Indo-European is not a language family known for discontinuous morphology, but there are occasional examples. I can think of two:

  • The German and Dutch past participle formants, ge-en and ge-t, e.g. Dutch ge-nom-en, ge-werk-t.
  • The Ancient Greek present formant -N-an-, with infixed nasal before the last consonant of the root (where phonotactically possible), e.g. lambanō from the root lab.

What other examples of discontinuous morphology in Indo-European languages are there?

  • I guess you could consider ablaut roots discontinuous, like "s_ng" as in sing sang sung.
    – dainichi
    Nov 19, 2013 at 23:09
  • 1
    Arguably the French negative ne ... pas might be analysed as a discontinuous morpheme, but I agree that this would be stretching the definition a bit!
    – Colin Fine
    Nov 20, 2013 at 14:29
  • @ColinFine: You could say ne...pas is not that different from ge...t. On the other hand, then either...or would also count. The difference is that the position of ge...t is far more restricted: ge-stem-t, but il ne me l'a pas donné.
    – Cerberus
    Nov 20, 2013 at 20:13
  • IE is not really my field. With that caveat, I’d like to suggest proceeding by analogy. Find languages outside IE that exhibit rich discontinuous morphology, then look for typological parallels with IE languages. Tagalog is one such non-IE language, and it also happens to be verb-initial. Is it possible that verb-initial IE languages (like Irish) have innovated instances of DM?
    – neubau
    Nov 21, 2013 at 2:15
  • 1
    A quick Google search yields this: “the tense marker [on an Irish verb] may appear either as a suffix, as an initial mutation of the stem, or as a discontinuous morpheme realized with both initial mutation and a suffix (eg mhol+f+ainn ‘I would praise’).” (McAuley ed, “The Celtic Languages” p. 67)
    – neubau
    Nov 21, 2013 at 2:16

3 Answers 3


There is modern English slang that is in-fucking-credible.

Further you could say various forms of reduplication are discontinuous, such as the Greek perfect, where reduplication and perfective suffix must come together or not at all, similar to Dutch/German ge-werk-t. Present stem lu-, perfect stem le-lu-k-, where le...k- together mark the perfect stem. A few verbs also have reduplication in non-perfective stems, but then many Dutch/German words also have ge- in non-past-participles, like ge-voel. To both ge...t and le...k- applies that they are perhaps best considered two morphemes each.

Seperable verbs might also count as discontinuous, but they are certainly not single morphemes.

There is a Dutch h + reduplicated vowel that are sometimes informally used as an infix for emphasis, mainly in speech:

KPN dwingen om de post op zondag te bezorgen? Dat kan toch niet, wat zei ik nou net: KPN is een ondernéhéming, geen staatsbedrijf.

"Force KPN (postal service) to deliver the post on Sundays? You know that isn't possible, what have I just told you? KPN is an enterpríhíse, not a government agency." However, I'm not sure whether this hV should be considered a morpheme rather than just a phonetic bit of...something.

  • The Greek perfect's a good example; I'd never thought of that as a discontinuous morpheme.
    – TKR
    Nov 20, 2013 at 18:29
  • @TKR: One more thing to consider: Ablaut can change the meaning of a morpheme. Or you could say it is a discontinous change, where it changes something in the middle of a morpheme. The same applies to compensatory lengthening. E.g. Greek phên-ai = "to show" (act. aor.), phan-ai = "to appear" (pass. aor.).
    – Cerberus
    Nov 20, 2013 at 18:45
  • Yes, certainly ablaut and other root changes can be seen in that way. What I'm really after (though this wasn't clear from the question) is examples of discontinuous inflectional or derivational morphemes, though.
    – TKR
    Nov 20, 2013 at 19:08

With Lithuanian reflexive verbs, the particle showing reflexivity goes at the end. However, if the verb has a prefix like pa-, iš- etc., the reflexive particle comes between this prefix and the stem. For example:

aiškinti ‘to explain’

paaiškinti ‘to explain a little’

aiškintis ‘to explain oneself’

pasiaiškinti ‘to explain oneself a little bit; (reciprocal) to resolve mutual disagreements’

Is there a case for calling –si- an infix? It seems like a simple question of prefix ordering. But still, in languages like Latvian or Russian with generally similar verbal morphology, this movement doesn’t take place – the Latvian equivalent of the above series would be ‘skaidrot, paskaidrot, skaidroties, paskaidroties’. Or looking at it another way, perhaps we could say that Lithuanian is averse to morphological discontinuity, since it stacks the prefixes in front.

  • This is an interesting affix-ordering phenomenon, but it isn't really about discontinuous morphology, since the reflexive affix is a different morpheme from the verbal prefixes.
    – TKR
    Nov 20, 2013 at 3:42
  • Another infix is in the imperfective stem of some Russian verbs, but again, this is not really an example of a discontinuous morpheme.
    – Colin Fine
    Nov 20, 2013 at 14:27

Some Lithuanian verbs show a kind of vowel grading between past and present tense combined with tense/person suffixation

eg KIRSTI (to cut)

Pres sg 1st, 2nd, 3rd person: kertu kerti kerta

Past sg 1st, 2nd, 3rd person: kirtau kirtai kirto

Since both elements are obligatory for the distinction present vs past you could argue that they are discontinous morphemes,


Pres sg 1st, 2nd, 3rd person: e_u e_i e_a

Past sg 1st, 2nd, 3rd person: i_au i_ai i_o

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.