Incrementalism and Ambiguity - Structural Ambiguity | LING 499A, Study notes of Linguistics

Material Type: Notes; Class: PSYCHOLING: APPROACHES; Subject: Linguistics; University: University of Maryland; Term: Unknown 2006;

Typology: Study notes

Pre 2010

Uploaded on 02/13/2009

koofers-user-ci5
koofers-user-ci5 🇺🇸

5

(1)

10 documents

1 / 22

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
1
Incrementality
and
Ambiguity
LING499A
Oct 2006
Structural ambiguity
The horse raced past the barn [… fell]
The man gave the boy the dog […bit a cookie]
The software manufacturers sell nowadays […is overpriced]
Put the frog on the napkin […into the box]
The students knew the answer [… was in the back of the book]
While the farmer was hunting the deer [… ran into the forest]
pf3
pf4
pf5
pf8
pf9
pfa
pfd
pfe
pff
pf12
pf13
pf14
pf15
pf16

Partial preview of the text

Download Incrementalism and Ambiguity - Structural Ambiguity | LING 499A and more Study notes Linguistics in PDF only on Docsity!

Incrementality

and

Ambiguity

LING499A

Oct 2006

Structural ambiguity

  • The horse raced past the barn [… fell]
  • The man gave the boy the dog […bit a cookie]
  • The software manufacturers sell nowadays […is overpriced]
  • Put the frog on the napkin […into the box]
  • The students knew the answer [… was in the back of the book]
  • While the farmer was hunting the deer [… ran into the forest]

Dealing with indeterminacy

  • Syntactic incrementality
    • How to deal with choices??
      • Use the analysis that satisfies particular structural metrics (Frazier & Fodor 1978; Kimball 1973)
      • Use other kinds knowledge (e.g. Altmann & Steedman 1988; cf. also Sedivy et al. 1999 on what kind of info is apparently available)
      • Do what’s worked best in the past -- frequency-based guidance

Actually: two problems

  • Remember generation and selection ??
    • How to generate candidate structures
    • How to select among alternatives
  • So, we have to sharpen the question, about how the parser deals with alternatives Garden-path theory

The Garden Path Theory

Strong contextual biases are ineffective (Ferreira & Clifton, 1986) John worked as a reporter for a newspaper. He knew a major story was brewing over the mayor scandal. He went to his editors with a tape and some photos because he needed their approval to go ahead with the story. He ran a tape for one of his editors, and he showed some photos to the other. (a) The editor played the tape agreed the story was big. (b) The editor played the tape and agreed the story was big. The other editor urged John to be cautious.

The Garden Path Theory

Strong plausibility biases are ineffective (Ferreira & Clifton, 1986) (a) The defendant examined by the lawyer turned out to be unreliable. (b) The evidence examined by the lawyer turned out to be unreliable.

The Garden Path Theory

  • Ignoring argument structure information (Mitchell, 1987):
    • After the audience had applauded the actors/ sat down for a well-deserved drink.
    • After the audience had departed the actors/ sat down for a well-deserved drink.
  • Slowdown in first display in depart condition; slowdown in second display in applaud condition.
  • This study much criticized in later work by others … ideas why? Trueswell, Tanenhaus, & Garnsey (1994)

But…

(Clifton et al., 2003)

Ambiguity Resolution in Sentence Processing: Evidence against Frequency-Based Accounts

Pickering, Traxler, and Crocker

Journal of Memory and Language 43, 447- (Hana’s Presentation)

Study Aims

• How does the processor choose the initial

strategy for resolving syntactic ambiguity?

• Hypothesis : Frequency-based

processing accounts are the strategy used

to resolve syntactic ambiguity.

Frequency-Based Processing

Accounts

  • People adopt the

analysis that has

most frequently

turned out to be

correct in the past.

  • Based solely on

subcategorization

preferences of the

verb

  • People adopt the

analysis that is most

likely to be correct

given the information

that is available at the

point of ambiguity.

  • Based on preferred

completions for

sentence fragments

Serial Lexical-Guidance Account Serial-Likelihood Account

(1a) The athlete realized her potential one day might make her a world-class sprinter. (1b) The athlete realized her exercises one day might make her a world-class sprinter. If readers adopt an object analysis first, they would realize that (1b) is less plausible before the word “might” than (1a) is. This should show up as either increased reading time and/or more regressive eye-movements in (1b) as opposed to (1a) before the word “might.” Readers find it more difficult to give up a plausible analysis that turns out to be wrong and easy to give up an implausible analysis. Therefore, if readers adopt an object analysis first, they will find it harder to adopt the correct sentential-complement analysis in (1a) and easier to switch analyses in (1b) This should show up as either increased reading time and/or more regressive eye-movements after disambiguation (after “might”) in (1a) as opposed to (1b).

The young athlete realized her

potential/exercises one day might make her

a world-class sprinter.

Noun Region Postnoun Region Verb Region Postverb Region

EXPERIMENT 1 RESULTS

EXPERIMENT 2 RESULTS

Experiment 3

(3a) While the pilot was flying the plane that had arrived stood over by the fence. (3b) While the pilot was flying the horse that had arrived stood over by the fence. The ambiguity here is whether “the plane” or “the horse” is the object of flying or the subject of “stood.” For both (3a) and (3b) “was flying” turns out to be intransitive. Both frequency-based accounts predict that readers will assume that “plane/horse” is the subject of the sentence therefore will not run into any problems such as the implausibility of flying a horse. However, if neither frequency-based account is correct, and readers first assume that plane/horse is the object of “was flying,” then a crossover pattern similar to those found in Experiments 1 and 2 is predicted. There will be more difficulty with (3b) than (3a) before “stood” and more difficulty with (3a) than (3b) after “stood.” (3a) While the pilot was flying the plane that had arrived stood over by the fence. (3b) While the pilot was flying the horse that had arrived stood over by the fence.