Apologia, using h instead of yhy

[vpbr...@gmail.com]

I was questioned on my usually using "h" instead of " ' " in lojban,
so I would like to explain my reasoning and usage preference,
and I invite constructive criticism, since I am surely open to reason in reconsidering the issue.

First, like every supreme court justice nominee I have ever heard,
I understand the great value of "stare decicis", "let what is decided stand".
Zamenhof laid down the unchangeable Fundamento to avoid the chaos
of endless amendments by creative conlangers.
Still which "decisis" are the ones that must "stare"?

From 1955, TLI loglan had no kind of "h" or "x" until 1981,
when the former was added as a regular consonant phoneme,
and the latter was its special-case colleague.
This is still true of TLI Loglan.
By 1989, LLG loglan/lojban arose with a regular consonant "x" and a special " ' ".

In CLL1, the pronunciation is canonical, even if inexact,
but for the orthography there is a standard form,
as well as two alternatives that seem to be accepted (Cyrillic and Tengwar),
plus the International Phonetic Alphabet, IPA, used as the standard by which you define everything else.
That is four orthographies.

In the usage of other lojbanists I see experiments in orthography,
e.g. the grave accents placed on accented vowels,
or the underdots or over-breves placed on semivowels, or even a few people using h.
I think my usage is not far from the mainstream,
and I coexist happily with those who write differently.

My actual extremest position is to use IPA in a broad transcription when your fonts support it,
but to do CLL with h when ASCII-only is required.
The IPA is a very widely accepted international standard.
An IPA dress for lojban does not look that strange or different; see this example.

Now, which spelling is better and why?

In CLL1 3.3 the apostrophe, period, and comma are grouped together as characters that divide syllables,
but the period and comma are silent ways of separating words or syllables,
while the apostrophe is an audible way to separate syllables
in precisely the same way that "t" separates the syllables in "mlatu".
Nothing special there.

In the same section, we read
    The letter "h" is not used to represent this sound for two reasons:
    primarily in order to simplify explanations of the morphology,
    but also because the sound is very common, and the apostrophe
    is a visually lightweight representation of it.

The sound is not so very common as to justifying it being a special case.
In 130000 words of lojban I can quickly lay my hands on, the letter frequencies are as follows.
55421 i
46986 a
43758 u
38175 o
36048 l
27341 e
26722 n
21790 h/'
20522 c
17803 s
16437 r
14279 m
13755 t
13551 k
10828 d
 9181 b
 8832 p
 7037 j
 5199 g
 5017 f
 4713 y
 4250 z
 4130 v
 3530 x
The h trails behind all the basic vowels and behind the consonants l and n, too, in frequency of use.

Even the measured time it takes to pronounce h, 77-109 msec for me,
similar in published results for other languages, is not like quicksilver.
In this respect, h is one of the faster fricatives, but in the middle of the pack for vowels, stops, and liquids.
So, why does it need a "lightweight" graphical representation?

What about the argument that spelling with apostrophe instead of h
simplifies explanations of morphology?
Except for "e" and "o" having the same phonology rules,
every other pair of letters differs in its allowed usages -- they are all special cases.
The h is far more constrained than the other consonants in its usage,
but it appears in every type of word except the gismu.

I think the principal rule that makes h seem special is the constraint
that brivla have a consonant pair in the first five letters of the word, after excluding h and y.
Still, if in this respect h is not a real consonant and y is not a real vowel,
then why is y allowed in the alphabet, but h not?

In terms of practical convenience, it is nice to be able to search in an editor for whole words
and have the editor software agree with you about what characters occur in words.
Typing /[a-z]+/ is much nicer than /[a-z',]+/.
And who wants to have to hack emacs syntax tables to search for words?

So, after I come to the conclusion that h/' ought to be in the alphabet as much as y or any consonant,
I think about how best to represent it.
The answer to that depends on past usage in other languages and on our desire to take advantage of
habit and familiarity to assist those learning lojban.
All the languages with latin alphabets that I know of that use the h sound
also use the h grapheme to represent it.
French lacks the sound so it uses the letter as a separator.
Spanish has only x which is written j.
Ancient Greek used to have a rough breathing sound represented by
a left-side arc or the left half of capital HTA,
  while the apostrophe-looking mark represented the lack of an h at the start of a word.
Modern Greek has no h sound, just a x.

The use of "h" for the unvoiced glottal fricative seems like a slam dunk choice to me.

mihe la bremenli

[Remo Dentato]

I've been away from lojban from quite some time but I'm still following the language development.
For what is worth, I believe using ' rather than h has proven itself not to bring any advantage.

The reasons brought up in CLL don't really stand, imho.  I don't see how explaining morphology is any simpler or what harms causes 'h' being "heavier" than '.

As you rightly pointed out, it makes writing program that parse lojban text (slightly) more complex that it should be.
Also on any computer system we *have* to use h in identifiers as the single quote has a special meaninge (e.g. jbofihe).
Also, at least to me, the quote visually breaks the continuity of the text and it requires much more focus and attention to read.

By the way, Italian (my mother tongue) has no /h/ sound and this makes difficult for me to correctly pronounce ' . The letter 'h' has no phonological value (is never pronounced) but it's used sa a mark to distinguinsh (in writing) words with the same sound:

   hanno -> they have

   anno -> year

or change the pronunciation of c and g before i and e:

  CIao     tʃ

  CHIave k

  

  GIoco  dʒ

  GHIro  ɡ

That said, I think that slipping toward h and leaving ' as an historical feature would only be beneficial for the language.

muho mihe la remod.

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to loj...@googlegroups.com.
Visit this group at https://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.

[uakci]

also, my 3 cents:

<'> can be used for any other unused fricative sound (as described in the CLL), so it needn't be pronounced /h/ all the time, especially when one's native language doesn't have /h/. then, if we picked <h> to symbolize some non-/h/ sound, would it make sense anymore? (I'm not sure about that, but I guess that the founding fathers of the CLL were sure that using the grapheme <h> would mess things up.)

—mi'e la rupnu be li pi no ci be'o se ju la uakci

[Remo Dentato]

I agree that the initial idea was for ' to be replaceable by other unvoiced fricative sound. Quoting literally CLL 3:

-----

The apostrophe is included in Lojban only to enable a smooth transition between vowels, while joining the vowels within a single word. In fact, one way to think of the apostrophe is as representing an unvoiced vowel glide.

As a permitted variant, any unvoiced fricative other than those already used in Lojban may be used to render the apostrophe: IPA [θ] is one possibility. The convenience of the listener should be regarded as paramount in deciding to use a substitute for [h].

-----

I wonder, however, how many lojbanists nowadays would recognize a ' pronounced as θ rather than h . If, as CLL says, the convenience of the listner should be the first priority, then having ' fixed as  h  might be a good solution.

[Ilmen]

The breathy-voiced glottal fricative [ɦ] has also been suggested (maybe even used unconsciously) as a possible realization of ‹'›. This sound exists in at least some dialects of English as an allophone of ‹h› between vowels (as in "behind" for example).

—Ilmen.

To unsubscribe from this group and stop receiving emails from it, send an email to lojban+un...@googlegroups.com.

[Remo Dentato]

Now that you mention it, I fear that's the way I've always pronounced '.

What would you suggest, Ilmen?

[Ilmen]

On 24/06/2017 18:50, Remo Dentato wrote:

Now that you mention it, I fear that's the way I've always pronounced '.

What would you suggest, Ilmen?

Hmm, keeping pronouncing it as [ɦ]? :-)

To unsubscribe from this group and stop receiving emails from it, send an email to lojban+un...@googlegroups.com.

[Remo Dentato]

:)

Actually, I meant: "what do you think of using h more and more instead of the current quote ' character?"

[MorphemeAddict]

For the reasons given in the Apologia, I'd welcome a change from "'" to "h" IF it were essentially community-wide. It's only the fringe aspect of its use that makes it unappealing. 

stevo 

[Mark E. Shoulson]

I've been saying essentially this for many years.  Thanks.

~mark

On 06/23/2017 06:34 PM, vpbr...@gmail.com wrote:

I was questioned on my usually using "h" instead of " ' " in lojban,
so I would like to explain my reasoning and usage preference,
and I invite constructive criticism, since I am surely open to reason in reconsidering the issue.

.....

[Mike S.]

Nice arguments; it's been hashed for decades.  The Lojban apostrophe is the symbol for a certain sort of semi- or degenerate-consonant sound that is only allowed to appear between vowels, and not near another consonant or a pause.  There is a certain logic to it.  The distribution of <'> is quite distinct from that of other consonants.  Also, there is a competing argument, not mentioned on this thread, that <h> should cover /x/.

Looking at the bigger picture, if there was one thing I could change about Lojban orthography, it'd be the use of <.> as a phonological segment, and the need to place it front of cmevla.  That's a worse aesthetic offense than the Lojban apostrophe ever was; IMHO it's downright jarring.  One might have the wish to back up and change a few things at once.  Most likely it's never going to happen though, since nothing about Lojban orthography is broken in a fundamental way.  It works.  It just pokes a finger in the eye of Latin-spelled languages, especially with that abominable initial <.>.

BTW, I agree that breathy-voiced h for <'> is the way forward.

Best,

-Mike

--

You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to loj...@googlegroups.com.
Visit this group at https://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.

--

co ma'a mke

Xorban blog: Xorban.wordpress.com
My LL blog: Loglang.wordpress.com

[guskant]

Above all, the CLL gives "the standard Lojban orthography" (Chapter 3, Section 1) and some "oddball orthographies" (Chapter 3, Section 12), but it does not interdict any other orthographies. I recognize that the so-called "standard orthography" is only a conventional representation of Lojban texts, and any other unambiguous orthographies are accepted. The convention may change depending on time, place or community.

I accept both {'} and {h} for the vowel separator of Lojban, but I would like to make clear that the separator {'} i.e. {h} has a significant difference from consonants in morphology. 

Le vendredi 23 juin 2017 22:34:57 UTC, Vincent Broman a écrit :

My actual extremest position is to use IPA in a broad transcription when your fonts support it,
but to do CLL with h when ASCII-only is required.
The IPA is a very widely accepted international standard.
An IPA dress for lojban does not look that strange or different; see this example.

Now, which spelling is better and why?

I don't agree to using IPA for written expression of Lojban texts, because a phoneme corresponds to some IPA characters. If IPA is used for written texts, both [sanga] and [sɑŋgɑ] represent {sanga}. Such a variety brings difficulty in searching for words in a text.

 

In CLL1 3.3 the apostrophe, period, and comma are grouped together as characters that divide syllables,
but the period and comma are silent ways of separating words or syllables,
while the apostrophe is an audible way to separate syllables
in precisely the same way that "t" separates the syllables in "mlatu".
Nothing special there.

Here are differences between the vowel separator {'} i.e. {h} and consonants.

1. consonants can follow or be followed by another consonant, for example {st} and {ts}, while {'} i.e. {h} cannot; 

2. consonants can be the first part of a word, for example {t} in {tavla} and {tu}, while {'} i.e. {h} cannot.

{'} i.e. {h} is therefore morphologically a vowel separator, not a consonant.

Because it is only a vowel separator, if a Lojban orthography use, for example, {w} {j} {3} for IPA's [w] [j] [ʒ] respectively, it does not need any letter for a vowel separator. in such an orthography, 

{coi fi'i .i ui e'u do lebna pa lo titspi .i ji'a tcati e'u}

will be 

{coj fii .i wi eu do lebna pa lo titspi .i 3ia tcati eu}.

Similarly, {ii} will be {ji}, and {ji} will be {3i} in that orthography.

In Lojban, any succession of plural vowels like [ii] [eu] are prohibited, so {fii} {eu} {3ia} in that orthography are necessarily recognized as [fihi] [ehu] [ʒiha]. There is no ambiguity. 

# By the way, the consonant separator {y} for [ə] cannot be omitted even in that orthography above. For example, {jesymabru} must be distinguished from {je smabru}.

 

So, after I come to the conclusion that h/' ought to be in the alphabet as much as y or any consonant,
I think about how best to represent it.

As I said at the beginning of this message, I accept both {'} and {h}.

The advantanges of {'} are:

- simplicity in handwriting, and 

- recognizability of a vowel separator among consonants in reading texts; the difference of them are explained above.

The advantages of {h} are:

- simplicity in typing and in treating in programming scripts, and

- harmony with the consonant separator {y}; {h} is just like an upside-down of {y}, and I like that symmetry.

mi'e la guskant

 

[gleki.is...@gmail.com]

Em sábado, 24 de junho de 2017 12:18:40 UTC+3, remod escreveu:

As you rightly pointed out, it makes writing program that parse lojban text (slightly) more complex that it should be.

So slightly that it's just negligible. By the way, most used parsers like http://camxes.lojban.org and derivations of Masato's implementation of xorxe's morphology have supported "h" from the very beginning.

[gleki.is...@gmail.com]

Em sábado, 24 de junho de 2017 01:34:57 UTC+3, Vincent Broman escreveu:

Now, which spelling is better and why?

la zbalermorna :P 

[Remo Dentato]

Let's not forget all the work and time spent on on larlermorna :) To me, the most logical choice for lojban. Not too mention that it could lead to beutiful calligraphy being born from the Elian script

At a certain point in time I suggested introducing an oblique grid for wovels, but it never cought up :)

--

[MorphemeAddict]

I'd rather see alphabets in which the letters are simple, yet as different from each other as possible, not based on a pattern or order or similarity. 

stevo

[Ilmen]

I personally have no problem with Lojban's «'» letter. Many languages written in the Latin alphabet uses the apostrophe as one of their letters, standing for the glottal stop sound.

I'm not either against the usage of «h» in stead of «'» if people prefer writing so. The javascript Camxes parsers have been allowing alternative characters such as «h», «ʃ», «ʒ» and «w» since several years ago.

The only disadvantage of having a diversity of charsets I can think of is that it can make searching for pieces of Lojban text harder.

To unsubscribe from this group and stop receiving emails from it, send an email to lojban+un...@googlegroups.com.

[Gleki Arxokuna]

2017-06-25 15:07 GMT+03:00 Ilmen <ilmen....@gmail.com>:

I personally have no problem with Lojban's «'» letter. Many languages written in the Latin alphabet uses the apostrophe as one of their letters, standing for the glottal stop sound.

I'm not either against the usage of «h» in stead of «'» if people prefer writing so. The javascript Camxes parsers have been allowing alternative characters such as «h», «ʃ», «ʒ» and «w» since several years ago.

The only disadvantage of having a diversity of charsets I can think of is that it can make searching for pieces of Lojban text harder.

That's actually been dealt with. E.g.

https://la-lojban.github.io/sutysisku/en/#sisku/coho_cuhi_ba'a_cu'i 

--
You received this message because you are subscribed to a topic in the Google Groups "lojban" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/lojban/czenfDaq3wc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to lojban+unsubscribe@googlegroups.com.

[iesk]

إی پِهو رُ دُ نا شیسکا لُ لُژبُ فیهُ لِرفو لُ شیزرا گیهِ ژائ شفیپو إی پاؤ نائ لُ فادنی شو ناهِ بانزو موهی ما واؤ زُهُ؟

(.i pe'u ro do na ciska lo lojbo fi'o lerfu lo cizra gi'e jai cfipu .i pau nai lo fadni cu na'e banzu mu'i ma vau zo'o)

موهُ میهِ لا ئِسکُ

[Mike S.]

(In the spirit of tinkering; not necessarily to be taken as a serious proposal...)

On Sat, Jun 24, 2017 at 11:47 PM, guskant <gusni...@gmail.com> wrote:

Because it is only a vowel separator, if a Lojban orthography use, for example, {w} {j} {3} for IPA's [w] [j] [ʒ] respectively, it does not need any letter for a vowel separator. in such an orthography, 

Your basic idea is good, but it's egregious to use a digit as a letter -- especially when there are letters left unused in the Latin alphabet (h, q)!  Instead one could do this:

For [j] use {j}.
For [w] use {w}.
For /ʒ/ use {x}.

For /x/ use {h}.

An approach I like better is this:

For [j] use {y}.
For [w] use {w}.

For /ə/ use {x}.

For /x/ use {h}.

For /ʒ/ continue to use {j} so {.lojban.}'s spelling stays the same.

 

{coi fi'i .i ui e'u do lebna pa lo titspi .i ji'a tcati e'u}

will be 

{coj fii .i wi eu do lebna pa lo titspi .i 3ia tcati eu}.

Similarly, {ii} will be {ji}, and {ji} will be {3i} in that orthography.

{coy fii .i wi eu do lebna pa lo titspi .i jia tcati eu}

{doy .horhes. peipey lo .lojban. tikxgaw to mua mia toy cu lojxpre ji cu cizxpre zoosay}

Using {q} for /ʔ/:

{coy fii qi wi qeu do lebna pa lo titspi qi jia tcati qeu}

{doyq horhesq peipey loq lojbanq tikxgaw to mua mia toy cu lojxpre ji cu cizxpre zoosay}

Ignoring the comma, this system employs all 26 basic Latin letters, plus the space as a word separator.

mi'e la guskant

coo mieq mayk(q)

[Jorge Llambías]

On Sun, Jun 25, 2017 at 4:48 PM, Mike S. <mai...@gmail.com> wrote:

{doyq horhesq peipey loq lojbanq tikxgaw to mua mia toy cu lojxpre ji cu cizxpre zoosay}

je

mu'o mi'e xorxes

Libre de virus. www.avg.com

[Pierre Abbat]

On Saturday, June 24, 2017 1:53:04 PM EDT uakci wrote:
> also, my 3 cents:
> <'> can be used for any other unused fricative sound (as described in the
> CLL), so it needn't be pronounced /h/ all the time, especially when one's
> native language doesn't have /h/. then, if we picked <h> to symbolize some
> non-/h/ sound, would it make sense anymore? (I'm not sure about that, but I
> guess that the founding fathers of the CLL were sure that using the
> grapheme <h> would mess things up.)

There are two words in which <'> was borrowed from /θ/:
fu'arka (rune — not entered because I'm not sure what the place structure
should be)
abata'adj (an alphabetical order of Arabic script, the other being la .abgad.)
There may be others, but I didn't find them.

On Saturday, June 24, 2017 10:36:26 PM EDT Mike S. wrote:
> Nice arguments; it's been hashed for decades. The Lojban apostrophe is the
> symbol for a certain sort of semi- or degenerate-consonant sound that is
> only allowed to appear between vowels, and not near another consonant or a
> pause. There is a certain logic to it. The distribution of <'> is quite
> distinct from that of other consonants. Also, there is a competing
> argument, not mentioned on this thread, that <h> should cover /x/.

The distribution of <y> is also distinct from that of other vowels; it does
not occur within morphemes of a brivla, nor at the end of a word except in
cmavo. However, that of <'> is more different than that of <y>. Also <y> is
the only vowel descender, but <h> is not the only consonant ascender.

> Looking at the bigger picture, if there was one thing I could change about
> Lojban orthography, it'd be the use of <.> as a phonological segment, and
> the need to place it front of cmevla. That's a worse aesthetic offense
> than the Lojban apostrophe ever was; IMHO it's downright jarring. One
> might have the wish to back up and change a few things at once. Most
> likely it's never going to happen though, since nothing about Lojban
> orthography is broken in a fundamental way. It works. It just pokes a
> finger in the eye of Latin-spelled languages, especially with that
> abominable initial <.>.

I find capitalizing syllables in the middles of words more jarring than
periods at the beginnings of words or apostrophes pronounced /h/. I much
prefer the acute accent. Lojban, though, isn't the only language that does
unusual things with capitals. Saanich is written entirely in capitals.

Pierre
--
lo ponse be lo mruli ku po'o cu ga'ezga roda lo ka dinko

[Michael Turniansky]

On Sun, Jun 25, 2017 at 10:32 AM, iesk <pa....@gmx.de> wrote:

إی پِهو رُ دُ نا شیسکا لُ لُژبُ فیهُ لِرفو لُ شیزرا گیهِ ژائ شفیپو إی پاؤ نائ لُ فادنی شو ناهِ بانزو موهی ما واؤ زُهُ؟

  Just for grins, I put this in to Google Translate, which decided it was Urdu meaning "Why do you tell me what I'm doing? I'm sorry, I'm sorry",which I think is pretty funny and apropos.

            --gejyspa

[Remo Dentato]

My suggestion would be to continuining doing what most of us are doing: reading  '  wherever a h is found (also considering that the selma'o names, even if they are not lojban words, use h to signify uppercase ' ).

I don't see any harm in doing it.

[iesk]

Brilliant!

[MorphemeAddict]

The Persian translation was interesting too. 

stevo

--

[vpbr...@gmail.com]

The problem of multiple IPA characters appearing for allophones of a single phoneme, e.g. "sanga" and "sɑŋgɑ", need not be a problem if you use a "broad transcription", i.e. if you neglect unimportant distinctions. This is the sanctioned principle that allows the IPA to be almost the same as CLL orthography.
You can write "sanga" even if you say "sɑŋgɑ".

mihe la bremenli

[MorphemeAddict]

IOW, use IPA for phonemes instead of allophones, IIUYC. 

stevo

--

[Vincent Broman]

stevo,

Yes. The level of broadness/narrowness in the transcription depends on
your practical needs. If you need phonemes, then phonemes are what you use.
The IPA supports recording an insane amount of precision, so you have to
draw your line somewhere on how much precision the situation calls for.

Vincent Broman, mihe la bremenli

> <https://lh3.googleusercontent.com/-QhCJoJuyYo4/WU2TSwKPGII/AAAAAAAAAK0/FGRJOHGog04BlOq-NFyVj9__miCOT0E_gCLcBGAs/s1600/ipa-example.png>

> send an email to lojban+un...@googlegroups.com
> <mailto:lojban+un...@googlegroups.com>.

> To post to this group, send email to loj...@googlegroups.com

> <mailto:loj...@googlegroups.com>.

> Visit this group at https://groups.google.com/group/lojban

> <https://groups.google.com/group/lojban>.

> For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.
>
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "lojban" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/lojban/czenfDaq3wc/unsubscribe.

> To unsubscribe from this group and all its topics, send an email to

> lojban+un...@googlegroups.com
> <mailto:lojban+un...@googlegroups.com>.

> To post to this group, send email to loj...@googlegroups.com

> <mailto:loj...@googlegroups.com>.

[melo...@gmail.com]

Personally, it bothers me far more to see that ⟨h⟩ is used as an uppercase of ⟨h⟩ or ⟨'⟩ for selma'o.