What do RNN Language Models Learn about Filler–Gap Dependencies?

Ethan Wilcox, Roger Levy, Takashi Morita, Richard Futrell


Abstract
RNN language models have achieved state-of-the-art perplexity results and have proven useful in a suite of NLP tasks, but it is as yet unclear what syntactic generalizations they learn. Here we investigate whether state-of-the-art RNN language models represent long-distance filler–gap dependencies and constraints on them. Examining RNN behavior on experimentally controlled sentences designed to expose filler–gap dependencies, we show that RNNs can represent the relationship in multiple syntactic positions and over large spans of text. Furthermore, we show that RNNs learn a subset of the known restrictions on filler–gap dependencies, known as island constraints: RNNs show evidence for wh-islands, adjunct islands, and complex NP islands. These studies demonstrates that state-of-the-art RNN models are able to learn and generalize about empty syntactic positions.
Anthology ID:
W18-5423
Volume:
Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP
Month:
November
Year:
2018
Address:
Brussels, Belgium
Editors:
Tal Linzen, Grzegorz Chrupała, Afra Alishahi
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
211–221
Language:
URL:
https://aclanthology.org/W18-5423
DOI:
10.18653/v1/W18-5423
Bibkey:
Cite (ACL):
Ethan Wilcox, Roger Levy, Takashi Morita, and Richard Futrell. 2018. What do RNN Language Models Learn about Filler–Gap Dependencies?. In Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 211–221, Brussels, Belgium. Association for Computational Linguistics.
Cite (Informal):
What do RNN Language Models Learn about Filler–Gap Dependencies? (Wilcox et al., EMNLP 2018)
Copy Citation:
PDF:
https://aclanthology.org/W18-5423.pdf
Data
Billion Word BenchmarkOne Billion Word Benchmark