Posted by & filed under Podcast.

We get right to it talking about suprasegmentals: mainly stress, phonation, tone, and nasalization.  After a long and fascinating (if incomplete) discussion, we finally get around to talking to DJP’s Kamakawi.

Top of Show Greeting: Kinál

Links and Resources:

Featured Conlang: Kamakawi


Comment thread on #30

Some comments we mention specifically:


A short introduction to danish numbers.

In danish “halv fem” (half five) is 4:30. Oh yeah, now Bianca found swedish ;)

There’s also a special word for ‘one and a half’ in danish. It’s “halvanden” (half-second), and it has rarely used counterparts: “halvtredje” (half-third = 2.5) “halvfjerde” (half-fourth = 3.5).

And this brings me to the vigesimal part of the numbersystem. The names of the tens 50-90 are derived from the number of scores. For example 60 is “tresindstyve” which breaks down to “three-times-twenty”, though this is a rather conservative way of saying it; it is normally* shortened to “tres”.

Now, 50 is “halvtreds(indstyve)” (half-third-times-twenty = 2.5 * 20 = 50).
70 is “halvfjerds(indstyve)” (half-fourth-times-twenty = 3.5 * 20 = 70).
80 is “firs(indstyve)” (four-times-twenty = 4*20 = 80).
90 is “halvfems(indstyve)” (half-fifth-times-twenty = 4.5 * 20 = 90).

*”normally” means always, except in ordinal numbers where you’d have to say “tresindstyvende”. And these are used a bit more than the english equivalents because we have flipped tens and ones (as in German).

{A relevant illustration}


Hi guys, some random thoughts apropos this episode:

Is it possible that Chinese’s writing system was a sufficiently strong influence to prevent numbers like 12 and 13 from being subjected to the forces of historical linguistics? I mean, would the fact that 12 was written [ten][two] be enough to stop the spoken form drifting away?

Half five: that’s just us Brits lazily omitting “past” :) As Bianca pointed out, Swedish would have it meaning “halfway to five”, and Norwegian, Danish, Dutch and German are the same. Similarly, Jameld has “fëfjel” (five-half), which is also used to mean “four and a half”.

William mentions not needing to use a plural marker after a numeral in Hungarian. Finnish also does that; IIRC it does funky things with the partitive and the verb stays in the singular. Which is fun. Russian is glorious: it uses the genitive singular with 2, 3, 4 and numerals ending in those digits, and the genitive plural for 5–9. So you get one book, two/three/four of book, five/six of books… , 21 book, 22 of book, 25 of books and so on. Truly inspirational :-)

22 Responses to “Conlangery #33: Suprasegmentals”

  1. Anthony Docimo

    So switching would be, using your example, “A man walked into a bar, and the horse kicked him”…
    …in English (without switching) it would be a man as the subject of both.
    …in a switching language, the man is the subject of the first, but the second could have the subject (recipient) be the bar….or an undisclosed individual not yet mentioned.

    • admin

      Not a great example. Your English actually has the horse as the subject of the second sentence. We only assume the subject carries over if the second clause doesn’t name the subject.

      • Anthony Docimo

        My bad. I thought the man remained the subject – even if he becomes a patient – because he’s the one things are happening to. (he’s entering, and he’s being kicked)

        • Ossicone

          Nope. For that to work you’ve have to put the second bit in the passive like you just did.
          ‘He walked into a bar and (he) was kicked.’

  2. Bryan Parry

    Re: Glottal stops and tone in certain American Languages

    I read before that Old Chinese (or pre-proto-old or whatever) had their syllable endings other than r,n,N collapse into glottal stops, and these glottal stops disappeared but left the effect of different tones.

    • Bryn LaFollette

      The process of Tonogenesis in Chinese is a bit more complex, in that there were a variety of factors that led to the respective development of each phonemic tone contour, but the basic premise is of course similar to what happened in the Americas and elsewhere. For anyone interested in the Chinese case, here’s a pretty accessible, and up to date account: (However this appears to simply be an essay, not a refereed publication or anything, so caveat emptor! 🙂 )

      In terms of historical change, the presence of various phonetic characteristics of consonants in either the onset or coda of a syllable can end up having all sorts of effects on the nucleus over time. Just as one example, the general consensus is that Proto-Indo-European has been pretty definitively reconstructed as having as few as two (though possibly three) phonemically distinctive vowels. Yet, it seems that from this very minimal inventory arose some truly massive vowel phoneme inventories, cross-linguistically speaking.

  3. David J. Peterson

    That Danish system is, indeed, completely crazy, but it’s not vigesimal: it’s decimal. The only difference from English is that they name their tens (from 50-90) based on scores. French, on the other hand, is truly vigesimal from 60-99. That is, 60 is soixante, 61 soixante-un, etc., but when you get to 70, it’s 60-10 (soixante-dix), 60-11 (soixante-onze), 60-12 (soixante-douze), etc. 80 is where you get a new base (quatre-vingt), but then from 81-99, it does the same thing: 80-10 (quatre-vingt-dix), 80-11 (quatre-vingt-onze), etc.

    Thanks for featuring Kamakawi! Wanted to add a couple notes on a couple things:

    (1) Pronouns: You guys were too kind, in this case. There is absolutely no excuse for the pronoun system, as it is (regarding gender in the third person, not regarding number). The reason the third person pronouns have so many forms is because Kamakawi began as a Mary Sue language (is that a term we use? If it isn’t, it should be), and I wanted a distinction for everything I could think of. Once I encoded it in the orthography, the forms stuck. If I were to do it over, there’d only be two forms: pea (specific) and kou (general), and the latter would derive from an earlier honorific, and might double as a second person pronoun.

    (2) Audio Files: I got your audio files right here (do your research!). ~;p

    (3) Orthographies: There are two orthographies for Kamakawi: One native, one imported. This has to do with the cultural context. The Kamakawi have lived on the Kamakawi Islands for quite some time, but in recent history, Zhyler speakers moved from the main land and colonized a large island to the northeast (rocky and not very hospitable, which is why the Kamakawi ignored it, for the most part). They brought significant technological advances to the Kamakawi, and brought them a “civilized” writing system that they insist on using, rather than learning the Kamakawi “picture writing”. The native orthography is described here, and the imported one (think Pinyin. It’s a modified version of the Zhyler orthography) is described here.

    And kind of a general comment:

    (4) Syntax: Back when I was starting out, I had the same view of language and linguistics as most ling. undergrads: syntax trees, morphemes, discrete phonemes, etc., and my languages (and their descriptions) reflected that (and suffered for it). Over the years, though, I’ve come to share the opinion that (I think) many in the field are beginning to adopt: that there really is no formal difference between morphology and syntax. Some have taken this to mean that there is no lexicon, and that word-internal rules look pretty much like syntactic trees (see Distributed Morphology). I’ve taken the opposite route to the same conclusion: the lexical machinery that derives words also derives larger units. The goal of both approaches (or one of them, anyway), though, is to eliminate the formal distinction between morphology and syntax. This is why I don’t fret if I don’t see a section describing the “syntax” of a given conlang in someone’s grammar: they’re on the right track. Besides, usually what you see in those sections is just some comments on word order and how to form questions or topicalize a noun… Such things would probably be better discussed elsewhere.


    Keep up the good work!

    • wm.annis

      Regarding point (4), I’m not sure I’m quite prepared to commit to the idea that there is no formal difference between morphology and syntax. Regardless, I’m curious to know where you think a discussion of, say, topicalization and word order [i]should[/i] go, if not in the exiled chapter on syntax?

      • David J. Peterson

        It depends on the language, of course, but usually word order is at least partially dependent upon predicate choice. I’d discuss it there. Or when discussing the order of other elements (e.g. one also has to talk about the ordering of adjectives and nouns, nouns and adpositions, etc.). It probably should go with the discussion of verbal morphology.

        As for topicalization, it depends how topicalization is achieved. If it’s via some particle, it should probably go with the discussion of like particles. If it’s purely word order, I’d put it with verbal morphology, again. It all depends on how the language works.

  4. Bryn LaFollette

    On the topic of Stress and Pitch Accent, something that might be interesting to the conlanger is the cool ways in which lexical, which is to say phonemic, accent can be used. While some languages’ patterns are fixed (e.g. stress in Finnish is always on the initial syllable, in Quechua it’s always on the penult), many have so-called “free” systems and can use accent phonemically in interesting ways. While there is often a default pattern to where accent can fall when not otherwise specified in these languages, it can be lexically specified pretty much anywhere on the word. But, the cool examples, in my opinion, are the ones in which the accent placement is used for, say, distinction of grammatical category. One need look no further than English for contrasts like CONvict vs conVICT. (i.e. noun vs verb, although, of course, at this point this distinction isn’t exactly productive in English anymore, and is somewhat confused by English’s new to-hell-with-contrastive-category morphological ability to just use any noun as a verb, but hey, what can you do?), but the cool thing here is the idea of specific patterns used to mark grammatical distinctions.

    If anyone is really (and I mean really) into learning how stress works, you could check out Bruce Hayes’s excellent book Metrical Stress Theory, but this may be a little over the top in terms of the level of detail most care about, but it does introduce the really interesting concepts such as Iambic vs Trochaic prosody.

    Also, the notion that might be worth playing around with, is that accent systems are really just a sort of abstract pattern that gets laid over a given prosodic unit, which can be a word or whole phrase, but the phonetic realization of this accent may be rich open ground to play with! Although it’s not really entirely accurate to say this, pitch versus stress accent could be looked at as different realization strategies for an accent pattern. There are good examples of lexical pitch accent patterns, like Japanese, where the pitch accent isn’t really a marking of tone difference on a single syllable, but rather a sort of pivot point upon which a larger pattern is anchored. In Japanese, at least in Tokyo, the basic pattern is that the first syllable (and I’m using syllable for accessibility even though the unit is more accurately the mora) will be low tone with all syllables after up to (and including) the one marked with the pitch accent will be high tone, then all syllables following will be low. The interesting thing is that not all words are marked with a pitch accent syllable, and certain particles behave as if they are part of the word for purposes of realization of this pattern. So, in a word with no pitch accent you’ll have a pattern of low tone first syllable then high to the end of the prosodic phrase, but in a case with accent on the last syllable, the word would be identical to one with no accent, except that a in the former case certain following grammatical particles will be low tone, but they would be high tone still in the latter case. And, of course, if accent is on the first syllable, then the word comes out with high tone first syllable, with all following low tone. The point of this excessive description is that one might imagine a naturalistic hypothetical language that, say, replaced this high-low tone contrast with instead Modal vs Creaky Voice or pharyngealized/non-pharyngealized vowels or who knows what!

  5. CMunk

    Thank you for taking up my feedback!

    To answer some of your questions: Yes, it is C-Munk; Yes, I did write it while I was listening; and yes, your Danish pronounciation was way off. And to be honest anything else would have surprised me 😉

    Talking about suprasegmentals, Danish has stød.

    • pá mamūnám ontā́ bán

      Russian also has a special word for one and a half, “полтора́” (m/n) “полторы́” (f), which weïrdly also only has one oblique case form – “полу́тора”. “Полтора́” is also used in the compound number for 150 – “полтора́ста”.

  6. pá mamūnám ontā́ bán

    Czech doesn’t have a pitch accent, the only Slavic languages that do are Serbo-Croatian and Slovene; as do both Latvian and Lithuanian. I think Welsh has some sort of pitch accent too, but it doesn’t affect meaning or anything, in which case it might be on the way out – or I could be completely wrong.

  7. pá mamūnám ontā́ bán

    Also, Portuguese does nasalise the high vowels /i/ and /u/ but keeps the nasal consonant, at least in European Portuguese which I’m most used to hearing, e.g. cinto [sĩntu] “belt” and fungo [fũngu] “fungus”.

    Portuguese has nasalisation all over the place: [ɐ̃], [ẽ], [ĩ], [õ], [ũ]. Some dialects even nasalise the approximants /w/ and /j/ in nasalised diphthongs, e.g. cão [kɐ̃w̃] “dog” and cães [kɐ̃j̃ʃ] “dogs”.

  8. MBR

    I’ve been playing with a suprasegmental that I’ve been calling “cookie voice” for one of my fantasy languages, Kûzurl (the circumflex marking this phonation type). It’s essentially a death metal growl. So far, it’s not very practical unless it imparts harsh voice on the vowel in the following syllable. I’ve been googling my butt off looking for examples of strident voice to compare and contrast, but sadly, I come up empty handed.

  9. MBR

    I just found an in-the-wild example of double aspiration in my Greek textbook: φθόγγον, which is glossed as “the sound.” Since we’re learning Koine Greek, though, we pronounced it [fθɔŋgɔn]

    • Wm Annis

      The UCLA Phonetics Lab web site has some nice examples. For example the very first word list for Armenian (there are MP3s and web pages of the word lists) has an aspirated cluster in the second word example.

  10. Daeiribu

    French nasalised its high vowels but then lowered them (or centralized them, I’m not an expert). I just wanted to point this out as an option to do something other than nasalizing e’s and a’s, but not i’s and u’s.

    • admin

      Absolutely. In fact, inspired by French, I have added a similar historical change in Pahran (my current work-in-progress), which gives me a second source for mid vowels (Pahran grows from a three-vowel system to a nine-vowel system with Turkish-esque vowel harmony). When I get to deriving dialects from Middle Pahran, one of the main differences will be whether or not nasalization is lost, leaving behind long mid vowels.

  11. Qwynegold

    So how do you pronounde that word? Because you keep saying “supersegmental”. And Latec and Lick instead of LaTeX and LyX. :S

    Btw, isn’t it a myth that high vowels repel nasalization? Because that’s what happens in Europe, but then you have all these languages that have a nasal counterpart for every oral vowel.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.