The role of paradigm structure in Hungarian morphology and phonology with typological comparisons

research sponsored by NKFIH (#139271)

home description publications software

description

PDF

summary and aims

We plan to study and analyse morphological and morphophonological phenomena (a) in Hungarian, (b) in some Uralic languages and (‌‌​c) from a general typological perspective which — due to the properties of their morpho(phono)logical patterns — cannot be properly described using models of morphology that presuppose agglutination as the basic principle for building words (e.g. the strict Item-and-Arrangement and Item-and-Process approaches). The aims of the research are two-fold and thus our expected results are also dual, empirical and theoretical. (i) On the one hand, we want to give an exhaustive, observationally precise and quantitative (i.e. frequency-based) description of the phenomena (some of which have never been sufficiently explored) with special emphasis on lexical variation, vacillation, paradigm gaps and templatic behaviour. (ii) On the other hand, we plan to use formal models, and possibly develop computational implementations (based on already available Word-and-Paradigm analogy-based models) which — as opposed to the frameworks designed for pure agglutination — are more suited to the analysis of the non-agglutinative patterns discovered and described.

fundamental question

The basic question of the research is to determine what kind of linguistic models are capable of adequately describing and explaining the complex phenomena of the derivational and inflectional morphological patterns and the related morphophonological regularities in Hungarian and other typologically significant Uralic languages. The following questions arise: What general linguistic/cognitive/perceptual principles do these morpho(phono)logical patterns manifest? What explains variation attested in these patterns and what are the differences and similarities between its two types, lexical variation and vacillation? What are the connections between individual morphological patterns and what roles do their type and token frequencies play? How and to what extent can we predict where variation and paradigm gaps occur in morphology? How much is templatic behaviour (e.g. restrictions on syllable count) characteristic of the grammars of the languages examined?

The starting hypothesis is that (i) there are many non-agglutinative morpho(phono)logical patterns in Hungarian and the Uralic languages we plan to analyse; and these patterns can be described and explained by (ii) direct reference to surface patterns, i.e. without referring to underlying representations, (iii) without assuming strict agglutination, i.e that words can be exhaustively and unambiguously decomposed into parts, (iv) by referring to full word-forms and their paradigms (in a word-and-paradigm model) and (v) that a given morphological pattern and its variability can be explained with its degree of similarity to other patterns and the (type or token) frequencies of the patterns (in an analogical approach).

relevance

The research will (i) (by virtue of the expected empirical and experimental results) extend and deepen our (often scanty) knowledge of some areas of Hungarian morphology and morphophonology — especially of variation and paradigm gaps; (ii) discover the non-agglutinative properties of the morphological pattern of some cognate languages and thereby identify their status in a general typological space and (iii) from a theoretical perspective it is a further contribution to the study of the linguistic system and its mechanism by developing an analogical and word-based morphological analysis based on language use, which in our view gives a cognitively more realistic account of the native speakers’ linguistic abilities than an approach relying on abstract underlying representations and rules/constraints (like various versions of generative grammar).

The uniqueness of the research consists in the fact that (a) we will demonstrate that — as opposed to the general view found in the Hungarian and international literature according to which Hungarian is essentially agglutinative (and therefore can be adequately described without reference to the paradigm) — in fact, Hungarian morphology has an abundance of non-agglutinative patterns; (b) it will fill gaps in and — since essentially corpus-based — it will significantly increase our knowledge of the morphologies of Uralic languages (whose morphologies are generally considered to be agglutinative too) compared to available descriptions. A further descriptive advantage is that by charting the non-agglutinative morphological patterns in these languages it will be possible to determine their status within a general typological space more precisely.

for the general audience

Hungarian and Uralic languages are generally considered to be “agglutinative” in that they build words from easily separable and identifiable minimal parts which have a constant meaning recurring in different words. The word szelíd-ít-het-ő-ség-ük-ről ‘about their tamability’ is an example where all minimal parts are meaningful and easy to separate. In the agglutinative “ideal type” every word is supposed to be like this and the grammar has to determine (i) these minimal parts (their form and meaning) and (ii) the rules that specify how to combine them into words. If we determine all the parts and all the rules, we have described the morphology of the language we describe. In this view of morphology there is no need to refer to whole word forms since by identifying the parts and the rules we have determined what words are possible.

Our descriptive aim in this research is to show and describe that Hungarian and the Uralic languages have plenty of phenomena that cannot be analysed in this way (Hungarian nicknames are an example because they cannot be described by referring to their parts only —the result must be two syllables, it does not matter how many syllables the “original word” has: e.g. Zsolt ~ Zsolt-i, Ferenc ~ Fer-i, Orsolya ~ Ors-i. Our theoretical aim is to develop and employ morphological models that can handle the non-agglutinative patterns because they can refer directly to full words forms and their similarities (their degree of similarity and frequency) and their connections with other forms in their paradigm (their “morphological network”) without accepting exhaustive decomposability that the simplistic agglutination-based model described above assumes.

home description publications software

description.txt · Last modified: by 127.0.0.1