[lojban] ambiguity in lojban

["Chris Capel" <pdf23ds@gmail.com>]

What follows is a version of this[1], except slightly modified to
remove explanations that are superfluous for readers of the list.

[1] http://pdf23ds.net/2006/06/20/on-the-ambiguity-of-lojban/

Many say that Lojban is an unambiguous language. This is true in many
respects, and is probably more true for Lojban than for practically
any other language. Lojban has an unambiguous phonology (when
pronounced clearly and without error) and an unambiguous grammar
(again, when used correctly). In its primary word list (both for gismu
and cmavo), it avoids polysemy, the assignment of multiple meanings to
a word.

Now, this lack of ambiguity, as far as it goes, makes for extremely
fascinating discussions about cmavo and their equivalents in other
languages. But what it does *not* do is substantially decrease the
occurrence of ambiguity in language as it's actually used.

Only a minority of the occurrences of ambiguity in normal English are
of the sort not found in Lojban. Misinterpretations of spoken lanugage
rarely happen. If a hearer has trouble, it's usually the case that
they can't form any sensible interpretation, and when they do form an
interpretation it's very likely the right one. (Music lyrics are
different, because so much information is lost due to the difficulty
of clearly hearing the words, the often unnatural pronunciation, and
the loss of most syllable stress.) Ambiguous grammar hardly every
causes any confusion in English.

Polysemy is a bit more complicated. First, it's the case that a
"meaning" of a word has a dimension. The meaning narrows down what's
being communicated to a certain range, and a particular instance can
be anywhere in that range. A "table" can mean a dining room table or a
coffee table or a cafe table or a workshop table, etc. That range of
meaning is a big blob in meaning-space. But it's not polysemy.
Polysemy is when a word has two or more completely separate blobs of
meaning, or when it has two or more *basically* separate blobs of
meaning that have a little bit of overlap. Now, Lojban does a lovely
job of having very precise meanings in its cmavo. The meaning blobs
are smaller than they are in English. Gismu also successfully avoid
having much polysemy. English has orders of magnitude more in its
common words. On the other hand, gismu have very wide extents in their
definitions--their meaning blobs are quite large. (This is because
there aren't many gismu. Lujvo can be formed, but Lojban is young, so
there aren't many yet.)

Polysemy rarely, rarely causes confusion in English, though it is the
source of a lot of humor. For instance, if I say "table" and mean
"dining table", it's going to be extremely rare that you'll understand
it as "actuarial table". No, the majority of misunderstandings in
English are due to confusion about *which part* of the meaning is
intended. I say "table" and mean "dining table"; you hear "table" and
understand "coffee table". I don't say "dining table" because I think
our shared context would make it redundant to say "dining". But the
context as you understand it makes it so "coffee" is the more sensible
interpretation.

This type of confusion is not avoided by Lojban, except in a few of
the more mechanical type words. Because this is also the most common
type of misunderstanding, by far, Lojban will not significantly reduce
misunderstanding during communication.

There's one more type of confusion that I want to address. I think
this kind of confusion is the most common in deep discussions between
ideologically similar people. Assertions often have subtle
implications that depend greatly on an individual's worldview. Whether
one agrees with, or even sees, these different implications depends on
one's own worldview. Much mistunderstanding is simply seeing different
implications of assertions. A person asserts X and sees implications Y
and Z. Another person reads X and sees implications V and W. They
disagree vehemently with X, because of V and W, though they agree with
X itself, and probably even Y and Z. This is similar to the previous
type of misunderstanding, except that most of the time the difference
in the extent of meaning can't be tied down to one word, or often,
easily tied down at all.

Now, I don't expect that the existence of these different types of
misunderstandings is controversial. But I additionally assert that the
more subtle misunderstandings constitute the vast majority of
ambiguity as it actually contributes to misunderstanding. Thus, I
think it's not possible to say that Lojban is even remotely free of
ambiguity, (without tightly restricting what one means by "free of
ambiguity"). And I will continue to be aggravated when people do so.

Many of the older texts mention that Lojban might be uniquely suitable
for human-computer communication. The idea is that since computers are
so bad at resolving the meaning of polysemous words, and at resolving
ambiguity in grammar, (two things that humans are extremely good at,)
that the absence of these things will make it much easier to allow
computers to understand human speech. But the fact is that computers
that don't have a flexible enough language model to resolve
ambiguities from polysemy and grammar ambiguity are unlikely to be
nearly subtle enough to correctly resolve ambiguities in extent. So
creating a language like Lojban for that purpose (which was never the,
or even a, primary purpose for the language, as I understand) is
analogous to taking a course in mechanical engineering and building a
shovel cleaner before starting the work on digging a tunnel to China.
Any application in which some restricted subset of English could be
used, except for problems with polysemy and grammar ambiguity, is an
application that probably doesn't need to use a natural language at
all.

Chris Capel
--
"What is it like to be a bat? What is it like to bat a bee? What is it
like to be a bee being batted? What is it like to be a batted bee?"
-- The Mind's I (Hofstadter, Dennet)


To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.