Baxter's Middle Chinese reconstruction has 拉 as *lop, which surprised me given that the Cantonese reflex is laai1.

Other characters with the rime -op such as 合, 答 and 雜 end up with the rime -ap or -aap in Cantonese, so what happened to 拉?

I imagine it's similar to what happened to 內 (MC *nop), but why didn't it happen with other syllables with the same rime?

  • 1
    I would speculate that it's similar to the case of 内 as well. In the case of 内, there were two pronunciations and one prevailed in the end. As to 拉, even in today's mandarin, you can see two pronunciations, la1 and la4. The la4 pronunciation reflected the Middle Chinese *lop, but somehow in Cantonese, this pronunciation lost to the other one.
    – user58955
    Aug 9, 2014 at 14:34
  • You should ask this on linguistics.se
    – Semaphore
    Aug 9, 2014 at 15:27
  • Great question... I was actually going to ask this myself too but you beat me to it. :)
    – Claw
    Aug 11, 2014 at 16:31

2 Answers 2


The qu4 去 tone class in Middle Chinese is generally understood to derive from an OC suffix –s. Sagart regards the whole class as deriving from this process (Roots of OC, p. 131). This results in word pairs of plain root and root + s that in Middle Chinese and later differ by tone.

If the –s is applied to a root that ends in a stop, it seems to efface that stop in later developments. Sagart, citing Baxter, gives these examples (p. 54):

nei4 內 *nups > nwojH ‘inside’, compare na4 納 *nup > nop ‘to put in’

dui4 對 *tups > twojH ‘to answer’, compare da2 答 *tup > top ‘to respond’

In Cantonese, in both these cases, the final stop has been retained in the second (unsuffixed) word, but has disappeared in the first. (In Mandarin you don’t see it anywhere, because all final stops are gone.)

The –s suffix seems to have had a variety of functions. Often it seems to form exoactive or causative verbs – as well as Sagart, Schuessler discusses this at length in the introduction to his etymological dictionary.

Could the Cantonese form of 拉 be the reflex of a word with this suffix, which caused the final stop to be lost?

  • Are you proposing that 拉 originally had an *-s final, and that the expected form in Cantonese should have had the reflex of 去 tone, but (unpredictably) ended up in the first tone instead?
    – jogloran
    Aug 10, 2014 at 8:20
  • Yes, that's right.
    – neubau
    Aug 10, 2014 at 11:45
  • Or more likely, there was a pair of related words, one with *–s final and one without. Somehow in Cantonese it was the suffixed word that was generalized, while elsewhere the other was, so the final stop was retained. (Just a guess!)
    – neubau
    Aug 10, 2014 at 15:41
  • This is an interesting hypothesis, though if it were the case, we'd find some documentation for it. For example, both 內 and 對 are documented in 《康熙字典》 as 去 tones (內:奴對切,餒去聲;對:都內切,𠀤音碓), with the alternative 入 pronunciation being listed for 內 (奴荅切。同納。). The entry for 拉, however, only recognizes the 入 form (落合切,𠀤音菈。讀與臘近。). If the modern pronunciation came from the OC *-s final, you'd think the 去 form would have made an appearance by the time 《康熙字典》 was compiled.
    – Claw
    Aug 11, 2014 at 22:08
  • Right, which is why I'm reluctant to accept this answer even though it's a good guess.
    – jogloran
    Aug 12, 2014 at 2:38

The construction ignores the lanaguage changes on various places. It is for reference only. The entering tone tends to lose when vowel lenghtens. When 'a' lengthens to 'aa' the p is dropped. This happened when Old Chinese evolved into Middle Chinese. So do the evolution of Mandarin.

  • I didn't downvote this, but this is obviously untrue: there are plenty of Cantonese syllables ending in -aap, one of which I included as an example...
    – jogloran
    Aug 10, 2014 at 5:52
  • No, you ought consider the initial as well. It is a pitfall without practising the actual pronunciation and just look at rhyme theortical value. Some combinatios of initial and final do not exist at all.
    – OmniBus
    Aug 10, 2014 at 7:27
  • Again I don't think that's true -- the combination laap exists in 蠟, 臘 and others. I do concede that I can't think of an example that appears in tone 1, but my question would then be whether this change is regular, and if so, what drives the change.
    – jogloran
    Aug 10, 2014 at 7:41
  • Again I have to point out that this is not true. The length 'a' in 蠟 is surely much shorter than 拉. It is better pronounce it and make comparison rather than just looking at these symbols.
    – OmniBus
    Aug 10, 2014 at 9:49
  • 1
    @OmniBus I think you misunderstand what jogloran means by 'strict formula'. Sound changes are generally regular. The diversity of languages comes about because different sound changes occur among different groups, but the changes themselves are generally fairly regular and consistent. This notwithstanding, exceptions can be found. In this case, the Cantonese pronunciation of 蠟 and 臘 came about via regular sound change from OC *lop, but this did not occur for 拉, so jogloran is asking why this is the case; often the answer uncovers something interesting about the history of the language.
    – Claw
    Aug 11, 2014 at 23:04

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.