Re: replies re. ka & mamta be ma

[ucleaar <ucleaar@ucl.ac.uk>]

Jorge:
> You said that {kau} is not essential.
> I said that that is true of every cmavo.
> You said that surely not about most members of SE, LE, PA.
> I maintain that yes about each member separately, I am NOT
> saying we can do without all (or most) of them at the same time.

I was meaning: Suppose we tried to make the grammar as small as
possible (without reducing expressive power), whittling it down
from its present size - in such a case, {kau} could go.

> > > i ko pilno lu le jufra poi mi ca co'a [> caho] cusku ke'a li'u
> > I note with pleasure your non-literal use of {ko}. I very
> > much doubt anyone will have been admonishing you for this
> > flagrant breach of prescribed usage.
> I'm not really being all that non-literal. I'm just eliding the
> condition "in case you want to avoid the use of {dei}".

Aha - but if you didn't elide the condition, you'd be saying:

  I "command" you to make it the case that if you want to avoid the
  use of {dei} then you use sentence S.

Whereas what you want to say is:

  If you want to avoid use of {dei} then I "command" you to
  make it the case that you use sentence S.

The latter would be a literal use of {ko}, but I don't think
it's doable without great circumlocution.

> I clearly stated that {la'e di'u jetnu sera'a ro cmavo}.
> You seemed to interpret it as {piro loi cmavo}, but that's not
> what I said.

I still don't think it's true, even on the distributive
interpretation. "Every cmavo is not needed" entails "There
is no cmavo that is needed", which surely means we could
get by with a cmavoless language. Someone (probably me, I
predict from experience) is reasoning wrong.

> > been trying to make
> > that sohe loi cmavo se nitcu, but na ku sohe lo cmavo se
> > nitcu?
> That {piso'e loi cmavo cu se nitcu} may or may not be true,
> I'm inclined to think it's not, but could well be wrong.
> (Of course, I'm taking "need" in your absolutist way, in the
> sense that you need it to express something otherwise
> inexpressible, not that you need it to make the use of the
> language easy.)

We agree on this.

> > Are you also saying sohe lo cmavo ku na ku se
> > nitcu? (which is different, unless {sohe} means "over 50%")
> I fail to see the difference. I know that with {ro} and
> {su'o} the claims are very different, but with {so'e} they
> seem to be the same.

Well, say {sohe} means "getting on for all" - then scopewise
it behaves pretty much like {ro}. Say, for sake of argument,
{sohe} is "75% of". Then "75% of camvo are unneeded" is very
different from "It's not the case that 75% of cmavo are needed".

> > Am I right in thinking you'd say I need suho loi kidney
> > but not suho lo kidney?
> At the risk of falling back into an old discussion, I'd say
> {do nitcu xe'e su'o lo do kidney}. I don't know much about
> these things, but I don't suppose that there is any one
> preferred kidney that you need, unless you only have one.

I don't want to reopen that discussion either. But I don't
recall you claiming there were things that could not be
said without {xe'e}, so I'd ask you to express the above
in a way such that I know how to translate it into pred
calc.

> {su'o loi kidney} is "at least one mass of kidney", and
> {su'o lo kidney} is "at least one kidney", I would tend to
> never use the first.

I shd have said {pisuho loi kidney}.

> > [Incidentally, any idea what {lo sohe broda} could possibly
> > mean?]
> Nope. The so'V series is not really a series. {so'a} and {so'e}
> are quite different from the other three, because they need
> a reference (ro) while the others are subjective but without a
> reference (so'i, so'o, and so'u can each be equal to ro).
> Since the inner quantifier of {lo} is always {ro}, anything that
> can't be equal to {ro} makes little sense there (so'a, so'e,
> me'iro, etc).

To me this is a bug in the grammar - one of those constructions
that yields bona fide gobbledygook.

> > > > > i zo do'e joi zo poi ka'e basti zo fi'o
> > > > How would that work? (E.g. if a selbri has 3 fiho modals)
> > > i <<lu do'e da poi broda ku'o do'e de poi brode ku'o do'e di poi brodi
> > > li'u>> cu basti <<lu fi'o broda da fi'o brode de fi'o brodi di li'u>>
> > I defer to your greater knowledge of Lojban,
> Very unwise of you.
> > but this is not how I thought {fiho} works. In your version with
> > {dohe}, the semantic relationship of the sumti places filled by da,
> > de and di are not specified.
> While in the {fi'o} version they are? I don't see what makes one
> more specific than the other, if you include the appropriate poi
> clause.

The poi clause gives you info about the sumti, but not about the
relationship holding between the sumti and the other sumti. It's
the difference between.

   tavla bau lo jbobau
   tavla dohe da poi jbobau

The former tells you the talking is done in Lojban, while the
latter tells you only that the talking somehow involves Lojban.

> > I don't even know if they have to
> > be different sumti-places. It depends on whether
> >   klama fo da ku fo de .i
> >   klama bai da ku bai de
> > are grammatical.
> Grammatical they sure are, at least in the sense of parsable
> (except for the {ku}, that should not be there, since {da} is already
> a closed sumti).

Another bugoid, I reckon. English doesn't allow it, at least not with
complements. Does anyone happen to know if in, say, Latin or Russian,
or Finnish or whatever, you can have two accusatives (or two words
in the same case) in the same clause? If it's possible, what does it
mean? Is there an implied "&"?

> > In contrast, I thought in {broda fiho brode da}, da is a sumti
> >  of broda, but not necessarily of brode, and
> > brode serves simply to identify the semantic relationship
> > between the broda selbri and the fiho sumti.
> Yes, all that is true. But how is that semantic relationship any
>  more specific than with {do'e}+appropriate explanatory sub-clauses?
> In other words, is there anything that {fi'o} allows you to say,
> that you couldn't express without it?

The answer to the latter question depends on whether there are ways
of doing without BAI in general. I suppose there are, in which case
{fiho} is dispensable (under the strict criterion of necessity).
As for the former question, let us imagine we wish to classify
plants by what they make. A plant producing threads of cloth could
be:
    lo spati be fiho zbasu lo bukpu cilta
>From that you can at least surmise that what is meant is not, say,
plants that look like cloth threads. In contrast, from
    lo spati be dohe da poi bukpu cilta
you know only that the referent is something that somehow has to do
with cloth threads.

> > > To avoid the use of {goi}, you can simply use a lujvo
> > > meaning "assign",
> > I doubt it. The lujvo wd make a claim about the way the
> > world is, which is wholly different from assigning a value
> > to a variable.
> Ok, use the full sentence: "From now on I will use "ko'a" to
> refer to the referent of "<sumti>". No {goi} required.

Okay, I might quibble that your revised version would still be an
assertion, but it would be a mad speaker who would lie about such
a thing, so in practice it wd work.

It wd be interesting to see a conlang with such a small grammar
that such verbosity was required.

> > > or you can avoid pronouns altogether, which are nothing but
> > > convenient optional add-ons anyway ;)
> > I don't think so. Assignable pronouns are virtually the only way to
> > guarantee continued constancy of reference. If I keep
> > on using {le nanmu} there's no assurance I'm talking about
> > the same bloke, but if I use {koha} you can be 100% sure.
> You use {le nanmu} the first time. Then use "{le nanmu} that I just
> mentioned", then "{le nanmu} mentioned twice before", etc. If wordiness
> is not a problem, you can make sure your audience is 100% sure which
> one you mean. (And 100% bored with the speech.)

I suppose a pretty failsafe method would be to say "le nanmu referred
to by the 1074th word of this utterance". You'd just have to count
and remember every word in the right sequence.

Okay then: anaphors are an add-on convenience. But let us make
a distinction between conveniences for avoiding the merely cumbersome
from conveniences for avoiding the impossibly cumbersome.

> ki'o  =  ,  =  ,000  =  1,000  (by convention).
                 ^^^^^^^^^^^^^^
Didn't know that bit. Where do you find this stuff out?

> > > > i le ma tadji cu ckaji le ka jicmu
> > > The more general method is more basic. But that's a separate
> > > issue.
> And if I ask which method is more general, I suppose it is the
> more basic one?  :)

No, the more general method is the one used in the most constructions
(so long as it is used with the same grammar in each construction).
I can't think of an actual example offhand, so here's an imaginary
one. Suppose we had two ways of expressing numbers, but one way
could be used in all contexts where PA can, but the other way worked
only when descriptored by {li}. The first way would be the more
general and the more basic.

> > Second, try doing {ro nanmu nelci lei ninmu} without
> > using LE or POI.
> ko'a ninmu gunma  i roda zo'u da nanmu nagi'a nelci ko'a

Well done. Does that differ from {ro nanmu nelci *loi* ninmu}?

> > Third, how does this show why you disagree
> > with me (an attitude you adopt, I suspect, largely for the
> > pleasure of being disputatious)?
> Partly, I admit, but I do disagree that there is anything
> fundamental about certain cmavo. If there was, the language would
> be very poor indeed.

Every lg needs a word/morpheme for "1", but doesn't need one
word/morpheme for "7582342".

I think one of the things that attracts many people to conlangs is
the search for what it is that is essential for a language. More
conlangs ask "what can we get rid of?" than ask "what useful gizmos
can we add?". Both questions are important, and Lojban does a
pretty good job at both shedding a lot and adding nice touches like,
say, {muho} (end of turn signal).

> There are many ways of expressing the same idea. That holds for every
> language, including Lojban.

And so it follows by my reasoning that you cd get away with having
only one way.

> Certainly the language needs cmavo, and certainly some cmavo are
> more useful than others, but I wouldn't call any one cmavo essential,
> nor assume that there is a minimal set such that all others are
> add-ons, and if you remove one from the minimal set then suddenly
> you condemn a whole bunch of ideas to inexpressability.

Is lahe the "condemn"-clause what you wouldn't assume? (Your sentence is
ambiguous: do you mean "and *that* if you remove"?)

I don't understand how you can claim what you do. Suppose you had to
devise a notation for all numbers. You could use as many symbols as
there are numbers. Or you could use fewer symbols but add a grammar
for interpreting combinations of symbols. As you use fewer and fewer
symbols you'd come to an irreducible minimum. One symbol alone will
not suffice. I guess some mathematician has worked out how few will
suffice.

> And I don't think {kau} is any more of an add-on than most other
> cmavo, which is what started this thread anyway.

What can be said with {kau} can also be said with {da} in the appropriate
prenex. Not all of what can be said with {da} can be said with {kau}.
This is why I said (& maintain) that {kau} is an add-on.

Once we have learnt how to manage without {kau}, then we can decide
if it is useful and convenient. I'm happy to accept that it is. It
may be glico (& I am rather inclined to suspect it is), but that
doesn't make it mabla.

---
And