It seems like Indo-European languages are always stuck between throwing away complicated fusional grammar (like English) or retaining most of it (like Russian). Are there any Indo-European languages that reanalyzed the verb system into agglutination with series of suffixes rather than trying to stuff complex tense, aspect, and person information into a single arbitrarily chosen suffix?

Or is it generally a rule that languages evolve away from agglutination and never towards it? It seems, for example, that Modern Japanese is a bit more fusional rather than agglutinative due to certain sound changes - for example in Classical Japanese "ikitari" went simplifying into "ikita" (which prevented further addition of affixes since -ta is not a verb citation form ending) and then fused into "itta", which obscures the root "ik", thus becoming somewhat fusional.

  • 4
    Are you only interested in languages which are agglutinative in the verbs? I think Armenian is fairly agglutinative, but mainly in the noun phrase, and not much in verbs. I think Persian is fairly agglutinative in its verbal system. May 7, 2013 at 4:42
  • Yes, the verbs. German, for example, agglutinates a lot in the nouns but noun agglutination is an easy gluing.
    – ithisa
    May 7, 2013 at 10:37
  • 6
    @EricDong: Don't confusde agglutination with noun compounds. For German to have agglutination in its nouns would mean, for example, that it had completely separate plural and case morphemes either or both of which may be present or absent but in a fixed order. May 7, 2013 at 22:32
  • 2
    To clear up misconceptions about Armenian I see mentioned here: it is not agglutinative. Armenian uses mostly postpositions, but so do many other IE languages, and there is also German entlang etc and arguably English ago. It also has more IE case system, including locative and ablative. And like in the core (IE) Balkansprachbund the definite article follows the noun and is suffix-like. It is is also fairly regular. So you end up with "(I) car-LOC am" for "I am in the car", which seems agglutinative. Sep 3, 2016 at 12:54
  • 1
    However, it is not extensible, it cannot get deeper than that first level of case ending/suffix. In fact it the article is even lost. The postpositions still require case and are separate words, as is the copula - not suffixes. There was intense contact with Hurro-Urartian and other Caucasian languages so like Hittite or Lydian it could have acquired the thin veneer of agglutination but Armenian is not agglutinative. Sep 3, 2016 at 12:55

4 Answers 4


Among ancient languages, there are certain traits of aggluttination in Hittite (especially in its verbs of mi-class).

If you are interested in living languages, then one of the possible suggestions would be verbs of the regular -ar or -er types in modern Swedish. The -ar- type is a type of verbs of regular declention taking, like the rest of regular Swedish verbs, a similar ending (here, -ar) for all and any personal forms in Present.

Consider e.g. the verb att svara (to answer):

     Sg.                          Pl.

1.  jag svarar                        vi svarar

2.  du svarar                         ni svarar

3.  hen svarar                        de [dom] svarar

NB: hen is a new coinage, presumably a calque from the Finnish hän. Both new Swedish and the Finnish words mean the same: (s)he. The actual and more exact gender-specific pronouns in Swedish are hon for 'she' and han for 'he'.

Now let's make a Past Tense form:

   Sg.                          Pl.

1.  jag svarade                 vi svarade

2.  du svarade                  ni svarade

3.  hen svarade                 de [dom] svarade

Finally, the Passive Form would be

svaras (it is answered = ´svara+ s ') and svarades (it was answered/answering = 'svara+*de*+s').

But there is more of agglunativity in Swedish nouns, though:-)

Consider e.g. the word eld (fire).

The plural form is eld+ar (yes, unlike -or, the -er and -ar endings are applicable for nouns and verbs, although the adjectives have a different plural - or, actually, emphasised - form).

The postpositional definite article is -na; hence, 'the fires' will be

eld+ar+na => eldarna

UPD: the -na here conveys the meaning of both plurality and definitiveness, and can be divided into -n- (postpositive plural definite marker) + a (emphasised form of a noun applicable in plural only).

So, this form of word can be further structured as eld+ar+n+a => eldarna

Morover, there is a common Germanic survival of the Genitive case, which is - suprisingly! - s.

Hence, 'of [the] fires' (like in 'light of [the] fires') will be

[[eld+ar+n+a+s] ljus] => eldarnas ljus

  • 3
    I would like to see an example of Swedish noun agglutination too please. Especially since it's clear the OP was confusing agglutination and noun compounding. Jun 7, 2013 at 1:02
  • 2
    I've added up some information.
    – Manjusri
    Jun 7, 2013 at 5:45
  • 5
    Nevertheless, Swedish is still not an agglutinative language in the strict sense (compare for instance with Finnish). I guess the answer is that that there is no IE agglutinative language, but there are languages that demonstrate certain agglutinative features. PS: This comment is also true for my suggestion below, regarding Lydian
    – Midas
    Jan 23, 2015 at 14:31
  • 2
    One possible problem with calling Scandinavian verb endings agglutinative is the passive suffix -s: in the present tense, it replaces (or elides) the conjugational ending -r, whereas in the past tense, it is added on to the past-tense ending -de. This would suggest that -s has to be analyzed as the "present-passive" suffix in one tense, but simply as the "passive" suffix in the other tense, and this might be enough of a basis to term -s and -r as "fusional" affixes.
    – user8017
    Feb 8, 2015 at 14:47
  • 1
    @Midas How would you define an agglutinative language in the strict sense? As far as I can tell, agglutinativity is not a property of languages but of specific morphemes, and some languages simply have more such morphemes than others. Finnish has some arguably non-agglutinative affixes (e.g. noun plurality is expressed with different affixes depending on the case form), and Estonian is noticeably less agglutinative overall than Finnish, possibly less so than many IE languages.
    – user8017
    Feb 8, 2015 at 14:53

The Lydian language was distinct from all other Indo-European languages for its agglutinative features. I don't know all the details, but there was an extensive usage of infixes. Generally, all Indo-European Anatolian languages had some kind of agglutinative qualities, apparently because of the substratum languages that were agglutinative e.g. Hattic and Hurrian. Nevetheless, that does not qualify them as agglutinative in the same sense as Korean, Japanese, Turkish, Basque, Berber and so on. On the other hand Tocharian languages changed their IE infectional structure into the agglutinative morphosyntactic type with multi-morphemic endings and suffixes e.g. Toch. B cämp-am-ñe-tstse 'to have the ability'.

As a conclusion, there are very few IE-languages that can be counted as having agglutinative features. As a native Swedish speaker I do not consider myself speaking a language with agglutinative qualities in the same extend as when I compare to Lydian, Tocharian or languages from the Uralic, Caucasian and Altaic language families.


As mentioned, modern Scandinavian (Swedish, Danish, Norwegian) seems like a fairly straightforward example of agglutinative IE languages: the only major exceptions to the agglutination in these languages (that I can think of) are the verbs whose past tense is formed with ablaut rather than suffixes.

Armenian (or at least Western Armenian -- I haven't studied Eastern) doesn't seem especially agglutinative in its noun or verb system. Although its nouns do not have gender classes and the plural ending -(n)er is used for almost all nouns (I don't recall any exceptions at the moment), many words have irregular case endings. Its verb system has three main conjugation classes, the personal endings are not the same in the past and present tense, and there are many different patterns in the formation of past-tense stems.


Persian is classified as an agglutinative language. Also, it seems more likely that rather than simply evolving away from inflection to analytic properties, this only occurs when languages come into contact. Take English, for example. Old English had five cases (if I recall correctly) and lost all but the genitive after hundreds of years of contact with French. Likewise, Mandarin Chinese, being a widely spoken language, would have had to be learned by millions of new speakers who, in practise, would not all have used the language correctly. This is why major languages tend toward analytic properties - inflection is used to a minimum by new speakers. In contrast, isolated groups tend toward languages which have a higher morpheme-word ratio. Australian Aboriginal languages are mostly agglunating; these are used only in small groups of people over large areas, meaning there are few second-language speakers. North American languages are another example, as are the Uralic languages, and even Proto-Indo-European, which was theoretically spoken only in a comparatively small region of the Caucasus. Languages can become more inflected through the process of grammaticalisation. A hypothetical example of this would be if English speakers began using prepositions as prefixes, and after many generations these prefixes lost all meaning as independent morphemes, simply becoming locative cases. Therefore, languages (when isolated from contact with other tongues) could in theory tend toward inflection rather than using analytic structures to convey meaning. If I'm incorrect about any of this, correct me. Hope this helped.

  • 2
    Can we be sure that 'contact with French' was the cause of English casting off its case-system? Why not just simplification, or increasing the use of prepositions and word-order to avoid misunderstanding between dialects? Jan 27, 2015 at 16:04
  • 1
    Do you have any evidence for this theory? I've never heard anyone suggest that the level of inflection in a language depends on how isolated it is, and frankly that doesn't make much sense.
    – curiousdannii
    Jan 31, 2015 at 1:12
  • 1
    Well, just one example: Icelandic versus all the other living Germanic languages. Other examples to follow! Feb 2, 2015 at 9:44
  • 1
    @curiousdannii for more on morphological complexity and case studies of its relationship to isolation, see Sampson et al., Language Complexity as an Evolving Variable. Sep 4, 2016 at 23:32

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