Transcript PSY 369: Psycholinguistics - the Department of Psychology at

PSY 369: Psycholinguistics
Language Acquisition:
Morphology
Language explosion continues

Morphology


Typically things inflections and prepositions start around
MLU of 2.5 (usually in 2 yr olds)
Wug experiment (Berko-Gleason, 1958)
Quic kTime™ and a
TIFF (LZW) decompress or
are needed to see t his pic ture.
Quic kTime™ and a
TIFF (LZW) decompress or
are needed to see t his pic ture.
Quic kTime™ and a
TIFF (LZW) decompress or
are needed to see t his pic ture.
Here is a
wug.
Now there are
two of them.
There are two
_______.
Acquiring Morphology
I holded the baby rabbits.

Children sometimes make mistakes.

This is ungrammatical in the adult language


its existence in child language shows that children are not
simply imitating what they have heard.
In this case, what they produce is not in their input.
Acquiring Morphology
I holded the baby rabbits.

Why do they make errors like these?



In the case at hand, we have what is called
overregularization
The verb hold has an irregular past tense form, held
Because this form is used, the regular past tense-that with -ed-- is not found (*hold-ed)
Where

Quic kTime™ and a
TIFF (LZW) decompress or
are needed to see t his pic ture.
Quic kTime™ and a
TIFF (LZW) decompress or
are needed to see t his pic ture.
Remember that regular forms require no stored
knowledge (wug test)

Whereas something must be memorized with irregulars

Examples:



Horton heared a Who
I finded Renée
The alligator goed kerplunk
Acquiring Morphology

Stages in the acquisition of irregular inflections


With regular verbs, the default form -ed is used
With irregulars, lists associating the verb with a particular
form of the past tense have to be memorized:


Past tense is -t when attached to leave, keep, etc.
Also patterns of stem-changing
Acquiring Morphology

Stages in the acquisition of irregular inflections
Step
1
2
3
4
5
Description
No inflection
Adult form
Overregularization
Transition
Adult form
Noun
Man
Men
Mans
Mens
Men
Examples
Verb
Adjective
Go
Bad
Went
Worse
Goed
Badder
Wented Worser
Went
Worse
Acquiring Morphology

Stages in the acquisition of irregular inflections
 On the face of it, learning these morphological
quirks follows a peculiar pattern:




Early: correct irregular forms are used
Middle: incorrect regular forms are used
Late: correct forms are used again
Why do we find this type of pattern?

Memory and rules
Frequency

It is possible to predict which verbs will be subject to
overregularization

We find here a frequency effect


The more often an irregular form occurs in the input, the less
likely the child is to use it as an overregularization
This is evidence that some part of overregularization occurs
because of memory failures

Something about irregulars is unpredictable, hence has to be
memorized
Rules

The use of overregularized forms starts at around the
same that that the child is beginning to apply the
default -ed rule successfully



Early: All forms-- whether regular or irregular-- are
memorized
Middle: The regular rule is learned, and in some cases
overapplied
Late: Irregulars are used based on memory, regulars use the
rule (the idea is that if the word can provide its own past
tense from memory, then the past tense rule is blocked)
An interesting pattern

Overregularization doesn’t replace correct forms

Instead it replaces errors

Early stages



74% correct (e.g. he held it)
26% bare form (e.g. he hold it)
Middle stages: Overregularization is not very frequent
(typically less than 5%)



89% correct
9% bare form
2% overregularized
Summary

Overregularization looks at first like children
are moving backwards



On closer examination, the child’s overall
performance is improving
The pattern of overregularization provides a
window on the process in which the child
(over)generalizes a rule
So how is the rule learned (learnt?)
What kind of “teaching” do kids get?

What kind of feedback is available for learning?

Are the kids even aware of mistakes?

The children are apparently aware of the fact that their
forms are strange:




Parent: Where’s Mommy?
Child: Mommy goed to the store
Parent: Mommy goed to the store?
Child: NO! Daddy, I say it that way, not you
Positive and negative evidence

Kinds of feedback


Positive evidence: Kids hear grammatical
sentences
Negative evidence: information that a given
sentence is ungrammatical

Kids are not told which sentences are ungrammatical
(no negative evidence)

Let’s consider no negative evidence further…
Negative evidence

Negative evidence could come in various
conceivable forms.


“The sentence Bill a cookie ate is not a sentence in
English, Timmy. No sentence with SOV word order is.”
Upon hearing Bill a cookie ate, an adult might
 Offer negative reinforcement
 Not understand
 Look pained
 Rephrase the ungrammatical sentence grammatically
Kids resist instruction…
McNeill (1966)



Child: Nobody don’t like me.
Adult: No, say ‘nobody likes me.’
Child: Nobody don’t like me.
[repeats eight times]


Adult: No, now listen carefully; say ‘nobody likes me.’
Child: Oh! Nobody don’t likes me.
Kids resist instruction…
Cazden (1972) (observation attributed to Jean Berko Gleason)







Child: My teacher holded the baby rabbits and we patted them.
Adult: Did you say your teacher held the baby rabbits?
Child: Yes.
Adult: What did you say she did?
Child: She holded the baby rabbits and we patted them.
Adult: Did you say she held them tightly?
Child: No, she holded them loosely.
Negative evidence via feedback?

Do kids get “implicit” negative evidence?


Do adults understand grammatical sentences and not
understand ungrammatical ones?
Do adults respond positively to grammatical sentences
and negatively to ungrammatical ones?
Negative evidence via feedback?
Brown & Hanlon (1970):




Adults understood 42% of the grammatical sentences.
Adults understood 47% of the ungrammatical ones.
Adults expressed approval after 45% of the
grammatical sentences.
Adults expressed approval after 45% of the ungrammatical
sentences.
This doesn’t bode well for comprehension or approval as a source of
negative evidence for kids.
Negative evidence via feedback?

Complete: consistent response, indicates
unambiguously “grammatical” or
“ungrammatical.”

Partial: if there is a response, it indicates
“grammatical” or “ungrammatical”

Noisy: response given to both grammatical and
ungrammatical sentences, but with
different/detectible frequency.
Negative evidence via feedback?
Marcus (1993)

Suppose response R occurs



Kid gets response R to utterance U, there’s a 63%
chance (20/32) that U is ungrammatical.


20% of the time for ungrammatical sentences,
12% of the time for grammatical sentences.
Guess: ungrammatical, but 38% chance of being wrong.
Ok, but nowhere near good enough to build a
grammar.
Lacking confidence

This is a serious task, a kid’s going to want to be
sure.


Based on R (20%-12% differential)


Suppose kid is aiming for 99% confidence (adults make
at most 1% speech errors of the relevant kind—pretend
this reflects 99% confidence).
they’d have to repeat U 446 times (and compile feedback
results) to reach a 99% confidence level.
This sounds rather unlike what actually happens.
In a way, it’s moot anyway…

One of the striking things about child language is
how few errors they actually make.



For negative feedback to work, the kids have to make the
errors (so that it can get the negative response).
But they don’t make enough relevant kinds of errors to
determine the complex grammar.
Pinker, Marcus and others, conclude that much of
this stuff must be innate.

But this isn’t the only view. There is a raging debate about
whether there are rules, or whether these patterns of
behavior can be learned based on the language evidence
that is available to the kids