Re: [lojban] Lojban wall of complexity (beginner thoughts)

[Pierre Abbat]

On Tuesday 24 July 2012 17:58:04 Bruno Durin wrote:
> Coi rodo
>
> I am a beginner and I am trying to summarize some thoughts and ideas about
> my experience of my first steps in Lojban. So please bear in mind that part
> of, or maybe all, of what I am saying may be moot because it's only the
> reflect of my low level in Lojban. Please note also that English is not my
> mother tongue (I'm French). I also send this message on this mailing list
> because to my mind it falls in a broader scope that just beginner
> questions.

There is a lojban-fr mailing list you may be interested in. There's also a
recording somewhere of several of us at last year's jbonunsla singing "ni'o
le bi'u cmalu bloti ze'epu naku xe klama". We weren't sure what to do
with "Ohé matelot navigue sur les flots", though.

> So let's imagine I would like to translate the following small text on
> classes to describe what is object programming (this is a very bad
> description, but anyway)
> "Classes are composed of members and methods. Two classes can be linked by
> an inheritance relationship. A class has a constructor, whose role is to
> initialize an instance of the class and a destructor, whose role is clean
> the instance before memory desallocation."
> So how do I decide which part of Lojban to use? Cmavo like "I define", "I
> state", etc.? Or should I look for gismo that express the relationship
> (something like: "members" pagbu "class" .i "methods" pagbu "class")? Or
> use set constructs with "du" or "mintu" and mex to express the fact that
> class = set {members, methods} which could also be a valid way to define
> class at the level of details of my silly example? To which extend should I
> translate computer science words? Use standard morphology as for
> "spaghetti" translation to Lojban? Use "klesi" for class or "le'a" to
> express category? etc. At this stage it seems to me that it's not only a
> question of my better learning Lojban, it's also a question of building new
> vocabulary, of knowing how to choose a construct rather than another
> (finding the right style in Lojban to write about a given field --here it's
> computer science--, as there is a specific style to follow when writing
> mathematics (theorems for example) in English), of understanding thoroughly
> Lojban design not to spoil it (for example by using systematically mex
> constructs to write formal definitions that have nothing to do with
> mathematics, as I could do with "class = set {members, methods}"). To what
> extend do you have to be a linguist and/or a major contributor of Lojban
> (who I am not, not even a computer scientist) to achieve this goal? Do I
> have to wait for Lojban texts on computer science (or juridic texts, which
> is another area where I find Lojban very interesting) to be written before
> writing some myself or am I able to contribute?

This computer-science term "class" is definitely not "klesi". A possible word
for the concept is "gaurtarmi", "a form of instances that do things". Methods
and members could be then gaurpau fa'u cadypau. "lo gaurtarmi cu cerda fi lo
drata gaurtarmi" sounds fine to me. You can indicate a definition
with "ca'edai" (it's a definition, but I didn't define it).

Pierre
--
Jews use a lunisolar calendar; Muslims use a solely lunar calendar.

[la .lindar.]

Regarding the state of most things, especially the website, the problem is that we simply don't have the man-hours to improve anything. We're focused on so much else that it's at the bottom of our list on things to do with what resources we have. There are *thousands* of pages on the tiki and eventually somebody (qualified enough to judge) will have to clean it up. However, we simply don't have enough people to allocate time to that.

[Bruno Durin]

Thank you for your answer!

Le mercredi 25 juillet 2012 04:44:28 UTC+2, Pierre Abbat a écrit :

There is a lojban-fr mailing list you may be interested in. There's also a
recording somewhere of several of us at last year's jbonunsla singing "ni'o
le bi'u cmalu bloti ze'epu naku xe klama". We weren't sure what to do
with "Ohé matelot navigue sur les flots", though.

I was aware of the French mailing list but I think there is less people looking at it than the English one (so less potential answers to my questions, either by asking directly or by googling archives). By the way, I've only managed to get the archive (http://mail.lojban.org/lists/lojban-fr/) that stops in late 2011 with some messages talking about resurrecting the list. So what is its status? Is is still active?

 

This computer-science term "class" is definitely not "klesi". A possible word
for the concept is "gaurtarmi", "a form of instances that do things". Methods
and members could be then gaurpau fa'u cadypau. "lo gaurtarmi cu cerda fi lo
drata gaurtarmi" sounds fine to me. You can indicate a definition
with "ca'edai" (it's a definition, but I didn't define it).

Thanks for the help. I understand the word you suggest and the example you give. In a non artificial language I would have looked in a dictionary and even with a low level in the language I could have build a clumsy but hopefully understandable sentence. Here you made the word for me. To your mind, is a very good knowledge of the grammar and basis words (gismo and cmavo) enough to be able to build new words as you did or are only linguists have the experience and intuition to do this? I don't know whether you are a linguist, I just assume you are, as founders of lojban.

Why klesi could not be used, maybe mixed with skami in an other tanru, or simply with x2 always dropped and x3 giving directly the definition as in "Point is a class with 2 numbers, coordinate x and coordinate y"? Point would klesi1, 2 numbers, coordinate x and coordinate y, klesi3. I am not trying to convince you klesi is the right word, I'd like to understand why you dismissed it. Another idea would be the lojban meaning "a form of instances containing things", highlighting the members instead of the methods as your definition does.

If we imagine that a world for "class" is eventually in the "official dictionary" of lojban in 2050, when lojban is spoken by millions of people (let's dream a bit...), would it be the result of specialist discussions during several years? or the people writing the most on the subject kind-of imposing their choice?

Sincerely,

Bruno

[Bruno Durin]

Something I could do as an unqualified person is to make the list of the questions for which I was starving for an answer and the pages I found useful in tiki. If several beginners accept to do this, this could allow a qualified person to focus on these pages. There is also the option of making a separate tiny beginner focused site, but I already find that there is two many lojban sites, this would make resources even more scattered.

Sincerely,

Bruno

[Pierre Abbat]

On Thursday 26 July 2012 16:33:49 Bruno Durin wrote:
> looking at it than the English one (so less potential answers to my
> questions, either by asking directly or by googling archives). By the way,
> I've only managed to get the archive
> (http://mail.lojban.org/lists/lojban-fr/) that stops in late 2011 with some
> messages talking about resurrecting the list. So what is its status? Is is
> still active?

It's still there, for any francophone lojbanist to send messages to other
francophone lojbanists. The last message (ignoring the monthly reminders,
which on my computer get filtered somewhere else) was sent by Robin in
English to the announce list, which forwards to several other lists.

> give. In a non artificial language I would have looked in a dictionary and
> even with a low level in the language I could have build a clumsy but
> hopefully understandable sentence. Here you made the word for me. To your
> mind, is a very good knowledge of the grammar and basis words (gismo and
> cmavo) enough to be able to build new words as you did or are only
> linguists have the experience and intuition to do this? I don't know
> whether you are a linguist, I just assume you are, as founders of lojban.
> Why klesi could not be used, maybe mixed with skami in an other tanru, or
> simply with x2 always dropped and x3 giving directly the definition as in
> "Point is a class with 2 numbers, coordinate x and coordinate y"? Point
> would klesi1, 2 numbers, coordinate x and coordinate y, klesi3. I am not
> trying to convince you klesi is the right word, I'd like to understand why
> you dismissed it. Another idea would be the lojban meaning "a form of
> instances containing things", highlighting the members instead of the
> methods as your definition does.
> If we imagine that a world for "class" is eventually in the "official
> dictionary" of lojban in 2050, when lojban is spoken by millions of people
> (let's dream a bit...), would it be the result of specialist discussions
> during several years? or the people writing the most on the subject kind-of
> imposing their choice?

You could look up the word in an English dictionary, or a Hungarian
dictionary, or a Russian dictionary and find what you want. If you looked it
up in a Tzutujil or Seri or Navajo dictionary, you would probably come up
empty.

The definition of "klesi" (x1 is a class[...] within x2 with defining property
x3) doesn't fit the computer definition of "class" at all. A CS class is a
type (variables of the type can be created) with some associated methods
(functions) and members (like a structure, not a set). The reason I
highlighted the methods is that without them, a class is just a struct
(although in C++, a struct is just a class with the default being public). An
example of a klesi is the set of languages in which the subject usually
precedes the verb.

Several languages, such as Hungarian and Hebrew and English, went through a
phase in which lots of new words were coined, thus making it possible to talk
in those languages about lots of kinds of things that couldn't be discussed
before. Many of those words (such as "catillate") are now forgotten, but many
remain. Lojban could be at the beginning of such a period.

Pierre
--
The Black Garden on the Mountain is not on the Black Mountain.

[Bob LeChevalier, President and Founder - LLG]

Bruno Durin wrote:
> Thanks for the help. I understand the word you suggest and the example
> you give. In a non artificial language I would have looked in a
> dictionary and even with a low level in the language I could have build
> a clumsy but hopefully understandable sentence. Here you made the word
> for me. To your mind, is a very good knowledge of the grammar and basis
> words (gismo and cmavo) enough to be able to build new words as you did
> or are only linguists have the experience and intuition to do this? I
> don't know whether you are a linguist, I just assume you are, as
> founders of lojban.

As one of the main founders of Lojban, I freely admit to not being a
linguist - never a course in the subject, though I've read a few books
now - but after starting on Lojban. Indeed, since you are bilingual in
English and French (and perhaps others), you are more of a linguist than
I am. (I have some rudimentary Russian, but not nearly as good as your
English, and I didn't even know that when Lojban started.)

It was expressly our intent that ordinary language users make up new
words as needed to express themselves. lujvo compounds supposedly
convey much of their meaning simply by decomposing the compound into its
source roots. There might be some confusion about the place structure
you intend, but that can be clarified by interaction (or in a written
work, one can attach a list of coined words and the intended place
structure).

An ad-hoc "Type 3" fu'ivla borrowing can be borrowed based on a word
from your own language, by lojbanizing the spelling and pronunciation,
and attaching a categorizing rafsi on the front. There are simple rules
for this that prevent word-formation problems, so that it can be done
with little linguistic skill.

ad-hoc words will work for communications. Those words which people
find useful will be reused by others, though possibly modified in form
or place structure. Ideally, a dictionary would include only those
words used by many people, though in starting the language that hasn't
been practical.

So, go ahead and invent words, and just be prepared for someone to ask
you to clarify, if they are unclear in meaning. Even the most skilled
among us will sometimes make poor words, but language is in nature
interactive and we learn from making such mistakes.

lojbab
--
Bob LeChevalier loj...@lojban.org www.lojban.org
President and Founder, The Logical Language Group, Inc.

[Bruno Durin]

Le vendredi 27 juillet 2012 04:03:33 UTC+2, Pierre Abbat a écrit :

It's still there, for any francophone lojbanist to send messages to other
francophone lojbanists. The last message (ignoring the monthly reminders,
which on my computer get filtered somewhere else) was sent by Robin in
English to the announce list, which forwards to several other lists.

Ok, I'll definitely subscribe.

 

You could look up the word in an English dictionary, or a Hungarian
dictionary, or a Russian dictionary and find what you want. If you looked it
up in a Tzutujil or Seri or Navajo dictionary, you would probably come up
empty.

This is a great example. And in these last languages I'd start to build some new word, from other words of the language or by borrowing foreign words or maybe another mechanisms.

And in lojban we have exactly the first 2 mechanisms, lujvo and fu'ivla. Understood.

 

The definition of "klesi" (x1 is a class[...] within x2 with defining property
x3) doesn't fit the computer definition of "class" at all. A CS class is a
type (variables of the type can be created) with some associated methods
(functions) and members (like a structure, not a set). The reason I
highlighted the methods is that without them, a class is just a struct
(although in C++, a struct is just a class with the default being public). An
example of a klesi is the set of languages in which the subject usually
precedes the verb.

I think I've understood, klesi is a class/category etc. defined by a predicate (klesi3), not by an enumeration of its content.

 

[Bruno Durin]

Le vendredi 27 juillet 2012 18:31:19 UTC+2, lojbab a écrit :

As one of the main founders of Lojban, I freely admit to not being a
linguist - never a course in the subject, though I've read a few books
now - but after starting on Lojban.  Indeed, since you are bilingual in
English and French (and perhaps others), you are more of a linguist than
I am.  (I have some rudimentary Russian, but not nearly as good as your
English, and I didn't even know that when Lojban started.)

I naively thought that all major contributors were linguists!

 

So, go ahead and invent words, and just be prepared for someone to ask
you to clarify, if they are unclear in meaning.  Even the most skilled
among us will sometimes make poor words, but language is in nature
interactive and we learn from making such mistakes.

You and Pierre Arbat have greatly clarified the way to go. I just use whatever I like in lojban and my design will be evaluated through the use and feedback of other people.

This is something a bit new for a French native as we have a very strong tradition of normalization by an authority (the French Academy dates back to something like the 18th century if I remember well). As a recent example, a lot of technical words these last 50 years are born in English and the use of the English word have been officially banned and replaced by French substitutes (like "courriel" for "e-mail" because "mail" in French is "courrier"). Of course not everybody uses them, I guess that people with no technical background are more inclined to use French equivalent that people that see the English terms at work or in the course of their hobby.

Anyway, thanks for the clarifications! I go back to the study of the reference grammar...

[Pierre Abbat]

On Friday 27 July 2012 16:03:10 Bruno Durin wrote:
> You and Pierre Arbat have greatly clarified the way to go. I just use
> whatever I like in lojban and my design will be evaluated through the use
> and feedback of other people.
> This is something a bit new for a French native as we have a very strong
> tradition of normalization by an authority (the French Academy dates back
> to something like the 18th century if I remember well). As a recent
> example, a lot of technical words these last 50 years are born in English
> and the use of the English word have been officially banned and replaced by
> French substitutes (like "courriel" for "e-mail" because "mail" in French
> is "courrier"). Of course not everybody uses them, I guess that people with
> no technical background are more inclined to use French equivalent that
> people that see the English terms at work or in the course of their hobby.
> Anyway, thanks for the clarifications! I go back to the study of the
> reference grammar...

I'm not a bat :)

I see English computer terms in Spanish sentences and it just sounds wrong to
me. I'd rather see "logicial" than "software". I've used "corriel" in Spanish
with no problem. I've also used "vepecista" (mail-order seller), which I had
to explain, but the acronym "VPC" makes equally good sense in Spanish and
French.

There is an Icelandic academy, which tries hard to keep non-Germanic
influences out of the language, but the word "fíll" (elephant) is from
Arabic. It entered the language in the Old Norse stage, so it's grandfathered
in.

There is an authority of the Lojban language, namely the BPFK, but we're
currently concerned with updating the CLL, resolving inconsistencies,
defining unclear cmavo, and figuring out exactly what word forms are valid,
rather than forbidding words (on any basis other than phonotactic or
morphological) or coining new terms.

My opinion, which I've broken a few times, is that fu'ivla should be taken
from:
*The six source languages, e.g. makpapi (poppy) from Russian and English.
*Taxonomic names, e.g. tarksako or tarsako (depending on morphology)
(dandelion), from the genus name.
*Sounds, e.g. kuerporuile (whippoorwill) from the Spanish and English words,
which are both imitative. (bois-pourri in French)
*Local languages, e.g. turdunu (bullroarer) from some Australian language. I
tried "turndunu" first, but that's a type-3.

Pierre
--
sei do'anai mi'a djuno puze'e noroi nalselganse srera

[Gleki Arxokuna]

Why from Spanish and English only? All languages have imitative words.

[Pierre Abbat]

On Saturday 28 July 2012 02:15:13 Gleki Arxokuna wrote:
> Why from Spanish and English only? All languages have imitative words.

Because I know that word in English, Spanish, French, biological Latin (not
imitative), and nothing else. Btw the general word for that kind of bird
is "ctecmocpi". The Latin genus name translates to "kanbycko"
or "kanbyladycko" and is a misnomer.

Pierre

--
Don't buy a French car in Holland. It may be a citroen.

[Bruno Durin]

Le samedi 28 juillet 2012 00:19:45 UTC+2, Pierre Abbat a écrit :

On Friday 27 July 2012 16:03:10 Bruno Durin wrote:
> You and Pierre Arbat have greatly clarified the way to go. I just use

I'm not a bat :)

Sorry for misspelling your name!

 

I see English computer terms in Spanish sentences and it just sounds wrong to
me. I'd rather see "logicial" than "software". I've used "corriel" in Spanish
with no problem. I've also used "vepecista" (mail-order seller), which I had
to explain, but the acronym "VPC" makes equally good sense in Spanish and
French.

Sure that's opinions are going to differ greatly from one to another person about this French names (and about somewhat artificial words imposed from an authority).

I think the way the French worlds that I was talking about has been imposed look like we fear to overwhelmed by English culture and as if we were defensive.

When French culture was strong and noblemen (not sure that's the word, I mean "people from aristocracy") used French words that became in English the basis for a lot of cooked food (for example, veal (from "veau") instead of calf the living beast, pork (from "porc") instead of pig), there was not such a resistance from adopting the French sounding words.

 

There is an Icelandic academy, which tries hard to keep non-Germanic
influences out of the language, but the word "fíll" (elephant) is from
Arabic. It entered the language in the Old Norse stage, so it's grandfathered
in.

Interesting example! I keep on thinking that such linguistic "protectionism" only makes the language poorer.

 

There is an authority of the Lojban language, namely the BPFK, but we're
currently concerned with updating the CLL, resolving inconsistencies,
defining unclear cmavo, and figuring out exactly what word forms are valid,
rather than forbidding words (on any basis other than phonotactic or
morphological) or coining new terms.

My opinion, which I've broken a few times, is that fu'ivla should be taken
from:
*The six source languages, e.g. makpapi (poppy) from Russian and English.
*Taxonomic names, e.g. tarksako or tarsako (depending on morphology)
(dandelion), from the genus name.
*Sounds, e.g. kuerporuile (whippoorwill) from the Spanish and English words,
which are both imitative. (bois-pourri in French)
*Local languages, e.g. turdunu (bullroarer) from some Australian language. I
tried "turndunu" first, but that's a type-3.

That's exactly the kind of info that is very useful for me, Thanks a lot! Here are 4 rules that I can apply to build words in lojban, and I'll stick to them until I know enough to have an opinion about them.

Bruno