Lecture 5- Speech processing and higher level sentence processing Flashcards
What is parsing?
A deep structure based on the surface structure (e.g. The words you’re u have identified and the order they occur in)
What is a parasodic cue
Putting stress on different words creates different meanings
Guides the reader to correct info
This happens only in spoken Lang- not written
How does a parser work
Meaning of sentanses can only be fully understood after the syntax of a sentence is parsed
How does the language acquisition devise link to parsers
Linguists imagine there is a specific module for how the brain generates syntactic integration - parsers
These are closely related to the LAD
LAD sets parameters for the parser
- so parser can use the grammar as grammar is universal
What is meant by universal grammar
Limited amount of grammar is available to us
Briefly explain the garden path model - Frazier and rayner 1982
Just one possibility considered based on grammatical rules only- sentence has to be repartee (re read) if wrong
Briefly explain the constraint satisfaction model McDonald 1994
Multiple possibilities considered- not all activated equally (all available invo taken into account to activate possibilities)
Semantic properties and frequencies of verbs in syntactic structures are taken into account
Briefly explain the unrestricted race model - van gompel 2000
Similar to garden path
But structures are considered based on both semantic and syntactic evidence
In general- structure which is easier to model wins and is preferred
Briefly explain the good enough processing model Ferreira 2002
Compatible with any model but says that result of parsing isn’t always well elaborated of correct
How does the garden path model work
The parser chooses one preferred interpretation- these are dependant on the syntactic rules : minimal attachment and late closure
Describe the 2 syntactic rules of the garden path mode
Late closure : what happens in sentences
- always want to end the main clause unless you can’t 1 while building your deep structure you don’t want to end a clause
Minimal atttchmemt: attaching incoming material into the phrase marker being constructed using the fewest nodes consistent with the well- formedness of language
- aka- should build/ attatch incoming material with the fewest nodes (words/ verb/ adverb) as possible
what happened in Allopenna, Magnuson & Tanenhaus (1998) study
Eye tracking: Where do participants look as you are telling them:
“Pick up the…
a) beaker
b) beetle
c) speaker
d) carriage
- the COHORT model would not predict any fixations on the speaker
The TRACE model does, and it even gets the timing right!
what I parsing and what aids this ?
Parsing: generating a deep structure based on the surface structure (i.e. the words you have identified and the order in which they occurred).
In spoken language, prosodic cues may help.
what is a parser and how is it related to linguisties
Linguists imagine that there is a specific module in the brain doing this – they call it the parser. This is very closely related to the idea of a Language Acquisition Device. The LAD essentially sets the parameters for the parser.
how does a parser work
The meaning of a sentence can only be fully understood after the syntax of a sentence is parsed
But does that mean that semantic information cannot influence the parser?
How does the parser deal with ambiguous information?
which model was proposed by Frazier and Raynrs 1982 study
Garden path setances for syntactic ambiguity
what are the strengths of the garde path model
Fits a lot of evidence, e.g. from eye tracking
No evidence for the competing theories at the time:
Parallel parsing (all alternatives are considered at the same time)
Minimal commitment (the sentence structure isn’t built until you get to the end of the sentence)
A lot of the time, the principles of minimal attachment and late closure are applied even if they don’t make sense semantically:
Readers take longer to read “After the child had sneezed the doctor prescribed a course of injections.” (van Gompel & Pickering, 2001)
Readers slow down at “the doctor” because they want to attach it as a direct object to “sneezed”.
But they have to stop themselves since you can’t sneeze someone!
what are the weaknesses of the garden path model
Listeners shouldn’t take context into account
Spivey, Tanenhaus, Eberhard, & Sedivy (2002)
Task: Look at an array of objects while listening to sentences referring to those objects
what did Spivey, Tanenhaus, Eberhard, & Sedivy (2002) research
the garden path model
-Task: Look at an array of objects while listening to sentences referring to those objects
what should pps look at in Spivey, Tanenhaus, Eberhard, & Sedivy (2002) study and what shouldn’t they
’ put the apple on the towel in the box’
- should look at apple (referant) then box (goal)
but if look at theapple on the towel as one phrase look at the apple and the towel in the experi
What are the results from Spivey, Tanenhaus, Eberhard, & Sedivy (2002) study
while building what listening to - take visual context into account
- Listeners take the context into account:
If there is only one referent (apple), they attach “on the towel” as the goal
Lots of looks at the towel
If there is more than one referent (one apple on a towel and one apple on a napkin) they attach “on the towel” to the referent
Very few looks at the towel
Briefly explain the constraint satisfaction model
- connectionist model
- version of the parallel approach
- considers many activation’s
- more/ less likely to be interpretated = more / less activation
more activation takes more time (less favoured??) - parser considers both semantic and syntactic info straight away
- most likely inter gets most activation (but alternative interps cause inhibition which lowers activation)
- if the most likely interp fails- results in shift in activation which takes time to resolve
what are the strengths of the constraint satisfaction model
- can explain the effect of semantic prosessing
- assumes some verbs are based towards specific syntactic structures- eg read/read
which of these sentances is easier to read and why
The professor read the newspaper had been
destroyed.
B. The professor read the newspaper during his break
.
B easier than a
because varies as 2nd focusses on newspaper as direct object due to subordinate clause
what is the word believe normally followed by
a subordiante clause
- which makes it easier to understand
which of these sentances are easier to understand
A. The professor read the newspaper had been
destroyed.
B. The professor read the newspaper during his break.
C. The professor believed the newspaper had been
destroyed.
D. The professor believed the newspaper
C easier than A
as reader bias towards subordinate clause
believe doesnt need a comma
describe altman and Kamides 1999 study researching the constraint satisfaction model
- eye tracking study
- uses a visual world paradigm- looking at where people look when listening to a sentance
- pps head ‘ the boy will move the cake’ or the boy will eat the cake
what was anticipated from altman and Kamides 1999 study
will pps anticipate cake as the semantically appropriate object for the word eat
at what point in the sentance will they look at the cake- when said the word eat of the word cake
any of the objects are appropriate for the control word move so can look at any
What were the results from the visual world paradigm in altman and Kamides 1999 study
- pps look at the word cake quicker in the eat study than move
- shows even before the word cake - they look at cake - usually starts at the word eat (determinant)
this shows semantics play a part ( as we have prior knowledge of the word eat relating to food ie cake) - so more likely to look at a cake than a bowling ball
What are the weaknesses of the constraint satisfaction model
- too general in its predictions
(obvs semantics and context play a part - but how do they interact) - maybe a garden path model with a fast semantic processing step after the syntactic processing can account for the results too
what are the 3 types of sentances suggested by Van Gompel in the unrestricted race model study ?
Globally ambiguous sentences: The burglar stabbed
only the guy with the dagger during the night.
• Meaning is not constrained by semantics: Both verb-phrase
attachment (the burglar uses the dagger) and noun-phrase
attachment (the guy has the dagger) are plausible.
B. Verb-phrase attachment: The burglar stabbed only the
dog with the dagger during the night.
• Syntactically ambiguous but disambiguated by semantics: dogs
don’t have daggers.
C. Noun-phrase attachment: The burglar stabbed only
the dog with the collar during the night.
• Syntactically ambiguous but disambiguated by semantics: you
can’t use collars to stab.
pps had to read these aloud
describe the unrestricted race model
suggested by Van Gompel 2000
unrestricted in terms of :
- syntactic structure and semantic info into account at the same time (like in the constraint satis)
initially all interpretations were built in parralel (race between poss interps)
- once one interp wins race- the other interps are ignored
if the preferred interp has to be abandoned this will trigger an extensive reevaluation of the sentence
what predictions would be made using the evidence by Van Gopels study - when considering the Garden path model, constraint satis model and the unrestricted race model ?
Garden-path model:
• All sentences are globally ambiguous syntactically, so
they should take the same time to process
– Constraint-satisfaction model: • In the sentences that are semantically constrained (B and C), it should be easier to choose the preferred interpretation, so they should be read faster than the unconstrained sentence (A)
– Unrestricted race model:
• Sentence A should be fastest since either of the two
interpretations can be picked
• Sentences B and C should be slower since people might
occasionally have to re-analyse these sentences if they
picked the wrong analysis initially
what are teh strengths and weaknesses of the Unrestiricted race model
Strengths • Combines the best elements of the gardenpath and the constraintsatisfaction model • Explains a lot of the experimental evidence so far Weaknesses • Relatively new, not all predictions have been tested yet.
what did van gompels model find
• The evidence (from van Gompel et al., 2001) matches the predictions of the Unrestricted Race Model
Sentence A should be fastest since either of the two
interpretations can be picked
• Sentences B and C should be slower since people might
occasionally have to re-analyse these sentences if they
picked the wrong analysis initially
what is Good enough processing model
- this fits with any model
assumes u dont jhave to process sentances to the extent which you should to be able to get an indepth analysis
Sentences aren’t always parsed to the point
where a fully-formed and correct
interpretation is available
• Processing just happens until the result is
good enough to get on with the task at hand
– Strategies for reducing working memory load
include the use of heuristics (e.g. the noun that
comes after a verb is usually the direct object of
that verb)
– Listeners and readers may adapt their
processing strategy based on task demands
what is the model about discourse processing called ?
standard pragmatics model
what is the standard pragmatics model
- working syntactic representation of sentences and meanings of sentances inside the sentance -> what the speaker is trying to say (pragmatics)
- pragmatics relate to the intended rather than literal meaning
eg figurative language (lang not suppose dto be taken literally ) and metaphors (words used figuratively describer what something means/ resembles)
eg its raining cats and dogs
what are grices 3 stages in the standard pragmatics model
The literal meaning is accessed (assessment)
– The reader or listener decides whether the
literal meaning makes sense in the context in
which it is read or heard ( context evaluated)
– If it doesn’t, then the reader or listener searches
for a nonliteral meaning of the sentence that
does make sense in context ( if doesnt make sense to context - find another meaning which does)
what is the co-operative principle according to grice 1975
- pps in a conversation agree to cooperate
- by saying things appropriate to convo
- and not ending the convo unilaterally
- violations of this principle can be informative too
#
give examples of violations to the cooperative principle- Grice 1975
“Be relevant”
– Joe asks Sue “How was your day?”
– Sue says “The train was late.”
• Either Sue’s reply is irrelevant (non sequitur) OR
• The train’s lateness is relevant to Sue’s day
• “Be as informative as required—and no more“
– Joe asks Sue “Where are the car keys?”
– Sue says “Somewhere” OR
– Sue says “See that blue couch over there with the
dirty spot on the arm of it. If you go over there and
look next to it there is a wooden side table. On the
side table is a small green bowl with a bunch of
stuff in it. The keys are one of the things in that
bowl.”
Tell the truth”
– Joe asks Sue “Could you tell me what time it is?”
– Sue says “No” (even though she is wearing a watch)
• “Be clear”
– Joe says to Sue “The project isn’t progressing very
well, could you fix it?”
– Joe doesn’t give Sue any useful suggestions.
• Notice that all these violations carry an implicit
(often aggressive) pragmatic meaning that is
not contained in the sentence structure!
what is meant by common ground- grice 1975
people generally tend to conform to the cooperativeness principle
- speakers listen to and work together to ensure mutual understanding by operating under shared knowlegde - CG is the info and knowledge they share
what can failure to speak under common ground lead to
Miscommunication
what does Keysar et al 2000 suggest about common ground
• Keysar et al. (2000) argued that keeping on common
ground can be effortful, instead:
– Listeners use a rapid and non-effortful egocentric heuristic
• Consider ungrounded objects from one’s own perspective as
potential referents
what is the strengths and limitations of common ground and the egocentric heuristics
Progress
• The distinction between common ground and egocentric heuristic accounts is oversimplified
– General expectation for use of common ground and cooperativeness principle
• May be better explained by an expectation of consistency
– Processing limitations sometimes prevent listeners from focusing on only the common ground
Limitations
• Many experimental situations are highly artificial and not ecologically valid
• Use of common ground is more likely with people we know well than strangers in the lab
• Processing limitations may put a cap on use of common ground
