It's generally accepted that languages go through a cycle of changes to their morphological type.

English is losing its inflectional endings and becoming more isolating/analytic.

But what about the big well known isolating/analytic languages, are any of them showing signs of moving to the next stage in the cycle? Adopting something like endings that might be counted as agglutination or inflection?

I'm talking about languages like Mandarin, Cantonese, Vietnamese, and Thai.

Is there some research about this or is it maybe that changes at this point in the cycle are not so easy to spot as the changes happening at the point in the cycle where English lies?

  • Err. Embarassingly I cannot give citations to specific literature, but I have heard it discussed that Classical Chinese was more isolating than contemporary Mandarin. One specific example that is often discussed is the particle le, which once was an adverb(?) of some sort and is now a perfective aspect marker. – Aaron Sep 22 '11 at 2:36
  • 1
    it is now definitely an aspect marker. So, it is no longer optional (like an adverb) but obligatory in certain contexts. It also, as far as I can tell, occurs obligatorily post-verbally. I know even less about Chinese phonology than about its morphosyntax, but if le is not already a suffix, the conditions are ripe for it to become one. – Aaron Sep 22 '11 at 3:20
  • 1
    Tachi I believe is a native Japanese morpheme, but it does behave remarkably like Chinese -men: glow.ling.nthu.edu.tw/3-4-2Kurafuji.pdf – ROBOKiTTY Sep 16 '12 at 3:09
  • 3
    It's not generally accepted (at least, in linguistics) that there is such a typological cycle. Please see my comment (and yours!) at 'What is the evidence of the typological cycle theory?'. – Gaston Ümlaut Sep 16 '12 at 23:33
  • 1

Yes, there is some evidence that some of the archetypal isolating/analytic languages of South-East Asia may be developing more morpholexical complexity. However, it seems that there are only early signs, in comparison to the evidence for English becoming more analytic - that's been happening for hundreds of years so it's pretty easy to spot by now.

It's important to consider that although many typical isolating languages seem to lack inflectional morphology, they may still show quite a lot of derivational morphology. In this paper by Enfield (2005), there is some brief discussion of derivational morphology in Mon-Khmer languages, noting that while for some languages, like Khmer, the derivational morphology is visible but no longer productive, in other languages, such as Semelai, the system is quite rich and productive.

Enfield also discusses the resources used by some languages of mainland South-East Asia to create elaborative, alliterative or rhyming expressions, noting that these frequently give rise to morphologically complex structures, and in particular that "a huge system of such patterns is found in Vietnamese (Thompson 1987 [1965]), despite its status as an archetypically morphology-poor language". He also adds that "the productivity and internal complexity of elaborative morpholexicon in MSEA languages should weaken claims that these languages lack morphology. One just has to know where to look."

One example given from Lao is qêêk5-lêêk5 'lying askew like someone asleep in an awkward position'.

I don't know if there's any evidence for the derivational morphology in these languages becoming more complex over time, but for some it certainly doesn't seem to be going away. As far as more inflectional-type morphology goes, there was some work comparing grammaticalization in Mandarin Chinese and Thai, two typologically very similar languages, both at the extreme end of the 'isolating' part of the morphosyntactic structure continuum. In this paper by Post (2007), various types of constructions are examined to assess the depth of the grammaticalization of certain features, given that compounds and polysyllabic constructions are common in both languages and "some affixlike behaviour has been observed". Post finds that, for Chinese and Thai constructions of comparable form and function:

  • "analogous terms grammaticalizing within these constructions have achieved a greater extent of structural adjustment as functors in Chinese than in Thai, i.e. are more 'deeply' grammaticalized", and;
  • "processes of compounding are more extensive in Chinese than in Thai, as measured by frequency of types and tokens as well as factors such as analyzability and semantic drift".

One example is the evidence for becoming generalized as a focused object-marker in Chinese.

Post suggests that the points noted above:

"appear to be symptomatic of a shift in morphosyntactic typology, on the Chinese side more so than in Thai, away from an extreme isolating, basically monosyllabic prototype towards a more concatenative structure. It is possible that both languages are shifting in this direction, and Chinese simply started earlier, or has moved faster, than Thai."

He then goes on to suggest that there might be a stronger correlation between more compounding and 'deeper' grammaticalization; the obsolescence of basic source lexemes for the grammaticalizing forms (the source lexemes typically being replaced by compounds) might actually encourage the partially grammaticalized versions to shed their lexical-like attributes, and allow behavioural restrictions to become more firmly established.

There are other hints of similar observations around, but these sorts of things can be very difficult to tease out in the early (or mid-term) stages, and require a lot of data, so I think it will be some time before we know for sure whether some of these languages are becoming more agglutinating. Seeing as there is really no such thing as a purely isolating language, we also can't know for sure whether some of these languages are on their way to becoming more or less isolating - they might not be ready to move on to agglutinating if they're not done with the isolating stage!


Erhua [1] is a very pervasive example of derivational morphology in Mandarin, most often associated with northern dialects. It's commonly described as a diminutive suffix. Although stereotypically thought of as a feature of northern Mandarin, variants also exist in non-rhotic forms in other varieties [2].


dōngxī [tʊ̃ŋ˦.ɕi˦] 'east (and) west'

dōngxi [tʊ̃ŋ˦.ɕiˑ] ~ dōngxīr [tʊ̃ŋ˦.ɕiɻ˦] 'thing'

The previous example hints at possibly both derivational and inflectional morphology in Chinese, through the neutral tone. The majority of morpho-syllables in Chinese have citation tones, with few that are underlyingly toneless. However, some seem to surface without tone in certain situations. Some Mandarin examples are as follows:

(Cf. Xiaonan Susan Shen, Mandarin neutral tone revisited)

  1. Always toneless

    de /tə/ 'subordinating particle; genitive particle; nominal modifier'

    le /lə/ 'perfective aspect particle'

    ne /nə/ 'interrogative particle; emphatic particle'

    ba /pa/ 'jussive particle'

  2. Sometimes toneless

    guò /kwɔ˥˩/ 'to pass'

    Toneless after a verb:

    (qù /tɕʰy˥˩/ 'to go')
    [wɔ˨˩.tɕʰy˥˩.kwɔˑ] 'I've gone there before.'
    Contrast: [wɔ˨˩.tɕʰy˥˩.kwɔ˥˩] 'I (will) go (and) pass (through).'

    zǐ /tsɨ˨˩˦/ 'child'

    Toneless after a nominal:

    (pán /pʰan˩˥/ 'disk-shaped object')
    [pʰãn˩˥.tsɨˑ] 'plate, saucer'
    (shī /ʂɨ˥/ 'lion')
    [ʂɨ˥.tsɨˑ] 'lion'
    Contrast: [ʂɨ˥.tsɨ˨˩˦] 'lion child'

One generalization here is that toneless morpho-syllables are semantically weak, are phonetically reduced, and generally behave like clitics/affixes/endings. In any case, they are grammaticalized.

Old Chinese might also have been less isolating than commonly thought. I think the consensus today is that Old Chinese derivational morphology, e.g. bound prefixes and suffixes, accounts for tonal correspondence in some word pairs in Chinese today [3] that are comparable to vestiges of PIE ablaut in English, like sing/sang/sung/song. So it might be that Sinitic languages have not moved much on the isolating-synthetic continuum.

It's also important to keep in mind the nature of Chinese writing, which leads many to assume every character is a morpheme/syllable. Abbreviated characters fused from multiple characters have been attested since Old Chinese, and they may not have always represented monosyllabic morphemes. To illustrate with an analogy, in phonemic scripts, even though the ampersand is one symbol, and it can be read as a single phoneme in most Romance languages today, it would be wrong to conclude that the symbol always represented one sound.

All in all, however, the evidence we have suggests Old Chinese was highly isolating, Middle Chinese was highly isolating, and modern Sinitic languages are still very isolating, but less so.

  • "they may not have always represented monosyllabic morphemes" ... is there a single concrete example of this at all? This just sound so out of the blue. – Apprentice Queue Jun 2 '17 at 4:08
  • 1
    It's posited that many Old Chinese morphemes were morphologically and phonologically complex, and some were sesquisyllabic or broke into bisyllabic morphemes. Polysyllabic characters, many contractions, are found in the oracle bone script. Concrete examples are hard to come by, because it's not easy to make them fit all reconstructions. One might work with Zhengzhang but not Baxter-Sagart, for instance. But one example might be sesquisyllabic 角 (Baxter *C.[k]ˤrok) splitting into something like *krok lok and then contracting to *kᵊlok and finally to *kɔk. – ROBOKiTTY Jun 2 '17 at 17:31
  • 1
    Another example with both one-character and two-character examples surviving is 頭 > 髑髏, but the correspondence is somewhat far-fetched in Baxter, where it would be *m-tˤo > *duk lug, but in Zhengzhang, it's *doo > *doog roo. – ROBOKiTTY Jun 2 '17 at 17:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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