*old response on pc as the ultimate authority who gets listened to %^)

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

>Date:         Fri, 5 Jan 1996 14:30:51 -0800
>From: "John E. Clifford" <pcliffje@CRL.COM>
>Subject:      tech:logic matters
>To: Bob LeChevalier <lojbab@ACCESS.DIGEX.NET>
>
>&
>Okay.  So {ro da broda} entails {da broda}.  But all other uses of {ro}
>don't entail existence - {ro da poi kea broda cu brode} and {ro broda cu
>brode} entail neither {da broda} nor {da brode}.  One ought to be
>encouraged, I think, to not use poi clauses with da, and to instead use
>logical connectives.
>
>pc:
>Well, in fact, _ro da poi broda_ was introduced ages ago exactly to
>carry the implication that there are brodas.

It was??? poi was JCB's jia which dates back to 3rd edition L1, if not
earlier, and I figured it was never really "introduced".

Now the meaning of "roda poi broda" was never discussed in L1, so
perhaps that particular discussion came later.  Can you at some point
find a reference?  (It seems more or less moot to me, since if indeed
roda broda implies da broda then this analysis would follow to me).

>The other expressions, _ro broda_ and _ro lo broda_ are, I think, up for
>grabs, but Cowan seems to have appropriated all of these distinct
>expressions for some other set of distinctions.

"ro lo broda" is intended to be a (restrictive?) quantification of "lo
broda".  If "lo broda" does not entail existence of broda, then "ro lo
broda should not".

"ro broda" is based on JCB's "se sorme", and hence is a "naturalism" or
"colloquialism".  JCB said that his natural instincts demanded that it
exist as an abbreviation, but never said what it was an abbreviation of.
I have always favored having it be the loosest and sloppiest of
meanings, as colloquialisms usually are.  For a long time that loosest
meaning was "ro lo broda" that did not claim existence.  But the
discussion of the last years convinced me that to truly be colloquial,
it must be vague, and thus I accepted that it may actually be based on
"le" and not veridical - "le ro broda", or "ro le ro broda".  It SHOULD
be either the vaguest/broadest plausible meaning, or an abbreviation for
a 'most "useful" meaning that cannot be expressed in a few words', such
as the "if x exists, then all x" interpretation.  I can live with
either, but feel that the same analysis should apply to "ro broda" as to
"ze mensi" (which I am ever more convinced is "le ze mensi" some
specific but indefinite 7 sisters who are mutually so).

>And I, of course, think that _ro_ should be treated as uniformly as
>possible, which would mean requiring the _ganai _gi_ construction to get
>the "modern" interpretation.

Not clear which this is, and for which constructs it should apply.

>(In fact, the current theory treats the standard quantifiers of first
>order logic as degenerate cases of restricted quantifiers -- restricted
>to everything -- as they historically are, and the considers the full
>range of possible quantifiers (= second order relations satisfying
>certain restrictions) and goes for a more general notion than usual,
>beyond the four traditional quantifiers (AEIO in Scholastic terms) and
>absorbing all the interpretations of those types as well as cases like
>the "modern" forms but less satisfying systematically:  "many," "most,"
>"denumerably many," and so on.  Lojban actually looks pretty good in
>this respect -- no text I've looked at even talks about "enough" (a
>three-place relation so a two place quantifier as a determiner).

I didn't follow the prelude to the last sentence, but I know we did
discuss "enough" at some point, but perhaps not in any prescriptive
text.  I agree that there is an implied place missing.  At which point
it becomes obvious that this is another "sloppy colloquialism".  I
certainly would not claim to know what the logic of "enough" quantifier
with a "da" variable should mean.

>The Lojban treatment of these all needs some tidying up, however.)  I
>applaud &'s last suggestion but think it is unreasonable to expect, for
>-- as Carter shows pretty clearly -- many people are terribly fond of
>the "modern" universal (though they, like Carter, greatly overestimate
>how often they need to use it).

I think that the avoidance of "roda" and indeed anything that hints at a
prenex is the avoidance of exactitude in specifying the universe of
discourse.  We are uncomfortable with the idea of saying "ro merko
prenu" (All Americans) because we have had it drilled into us that there
is virtually NOTHING that can be predicated about "all Americans" except
that they are all Americans.  We want to be able to use some kind of
implicit unstated and not-necessarily in-mind set of restrictions, and
all our exposure to logic and prenexes says that ro means ro.  "se
sorme" "ze mensi" constructs as I have recently understood them, fit
that line of reasoning, and consistency makes me want "ro mensi" to work
like "ze mensi" - if I say "All sisters like each other", I am saying
that sisters of each other like each other, and not that Mary the sister
of John and Kate the sister of George necessarily like each other.
There is an implicit restriction based here on an unfilled place of
mensi.  (In other cases, it may be based on some restrictive criteria
that are not places of the selbri, but no examples come to mind on the
fly).

"ro lo broda" is explicitly equivalent to JCB's old universal veridical
"lea".  Indeed, that was the starting point when we introduced "lo", the
veridical "lea", but with different default quantification because pc
and I agreed that universal statements simply were not all that useful
or common.  "ro lo" should mean what "lea" meant.  But again, I think it
should be consistent with what "ci lo broda", or maybe vice versa.

If we respect the need for people to have "ro lo broda" and "ro broda"
be somewhat abbrevatory and related to the their non-universal
varieties, I don't much care how the semantics settle (I don't use "ro"
much at all for anything, I am so antagonized by universal statements.
That is why we have da'a - I am apparently more extreme than formal
logic in that I insist that the superlative is not zmadu roda, but zmadu
da'ada, because it is not more than itself).  I agree that IFF we can
agree on the semantics, then Cowan should include same in the
refgrammar, and that if no consensus exists, he should state what
limited alternatives have been considered valid.

>&
>should we propose that to CAhE we
>add two cmavo for "true of the real world" and "true not necessarily
>of the real world", with the former being the default?
>pc
>In fact, we have those already, _ca'a_ and  _ka'e_.  Maybe we
>need something explicitly about stories (my list still has _ci'a_
>free, but is badly out of date, I suspect)

At one point we had an analysis of attitudinals as to whether they
referred to the real world or to some internal mental state.  This
sounds a little like that distinction.  I wonder what word we used ...
I will agree that a lot of this "real world" vs.  "imagined world" might
go away if we had such a marker.  But the pragmatic default will
probably turn out to be the non-real world %^)

I sense that this is different than ca'a/ka'e, but can't figure out why.
I would support an added cmavo if that seems to be the appropriate
solution.  I do NOT have Cowan's aversion to using ci'a for anything.

>&
>{xee} meant "any old broda whatsoever", which I think amounts, in
>effect, to the same thing as "the quantification on this sumti does
>not belong in the bridi the default rules would locate it in".
>I vaguely remember {xee} being in LE or PA, but LAhE is a better
>selmao for it.
>
>pc:
>I don't see that reading: "any old broda whatever" just draws from
>whatever the established interpretation of _broda_ is, it says
>nothing about how that is established and does not obviously
>change that establishment.  I think x wanted to have a leaper, that
>would guarantee that the _broda_ down in a _nu_ clause was
>drawn from the external assignment.  So, the goal is right (though
>more directed) but the reading is questionable.

Nobody ever wrote the bloody thing up into a proposal and summarized it
for all of us that could not follow the voluminous debate, or we would
KNOW what was intended .uinai.  This was true of just about everything
in the "any" debate, which in turn is why it never resolved anything.
SOMEONE needs to summarize whatever consensuses may have formed.

>&:
>> We need (in the current depressing state of the language)
>why depressing?
>
>pc:
>Because it is getting further and further from a natural language
>commonality (if not universal) and so harder and harder to say
>normal things.  And all for no visible purpose, either practical or
>theoretical.

One reason why I want "ro broda", and "ro lo broda" to be naturalistic.
I gave up on the idea that people who talk Lojban would "talk formal
logic" probably before I started.

I want the language to ALLOW people to talk formal logic easily, but I
also want naturalistic alternatives in there.  That allows a test of
subportions of the logical discourse to be used in Lojban.  One can
avoid the nuances of quantificational logic (which I do, having flunked
the subject), while using the concepts of predicate logic (which I did
fine with).  Given that most people will NOT have mastered formal logic
it is better to start with subportions that are more easily learned (and
internalized) than others.


>&
>You're probably right, but we ought to agree on what this new LAhE
>would mean. We already agree that this new LAhE is not needed to
>make the hitherto unsayable sayable [unlike the new CAhA, which
>would make the unsayable sayable]; so by normal standards, the new
>LAhE shouldn't be adopted, but your advocacy is given privileged
>attention.
>
>pc:
>I am not clear what the new CAhA says that can't be said already (see
>above, with appropriate added notes, from BAI, I think).  As for the
>LAhE -- which I would advocate, remember, only because the natural
>solution got lost somewhere between 1988 and now and has apparently
>become unredeemable at this point (a strong argument for carrying on the
>kind of conversations we have, that are said to never come to a
>conclusion -- if we had argued this a few years ago, we would not have
>ever had the gawdawful change) -- the plea is just that, while we CAN
>say what we want, to do so involves profoundly complex locutions for
>saying what we have, at first glance (but not at about third) simple
>expressions for:  real, possibly insuccessful, hunting, for example.
>And this is not a minor problem, since, if you look at the literature,
>it turns out that all the logical and semantic vocabulary ("entail,"
>"true," etc.) and most of the metascientific ("cause," "explain,"
>"predict," etc.) are of that intentional/intensional/propositional group
>that create contexts of the problematic sort.  So great slabs of one
>line of justification for Lojban has just gotten markedly (factor of 2+)
>more complex.

I'm not sure here what you say has gotten lost.  I still favor
naturalistic forms over formal ones. tu'a strikes me as a way to keep
naturalism while acknowledging the reality of sumti raising.  A possibly
irrealis "lo" also does.

I think one early root of the problem came from our trying to resolve
the multiple interpretations of "John seeks a bicycle or a fish", which
ran afoul of multiple intensionality and existence problems.  Cowan and
I wanted to claim on sci.lang that Lojban could handle all varieties of
meaning unambigiously (I'm not sure that we actually did so claim).

Otherwise, whatever may have happened to "ro" has been largely through
non-use and non-decision.  As I said - I avoid universals like the
plague.  Debate on le vs. lo almost always centered around unicorns, but
my one other major goal was that "lo" NOT become identical to the simple
English indefinites a/some.  "quantifier+broda" has never been much
discussed until recent times - it was always assumed to be short for
"quantifier lo broda" because of lack of thought and the desire to NOT
have hidden bound variables (as equating it to "quantifier da poi broda"
would have).  We don't want "non-logical" usages to imply a logical
structure.  I think a lot of the problem with logical usages in English
is that we CANNOT tell whether a usage is intended to be logically
rigorous or not - hence rules like that of double negation become
prescribed.  In Lojban, using a bound variable or a prenex is prima
faciae evidence of intent to be formally logical (and it may be the ONLY
such evidence - virtually all other usages have proven to have
natuiralistic approximations of some kind built into them).

Otherwise, debates over the x2 of "sisku", and possible indirect
influences of Carter's language on Cowan on his thinking (I never
understood, much less bought, Carter's arguments on sets and extensions
and their interaction with predicates.)  I opposed the change in sisku
and lost, but am happy to be restoring it to naturalistic meaning
whenever I update the gismu list if I haven't done so already.

Finally, if there has been slippage away from naturalism, it MAY have
been from Jorge and And carrying on their highly unnatural debates at
infinitum.  Since Jorge has been the major exemplar of the language on
net, his bent towards formalism MAY have exercised some force away from
naturalism.  But I can't say since I don't read the net texts.

The "any" debates may have changed some minds about the complexity of
problems, and maybe even the answers to them.  But since no one ever
wrote any proposals or conclusions, I assert that NOTHING has changed
from before that debate - everything is still up in the air.

So I am not sure I see that Lojban has left the "natural mode".
Certainly, the Lojban Central dialect of the language is totally
unaffected by all the debates of the last 2 years, and is probably still
close to that of 1988, with the exception of my Russian experiences
having shifted me heavily towards perfective tenses over aorist ones.

>As for my special clout, sure!, again.  If I had it, we wouldn't be in
>this jam in the first place -- or would be out of it by now and into a
>very different one, the one I thought we were in originally (specifying
>external referents within opaque contexts), which we can deal with by
>careful forethought, to be sure.

pc - let it be perfectly clear.  During this prescriptive phase, on
matters logical you have ENORMOUS if not ULTIMATE clout.  When you make
a definitive decision.  Prior to your getting on net, we only had a
chance to get decisions from you by cornering you on the phone going
into depth possibly with the use of a stack of Lojban List printouts.
And I was always pretty careful to force you into a commitment to an
answer in spite of your wanting to change your mind.  I wanted you on
the net because it was clear to me that I did not understand the issues
that were being raised enough to translate them into design decisions or
you to make.  There has been NO significant changes to the logical
design of the language since the tense and negation papers were written
(and you were integral to those discussions), except the sumti-raising
tu'a/jai constructs, and we consulted with you heavily on them as well.
Everything else has been, as far as I am concerned, inconclusive debate
BECAUSE you have made no rulings.

WHEN you decide that something about the logical aspects of the language
is going counter to what you believe it should be and state your ruling
definitively, I am inclined to back you without question, perhaps only
pausing long enough to let one round of people trying to talk you out of
it (and to let Nora see it and get her two cents in since she is the
logical one around here) and if you hold firm, then that is it.

Go through the refgrammar and if something violates your sense of the
way the language should work in matters of logic, and again, people get
one chance to try to get you to change your mind.  The important thing
is that the refgrammar state the meaning of a construct if it has been
determined, and that if it has not, that the range of plausible choices
be stated as clearly as is known.

If you think we need some new cmavo for any feature, say so, preferably
by writing it up as a cmavo list line.  Cowan will assign an actual
value to it from the available list, and it will become part of the
language.

And if you feel that we cannot decide, and that it is acceptible
logically to let 1000 flowers bloom - well you applied that metaphor to
Loglan, and it has been an underlying philosophy of my design efforts
ever since.  Something you feel tolerates indecision needs not to be
decided, and it won't be.

In short, pc, *decide*, and it shall be decided.  Possibly up to the very
day that Cowan says the book is done insofar as he is concerned.  I am
pretty sure that all who are in the broadest sense associated with
"Lojban Central" feel this way.

(If you feel this puts a burden on your shoulders, I hope you enjoy it.
You have been a founder and leader of the Lojban project from its
inception, and the rest of us have always had you in the position that I
have hopefully made clear.)

>And:
>{lo nu la djordj ualas cu merko gugde ralju cu na fasnu} means {da zou
>da nu la djordj ualas cu merko gugde ralju kei gie da na fasnu}, which
>is false, since {no da fasnu gie na fasnu}.

No!  NO!  NO!NO!  NO! a thousand times NO!  It is this argument exactly
that causes me to reject "lo broda" = "da zo'u da broda".

We looped the loop on this debate a few years ago over zasti/na zasti as
well, and indeed that was probably when the inequality of the two
constructs became cast in concrete.  We changed the place structure of
"xanri" so it would mean something meaningfully distinct from "na zasti"
at the same time.

Iain:
>In another post you say something about making it difficult to say
>simple things.  I contend that one of the things that Lojban in
>particular highlights is that things that we express "simply" in
>natlangs frequently aren't at all simple.

Indeed.  BUT, we have historically created constructs that acknowledged
the complexity but were naturalistically simple in order to handle these
non-simple constructs.  Most of the bells and whistles in the language
are just such simplifications.  The debate seems to be over whether we
have covered all important hidden complications, and the upshot seems to
be that intensionality and opacity/transparency and whatever lambda is
are still not well covered in a natural way.  We have a proposal on the
table for lambda.  I'm hoping that this message if nothing else, gets
some proposal formalized on the other questions so we can move on.

lojbab