My understanding of morphology is that a word is taken and many different words are glued to it.

Is not this true for both agglutinative and polysynthetic languages? Or what is the finer level of distinction?

  • 2
    Can you cite which reference claimed this definition? It's a useless definition and you should stop using such a poor reference. Commented Oct 31, 2014 at 23:28
  • the sentence "a word is taken and many different words are glued to it" is no citation, just the citation of my inner voice that tries to make sense of two definitions read in a very old looking book which pages smelt like cheese from the prussian ära.
    – meireikei
    Commented Oct 31, 2014 at 23:31
  • Analytic languages don't have endings. Inflectional languages pack several bits of information into a single ending. Agglutinative languages build up endings from a series of atomic pieces. Polysynthetic languages join multiples parts of speech into a single word, typically incorporating nouns into their very complex verbs. Commented Oct 31, 2014 at 23:33
  • 3
    So read the definitions again and don't format your misunderstandings as quoted definitions. Commented Oct 31, 2014 at 23:34
  • so the difference between agglutinative and polysynthetic lang. is like this? agglutinative: noun + piece1 + piece2 ....piece51, polysynthetic: piece1+noun+piece2+adjective+piece3+verb+piece4
    – meireikei
    Commented Oct 31, 2014 at 23:37

2 Answers 2


"A word is taken and many different words are glued to it" — that's wrong for both agglutinative and polysynthetic languages.

In agglutinative languages, a string of affixes is "glued" to a root, each affix with its own grammatical meaning, an affix doesn't combine several grammatical meanings, like in Latin 'pueris' (from boys) the affix '-is' means plural + ablative case at the same time. An example of a word of an agglutinative language:

Turkish evlerimden: ev-ler-im-den "from my houses" - HOUSE-plural-my-ablative.case

In polysynthetic languages there are also many morphemes in a word, so that noun, verb and adverb morphemes can combine in one word in such a way that a whole English sentence is needed to translate such a word:

Yupik tuntussuqatarniksaitengqiggtuq: tuntu-ssur-qatar-ni-ksaite-ngqiggte-uq "He had not yet said again that he was going to hunt reindeer." REINDEER-HUNT-future-SAY-negation-AGAIN-third.person.singular.indicative

Polysynthetic languages are often agglutinative.

  • can the affixes within a polysynthetic word/sentence(?) be both a) like in turkish: one affix one grammar category b) like latin: one affix several grammar categories ?????
    – meireikei
    Commented Oct 31, 2014 at 23:53
  • now that you explained it, it is just a guess, but there is a sequence of increased order: 1. polsynthetic 2. agglutination 3. inflection 4. analytic . maybe the first languages uttered were polysynthetic because it is a good way of putting a lot of brain storm into one unit without having yet established more regulated structures.
    – meireikei
    Commented Oct 31, 2014 at 23:56
  • Yes, they can. In the Yupik example above, the last affix -uq combines third person + singular + indicative, on the other hand there are polysynthetic languages like Chukchi which are highly agglutinative with 1 grammar meaning per 1 affix.
    – Yellow Sky
    Commented Oct 31, 2014 at 23:57
  • 2
    @meireikei - Polysynthesis, agglutination and inflection are not opposed to one another, a language can combine them all, English is an example of such a language, and it's also a very analytic language.
    – Yellow Sky
    Commented Nov 1, 2014 at 0:08
  • This might be a broader question than the one in the original post, but what makes it impossible to segment that Yupik form in ways that make it seem much less polysynthetic? For example: tuntussurqatar niksaitengqiggtuq (separating the "reindeer hunt" part from the "say again" part), or tuntussur qatar (if I understand correctly, qatar is the future-tense morpheme in Yupik, and in English, the future-markers will / would are transcribed as separate words).
    – user8017
    Commented May 12, 2015 at 7:00

Agglutination is a form of inflection. So is fusion (aka amalgamation).
The major difference is that agglutinative paradigms are one-dimensional,
while fusional paradigms are multi-dimensional. Consequently one fusional inflection can refer to
many categories (e.g, Latin -tis '2nd person plural subject of verb in present tense, active voice, indicative mood), whereas one agglutinative inflection refers to one category (e.g, Turkish '-im_ '1st person subject' and -iz 'plural subject' -- -imiz together means '1st person plural subject').

Languages with a lot of inflections are called Synthetic languages.
Their inflection may be either agglutinative or fusional.

  • Turkish is an example of an agglutinative synthetic language.
  • Latin is an example of a fusional synthetic language.

Languages that have so much inflection that there is no simple way
to distinguish an inflected word from a clause are called Polysynthetic languages.

  • Lushootseed is an example of a polysynthetic language.

The inflection in a polysynthetic language may be agglutinative or fusional or compound-root.
Generally there's some of each, but there are an awful lot of complex paradigms.

  • In the case of Latin -tis, the relevant variables are "active" (because the 2pl. passive suffix -mini is entirely different) and "indicative" (because the 2pl. imperative -te isn't formed by the regular addition or subtraction of any suffix). By contrast, it seems to me that "2nd person", "plural (subject)" and "present" are not separate variables in Latin to begin with (i.e., they do not have dedicated morphemes anywhere else in the language), so they don't seem relevant for assessing whether -tis is an agglutinative or fusional affix.
    – user8017
    Commented May 12, 2015 at 6:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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