6

We've had questions about inflected languages moving towards analytic morphology and about isolating languages moving to agglutinating morphology but we haven't yet investigated the third case.

In the typological cycle theory languages slowly move from one morphological type to the next in a fixed cycle.

What are some examples of agglutinating languages such as Finnish, Hungarian, or Turkish merging some agglutinations into inflections?

If not in these, other agglutinative languages or partially agglutinative ones like Japanese verbs. Also if I'm a bit off in the kinds of changes, correcting me on that would also be interesting (-:

Aaron
  • 3,612
  • 22
  • 25
hippietrail
  • 14,687
  • 7
  • 61
  • 146
  • I think it's a mistake to think of languages as being 'agglutinative' or 'inflecting'. Rather, this is a continuous scale that languages move backwards and forwards along over time. There may be languages that have moved to relatively extreme positions on these scales, but they are rare. If you look at some references on grammaticalisation you'll see discussion of these issues. There's a fair discussion on Wikipedia and lots of refs on the web. – Gaston Ümlaut Oct 06 '11 at 10:53
  • Well much of the things in linguistics which have been given names has this issue as was covered in my question about the concept of the phoneme for instance. They are the terms used by linguists though. – hippietrail Oct 06 '11 at 11:47
  • Yes these terms are used by linguists but not the notion that languages can be clearly categorised into one of these four morphological categories. Rather, most languages show varying degrees of the two parameters of synthesis and fusionality in different parts of their grammar. – Gaston Ümlaut Oct 06 '11 at 12:20
  • Additionally, I have not heard of any linguists who subscribe to this 'typological cycle theory'. – Gaston Ümlaut Oct 06 '11 at 12:22
  • @Gaston: Really? Then where did this come from if not from linguists? It seems another question on the topic is in order. – hippietrail Oct 06 '11 at 17:01
  • I believe it was Von Humboldt, in the 19th century. There's a discussion here. But (as that article indicates) it's not something that's widely accepted now. – Gaston Ümlaut Oct 07 '11 at 00:58

2 Answers2

8

Navajo would be one to look at. It is still clearly an agglutinative language, but it's well along the pathway that's hypothesized to lead to inflecting morphology.

The Wikipedia article — which FWIW is pretty decent — puts it this way: "Navajo is an agglutinating, polysynthetic head-marking language, but many of its affixes combine into contractions more like fusional languages."

To pick out an example from Wikipedia: the underlying morphemes di-'a-ni-sh-ł-bąąs, after various phonological rules have been applied, give you the word di'nisbąąs. Just looking at the surface form, you can't segment it cleanly into six morphemes. (There's no clean segment of the surface form corresponding to the underlying -ł-, for instance.) At best you can split it into three "chunks," some of which correspond to several morphemes and contribute several grammatical features to the word.

Eventually, the hypothesis goes, speakers will cease thinking of the word as having six morphemes, and start thinking of those three "chunks" as morphemes in their own right. Keep this up over a few more centuries and you'll have a truly inflecting language, where each word is made up of just one or two "chunks" and each of the "chunks" has a complex meaning like "third person singular past tense."

Leah Velleman
  • 4,585
  • 22
  • 35
6

A nice pair to look at is Finnish and Estonian. They are (very) closely related but due to a relatively low number of mostly sound changes, Estonian has quite moved towards inflection, while Finnish remained agglutinative. For starters, you might want to take a look at H. Metslang's paper A General Comparison of Estonian and Finnish, from which I took the following examples:

part.sg. : part.pl. → Est. lampi : lampe, Finn. lamppu.a : lamppu.j.a

part.pl → Est. pere.sid, Finn. perhe.i.tä

There's a paper by I.I. Popescu and G. Altmann 2008, Hapax Legomena and Language Typology, Journal of Quantitative Linguistics 15/4, 370–78. Unfortunately, they don't give Finnish and Estonian data but if anyone here is good enough at statistics to calculate it, I'd be very interested to find out the results, and how to achieve them.

kamil-s
  • 1,882
  • 16
  • 17
  • 1
    I heard that they were similar, which comes in handy since I'm learning Finnish. But do you mind expanding a little bit? It's ok to take inspiration from the paper you linked (or other books too). I'm asking this because if the paper disappears, your answer might "deteriorate" in terms of quality. Also, remember to quote the entire title and author in your answer of any book/paper (this is for attribution). :) – Alenanno Mar 06 '12 at 10:25