Sentence processing

Question 1

Parsing:

Answer 1

Analyzing an input string of words into a syntactic structure

Question 2

Sentence processing:

Answer 2

Includes parsing, general term

Question 3

The parser/ processor

Answer 3

Mechanism in the human brain that performs speech processing functions such as parsing

Question 4

The garden path model

Answer 4

• A syntactic parsing principle
• ambiguous region causes confusion after a disambiguating region that breaks expectations follows it
• misleading sentences that lead to one expectation (the single structure), which is shown to be incorrect (requires reanalysis)
• a two stage model
• processor constructs on structure first as words come in, later has to reanalyze if needed in the second stage
• processor builds the simplest structure compatible with input, which includes late closure and minimal attachment

Question 5

Late closure

Answer 5

Attach incoming GP material within phrase you are working on
The lowest possible nonterminal node dominating the last item analyzed
In the GP model, you close off phrases earlier than you need to

Question 6

Minimal attachment

Answer 6

Attach incoming material into the phrase marker being constructed, using the fewest possible nodes (while respecting the rules of the language)
You mentally construct one thing when you hear a GP sentence, but it ends up defying the simplest solution
Minimal attachment is violated when an extra structure is added without evidence/ expectation

Question 7

How does reanalysis proceed?

Answer 7

!! Selective reanalysis
- the parser uses all available information to locate the source of error/ confusion.

Question 8

Predictions in GPT

Answer 8

Varies depending on ease of recovery/ reanalysis

Some sentences make it easier to reanalyze from the disambiguating cue

Question 9

Pragmatics

Answer 9

Branch of linguistics studying how language is used, and how it is integrated into a context (linguistic and non-linguistic)

Question 10

Presuppositions

Answer 10

Something that is being assumed, or presupposed by an utterance, rather than being explicitly stated

Even if you negate a sentence, the presupposition (assumption) remains

Question 11

Presuppositions and relative clauses

Answer 11

Restrictive relative clauses serve to pick one entity out of a set
Tends to presuppose that there are multiple of something, for example

Question 12

Processing in context slides

Answer 12

slides 55, 56

Question 13

Does the visual context affect how we process sentences in real time?

Answer 13

Yes
Eye movement is reflects our processing of a sentence

Question 14

Pragmatic context

Answer 14

Whether presuppositions are supported by the current discourse

Question 15

Visual context

Answer 15

The visual scene can support the use of a modifier like “on the towel” in the eye tracking task.

Question 16

Can the frequency with which a word appears in a certain
construction influence processing?

Answer 16

Frequency as determined though a corpus search or a
completion task.

Question 17

Lexical preferences, parsing, and plausability

Answer 17

• The bias of a verb toward an NP or clausal (S) object matters in sentence parsing.
• Verbs that are biased toward a clausal object show less evidence of a garden path effect.
• Plausibility of an NP as an NP direct object did not influence parsing significantly (in this study!)

Question 18

Reduced Relative clause ambiguities

Answer 18

skip

Question 19

Digging-in effects

Answer 19

When multiple constraints conspire, the disfavored interpretation is inhibited

After the Martians invaded the town was evacuated.
VS
After the martians invaded the town that the city bordered was evacuated.

Question 20

Constrained-based accounts:

Answer 20

Model that proposes:
- that multiple structures/ representations are built in parallel and weighted over time
- all sources of information are used simultaneously
- strongly interactive models

Question 21

Problems for constraint-based models

Answer 21

• Constraint-based models typically predict longer processing times for balanced ambiguous regions.
• Typically these are not found in experiments of sentence processing (See Staub & Clifton, 2008 for a review).
  »> BUT cf. Green & Mitchell (2006) - how often are ambiguities truly balanced?

Question 22

Hybrid accounts

Answer 22

Alternative theory for processing ambiguous sentences, alternative to the garden path model and constraint-based model