Agglutinative folk?

Jun 3, 2014 by


Several nationalistic YouTube videos, chiefly from “Hungarian History Productions”, speak of Ural-Altaic peoples or civilizations as “agglutinative folk” (see, for example, here and here). However, a correlation between Ural-Altaic languages—let alone ethnic groups associated with those languages—and agglutinative morphology is tenuous at best. But first a few words about the supposed Ural-Altaic language family. Originally proposed in the 19th century, the hypothesis that Uralic languages (including Finnic, Ugric, and Samoyedic languages) and Altaic languages form a family enjoyed wide acceptance among linguists into the mid 20th century. However, since the 1960s, it has been rejected by most scholars. In fact, most linguists today doubt the validity of the Altaic grouping, which links the Turkic, Mongolic, and Tungusic languages into a single family unified by common descent. Starting in the 1990s, a renewed interest in a relationship between the Uralic and Altaic tongues captured the imagination of some scholars in the context of the Eurasiatic hypothesis, but most linguists view the relationship as one of contact and mutual influence rather than common origin. In fact, Uralic languages may have a closer genetic affinity to Indo-European than to Altaic languages. Therefore, the most widely adopted view is that Ural-Altaic is a geographical rather than linguistic label. But in either case, could it be true that the languages designated as “Ural-Altaic” are associated in some deep, ethno-cognitive way with agglutinative structure?

In linguistics, the term “agglutinative” refers to a model of morphology, generally opposed to “isolating” and “fusional” (the latter is also often referred to as “synthetic”, although this term has been used in several confusing ways).* In the isolating model, grammatical features such as tense, aspect, or agreement (on verbs) and number and case (on nouns) are expressed through free-standing words. A classical example of the isolating model is Mandarin Chinese, where the plural form of the noun ‘dog’ is achieved not by adding a suffix—as in the English dogs—but by adding a quantifier (‘several’) and a classifier (‘item’), as in ‘several item dog’ (cf. the English several grains of rice). In the agglutinative and fusional models, such grammatical features are expressed through morphemes—suffixes or prefixes—that attach to the verb or the noun. The main differences between the agglutinative and fusional models concern how many features each morpheme expresses and how much interaction occurs between the morphemes. In the agglutinative model, each morpheme expresses one grammatical feature and there is little interaction between morphemes. In effect, the morphemes are strung one after another into long strings that are words. In the fusional model, a morpheme typically expresses several grammatical features at once; moreover, when morphemes attach to the stem, they affect the pronunciation of each other.

Let us examine how this works in an agglutinative language, Turkish, and a fusional language, Russian. Take, for example, the genitive (“possessive”) plural form of ‘dog’. In Turkish it is a three-morpheme köpek-ler-in (hyphens indicate breaks between morphemes), whereas in Russian it is a two-morpheme sobak-0 (the “0” indicates a morpheme that is not pronounced, in contrast with the citation form sobak-a and other forms). The difference in the number of morphemes comes from the fact that in Turkish the plural number and the genitive case are expressed by separate morphemes, –ler and -in, respectively, whereas in Russian both things are expressed by a single morpheme. Moreover, in Turkish the plural is always -ler and the genitive is always –in (except for adjustments due to vowel harmony, which are ubiquitous in the language and do not concern these specific morphemes). In Russian, in contrast, the genitive plural may be expressed by the zero morpheme, as in the example above; by -ov, as in kot-ov ‘of (male) cats’; or by -ej, as in myš-ej ‘of mice’—and the choice idiosyncratically depends on the noun itself. If you want to speak Russian, you just have to learn which nouns go with “0”, or -ov, or -ej, and the same for all other number/case combinations.

While it is true that “Ural-Altaic” languages tend to be agglutinative in their morphology, it is not true of all of these languages, nor has it always been true. For example, Proto-Finno-Ugric—the ancestral form of Finno-Ugric languages—was mostly agglutinative, modern Finno-Ugric languages combine elements of the fusional and agglutinative models. Over long periods of time, agglutinative languages tend to morph into fusional ones, as strings of morphemes become grammaticalized as single morphemes. For example, a number morpheme and a case morpheme may with time become fused a single morpheme expressing both number and case at once. On the other hand, the formerly independent words of an isolating language are sometimes reanalyzed as attachable (or “bound”) morphemes; for instance, a free-standing auxiliary verb may become a tense or aspect suffix, which is what happened to aspectual suffixes -le and ‑guo in Mandarin Chinese.


More importantly, it is wrong to talk about Ural-Altaic groups as “the agglutinative folk” because the agglutinative model is found in many diverse tongues around the world, including languages spoken in Europe (Basque, isolate), Asia (Kannada, Dravidian), Africa (Swahili, Bantu), and the Americas (Cree, Algonquian). Languages in the three major language families indigenous to the Caucasus region—Abkhaz in Northwest Caucasian family, Lezgin in Northeast Caucasian family, and Georgian in Kartvelian family—are mostly agglutinative as well. Other agglutinative languages include Burmese (Tibeto-Burman), Japanese and Korean, Squamish (Salish), Quechua, and Greenlandic (Eskimo-Aleut). Aboriginal Australian languages in the Pama-Nyungan family are agglutinative as well. Would the Hungarian nationalists posting their “agglutinative folk” videos on YouTube want to include all of those groups within their “folk”? I doubt it.


The same vast geographical spread characterizes the other morphological models as well. For example, isolating languages are found in Africa (Hausa), Asia (Vietnamese), Oceania (Rapanui), and the Americas (Kipea). In fact, all deep grammatical typology properties seem to be widely scattered across the world, often irrespective of language families. At the same time, neighboring languages, even if they are related by descent, often show deeply contrasting grammatical patterns. Take, for example, the property of whether the verb precedes or follows the object (‘drink beer’ or ‘beer drink’). As the map on the left illustrates, both Verb-Object and Object-Verb languages can be found on every continent and in a number of different language families.


Yet, in any relatively restricted area—as illustrated here with western sub-Saharan Africa—languages of both types can be found. This pattern of spatial distribution obtains also with the property of using prepositions (‘in school’) or postpositions (‘school in’), adjective-noun (‘big house’) or noun-adjective (‘house big’) order, and of preferentially using suffixes or prefixes. According to Mark Baker’s Atoms of Language, languages that are syntactically polysynthetic—meaning that verbs exhibit agreement not only with the subject, but also with the object and with indirect object (if present)—can be found in the US Southwest (Southern Tiwa, Kiowa Tanoan family), Mexico (Nahuatl, Uto-Aztecan family), and Central Chile (Mapudungun, Araucanian), as well as in Australia (Nunggubuyu, Gunwingguan family), Northeastern Siberia (Chukchi and Koryak, both from the Chukotko-Kamchatkan family), and India (Sora, Munda family).


Another good example of how neighboring languages can differ radically in their typological properties concerning such properties as word order comes from languages of the Atlantic Europe: Basque, English, Welsh, French, and Spanish. Note that not only are they spoken in the same region, but also two of them—French and Spanish—are close relatives in the Romance grouping, and four out of five—all but Basque—are Indo-European. However, when it comes to word order possibilities, the five languages are remarkably different. Basque stands out by having the object precedes the verb (OV); see the map with green and red dots above. English differs from the others in that the verb follows rather than precedes the negation marker and certain adverbs. For example, in English one says John often eats chocolate, whereas the French say Jean mange souvent le chocolat, literally, ‘John eats often the chocolate’: in English the verb eats comes after the adverb often, whereas in French the order is reversed. Welsh is unlike the other languages in the list in that its verbs come in the beginning of the sentence: Mi brynes i gar newydd is literally ‘Bought I car new’ (a non-literal English translation being ‘I bought a new car’). Finally, French and Spanish differ as to whether the subject may be omitted, if understood from context or the ending on the verb: while Spaniards are perfectly happy with Hablamo español, the French would cringe at the ungrammatical *Parlons français (unless understood as ‘Let’s speak French’). Each of the five languages, though distinct from its neighbors (and in some cases, its relatives) finds typological twins among unrelated languages spoken in far-flung regions of the world. The global spatial distribution of OV languages, similar to Basque, has been discussed and mapped above. English patterns in the relevant respect with Indonesian, an Austronesian language, and Èdó, a Kwa language spoken in Nigeria. Verb-initial sentences, like in Welsh, are found also in Zapotec, spoken in Mexico, and Niuean, spoken on the remote Pacific island of Niue.


*The reason I talk here about “models” rather than “languages” here is that most languages combine elements of these idealized models. For example, English has elements of isolating morphology (e.g. auxiliary verbs to express tense and aspect) and elements of fusional morphology (e.g. the suffix -s in John laughs expresses at once the present tense and the third person singular subject agreement). Even Mandarin Chinese, often held to be the perfect example of an isolating language, has elements of agglutinative morphology in the form of aspectual suffixes -le and ‑guo.

Subscribe For Updates

We would love to have you back on Languages Of The World in the future. If you would like to receive updates of our newest posts, feel free to do so using any of your favorite methods below:

  • Ivan Derzhanski

    That Hausa has come across as isolating in Chapter 20A in WALS is a strange artefact of how the study’s been done. The language has no morphological case, and its tense-aspect-mood (as well as subject person-number-gender) formative is written separately from the verb, but look at how nouns and adjectives form their plurals.

    • Well, it’s dangerous use the WALS database to make conclusions beyond what is represented. They never claim that Hausa is isolating. In fact, they do not treat “isolating” as a primitive category at all…