Docstoc

Lost on the garden path Exploring misinterpretation and “good

Document Sample
Lost on the garden path Exploring misinterpretation and “good Powered By Docstoc
					  Lost on the garden path:
 Exploring misinterpretation
and “good enough” language
         processing
            Kiel Christianson
    Dept. of Educational Psychology
          & Beckman Institute
                Collaborators
   Fernanda Ferreira
   Carrick Williams
   Andrew Hollingworth
   Rose Zacks
   Tim Slattery
   Susan Garnsey
   Laura Matzen
   RAs in my lab (Kent Lee, Jeong Ah Shin, Ji Kim,
    Jung Hyun Lim, Heeyoun Cho)
     So we don’t get lost ourselves,
             a brief map
1.   What are garden path sentences?
        And why are they interesting?
2.   Why worry about interpretation?
        And why haven’t other psycholinguists until recently?
3.   Basic data
        Christianson, Hollingworth, Halliwell, & Ferreira (2001)
        Christianson, Williams, Zacks, & Ferreira (in press)
4.   Recent data
        Christianson & Slattery (2005, in prep)
        Christianson (still running!)
5.   Some semblance of a conclusion, I hope…
        A working definition of “good enough”
        Parsing, processing, and interpretation
        Implications
        What relevance to SLA?
   Theoretical: Do L2 speakers parse L2
    same as L1 speakers do?

   Pedagogical: Misinterpretations can be
    informative wrt mental representations
    – You don’t know for sure unless you ask!
        Garden path sentences
   Sentences that lead the human sentence
    processor (HSP) to construct an initial
    syntactic structure, which turns out to be
    incorrect, and thus requires syntactic (and
    semantic) reanalysis.
        Example


While
While Anna
While Anna dressed
While Anna dressed the
While Anna dressed the baby
While Anna dressed the baby spit
While Anna dressed the baby spit up
While Anna dressed the baby spit up on
While Anna dressed the baby spit up on the
While Anna dressed the baby spit up on the
 bed.
While Anna dressed the baby spit up on the bed.

 the baby = ambiguous noun phrase (ambiguous
  region)
 spit up = disambiguating verb (disambiguating
  region)
    Why use sentences like this?
   They induce difficulty and observable
    slow-downs in processing that is normally
    smooth and fast
    – Point is to observe how the system is
      perturbed, and how it recovers


   Not all suffer from “mistakes”
    – Put the book on the shelf in my backpack.
        Frazier & Rayner (1982)
   The “garden path theory” of syntactic parsing
    – Eye-tracking used to measure how people read such
      sentences
    – Predictable patterns:
        Longer fixations (reading times) on disambiguating verb
        Regressive eye movements to ambiguous NP and
         subordinate verb (dressed)
   Serial, modular model
    – one parse at a time, just syntax first
    – (But this architecture isn’t crucial for assumptions that
      follow.)
         Traditional assumptions
     (no matter what parsing model)
    Garden path sentences can be handled
     one of two ways

1.   Mis-parse is recognized by the HSP,
     revision is undertaken; if not successful,
     processor gives up and interpretation is
     not achieved
2.   Ambiguity/mis-parse isn’t noticed at all;
     person just keeps reading
Questioning traditional assumptions
 Does the mis-parse HAVE to be reanalyzed
  syntactically?
 Does the interpretation HAVE to be revised?
    – Automatic?
    – MacDonald et al. (1994): There might be situations
      in which “the communicative goals of the listener can
      be achieved with only a partial analysis of a sentence,
      but we view these as degenerate cases” (p. 686).
        (An assumption made by proponents of both serial and
         parallel models of parsing)
               “Good enough”
             sentence processing
   Ferreira & Henderson (1999); Christianson, et al
    (2001); Ferreira, Christianson, & Hollingworth
    (2001); Ferreira, Bailey, & Ferraro (2003);
    Christianson, et al (in press)

   Loosely defined as processing in which the HSP
    settles for a parse that is in some way
    incomplete or underspecified, resulting in an
    interpretation that is not faithful to the input.
So why worry about interpretation?
“The central problem for future theories of
  sentence processing is … the development
  of theories of sentence interpretation.”
                     --Frazier (1998)

(Besides, isn’t the whole point of language
  to derive meaning?)
      How do we go about studying
            interpretation?
   Traditionally, we don’t.
    – comprehension question for every 4th sentence or so,
      just to make sure they’re not zoning out

    While the man hunted the deer that was brown and
     graceful ran into the woods.
    Q: Was the deer brown? OR
       Was the deer in the woods?
    Key Q (never asked): Was the man hunting the deer?
What happens to the interpretation
generated by the initial mis-parse?

 Does it linger?
 Does it just disappear?
 Can it block a full reanalysis?
 Can it cause interpretive difficulties even
  after the rest of the sentence is read?
         Christianson, Hollingworth,
         Halliwell, & Ferreira (2001)
   What happens to that original, incorrect interpretation
    derived from the initial, partial, and ultimately incorrect
    parse?
    – If syntax (and, it is generally assumed, consequently semantics)
      fully reanalyzed, it should not influence final interpretation


   Major assumption: If interpretation is incorrect, then full
    reanalysis has not taken place.
    – Syntactic representation remains incomplete, and thus the
      interpretation is incorrect
    – Might be too strong: Maybe syntax OK, semantics never fixed
                   Expt. 1b
(1a) While Bill hunted the deer (that was brown
  and graceful) ran into the woods.

(1b) While Bill hunted the deer (that was brown
  and graceful) paced in the zoo. (implausible)

(1c) While Bill hunted the pheasant the deer (that
  was brown and graceful) ran into the woods.
  (non-GP)
    How to judge interpretation?
   Radical: Just ask.

Q: Did Bill hunt the deer?

    Yes=INCORRECT No=CORRECT
                  Results Expt. 1b
Also gathered
confidence
ratings;
No diff. in any
condition
in any expt.
VERY
confident.
                     Expt. 2
   Maybe no reanalysis at all?
   Maybe just inference (despite the length of
    ambiguous region effect in 1b)?

(2a) While Bill hunted the brown and graceful
  deer/the deer that was brown and graceful ran
  into the woods.

(2b) The brown and graceful deer/the deer that
  was brown and graceful ran into the woods
  while Bill hunted.
       Another question, too
Did Bill hunt the deer?
(subordinate clause question)

OR

Did the deer run into the woods?
(matrix clause question)
Expt. 2 results
                     Expt. 3
   So far, baseline inference, but syntactic
    manipulations push effect around above
    and beyond inference.
    – Conclusion: Syntax not fully reanalyzed
    – Yet…Wouldn’t it be nice to find a syntactic
      structure that, if fully reanalyzed, would NOT
      ALLOW THE INFERENCE?
   Reflexive absolute transitive (RAT) verbs
              RAT verbs
While Anna dressed the baby that was cute
 and cuddly spit up on the bed.

If fully reanalyzed, Anna CANNOT be
 dressing the baby; must be dressing
 HERSELF.
Results Expt. 3a-b
                  Conclusion
 “Good enough” sentence processing
 Syntactic parse not fully reanalyzed
    – If it is, it’s not mapped onto semantics
   Processor happy with incomplete analysis
    as long as it is plausible.
    – Likely: “the deer” overtly serves as subject of
      matrix clause, remains syntactically present as
      object of subordinate.
          Older vs. younger readers
   Christianson, Williams, Zacks & Ferreira
       (in press, Discourse Processes)
 Perhaps misinterpretation effect larger for
  older readers?
  – Caused by decrement in inhibitory control in
    older folks (Hamm & Hasher, 1992; Hasher,
    Zacks, & May, 1999)
  – Older readers might even be worse at
    inhibiting initial incorrect parse.
                             Expt. 1
   OPT verbs Garden path Structure (subordinate-main clause order)
    – While the man hunted the deer that was brown and graceful ran into
       the woods.
   Non-garden path Structure (main-subordinate clause order)
    – The deer that was brown and graceful ran into the woods while the
       man hunted.
   Q: Did the man hunt the deer?


   RAT: Garden path Structure (subordinate-main clause order)
    – While Anna dressed the baby that was small and cute played in the
       crib.
   Non-garden path Structure (main-subordinate clause order)
    – The baby that was small and cute played in the crib while Anna
       dressed.
   Q: Did Anna dress the baby?
Results Expt. 1
Reading span correlations
                    Expt. 2
   Maybe olders more likely to infer
    (Hartmann & Hasher, 1991)
    – OPT verbs allow inference; RAT do not
 If so, should see exaggerated effect in
  plausible conditions for older readers
 Also manipulated length of ambiguous
  region to see if longer-held interpretations
  harder to inhibit
                          Sentences
   Long Ambiguous Region -- Plausible/Implausible
   Garden path Structure (subordinate-main clause order)
    – While the man hunted the deer that was brown and graceful ran into
       the woods/paced in the zoo.
   Non-garden path Structure (main-subordinate clause order)
    – The deer that was brown and graceful ran into the woods/paced in the
       zoo while the man hunted.


   Short Ambiguous Region -- Plausible/Implausible
   Garden path Structure (subordinate-main clause order)
    – While the man hunted the deer ran into the woods/paced in the zoo.
   Non-garden path Structure (main-subordinate clause order)
    – The deer ran into the woods/paced in the zoo while the man hunted.
              Results, Expt. 2
 Main effects of ambiguous NP length,
  plausibility, and age
 BUT:
    – Age did not modulate the effect of plausibility
      (F1<1; F2<1) nor did it influence the effect of
      ambiguous NP length (F1<1; F2<1).
    – As in Expt 1, age didn’t interact with sentence
      structure, either.
      Not inhibition or inference
   As far as we can tell, anyway
    – Maybe no inhibition required? Never an
      alternative full parse/interpretation
      constructed?
        Very “good enough-y”
   If inference not the issue either why
    better at RATs than OPTs?
        Maybe they aren’t….
                    Expt. 3
   RAT sentences

    While Anna dressed the baby that was
    cute and cuddly spit up on the bed.

   Another question: Did Anna dress herself?

    (Answer should be YES!)
           Results Expt. 3




(nGP: The baby…spit up…while Anna dressed.)
WM correlations with wrong answer
          rate in Expt. 3
                 Explanation
   OPT verbs: Two ways to answer “Did the
    man hunt the deer?”
    – Recall verbatim and figure out
    – Recall propositional content
       The man hunted the deer (initial parse)
       The man hunted [SOMETHING unspecified]
    – Congruent with OPT verbs
    – Olders more likely to rely on “gist”
      (=propositional) rather than verbatim content
                           RAT
   Propositional content
    – Anna dressed the baby (initial)
    – Anna dressed [SOMETHING specific]
        But what?
   Semantics of RAT verbs don’t allow congruency
    of propositional content
    – “dress” doesn’t allow unspecified interpretation
   In order to get reflexive reading, must reactive
    syntax to establish government relation and co-
    indexation
                   WM tie-in
 Olders with less WM resources unable to
  reactivate the syntactic structure required
  to get the reflexive reading.
 Processor may settle on good enough
  interpretation, but to answer the question,
  you need more than that
    – If not enough working memory available to
      either keep working on structure or recall,
      recompute, and revise, then stuck.
    Christianson & Slattery (2005)
   No one has ever looked to see if garden
    paths affect SUBSEQUENT reading
    – Why should they? Recall traditional
      assumptions.
   If “good enough” processing takes place,
    should see people moving on to read
    subsequent text before they’ve completed
    a full reanalysis.
                       Method
   Context AFTER garden path sentence (eye-tracking)

opening region S1     | ambiguous NP1 region
While the man hunted(,) | the deer that was

                 |   disambiguation
large and brown | ran into the woods.

      opening region S2 |    NP2 region
    The man was hunting | a deer (bear) in the
    woods.
                     Results
   Clear classic GP effects
    – First pass time
       ME of struct. on NP1; 72ms longer when non-GP
       ME of struct. on disambiguation; 56ms longer
        when GP
       ME of NP2 on NP2; 158ms longer when
        mismatched
                New Results
   Go Past time (includes re-fixations
    after leftward regressions)
    – ME of struct. on disambiguation; 264ms
     longer when GP
    – ME of struct. on NP2; 86ms longer when GP
    – ME of NP2 on NP2; 248ms longer when
     mismatched
    – Marg. ME (p = .081) by P of struct. on S2
     opening region
                    Summary
   Robust GP effects in early and late measures

   Clear indication that readers moved on to S2
    before structural work on S1 was completed

    Lack of interaction suggests that processes
    related to structural revision and lexical content
    are separate. S1 ambiguity lingers into S2 &
    amplified by NP2, irrespective of match.
What is “good enough” processing?
   NOT “shallow” parsing
    – In other words, not just lack of effort
    – Confidence ratings; downstream effects of GP structure

   Results in SOME kind of underspecified representation
    – Which representation (syntax, semantics, both, other)?

   Underspecification likely result of Incomplete Processing
    (=good enough)
    – Interpretation formed before all sources of information are
      available (some sources slowed by computational demands)
    – Processor moves on (even if some processes are still running)
    Christianson (in preparation)
   Change detection paradigm
       (Sanford, et al., 2005)
    – Memory for text based on representation
      constructed for it.
    – Changes to text that are consistent with
      representation should be harder to detect.
The cookout was going well so far. While
 Tom grilled the hot dog that was long and
 fatty began to burn. The burgers sure
 looked good, though.
The cookout was going well so far. While
 Tom grilled the hot dog that was long and
 fatty it began to burn. The burgers sure looked
 good, though.
                   Conditions
   Garden path vs. non-garden path
    (comma)
    – While Tom grilled, the hot dog that was long
      and fatty began to burn.
   NP-it vs. it-NP
    – While Tom grilled it the hot dog that was long and
      fatty began to burn.
          Results, Expt. 1




Sig. ME of structure & order; Sig. INTERACTION
                    Summary
   People more sensitive to changes in GP
    sentences
    – NOT “shallow” processing; processor notices
      the ambiguity
    – Change acts like question in Christianson, et
      al (in press) and NP2 in Christianson &
      Slattery
        Spurs processor to resolve lingering structural
         problem by some means, because that information
         becomes critical for interpretive task
   However, significant interaction (p = .018)
    suggests that in GP condition, sometimes
    the partial reanalysis proposed by
    Christianson et al (2001) DOES take place
    – Two “hot dogs” in representation, congruent
      with addition of “it” in DO position of
      subordinate clause
                    Conclusion
   Good enough processing results in
    interpretations not faithful to the content
    – Not previously noticed by researchers because right
      questions not asked
    – Not usually noticed by people because usually not
      critical for integration of later material (often even
      incorrect interpretation can be plausibly maintained in
      context)

   Good enough, not just shallow
    – Processor actively tries to resolve, but may move on
      because resources are limited, and input is not
    – (The “Life is short!” model of sentence processing)
    Implications (psycholinguistic)
   Suggests different mechanisms for parser and
    processor
    – Parser worried about getting a licit syntactic structure
      (but might truncate the parse, too)
    – Processor worried about getting a plausible,
      contextually consistent interpretation
    – Parser might be slowed down by ambiguities
    – Processor might run ahead and not check final parse
      unless underspecified representation results in an
      interpretation that doesn’t fit in context
          Implications (general)
 Extent to which parser keeps working or
  processor can look back at results probably
  depends on STM capacity
 STM or other individual differences likely
  predictive of eventual interpretation accuracy
 Over-reliance of processor on top-down
  (semantic, discourse) information (perhaps
  compensatory) might accentuate
  misinterpretations (whether it affects syntactic
  parse or not)
    – Older readers, L2 readers, struggling readers, young
      readers
   Good enough usually good enough, but
    not always.

   Misinterpretations informative for theorists
    – can be predicted and manipulated consistently
      enough to be exploited in reading research
      and instruction (e.g., to increase meta-
      linguistic awareness)
Thank you!

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:3
posted:1/8/2012
language:English
pages:67
jianghongl jianghongl http://
About