Acquisition of Inflectional Morphology in Artificial Neural Networks With Prior Knowledge

TL;DR

This study investigates how prior knowledge of one language's morphology affects neural network learning of inflectional rules in a second language, revealing that relatedness and morphological type influence learning ease.

Contribution

It provides new insights into how artificial neural networks transfer morphological knowledge across languages with different typologies and relatedness levels.

Findings

Related languages are easier for models to learn.

Prefixing and suffixing languages influence each other's learning difficulty.

Agglutinative source languages simplify learning of other inflectional morphologies.

Abstract

How does knowledge of one language's morphology influence learning of inflection rules in a second one? In order to investigate this question in artificial neural network models, we perform experiments with a sequence-to-sequence architecture, which we train on different combinations of eight source and three target languages. A detailed analysis of the model outputs suggests the following conclusions: (i) if source and target language are closely related, acquisition of the target language's inflectional morphology constitutes an easier task for the model; (ii) knowledge of a prefixing (resp. suffixing) language makes acquisition of a suffixing (resp. prefixing) language's morphology more challenging; and (iii) surprisingly, a source language which exhibits an agglutinative morphology simplifies learning of a second language's inflectional morphology, independent of their relatedness.