cultural gismu, LogFlash, etc.

[Logical Language Group <lojbab@ACCESS.DIGEX.NET>]

>From: Paolo Cuzzi <MC7926@MCLINK.IT>
>Subject:      Ethnic Gismu, Learning Lojban

>First, I browsed the gismu list, and I noticed that ethnic-related terms
>are somwtimes puzzling.  For instance, no gismu (nor lujvo, as far as I
>could see) is avaible to express 'Italian' and related concepts

latmo covers, in Ancient times, the Roman empire; in modern times, Rome
IS the capital of Italy.

Jorge mentioned "talno", which was a TLI Loglan word that stayed on in
mine and a couple of other old-timers' dialects after we remade the
words.  However in JCB's version of the language, talno was truly an
anomaly, with no plausible explanation as to why it was added, other
than the fact that someone once wanted to make a sentence about Italy,
and coined the word, and happened to do so before JCB wrote his
dictionary.  JCB also included a gismu for "billiards" for the same
reason.

Many of JCB's idiosyncratic gismu survived into Lojban, and a desire to
justify them often led to rather extreme or unreasonable criteria for
others. talno could not be justified under any criteria that would not
have called for us to add several dozen other culture words.

A major reason for "talno" still surviving in peoples' minds these days
is that the logical fu'ivla forms for the word involve the use of
"talno" as the lojbanized root:  bangrtalno, gugdrtalno, etc.

>, while you can find words for 'Greek' or 'Palestinian'.

Like Latin, Greek is in there as a major ancient civilization that
happens to also have an incarnation as a modern country.  So is
Hebrew/Israeli.  "Palestinian" was an attempt to steer clear of cultural
biases in including the latter.

>Which are the inclusion criteria?

Jorge responded:
>There are no clear criteria.

Which is not true - there were criteria, but they were followed to
arbitrary levels based on Loglan/Lojban history, conflict with other
gismu, lack of careful research, and perhaps also some cultural bias.
This has been covered in much detail in precious discussions:  the
intent was to start with the main source languages used for
Loglan/Lojban, and to stop at some reasonable level of nation/language
naming.  What is "reasonable" is alas altogether too arguable; we went
unreasonably far in including too many Arabic countries, and stopped
short of reasonableness in the case of Spanish-speaking countries.

Italian (as well as Greek and Hebrew and Sanskrit) would not have
qualified under the source languages criteria, but were instead
qualified as "sources of Western civilization".

>In other cases, the terms seem to have been chosen with some
>carelessness (is it an English word?):

Much care, but not always perfect knowledge.

>e.g., kisto means both 'Pakistani' and 'Pashto', while pashto-speakers
>are a minority in Pakistan, with no special relationship with the Urdu-
>speaking majority (Pashto is mainly spoken in Afghanistan, where second
>language is a variety of Persian - Dari -rather than Urdu).

But in this case we were reasonably aware of what we were doing.  Yes,
but Pakistan IS named after Pashto.  Hence, as a NAME, if one were to
talk about "the" Pashto-country, Pakistan is at least as justifiable as
Afghanistan as the interpretation.  Since we do also have Urdu
represented, you could also use "Urdu-country".  Which term the people
of Pakistan would prefer is not something that I intend to decide, but
to default to the one closest to the name of the country seems
appropriate.

The real issue is that of overlapping words identifying cultures with
words identifying countries and words identifying languages.  To do so
cuts the number of words you need, and risks some cultural bias, but the
alternative is a proliferation of words that, as been generally
conceded, are not going to be heavily used except by members of the
named culture.

>Second, I just began using the Logflash program.  I think it is really
>useful for an absolute begineer like me.

Thank you.

>I first accepted the proposed New Word mode, which is anyway too
>indulgent with the learner (I was credited to know more than one hundred
>gismu, while I just possessed a dozen words or so).  Gaining mode seems
>really more effective.

Words that are "Under Control" are NOT considered to be "known" in
either New Word or Gaining Control modes.  You do not know the
vocabulary until you get 97% correct on a large lesson in Maintenance
mode.  In general, each time you go through the entire set of gismu,
your error percentage will be cut in half, until you reach that level.

New Word mode is intended for people who really aren't trying to master
the vocabulary (i.e. dilettantes), people who expect to take several
months or years to follow/learn the language before mastering it, and
people who simply cannot muster more than 15-20 minutes a day to do
LogFlash.  It admittedly is NOT going to work in teaching you the words,
and you do not "know" the words when you have gone through them in this
mode.  The point is to get through the entire vocabulary quickly, so that
1) you have a sense as to the full scope of the gismu vocabulary
2) are reasonably likely to recognize some words in running Lojban text
even when you haven't really learned them, enough to perhaps pick up the
subject of a text to decide if you want to read it more carefully.

New Word mode was NOT part of the original LogFlash algorithm.  But
whereas I learned the gismu list quite effectively with 40-word Gaining
Control new-word-lessons, spending an hour a day for around a month to
get through the list the first time, many later people who were unable
to put in this much time were doing 20 minute lessons, and thus shrunk
the size of their lessons down to only 10 words at a time.  At that rate
it would take over 4 months to get through all the words, by which time
you have forgotten the ones from the beginning.  Added to this, these
people would often be doing lessons at work during lunch time, and
skipping weekend sessions, and might miss several days in a row for
arbitrary reasons.  After a gap like this, their error pile word count
would suddenly jump, often up to several hundred words, requiring them
to take a few hours for one lesson (if they actually did this, then the
size of the "bump" in time required tends to decrease, but more often
people would just reset all their words back to the bottom and start
over, in which case they never got done).

Thus we added the "Under Control" lesson, which consists of words
promoted to that level 5 sessions before, along with a random scattering
of earlier words, to minimize this loss over long learning periods.  And
we added New Word mode as a way to get people to get through the entire
vocabulary more quickly, and keep their lessons sizes up above 20.
Finally we inserted an option that lets you set a maximum size on a
lesson so that no one feels compelled to reset the words when they get a
large error pile accumulation.

None of these options are necessary if you are going to be diligent
about studying every day or nearly so in Gaining Control mode, and are
studying enough words that the lessons are challenging you.  LogFlash
works best when you do enough words in a lesson that your typical
new-word error count is over 10 (and ideally as high as 20), so that the
error drill requires thought and not just typing speed.  When the error
count is under 5 or 6, LogFlash error drills will help very little.

>It's incredible how more difficult is to remember lojban equivalents of
>English words than the other way round.

Recall of new language words is indeed more difficult than recognition
of those words.  Another reason for the New Word mode, which puts this
study off until you have seen all the words.

>In many cases, only consonants are clearly remembered.  I personally
>tend to forget final vowels, perhaps as a reflect of my Italian-speaking
>habits (since final vowels are not functional in Lojban; I noticed a
>similar problem in learning some Swahili).

English speakers also have final vowel problems in Recall.

lojbab
----
lojbab                                                lojbab@access.digex.net
Bob LeChevalier, President, The Logical Language Group, Inc.
2904 Beau Lane, Fairfax VA 22031-1303 USA                        703-385-0273
For the artificial language Loglan/Lojban, see powered.cs.yale.edu  /pub/lojban
    or see Lojban WWW Server: href="http://xiron.pc.helsinki.fi/lojban/";