More lessons learned about how to design and learn a language

[Logical Language Group <lojbab@GREBYN.COM>]

My experiences in communicating in Russian with my children have given me
added insight possibly relevant to both the Lojban community and to the
conlang community at large.

Russian being a language with no less than 6 cases, not all of which I
know very well (even if they were totally regular), I have found that my
speech error rate is extremely high in use of the cases.  I have also
found that this seems to make almost no difference to my children insofar
as understanding goes.

We aren't talking minor errors here, by the way - I am incredibly
slipshod in my use of the language.  So much so that a couple of people
have said that I'm not speaking Russian at all, but rather some kind of
pidgin Russ-american.  I can't even start to understand real [adult]
Russian speech at fluent rates of speed, either - my vocabulary remains
much too small.  My speech is perhaps 1/3 of fluent speech rates at best,
even when I am reading prepared text with accents marked I don't do much
better.

This isn't just incidental usage, by the way.  Conversation in this
household has been entirely in Russian for the last month except between
my wife and myself, and when people visit or telephone (and a number of
Russianisms seem to slip in there, too).  My slowness at picking up the
language when I am in a total immersion situation many hours a day have
only confirmed in my mind how much difficulty I have in learning
languages.

But my kids understand me, and think I speak Russian VERY WELL [even
when I read then stories in Russian and have to guess the accents on
words - Russian is not especially regular in stress to my mind].  And
they know enough to be able to tell the difference.  This isn't just
happiness that I've made an effort, or a desire to please papa; they
distinguish my level of speech from Nora's (she has studied roughly 2/3
of my level of vocabulary and grammar, but spent much more time
listening to tapes and she HAS spoken another language conversationally:
French),

I have come to believe that kids just have a far different threshold of
what constitutes "good" language use:  if you are understood most of the
time, especially when you can get clarification when you don't
understand, this is fine.  There is a high enough degree of redundancy
in the language that my error rates do not destroy communication.

As backup data, I have to note that my son, Avgust, has a bit of a
speech impediment - his 'l's and 'r's are rarely distinguishable, and
likewise for me his "s"s and "sh"s and may of his palatalizations.  His
sister seems to have no problem understanding him in spite of these
problems, and I usually don't provided that the words are ones that I
know (if I have to go to the dictionary, I can rarely figure out the
sounds well enough from his speech to come close to the spelling).

My kids also make a good number of grammatical errors, especially in
case selection, that even I notice.  Some verbs lose their endings, and
they make frequent apparent errors in using some case structures
regularly that are apparently quite common in Russian (e.g., use of
dative "subject" with several verbs of attitude/intension).  I make
these errors too, but I'm quite sure they didn't learn it from me, because
I think they were saying such things back in Moscow when I as yet could
hardly make out what they were saying and relied on the translator for
much of our interaction.

Where I have limits in communication is in vocabulary, vocabulary,
vocabulary.  My 1000 or so words of Russian just isn't enough for the
task of communication even at the 6 year old level, and my vocabulary
isn't growing nearly fast enough to hope for fluency in a matter of
months.  I thus have gotten very fast at using the dictionary, and
sometimes just paraphrase around the subject.

My conclusions:  The argument over the "difficulty" of the accusative
endings of Esperanto for English speakers is irrelevant.  This is a
problem in learning Russian, too.  The error, when you do make it,
rarely if ever seems to cause any difficulty in understanding, and
meanwhile the influence of hearing people do it correctly has (very)
slowly improved my skill.  It just plain doesn't matter that much in
speech, and may not even in writing unless maybe you are writing poetry.
(I make far fewer errors of this type in my rare writing efforts in
Russian - the problem for me is largely one of trying to speak at
communicative rates of speed.)

I will go so far as to say that I now think that as long as the grammar
rules are regular, reasonably few in number, and can be explained in
terms that you can understand, that there is no conlang grammar that is
too hard.  Explaining in terms you understand, in my mind includes
explaining at least all first-order interactions between rules:  the
trickiest part of Russian for me has been remembering what rules are
most important, and how imposition of one rule may affect whether
another comes into play.

I also have had little trouble with the Russian phonology, although it
makes a number of distinctions that are not made in English.  Both my
kids understand what I say quite well - I've never noticed any failure
of communication that was directly traceable to mispronunciation of a
strange phoneme.  Some adult Russian speakers have even said that my
accent is acceptable [although I'm sure quite noticeable].  So I suspect
that if the phonology has reasonably regular rules (as regular as
Russian is probably sufficient), it will not impose much of a burden on
learning.

But vocabulary remains the big pain.  Cognates help some, but not that
much, from what I've found.  When most of the words do not have an
English cognate, I find that the few that are cognates feel more like
surprises than useful crutches.  I don't trust my memory with these
cognates, unless they are the kind of special vocabulary words that are
likely to be borrowings from another language.  I seem to do much better
with words that have Lojban cognates, or even non-cognate word parts
that I can map to Lojban or English after the manner of the Loglan
word-making algorithm, or some other indirect or unusual memory hook to
hang the word on.

This of course makes me feel much better about the Lojban vocabulary's
learnability, of course.  I'm finding the techniques working well on
another language that wasn't "designed".

The main lesson all this has taught me, though, is that people trying to
learn another language, whether a conlang, or a natlang, should STOP
AIMING FOR PERFECTION in their self-expectations.  Indeed you should not
worry if you hardly know where to start in communicating.  Just DO IT!!!
Express yourself, and let others give you feedback on whether they
understand you, and where they had problems.  I suspect that with most
conlangs, you'll reach a reasonable communicative written level of skill
quite quickly (not a literary or formal translation skill level, but one
that allows you to communicate what you essentially need and want to in
real life).

I think this also helps me as a reviewer of other people's writing
efforts in Lojban.  I need to review at two levels:  first of all, did I
understand what the speaker was essentially trying to say (not worrying
the details - did I get the essential information?), and only then
worrying about such things as grammar errors and word-choices and
culturally-biased expressions, etc.  The latter I can mostly save for
the people who can write in Lojban almost as well as I can (or better).

Using this standard, my suspicion is that Lojban has been demonstrated
learnable for people of several native languages in very short order.
(I think Nick Nicholas said to me a few months ago that it may take a
few weeks to be able to write understandably, and maybe 3 months of
serious effort to be good enough to be among the leaders of the project.
He and a couple of others have actually done better than this, but given
my experiences as to the minimum standard necessary for Russian
communicative competence, I now think that this is true of most
everybody.) 3 or 4 attempts of writing a paragraph of text, and getting
useful commentary on errors that are IMPORTANT to human communicative
understanding (as opposed to what is important for unambiguous computer
understanding) will make anyone a Lojbanist.

Funny, now that I think about it, I suspect that our conversation
sessions already practice this.  We make a lot more errors in practice
when speaking the language in Lojban conversation sessions.  But we
concentrate on different types of errors than I seem to notice in
writing, and ignore most others, except for pedagogical purposes.

lojbab