Ling 240

Review for Exam 1

Questions:1) What is it that we know when we

know a language? 2) How do children acquire that

knowledge?

Linguistics as Cognitive Science

mental grammar

Mental Grammar:

the system of rules stored in the mind of a speaker that generates the words and sentences of that speaker’s language.

What do you know about your language?

Example 2b

cat catsdog dogstick tickspole polescar cars

[s][z][s][z][z]

Arguments for Mental Grammar

Mental grammar vs.

memorizing or storing the list of possible sentences

Argument for Mental Grammar #1

• The number of sentences we are capable of using is too large to store them all

Argument for Mental Grammar#2

BUT, it’s not just that the possible number of sentences is enormous

It’s infinite.

Because we can do this:

This is the house [that Jackie built].This is the door [that leads to the

house [that Jackie built]].This is the mold [that grew on the

cheese [that fell out of the fridge [that stood in the house [that Jackie built]]]].

Recursion

We can produce (and understand) potentially infinitely long sentences.

Therefore, there are an infinite number of sentences.

Argument for Mental Grammar #2

Given that our brains are finite, there is no way we can store all of the possible sentences of our language

Is there anything that we do store (memorize)?

Lexicon

Lexicon = internal dictionary

WordsPronunciationMeaning

Mental Grammar

LEXICON RULES

The linguist’s goal

Build a model of the speaker’s knowledge of language

Methodology: Scientific Method

• Examine data• Make generalizations (describe

observable patterns)• Make hypotheses to explain observed

generalizations• Test predictions of hypotheses against

more data

Pronunciation of the plural

[s] [z]cats dogsticks polestrips cars

tabstoesdoves

If it’s a rule, what’s the rule?

Pronunciation of the plural

[……….. X]N

If X is voiceless, then plural => [s]If X is voiced, then plural => [z]

Hypothesis A: Complex forms stored in memory

LexiconCat [kæt]Cats [kæts]DogDogs

PREDICTIONS?

Hypothesis B: complex forms derived by rule

LexiconCat [kæt]Dog

Rules

N + PL = Npl

PL → [s] if […-vc]N→ [z] if […+vc]N

PREDICTIONS?

Evidence for Rules

• Novel forms

SmickPlural: Smicks [s]

“s” after [r]

Doors [z]Horse [s] NOT [z]

Rule only applies to the plural “s”

Conclusion: mental rule

Mental rule or physical necessity?

Also, evidence from other languages

English: toes

Spanish: libros

Also, evidence from other languages

English: toes [toz]

Spanish: libros [libros]

The linguist’s job

Build a model of the speaker’s knowledge of language.

Discover what the rules are

But what counts as “grammatical”?

Grammaticality = meaningfulness?

One often-heard statement:A sentence is grammatical if it makes sense. If we can understand the meaning, then the sentence is grammatical.

• Remember: The job of the linguist is to determine what the rules are that allow people to speak and understand novel sentences in their native language

• Study of the human mind

• Linguists study how people really speak (describe patterns that exist in the data)

• They do not try to tell people how to speak

From this point on, grammatical = native speaker would say

innateness

How do children learn language?

• Parents teach them?

• Imitation? (i.e., copy & memorize)

How do children learn language?

Child: Nobody don’t like me.Mother: No, say “nobody likes me.”Child: Nobody don’t like me.

x 8Mother: Now listen carefully; say,

“nobody likes me.”Child: Oh! Nobody don’t likes me.

a) manu-fuckin-facturer (manufacturer)

b) cali-fuckin-fornia (California)

c) *Chom-funckin-sky (Chomsky)

d) *ele-fuckin-phant (elephant)

cf. http://www.cafepress.com/tvtee.41728115

How do children learn language?

How do children learn language?

• Not taught• Not conscious

A Paradox

Language is unbelievably hard to figure out– Engineers and computer scientists trying to

solve problems of machine translation and AI for 50 years

– Linguists studying grammatical principles underlying language

Innateness Hypothesis

The human brain contains a genetically determined specialization for language.

Language learning = natural ability

Part of human biology

• Just like the ability to walk

• Just like using sonar to echolocate is part of bat biology

Innateness Hypothesis

Mental Grammar

Innate part Learned part

Arguments for innateness

• Uniformity throughout the species• Species-specificity• Poverty of the stimulus: children create

grammar beyond what they have evidence for

• Critical period

What could a baby possibly have in her brain that could help her acquire any language that she might end up being exposed to?

=> Universal Grammar (UG)

Language Universals

How are all languages alike?What is the range of possible differences?

This is the research program

phonetics

Chart

Review questions…

• What are the features of the following sounds:

1. [t]2. [∂]3. [ß]4. [dΩ]5. [˜]6. [h]7. [w]8. [®]

iɪ

eɛ

uʊ

oɔ

a

ʌə

Front Central Back

High

Mid

Low

round

tense æ

Diphthongs: two-part vowels (cf. monophthongs)

1. [åI] bite, aisle, choir, island

2. [å¨] brown, doubt, loud, hour

3. [øI] boy, rejoice, annoy, poison

4. [o¨] boat, grow, though, over

5. [eI] bait, reign, great, they, gauge

phonology

Is the pronunciation predictable?

[th] [‰] [÷]

timetoptable

waterlittlebutternotable

buttonkittencarton

(cf. notation)

Patterns

[th] occurs at the beginning of stressed syllables.

[‰] occurs between two vowels

[÷] occurs in the middle of words before a syllabic nasal.

Articulatory Processes

• Sounds are pronounced differently in different environments

These are RULES in mental grammar

/t/

[th]

[t]

[‰]

[÷]

In what sense are these all the same “sound”??

Lexicon

/wåt\r/

Rules

/t/ -> [‰] / V_ V

Output

[wå‰\r]

Phonemes Allophones

Phoneme

• the abstract representation of a sound. • the way the sound is stored in word in

the mental lexicon

Allophone

• how the sound is actually produced in a given environment

• an instance of a phoneme• a member of a more general category

Inventory of Phonemes

First step: look for minimal pairs• if difference between sounds causes a

difference in meaning => contrastive=> different phonemes

Second step:If no minimal pairs, look for a pattern(distribution of the two sounds)

– What environments does sound X occur in?

– What environments does sound Y occur in?

Inventory of Phonemes

steps in finding out a generalization

• List the environments for the target phones

• Generalization (e.g., x occurs in environment Y) and distribution (x and z are in complementary/contrastive distribution)

• Formulate a rule that derives the pattern

Exercise: making generalizations about the environments

can /k(æ)n/I can ask [åI kn æsk]I can see [åI kn si]

I can bake [åI km beIk]I can play [åI km pleI]

I can go [åI k˜ go¨]I can gather [åI k˜ gæ∂\®]

Generalizations

[m] occurs before a bilabial consonant

[˜] occurs before a velar consonant

[n] occurs everywhere else (elsewhere)

Generalizations => Rules

/n/ becomes [m] before a bilabial consonant

/n/ [m] / ___ (bi)labial consonant

/n/ becomes [˜] before a velar consonant

/n/ [˜] / ___ velar consonant

elsewhere /n/ is pronounced [n]

/n/ [n] / elsewhere

Perceiving VOT

‘Categorical Perception’

Categorical perception

• The mind imposes discrete, abstract categories which do not exist in the physical world

• Gradations in the phonetic categories are ignored

Is phoneme discrimination innate?

• Hypothesis A: Innate ability to discriminate all possible phoneme distinctions (in any language)

• Hypothesis B: children learn to discriminate the contrasts in their language

• Hypothesis C: Something in between…some contrasts are innate, others learned

High Amplitude Sucking Procedure

• Initially sucking rate increases (novelty)

• Then decreases• Decline in response

= habituation

LSCP Infant Lab

Kuhl & Miller, 1978

010

2030

4050

6070

8090

100

0 10 20 30 40 50 60 70 80

Voice Onset Time (ms)

Perc

ent l

abel

led

[b]

English SpeakersChinchillas

Universal Phonetic Space

Universal phonetic inventory

French

EnglishJapanese

Testing Across the Lifespan

• HAS works only with infants from birth to 4 months

• Adults & children can be asked to detect a change

• The Conditioned Head Turn Paradigm can be used for 6-12 month-old infants

Results

Werker, 1995

First proposal: Maintenance/loss Hypothesis

• role of experience is to maintain perceptual sensitivities

• lack of exposure leads to loss of perceptual ability

Alternative Hypothesis: Functional Reorganization

• no absolute hardware changes in auditory system

• development of a linguistic system that imports a subset of the contrasts from the auditory system