Re: [lojban] idea: selma'o fu'ivla

On 27 July 2013 11:06, Wuzzy <almikes@aol.com> wrote:

> I'd probably define these in Lojban with {x1 se cmavo zo bai .e
> zo ...} and list all the elements + {fo lo lojbo}. The concise
> definition would be be {x1 se cmavo zo bai ra'u fo lo lojbo}.

Almost okay, I would change “lojbo” to something which explicitly
refers to the language. Like “lo banjubu'o”, “lo jbobau” or “la
lojban.”.

Your call, but to be honest, I believe this to be unnecessary as cmavo4 is already a language by the definition of {cmavo}, therefore saying {lo lojbo} in cmavo4 must refer to the lojbanic language and not some merely something lojbanic in nature. This gives rise to the following example: "I'm playing a fighting game" translated to {.i mi kelci lo damba se kelci} is considered redundant because you're overspecifying kelci2. Rather than use kelci twice, we'd say something like {.i mi kelci (la'e|tu'e|zo'e ne) lo nu damba} (or the tanru {.i mi damba kelci}).

I am not sure, but I think that “mu'a” instead of “ra'u” would be more
logically here.

{mu'a} is better, I agree.

Here is my example short definition for BAI:
“$x_1$ se cmavo zo bai mu'a fo lo banjubu'o”

.i'e

Additionally, since the “official” selma'o name is so important, we
should include it in the “notes” field:
Notes: “la'o sy WhATEVER sy cu drata cmene $x_1$” where “WhATEVER” is of
course the “official” selma'o name.

You've made one of the most common mistakes: using la (or one of its friends) in cmene1. This isn't *wrong*, but you're saying that the named-thing is the name. This page jokingly explains this http://www.lojban.org/tiki/White+Knight's+Song+Gotcha

I chose the “sy” delimeter for “selma'o”. Also “sy”
can’t be a selma'o name, because “sy” is already in BY. So this
template should work for all selma'o imaginable and therefore is usable
for automation.

Why not use {selma'o} as the delimiter? It's more obvious, for one,

Thus, I'd write the notes as {.i zoi selma'o BAI selma'o ji'a cmene $x_1$}.

> > UI cmavlrui
> the [u] in {ui} is a consonant, so {cmavrui} is fine.

But it isn’t a stage-3 zi'evla and I want to start with stage-3 zi'evla
for all selma'o. It’s just pure luck that “cmavrui” is at
least morphologically correct, it’s a stage-4 zi'evla. You don’t have
this luck all the time. “cmavri” (for selma'o I), for example, would be
a lujvo. And although the “u” in “ui” is spoken like a consonant, in
morphology, that “u” is still considered a vowel IIRC.

In the old morphology, i.e. the one used by vlatai, the tool jbovlaste uses to determine what class of word something is, yes, it is considered a vowel. But the old morphology is awful for the simple reason that it allows {straeiuaoeiuae} as a valid word.

The modern morphology's principal axiom is that every syllable must begin with a vowel. Thus, when examining a string such as {.aui}, does it break up into {.au-i} or {.a-ui} ? The former isn't possible as the second syllable would begin with a vowel, i.e. /i/. (The first syllable begins with the special consonant {.}.)

All that being said, {cmavrui} in the modern morphology is a type-3 zi'evla.

So far, so good. But now I noticed that the coice “l” + “r” was actually
a bad one. “cmavlra”, “cmavlri” and “cmavlrui” aren’t stage-3 zi'evla,
they are stage-4. Whoops.

I have changed my mind and suggest this instead:

If the selma'o starts with a vowel, use the “r” hyphen and append “xr”
to it and after that the lower-cased selma'o name. So:

“cmavrxrui” → “cmav” + “r” + “xr” + “ui” → clearly selma'o UI
“cmavrxri” → “cmav” + “r” + “xr” + “i” → clearly the selma'o I
“cmavrxra” → “cmav” + “r” + “xr” + “a” → clearly the selma'o A

You can't break “cmavrxrui” down into “cmav” + “r” + “xrui” and
come to the conclusion that it stands for the selma'o XRUI. Because no
cmavo is allowed to start with two consonants.

Frequently, I see people (primarily anglophones :P ) complain that type-3 zi'evla are hard to pronounce. For me, that generally isn't the case, but here you've really done something awful, zo'o. {cmavrxrui} is *REALLY* hard to say. That's five consonants in a row. Yes, the /r/ is syllabic, which makes it easier, but /xrui/ is *not* easy.

My suggestion is to lower the complexity of the consonant cluster, instead using a suffix, namely "vo" from the end of "cmavo", such that:

I: cmavrxivo

A: cmavrxavo

and

UI: cmavrui, since in the modern morphology this is a perfectly acceptable type-3 zi'evla.

1) One cannot think that it's selma'o IVO, or AVO as each are composed of two separate words. Furthermore, the presence of the -vo suffix is indicative enough of the fact the /x/, is not a part of the selma'o name.

2) One cannot think that it's a descriptive brivla, since there is no consonant cluster in the first five letters of the zi'evla-tail.

So I suggest these new slightly changed building rules for selma'o
zie'vla:

selma'o name contains “Y” → use brivla
starts with other vowel → “r” hyphen + “xr” + selma'o
starts with “R” → “n” hyphen + selma'o
starts with other consonant → “r” hyphen + selma'o

Here's my version of the flowchart:

selma'o name contains y -> use brivla

starts with other vowel -> r-hyphen, x-hyphen, selma'o, vo-suffix.

starts with r -> n-hyphen, selma'o

starts with other consonant -> r-hyphen, selma'o

Now the “rxr” idea is much better than the “lr” idea. It shares the
spirit of the “x” convention (thanks for pointing it out) because it’s
very similar to it—it's basicly just the “x” convention with an “r”
appended.

-rxr- would have been good if it weren't the source of an incredible consonant cluster (5 in the case of UI !).

.i mi'e la tsani mu'o