Received-SPF: none (google.com: lojbab@lojban.org does not designate permitted sender hosts) client-ip=68.230.241.214;
Message-ID: <538215D3.5050703@lojban.org>
Date: Sun, 25 May 2014 12:09:55 -0400
From: "Bob LeChevalier, President and Founder - LLG" <lojbab@lojban.org>
Organization: The Logical Language Group, Inc.
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:24.0) Gecko/20100101 Thunderbird/24.5.0
MIME-Version: 1.0
To: lojban@googlegroups.com
Subject: Re: [lojban] Balningau: The Great Update
References: <53800942.9030407@lojban.org>	<CAO7bV+j4=md2oKaN_i6u68YS7hEzjt_iUkqGkiADRCz21NF_Ww@mail.gmail.com>	<1400948010.93944.YahooMailNeo@web181103.mail.ne1.yahoo.com>	<CAO7tK2cuBS+5bqDuf_fFYid5fRHepXK_S6_9cebqSZ0gZcBgMw@mail.gmail.com>	<53D0F569-A434-4953-9A1A-3F8D2001A8BB@yahoo.com> <CAO7tK2cjetXAJ+D8zjNqRpseindts2=hRkNJrh5giO6rYAvY0Q@mail.gmail.com> <5380F63E.3070006@gmx.de> <5wnJ1o00G56Cr6M01wnKXB> <5381C5B0.1000304@lojban.org> <6BKF1o01b56Cr6M01BKG8J>
In-Reply-To: <6BKF1o01b56Cr6M01BKG8J>
Reply-To: lojban@googlegroups.com
Precedence: list
Mailing-list: list lojban@googlegroups.com; contact lojban+owners@googlegroups.com
Sender: lojban@googlegroups.com
Content-Type: text/plain; charset=UTF-8; format=flowed
X-Spam_score: -0.0
X-Spam_score_int: 0
X-Spam_bar: /

On 5/25/2014 7:18 AM, selpa'i wrote:
>>> Lojban is marketed as "simple and easy to learn".
>>
>> By the standards of languages, it is.  Of course, in point of fact, if
>> you are trying to learn Lojban to a level of fluency that requires 30000
>> distinct concepts to be labeled clearly (i.e 30000 brivla with known
>> place structures), you probably won't find that to be "easy", no matter
>> how regular the lists are.
>
> What's easier to memorize: a gismu that follows a common pattern or one
> that doesn't?

Neither.  Easiest is a gismu that you use, and you tend to memorize only 
the places that you actually use and that you hear others use.  25 years 
now, and I never made any effort to learn place structures systematically.

For a language user, rather than a designer, why would you choose to 
memorize a gismu (place structure) that follows a common pattern but 
which never actually comes up in your conversation, over a useful brivla 
where the place structure matters?

> When there is a semantic group of, say, 20 gismu

what makes them a semantic group?  Such considerations were meaningful 
when we were trying to initially figure out place structures,  But now, 
with the language complete, I would try to avoid grouping words 
semantically (as I said before, usually such "grouping" is really on the 
x1 of the gismu and not on the gismu itself - otherwise all words with 
"under conditions" places are equally a "semantic group" as all words 
with a type of animal in x1.  But who would try to memorize all gismu 
with an "under condition" place?

Sometimes a semantic association might make memorizing a gismu or its 
place structure a little easier, but which such associations are 
important is purely individual.

>> Meanwhile, the very first change you make to the existing gismu list, no
>> matter how regularizing it may seem inherently makes the language HARDER
>> to learn because you have potentially invalidated all prior use of that
>> gismu both as an individual words and as a component in a lujvo.
>
> No.
>
> First of all, lujvo don't change, since they mean what are defined to mean.

Who decides what they are defined to mean?  Most lujvo are invented and 
used ad hoc with no one bothering to define them or their place 
structures. When their place structures are defined, as likely as not it 
will be by someone who did not coin the word, and perhaps someone who 
does not know how it has been used.  This has been especially true since 
a distinction was realized between making place structures according to 
some system/rules vs more ad hoc methods (which might include basing 
them on arbitrary semantic groupings as you wish to do for gismu).

In reality, the meaning of lujvo is not and cannot be prescribed.  And 
to a large extent, gismu cannot either.  Words mean what people actually 
use them to mean, not what someone writes that they should mean in a 
dictionary.  Thus gismu place structures should only be changed if 
someone observing many years and many Lojbanists using a place structure 
different from that prescribed.

> Secondly, when a speaker always has to skip around a place (e.g. {broda
> fi ko'a} for skipping x2) because they never need that place, then that
> is an annoyance.

Tough.  Be annoyed.  Or perhaps start using the place you've been 
skipping - you know: allowing the language to structure the way you 
think about things.  The language was after all originally designed to 
test the Sapir-Whorf hypothesis.

Of course it would be easier to coin a lujvo with the place omitted or 
in a different place, and memorize that instead of the gismu with the 
place structure you dislike.  In actual usage, gismu should not be more 
privileged than lujvo.

Which incidentally brings up another matter.  To some extent, the place 
structures were NOT intended to reflect to most useful form for use as a 
bare gismu, but rather, we were trying to think about how words would be 
used in combination, and especially in lujvo.  And we were also trying 
to use the place structures as a defining tool.  Most people won't use 
"under conditions" places or "by standards" places in most of their 
usage.  But it is useful to embed in a predicate word referring to a 
liquid, that the conditions determine whether it will in fact act as a 
liquid.  And whether something is "good" or not depends on the standard 
(morality? or perhaps benefit), and possibly the person doing the 
evaluation because goodness itself is subjective.  One could claim these 
places aren't needed because in natlangs, they seldom are mentioned. 
But Lojban is NOT a natural language, and we don't rely on natural 
language conventions if possible.

Your "semantic groupings" sound very much like your personal natural 
language conventions.  I rather suspect that a native speaker of a 
language quite unlike yours would consider different "semantic 
groupings" more important than yours, and perhaps will find places 
useful that you prefer to skip, because in their native language, 
different assumptions about the world have shaped the meaning of words.

>> You
>> have made someone who knew that gismu less knowledgeable (and whether
>> you have experienced it or not, the relearning of language changes is
>> among the more difficult parts of learning a new language.
>
> They need to relearn the gismu, yes,

Why should they?  And in particular, what gives a newbie like you the 
right to tell them that they should?

>but it will be a simpler definition.

Only according to your specific, natural-language-biased assumptions.

We've had the debate before, countless times, over place structure 
minimalization, maximalization, strongly regularized, etc. in cluding 
several times before the gismu list was baselined.  Different speakers 
were involved, and results were, umm, inconsistent, as you can see by 
the fact that you find the current set inconsistent.

It is pure hubris on your part to think that you, and your group of 
fellow travellers are more insightful than the people who came before you.

I learned better a long time ago (actually I learned it while still 
working on TLI Loglan before the split occurred - yes, they also had 
debates about "regularizing the list", and JCB himself was one of the 
worst - go look at the TLI word list and see what sorts of lujvo he made 
using madzo (our zbasu) and durzo (our gasnu) many of those came about 
by simplifying and regularizing, and incidentally ignoring the original 
meaning of the gismu in favor of a meaning implied by an English gloss) 
http://www.loglan.org/Loglan1/app-f.html).

> Also, if we also take future Lojban speakers into account,
> it's more desirable (in my opinion) to hand them a consistent and
> easy-to-learn gimste,

in *your* opinion consistent, and in *your* opinion easy to learn.  Of 
course you have no actual basis for that opinion, only some untested 
assumptions about what sorts of things make learning easier, and place 
structures more consistent.

  and they *won't* have to relearn anything.

Of course they will.  You think you will be the last person to come 
along and argue for a new improved gismu list?  This comes up every few 
years.  And if we ever said "yes" to a single one, we surrender all 
moral authority to oppose the next dozen attempts.

This one I also learned a long time ago, luckily schooled by other 
Lojbanists.  The original set of rafsi assignments was based on word 
frequencies and usage of gismu in proposed lujvo up through 1991.  I did 
a nice systematic study and assigned rafsi to give the shortest and best 
words based on the then-current list.

In 1994 after much more usage, I did the same analysis, finding that a 
couple hundred rafsi should be reassigned.  But the list was baselined, 
so I put it to a BPFK like committee.  They rejected most of my 
proposals, and I am glad they did.  Try reading any pre-1994 Lojban, and 
you'll find it rather hard, simply because a few percent of the rafsi 
changed.

Now envision your future Lojban student trying to read any of the 
megabytes of Lojban text in the current corpus.  You want to throw out 
25 years of usage history by hundreds instead of just 5 years of history 
by a couple dozen.

>The
> number of people right now who are fluent in all the sumti places of
> every single gismu is close to zero.

Good.  I wouldn't want them to have wasted so much time trying. 
Assuming "fluent" is meaningful the way you used it there.

Let is say that they are.  Then the issue is that number of people who 
are "fluent" in all the sumti places of all the lujvo in jbovlaste 
(which is of course a small subset of all of those in the corpus) is 
even more certainly zero.

Knowing all of the place structures is only slightly more useful than 
memorizing the OED (or similar large dictionary).  "slightly" more 
because gismu place structures are useful in highly analytical 
lujvo-making according the the jvojva "rules" in CLL. (but those rules 
are merely a convention, and one I don't think has been carefully followed).

> What the gismu are about remains
> the same, and some details which most people never even got familiar
> with are adjusted. The practical impact is much less drastic than you
> make it sound.

The practical impact is that the resistance to revising the gismu list 
every time some new reformer like you comes along goes away.  And old 
Lojban text in invalidated to some unknown extent.  And the result isn't 
really any better than the old list because people shouldn't be wasting 
their time memorizing all of the gismu place structures

(there is somewhat more limited benefit to knowing all gismu at the 
keyword level, and most or all of the rafsi, because it tells you how 
the wordspace is filled and how rafsi-space is filled and thus makes it 
easier to decode a new lujvo that you don't know the meaning of.  If you 
know the rafsi for sralo, you won't accidentally interpret that rafsi as 
meaning something else, and since rafsi space is so crowded, knowing 
some of the rafsi makes it enormously easier to learn the rest, merely 
by elimination.

No such factor motivates place structure memorization.  You learn them 
by using them, as you need them.

>>> The simpler and the more consistent the gimste, the higher are the
>>> chances for the average person to learn Lojban  and the more pleasant it
>>> is to be a user of this language, I believe.
>>
>> Those claims are merely that: claims.  Unsupported by actual evidence.
>
> Experience, both my own and those of other jbopre I interact with. We
> use the language daily,

Whoopie.  Anecdotal evidence based on personal experience.  Let me know 
when a linguistics journal accepts your paper based on that "experience".

>and making slight adjustments in gismu place
> structures results in a big increase in pleasantness of use. This may
> not be the case for you, but it is for some.

And why should your personal aesthetics preferences count more than mine?

>> And it is important that this be so, because too much semantic emphasis
>> on the x1 risks losing the predicate nature of the language.
>
> But they are not limited to x1. And they couldn't be, since Lojban
> sometimes puts the experiencer in x2 and someimes in x1. We are able to
> look past x1 and figure out what a gismu is about.

Sometimes.  And sometimes the semantic experiencer is in x3 or x4 or x5. 
  And probably in some lujvo, in x8.

And it is just as fundamental to understanding Lojban conceptually that 
a beginner be able to cope with an experiencer in x8 as in x1 or x2. 
(Of course a beginner is far less likely to run into such a word these 
days.)

>>> That way it becomes much easier to get related gismu to align.
>>
>> All gismu are "related".
>>
>> Why should some "align", and not others
>>
>>> For example, all the gismu about emotional states could go together.
>>
>> They could.  But what about the words that you don't recognize to be
>> about emotional states
>
> I think we can agree that {klama} is not an emotion, whereas {badri} is.

klacni (or maybe klaselcni) would probably be an emotion (the emotional 
reaction to going somewhere, which reaction might be dependent on the 
route and means).  It might or might not have a place structure similar 
to badri.  cricni is even more recognizably an emotion (English gloss 
"loss") and even more likely to have a different place structure.  But 
we aren't going to try to systematize all lujvo that are used to talk 
about emotions, so why do so for gismu.

Gismu are NOT semantically privileged in Lojban.  They are 
morphologically privileged in having rafsi, but not in any other way. 
And most people learn a lot of gismu relative to lujvo when first 
starting, but that is likely an artifact of how the language was 
designed.  I rather suspect that Robin's kids recognize or attach 
significance to gismu vs lujvo, and you probably shouldn't either, but 
no one has written textbooks that reflect this fundamental truth.

>>> As part of the revision, I would also like to define each gismu well
>>> enough that we can come up with examples filling *every* sumti place of
>>> every gismu.
>>
>> That would be an interesting but probably unproductive challenge.
>
> Then you aren't aware of one of the most common requests by beginners I
> hear. They want examples, they want to know how to use a gismu. (to
> which you will reply again "it's used how people want to use it")

I am quite aware of the requests of beginners.  I have after all been 
teaching the language longer than anyone else.  And, I don't reply that 
way.  If someone is a beginner, I wouldn't be trying to explain the 
language using gismu that you don't know how to use.  Beginners aren't 
going to be able to use the whole language with facility.

>>  > so we should either figure out what it means
>>
>> Worthwhile goal, but not a very high priority one when we have cmavo
>> that people haven't figured out.
>
> Not a high priority for you, but for many others. Lojban is not only
> made up of cmavo.

The rest of the language has been defined to a certain level.  But be 
that as it may, we cannot adopt everyone's priorities.  Nor can we 
respect them if they run counter to our own.

>> Of course, since the language is intended to "go feral" and cease to be
>> under prescriptive control, it can be argued that we already have
>> conceded that time-free sense in incorrect.  Or perhaps any apparent
>> rule which does change through usage is therefore not-a-rule.  In which
>> case we may never know "the rules of the language" as long as there are
>> Lojban users.  (This tension may indicate why many Lojbanists like the
>> idea of the community deciding what the language is through usage, while
>> at the same time want a perpetual BPFK around to codify usage questions
>> prescriptively (which is arguably exactly the opposite of community
>> decision.)
>
> I know you want usage to decide (but since you want usage to decide,
> wouldn't you accept it if people simply started using different gismu
> place structures?).

Guess what?  Neither you nor I have a choice if people choose to not 
follow the language prescription.  They can even use TLI Loglan; I have 
no say in the matter.

> However, a *lot* of people prefer a centrally
> defined language. They want clear semantics, they want clear rules. If
> the community wants it, then why should they not get it?

Because it is an impossibility, for one reason.  "Clear semantics" is an 
oxymoron.

But even ignoring that, if some people want a strongly prescribed 
language and others do not, we have a fundamentally intractable 
contradiction, and cannot please everyone.  So we follow the concepts 
under which the project was started and under which it has survived 25 
years.

If prescriptivists want to prescribe up a storm, they can try, but not 
as part of LLG, and we would prefer that they not try to pretend that 
they are working on Lojban.  (Again, we have no way to stop someone from 
doing so, but we certainly won't offer help or encouragement.)

I can't please your ilk, and I'm not inclined to try, even if I didn't 
have that LLG members' motion directing me not to do so.

You got yourself onto the TLI Academy (JCB's likely turning over in his 
grave about that) - they accept the possibility of prescribing 
everything.  Good luck over there.

> Afterall, who, if not the community of users, keeps Lojban alive?

*You* aren't the "community of users", and you and your friends are only 
a tiny subset of that community, if what you are using still fits the 
label "Lojban".  Lojban has stayed alive for many years before you came 
along, and will stay alive just fine without you, and might even do 
better, since more people will understand that we aren't going to 
support or even cooperate with every splinter group that announces itself.

> Before you call this another claim without evidence, here is a thread
> from 2010 about just that:
>
> https://groups.google.com/forum/#!topic/lojban/xn8hCt3Aagw

I don't see anything under that topic even slightly relevant to the 
current discussion.  It seems to be about whether people use the machine 
grammar in their heads when speaking Lojban.  Though indeed most of the 
claims in said discussion were indeed without evidence.

> I hope we don't have to repeat that thread again.

I didn't get involved the last time, but why do you think I care enough 
about your opinion to bother?

lojbab

-- 
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 lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.