Lepcha is a language of the Indian state of Sikkim and nearby regions; it’s a Sino-Tibetan language, but its exact relationship to the others is apparently in dispute. It’s spoken by the Lepcha people (self-designation: Róng), and it has its own script dating to the 17th century. Now there’s a proposal for encoding the rather complex script; you can read all about it at abecedaria.
Addendum. Since the discussion in the comments is almost entirely about morae (or, if you prefer, moras), I’ve added “mora” to the entry title. The discussion is over my head but very interesting: thanks, Tim and Suzanne!


  1. I saw that too, reading the Unicode list recently… what interested me more, though (I think I’d seen the previous proposal for Lepcha) is this proposal for encoding Vai. What particularly caught my attention was that one of the references cited is a book by Solomana Kante in N’Ko, and a page from it is reproduced (page 15 of the pdf), so you can see one minority script being used to describe another.

  2. Hi TIm,
    What I don’t understand about the Vai Proposal is this line in the second paragraph of page 2.
    “(Strictly speaking, the writing system is
    based on the mora, as a syllable may be written with up to two characters.)”
    Are all writing systems in which a syllable is written with two symbols now called moraic? I have seen this trend creeping in lately but I can’t make any sense of it. I thought a language had a moraic phonology or it didn’t.

  3. Well, you’re probably at least as well informed as I am, but what I think they mean is that bimoraic syllables are written with two characters, so the number of characters is equal to the number of morae represented. Certainly it looks like long vowels work that way, and this VAI SYLLABLE NG is described as a “syllable final” – if it’s the only allowed coda consonant, the principle of one character per mora holds. The situation’s then very similar to Japanese kana, which is the main thing I’ve heard called a moraic script. But I don’t know anything about Vai phonology beyond what’s in that proposal, so this is somewhat speculative.
    So, anyway, no, I don’t think it’s meant to imply that any writing system where one syllable may be written with two characters is moraic.

  4. Everson’s definition makes it look like that. Henry Rogers new book on writing systems(Amazon has a review) seems to be using it that way. Here are his course notes.
    He lists Mayan, Cree and Cherokee as Moraic. Now Bill Poser is using the term also.
    Read Bridget’s notes which she says are from Poser but he hasn’t pubished them.
    It’s a meme!
    I watched native speakers write Cree, sometimes they just left out the final consonant because it didn’t seem important – hardly a mora.

  5. I’m unclear about something here. I’m not a linguist, but I thought the term mora had more or less been restricted to the phonological elements of a spoken language, rather than to writing systems. Am I wrong? As an obvious example, the Japanese use of kanji wouldn’t suggest a moraic pattern to the spoken language unless you’d been raised thinking of it that way. Are we really using the term mora for writing systems now? Seems like trouble. Writing system and the spoken languages they represent never have more than a tenuous, very contingent kind of relationship.

  6. I agree with you, Matt. Read Gary Feng’s response to Poser on Shadow.

  7. I’m not able to check at the moment, but the numberals seem very similiar to Mobgol numerals. I think that they may be generic Buddhist.

  8. “Mongol.”
    I’d hate to get a hapax legomenon floating out there to confuse future generations. We do not want the Mobgols to join the Moops.

  9. Yes, the numerals are very close. A few are turned 90 degrees, etc. The 8’s are identical, and look like an upside down 7.
    “Numberals” is probably already in the OED, so I won’t apologize.

  10. Numberals gets 931 hits in google!

  11. “Mobgols” gets 4. “Moops” gets 12300.

  12. Matt: To be honest, I’ve only heard the term once, from Poser (talking about kana), until this and Suzanne’s recent postings. So it’s a new phenomenon, using mora to refer to writing systems, and still a minority one. But at least with respect to kana, I think it makes a certain amount of sense.
    (I should note that this is all just the results of my thinking about Poser’s offhand comment about kana – it may not reflect the views of actual theorists.)
    If you look at kana, basically you’ve got characters for V and VC syllables (which are 1 mora each) and characters that let you make long or closed syllables from them ( -n, small tsu, and the length mark in katakana). Obviously “mora” is a phonological notion which is not necessarily going to have an absolutely straightforward relation to the written language, but I don’t see that that’s any less true of “syllable”, which has been used forever.
    The point is, if you say a script is a “syllabary” that suggests one character per syllable, and if you look at kana, you don’t get that. You do get one character per mora, mostly (exactly? can you think of any exceptions?), which suggests that “moraic script” might be a better term. And the mora clearly is a salient concept in Japanese phonology.
    Obviously this doesn’t apply to kanji, which aren’t necessarily either moraic or syllabic.
    Suzanne: Well, I don’t know, does the use of the term “mora” imply it’s a big deal to speakers? Clearly, writing systems don’t always fit perfectly into these conceptual categories we make for them, but if Cree syllabics basically consists of short open syllables and finals you can add to them, seems like under this classification it’s closer to a moraic script than a syllabary, even if finals aren’t universally applied.
    Mayan, that’s a interesting thought, seems to fit, more or less… do you know much about the Maya script? I recently found a good introduction here.

  13. I think the Lepcha numerals resemble the Mongolian less than they do the Tibetan (which they’re both probably descended from – the whole Lepcha script’s closely related to Tibetan, and I’m sure you’re more familiar than me with the history of Mongol-Tibetan relations). And the Tibetan’s recognizably related to e.g. Devanagari, and so on.

  14. Tim,
    If the final consonant does not have a unit of length that compares to the previous CV unit, is it a mora?
    When the term “syllabic” is used for writing systems it doesn’t have to mean that the whole syllable is represented every time but simply that the basic unit of representation must be a pronounceable syllable not an abstract phoneme. It was always recognized that this might involve some analysis or an organization of smaller units into a syllable. It is a continuum of organization.
    Some linguists are acting like they have just discovered this and no one else ever knew it. What I see is a demand for a label for each type of syllabic construction.
    CVC is the whole syllable, C-VC is called onset and rime, so what is CV-C, or CV-V. CV-C was called syllabic and final, CV-V was handled by a mark (so an analytic feature). It seems now that it is called moraic, whether it is a mora-timed language or not. Hmmm.
    Mayan is a little different again. CV1-CV1 to represent the phonology of CVC.NA-NA for NAN, So each graphic unit must represent a pronounceable syllable. Okay if we need some labels let’s find them, but mora? Confusing, no?

  15. If the final consonant does not have a unit of length that compares to the previous CV unit, is it a mora?

    Mmmm… you should really ask a phonologist, but I don’t think the distinction is necessarily one of length. A mora is a unit of prosodic weight, so whether a coda consonant is moraic is a matter of whether it makes a syllable heavy. Syllable weight is something that affects various phonological processes, particularly stress systems. Here’s Dirk Elzinga summarizing the area on the CONLANG list.

  16. Thanks,that was a good discussion on Conlang by Dirk. I read a few posts and overall it sounds as if there is more than one point of view. However, now I am up on all the terms like wts etc. etc.
    Then I put the conlang post into blogger to reformat and accidently posted – oh well. Thanks for a good little discussion Tim and LL for hosting it.

  17. Michael Farris says

    Just a quick note for what it’s worth.
    Many years ago I did a field methods course where Japanese was the in-class language, I (thought I) knew a little about Japanese phonology, but everything I thought I knew ahead of time turned around and bit me on the ass (also with my outside-of-class language).
    Anyway, I’d sort of heard of the Japanese mora and kept waiting for it to appear as something that had to be dealt with in the phonology and it never did.
    I don’t know if it was just the informant we had*, but mora were pretty clearly not being used in speech timing, though there was the intriguing difference between long vowels (as in gyuunyuu) and vowel clusters (as in mizuumi).
    A few years later I read that while mora have some psychological for Japanese speakers (because of the script and it’s metrical importance for poetry) but that modern spoken Japanese was more syllable based.
    From Suzanne McCarthy’s comments, it sounds like mora is a new toy of a lot of linguists and they’re gonna play with it till they wear it out.

  18. Michael:
    Modern spoken (standard) Japanese has, AFAIK, very few phonological processes that pay attention to syllabicity, but several that have to do with moras, particularly accent placement and constraints on minimal words. In fact, there is a “native” Japanese word for mora, namely haku, literally ‘beat’ (scarequotes because this is a Sino-Japanese loan).
    On the other hand, syllabicity trumps moraic weight in some non-standard dialects. Off the top of my head I can’t remember which, exactly, but I think the Okinawan versions of Japanese (i.e., as opposed to ryukyuan) is one such dialect.

  19. Sproat, 2000, makes a reference to Horodeck, 1987, p.33 for labeling Japanese a “moraic” writing system. It is the recent expansion to any system with a CV-C grapheme pattern that has surprised me.

  20. Michael Farris says

    Russell, the speaker we had claimed (for different reasons) both Tokyo and Osaka as home towns, more or less equally. That’s probably why we ditched pitch (for that speaker) as a significant factor since the pitch on solicited words tended to vary from repitition to repitition.
    I’m not sure how I was expecting morae to make their appearance, but that speaker’s Japanese (as presented during elicitation sessions) was syllable based enough so that anything else never entered our deliberations (final n tended to be realized as the nasalization of the previous vowel, hardly in line with the standard description of final n as equal in length with a syllable …)

  21. Re: addendum – To be fair, we did have a few posts about Lepcha numerals. All this stuff about morae pretty much goes over my head too – I more or less know what they are, but I have difficulty answering theoretical questions.
    I thought I might just mention something, for those unfamiliar with morae. The first place I encountered the term was in relation to Japanese poetry. Everyone knows that a haiku is 17 syllables wrong, but in fact this isn’t necessarily true – a haiku is (prototypically) 17 morae long, and so may have fewer syllables.

  22. Tim,
    You mentionned the recent Vai proposal and I jumped head first into the whole issue of morae. I was putting off posting about Vai on abecedaria until I researched more on morae, – there you have it.
    I don’t know Japanese or Vai, but Cree, that is another thing altogether – if someone comes along and classifies Cree Syllabics as a moraic writing system then I want to make sure I understand what that means.
    And I do apologize profusely for saying LL instead of LH. At least I didn’t type LOL. I am prodigiously bad on the keyboard. Now back to More – Thomas More.

  23. BTW I did look at N’ko but too briefly. I will have to go back and give it the attention it deserves.

  24. Hey, I wasn’t objecting — I love it when my posts go off in unexpected directions!

  25. Thanks

  26. As an obvious example, the Japanese use of kanji wouldn’t suggest a moraic pattern to the spoken language unless you’d been raised thinking of it that way. Are we really using the term mora for writing systems now? Seems like trouble. Writing system and the spoken languages they represent never have more than a tenuous, very contingent kind of relationship.

  27. Having studied sanskrit script, prakrit scripts, and tibetan script (I am more of a proficient in writing systems than in spoken language), I agree that Lepcha script is more rightly emanated from Tibetan script than from the Mongolian. While the Lepcha and Mongolian are obviously related (considering they both come from devanagari sources), it makes more sense that the Lechpa utilized scriptual hints from a closer neighbor – the Tibetans (esp. through the spread of Buddhism to non Indo-European speaking areas).
    While the Lepcha script is a bit complex for unicode, I am always pleased to see devanagari script cousins, as these writing systems are so beautiful. Yet, it makes sense that it would become complicated – it always gets hairy when you take a sino-tibetan language and apply an indo-european script. 🙂

Speak Your Mind