1/136
Looks like no tags are added yet.
What is linguistic competence?
A speaker-hearer's knowledge of their language.
What does competence allow us to do?
Make grammaticality judgements, even for sentences we have never heard before.
What is linguistic performance?
The actual use of language in concrete situations.
Why might performance differ from competence?
Because of memory limitations, hesitations, errors, distractions, and other processing constraints.
Which aspect of language is primarily of interest to linguists?
Competence.
Which aspect of language is primarily of interest to psycholinguists?
Performance.
Grammaticality judgements
Not learned exactly but implicit knowledge is applicable to new grammatical scenarios
Sensicality judgements
whether something makes sense - and you understand that it does/doesn't
Rules = finite → sentence number = infinite (rule governed creativity)
What is a phrase structure tree (tree diagram)
A diagram that represents the hierarchical grouping of words into syntactic categories within a sentence.
What do phrase structure trees show?
How sentence constituents (words and phrases) relate to one another hierarchically.
Do phrase structure trees imply that speakers mentally construct these exact trees during language processing?
No. They are theoretical representations, not literal mental structures.
What kind of structure do people generate during parsing
A representation of "who did what to whom" in the sentence.
What is a constituent?
A word or group of words that functions as a single unit within a sentence.
What does NP stand for
Noun Phrase
What does VP stand for
Verb Phrase.
What does S stand for in a phrase structure tree?
Sentence
Parsing
making sense of sentence - which way to break it up/ interpret it
What is the question of encapsulation in parsing?
Whether different sources of knowledge (e.g., syntax, semantics, discourse) operate separately or interact during sentence processing.
What do modular accounts of parsing propose
That syntactic information is processed first and independently of other information.
What is serial processing
A step-by-step process in which syntax is analysed first, followed by semantic and other information.
Why are modular models called two-stage models?
Because they first process syntax alone, then incorporate additional information such as semantics.
What do interactive accounts of parsing propose?
That syntactic, semantic, and other sources of information are used simultaneously during processing.
What is interactive processing?
The idea that all available information is considered at the same time during sentence comprehension.
How do researchers test parsing models?
By using sentences that can be misinterpreted (temporarily ambiguous sentences) and measuring processing time.
What does a longer processing time for one interpretation suggest?
that a different structure was initially built and then had to be revised.
What is the Garden-Path Model?
A modular, two-stage model of sentence parsing proposed by Frazier (1987).
What does Stage 1 of the Garden-Path Model involve?
Parsing based solely on syntactic information, without considering meaning or context.
What is Minimal Attachment?
The principle that the parser chooses the simplest syntactic structure with the fewest nodes.
What is Late Closure?
The principle that new words are attached to the phrase or clause currently being processed whenever possible.
If Minimal Attachment and Late Closure conflict, which takes precedence?
Minimal Attachment.
What does Stage 2 of the Garden-Path Model involve?
Reanalysis of the sentence if the initial parse is incompatible with semantic, thematic, or later syntactic information.
What is reanalysis?
Revising the initial interpretation of a sentence when later information shows it is incorrect.
What are constraint-based models of sentence processing?
Interactive models in which all relevant sources of information can influence parsing immediately
What are "constraints" in constraint-based models?
Sources of information that help guide parsing, such as syntax, semantics, discourse, and frequency information.
What types of information can be used immediately in constraint-based models?
Semantic, syntactic, discourse, and frequency information.
How do interactive accounts differ from modular accounts?
Interactive accounts use multiple sources of information simultaneously, whereas modular accounts prioritise syntax first.
How are syntactic analyses processed in constraint-based models?
Multiple possible syntactic analyses are generated in parallel.
What determines the activation of a particular syntactic analysis?
The amount of support available from relevant constraints at that moment.
When is sentence processing easy according to constraint-based models?
When one analysis is strongly supported over the alternatives.
When is sentence processing difficult according to constraint-based models?
When multiple analyses receive similar levels of support and compete with one another.
What is meant by incremental processing in reading?
Each incoming word is processed almost immediately as it is encountered.
What is a key challenge for semantic processing?
Determining the correct interpretation of a word within its context.
Why can word meaning be difficult to determine?
Because there can be multiple mappings between words and concepts.
What question do researchers ask about semantic activation?
What semantic information becomes activated when a word is read.
Why is semantic processing more complex than syntactic parsing?
Because there are potentially vast numbers of possible meanings and interpretations.
What example does Pinker use to illustrate the complexity of semantic processing?
10-word sentence with 20 choices per position could generate approximately 10201020 combinations.
Why must the semantic processor be flexible
To cope with the huge variety of possible inputs and interpretations.
Why must semantic processing be quick?
Because readers and listeners interpret language rapidly as it unfolds.
What is lexical ambiguity?
When a word has more than one possible meaning.
What is a homonym?
A word with two unrelated meanings or interpretations.
What does the selective access model propose?
Context restricts access to only the contextually appropriate meaning.
What does the ordered access model propose?
Meanings are accessed according to their frequency, then checked against context.
What does the parallel access model propose?
All meanings of an ambiguous word are activated simultaneously.
What is the key difference between selective and parallel access models?
Selective access activates only the context-appropriate meaning, whereas parallel access activates all meanings.
What did older experiments on lexical ambiguity find?
Some evidence for multiple (parallel) access and some evidence for selective access.
Why were findings from older lexical ambiguity experiments inconsistent?
results depended heavily on the type of offline task used.
What method did Swinney (1979) use to study lexical ambiguity?
Cross-modal priming.
What is cross-modal priming?
A technique where participants hear a sentence and respond to visually presented target words to measure which meanings are activated.
Which ambiguous word was used in Swinney's experiment?
"Bug" (insect vs spying device).
what target words were used in Swinney's study?
ANT (related to insect meaning), SPY (related to surveillance meaning), and SEW (unrelated control).
What was found at Time Point 1?
Facilitation occurred for both ANT and SPY, showing both meanings were activated.
What was found at Time Point 2?
facilitation occurred only for ANT, the contextually appropriate meaning.
What did Swinney's results suggest about lexical ambiguity resolution?
All meanings are initially activated, then context selects the appropriate meaning.
Which model of lexical access is supported by Swinney (1979)?
Parallel access (multiple meanings activated simultaneously).
What type of homonyms were investigated by Rayner & Duffy (1986)?
Biased homonyms and balanced homonyms.
What is a biased homonym?
A homonym with one meaning that is much more common than the other.
hat is a balanced homonym?
what is a balanced homonym?
What example of a biased homonym was used in the study
port (more commonly a drink than a harbour).
What example of a balanced homonym was used in the study
Coach (e.g., bus/carriage vs instructor).
What did the researchers measure
Eye movements and reading times during sentence comprehension.
What happened when readers encountered biased homonyms
The dominant (more frequent) meaning was activated most strongly.
What happened when readers encountered balanced homonyms?
Both meanings competed because they had similar frequencies.
Why were balanced homonyms harder to process?
Because multiple equally probable meanings competed for interpretation.
What did Rayner & Duffy's findings suggest about lexical access?
Word meaning frequency influences which interpretations are activated.
What two factors are important in newer research on lexical ambiguity
Meaning frequency and context.
What is the Reordered Access Model?
A hybrid model combining aspects of exhaustive (parallel) and selective access.
Who proposed the Reordered Access Model?
Duffy et al. (1988).
What is a "contextual boost"?
Prior context increases the activation level of a contextually relevant meaning.
Does context completely determine meaning activation?
No. Context boosts relevant meanings, but meaning frequency also influences activation.
How does meaning frequency affect lexical access?
More frequent meanings start with higher activation levels and are accessed more easily.
What does the Reordered Access Model predict for biased homonyms?
The dominant meaning is usually accessed first unless context strongly boosts the subordinate meaning.
What is the subordinate bias effect?
lower processing when a biased homonym must be interpreted using its less frequent meaning.
Lexical polysemy
Words with multiple senses - two different interpretations that are still semantically related
Metonymy
most noticeable property of an idea is used to describe the whole entity - watering down a complex concept to make a more simple sentence - literal vs figurative
Unranked parallel
- no order/hierarchy when activating words
Ranked parallel
activation depends on how frequently word is used
If figurative sense of word/concept is known to pps it is recognised as fast as literal is
Why is a simple unranked parallel model problematic?
Because many words have a large number of related senses, making activation of every meaning inefficient.
What is the alternative to activating all possible meanings?
Activation of a single, abstract underspecified meaning.
What is an underspecified meaning?
A general semantic representation shared by all related senses of a word.
Does the underspecified meaning model involve immediate selection of a specific sense?
No. A specific sense is selected later.
What happens after an underspecified meaning is activated?
A context-driven homing-in stage identifies the appropriate specific meaning.
What role does context play in the underspecified meaning model?
Context helps narrow down the correct interpretation rather than acting as a judge between competing meanings.
Why does the underspecified meaning model reduce errors
Because the wrong sense is rarely assigned initially.
Why is sense assignment considered efficient in this model?
Because it is a simple process that can occur very quickly.
How does the model explain figurative language processing?
All related senses are processed the same way initially, so figurative language requires no special mechanism.
What is pragmatics?
The study of the difference between the literal meaning of words and what a speaker actually intends to communicate.
What is the principle of compositionality?
Sentence meaning comes from word meanings and how they are combined.
What is coherence?
The expectation that a text forms a meaningful whole.
What is cohesion?
Consistent links between ideas, characters, objects, or events across a text.