Chapter 11: Language Flashcards
how many words are in the vocabulary of an average American adult?
42,000
how many messages can the average American adult produce?
an infinite number of messages
Psycholinguistics
the subfield of psychology concerned with how we learn, understand, and produce language
examples of other species communciating
- Ants communicate using the transfer of chemicals called pheromones
- Bees communicate information about their food source using a complicated waggle dance
- Male Campbell’s monkeys have their own proto-syntax (different sequences of sounds signal different information) with 3 basic sounds that they can combine to signal different inforamtion
can animals communicate complex information?
Animal communication tends to be highly limited
what’s the most fundamental difference between human and animal language
the presence or absence of grammar for combining words
Productivity/digital infinity
a feature of human language in which an effectivity infinite number of grammatical sentences are possible
Noam Chomsky on language’s uniqueness
language’s infinite expressive ability is what makes it so unique
Alex the Parrot
was trained to produce some remarkable linguistic behaviours, including describing abstract concepts. However, he was never able to generate a novel, multiword sentence using grammar
Washoe the chimp
was able to learn up to 350 signs and combine words to create new ones, but he never reached any true linguistic competency such that he could generate novel, multi-sign sentences
can animals produce true language?
Animals are unable to produce true language even with extensive training
Behaviourists like B.F. Skinner on language acquisition
propose that all language is learned through reinforcement and modelling
Noam Chomsky on language acquisition
there is an innate capacity to learn language that is present prior to any actual language experience
universal grammar
a theorized set of syntactic linguistic rules that are present across all natural human languages
what gene is thought to be involved in language?
FOXP2
why do researchers think FOXP2 is involved in language?
is involved in language because children who have mutations in this gene often suffer from developmental verbal dyspraxia
Developmental verbal dyspraxia
a disorder that affects the ability to pronounce syllables and words
FOXP2 knockout in female mice
they don’t generate high-frequency vocalizations in response to their pups
FOXP2 knockout in songbirds
affects their ability to learn and imitate their characteristic songs
Poverty of the stimulus
a proposed phenomenon that states that there is insufficient data for children to learn the rules of grammar based on experience alone
Pidgin
a quasi-language that does not contain full grammar, typically generated by adult immigrants to a location with a different language
Creole
a fully expressive novel language consisting of a combination of two pre-existing languages, typically created by the children of immigrants who are exposed to their parents’ language alongside that of their current resilience
what is a creole an example of
people acquiring grammar without sufficient stimuli
Deaf isolates
people who can’t hear but are not exposed to any real sign language
do deaf isolates communicatie?
They will develop their own sign language, demonstrating an innate capacity to learn a language
when does language learning begin?
before birth; Infants’ sucking responses in the womb demonstrate a preference for their mothers’ voices and their native language
language abilities 0-3 months
cooing
language abilities 4-8 months
babbling (consonant & vowel)
language abilities 1-2 years
two word phrases
language abilities 2-3 years
explosion of word knowledge; 2-3 word telegraphic speech
language abilities 8 months- 1 year
single words
language abilities 3-4 years
complex multiword speech
language development in different environments
The regularity of linguistic development persists even when there’s a high degree of linguistic difference in environments
Child-directed speech (CDS) or infant-directed speech (IDS)
speech that is tailored to a young infant or child
Motherese/ parentese
an infant-directed way of speaking that involves sing-song like speech cadences, exaggerated vowel pronunciations and repetition
do infants show a preference for types of speech?
they prefer IDS
head-turn task
a behavioural task used in infant language testing where babies are conditioned to turn their head when they hear a change in sound
Liu et al., 1993 motherese experiment
assessed the abilities of 6-12 month old babies to discriminate between different speech sounds using the head-turn task. Found that the language abilities of infants were positively correlated with their mother’s use of elongated and open vowel sounds, typical of motherese
are IDS & motherese necessary for language development?
no
phonemes
the smallest unit of speech that can change the meaning of the world
phonemes for apple
ah, p, l
morphemes
the smallest meaningful units of speech, which have to convey some meaning either on their own or in combination with other units of speech. This includes prefixes and suffixes
morphemes for apples
apple, s
Pollack & Pickett, 1964 context and speech experiment
They found that people could correctly identify the word less than half of the time, but when several words from before or after the target word were included, people were better at identifying the word. Demonstrates that context plays an important role in speech perception
phonemic restoration effect
a perceptual phenomenon in which sound that is missing or obscured is still perceived if it is highly predictable
who discovered the phonemic restoration effect?
Warren, 1970
what type of processing is phonemic restoration effect based on?
top-down
Warren, 1970 phonemic restoration effect experiment
played participants a recording with a phoneme covered with a cough. Found that most participants didn’t notice any sounds missing
McGurk effect
when we view the visual articulations of one phoneme while hearing the auditory signal consistent with another. Demonstrates that our brain uses additional info besides the speech stimulus to determine what phoneme is being said
main challenge to speech segmentation
The fact that we do not pause between words
Saffran et al., 1996 artificial language and preferences experiment
Found that babies showed a clear bias towards nonwords and that they encoded the frequency with which different sounds appear together, allowing them to identify words
lexical processing
determining the meaning of individual words
homophones
words that are pronounced identically but have different meanings
homographs
words that are spelt identically but have different meanings
why do we prefer commonly used words?
because they are processed quicker
context is important for
disambiguation
Lexical decision task (LDT)
a psychological task in which a participant makes a quick decision about whether a given string of letters is a legitimate word or not
Scarborough et al., 1977 commonly used words experiment
used the LDT and found that people were faster at recognizing strings as real words when those words were more common
Swinney, 1979 LDT and homophones experiment
Had participants perform a LDT while listening to one of the versions of the homophone and found that participants’ response times were faster in the LDT when related words were presented together.
are both meanings of the homophones activated?
both meanings are briefly activated despite the surrounding context favouring one meaning
parsing
breaking up languaging into its constituent parts
garden-path sentence
a sentence that tends to induce the wrong parsing
clause
a group of words that express a full idea of someone or something
subject
someone or something
predicate
the verb
syntax-first approach
a theory of language parsing that holds that the parsing of a sentence is first derived based on principles of grammar alone without regard to the meaning of the words, except to determine which grammatical category they belong to
late closure
a tendency when parsing to attach incoming words to the current phrase
Trueswell et al.,1994 eye tracking and prosody experiment
presented participants with sentences that contained parsing ambiguity and tracked their eye movements to see whether they had to go back and reanalyze each sentence. Found that the meaning of words affected parsing.
Tanenhaus et al., 1995 visual input and parsing expeirment
presented participants with sentences that contained parsing ambiguities, and tracked their eye movement while also showing them visual stimuli. Found that the visual information affected parsing behaviour.
prosody
the patterns of stress and intonation (pitch) of the speaker, which can convey critical information
punctuation and prosody
Punctuation can be seen as a limited version of prosody
discourse processing
The ability to understand language that is at least several sentences long
anaphoric inference
an inference that connects a reference to an object or a person in one sentence to an object or person in a different sentence
causal inference
an inference about the causal relationship between information in one sentence regarding information in another sentence
backward inference
a kind of inference in discourse processing in which previous information is needed in order to process current information
elaborative experiment
a kind of inference in discourse processing in which the inferred information is not necessary in order to properly understand the text
when are participants more likely to report information as novel?
if it was consistent with their expectations
when do people engage in inferential reasoning?
at the single-word level as they are reading
instrumental inference
a form of elaborative inference in discourse processing in which the tool or instrument that is typically used to perform the task is inferred from the task
who discovered instrumental inferences
Singer, 1979
when do people engage in online elaborative inferences?
if they have enough context
example of the importance of context in inferences
participants understood a text much better when they were told it was about doing laundry because it allowed them to relate the text to their prior experiences
Neurolinguistics
a branch of linguistics concerned with the relationship between linguistic behaviour and the structures of the brain
Wernicke’s area
- critical for language comprehension
- Located in the temporal lobe
- The first region of the cortex to receive auditory information after the thalamus
Broca’s area
- critical for language production
- Located in the frontal lobe
recent studies have provided evidence that language comprehension really involves___
a broader portion of the temporal lobe
Broca’s area is also involved in ___
language comprehension, particularly the presence of complex sentences
are other areas involved in speech production?
Broca’s area might not be the only factor in speech production
Arcuate fasciculus
a band of fibres in the brain that connects Broca’s and Wernicke’s areas
Arcuate fasciculus in other species
Is absent in many primates, suggesting it might be involved in humans’ unique language capacities
hemisphere specialization and language processing
Language production depends heavily on the left hemisphere but the right hemisphere is involved as well, especially in discourse processing
patients with right-hemisphere damage
have language deficits. They speak the same number of words, but their speech is less informative and coherent
Beeman et al., 2000 hemisphere specialization and naming task experiment
Found a reduced naming time for the left visual field, suggesting that the right hemisphere might be involved in elaborate processing
recent studies found pronounced activation in what area during comprehension tasks?
right temporal lobe
Linguistic relativity
the theoretical perspective that the language someone speaks affects other areas of cognition
who is linguistic relativity attributed to?
Edward Sapir & Benjamin Lee Whorf
Whorf linguistic relativity studies
studied Indigenous languages and found that they had many words for snow and thus perceived it differently. However, they didn’t have words for units of day and this influenced their behaviour. These studies are largely discredited
Linguistic universalists
the theoretical perspective that the language someone speaks doesn’t affect other areas of cognition
Winawer et al., 2007 linguistic relativity study
had Russian speakers (who have two distinct names for blue) perform a colour-matching task and found that they performed the task 10% faster when the non matching patch was in a different colour category than the matching patch compared to when they were both in the same colour category. This didn’t hold for English speakers.
Natural language processing (NLP)
a subfield of artificial intelligence concerned with machines understanding and producing languages
Turing test
a test proposed by Alan Turing to determine whether a machine can think based on fooling a human conversing with the machine that it is another human
does AI pass the Turing test?
To date, no AI does very well at this task
Sequence-to-sequence learning
a type of machine learning task in which both the inputs and outputs are sequences, such as strings of words represented numerically
what distinguishes humans from other species?
Language
what is language?
A shared symbolic system for purposeful communication
symbolic
there are units to reference something else
shared
it is common among a group of people
purposeful
to communicate and translate thoughts
morphology & number of language speakers
Morphology (complexity) decreases with languages spoken by more people
lexical tones and climate
Lexical tones (tonal languages) are partly determined by climate. Tonal languages are less prominent in dry climates because dry air makes it harder to have the vocal control to produce tonal sounds
language and gender inequality
Countries with gendered languages experience higher gender inequality
women vs. men’s use of language
Women use more adjectives and first-person plurals than men (we need to hurry) and use a reverse accent more than men (end statements like questions)
aphasia
Impaired language function from brain injury
broca’s aphasia
- Expressive aphasia
- Intact language comprehension
- Impaired speech production and articulation
cause of Broca’s aphasia
damage to Broca’s area (left inferior frontal gyrus)
Patient Tan
- Could only speak one syllable (Tan)
- Still tried to communicate via gestures, tone, inflection
- A large lesion in Broca’s area (left inferior frontal gyrus)
speech in patients with Broca’s aphasia is ___
Mostly just nouns and verbs
writing in patients with Broca’s aphasia is ___
affected in a similar manner
the severity of Broca’s aphasia depends on ___
the extent of the damage to Broca’s area
Wernicke’s aphasia
- Comprehension, both written and spoken
- Language content is not meaningful nor comprehensible
- Characterized by paraphasias
what causes Wernicke’s aphasia
damage to Wernicke’s area (posterior superior temporal lobe), typically left hemisphere
what are the types of paraphasias
- verbal
- phonemic
- neologisms
verbal paraphasia
- substituting a word with something semantically-related
- Shares meaning with the intended word
- Ex. swapping term brother with sister
phonemic paraphasia
- swapping or adding speech sounds
- Shares sounds with the intended word
- Ex. calling crab salad sad cralad
neologisms
- using a made-up word
- Ex. mansplain
conduction aphasia
- Production intact
- Comprehension intact
- Impaired repetition
cause of conduction aphasia
damage to the arcuate fasciculus (neural pathway that connects Broca’s and Wernicke’s areas)
lateralization of language
- Aphasias indicate that language is left-lateralized
- Lateralization is not fully understood nor linked to handedness
- Broader aspects of language are supported by the right hemisphere
updated classical language model
- Distinction between language production and comprehension
- Language function relies on a set of neural circuits
dorsal stream of language
for matching sounds to movement (speech production)
ventral stream of language
for matching sounds to meaning (speech recognition)
nurturist view of language
language is acquired through the same mechanisms as skill learning
naturist view
we are born with the innate capacity to learn language
what view of language acquisition is favoured?
naturist view
the innateness hypothesis
Grammar syntactic structure, is separate from semantic meaning
Language Acquisition (LAD)
abstracted entity that supports language
types of support for the innateness hypothesis
- convergence
- uniformity
- poverty of the stimulus agent
convergence
children are exposed to different learning situations, yet convergence on the same grammar
uniformity
children have a similar language-learning trajectory
poverty of the stimulus agent
the linguistic environment of a child is not sufficient enough to allow that child to learn a language via reinforcement, rules or imitation.
cons of the poverty of the stimulus agent
- Can’t address the following questions:
1. What information is innate?
2. How can you disprove this argument?
3. How do you provide a complete account of all the linguistic data available to a child? - Adult reformulations of children’ speech target the structure but not meaning
long-term effects of parentese
The better infants are at distinguishing the phonetic units, through parentese, the better complex language skills they have years later
how many English phonemes are there?
a few dozen
syntax
rules that govern how words are arranged in a sentence
semantics
the meaning
language comprehension
Understanding the message from language
phonological ambiguity
Determining phonemes depends on the audio signal, often noisy
lexical ambiguity
A single-word form can refer to more than one different concept
Laurel vs. Yanny is an example of
phonological ambiguity
how is lexical ambiguity resolved?
context
cross-modal priming task
- Bug primes spy (context inappropriate) just as much as ants (context related) at shorter SOA
- Bug primes only ants and not spy at longer SOA
- Both meanings are initially retrieved, but the contextually inappropriate meaning is quickly discarded
constraint-based models
constraints are used to resolve ambiguity (ex. Semantic and thematic context, Expectation, Frequency)
reading
a newer form of language (5,500 years old)
surface dyslexia
- impaired at producing irregular words (25% of English words)
- Reading happens letter-by-letter
- Difficulty matching words to the mental dictionary
phonological dyslexia
- impaired at reading non-words or new words
- Reading happens by comparing whole words to a mental dictionary (lexicon)
- Difficulty reading letter by letter
dual route model of reading
Route 1: accessing full words through the mental lexicon
Route 2: reading letter by letter (grapheme-phoneme conversion)
nativism
language and thought are independent
linguistic relativity
language and thought are interconnected
language of thought thypothesis
Mentalese, an innate non-spoken to represent all conceptual content and propositions to create thought
what type of view is the language of thought?
nativist
Sapir-Whorf Hypothesis
Linguistic determinism states that a person’s thoughts are determined by language
Hopi tribe experiment
Hopi Indians don’t have a word for time, so do they not think about time?. Evidence suggests that they do
colour matching and Berinmo tribe
- Berinmo tribe: the colour boundary between English blue and green doesn’t exist
- English speakers: the colour boundary for Berinmo words nol and wor don’t exist
- Berinmo tribe members were impaired at blue/green pair trials
- English speakers were impaired at nol/wor pair trials
- Demonstrates that language shapes colour memory
colour matching and Dani tribe members
- Participants were either English language speakers (11 words for colour) or Dani tribe members (2 words for colour)
- Test 1: named colour patches; two groups performed this differently
- Test 2: match/categorize learned colour patches; both groups performed equally
- Accessing colour categories without language labels didn’t change across language
bilinguals
All individuals who use more than one language
L1
first (native) language
L2
second language
how many people in the world are bilingual?
Estimates range from 50-70%
% of English-French bilinguals in Canada vs. Quebec vs. Montreal
- 18% of people in Canada are English-French bilinguals
- 46% of people in Quebec are English-French bilinguals
- 70% of people in Montreal are English-French bilinguals
bilingualism in the U.S.
- 22% of Americans speak a language other than English at home
- Spanish is the most common second language
- Bilingualism has almost doubled since 1980
the traditional view of psycholinguistics
- Most research on language and cognition examined only speakers of a single language (typically English)
- Monolinguals were the model subjects of the study
- Only the native language could provide an ideal basis for understanding the nature of the language system
bilinguals on the traditional view of psycholinguistics
On this view, bilinguals have been considered a special group of language users
accented speech and L2 learners
The older individuals were when first exposed to the L2, the more accented their speech is perceived to be
grammaticality and L2 learners
Individuals’ ability to correctly judge the grammaticality of sentences decreases with the age they were exposed to the second language
new attitude about bilingualism
- Language learning occurs at all ages and language processes are dynamic
- Bilingualism provides a lens for examining aspects of cognition that are being obscured by studying monolinguals alone
3 key discoveries about bilingualism
- Bilinguals are not two monolinguals in one. Both languages are active and competing
- The bilingual’s language system is permeable in both directions. Critically, the L1 changes in response to learning and using an L2
- Not all bilinguals are the same. Bilinguals differ by virtue of where they live and the demands that are placed on them to use each language
parallel activation
Both languages are active regardless of the requirement to use one language alone: in reading, listening, or planning speech
evidence for parallel activation
Many studies have demonstrated that bilinguals recognize cognates (the same word in both languages with the same meaning) more quickly but homographs (the same word in both languages but different meanings) more slowly than control words. Monolinguals do not show these effects.
picture naming task
Naming images out loud
findings of picture naming task
Found cognate facilitation effects, demonstrating that lexical information is activated in the target and non-target languages
Libben & Titone, 2009 study method
tracked French-English bilinguals’ eye movements during reading.
fixation
length of the first time the eye fixates on the target
how is fixation represented?
by the size of the dot
what words elicit more fixation?
more difficult words
saccade
the transition between words
how are saccades represented?
the lines
regression
going back to revise something previously fixated on
total fixation duration
length of all eye fixations on a target
Libben & Titone, 2009 findings
- In the early stages of comprehension, there is parallel activation regardless of sentence constraint (two languages are active)
- In the late stages of comprehension, parallel activation is resolved for contexts that provide a high semantic constraint (two languages are no longer active)
Semantic Relatedness Task
Clapping if two words are semantically related (related in meaning)
Morford et al., 2011 ASL semantic relatedness findings
found that bilinguals are faster to judge English when the ASL converges and slower when it conflicts. Monolinguals don’t show these effects
Lexical Decision Task
Clapping if you see a word in x language
Bice & Kroll, 2015 study aims
examined the cognate effects in monolinguals and L2 learners of Spanish.
event-related potentials
voltage fluctuations that are time-locked to an event
Bice & Kroll, 2015 findings
- Behaviourally, there was no cognate effect for either group
- ERPs show that the native language is affected by the second language (Reduced N400 for cognates)
verbal fluency task
Say as many words that fit into a category as you can in 30 seconds
Linick et al., 2009 context and L2 learning
compared to classroom learners, immersed learners produced fewer L1 exemplars. The L1 is suppressed while living in an L2 context
Dussias & Sagarra, 2007 L2 and context
the influence of L2 on L1 is not limited to words. It’s also about grammar
bilingual experience is characterized by:
- Predominant language
- Habits of language use
- Contextual linguistics diversity
linguistic diversity of montreal
mixed French and English
linguistic diversity of ottawa
French on one side of the river and English on the other
linguistic diversity of Toronto
English with various pockets of other languages
French-English bilinguals who report greater linguistic diversity:
- Show a higher reliance on contextual clues
- Exhibit higher connectivity between regions implicated in monitoring such as the anterior cingulate cortex and the putamen
what brain regions are involved in monitoring?
the anterior cingulate cortex and the putamen
competitive contexts
L1 & L2 are compartmentalized
cooperative contexts
L1 & L2 are integrated
Beatty-Martinez & Dussias, 2017 study method
measured ERPs for bilingual speakers who use unilingual and codeswitched sentences
Beatty-Martinez & Dussias, 2017 study findings
Non-code switchers process the cost regardless of the type of code switch
For code switchers, only rare code switches elicit processing costs
processing costs depend on
the nature of the bilingual
code switches depend on
distributional regularities. Some switches are more common than others. Larger N400 for rarer code switches, demonstrating that they have more difficulty in processing
large N400 represents
difficulty processing
