Language & Development - Block 1

Uniquely human

Language in all its complexity is _________ _____

Intro facts about language

All human groups on earth have language

All languages are equally expressive

All languages are easily learned by children

A critical evolutionary step in human evolution

Even though it comes easily to most of us: People may lose this ability, not have fully developed it, or may not be able to learn it

Natural language

Communication system of arbitrary signs

Sign components

Spoken/written form, world reference, concept

Arbitrary

No transparent or intrinsic relation between the sign and its reference, not iconic

Motor way signs

These are NOT arbitrary

Communication system

Set of shared rules that establish a systematic mapping between words and the world

Rules are implicit and established by common use

9 Characteristics of a natural language (Hockett)

Semanticity, Arbitrariness, Discreteness, Displacement, Productivity, Traditional Transmission, Duality of Patterning, Prevarication, Reflexiveness

Semanticity

language conveys meaning or referring to something

Displacement

We can talk about things that aren't present or factual

Cultural Transmission

Languages are learned in a community of native speakers

Discreteness

Linguistic units (sounds, words) are perceived as distinct from one another rather than as a continuum

The sounds "R-OO-M" in room, distinct from those in LOOM

Productivity

New words can be created, new meanings can be generated, recursive rule*

Recursive rule

A rule for a sequence in which one or more previous terms are used to generate the next term. It can be re-applied to its output

Duality of patterning

A handful of meaningless sounds can be combined to produce thousands of meaningful words

Combination of units at several levels

Example: word + "-ness" = sadness, lightness, craziness

Word parts can combine into other words and combine into sentences (syntax)

Learnable structured patterns

Combinations follow language-specific patterns (rules) giving rise to:

No

Is Morse code a natural language distinct from spoken languages like English?

No (2)

Is music a natural language?

Comprehension

Mapping word sounds to stored meanings involves:

1. Recognizing sounds and words (cat, not pat or pal)

2. Accessing/retrieving the meaning of the word (cat = domestic animal)

3. Combining words into a message (establishing relationship between words, lion killed the cat vs lion was killed by the cat)

Damage to Broca's area

Mr. Landry doesn't know which animal died when stating “The leopard was killed by the lion.” He only understands nouns and verbs, not other words or endings.

Production

Fluent speech takes longer to appear in learning —> speaking is harder than comprehending

Speaking involves

1. Retrieving words from memory

2. Putting words together

3. Fine grained motor planning

Koko the gorilla

Example of displacement; Koko is a gorilla that has been taught sign language

Arbitrary

Primate alarm calls for snakes, eagles, and leopards in vervet monkeys are learned or inherited. Do not represent the sounds these predators make. Referential calls, not just a behavioural reflex

Chimpanzee Vicky

Animals do not have complex systems in the world but can they learn a human language?

This animal was raised in a family alongside a child and with intensive speech therapy, she learned to say 4 words.

Human vocal anatomy

Longer oral cavity and much lower larynx, compared to orangutan and chimpanzee

Whashoe (chimp) and Koko (gorilla)

These animals went through intensive practice and training with modified ASL. But not very clear that signs were spontaneous because the researchers are showing her the sign beforehand.

Lana (bonobo)

Another animal taught with visual signs. She pushes keys to get things

Kanzi the Chimp/bonobo

Spontaneously learned looking at his mother

Best example of apes' cognitive abilities in optimal conditions (make tools and plan complex actions; knows hundreds of words; understands sentences, word order, and complex phrases like a 2 year old, suggest something a lot closer to human languages)

Productivity in expressing thought

What feature of human languages is absent in Kanzi?

Reference, productivity in expressing thought, analyzing sentences, or use of arbitrary signs

Non-primates

They can tell us how much communication systems evolved independently of human lineage.

Dolphins

In the wild: sociable, learning from parent to child, signature clicks, creative behavior, cross-species cooperation

In captivity: capable of high cognitive abilities, can categorize objects, absence, and presence of objects, complex concepts like under and thru, artificial language with word order variation

Summary of animal studies

28%

Mammals are born with 90% of brain mass, whereas humans are both with:

This allows for opportunity to learn and wire connections

Social-communicative urge

Young babies naturally cooperate with others, whereas chimps do not 1st other factor operating thru evolution

Joint attention and understanding of others communicative intentions

We have an innate cognitive capacity for these: although dogs can do this, they lack other necessary characteristics they can't do elbow cross pointing though (3 year olds can)

They are domesticated

Why can dogs understand pointing?

Do animals have communication systems similar to human languages?

If you look at actual properties of human languages, other species appear limited in

• The extent they communicate in nature
• The extent to which they can learn human languages

Innate modular view

• Language is an instinct

• Innate knowledge of universal language principles

• Innate language organ guiding acquisition: a separate brain system, independent of other cognitive abilities

• Based on poverty of stimulus argument

poverty of the stimulus argument (Chomsky)

the argument that the linguistic environment to which a child is exposed is not good enough to enable language acquisition on its own

Take a recursive rule:

• Children cannot learn this rule from input: They do not encounter sufficient evidence to learn these rules. They must analyze the sentence structure, and this is not taught

• They could not learn this rule by stimulus-response conditioning

Conclusion: universal structural principles are innate

Experience-based interactive view

• Linguistic and non-linguistic input is rich and structured (Child must map one to the other and social interaction gives you feedback)

• Children cannot learn extract rules/patterns over time (Sentence structure can result from learning: sensitivity to regularities in the input)

• There is a general innate capacity to learn from experience, but not language specific principles (This innate capacity may include instinct to communicate and innate computation power — domain general, not language specific)

Innate vs experience

photo

Nativists

If language is modular, we should find dissociations between language and other cognitive abilities (e.g., visuo-spatial, IQ)

Williams Syndrome

a disorder characterized by impairments of spatial cognition and IQ but superior linguistic abilities Nativist claim: evidence of intact language vs impaired cognition (developmental delay)

Specific language impairments

Impairments in distinguishing sounds, sentence understanding, producing well-formed sentences

Nativist claim: typical cognition vs impaired language at several levels

Criticisms of nativist claims - Williams syndrome

• Language rules are not intact but impaired, delayed language development
• Genes can have cascading effects in many cognitive effects

Criticisms of nativist claims - Specific language impairments

• They are not language specific: working memory and sound processing deficits may be the problem (Because small words like by, a, was, and endings like -ed are not phonologically stressed, children have difficulties extracting structural patterns in the input)

Can humans invent a language?

Nativists: if there are innate language principles, children should invent a language with little input

Feral children and cases too varied to make conclusion

What happens to children deprived of linguistic input?

Nicaraguan Sign Language (NSL)

Evidence for sensitive period!!! Sign language developed by deaf children in Nicaragua Two generations of signers were observed: First group of older students arrive with their home signs (less consistently used, oriented to global meaning without sub units) Younger signers learnt this rudimentary system, but: conventionalized and regularized the rules along with adding subtler distinctions to express more info more efficiently

Over generalization error

Mouses instead of mice

Sensitive period: nature vs nurture

Nativist:

• Young signers create a regular system with little input due to innate principles

• Maturation of language system is innately determined

Non-nativist:

• Sensitive period is due to initial brain plasticity

• Sensitive period is not language specific but found in other domains

• The brain extracts generalizations from patchy input for all types of input

• Learner needs the right kind of input

What appears to be innate?

• Instinct or social pressure to communicate.
• Developmental constraints: neural plasticity and brain growth along with experience.
• The ability to extract patterns and generalise from experience.

Sentential markers

English: train arrive-d (tense) Turkish: train arrive-di (evidence type)

Vocabulary differences

color terms (light and dark vs many other terms)

motion verbs (manner verbs vs path verbs)

Manner verbs tell how the action happens (the gait, style, or physical characteristic of the movement).

• Examples: run, swim, stumble.

Path verbs tell where the object goes (the trajectory or direction of the movement).

• Examples: enter, descend, cross.

Can language-specific characteristics shape our mental representations?

Universal (innate, modular view)

• No. All humans have similar representations
• Languages differ in the way they map the same thoughts into words

Linguistic relativity (learning, interactive view)

• Maybe. Language may influence our mental representations at least in some circumstances

Summary for this unit

Munsell color system

a color naming system based on hue, saturation, and value

Berinmo Language

Language with five basic color words.

English speakers were better for blue/green, but these speakers were better for nol/wor

Color names help us discriminate colors in memory

What does this study show?

note: but can we infer from this that color names typically alter our ability to discriminate colors and whether this effect generalizes to other situations

Winawer et al. (2007)

Evidence for Linguistic Relativity: Colour Perception

• In general, speakers are faster between colour categories than within

• Ie. if ppts are presented with 3 patches, they are faster to distinguish between two green and one red, than two dark green and one light green

• Russian speakers: have two different categories for blue: light blue (goluboy) and dark blue (seigney)

• English speakers: have one category for blue

• So Russian speakers should be faster than English speakers when one square is from the light blue category and two are from the dark blue category

• For Russian speakers, they cross a colour category, for English speakers, they don't

• Task: carried out under different levels of itnerface

  • 1) Task without any interference

  • 2) Rehearsing Digits: intended to inhabit language - keeping the phonological loop full so ppts not able to engage language as much

  • 3) Maintaining a visual pattern: takes up processing but not specifically language (control)

Results For the no interference condition:

• Russian speakers significantly faster when the colours were across categories than within colour categories

• English speakers: showed no significant difference between the two, were just as slow whether it was light blue/dark blue or two shades closer together

For the spatial interference condition

• Same pattern

For the Verbal Interference Condition

• The effect disappears, Russian speakers equally slow in both conditions

Conclusions

• Language CAN affect perception, the way people see the world seems to be different in different languages

Gilbert et al. (2006)

• Manipulating which brain hemisphere receives visual info first

• Left hemisphere is where words are stored

• Task: identify odd one out fast

• Results: language effect (English vs Russian) only when LH is first

Conclusions from Gilbert et al. (2006)

• Color words do not permanently alter perception

• We see an influence of language on some perception tasks

  • When color words are quickly activated in the left hemisphere

  • When people can rely on words to perform a memory or discrimination task

• Consistent with linguistic relativity or the Whorfian hypothesis!

Two types of motion verbs

photo

Gennari et al. (2002)

Study showing similar memory across languages.

Papafragou et al (2008) and Gennari results

What do these motion studies tell us?

Attention allocation depends on the task: observation vs description

• Describing elicits language-specific pattern of looks

• Simple observation does NOT seem to involve linguistic meaning, unlike color names

• But language might be used to memorize events

Conclusions

• Language may influence some cognitive representations depending on task and stimulus types (colors vs events) (e.g., language may help memory or perceptual discriminations)

• Evidence consistent with language relativity view

• Habitual or learned words-to-world associations can be useful if this makes task performance more efficient

• Against universal view

Review

All tutorial notes and studies