Sunday, July 10, 2011

Questions and Answers

Peter Cyrus e-mailed me the following, and I'll try to answer his questions one at a time:

I have a few other questions and comments:

1) I find it confusing that you define liquids and nasals as
"non-consonants".  Why don't you just call the two group "obstruents"
and "sonorants", and put vowels in a third group?  And isn't your
morpheme shape really (S/Z)Ob(Son)V(Son)?

The terminology for this is a challenge, at least for me.  Loglan just has consonants and vowels, so I originally simply defined aeiouywrlmnq as vowels, for the purpose of defining morpheme shape.  Then I realized that it was a little misleading, because of course CV shapes can't include anything but the aeiou as the V. Not to mention the fact that every morpheme has to have at least one vowel in it to be pronounced. So what I've done is fail to keep the morpheme shape rule from getting mixed up with permissible letter combination rules.  

So if we divide everything up into Consonants and Nonconsonants, the morpheme shape is clear.  All morphemes are in the shape nCnN.  A special case of that is the grammar word shape, which is CN.  Now, none of that is taking into account the differences between the nonconsonants and the limits on how they can be combined and still be easily pronounced.  If you or anybody else can figure out a better way to express that, I'd love to clarify it on the Wiki.
2) I didn't understand "A vowel followed by a consonant signifies the
end of a morpheme".  At first, I thought it meant that words all end
in a vowel, a real vowel.  But that doesn't seem to be the case.  Do
you mean that each Consonant (Obstruent) starts a new morpheme?

Each consonant cluster starts a new morpheme.

3) You seem to be missing the ZH sound as in beige or treasure.  Is
that because there's no letter for it?

Yes. I didn't want to have more than the basic 26 letters.  I originally tried using the Loglan phonology, which indicates /tS/ with 'tc' and /dZ/ ad 'dj', thereby having c and j for /S/ and /Z/, but that I ended up rejecting for esthetic reasons more than anything else.  So I didn't have enough letters to reserve one for /Z/.  I know that leaves the phonology a little assymetric, but a look around the world at the big languages seems to show that many languages have either /Z/ or /dZ/, but few have both (even English just barely has /Z/).  And many have both /S/ and /tS/.  And my own subjective feel is that I can easily distinguish minimal pairs in the latter case, but the /Z/ vs. /dZ/ distinction is harder.

4) Attributive adjectives need SA, after the Chinese model.  But
compound words don't - again, like Chinese DE.  What does that SA earn
you?  In English, we distinguish "black bird" from "blackbird" by
stress, and so can you.

Here we have the thing about terseness vs. preciseness.  In spoken Ceqli, you include as much as you need to for clarity.  And you can certainly say blu fawl as distinguished from blufawl.  But blu sa fawl makes for complete clarity.  Also, sa, again like Chinese, can show a modifying phrase.

To go paydey kom sa pan.  The bread I ate yesterday.

You see, Loglan insists on preciseness, slowing down the process.  Ceqli permits preciseness, as needed.
5) Are ZAM and ZMA examples of a system of oppositions?

Yes, but not a good example:)  Inversions should be more clearly distinguishable than that.  So I try to pick vocabulary words that can be clearly inverted when their opposites are likely to be common.

6) Do you pronounce XYEN different from XEN?  That's subtle.  In most
languages, palatal sounds can't be palatalized.  Chinese SH is
retroflex (in Putonghua), but most SHs are palatal, as in English

Yes, it is subtle.  When y is hard to deal with, as in this case, it's permissible to pronounce it as i, but you still follow the stress rules as tho it were a y.

Xyen - shee-EHN.  Xien would be pronounced SHEE-ehn.
7) Your E is mid-open, while your O is mid-closed - that's odd.
Mid-open E as in English "bet" is hard for English speakers to
pronounce at the end of a syllable, since it's a checked vowel in
English, and that letter is reduced (usually to schwa) when final and
unstressed in most other languages - it's not stable.  Meanwhile, many
languages would consider O as in English "boat" to be a diphthong : OW
in your orthography. The mid-open O is like English "ball" or even
"bought".  It seems to me you either want to make both of them
mid-open, never final, and add the OW diphthong, or you want them both
mid-closed and pronounce the E as in English "bait", losing the EY

I don't intend the vowels to be that specifically defined, but the default is the Esperanto values.  Only exception, sort of, is that I emphasize the shortness of the 'e.' That's to clearly distinguish it, as you say, from the ey.  As I understand it, Japanese e behaves pretty much the same way.  I'm aware that it tends to turn to schwa, so I'm trying to avoid it as an unstressed end-letter as much as possible.

8) Go-Zi-Da are your 1st-2nd-3rd pronouns, but Ci-Ca-Cu are your
1st-2nd-3rd demonstratives?  Seems odd.

Yes.  It was a judgment call.  I went with the i-suggests-closeness concept for the here/there/yonder thing, and since I already had the pronouns, I decided to not try to make them suggest each other.

9) Bi and Sta are "ser" and "estar", or "desu" and "arimasu"?  What
does that gain you?

Actually, bi is equals, and I shouldn't have included it in the pictures thing.  I'll take it out in the revision.  Ceqli is like Loglan in that all predicates behave as verbs, so Go jino means I am-a-man.  So there really isn't any word for ser in that sense, except for identity sentences like Go bi Janzo.  If a sentence is reversible without changing meaning, bi is appropriate.

But that points out a glitch I could use some help with.  If the answer to 'what is that?' is

Cuba jino.  That is a man.

What is the Ceqli form for the question?  "Da ka" seems wrong, because 'ka' isn't a predicate.  "Da bi ka?" uses the bi, but the bi isn't called for in the answer. Loglanists/Lojbanists, how is this handled?  I forget.

Hold everything.  I just looked it up.  Loglan has a special 'what' word, 'he', which means "is/does what," so it's a form that behaves as a predicate, despite its shape.  So.

Ta he?  means what is that?

Won't work in Ceqli, because ka is an argument, not a predicate.  I really don't want to use 'bi,' because logically, it's a question that calls for an argument, not a description.

If you want to say

Who/which is that

In Loglan, you do use bi, but with the word 'hu,'  which is an argument.

So if I stick to the Loglan form, which I try to do as much as possible, the only think I can say is

Cuba bi ka?  Who/which is that?

So maybe I need a 'blank' predicate of some sort, that calls for another predicate as an answer, like Loglan does.

Cuba bi ka?  Which is that?  Asking which of a group of known things something is, and expecting a name or a to+predicate as an answer.

Loglanists, help me out here.  Is there any problem with saying:

Cuba ka?

And relying on the fact that the lack of a 'bi' shows that I'm expecting a predicate for an answer?  Or do I need an equivalent to 'he,' to make it clear?  And then will the 'he' word have any other use except in questions like this?  Or should I have an "-um" type thing I can attach that turns the 'ka' into a predicate? And if I did, would it have any other use?

I need help here, most especially from Stivzo.

Hold the presses!  I just thought of Loglan 'me,' which is a prefix that turns a name into a predicate, as in 'melakraisler.'  Why not use that system to turn 'ka' into a predicate?  Just thinking out loud here.  If I used, say 'swa' for that.

Cuba kaswa?  What is that?  Cuba xyen.  This is-a-dog.

Then how the hell do I say what's in the house?

Ka dan dom?  But that calls, again, for an argument, not a predicate.

Kaswa dan dom?  Doesn't do it, because here, 'kaswa' is a predicate, and it doesn't parse, does it?

Now I know I can say 

Dan dom sa ba kaswa?

The thing that is in the house is what?

But that's a lot less natlang sounding than I want.

But sta does mean estar, but it also means 'located at,' so to speak.

Go sta cwaq sa dorm.  I inthebed sleep.
That's enough, maybe too much.

Not at all.  Great set of questions.  You've helped me clarify my thinking.  Keep it up!


  1. So if you used a different writing system, you might have different phonemes?

  2. Maybe. I'd probably have a /Z/, if there was room. It's a judgement call. If I'd gone with the dj and tc system of Loglan, I'd have two extra letters to fool with. OTOH, djV and tcV wouldn't be available as grammar words then. And I figure the borrowing possibilities from English, German, and Chinese make having a /N/ especially desirable. I have kept in the back of my mind that a /Z/ can always be added at a later time if technology makes it easy enough. Going over the 26 letters makes for all kinds of problems in doing the actual entering, not to mention Google searching and alphabetization.