You are on page 1of 12

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/251712832

Text and Context in Early Language Comprehension

Chapter · January 1983

CITATIONS READS

13 135

2 authors, including:

Robin N Campbell
University of Stirling
62 PUBLICATIONS   503 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

History of Language Acquisition research View project

Explanations for drawing development between 4 and 7 years of age. View project

All content following this page was uploaded by Robin N Campbell on 25 April 2015.

The user has requested enhancement of the downloaded file.


Text and Context in Early Language Comprehension1

Robin N Campbell & Theresa Bowe Macdonald

Understanding is a mysterious process in which the information carried by an utterance or


inscription – by units of text – is adapted to and integrated with the listener's knowledge
based on prior discourse and the current situation – a complex structure which we will refer
to, rather globally, as a context. The outcome of this process is, generally, the reception of
what we shall call a message. The process of adaptation involves the fitting of both text to
context and context to text. For the text may guide the listener's attention to elements of his
current situation or to aspects of his world knowledge not so far implicated by the preceding
discourse thus modifying the context. Equally, the context may enable the listener to decide
between competing inter-pretations of the text or, indeed, to provide an interpretation of a
defective text containing illegible, impossible or unknown elements. Of course, texts are
themselves characterized by considerable redundancy, at all linguistic levels, and this
redundancy will enable a competent listener to guess the values of illegible, ambiguous or
unknown elements with a high probability of success even without the assistance of context.
Thus a listener may be able to make a guess at the intended message on the basis of the text
alone. We shall refer to such guesses as textual projections. On the other hand, contexts may
provide strong indications that a certain message is about to be communicated (think of the
mumbling foreigner advancing towards you with outstretched hand). In such cases the
‘listener’ may be able to guess the intended message on the basis of context alone. We shall
refer to this sort of guess as a contextual projection.

It therefore appears that in ordinary language comprehension text and context carry out
complementary functions and that the processes of textual and contextual projection are
symmetrical and interdependent. This general view of the nature of understanding is hardly
original.The first two essays in Ziff (1972) constitute a lengthier and more persuasive
presentation of a roughly similar view.

In this paper, we shall be concerned with the understanding of spoken texts containing
ambiguous words, or homophones as they are sometimes called. To make use of an example
from Donaldson (1978, p. 17) a child, who is asked by his teacher to ‘sit there for the
present’, has been presented with an utterance containing such an element, the homophone
present. Now it may be that the child knows only one meaning for present – 'gift' – or it may
be that the second meaning, though known to him, is one that comes less readily to mind,
through lack of use. In either case, textual projection will lead him to expect a gift. On the
other hand he may be sufficiently apprised of the ways of schools and teachers to know that
he is merely being told where to sit: in this case, contextual projection will have carried the
day and secured a valid interpretation. In the anecdote recounted by Donaldson, the
inappropriate textual projection prevailed and a grave disappointment resulted.

However, when we examine the experimental literature dealing with language


comprehension in young children, it seems that if there is any asymmetry in the relationship
between the processes of textual and contextual projection, it is textual projection that plays a
secondary role. For example, a number of studies with quantifiers (Donaldson and Lloyd,
1975; Donaldson and McGarrigle, 1974), relational terms (Donaldson and Balfour, 1968;
Carey, 1978) and prepositions (Clark, 1973; Hoogenraad et al., 1978) provide striking
instances of misunderstanding in preschool children, which are best explained not by
supposing that the child has assigned wrong values to these terms in analysing the texts
containing them, but by supposing that the results of such analyses are insufficient to
determine a message and, hence, a response. Consequently, the child is thrown back on the
resources of contextual projection, (commonly slender in these experiments) or, failing that,
of various standby procedures and response biases in order to produce his response.
Moreover, McGarrigle and Donaldson (1976) have argued that in situations such as the
number conservation tasks explored by them, where the results of textual and contextual
projection conflict (in their terms, where text and context are ‘uncoupled’), it is contextual
projection which prevails, the results of textual projection being set aside or, at least,
distorted.

The dynamical picture which this body of work suggests is one in which the child's
knowledge of the context and his expectancies based on that knowledge set inflexible prior
limits to the values that may be assigned to texts or elements of texts: the process of
comprehension is therefore only partially adaptive and functions accurately only when the
value that the speaker intends should be assigned to his text falls within these limits. There is
thus no significant adjustment of context to text. We shall refer to instances of
misunderstanding that arise in this way as cases of context-based misunderstanding.

The work that we will present here consists of some observations and an exploratory study
which provide evidence that, under certain circumstances, the child's comprehension may
exhibit an opposite asymmetry to the one we have just discussed, in which the results of
contextual projection are ignored or set aside when they conflict with the results of textual
projection. We shall call instances of misunderstanding that arise from this sort of process
text-based misunderstanding. After presenting our results we will discuss, very briefly, the
possibility of reconciling our findings with those of the scholars cited above -with whom we
are not in disagreement.

Our attention was first drawn to this phenomenon of text-based misunderstanding several
years ago when we encountered some very striking responses in a study of botanical
vocabulary. These came from three children interviewed in succession under more or less the
same circumstances. After 10-15 minutes of interview in which the elements of their
botanical vocabularies were elicited and their ranges of application discussed we showed the
children a picture of some pine cones on a twig and asked, ‘What are these things?’ The
children's responses and the ensuing discussion are given below:

Observation 1: Michelle (four years, six months)

What are these funny things? – Cones.

Cones. That's funny. Where d'you find cones? – (No response).

Where d'you get cones? – Off-off the cafe.

Off the cafe? What cafe? – (No response).

A cafe that you go to? – Uh-huh.

Whereabouts? – Causewayhead.

Causewayhead? Do they have cones there? – Uh-huh and wafers.

Aaaah! Ice-cream cones. But these are not ice-cream cones, are they? – No.

What kinds of cones are they? – They're brown cones.

Can you eat them? – No.

Why not? – Because they would poison you and make you dead.

Observation 2: Gillian (four years, three months)

What are these things? – Cones.

Where d'you get cones? – In the shop.

What shop? – Daddy's shop (an ice-cream shop).

Observation 3: Becky (three years, nine months)

What are these things? – Things to go in your hair.

Where do you find things like that – I don't know.


(Repeated). What are these things? – Comb. Comb your hair.

Now, of course, of the various meanings the word cone can take in English, undoubtedly
'ice-cream cone' is the one most familiar to young children. The most plausible explanation
of the first two observations is then that textual projection of ‘Where do you find/ get cones?’
attaches such a high probability to the meaning ‘ice-cream cone’ that the information
obtainable from contextual projection is ignored, overruled or simply not extracted.

In the case of the third observation a simpler explanation is possible. It could be that this
child has observed (at the time of original learning, perhaps) that pine cones exhibit a strong
functional similarity to combs, at the same time assimilating the phonetic form (ko:n) to
(ko:m). If we allow this assumption, there is no contextual inconsistency attaching to her
answer. The cones are simply being identified by means of their mistakenly supposed human
function, just as berries might be (correctly) identified by remarking that they are ‘things to
eat’. The oddity of this observation is then (under this explanation) traced back to an
idiosyncratic assimilation at the time of original learning. If we reject this ad hoc explanation
and look for one along the lines of the explanation offered for the first two observations, then
we are little better off, since the story we must tell is an exceedingly unlikely one, namely,
that the pine cones elicit the form (ko:m) as a response but that this is too faint or obscure to
be retrieved. The link from (ko:m) to ‘things to go in your hair’ is next excited, being
stronger than the link to ‘things you find in the woods’ – that is to say, favoured by textual
projection – and the response produced. Although this type of observation is unreplicable
(except by pure luck) we have negative evidence that suggests it is the simpler sort of
explanation which is correct. In an attempt to replicate the first type of observation with 14
children of mean age 3;10 (range 3;2 to 5;0 years), and using five ambiguous words (bat,
hair/hare, cone, letter, bulb) as ‘targets’, we obtained 70 responses to questions of the same
type as ‘Where do you get/ find cones?’. In no case did any child produce a response fitting
our second, complex explanation of Observation 3. On the other hand there were many
responses that showed idiosyncratic assimilation, e.g. bats were called frogs, mice, owls;
hares were called kangaroos; cones were called flowers and bulbs (that is, e.g. daffodil bulbs)
were called mushrooms and apples. It therefore seems unlikely that we need to consider this
remarkable explanation of Observation 3 further and that it is the first, simpler explanation
which is correct.

In an attempt to show further evidence of text-based misunderstanding at work we carried


out another experiment using a sample of 24 children from a local playgroup. The sample
contained 14 girls and 10 boys, and consisted of three groups of eight children (mean ages
3;8, 4;3 and 4;9). In this study the children were told the following story:

One morning, Jane and her Mummy went to see Jane's auntie who works in the post
office. She would like to work in the big post office but she works in a branch (1).
When they got home it was lunchtime. ‘We must set the table,’ said Jane's Mummy,
‘Can you put the leaf up and put the tablecloth on?’(2) That afternoon, Jane and her
Mummy and Daddy went for a picnic in the car. As they were driving along they saw
a hare run across a field.(3) They stopped the car and got out to look but the hare had
gone. The field was full of corn and they saw a little mouse sitting on one of'the ears.
(4) Then they got back into the car and drove to the seaside. When they got there they
went tor a walk along the quay.(5) They saw a big buoy with orange and white stripes
floating in the water. ‘What's that buoy made of?’ asked Jane. ‘The buoy's made of
rubber and full of air,’ said her Mummy.(6) The sea was rather rough so they went
for a walk in a wood full of fir trees. Underneath the trees were lots of cones. They
had fallen from the branches. (7) At the far side of the wood was a castle. ‘Look at
this castle,’ said Jane's Daddy. ‘The oldest wing is over 500 years old.’(8) They went
into the castle. It was very dark inside and there were bats flying around.(9) Soon
they got back into the car to go home. They got held up behind a lot of other cars, all
going very slowly. ‘I hope we get out of this jam soon,’ said Jane's Daddy.(10)

The underlined words are all, once again, ambiguous and are used here in their unfamiliar,
secondary senses. At the points (1) - (10) in the course of the story the child was questioned
in such a way as to establish his under-standing of the text containing the ambiguous word.
Children were first of all encouraged to draw, or to complete a picture, by drawing ‘the hare
running across the field’ (in this case a drawing of a field was provided) or the like. In each
case an interpretation which followed the primary, familiar meaning of the word would be
judged as fantastic or grotesque by an adult. That is, a text-based comprehension process is
always inadequate and leads to misunderstanding. The children's responses were examined
and classified as follows:

(1) Secondary responses. These were correct (adult-like) interpretations based on knowledge
of the secondary meaning of the word.

(2) Primary responses. These were incorrect and grotesque text-based interpretations based
upon the primary meaning, e.g.

Observation 4: Caroline (three years, eleven months)


Can you draw the hare running across? – (Draws mass of hair).

What does a hare look like? – (Touches hair).

Do you think it would be running across a field? – Yes. . . . . .

They went for a walk along the quay...Can you draw the quay? – (Drawing obscure).

What sort of thing is a quay? What's a quay for? – For opening doors.

Do you think they could walk along a quay? – (Nods).

Observation 5: Elizabeth (four years, four months)

Can you draw the hare for me? – I can't draw hair.

Why can't you draw it? – Too little.

Why do you think a hare was running across a field? – Don't know.

Have you ever seen a hare run across a field? – (Nods).

What did the hare look like? – Don't know.

What sort of thing is a hare? – (No response).

Where do you normally see hares? – On your head.

Do you think it was a hair from your head running across a field? – Yes.

That's a bit funny, isn't it? – (Nods).

Some representative (and amusing) drawings are shown in Figure 1.

Responses to the word buoy were all primary responses, suggesting that there is no conflict
with context in this case. That is, young children may not regard it as odd that a boy should
have orange and white stripes, be made of rubher and full of air! Some support for this view
is provided by the following case:

Observation 6: Alastair (four years, eight months)

Can you draw the big rubber buoy? – Yeah. I here's his body. (Draws – see Figure 1)

What was the buoy made of? – Rubber.

Do you often get buoys made of rubber? – Yeah. So'm I made of rubber, and there's
bones inside there.
(3) Inventions. These were contextually-consistent interpretations not based on knowledge of
the secondary meaning of the word, e.g.:

Observation 7: Lisa (four years)

What did Jane have to do? – Put the tablecloth on.

What did she have to do before she put the tablecloth on? – Don't know.

I think she had to put the leaf up. – She put the leaf up.

Do you know what the leaf of the table is? – (No response).

Which bit is that? – The cloth.

(4) Don't know. These were uninformative responses of the usual sorts, e.g. I don't know, I
can't remember, no response, etc.

(5) Unclassifiable.

There were also interesting subcategories of types (2) and (3), called blends. A primary
blend occurs when the child's interpretation, though text-based, shows some moderating
influence of contextual projection, e.g.:

Observation 8: Donna (four years, seven months)

Do you think you could draw the castle? – (Draw, see Figure 1).

Can you draw the old wing as well? – (No response).

The castle in the story has an old wing. – (Draws).

What do you think the wing was for? – (No response).

Do castles generally have wings? – (Nods).

And what are the wings for? – (No response).

What do you think that wing's made of? – (No response).

(Repeats question). – Wood.

What's the rest of the castle made of? – Stone.

Why's the wing made of wood? – (No response)

Are there lots of rooms in the castle? – (Nods).

How about the wing? Does that have lots of rooms? – (Shakes head)

Why not? – (No response).

What else has wings as well as a castle? – A bird and a bat.

Do you think it's the same sort of wing as the castle had? – (No response)

What sort of wing do you think the castle had? – Wooden.

What are a bird's wings made of? – Feathers.

What are a bat's wings made of? – (No response).

The point about this observation is that Donna insists that the wing should be made of
wood rather than feathers, showing (presumably) awareness that it is impossible for a stone
building to sprout such an organ from its flanks!

Inventive blends, on the other hand, are contextually consistent but show the influence of
the primary meaning of the word, that is, of textual projection. Examples of this kind of
response are relatively abundant, e.g. 'snake' and 'grass' as an interpretation for the hare
running across the field, 'mud' and 'sand' for the traffic jam.

This analysis left rather a large number of unclassifiable responses. However, in many
cases we could narrow down the possibilities to just two categories. For example, it was quite
often the case that a response was evidently either a secondary or invention but that it was
impossible to say which (this occurred particularly with words where the secondary
interpretation was guessable with high probability, e.g. branch, leaf). Minimally ambiguous
cases like these were allocated half to one category and half to the other. The residue of cases
formed the fifth category of unclassfiable responses. The results of this classification are
shown below:

Table 1: Incidence of response types by age groups

Response types

(1) (2) (3) (4) (5)

Age groups Secondary Primary Invention Don't know Unclassifiable

Mean age: 4;9 32% 17% 24% 19% 8%

Mean age: 4;3 18% 40% 11% 27% 4%

Mean age: 3;8 23% 33% 8% 13% 23%

All groups 25% 31% 15% 19% 10%

As expected, there are a good many cases (31%) of primary responses, confirming the
existence of text-based comprehension processes in children of this age. Of course, this
figure, although high, is by no means overwhelming: a good many children managed to
integrate text and context adaptively. Moreover, there were a small number of cases,
increasing across the age-groupings, of successful interpretation which involved inventive
resolution of the conflict between text and context. An important qualification must be made
about such inventions: we cannot conclude from a successful invention that the child has
learned anything new about the meaning of the word. A good example of this is given below;

Observation 9: Nerissa (three years, eight months)

Where is it that Jane's auntie works? – In the post office.

What sort of post office? – (No response).

Did she work in the big post office? – (Shakes head).


What sort of post office did she work in? – A little one.

It says in the story that she works in a branch. Is that right? – (Shakes head).

Why is that not right? – (No response).

Do you think she could work in a branch? – (Shakes head).

Why not? – Because all her things would fall out.

What would it fall out of? – The branches.

Where are the branches? – On the tree.

Notice that, although Nerissa (a) makes a contextually plausible response and (b)
recognizes that a primary response is impossible, there is no evidence that she has attached
any new value to branch as a result of this experience. That is, although there is fitting of text
to context here, this assimilation is global and not articulated to the extent that the impossible
element (branch) acquires a new value.

Our discussion of these results will be brief, since our goal here has been simply to
establish the existence of a certain sort of misunderstanding. It seems to us that we have
shown that young children will, under some circumstances, be guided by their knowledge of
words to mistaken interpretations which are strongly in conflict with the current context and
with their knowledge of the world. If we consider this type of misunderstanding along with
the more familiar sort where contextual projection leads them to a message which is
inconsistent with the text, it may be concluded that what is characteristic of the young child
is a difficulty in integrating the two processes of textual and contextual projection, on the
relatively rare occasions when they yield conflicting results. It is important to grasp the point
that in our demonstrations, just as in those by the other scholars referred to above, it is only
by contrivance that the child's comprehension is made to fail. In general, it may be very
efficient to favour the stronger projection, be it textual or contextual. After all, there must be
many occasions in the life of a child when his contextual expectations are weak or absent and
where he must be guided by his understanding of our words.

Equally there will be many occasions where our words are not understood and he must be
guided by contextual projection. A child who stopped short of inferring a message whenever
the information available to him from textual or contextual projection was deficient or
conflicting would require frequent instruction and supporting explanation. It is plain that
young children are not like this. Rather, they enthusiastically construct messages in the most
unpromising circumstances. Accordingly, the sorts of misunderstanding that we and the other
scholars referred to earlier have uncovered are perhaps to be regarded as the inevitable
consequence of what is, from the point of view of parent or teacher, an extraordinarily
efficient system of communication.

References

Campbell, R. N. & Bowe, T. 1977. Functional asymmetry in early language understanding.


In G. Drachman (ed.) Salzburger Beitrage zur Linguistik, Band 4 (Akten der 3.
Salzburger Jahrestagung fuer Linguistik), 13-27. Salzburg. Wolfgang Neugebauer.

Campbell, R. N. & Macdonald, T. B. 1983. Text and context in early language


comprehension. In Donaldson, M. C., Grieve, R. & Pratt, C. (eds.) Early Childhood
Development and Education, pp. 115-126. Blackwell Books.

Carey, S. 1978 ‘Less’ may never mean more. In Campbell, R. N. & Smith, P. T. Recent
Advances in the Psychology of Language, Volume 1, 109-132.

Clark, E. V. 1973. Non-linguistic strategies and the acquisition of word meanings. Cognition
2, 161-182.

Donaldson, M. C. & Lloyd, P. 1975. Sentences and situations: children’s judgments of match
and mismatch. In Bresson, F. (ed.) Current problems in psycholinguistics. CNRS.

Donaldson, M. C. & McGarrigle, J. 1974. Some clues to the nature of language acquisition.
Journal of Child Language 1, 185-194.

Donaldson, M. C. & Balfour, G. 1968. Less is more: a study of language comprehension in


children. British Journal of Psychology 59, 461-471.

Donaldson, M. C. 1978. Children’s Minds. Fontana.

Hoogenraad R., Grieve, R., Baldwin, P. & Campbell, R. N. 1978. Comprehension as an


interactive process. In Campbell, R. N. & Smith, P. T. Recent Advances in the
Psychology of Language, Volume 1, 163-186.

McGarrigle, J. & Donaldson, M. C. 1976. Conservation accidents. Cognition 3, 341-350.

Ziff, P. 1972. Understanding understanding. Cornell U.P.

1
The research reported here was carried out as part of an SSRC granted project ‘Semantic
Development in Young Children’ in collaboration with Robert Grieve and Robert
Hoogenraad. The original report appeared as Campbell & Bowe (1977) and was republished
with some minor modifications as Campbell & Macdonald (1983).

View publication stats

You might also like