Methods of Modern Psycholinguistics

Methods of Modern Psycholinguistics

I. Behavioral methods

A. Latency (response time) measures

for example --

lexical decision (decide if a letter string is a word or not)

B. Accuracy measures

for example –

grammaticality judgments

C. Memory measures

for example --

test for the “memory” of an inferred entity in a discourse

D. Eye tracking

II. Studies of neurologically impaired populations

A. Brain damage (resulting from a cerebrovascular accident (CVA) or trauma)

1. CVAs

thrombosis – occlusion of blood from blood clot

embolism – occlusion of blood flow from junk floating in arteries

aneurysm – weakness in blood vessel wall

hemorrhage – blood seeps into cortical tissues (often caused by a ruptured aneurysm)

------

hematoma – mass of blood encapsulated by compressed cerebral tissue

2. Trauma

gun shot wounds, accidents, etc

3. Types of language-specific deficits that can result

aphasia – an acquired deficit in speech

alexia – an acquired deficit in reading

agraphia – an acquired deficit in writing

B. Other syndromes

e.g., Williams syndrome, dyslexia, etc.

III. Invasive brain-based measures

A. Cortical stimulation during neurosurgery

directly stimulate neurons in brain using

a microelectrode

B. Electrocorticogram prior to neurosurgery

place electrodes on top of cortex (but

underneath the scalp)

IV. Non-invasive brain-based measures

1. PET (Positron Emission Tomography)

How it works:

Formation of image depends on distribution of inhaled/injected isotopes that are bound to water, glucose, or neurotransmitters.

Rationale:

The brain uses oxygen (in water) and glucose as fuel. The most active brain region take up more oxygen and glucose, leaving these areas with the most radioactivity.

Resolving capacity:

spatial resolution – 5 to 10 mm (good!)

temporal resolution – several minutes (bad!)

2. Magnetic resonance imaging (MRI)

structural MRI

functional MRI (fMRI)

How it works:

Formation of image depends on radio frequencies emitted by the perturbation of hydrogen atoms from their “orbits.” Perturbations are induced by a strong magnetic field.

Rationale:

Thick blood vessels have more hydrogen than thin blood vessels (this provides a meaure of cerebral blood flow, or CBF). Blood flow to a brain region is a function of the activity of that region.

Resolving capacity:

spatial resolution – 1 to 10 mm (good!)

temporal resolution – second (pretty bad for language!)

3. ERPs (Event-related brain potentials)

How it works:

Place electrode on scalp, measure changes in electrical voltage over time. Measures activity of large groups of synchronized neurons in brain.

Rationale:

None needed – you are directly measuring the brain’s electrical activity

Resolving capacity:

spatial – 1 cm at best (bad!)

temporal – 1 millisecond (outstanding!)

Language Acquisition

Amazing empirical fact to account for:

By the age of 3, EVERY neurologically intact child has mostly mastered his/her parents’ language.

The rate of acquisition is constant across cultures, socio-economic background, and intelligence levels

Explanations of this empirical fact:

1950’s – language is learned through imitation

1960’s – language is innate (language acquisition device, or universal grammar)

More recently -- a recognition that learning, innate elements, and social factors all pay important roles in language acquisition

I.Language learnability

Question: To what extent can language acquisition be attributed to learning alone?

B.F. Skinner believed that language was acquired by general learning mechanisms such as conditioning and reinforcement.

Learnability theory:

Is it logically possible for a child to acquire a language through learning alone, by the age of 3?

To answer this question, we need to know exactly what sorts of INPUT the child has that might help him/her learn a language.

II. Descriptive evidence: Stages of acquisition as reflected in children’s speech production

A. Pre-babbling

1. 0 – 6 weeks “vegetative sounds” (cry, burp, sucking sounds)

2. 6 weeks: cooing

3. 16 weeks: laughing

B. Babbling (5-10 months)

1. nonsyllabic babbling (5-7 months)

Baby begins to play with sounds (clicks,

humns, hisses, smacks)

2. syllabic babbling (7-8 months)

Baby begins to produce real syllables

“bababab” “deedeedee” “neh neh neh”

3. gibberish babbling (8-12 months)

Baby begins to mix syllables

“da-dee” (probably doesn’t mean “Daddy”)

“new-nee”

III. Experimental evidence

A. Sounds (phonology)

“Prosody” – pitch (intonation) and

rhythm of language

1. Mehler (1980s):

Question: Can 4-day-old babies discriminate between their mother’s language and other languages?

Learning about words and meaning

A. Acquisition of vocabulary

Most of us know ~80,000

Assuming you are 20 years old, that’s an average of 4 word a day, each day of your life

B. Learning about word meanings

Aitchison (1987): 3 stages of learning

1. Labeling – the realization that a particular combination of sounds “means” or “refers to” something specific

6-12 months – “words” but not a lot of labeling

12-24 months – child “gets it” and labels like crazy

2. Packaging – understanding a word’s “scope of reference”

WordPossible Reference

DOGall animals

all animals with 4 legs

all animals that are house pets

all animals that have black and

white spots

all animals that have a ferocious bark

all animals that smell bad

Child’s wordFirst referentunderextension

Underextensions

“doggie”poodleonly poodles are

dogs

“whitesnowonly snow is white

Overextensions

“mooi”mooncakes, letter 0

“fly”flyspecks if dirt,

small insects,

Theoretical accounts of what is learned, and of extension errors:

a. Clark: semantic feature hypothesis

- the meaning of word is comprised of a set of “semantic features”

- when there is a mismatch in the features the child assumes and the adult assumes, extension errors occur

overextensions occur when a feature set is

incomplete; underextensions occur when

the child has to many features

- semantic development involves getting the feature sets just right

b. Bowerman: prototype hypothesis

- a prototype is an average member of a category

- lexical development is a process in which the child develops prototypes that are adult-like

3. Networking – the development of a “semantic network” relating to word knowledge

For example, the child must learn subordinate/ superordinate links that are implicit in categorization

Second-language (L2) learning and bilingualism

I. Preliminaries

Definitions

simultaneous bilingualism: child learns L1 and L2 at the same time

sequential bilingualism: person learns L1 first, L2 at some delay

Types of learning environments

traditional classroom: direct translations of L1 to L2, explicit instruction in grammar; L1 is used a lot

immersion: learners are taught exclusively using L2

submersion: you are living among native speaker of the L2

II. Does bilingualism help you or hurt you, or neither?

A. simultaneous learning in childhood

Bilinguals might have some minor disadvantages on some tasks (for example, it takes them longer to decide if a letter string is a word or not, for either language); these deficits are quite limited.

B. later in life

bilingualism might inhibit “cognitive aging effects” in middle-aged and older people

Bialystok et al (2004): effects of bilingualism on performance on the “Simon Task,” which measures a person’s ability to focus their attention on some task (or, “distractibility”)

distractibility increases with age (performance goes down on the Simon task as a person ages)

II. Effects of age on language learning

Common belief: The older you are, the more difficult it is.

“There is no doubt that children learn language more naturally and efficiently

than adults.” (Kuhl, 2000)

In fact, there is some truth to this:

Success at ultimate attainment of a language seems to decline (perhaps linearly) with age.

A. The questions

1. What aspects of the language become harder to learn as you get older?

phonology: often very hard to master as an adult

words/meaning: uncertain if there are age effects for word learning (rarely studied), although recent evidence suggests there might be

syntax/grammar: it’s often claimed that it is very difficult to acquire L2 syntactic rules that are different from the L1

2. Why is age associated with a decline in L2 learning ability?

Standard theoretical explanations

a. Effects of maturation (“critical period” for language learning)

Lenneberg (1967): there is a biologically defined “critical period” for learning language; once this “window of opportunity” ends, it becomes impossible to attain native-like proficiency in a language

When does the window close? Puberty.

b. Effects of experience with a first language (L1)

Perhaps experience with a first language interferes with learning a second. Language learning is harder as you age, because you have increasingly more experience with your L1.

Kuhl, 2000: speech sounds

Ellis & Lambon Ralph, 2000: words

Note: Both explanations implicate a reduction in neural plasticity.

However, there is no direct evidence that your brain becomes dramatically “less plastic” as you age; or, more specifically, that your brain responds less quickly to a new language when you are old than when you are young.

Reading

I. Learning to read (for an alphabetic script like English)

phonological awareness – awareness of how sound maps onto graphemes

very important for beginning readers (phonic method works a lot better than whole-word method of teaching reading)

During normal development, awareness seems to begin with small units (individual letter-sound mappings, e.g., 1 letter = 1 phoneme), and then progresses to larger units (e.g., syllables)

Phoneme-letter awareness seems to best predict subsequent reading skill.

SO: Right from the start, sound and writing go hand in hand.

II. “Dual route” model of normal reading

Graphemes

directphonological

“visual”recoding

route

Word recognition/

lexical access

Definitions:

grapheme-to-phoneme (spelling-to-sound) “rules”

regular words – those that correspond to the most common patterns

(e.g., steam, hint)

irregular words – those that do not correspond to the most common patterns

(e.g., steak, pint, yacht, aisle)

For normal reading, there seems to be a “race” between the two routes, which both occur; the speediest “wins”

III. Alexia (in the U.K., acquired dyslexia)

A. Syndromes

1. “peripheral” alexias (affect visual processing system)

a. spelling alexia

takes forever to read a word; each letter is processed individually, one at a time

2.”central” alexias (specifically affect reading system)

a. surface alexia (reading by sound)

selective impairment in ability to read irregular (exception) words (e.g., yacht, pint)

pint  spoken o that it rhymes with “hint”

steak  steek

yacht  yached

But, comprehension of these words is normal. mostly

b. phonological alexia (reading by sight)

selective impairment in ability to read pronounceable nonword (pseudowords)

can’t read aloud “sleeb”, although they can read “sleep”

c. deep dyslexia

semantic reading errors (paralexias)

Written word …Read aloud as …

daughtersister

roseflower

etc

IV. Dyslexia (in the U.K., developmental dyslexia)

very common (depending on how you define it, 5-10% of the population)

Relationship between dyslexia and alexia is not clear. (I think they are very distinct.)

Possible causes:

1. phonological processing deficit

a. lack of phonological awareness

b. poor ability to segment speech stream into phonemic units

2. visual processing deficit

a. problem processing closely spaced, highly detailed visual objects (like letters)

Treatments:

1. Try to increase phonological awareness.

2. Try to improve ability to segment speech stream into phonemes (e.g., Fast ForWord)