Kim wood 4

Language Development Chapter Nine

Is Acquisition Innate?

• Biological Basis for language? – It’s unique: only humans have it

– It’s ubiquitous: normally, all humans have it

– Kids can learn any language

– Learners clean up noisy input

– It has a critical period (suggesting maturation)

– Kids are proficient by age 3-4 without formal instruction across cultures

– Not correlated with technology • US English not more complicated that Old English

So, What is Innate?

• Words aren’t, they still need to be learned

• Syntax is: Universal Grammar (UG)

– Humans born with innate knowledge of all possible grammars

– Acquisition is figuring out which grammar to use

– Critical period: Adults lose access to UG

Poverty of the Stimulus Chomsky: “To come to know a human language would be an

extraordinary intellectual achievement for a creature not specifically designed to accomplish this task” because: (i) “A human language is a system of remarkable complexity;” and (ii) ”A normal child acquires this knowledge on relatively slight exposure and without specific training.”

• Input underdetermines grammar – Finite input but needs to generalize to infinite set

– Infinite number of languages consistent with input

• Degenerate input: – both grammatical and ungrammatical utterances

• Little direct negative evidence

McNeill (1966)

Child: Nobody don’t like me.

Mother: No, say “Nobody likes me.”

Child: Nobody don’t like me.

Mother: No, say “Nobody likes me.”

Child: Nobody don’t like me.

Mother: No, say “Nobody likes me.”

Child: Nobody don’t like me.

[dialogue repeated five more times]

Mother: Now listen carefully, say “Nobody likes me.”

Child: Oh! Nobody don’t likes me.

Braine (1971) Child: Want other one spoon, daddy. Father: You mean, you want the other spoon. Child: Yes, I want other one spoon, please Daddy. Father: Can you say ‘the other spoon’? Child: Other…one…spoon Father: Say ‘other’ Child: Other Father: ‘Spoon’ Child: Spoon Father: ‘Other spoon’ Child: Other…spoon. Now give me other one spoon

“I mean I, I mean I dunno I mean I just, I just ha=, I just had to tell you because I mean it really is pissing me off the way f**king Angela treats me”

How Grammatical is Actual Speech?

Variability & Complexity of Input

• Different voices have different characteristics: e.g., pitches, tones, accents, ways of phrasing things

• Kids have to abstract across incredible variability to get even the phonology right

• Have to get phonology right before learning words, which you need to figure syntax out

Logical Problem of Acquisition

• Gold (1967): Not certain that particular class of grammars could be inferred from only positive examples

– Mistaken inference: Language can’t be learned

• Logical argument: If we know X, and X can’t be learned, then X has to be innate

Crain & Thornton

• Structure dependence – Language reflects structure, not just word order

• Error avoidance

– Pronouns

– Wanna contraction

– Negative polarity items (e.g, ever, any)

• Downward entailment

• C-command

Structure Dependence

• Grammars deal with abstract structural categories and units, not concrete positional or statistical information – E.g., grammatical rules are based on notions like

“auxiliary verb” and “noun phrase” rather than “next word” or “words with more than 3 syllables”

• This information isn’t marked on the input, and would thus be difficult to learn

Structure Dependence

Question Formation

Declarative Interrogative (a) Susan must leave. Must Susan leave?

(b) Harry can swim. Can Harry swim?

(c) Mary has read the book. Has Mary read the book?

Statistics: just move the first verb to the front of the sentence

The lawyer who is suing me is mean

*Is the lawyer who suing me is mean

Gonna Contraction

• I’m going to dance -> I’m gonna dance

• I’m going to the dance -> *I’m gonna the dance

• I’m going two directions-> *I’m gonna directions

• That’s where I’m going too -> *That’s where I’m gonna

Parameter Setting

• Allows for cross-linguistic variation in UG

• Acquisition is setting about 10-60+ binary parameters to get appropriate grammar

• Default settings, overwritten by contradictory experience

• 30 parameters gives us 2 30

= 1,073,741,824 different possible grammars

• If we checked 1/sec from birth it could take 34 years to get it right (Clark, 1990) (of course, we don’t have to check all possible grammars)

Examples of Parameters

• Null Subject (pro-drop)

– English: no; Spanish: yes

• Headedness (e.g., verb-compliment order)

– Head initial: English “the boy [kicked the ball]”

– Head final: Japanese “the boy [the ball kicked]”

“The setting of the head parameter should be obvious in the primary data. Does the head come before or after the complement”

Pro-Drop in English?

Robert: Going to the store now. Need anything?

Andrea: Nope. Have fun.

Robert: Will do, thank you.

Andrea: Love you.

Robert: Love you too.

Andrea: Oh, don’t forget the coupons.

Robert: Already got them. See you later.

Andrea: Yeah, see you later.

Do Parameters Really Help? “The setting of the head parameter should be obvious in

the primary data. Does the head come before or after the complement?”

• Core aspects of syntax aren’t directly marked in the input

• So kids need to identify abstract linguistic patterns in order to set parameters

• If they already know the heads, complements and have determined their order, why do they need a parameter?

Biological Evolution of UG?

• Many people say that language is literally an organ or instinct

– Chomsky once argued that having language is like growing an arm or having a liver; “The Language Organ” by Anderson & Lightfoot (2002)

– Piatelli-Palmarini draws analogies to immune system

– Pinker suggests human story telling is biologically equivalent to spider’s web spinning ability

Biological Evolution of UG?

• Natural selection of a dedicated language “organ” in our brains is implausible

– We branched off from other primates maybe 2-6 million years ago

– At most, humans have probably only had language for about 100k years

– Conversely, many vertebrate organs have evolved over 500 million years, cerebral cortex over 200-300 million years, and the visual system over 100 million years

Biological Evolution of UG?

• If language is an organ, shouldn’t we be able to see it? Imaging data suggest:

– No specific language area, widely distributed activation

– Areas also active during non-linguistic processing

– Same brain waves elicited by syntactic and musical violations

• Not to say that languages haven’t evolved (NSL evolved very quickly), but UG as a independent biological organ questionable

A Different Solution

• Language evolves through cultural transmission (Tomasello, 1999)

• Goal isn’t grammar identification, but communication (Seidenberg, 1997)

• Kids start out with concrete utterances, build up to grammatical generalization (Tomasello)

– Learnability not a problem because not learning 1 of 1,000,000,000 grammars, just what you hear

• But why such strong innate biases to use language? Strong cross-linguistic universals?

Bates and MacWhinney (1988)

• “The human capacity for language could be both innate and species-specific, and yet involve no mechanisms that evolved specifically and uniquely for language itself. Language could be a new machine constructed entirely out of old parts. The universal properties of grammar may be indirectly innate, based on interactions among innate categories and processes that are not specific to language.”

What’s Wrong with Emergentism

• Doesn’t explain why all kids pick up language relatively effortlessly, whereas chimps don’t; Chimps have similar social cognition to humans

• Relies heavily on abstraction to get from kid grammars to adult ones, but how does “abstraction” work?

Sign Language

• Used primarily by hearing-impaired people

• Uses a different medium: hands, face, and eyes (rather than vocal tract or ears)

• NOT derived from spoken language

– ASL is a completely different language than English, much closer to French SL

Sign Language

• Maybe around 300 signed languages around the world, all are different

• Within a SL, there are different dialects, people have accents

• Has all of the grammatical properties of spoken languages (e.g., pronouns, verb tenses, grammatical structure)

Sign Language

Structure of Signed Languages

• Five Basic Parameters – Shape of the hand

– Place of articulation (location)

– Movement

– Palm Orientation • Region of the hand contacting body

• Orientation of the hand to body

• Orientation of hands to each other

– Facial Expressions

Unique Challenges

• 90% of congenitally deaf children have hearing parents

• Thus, many deaf children aren’t exposed to any language (signed or spoken) until they go to school

• Provides naturally occurring data to study the role of language deprivation and age of exposure on language development

Newport & Supalla (1987)

• Study of deaf children of hearing parents, only exposed to language at school

– Exposure before 4 = native proficiency

– Exposure from 4-6 = errors, but near-native

– Exposure after 6 = more errors in closed-class items, more frozen forms

– Exposure after 12 = much higher error rate and variable error types

Singleton & Newport (2004): Simon

Input: • Parents were late learners of ASL

• Used ASL verb inflections only 60% of the time

• In school, only exposed to a signed English system

Output: • As good as “native of native” children on most aspects of

ASL inflection

• Simon significantly surpasses the performance of his parents

• Simon does not acquire the “noise” of his parents; he regularizes the irregular input from his parents.

• Goldin-Meadow & Mylander (1998)

– Spontaneous sign systems in deaf American and Chinese children of hearing moms

– Found that US kids gestures didn’t resemble their mothers, but did resemble Chinese kids gestures

Nicaraguan Sign Language

• Until 70’s Nicaragua had no schools for the deaf and no signed language

• ‘77- ‘83: oralist school created

• By ’86 signed pidgin emerges

• New children added to community at variety of ages

Photos: Judy Kegl

Nicaraguan Sign Language

• Where does the structure of the language come from?

• Little contact with other sign systems

• Do younger children create the language or older more experienced kids?

• Age matters: younger learners are more fluent

• Input matters: those who heard less mature language (cohort 1) are less fluent

Senghas & Coppola (2001)

Senghas & Coppola (2001), p. 327

• “…the second cohort did not reproduce the language as it was produced by their first-cohort elders; rather, they changed the language as they learned it.”

• Arguably, all language works this way, or we’d never see diachronic change

• Along with Singleton & Newport, indicates that learners clean up noisy input, apply structure

Al-Sayyid Bedouin Sing Language

Sandler et al. (2005)

• Spontaneous sing language used by the Al- Sayyid Bedouin in Israel

• 4% of 3500 are congenitally deaf vs. 0.1% here

• ABSL is about 70 years old and on 3rd generation, typologically different from surrounding languages (e.g., Arabic, Hebrew, ISL)

General Conclusions

• Children can acquire language in wide range of environments

• When the input is chaotic, they structure it • When language is absent they create it

Why?

• Humans have a ‘linguistic nature’ that rescues learners from inadequacies in environment

• Language is a result of the developing human brain • Language acquisition appears to be supported by

powerful innate biases and mechanisms - Still a lot of theoretical debate about their nature