ruminations on bangu, place structures, and corners

[Logical Language Group <cbmvax!uunet!GREBYN.COM!lojbab>]

While slowly reading my backed-up mail queue, I noted a non Lojban List
discussion last month (Mark and Ivan) about what a good Lojban word for
'natural languages', I think you can eliminate the problems of 'rarna
bangu' by making it "rarna farvi bangu" (naturally-developed language).

However, this question prompted a somewhat different one in my mind.

I started with a comment in this discussion about "national languages"
being inappropriate for the concept.

This brought to mind that Lojban has a finer division of the semantics
of people-groupings than many or even most languages.  We have natmi
(nation/ethnos) distinct from gugde (country) distinct from jecta
(polity) distinct from kulnu (culture) distinct from cecmu (community).
These relate a group of people and their common characteristic.

bangu is also one of these people-grouping words, in a sense.  The
defined place structure for bangu is "x1 is the language of people x2".
It defines the group of people by their common language.  But in light
of this discussion, I now see two readings of this statement, and am
unsatisfied.

1. "of people x2" might refer to any speaker of the language, with any
degree of fluency (or even with a specified minimum degree of fluency)

2. "of people x2" might refer to the group of people who are native
speakers

3. "of people x2" might refer to a community of speakers who use a
language regularly in daily life

All are problematic as the word is typically used.

Most English (and probably other language) usages of the form "language
of _____" where "_____" indicates either individuals or a group, are
referring to the native speakers.  If this is primary, though, then we
restrict bangu only to natural languages, and Lojban is not a bangu, and
Esperanto is one only to the extent that there are a couple of native
speakers.

Definition 3 is the most important alternative interpretation, and may
be most common when we turn the phrase around (se bangu).  When we use
"Russians" to refer to people who speak Russian, we are not necessarily
referring only to those who speak the language natively, since I think
that someone could be considered a member of this community after
acquiring it fluently in later life, and adopting Russian as his primary
tongue.  We certainly here in the US frown on distinguishing or
excluding immigrants who successfully adopt the English language as
their primary tongue in place of their native language (and what of a
child who acquires English as a second language while still young, and
essentially has lost his native language by adulthood by sheer non-use).

This may suggest a fourth possibility, which is that buying into a
language to the degree necessary to be part of its 'community' means
buying into the 'culture' of those who speak the language.  That there
is a tie between language and culture is inarguable even discounting the
determinism proposed by the Sapir-Whorf Hypothesis.  It might be that
this sense of culture is distinct from ethnos (se natmi) or even the
more general culture of a people (kulnu).  But to the extent that every
natural language >IS< associated with an ethnic culture, "selnatmi
bangu" may indeed be appropriate for 'natural language', and indeed this
4th possibility is probably adequately covered by the tanru (and
associated lujvo) "bangu natmi".

Using this argument, possibility 3 can be covered by "bangu cecmu",
though I'm not sure I'm satisfied by this.

We still haven't dealt with the disinction between meaning 1 and all of
the rest.  I am not a native speaker of Lojban, only barely a part of a
Lojban speaking 'community' in that we had a sustained conversation group
here in DC until Athelstan's accident.

One also has to deal with the question of the importance of 'speaking':
does a mute person who uses 'Signed English' (as opposed to ASL) belong
to le se bangu be la gliban?  How about someone who merely reads and
writes a language, possibly fluently, but does not speak it?  (A lot of
Lojbanists are likely to become active members of a written Lojban
community long before they become speakers of any degree of fluency, and
I presume that many or even most Esperantists may be members of the
written language community without being part of the speaking
"community" such as there is one.)

It is possible that meaning 1 is covered by cusku/ve cusku - loi cusku
be fo la lojban. are those who have expressed meaningful expressions
communicatively in Lojban.

The only I cannot seem to paraphrase simply is meaning 2. But if le se
bangu only refers to native speakers, then Lojban isn't a language of
any kind.  Some linguists indeed feel this to be the case - that a
language must have native speakers to be a language (others require an
active native-speaking community using it as their primary tongue, and
hence exclude Esperanto even with its few native speakers).

If we limit ourselves in languages to only those spoken by native
speakers, we have chosen one of the most constrained definitions, and we
make talk of "artificial bangu" or "planned bangu" to be poor metaphors,
since the x2 of bangu that sticks with the tanru has no legitimate
value.

We haven't yet addressed one other, more devastating problem:  does
'bangu' exclude computer languages, mathematical language, etc.  We use
the term language for a communicative system which may not be spoken, or
speakable by anyone, and which is appropriate as a 'native' or 'primary'
language.  Thus "of people x2" may simply be an inappropriate place
for bangu.

One possibility, but I'd like to see others' ideas:

x1 is a language used by x2 to communicate x3

(which ties more strongly to cusku and meaning 1 above)

We might also include a native speaker place recognizing that it might
have a null sumti value for a language such as Lojban:

x1 is a language associated with native-speaking community x2

(This may open up a can of worms about null-values for sumti which is
unrelated to the main point of this message.  However, it seems to be
plausible that null-values can be reflected in the semantics of a
predicate if that null value is significant to the truth of the bridi.
In other words, using this place structure, "la lojban. bangu noda" may
be perfectly true, and even the elliptical "la lojban. bangu zo'e" may
be true if we allow "noda" as a permitted value to be represented by
"zo'e".)

If you understood and bought that last parenthetical argument, how about

x1 is a language of native speakers x2, used by x3 to communicate x4

which I think covers all bases if x2 is permitted to be noda.

("noda" ellipsis may turn out to be a necessary part of the language
as used by people communicatively.  Otherwise, given tensed sentences,
you result in Latin being a "pu bangu" but not a "ca bangu" - it no longer is
a language because there are no longer any native speakers.

Now what kind of corner have I worked us into?  And how does it affect
ancient discussions about "klama" vs. "litru"?)

Nora opines that there is no particular problem.  Latin is a bangu even
in the present tense: it is the language of "loi pu latmo prenu".
Lojban can also be a bangu to those of us who believe it viable because
it is the >potential< language of speakers - it is innately capable of
being a native language.  This may be a way around the "noda" problem.

There seems to be three different types of relationships that lead to a
"noda" value for a sumti.

The first is a >specific< relationship which explicitly has no value in
the sumti place that makes it true.  My yesterday going to the store
from San Francisco has "noda" for route and for means because I DIDN'T
go to the store from San Francisco, and wasn't even IN San Francisco.
It is still nonetheless relevant to "going" that there be a route and a
means, and if I had done the hypothetical going, there would have been
values.  Such a claim is false unless I explicitly use noda, or (maybe)
there is a darn good reason why a listener could accept "noda" as the
elliptical value if I omitted it.

The second is a situation like Lojban and bangu, where you have a
relationship that is being evaluated at a point in time, but the nature
of the relationship is that there is, or potentially is, a value that
makes it true at some other time.  The place is still appropriate
because we've defined going to include a route and a means.

The third is a situation where there is no value that could fit in the
place because it makes no sense that there be such a value.  "la lojban.
zutse" = "Lojban sits" - since it is not in the nature of languages that
they have a meaningful capability of sitting on anything, the x2 of
zutse which is the thing "sat on" is "noda".  Indeed the metalinguistic
negator "na'i" could be used on either "la lojban." or "zutse" because
on of the two words used is inappropriate in a way that transcends
simply truth or falseness.

The other situation is that addressed by the klama/litru discussion.
"litru" which has a place for route, but none for origin or destination,
does NOT claim that there is no origin nor destination.  A Lojban bridi
describes a particular relationship among sumti - we define the bridi in
terms of what it relates and how they are related.  "litru" defines a
relation between a traveller, a route, and a means.  Whether there may
be other things like an origin or a destination are irrelevant to the
claim, because those roles are not defined to be part of the relation.

The argument over whether we need both "klama" and "litru" in the
language is really one over whether it is useful to talk about
travelling without inferring an origin and destination, and secondarily
as to what the mapping is from English words to Lojban selbri.

Now.  Did I indeed talk myself into a corner, and then did I
successfully work myself out of it?  %^)  And what should the place
structure of bangu be?

lojbab

Oh, and are computer languages and amthematical languages bangu?  If not,
how about someone devising tanru/lujvo.