Related Experiment Videos
Garden path repair: diagnosis and triage
1Program in Linguistics, Graduate Center, City University of New York, 365 Fifth Avenue, New York, NY 10016, USA. jfodor@gc.cuny.edu
Language and Speech
|February 24, 2001
Summary
This study investigates "triage" in sentence parsing, a proposed function to assess reanalysis effort. Evidence suggests human sentence parsers lack this triage capability, relying instead on basic diagnosis.
Area of Science:
- Psycholinguistics
- Computational Linguistics
- Cognitive Science
Background:
- Garden-path sentences present parsing challenges, requiring reanalysis.
- Previous models proposed a "basic diagnosis" system for resolving ambiguity.
- Meng and Bader (2000b) introduced a "triage" function to predict revisability.
Purpose of the Study:
- To determine if "triage" is a component of human sentence parsing.
- To establish the scope and function of a potential triage mechanism.
- To differentiate between models with and without triage capabilities.
Main Methods:
- Comparison of four distinct hypotheses regarding the scope of triage.
- Identification of empirical predictions distinguishing "triage" from "basic diagnosis" models.
- Analysis of existing evidence on human sentence processing.
Main Results:
- A powerful triage component could potentially resolve parsing failures.
- Limited local triage may be ineffective if cue effectiveness varies.
- An intermediate triage hypothesis offers better discrimination for revisions.
- The most psychologically plausible triage variant is difficult to distinguish from no triage.
Conclusions:
- Current evidence suggests the human sentence parser does not possess a triage capability.
- The "basic diagnosis" model, without triage, may better account for observed parsing behavior.
- Further empirical research is needed to definitively establish the presence or absence of triage in parsing.