PsycholinguisticsPsycholinguisticsOr what I did in my MScOr what I did in my MSc

Natasha DareNatasha Dare

This talkThis talk What is psycholinguistics?What is psycholinguistics? The origins of languageThe origins of language Some of the major areas of researchSome of the major areas of research

Famous experimentsFamous experiments TerminologyTerminology State of the artState of the art

Particularly what is being done at EdinburghParticularly what is being done at Edinburgh When language goes wrongWhen language goes wrong

Not controversial Not controversial

What is psycholinguistics?What is psycholinguistics? Psychological processing of languagePsychological processing of language

Part of cognitive sciencePart of cognitive science Input from neuroscience, informatics and Input from neuroscience, informatics and

linguisticslinguistics Why is this interesting/important?Why is this interesting/important?““No man is an island, entire of itself; every man is No man is an island, entire of itself; every man is

a piece of the continent, a part of the main.”a piece of the continent, a part of the main.”

John Donne, Meditation XVIIJohn Donne, Meditation XVII

Origins of language - Child Origins of language - Child languagelanguage

Infants very quickly learn about languageInfants very quickly learn about language 3-day old neonates prefer the sound of their mother’s voice 3-day old neonates prefer the sound of their mother’s voice

(DeCasper & Fifer, 1980)(DeCasper & Fifer, 1980) 4-day old neonates prefer listening to their parents’ language 4-day old neonates prefer listening to their parents’ language

(Mehler & Dupoux, 1994)(Mehler & Dupoux, 1994) 1-month old babies can distinguish between speech sounds 1-month old babies can distinguish between speech sounds

(Eimas, Miller, & Jusczyk, 1987)(Eimas, Miller, & Jusczyk, 1987) 6 weeks: cooing6 weeks: cooing 6-9 months: babbling6-9 months: babbling 12 months: initial word use12 months: initial word use 18 months: vocabulary explosion of 40 new words per 18 months: vocabulary explosion of 40 new words per

weekweek 24 months: short sentences24 months: short sentences 36 months: 90% intelligible36 months: 90% intelligible

Origins of language - Animal Origins of language - Animal communicationcommunication

Many animals have complex communication systemsMany animals have complex communication systems Vervet monkeys: leopard vs snake vs eagle Vervet monkeys: leopard vs snake vs eagle Chaffinches: combined territorial and mating calls – an initial trill to Chaffinches: combined territorial and mating calls – an initial trill to

deter males, and a final flourish to attract femalesdeter males, and a final flourish to attract females But is this true language?But is this true language?

Animal communication is holisticAnimal communication is holistic Human language is compositionalHuman language is compositional

Can see origins of human language in animalsCan see origins of human language in animals e.g. teeth baring ritual = threat -> symbolic actse.g. teeth baring ritual = threat -> symbolic acts courtship gift = attention to third entity -> reference courtship gift = attention to third entity -> reference

““It is nothing other than words which has made us human”It is nothing other than words which has made us human”PavlovPavlov

How did human language develop?How did human language develop? Robotic agents to simulate evolution, especially Robotic agents to simulate evolution, especially

emergence of regularitiesemergence of regularities 5 vowels most common (Latin)5 vowels most common (Latin) Maximal acoustic distanceMaximal acoustic distance De Boer (1997) - regularities emerge spontaneously De Boer (1997) - regularities emerge spontaneously

after 3,000 games after 3,000 games Pidgins -> creolesPidgins -> creoles

Pidgins are formed when communities are Pidgins are formed when communities are deliberately mixed e.g. Ruso-Norskdeliberately mixed e.g. Ruso-Norsk

Creoles develop from pidgins but have full syntax Creoles develop from pidgins but have full syntax and native speakersand native speakers

Tok Pisin: Tok Pisin: sapossapos = if, = if, bilongbilong = possessive = possessive

ie

aou

front backhigh

low

The problemThe problem

δϖσlexicon

Bilingualism

Spoken word comprehension

Written word comprehension

ComprehensionProduction

Spoken word productionWritten word production

Dialogue

/Discourse

Some linguistics termsSome linguistics terms SemanticsSemantics

MeaningMeaning SyntaxSyntax

Specialised term for grammar – word orderSpecialised term for grammar – word order PhoneticsPhonetics

Sounds of language e.g. Sounds of language e.g. thinthin = [ = [θθIn]In] PhonologyPhonology

Language-specific phonetics e.g. rhotic [r] in Language-specific phonetics e.g. rhotic [r] in paparrtyty in USA in USA MorphologyMorphology

Words and word formationsWords and word formations Sometimes change the underlying meaning (re-)Sometimes change the underlying meaning (re-) Sometimes don’t (-s)Sometimes don’t (-s)

PragmaticsPragmatics Language useLanguage use

Auditory word recognitionAuditory word recognition

What are the sound representations used to What are the sound representations used to access the lexicon (mental dictionary)?access the lexicon (mental dictionary)?

Very difficult taskVery difficult task We understand 20 phonemes per secondWe understand 20 phonemes per second We can recognise words in context 200 ms after We can recognise words in context 200 ms after

onset (Marslen-Wilson, 1984)onset (Marslen-Wilson, 1984) Miller and Jusczyk (1989): invariance (phonemes Miller and Jusczyk (1989): invariance (phonemes

sound different in different contexts) and sound different in different contexts) and segmentation (speech slurs words together)segmentation (speech slurs words together)

Assimilation of sound properties from other wordsAssimilation of sound properties from other words Co-articulation of words Co-articulation of words

Auditory word recognition cont.Auditory word recognition cont. Two cues: uniqueness point and contextTwo cues: uniqueness point and context

/t/ /tr/ /tre/ /tres/ /tresp/ /trespass//t/ /tr/ /tre/ /tres/ /tresp/ /trespass/ uniqueness pointuniqueness point Cohort model (Marslen-Wilson, 1989)Cohort model (Marslen-Wilson, 1989)

Bottom-up: Access a cohort of possible words, Bottom-up: Access a cohort of possible words, select one, integrate its properties into current task select one, integrate its properties into current task

TRACE model (McClelland & Elman, 1986)TRACE model (McClelland & Elman, 1986) Interactive: Context (top-down) and acoustic signal Interactive: Context (top-down) and acoustic signal

(bottom-up) both cause one candidate to be (bottom-up) both cause one candidate to be selectedselected

Visual word recognitionVisual word recognition

How do we know if a letter string is a lexical item? How do we know if a letter string is a lexical item? Two levels of checks: phonetic constraints on impossible non-Two levels of checks: phonetic constraints on impossible non-

words (words (mxbtmxbt), phonological/semantic constraints on possible ), phonological/semantic constraints on possible non-words (non-words (kstreemkstreem//glubglub))

What factors affect the ease of recognition?What factors affect the ease of recognition? Frequency: Frequency: havehave easier than easier than jadejade Neighbourhood: Neighbourhood: minemine easier than easier than muchmuch Length: Length: bankbank easier than easier than discriminatediscriminate

Lexical ambiguityLexical ambiguity Homophones: Homophones: knightknight//nightnight Homographs: Homographs: leadlead Meanings versus senses: Meanings versus senses: bankbank versus versus filmfilm

Visual word recognition cont.Visual word recognition cont.

Connectionist modelling of orthography (letters) -> Connectionist modelling of orthography (letters) -> phonology (sounds) e.g. split-fovea model (Shillcock, phonology (sounds) e.g. split-fovea model (Shillcock, Ellison, & Monaghan, 2001)Ellison, & Monaghan, 2001) Neurally inspiredNeurally inspired Distributed network of layers of nodesDistributed network of layers of nodes Words are represented by patterns of Words are represented by patterns of activation between associated words and soundsactivation between associated words and sounds Models are trained with word-pronunciation pairsModels are trained with word-pronunciation pairs They learn by reducing the error between the They learn by reducing the error between the actual and desired outcomeactual and desired outcome

Dyslexias – developmental (Dyslexias – developmental (necisarynecisary), surface (), surface (mintmint)),, phonological (phonological (kintkint), deep (), deep (sympathysympathy -> -> orchestraorchestra))

/a//o/ /l/ /g/ /d/

o d a l g

output

input

Sentence comprehensionSentence comprehension How do we parse sentences (build up syntactic How do we parse sentences (build up syntactic

representations) given the meanings of words?representations) given the meanings of words? Phrase structure rules -> tree diagrams of Phrase structure rules -> tree diagrams of

sentencessentences S -> NP VP VP -> V NP NP -> Det NS -> NP VP VP -> V NP NP -> Det N

S

NP VP

Det N V NP

Det N

The dog bit the man

How do we access this seemingly How do we access this seemingly effortless task?effortless task? Ambiguity – two or more possible structuresAmbiguity – two or more possible structures

““enraged cow injures farmer with axe”enraged cow injures farmer with axe” Reassessment using garden-path sentences Reassessment using garden-path sentences

(Frazier, 1987)(Frazier, 1987) The horse raced past the barn fell – reduced The horse raced past the barn fell – reduced

relative clauserelative clause

Sentence processing cont.Sentence processing cont. Eye-tracking and self-paced reading: longer time spent Eye-tracking and self-paced reading: longer time spent

looking at a word = greater processing difficultylooking at a word = greater processing difficulty Does semantics help with parsing i.e. are we garden-pathed Does semantics help with parsing i.e. are we garden-pathed

when there are strong semantic pointers to one interpretation?when there are strong semantic pointers to one interpretation? Ferreira and Clifton (1986): Ferreira and Clifton (1986): The defendant/evidence examined The defendant/evidence examined

by the lawyer turned out to be unreliable by the lawyer turned out to be unreliable Although Although evidenceevidence disambiguates disambiguates examinedexamined, eye-movements , eye-movements

showed that semantic information did not stop people from being showed that semantic information did not stop people from being garden-pathedgarden-pathed

But how much do we actually parse at all?But how much do we actually parse at all? Christianson, Hollingworth, Halliwell & Ferreira (2001): Christianson, Hollingworth, Halliwell & Ferreira (2001): While Bill While Bill

hunted the deer ran into the woodshunted the deer ran into the woods Participants thought that the deer ran into the woods and that Bill Participants thought that the deer ran into the woods and that Bill

hunted the deer hunted the deer

DiscourseDiscourse

How do we maintain coherence across longer texts?How do we maintain coherence across longer texts? InferencesInferences

Logical: Logical: Toby is a bachelorToby is a bachelor – Toby is a man – Toby is a man Bridging: integrating new with previously given information. Bridging: integrating new with previously given information.

Uses anaphor e.g. Uses anaphor e.g. John gave Bob the book. He liked it very John gave Bob the book. He liked it very muchmuch – who do – who do it it and and hehe refer to? refer to?

Elaborative: extending what is in the text to world knowledgeElaborative: extending what is in the text to world knowledge Tend to only remember gist/important points e.g. Tend to only remember gist/important points e.g.

passages with personal significancepassages with personal significance

“ “One night there flew over the city a little Swallow. His friends had gone away to Egypt six One night there flew over the city a little Swallow. His friends had gone away to Egypt six weeks before, but he had stayed behind, for he was in love with the most beautiful Reed. He weeks before, but he had stayed behind, for he was in love with the most beautiful Reed. He had met her early in the spring as he was flying down the river after a big yellow moth, and had had met her early in the spring as he was flying down the river after a big yellow moth, and had been so attracted by her slender waist that he had stopped to talk to her.been so attracted by her slender waist that he had stopped to talk to her. "Shall I love you?" said the Swallow, who liked to come to the point at once, and the Reed "Shall I love you?" said the Swallow, who liked to come to the point at once, and the Reed made him a low bow. So he flew round and round her, touching the water with his wings, and made him a low bow. So he flew round and round her, touching the water with his wings, and making silver ripples. This was his courtship, and it lasted all through the summer.”making silver ripples. This was his courtship, and it lasted all through the summer.”

Oscar Wilde, ‘The Happy Prince’Oscar Wilde, ‘The Happy Prince’

Language productionLanguage production 3 parts: conceptualisation, formulation, 3 parts: conceptualisation, formulation,

articulationarticulation Pre-verbal messagePre-verbal message Concept -> linguistic formConcept -> linguistic form Articulatory planningArticulatory planning Motor executionMotor execution

Although we are very good at this (1/1000 words Although we are very good at this (1/1000 words is an error), errors are very revealingis an error), errors are very revealing Spoonerisms: exchange of initial consonantsSpoonerisms: exchange of initial consonants““The Lord is a shoving leopard to his flock”The Lord is a shoving leopard to his flock”““I’d rather have a bottle in front of me than a frontal lobotomy”I’d rather have a bottle in front of me than a frontal lobotomy” Freudian slipsFreudian slips““A Freudian slip is when you say one thing but mean your mother”A Freudian slip is when you say one thing but mean your mother”

Language production cont.Language production cont.

Tip of the tongue statesTip of the tongue states Temporary inability to retrieve word despite ‘feeling of knowing’Temporary inability to retrieve word despite ‘feeling of knowing’ Can retrieve partial information e.g. gender, initial phoneme, Can retrieve partial information e.g. gender, initial phoneme,

number of syllablesnumber of syllables Blockers/pop-upsBlockers/pop-ups

To formally renounce the throneTo formally renounce the throne begins with begins with aa, like , like abductabduct

Suggests that there is a separation between Suggests that there is a separation between syntax/semantics (lemma) and morphology/syntax/semantics (lemma) and morphology/phonology (lexeme)phonology (lexeme) Levelt (1992)Levelt (1992) Dell (1986)Dell (1986)

The lexiconThe lexicon How do we know what words mean?How do we know what words mean? Different methods proposedDifferent methods proposed

Defining attributes (ISA links)Defining attributes (ISA links) animal superordinateanimal superordinate bird fish bird fish canary robin salmon subordinatecanary robin salmon subordinate Exemplars/typicality effectsExemplars/typicality effects

Concepts partly based on perception e.g. brown, Concepts partly based on perception e.g. brown, stickysticky

But not the whole story e.g. But not the whole story e.g. thethe, , himhim

breathes/skin/skeleton

gills/swims/fins

wings/flies/feathers

The lexicon cont.The lexicon cont. Semantically similar words are interchangeable Semantically similar words are interchangeable

in sentencesin sentences The child slept on the bedThe child slept on the bed The dog walked on the carpetThe dog walked on the carpet Semantic networks (e.g., Burgess & Lund’s Semantic networks (e.g., Burgess & Lund’s

HAL)HAL) Meanings come from other words (like a dictionary)Meanings come from other words (like a dictionary) Distances between words in network show Distances between words in network show

relatedness, with 140,000 dimensional spacerelatedness, with 140,000 dimensional space kitten

lion tooth

oystercar eye

facefoot

leg

DialogueDialogue Two or more people – turn-taking, feedbackTwo or more people – turn-taking, feedback

Difficult to study as fewer paradigms, noisier data, harder to Difficult to study as fewer paradigms, noisier data, harder to control control

Commonly use tasks or games to elicit controlled languageCommonly use tasks or games to elicit controlled language Audience design: do speakers tailor their utterances to Audience design: do speakers tailor their utterances to

the listeners? the listeners? Common ground: do listeners use information known Common ground: do listeners use information known

only to themselves?only to themselves? Perspective-taking: do speakers and listeners Perspective-taking: do speakers and listeners take each other into account?take each other into account?

Schober and Clark (1989) tangram matching taskSchober and Clark (1989) tangram matching task Matchers 99% accurate, overhearers 78% Matchers 99% accurate, overhearers 78%

Dialogue cont.Dialogue cont. Alignment (Pickering & Garrod, 2004)Alignment (Pickering & Garrod, 2004)

Linguistics representations used by interlocutors become Linguistics representations used by interlocutors become aligned at multiple levels via primingaligned at multiple levels via priming

Lexical: use the same referring expressionsLexical: use the same referring expressions Syntactic: ‘Syntactic: ‘AtAt what time do you close?’ ‘ what time do you close?’ ‘AtAt 9’ 9’ Accent and speech rateAccent and speech rate

Alignment permeates throughout levelsAlignment permeates throughout levels All happens automaticallyAll happens automatically

Prosody, disfluencyProsody, disfluency What information do they give the listener?What information do they give the listener?

When Roger leaves the house is dark/it’s darkWhen Roger leaves the house is dark/it’s dark Kjelgaard and Speer (1999): when syntactic and prosodic cues Kjelgaard and Speer (1999): when syntactic and prosodic cues

matched, listeners’ comprehension was facilitated matched, listeners’ comprehension was facilitated Does it follow that they are produced intentionally by the Does it follow that they are produced intentionally by the

speaker?speaker?

BilingualismBilingualism Do we have separate language systems for Do we have separate language systems for

each language?each language? One system – semantic priming One system – semantic priming produces facilitation between languagesproduces facilitation between languages Two systems – aphasia can affect one Two systems – aphasia can affect one language onlylanguage only Does age/proficiency explain these contradictions? Does age/proficiency explain these contradictions?

Problem: no standard, widely-used proficiency testProblem: no standard, widely-used proficiency test Hard to generalise across resultsHard to generalise across results Country/community/family/colleagues all have effectsCountry/community/family/colleagues all have effects

Most likely is one semantic store, two lexiconsMost likely is one semantic store, two lexicons

semantics

L1 L2

Final factsFinal facts 6,912 known living languages in the world6,912 known living languages in the world 896,190 words in English (correct as of 896,190 words in English (correct as of

yesterday)yesterday) Mandarin has 1,075,000,000 speakersMandarin has 1,075,000,000 speakers Most popular word is ‘ok’Most popular word is ‘ok’

““When ideas fail, words come in very handy”When ideas fail, words come in very handy”GoetheGoethe