Transcript Powerpoint format - rufla

Rutgers Forum for
Language Acquisition
November 5, 2002
1
Influence of Context on
Verbal Inflection:
An experimental study with
Spanish-speaking children
María Blume
Cornell University
2
1. Introduction
 What do children know about the interface
between Pragmatics and Grammar?
 Pragmatics
= Context
 Grammar = Tense and Aspect Verbal Inflection
3
Examples
(1)
Adult: ¿Qué hace?
What do-Present-3-sg
'What does it do/ what is it doing?'
Child: Nadando
Swim-Present Participle
'Swimming'
(2)
Adult: ¿Qué
estás
What
be-Present-2-sg
'What are you doing?'
haciendo?
do-PP
Child: Jugar.
Play-Infinitive
'Lit. To play/ Playing'
4
2. Pragmatics:
Use of language in context.
Communication: 2 important aspects
Decoding
Interpretation
Interpretation. Example:
(3) Can
you open the window?
Usual pragmatics studies look at issues like
politeness, requesting, etc.
5
2.1 Decoding : The missing
information problem
But pragmatics also deals with filling in the gaps needed
for correct decoding. Pragmatic context (the situation) is
necessary to find the right referents.
(4)
I'll come tomorrow.
•I who?
•come where?
•tomorrow what date?
(5)
Betsy's gift made her very happy
•Betsy who?
•her who?
Examples from Sperber & Wilson (1995: 10)
6
But in other cases, like in (6) which is equivalent to
Spanish sentence (1), the context is necessary to
understand the meaning of the sentence itself
(6)
Speaker 1: What is the doggy doing?
Speaker 2: Swimming.
Swimming who, when?
(person & number features plus tense features)

Note that this answer is ungrammatical in isolation
7
 These answers require an interaction between
morphosyntax and discourse.
 Their missing tense and aspect features have
to be recovered from the preceding question.
 A Spanish-speaking child has to realize that
although Spanish grammar requires most
verbs to be inflected, the inflection can be
dropped in particular cases such as these, in
which the discourse context crucially provides
the missing information.
8
3. Why Spanish?
(7) Question:
¿Qué está haciendo el
perro?
What is doing the dog?
'What is the dog doing?'
Possible answers:
A: está comiendo
is eating
'it is eating’
B: come
eats
'it eats'
C:
D:
comiendo
eating
‘eating’
comer
to-eat
to eat
comiendo 'eating' Missing
person, number & tense
features.
comer 'to eat' Missing
person, number, tense &
aspect features.
9
3.1 Ambiguity of Imperfect
aspect problem
 Imperfect aspect forms in Spanish can have
two different interpretations:
(8) Question: ¿Qué hace el perro?
What does the dog?
'What does the dog do?'
Habitual: What does the dog usually do? (√ English)
 Ongoing activity: What is the dog doing?
(X English).

10
Questions in Spanish and their
possible answers: Progressive
questions
 Ongoing activity (Now)
 ¿Qué está haciendo el perro?
 A. Come
 B. Está comiendo
 C. Comiendo
 D. Comer
--> √
--> X
--> √
--> √
--> X
11
Questions in Spanish and their
possible answers: Imperfect
questions
 Ongoing activity (Now)
¿Qué hace el perro?--> X
A. Come
--> X
 B. Está comiendo --> √
 C. Comiendo
--> √
 D. Comer
--> X

 Habitual activity
(Usually)
¿Qué hace el perro?
A. Come
 D. Comer

--> √
--> X
12
4. What does the child need to
know to produce these
answers?
 This is a question, it requires an answer.
 Meaning of verbs.
 Which is the form of the verbs in
Spanish: root + TV cant-a
 Which of these formal features are
realized in Spanish. (Person, number,
tense, aspect)
13
Where are they realized?
Imperfect forms: all in the main verb
• Tú com-e-s
Progressive tenses: person, number
and tense on the auxiliary, aspect on
the main verb.
• Tú est-á-s com-ie-ndo
 In what order: first tense-aspect, 2nd
person-number
14
 Which particular morphemes
correspond to the formal features in
Spanish.
 When can you forget all this and drop
features, i.e., use an Infinitive or a
Present Participle.
 When you use a non-finite form , the
subject of the utterance has to be null,
as in (9)
15
(9) Question:
Answer C:
Answer C':
Answer D:
Answer D':
¿Qué está haciendo el perro?
'What is the dog doing?'
comiendo
√
'eating'
él comiendo
X
'it eating’
comer
√
'to eat’
él comer
X
'he to eat'
16
5. Influence of Context in the
Answer’s Meaning
 When the question is ambiguous (Present
Imperfect and Past Imperfect)
 What is the effect of the context?
 Discourse
context: The preceding question
 Referential context: Story book (which showed
the characters doing the activities so it favored
an ongoing-activity interpretation).
17
6. Method: 2 experiments.
Elicited Imitation and Elicited
Production
 All experiments were administered orally.
 Subjects
 27
children, ages from 2 years 1 month to 3
years 11 months.
 14 adults (adults did Grammaticality Judgment
instead of Elicited Imitation)
18
6.1 Elicited Imitation: Discourse
Context Only
 The child heard a question and an answer and
was asked to imitate the answer exactly as the
experimenter said it.
 Children imitate exactly the structures they
have acquired, and change the structures they
have not yet acquired.
19
To imitate a sentence, the child, has
to process it, understand it, save it
in his/her memory and produce it.
(10) Adult: The red beads (Ø) and brown beads
are here.
Child: Brown beads here an' a red beads here
(2;03;03)
(11) Adult: The owl eats candy and (Ø) runs fast.
Child: Owl eat candy…owl eat the candy
and…he run fast
(2;04;03).
Slobin and Welsh (1973)
20
My experiment:
 Pretraining session: The game was called el juego del
lorito (‘the parrot game’).
 Animals sounds. What kind of does sound the parrot
make? Do you want to be the parrot?
 The puppet.
 The child heard a question, by the puppet, and an
answer, by the experimenter, and was asked to repeat
the answer.
 The child heard the target utterance at most twice.
21
Video 1
 Scoring
 Correct
or Incorrect Imitation
 If the imitation was incorrect, what did the child
change?
22
Results for Elicited Imitation:
Incorrect Imitation of
Progressive
23
24
Interpretation
 They can produce all answers
 Imperfect
= Come ‘eats’
 Progressive = Está comiendo ‘is eating’
 Present Participle = Comiendo ‘eating’
 Infinitive = Comer ‘to eat’
25
Adult results:
26
27
Comparing adults and children.
 Adults show an effect for discourse context.
 Imperfect
Questions: Both Imperfect and
Progressive are high.Present Participles are
low.
 Progressive Questions: Imperfect declines.
Present Participle increases.
 Infinitives are accepted with all types of
questions.
 Children do not show effect for type of
question.
28
Why don’t they imitate the
Progressive well?
 They are changing it to Present Participle
 Question:
What is the hen doing?
 Answer: She is sweeping.
 Child’s Imitation: Sweeping
 These forms seem to be equivalent for them.
29
Examples
(12) E: ¿qué hace la vaca?
'What does the cow do?/What is the cow doing?’
E: está cantando.
'(it) is singing’
Child: cantando.
'singing'
(AJ091295, 3;00;12)
30
(13) E: ¿qué estaba
haciendo
el
pollito?
'what was the chicken doing?’
E: hablando.
'talking’
Child: hablando estaba (CC090996, 3;10;25)
'lit: (it) talking was/(it) was talking’
(14) E: ¿qué hace la oveja?
'what does the sheep do?/What is the sheep
doing?’
E: paseando. (AV012796, 2;07;23)
'strolling'
Child: …estaba la oveja paseando.
'lit. was the sheep strolling/the sheep was strolling
31
6.2 Elicited Production:
Discourse and Storybook
 What children produce naturally when given an
option. (Closest to natural speech)
 Constraint the context to have the subject
produce a particular answer.
 Allows experimenter to collect a large set of data
about a particular form
32
My experiment
 Subject questions: Discourse only, without
book
 All other: Storybook
33
Procedure
 Pretraining: Making the child tell the story.
 Eric Hill’s “Spot’s big book of words”.
 Animal names.
 Crocodile = Dragon, Dinosaur
 Hippopotamus = Bear, Cow
Hill, Eric (1998) "Spot's big book of words". New York: G. P. Putnam & Sons.
34
Felicity Conditions: Eliciting
Past Tense
 Pilot studies, Thornton (1996)
 She argues that adults have better capacity
than children to accommodate pragmatic
infelicities.
 This was not true with my experiments.
35
Examples:
36
Video 2
 Scoring
 Did
the verb in the answer had the same tense
an aspect as the verb in the question?
 If not, what form was used?
37
Results for Elicited Production:
Questions with Storybook
(Ambiguous Imperfect and
Progressive)
 With the book (extensive
context) children match more
the progressive forms than
the imperfect ones.
 They are clearly marking the
ongoing activity
interpretation.
 They don’t match with
imperfect because they are
answering with Present
Participles or Progressive.
38
Comparison with adults
 Adults show the contrary
pattern. They match
more with Imperfect than
with Progressive.
 In general adults tend to
match the question’s
form.
 Less match with
Progressive because
they allow Present
Participles.
 More attention to
discourse than children.
39
Results for Elicited Production:
Questions without Storybook
(Habitual Imperfect and
Progressive)
 Children pay more
attention to the
discourse context
without the book.
 Here they do match with
the Imperfect, since it
has habitual meaning
and it cannot be
answered with a Present
Participle.
40
Comparison with adults.
 Here adults and children
show the same pattern.
 However, note that
adults are producing
more matching with
Imperfect than children.
41
General Results
 They notice the difference between an
ambiguous and a non ambiguous imperfect
question.
 When it is ambiguous, they answer with an
ongoing-activity form (Progressive or Present
Participle). (Effect of the Book).
 When it is habitual, they answer more with an
Imperfect form. (Discourse context alone)
42
The 2 ;00 to 2;05 children
 It was very difficult to make the younger group
answer the questions about themselves.
 Three possible explanations.
 These
questions were more difficult for them
due to lack of extensive context.
 They may have had difficulty imagining
themselves at different times and situations.
 They may have just been uninterested in these
questions since they liked the storybook so
much that they wanted to continue reading it.
43
Development with Ambiguous
Imperfect.
 Progressive is very high
from almost the
beginning and it
decreases with age.
 Use of Imperfect
increases. More
attention to discourse
context.
 2;00-2;05 group
produced few
Progressive forms, and
no Past Progressives.
 They do not show adult
pattern
44
Development with Habitual
Imperfect
 Progressive remains
constant (except for
younger group).
 Increase in the use of
Imperfect.
 Same pattern as adults
starts at age 3;00-3;05.
45
Incorrect answers to habitual
questions by age group.





Age Group
2;00-2;05
2;06-2;11
3;00-3;05
3;06-3;11
Present
Past
2 (22.2%)
0 (0%)
3 (37.5%)
4 (66.7%)
3 (25%)
4 (33.3%)
2 (18.2%)
1 (14.3%)
46
Errors with Habitual Questions
(15) E: Do you go to your
grandmother’s house?
E: What do you do
at your
grandmother’s
house?
Child: eating
(DR050398, 2;02;25)
(16) E: listen, and what else
did you do when you were
a baby?
Child: (I) was doing […]
drawing with a-with some
(17) E: do you sometimes
go to the park, [Child’s
name]?
Child: mmh.
(affirmative sound)
E: What do you do at
the park?
Child: playing.
(AP031597; 3;03;25)
47
Why do they answer habituals
incorrectly?
 Two possibilities:
 They do not understand the question asks
about habitual activities.
 They think one can answer habitual questions
with a Progressive form (they think it is
synonymous with Imperfect)
48
6.3 Comparing both tasks
 Both adults and children show effect of
referential context (book) by producing more
ongoing-activity forms (Progressive and
Present Participle).
 Adults show effect of discourse context by
matching the question’s form. Children only
show this effect with habitual questions.
 Adults reject Present Participle answers to
Imperfect questions, but produce them in the
Elicited Produsction task (23%)
49



Children produce and imitate Present Participles
very well with all question forms. They are the
favorite forms.
Regarding infinitives, both adults and children
accept them, but they don’t produce them much
(adults up to 4.7% in total, children up to 10.9% in
total).
Children develop in:
• Their awareness of discourse context.
• Their knowledge about habitual questions.
• Their production of Progressive forms, especially
Past Progressives (imitated but not produced by
2;00-2;05 group). They imitated them better when
the question was in the Past Progressive.
50
References
Blume, M. (2002) Discourse-morphosyntax interface in Spanish nonfinite verbs: A comparison between adult and child grammars.
Ph.D. Dissertation, Cornell University.
Hill, E. (1998) Spot’s big book of words. New York: G. P Putnam &
Sons.
Slobin, D. I. & Welsh, C. (1973) Elicited imitation as a research tool in
developmental psycho-linguistics. In C. A. Ferguson & D. I. Slobin
(eds.), Studies of child language development. New York: Holt,
Rinehart and Winston.
Sperber, D. and Wilson, D. (1995) Relevance: communication and
cognition. Oxford: Blackwell.
Thornton, Rosalind (1996) Elicited Production. In McDaniel, Dana;
McKee, Cecile, and Smith Cairns, Helen Methods for assessing
children's syntax (pp. 77-102). Cambridge, MA: MIT Press.
51
Acknowledgments
 Prof. Liliana Sánchez & Prof. José Camacho, Rutgers
University.
 The Cognitive Studies & European Studies Programs,
and CISER at Cornell University.
 Prof. Barbara Lust, Cornell University.
 Sarah Callahan, Makeba Parramore and Smriti Shetty.
 All the children, parents, teachers and adult subjects
that allowed me to work with them at Peru and Spain.
52