Related Experiment Videos
The role of prediction in construction-learning
Adele E Goldberg1, Devin M Casenhiser, Nitya Sethuraman
1Linguistics, East Pyne Hall, Princeton University, Princeton, NJ 08544, USA. adele@princeton.edu
Journal of Child Language
|July 28, 2005
Summary
Prediction motivates learning semantic generalizations in language. Argument constructions are as reliable as verbs for predicting sentence meaning, making them essential for language comprehension.
Area of Science:
- Cognitive Science
- Linguistics
- Computational Linguistics
Background:
- Non-linguistic categorization relies on predictive functions.
- The role of prediction in learning semantic generalizations about argument structure constructions is under-explored.
Purpose of the Study:
- To investigate the role of prediction in learning semantic generalizations for argument structure constructions.
- To compare the predictive validity of argument constructions versus verb morphology in determining sentence meaning.
Main Methods:
- Corpus analysis of linguistic data.
- Statistical analysis of cue validity and category validity.
Main Results:
- Argument constructions demonstrate high cue validity, comparable to verb morphology, in predicting sentence meaning.
- Argument constructions exhibit greater category validity than verb morphology.
- Many verbs possess low individual cue validity, highlighting the importance of constructions.
Conclusions:
- Prediction is a key driver for learning semantic generalizations in argument structure constructions.
- Argument constructions are crucial, reliable predictors of sentence meaning, often more so than individual verbs.