Julius Steuer

We present a cross-linguistic study of vowel harmony that aims to quantifies this phenomenon using data-driven computational modeling. Concretely, we define an information-theoretic measure of harmonicity based on the predictability of vowels in a natural language lexicon, which we estimate using phoneme-level language models (PLMs). Prior quantitative studies have heavily relied on inflected word-forms in the analysis on vowel harmony. On the contrary, we train our models using cross-linguistically comparable lemma forms with little or no inflection, which enables us to cover more under-studied languages. Training data for our PLMs consists of word lists offering a maximum of 1000 entries per language. Despite the fact that the data we employ are substantially smaller than previously used corpora, our experiments demonstrate the neural PLMs capture vowel harmony patterns in a set of languages that exhibit this phenomenon. Our work also demonstrates that word lists are a valuable resource for typological research, and offers new possibilities for future studies on low-resource, under-studied languages.

Research on the cognitive plausibility of language models (LMs) has so far mostly concentrated on modelling psycholinguistic response variables such as reading times, gaze durations and N400/P600 EEG signals, while mostly leaving out the dimension of what Mahowald et al. (2023) described as formal and functional linguistic competence, and developmental plausibility. We address this gap by training a series of GPT-like language models of different sizes on the strict version of the BabyLM pretraining corpus, evaluating on the challenge tasks (BLiMP, GLUE, MSGS) and an additional reading time prediction task. We find a positive correlation between LM size and performance on all three challenge tasks, with different preferences for model width and depth in each of the tasks. In contrast, a negative correlation was found between LM size and reading time fit of linear mixed-effects models using LM surprisal as a predictor, with the second-smallest LM achieving the largest log-likelihood reduction over a baseline model without surprisal. This suggests that modelling processing effort and linguistic competence may require an approach different from training GPT-like LMs on a developmentally plausible corpus.

This study presents an analysis of diachronic linguistic changes in English scientific writing, utilizing surprisal from transformer-based language models. Unlike traditional n-gram models, transformer-based models are potentially better at capturing nuanced linguistic changes such as long-range dependencies by considering variable context sizes. However, to create diachronically comparable language models there are several challenges with historical data, notably an exponential increase in no. of texts, tokens per text and vocabulary size over time. We address these by using a shared vocabulary and employing a robust training strategy that includes initial uniform sampling from the corpus and continuing pre-training on specific temporal segments. Our empirical analysis highlights the predictive power of surprisal from transformer-based models, particularly in analyzing complex linguistic structures like relative clauses. The models’ broader contextual awareness and the inclusion of dependency length annotations contribute to a more intricate understanding of communicative efficiency. While our focus is on scientific English, our approach can be applied to other low-resource scenarios.