Context Accommodation in Human Language

Download Report

Transcript Context Accommodation in Human Language

Context Accommodation in
Human Language Processing
June 2010
Jerry T. Ball
Senior Research Psychologist
711th HPW / RHAC
Air Force Research Laboratory
DISTRIBUTION A. Approved for public release; distribution unlimited.
Theoretical Alignment
•
Cognitive Linguistics
– No autonomous syntax
– Grammatical categories are semantically motivated
•
Construction Grammar
– Constructions at multiple levels of idiomaticity
– No sharp distinction between lexicon and syntax
•
X-Bar Theory
– Prior to introduction of functional heads
•
Simpler Syntax
– Flat syntax trees
Distribution A
2
Theoretical Foundations
Language Representation and Processing
•
Double R Grammar
– Cognitive Linguistic theory of the grammatical
encoding of referential and relational meaning
•
Double R Process
– Psycholinguistic theory of the processing of
English text into Double R Grammar based
representations
•
Double R Model
– Computational implementation using the ACT-R
cognitive architecture and modeling environment
www.DoubleRTheory.com
Distribution A
3
Research Goals
Develop models of Human Language Processing…
•
Functional
– Use in real applications
– Synthetic Teammate prototype
•
Cognitively Plausible
– Adhere to well-established cognitive constraints
– Don’t use computational techniques that are not
cognitively plausible
• tokenize  tag part of speech  syntax
processing  semantic processing  …
Distribution A
4
Constraints on Human Language
Processing
•
Visual World Paradigm (Tanenhaus et al. 1995)
– Subjects presented with a visual scene
– Subjects listen to auditory linguistic input describing scene
•
Immediate determination of meaning
– Subjects look immediately at referents of linguistic
expressions, sometimes before end of expression
•
Incremental processing
“the green…”
•
Interactive processing (Trueswell et al. 1999)
– Ambiguous expressions are processed consistent with scene
“put the arrow on the paper into the box”
Distribution A
5
Cognitively Plausible Mechanism
•
Pseudo-deterministic, serial processing mechanism
with context accommodation operating over a parallel,
probabilistic constraint mechanism
– Parallel, probabilistic constraint mechanism
proposes best alternative given current context
– Processor proceeds as though it were serial and
deterministic, but accommodates the subsequent
input as needed
– Integrates the advantages of parallel processing
with an essentially serial processing mechanism
Distribution A
6
Pseudo-Deterministic HLP
•
Presents the appearance and efficiency of a serial,
deterministic processor, but…
•
Relies on parallel, probabilistic constraint
mechanism for making the best choice at each
choice point
•
Relies on non-monotonic context accommodation
mechanism to make modest adjustments to the
evolving representation given the current context
•
•
•
Limited lookahead, delay and underspecification
Limited parallelism within serial mechanism
No backtracking
Distribution A
7
Context Accommodation
• If current input is unexpected given the prior
context, then accommodate the input
– Adjust the representation
• Override
• Block
• Function Shift
– Coerce the input into the representation
• Head of nominal need not be a noun!
• Head of clause need not be a verb!
Distribution A
8
Context Accommodation
• Related to Lewis’s notion of Limited Repair Parsing
– “The putative theoretical advantage of repair
parsers depends in large part on finding simple
candidate repair operations” (Lewis, 1998)
– “Lightweight” repair – may be no additional cost
relative to processing without accommodation
•
Part and Parcel of normal processing – not
reanalysis
•
Non-monotonic
Distribution A
9
Types of Accommodation
• Coercion
– “the running of the bull” – head of nominal
• “running” construed objectively, arguments
not expressed (“of the bull” functions as a
modifier)
– “a Bin Laden supporter”
• Proper Noun functions as modifier
– “the newspaper boy porched the newspaper” –
nonce expression (H. Clark 1983)
• “porched” construed as transitive action
Distribution A
10
Types of Accommodation
• Override
– Single word vs. Multi-Word Expression (MWE)
• “kicked…”  transitive verb
– “kicked the bucket”  idiomatic expression
• “take…”  transitive verb
– “take a hike” “take five” “take time” “take
place” “take out” “take my wife, please”
“take a long walk off a short pier” …  many
idiomatic expressions
• Not possible to carry all forward in parallel
Distribution A
11
Types of Accommodation
• Grammatical Function Shift
– “he gave it to me”
• direct object (initial preference due to inanimacy)
– “he gave it the ball”
• direct object (initial preference)  indirect object
– “he gave her the ball”
• indirect object (initial preference due to animacy)
– “he gave her to the groom”
• indirect object (initial preference)  direct object
Distribution A
12
Types of Accommodation
• Grammatical Function Shift
– “he said that…”
• In context of “said”, “that” typically functions as a
complementizer
– But subsequent context can cause a function shift from
• complementizer
– “he said that she was happy”
• To nominal specifier to
– “he said that book was funny”
• To nominal head
– “he said that.”
Distribution A
13
Types of Accommodation
• Grammatical Function Shift
– “pressure” vs.
“pressure valve” vs.
“pressure valve adjustment” vs.
“pressure valve adjustment screw” vs.
“pressure valve adjustment screw fastener” vs.
“pressure valve adjustment screw fastener part” vs.
“pressure valve adjustment screw fastener part number”
• Serial nouns (and verbs) incrementally shift from head to
modifier function as each new head is processed
• Functions like lookahead, but isn’t limited
Distribution A
14
Types of Accommodation
• Modulated Projection
– “the rice” vs. “rice”
– “the” projects a nominal and functions as a specifier
– In the context of “the” “rice” projects a head which functions
as the head of the nominal
– When there is no specifier, “rice” projects a nominal as well as
a nominal head
Nominal
Nominal
+
spec
head
rice
the
“the rice”
Distribution A
head
vs.
head
rice
“rice”
15
Types of Accommodation
• Grammatical Feature Blocking and Overriding
– “A few books” = indefinite + plural
override
• A (sing, indef) few (plural) books (plural, indef)
– “The books” = definite + plural
block
• The (def) books (plural, indef)
Distribution A
16
Types of Accommodation
• Grammatical Feature Blocking and Unsetting
– “He has given me the book” = active + perfect
block
• He has (active) given (passive, perfect) me the
book
– “He has been given the book” = passive + perfect
unset
• He has (active) been (de-act, perfect) given
(passive, perfect) the book
Distribution A
17
Summary of
Context Accommodation
• Context Accommodation is part and parcel of normal
processing
– Non-monotonic
– Not reanalysis
•
Processor proceeds as though it were deterministic, but
accommodates the input as needed
•
Gives the appearance of parallel processing in a serial,
pseudo-deterministic mechanism
Distribution A
18
Computational Implementation
no airspeed or altitude restrictions 
no
negative
“no” projects an object referring expression and functions as
the specifier of the object referring expression
Distribution A
19
Computational Implementation
no airspeed or altitude restrictions 
airspeed
singular
inanimate
“airspeed” is integrated as head of the object referring expression
projected by “no” – in parallel an object-head is projected to support
a more complex object referring expression
Distribution A
20
Computational Implementation
no airspeed or altitude restrictions 
function overriding
The processing of “or” is delayed until the word after “or” is processed.
In the context of “or” and “airspeed”, “altitude” is conjoined with “airspeed”
into a conjoined object head.
The conjoined object-head overrides the previous head.
Distribution A
21
Computational Implementation
no airspeed or altitude restrictions 
plural
(feature overriding)
function shifting
In the context of an object head, the previous head is shifted into
a modifier function so that “restrictions” can function as the head
Distribution A
22
Computational Implementation
his… book
Two referring expressions
projected!
distinct bind indexes
his
his projects poss-obj-spec and higher level obj-refer-expr
his is reference point for higher level obj-refer-expr
Distribution A
23
Computational Implementation
his book…
book
book integrated as head of higher level obj-refer-expr
Distribution A
24
Computational Implementation
her… books
Only one referring expression
projected!
her
Distribution A
Compare –
I like his
I like her
her projects obj-refer-expr (not poss-obj-spec)
25
Computational Implementation
her books…
Second referring expression
is projected!
books
books integrated as head of higher level obj-refer-expr
Distribution A
26
Computational Implementation
hers…
Head of higher level object referring expression is implied
Number of higher level referring expression is unspecified!
hers is nice
hers are nice
implied
hers
Distribution A
27
Conclusions
•
Context accommodation, combined with
parallel, probabilistic selection of
alternatives, makes a serial, deterministic
processor feasible
•
HLP is Pseudo-Deterministic
– Serial, deterministic (incremental), but…
– Non-monotonic (context accommodation)
– Depends on highly context sensitive,
parallel, probabilistic constraint
mechanism (interactive)
Distribution A
28
Questions?
29
Questions?
Ball, J. (2007a). A Bi-Polar Theory of Nominal and Clause Structure and Function.
Annual Review of Cognitive Linguistics.
Ball, J. (2007b). Construction-Driven Language Processing. Proceedings of the 2nd
European Cognitive Science Conference.
Ball, J., Heiberg, A. & Silber, R. (2007). Toward a Large-Scale Model of Language
Comprehension in ACT-R 6. Proceedings of the 8th International Conference on
Cognitive Modeling.
Heiberg, A., Harris, J. & Ball, J. (2007). Dynamic Visualization of ACT-R Declarative
Memory Structure. Proceedings of the 8th International Conference on Cognitive
Modeling.
Distribution A
30
References
Bever, T. (1970). The cognitive basis for linguistic structures. In J.R. Hayes (ed.),
Cognition and Language Development, 277-360. New York: Wiley.
Biber, D., Conrad, S. & Leech, G. (2002). Student Grammar of Spoken and Written
English. Essex, UK: Pearson Education Limited.
Christianson et al. (2001). Thematic roles assigned along the garden path linger.
Cognitive Psychology, 42, 368-407.
Clark, H. (1983). Making sense of nonce sense. In G. Flores d’Arcais & R. Jarvella (Eds.),
The Process of Language Understanding, 297-331. New York: John Wiley.
Crocker, M. (1999). Mechanisms for Sentence Processing. Garrod & Pickering (eds.),
Language Processing, London: Psychology Press.
Culicover, P. & Jackendoff, R. (2005). Simpler Syntax. Oxford: Oxford University Press.
Gibson, E. (1991). A computational theory of human linguistic processing: Memory
limitations and processing breakdown. Unpublished doctoral dissertation, Carnegie
Mellon University, Pittsburgh, PA.
Gibson, E. & Pearlmutter, N. (2000). Distinguishing Serial and Parallel Parsing. Journal of
Psycholinguistic Research, 29, 231-240.
Distribution A
31
References
Henderson, J. (2004). Lookahead in Deterministic Left-Corner Parsing. Proceedings of
the Workshop on Incremental Parsing: Bringing Engineering and Cognition Together.
Barcelona, Spain.
Huddleston, R. & Pullum G. (2005). A Student’s Introduction to English Grammar. NY:
Cambridge University Press.
Huddleston, R. & Pullum G. (2002). The Cambridge Grammar of the English Language.
NY: Cambridge Unversity Press.
Lewis, R. (1998). Reanalysis and Limited Repair Parsing: Leaping off the Garden Path. In
Fodor, J. & Ferreira, F. (eds). Reanalysis in Sentence Processing. Boston: Kluwer
Academic.
Lewis, R. (2000). Falsifying serial and parallel parsing models: Empirical conundrums
and an overlooked paradigm. Journal of Psycholinguistic Research, 29, 241-248.
Marcus, M. (1980). A Theory of Syntactic Recognition for Natural Language. Cambridge,
MA: The MIT Press.
MacDonald, Pearlmutter & Seidenberg (1994). The lexical nature of syntactic
ambiguity resolution. Psychological Review, 101, 676–703.
Distribution A
32
References
O’Grady, William (2005). Syntactic Carpentry, an Emergentist Approach to Syntax.
Mahway, NJ: LEA.
Quirk, R., Greenbaum, S., Leech, G. & Svartvik, J. (1985). A Comprehensive Grammar of
the English Language. Essex, UK: Pearson Education Limited.
Trueswell, J. & Tanenhaus, M. (1994). Toward a lexicalist framework for constraint-based
syntactic ambiguity resolution. In C. Clifton, K. Rayner & L. Frazier (Eds.), Perspectives
on sentence processing, pp. 155-180. Hillsdale, NJ: LEA.
Tanenhaus, M., Spivey-Knowlton, M. Eberhard, K. & Sedivy, J. (1995). Integration of
visual and linguistic information in spoken language comprehension. Science, 268,
632-634.
Trueswell, J. Sekering, I., Hill, N. & Logrip, M. (1999). The kindergarten path effect:
studying on-line sentence processing in young children. Cognition, 73, 89-134.
Distribution A
33