What is psycholinguistics?

The three primary processes investigated in psycholinguistics

•Language Comprehension

•Language Production

•Language Acquisition

 

Psycholinguistics is a branch of cognitive science

 

 

What will be covered in this class?

• How do we produce and recognize speech?

• How do we perceive words, letters, and sentences?

• How do we learn and recall information from texts?

• How can we improve texts to make them easier to understand?

• How does the brain function to process language?

• What are the causes and effects of reading disabilities?

• Is there language in other species?

Central themes in psycholinguistics

1) What knowledge of language is needed for us to use language?

Tacit (implicit) knowledge vs. Explicit knowledge

• tacit: knowledge of how to perform something, but not aware of full rules

• explicit: knowledge of the processes of mechanisms in performing that thing

 

 

2) What cognitive processes are involved in the ordinary use of language?

How do we understand a lecture, read a book, hold a conversation?

Cognitive processes: perception, memory, thinking, learning

 

Some definitions of basic components of language:

Semantics: The meaning of words and sentences

Syntax: The grammatical arrangement of words in a sentence or phrase

Phonology: The sound pattern of language

 

Pragmatics: How language is used in a social context

Examples from psycholinguistics

 

Parsing garden path sentences

The novice accepted the deal before he had a chance to check his finances, which put him in a state of conflict when he realized he had a straight flush.

 

 

 

 

 

 

 

1) The defendant examined by the lawyer turned out to be unreliable

 

 

2) The evidence examined by the lawyer turned out to be unreliable

 

 

The process of parsing is the process of making decisions

The effect of prior knowledge on comprehension

 

 

The procedure is actually quite simple. First you arrange things into different groups. Of course, one pile may be sufficient depending on how much there is to do. If you have to go somewhere else due to lack of facilities, that is the next step; otherwise you are pretty well set. It is important not to overdo things. That is, it is better to do too few things at once than too many. In the short run this may not seem important, but complications can easily arise. A mistake can be expensive as well. At first the whole procedure will seem complicated. Soon, however, it will become just another facet of life. It is difficult to foresee any end to the necessity for this task in the immediate future, but then one never can tell. After the procedure is completed, one arranges the materials into different groups again. Then they can be put into their appropriate places. Eventually they will be used once more, and the whole cycle will then have to be repeated. However, that is part of life.

 

 

 

 

 

 

 

Bransford & Johnson, 1973

Recall:

No context: 2.8 idea units out of a maximum of 18

Context afterwards: 2.7 idea units

Context before: 5.8 idea units

Child language development

How many words do you know?

Hint: Dictionary has about: 450,000 entries

 

 

Test high school graduates: How many words do they know?

About 45,000 english words

About 60,000 including names and foreign words

The average six year old knows about 13,000 words.

Learning about 10 words per day since age 1. (One every 90 minutes)

 

 

How much do we have to teach children to learn language?

Do you have to teach a child to walk?

Is it the same way of learning a language?

 

 

My teacher holded the baby rabbits and we patted them

I eated my dinner

A brief history of psycholinguistics

Wilhem Wundt (early 1900s)

Interest in mental processes of language production

• Sentence as the primary unit of language

• Speech production is the transformation of complete thought processes into sequentially organized speech segments.

 

 

Behaviorism (1920s-1950s)

• Rejected the focus on mental processes

• Measurement based on objective behavior (primarily in lab animals)

• How does experience (reward and punishment) shape behavior?

B. F. Skinner:

Children learn language through shaping (correction of speech errors)

Associative chain theory:

A sentence consists of a chain of associations between individual words in the sentence

What’s wrong with the behaviorist approach?

Noam Chomsky (1950s - present)

1) Colorless green ideas sleep furiously

2) Furiously sleep ideas green colorless.

 

3) George picked up the baby

4) George picked the baby up.

 

Almost every sentence uttered is a new combination of words

The Poverty of stimulus argument:

There is not enough information in the language samples given to children to account for the richnes and complexity of children’s language

The pattern of development is not based on parental speech but on innate language knowledge

 

Linguistic Diversity vs. Linguistic Universals

Linguistic diversity

There appears to be a lot of diversity among languages

Even within languages there is diversity

When are two languages different?

We speak the same language if we can understand each other

Exceptions: Norwegian and Swedish

Cantonese and Mandarin

 

Dialects within languages: The myth of pure language

How/why do languages change?

 

 

Why does there seem to be a "correct" English?

 

 

Members of the dominant (most powerful) sub-culture tend to speak one dialect and may punish those who do not

 

 

 

Linguistic Chauvinism

Belief that one’s own language/dialect is the best of all possible languages

Black English Vernacular (BEV)

Study by William Labov

Interviewed African-American street youth

You know, like some people say if you’re good an’ sh*t, your spirit goin’ t’heaven . . . ‘n if you bad, your spirit goin’ to hell. Well, bullsh*t! Your spirit goin’ to hell anyway, good or bad.

[Why?]

Why? I’ll tell you why. ‘Cause, you see, doesn’ nobody really know that it’s a God, y’know, ‘cause I mean I have seen black gods, white gods, all color gods, and don’t nobody know it’s really a God. An’ when they be sayin’ if you good, you goin’ t’heaven, tha’s bullsh*t, ‘cause you ain’t goin’ to no heaven, ‘cause it ain’t no heaven for you to go to.

 

• Place holders: "There" vs. "It" in the copula

• Copula: "Is", "Was" optional

• Negatives: "You ain’t goin’ to no heaven"

 

 

BEV just as linguistically complex as Standard American English

We don’t see/understand the complexity in other languages

 

Moral: All languages seem to permit as wide range of expressions as others

Linguistic Universals

What is in common with all languages?

Sentences are built from words based on the same physiological processes

• All languages have words

• All humans have ways of making sounds.

• Languages tend to use a small set of phonemic sounds

• Phoneme: The minimal unit of sound that contributes to meaning

How many phonemes in a language?

• English: 40 phonemes

• Range: Polynesian 11 to Khoisan 141

 

 

Discreteness

Messages in human language (e.g. speech sounds) are made up of units of which there is a discrete (limited) number

 

Arbitrariness

The relationship between meaningful elements in language and their denotation is independent of any physical resemblance between the two.

Words do not have to look or sound like what they describe

Openness

• New linguistic messages are created freely and easily

• Languages are not constrained in a way so that there are a limited number of messages that can be created.

Linguistic Productivity: The ability to understand and create an unlimited number of sentences

 

The question studied by psycholinguists is "how to characterize and account for the creativity to construct and create an infinite number of sentences given the limited capabilities of the human brain"

 

 

 

Duality of Patterning

Language involves relating two different kinds of patterns or forms of representation

• A phonological system

• A semantic system

These two systems use very different types of codes, although there is a phonological representation for each item in the semantic system

Phrase structure

Information on how a sentence is grouped into phrases.

The quiet boy ate the red apple

 

A set of Phrase Structure rules:

PS 1 S (sentence) -------------> NP + VP

PS 2 NP (noun phrase)-------------> det + (adj) + N

PS 3 VP (verb phrase) -------------> V +NP

PS 4 N (noun) -------------> boy, dog, man, book

PS 5 V (verb) -------------> ate, broke, kissed

PS 6 adj (adjective -------------> quiet, red, happy, wormy

PS 7 det (determiner) -------------> a, the

We use "lexical-insertion rules" to put words into the structure.

Phrase-structure rules provide a good account of phrase-structure ambiguity.

They are broiling hens

 

Morphology

Morphology is the component of grammar that builds words out of units of meaning (morphemes)

 

A morpheme is the smallest meaningful unit of language

 

How many morphemes?

bird

firetruck

undereducated

unmicrowaveability

 

Insights from American Sign Language (ASL)

Unlike speech, signs are expressed in visual or spatial form

Do a lot of the same grammatical concepts hold?

 

 

Arbitrariness

ASL possesses iconicity

signs can represent objects or actions to which they refer.

 

 

However, the degree of iconicity has declined over the years

 

Duality of Patterning

signs are composed of smaller elements that are meaningless

Example: 3 parameters

• 19 values of hand configuration

• 12 values of place of articulation

• 24 values of movements

 

Meaningless patterns can be combined in various ways to from ASL signs.

 

 

 

What about "openness" and "discreteness" within ASL?

Transformational Grammar (Chomsky 1950s)

 

Language: an infinite set of well-formed sentences

Grammar: A finite set of rules that generates sentences in the language

 

 

How do we know that a grammar is a good theory of language?

Three criteria:

Observational Adequacy:

A grammar is observationally adequate if it generates all acceptable sequences and no unacceptable sequences.

 

Descriptive adequacy:

A grammar must also explain how a sentence relates to other sentences that are similar & opposite in meaning.

The ball was caught by John

John caught the ball

The ball was not caught by John

 

Explanatory adequacy

It is possible for multiple grammars to attain observational and descriptive adequacy.

Which is the correct/best one?

Children learning language are presented with many samples of language and must determine the grammar from these samples.

 

There must be some innate language constraints that help children determine the correct grammar.

 

 

There exist Linguistic Universals that are common to all languages

 

The fact that there are similarities in languages is based on the fact that languages are determined by the nature of the mental structures and processes which characterize human beings

 

A Grammar must explain the role of linguistic universals in language acquisition

 

Deep and Surface structure

Deep structure: The structure of the sentence that conveys the meaning of the sentence.

 

Surface Structure: The superficial arrangement of constituents

 

Deep structure ambiguity: A single surface structure that is based on two different deep structures

Flying planes can be dangerous

Phrase structure rules would not be able to account for the differences in meaning

 

Sentences can have similar phrase structure, although their underlying structure is different:

John is easy to please

John is eager to please

 

Sentences can different surface structure, but similar deep structure

Arlene played the tuba

The tuba was played by Arlene

Transformational Grammar

A two part process to derive a sentence

1) Use Phrase-structure rules to generate the underlying tree structure (deep structure)

2) Apply a sequence of transformational rules to the deep structure to generate the surface structure of the sentence

Transformations occur by adding, deleting or moving constituents

John phoned up the woman

John phoned the woman up

Phrase structure approach: Two different rules

VP --> V + (particle) + NP

VP --> V + NP + (particle)

Each sentence is derived separately, using different phrase structure rules.

Transformational grammar approach: One rule

V + particle + NP --> V + NP + particle

John phoned up the interesting woman

John phoned the interesting woman up

John phoned up the woman with the curly hair

John phoned the woman with the curly hair up.

Restrictions on transformations

The particle-movement transformation can not be applied to pronouns

John called them up

*John called up them

 

 

Example 2: Passive transformation

NP1 + V + NP2 --> NP2 + be + V + en + by + NP1

Arlene played the tuba

The tuba was played by Arlene

 

Psychological Reality of Transformational Grammar

If using language is a process of converting the deep structure to the surface structure, then the number of transformation rules applied should affect how long it takes to process a sentence.

However, experiments do not consistently show that this holds true

 

Current theories of grammar

Lexical-Function Grammar

Made up of three components: a constituent structure, a functional structure, and lexical entries

 

Constituent Structure: Similar to phrase structure

Functional Structure: All the information needed for semantic interpretation

John told Mary to leave Bill

Predicate tell (subj, obj, V-comp)

Tense Past

Subj John

Obj Mary

V-comp predicate leave

subj Mary

obj Bill

 

Lexical Entries

Lexical entries contain information about:

• the forms of the word

• the kinds of sentences into which they fit,

• arguments and semantic roles

Mary kissed John

John was kissed by Mary

Entry for "kiss" includes

underlying semantic structure

kiss: (agent, patient)

Forms of the word

kiss: agent = subject: patient = object

(be) kiss: agent=object: patient = subject

 

 

 

 

 

Major significance of LFG

Most of the explanation of how we process language is based on the lexicon (where we store information about words) .

 

 

Government-Binding Theory or Universal Grammar

Chomsky’s view of innate grammatical mechanisms.

In GB theory, grammar is modular.

Grammar due to interaction of several independent subsystems, or modules.

Each module is fairly simple and performs part of the task

But all modules interact in order to constrain the rules made by the other modules in the grammar

 

Implications

We all inherit a universal grammar that can be set to different parameter values.

These parameter values correspond to different languages.

As we get experience with a language, we acquire these parameter values, and thus the language upon which it is based.

Research methods in Psycholinguistics

How do we observe, collect information on phenomena related to psycholinguistics?

Naturalistic Observation

Observing information in a non-experimental setting

Slips of the tongue

Phonological switching: Crushing blow --> Blushing crow

semantic replacements: blond eyes for blond hair

 

Language Acquisition

 

The use of language over time

 

Data from naturalistic observation

Rich, but hard to analyze

 

Controlled experiments

Goal: test an empirical hypothesis

Hypothesis: A chapter will be easier to understand if each section starts with a summary of what will be said.

 

Independent Variable: Variable that is manipulated to test the hypothesis.

 

Dependent Variable: Variable representing the behavior we want to measure

 

Control Variables: Other variables we need to control in order to see the effect of the independent variable

 

Subjects: Who is going to participate in the experiment?

 

 

Analysis: How do we know if there are differences bewteen the two chapters?

 

The Human Information Processing System

What psychological mechanisms are involved in using language?

The Sensory store

Processes incoming information from the environment

• Individual sensory stores for each sense

• Information retained for a short duration

The visual sensory store

Experiments by Sperling (1960)

X M R K

C N J P

V F L B

 

The partial report technique

Auditory sensory store

Experiment by Darwin, Turvey & Crowder (1972)

3 digits or letters auditorally presented to each ear and center at the same time

 

 

 

 

 

 

 

 

 

What is the use of the sensory store?

It maintains information long enough so that we can do additional processing to it.

Working memory or short term memory (STM)

STM used to describe the fact that it holds information for a short time, while working memory refers to the processing capacity.

STM works as a temporary holding place for intermediate decisions.

Limited in size.

Chunking

 

Working memory: there is a limited amount of processing capacity that you can use as you perform a problem

Long term memory

Knowledge of how to do things, things we have learned, grammar rules, personal memories.

All knowledge that is not active.

Information that becomes active is retrieved from LTM and put in STM.

Anything we learn is first processed in STM and some of it is put into LTM

 

 

Episodic vs. Semantic Memory distinction

Semantic memory

• Organized knowledge of words, concepts, symbols and objects. motor skills, general knowledge, spatial knowledge , social skills.

• All information is organized semantically, but not tagged based on when it was learned.

Episodic memory

• Holds traces of events specific time and place.

• Memory of personal experiences.

 

Interaction between semantic and episodic memory

 

What does the organization of the information processing system have to do with language processing?

Pattern Recognition

 

 

 

 

Parsing/understanding sentences in working memory

This is a long sentence and yet somehow you can keep it all in working memory

 

 

The organization of Long Term Memory

That cat plays really cool jazz

Serial vs. Parallel Processing

Serial processing: One process working at a time

Parallel Processing: Multiple processes working at a time

In a serial model of language processing, individual modules would work one at a time to process the information.

A parallel model would say that the processes happen at the same time.

 

Parallel models as neurally inspired models of cognitive processes

 

Top-Down vs. Bottom-Up processing

Cognitive processing occurs at levels

Bottom-up processing is done in such a way that all processing occurs starting from the lowest level processes and proceeds onto the higher level processes

Higher level processes do not influence any of the processing done at the lower levels

 

Top down processing: Information at the higher levels influences processing at the lower levels.

Advantages and disadvantages of Top-Down processing

Automatic vs. Controlled processes

We have a limited amount of processes that we can do at a time.

Controlled processing: Processes that require a substantial amount of cognitive processing.

Automatic processing: Processes that do not require a substantial amount of cognitive processing.

The role of practice in automatic processing

 

The Stroop effect

 

 

 

 

 

 

Putting it all together: Cognitive processes in action

 

The novice accepted the deal before he had a chance to check his finances, which put him in a state of conflict when he realized he had a straight flush.

The Internal Lexicon

How are words stored? What are they made up of? How are word related to each other? How do we use them?

 

Internal lexicon The representation of words in long term memory

 

Lexical Access: How do we activate the meanings of words?

 

 

Aspects of Meaning

Reference: The relationship between words and things in the world

Things in the world are referents of a word

My dog has fleas

My dog is from Mars

 

But not all reference can be mapped to concrete things

Abstract words: Love, Justice, Equality

Non existent objects: Unicorn, Martians

 

 

Meaning is not restricted to the real world, but also imaginary worlds

Sense: The relationship of a word with other words in the language

Student at NMSU vs. Undergraduate at NMSU

 

Synonymy (same meaning)

Car Automobile

Antonymy (opposite meaning)

Happy Sad

Incompatibility (do the words contradict each other?)

John is happy vs. John is sad

Hyponymy (are they part of the same class?)

A dog is an animal, Bowser is a dog,

 

 

 

 

 

Denotation vs. Connotation

Denotation: The objective meaning of the word

Connotation: The aspect of the meaning beyond its explicit meaning

Bachelor Spinster

Hungry Starving

 

 

The Mental Representation of Meaning

The representation of the meaning of a word is based on the semantic features of that word

 

We acquire the meaning of a word by learning its semantic features

Children make semantic mistakes

 

Verbs of possession.

We understand more than the meaning, we have knowledge of the relations between these words

sold vs. paid

give vs. receive

lose vs. find

Prototypes:

Some members of a category are better instances of the category than others

Apple vs. pomegranate

What makes a prototype?

More central semantic features

What type of dog is a prototypical dog

What are the features of it?

We are faster at retrieving prototypes of a category than other members of the category

Semantic Networks

Words can be represented as an interconnected network of sense relations

• Each word is a particular node

• Connections among nodes represent semantic relationships

 

 

 

Mental models:

A model/understanding of how the world works and how pieces of textual information fits in with it.

John is sitting in a chair. That chair is on a table. The table is blue and round. John has red hair.

 

 

 

The structure of the Internal Lexicon

How do these pieces of semantic information relate to each other?

Semantic verification task

An A is a B

An apple is a fruit

A robin is a bird

A robin is an animal

A dog has teeth

A fish has gills

A fish has feathers

An apple has teeth

NMSU is in New Mexico

Harvard is in California

 

Use time on verification tasks to map out the structure of the lexicon.

 

Models of the Lexicon

Collins and Quillian Hierarchical Network model

Lexical entries stored in a hierarchy, with features attached to the lexical entries

Representation permits cognitive economy

Testing the model

Sentence Verification time

Robins eat worms 1310 msecs

Robins have feathers 1380 msecs

Robins have skin 1470 msecs

 

A category size effect: Subjects do an intersection search

 

Problems with Collins and Quillian model

1) Effect may be due to frequency of association

2) Assumption that all lexical entries at the same level are equal

The Typicality Effect

Which is a more typical bird? Ostrich or Robin.

A whale is a fish vs. A horse is a fish

 

Major conclusions of the model:

1) If a fact about a concept is frequently encountered, it will be stored with that concept even if it could be inferred from a more distant concept.

2) The more frequently encountered a fact about a concept is, the more strongly that fact will be associated with the concept. And the more strongly associated with a concept facts are, the more rapidly they are verified.

3) Verifying facts that are not directly stored with a concept but that must be inferred takes a relatively long time.

 

Spreading Activation Models (Collins & Loftus)

• Words represented in lexicon as a network of relationships

• Organization is a web of interconnected nodes in which connections can represent:

categorical relations

degree of association

typicality

 

Retrieval of information

• Spreading activation

• Limited amount of activation to spread

• Verification times depend on closeness of two concepts in a network

 

Context effect in spreading activation models

Present either: Murder is a crime or Libel is a crime

Then get verification time for Robbery is a crime

Subjects faster when they see Murder than Libel. Why?

 

 

Advantages of Collins and Loftus model

• Recognizes diversity of information in a semantic network

• Captures complexity of our semantic representation

• Consistent with results from priming studies

Lexical Access

What factors are involved in retrieving information from the lexicon?

Semantic Priming

Meyer & Schvaneveldt (1971) Lexical Decision Task

Prime Target Time

Nurse Butter 940 msecs

Bread Butter 855 msecs

Evidence for associative spreading activation

 

Ratcliff and McKoon (1981)

Subjects study and memorize The doctor hated the book

Task: "Was this word from the sentence you memorized?"

Prime Target Time

None Book 667 msecs

Doctor Book 624 msecs

 

 

 

Word Frequency

Does word frequency play a role in lexical access?

 

Lexical Decision Task:

gambastya, revery, voitle, chard, wefe, cratily, decoy, puldow, raflot, oriole, vuluble, booble, chalt, awry, signet, trave, crock, cryptic, ewe, himpola

mulvow, governor, bless, tuglety, gare, relief, ruftily, history, pindle, develop, gardot, norve, busy, effort, garvola, match,sard, pleasant, coin, maisle

 

Lexical Decision is dependent on word frequency

 

 

Eyemovement studies:

Subjects spend about 80 msecs longer fixating on low-frequency words than high-frequency words

 

Morphological Structure

So we strip off the prefixes and suffixes of a word for lexical access?

Decision = Decide + ion

 

Lexical Decision Tasks:

Snodgrass and Jarvell (1972)

Response times greater for affixed words than words without affixes

Evidence for a stage where prefixes are stripped.

 

Taft (1981)

Lexical Decision times shorter for prefixed words (remind) vs. pseudoprefixes (relish)

 

But this does depend on strategy?

If you see many prefixed words you take longer with the pseudoprefixes, but if you see only 10% prefixed words, you don't take any longer

 

Lexical Ambiguity

Words can have multiple interpretations

The role of frequency of meaning

Hogaboam and Pefetti (1975)

Task, is the last word ambiguous?

The jealous husband read the letter

The antique typewriter was missing a letter

Subjects are faster on the second sentence. Why?

Evidence for more than one node for each word in semantic memory

 

The role of prior context

Swinney (1979)

Auditory presentation:

Rumor had it that, for years, the government bulding has been plagued with problems. The man was not surprised when he found several spiders, roaches and other bugs in the corner of his room.

Lexical Decision task

Context related: ant

Context inappropriate: spy

Context unrelated sew

Within 400 msecs of hearing "bugs", both ant and spy are primed, but after 700 msecs, only ant is primed

Why activate all aspects of meaning for a word?

Retention of lexical items

How long does it take for studied words to be forgotten?

Wickelgren (1975)

Presented subjects with a sequence of words to study

Examined the probability of recognizing words over 14 days

Performance systematically decays over time

Negatively accelerated decay

 

 

Bahrick (1984)

Student's retention of spanish-english vocabulary items from 0 to 50 years

 

 

 

Power law of decay

Review on the internal lexicon

Aspects of meaning:

Reference and Sense

Denotation and Connotation

 

What is the mental representation of meaning?

 

Models of the Lexicon

Hierarchical Network Model

Spreading Activation Model

 

What factors are involved in retrieving information from the lexicon?

Semantic Priming

Word Frequency

Morphological Structure

Lexical Ambiguity

Retention of lexical items