The paper presents four models submitted to Part 2 of the SIGMORPHON 2021 Shared Task 0, which aims at replicating human judgements on the inflection of nonce lexemes. Our goal is to explore the usefulness of combining pre-compiled analogical patterns with an encoder-decoder architecture. Two models are designed using such patterns either in the input or the output of the network. Two extra models controlled for the role of raw similarity of nonce inflected forms to existing inflected forms in the same paradigm cell, and the role of the type frequency of analogical patterns. Our strategy is entirely endogenous in the sense that the models appealing solely to the data provided by the SIGMORPHON organisers, without using external resources. Our model 2 ranks second among all submitted systems, suggesting that the inclusion of analogical patterns in the network architecture is useful in mimicking speakers’ predictions.

Cet article présente la conception et le développement de Demonette2, une base de données dérivationnelle à grande échelle du français, développée dans le cadre du projet ANR Démonext (ANR-17-CE23-0005). L’article décrit les objectifs du projet, la structure de la base et expose les premiers résultats du projet, en mettant l’accent sur un enjeu crucial : la question du codage sémantique des entrées et des relations.