1

Hungarian is often used as the prototypical example of a heavily agglutinative, synthetic language, and with regards to noun declension and derivational morphology this is doubtless true; Hungarian nominal morphology has a fairly regular agglutinative paradigm with an invariant plural marker -Vk-. However, the verbal conjugation system seems to be fairly fusional, with three or so distinct sets of endings that are not formed through agglutination.

Present Indefinite: -ok, -sz, -Ø, -unk, -tok, -nak

Present Definite: -om, -od, -ja, -juk, -tájok, -ják

Past Indefinite: -tam, -tál, -ott, -tunk, -tatok, -tak

Past Definite: -tam, -tad, -ta, -tuk, -tátok, -ták

Conditional endings seem to be derived from the present endings, with the addition of a suffix -nV- directly after the root. However, there is variation in the vowel and the 2nd person conditional ending seems to come from the past ending, rather than the present ending.

Subjunctive endings mostly seem to derive from a combination of a -s- suffix, followed by Present endings, with a different marker (-on) in the third person singular indefinite subjunctive.

Considering that these 4-6 distinct sets of endings within the same paradigm, whereas ancient IE languages tend to have between 2 and 3 in the active ("primary", "secondary", and "perfect"), can it not be stated that Hungarian is more (or at the very least as) agglutinative than classical IE languages?

3
  • fusional/agglutinative is not a strict divide, but more of a continuum and, compared to its Indo-European neighbours, Hungarian certainly seems agglutinative, even if it doesn't when compared to e.g. Georgian – Tristan Jan 15 at 10:39
  • The main point of agglutination is the properties of the morphemes border: in the agglutinative languages there's no sound changes at the morphemes border, in the fusional languages the changes are often, e.g. Polish, fusional, ka “hand” Nom., ki Gen., but ce Dative. So what will you say about the characteristics the morphemes border in Hungarian? Are there sound changes? Naturally, I mean the changes in the root morpheme. – Yellow Sky Jan 15 at 15:19
  • @Tristan Indeed. It is a spectrum. However, it seems that at least in verbs, Hungarian is more "fusional" – or at least as fusional as – it's IE neighbors – it has more sets of distinct verb endings, formed by suppletion of the morpheme (ie -ok > -om) rather than concatenation of an extra marker. Compare Spanish, where the future is generally formed with the addition of a suffix -rV-, then the normal present verb endings : a more "agglutinative" seeming method than Hungarian's four distinct sets of endings. – user8606 Jan 16 at 10:12
2

You're completely correct that Hungarian verb conjugation is quite fusional, but even there it's at most on par with IE languages. You mention Spanish, which does have a relatively neat TAM marker + personal ending system, but there are still three sets of endings and numerous classes of stem alternations. Classical IE languages can be a lot worse; Ancient Greek distinguishes personal endings on at least athematic/thematic, active/middle, primary/secondary and imperative/non-imperative axes, while the verb itself has up to six unpredictable stems.

In the end it doesn't matter much: the rest of Hungarian morphology is vastly more agglutinative than what you find in IE languages. Even the rest of verbal morphology, with a number of borderline inflexional suffixes such as potential -hAt and causative -(t)At has little equivalent in IE languages, and words such as hypothetical ír-ogat-tat-hat-ná-tok (write-freq-caus-pot-cond-2pl) certainly don't.

1

This raises a factual question, namely whether Hungarian is often used as the prototypical example of a heavily agglutinative, synthetic language. In my experience, Turkish is the "prototypical" example of an agglutinative language, but I suspect that the prototype depends very much on the literary genre where the concept is invoked. The primary distinction ought to be in terms of intended audience, that is, are you looking at popular-press claims, or technical, peer-reviewed claims. A secondary division should be made in terms of when the claim is made (e.g. books of the 1950's vs. blogs of the 2020's). The concept arises from 19th century typological studies rooted in evolutionary ideas of the Romantic movement, and has not been extinguished by contemporary research. There is a paper in IJAL by Joseph Greenberg, "A Quantitative Approach to the Morphological Typology of Language", which discusses some of the history of these terms. He also points to Sapir's work on the topic, which begins to disassemble the typology, and replace the vague concepts with more specific grammatical properties. Greenberg then proposes a series of indices for classifying "degree-of-X", where Swahili and Yakut (in a small language survey) have a high agglutination index, and Sanskrit and Swahili have a high synthesis index.

There isn't much agreement about what makes a language "agglutinative", and what the alternative is – usually, agglutinative and fusional languages are "more similar" compared to "analytic" or "isolating" languages. What is at stake is whether single "words" have very many form-variants (tenses, moods, persons, numbers, etc) or very few (the relevant concepts are expressed by word-combinations). Agglutinative languages are traditionally distinguished from fusional languages by agglutinative languages having a more transparent form/meaning relation, thus in Turkish, there is a plural affix and case affixes, whereas in Indo-European languages you have to memorize combinations that express case-gender-number. The more a language binds multiple semantic properties into a single affix, the less "agglutinative" is is. A second less-popular feature of agglutination is shape-invariance, that there should be only one context-free exponent of an expressed semantic property. On the one hand, this correctly distinguishes portmanteau gender-number-case complexes in Indo-European from Quechua and Turkish (which happen to lack gender markers), but on the other hand would exclude Turkish because there are two surface allomorph of the plural in Turkish, [lar] and [ler], and even more for the 3rd person possessive suffix in Turkish (8 surface variants, according to phonological context). What confounds the taxonomy of languages into agglutinative vs. other kinds of languages is the inability to distinguish morphologically-governed allomorphy from phonetic variants arising from phonological rules.

Here is an empirical question whose answer is not well-known: are there any pure agglutinative languages? A "pure agglutinative language" would be one which has a single, invariant affix on a word that expresses a syntacto-semantic property. If a language has "nominative" and "accusative", there is only one affix, for all persons and numbers; if there is number marking, everything with number marking has the appropriate number marker and there is no difference between how things are marked in verbs and nouns, or 1st person and 3rd person. I suspect that the few if any languages that are "purely agglutinative" are in fact isolating languages that have recently changed a phrasal clitic into a general affix. The justification for including Hungarian as an agglutinative language is that on the continuum of agglutination, it is "more agglutinative" than not, but there is no linguistic excuse for holding Hungarian to be "prototypical".

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.