Transcript Stackhouse & Wells model (1997)

Language and Cognition
Colombo, June 2011
Day 3
Child Language and Disorders
Plan
• Questions of innateness
• Modelling language processing in children
• Acquisition of syntax
Why study language development?
• Acquisition of a complex cognitive system:
understanding how children do this means we are
understanding something substantial about how the
human mind works
• Species-specificity: determining what it is that makes
humans unique (if we are unique)
• Some children have difficulties acquiring language
even in the absence of other apparent learning
difficulties
• Cultural differences in language use can lead to
difficulties in mainstream education
Issues in the study of language
development
• Nature vs nurture: does the child acquire
language from the environment, or is it
genetically pre-programmed?
• Too simplistic: it’s certainly some combination
of these two factors
• The debate now is more concerned with the
“nature of nature”; what / how much is
innately specified, and how much must be
acquired
Is language innate?
Is language innate? Some arguments
• no negative evidence
• species-specificity
• specified neurological and genetic
underpinnings
• speed of language acquisition: around 2
years for most of the groundwork (cf.
shoelace-tying)
• critical periods
No negative evidence
Father: where is that big piece
of paper I gave you
yesterday?
Abe: Remember? I writed on
it.
Father: Oh, that’s right. Don’t
you have any paper down
there buddy?
Species-Specificity
“. . . it is now widely recognized that these efforts have
failed, a fact that will hardly surprise anyone who gives
some thought to the matter. The language faculty
confers enormous
advantages on a species
that possesses it. It is
hardly likely that some
species has this
capacity but has
never thought to use
it until instructed by
humans. That is
about as likely as
the discovery
that on some
remote island
there is a
species of bird
that is perfectly
capable of flight
but has never
thought to fly until
instructed by
humans in this
skill.”
(Chomsky 1988).
Localization of function
• Phrenology
– Gall, Spurzheim, early 1800s
• Different cognitive functions can be localized
to different parts of the brain
• Level of development of a particular function
is reflected in skull formation
• The sad tale of Phineas Gage
• Dissociation of language from other cognitive
faculties
Localization
• Paul Broca (1861): patient ‘Tan’ • Carl Wernicke: patients with
posterior lesions in the left
• Slow, effortful, nonfluent
hemisphere
speech with many omissions;
but good comprehension
• comprehension is impaired
but speech is fluent
• on parle avec l’hemisphere
gauche
Genie: The Wild Child
Genie: The Wild Child
•
•
•
•
Locked up in a room without language
Discovered at age 13 years
Could she recover language?
Is there a critical period for learning language?
Special cases
Johnson & Newport: Critical Period Effects for L2
Cerebral plasticity and language
dominance
• Damage to language areas in young children may be
associated with a shift in language function to the right
hemisphere
• Plasticity is greater in younger brains – pathways are still
being formed – “Equipotentiality” (Lenneberg 1967)
• But the LHS seems ‘specialized’ for language – there are
usually some functional differences in cases of early childhood
hemispheric shifts
• Language dominance appears to be established before birth
– Planum temporale asymmetries are apparent as early as the third
trimester
– Early hearing screenings show a right-ear advantage for linguistic
stimuli
Critical period
•
•
•
•
Evidence from Wild Children
Evidence from L2 acquisition and attainment
Evidence from plasticity
Additional evidence: children who are not
exposed to sign language
Developmental sequences
What we study
• What’s the nature of the Language Acquisition Device?
• How does it interact with environmental and social
influences to result in development of language?
An example: Stackhouse & Wells (1997)
Input
• 'Physical sound wave' - a sound wave, whether speech or
non-speech, occurs in the environment.
• 'Peripheral auditory processing' - the ear notices that a sound
has been heard.
• 'Speech/non-speech discrimination' - the sound heard is
classified as being either speech or a non speech sound.
• 'Phonological recognition' - speech sounds are classified as
being part of a known language. 'Like tuning a radio until you
reach a channel where you recognise the language.‘
• 'Phonetic discrimination' - unusual speech sounds are
processed here. This is used when speech sounds differ from
the expected 'norm', for example, when processing different
accents and dialects.
Representation
• 'Phonological representation' - whole words
are stored according to how they sound
• 'Semantic representation' - the meanings of
words are stored here
• 'Motor program' - the motor instructions
required for speech muscles to produce the
necessary sounds for words
Output
• 'Motor programming' - allows the production
of words not previously known: copying a
nonsense word such as 'short'; this enables
the learning of new words.
• 'Motor planning' - allows for factors about
how a word will be said, for example, quickly,
loudly or with specific intonation.
• 'Motor execution' - the speech organs are
activated and a word is articulated.
Questions
• How do children develop such a system?
• How much of this framework do they bring to
the language acquisition situation? (i.e., how
much is innate?)
• How much do they have to learn?
• HOW do they learn this?
Principles and parameters
• Universal Grammar: The innately specified
principles and properties that pertain to the
grammars of all human languages
• Linguistic theory: A hypothesis about
Universal Grammar
• Principles and Parameters: one such linguistic
theory
Principles and parameters
• Universal Grammar consists of
– Principles: accounting for the similarities between
languages
• E.g. Structure dependency, (extended) projection principle
– Parameters: accounting for variation between languages
Learnability
• The Principles and Parameters hypothesis can
account for
– the specific ways in which (the grammars of)
languages can differ and
– the speed with which children acquire their
language
• under this hypothesis, the child only has to choose
from among a narrowly restricted set of values in each
of a limited number of innate parameters
• constrains the hypothesis space
• learning a language is reduced to parameter setting and
lexical learning
Acquisition of negation
• Data from English, German & French
• Stage 1 (about 24 months): Neg + sentence
No the dollie sleep.
Nein ich putt mache.
No I kaputt make
‘I didn’t break it.’
Pas la poupee dormir.
Not the doll sleep.
Acquisition of negation
• Stage 2 (about 28 months)
• Constructions with negative marker but no
auxiliaries
The dollie no sleep.
Ich mache das nich. (adult negation)
I do
that not
La poupee dort pas. (adult negation)
the doll sleep not
Acquisition of negation
• Stage 3 (about 36 months)
• Negation with auxiliaries
I didn’t/can’t do it. (adult negation)
• Why does it take English children longer to
acquire adult-like negation than it takes
German or French children?
A negation parameter
• Either:
– Any verb can carry negation, OR
– Only auxiliary verbs can carry negation
•
•
•
•
French, German: Any verb OK
English: only Aux OK
Aux is acquired late (Brown: 29-50 months)
So in languages where Aux is required to carry
negation, adult-like negation forms will also be
acquired late
Question formation – adult grammar
Question formation – adult grammar
Question formation – adult grammar
Question formation – adult grammar
Question formation – adult grammar
Acquisition of questions
• Stage 1: sentence with external question
marker
Mommy eggnog?
Where milk go?
(Boy eat?)
(What boy eat?)
Acquisition of questions
• Stage 2: Subject-Aux inversion in yes/no
questions, but not in wh-questions
Does the kitty stand up?
Oh, did I caught it?
Where the other Joe will drive?
Why kitty can’t stand up?
(What the boy eat?)
(What the boy did eat?)
Acquisition of questions
• Stage 3: subject-auxiliary inversion in wh
questions
What did you doed?
What does whiskey taste like?
(What did the boy eat?)
Back to the parameter
• Only aux / any verb can carry negation
• Reformulate: distinction between auxiliaries and lexical
verbs in terms of their distributional properties
– English: yes
– French, German: no
• In French or German, the lexical verb can be inverted to
form a question
– Dort la poupee?
– *Sleeps the doll?
-- Schläft die Puppe?
– La poupee dort pas
– *the doll sleeps not
-- Die Puppe schläft nicht
• In French or German, the lexical verb can carry negation
• In English, these constructions will not look like the adult
equivalents until the auxiliary system is acquired
Parameters
• So that parameter captures a lot of cross-linguistic
variation, as well as some facts about language
acquisition
• The “verb movement parameter”
• Lexical verbs can either move, or they can’t
• If they can (French, German), adult-like negation and
question formation will be acquired earlier
• If they can’t (English), adult-like negation and
question formation will not be acquired until the
auxiliary system matures
Our questions:
• How do children develop such a system?
– Interaction between principles (universals) and parameters
(limited variation, determined by exposure to linguistic
environment)
• How much of this framework do they bring to the language
acquisition situation? (i.e., how much is innate?)
– The principles are innate; the parameters are present but unset;
possibly also some specific statistical learning procedures, an
organizational framework, structural properties of the lexicon
• How much do they have to learn?
– Which way parametric properties go; lexical and phonological
properties of the system
• HOW do they learn this?
– Exposure, statistical algorithms, specific (?) learning mechanisms,
domain general (?) learning mechanisms