Lopez, Adam

Goldwater, Sharon

McCurdy, Kate

McCurdy, Katherine

2024-02-12T10:15:26Z

2024-02-12T10:15:26Z

2024

Morphological generalization, or the task of mapping an unknown word (such as a novel noun Raun) to an inflected form (such as the plural Rauns), has historically proven a contested topic within computational linguistics and cognitive science, e.g. within the past tense debate (Rumelhart and McClelland, 1986; Pinker and Prince, 1988; Seidenberg and Plaut, 2014). Marcus et al. (1995) identified German plural inflection as a key challenge domain to evaluate two competing accounts of morphological generalization: a rule generation view focused on linguistic features of input words, and a type frequency view focused on the distribution of output inflected forms, thought to reflect more domain-general cognitive processes. More recent behavioral and computational research developments support a new view based on predictability, which integrates both input and output distributions. My research uses these methodological innovations to revisit a core dispute of the past tense debate: how do German speakers generalize plural inflection, and can computational learners generalize similarly? This dissertation evaluates the rule generation, type frequency, and predictability accounts of morphological generalization in a series of behavioral and computational experiments with the stimuli developed by Marcus et al.. I assess predictions for three aspects of German plural generalization: distribution of infrequent plural classes, influence of grammatical gender, and within-item variability. Overall, I find that speaker behavior is best characterized as frequency-matching to a phonologically-conditioned lexical distribution. This result does not support the rule generation view, and qualifies the predictability view: speakers use some, but not all available information to reduce uncertainty in morphological generalization. Neural and symbolic model predictions are typically overconfident relative to speakers; simple Bayesian models show somewhat higher speaker-like variability and accuracy. All computational models are outperformed by a static phonologically-conditioned lexical baseline, suggesting these models have not learned the selective feature preferences that inform speaker generalization.

https://hdl.handle.net/1842/41429

http://dx.doi.org/10.7488/era/4161

en

The University of Edinburgh

Dankers, V., Langedijk, A., McCurdy, K., Williams, A. and Hupkes, D. (2021), Generalising to German Plural Noun Classes, from the Perspective of a Recurrent Neural Network, in ‘Proceedings of the 25th Conference on Computational Natural Language Learning’, Association for Computational Linguistics, Online, pp. 94–108. URL: https://aclanthology.org/2021.conll-1.8

McCurdy, K., Goldwater, S. and Lopez, A. (2020), Inflecting When There’s No Majority: Limitations of Encoder-Decoder Neural Networks as Cognitive Models for German Plurals, in ‘Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics’, Association for Computational Linguistics, Online, pp. 1745–1756. URL: https://www.aclweb.org/anthology/2020.acl-main.159

McCurdy, K., Goldwater, S. and Lopez, A. (2022), ‘Regularization or lexical probabilitymatching? How German speakers generalize plural morphology’, Proceedings of the Annual Meeting of the Cognitive Science Society 44(44). URL: https://escholarship.org/uc/item/61v0r6f1

McCurdy, K., Lopez, A. and Goldwater, S. (2020a), Conditioning, but on Which Distribution? Grammatical Gender in German Plural Inflection, in ‘Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics’, Association for Computational Linguistics, Online, pp. 59–65. URL: https://www.aclweb.org/anthology/2020.cmcl-1.8

McCurdy, K., Lopez, A. and Goldwater, S. (2020b), Modeling grammatical gender and plural inflection in German, in ‘Proceedings of the 26 Architectures and Mechanisms for Language Processing Conference (AMLaP)’, Universität Potsdam. URL: https://amlap2020.github.io/a/261.pdf

McCurdy, K. and Serbetçi, O. (2017), Grammatical gender associations outweigh topical gender bias in crosslinguistic word embeddings, in ‘Presented at WiNLP (Women in Natural Language Processing)’, Vancouver, Canada. URL: http://www.winlp.org/wp-content/uploads/2017/finalₚ𝑎𝑝𝑒𝑟𝑠₂017/46ₚ𝑎𝑝𝑒𝑟.𝑝𝑑𝑓

Soulos, P., Hu, E., McCurdy, K., Chen, Y., Fernandez, R., Smolensky, P. and Gao, J. (2023), Differentiable Tree Operations Promote Compositional Generalization, in ‘Proceedings of the 40th International Conference on Machine Learning’, Proceedings of Machine Learning Research, PMLR.

German plural system

Morphological Generalization

German plural inflection

rule generation

type frequency

static phonologically-conditioned lexical baseline

Rules, frequency, and predictability in morphological generalization: behavioral and computational evidence from the German plural system

Thesis or Dissertation

Doctoral

PhD Doctor of Philosophy