BOX 5
195
196
Conclusion
This chapter has introduced you to the complex nature of learning grammar as an
infant, and to various theoretical explanations of how acquisition of grammar
might be achieved. These explanations can be aligned to broadly nativist and
empiricist accounts of development. Chomskys Universal Grammar and Pinkers
dual route explanation of inection use suggest that children are born with innate
cognitive mechanisms that enable them to acquire grammatical competence on
exposure to the linguistic environment. In contrast, single route theories of
inection use and distributional accounts of syntax development suggest that
children can acquire representations of grammatical forms and how to use them
simply from exposure to the linguistic environment.
The balance of evidence presented in this chapter favours the more empiricist
accounts. Single route theories appear to offer a more credible account of
development across languages, and distributional accounts of syntax have the
support of evidence from modelling studies. This theory is also consistent with
observations of childrens language production and the nature of parental speech
to children. However, that is not to say that the idea of there being some form of
innate predisposition to acquire language has been rejected. Theoreticians are in
broad agreement that the homo sapien comes equipped with specialized
machinery that permits the acquisition of language. The days of tabula rasa
197
198
empiricism are long gone. Discussions are concerned with identifying the nature
of the skills that the child brings to language learning, and whether those skills are
specically related to language, or part of a more general cognitive ability. The
idea of a Universal Grammar, a detailed, innate knowledge of all types of
grammatical rules, is still being considered and assessed, as is the idea of children
having the ability to extract grammar from their environment as a result of the
nature of their neural systems.
Throughout this chapter you should also have gained a sense of how methods
of studying childrens grammatical development have developed over time. Early
diary studies were superseded by the use of recording equipment and commonly
agreed methods of systematic transcription. However, as these methods only have
the potential to inform discussions of expressive language use, experimental
techniques have been used to assess the nature of childrens expressive and
receptive language abilities (for example, see Research summary 2 and Research
summary 1 respectively). Finally, advances in cognitive science have led to the
development of connectionist models that enable theorists to test the viability of
their theories regarding the cognitive architecture needed to realize acquisition of
grammar. As methods continue to develop in response to the theories,
psychologists will move closer to a resolution to debates regarding the nature and
extent of innate knowledge in childrens grammatical development.
Further reading
Aitchison, J. (1998, 4th edn) The Articulate Mammal: an introduction to
psycholinguistics, London, Routledge.
References
Bellugi, U., Bihrle, A., Jernigan, D., Trauner, D. and Dougherty, S. (1990)
Neuropsychological, neurological and neuroanatomical prole of Williams
Syndrome, American Journal of Medical Genetics, vol. 6, pp. 11525.
Bishop, D. V. M. (1997) Cognitive neurospsychology and developmental
disorders: uncomfortable bedfellows, The Quarterly Journal of Experimental
Psychology, vol. 50, pp. 899923.
Braine, M. (1963) On learning the grammatical order of words, Psychological
Review, vol. 70, pp. 32348.
Braine, M. D. S. (1976) Childrens rst word combinations, Monographs of the
Society for Research in Child Development, vol. 41, Chicago, IL, University of
Chicago Press.
Brown, R. (1973) A First Language: the early stages, Cambridge, MA, Harvard
University Press.
199
200
Reading
Reading A: An extract from The Articulate Mammal
Jean Aitchison
Chomskys later views: setting switches
Suppose children knew in advance that the world contained two hemispheres, a
northern and a southern. In order to decide which they were in, they simply
needed to watch water swirling down the plug-hole of a bath, since they were
pre-wired with the information that it swirled one way in the north, and another
way in the south. Once they had observed a bath plug-hole, then they would
automatically know a whole lot of further information: an English child who
discovered bath-water swirling clockwise would know that it had been placed in
the northern hemisphere. It could then predict that the sun would be in the south
at the hottest part of the day, and that it would get hotter as one travelled
southwards. An Australian child who noticed water rotating anticlockwise would
immediately realize the opposite.
This scenario is clearly science ction. But it is the sort of situation Chomsky now
envisages for children acquiring language. They are pre-wired with a number of
possible options which language might choose. They need to be exposed to
relatively little language, merely some crucial trigger, in order to nd out which
route their own language has chosen. Once they have discovered this, they
automatically know, through pre-programming, a considerable amount about
how languages of this type work.
Let us consider how Chomsky hit on such an apparently bizarre idea.
Learnability remained Chomskys major concern. How is language learnable,
when the crumbs and snippets of speech heard by children could not possibly (in
Chomskys view) provide sufcient clues to the nal system which is acquired? ...
The learnability problem has also been called the logical problem of language
acquisition: how, logically, do children acquire language when they do not have
enough information at their disposal to do so?
The logical answer is that they have an enormous amount of information prewired into them: the innate component must be considerably more extensive than
was previously envisaged. Children, therefore, are born equipped with Universal
Grammar, or UG for short: UG is a characterization of these innate, biologically
determined principles, which constitute one component of the human mind the
language faculty (Chomsky 1986: 24). This is a distinct system of the mind/brain
(1986: 25), separate from the general intelligence...
In 1986, then, Chomsky viewed UG and language as something like an orchestra
playing a symphony. It consisted of a number of separate components or
modules, a term borrowed from computers. Chomsky noted: UG ... has the
modular structure that we regularly discover in investigation of cognitive systems
(1986: 146). Within each module, there were sets of principles. Each principle was
201
202
We are therefore dealing with a system of unifying principles that is fairly rich in
deductive structure but with parameters to be xed by experience (Chomsky
1980: 66). The interlocking nature of the system will ensure that minor alterations
will have multiple consequences: In a tightly integrated theory with a fairly rich
internal structure, change in a single parameter may have complex effects, with
proliferating consequences in various parts of the grammar (Chomsky 1981: 6).
In particular, a few changes in parameters yield typologically different languages
(Chomsky 1986: 152). This whole idea has become known as the principles and
parameters or P and P approach ...
Children have relatively little to do in this new system: We view the problem of
language acquisition as ... one of xing parameters in a largely determined
system (Chomsky 1986: 151). Indeed, many of the old rules which children had
to learn just appear automatically, because the principles underlying them are
there already. ... If this minimal effort by the child is correct, then it makes sense
to think of the language system as a mental organ, which grows mainly by itself,
in the same way that the heart grows in the body. Chomsky is increasingly
concerned to understand the principles which underlie this growth.
Chomsky, N. (1986) Knowledge of Language: Its Nature, Origin and Use, New
York: Praeger.
203
204
Chapter 5
Contents
Learning outcomes
207
Introduction
207
208
209
211
213
217
217
219
222
Executive dysfunction
225
Conclusion
228
References
228
Learning outcomes
After you have studied this chapter you should be able to:
1 provide a working denition of executive function;
2 evaluate the developmental signicance of executive function;
3 describe the development of one aspect of executive function in children;
4 identify some key implications of executive dysfunction.
Behaviour
The public side of
human life:
anything that a
person does that
is observable by a
third party.
Experience
The private side
of human life:
internal cognitive
and emotional
processes that
cannot be
observed from the
outside.
Introduction
One way of thinking about the developmental story that we have told so far in
this book is in terms of the progressive organization of childrens behaviour and
experience.
Chapter 1 discusses ways in which infants organize their understandings of the
world in terms of categories. Chapter 2 shows how this organizational framework
underpins, and is developed by, language. Chapter 3 examines ways in which
brain development can be seen in terms of increasing levels of organization in
both structure and function. The theme is extended further in Chapter 4 by
looking at how language itself develops into an organized system of linguistic
structures. Why is this developmental trajectory towards organization in the
growing child of such interest to psychologists?
One answer to this question is that organization of behaviour and experience is
a necessary pre-condition for intelligent, aware, planned action. It is a precondition for children having control of their own actions. If a child (or indeed a
person of any age) has no way of organizing their understanding of the world
around them, then they will be unable to behave appropriately or effectively.
Imagine, for a moment, being unable to comprehend that the next person you
meet is a person, that they have things in common with other people, and that
they differ in signicant ways from things that are not people. Alternatively
imagine knowing that people differ in signicant ways from each other, and from
things that are not people, but not being able to control your actions in a way that
acknowledges these differences. In both cases there would simply be no basis for
anything other than chaotic behaviour.
The term executive function (EF) is used to refer to the set of high-level
cognitive functions that enable people to plan, initiate and carry through goaldirected behaviour in an organized and thought out way. This denition reects
the contrast between high-level cognitive functions such as integration, synthesis,
planning, organizing and so forth with more basic, low-level cognitive functions
such as processing auditory, visual and tactile sensations. This distinction maps
onto the model proposed by Fodor in Chapter 3, Section 3, which sees the lowerlevel cognitive functions as being organized in a modular fashion in such a way as
to feed the higher-level, and more global executive functions.
207
208
It is Debbies fth birthday party. A group of children stand facing her mother at
the front of the room. She is leading them in a game of Simon says. In this game
she models simple actions, like putting her hands on her head, or jumping up and
down. The children must either copy these actions, or ignore them. Each action is
accompanied by a verbal instruction. If Debbies mother says Simon says, do this
then the children must copy the action that follows. If she says only do this then
the children must stay still and not copy her.
Debbies mother says Simon says, do this and sits on the oor. All the children
sit down as quickly as they can. She then says Simon says, do this, and jumps up
again. All the children follow. Immediately she says do this and sits down again.
Half the children sit down. The other half put their hands over their mouths and
giggle. The children who have sat down quickly realize their mistake and jump to
their feet again.
The behaviour of the children who mistakenly sat down may be attributed to a
number of factors. They may not have listened to or understood the instructions
of the game, for example. But rst and foremost their behaviour can be attributed
to a failure of executive function. Their behaviour was automatic and
momentarily beyond their conscious control. They acted rst, and thought later.
Note that within seconds of sitting down (perhaps even in the same moment as
they sat down) the children probably realized their mistake. Conversely, the
children who ignored the adult model were able to resist what was most likely a
strong urge to sit down. Their behaviour was guided by a greater level of
executive control.
There are several things to note about this example that will help you to relate
the abstract concepts that will be covered in this section of the chapter to
everyday experiences (see Box 1).
BOX 1
2 Use simple actions, for example sit down, jump up, touch your nose.
3 Maintain a quick pace, so that the actions come thick and fast.
4 Lull the children into a false sense of security. Perform many repeated Simon says
209
210
.
.
.
.
.
.
.
The Simon says example is most relevant to the last two situations listed above,
where conscious control of behaviour is necessary, and where there is a need to
overcome a strong, habitual response. The adult leads the children into a
sequence of behaviour in which they are behaving automatically. The actions are
simple so the children do not need to engage in effortful processing in order to
copy them. They may start to watch others, become distracted by other thoughts
and experiences, and forget to monitor what the adult is saying. Suddenly, and
with minimal warning (by means of a minor alteration to the verbal cue), the
automatic response is no longer the correct one. In order to succeed the children
have to inhibit the habitual response of copying.
An early example of the distinction between automatic and controlled actions
in the psychological literature was given by one of the founding fathers of
psychology in the USA, William James. He described how he went upstairs to
change his clothes and then found that he had got into bed instead. This kind of
automated, unmonitored action can be contrasted with actions that are preceded
by an additional conscious element (James, 1890, p. 522). Note how the children
who succeed at Simon says need to attend to the precise wording of the
instructions, and have to monitor their behaviour consciously in order to avoid
slipping into thoughtless imitation.
t
Activity 1
Slips of action
Allow about
10 minutes
This activity illustrates the distinction between automatic, habitual behaviour on the one hand, and
consciously planned behaviour, on the other. It will enable you to use this distinction to analyse
familiar everyday mistakes.
Identify examples from your own experience of slips of action, when something that you
meant to do was overridden by automatic, habitual behaviour. One example might be dialling
a familiar telephone number, only to realize when the person answers that you had meant to
telephone someone else.
Comment
Here are some examples that we came up with:
.
nding yourself taking the turning that you usually take every day on your way to work,
when you were meant to be driving somewhere else;
going upstairs to fetch something, getting distracted by something else, and going back
downstairs without the thing that you went to get in the rst place;
clicking the OK button on the computer to ignore, as usual, a warning when this time you
should have pressed Cancel. You realize this as your nger clicks the mouse button. Your
whole assignment has been deleted!;
making coffee for four people. One of them does not take milk. You intend to leave one
cup black, but you end up putting milk in all of them.
Everyday mistakes such as these appear to result from strongly automated behaviours taking
over from a conscious plan to do something else. They are examples of momentary failures of
executive function. The mistakes children make in the Simon says game also result from a
strongly automated set of behaviours (copying the adult again and again) overriding conscious
planning.
It is important to note that for an act to be automatic it does not necessarily need
to be simple, such as copying an adult putting one hand in the air. Driving a car is
a complex skill but it is one which can become almost automatic if it is sufciently
well learned. So, what does distinguish controlled actions from automatic actions?
Controlled actions involve three important features:
1
2
211
212
Motor cortex
Premotor
cortex
Prefrontal
cortex
Phineas Gage
In 1848 Phineas Gage, a railway worker in the USA, was using an iron rod to tamp down
gunpowder for blasting when a spark caused an explosion and the rod was propelled
straight up through one eye socket and both left and right frontal lobes (see Figure 2).
Initial interest was focused on the remarkable fact that he had survived this injury at all,
and the wider implications of the case were not appreciated until a follow-up report 20
years later. In this report, Harlow drew attention to the fact that whilst Gage had
indeed recovered from the accident with his speech, learning and memory intact, he
was left with such a profound impairment of social behaviour and personality that
according to friends and co-workers he was no longer Gage. Changeux quotes from
Harlows account of Gage:
He is tful, irreverent, indulging at times in the grossest profanity ... impatient
of restraint or advice when it conicts with his desires ... capricious and
vacillating, devising many plans of future operations, which are no sooner
arranged than they are abandoned in favor of others.
(Harlow, quoted in Changeux, 1985, pp. 1589)
Cognitive exibility
Cognitive exibility refers to the aspect of executive function that enables people
to think and behave appropriately according to the changing needs of a complex
environment, and in line with their plans and goals. One of the hallmarks of
mature human behaviour is its exibility and responsiveness to the constantly
changing features and demands of the environment, particularly of the social
213
214
executive function
215
216
t
Activity 2
Allow about
5 minutes
This activity gives you practical experience of doing the Tower of London task and illustrates how it
measures planning.
Look at Figure 3. Moving one ball at a time, what is the minimum number of moves that is
needed to change the pattern of balls in the initial position into the pattern shown as goal
position no. 1? Starting from the initial position in each case, work out the minimum number
of moves needed for goal positions no. 2 and no. 3.
Initial position
Goal position
(no. 1)
Goal position
(no. 2)
Goal position
(no. 3)
Comment
Turn to the end of the chapter to nd out the answers. The real task is somewhat easier to do
because it involves real pegs and balls that can be moved around (or ones that can be moved
around on a computer screen). Using static pictures, as you have been, places an additional
load on working memory, because for every move that you make you must actively hold in
your memory the position to which each ball has been moved.
As you can see from Activity 2, working memory is related to planning in this
view of executive function. Working memory refers to that aspect of remembering
that involves holding things actively in mind in support of ongoing plans of
action. For example, in the previous example of conversing with someone,
working memory would be used to maintain a representation of what the other
person is saying, or has just said, while planning a response to them. Working
memory is in itself a huge area of psychological research, which is beyond the
scope of this chapter. It is important to recognize, however, that it can be
regarded as a component of executive function.
For much of the twentieth century, research on executive function centred almost
exclusively on adults. This was mainly because the prefrontal cortex was thought
to become functionally mature only late in development, around adolescence
(Luria, 1973). However, it has become increasingly clear that the onset of the
development of executive function occurs much earlier than was previously
thought. This has become apparent following the appearance of more
appropriate tools for studying it at earlier ages. In this section we will look at the
course of one aspect of early executive function development; inhibitory control.
t
Activity 3
Allow about
5 minutes
This activity encourages you to reect on the complexity of your everyday environment.
Consider your sensory world. Look around the environment you are in. Try to observe
everything you can see. Then sit back, shut your eyes and listen to everything that you can hear.
After that, focus on what is touching you. Attend to each part of your body in turn. What can
you feel? It takes a few minutes to get the most out of this activity, so do not hurry it.
Comment
The richness and complexity of the assaults on your senses are considerable. However, it is
hard to experience this complexity to its full extent because the mature human mind is so
expert at ltering out irrelevance, and building stimuli into simpler, more meaningful (and
organized) patterns. If this ltering did not take place, and you gave equal weight to all
incoming sensory information, it would be impossible to behave in anything but a chaotic
manner.
Your current plan of action probably revolves around reading this section of the
chapter. In order to do this successfully you have to be able to ignore most of the
217
218
sensory world around you. You have to be able to prioritize the meaningful
stimuli the print that makes up words on the page and inhibit responses to
irrelevant stimuli, in order to enact your plan of reading the current paragraph,
and to achieve the goal of nishing it. What if you were unable to do this? What if
you could not give the words on the page any greater priority in cognition than
the grain of the wood on the table on which your book is resting? What if you
were unable to inhibit responses to the sound of the computer fan as it whirrs in
the background, or the feel of your feet on the carpet?
If you were unable to inhibit responses to stimuli that do not relate to the task
that you have planned to do, then it would probably be impossible to complete it
and achieve your goal. You would be drawn from one stimulus to another, in a
haphazard fashion, and it would be impossible to undertake any coherently
organized action. This is a rather extreme way of conveying the point, but young
children and people with executive function related disorders of inhibitory
control do have difculty in prioritizing their response to task-related stimuli, and
do have difculty in inhibiting responses to what are referred to as prepotent
stimuli.
A prepotent stimulus is a stimulus that draws a persons attention towards it,
and which seems to cause the person to behave in a particular way (the prepotent
response). Prepotency is a very important feature of effective everyday
functioning. It is to be hoped, for example, that a red trafc light would draw a
drivers attention towards it, and cause the driver to behave in a certain way. The
sight, smell and feel of the mothers breast are the most likely prepotent stimuli
for the young breast-feeding infant.
In the course of typical development it is possible to observe infants and young
children being distracted by inappropriate prepotent stimuli. By inappropriate
we mean stimuli that are nothing to do with the childs current plan of action. For
example, one might observe an 8-month-old infant catch sight of a toy on the
other side of the room and begin crawling towards it. It is clear to an observer that
they are enacting a plan to get the toy, but halfway across the room the infant
notices a scrap of paper on the oor. This seems to capture their behaviour and
their attention. They pick it up, sit down and inspect it. The original plan is now
lost and they have been catapulted onto another stream of behaviour, which
might involve another plan, which might itself get interrupted by another
prepotent stimulus, and so on and so forth. This executive function analysis of a
familiar scene offers one explanation of why infant behaviour sometimes appears
somewhat haphazard and disorganized to an adult onlooker according to this
view it is because executive functions are as yet undeveloped.
One aspect of child development that psychologists have become interested in,
then, is the way in which children develop an ability to inhibit responses to
stimuli that are nothing to do with their current plan of action. Put another way,
this amounts to an ability to prioritize responses to task-relevant (as opposed to
task-irrelevant) stimuli. When children begin to be able to do this, their behaviour
becomes less haphazard, and progressively more strategic and organized.
t
Activity 4
Allow about
15 minutes
This activity allows you to do the Stroop task and gives you a way of experiencing prepotency directly.
Get three coloured pens perhaps a blue one, a red one and a black one. Write down a list
of fteen words as follows (using the names of the colours that you have chosen). Note that
BLUE
RED
BLACK
RED
RED
(continue writing each word in the wrong colour, sometimes writing RED in
BLUE
BLACK
BLUE
RED
BLACK
BLACK
BLUE
RED
BLUE
BLACK
Now, out loud, and working as accurately and as a quickly as you can, call out the colour of
the ink in which each word is written. What happens?
Comment
You should nd that after a few words you start to get confused, wanting to call out the word
that you are reading, rather than the colour of the ink in which it is written. The meaning
associated with the word is acting as a powerful prepotent stimulus. You have to inhibit
everything you have learned about words and their meaning in order to call out the colour of
the ink. You could take two measures of performance from this, each of which would give
some information about inhibitory control: speed of completion, and number of errors made.
Why do you think that this task is not suitable for children of, say, 3 years of age?
The Stroop task tends to be used with older children and adults because of the
demands it makes on literacy skills, which in themselves are not a component of
219
220
The Handgame
The Handgame is a task that has the same basic structure as the Stroop task from
Activity 4. The child must inhibit a prepotent response in order to execute a ruleguided action. Specically, the child is rst asked to imitate two hand actions (making a
st and pointing a nger). Then, in the conict condition, children must make the
opposite responses (making a st when the experimenter points their nger and vice
versa). This involves:
.
inhibiting the prepotent response to imitate; and
.
performing an action guided by the rule do the opposite of what the experimenter is doing.
The measure of executive function that this task provides is the number of errors in the
conict condition. The task is based on work by Luria and has been used with preschool children. Other variants of this task include Lurias Knock/Tap game (in which
the child must knock when the experimenter taps the table and vice versa), the
Opposite Worlds task (in which school-aged children are asked to say one when they
see a 2 and to say two when they see a 1), and the Day/Night Stroop task (Gerstadt
et al., 1994; see Figure 4) in which children are instructed to say the word day when
shown a line drawing of the moon and stars, and night when shown a line drawing of
the sun. You should be able to see how these tasks relate conceptually to the Stroop
task, and how they are better suited to minds that are not yet at the stage of having
overlearned (automated) the ability to read words.
Commission
errors
Giving a response
when no response
is required (on the
Go/NoGo task
this is when the
child Goes on
the NoGo
stimulus).
Commission
errors can be
interpreted as a
failure to inhibit.
Omission errors
Missing a
response when
one is required
(on the Go/NoGo
task this is when
the child does not
Go on the Go
stimulus).
Omission errors
can be interpreted
as a failure to
attend to the task.
221
222
Perseveration
The continuous,
insistent
performance
(repetition) of a
behaviour.
The most widely used measure of executive function is Grant and Bergs (1948)
Wisconsin Card Sorting Test (WCST). The WCST involves a set of four target cards,
and a deck of stimulus cards, each with a picture that varies in colour, shape, and
number (for example, two blue stars, or three black squares; see Figure 5). The child is
asked to sort each card onto one of the target cards, and is told whether or not they
have placed the card correctly. Feedback is given after each card is sorted. By a process
of trial and error the child should be able to work out the rule according to which they
should sort the stimulus cards (for example, by colour or by shape). After six
consecutively correct responses the sorting rule is changed and the child must discover
the new rule. The two key measures from the WCST are (i) the number of rules for
which a child achieves a run of six correct trials; and (ii) the number of post-rule-change
trials for which the child makes a perseverative error (that is, continuing to sort
according to the previously correct, but now incorrect rule). Like other traditional
executive function tasks, the WCST actually taps into several different aspects of
executive function and so provides an index of overall executive function development.
For example, because the same set of cards is used for each sorting rule (and the child
is not told that the rule has changed), it is unclear whether perseverative errors reect
a failure to shift to the newly relevant rule (thus demonstrating cognitive inexibility) or
a failure to inhibit a previously reinforced response to a specic card type (thus
demonstrating a failure of inhibitory control).
223
224
Target card
Subject tries
Shape
Wrong
Colour
Wrong
Number
Right!
n
As executive function develops, so childrens abilities to learn new skills improve,
and they are increasingly able to behave in a planned, strategic and organized
manner. They are able to stay on-task longer. They are able, when necessary, to
override habitual responses to prepotent stimuli. They become more skilled and
exible in orchestrating elements of their thinking and behaviour, and they are
able to engage in increasingly sophisticated planning and decision making.
Inhibitory control is only one component of this developmental trajectory, but it is
of fundamental importance. One way to assess this importance is to look at the
implications for child development of a failure to develop typical levels of
inhibitory control. This is the topic of Section 4.
.
.
Executive dysfunction
This section looks at executive function from a different perspective. What might
the consequences be if the development of executive function were impaired?
This is an important question to ask for two reasons. The rst of these is to
understand the experiences and needs of those children and adults who are
directly affected. The second reason is that investigating the dysfunction of
cognitive processes is one way of increasing psychological knowledge about how
the processes work, and of nding out about what contribution they make in
typical development. We will focus on problems in the development of inhibitory
control, but you should bear in mind that executive dysfunction can also take the
form of cognitive inexibility, planning difculties, and problems associated with
working memory.
t
Activity 5
Allow about
10 minutes
This activity encourages you to consider the developmental consequences of poor inhibitory control.
What would you predict to be the implications for a child of a failure to develop typical levels
of inhibitory control? Note down possible consequences of under-developed inhibitory
control with respect to cognitive, social and emotional development.
Comment
If you were to compare a group of children with poor inhibitory control with a group of
typically developing children of the same ages, you might expect the former group to be, on
average:
.
more distractible;
less able to undertake tasks and learn skills that require sustained levels of concentration;
more socially awkward (perhaps saying things at inappropriate moments; blurting out);
225
226
less able to regulate emotions (more prone to mood swings and ts of anger);
This is by no means an exhaustive list. You may have come up with more ideas than this. The
activity should give you some indication of the importance of inhibitory control in child
development.
227
228
Conclusion
Answers to Activity 2
Goal position 1: two moves.
You would move the grey ball from peg 2 to peg 3 and the blue ball from peg 1 to
peg 2.
You would move the grey ball from peg 2 to peg 3; the blue ball from peg 1 to peg 2;
the black ball from peg 1 to peg 2 and the grey ball from peg 3 to peg 1.
You would move the blue ball from peg 1 to peg 2; the black ball from peg 1 to peg 3;
the blue ball from peg 2 to peg 1; the grey ball from peg 2 to peg 1 and the black ball
References
Baker, L. and Brown, A. L. (1984) Metacognitive skills and reading, in Pearson, P.
D. (ed.) Handbook of Reading Research, pp. 35394, White Plains, NY, Longman.
Barkley, R. A. (1997) Behavioral inhibition, sustained attention and executive
functions: constructing a unied theory of ADHD, Psychological Bulletin, vol. 121,
pp. 6594.
Brophy, M., Taylor, E. and Hughes, C. (2002) To go or not to go: inhibitory
control in hard-to-manage children, Infant and Child Development, Special Issue
on Executive Functions and Development, vol. 11, pp. 12540.
Brown, T. E. (1999) Does ADHD diagnosis require impulsivity-hyperactivity?: a
response to Gordon & Barkley, ADHD Report, vol. 7, pp. 17.
229
230
Chapter 6
Understanding minds
Emma Flynn
Contents
Learning outcomes
233
Introduction
1.1 The importance of understanding minds
1.2 Understanding minds and egocentrism
233
233
234
235
236
238
241
243
244
245
246
248
251
252
241
5.2 Communication
253
5.3 Gender
5.4 Summing up the signicance of social factors
254
255
Conclusion
256
References
256
UNDERSTANDING MINDS
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 understand and dene what is meant by the term theory of mind;
2 understand the rationale underlying the tests used to establish whether
young children have developed the ability to reect on another persons
mind;
3 recognize what a false belief is and why it is important;
4 understand the difference between rst-order and second-order theory of
mind;
5 appreciate that the development of a theory of mind allows other skills to
develop;
6 identify the relationship between skills which develop early in infancy and
a subsequent appreciation of minds;
7 understand the relationship between childrens cognitive development and
their social environment;
8 understand the relationship between childrens communication and the
development of their theory of mind.
Introduction
233
234
Theory of mind
The ability to take
the perspective of
others and
understand their
mental state, i.e.
beliefs and
desires.
UNDERSTANDING MINDS
Psychologists interest in the development of theory of mind skills began, not with
human beings, but with research on chimpanzees. In a study which has since
become a classic (Premack and Woodruff, 1978), an adult chimpanzee called
Sarah was shown a series of videotaped scenes in which a human actor struggled
with problems of different kinds. Some of the problems appeared simple, such as
trying to get hold of bananas that were out of reach. Others were more complex,
such as trying to escape from a locked cage. For each scene Sarah was presented
with two photographs that showed different actions, only one of which was an
effective solution to the problem. So, for the scene in which the actor was locked
in a cage, Sarah was presented with a photograph of a key and a photograph of a
solution to a different problem, for example, a stick to get the out-of-reach
bananas. On the overwhelming majority of occasions (21 out of the 24) Sarah
selected the correct solution. Premack and Woodruff concluded that the
chimpanzee understood the actors purpose and therefore understood the
contents of the actors mind.
t
Activity 1
Allow about
10 minutes
This activity will help you identify possible explanations for Sarahs behaviour.
From what you have just read of the experiment involving Sarah, do you think the results
indicated that the chimpanzee understood the actors purpose and the contents of the
actors mind? What other interpretation might there be?
Comment
The philosopher Daniel Dennett (1978) has provided an important line of criticism of
Premack and Woodruffs conclusions. Dennett suggested that their evidence did not show
that chimpanzees have an understanding of other peoples intentions, desires or knowledge.
Sarah could have shown that level of success on the task without reecting on the actors
mental states at all. Instead, she could have solved the problem by drawing on her knowledge
of the associations between objects in the real world, what is sometimes referred to as the
external contingencies of objects. So, for example, on seeing the actor in a cage, rather than
impute the actors mental state, i.e. her desire and intention to escape from the cage, Sarah
could see the cage and match this to its common partner, a key, as shown in one of the
solution photographs.
235
236
6 to 9 year olds
(n = 24)
92%
Extensive research on the development of theory of mind has been carried out
since Wimmer and Perners study. Improvements have been made to the tasks,
such as reducing the linguistic demands (for example, using simpler test
questions) which has resulted in enhanced performance particularly for younger
children. However, regardless of these modications, the age of 4 years still
remains as the point at which the critical change in childrens understanding of
false beliefs appears to occur. Before 4 years, children rely on the current state of
reality to answer questions in the unexpected transfer task. From about 4 years
on, children begin to hold representations of peoples mental states, reecting on
the belief of the story character and recognizing that that belief is false.
UNDERSTANDING MINDS
BOX 1
Figure 1
The unexpected
transfer task
(adapted from Frith,
1989, p. 160).
This is Anne.
237
238
Much of the research on the development of theory of mind has been based on
samples of children in Europe and North America. How do children in contrasting
cultures fare on the unexpected transfer task? Avis and Harris (1991) investigated
this with children of the Baka tribe, who live in the rainforests of south-east
Cameroon. The Baka are non-literate pygmies who have a hunter-gatherer
lifestyle with different values from those which predominate in Western societies.
The researchers acted out a version of the unexpected transfer task, adapted to be
more appropriate to the Baka children. In this version a member of the group
cooks some mangoes and leaves them in a cooking pot. While the cook is absent
the mangoes are transferred to a second, closed pot. The children were then
asked where the cook would look for his mangoes when he returned. The 5 and
6 year olds judged that the cook would look for the mangoes in the cooking pot.
Younger children made reality-based errors, saying that the cook would look for
them in the closed pot. Avis and Harriss results are very similar to those found by
Wimmer and Perner and other researchers, and they suggest that the
development of an understanding of false beliefs is not specic to one culture and
that important changes occur at a similar age for different children around the
world.
Metarepresentation
The ability to hold
a number of
representations of
a situation in
mind at the same
time.
The evidence presented so far indicates that before the age of about 4 years
children are unable to appreciate that other people can hold an incorrect
representation of the world. Young children are realists and believe that everyone
in the world behaves according to how the world truly is, rather than how each
individual represents the world to be. At about 4 years, children appear to
become able to hold a number of representations concerning a situation in their
mind at the same time that is they can meta-represent. For example, in the
unexpected transfer task 4 year olds can hold a representation about the true state
of reality (the ball is in the box) and also a different representation of a story
characters beliefs about the state of reality (the ball is in the basket). They can
compare and contrast these different representations and appreciate that
someone can hold a representation about the world that is incorrect.
False beliefs are a common feature of everyday life, for example searching for
something in one location when it is actually in another. Before the age of 4 years,
children observing such a scene would not appreciate why such mistakes occur.
But if children do not understand other peoples false beliefs, can they
nevertheless understand their own inaccurate beliefs about the world? This
question has been addressed using another ingenious procedure known as the
Smarties task, for reasons that will become apparent from the description in
Box 2. Less entertainingly, it is also referred to as the deceptive box test.
UNDERSTANDING MINDS
BOX 2
Figure 2
test.
Gopnik and Astington (1988) found that children older than about 4 years were able to
appreciate that Sooty would hold a false belief about the contents of the tube that it
contained sweets. However, younger children who are reality-biased, said that Sooty
would say there were pencils in the tube. Ingeniously, the researchers then asked the
children about their own, previous false belief, When you rst saw this tube and it was
all shut up like this, what did you think was inside? (Frame 5). Again, children of 4 years
or older were able to reect on their own mental states, and even though they now
knew the tube contained pencils, answered that they had previously believed that it
contained sweets. By contrast, the younger children stated that originally, before the
239
240
tube had been opened, they had believed that it contained pencils. These ndings
suggest that as well as being unable to reect on other peoples false beliefs, young
children remain reality-biased and are unable to reect on their own, previous false
beliefs.
As is often the case with research of this sort, questions arise as to whether the results
might be affected by the experimental procedures or other factors. So, further studies
using the Smarties task have concluded that the young childrens responses are not
attributable to difculties in understanding the specic moment of time to which the
question refers. This was established by asking, What did you think was in the box
before I took the top off? (Lewis and Osborne, 1990). Nor were the results due to the
childrens embarrassment at admitting their original ignorance. Wimmer and Hartl
(1991) introduced children to a silly puppet, who always made mistakes. Adding this
element to the design of the experiment enabled the researchers to work out whether
childrens incorrect answers were caused by their embarrassment at admitting their
previous ignorance, rather than lack of theory of mind skills. When the silly puppet was
introduced, if children were simply embarrassed about getting the question wrong,
they would fail questions about their own previous false belief by stating that they
always knew the contents of the tube. However, they would be expected to pass
questions regarding the false belief of the silly puppet, and to state that the puppet
previously believed that the tube contained Smarties since they would not be expected
to feel embarrassed about the silly puppets false belief. However, young children
without theory of mind skills were just as likely to get the false belief question wrong,
irrespective of whether they were asked about their own previous false belief, or the
false belief of the puppet.
n
The overall conclusion appears to conrm that before the age of about 4 years
children rely on reality to predict their own and other peoples behaviour.
However, from about 4 years onwards, they begin to appreciate that a
psychological level of information (the beliefs that people hold about the world)
dictates peoples behaviour and that sometimes these beliefs are wrong.
UNDERSTANDING MINDS
.
.
.
.
Theory of mind is measured using tests which involve false beliefs, for
example the unexpected transfer task and the deceptive box test.
Before about 4 years of age, children rely on reality to address questions
of false belief. After 4 years, children are able to reect on different
peoples representations of the world.
Research suggests that the development of theory of mind occurs across
cultures and at roughly the same age across the world.
Research has shown that not only are young children unable to reect on
the mental states of other people, they are also unable to reect on their
own previous false beliefs.
241
242
Second-order
theory of mind
The ability to
attribute beliefs
about beliefs or
beliefs about
intentions.
Thus far, this chapter has been about rst-order theory of mind, that is the
ability to appreciate and reect on the contents of another persons mind. Secondorder theory of mind is the ability to attribute beliefs about beliefs, or beliefs about
intentions.
Figure 3 illustrates the difference between rst-order and second-order theory
of mind skills. The boy on the left is simply thinking about taking the girls apple;
he is not using theory of mind skills. The girl, by contrast, is using rst-order
theory of mind skills as she is thinking about the boy on the lefts desire and
intention (mental states) to steal her apple. Unfortunately for her, the boy on the
right is using second-order theory of mind skills. He realizes that he can steal the
girls apple, for he believes (correctly, as it happens) that she is preoccupied with
thinking about the other boys intention to steal her apple.
Second-order theory of mind skills are believed to develop between the ages of
6 and 8 years.
A lie or a joke?
Do children need to have attained second-order theory of mind skills in order to
be able to distinguish lies from jokes? This question was answered by Sullivan
et al. (1995) in a study in which 48 children, whose ages ranged from 5 to 9 years,
were told four brief stories. Two of the stories assessed the childrens secondorder ignorance (for example, does John know that Mary knows X?) and secondorder belief understanding (for example, what does John think Mary thinks?). The
UNDERSTANDING MINDS
other two stories assessed the childrens ability to discriminate between a lie and a
joke. In one story a boy who did not clean his room lies to his mother by saying, I
did a really good job cleaning my room. In a second story a boy does not nish
eating his dinner and jokingly says to his mother, while they both sit at the table, I
did a really good job nishing my peas. At the end of each story the participants
were asked whether the story character was lying or joking. The results showed
that children typically were able to distinguish a lie from a joke only after they
could attribute second-order ignorance, but before they could attribute secondorder false belief. Therefore, the skill that appears to be crucial when
distinguishing a joke from a lie is the ability to appreciate and reect on what
different people know or do not know, rather than being able to reect on the
knowledge that people have about other peoples minds.
Such conclusions seem to make sense, since jokes differ from lies in terms of
the knowledge of the listener and the speaker. With jokes, the speaker knows that
the listener also knows the truth. In the example, both the boy and his mother
could see the peas remaining on his plate when he said, I did a really good job
nishing my peas. However, with lies the speaker and listener have different
knowledge and the speaker is aware of this difference; the boy who did not clean
his room knows that his mother does not know he has not cleaned his room when
he says, I did a really good job cleaning my room. Therefore, understanding lies
and jokes requires an understanding of another persons knowledge or ignorance
of facts rather than an understanding of someones beliefs about beliefs.
243
244
t
Activity 2
Allow about
5 minutes
This activity will help you to appreciate some of the difculties in establishing cause and effect.
Would you say that Astington and Jenkins conclusion was justied?
Comment
One problem with research that shows a correlation between two variables is that it is
difcult, if not impossible, to identify with any certainty whether an increase in one variable
causes an increase in the other. In this case, do enhanced theory of mind skills lead to an
improvement in social interactions, or are children who are more socially skilled more likely to
develop a theory of mind before those who are less socially skilled?
3.3 Bullying
The previous example of the relationship between theory of mind skills and
childrens interactions during play is positive in every respect. It would be
reassuring to think that the ability to appreciate the feelings and thoughts of
others would make children more altruistic and sensitive to those around them.
However, this is not always the case. The development of a theory of mind may
also bring with it an improved capacity for children to bully others and also to lie
and deceive. To be sure, children may lie, deceive or bully before they
understand mental states, but once they have developed a theory of mind they
become more sophisticated in the strategies that they can use. By being able to
appreciate another persons mind, bullies are better able to identify their victims
weaknesses and vulnerabilities and use this understanding to rene their bullying
strategies.
By contrast with the popular stereotype of an oash bully lacking in social
skills and understanding, bullies may be manipulative experts in social situations,
organizing gangs and using subtle, indirect methods to bully. One study
UNDERSTANDING MINDS
The Sally/Anne task and the Smarties task have been devised by psychologists as
techniques for studying the emergence of theory of mind in controlled,
experimental settings. But they are not the only way of gaining access to this
aspect of childrens development; this and the following section consider three
other methods.
1
245
246
t
Activity 3
Allow about
10 minutes
This activity asks you to analyse short extracts of childrens talk to identify patterns in their use of
desire-based and belief-based terms.
Look at the following six examples from four of the children in the study by Wellman and
Bartsch and identify which contain desire-based terms and which belief-based terms. Note
the ages of the children at the time of the recording: does any pattern emerge?
1 Adam (2 years 7 months)
Adam: Eat mommy.
Mother: Eat?
Adam: Yeah.
Mother: Im not hungry.
Adam: Want spoon?
Mother: No thank you.
Adam: OK. You dont want a spoon. You dont want a spoon.
2 Adam (3 years 3 months)
Adam: Can I put dis in de mail? Can I put my head in de mailbox ... so de mailman can know
where I are?
3 Ross (2 years 6 months)
Ross: He scratched me.
Adult:
Didnt it hurt?
Ross:
Adult:
The boy hurt you? How did the boy hurt you?
Ross:
UNDERSTANDING MINDS
Ross:
Adult:
Abe:
No dont be sad. I thought I would, cept I didnt see one for you.
Comment
Extracts 1 and 3 contain desire-based terms You dont want a spoon and I want a bandaid. Extracts 2, 4, 5 and 6 contain belief-based terms so de mailman can know where I are, I
know he could rip it, I thought I would and You dont know where the pieces go. The two
examples of desire-based terms come from Adam and Ross when they were younger by
several months than in the other extracts.
This hint of an age trend across these examples was borne out by the fuller
analysis of the data. Wellman and Bartsch found that genuine reference to the
subjective mental state of desire occurred around 2 years of age. At this age, not
only do children refer to their own desires but they are also able to refer to other
peoples desires, as Adam does in Extract 1. This use is distinct from some early
uses of the word want I want this, I dont want to that may not truly reect a
childs understanding of mind, but rather may simply be a tool they have learned
to use in order to obtain desired objects or not to have to undertake undesired
activities. These children may have no psychological understanding of the
psychological state desire, but may simply be using want because they have
built up a conditioned response which achieves desired goals. The use of
genuine reference at the beginning of this paragraph distinguishes those learned
responses from uses which indicate a true understanding of the mental state of
desire.
The Sally/Anne and Smarties tasks provide ways of establishing whether
children really do understand the psychological consequences of mental states
rather than simply rote learning mental state terms. Just as with analyses of
everyday language, studies employing these tasks show that children pass tests
for an understanding of desires before they pass tests for false belief
understanding. Research summary 1 provides an example of one such study by
Repacholi and Gopnik (1997).
247
248
RESEARCH SUMMARY 1
n
As can be seen from these examples, both naturalistic observations and
experimental paradigms have shown that there is a critical shift at about 18
months as children begin to appreciate that different people can have different
desires. It is likely that this understanding is an important stepping stone to
understanding false beliefs.
Joint attention
A notable feature of all infants experience is the conversation-like exchanges that
they get into with adults, typically from the age of 2 months. Child and adult
become the focus of each others attention and take turns in exchanges of looks,
mouth movements, noises and other recognizable expressions; psychologists call
this dyadic joint attention.
From about 9 months onwards, these sorts of exchanges take on a third
dimension, as objects typically toys are introduced into the conversation; this
is triadic joint attention.
UNDERSTANDING MINDS
n
Understanding intentions
The example of the bus queue, which began this chapter, illustrated that in order to
understand the intentions of another person it is necessary to impute desires and
goal states to their actions the people are standing, waiting, because they are
hoping that a bus will arrive. In order to understand that people have minds and
that their minds dictate their behaviour, it is necessary to appreciate that behaviours
are purposeful. If behaviours are not seen as purposeful then there would be no
need to try to explain them because they would appear to be unsystematic and
therefore meaningless. When do children show evidence of beginning to
understand the intentions behind peoples actions? Research summary 3 describes
a piece of research by Meltzoff (1995) which addresses this question.
249
250
RESEARCH SUMMARY 3
(a)
(b)
Figure 4 (a) The human experimenter attempts to pull the dumbbells apart but his ngers slip off the
end. (b) The mechanical arms and ngers located in front of the human experimenter fail in a similar
fashion to pull the dumbbells apart (from Meltzoff, 1995, p. 844).
UNDERSTANDING MINDS
.
.
t
Activity 4
Allow about
10 minutes
This activity will help you think about experiences that might encourage an understanding of other
peoples minds.
Think of some experiences within a young childs day-to-day environment that might
encourage the development of an understanding of other peoples minds. What might assist
them to stand in another persons shoes? And how, in turn, might these different factors
affect childrens cognitive development?
Note down your thoughts and review them as you read the rest of this section.
251
252
5.1
Social interaction
Low language
ability group
High language
ability group
0.3 (n = 13)
2.9 (n = 9)
1.4 (n = 15)
2.9 (n = 17)
3.2 (n = 4)
3.6 (n = 10)
t
Activity 5
Allow about
5 minutes
This activity will give you further experience of analysing and interpreting data.
Study the scores shown in Table 2. What conclusions can you draw from these results?
Comment
Where children have no siblings there appears to be a relationship between their level of
language ability and their competence at theory of mind tests those with lower language
ability do less well. However, when children have two siblings there is a clear indication that
this difference is reduced the presence of siblings appears to compensate for slower
language development in developing false belief understanding.
If siblings make a difference in this way it would seem plausible that their age
might be of signicance. When Ruffman et al. (1998) examined the results of 444
English and Japanese children whose ages ranged from 3 years and 1 month to 6
UNDERSTANDING MINDS
years and 11 months, they found that the number of theory of mind tests that they
passed increased with the number of older siblings a child had. However, there
was no connection between childrens success on the tests and either the number
of younger siblings or the gender of the siblings (whether older or younger). How
might the older siblings facilitate a childs theory of mind development? Ruffman
et al. point to activities such as pretend play, deception, teasing and talk about
feelings, all of which would feature less in interactions with younger siblings.
5.2 Communication
The role of language use in the family
What is it about the interactions with older siblings and adults that has a positive
effect on young childrens theory of mind development? Ruffman et al. (1999)
investigated whether mothers in some way facilitate false-belief understanding in
their children and, if so, what elements of their behaviour aid this facilitation. The
researchers used a questionnaire to ask the mothers of 64 children aged 3 and
4 years about their own education and occupation, the number of younger and
older siblings the child had, the time the child spent with them and how they
dealt with disciplinary situations involving their child. Three factors were found to
relate to the childrens belief understanding: the age of the child, the number of
older siblings, and the number of times mothers said they would respond to
disciplinary situations by asking the child to reect on the victims feelings, for
example Imagine how Billy feels when you hit him, or How would you feel if
Billy hit you? There was no effect for mothers who engaged their child in general
discussion or exploration of the disciplinary issues, or who simply reprimanded
the child without discussing the situation. This suggests that the way a mother
approaches disciplining her child can have an effect on the childs understanding
of feelings and mental states. It seems that asking children to reect on mental
253
254
Late signers
(n = 32)
Mean score on
theory of mind
tests
1.42
0.34
0.35
1.30
Standard
deviation
0.61
0.65
0.86
0.35
The data in Table 3 indicate that the children in the late-signing group showed
less well-developed theory of mind skills than either the native-signers (p <0.001)
or a sample of 4-year-old hearing children (p < 0.001). It would appear that access
to early conversation through whatever medium is an important factor in the
development of a theory of mind.
5.3 Gender
There is much research to show that the gender of a child can affect the types of
interaction they have with their environment. For example, mothers tend to talk
about emotions more to their 2-year-old daughters than to their 2-year-old sons
(Dunn et al., 1987). Older siblings tend to mention feeling states more frequently to
girls than boys (Brown et al., 1996). Girls also tend to use words which signify
emotional states earlier and more frequently than boys (Cervantes and Callanan,
1998). This suggests that girls might succeed on tests of false belief understanding
UNDERSTANDING MINDS
before boys. Evidence from nearly 1,500 children, aged between 2 years and 4
months and 6 years and 3 months, indeed showed a slight advantage for girls on
false belief understanding tests (Charman et al., 2002). However, this was only so
for children younger than 4 years and 8 months, not for children older than that,
and in any event the researchers concluded that if there is an age-specic
advantage for girls in the acquisition of false belief understanding it is only a very
weak effect.
255
256
Conclusion
At the beginning of this chapter you were asked to imagine what the world would
be like if we were not able to appreciate the minds of other people. It should now
be clear that a theory of mind is an exceptionally important ability that allows us
to perform complex behaviours and draws upon some fundamentally important
skills. Before the age of about 4 years, children do not fully appreciate other
peoples mental states and so they are unable to use this rich source of
psychological information to predict a persons behaviour. Young children
believe that people behave according to how the world really is, rather than how
each person believes the world to be. However, research using tests of false belief
understanding has shown that, at about 4 years, children begin to appreciate the
mental states of others. This change in cognitive skills appears not to be restricted
to Western cultures.
Having a new realm of information upon which to draw allows children to
develop many other skills, including improved interactions with others and the
appreciation of intentions in lies, jokes, and ironic or sarcastic comments. The
development of a theory of mind also provides children with the tools to bully
and lie more effectively.
The development of theory of mind skills draws upon many cognitive abilities,
including joint attention and the understanding of desires and intentions. Being
able to share ones perceptual focus with another person and being able to
appreciate the desires and intentions of others by watching their actions appear to
underpin the development of an understanding of false beliefs. Furthermore, the
environment within which a child lives can also affect theory of mind
development. The interactions that children have with their siblings, friends,
parents and other adults affect how they see other people. From about 4 years of
age a child sees other individuals not just as people, but as people who have
minds of their own.
References
Astington, J. and Jenkins, J. (1995) Theory of mind development and social
understanding, Cognition and Emotion, vol. 9, pp. 15165.
Avis, M. and Harris, P. (1991) Belief-desire reasoning among Baka children:
evidence for a universal conception of mind, Child Development, vol. 62,
pp. 4607.
Brown, J., Donelan-McCall, N. and Dunn, J. (1996) Why talk about mental states?
The signicance of childrens conversations with friends, siblings and mothers,
Child Development, vol. 67, pp. 83649.
Cervantes, C. and Callanan, M. (1998) Labels and explanations in motherchild
emotion talk: age and gender differentiation, Developmental Psychology, vol. 34,
pp. 8898.
UNDERSTANDING MINDS
Charman, T., Baron-Cohen, S., Swettenham, J., et al. (2000) Testing joint
attention, imitation, and play as infancy precursors to language and theory of
mind, Cognitive Development, vol. 15, pp. 48198.
Charman, T., Ruffman, T. and Clements, W. (2002) Is there a gender difference in
false belief development?, Social Development, vol. 11, pp. 110.
Dennett, D. (1978) Beliefs about beliefs, Behavioral and Brain Sciences, vol. 1,
pp. 56870.
Donaldson, M. (1978) Childrens Minds, Glasgow, Fontana.
Dunn, J., Bretherton, I. and Munn, P. (1987) Conversations about feeling states
between mothers and their young children, Developmental Psychology, vol. 23,
pp. 1329.
Frith, U. (1989) Autism: explaining the enigma, Oxford, Blackwell.
Gopnik, A. and Astington, J. (1988) Childrens understanding of representational
change and its relation to the understanding of false belief and the appearancereality distinction, Child Development, vol. 59, pp. 2637.
Jenkins, J. and Astington, J. (1996) Cognitive factors and family structure
associated with theory of mind development in young children, Developmental
Psychology, vol. 32, pp. 708.
Jenkins, J. and Astington, J. (2000) Theory of mind and social behavior: causal
models tested in a longitudinal study, Merrill-Palmer-Quarterly, vol. 46,
pp. 20320.
Lewis, C., Freeman, N., Kyriadidou, C., Maridaki-Kassotaki, K. and Berridge, D.
(1996) Social inuences on false belief access, Child Development, vol. 67,
pp. 293047.
Lewis, C. and Osborn, A. (1990) Three-year-olds problems with false belief:
conceptual decit or linguistic artifact?, Child Development, vol. 61, pp. 151419.
Meltzoff, A. N. (1995) Understanding the intentions of others: re-enactment of
intended acts by 18-month-old children, Developmental Psychology, vol. 31,
pp. 83850.
Piaget, J. (1932) The Moral Judgement of the Child, New York, Harcourt Brace.
Piaget, J. and Inhelder, B. (1968) The Psychology of the Child, New York, Basic
Books.
Premack, D. and Woodruff, G. (1978) Does the chimpanzee have a theory of
mind?, Behavioral and Brain Sciences, vol. 1, pp. 51526.
Repacholi, B. M. and Gopnik, A. (1997) Early reasoning about desires: evidence
from 14- and 18-month-olds, Developmental Psychology, vol. 33, pp. 1221.
Ruffman, T., Perner, J., Naito, M., Parkin, L. and Clements, W. (1998) Older (but
not younger) siblings facilitate false belief understanding, Developmental
Psychology, vol. 34, pp. 16174.
Ruffman, T., Perner, J. and Parkin, L. (1999) How parenting style affects false
belief understanding, Social Development, vol. 8, pp. 395411.
257
258
Sullivan, K., Winner, E. and Hopeld, N. (1995) How children tell a lie from a
joke: the role of second-order mental state attributions, British Journal of
Developmental Psychology, vol. 13, pp. 191204.
Sutton, J., Smith, P. and Swettenham, J. (1999) Social cognition and bullying:
social inadequacy or skilled manipulation?, British Journal of Developmental
Psychology, vol. 17, pp. 43550.
Wellman, H. and Bartsch, K. (1994) Before belief: childrens early psychological
theory, in Lewis, C. and Mitchell, P. (eds) Childrens Early Understanding of
Mind: origins and development, pp. 33154, Hove, Lawrence Erlbaum Associates.
Whiten, A. (1991) Natural Theories of Mind: evolution, development and
simulation of everyday mindreading, Oxford, Blackwell.
Wimmer, H. and Hartl, M. (1991) Against the Cartesian view on mind: young
childrens difculty with own false beliefs, British Journal of Developmental
Psychology, vol. 9, pp. 12538.
Wimmer, H. and Perner, J. (1983) Beliefs about beliefs: representation and
constraining function of wrong beliefs in young childrens understanding of
deception, Cognition, vol. 13, pp. 10328.
Woolfe, T., Want, S. C. and Siegal, M. (2002) Signposts to development: theory of
mind in deaf children, Child Development, vol. 73, pp. 76878.
Chapter 7
thinking
Contents
Learning outcomes
261
Introduction
1.1 Cognition and language in context
261
261
scientic knowledge?
262
268
268
270
271
274
278
280
281
282
286
292
Conclusion
297
Further reading
298
References
298
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 understand the difference between generative and reproductive
knowledge;
2 distinguish between the ways in which children think in mathematics and
science;
3 understand the importance of situations, and not just of computations, in
problem solving in mathematics;
4 assess the impact of cultural settings on mathematical and scientic
thinking;
5 discuss possible connections between the development of intelligence and
learning science and mathematics;
6 compare different theories about how children progress in their
mathematical and scientic thinking and the implications of this analysis for
education.
Introduction
261
262
t
Activity 1
Allow about
5 minutes
This activity will help you to appreciate the difference between reproductive and generative learning.
Study the numbers below, which are in Japanese but written phonetically in English.
1
ichi
10
ju
20
ni ju
ni
11
ju ichi
21
ni ju ichi
san
12
ju ni
22
ni ju ni
shi
13
ju san
go
14
ju shi
roku
sichi
hachi
ku
How do you say 15 in Japanese? How do you say 43? How do you say 67? Do you think you
could count to 99? How did you get your answers?
Comment
Your performance in Activity 1 can show you that you know more facts about Japanese
numbers than you were taught. Learning mathematics means understanding a system that
goes beyond the examples you learn from. For this reason mathematics learning is said to be
generative, that is, learning the system allows you to generate new facts that you were never
taught about. If mathematical and scientic knowledge were simply reproductive (the
repetition of what the person was taught), you could not have answered the questions in
Activity 1.
t
Activity 2
Generating numbers
Allow about
5 minutes
This activity encourages you to consider the need to look at learning processes (how children come to
understand something) rather than just at the outcomes (what they are able to do).
Look at Figure 1. It contains numbers written by four children aged between 5 and 6 years of
age. They had not been taught how to write multi-digit numbers in school. However, they
see numbers around them and do their own analysis of how numbers should be written. Can
you gure out how each one generates the writing of numbers? Which children seem to
think similarly about writing numbers? What similarities and differences are there?
(9)
(5)
(25)
(9)
(45)
(5)
(100)
(25)
(40)
(123)
(100)
(109)
(103)
(1000)
(109)
(2000)
(1000)
(1002)
(2009)
(1009)
(2568)
(1598)
one hundred
eight
two hundred
one hundred
and nine
five
one thousand
twenty-five
one thousand
and five
forty
two thousand
and fifty-three
263
264
Comment
In order to understand childrens learning of mathematics, it is necessary to analyse their
productions the way they count, write numbers, solve problems and so on and try to
gure out how they think. It is not enough to try to nd out what number facts they know or
do not know. Consider rst Alices production. Alice seems to have the idea that, for each
number word, you put down a digit. She does not know which digits she should put down for
some words for example, twenty, hundred and thousand so she uses one line for twenty
and two number-like shapes for hundred and thousand, respectively. Although not correct,
her writing of numbers is not random and can be understood if you crack her system. Luke
and Susanna seem to use similar systems. It is quite likely that they have learned how to write
two digit numbers so they get these right (if you disregard Lukes inversion of the digit 2). For
three digit numbers, they have created a system: they write each number in sequence, as if
they wrote the words in sequence. Megan succeeds in keeping three digit numbers within the
conventional writing but nds it hard to do so with four digit numbers. For these three
children, number learning is not a simple sequence to be memorized, with the larger ones
being learned later: they succeed with 2,000, yet do not write smaller numbers correctly. They
have a system, but it happens not to be the one that adults use.
It may seem that this notion of generative knowledge applies more readily to
mathematical than to scientic knowledge. It could be argued that there are many
scientic facts about the world that children can be taught in school and they will
learn them without difculty and this is to some extent true. However, scientic
information is often considered to be something that is transmitted in school as
mere fact, when this information is not only that. Often what people call
scientic fact is important because it reects a way of thinking. Try Activity 3
now and think about facts and ways of thinking in science.
t
Activity 3
Allow about
10 minutes
This activity will help you to consider the distinction between understanding a fact, and showing
awareness of the underlying principles that explain that knowledge.
It has been known now for a long time that the world is round and not at, as it was thought
in the past. This is a fact. But is this a mere fact or is it a way of thinking? This is what
Nussbaum, Novack and their colleagues at the University of Cornell decided to gure out
(Nussbaum and Novack, 1976; Nussbaum, 1985).
Look at the picture of the world in Figure 2. It illustrates the idea that there are two girls, one
who lives at the North Pole and one who lives at the South Pole. They have two bottles each,
one with a cork in and one that is open. Their bottles are half lled with juice. Can you draw
in the juice in their bottles?
Look at the second picture of the world in Figure 3. It shows three boys, one living at the
North Pole, one living on the Equator, and one living at the South Pole. They each have a ball
in their hand. What will happen to the ball if they drop it? Can you draw in the path the ball
will follow?
Comment
As with mathematical knowledge, scientic knowledge is not the accumulation of facts but is
dened by ways of thinking. Most children in primary school will know and tell you (if asked)
that the world is round. Yet, their thinking may be governed by the idea that the world is at,
with the sky above and the ground below. This is what their drawings of people on earth and
of rain clouds suggest (see Figure 4). Their answers to the questions in Activity 3 (Figure 5)
also indicate a way of thinking that is consistent with a at world conception.
(a)
(b)
Lucas
Frances
265
266
Lucas (6 years) and Frances (8 years) were asked (following Nussbaum and colleagues) to
imagine that they were astronauts and were looking at the world from a distance. What would
the world look like? The answer to this question typically is round or like a ball or something
of the sort and both Lucas and Frances answered in this way. They were then asked to draw
the world as it would be seen from the spacecraft. After this was accomplished, they were
asked to put in some people, some clouds and some rain in the picture. The children were
also asked to indicate the North and the South Pole as a preparation for the next activity.
Both drew a round world and both decided to put in some countries. Lucas put in a rather
good outline of South America and England plus what he thought was not a good outline of
Australia, but he wanted to have it in anyway (see Figure 4a). Frances put in some countries
but was not sure what they were supposed to look like (see Figure 4b). As you can see, apart
from these similarities, their drawings of people on earth, clouds and rain differ in signicant
ways. For Lucas, each country had its own cloud on top and its rain falling from top to
bottom. His round world contains people all standing in one direction with clouds above and
not all around. Even though 6-year-old children have been told that the world is round and
draw it round when asked to, they place people, clouds and rain in a way that is more tting
with a belief in a at world. They have learned the fact but not the way of thinking that goes
with it. For Frances, the conception of the round world prevails in this drawing; she also put in
some countries, each one with a rain cloud but the clouds are positioned all around the world.
INTERVIEWER TO FRANCES:
FRANCES:
Yes. [see (4b)] (Later, as she draws the South Pole) ... I used to think that people
would fall off the earth. I didnt know why they wouldnt fall off when they were
here.
No (laughing).
My father told me. Its because of gravity. (She draws the rain falling on the South
Pole and remarks that Isnt it funny to think of rain falling this way?)
The drawings by the children are consistent with the solutions they give to the
problems posed in Activity 3 presented in Figure 5. Lucass solution is typical for
children of his age. The drink in the uncorked bottles owned by the little boys at
the South Pole and at the Equator falls out of the bottles. The balls also fall the
same way as the liquid. Frances thought that the liquid would stay in the bottles
but was quite doubtful when giving her answer about the little girls liquid at the
South Pole. When asked about the ball, she commented it would go right down
into space. Frances seemed to alternate between a round and a at world way of
thinking.
So, what can be seen from this example is that knowledge of mathematics and
science is about knowing how to think appropriately about a problem, rather
than repeating facts that have been learned by rote. Moreover the example shows
that although children often know particular facts, they can be seen to apply their
knowledge in a way that betrays a lack of understanding about the underlying
laws that account for those facts.
We will now consider childrens development in understanding these laws in
relation to specic aspects of mathematical and scientic reasoning. We will turn
to mathematical reasoning rst, and consider the question of whether children are
always susceptible to these kinds of errors, or whether mathematical
understanding is mediated by certain contextual factors.
267
268
From the examples in Section 1 you might conclude that children are likely to nd
the acquisition of mathematical understanding difcult. This is because they need
to think about the facts and procedures that they know in a systematic way in
order to reveal their underlying uniting principles and rules in a way that will
enable them to make predictions about their environment. This type of reasoning
seems especially sophisticated. Jean Piagets studies of mathematical reasoning
appear to support this conclusion. However, subsequent research that has
explored the role of factors that appear to mediate childrens understanding of
mathematical concepts tells a slightly different story.
Concrete
operations
The stage at
which children
can think logically
about objects and
events in their
environment, and
represent them
symbolically.
However, they are
still unable to
generate
hypotheses
systematically
combining
possible different
situations and
outcomes.
coins would have increased the quantity. Through their own interactions with
objects, by changing displays in several ways and checking the results of such
changes, they would later on come to understand the invariance of number.
The discovery of the invariance of number in spite of spatial displacements that
make sets look larger or smaller (around the age of 6 or 7) was considered so
important by Piaget that it was taken as a mark of a new stage in childrens
intellectual development: the stage of concrete operations.
This was not the end point of development, though, for at this stage childrens
mathematical understanding was, according to Piaget, restricted to one-variable
problems for which addition and subtraction are sufcient. If 7-year-old children
have to solve a problem that requires establishing a proportional relationship
between two different variables, they are not able to cope. An example of
proportional relations is given in Activity 4 below.
t
Activity 4
Proportional relationships
Allow about
5 minutes
This activity will encourage you to test your understanding of the nature of proportional reasoning.
Here is an eel:
Comment
The information in the rst two examples should have indicated to you that there is an
underlying ratio of eel length to amount of food of 1 cm: 2 g, and this should enable you to
work out that Ernie needs 10 g of food each day. In a study by Inhelder and Piaget (1958)
children were told that eels have to be fed amounts of food that vary with their size. They
269
270
were asked to look at some values in both variables length of the eel and amount of food it
needed and then say how much food an eel of a certain length should receive. Children at
age 7 normally suggested higher values for longer eels but did not derive the values
systematically. In contrast, older children (about 1113 years of age) tried to establish a
relationship between the two variables and to use this relationship to deduce how much food
the longer eel would need.
Formal
operations
The stage at
which children
are able to reason
in an abstract way
without reference
to concrete
experience. They
can tackle
problems in a
systematic and
scientic manner
and are able to
generate
hypotheses about
the world based
on their
accumulated
knowledge.
Piaget and his colleagues (Inhelder and Piaget, 1958; Piaget et al., 1968)
considered the achievement of proportional reasoning as another landmark in the
development of intelligence. Proportional reasoning involves the recognition that
the relationship between two variables remains the same (invariant) although the
values in both variables are changed. This ability to establish an invariant relation
among relations is a higher order operation that, according to Piaget, indicates the
achievement of a new level of thinking, that of formal operations.
problem solving, they do not directly determine what a person will actually do.
Other aspects of the problem situation need consideration.
Gerard Vergnaud, a French developmental psychologist who studied with
Piaget, synthesized these ideas in his theory, which is known as the theory of
conceptual elds. According to Vergnaud (1985), in order to analyse mathematical
concepts, the invariant properties of the concept must be considered, as Piaget
proposed, and also the situations that give meaning to the concept and the symbols
used in its representation. He further proposed that in mathematics one should
think not of isolated concepts, but of conceptual elds where the different concepts
are connected to the same core invariants, situations and symbolic representations.
For example, he distinguished between the conceptual eld of additive reasoning
and the eld of multiplicative reasoning. Although there are connections between
these two domains of reasoning, Vergnauds view was that the differences between
them are important enough to require that childrens understanding in these two
domains must be analysed separately. We will use additive reasoning as an
example to illustrate our discussion of mathematical development.
t
Activity 5
Allow about
5 minutes
This activity will encourage you to reect a little on the nature of addition problems, and in particular
Read the arithmetic problems presented below. They have different levels of difculty for
young children. Try to gure out which one is easier, which is more difcult and why.
Problem 1
John had some marbles. He played with a friend and won four marbles. Now he has nine.
How many did he have before the game?
Problem 2
Mary had nine sweets. She gave four to her sister. How many does she have now?
Problem 3
Paul had nine buttons in his pocket. His pocket had a hole and some fell out. Now he has four
buttons. How many buttons did he lose?
271
100
Comment
Look at Figure 7 and check whether your predictions
were correct.
80
Percentage correct
272
60
40
20
Problem type
s
In view of the strong connection between addition and subtraction, it is now
considered more appropriate to discuss the development of additive reasoning as
a whole rather than the development of each of its parts (addition and
subtraction) in isolation. Much research has been carried out comparing the
difculties that children face in solving addition and subtraction problems
(Carpenter and Moser, 1982; Riley et al., 1983). This research shows that children
may know how to solve a particular numerical computation that is, they may
know, for example, that 9 4 = 5, but may still not be able to solve problems that
require just that computation. The understanding of addition and subtraction does
not depend only on knowledge of number facts but also on the childrens ability
to analyse the situations (Brown, 1981; Vergnaud, 1982).
The simplest additive reasoning situations are related to questions in which
elements are added to or taken away from groups or sets (referred to as
transformations). A related and similarly easy type of problem has to do with
joining or separating two sets and asking about the results of this union/
separation, for example In a family there are three girls and two boys. How many
children are there altogether? Problems that involve the comparison of two sets,
like Mary has ve books; Tom has three books; how many more books does
Mary have than Tom? are rather more difcult.
The analysis of the situation described in the problem is not sufcient to
characterize its level of difculty. The information that is unknown is also
important. For example, a transformation problem is very easy when the
information to be calculated is the result of the transformation (e.g. in Problem 2
of Activity 5, the calculation is 9 4 = ?). In this case, the action in the story and
the operation to solve the problem are directly related: a transformation that
increases the number will be solved by addition, and one that decreases the
number will be solved by subtraction.
In contrast, when the initial situation is unknown and must be calculated on the
basis of information about the transformation and its end result (for example, in
273
274
they relied mostly on written procedures, which proved much less efcient: their
rate of correct responses was 74 per cent to word problems and 37 per cent to
computation exercises.
The comparison between their successful performance in the streets and their
rather high rate of failure when solving computation exercises suggests that
symbolic systems are not merely accessories to peoples reasoning. They mediate
complex reasoning. In other words they are part of the reasoning process and as
such change the nature of the activity by their use (Luria, 1979). In a later study,
Nunes et al. (1993) analysed the inuence of symbolic systems not only on the
rates of correct and incorrect responses but also on the size of the errors made by
children when using either oral or written arithmetic. Three different error bands
were dened for this comparison: errors that fell within 10 per cent of the value of
the correct answer (for example, between 18 and 22 if the correct response to the
problem was 20); errors that were larger than 10 per cent but did not differ by
more than 20 per cent from the correct answer; and errors that differed from the
correct answer by more than 20 per cent. The percentages of correct responses
and errors in each error band for oral and written addition and subtraction are
presented in Figure 8. The gure clearly shows that written arithmetic led to larger
errors, a tendency that was supported by a statistical analysis of the association
between type of strategy used in solving the problem (oral versus written) and
error band.
100
Oral procedure
Written procedure
Percentage observed
80
60
40
20
0
(a)
Correct
Within
10%
Type of responses
Correct
(b)
Within
10%
Type of responses
Figure 8 The percentage of correct responses and errors for oral and written addition and
subtraction (data from Nunes et al., 1993, p. 47). (a) Percentage of correct responses and of errors
within each error band in addition problems. (b) Percentage of correct responses and of errors
within each error band in subtraction problems.
275
276
children seemed to keep the meaning of the problem in mind. For example, in
the division problem below, the boy clearly keeps in mind the fact that he is
trying to gure out how many marbles each of the ve children will get:
Child: F. Condition: Word problem. Computation: 75 7 5.
F: If you give ten marbles to each (child), thats fty. There are twenty-ve
left over. To distribute to ve boys, twenty-ve, thats hard (Experimenter:
Thats a hard one.) Thats ve more for each. Fifteen each.
(from Nunes et al., 1993, p. 43)
The references to marbles and children are clear throughout the problem solving
procedure. In contrast, in written arithmetic, references to the problem and even
to the relative values of digits are set aside. For example, when solving in written
arithmetic the same computation, 75 divided by 5, the digit 7 is spoken of as
seven, not seventy, which would take into account its relative value. This loss of
meaning in written arithmetic is probably one of the reasons for childrens
acceptance of responses to computations that would seem, under other
circumstances, unacceptable, as illustrated by this example:
Child: Ev. Condition: Computation exercise.
Computation: 100 7 4.
[After attempting unsuccessfully to solve the exercise on paper, Ev.
claimed that it was impossible. She rst attempted to divide 1 by 4, which
she decided was not possible, then to divide 0 by 4, and nally gave up.
The examiner asked for a justication.]
Ev: See, in my head I can do it. One hundred divided by four is twentyve. Divide by two, thats fty. Then divide again by two, thats twentyve. (She proceeded here by factoring; two successive divisions by 2
replace the given division by 4.)
(from Nunes et al., 1993, p. 43)
Such examples illustrate the relative difculty of written arithmetic for these
children. This is further illustrated by a study conducted by Nunes (1993)
described in Research summary 1.
RESEARCH SUMMARY 1
The participants in the study were distributed randomly to either an oral or a written
condition of solution that is, either they were not given paper and a pencil and had to
solve the problem in their heads or they were asked to write the information down
Especially informative was the performance of some children who had been assigned to
the written condition. They often made mistakes when solving the problem but then,
when attempting to explain their answer orally, realized their mistake and clearly
demonstrated their competence in understanding the cancellation of prots and debts.
One such example is presented below. It seems that the boy was limited by a poorly
learned written practice but, given the opportunity to think without having to
manipulate the written symbols, he was well able to make the appropriate inferences
about positive and negative numbers.
Problem: Seu Pedro (the farmers name) started the season with a debt of
10 cruzados (Brazilian currency at the time). He planted manioc
and beans. He gained 10 on the manioc and 20 on the beans.
What was his situation at the end of the season?
JC:
(19 years old, sixth-grader, assigned to the written condition) wrote line 1
without indicating whether the number was positive or negative, and
line 3 without a sign. Added all the numbers, obtaining 40, which was
written on line 4.
on paper.
I:
Why?
JC: He had a prot of 10, paid the 10 he owed. Then he still has his prot
from the beans, he has 20.
277
278
t
Activity 6
Allow about
5 minutes
This activity will encourage you to reect on the key differences between oral and written arithmetic
for these children, and thereby identify what kinds of factors might increase the ease or the difculty
of learning mathematics in school.
Consider the examples of oral and written arithmetic above and the account given in
Research summary 1. Why do you think that the children nd oral arithmetic easier than
written arithmetic?
Comment
One reason why oral arithmetic practices appear to be more successful for children than
written forms of mathematical problem solving may be because this form of reasoning appears
to retain the human sense of the problem. The problem retains its everyday, practical
meaning and is thought about with reference to quantities. However, by presenting the
problem on paper, its human sense is lost. Moreover, the process of translating the elements
of the problem into correct mathematical notation also presents the children with an
additional step in which errors and confusions can be introduced.
equal pretend cakes (cardboard rectangles) into halves, cutting one along the
height (resulting in two smaller rectangles) and the other along the diagonal
(resulting in two triangles). They ensured that the children realized that the two
parts of each cake were of the same size. They then asked the children: if one
child ate one of the rectangular parts and another child ate one of the triangular
parts, would they be eating the same amount of cake?
If the children understood the idea of half as the result of an exact division into
two, they should conclude that halves of two equal wholes indicate the same
amount irrespective of their appearance. However, children appeared to judge
the parts on their appearance and did not rely on the underlying logic of dividing
the whole. In a recent study (Nunes et al., 2002), we conrmed Piagets results
using a paper-and-pencil task: 45 per cent of the Year 4 children (aged about 8
years) and 20 per cent of the Year 5 children (aged about 9 years) did not
recognize that the two different looking halves were equivalent amounts of cake.
For concrete representations to be successful, there must be visual similarity
across each of the portions of the whole.
Representing fractions through conventional notation does not appear to make
the task any easier. Children have a great deal of difculty realizing that 18 is a
smaller number than 16. Ordering fractions in this way is much more difcult for
them than ordering whole numbers. By the age of 8 or 9, children have no
difculty at all in ordering numbers up to 100 or 1,000 by their size or indicating
the place of a whole number in a number line. In contrast, Kerslake (1986) found
that only one of fteen youngsters in the age range 1214 could correctly place 23
on a number line. Kerslakes study was carried out with a small sample, but the
results are in line with other research. Mack (1993) reports that the vast majority of
her sample of 12 year olds in the United States indicated that 16 was a smaller
fraction than 18 because 6 is smaller than 8. However, Mack reports that this was a
difculty observed only when the youngsters were presented with the problems
symbolically. If asked to imagine that a pizza was divided into six pieces and an
identical pizza was divided into eight pieces, and then to indicate which pieces
would be larger, those obtained from the rst or the second pizza, the youngsters
had no difculty in giving the correct answer.
Our recent investigations bring support to these results. We presented 142
youngsters in the age range 8 to 10 years with two comparable problems about
fractions. In one problem, the youngsters were told that two boys would share
fairly one pie and three girls would share fairly an identical pie. Most of the
youngsters correctly indicated that each boy would receive more pie than each
girl, with rates of correct response varying from 8996 per cent between the 8 and
the 10 year olds. When the youngsters were asked to compare the fractions 12 and
1
3, the rate of correct responses dropped to 11 per cent, 31 per cent and 59 per
cent, respectively, for the 8, 9 and 10 year olds. These results suggest that the way
in which number problems are represented is another powerful inuence on the
solutions that children will arrive at.
279
280
So far you have seen how the study of childrens mathematical development has
demonstrated that (1) the logical aspects of a concept (2) the situations that give
meaning to a concept and (3) the way that a problem is represented mediate
childrens understanding of mathematics and consequently affect their success.
These three aspects of concepts are also major determinants in childrens
acquisition of scientic concepts. However, there is an important difference in
childrens understanding of mathematics and science. Childrens everyday
experiences with mathematical concepts and representations set them on the right
track to understanding concepts in mathematics. In contrast, childrens everyday
experiences in their environment often conict with their learned knowledge of
scientic concepts. This section discusses obstacles that children face when their
everyday conception of the world conicts with their scientic ideas.
t
Activity 7
Allow about
5 minutes
This activity illustrates that it may be possible to show an implicit appreciation of principles relating to
physics, without being able to demonstrate an explicit awareness of them.
Find a room or other open space and place a box or bin in the centre. Make a paper ball or
nd a similar object. Now move quickly past the box and drop the ball into it as you pass by.
Can you do it?
Now, imagine you are running across a room and want to drop (not throw) the ball into the
box. Where should you be when you drop the ball in order for it to go into the box? Look at
Figure 9 and choose which picture correctly represents where you should be when you drop
the ball.
Figure 9 Where
should you drop the
ball if you want it to
fall into the box?
(a)
(b)
(c)
Comment
You should nd that the physical task is quite easy. However, it is likely that you have a wrong
idea of where you needed to be to get the ball in the box. The correct answer is to release
the ball before you get to the box (picture a). Although you may have found the question
difcult to answer because it demands an explicit awareness of the physical forces acting on
the ball, intuitively you knew when to release the ball when you were required to do so for
real.
The task you tried in Activity 7 was studied by McClosky et al. (1980), who asked
college students in physics and other students to say where they should drop the
ball from if they were running across a room and wanted to make sure it landed
in the box. The students were asked to answer by choosing the correct picture
from a set of three similar to the ones in Figure 9.
Physics students were signicantly more accurate in answering this question
although, as McClosky and his colleagues suggested, they were unlikely to be
281
282
more successful in actually dropping the ball into the basket. Whereas 73 per cent
of the physics students chose the picture in which the stick gure dropped the
ball before reaching the target, only 13 per cent of the non-physics students made
this correct choice. Among the physics students, the answers were divided
between the correct choice and the picture where the ball was dropped directly
above the box (27 per cent of the physics students chose this alternative). In
contrast, 80 per cent of the non-physics students chose the picture of the stick
gure dropping the ball directly over the target and there were also choices in
which the ball was dropped after the gure had already run past the box (7 per
cent). These results seem to indicate that physics students were thinking about the
action differently from the non-physics students with respect to force and motion.
The physics students took into account the fact that the ball was already in motion
when carried across the room by the running stick gure and thought of its path
as it was dropped as the result of this initial force and direction plus the effect of
gravity, whereas the non-physics students disregarded the motion of the ball
while it was in the stick gures hand, and thought only about gravity.
This illustrates a key point about understanding scientic concepts: while
people can operate successfully within the physical world, they often do so
without a genuine understanding of the physical principles that underpin it.
Scientic learning is often a challenge to everyday understanding of cause and
effect.
everyday life with a world that is continuous, in which objects are solid and
undivided. Yet, to understand many of the changes they observe in the world, it is
necessary to develop a way of thinking that describes solid objects as bundles of
particles, molecules, or atoms that is, discontinuous elements that are
somehow kept together.
Piaget and Inhelder (1974) were pioneers in the investigation of childrens
understanding of the particulate nature of matter. They set out a pattern of
investigation by pointing out that it is when children have to understand change
that they come to invent an atomic theory about the world. Piaget and Inhelder
asked children to explain what happened to sugar when it was put into water and
then stirred. Whereas the younger children seemed to believe that the sugar
somehow disappeared, those at the ages of 1113 were aware of the fact that if
the taste of sugar remained in the water, then the sugar itself must still be present
in some form. This permanence of a property of sugar its taste contradicted
the apparent disappearance of sugar from the viewpoint of the older children. In
order to eliminate the contradiction between the disappearance of sugar and the
preservation of one of its properties the sweet taste the older children
invented an atomic theory about physical quantities.
Many studies have rened Piagets original ideas about childrens
understanding of the particulate nature of matter. Driver et al. (1985) and Hesse
and Anderson (1992), for example, studied other chemical transformations such
as combustion (what happens when wood is burned?) and oxidation (rusting and
cleaning bits of iron) and also investigated pupils conceptions when they
observed other changes of state, such as evaporation. An example of the type of
responses observed by Driver et al. (1985) is presented in Figure 10, where one
child is shown attempting to illustrate what happens to particles in each of the
three states of matter: solid, liquid and gas. The gure illustrates the difculty of
moving from a world of
observables to a conceptual
world of non-observed entities.
It also illustrates childrens own
conceptions of changes of state
rather than taught solutions.
A third type of difculty lies
in the everyday world view of
things as stable, with properties
that are part of the very matter
of which they are made. As
discussed earlier, Piaget and his
colleagues hypothesized that
children develop a theory about
the particulate nature of matter
in their attempt to explain the
sweetness of the water after the Figure 10 A drawing of an 11 year olds
disappearance of sugar: if a
representation of the three states of matter: solid, liquid
property of sugar, sweetness,
and gas (Driver et al., 1985, p. 145).
283
284
still exists in water, then the sugar must still be there. A conception of the world
that only deals in terms of matter and its properties leaves a range of phenomena,
such as light, heat and electrical current, unexplained. Chi and her colleagues
(Chi and Slotta, 1993; Chi, 1997) pointed out that not all scientic ideas are based
on the notion of matter: many important scientic ideas are based on the notion
of processes. These researchers suggest that the main obstacle to childrens
understanding of these scientic ideas is that they need to switch categories, from
matter to processes.
Examples of this kind of difculty are easy to nd (Reiner et al., 2000). For
example, children may have a natural tendency to treat heat and temperature as
substances with the essential property of hotness. Certainly young children nd it
hard to think of heat and cold as existing independently of objects and thus treat
temperature as a property of objects. Similarly, Clement (1982) has provided
evidence that young children treat forces as properties of objects. In the case of
inanimate objects, children see forces as acquired properties that cause
movement; and think that objects stop when this acquired property dissipates in
the environment (Vosniadou et al., 2001).
A fourth obstacle to the development of scientic thinking is the fact that many
scientic concepts represent intensive rather than extensive quantities. Extensive
quantities, such as height and weight, are measured through the simple
application of a unit of measurement that is repeated until the number of units
completely describes the quantity. For example, when length is measured in
centimetres, the value that is obtained by this measurement is equivalent to the
number of times 1 cm that is the unit can be tted onto the object measured.
In contrast, intensive quantities, which include density, speed and force, are
measured through a ratio between two other measures. So, density is measured as
a ratio of mass to volume. Similarly, speed involves a relation between distance
and time, and force involves a relation between mass and acceleration.
An everyday intensive quantity is taste. For example, how sweet a glass of
lemon juice will taste depends on how much sugar you add to the lemon juice
and how much lemon juice there is in the glass. There is a direct relation between
the amount of sugar and the sweetness: the more sugar you add, the sweeter the
lemon juice will taste. There is also an inverse relation between the amount of
lemon juice and sweetness: the more lemon juice there is in the glass, the less
sweet it will taste. The results of the investigation reported in Research summary 2
illustrate the difculty of inverse relations for children.
RESEARCH SUMMARY 2
Intensive quantities
Nunes et al. (2002) conducted an investigation of childrens understanding of inverse
relations. The children (142 in total) were shown pictures of glasses of lemon juice, as
those presented in Figure 11.
The children were asked rst whether they thought that the lemon juice would taste
the same in the two glasses. Those children who thought that the taste would be
different were then asked to say which would taste sweeter. The percentages of
correct responses in Table 1 refer to the total number of children by age for both
questions.
Table 1 Percentage of children who correctly responded to the questions
about the relative sweetness of the two glasses of juice
Age in years
7
44%
70%
60%
44%
40%
51%
Two results should be pointed out from the data in Table 1. First, there were no
signicant differences between the age groups; so the small improvement in percentages
with age could be due to chance. Thus the problem is difcult even for children at age 9.
Second, quite a few 8 year olds realized that the taste would be different in the two
glasses but made the mistake of thinking that the more juice, the sweeter it would taste.
Rather than thinking of an inverse relation, they assumed that the amount of juice and
the amount of sweetness would be directly related. In fact, when children are asked to
respond to situations where all that matters is the variable that is directly related to the
intensive quantity, their performance is perfect. When we asked them to say which
lemon juice would taste sweeter when the two glasses had the same amount of juice and
different amounts of sugar, the rate of correct responses was close to 100 per cent.
These results were replicated in an earlier study by Desli (unpublished).
285
286
The intensive quantities that occur in science are usually more complex than
those encountered in everyday life because, as we showed earlier, they require
distinctions that people do not make in everyday life. It is likely that science
teaching involving intensive quantities could prot from the use of analogies with
simpler intensive quantities used in everyday life, in the same way that
mathematics teaching seems to prot from the use of everyday situations as
models for teaching about mathematical concepts and operations. This is a
reasonable hypothesis, but further research is needed to test it.
Comment
Vygotsky (1978) attempted to explain this difference between everyday and scientific
concepts by suggesting that people learn scientific concepts by going from the
general to the particular (from rules, to specific instances of that rule), but that
everyday concepts are learned by going from the particular (specific instances of an
event) to the general (generating the rule). Thus in school children begin learning
about density with the rule that in order to compare the density of two materials
they must keep their volume the same; then they learn the rule that objects with a
mass that is greater than that of an equivalent volume of water will be of greater
density and thus will sink. In contrast people do not start from rules in everyday life:
they attempt to make generalizations on the basis of sets of observations. Such
generalizations will be limited by their experiences, which may not provide the basis
for the most parsimonious rule.
A second formal difference is that scientists strive for consistency and are (or
should be) willing to reject their explanations if they are inconsistent with
available evidence or with other valid theories. In everyday life people may not
be so ready to use consistency as a criterion for validating their knowledge. For
example, Hatano (1993) has observed that young children can respond correctly
to some questions about heat, but give different answers to other problems which
are inconsistent with their earlier responses. If asked what they would do in order
to cool the water in a bath, even young children will say that they would add cold
water to it and thereby treat the resulting temperature in the bath as being
somewhere between that of the hot and the cold water. However, when asked
what is the temperature of the water in a container into which one litre of water at
the temperature of 60 degrees and one litre of water at the temperature of 20
degrees were poured, the most frequent answer given by young pupils is 80
degrees. Whereas in the rst problem about the temperature of the water in the
bath the resulting temperature is intermediate between hot and cold, in the
science classroom problem the resulting temperature is viewed as the sum of the
temperature in the two volumes of water. On the basis of a different set of
examples, di Sessa (1993) also argued that childrens conceptions lack the
coherence of scientic theories. However, there is disagreement on this issue.
Some researchers suggest that childrens theories are coherent even though they
are less powerful than adults theories (Posner et al., 1982).
The physicists conception of the world differs from that of the child.
Some of the difculties involved in applying learned scientic concepts to
everyday ways of thinking about the world result from having to make
distinctions that people would not ordinarily need to make. Other
hurdles come from the fact that scientic concepts often require
reasoning about non-perceptible aspects of the physical world.
Children may experience difculty in learning about scientic concepts
because the knowledge and conceptions they already have conict with
the concepts being introduced by the teacher. One of the generic
differences between naive conceptions and scientic views is that
children often explain events in everyday life by referring to properties of
objects, related to their matter, whereas the scientic view treats the same
phenomenon as a process.
Many scientic concepts represent intensive, rather than extensive,
quantities. Intensive quantities are measured through a ratio between two
other measures, whereas extensive quantities are measured directly,
through the repeated application of a unit to the dimension that is being
measured. Intensive quantities have different properties from extensive
quantities and involve reasoning about inverse relations.
Scientists operate on the principle of parsimony, that is explaining the
largest number of phenomena with the smallest set of assumptions.
Vygotsky suggests that reasoning can be more parsimonious when
concepts are acquired from the general to the particular than when they
are acquired from the particular to the general. Scientists also strive for
consistency. This is not necessary in everyday life.
287
288
correlation relied on those items which related to reasoning that Inhelder and
Piaget expected to develop only at the level of formal operations, whereas those
that can be solved at the concrete operational level were not related to success in
science.
t
Activity 8
Understanding correlations
Allow about
5 minutes
This activity will help you to understand the importance of reecting on the nature of the evidence
that theoretical claims are based on.
What problems can you see in interpreting Piburns results?
Comment
Piburns results cannot be used on their own to infer that propositional reasoning ability
contributed to the students attainment in science. This is because their grades and their
propositional reasoning ability were obtained at the same point in time. While there may be a
relationship it is not possible to identify the direction of that relationship: it could equally be
the case that the students science grades contributed to their propositional reasoning ability.
To determine the direction of any relationships a longitudinal study is required, in which the
science learners are tested for their propositional reasoning before they start the science
course.
The second prediction from Inhelder and Piagets theory is about the
generalization of formal operational reasoning across contexts. The formal
operational stage was conceived by Piaget (Piaget and Garcia, 1971) as a stage in
which the operations of thought (such as propositional reasoning) become
abstract and therefore independent of their content. This means, according to
Piaget, that thinking can now function in a formal way, which makes it possible
for children to learn about any scientic domain without the development of new
operations of reasoning. This prediction about the generalization of the structures
of thought has met a growing degree of scepticism.
Specically with respect to propositional reasoning, much work has been
carried out that indicates its susceptibility to inuences from the content of the
proposition. For example, Cheng and Holyoak (1985) and Girotto et al. (1988)
looked at young peoples reasoning about several types of ifthen propositions.
They have suggested that people reason about these propositions not in a formal,
content-free way, as originally suggested by Piaget and his colleagues, but rather
they rely on pragmatic schemas that involve both formal and content specic
elements. For example, when the ifthen proposition is a prohibition people are
likely to make inferences that differ from those made when the proposition is a
promise. Thus the content of a proposition has a clear inuence on the inferences
that are made even if the propositions have the same form. Activity 9 presents an
example of the contrast between prohibitions and promises.
289
290
t
Activity 9
Allow about
5 minutes
This activity will help you to see how the content of a proposition can affect the way that people
reason about it, as they often rely on their everyday experiences to guide them to an appropriate
conclusion.
Here are two types of ifthen propositions: promises and prohibition/permissions.
1
If you mow the lawn, I will give you 5. You didnt mow the lawn. Do I give you 5?
If you drive a lorry, you must not drive through the city centre. You drive a car. Can you
drive through the city centre?
Comment
In the rst proposition, you probably concluded that you were not given 5 because you did
not mow the lawn. Although this conclusion seems logical based on your everyday
experiences of promises like this one, in fact it is not the logical conclusion. Just because you
did not do the task, it does not follow (in strict logical terms) that you will not be given 5. In
fact, the proposition does not specify what will happen in this circumstance, only what will
happen when you do mow the lawn. Similarly, in the second, prohibitive statement, you cannot
logically conclude that you can drive through the city centre. The proposition simply states
that you cannot drive through the city centre if you are driving a lorry. Although your common
sense tells you that you can imply that you are permitted because you are not explicitly
prohibited (they would have mentioned it if I was not!), in fact you cannot come to that
conclusion on the basis of the proposition alone. From these examples you can see that when
reasoning about promises or prohibitions, people rely on their cultural knowledge to draw
their conclusions, rather than on strict logical deduction.
formulae, the terminology, and yet still maintain their previous conceptions. This
difculty has been illustrated many times, for example, when instructed students
are interviewed about heat and temperature or forces, as discussed earlier. It is
often identied by teachers as a difculty in applying the concepts learned in the
classroom; students may be able to repeat a formula but fail to use the concept
represented by the formula when they explain observed events.
Piaget suggested an interesting hypothesis relating to the process of cognitive
change. Cognitive change was expected to result from the pupils own intellectual
activity. When confronted with a result that challenges their thinking that is when
faced with conict pupils realize that they need to think again about their own
ways of solving problems, regardless of whether the problem is one in mathematics
or in science. Conict was hypothesized to bring about disequilibrium and would
result in the setting off of equilibration processes that would ultimately produce
cognitive change. For this reason, according to Piaget and his colleagues, in order
for pupils to progress in their thinking they need to be actively engaged in solving
problems that will challenge their current mode of reasoning.
However, Piaget also pointed out that young children do not always discard
their ideas in the face of contradictory evidence. They may actually discard the
evidence and keep their theory. Activity 10 describes a science lesson where
children aged about 6 were asked to observe whether objects oated or sank in
water and later in the afternoon were interviewed about what they had learned in
their science lesson.
t
Activity 10 Do heavy things sink and light ones oat?
Allow about
10 minutes
This activity will encourage you to reect on the problematic nature of supporting conceptual change
in children.
Read the transcript below and note how Billy integrates the evidence that he collects (his
observation of the oating carton) with his existing theory that heavy things sink.
The rst of the interactions was recorded during a science lesson in which children were
asked to make predictions about whether some things would oat or sink when put in a
basin full of water. The second is an excerpt of an interview carried out with one of the pupils
after the lesson.
TEACHER: You have to tell us what youve chosen (to place in the water).
BILLY: Ive got a milk.
T: Youve chosen a carton of a milk.
B: And its a bit heavy.
T: Yeah!
B: And you can drink out of it and eh ...
T:
T: Its a carton. Whats a carton made from? ... Do you know what cartons are made from?
Louisa?
291
292
Cardboard.
T:
B:
Sinks.
B:
It oats.
T:
B:
I:
B:
It was a bit light ... its light ... because its little and light.
Comment
What do you think of Billys conclusion after the lesson: did he discard his explanation for why
some things oat and others sink? What does this imply about childrens reaction to conicting
evidence? It would seem that Billys theory that heavy things sink remained intact, despite the
heavy milk carton oating. To account for this, he simply re-categorized the carton as light in
his explanation. In this way, Billys theory remained unchanged but his view of the evidence
was altered.
Comment
Although the work on peer collaboration can be seen as initially stemming from the
Piagetian ideas about the processes of conflict and equilibration, much of the interest
that currently surrounds peer collaboration in problem solving is motivated by other
views of conceptual development. Within these other perspectives, motivated by
researchers such as Vygotsky, Luria and Bruner, conceptual change is viewed as
resulting not solely from the childs own activity but also from the childs interactions
in a social world.
293
294
A new concept was developed by Vygotsky in order to address the social nature
of cognitive change: the concept of the zone of proximal development. The zone
of proximal development is the difference between what a child can accomplish
in solving a problem working independently and what can be accomplished by
the same child with adult help.
More generally, the concept refers to an interactive system within which
people work on a problem which at least one of them could not, alone,
work on effectively. Cognitive change takes place within this zone, where
the zone is considered both in terms of an individuals developmental
history and in terms of the support structure created by the other people
and cultural tools in the setting.
(Newman et al., 1989, p. 61, our italics)
out by the children. She thus structured the activity so that the children could mix
the chemicals, compare results, and discuss what they had seen without being
unduly pressured by the planning of the mixing of chemicals. The planning of the
experiment was off-loaded by the teachers assignment of chemicals (one to
each child) so that they could (as a pair) effectively generate a full table of mixes
and use the results to come to conclusions. This example illustrates how pupils
were allowed to use some aspects of a system of knowledge without its full
appreciation. Thus although the design of the experiment may have been beyond
their level of cognitive development, they were still able to participate in a
research activity, recording observations, comparing results, and trying to come to
conclusions that would be relevant to the learning of chemical mixtures.
t
Activity 11
Allow about
10 minutes
Whats science?
Whats science?
TEACHER: Whats science? Does anybody have any ideas what science is? ... Marcia.
MARCIA:
TEACHER: Like when you work with the sand. What sort of ... What sort of sand work is
science work, because sometimes I get you to work with the sand for number,
dont I?
BILLY:
Em ...
TEACHER: What sort of things do I ask you to do of ... say ... if it was science and I asked
you to go and work with the sand? What sort of things would you be doing?
CAIN:
Measuring work.
TEACHER: Em ...
CAIN:
I know.
TEACHER: Sh ... we have hands, dont we? (Shushing Cain, who started the whole thing.) Paul.
PAUL:
TEACHER: Thats more a number job, isnt it? Excuse me, Sarah. ... Jodie.
JODIE:
295
296
TARA:
Adds up.
RICHARD: If you could see whether a thing was sinking or oating in the water.
CAIN:
Being quiet.
TEACHER: Tara.
TARA:
TEACHER: Science is sort of what we call investigating. When were seeing whats
happening ... you were going to say that were you, I thought you might be about
to say that. When you think ... I wonder if such and such happens? I wonder
what will happen if? Those are the sorts of science questions. I wonder ...
Comment
Working in the zone of proximal development means developing some common ground for
understanding science as children gain experience under the teachers guidance. The transcript
is an illustration of just how difcult it is for children to gain an understanding of what science is.
They participate in concrete activities that are not direct illustrations of what science is. The
teachers view of science is about asking questions and testing ideas in a systematic way. The
children only understand science as a label that describes activities that they do in school. The
teacher could orientate the children to her viewpoint by setting up a session such that the
children have to generate some I wonder questions for themselves which the teacher could
guide them through answering.
There is some support for the idea that, as childrens reasoning becomes
more sophisticated, they may become better learners in science.
There is little support, however, for the idea that young people reach a
stage of formal operations where sophisticated forms of reasoning, such
as propositional reasoning, are easily applied to any content.
The search for what promotes cognitive changes in the development of
mathematical and scientic concepts has generated at least three main
hypotheses that are alternatives to the traditional way of teaching:
discovery learning, in which children are active but work mostly on their
own; collaborative learning, in which peers work together to solve
problems; and the model of learning as appropriation, in which a system
of knowledge is progressively mastered as the learner uses it effectively,
although initially without a full appreciation of what the system does for
him or her.
Conclusion
297
298
Further reading
References
Behr, M., Lesh, R. Post, T. R. and Silver, E. A. (1983) Rational number concepts, in
Lesh, R. and Landau, M. (eds) Acquisition of Mathematical Concepts and
Processes, pp. 91126, New York, Academic Press.
Brown, M. (1981) Number operations, in Hart, K. M. (ed.) Childrens
Understanding of Mathematics: 1116, pp. 2346, London, John Murray.
Bryant, P., Christie, C. and Rendu, A. (1999) Childrens understanding of the
relation between addition and subtraction: inversion, identity and
decomposition, Journal of Experimental Child Psychology, vol. 74, pp. 194212.
Carpenter, T. P. and Moser, J. M. (1982) The development of addition and
subtraction problem-solving skills, in Carpenter, T. P., Moser, J. M. and Romberg,
T. A. (eds) Addition and Subtraction: a cognitive perspective, pp. 924, Hillsdale,
NJ, Lawrence Erlbaum Associates.
Carraher, T. N., Carraher, D. W. and Schliemann, A. D. (1985) Mathematics in the
streets and in schools, British Journal of Developmental Psychology, vol. 3,
pp. 219.
Cheng, P. W. and Holyoak, K. J. (1985) Pragmatic reasoning schemas, Cognitive
Psychology, vol. 18, pp. 293328.
Chi, M. T. H. (1997) Creativity: shifting across ontological categories exibly, in
Ward, T. B., Smith, S. M. and Vaid, J. (eds) Creative Thought: an investigation of
conceptual structures and processes, pp. 20934, Washington, DC, American
Psychological Association.
Chi, M. T. H. and Slotta, J. D. (1993) The ontological coherence of intuitive
physics, Cognition and Instruction, vol. 10, pp. 24960.
Clement, J. (1982) Students preconceptions in introductory mechanics,
American Journal of Physics, vol. 50, pp. 6671.
Desli, D. (1999, unpublished) Childrens understanding of intensive quantities,
PhD thesis, Institute of Education, University of London.
di Sessa, A. (1993) Toward an epistemology of physics, Cognition and
Instruction, vol. 10, pp. 105225.
Doise, W. and Mugny, G. (1984) The Social Development of the Intellect, Oxford,
Pergamon.
Driver, R., Guesne, E. and Tiberghien, A. (1985) Childrens ideas and the learning
of science, in Driver, R., Guesne, E. and Tiberghien, A. (eds) Childrens Ideas in
Science, pp. 14569, Milton Keynes, Open University Press.
Erickson, G. and Tiberghien, A. (1985) Heat and temperature, in Driver, R.,
Guesne, E. and Tiberghien, A. (eds) Childrens Ideas in Science, pp. pp. 5284,
Milton Keynes, Open University Press.
Gelman, R. and Meck, B. (1992) Early principles aid initial but not later
conceptions of number, in Bideaud, J., Meljac, C. and Fishcher, J.-P. (eds)
Pathways to Number: childrens developing numerical abilities, pp. 17189,
Hillsdale, NJ, Lawrence Erlbaum Associates.
Ginsburg, H. P., Posner, J. K. and Russel, R. L. (1981) The development of mental
addition as a function of schooling and culture, Journal of Cross-cultural
Psychology, vol. 12, pp. 16379.
Girotto, V., Light, P. H. and Colbourn, C. J. (1988) Pragmatic schemas and
conditional reasoning in children, Quarterly Journal of Experimental Psychology,
vol. 40, pp. 46982.
Groen, G. J. and Resnick, L. B. (1977) Can pre-school children invent addition
algorithms?, Journal of Educational Psychology, vol. 69, pp. 64552.
Hatano, G. (1993) Childrens concepts of heat and temperature, paper presented
at SRCD Biennial Meeting, New Orleans, March 1993.
Hesse, J. J. and Anderson, C. W. (1992) Students conceptions of chemical
change, Journal of Research in Science Teaching, vol. 29, pp. 27799.
Howe, C., Tolmie, A. and Anderson, A. (1991) Information technology and group
work in physics, Journal of Computer Assisted Learning, vol. 7, pp. 13343.
Howe, C., Tolmie, A. and Rodgers, C. (1992) The acquisition of conceptual
knowledge in science by primary school children: group interaction and the
understanding of motion down an incline, British Journal of Developmental
Psychology, vol. 10, pp. 11330.
Inhelder, B. and Piaget, J. (1958) The Growth of Logical Thinking from Childhood
to Adolescence, London, Routledge.
Kerslake, D. (1986) Fractions: Childrens Strategies and Errors: a report of the
strategies and errors in secondary mathematics project, Windsor, NFER-Nelson.
Kieren, T. (1988) Personal knowledge of rational numbers: its intuitive and
formal development, in Hiebert, J. and Behr, M. (eds) Number Concepts and
Operations in the Middle-grades, pp. 5392, Reston, VA, National Council of
Teachers of Mathematics.
Lave, J. (1988) Cognition in Practice: mind, mathematics and culture in everyday
life, Cambridge, Cambridge University Press.
Luria, A. (1979) Curso de Psicologia Geral, Rio de Janeiro, Civilizacao Brasileira.
299
300
Reiner, M., Slota, J. D., Chi, M. and Resnick, L. B. (2000) Naive physics reasoning:
a commitment to substance-based components, Cognition and Instruction,
vol. 18, pp. 134.
Riley, M. S., Greeno, J. G. and Heller, J. I. (1983) Development of childrens
problem-solving ability in arithmetic, in Ginsburg, H. P. (ed.) The Development of
Mathematical Thinking, pp. 153200, New York, Academic Press.
Vergnaud, G. (1982) A classication of cognitive tasks and operations of thought
involved in addition and subtraction problems, in Carpenter, T. P., Moser, J. M.
and Romberg, T. A. (eds) Addition and Subtraction: a cognitive perspective,
Hillsdale, pp. 3959, NJ, Lawrence Erlbaum Associates.
Vergnaud, G. (1985) Concepts et schemes dans une theorie operatoire de la
representation, Psychologie Francaise, vol. 30, pp. 24552.
Vosniadou, S., Ioannides, C., Dimitrakopoulou, A. and Papademetriou, E. (2001)
Designing learning environments to promote conceptual change in science,
Learning and Instruction, vol. 11, pp. 381419.
Vygotsky, L. S. (1978) Mind in Society: the development of higher psychological
processes, Cambridge, MA, Harvard University Press.
301
Chapter 8
A socio-cognitive perspective on
Contents
Learning outcomes
305
Introduction
305
306
306
309
310
312
313
313
315
315
315
316
317
317
318
318
320
320
321
322
323
325
326
Conclusion
327
References
328
Learning o u t c o m e s
After you have studied this chapter you should be able to:
1 describe Piagets views on the signicance of authority, peer collaboration
and individual activity for cognitive development;
2 describe Vygotskys position on the role of social interactions in cognitive
development;
3 describe Meads theory of symbolic interactionism;
4 explain how these theorists socio-historical contexts may have inuenced
their thinking;
5 dene the concept of socio-cognitive conict and describe its role in
fostering cognitive growth;
6 explain why intersubjectivity is signicant in understanding teaching;
7 dene didactic contract and explain its signicance;
8 discuss how Doises four levels of analysis offer ways into the study of
learning and development.
Introduction
305
306
dominant cultural ideologies of Europe, Soviet Russia and the United States and
consider some of their implications for the ways in which education and learning
are conceived and structured. We then bring in new ways of thinking about
learning and teaching situations, illustrated and supported with research
evidence, to build a richer understanding of the social and psychological
processes involved. A particular focus throughout the chapter is on the mutual
understandings that participants in such situations hold about the nature of the
social relationships supporting teaching and learning interactions; what we
describe as the didactic contract.
behaviourist view, arguing instead for the importance of the childs own activity
in the construction of knowledge. Educational policy was, to some extent,
inspired by the Piagetian position. One important development was that
discovery learning became a fashionable approach, in which children were
offered richly resourced learning environments, and the role of the teacher was
reworked as a facilitator of childrens own constructive endeavours, rather than
purely being a transmitter of knowledge.
However, despite the progressive label attached to such approaches, and the
marshalling of Piagetian ideas as a counter to the behaviourist theory of learning,
the individualistic constructivist perspective throws up some serious issues that are
also open to critical analysis. Perhaps the most damaging of these is the potential
that Piagets theories offer for ethnocentric interpretations of delays in childrens
development in different socio-economic and cultural circumstances. For example,
it is often the case that children in socially disadvantaged groups, or from countries
where there is much poverty and hardship, are found to be delayed in their
development relative to their more advantaged peers with better standards of
living. Given such ndings, the view that development is essentially endogenous
(i.e. springing from within the child) inevitably leads to the suggestion that such
differences are due to inherent decits in these cultural groups. Highlighting this
problem raises the issue of how the relation between culture and the development
of thinking can be reconciled with the Piagetian tradition.
Piaget recognized this issue, but he gave little direct attention to the problem,
and neither have those who have built on his theory in orthodox ways. However,
as we will show during the course of this chapter, the question is closely aligned
with a related issue that Piaget did consider to be of great importance; the relation
between the development of the individual and what we will call in this chapter
authority. By this we mean the inuence of those who can exert power and
direction over learners in various ways, by their control of resources,
communication, time and routine, by virtue of their greater age or greater
knowledge, or by holding positions of respect, such as teacher, lecturer or school
head. Perret-Clermont (1996) has suggested that the socio-historical context of the
rst half of the twentieth century and Piagets own personal biography were
sources of the ethical beliefs he came to adopt. These in turn had a major
inuence on the role he attached to authority in his theory and the image he had
of rational thought and what it could achieve.
Piaget lived in Europe through the blind violence of the First and Second World
Wars, which he blamed on the prevailing religious and cultural traditions and their
autocratic use of authority over people. Perhaps in reaction to this experience, he
rejected the notion of authority and placed his faith in rationality and the
autonomy of the individual in non-hierarchical social relationships with peers. In
line with this way of thinking, he focused his efforts on identifying the conditions
that could lead a child towards developing autonomous, critical and rational
thought. As a crucial foundation for this, he placed importance on a capacity to
reect on ones own actions and to apply logical reasoning to ones own behaviour
as well as to problems in the world. Piaget doubted whether the external,
authoritative inuences of education and other social institutions could ever do
307
308
more than just load children with cultural constraints. This baggage he saw as only
giving pre-formed answers to questions that children had neither formulated for
themselves, nor attempted to answer in the course of individually developing their
own ways of thinking. At the heart of this problem, Piaget believed, lies the
asymmetry of status between adult and child, which he saw as constraining
children from developing the capacity for autonomous, reective thought.
This line of thinking is often seen as demoting culture to just a collection of
beliefs, social rules and constraints rather than seeing it as providing a legacy of
meanings, understandings, narratives and social structures transmitted from one
generation to the next. This creates a major problem for seeing how formal
teaching can play any kind of role in fostering cognitive development, especially
when taken in combination with Piagets stance that authority inhibits the
development of independent thought and reection. Indeed, Piaget saw formal
teaching as an obstacle and a hindrance to cognitive development; for him,
interaction between peers was much more important. Piagets positive view of
peer interaction as a fundamental facilitator of cognitive development is an
important point, but one that is rarely noted by commentators, who have tended
to portray social processes as insignicant in Piagetian theory. And when Piaget
himself stressed the importance of peer interactions, he seems to have been
unaware that these commonly occur in social settings organized by adults.
Under the banner of active learning, some teachers have taken on these three
strands of Piagets thinking (i.e. negative towards authority and positive towards
peer collaboration and individual activity) and have concluded that, apart from
providing a rich environment, they need to leave children to learn entirely on
their own. In contrast to the educational ideology that children learn best by
imitating correct models, teachers who wish to be Piagetian seem to feel that
they have to stay backstage. Although teachers direct experience is likely to tell
them that an educational environment needs to be carefully structured if it is
going to engage childrens interest, the received Piagetian view is that
organization must come from the children. From this perspective, the teacher
must stay in the background, so that their authority and knowledge do not hold
back the child. This creates a tension, for example when children might justiably
ask If the teacher knows the right answer, why not give it to us rather than
expecting us to discover it for ourselves?.
This approach can be contrasted with more traditional teaching situations,
where teachers are likely to seem omnipresent. In such contexts, teachers often
start sentences which students are meant to complete (e.g. The name of the
largest lake in Switzerland is ...?), and students are then rewarded if they give the
correct response. This is a method which is thought to encourage students
participation, but in fact the ultimate aim is for the child to give a response that
matches that of the adult. Yet a common theme in both traditional and progressive
approaches is that the nal aim is for the child to achieve the adults answer. Is
this expectation a type of cultural ethnocentrism, or simply a lack of awareness of
cultural constructions?
309
310
individual development was located within, even subservient to, broader sociocultural development. Second, he also emphasized the role of technology,
through his continued reference to tools (in a broad sense, he was referring to
what we can call cultural tools of thinking as well as physical artefacts). Third, he
saw mental development as springing from the use of tools embedded in social
activities. His position was broadly aligned with the Marxist view that the
development of thought is the outcome of both material conditions and peoples
activities and struggles.
However, despite these broad differences between the two theorists, there is a
parallel with Piagets reaction to the context in which he was located. Vygotsky also
stressed the individual, constructive aspects of development. He did not see the
child as simply a passive recipient of cultures transmission of tools, but also as a
creative builder of their own thinking. This was not consonant with the prevailing
ideology of Soviet Russia, and Vygotskys work was not widely published for many
years, so it is for this reason that Vygotskys work is often seen as following Piaget,
although in fact the two theorists were to some extent independently developing
their ideas at much the same time. Another point of agreement between the two
lies in their recognition that biological maturation also plays a role, albeit a
secondary one, in making certain developments possible on the cognitive plane.
On balance, Piaget saw the control of development as being essentially inside
the child, while Vygotsky saw the childs development as being the internalization
of outside inuences or social co-ordinations. In his theory, Piaget justied his
position by seeing the logical progression of development as arising from
childrens reections on their own actions, culminating in the achievement of
formal operational thought, that is, fully abstract and hypothetico-deductive
reasoning. For Vygotsky, however, this achievement was the result of a social,
rather than an individual process.
appropriate ways by signicant others. For the child, seeing a hammer used in
specic ways by another person and hence in a social context, imbues the object
with social signicance. Through such repeated experiences whereby objects,
actions and language come to acquire social meanings, the child also begins to
form simultaneously a representation of the self and of what Mead called the
generalized other. This means that the child becomes able to see themselves and
their actions from the perspective of another person, and as a result, begins to
internalize meanings and values. According to Mead, it is this process of
internalization that allows childrens thinking to advance from the level of
immediate experience to a level of self-reection. Mead believed that this is an
important part of children forming a conscious concept of self, and he believed
that the play of young children, in which role-playing other people and their
activities tends to be common, was a primary vehicle for this development. This
stands in contrast to Piagets view of the role of other people. According to Piaget,
play is at rst solitary, and then later (as two or more children play together)
becomes parallel play, where the players merely co-ordinate their behaviour, but
do not truly co-operate. Only in Piagets nal stage of development do they
properly take account of other peoples real or potential perspectives.
The reason why these different interpretations of the role of play are important
is that play can be seen as an excellent example of a form of social activity that is
not bound by external authority structures, nor by imposed power relations.
Although at times one child or another inevitably sets the agenda for where the
play is going, in general, play is an activity that children engage in for its own
sake, not for its purpose in relation to some dened outcome. The signicance of
such reciprocal peer relations for development will be explored in more detail as
this chapter progresses.
.
.
311
312
development of thought
These three theorists positions on the development of thought and its relation to
social processes became widely disseminated throughout the academic
communities of Europe and the United States. This led to a recognition of
parallels and disagreements among the theories, which formed the background to
the emergence of a new genre of research in the 1970s. This was strongly
empirically based, and set out to address the new questions posed by the
theoretical debates about the extent to which cognitive development is
inuenced by social factors. In particular, interest focused on how children
negotiate and make use of situations involving conicts of ideas to make progress
in their development. This new line of enquiry took a scientic approach, using
standardized tasks and situations, often in laboratory settings, to explore how
children resolved the intellectual and social problems posed by conict. This
recognition that an intellectual conict requires both an intellectual solution and a
social solution is central to the ideas explored in this chapter. If someone
disagrees with you, and you each want to do more than simply register the
difference of opinion (i.e. seek some kind of resolution) you need to do the
intellectual work of analyzing the nature of the dispute and do the creative work
of coming up with something that can satisfy both parties. But, and this is a crucial
point, you also have to do equivalent work towards resolving the social
disjunction that occurs in conict situations.
Since the 1970s, the issue of the role of socio-cognitive conict as a stimulus for
development has inspired an impressive body of research. Findings from this
research have shown clearly how children, adolescents and adults have a capacity
to benet, in terms of developing their modes of thought, from situations in
which the joint resolution of cognitive challenges is required (Perret-Clermont,
1979, 1996; Doise and Mugny, 1981; Perret-Clermont and Nicolet, 1986, 2002;
Gilly, 1989; Howe et al., 1990, 1995; Azmitia, 1996; Carugati and Selleri, 1996;
Carugati, 1997; Littleton and Light, 1999; Schwarz et al., 2000).
The rst important nding was that if children work together to solve a
problem, they are often able, by co-ordinating their actions, to arrive at cognitive
solutions that neither of them are able to reach alone. This shows the limitations of
looking solely at what children can achieve by working on their own, and how
such an approach can lead to an underestimation of childrens levels of cognitive
ability. This issue is especially important when childrens abilities are being
assessed on the basis of their performance in an educational context. It is also
worth noting that the asymmetry of power relations that exists in teacherstudent
interactions is not present in this sort of peer collaboration. We will return to this
point in Section 6.
The second important nding was that when children work together in
particular ways on tasks that they cannot initially solve alone, they can, even in
the short term, become able to tackle tasks of a similar level of difculty on their
own. In this way, cognitive tools that are rst constructed jointly to solve
313
314
the classic Piagetian conservation of liquid task, a child is rst asked to conrm
that the amounts of liquid in two identical containers are, indeed, the same. Then
the liquid from one container is poured into another container with different
proportions and the child is asked whether the two amounts of liquid are the
same. Children nd it much easier to solve this problem correctly if the task is
presented as a fair shares reward of a drink to each child for their participation in
the experiment. Here, the task of judging whether the two containers hold the
same amount is supported by the notions of fairness and equity. This particular
social marking is therefore based on the concept of distributed justice. But, of
course, social marking is itself subject to meaning construction and how it
operates in a particular situation will depend on how the children make sense of
the social rules that the adult intends to appeal to. For example, evoking the
concept of distributed justice is more effective if children are placed in a cooperative situation than if they are expected to compete with each other. Research
that has made use of rules to do with the right to have equal quantities of objects
or drink have conrmed this phenomenon.
In this section, then, we have shown two broad patterns of ndings regarding
the social context of experimental conditions. The rst contextual effect that we
have found is that the cognitive demands that a task makes of a child, and the
childs chances of reaching a solution, do not simply reside in the task itself. They
are made harder or easier by the social context of the child, the task and other
people. A degree of social conict around possible routes to solving the task can
make the nding of a correct solution easier. Once such a joint solution has been
found, through the resolution of the socio-cognitive conict, it can then be
available for the child to apply to new tasks with similar demands when working
independently. In this way the childs cognitive development has been fostered.
The second important contextual effect that we have highlighted concerns the
match of the social context to the rules that have to be applied to the task in
hand. For example, as we illustrated in respect of the Piagetian conservation task,
a match between task and social context also helps children to come to a
successful solution. In the favourable circumstances outlined above, then,
children can perform at higher levels than those predicted by theoretical positions
that only offer individualistic interpretations of cognitive development, such as
those of Piaget.
315
316
there is a trick!. Other children started the test, but then turned to the researcher,
asking Now, should I tell her that there is a trick or not?. Further, Grossen
observed that children who were non-conservers (i.e. they believed that moving a
quantity of liquid from one container to another of different proportions changed
the amount of liquid) tended to produce correspondingly non-conserving
responses from the children that they were testing. Conserving children, on the
other hand, tended to elicit conserving responses from their partners.
.
.
In the 1980s, research looking into the links between context, teachers, students
and subject area introduced the concept of a didactic contract, a special case of
Rommetveits communication contract, to refer to the ways of behaving that
teachers and students adopt and the rules that they follow (Brousseau, 1980;
Chevallard, 1985). There are two sides to this contract. On the students side there
are expectations of what it is to be a learner, and on the teachers side there are
expectations of what it is to be an effective teacher. This implicit contract reects
the knowledge and understandings of what one has to comply with in
educational processes in everyday school life.
317
318
Leoni (1990) has also shown how younger children are more successful at solving
problems when the tester is a psychologist who is making a game out of a
problem-solving situation than if the tester is a teacher. In contrast, children who
have had more experience of school are more likely to succeed if the task is
presented as a school task and not as a game.
319
320
school life
In the preceding sections, we have referred to several research ndings that have
highlighted how the social contexts of learning and testing, and how children
interpret these contexts, can affect the levels of performance that children show.
We have also suggested that this affects the opportunities that they have to
advance the levels of their thinking. However, within a conception of cognitive
development that stresses the individual origins of cognitive ability, the impact of
such studies may be considered limited. They may be viewed, for example, as
showing that there are additional variables that affect development, but may not
be considered as signicant in terms of further implications for theory. However,
if social dynamics are considered to be at the centre of the construction of
cognitive capacities, this suggests that cognitive development is a much more
complex phenomenon. It suggests that these issues of communication,
interpretation and negotiation are crucial to the meaning construction that goes
on in teaching and learning situations. From this perspective, it seems clear that it
is not enough to just study these processes in abstracted, research laboratory
situations, but that it is important to study concrete learning contexts in the
everyday life of schools, students and teachers.
conversations, as a result, are structured in very different ways from most other
situations in everyday life.
The implicit assumption that guides most classroom activity is that the teacher
poses the questions and knows the right answers. On this basis, a teacher
evaluates the students responses and then uses this information to judge the
students level of competence. Where a group of children show that they
understand and accept this basic, underlying social rule, the ow of
communication in their classroom is then well regulated on the basis of this
consensus. This is generally seen as best tted to meeting the mutual expectations
for good teaching. Edwards and Mercer (1987) and Mercer (1995) have shown
how, in classes of 810-year-old students, classroom activity and the teachers
language use build shared understandings about these basic social rules which
students then use to attribute meaning to their interactions. The teacher uses
language to point up and mark these implicit understandings, so that the way talk
is organized and used in the classroom by the teacher develops a shared
vocabulary which xes the respective roles of teacher and student. This sets up
the particular perspective in which the teacher is expected to lead the student
along a path to the acquisition of new concepts and the reorganization and
extension of knowledge.
321
322
based not just on a reective analysis of the answer but also on the childs
perception that they need to give a different, more correct response.
A specic type of classroom dialogue that has attracted the interest of
researchers is where groups of children seem to be able to develop new
understandings through externalized reasoning where they all participate in
more of a true debate. In this sort of situation, the teachers interventions take on
a different meaning. Their role can be to get closer to understanding the students
modes of thought, rather than encouraging the student to better understand the
teachers thought processes. Rather than guide them along a preconceived path,
the teacher may facilitate the students discourse and cognitive processes to
support the development of new understandings (Mercer, 1995).
These insights have led to a major shift in perspective that now puts at centre
stage the relationship between studentteacher debate and the development of
thinking (Pontecorvo, et al., 1991; Pontecorvo, 1993; DeGroot and Schwarz, 2003;
Schwarz et al., 2003). Seeing the classroom situation as a context in which
characteristic types of conversational events take place, where the development
of modes of thought are either facilitated or hindered, points up the reciprocal
relation between the social milieu and the knowledge of the child (Barth, 1994).
The shared cultural practices of school life, the objects that they involve and the
communication patterns within them all contribute to a referential framework
(Resnick, 1991) for the teaching and learning process.
rarely use arithmetic approaches even when these have been studied and learned
in the school setting.
As we saw in Chapter 7, research on this theme was carried out in Recife in
Northern Brazil by Nunes et al. (1993) who studied the mathematical abilities of
street children. At ages of 6 or 7, the researchers found that these children are
already surviving independently by re-selling oranges, or doing small jobs, and
some of them quickly become highly accurate in working out monetary
transactions and giving the correct change. Yet these same children systematically
fail in mathematics in their rst year of schooling and tend to be rapidly rejected
by the educational system. Why is it that they are unable to transfer the
knowledge acquired from their street life to school tasks? On the streets, children
learn highly complex algorithms to deal efciently with the buying and selling
situations that they encounter, situations that demand accurate calculation. In the
school situation, the teacher does not know that the students already possess
these practical techniques and when they spontaneously use them in class the
teacher tends to treat their strategies as clumsy or incorrect. This then contradicts
the childrens experiences and does not allow them to reect on the strengths and
limitations of these strategies (in large part oral) that they use in their commercial
transactions.
The question thrown up by this research is the nature and extent to which there
is a link between the context in which one acquires knowledge and the context in
which it is used or transferred, since the desired transfer of knowledge and skills
often fails to occur. There can be many reasons for this; perhaps the teacher fails
to understand the signicance of this aspect, or fails to clarify it, or perhaps the
students themselves fail to grasp the relations between the knowledge they have
gained in one situation and that required in another. It may also be the case, at
times, that knowledge is context-bound without any of the participants being
aware of this. Thus we have to consider further the three-way relations in the
triangle of teacher, student and knowledge.
Student
Teacher
323
324
well as a student. We have shown so far in this chapter how the development of
the individuals knowledge cannot be easily isolated from the other two points,
the teacher and the object of knowledge itself, since all of these are situated
within an institutional and cultural context. We have also tried to show how much
there is to be gained in enriching understandings of cognitive development by
considering all three points together. Generally, trainee teachers see knowledge
as being their focus, so, for example, in mathematics they may concentrate on
how to design a series of lessons, how to demonstrate a formula or how to
choose examples to be worked through. We would argue that this is a narrow
view of teaching and that to study this aspect unquestioningly, without also
considering the other points of the triangle and their relations, is too limited and
reductionist.
As far as the teacher is concerned, theoretical and research interest has mainly
tended to focus on the socio-emotional aspects of the teacherstudent relationship
rather than on the ways in which the teacher represents the childs points of view.
It has also tended to see teaching in relation to the evaluation of student abilities. It
is only very recently that studies have been conducted on how teachers think
about intelligence, learning and development, and their mental models of students
(for an exposition of the dynamics of causal attributions see Monteil, 1989; with
respect to the educational implications of social representations, see Mugny and
Carugati, 1985; Selleri et al., 1994; Carugati and Selleri,1996). However, little is
known about the reasons why teachers nd it difcult to decentre from their own
points of view.
The study of this teacherstudentknowledge triangle in teaching situations is
becoming increasingly important. How do teacher-trainers, teachers themselves
and psychologists studying the triangle see it functioning? The implicit didactic
contract seems to be that the teacher sets tasks that correspond with what he or
she wishes the students to know. Because students come to understand what they
are expected to know, they then expect to be given exercises that will lead to
success in the task. One can talk of a genuine micro-culture, with slogans such as
You have to do this, because it is how you learn properly; If you do it in this
way, you get a good mark; If you do this, it means that you have understood. In
other words, students responses and the roles that they take on are not
spontaneous constructions, but are the product of particular teaching practices.
So how can these issues around the meanings of learning situations and the
corresponding expected behaviour(s) be addressed? A crucial theme here is the
modes of constructing shared understandings between teacher and students that
support or do not support students in discovering the knowledge that the teacher
wishes to transmit. In what circumstances do children genuinely learn what one is
seeking to teach them? The school system and the life of the classroom are usually
sufciently based on well-oiled daily routines to allow students to respond
correctly without having properly understood. Students do not often see the aim of
schoolwork as being to gain understanding, but rather to pass tests successfully.
For students, essential achievements are to gain a satisfactory mark for homework,
to meet the teachers expectations and to show that they know what they ought to
know. To understand is an optional bonus! This is often what happens if, when a
teacher says try to understand, the student hears this as try to understand how
you have to answer, so that I believe that you have understood! Some students
even gain good qualications without ever having properly experienced the
satisfaction of having truly understood something. We believe that to understand is
an autonomous act, internal and free, which does not depend on hierarchical
structures, nor can it be prescribed or demanded. So there is a risk that this level of
deep comprehension, if it has not been experienced, may end up playing no part
in the students eld of experience, or even in their expectations.
The methods and scientic culture of modern social psychology aim to focus
attention on and develop understanding of the ways in which the individual and
the collective are connected. Willem Doise, a social psychologist who has been
concerned with this chapters topic of interest, has differentiated four levels of
analysis which can serve as points of entry for studying these connections (Doise,
1982). In this section we make use of these four levels to organize what we have
already covered in this chapter, and to see what insights they can offer into the
learning process.
We hope that our preceding arguments have shown convincingly that cognitive
development is not just the result of a simple maturation of an innate potential
nor is it a simple behaviourist stimulus-response process. It does not result solely
from social transmission, pouring knowledge into empty vessels. The recent
research that we have discussed in this chapter is beginning to reveal the
complexity of the processes that support the development of modes of thinking.
That one person can appropriate the knowledge of another, that each generation
can build on the experience of the previous one, that cultural life can shape and
transform thinking (Bruner 1990, 1991) and that it is possible for new, creative
thoughts and ideas to emerge ... these are phenomena that seem miraculous
when examined in detail.
325
326
Summary of Section 7
.
.
.
Doise has proposed that the social nature of cognition can be analysed at
four different levels. These four levels focus on different objects of
concern.
Level 1 is the level of the individual engaged in constructing their modes
of thought and their knowledge, trying to make sense of their experience.
Level 2 is the level of actual relationships, interaction and discourse, for
example, studentteacher or peerpeer expectations. It focuses on, for
instance, the construction of intersubjectivity and mutual expectations.
Level 3 refers to roles and statuses that frame and constrain the sorts of
behaviour that are considered proper for teaching and learning in
different contexts.
Level 4 is the level of national policies and ideologies, systems of social
representations, that dene situations (such as schools) in specic ways.
These may often be dictated by professional training and traditions, and
by government policies that dene situations in specic ways. But these
are always reinterpreted by the participants, for example by teachers and
students in specic schools and situations.
These levels are not independent: one level can affect or transform
another.
Conclusion
In this chapter we have set out to show how communication contracts and
didactic contracts, both of which are largely implicit, govern the transmission of
understandings and the conditions of learning. Learning lies at the heart of
specic ways of relating which are closely interlinked with the rest of the daily life
of the class. Learning is also linked more generally with the ways in which the
school operates so as to organize understandings and the dynamics of teacher
student interaction. We have also explained the importance of the ways in which
teachers and students construct their own meanings for these practices and
traditions. The transmission of experience and knowledge from one generation to
the next is seen as something that can be appropriated and interpreted. It can
become both a personal and also a collective acquisition, and is thus not a
327
328
Acknowledgement
This chapter is an updated and modied version of the chapter published in
Italian (Carugati, F., and Perret-Clermont, A. -N. (1999) La prospettiva psicosociale: intersoggettivita` e contratto didattico, in Pontecorvo, C. (ed.) Manuale di
Psicologia delleducazione, pp. 4166, Bologna: Il Mulino.) and an abridged
version of the original paper is in press in French.
References
Azmitia, M. (1996) Peer interactive minds: developmental, theoretical, and
methodological issues, in Baltes P. B. and Staudiner, U. M. (eds) Interactive
Minds: life-span perspectives on the social foundations of cognition, pp. 13361,
Cambridge, Cambridge University Press.
Barth, B. M. (1994) Le Savoir en Construction: former a` une pedagogie de la
comprehension, Paris, Retz.
Baucal, A., Muller, N., Perret-Clermont, A.-N. and Marro, P. (2002) Nice designed
experiment goes to the local community, paper presented at the Fifth Congress of
the International Society for Cultural Research and Activity Theory, Amsterdam,
June 2002.
329
330
Howe, C., Tolmie, A., Greer, K. and Mackenzie, M. (1995) Peer collaboration and
conceptual growth in physics: task inuence on childrens understanding of
heating and cooling, Cognition and Instruction, vol. 13, pp. 483503.
Hundeide, K. (1985) The tacit background of childrens judgements, in Wertsch,
J. V. (ed.) Culture, Communication and Cognition: Vygotskian perspectives,
pp. 30622, Cambridge, Cambridge University Press.
Hundeide, K. (1988) Metacontracts for situational denitions and for presentation
of cognitive skills, Quarterly Newsletter of the Laboratory of Comparative Human
Cognition, vol. 10, pp. 8591.
Hundeide, K. (1992) The message structure of some Piagetian experiments, in
Wold, A. H. (ed.) The Dialogical Alternative: towards a theory of language and
mind, pp. 13956, Oslo, Scandinavian University Press.
Labov, W. (1972) The study of language in its social context in Giglioli, P. P. (ed.)
Language and Social Context, pp. 283307, Harmondsworth, Penguin Education.
Levy, M. and Grossen, M. (1991) Contrat experimental et acte de questionnement:
deux illustrations empiriques de larticulation entre processus et activite cognitive
de lenfant dans une situation de test piagetienne, Bulletin de Psychologie, vol. 44,
pp. 22938.
Light, P. and Perret-Clermont, A.-N. (1989) Social context effects in learning and
testing, in Sloboda, J. A. (ed.) Cognition and Social Worlds, pp. 99112, Oxford,
(Oxford Science Publications), Oxford University Press.
Littleton, K. and Light, P. (eds) (1999) Learning with Computers: analyzing
productive interaction, London, Routledge.
Mead, G. H. (1934) Mind, Self and Society, Chicago, Chicago University Press.
Mercer, N. (1995) The Guided Construction of Knowledge: talk amongst teachers
and learner, Clevedon, Multilingual Matters.
Monteil, J. M. (1989) Eduquer et Former, Grenoble, Presses Universitaires de
Grenoble.
Mugny, G. and Carugati, F. (1985) lIntelligence au Pluriel, Cousset, Delval.
Nicolet, M. (1995) Dynamiques Relationnelles et Processus Cognitifs, Lausanne,
Delachaux et Niestle.
Nunes, T., Schliemann, A. D. and Carraher, D. W. (1993) Street Mathematics and
School Mathematics, Cambridge, Cambridge University Press.
Perret, J. F. and Perret-Clermont, A.-N. (2001) Apprendre un Metier dans un
Contexte de Mutations Technologiques, Fribourg, Editions Universitaires Fribourg.
Perret-Clermont, A.-N. (1979) La Construction de lIntelligence dans lInteraction
Sociale, Berne, Peter Lang.
Perret-Clermont, A.-N. (1996) La Construction de lIntelligence dans lInteraction
Sociale (4th edn), Peter Lang, Berne.
331
332
ACKNOWLEDGEMENTS
Acknowledgements
Chapter 1
Table
Table 1: Younger, B. and Gotlieb, S. (1988) Development of categorization skills:
changes in the nature or structure of infant form categories?, Developmental
Psychology, (24), pp. 61119, The American Psychological Association, Inc.
Figures
Figure 1: Harold Taylor/Oxford Scientic Films; Figure 3: Younger, B. and
Gotlieb, S. (1988) Development of categorization skills: changes in the nature or
structure of infant form categories?, Developmental Psychology, (24), pp. 61119,
The American Psychological Association, Inc.
Chapter 2
Text
Reading A: Christophe, A. and Morton, J. (1998) Is Dutch native English?
Linguistic analysis by 2-month-olds, Developmental Science, vol. 1 (2), pp. 215
19, Blackwell Publishers Ltd; Reading B: Harris, M. and Chasin, J. (1999)
Developments in early lexical comprehension: a comparison of parental report
and controlled testing, Journal of Child Language, 26, pp. 45360, Cambridge
University Press.
Figures
Figure 2: Johnson, E. K. and Jusczyk, P. W. (2001) Word segmentation by 8month-olds: when speech cues count more than statistics, Journal of Memory
and Language, vol. 44, p. 555, Academic Press; Figure 3: Johnson, E. K. and
Jusczyk, P. W. (2001) Word Segmentation by 8-month-olds: when speech cues
count more than statistics, Journal of Memory and Language, vol. 44, p. 557,
Academic Press; Figure 5 and Figure 9: Fenson, L., Dale, P., Reznick, S., Bates, E.,
Thal, D. and Pethick, S. J. (1994) Variability in early communicative
development, Monographs of the Society for Research in Child Development,
vol. 59 (5), p. 42, The Society for Research in Child Development, Inc., University
of Michigan; Figure 6: Elman, J., Bates, E., Johnson, M., Karmiloff-Smith, A., Parisi,
D. and Plunkett, K. (1996) Rethinking Innateness: a connectionist perspective on
development, MIT Press; Figure 7: Jones, S. et al. (1992) Human speech and
language, The Cambridge Encyclopedia of Human Evolution, Cambridge
University Press; Figure 8 and Figure 10: Fenson, L., Dale, P., Reznick, S., Bates,
333
334
Chapter 3
Text
Reading A: Mehler, J. and Dupoux, E. (1994) The organ of language, What
Infants Know: the new cognitive science of early development, pp. 14650,
Blackwell Publishers Ltd; Reading B: Karmiloff-Smith, A. (1992) Beyond
Modularity: a developmental perspective on cognitive science, MIT Press.
Figures
Figure 4: Reprinted by permission of the publisher from The Postnatal
Development of the Human Cerebral Cortex vols I-VIII by Jesse LeRoy Conel,
Cambridge, MA.: Harvard University Press, Copyright #, 1939, 1975 by the
President and Fellows of Harvard College; Figure 5: Photo courtesy of Leslie
Tucker, Centre for Brain and Cognitive Development, Birkbeck College; Figure 6:
Scott Camazine/Science Photo Library; Figures 7 and 8: Neville, H. J. et al. (1998)
Cerebral organisation for language in deaf and hearing subjects: biological
constraints and effects of experience, Proceedings of The National Academy of
Science, U.S.A., vol. 95, # Copyright 2002, The National Academy of Sciences,
U.S.A.; Figure 9: Reprinted from Brain and Language, vol. 61, p. 351, Reilly, J. S.
et al., Narrative discourse in children with early focal brain injury, Copyright
(1998), with permission from Elsevier.
Chapter 4
Text
Reading A: Aitchison, J. (1998, 4th edn), A blueprint in the brain, The Articulate
Mammal: an introduction to psycholinguistics, Routledge/Taylor & Francis.
Chapter 6
Figures
Figure 1: Frith, U. (1989) Figure 3: The false belief paradigm, Autism: explaining
the Enigma, Blackwell Publishers Ltd; Figure 3: Whiten, A. (1991) Natural
Theories of Mind: evolution, development and simulation of everyday
mindreading, Blackwell Publishers Ltd; Figure 4: Meltzoff, A. N. (1995)
Understanding the intentions of others: re-enactment of intended acts by 18month-old children, Developmental Psychology, vol. 31, (5) Copyright # 1995 by
the American Psychological Association. Reprinted with permission.
ACKNOWLEDGEMENTS
Chapter 7
Text
The observations in Activities 10 and 11 were collected by S. Allen in a small-scale
study for an Open University course and are reprinted with permission.
Figure
Figure 10: Driver, R. (1985) Beyond Appearances: the conservation of matter
under physical and chemical transformations, in Driver, R., Guesne, E. and
Tiberghien, A. (eds) Childrens Ideas in Science, Open University Press.
Cover photographs
Getty Images (Photodisc).
335
336
Name index
Aitchison, J. 172, 2012
Anderson, C. W. 283
Arterberry, M. E. 42
Asher, S. R. 227
Astington, J. 239, 243, 244, 252
Avis, M. 238
Azmitia, M. 312
Baer, R. A. 223
Baker, L. 214
Baldwin, D. A. 77
Barkley, R. A. 226, 227
Barlow-Brown, F. 107
Barrett, M. 107
Barrett, M. D. 85, 88
Barth, B. M. 322
Bartsch, K. 246, 247
Bates, E. 85, 93, 106, 107, 110, 111
Baucal, A. 316
Bavelier, D. 140
Behl-Chadha, G. 36
Behr, M. 278
Bell, M. A. 146
Bell, N. 316
Bellugi, U. 189
Benedict, H. 107, 110
Benigni, L. 107
Berg, E. A. 223
Bertoncini, J. 98
Bickerton, D. 136
Bishop, D. V. M. 190
Bomba, P. C. 30, 53
Bornstein, M. H. 42
Borton, R. W. 124
Bosch, L. 104
Braine, M. D. S. 195, 197
Bretherton, I. 107, 110
Brooks, S. 107
Brophy, M. 227
Brousseau, G. 317
Brown, A. L. 214
Brown, J. 254
Brown, M. 272
Brown, R. 170, 174, 175, 179, 181,
184, 193
Brown, T. E. 227
Bruner, J. S. 63, 734, 76, 87, 150,
293, 325
Bryant, P. 273
Burgess, P. 213
Butterworth, G. 41, 77
Callanan, M. 254
Camaioni, L. 107
Carpenter, T. P. 272
Carraher, T. N. 274
Carugati, F. 312, 313, 324
Casey, B. J. 131, 222
Cavanagh, J. C. 214
Cervantes, C. 254
Changeux, J. -P. 124, 132, 213
Charman, T. 249, 255
Chasin, J. 78, 10611
Chelune, G. J. 223
Cheng, P. W. 289
Chevallard, Y. 317
Chi, M. T. H. 284
Choi, S. 89
Chomsky, N. 11, 14, 64, 135, 137,
143, 144, 160, 165, 172, 192, 194,
197, 2013
Christophe, A. 67, 69, 98104
Clement, J. 284
Cowan, M. W. 122
Crago, M. 189
Csibra, G. 147, 148
Cutler, A. 102
Dale, P. 106
DeCasper, A. J. 66, 67, 120
DeGroot, R. 322
Dehaene-Lambertz, G. 99
Deloache, J. S. 48
Demetras, M. J. 193
Dennett, D. 235, 236
Desli, D. 285
di Sessa, A. 282, 287
Diamond, A. 146, 221
Doise, W. 19, 292, 305, 312, 313,
325, 327
Donaldson, M. 234, 318
Donaldson, M. 1718
Dore, J. 79, 107
Drewe, E. A. 221
Driver, R. 283
Dunn, J. 215, 254
Dupoux, E. 67, 100, 129, 1547
Edwards, D. 321
Eimas, P. D. 34, 35, 40, 434, 48, 52,
53
Elbers, E. 316
Elman, J. 129, 133, 196
Emiliani, F. 326
Erickson, G. 282
Fantz, R. L. 28
Fearing, D. D. 47
Fenson, L. 78, 79, 83, 84, 92, 93,
106, 107, 109, 110, 111
Fifer, W. 67
Fodor, J. A. 14, 1278, 129, 135, 156,
157, 158, 159, 207
Fox, N. A. 146
French, R. M. 51
Frith, U. 237
Furrow, D. 196
Fuster, J. M. 119, 145
Gage, P. 21213
Garcia, R. 289
Gelman, R. 160, 278
Gerstadt, C. L. 220, 221
Gilly, M. 312, 326
Ginsburg, H.P. 274
Girotto, V. 289
Goldfield, B. A. 86, 111
Goldfield, E. C. 133
Goldman-Rakic, P. S. 119, 145, 146
Goldstein, K. 212
Gopnik, A. 54, 56, 89, 239, 247, 248
Gopnik, M. 189
Gottlieb, S. 30, 31, 32, 33
Graham, S. 47
Grant, D. A. 223
Groen, G. J. 273
Grossen, M. 31516, 319
Gunzi, S. 110
Hanlon, C. 181, 193
Harlow, J. M. 21213
Harris, M. 74, 75, 778, 78, 84, 85,
86, 87, 88, 89, 10611
Harris, P. 238
Hart, B. 89
Hartl, M. 240
Hatano, G. 286
Hebb, D. 133
Hesketh, S. 99
Hesse, J.J. 283
Holyoak, K. J. 289
Howe, C. 2923, 312
Hughes, C. 215, 221, 227
Hundeide, K. 316, 318
NAME INDEX
Maurer, D. 124
Mead, G. H. 305, 31011
Meck, B. 278
Mehler, J. 67, 98, 129, 1547
Meltzoff, A. N. 54, 56, 124, 249, 250
Mercer, N. 321, 322
Mervis, C. B. 107
Millikan, R. G. 43
Mills, D. M. 132
Milner, B. 119, 145
Molinari, L. 326
Monteil, J. M. 324
Moon, C. 98
Morton, J. 67, 69, 98104
Moser, J. M. 272
Mugny, G. 19, 292, 312, 324
Naigles, L. 1789
Nazzi, T. 98, 102
Nelson, K. 93, 107
Neville, H. J. 130, 132, 139, 140, 158
Newman, D. 294
Nicolet, M. 312, 313
Norgate, S. H. 78
Novack, J. D. 264
Nunes, T. 270, 274, 275, 276, 277,
279, 285, 323
Nussbaum, J. 264, 266
Oakes, L. M. 43, 47
Oakley, Y. 107
OLeary, D. D. M. 129
Oller, D. K. 81
Oppenheim, R. 122
Osborne, A. 240
Parker, J. G. 227
Perlmutter, M. 214
Perner, J. 221, 238
Perret, J. F. 328
Perret-Clermont, A. -N. 292, 307,
312, 316, 318, 319, 322, 328
Petersen, S. E. 130
Pethick, S. J. 106
Piaget, Jean 13, 14, 17, 146, 158,
159, 160, 221, 234, 255, 268, 278
280, 283, 288, 289, 291, 292, 294,
297, 305, 314, 315, 318
Piburn, M. D. 288, 289
Pinker, S. 1358, 143, 144, 169, 172,
180, 182, 187, 189, 193, 194, 197
Plunkett, K. 80, 170, 184, 190
Pontecorvo, C. 322
Posner, G. J. 287
Poulin-Dubois, D. 43, 47
Premack, D. 235, 236
Prince, A. 182
Quinlan, P. T. 133
Quinn, P. C. 30, 34, 35, 37, 38, 40,
434, 47, 48, 49, 51, 52, 53
Rakic, P. 129
Rakison, D. 41, 43
Reed, J. J. 274
Reilly, J. S. 1402
Reiner, M. 284
Repacholi, B. M. 247, 248
Resnick, L. B. 273, 322
Reznick, J. 86, 106, 111
Richards, D. 65
Rijsman, J. 313
Riley, M. S. 272
Rommetweit, R. 315, 317, 318
Roth, I. 16, 18
Rubia, K. 221, 2267
Ruffman, T. 2523
Rumelhart, D. 183, 185
Saljo, R. 318
Schubauer-Leoni, M. L. 315,
31718, 322
Schwarz, B. B. 312, 322
Sebastian-Galles, N. 104
Seidenberg, M. S. 190
Selleri, P. 312, 313, 324
Seuss, Dr 66
Shallice, T. 160, 190, 213, 215, 216
Shatz, C. J. 129
Siqueland, E. R. 30, 53
Skinner, B. F. 14, 159
Slotta, J. D. 284
Smedslund, J. 313, 316
Snow, C. E. 196
Snyder, L. 110
Spelke, E. 160, 177
Spence, M. J. 66, 67, 120
Spencer, J. 40
Stetri, A. 124
Stiles, J. 148
Stiles, L. 142
Sullivan, K. 242
Sutton, J. 245
Swanson, J. M. 226
Szagun, G. 1889
Szeminska, A. 268
337
338
Tardif, T. 89
Tatham, M. 68
Tees, R. C. 81
Thal, D. 106, 142
Thatcher, R. W. 147
Thomas, M. S. C. 190
Tiberghien, A. 282
Tomasello, M. 107, 194, 196
Trauner, D. 148
Tucker, L. A. 148
Turkewitz, G. 124
Tzourio, N. 131
Vergnaud, G. 271, 272, 274, 280
Volterra, V. 107
von der Malsburg, C. 132
Vosniadou, S. 284
Vygotsky, L. 13, 17, 19, 251, 255,
286, 287, 293, 294, 297, 305,
30910, 311, 321
Waxman, S. R. 43
Wellman, H. 246, 247
Welsh, M. C. 223
Werker, J. F. 81
Whiten, A. 242
Williams, D. 127
Wimmer, H. 238, 240
Wood, D. 19
Woodruff, G. 235, 236
Woolfe, T. 254
Wyndhamn, J. 318
Yeeles, C. 107
Younger, B. A. 30, 31, 32, 33, 47
Zittoun, T. 328
SUBJECT INDEX
Subject index
A-not-B task 146, 211, 222
action
slips of 21011
109, 111
316
amalgams 174
animals
structure in 26
development 54
367, 47, 48
procedure 412
see also cat categories; dog
categories; horses
antisocial behaviour,
communication skills and
executive function 21516
aphasia, and cognitive modularity
1546
arithmetic, oral versus written
2745
ASL (American Sign Language),
neuroscientific studies of deaf
participants and 13940
attention deficit hyperactivity
disorder (ADHD) children, and
inhibitory control 12, 2267
39
3078
autism 127
auxiliary verbs
138
selectionism 124
98104
behaviourism
and domain-general
development 159
104
boys
84
development 2545
1011, 11561
anatomy 83
influences in 14
11618
14, 13545
prefrontal cortex
brain damage
1402, 143
148
256
cat categories
3940
novelty-preference method
29, 30
of humans 52, 53
36
categorization
and language 23
mind 244
339
340
1546
modules 12535
196
children testing children 31516
chimpanzees
classroom 3235
thinking 28896
life 3205
30910
socio-cognitive perspective on
13, 3257
31011
development
1547
function in 20911
communication
168
concept formation 9
conceptual primitives 44
development 269
conflict
cognitive 19
connectionist computational
modelling 149
196, 198
brain 134
of vocabulary learning 80
reasoning 2689
2867
constructivism
context-bound words
actions 211
language 136
cues
development 54
sequential touching
morphological 172
learning 778
cultural practices
arithmetic 2748
cultural tools 17
culture
254
deaf people
language 136
neuroscientific studies of
13940
dendrites
1201, 122
developmental cognitive
neuroscience
brain imaging methods 1302
and language development
13842, 143
childrens interpretation of
adults questions 31819
the students view of being a
student 31718
SUBJECT INDEX
marking 314
distributional approaches to
dog categories
3940
novelty-preference method
29, 30
36
domain-generality, and
1601
128, 159
303
and memory 34
specificity 160
categorization 446
mind 248
EEG (electroencephalogram)
studies 131
egocentric speech 17
1978
development 124
English language
meaning 89
67, 98100
130
phonology 1678
69
environmental influences
on brain development
1324
1267
on grammatical development
2515, 256
epigenesis
modularization 12930
and innateness 11618
and language development
1434
1267
ethnocentrism
131, 140
20728
in clinical psychology 21113
and cognitive flexibility 21314,
21415, 221, 223
dysfunction 2257
importance in child
development 208
experience
organization of 207
experience 878
content 93
23840
mind 2423
2478
and the unexpected transfer task
2368
familiarization/novelty-preference
method
and infant categorization 2830,
47, 50
and memory 34
49
development 2523
173
variations in development
rates 913
learning to say words 815
linking sounds to meaning 738
meaning of 8591
development of word
meaning 889
341
342
development 2545
and vocabulary content 93
global/superordinate categories
and infant categorization 35, 37,
412, 47, 50
single-process model of 43,
48
Go/NoGo tasks 221, 222, 227
grammar 11, 165203
defining grammatical terms 167
grammatical development
17381
diary studies of 1956
early combinations 174
empiricist theories of 165,
187, 1946, 1978
and first words 173
nativist theories of 65, 187,
1934, 197
to full blown grammar
17981
and utterance length 1749
word endings 18290
grammatical errors and braindamaged children 1412
and the nature of spoken
language 16673
learning morpho-syntax
1712
phonology 166, 1678
and Pinker on the innateness of
language 1367
plural forms 169, 182, 184, 188
see also morphology; nouns;
syntax; UG (Universal
Grammar); verbs
grammatical morphemes 168
habituation/dishabitation sucking
method, testing infants
perception of languages 98104
the Handgame 220, 221
hearing loss 126
heat and temperature, and
childrens understanding of
science 282, 284, 2867, 290
Hebb rule 133
hiding from policemen task 18
horses, infant categorization of 43,
52, 53
humans, infant categorization of 50,
513
hybrid stimuli, and infant
categorization 40
SUBJECT INDEX
interactionism 311
subtraction 273
mind 241
recognition 99
Japanese language
language
13545
and categorization 23
28, 57
546
single-process model of 43
understanding 261
language development 17
2534
see also first words; grammar;
speech; vocabulary
256
MacArthur Communicative
Development Inventories
91
development 93
26180
28896
conceptual change in
individuals 2902
classroom 3223
270
2745
situations 2701
2714
21416
and re-cognition 26
mind 238
brain) 12535
modularization 12732,
15861
nativist theories of 65
182, 185
196
mothers
environment 127
multiplicative reasoning,
343
344
nativism
defining 64
129, 1589
(UG)
patterns in 133
influences; nativism
38
neural activity 130
neuroimaging 116
neurons
dendrites of
development 1201
neuropsychology 125
neurotransmitters
nouns
106
categorization of 303
categorization of 48, 49
novelty-preference see
familiarization/noveltypreference method
object examination, and infant
categorization 47
object names
acquiring 778
and early comprehension
vocabulary 108, 109, 110, 111
147
control 221
peer relations
313
of constructivism 16
formation by 445
perceptual categorization 278
perceptual learning models, of
infant categorization 44, 48
categorization 44, 46
phonetic contrasts 81
phontactics 167
2812
pidgins, and the development of
language 136
21416
play
categorization 42
778
(PET) 1301
119, 120
1459
1479, 222
children 226
119, 120
SUBJECT INDEX
boundaries 72
303, 38
quantitative enrichment 43
question formation
language 1378
31819
reading
214
skill 223
reduplicated babbling 81
content 93
mind 244
326
Russian language, and infant
of mind 241
scaffolding 19
2867
2812
second-order theory of mind 2413
selectionism 1245
and the Hebb rule 133
language 166
syntax 1701
256
sign language
language 1367
development 254
1839, 197
skill acquisition
development 19
314
3257
changes in 54, 56
104
50
speech
63, 64, 73
834, 94
17381
recognizing 6573
6872
345
346
research 326
perception of 1023
268, 2713
6972
symbolic interactionism, Meads
theory of 31011
symbolizing 8
synaesthesia 125
synapses
121
1978
and grammatical development
180
developing 24551
2515, 256
understanding intentions
24950
2434
23840
understanding
transducers
128
problems 272
boundaries 6972
categorizing 25
of mind 2489
verbs
106
176
vocabulary
91, 92
variations in content 93
development 913
see also first words; object
names; speech
vocabulary spurt
174
teachers
and classroom dialogue 3202
and the developmental triangle
3234
classroom 323
constructivism 310
childrens understanding of
U-shaped development in
100, 102
Williams Syndrome
(WCST) 2234
morphology
word order see syntax
words see first words
variegated babbling 81
SUBJECT INDEX
21416
347