An Information-Theoretic Characterization of Morphological Fusion

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10115–10120 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 10115 An Information-Theoretic Characterization of Morphological Fusion Neil Rathi Palo Alto High School neilrathi@gmail.com Michael Hahn∗ Stanford University SFB 1102, Saarland University mhahn2@stanford.edu Richard Futrell∗ University of California, Irvine rfutrell@uci.edu Abstract Linguistic typology generally divides syn- thetic languages into groups based on their morphological fusion (von Humboldt, 1825). However, this measure has long been thought to be best considered a matter of degree (e.g. Greenberg, 1960). We present an information- theoretic measure, called informational fusion, to quantify the degree of fusion of a given set of morphological features in a surface form, which naturally provides such a graded scale. Informational fusion is able to encapsulate not only concatenative, but also nonconcatenative morphological systems (e.g. Arabic), abstract- ing away from any notions of morpheme seg- mentation. We then show, on a sample of twenty-one languages, that our measure reca- pitulates the usual linguistic classifications for concatenative systems, and provides new mea- sures for nonconcatenative ones. We also eval- uate the long-standing hypotheses that more frequent forms are more fusional, and that paradigm size anticorrelates with degree of fu- sion. We do not find evidence for the idea that languages have characteristic levels of fusion; rather, the degree of fusion varies across part- of-speech within languages. 1 Introduction Traditional morphological typology divides syn- thetic languages into two distinct groups, aggluti- native and fusional (von Humboldt, 1825). Agglu- tinative languages have morphemes which can be separated into identifiable parts corresponding to single features. For example, the Hungarian form embereknek can be separated into a root and two suffixes, each of which expresses a single morpho- logical feature: ember-ek-nek (person-PL-DAT). On the other hand, fusional languages express multi- ple features in a single morpheme, such as Latin serv¯ıs (servant-DAT.PL), where the suffix –¯ıs in- dicates the dative and plural simultaneously and *Equal contribution by MH and RF. cannot be analyzed into parts that individually cor- respond to the genitive or plural features (Brown, 2010; Plank, 1999). Linguistic typologists have long recognized that this distinction is more of a spectrum than a cate- gorical distinction, with Greenberg (1960) defining an ‘index of agglutination’ metric to determine the degree to which a language is agglutinative across its morphological paradigms. Interestingly, the no- tion appears to be graded even within a language. For example, the Latin adjectival feminine genitive plural suffix is –¯arum, where the thematic vowel ¯a corresponds weakly to the feminine. Here, we provide an information-theoretic char- acterization of the degree of fusion of any given form in a language, naturally providing a graded measure. Our core intuition is that a form which expresses a given set of features can be classified as fusional if it cannot be predicted given the forms for other sets of morphological features (i.e. the “rest of the paradigm”). For example, the Latin end- ing –¯ıs in Table 1 is almost entirely unpredictable from the rest of the …

Content truncated to 500 words