[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: le nu pilno (Was: coi...)
coi goran.
>> .i mi pacna lenu le mi bruna cu na merko kulnu catke kei li pa
>
>.i mi na jimpe .i catke ki'a? .i mi na kakne co zgana lenu zo catke .oi cu
>vanbi di'u .i le mi menri se jimte .ije'i? de'u klulacri
.i naje
.i mi pu srera lacri le glico kulnu
.i zoi gy "push" gy smuni lenu tu'a da bapli lenu de lebna di da
.i lu merko kulnu catke li'u trocytanru fo lenu bapli lenu lei la nePAL.
prenu cu lebna lei merko kulnu
.iki'ubo so'i le merko ke pacna jenmi cu djica lenu ctuca fo lei cmaci
.iku'i so'i le merko ke pacna jenmi cu djica lenu ctuca fo lei merko lijda
.a lei merko kulnu .a lei merko citka gi'e jinvi ledu'u lei drata kulnu cu
mleca lei merko kulnu leka xamgu
.i .u'u mi troci lenu jbosku lei nandu seni'i lenu mi binxygau lei nandu
sidbo lei traji frili vau .oi
co'o mi'e kris
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: i.alexander.bra0125@oasis.icl.co.uk
Subject: TECH: RE: do djica loi ckafi je'i tcati
la kris. bogart. cusku di'e
> Apparantly I was on vacation or simply not paying attention when this got
> resolved. .u'u Sorry.
No regrets necessary - I don't think we've got all the wrinkles out of
this yet.
la xorxes. cusku di'e
> On the other hand, it can be thought in a different way and it doesn't work.
> I'll change to {nitcu} instead of {djica} to avoid having to use {tu'a}.
Ummm... {nitcu} is one indication we haven't got all the answers yet.
Being _able_ to use {tu'a} gives us a way of making a distinction
which isn't easy otherwise.
> What is the meaning of: {mi nitcu lo tanxe}?
> Is it "I need something which is a box", or is it "there exists at least
> one box such that I need it"?
I believe it has to be the latter.
> > 9) "do djica tu'a loi ckafi ji loi tcati lu'u"
> Well, I agree this is a solution, but I don't think it is the right general
> solution. In the case of {djica}, it would seem that {tu'a} has to be used
> anyway to avoid illegal sumti raising, so that the example is a bad one. When
> there is no sumti raising, e.g. {do nitcu lo tanxe ji lo dakli}, using {tu'a}
> seems wrong.
I look at it from the opposite point of view. I took examples like these
as indications that there is some sumti raising going on. But there's
a nagging suspicion at the back of my mind that this isn't the whole answer.
We have several gismu, one of whose places may need to be filled with
something of the form "any-old-<x>". Last I remember, there were three
different styles of place definition for these - {sisku} takes a property,
{djica} takes an event, and {nitcu} takes a concrete sumti. It could be
that all gismu of this type (and I'm not sure how you spot them all)
need to take an abstraction, at least as an option, in which case {tu'a}
works. And in that case any such gismu which didn't allow an abstraction
would not be usable to express the any-old-<x> case.
Or maybe there's something else going on. Natlangs seem to avoid the
issue, or use constructions like any-<x>-whatever to emphasise the point.
But I don't see how you carry that over into a logical language.
> {do nitcu lo tanxe ji'e dakli} may be all right, depending on
> what is the answer to my question above, but an appropriate sumti connective
> would be nice too.
> Since it would be very simple to allow BAIs to work like that (they're already
> allowed in forethought form, so why not in afterthought also?), I don't see
> any reason not to.
> At least {mau}, {me'a}, {du'i}, {li'e}, {pa'a}, {fa'e}, {ba'i} and {do'e} can
> be given good use in this function.
This appears to be off into hyperspace. I hope I'm misunderstanding you. :-)
mu'o mi'e .i,n.
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
mi cusku di'e
> > Since it would be very simple to allow BAIs to work like that (they're
> > already allowed in forethought form, so why not in afterthought also?),
> > I don't see any reason not to.
>
> > At least {mau}, {me'a}, {du'i}, {li'e}, {pa'a}, {fa'e}, {ba'i} and
> > {do'e} can be given good use in this function.
i di'e la'edi'u pinka la i,n
> This appears to be off into hyperspace. I hope I'm misunderstanding you. :-)
Let me explain. The problem is that the subject shifted a bit from the
original question.
I wanted to say "I need the box or the bag (I don't care which)".
(I'm changing from "a" to "the" so as not to confuse with the other issue.
This time it's about a specific box and a specific bag.)
{mi nitcu le tanxe .a le dakli} is wrong, because if I need the box, but
not the bag, the sentence is true. In fact, I'm not claiming anything
about "I need the box" or "I need the bag" separately, so I can't use
logical connectives that by definition can be split into two sentences.
My proposed solution was to say {mi nitcu du'igi le tanxe gi le dakli},
and I wanted to be able to say the same thing using afterthought mode
instead of forethought: *{mi nitcu le tanxe du'ibo le dakli}
The other BAIs that I listed too can be useful as connectives.
Jorge
Date: Thu, 15 Sep 1994 00:23:46 -0400
From: Logical Language Group <lojbab>
Subject: bogus file processing for gismu list needed ASAP
The gismu list is in our directory along with the list of bogus entries that
you omitted the first time. Can you do the KWIC run for those entries now.
As I posted to the list, I am 85% done. Specifically, i have 1231 lines
left out of 8931 (hmm, guess that is 86%), and I have done more than 400
lines a day for the last couple of sessions. Thus I think I will be ready
for the bogus kwic file by Sunday or Monday. Nora is on vacation next week,
and will be at home, so I expect an especially productive week next week,
hence I may be able to get done or close to done with all gismu files by the
end of next week.
(I have already looked at the cmavo list files and they should go much
quicker. There is a limit to what we can do with E-order for the cmavo list,
but the KWIC-index seems to have maximized it. Lots of weeding to do,
but weeding is the fast part of the job, as it has turned out.)
lojbab
From: i.alexander.bra0125@oasis.icl.co.uk
Subject: TECH: Any old thing whatsoever (was RE: do djica loi ckafi je'i
tcati)
Subject: TECH: Any old thing whatsoever (was RE: do djica loi ckafi je'i
tcati)
cu'u la'o gy. Randall Holmes .gy.
> This seems only to go to you via my reply function, and I don't know
> how to e-mail the whole list; I'll try to do this in the cc: line at
> the end of this message.
It sounds from Mark's comment as if it got through.
> The same problem arises in TLI Loglan; the paradigmatic example which
> caused a lot of discussion was "I am waiting for a taxi". The
> difficulty seems to be that the logical form of the sentence is an
> illusion; there is no box referred to in "I need a box", and there is
> no taxi referred to in "I am waiting for a taxi" (there need not even
> exist any boxes or taxis meeting your requirements for the statement
> to be true). The context is "referentially opaque", in Quine's
> terminology, and the object of the sentence, if it has one, is some
> kind of "intensional" object (something on the order of a concept of a
> box or taxi).
> --Randall Holmes
> ("logician in residence", TLI)
Yes, but *what* sort of "intensional object"?
This attemps to identify the problem, but I'm not sure how much
it helps us find the solution.
cu'u la mark. clsn.
> I recall we went through this discussion once before; in fact it was
> spurred on by a similar discussion regarding TLI Loglan regarding taxis
> (mentioned by Randall Holmes here, I see). The answer there (our analogous
> version of JCB's I think, and I liked it) was "loi tanxe". This works. I
> need [some part of] the mass of things that are boxes. Possibly "lei
> tanxe" if you want to admit something that isn't a box but turns out to be
> what I meant anyway. I don't think we need a new quantifier for this one;
> massification works (unless massification was rethought and redefined since
> the last time this question came through and I missed it). I'll try to
> find quotes from the last time.
(I'm with Jorge on this one.)
Sorry, Mark, I didn't really buy this the last time round,
and I think I understand better why now.
mi nitcu loi tanxe
means
There is some part of the mass of things that are boxes
that I need.
In other words, it suffers from the same problem as the {lo} version.
It's more difficult to think of examples where you would actually
want to say this, but I firmly believe that it has to work this way.
Massification is irrelevant.
Jorge's {xe'e} = "any":
This strikes me as pretty dubious semantically (even more
problematical than {po'o} = only). "Any" is kind of ambiguous
between "all" and "one".
xe'eti ka'e se pilno
Any of these will do.
might as well be
roti ka'e se pilno
All of these are usable.
Each of these is capable of being used.
(I've previously on occasion advocated translating {ro} as
"each" rather than "all". It means the same in the simple cases,
and helps demonstrate problems similar to the one we're discussing
here in the more complicated ones.)
The most promising candidate for an answer seems to be that
there is some sort of abstraction which has been elided.
(This is presumably something like what Quine has in mind
as an "intensional" object. (Haven't read Quine - should get
to the top of the to-do list sometime in the next ten years.) :-) )
But is it a potentially different abstraction in each case,
or is there conceivably a single general purpose one which
would cover them all?
{tu'a} appears to serve in all the cases we've considered
(except perhaps {sisku}, and I seem to remember John Cowan
having second thoughts about the use of {ka} for that one),
although it's just about as vague as it could possibly be.
Or should it perhaps be a Lojban abstraction,
{su'u <gismu> kei beleka broda} for some suitable {broda}?
A single one, or individually chosen?
cu'u la'o gy. Chris Bogart .gy.
> Reading this over, I'm unsure whether I really want to claim "ka" is just
> the right word; suppose there were a new abstractor with the same grammar
> that filled this function -- I have no idea how you'd define it in English,
> though.
Quite.
> But we shouldn't *require* nitce to take an abstraction because it's still
> useful to be able to say "mi nitce lo tanxe" if it is in fact a specific box
> you are referring to, rather than just a box in general.
Well, yes and no.
In theory, even if {nitce} (etc.) takes an abstraction, you can always
adjust the quantification to achieve that effect, e.g.
da poi tanxe zo'u: mi nitce tu'a da
In practice, however, that's admittedly not very convenient.
Sorry, no answers today, just questions.
Yours,
Puzzled of Reading.
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
In-Reply-To: (Your message of Tue, 13 Sep 94 19:53:50 EDT.)
(A)
Jorge:
> Maybe what we need is a new quantifier "any" in selma'o PA.
> Let's call it {xe'e}.
>
> xe'eti ka'e se pilno
> Any of these will do.
>
> mi na vecnu fo xe'eda
> I won't sell at any price. (But I may sell at some price :)
>
> do ka'e cuxna paxe'e selska poi xekri
> You can choose any colour, as long as it is black.
[^^ - do we have an anglophile here?]
Wouldn't plain universal quantification suffice for these?
For every x, x is one of these, x will do
a price, I won't sell at x
a colour, you can choose x
(B)
My contribution to the needing a box, waiting for a taxi, seeking
a unicorn debate is to suggest use of a NU cmavo meaning 'hypothetical
event' - "xuhu", say. Then we could have:
I need it to be the case (though it might not be the case) that there
is x, x a box and I have x.
I wait for it to be the case (though it won't necessarily be the case)
that there is x, x a taxi.
Galadriel wants it to be the case (though it might not be) that there is x,
x a unicorn & Galadriel finds x
where "it to be the case (tho it might not be)" = "le xuhu"
I make this contribution having completely forgotten what John Cowan
& Iain Alexander ha d to say on the matter a fair while back.
----
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi
je'i
la lojbab cusku di'e
> I don't see the problem. If you want exactly one box, you want
> 'pa lo tanxe' or 'pa tanxe'. That means 'any box' unless I have
> missed the point of this discussion.
{mi nitcu pa tanxe} means "there exists exactly one box such that
I need it". That's not what I usually mean by "I need a box".
la i,n cusku di'e
> Jorge's {xe'e} = "any":
>
> This strikes me as pretty dubious semantically (even more
> problematical than {po'o} = only). "Any" is kind of ambiguous
> between "all" and "one".
We can make it as unambiguous as we care to. "Any one of all" is
not the same as "each" and is not the same as "one".
(I would prefer that {xe'e} not be restricted to one, but this would
probably be the default. Then {paxe'e}, {rexe'e}, etc. for "any one",
"any two" etc.)
> xe'eti ka'e se pilno
> Any of these will do.
>
> might as well be
>
> roti ka'e se pilno
> All of these are usable.
> Each of these is capable of being used.
Yes, because the English expression in this case is somewhat ambiguous,
but the meanings are different. In the first case I say that only one can
be used, but it can be any of these. The "innateness" of {ka'e} complicates
the issue, though. How about
mi ba dunda xe'eti do
I will give you any (one) of these.
It is different from
mi ba dunda roti do
I will give you each of these.
and different from
mi ba dunda pati do
I will give you one of these. (Guess which one.)
> (I've previously on occasion advocated translating {ro} as
> "each" rather than "all". It means the same in the simple cases,
> and helps demonstrate problems similar to the one we're discussing
> here in the more complicated ones.)
I agree. "Each" is much better, because "all" can suggest a massification
that is not there.
BTW, what is so problematic about {po'o}? I've seen this complaint a few
times, but never read any description of what the problem might be.
And now that you mention it, it may well be that an attitudinal of the
{po'o} type could be used instead of a PA, as I proposed.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
la and cusku di'e
> (A)
> Jorge:
> > Maybe what we need is a new quantifier "any" in selma'o PA.
> > Let's call it {xe'e}.
> >
> > xe'eti ka'e se pilno
> > Any of these will do.
> >
> > mi na vecnu fo xe'eda
> > I won't sell at any price. (But I may sell at some price :)
> >
> > do ka'e cuxna paxe'e selska poi xekri
> > You can choose any colour, as long as it is black.
> [^^ - do we have an anglophile here?]
>
> Wouldn't plain universal quantification suffice for these?
> For every x, x is one of these, x will do
> a price, I won't sell at x
> a colour, you can choose x
The {ka'e} ones are tricky, because of the "innateness", so I would change
to {kakne} for those examples.
The difference between
do kakne le nu cuxna ro selska
You can choose every colour.
and
do kakne le nu cuxna paxe'e selska
You can choose any one colour.
is that in the first one you allow to choose more than one colour,
while in the second one you don't. Of course,
do kakne le nu cuxna pa selska
You can choose one colour.
makes you ask "which one?". (In this case black, but that was only to
confuse the issue even more.)
The selling example is confusing because of the negative, let's start
with
mi vecnu fo xe'eda
I sell at any price.
which obviously is not the same as
mi vecnu fo roda
I sell at every price.
so the negation of each of them is different.
> (B)
> My contribution to the needing a box, waiting for a taxi, seeking
> a unicorn debate is to suggest use of a NU cmavo meaning 'hypothetical
> event' - "xuhu", say. Then we could have:
> I need it to be the case (though it might not be the case) that there
> is x, x a box and I have x.
You can do this with the existing abstractions:
mi nitcu le nu mi ponse lo tanxe
I need the event that I possess a box.
in no way implies that "le nu mi ponse lo tanxe" should actually ever
be the case.
But I don't like solving it by kicking the problem into abstraction space.
We'd end up saying
mi nitcu tu'a lo tanxe
I need some abstraction about a box.
instead of saying what we really mean. Why should we make it so difficult
to make a distinction that can be made clearly in other languages?
Jorge
Date: Fri, 16 Sep 1994 00:10:49 -0400
From: Logical Language Group <lojbab>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi je'i
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
JL>la lojbab cusku di'e
JL>
JL>> I don't see the problem. If you want exactly one box, you want
JL>> 'pa lo tanxe' or 'pa tanxe'. That means 'any box' unless I have
JL>> missed the point of this discussion.
JL>
JL>{mi nitcu pa tanxe} means "there exists exactly one box such that
JL>I need it". That's not what I usually mean by "I need a box".
No, I don't think that is a correct translation. It means "I need exactly
one out of the set of things that 'box'. It is not specific as to which of
the set of things is needed, merely thatthere is a single thing needed,
and it veridically is a box. (I can't remeber what it means if there is
no such thing as a box, cf. "I need a unicorn", but it has been discussed.)
I think your translation is expressed by "pa da zo'u da tanxe gi'e se nitcu mi"
"lo" as we have defined it is non-specific as to what member(s) you select if
you select a specific number of them less than 'all'.
TLI Loglan does not have a "lo" - their closest equivalent "lea" is pretty
much equivalent to "rolo", but I do not believe there has been any discussion
of usage with a non-"ro" quantifier before the "lea" (it is probably permitted
by their grammar, but it is not necessarily defined in meaning). TLI
Loglan has only "le" for individuation, and "le" is indeed specific in both
versions of the language. Of course "le" is also intensional and thus a
speaker can claim that any useful occurance of a box that turns up just
happens to be 'the' box that he had in mind.
You might also be able to do something with "pa lu'a roda poi tanxe (or
ro lo tanxe, or ro tanxe).
But I still think we, unlike TLI don;t really have a problem with "lo",
and we SHOULD like TLI, use "loi" (which in TLI Loglan is "lo" for the
benefit of R Holmes).
lojbab
X-Sender: cbogart@teal.csn.org
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Fri, 16 Sep 1994 00:07:22 -0600
From: cbogart@csn.org (Chris Bogart)
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
X-Mailer: <PC Eudora Version 1.4b17>
>I don't see the problem. If you want exactly one box, you want
>'pa lo tanxe' or 'pa tanxe'. That means 'any box' unless I have
>missed the point of this discussion.
The discussion hinges on the fact that all lojban sumti can be quantified in
the prenex without any change in meaning. "mi nitce pa tanxe" means the
same as "da poi pa tanxe zo'u: mi nitce da". I.e. There's a box out there
somewhere, and I need it. It implies there is some box in existence, and
you're saying that you need that particular box.
But suppose I need a huge box to put my house in. No such box exists. I can
look disapprovingly at my big, unboxed, house, and say "I need a box", but
it would be wrong to say "da poi pa tanxe zo'u: mi nitce da", which claims
not only that I need a box, but that such a box already exists!
Or suppose I've got a bunch of materials for making boxes, and I say "I need
a box" with the intention of building my own. There does not exist some box
X, such that I need X. There may not even be a future box "le ba tanxe"
because I may fail in my attempts to build one.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Chris Bogart <cbogart@CSN.ORG>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
>No, I don't think that is a correct translation. It means "I need exactly
>one out of the set of things that 'box'.
So how do you make the distinction in Lojban between Quine's examples of "I
want a sloop" and "There is a sloop I want"?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi
je'i
la lojbab spusku di'e
> JL>{mi nitcu pa tanxe} means "there exists exactly one box such that
> JL>I need it".
>
> No, I don't think that is a correct translation. It means "I need exactly
> one out of the set of things that 'box'. It is not specific as to which of
> the set of things is needed, merely thatthere is a single thing needed,
> and it veridically is a box.
Suppose there are three boxes of different sizes, and I only need the biggest.
Does {mi nitcu pa le ci tanxe} mean that exactly one of the three boxes
(the biggest) is needed by me, or that I need any one of the three?
> I think your translation is expressed by "pa da zo'u da tanxe gi'e se nitcu
mi"
Yes, but I thought that {mi nitcu pa tanxe} means exactly that.
Does {mi ponse pa tanxe} mean the same as {pa da zo'u da tanxe gi'e se ponse
mi}?
If so (as I think) then why should it work for {ponse} and not for {nitcu}?
> "lo" as we have defined it is non-specific as to what member(s) you select if
> you select a specific number of them less than 'all'.
I agree that it is non-specific. The issue is whether it is "identifiable"
or "non-identifiable" (probably the wrong technical terms).
> But I still think we, unlike TLI don;t really have a problem with "lo",
> and we SHOULD like TLI, use "loi" (which in TLI Loglan is "lo" for the
> benefit of R Holmes).
I don't think {loi} works in the sense of "any whatsoever".
mi nelci loi xruli
I like flowers
doesn't claim that I like any flower whatsoever, does it?
Why should
mi nitcu loi xruli
I need flowers
mean that I need any flower whatsoever?
I don't think there's any problem with {lo} either. The only problem is
that we don't have any easy way to refer to that sense of "any". (Something
like "ajn" in Esperanto.) I think that either a PA with that sense, or an
attitudinal like {po'o} could be the answer.
Abstractions work too, in the sense that I can say:
mi nitcu le nu mi ponse lo tanxe
I need that I have a box.
and I suppose this is the sort of problem that made {djica} unable to
take a simple object. But I don't like that solution, because it is either
too complicated, forcing you to use {ponse} or some other relationship
when you don't want to, or it is too ambiguous {mi nitcu tu'a lo tanxe}.
There has to be a way to say: "I need any box whatsoever."
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: do djica loi ckafi je'i tcati
la djer cusku di'e
> I rather liked my first suggestion, as corrected by jorge:
>
> mi nitcu le su'u me le taxpu'i me'u da kei
> I want the in-mind abstracton about boxing things.
I think I muddled things more if this comes out of my comment.
{me le taxpu'i} is something specific to someone who puts thing in boxes.
I think {mi nitcu le su'u taxpu'i [da]} is approximately what you want
"I need the abstraction of boxing something"
But I would use simply {nu}: {mi nitcu le nu taxpu'i da} "I need to
box something".
This is sidestepping the issue, though. I still would like to say
"I need a box" without having to specify what I need it for.
> And now I am beginning to be haunted by a Beatles lyric:
> "Something in the way she walks.."
> How do you say that "something" in lojban when you don't really want it
> instantiated to anything known?
I like lojbab's {le su'u ko'a cadzu}.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi
je'i tcati)
la lojbab cusku di'e
> A major reason why "loi tanxe" should work in Lojban is that in Lojban,
> singular plural is a marked distinction, AND that all nouns can be considered
> as mass nouns.
Nobody says {loi tanxe} doesn't work. What I'm saying is that it doesn't
mean "any box", it means "[part of] the mass of boxes".
So, I could say
mi pilno loi tanxe le nu vasru lei mi cukta
I use boxes to keep my books
there's no problem with {loi tanxe}, it works, as a mass of boxes.
> You do not, in English, typically say "I need *a* water", you say "I need
> water", and leave the quantity of the mass to be determinable from context.
Exactly. Just like in Lojban
mi nelci loi djacu
I like water (not "I like any water")
Now, how do we say "I need water"? It can't be {mi nitcu loi djacu}, because
that says that there is some part of the mass of water that I need, and
that's not what "I need water" means.
> If you need to be specific, you say something like "I need a glass of water",
> or "I need a liter of water". One way to do this in Lojban, for boxes is
> "mi nitcu pa selci poi tanxe" "I need one indivisible subunit of the mass
> of boxes."
It's the same problem. Is {pa selci} one (certain) selci, or is it any selci?
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi
je'i
> Back to present, Mark Shoulson reporting.
>
> That said, it looks to me like there may be more than one thing at work
> here. On the one hand, things like "I like tennis"" or "I like women"
> should use "loi".
Agreed.
> But "I need a box" (and possibly tho not necessarily I need a taxi) may be
> different, since you're not referring in general... precisely because *NOT*
> "any of" the mass will do.
I think I may be changing my mind.
What confused me was the quantification of loi as "part of the mass of..."
{mi nitcu loi tanxe} could well mean that the _whole_ mass of boxes is
such that I need it, and since the properties of the individuals are
also the properties of the mass, then as long as I need one of them
I need the whole mass. Actually, there isn't one box such that I need it,
but the mass has more properties than each individual, so we'd still be ok.
> A full box won't help you. Ormaybe "lei" will
> help there. I dunno..
It doesn't matter. You're saying that the mass of boxes has the property
that you need it, not that the components of that mass have the property.
This makes me wonder what's the point of quantifying {loi} with anything
other than {piro}, since anything that is true for one quantification
should be true for any other.
So {mi nitcu piro loi tanxe} might make sense for "I need a box", but
is very counterintuitive.
The new problem is that "I need a box" doesn't mean that I need any box
whatsoever. Maybe I need a big box, in which case "I need a box" is true,
but "I need any box whatsoever" is not. And I don't think {lei} helps,
it's not a matter of specific boxes...
> For consideration...
Enough for today, I'm going home.
Jorge
Date: Sat, 17 Sep 1994 00:41:07 -0400
From: Logical Language Group <lojbab>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
Cc: lojbab@access.digex.net, pcliffje@crl.com
As I posted a little while ago, our "lo" does not claim existence of the set
(non-empty, that is) meeting the descrription. Thus statements about
"lo (unicorns)" can be made and be meaningful - the unqiuantified form
describes at least one of a hypothetical set that veridically meets the
description, is non-specific as to which of such a set is to be selected.
Does your "lea" claim that the referenced set exists? Or do claims
about "lea (unicorns)" become isomorphic to claims about "lea (pegasi)"
because both are descriptions of the empty set (at least in the real world)?
My recollection is that "lea" gives ALL members of such a set (whether it
exists or not). Our "lo" just has a different quantifier and gives "at least
one" of the set.
pc has said that "lo (preda)" is not exportable to the prenex of the sentence;
it is not the same as a restricted quantified variable in the prenex
da poi preda zo'u: mi nitcu da
ba jio preda goi: mi cnida ba in TLI vocabulary per L1
whether lea (default quantifier "ra" which is Lojban "ro")
id the same as "rolo" depends on whether the referent of the description can
be empty and yet meanuingful. We ttried to define it to meet the logical
as well as the linguistic requirements by separating it from the quantified
version.
lojbab
Date: Sat, 17 Sep 1994 00:48:35 -0400
From: Logical Language Group <lojbab>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
Not having read Quine (and not being convoinced that I would understand
it if I did - if I had to be a logician to lead this project, I would
resign yesterday %^)
I beleive that the distinction is
mi djica lo bloti poi falnu se catke
I want a boat which is sail-pushed (I hope that will do for sloop in this
example)
which does not claim that such a boat exists
and
(pa) da poi bloti gi'e falnu se catke zo'u: mi djica da
There exists 1 thing that is a boat and sail-pushed such that: I want it.
which claims that such a boat exists AND that it is a specific boat.
(I left the "pa" out from before "lo bloti" in the firts example, if
singular is necessarily explicit in "a boat". Lojban does not require
singular/plural distinction if it is not important.
lojbab
Date: Sat, 17 Sep 1994 01:07:20 -0400
From: Logical Language Group <lojbab>
Subject: Re: Updated gismu list
Cc: lojbab@access.digex.net
Well, I sent you a cc of one message to Holmes that will give you some
of the ghist of the issue at hand. Traffic is running a little heavy on
the issue, but I am sticking my neck out as usual, in hopes that uyou
will eventually bail me out.
It is the same old issue of intensionality of "lo" - do all references to
veridical descriptions claim that the referent really exists, and if not,
can we avoid haviung the silliness that all claims about the empty set are
equivalent so that "I want a unicorn", "I want a balrog" and "I want the
Wizard of Oz" to all mean the same thing and have the same truth value.
I think we resolved things so that this is not the case - that "lo" is
intensional as to existence but veridical with respect to the description
applying to the referent, and that "lo" is non-specific (unless restriucted),
but I won't pretend to understand the magic that makes thinsg work.
Lots of references to Quine (Word and Object, I think - I have never read
Quine and don't think I have any of his books if I wanted to) in the
discussion.
If catching up on the old issues is going to be a long term problem, we need
to consider whether the priority should be to keep up on new issues over
catching up. Otherwise, you may sit there forever coasting along a year
behind our current issues, which probably won;t satisfy either of us.
I think that (when you have the computer to use) if you are reading and responding to stuff as it appears, or shortly thereafter, you will tend NOT to get
500K files to read. Even this very lively discussion going on is only a
half-dozen messages a day. (WE just got statistics that the 3 year average for
Lojban List is 4 messages per day).'
I think the main things that only you can provide are - for current issues
keeping us from getting ourselves up the logical creek without a paddle.
(Does that make you the paddle??? %^), and for old issues, 1) making sure
that if something really needs to be changed, we identify it befoire the
book is done and 2) eventually weed out what is right and wrong in the
discussions so that we can use them as answers when the topic inevitably
comes up again (at least 3 times in the last year in the case of ZAhO tenses).
lojbab
Date: Sat, 17 Sep 1994 01:28:13 -0400
From: Logical Language Group <lojbab>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi je'i
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
JL>This is sidestepping the issue, though. I still would like to say
JL>"I need a box" without having to specify what I need it for.
Then you want to use "le
tanxe" and not "lo tanxe" or "pa tanxe" which is related to "lo tanxe"
(specifically "pa lo tanxe".
"pa le tanxe" means one of 'the' boxes I have in mind. I may not be able
or willing to specify all the relevanmt properties of "le tanxe", but at
least it is possible to ask me if it isn;t clear.
Whenever you use "lo" you are, at least logically, not constarining the set
any more than is indicated by restrcitions you provide.
JL>Suppose there are three boxes of different sizes, and I only need the
JL>biggest.
JL>
JL>Does {mi nitcu pa le ci tanxe} mean that exactly one of the three boxes
JL>(the biggest) is needed by me, or that I need any one of the three?
I think the latter. You want
mi nitcu le pa le ci tanxe
JL>> But I still think we, unlike TLI don;t really have a problem with "lo",
JL>> and we SHOULD like TLI, use "loi" (which in TLI Loglan is "lo" for the
JL>> benefit of R Holmes).
JL>
JL>I don't think {loi} works in the sense of "any whatsoever".
JL> mi nelci loi xruli
JL> I like flowers
JL>
JL>doesn't claim that I like any flower whatsoever, does it?
JL>
JL>Why should
JL>
JL> mi nitcu loi xruli
JL> I need flowers
JL>
JL>mean that I need any flower whatsoever?
Because you have massified the set "lo'i xrula", and any portion of that
mass will suffice. Now this is logic - when you say "I want water", you
will not be satisfied by poisoned water, or water embedded in the crystalline
structure of a hydrated rock, etc. So pragmatics may give SOME limits here.
But logically at least, "loi xrula" refers to ALL flowers.
"lei xrula" refers to a more specific set of flowers, like "le xrula" would.
(I think that when the Trobriand Islanders consider all rabbits to be an
instance of Mr Rabbit, they do have some pragmnatic restruictions as well -
I doubt that they consider a dead, cut-up rabbit in their stew as being
EXACTLY the samne thing as the one sitting in the field, but linguistically
they may not make the distinction - I wouldn't know for sure, though.
Similarly, there are pragmatic restrictions on "lo djacu" or "lo tanxe".
I suspect that this would best be dealt with by restricting the universe of
discourse intensionally, in which case 'all flowers' does not refer to
every flower that ever existed anywhere, and 'water' doesn;t refer to the
scattered molecules in interstellar space.)
Given these kind of pragmatic restrictions, "lo"
can serve the purpose needed. But "lo broda" still is NOT the
same as "da poi broda" in that it doesn;t claim that "broda" really exists.
lojbab
Date: Sat, 17 Sep 1994 01:39:34 -0400
From: Logical Language Group <lojbab>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi je'i
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
JL>
JL>{mi nitcu loi tanxe} could well mean that the _whole_ mass of boxes is
JL>such that I need it, and since the properties of the individuals are
JL>also the properties of the mass, then as long as I need one of them
JL>I need the whole mass. Actually, there isn't one box such that I need it,
JL>but the mass has more properties than each individual, so we'd still be ok.
JL>
I THINK that the default quantifier "piso'u" on "loi" is needed BECAUSE
properties of different parts of the mass may not all be identical.
THus saying that "loi" has that quantifier means that the properties of
the individual are also the properties of PART of the mass. The mass as
a whole can have self-contradictory properties.
This may answer you other question - liking "loi xrula" does not necessarily
mean that you like ALL flowers, but that you like some unspecified portion
of all flowers. It is possible that you like ALL flowers and that you dislike
ALL flowers, if "ALL flowers" is piro loi xrula and you like at least one
flower and dislike at least one flower.
Statements about "piro loi ..." are just as universal as statements about
"ro da poi ...", but less useful because the opposite may also be true.
lojbab
From: Gerald Koenig <jlk@NETCOM.COM>
Jorge said this and a lot more:
There has to be a way to say: "I need any box whatsoever."
Jorge
__________________________________________________________________________
I think the below sentences translate this accurately. The style is
another question. I would say too that the word "any" does give rise to
a number of problems of ambiguity in English and carrying it over to
lojban might be a problem there.
1. roda tanxe da inaja mi nitcu da
For all X: X is a box implies I need X.
If X is a box, I need it.
2.roda poi tanxe ku'o mi nitcu da
For all X which is a box, I need X
I hope inaja is the correct form for logical implication as usually
represented in logic texts by an arrow --->. The if... then... form.
It is said of plays that they are not written, they are wrought. I think
the same could be said of lojban.
djer jlk@netcom.com
Date: Sat, 17 Sep 1994 02:14:15 -0400
From: Logical Language Group <lojbab>
Subject: G. Koenig on 'any'
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
GK>I think the below sentences translate this accurately. The style is
GK>another question. I would say too that the word "any" does give rise to
GK>a number of problems of ambiguity in English and carrying it over to
GK>lojban might be a problem there.
GK>
GK>1. roda tanxe da inaja mi nitcu da
GK>For all X: X is a box implies I need X.
GK>If X is a box, I need it.
I think I know what you were trying for, but that isn't what you got.
roda tanxe d
roda tanxe da (that is)
(Absolutely) Everything is a box for containing itself.
You wanted
roda zo'u tu'e da tanxe .inaja mi nitcu da
For all x: [x is a box => I need x]
GK>2.roda poi tanxe ku'o mi nitcu da
GK>For all X which is a box, I need X
Closer: just use zo'u - the prenex terminator, instead of "ku'o":
roda poi tanxe zo'u mi nitcu da
As you wrote it
nitcu has
"roda poin tanxe ku'o" as x1
"mi" as x2
"da" as x3
Every box needs me for purpose 'itself' (sumti-raised!) !!! %^)
In spite of this correction, I want to complement you on the great improvements
you've made in Lojban capability since the era when you were part of the LA
Group.
lojbab
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: TECH: Any old thing whatsoever (was RE: do djica loi ckafi je'
In-Reply-To: (Your message of Fri, 16 Sep 94 00:10:49 D.)
Lojbab:
> JL>{mi nitcu pa tanxe} means "there exists exactly one box such that
> JL>I need it". That's not what I usually mean by "I need a box".
>
> No, I don't think that is a correct translation. It means "I need exactly
> one out of the set of things that 'box'. It is not specific as to which of
> the set of things is needed, merely thatthere is a single thing needed,
> and it veridically is a box. (I can't remeber what it means if there is
> no such thing as a box, cf. "I need a unicorn", but it has been discussed.)
"I need one out of the set of boxes" is false if every box is such that
I don't need it.
"I need a box (any box)" is false is there exists a box such that I
don't need it *AND* I don't have one yet.
"I need every box" is false if there exists a box such that I don't
need it.
The first two are easy to distinguish. But how to capture the difference
between the last pair is the problem.
---
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: needing books
How do you distinguish, (preferably in non-pedantic usage):
I need there to be a specific book such that I have it.
There is a specific book such that I need to have it.
I need there to be x, such that x is a book & I have x.
There exists x, such that x is a book & I need to have x.
Yours in more than usual confusion,
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
In-Reply-To: (Your message of Thu, 15 Sep 94 15:01:46 EDT.)
Jorge:
> The difference between
> do kakne le nu cuxna ro selska
> You can choose every colour.
> and
> do kakne le nu cuxna paxe'e selska
> You can choose any one colour.
> is that in the first one you allow to choose more than one colour,
> while in the second one you don't.
Does this have something to do with the scope of quantifiers with
respect to modal operators?
You can choose every colour:
It is possible that for every x, x a colour, you choose x
You can choose any one colour
For every x, x a colour, it is possible that x is the one
colour you choose.
cf.
Everyone could marry me.
It is possible that everyone marries me
Anyone could marry me.
For everyone it is possible that they marry me/
For everyone it is possible that they be my one spouse
Or am I treeing up the wrong bark?
> The selling example is confusing because of the negative, let's start
> with
> mi vecnu fo xe'eda
> I sell at any price.
For every price it is possible that this is the (one) price I sell at
> which obviously is not the same as
>
> mi vecnu fo roda
> I sell at every price.
For every x, x a price, I sell at x
----
And
Date: Sat, 17 Sep 1994 09:49:12 -0700 (PDT)
From: "John E. Clifford" <pcliffje@crl.com>
Subject:
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
As usual, I do not yet know my way around here. Since generic passwords
do not seem to work at ftp.sc.yale.edu, what is the secret grip this time?
Golub may show up this afternoon -- or he may not, but i thought I'd get
a load of stuff anyhomw (M is away doing mothertending,so I have the
machine for a while).
pc>|83
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
la lojbab cusku di'e
> "lo tanxe" is NOT the same thing as "da poi tanxe";
This, I think, is to avoid problems in case no tanxe exists. But it doesn't
solve the "any" problem.
It seems to me that the problem with {nitcu} is different, and has little
to do with whether boxes exist or not.
> Now "mi nitcu pa tanxe", which is NOT restricted, does say that ANY member
> of the (unrestricted) set of things that 'are boxes' will satisfy your
> need.
Using that same logic, you would conclude that "mi ponse pa tanxe" says
that ANY member of the (unrestricted) set of things that 'are boxes' is
owned by you.
Let's say I have three boxes, one red, one blue, and one purple with
little pink flowers, and _any_ of them will serve for whatever purpose
they are needed.
Now you say {mi nitcu pa tanxe}
And I ask {xu do nitcu le xunre tanxe}
What should the answer be?
If {do nitcu le xunre tanxe} is true, and {do nitcu le blanu tanxe} is
also true, then it was false that {do nitcu pa tanxe}, because we've
shown that {do nitcu re tanxe} is true.
If {do nitcu le xunre tanxe} is false, and {do nitcu le blanu tanxe} is
also false, then we could go over the list for every existing box and
it would be false for all of them, then {do nitcu pa tanxe} would be
false, because we couldn't find any {pa tanxe} that made it true.
But there should be a true utterance to express that you need a (any) box,
without commiting to claiming that you need a (certain) box.
> I beleive that the distinction is
>
> mi djica lo bloti poi falnu se catke
> I want a boat which is sail-pushed (I hope that will do for sloop in this
> example)
> which does not claim that such a boat exists
Does {mi ponse lo bloti poi falnu se catke} claim that such a boat exist?
What's the difference?
Now for {loi}.
I think that {mi nitcu loi tanxe} might be at least part of the solution,
but I'm confused with the quantification of {loi}.
{piro loi tanxe}, the whole mass of boxes, has all the properties of its
members and then some more.
If I needed that red box over there, then it would be true that {mi nitcu
piro loi tanxe}, because since I need a member, I need the whole mass by
extension.
If I need any one box whatsoever, then the property of being needed by me
is not a property of any single one of the members, so it's not a property
that {piro loi tanxe} receives from them, but it well could be one of the
emergent mass properties. This makes some sense, if I need a box then I have
need of mass box, so {mi nitcu piro loi tanxe} is true.
It still doesn't tell me much, because it could be that what I need is that
red box there, so I didn't transmit the fact that any old box will do.
Now, other quantifications for masses confuse me. What do they really mean?
For example, can we say
pimu loi remna ka'e se jbena
Half of the mass of humans is innately capable of giving birth
Is it different from
piro loi remna ka'e se jbena
All of the mass of humans is innately capable of giving birth
The last one is true, because the mass inherits all properties of its
members, but then what does the other one say? And if it's true for
the whole mass, should it be true for 75% of the mass?
And supposing that {pimu loi remna} inherits only the properties of
half of all humans. Is it any arbitrary half, or is it certain half
(e.g. those humans that are female)?
And asks:
> If "waiting for a taxi" is "waiting for loi taxi", how do we say
> "we're waiting for two taxis". Does "reloi" do this?
No, and that's an excellent question. (I think reloi gives you two masses
of taxis.)
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: any and all
la djer cusku di'e
> There has to be a way to say: "I need any box whatsoever."
> __________________________________________________________________________
> I think the below sentences translate this accurately. The style is
> another question.
>
> 1. roda tanxe da inaja mi nitcu da
roda zo'u tu'e da tanxe inaja mi nitcu da tu'u
> For all X: X is a box implies I need X.
> If X is a box, I need it.
>
> 2.roda poi tanxe ku'o mi nitcu da
roda poi tanxe zo'u mi nitcu da
> For all X which is a box, I need X
Those say "I need every box".
But I need only one, so they can't be equivalent to "I need any box".
> I would say too that the word "any" does give rise to
> a number of problems of ambiguity in English and carrying it over to
> lojban might be a problem there.
I agree we shouldn't just have a word that means "any". What we need is
something to translate some of the things that can be said in English
using the word "any", and that seemingly can't be said in Lojban.
(There might be a way to say it using the existing words. If so, I'd
like to know what it is.)
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: needing books
la and reisku di'e
> How do you distinguish, (preferably in non-pedantic usage):
>
> I need there to be a specific book such that I have it.
mi nitcu le nu da poi cukta zo'u mi ponse da
Non-pedantic:
mi nitcu le nu mi ponse lo cukta
> There is a specific book such that I need to have it.
da poi cukta zo'u mi nitcu le nu mi ponse da
As non-pedantic as I can:
lo cukta zo'u mi nitcu le nu mi ponse ra
Or, (but lojbab disagrees):
mi nitcu lo cukta
> Jorge:
> > The difference between
> > do kakne le nu cuxna ro selska
> > You can choose every colour.
> > and
> > do kakne le nu cuxna paxe'e selska
> > You can choose any one colour.
> > is that in the first one you allow to choose more than one colour,
> > while in the second one you don't.
>
> Does this have something to do with the scope of quantifiers with
> respect to modal operators?
>
> You can choose every colour:
> It is possible that for every x, x a colour, you choose x
>
> You can choose any one colour
> For every x, x a colour, it is possible that x is the one
> colour you choose.
I think it does have a lot to do, but I can't write your second phrase
in Lojban using {kakne}.
> cf.
> Everyone could marry me.
> It is possible that everyone marries me
> Anyone could marry me.
> For everyone it is possible that they marry me/
> For everyone it is possible that they be my one spouse
>
> Or am I treeing up the wrong bark?
I think it's the right bark, keep treeing that it's helping a lot.
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: any and all
la xorxes cusku di'e
> Date: Sat, 17 Sep 1994 16:04:17 EDT
> From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
> Subject: any and all
> I agree we shouldn't just have a word that means "any". What we need is
> something to translate some of the things that can be said in English
> using the word "any", and that seemingly can't be said in Lojban.
> (There might be a way to say it using the existing words. If so, I'd
> like to know what it is.)
>
> Jorge
I think a solution which Lojbab suggested would work. When I'm
saying 'I need a(ny) box' or 'I need (any) two boxes' what I'm
really saying is that I need a set with a certain number of
members having certain properties. That can be simply expressed
in Lojban using {mei}:
mi nitcu lo tanxe pamei/remei
I need what amounts to a single box/ a pair of boxes
In another context I might say
mi nitcu le vu tanxe remei
I need that pair of boxes over there
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any and all
la veion cusku di'e
> I think a solution which Lojbab suggested would work. When I'm
> saying 'I need a(ny) box' or 'I need (any) two boxes' what I'm
> really saying is that I need a set with a certain number of
> members having certain properties. That can be simply expressed
> in Lojban using {mei}:
>
> mi nitcu lo tanxe pamei/remei
> I need what amounts to a single box/ a pair of boxes
{mi nitcu lo tanxe pamei} means practically the same thing as {mi
nitcu pa tanxe}. Is there an object, call it {pa tanxe}, call it {lo
tanxe pamei}, that is related to {mi} by the relationship {nitcu}?
If there is exactly one such object, then the sentences are true,
but that is not what "I need any box" means.
Jorge
From: "John E. Clifford" <pcliffje@CRL.COM>
Subject: Re: any and all
I came in late on this thread, but the examples I have seen, with nitcu,
do not look particularly "any" ish. In modal (intentional) and negative
contexts (both of which nitcu -- or at least "need" -- is) "any' behaves
like tight-scope "a"; "I need any box" amounts to just "I need a box",
provided that it is clear that it is not a particular box but any old one
will do. I suspect there is some infection in the example from the
notion of "all the boxes I can get" which might be expressed by "any
boxes you have" or so. I suspect that that is about some enough-ad of
boxes (drat, I wish I was remembering lexes again) or just "enough x such
that x is a box". "Any" is a pain: a context-leaper in some cases to
become a wide-scope universal (and that would be nice to have in Lojban,
if we do not have it already -- or is it the tight-scope one we need?)
and in other cases, as here, a tight-scope particular. Cf Vendler's
article in Encyclopedia of Philosophy (isthat where all this started?)
pc >|83
Date: Sat, 17 Sep 1994 22:33:40 -0400
From: Logical Language Group <lojbab>
Subject: Re: needing books
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
UC>How do you distinguish, (preferably in non-pedantic usage):
UC>
UC> I need there to be a specific book such that I have it.
UC> There is a specific book such that I need to have it.
UC> I need there to be x, such that x is a book & I have x.
UC> There exists x, such that x is a book & I need to have x.
1) mi nitcu lenu da zo'u da cukta gi'e se ponse mi
2) da zo'u tu'e da cukta .ije mi nitcu lenu mi ponse da
In 1), the "da" in the prenex within the "lenu" clause is not exportable
outside of that clause.
I think this covers your distinction, if I understand it correctly (no bets %^)
lojbab
Date: Sat, 17 Sep 1994 22:41:50 -0400
From: Logical Language Group <lojbab>
Subject: Re: needing 2 taxis
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
UC>If "waiting for a taxi" is "waiting for loi taxi", how do we say
UC>"we're waiting for two taxis". Does "reloi" do this?
No. Since masses are intended not to be used where singular plural
distinction is important, then "waiting for two taxis" is ALSO
"waiting for loi taxi" in a sense. Thuink again about how you ask for
a mass noun in English "I am waiting for some water" is the mass
expression whether you are waiting for 1 glass or 2 glasses.
To make the distinction clear you could wait "loi taxi pamei" vs.
"loi taxi remei", or you could wait for "pa/re selci poi taxi" which is
more or less the same as "pa/re [lo] taxi"
lojbab
Date: Sat, 17 Sep 1994 22:50:02 -0400
From: Logical Language Group <lojbab>
Subject: Re:
Cc: lojbab@access.digex.net
It is ftp.cs.yale.edu - not .sc. (the net is very unforgiving of typos
in addresses, and seldom tells you why)
You should be using anonymous ftp which has login name "anonymous"
and you email address as password.
You might see if your machine has a program called "ncftp" in addition to
"ftp". The former is a little smarter version - it automatically does
anonymous login, and remembers the last 20 ftp sites you have accessed as well
as the directories that you were last in - this is very useful, needless to
say. It also lets you 'page' short files or the beginnings of files so that
ypou can see what you are downloading or read them on-line if short enough.
BTW, just got our 3rd (of 3) Chinese person who is going staright to level 3.
Lojban seems to be attractive to such people. This person is an Esperantist,
so I am told, and is actually in PRC (the other two are in Taiwan, and both
have net access).
lojbab
Date: Sat, 17 Sep 1994 23:38:27 -0400
From: Logical Language Group <lojbab>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
JL>> Now "mi nitcu pa tanxe", which is NOT restricted, does say that ANY member
JL>> of the (unrestricted) set of things that 'are boxes' will satisfy your
JL>> need.
JL>
JL>Using that same logic, you would conclude that "mi ponse pa tanxe" says
JL>that ANY member of the (unrestricted) set of things that 'are boxes' is
JL>owned by you.
No. It says that exactly ONE out of the unrestricted set are owned by
me,
but gives no clue as to which of that set it is (it could be 'any' of them).
JL>Let's say I have three boxes, one red, one blue, and one purple with
JL>little pink flowers, and _any_ of them will serve for whatever purpose
JL>they are needed.
JL>
JL>Now you say {mi nitcu pa tanxe}
JL>
JL>And I ask {xu do nitcu le xunre tanxe}
JL>
JL>What should the answer be?
not "go'i".
Most likely something like "ri banzu"
JL>If {do nitcu le xunre tanxe} is false, and {do nitcu le blanu tanxe} is
JL>also false, then we could go over the list for every existing box and
JL>it would be false for all of them, then {do nitcu pa tanxe} would be
JL>false, because we couldn't find any {pa tanxe} that made it true.
It is not the case that there is a sentence isomorphic to "mi nitcu pa tanxe"
that will answer the question, changing only the quantifier.
We do not in English answer "I need a box" with "Do you need the red box?"
and so forth for every box known to exist. If we did then the answer might
very well be "no" to each such question, because it is not necessarily the
case that the specific box being referred to is'the' box that is needed.
Iwould be unlike to respond to "I need a box" in English OR Lojban with a
question involving predicate "need"/"nitcu". If you insist, then the
question "xu do nitcu pa lu'a le xunre tanxe ce le blanu tanxe ce le zirpu
tanxe". The answer to this might STILL be "no", though, if it is not the case that pragmatically, the first person decides that indeed 1 of those 3 is THE
one s/he wanted originally, but did not restrict in his original statement.
A question involving "banzu" is far more appropraite in response. This is
because the origoinal speaker was being non-specific, and you are in effect
trying to make him a liar by forcing him to decide that there was indeed
a specific one that was needed.
Now, in reality, the first speaker should never say "mi nitcu pa tanxe",
because it is very unlikely that just 'any' box will do. Indeed, I wouyld
go so far as to say that one should not make truth-critical statements
using "lo" any more than with "da", because very rarely in real life do we
specify all relevant restrictions.
Nick and John came upo with the answer to this by coining "voi" - where
"pada voi tanxe" parallels "le" in semantics. But this is another bound
variable and (may) claim existence. On the other jhand, it is then possible
to say "mi nitcu lenu pada voi tanxe cu co'e" (the quantificatiojn of the
"da" cannot be exported outside of the lenu clause). I guess youi can
even use pada poi tanxe, come to think of it. Perhaps another solution to
the original problem (which I am already not certain I remember).
JL>Now, other quantifications for masses confuse me. What do they really mean?
Not much, if you are getting into truth functional statements.
loi cifno lives in Africa, but also on every other continent.
There is no quantifier that could go on that "loi" other than "pisu'o" or one
of the other non-specific fracxtional quantifiers that would be meaningful.
"pimu" only works if EXACTLY half, not 1 more or less than half, of lions
live in Africa. In real life we seldom know quantifiers that exactly when
dealing with masses. (Again, I ask you to think of mass nouns in English,
and Spanish assuming they exist in Spanish. If you use quantifiers with
mass nouns, it is at the very least probably a highly marked usage that
will practically beg people to look for some deeper hidden structure to
your statement (an elided sumti raising or restriction, most likely).
JL> piro loi remna ka'e se jbena
JL> All of the mass of humans is innately capable of giving birth
JL>
JL>The last one is true, because the mass inherits all properties of its
JL>members, but then what does the other one say? And if it's true for
JL>the whole mass, should it be true for 75% of the mass?
No it is not true. The mass inherits all properties of its members, but the
mass as a whole does not EXHIBIT those properties.
You can truthfully say
"Water is frozen" (water exists in a frozen state) or
"Water isn't always frozen" (water exists in a non-frozen state)
(Yeah, I'll admit the former is rather strange - how about "Roses are red"
even though some are other colors.)
You cannot say "All water is frozen" or "All roses are red".
JL>And asks:
JL>
JL>> If "waiting for a taxi" is "waiting for loi taxi", how do we say
JL>> "we're waiting for two taxis". Does "reloi" do this?
JL>
JL>No, and that's an excellent question. (I think reloi gives you two masses
JL>of taxis.)
Consensus has been that any quantifier greater than "pa"/"piro" on "loi"
is nonsensical. I have propsoed some meanings in the past, but John C. has
not agreed with me, if I recall. If he has, he almost certainly has added it
to the appropriate paper (someone oughta check and see what his papers say
about this issue, BTW.)
lojbab
Date: Sat, 17 Sep 1994 23:49:12 -0400
From: Logical Language Group <lojbab>
Subject: Re: any and all
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
I think part of the trouble we are having is that in English we DON'T say
"I need any box." I cannot fathom what someone wouyld mean by this without
context. "I need a box" sometimes, but but not always, could mean "pa tanxe".
I doubt that it would be said if there did not exist a box that would
satisfy (so using da poi or da voi with any particular scope would not
inherently cause probalems). And in MOST cases, the spoeaker has some
unstated restrictions in mind, in which case "pa le tanxe" is also accurate.
And le (probably) does not claim existence, only in-mindedness.
The real probalem is that, as we have generally found in Lojban, very few people
ever really make statements that go over into quantificational logic very
well, because the quantificational versions (claiming existence and exactness
of numbers and restriction, etc) rarely accurately state the contextual
intent. This is why almost no Lojbanist makes statements about "roda", or
"ro" ANYTHING - we KNOW that universal quantification doesn't often work.
And statements that most people make as generalizations are really mass
statements or stereotypes, and these are clearly distingiuished from the
"ro" type statements.
lojbab
Date: Sat, 17 Sep 1994 23:54:52 -0400
From: Logical Language Group <lojbab>
Subject: Re: needing books
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
JL>Or, (but lojbab disagrees):
JL>
JL> mi nitcu lo cukta
I only disagree in that if And attaches significance to 'specific'
(which I failed to notice in my own response to him), then it should
be "le" rather than "lo" because "lo" is inherently non-specific
Hmm. I'll amend that. If we are not being pedantic, then "lo" may
possibly be specific because pragmatics will indicate that there are
unstated restrictions.
Hmm. Do we have an attitudinal that says "pedantic - non-pedantic" %^)
lojbab
Date: Sun, 18 Sep 1994 00:28:59 -0400
From: Logical Language Group <lojbab>
Subject: Hi, pc
Cc: lojbab@access.digex.net
Since he has finally joined Lojban List and contributed (thereby identifying
himself:
Let me welcome John Parks-Clifford (now John Clifford) aka "pc" to Lojban List.
pc is Vice President and co-founder of LLG.
He is former President of The Loglan Institute, Inc., former editor of
'The Loglanist', the journal of that organization, and the chief logician
of the Loglan Project from 1975 on (though understandably Randall Holmes
shares that claim now having 'stepped into the role' left after pc left TLI.
It is arguable that prior to Lojban, pc and James Cooke Brown WERE the Loglan
Project, and certainly were 'Loglan Central', although pc and I basically have
taken the position that the entire community was what made Loglan what it
was/is.
I would not have started the Lojban redesign without pc's technical expertise
in the language, and in logic (the current argument should show just how
lacking I am in the logical arena), and he has contributed to almost every
major technical and design decision in the project. In particular, he
is principally responsible for two of the most major enahncements Lojban
has over TLI Loglan - our detailed tense system and our handling of negation.
We owe pc a LOT of thanks.
(Now don't flood him with tons of new technical issues - he is still
plowing through the 500K of discussion of ZAhO from the past year, as well as
a couiple of other issues, and will probably be involved when we are ready to
take up Jorge/Veijo's proposals regarding center-embedding again. he is also
new to the net, and we can't count on him having the infinite time we wish
he had to deal with some of the open Lojban questions.
On the other hand, expect ME to go out on fewer limbs in the area of logical
expression in Lojban, like I have in the past few days. This is probbaly
for the better, since I don't always know what I am talking about, in spite
of having that mabla label of "Lojban Central", and because I really oughta
be spending more time on the dictionarya nd not trying to analyze unaanalyzable
logic questions %^)
lojbab
From: Randall Holmes <holmes@DIAMOND.IDBSU.EDU>
Subject: Massification and Boxes/Taxis
If you try using massification, you are still faced with the problem
that "I want a box" does not require as part of its truth conditions
that there be any boxes whatsoever; translating it as "I want a piece
of Mr. Box" doesn't work if there are no such pieces!!!
Consider as a test case "I want a unicorn"; this is a statement which
_can_ be true, and is clearly not about any piece of the mass of unicorns.
--Randall Holmes
Date: Sun, 18 Sep 94 09:25:00 -0600
From: Randall Holmes <holmes@diamond.idbsu.edu>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
Cc: LOJBAN@CUVMB.BITNET, pcliffje@crl.com
The set of unicorns does exist; it happens to be empty. Similarly,
"lea preda" always exists, but may be empty.
I don't think that dodges about empty sets will save you. Here is why not:
"I need a box which will hold my house (when there are no such boxes)"
is FALSE if it is interpreted as referring to "some element of (a restricted
subset of) the set of boxes"; if the set is suitably restricted, there is
no element of the restricted subset, so a statement about "some"
element of that set is automatically false. But this statement is
true; thus, in spite of its surface form, it does not contain an
existential quantifier over any set of boxes whatsoever!!!
I think that what is really needed by the speaker is a certain state
of affairs: I need that (there be a box big enough to hold my house
and I have it). Notice that this statement is not about any
particular box, nor does it presuppose the existence of a suitable
box. The form of the sentence in Loglan or Lojban would be such that
it would be clear that it was unreasonable to ask "which box?", which
is not true if the surface form of the English is followed.
--Randall Holmes
From: Randall Holmes <holmes@DIAMOND.IDBSU.EDU>
Subject: Even worse!
Has it occurred to any of you that "I need a box" can be true while
"I need box A" may be false of each specific box A (because some other
box would do just as well?) There is really no quantification over
boxes going on in this sentence!!!
--Randall Holmes
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: needing books
GK> I am going to take a fling at these, I hope it doesn't add to the
confusion which I definitely share. I'm going to number them too.
------------------------
And>
How do you distinguish, (preferably in non-pedantic usage):
(1). I need there to be a specific book such that I have it.
(2). There is a specific book such that I need to have it.
(3). I need there to be x, such that x is a book & I have x.
(4). There exists x, such that x is a book & I need to have x.
Yours in more than usual confusion,
And
----------------------
GK> For a specific book I want one of a kind. I hope that's what you
meant. I've never read it but it's a one word title. I choose the
original manuscript of Beowulf. I call it Beoualf. Another way of
talking about a specific book is to call it "the" book where "the"
is part of what is known as a definite description. Bertrand Russell is
the author of this version of things. I'll do this both ways.
"The crucial feature of Russell's account is that if a description 'the
P' is being correctly used-that is, if there exists a P and only
one-then it achieves the same effect as a proper name, in singling out a
unique thing in the world" ( courtesy of Richard Smith).
(1a). First try to put it in language that is closer to the language of
predicate logic and lojban:
I need the state of: both Beowulf is a book and I possess it.
mi cu nitcu lo za'i ge la zoi gy. beoualf gy. goi ko'e cukta gi me ponse
ko'e
I need the state: both [forethought and] the thing named [quote
nonlojban] beoualf, to which I assign the pronoun it2, is_a_book, and I
own it2.
Without the parser I would never have gotten this straight. It may not
be yet.
(1b). I need there to be a specific book such that I have it.
I need the state: I have "the" book.
This is Russell's "the". It means there is just one such book [of its
kind].
It is equivalent to a name, here the manuscript Beowulf.
mi nitcu lo za'i mi ponse lo pa cukta
I need the state: I possess what really is the one and only book.[of its
kind]
What I am trying to say by [of its kind] is that there is some universe
of discourse. When Richard Nixon used to say "I am 'the' President", as
he was so fond of doing, he didn't mean he was the one and only
president in the world. He meant he was the only current one in the
United States. Likewise lo pa cukta means something like original book
about Beowulf. All this is just my idea of how it ought to work.
There is an explanation of "paboi sumti" where the pa is adjacent to the
sumti in the lessons. I don't think it has been used this way, again its
my idea of how it ought to work.
(2.) There is a specific book such that I need to have it.
I need to possess the specific book.
.i mi nitcu lo nu mi ponse lo pa cukta
I need to possess "the" book. Russell's "the" again. Our "lo" is close
to the logical "description operator", known as TAU, which is an
equivalent form of Russell's "the". lo pa [broda]= "the" [broda].
You will have to look these things up to see where I'm coming from.
This is definitely not consensus lojban, but I don't see how else to do
it.
(3). I need there to be an x, such that x is a book & I have x.
I need the state: x exists, and x is a book and I have x.
.i mi nitcu lo za'i su'o pa da zu'o ge da cukta gi mi ponse da
I need the state: at least one object x exists and both x is_a_book
and I possess it.
(4). There exists x, such that x is a book & I need to have x.
.i su'o da zu'o ije da cukta ije mi nitcu da
There exists some x [end prenex] and that x is_a_book and I need
it.
Maybe "need to have" should be nitcu ponse or some such.
If I have worked these examples correctly, and. you owe me something.
This has got to get easier. I am going on vacation for about two weeks
so I won't be able to respond for a while but I would appreciate
feedback.
djer jlk@netcom.com
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: any
la djer cusku di'e
> There has to be a way to say: "I need any box whatsoever."
> __________________________________________________________________________
> I think the below sentences translate this accurately. The style is
> another question.
>
> 1. roda tanxe da inaja mi nitcu da
roda zo'u tu'e da tanxe inaja mi nitcu da tu'u
> For all X: X is a box implies I need X.
> If X is a box, I need it.
>
> 2.roda poi tanxe ku'o mi nitcu da
roda poi tanxe zo'u mi nitcu da
> For all X which is a box, I need X
Those say "I need every box".
But I need only one, so they can't be equivalent to "I need any box".
GK> I beg to differ. "I need every box" would be:
mi nitcu ro lo tanxe.
This is quite different from the structures above.
I think you're going off course on the " roda"
quantifier. Here are some quotes from THE LANGUAGE OF FIRST
ORDER LOGIC by Barwise & Etchemendy:
"Universal quantifier (upside down A).
This symbol is used to express universal claims, those we
express in English using such terms as 'everything, each thing,
all things, and anything'. It is always used in connection with
one of the variables u,v,w,x,.., and so is said to be a variable
binding operator. The combination Ax is read, "for every object
x," or (somewhat misleadingly) "for all x". {Footnote}:We
encourage students to use the first locution when reading
formulas, at least for a few weeks, since we have seen many
students who have misunderstood the basic function of variables
as a result of reading them the second way." {End footnote}
roda is our notation for the universal quantifier used in
connection with the variable x. It is to be
read then as "for every object x".
roda zo'u tu'e da tanxe inaja mi nitcu da
says "for every object x such that x is a box, it is implied
that I want that object." That object is one box and I want
it. Which one is not specified. There is no implication that a
box search is underway so that there is a recursive collection
formed of all boxes. You wouldn't do this with your statement.
I think that my statement can fairly be said to express your
statement: " I want any box whatsoever."
> I would say too that the word "any" does give rise to
> a number of problems of ambiguity in English and carrying it over to
> lojban might be a problem there.
I agree we shouldn't just have a word that means "any". What we need is
something to translate some of the things that can be said in English
using the word "any", and that seemingly can't be said in Lojban.
(There might be a way to say it using the existing words. If so, I'd
like to know what it is.)
Jorge
GK> I am not categorically against "any" in the language.
I would need to see some indispensible uses of "any". Remember,
I am new at this too, and conclusions we reach could be at
serious odds with the logic community. Queremos una lengua pura,
no una lengua cualquiera. We want a pure language, not any
language whatsoever. Now how do you say that in lojban?
djer jlk@netcom.com
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
> Date: Sun, 18 Sep 1994 09:25:00 -0600
> From: Randall Holmes <holmes@DIAMOND.IDBSU.EDU>
> Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
[...]
>"I need a box which will hold my house (when there are no such boxes)"
[...]
> I think that what is really needed by the speaker is a certain state
> of affairs: I need that (there be a box big enough to hold my house
> and I have it). Notice that this statement is not about any
> particular box, nor does it presuppose the existence of a suitable
> box. The form of the sentence in Loglan or Lojban would be such that
> it would be clear that it was unreasonable to ask "which box?", which
> is not true if the surface form of the English is followed.
>
> --Randall Holmes
mi nitcu le za'i mi ponse lo tanxe poi ka'e vasru le mi zdani
I need the state that I have a box which is capable of containing
my house
which might be simplified to
mi nitcu le za'i me lo tanxe poi ka'e vasru le mi zdani
Explicit quantification can be thrown in
mi nitcu le za'i me re tanxe
I need two boxes
{za'i} (state) could be replaced with {nu} (event)
mi nitcu le nu me re tanxe
In the case of 'I need a box' this might be simplified to
mi nitcu le nu tanxe
co'o mo'e veion
PS. just as I was writing this I received Gerald's posting with
similar structures.
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
I think we are arguing in circles. I will state the problem and its
proposed solutions as I understand it, maybe to confuse things more.
PART I: STATEMENT OF THE PROBLEM or "I need a box"
====================================================
Do we all agree that "I need a box" can have (at least) two meanings
in English, which in Quine's nomenclature are:
transparent: There exists a real, flesh and blood box (as it were), that
can be seen and touched, and is needed by me. (This is not
necessarily a specific box in the sense that the speaker is
not identifying it in any other way than saying that it is
a box, so it doesn't necessarily have to be {le tanxe}.)
opaque: I need that there be a box such that I can have it/use it/whatever.
The "normal" meaning in English is the opaque one for that phrase. When we
want to emphasize the transparent case we would say "I need certain box" or
something of the sort. (Which can still be the opaque case, if by "certain
box" all I mean is "certain kind of box".)
I hold that in Lojban {mi nitcu lo tanxe} has the tranparent meaning. This
is because Lojban predicates state relationships between objects/events.
In this case, the relationship {nitcu} is between that referred to as {mi}
and that referred to as {lo tanxe}: the speaker and a box (at least one).
{lo tanxe} does not specify which box it is, that would be {le tanxe}, but
it still refers to one of all the thing that are boxes, and the relationship
{nitcu} is claimed to hold between {mi} and that {lo tanxe}. This is Quine's
transparent meaning as I understand it.
Lojbab, on the other hand, said that it has the opaque meaning, but when
I translate And's:
> > There is a specific book such that I need to have it.
> da poi cukta zo'u mi nitcu le nu mi ponse da
> Or, (but lojbab disagrees):
> mi nitcu lo cukta
lojbab responds:
> I only disagree in that if And attaches significance to 'specific'
> (which I failed to notice in my own response to him), then it should
> be "le" rather than "lo" because "lo" is inherently non-specific
I don't think And was attaching the le/lo significance to it, so you are
agreeing with me, but you were disagreeing with the same claim in other
posts.
Now, if {mi nitcu lo tanxe} has the transparent meaning, how do we get
the opaque meaning that is the one we usually want?
(If it has the opaque meaning, then the problem is how do we convey the
transparent meaning, but since I don't believe that to be the case, I only
deal with the problem of trying to express the opaque meaning.)
PART II: THE SOLUTIONS(?) or Paraphrases, Masses and Anys.
============================================================
1 - Being very clear on what we mean.
One way is to paraphrase:
mi nitcu le nu mi ponse lo tanxe
I need to have a box.
or
mi nitcu tu'a lo tanxe
I need some abstraction to do with a box.
The problem I have with those is that in the first case I'm saying
more than I want and in the second I'm saying less. This may be
unavoidable, I don't know.
In any case, I don't think anyone disagrees that this is a possibility.
The question is whether we are satisfied with the long-winded approach,
or do we want some elegant way of saying what we mean that is short and
to the point.
2 - Massification.
What does {mi nitcu loi tanxe} mean?
First we need to know what masses mean:
> JL>Now, other quantifications for masses confuse me. What do they really mean?
>
> Not much, if you are getting into truth functional statements.
> loi cifno lives in Africa, but also on every other continent.
Then I take it {loi cinfo} is NOT "ANY part of the mass of lions", since
it is false that any such part lives in Africa. It is rather "SOME (certain)
part of the mass of lions".
(That is what I understood originally, but lately I was confused.)
> "pimu" only works if EXACTLY half, not 1 more or less than half, of lions
> live in Africa.
Ok, but it would mean "certain half", not "any half".
> In real life we seldom know quantifiers that exactly when
> dealing with masses.
The "exactness" is totally irrelevant to my question. {ji'ipimu} would
have done just as well.
> JL> piro loi remna ka'e se jbena
> JL> All of the mass of humans is innately capable of giving birth
> JL>
> JL>The last one is true, because the mass inherits all properties of its
> JL>members, but then what does the other one say? And if it's true for
> JL>the whole mass, should it be true for 75% of the mass?
>
> No it is not true. The mass inherits all properties of its members, but the
> mass as a whole does not EXHIBIT those properties.
What does it mean that it inherits their properties, then?
What properties does the WHOLE mass exhibit? Only those exhibited by ALL of
its members?
In any case, if it is true that
ji'ipimu loi remna ka'e se jbena
Approx. half of humans can give birth.
then clearly {ji'ipimu loi remna} means "a certain half", and not "any half
whatsoever".
So {mi nitcu loi tanxe} = "Certain part of the mass of boxes is needed by me",
is not the solution to "I need a box".
I had changed my mind because in {mi nitcu piro loi tanxe}, the property of
beeing needed by me could be an emergent property of the mass of boxes. But for
this to make any sense, it should be true that {piro loi tanxe} exhibit the
properties of every single box. Since it doesn't, I go back to my previous
opinion that {mi nitcu loi tanxe} doesn't work for this.
[Marginal note:
> Consensus has been that any quantifier greater than "pa"/"piro" on "loi"
> is nonsensical.
I agree that {re loi tanxe} is nonsensical. I could argue for {le re loi tanxe}
though... {piro loi blanu tanxe} and {piro loi xenru tanxe} for instance.
end marginal note]
3 - Please say something, anything...
"I need any box" is not completely equivalent to the opaque "I need a box".
The former implies the latter, but not viceversa. "I need any big box" also
implies "I need a box", and so on. In fact, "I need a box" could be very
restricted, and still have the opaque meaning: If I need a blue box of
size 10cm x 10cm x 10cm made of gold, I could still say "I need a box
(opaque sense)", but not "I need any box".
Using the proposed {xe'e}, "I need any box" would be {mi nitcu xe'e tanxe},
while "I need a box", if it doesn't mean any whatsoever, would have to
be {mi nitcu xe'e le tanxe} or something else.
PART III: CONCLUSIONS or We Are Still Where We Started
========================================================
* The only thing that really works, as far as I can tell, is to use
abstractions. Hopefully a less wordy solution can be found.
* Massification is not related to this issue, in my opinion.
* I still think something is needed to translate the "any whatsoever"
of English, but that won't completely solve the "I need a box" problem.
Anyway...
co'o mi'e xorxes
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: needing 2 taxis
la lojbab di'e la and spusku
> UC>If "waiting for a taxi" is "waiting for loi taxi", how do we say
> UC>"we're waiting for two taxis". Does "reloi" do this?
>
> No. [...]
>
> To make the distinction clear you could wait "loi taxi pamei" vs.
> "loi taxi remei",
A mass of taxi singlets vs. a mass of taxi pairs ?!
> or you could wait for "pa/re selci poi taxi" which is
> more or less the same as "pa/re [lo] taxi"
If you don't see any problem with "pa/re selci poi karce", I don't know
why you wouldn't just use "pa/re karce". Since the problem that the latter
has is shared by the former, why use such a convoluted expression at all?
The question whether it is any selci or certain selci is exactly the same
between any or certain taxi.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
(I would respond to all of lojbab's points, but I would be repeating
myself yet again, so I only reply to a few :)
> Now, in reality, the first speaker should never say "mi nitcu pa tanxe",
> because it is very unlikely that just 'any' box will do. Indeed, I wouyld
> go so far as to say that one should not make truth-critical statements
> using "lo" any more than with "da", because very rarely in real life do we
> specify all relevant restrictions.
So I should never say {lo remna cu mamta mi} because it is very unlikely,
(indeed outright false) that just 'any' remna will do. Is that really what
{lo} means?
Jorge
From: Chris Bogart <cbogart@CSN.ORG>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
>"lo tanxe" is NOT the same thing as "da poi tanxe"; likewise
Ahhh. Now I see why I was confused.
>Not having read Quine (and not being convoinced that I would understand
>it if I did - if I had to be a logician to lead this project, I would
>resign yesterday %^)
I only read the excerpt from Quine that Gerald posted recently; and it's
pretty opaque. But the "sloop" sentence seemed like a clearer example than
the boxes...
Your translations of the "sloop" sentences cleared things up for me; I no
longer think we need an "any" word.
Thanks!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any and all
lojbab:
> I think part of the trouble we are having is that in English we DON'T say
> "I need any box." I cannot fathom what someone wouyld mean by this without
> context.
The meaning we're trying to say in Lojban is:
"Of all the things that are boxes, I have the need to use or own or have at
my disposal one of them. I don't at all care which one it is, just any of
them all."
We could translate the whole phrase, but the question is, does {mi nitcu
pa tanxe} mean that? You say it does (at least some of the time you say that)
but I'm trying to convince you that it doesn't.
What I think {mi nitcu pa tanxe} means is:
"Of all the things that are boxes, there is one in particular that I need.
I'm not being specific about which one it is, but there is only one that
will do, even though I'm telling you nothing about which one it is. Only
one is in the {nitcu} relationship with {mi}"
So, it also makes sense to say {pa remna cu mamta mi} = "Of all the humans
that there are, one of them, and only one of them, is my mother. I'm not
telling you which one, just that it is one of all humans. Not just anyone."
> The real probalem is that, as we have generally found in Lojban, very
> few people
> ever really make statements that go over into quantificational logic very
> well, because the quantificational versions (claiming existence and exactness
> of numbers and restriction, etc) rarely accurately state the contextual
> intent.
Are you saying that since we'll get it wrong anyway, we should forget about
logic and use the language like we would any other?
I know that we will do that, in any case, but it's fun to make the effort
to try to use it logically.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: RE: do djica loi ckafi je'i tcati
> I think that what is really needed by the speaker is a certain state
> of affairs: I need that (there be a box big enough to hold my house
> and I have it). Notice that this statement is not about any
> particular box, nor does it presuppose the existence of a suitable
> box. The form of the sentence in Loglan or Lojban would be such that
> it would be clear that it was unreasonable to ask "which box?", which
> is not true if the surface form of the English is followed.
>
> --Randall Holmes
Totally agree. I would also like to say it in Lojban as simply as I can
in English.
Of course I can make the long claim: {mi nitcu le nu mi ponse lo tanxe}
Is there some way to make it shorter? e.g. {mi nitcu xe'e tanxe}
(BTW, it is hard to ask "which box?" in Lojban.)
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Even worse!
> Has it occurred to any of you that "I need a box" can be true while
> "I need box A" may be false of each specific box A (because some other
> box would do just as well?) There is really no quantification over
> boxes going on in this sentence!!!
>
> --Randall Holmes
Exactly!
That's precisely the difference I propose between {mi nitcu xe'e tanxe}
and {mi nitcu lo tanxe}.
For the latter to be true, "I need box A" has to be true for at least one A.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any
la djer cusku di'e
> > > roda poi tanxe zo'u mi nitcu da
> > > For all X which is a box, I need X
> > Those say "I need every box".
> I beg to differ. "I need every box" would be:
> mi nitcu ro lo tanxe.
That's the same as
mi nitcu roda poi tanxe.
(Since in this case non-existence is not a problem, I hope lojbab agrees.)
> GK> I am not categorically against "any" in the language.
> I would need to see some indispensible uses of "any".
Indispensibility is hard to determine... :)
> Remember,
> I am new at this too, and conclusions we reach could be at
> serious odds with the logic community. Queremos una lengua pura,
> no una lengua cualquiera. We want a pure language, not any
> language whatsoever. Now how do you say that in lojban?
Let's see:
ma'a djica lo curve bangu enai xe'e bangu
Anybody wants to try without {xe'e}?
mi'e xorxes
From: Randall Holmes <holmes@DIAMOND.IDBSU.EDU>
Subject: Chief logician?
Someone seemed to think that I might claim some such title with
respect to the Loglan project because of my relation to TLI. I
disclaim any such role. My job over here is mostly critical; since
the language is "ready for use", I can't do major engineering
modifications on it
From: Desmond Fearnley-Sander <dfs@HILBERT.MATHS.UTAS.EDU.AU>
Subject: Re: any
Hello. I am new to lojban and this list. I had intended to be a passive
observer, but feel constrained to contribute to the 'any' debate. My
interest in lojban springs from the fact that at the level of grammar its
aspirations are very much in sympathy with those of a programming language
we are implementing. The language is called dr.
A feature of dr is the fundamental role in it of what I call
*indeterminates*. For example, if a and b are indeterminates of the sort
number, then the *unquantified* sentence
a^2 - b^2 = (a-b)(a+b)
is true. a and b are *potential entities* of the sort number. This may be
the only information we have about them, or we may have total information
about them (such as that a=5 and b=3) or we may have partial information
about them (such as that a is positive). In each case our sentence remains
true: it is true by virtue solely of the fact that a and b are numbers. On
the other hand, in the absence of specific information about a and b, the
sentence
a^2 - b^2 = (a-b)^2
(though a perfectly acceptable sentence) is neither true nor false. It
becomes true in the presence of the information that b=0, and it becomes
false in the presence of the information that a=5 and b=3.
I believe that indeterminates in this sense play a fundamental role in
everyday reasoning as well as in mathematical reasoning. Ordinary language
accomodates indeterminates nicely. The use of 'a box' in the sentence "I
need a box." is an example. It is a way of referring to something whose
type is known, but about which we have no other information. Additional
information that may be given serves to pin down what is meant:
"I need a box."
"You mean a cardboard box?"
"Yes."
"Here's one from the attic."
"Great."
"What are you going to do with the box?"
>
The dialogue starts with a total indeterminate (a potential entity of the
sort box) and concludes with an entity that instantiates it.
I do not think that classical logic accomodates or is even compatible with
this notion --- I am going out on a limb here, and might be persuaded
otherwise. Indeterminates are not constants, and they are not variables,
they require a *typed* language and they do away with the need for
universal quantification.
It would be disappointing to me if lojban did not admit indeterminates in a
simple way, but that's what the debate seems to suggest. Am I wrong about
this?
I did not catch the beginning of the 'any' debate, so bear with me please
if I'm covering old ground.
Desmond FS
----------------------------------------------------------------------------
Desmond Fearnley-Sander
Department of Mathematics, University of Tasmania
GPO Box 252C, Hobart, Tasmania 7001, AUSTRALIA
EMAIL: dfs@hilbert.maths.utas.edu.au
PHONE: (002) 202445 (from in Australia)
+61 02 202445 (from outside Australia)
FAX: (002) 202867 (from in Australia)
+61 02 202867 (from outside Australia)
----------------------------------------------------------------------------
From: Desmond Fearnley-Sander <dfs@HILBERT.MATHS.UTAS.EDU.AU>
Subject: Re: any --- second thought
Please delete the phrase 'or is even compatible with' from my note. I
don't want to spark an irrelevant debate.
Desmond
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: needing books
Please substitute zo'u, the logical end prenex word; for zu'o the
activity abstractor, in my translations of .and's sentences. Then add a
tu'e right after. Sometimes perfection seems farther away than ever.
jlk@netcom.com
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: any answer
>From my previous post:
roda is our notation for the universal quantifier used in
connection with the variable x. It is to be
read then as "for every object x".
roda zo'u tu'e da tanxe inaja mi nitcu da
says "for every object x such that x is a box, it is implied
that I want that object." That object is one box and I want
it. Which one is not specified. There is no implication that a
box search is underway so that there is a recursive collection
formed of all boxes. You wouldn't do this with your statement.
I think that my statement can fairly be said to express your
statement: " I want any box whatsoever."
Jorge;
I do need to know whether or not you agree with the above. Have I
expressed the idea of "I want any box whatsoever" to your satisfaction
with my statement "roda zo'u..etc." or not? I'm beginning to feel like
we are caught in a Wharfian warp.
I'm off for vacation now.
djer jlk@netcom.com
Date: Mon, 19 Sep 1994 06:00:55 -0400
From: Logical Language Group <lojbab>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
JL>So I should never say {lo remna cu mamta mi} because it is very unlikely,
JL>(indeed outright false) that just 'any' remna will do. Is that really what
JL>{lo} means?
No. I was just discussing this with Nora. Since the default quantification
of "lo" is "su'o" outside, then statements about "lo remna" are true if at
least one of the members (non-specific) will make the sentence true. HOWEVER,
you can't pick which one (other than by restrictions), so "lo tanxe ka'e
vasru le zdani dinju is true if there is some box somewhere that is capable
of doing so. I suspect that the Vehicle Assembly Building at Kennedy Space
Center qualifies as such a box. Hence it is a true statement. Probably not
a useful statement.
The problem comes with statements about "lo unicorn", which doesn't exist.
1) They actually do exist - there have been photos published of 1-horned
goats ..., but these don't fit the definition of "ba'e le unikorn" that
most people refer to
2) In MOST statements about unicorns, the universe of discourse is not the
'real world', but aworld where unicorns DO exist. In a fantasy world, I need
a unicorn is a perfectly acxcetpable statement.
3) The logical content of "lo unicorn" is clearly tied to "if unicorns exist
then at least one of the set". But it is less clear what it means if there are
no unicorns. My inclination would be to say that such a statement immediately
throws the universe of discourse into a subjunctive 'fantasy' one, where there
IS at least one unicorn - the description is veridical in referring to a set of
properties that such a unicorn MUST have (unlike le unikorn, which needn't have
anything to do with unicorns), though it is somewhat intensional as the set of
properties being ascribed is a set that describes no such creature. Thus
you are saying: I need a se ckaji rolo ka unikorn, but do not claim that such
a se ckaji really exists.
lojbab
Date: Mon, 19 Sep 1994 06:22:32 -0400
From: Logical Language Group <lojbab>
Subject: Re: any
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
A quick skim of your article suggests that your new category of 'objects' may
match our intended usage for "lo", but I will have to look at it further, and
maybe ask more questions. Lojban "lo" permits quantification, but the extent
to which it is analyzable by traditional predicate logic is not complete.
lojbab
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any answer
la djer cusku di'e
> >From my previous post:
>
> roda is our notation for the universal quantifier used in
> connection with the variable x. It is to be
> read then as "for every object x".
>
> roda zo'u tu'e da tanxe inaja mi nitcu da
>
> says "for every object x such that x is a box, it is implied
> that I want that object." That object is one box and I want
> it. Which one is not specified. There is no implication that a
> box search is underway so that there is a recursive collection
> formed of all boxes. You wouldn't do this with your statement.
> I think that my statement can fairly be said to express your
> statement: " I want any box whatsoever."
>
> Jorge;
> I do need to know whether or not you agree with the above. Have I
> expressed the idea of "I want any box whatsoever" to your satisfaction
> with my statement "roda zo'u..etc." or not?
No. To my understanding, what you have expressed is "I need every box
there is".
You are saying that the relationship {nitcu} holds between {mi} and every
{da} which is a {tanxe}.
> I'm beginning to feel like we are caught in a Wharfian warp.
My feeling precisely.
> I'm off for vacation now.
Have a good time!
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any
lojbab:
> A quick skim of your article suggests that your new category of 'objects' may
> match our intended usage for "lo", but I will have to look at it further, and
> maybe ask more questions.
I think it doesn't. It certainly doesn't match the _actual_ usage.
Desmond:
> A feature of dr is the fundamental role in it of what I call
> *indeterminates*. For example, if a and b are indeterminates of the sort
> number, then the *unquantified* sentence
> a^2 - b^2 = (a-b)(a+b)
> is true. a and b are *potential entities* of the sort number. This may be
> the only information we have about them, or we may have total information
> about them (such as that a=5 and b=3) or we may have partial information
> about them (such as that a is positive). In each case our sentence remains
> true: it is true by virtue solely of the fact that a and b are numbers.
This is not the case for Lojban {lo}.
For example:
lo remna cu mamta mi
A human being is mother to me
is true. Not by virtue of the fact that {lo remna} is a human being, but because
of the fact that there is one human being that is in relationship {mamta}
with {mi}.
> On
> the other hand, in the absence of specific information about a and b, the
> sentence
> a^2 - b^2 = (a-b)^2
> (though a perfectly acceptable sentence) is neither true nor false.
Sentences with {lo} in Lojban are usually true or false.
For example:
lo remna cu kalte lo remna
A human hunts a human
is true only if there really is at least one human that hunts at least one
human. It's not a matter of giving values to each {lo remna}.
If "a" and "b" were replaced by {lo namcu} = "a number" in your sentence,
it would be a true sentence in Lojban, because there indeed exists at least
one "a" and at least one "b" that make it true.
> It
> becomes true in the presence of the information that b=0, and it becomes
> false in the presence of the information that a=5 and b=3.
That sounds like it might be more or less equivalent (at least for some
purposes) to Lojban {le}
le remna cu mamta mi
The human is mother to me.
is true if by {le remna} I mean the human who is my mother. In that sense,
you can say that it's neither true nor false in the absence of information
of what {le remna} is referring to, but that information is at least in
principle always obtainable (by asking the speaker who they meant by it).
>From what I understand, your "a" need not have a value obtainable even in
principle.
> I believe that indeterminates in this sense play a fundamental role in
> everyday reasoning as well as in mathematical reasoning. Ordinary language
> accomodates indeterminates nicely. The use of 'a box' in the sentence "I
> need a box." is an example. It is a way of referring to something whose
> type is known, but about which we have no other information.
I think something like that is what I meant by my proposal of {xe'e},
although I don't have it that clear in my mind.
> Additional
> information that may be given serves to pin down what is meant:
>
> "I need a box."
> "You mean a cardboard box?"
> "Yes."
> "Here's one from the attic."
> "Great."
> "What are you going to do with the box?"
> >
> The dialogue starts with a total indeterminate (a potential entity of the
> sort box) and concludes with an entity that instantiates it.
The first two mentions of "box" are indeterminate (one of the sort "box",
the other of the sort "cardboard box"). The last one is an actual box.
In my opinion, as things stand now, we can only refer to the first type
in Lojban within abstractions.
> I do not think that classical logic accomodates or is even compatible with
> this notion --- I am going out on a limb here, and might be persuaded
> otherwise. Indeterminates are not constants, and they are not variables,
> they require a *typed* language and they do away with the need for
> universal quantification.
I wouldn't know if they do away with it, but it would be nice to have them.
> It would be disappointing to me if lojban did not admit indeterminates in a
> simple way, but that's what the debate seems to suggest. Am I wrong about
> this?
I think you're right. But maybe it's just me :)
> I did not catch the beginning of the 'any' debate, so bear with me please
> if I'm covering old ground.
I don't think it is, but even if it were, we need to cover the same ground
many times before we, if ever, become familiar with it.
Jorge
From: Chris Bogart <cbogart@quetzal.com>
Subject: TECH: "any" & quantification
I've been following the "any" discussion avidly, but everytime I try to
participate I notice myself flip-flopping on the issue. It's very
confusing, and the only thing I've convinced myself of is that there's
something lojbab is saying that Jorge is missing and vice-versa.
Suppose (only for the sake of discussion!) we had a manditory particle
before every sumti in lojban, a choice of either "xe'e" meaning
referentially opaque, or "xa'a" meaning referentially clear. That makes the
"box" example easy to analyze: "xa'a mi nitcu xe'e lo tanxe" means "I need
a box" and "xa'a mi nitcu xa'a lo tanxe" means "There is a box I need".
Maybe it would help to make that assumption for a while just for the purpose
of exploring what the implications would be on other sentences.
xa'a pa remna cu mamta xa'a mi - One person is my mother
xe'e pa remna cu mamta xa'a mi - Only one person can be my mother (??)
xa'a mi nelci xa'a do - Some of us like some of youse
xe'e mi nelci xe'e do - ???
Am I correct in thinking that the current disagreement between Jorge and
Lojbab is whether unmarked sumti in real lojban are equivalent to sumti
marked with "xe'e" or "xa'a"? If so, let's do more translations with these
two markings and see which one comes out more like the way we think lojban
is currently defined.
Or could it be that the marking is only possible in certain place
structures, and it is meaningless to contemplate "xe'e mi nelci xe'e do"?
By the way, is "xa'a" as I've defined it the same as asserting existence?
(if "xa'a" isn't an unassigned cmavo and actually means something already,
the my apologies and could someone clued in on the "experimental list" pick
a better one?)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: general response on needing books
Thanks to Jorge, Lojbab, Djer & Randall for their replies
to my query, which I repeat here:
> How do you distinguish, (preferably in non-pedantic usage):
> (1a). I need there to be a specific book such that I have it.
> (1b). There is a specific book such that I need to have it.
> (2a). I need there to be x, such that x is a book & I have x.
> (2b). There exists x, such that x is a book & I need to have x.
None of the replies wholly answered my intended query (because
my intention was not in the least apparent from my wording).
Let me explain, with an altered example: "I need you to have
a book", instead of "I need to have a book". My idea was that
the +/- specific distinction (LE v. LO) cuts across the transparent/
opaque distinction. (1a-b) are +specific, (2a-b) are -specific, (1a,2a)
are opaque, (1b,2b) are transparent.
One of (1a-b) (or maybe, ambiguously, both) is, in Lojban:
(1a/b) Mi nitcu le/lo nu do ponse le cukta
One of (2a-b) (or maybe, ambiguously, both) is, in Lojban:
(2a/b) Mi nitcu le/lo nu do ponse lo cukta
I should clarify what I mean by (1a) (specific & opaque): I don't
necessarily have the referent of 'le cukta' in mind, but were it to
come to evaluating the truth of the lenu clause I would have a
specific referent of 'le cukta' in mind. "I need there to be a specific
book such that you have it" - in order to decide whether my need has
been satisfied, you have first to find out which specific book this is;
but I think it should be possible for me to say "X needs there to be
a specific book such that Y has it" without me having decided which
specific book this is.
Further remarks.
(i) My impression so far is that (1a/b) and (2a/b) really are ambiguous
in Lojban. (This is currently being disputed on the List.)
(ii) (2a/b) can be disambiguated by using "da (zohu)" either inside or
outside the lenu clause. BUT this is pedantic & not normal usage.
(iii) (1a/b) cannot be disambiguated.
A new suggested solution: LE & LO are transparent (wide scope, with
quantification/reference assignment outside their clause). But when
combined with a certain cmavo (e.g. Jorge's xehe) they are opaque
(narrow scope, with quantification/reference assignment outside their
clause). Note that under my proposal "xehe" is used in tandem with
LE/LO, not instead of them.
Incidentally, contra Jorge, I think "mi nitcu lo tanxe" is illicit sumti
raising: what you really need is some state of affairs to be the case,
such that its not being the case would be disadvantageous. A "less
wordy" locution would require us to not say what we mean.
I agree with Jorge (& others) that massification is not relevant.
My proposals don't solve the "any whatsoever" problem. Some but not
all of these problems are solved by using wide scope universal quantification:
You may read all of the books.
It is permitted that for all x such that x is one of the books
you read x.
You may read any of the books.
For all x such that x is one of the books it is permitted
that you read x.
But this, I feel, doesn't handle "You may read any three of the books"
(i.e. a total of three books, selected freely from the set of all of
the books). So maybe here a new quantifier is needed, as Jorge has been
arguing.
-----
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Chief logician?
Randall asks:
> Does anyone in the
> Lojban community realize that logical connectives applied to arguments
> produce problems of scope (usually handled implicitly in NL's)
> precisely analogous to those connected with quantification?
Logical connectives are never applied to arguments in Lojban. (In the sense
that all forms of logical connection are contractions of full bridi
connections.)
And yes, the debate on "any" was sparkled by a logical connection question:
mi nitcu le tanxe a le dakli
I need (the box OR the bag)
means:
(I need the box) OR (I need the bag)
It does not mean what we usually mean in English by "I need either of the
box or the bag".
> Consider
>
> John and James love Mary or Sally
>
> versus
>
> Mary or Sally is loved by John and James
>
> In the second sentence, but not in the first, it is clear that John
> and James love the same unspecified element of {Mary, Sally}; in the
> first sentence, they may love different elements of the set.
If you have two logical connectives in a Lojban sentence, I think the
first one binds tighter, so
la djan e la djeimyz prami la meris a la salis
expands to:
la djan e la djeimyz prami la meris
ija la djan e la djeimyz prami la salis
(John loves Mary AND James loves Mary)
OR (John loves Sally AND James loves Sally)
while:
la meris a la salis se prami la djan e la djeimyz
goes to:
la meris a la salis se prami la djan
ije la meris a la salis se prami la djeimyz
(Mary is loved by John OR Sally is loved by John)
AND (Mary is loved by James OR Sally is loved by James)
This is in reverse of the meaning you give for the English sentences,
but there is no ambiguity.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
> JL>So I should never say {lo remna cu mamta mi} because it is very unlikely,
> JL>(indeed outright false) that just 'any' remna will do. Is that really what
> JL>{lo} means?
>
> No. I was just discussing this with Nora. Since the default quantification
> of "lo" is "su'o" outside, then statements about "lo remna" are true if at
> least one of the members (non-specific) will make the sentence true.
Ok, we agree here. At least one member has to make the sentence true.
Then you contradict what you just said:
> HOWEVER,
> you can't pick which one (other than by restrictions),
If I can't pick which one, then {lo remna cu mamta mi} is false.
> so "lo tanxe ka'e
> vasru le zdani dinju is true if there is some box somewhere that is capable
> of doing so.
I agree.
> I suspect that the Vehicle Assembly Building at Kennedy Space
> Center qualifies as such a box. Hence it is a true statement.
Agreed. The point about "you can't pick which one" is inconsistent with
the rest of what you say.
> The problem comes with statements about "lo unicorn", which doesn't exist.
But this is a different problem. The "any" problem applies to boxes as much
as to unicorns.
> 2) In MOST statements about unicorns, the universe of discourse is not the
> 'real world', but aworld where unicorns DO exist. In a fantasy world, I need
> a unicorn is a perfectly acxcetpable statement.
Yes, but {mi nitcu lo pavyseljirna} doesn't mean the same that "I need a
unicorn"
usually does in English.
{mi nitcu lo pavyseljirna} means something like:
da poi pavyseljirna zo'u mi nitcu le nu mi ponse da
while "I need a unicorn" normally means:
mi nitcu le nu da poi pavyseljirna zo'u mi ponse da
Do you agree that those two sentences say something different from the other?
Jorge
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: TECH: "any" & quantification
In-Reply-To: (Your message of Mon, 19 Sep 94 13:56:28 CST.)
Chris:
> xa'a mi nelci xa'a do - Some of us like some of youse
> xe'e mi nelci xe'e do - ???
I reckon the 2nd example shd mean: Let x be any n [default: 1],
but no more than n, of us, and let y be any n [default: 1]
of you, & it is asserted that x likes y.
> Or could it be that the marking is only possible in certain place
> structures, and it is meaningless to contemplate "xe'e mi nelci xe'e do"?
I don't think it's meaningless, but 'irrealis' contexts (descriptions
of things not, or not necessarily, the case) are more likely to
call for "xehe".
---
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any
> >{mi nitcu lo pavyseljirna} means something like:
> >
> > da poi pavyseljirna zo'u mi nitcu le nu mi ponse da
> >
> >while "I need a unicorn" normally means:
> >
> > mi nitcu le nu da poi pavyseljirna zo'u mi ponse da
> >
> >Do you agree that those two sentences say something different from the other?
>
> Yes.
Alleluyah!
> But I think that "mi nitcu lo pavyseljirna" is more likely to mean
> the second expansion than the first.
By "more likely" which do you mean:
(1) In most contexts where that Lojban sentence appears, it means something
like the second expansion, but in some contexts it may mean something
like the first.
(2) It always means the same one of them, but you are not sure which, and
you suspect it is the second, that's why you say "more likely".
If (1) then you are throwing logic out the window, and saying that just
like in English "I need a box" can have two meanings (Quine's transparent/
opaque), the same happens in Lojban, and which meaning it is is determined
by context.
If (2) then the way we've been interpreting most other predicates is wrong.
(Unless you say that {nitcu} should be interpreted in a different way than
other predicates.)
> The first claims that unicorns
> exist, which "lo" does not.
But we're in that fantasy world where they do exist. Change it to boxes,
the problem remains.
> I'm not sure about truth-fucntional value
> of the second expansion - what the value is of "le nu [false statement]"
> is not clear.
Because there is no truth value for it. It is a sumti, and sumti don't have
truth values. I would even question that in "le nu [statement]", there is
any truth value for that [statement], since it is not being claimed.
> However your clarification may point to the fact that se nitcu should be an
> abstraction regardless about how we resolve the "lo"/"any" question.
This is what was done to {djica}, which is analogous to {nitcu}. But I don't
see why the transparent meaning should be forbidden. {mi nitcu le vi tanxe}
= "I need this box", makes perfect sense. And so does {mi djica le vi tanxe}.
> If so,
> then PLEASE separate the place structure issue and let's get it resolved -
> you have only a few days before I finsih the e-order gismu list. (I presume
> that the issue affects "nitcu" and "djica". Any others?
{cpedu} comes to mind, but there probably are others. I disagree that they
should be made to accept abstractions only.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Chief logician?
Randall says:
> There is a good reason for the precedence to go the other way,
> unless you also reverse the usual convention for implicit quantification:
>
> the point is that P[John and James] means roughly the same thing as
> "For all x in {John,James}, P[x]", and, similarly, P[Mary and Sally]
^^^
or, I think
> means roughly the same thing as "For some x in {Mary,Sally}, P[x]";
> where two of these connected arguments appear in a sentence, one has
> essentially the same problem one has with the usual form of implicit
> quantification as in
>
> Someone loves everyone
>
> versus
>
> Everyone is loved by someone
Yes, I see your point. In fact, the paper on connectives doesn't really
mention that case. It says that
broda ije brode ija brodi
is grouped from left to right:
(broda ije brode) ija brodi
and from there I generalized to assume that the first connective is
bound tighter.
Maybe it should be the other way around when the connectives are in
different terms.
There is still going to be counterintuitive cases, though:
da prami la djan e la djeimyz
Someone loves John and James
means the same as:
la djan e la djeimyz se prami da
John and James are loved by someone
in both cases, the quantification is: For some da; for all x in {John; James}.
> Of course, I know that the underlying "expanded" form of the sentence
> does not involve application of logical connectives to arguments;
> I'm a logician, remember?
Of course :)
But you did mention something about negation of arguments being allowed
in Loglan...
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Chief logician?
[I still haven't received the post by lojbab to which Randall
is responding here]
> On your comments about negation, I have problems with terminology, but
> as far as I understand you I probably disagree flatly. There is one
> concept of negation (the propositional connective) and then there are
> various other notions which NL's confuse with negation;
I agree.
> I would hate
> to think that you are importing NL confusions (more likely you are
> defining these other notions precisely and using them correctly and
> the only confusion is that you call them "negation" :-) ).
In fact, I think that's exactly what's going on.
For example, does Loglan have {na'e} = non-/other than ?
We can easily distinguish
ta na blanu tanxe
It is false that: that is a blue box.
ta na'e blanu tanxe
That is a non-blue box
and things like that.
I don't like calling {na'e} negation, and even less calling {to'e} negation
(to'e=opposite). But they are very useful.
> Explain by
> example what you mean by "metalinguistic" negation.
It's the answer to "Have you stopped beating your wife?"
Strictly logically, I think that {na} suffices, but that leaves the
wrong impression in some people.
> Of course I understand how logical connectives applied to arguments
> are eliminated! The difficulty arises in expanding sentences when
> there is more than one such "argument" in it, and an answer I received
> seems to indicate that you have an official solution to this (good! --
> so far as I know, TLI Loglan does not) but that it goes contrary to
> the natural analogy with implicit quantification (not so good --
> explained fully in another post)
I'm not sure if there really was an official solution, since that
particular case is not explicitly mentioned in the connectives paper.
I agree that the opposite order to the one I suggested may be better.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: "any" & quantification
la kris cusku di'e
> Suppose (only for the sake of discussion!) we had a manditory particle
> before every sumti in lojban, a choice of either "xe'e" meaning
> referentially opaque, or "xa'a" meaning referentially clear. That makes the
> "box" example easy to analyze: "xa'a mi nitcu xe'e lo tanxe" means "I need
> a box" and "xa'a mi nitcu xa'a lo tanxe" means "There is a box I need".
Excellent idea.
> Maybe it would help to make that assumption for a while just for the purpose
> of exploring what the implications would be on other sentences.
>
> xa'a pa remna cu mamta xa'a mi - One person is my mother
> xe'e pa remna cu mamta xa'a mi - Only one person can be my mother (??)
The second one is "Any one person is my mother".
Pretty meaningless, but consider a more contrived situation;
Say there are three women in front of us, and I tell you
xa'a pa le ci ninmu cu mamta xa'a mi
One of the three women is my mother.
That's clear. Now the other case is
xe'e pa le ci ninmu cu mamta xa'a mi
Any one of the three women is my mother.
(Because we are actors, and in a play we're doing, the role of
mother is played by any of them, so only one is my mother, but
any one.)
> xa'a mi nelci xa'a do - Some of us like some of youse
> xe'e mi nelci xe'e do - ???
Any (one) of us likes any (one) of youse.
Which doesn't say much. (It doesn't mean that every one likes everyone,
even though the English version could be read like that.)
> Am I correct in thinking that the current disagreement between Jorge and
> Lojbab is whether unmarked sumti in real lojban are equivalent to sumti
> marked with "xe'e" or "xa'a"?
Correct, as far as I understand it. I say unmarked should always be xa'a,
while lojbab (I think) says that for {nitcu} it should be xe'e and for the
others xa'a.
> If so, let's do more translations with these
> two markings and see which one comes out more like the way we think lojban
> is currently defined.
I can't think of any examples that don't involve {nitcu} or {djica}, or
maybe {cpedu}, for which the xe'e marking would be natural for {lo broda}.
> Or could it be that the marking is only possible in certain place
> structures, and it is meaningless to contemplate "xe'e mi nelci xe'e do"?
I'd say useless rather than meaningless in that case.
> By the way, is "xa'a" as I've defined it the same as asserting existence?
Does {mi viska lo pavyseljirna} assert existance, or the fact that there
are no unicorns makes the sentence false?
Your "xa'a" asserts existance as much as "lo" does, as far as I can tell.
> (if "xa'a" isn't an unassigned cmavo and actually means something already,
> the my apologies and could someone clued in on the "experimental list" pick
> a better one?)
I don't think there is such a list, at least I've never seen it.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: general response on needing books
And:
>> How do you distinguish, (preferably in non-pedantic usage):
>> (1a). I need there to be a specific book such that I have it.
>> (1b). There is a specific book such that I need to have it.
>> (2a). I need there to be x, such that x is a book & I have x.
>> (2b). There exists x, such that x is a book & I need to have x.
>
>None of the replies wholly answered my intended query (because my
>intention was not in the least apparent from my wording).
>
>Let me explain, with an altered example: "I need you to have a book",
>instead of "I need to have a book". My idea was that the +/- specific
>distinction (LE v. LO) cuts across the transparent/ opaque distinction.
I agree with that. But I don't think your examples really show that.
>(1a-b) are +specific, (2a-b) are -specific, (1a,2a) are opaque, (1b,2b)
>are transparent.
The way I (mis?)understood the Quine excerpt, all your examples are
transparent.
"I need that there be a book such that I have it" is the
transparentisation of "I need a book" in its opaque sense.
"There is a book such that I need it" is a simple re-wording of
"I need a book" in its transparent sense.
Also, you didn't make very clear the +/- specific distinction.
"A specific book" is a non-specific reference, just as "that nondescript
book" is specific.
>One of (1a-b) (or maybe, ambiguously, both) is, in Lojban:
> (1a/b) Mi nitcu le/lo nu do ponse le cukta
I would say:
Mi nitcu le nu do ponse le cukta
I need that you have the book.
Mi nitcu lo nu do ponse le cukta
There is an event of you having the book that I need
I would say that bringing "lo nu" into this discussion will complicate
things unnecessarily.
(1a) and (2a) are very similar
>> (1a). I need there to be a specific book such that I have it.
>> (2a). I need there to be x, such that x is a book & I have x.
I would translate both as
mi nitcu le nu mi ponse lo cukta
(Or maybe {lo steci cukta} for "a specific book")
>One of (2a-b) (or maybe, ambiguously, both) is, in Lojban:
> (2a/b) Mi nitcu le/lo nu do ponse lo cukta
I disagree, of course. The quantification does not go outside the "le nu".
So:
>> (2b). There exists x, such that x is a book & I need to have x.
Has to be:
da poi cukta zo'u mi nitcu le nu mi ponse da
The simple {mi nitcu le nu mi ponse lo cukta} never implies the existence
of a book, and I think even lojbab will agree with that.
>I should clarify what I mean by (1a) (specific & opaque): I don't
>necessarily have the referent of 'le cukta' in mind, but were it to come
>to evaluating the truth of the lenu clause I would have a specific
>referent of 'le cukta' in mind. "I need there to be a specific book
>such that you have it" -
Can we change "a specific book" to a truly specific reference like
"that book"? Otherwise, it is confusing to translate it with {le}.
Then, do you really mean to say "I need there to be that book such
that you have it". With a truly specific reference it doesn't make that
much sense.
>Further remarks.
>(i) My impression so far is that (1a/b) and (2a/b) really are ambiguous
>in Lojban. (This is currently being disputed on the List.)
Lojban is perfectly capable of distinguishing between them, if you are
as wordy as you are being in English in those examples.
The question is whether {mi nitcu lo cukta} is as ambiguous as
"I need a book", which can mean both. I hold that it has to have the
transparent meaning.
(And I am also certain that no matter what we end up deciding, it will
be happily used by everyone, including myself when I'm not paying
attention, with the opaque meaning.)
>(iii) (1a/b) cannot be disambiguated.
Yes they can. (1a) is {mi nitcu le nu mi ponse lo cukta}
For (1b), if you think {zo'u} is pedantic, you can say something like
{lo cukta cu zasti ije mi nitcu le nu mi ponse ra} = "A book exists,
and I need to have it".
>A new suggested solution: LE & LO are transparent (wide scope, with
>quantification/reference assignment outside their clause). But when
>combined with a certain cmavo (e.g. Jorge's xehe) they are opaque
>(narrow scope, with quantification/reference assignment outside their
>clause). Note that under my proposal "xehe" is used in tandem with
>LE/LO, not instead of them.
This is exactly my proposal. {xe'e} is a PA, so {xe'e cukta} means
exactly the same as {xe'e lo cukta}, and you can of course say
{xe'e le xukta}.
>Incidentally, contra Jorge, I think "mi nitcu lo tanxe" is illicit sumti
>raising:
Then you don't need {xe'e} at all. :)
> what you really need is some state of affairs to be the case,
>such that its not being the case would be disadvantageous.
So you say that "I need that book" doesn't have any meaning, other
than "I need the state of affairs where I have that book".
I agree that {nitcu} is like {djica}, so if one is illicit sumti raising,
the other is as well. I don't think either of them should be.
Is asking for a book also sumti raising? You really are asking for some
state of affairs to be the case where you have a book.
> A "less
>wordy" locution would require us to not say what we mean.
I disagree. {mi nitcu le vi cukta} = "I need this book" is as precise
as {mi nelci le vi cukta} = "I like this book".
The fact that "I need a book" usually has the opaque meaning is no reason
to prohibit {mi nitcu lo cukta} in its transparent meaning.
>My proposals don't solve the "any whatsoever" problem. Some but not all
>of these problems are solved by using wide scope universal
>quantification:
>
> You may read all of the books.
> It is permitted that for all x such that x is one of the books
> you read x.
> You may read any of the books.
> For all x such that x is one of the books it is permitted
> that you read x.
Right, but you have to be pedantic and quantify outside the abstraction.
{xe'e} saves you from having to do that in this case.
>But this, I feel, doesn't handle "You may read any three of the books"
>(i.e. a total of three books, selected freely from the set of all of the
>books). So maybe here a new quantifier is needed, as Jorge has been
>arguing.
This is how I would make the specific/non-specific and transparent/opaque
distinctions:
non-specific/transparent
mi nitcu re lo cukta
I need two books (only two books exist that I need).
non-specific/opaque
mi nitcu rexe'e lo xukta
I need any two books.
specific/transparent
mi nitcu le vi cukta
I need this book.
specific/opaque
mi nitcu rexe'e le ci cukta
I need any two of the three books.
(The last one is cheating, since it really is non-specific/opaque.
I think you can't have specific/opaque.)
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any? (response to Desmond)
la lojbab cusku di'e
> > lo remna cu mamta mi
> > A human being is mother to me
> >
> >is true. Not by virtue of the fact that {lo remna} is a human being,
> >but because of the fact that there is one human being that is in
> >relationship {mamta} with {mi}.
>
> We may be dealing with the idiosyncracies of individual predicates here.
> Replace mammta with "se bersa" and the answer is probably indeterminate,
> since you (probably) do not know whether you (will) have a son in a
> time-free sense.
Time is a different problem. Seen from a timeless perspective,
lo remna cu bersa mi
A human being is a son to me
is true if there is a human being that is (timeless) in relation
{bersa} with {mi}. This has nothing to do with statements like
a = b which are indeterminate if all we know is that a and b are
numbers, while a + a = 2a is true, even if all we know is that
a is a number.
lo remna cu bersa lo remna
is true, not indeterminate, but for a different reason than a + a = 2a.
The sentence is true because there are certain {lo remna} that make it so.
The equation is true for mathematical reasons, and as a consequence, it will
be true no matter how much more we restrict the value that a can take.
> >> On
> >> the other hand, in the absence of specific information about a and b, the
> >> sentence
> >> a^2 - b^2 = (a-b)^2
> >> (though a perfectly acceptable sentence) is neither true nor false.
>
> I think that this claim is a definition and not a given. You know they
> are numbers, and you know that there is at least one number assignment
> that could make it true (b=0). You lack specific information as to
> whether that (or any other specific value) is a permissible value of
> "b".
It is a definition if he claims that it is true. Since he explicitly says
that it is neither true nor false, it makes no restrictions on a and b.
> Pragmatic usage of "lo" has incomplete specification of necessary
> restrictions.
I really don't understand what that means.
> >Sentences with {lo} in Lojban are usually true or false.
>
> Is "mi nitcu lo [unikorn]' true or false?
If it is true that {mi} and at least one of those that [unicorn] are in
relationship {nitcu}, then it is true. This is different from the situation
withb the mathematical statements above.
> "In the absence of specific
> information" applies much more often to mathematical problems than to
> linguistic ones.
I don't understand that either.
> >For example:
> >
> > lo remna cu kalte lo remna
> > A human hunts a human
> >
> >is true only if there really is at least one human that hunts at least
> >one human. It's not a matter of giving values to each {lo remna}.
>
> Umm. Let me hedge this a bit. Remember that we have some modals that
> have significant truth-functional import, and some of them involve
> potentiality. We can translate "inflammable" by "jelca", not requiring
> explicit use of "ka'e". Is "lo remna cu jelca" true or false? - depends
> on the modalities.
You are bringing up things that have nothing to do with the quantification
problem under discussion. Write {ca'a} explicitly in all the sentences
that we've been using and the arguments don't change.
> >If "a" and "b" were replaced by {lo namcu} = "a number" in your
> >sentence, it would be a true sentence in Lojban, because there indeed
> >exists at least one "a" and at least one "b" that make it true.
>
> "'a' number" in the same sense as "I need 'a' box"???
{lo namcu} in the same sense as {lo tanxe} in {mi nitcu lo tanxe}, yes.
NOT "'a' number" in the same sense as "I need 'a' box" no, because usually
this last has the opaque sense where 'a' is closer to 'any' than to
'certain'.
a > b is neither true nor false in his system, because "a" and "b" don't
have any values assigned.
{lo namcu cu zmadu lo namcu} is true in Lojban, because there indeed is at
least a number that is greater than at least a number.
That is why his "indeterminates" are not at all equivalent to Lojban's {lo}.
> >> It
> >> becomes true in the presence of the information that b=0, and it becomes
> >> false in the presence of the information that a=5 and b=3.
> >
> >That sounds like it might be more or less equivalent (at least for some
> >purposes) to Lojban {le}
> >
> > le remna cu mamta mi
> > The human is mother to me.
> >
> >is true if by {le remna} I mean the human who is my mother. In that
> >sense, you can say that it's neither true nor false in the absence of
> >information of what {le remna} is referring to, but that information is
> >at least in principle always obtainable (by asking the speaker who they
> >meant by it). From what I understand, your "a" need not have a value
> >obtainable even in principle.
>
> Ask Shakespeare what he means by various passages in his plays.
I don't understand how this fits here either.
> >> I believe that indeterminates in this sense play a fundamental role in
> >> everyday reasoning as well as in mathematical reasoning. Ordinary language
> >> accomodates indeterminates nicely. The use of 'a box' in the sentence "I
> >> need a box." is an example. It is a way of referring to something whose
> >> type is known, but about which we have no other information.
> >
> >I think something like that is what I meant by my proposal of {xe'e},
> >although I don't have it that clear in my mind.
>
> I think that pragmatically, "lo" is used as a non-specific categorizer.
> I like the word "indeterminate" better than "non-specific", now that
> Desmond has brought it into the jargon.
Even if you call it indeterminate, {lo} has nothing to do with his
indeterminates.
> If I say "lo [unikorn] cu klama lo zarci", you do not know what unicorn
> I am talking about (much less what store).
No, but it will be true only if there is at least one of all unicorns and
a market that are {klama}-related.
If you say [a] cu klama [b]
where [a] and [b] are Desmond's indeterminates, ([a] is of type unicorn,
[b] of type market) then you can say that the statement is acceptable
but it doesn't have a truth value (unlike the Lojban statement) unless
you decide to give values to those indeterminates. Just like a = b does
not have a truth value.
> You only know that it is
> veridically a member of the class of unicorns, if such a member exists.
> "Unicorn" is serving as a 'type' for the sumti.
>
> Change the example from "lo [unikorn]" to "lo nanmu" and you may have
> the same situation.
I'm not sure I can follow what you say next. Changing to {lo nanmu}:
lo nanmu cu klama lo zarci
> If you require that "lo zarci" refer to a specific
> one store that merely hasn't been specified, rather than 'any' store in
> your "xe'e" sense,
Yes, that's what I require for the sentence to be true.
> then you do not know the truth value of the sentence
> unless you can say that for EVERY possible value of "lo zarci", it is
> true that at least one man goes there.
What? For AT LEAST ONE zarci. Definitely I don't require it for every zarci.
> Alternatively, you can say that (assuming that the sets exist), the
> statement means merely
>
> "su'oda poi nanmu ku'o su'ode poi zarci zo'u da klama de"
> There exists at least one man X, and at least one market Y such that: X
> goes to Y
That's exactly what it means, yes.
> But this really doesn't track with your "mamta" example above. Yeah, it
> works since there is indeed at least one human that is your mother, but
> there really is a little implication of specificity or you wouldn't
> argue so comfortably that it is true.
Again, what?
>
> But how do you evaluate a story:
>
> "lo nanmu cu klama co jibni lo ninmu .i le nanmu cu cpedu le ninmu lenu
> kansa klama le dansu nunsalci"
>
> "A man goes near a woman. And the man asks the woman to
> accompanyingly-go to the dance-celebration."
>
> Now what do you make of this? Is the first sentence inherently true
> because at least one man has at some time gone near a woman?
Yes. If you are telling a story, you would probably want to use {le} there,
but {lo} is ok too.
> If so, it
> makes "lo" rather useless.
Why?
> I think that there may indeed be a 'typing'
> going on here, and the 2nd sentence "le" is an instantiation that tells
> us that the first sentence WAS referring to a specific man and a
> specific woman.
Pragmatically, yes. Other than that, there's nothing to tell you that
the man you are talking about is the one that makes the first sentence true.
> hypothetical mode:
>
> IFF Desmond's concept turns out to be what we (want to) mean by "lo"
No, please no. It makes things very unintuitive. That's my opinion anyway.
> (and by extension "loi" and "lo'i", though the standard quantification
> values attached to those may tend to make them a little less
> problematical), would this resolve the issues of "I need a box"?
Maybe. But it would create infinitely many more problems.
> What
> new issues can you see it introducing? In particular, what actual
> Lojban usages that you can think of are incorrect and which are
> uncertain.
Probably every single sentence that used {lo} so far would mean
something different than intended.
Jorge
Date: Wed, 21 Sep 1994 02:00:28 -0400
From: Logical Language Group <lojbab>
Subject: Re: any
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
JL>By "more likely" which do you mean:
JL>
JL>(1) In most contexts where that Lojban sentence appears, it means something
JL> like the second expansion, but in some contexts it may mean something
JL> like the first.
JL>
JL>(2) It always means the same one of them, but you are not sure which, and
JL> you suspect it is the second, that's why you say "more likely".
JL>
JL>
JL>If (1) then you are throwing logic out the window, and saying that just
JL>like in English "I need a box" can have two meanings (Quine's transparent/
JL>opaque), the same happens in Lojban, and which meaning it is is determined
JL>by context.
JL>
JL>If (2) then the way we've been interpreting most other predicates is wrong.
JL>(Unless you say that {nitcu} should be interpreted in a different way than
JL>other predicates.)
(1), I think (it is 2am %^)
JL>> I'm not sure about truth-fucntional value
JL>> of the second expansion - what the value is of "le nu [false statement]"
JL>> is not clear.
JL>
JL>Because there is no truth value for it. It is a sumti, and sumti don't have
JL>truth values. I would even question that in "le nu [statement]", there is
JL>any truth value for that [statement], since it is not being claimed.
Staements about non-0existent things are meaningless (they may be defined
as false because they are meaningless, but this is definitional).
The present king of France is bald.
has problems being evaluated truth-functionally because if it is false, then
its logical negation (contradictory) must therefore be true.
Similarly, if unicorns do not exist, then "I need a unicorn", and "It is
false that I need a unicorn" are both incorrect statements (false?)
JL>> However your clarification may point to the fact that se nitcu should be a
JL>> abstraction regardless about how we resolve the "lo"/"any" question.
JL>
JL>This is what was done to {djica}, which is analogous to {nitcu}. But I
JL>don't see why the transparent meaning should be forbidden. {mi nitcu le vi
JL>tanxe}
JL>= "I need this box", makes perfect sense. And so does {mi djica le vi tanxe}
Does it? The (1) or (2) dichotomy above suggests that there is always a
sumti raising going on, and we are allowing the raising when there is no
scope problem, and not otherwise. That is, I think, different than the way
we have dealt with other raising questions (not necessarily wrong, but we
really oughta know what we are dealing with).
lojbab
From: Randall Holmes <holmes@DIAMOND.IDBSU.EDU>
Subject: Analogy
The relation between resolving meanings of multiple appearances of
logically connected arguments and implicit quantification is not precisely
one of analogy; the first is seen to be a special case of the second on
ananlysis, and so it would be better if they were handled in the same way.
Jorge remarks that he is not so certain that there really is an offical
line on how to resolve this kind of situation; I suspect in TLI Loglan
the solution is the one that I suggest (as soon as someone asks me about it).
--Randall Holmes
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: general response on needing books
In-Reply-To: (Your message of Tue, 20 Sep 94 20:22:57 EDT.)
Jorge:
> Also, you didn't make very clear the +/- specific distinction.
> "A specific book" is a non-specific reference, just as "that nondescript
> book" is specific.
"A certain book" would have been a better way of englishing it.
> (1a) and (2a) are very similar
>
> >> (1a). I need there to be a specific book such that I have it.
> >> (2a). I need there to be x, such that x is a book & I have x.
>
> I would translate both as
>
> mi nitcu le nu mi ponse lo cukta
>
> (Or maybe {lo steci cukta} for "a specific book")
You misunderstand my intention behind (1a), which I could rephrase
as "I need there to be a certain book such that I have it (but I
don't necessarily know which book this is)".
I'll return to this below.
> >Further remarks.
> >(i) My impression so far is that (1a/b) and (2a/b) really are ambiguous
> >in Lojban. (This is currently being disputed on the List.)
>
> Lojban is perfectly capable of distinguishing between them, if you are
> as wordy as you are being in English in those examples.
(1a/b) & (2a/b) were Lojban sentences. It is still in dispute whether
they're ambiguous.
> >(iii) (1a/b) cannot be disambiguated.
>
> Yes they can. (1a) is {mi nitcu le nu mi ponse lo cukta}
No. See below.
> >A new suggested solution: LE & LO are transparent (wide scope, with
> >quantification/reference assignment outside their clause). But when
> >combined with a certain cmavo (e.g. Jorge's xehe) they are opaque
> >(narrow scope, with quantification/reference assignment outside their
> >clause). Note that under my proposal "xehe" is used in tandem with
> >LE/LO, not instead of them.
>
> This is exactly my proposal. {xe'e} is a PA, so {xe'e cukta} means
> exactly the same as {xe'e lo cukta}, and you can of course say
> {xe'e le xukta}.
What do they mean, under your proposal?
> >Incidentally, contra Jorge, I think "mi nitcu lo tanxe" is illicit sumti
> >raising:
>
> Then you don't need {xe'e} at all. :)
I was suggesting a cmavo that marks whether quantification/reference
assignmetn takes place inside or outside the local abstracction.
> > what you really need is some state of affairs to be the case,
> >such that its not being the case would be disadvantageous.
>
> So you say that "I need that book" doesn't have any meaning, other
> than "I need the state of affairs where I have that book".
Right. Or it could be "mi nitcu lenu cukta", which I take to mean
"I need the book to exist".
> I agree that {nitcu} is like {djica}, so if one is illicit sumti raising,
> the other is as well. I don't think either of them should be.
I think they should both be. Needing and wanting both involve a
comparison of the existence/nonexistence of a state of affairs.
For you to persuade me that there isn't illicit raising you need
to givve me a definition of 'needing' & 'wanting' to back up
your view.
> Is asking for a book also sumti raising? You really are asking for some
> state of affairs to be the case where you have a book.
Right.
> > A "less
> >wordy" locution would require us to not say what we mean.
>
> I disagree. {mi nitcu le vi cukta} = "I need this book" is as precise
> as {mi nelci le vi cukta} = "I like this book".
No. For 'like' we could paraphrase 'contemplate with pleasure' (roughly),
& one can contemplate an object as well as a book. But a needee can
only be an event.
> The fact that "I need a book" usually has the opaque meaning is no reason
> to prohibit {mi nitcu lo cukta} in its transparent meaning.
Right, but it should be abjured as sumti raising, & as gobbledygook.
> I think you can't have specific/opaque.)
To evaluate the truth of "Lo cukta cu blanu" you examine every book,
and only if every book turns out not to be blue is the statement
false. To evaluate the truth of "Le cukta cu blanu" you identify
the referent of "le cukta" and then check whether it's blue.
"There is a certain book that I need to have": to evaluate the
truth of this you identify the referent of "a certain book"
(le cukta) & check whether I need to have it.
"I need there to be a certain book that I have": to evaluate the
truth of this you don't have to identify the referent of "a certain
book", but if you wanted to test whether my need had been satisfied
you would have to identify the referent of "a certain book".
This latter case is what I meant by 'specific & opaque': i.e. the
specific referent is established only within the local predication.
I'd be happy to use a term other than 'opaque' if you feel that
I'm misusing it.
Finally, I reiterate my cmavo proposals:
"xihi" - modifies LO/LE & indicates for LO that quantification
takes place in local abstraction & for LE that reference
is assigned in local abstraction. I assume that in the
absence of "xihi" quantification/ref.assignment takes
place at sentence level.
"xuhu" - "xuhu PA X cu blanu" indicates that PA things selected randomly
from the set containing only every X are blue, but no claim
is made about whether any additional X is blue.
-----
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any? (response to Desmond)
Desmond:
> >>> A feature of dr is the fundamental role in it of what I call
> >>> *indeterminates*. For example, if a and b are indeterminates of the sort
> >>> number, then the *unquantified* sentence
> >>> a^2 - b^2 = (a-b)(a+b)
> >>> is true. a and b are *potential entities* of the sort number. This may
be
> >>> the only information we have about them, or we may have total information
> >>> about them (such as that a=5 and b=3) or we may have partial information
> >>> about them (such as that a is positive). In each case our sentence
remains
> >>> true: it is true by virtue solely of the fact that a and b are numbers.
> Jorge:
> >>This is not the case for Lojban {lo}.
> >>
> >>For example:
> >>
> >> lo remna cu mamta mi
> >> A human being is mother to me
> >>
> >>is true. Not by virtue of the fact that {lo remna} is a human being,
> >>but because of the fact that there is one human being that is in
> >>relationship {mamta} with {mi}.
Desmond:
> I don't think that conflicts with lo remna being an indeterminate. A
> sentence involving an indeterminate may be true for various reasons. Your
> sentence is true in the presence of the information that humans have
> mothers.
No, that's not enough to make it true. One of all those that are humans
has to be my mother in order that the sentence be true.
> Everyone has that information so everyone agrees. (My "it is
> true by virtue solely of the fact that a and b are numbers" may be
> misleading. Add "and addition of numbers is associative, and ...".)
You can give a and b any values, and your sentence will remain true. If you
give different values to {lo remna} in my sentence, you get false statements.
The sentence is true because there is one remna in relationship {mamta}
with {mi}.
(Another difference is that you can have the same indeterminate "a" appearing
in various places in the sentence, while various {lo remna} in the same
sentence do not refer necessarily to the same one.)
But look at an example that is easy to translate:
a > b
is neither true nor false if all you know is that a and b are numbers.
lo namcu cu zmadu lo namcu
A number is greater than a number
_is_ true.
Then, you can't say that {lo namcu} = "a number" works like an indeterminate.
Do you agree?
> lojbab:
> >But how do you evaluate a story:
> >
> >"lo nanmu cu klama co jibni lo ninmu .i le nanmu cu cpedu le ninmu lenu
> >kansa klama le dansu nunsalci"
> >
> >"A man goes near a woman. And the man asks the woman to
> >accompanyingly-go to the dance-celebration."
> >
> This makes good sense to me. Could one put it like this? In the first
> sentence "a man" and "a woman " are indeterminates about which we have no
> information except their types (and all the agreed properties that go with
> their types). In the second sentence some more information about these
> same indeterminates is expressed. "the man" (and "le nanmu"?) tell us that
> whatever entity may have been referred to in the first sentence is also the
> entity referred to in the second.
Not in that Lojban sentence. {le nanmu} is not necessarily a nanmu that makes
the first sentence true. (Context suggest it is, though, but there are better
ways to translate that English sentence into Lojban.)
> English, though clear, is rather clumsy
> about this. Standard mathematical usage would be "m goes near w. And m
> asks w to ...." English would require "the first man" and "the second man"
> to refer again to "a man" and "another man".
In Lojban one could say:
ko'a klama le jibni be ko'e i ko'a cpedu fi ko'e ...
A goes near B. A asks B to...
Just like your mathematical example. Probably the {ko'a} series work well
as your indeterminates. They may or may not have previously assigned values
or types.
> I would be interested in the simplest lojban
> rendering of
> (1) A man is eating an icecream. The man is happy.
> (2) Two people are in a room. The man is happy.
> (3) A man may eat an icecream. A man may be happy.
> (4) A man may eat an icecream. That would make him happy.
You must realize that in different contexts, those sentences will mean
different things and may have to be translated differently.
Here's a possible rendering:
(1) le nanmu ca'o citka le biskruji i ny gleki
(2) le re prenu cu nenri le kumfa i le nanmu cu gleki
(3) lo nanmu ka'e citka lo biskruji i lo nanmu ka'e gleki
(4) le nanmu ka'e citka le biskruji i le nu ca'a go'i cu geirgau ny
Jorge
Date: Wed, 21 Sep 1994 13:35:04 -0700 (PDT)
From: "John E. Clifford" <pcliffje@crl.com>
Subject: Re: silence?
In-Reply-To: <199409211059.AA10281@access1.digex.net>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Silence implies assent, as the old saying goes. But mainly I am
overwhelmed by the volume (I know 4 a day is just an average but still,
20 a day is bound to scew the values soon). And I find that, by the time
I have formulated my response, someone else has done it more elegantly.
then, by the time I have formulated a defense against carping critics,
someone else has done that too. Eventually, I will weigh in to sum up on
the side of right and truth, but it seems that most of the problems take
care of themselves in pretty short order (burger and fries, small coke).
The right of "any" is raised sumti from an event abstract, of course. I
still haven't sorted out the "le/lo" problems yet, though. And the
hyphenated Aussie have haven't penetrated at all. pc>|83
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any? (response to Desmond)
la djan fanva di'e lo cizra mifra la gliban de'e
> > (4) A man may eat an icecream. That would make him happy.
>
> 4) lo -man (modal) -eat loi -icecream gi'e -happy.
All the others are debatable, because not much context is given,
but in this one you are ignoring the causality clearly present in
the English sentence.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: general response on needing books
And:
> What do they mean, under your proposal?
{paxe'e lo cukta} Any one book. Examining whether a relationship holds
for individual cukta does not help to decide whether the relationship
holds for {paxe'e cukta}.
{paxe'e le cukta} The same, but any of "the books", rather than any book
whatsoever.
> I was suggesting a cmavo that marks whether quantification/reference
> assignmetn takes place inside or outside the local abstracction.
That's not my xe'e then. I was confused.
> > I agree that {nitcu} is like {djica}, so if one is illicit sumti raising,
> > the other is as well. I don't think either of them should be.
>
> I think they should both be. Needing and wanting both involve a
> comparison of the existence/nonexistence of a state of affairs.
> For you to persuade me that there isn't illicit raising you need
> to givve me a definition of 'needing' & 'wanting' to back up
> your view.
ca'e ko'a nitcu ko'e
ijo tu'e ko'a claxu ko'e
i le nu go'i cu to'e mansa ko'a tu'u
ni'o
ca'e ko'a djica ko'e
ijo tu'e ko'a claxu ko'e
i ko'a gleki le nu la'e di'u cu sisti tu'u
Maybe my definitions are not very good, but they mainly say that {nitcu}
and {djica} are forms of {claxu} with more properties for x1.
Does {claxu} suffer from illicit raising as well?
> > I disagree. {mi nitcu le vi cukta} = "I need this book" is as precise
> > as {mi nelci le vi cukta} = "I like this book".
>
> No. For 'like' we could paraphrase 'contemplate with pleasure' (roughly),
> & one can contemplate an object as well as a book. But a needee can
> only be an event.
If 'like' is 'contemplate with pleasure', then 'want' is 'contemplate with
desire' and 'need' is 'contemplate with hunger'.
> > The fact that "I need a book" usually has the opaque meaning is no reason
> > to prohibit {mi nitcu lo cukta} in its transparent meaning.
>
> Right, but it should be abjured as sumti raising, & as gobbledygook.
I agree that if {djica}, {sisku}, etc have been so treated, so should
{nitcu}. Unfortunately, we lose the capability to say simple things
like "I'm looking for my umbrella".
> > I think you can't have specific/opaque.)
>
> To evaluate the truth of "Lo cukta cu blanu" you examine every book,
> and only if every book turns out not to be blue is the statement
> false.
Yes. I wish lojbab would agree that to evaluate the truth of
"lo cukta cu se nitcu" (ignoring raising for the moment) one should
follow exactly the same procedure.
> To evaluate the truth of "Le cukta cu blanu" you identify
> the referent of "le cukta" and then check whether it's blue.
Exactly.
> "There is a certain book that I need to have": to evaluate the
> truth of this you identify the referent of "a certain book"
> (le cukta) & check whether I need to have it.
That "a certain book" sounds non-specific to me. How come this
doesn't work with "that book" or "the book" or "my book"? I think
that we are confusing the specificity of the reference ("le cukta"
is a specific reference) with that of the referent (very likely
that I'm using the wrong words). In "a certain book", the referent
is specific, but the reference is non-specific.
I really don't see any significant difference between "there is a
certain book that I need to have" and "there is some book that I
need to have".
> "I need there to be a certain book that I have": to evaluate the
> truth of this you don't have to identify the referent of "a certain
> book", but if you wanted to test whether my need had been satisfied
> you would have to identify the referent of "a certain book".
The satisfaction of the need is irrelevant to the claim anyway.
> This latter case is what I meant by 'specific & opaque': i.e. the
> specific referent is established only within the local predication.
Could you give an example with a clearly specific reference "the book",
"this book", "my book", or something like that? Your example seems
non-specific to me.
> I'd be happy to use a term other than 'opaque' if you feel that
> I'm misusing it.
I could well have misunderstood the Quine passage, but according
to what I understand, all your examples are transparent.
> Finally, I reiterate my cmavo proposals:
> "xihi" - modifies LO/LE & indicates for LO that quantification
> takes place in local abstraction & for LE that reference
> is assigned in local abstraction. I assume that in the
> absence of "xihi" quantification/ref.assignment takes
> place at sentence level.
You mean that for {lo} the xihi-less quantification would be outside
the abstraction? I think that goes against current usage.
And what would local quantification for {le} mean?
{mi nitcu le nu mi ponse le vi cukta} is "I need to have this book".
What would {mi nitcu le nu mi ponse xi'i le vi cukta} mean?
> "xuhu" - "xuhu PA X cu blanu" indicates that PA things selected randomly
> from the set containing only every X are blue, but no claim
> is made about whether any additional X is blue.
This is one of the many meanings of "any". Do you think it is the most
useful? I think that if you change it to "only PA things" then you can
recover your meaning with {xu'u su'oPA}, and it would be close to what
I meant by {xe'e} (I think).
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any
I think we are converging!
> JL>= "I need this box", makes perfect sense. And so does {mi djica le vi
tanxe}
>
> Does it? The (1) or (2) dichotomy above
[that was (1) {mi nitcu lo tanxe} can have opaque/transparent meaning
according to context.
(2) It's always transparent (or always opaque, but I think
no one argued for always opaque.) ]
> suggests that there is always a
> sumti raising going on, and we are allowing the raising when there is no
> scope problem, and not otherwise.
Since I don't fully understand when sumti raising is acceptable and when
it isn't, I may or may not agree with that.
> That is, I think, different than the way
> we have dealt with other raising questions (not necessarily wrong, but we
> really oughta know what we are dealing with).
Yes. I think {djica} and {nitcu} should be treated alike. (And also {sisku},
which for some reason is treated differently, and maybe {cpedu}, and who
knows how many more.)
>
> lojbab
>
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Any old thing whatsoever (mi nitcu lo tanxe)
Mark Biggar says:
> Yes, but the meaning of "I need two boxes" can depend heavly on context
> as it can mean the same as "I need another box" in the case where you
> already have 1 box, it is full and you still got stuff to pack.
You're right. Complications keep cropping up...
> So
> How do you say "I need another box" with the same type of transparentness
> as when you say "I need a box"?
{mi nitcu re tanxe} is still good, because having it doesn't preclude
needing it. You still need the one that is packed full.
If you need the second "opaquely", you could say {mi nitcu le nu
mi ponse lo remoi tanxe} = "I need to have a second box".
Or something.
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: TECH: Transparency / Opaqueness
I'm just thinking (which is dangerous as ever :-).
If I have
(a) I'll read two books. [any two]
(b) I'll read two books, this and that.
((a) is the opaque case and (b) the transparent one)
and put it into Lojban, I get
(a) mi ba tcidu re [lo] cukta
(b) mi ba tcidu le [bi'u] re [lo] cukta ku ne ti .e ta
In (b) I think we must use {le} as I clearly indicate
two quite specific things in spite of the English surface
structure.
Now I'm wondering what would happen, if we said in
general that {le} is used for transparency and {lo}
for opaqueness. Already the implicit outer/inner
quantifiers seem to imply this: {su'o lo ro} vs.
{ro le su'o}, i.e. sampling vs. grabbing them all.
t mi nitcu le tanxe
I need the box(es)
o mi nitcu lo tanxe
I need [one or more] boxes
t mi nitcu le pa tanxe [ku poi ...]
I need the one box [which is ...]
t mi nitcu le pa lo tanxe
I need a certain one of all the boxes there are
mi nitcu le poi blanu ku'o pa lo tanxe
I need a certain blue box
le pa lo remna cu mamta mi
a certain human being is my mother
o mi nitcu pa [lo] tanxe
I need a box [any box of all the boxes]
mi nitcu pa lo tanxe ku poi blanu
I need a blue box [any blue one of all the boxes]
mi nitcu pa lo tanxe poi blanu
I need a blue box [any one of the blue boxes]
o (*) mi nitcu pa [lo ro] le ci tanxe
I need one of the three boxes [any one of all the three]
t mi nitcu le poi blanu ku'o pa le ci tanxe
I need the one which is blue of the three boxes
(*) This gave me some trouble before I realized that I can
insert {lo ro} without change of meaning. Even if there
are actually 4 boxes (or whatever I decide to call boxes),
the totality is veridical. I think that even generally it
might be useful to define an outer quantifier of the {le}
descriptor to include an elidable {lo ro}.
This way things would fit within the existing framework and
we would need no new cmavo (Jorge's xe'e). Secondly we could
crossreference {le/lo} quite nicely to Quine and remove even
more of the correspondence to the English {the/a} (or the
Indo-European definite/indefinite articles in general).
Now let's have all the resulting problems :-)
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
Date: Thu, 22 Sep 94 09:16:30 -0600
From: Randall Holmes <holmes@diamond.idbsu.edu>
Subject: Re: Analogy
Cc: lojban@cuvmb.cc.columbia.edu
So what is the official line? Does it coincide with the interpretation
using implicit quantification or does it reverse it? If it reverses it,
I really do suggest that you change it.
--Randall Holmes
Date: Thu, 22 Sep 1994 09:09:13 -0700 (PDT)
From: "John E. Clifford" <pcliffje@crl.com>
Subject: Re: silence?
In-Reply-To: <199409220521.AA10382@access4.digex.net>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sorry I was unclear. I meant "any" in the context of "need" (and other
intentional predicates - "want", "strive for", ...) is in -- and raised
out of -- an event abstraction and quantifies over the entities in that
world, not the real one. But the real entities will satisfy the tension
in the intention, so they are not irrelevant. This takes some writing up
and thinking through and I feel uncomfortable doing the thinking in print
(unlike some of the people out there, who shoot away until they hit
somehing -- even it it is only a superficial scratch). Maybe I'll get
over this in a while, but right now the pressure of just scanning the
latest takes up all my Lojban time. pc>|83
From: jimc@MATH.UCLA.EDU
Subject: Re: TECH: Transparency / Opaqueness
In-Reply-To: Your message of "Thu,
22 Sep 94 10:00:04 +0300."
<9409220701.AA00562@julia.math.ucla.edu>
Veijo Vilva <veion@XIRON.PC.HELSINKI.FI> writes:
> Now I'm wondering what would happen, if we said in
> general that {le} is used for transparency and {lo}
> for opaqueness. Already the implicit outer/inner
> quantifiers seem to imply this: {su'o lo ro} vs.
> {ro le su'o}, i.e. sampling vs. grabbing them all.
.i lo xrula cu te mlepurdi la veion lemi mrita'e
This, with the examples, is the best statement of the problem that I've seen
so far, and a very believable solution.
I'm not comfortable with the symbolism of Quine's opaque vs. transparent
metaphor;
I prefer a metaphor of formal vs. actual parameters of a Boolean valued
function,
similar to Desmond's usage with his "dr" language.
JCB originally defined "le" to mark a sumti as having an "in-mind" referent so
that the referent did not actually have to satisfy the predicate; he devoted
several paragraphs to "that man is [really] a woman: _le_ va nanmu cu ninmu.
Nonetheless, his emphasis is on the referent being specifically chosen by the
speaker. We should be sure that when the meaning of "le" is finally nailed
down, a counter-to-fact s-bridi will still be legal, but the counter-to-fact
possibility should not over-dominate the design choices.
The word "mass" has been used to refer to the referent of "lo" as Veion has
suggested it be used; I paraphrase the usage in these nearly equivalent ways:
lo broda is any member of the set lo'i broda (members really satisfy the
s-bridi)
The referent of "lo broda" is potentially any member of the set.
From the point of view where "lo broda" is written, different set members
are more-or-less interchangeable. This is my interpretation
of the "Mr. Rabbit" metaphor that JCB used to explain TLI "lo".
These differ from lo'e broda, a typical item actually satisfying broda. They
also differ from "lei" or "loi" which is the determiner used in Lojban for
"mass".
Often appearing in a discussion of masses is the "team" (bende). One can use
"loi" to say "loi se bende cu zbasu le dinju" (the members of the team, as a
mass, build the house", whereas it is not true that any one member does the
whole job, the members are not interchangeable as Veion's version of "lo" would
imply, the needed quantification is that all (ro loi) the members participate
rather than some (su'o lo), and it is not reasonable to say that a "set" does
the building so that "le bende cu zbasu..." is deprecated.
I hope this doesn't muddy the water excessively.
James F. Carter Voice 310 825 2897 FAX 310 206 6673
UCLA-Mathnet; 6221 MSA; 405 Hilgard Ave.; Los Angeles, CA, USA 90024-1555
Internet: jimc@math.ucla.edu BITNET: jimc%math.ucla.edu@INTERBIT
UUCP:...!{ucsd,ames,ncar,gatech,purdue,rutgers,decvax,uunet}!math.ucla.edu!jimc
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: remarks on gismu lexicology
A heterogeneous assembly of comments & queries
mostly concerning gismu lexicology.
(1) x2 of nitcu (need), djica (want) and cpedu
(request), and x3 of pikci (beg) should be an
event abstraction. "Need/want/ask/beg to have"
should be rendered "ponse zei nitcu/djica/cpedu/
pikci" (with x2 of the lujvo being x2 of ponse).
(2) Facki (find, discover) is redundant.
Facki is equivalent to cilre (learn) (except that
cilre has an x4 place for method, which is anyway
not inappropriate for discovering.)
(3) Sisku (seek) is redundant and too vague.
(x2 of sisku is in my list object/event/property.
I believe that this has been changed to just property,
but either way I don't understand why.) "Mi sisku do"
is either "Mi troci lenu mi { penmi [=encounter] / ponse
[=possess] / jitro [=control] } do" or "I try to learn where
you are" (don't know how to say that, but it involves
"mi troci lenu mi cilre" (-- can someone please remind me how
Lojban handles subordinate interrogatives, as in "I wonder
what you're reading"?)). These could be condensed into
penmi/ponse/jitro zei troci (with x2 of lujvo being x2 of
penmi/ponse/jitro).
(4) How to say "I search the pockets"? "mi zukte fe
le nu catlu le daski kei fi le nu mi penmi/ponse/jitro/kavbu"?
(Or with lujvo, "mi catlu zei zukte le daski le nu
mi penmi/ponse/jitro/kavbu".)
I assume "catlu" means not "look" but "inspect, examine".
(5) How to say "watch, heed, pay attention to"? 'Zgana'
doesn't seem right.
(6) simlu: x1 seems/appears to have property(ies) x2 to
observer x3 under conditions x4.
So "I seem blue" is "mi simlu le ka blanu"?
Then how to say "It seems to be raining, it seems that it
is raining"? I think we should be able to say "simlu fa
le duhu carvi" - that is, x1 of simlu is a duhu abstraction
and x2 is scrapped. "I seem blue" would be "simlu fa le duhu
mi blanu".
simlu: x1 (duhu) seems-to-be-the-case to observer x2
under conditions x3.
or, perhaps more usefully:
simlu: to observer x1 x2 (duhu) seems-to-be-the-case
under conditions x3.
(This latter order avoids need for 'fa' to postpose the
duhu clause, & lends itself as a translation of "it seems
to me that...".)
(7) galfi: x1 (event) modifies/alters/changes x2 into x3
stika: x1 (event) adjusts/changes x2 (ka/ni) in amount/degree x3
I think the x1 place of these should be abolished.
Galfi then becomes redundant with binxo:
binxo: x1 becomes/changes into x2 under conditions x3
And I think binxo should have an extra place:
binxo: x1 changes from belonging to category (ka) x2
into belonging to category (ka) x3
I think some new but related meaning should be found for galfi,
such as:
galfi: x1 evolves from (ka) x2 into (ka) x3 under conditions/
constraints x4 [e.g. natural selection]
(8) panci: x1 is an odor/fragrance/scent/smell emitted by x2
and detected by observer/sensor x3
sumne: x1 (experiencer) smells [transitive verb] x2;
x2 smells/has odor to observer x1
ganse: x1 [observer] senses/detects/notices/is aware of
stimulus x2 by means x3 under conditions x4
vrusi: x1 is a taste/flavor of x2
(a) Given sumne, why does panci have this x3 place? Suppose
I want to describe the smell of an unsmelt rose.
(b) Why does sumne lack a place for the odour?
(c) Why does 'vrusi' have no 'transitive' counterpart? I suppose
we could have:
vrusi zei ganse: x1 tastes taste x2 of x3
But in this case, why bother with having sumne?
(d) Unless I've misunderstood, I suggest dropping the x3 of panci,
and dropping sumne altogether, using panci zei ganse instead.
(9) Is there an agreed expression for look/appearance/
countance/visual stimulus, without there being an implied
perceiver?
(10) tirna: x1 hears x2 against background/noise x3
Could x3 be abolished, please? Otherwise, when there's
no background noise we'll have to remember to use "xohe"
(or whatever the sumti-abolishing cmavo is).
(11) Is there a standard expression for 'saliva'?
---
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: TECH: Transparency / Opaqueness
In-Reply-To: (Your message of Thu, 22 Sep 94 10:00:04 P.)
La Veion:
> (a) I'll read two books. [any two]
> (b) I'll read two books, this and that.
>
> ((a) is the opaque case and (b) the transparent one)
> and put it into Lojban, I get
>
> (a) mi ba tcidu re [lo] cukta
> (b) mi ba tcidu le [bi'u] re [lo] cukta ku ne ti .e ta
>
> In (b) I think we must use {le} as I clearly indicate
> two quite specific things in spite of the English surface
> structure.
> [....]
> This way things would fit within the existing framework and
> we would need no new cmavo (Jorge's xe'e). Secondly we could
> crossreference {le/lo} quite nicely to Quine and remove even
> more of the correspondence to the English {the/a} (or the
> Indo-European definite/indefinite articles in general).
It has been established (to my satisfaction, at any rate)
that LE/LO is +/-specific [Colin propounded this most lucidly].
It only relates to definiteness in
that only +specifics can be +/-definite.
I think that we do need a new cmavo & that LO/LE isn't
the same as transparent/opaque. You seem to miss the ambiguity
of (a).
I'll read any two books. - pick two items freely from the
set of all books, & it is asserted that I'll read them.
There are two books I'll read. - examine every book & if
you find at least two that I'll read, the assertion is
true.
I hope I didn't misunderstand you.
-----
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: general response on needing books
In-Reply-To: (Your message of Wed, 21 Sep 94 20:16:53 EDT.)
Jorge:
> > What do they mean, under your proposal?
>
> {paxe'e lo cukta} Any one book. Examining whether a relationship holds
> for individual cukta does not help to decide whether the relationship
> holds for {paxe'e cukta}.
>
> {paxe'e le cukta} The same, but any of "the books", rather than any book
> whatsoever.
Very very similar to my "xuhu". Good!
> > > I agree that {nitcu} is like {djica}, so if one is illicit sumti raising,
> > > the other is as well. I don't think either of them should be.
> >
> > I think they should both be. Needing and wanting both involve a
> > comparison of the existence/nonexistence of a state of affairs.
> > For you to persuade me that there isn't illicit raising you need
> > to givve me a definition of 'needing' & 'wanting' to back up
> > your view.
> ca'e ko'a nitcu ko'e
> ijo tu'e ko'a claxu ko'e
> i le nu go'i cu to'e mansa ko'a tu'u
> ni'o
> ca'e ko'a djica ko'e
> ijo tu'e ko'a claxu ko'e
> i ko'a gleki le nu la'e di'u cu sisti tu'u
> Maybe my definitions are not very good, but they mainly say that {nitcu}
> and {djica} are forms of {claxu} with more properties for x1.
> Does {claxu} suffer from illicit raising as well?
Claxu doesn't suffer from illicit raising, but it does lead to scope
ambiguities.
mi claxu lo cukta
perhaps can mean either:
Ex book(x) & not have(me,x) "There's a book that I lack"
not Ex book(x) & have(me,x) "I don't have any books"
Hopefully this second is not a possible meaning of the Lojban
sentence. However, it is what we usually mean by "I lack (any) books".
Given your not unreasonable interpretation of nitcu and djica as
being kinds of claxu, you are quite right that they don't involve
sumti raising. However I strongly urge that your interpretations
should be rejected. First, in English I can say "I want/need this
book" even if I already have it - I mean that I prefer having the
book over not having it, & am not asserting that I don't have it.
So for English, wanting/needing doesn't entail lacking.
Second, if the x2 of nitcu/djica is an event, we can translate
"I want/need you to go", which mean, roughly, I prefer your going
over your not going. Third, "want/need to have" can then be
rendered "ponse zei djica/nitcu". Fourth, wanting/needing +
lacking can be rendered "djica/nitcu zei claxu".
> > > I disagree. {mi nitcu le vi cukta} = "I need this book" is as precise
> > > as {mi nelci le vi cukta} = "I like this book".
> > No. For 'like' we could paraphrase 'contemplate with pleasure' (roughly),
> > & one can contemplate an object as well as a book. But a needee can
> > only be an event.
> If 'like' is 'contemplate with pleasure', then 'want' is 'contemplate with
> desire' and 'need' is 'contemplate with hunger'.
No. Like X = 'contemplate X & experience pleasure'. 'Contemplate with
desire' = 'contemplate X & experience desire for x'. This is indeed
djica, but it doesn't decompose analogously to nelci.
> > > The fact that "I need a book" usually has the opaque meaning is no reason
> > > to prohibit {mi nitcu lo cukta} in its transparent meaning.
> >
> > Right, but it should be abjured as sumti raising, & as gobbledygook.
>
> I agree that if {djica}, {sisku}, etc have been so treated, so should
> {nitcu}. Unfortunately, we lose the capability to say simple things
> like "I'm looking for my umbrella".
I sent off another posting covering this. My answer would be that
lujvo can be adopted to give briefer locutions. E.g. ponse zei
nitcu, with x2 being the thing possessed.
> > > I think you can't have specific/opaque.)
> >
> > To evaluate the truth of "Lo cukta cu blanu" you examine every book,
> > and only if every book turns out not to be blue is the statement
> > false.
>
> Yes. I wish lojbab would agree that to evaluate the truth of
> "lo cukta cu se nitcu" (ignoring raising for the moment) one should
> follow exactly the same procedure.
>
> > To evaluate the truth of "Le cukta cu blanu" you identify
> > the referent of "le cukta" and then check whether it's blue.
>
> Exactly.
>
> > "There is a certain book that I need to have": to evaluate the
> > truth of this you identify the referent of "a certain book"
> > (le cukta) & check whether I need to have it.
>
> That "a certain book" sounds non-specific to me. How come this
> doesn't work with "that book" or "the book" or "my book"? I think
> that we are confusing the specificity of the reference ("le cukta"
> is a specific reference) with that of the referent (very likely
> that I'm using the wrong words). In "a certain book", the referent
> is specific, but the reference is non-specific.
I am not trying to translate the English sentence; the English
sentence is merely an attempt to indicate the meaning I'm trying
to describe. I can't use 'that/this/my', because reference of
these is assigned at sentence level, not only in an inner bridi.
There is no way I can think of to say in English what I want to say.
If you will agree to define specificity as I did above (you do
say "exactly"), then would you agree that there is a difference
in meaning according to whether you have to identify the referent
in the local bridi or in the outermost bridi. My point is precisely
that there is a difference, & one worth making expressible in
Lojban.
> > "I need there to be a certain book that I have": to evaluate the
> > truth of this you don't have to identify the referent of "a certain
> > book", but if you wanted to test whether my need had been satisfied
> > you would have to identify the referent of "a certain book".
>
> The satisfaction of the need is irrelevant to the claim anyway.
Right. Yes.
> > This latter case is what I meant by 'specific & opaque': i.e. the
> > specific referent is established only within the local predication.
>
> Could you give an example with a clearly specific reference "the book",
> "this book", "my book", or something like that? Your example seems
> non-specific to me.
I can't, for the reasons given above.
> > Finally, I reiterate my cmavo proposals:
> > "xihi" - modifies LO/LE & indicates for LO that quantification
> > takes place in local abstraction & for LE that reference
> > is assigned in local abstraction. I assume that in the
> > absence of "xihi" quantification/ref.assignment takes
> > place at sentence level.
>
> You mean that for {lo} the xihi-less quantification would be outside
> the abstraction? I think that goes against current usage.
This is what I mean. I suppose it could be the other way around,
but this would be needed less often. Alternatively there could
be 2 cmavo, one for 'outermost' and one for 'local', with it
understood that if neither is used there is a potential
ambiguity.
> And what would local quantification for {le} mean?
> {mi nitcu le nu mi ponse le vi cukta} is "I need to have this book".
> What would {mi nitcu le nu mi ponse xi'i le vi cukta} mean?
Local reference assignment, not local quantification - I don't
think 'le' involves quantification. Your first example is
translated okay. Your second means, very roughly, "I need to
book-have", & in order to decide whether my need is satisfied
you have to ask "which book?". I agree that you can evaluate the
sentence without deciding whether my need is satisfied.
> > "xuhu" - "xuhu PA X cu blanu" indicates that PA things selected randomly
> > from the set containing only every X are blue, but no claim
> > is made about whether any additional X is blue.
>
> This is one of the many meanings of "any". Do you think it is the most
> useful? I think that if you change it to "only PA things" then you can
> recover your meaning with {xu'u su'oPA}, and it would be close to what
> I meant by {xe'e} (I think).
Ok, if you promise that "xehe suhore le cukta cu blanu" means
"at least (any) two of the books are blue" then I'll go along
with you. OK: Forget 'xuhu'. I now wholeheartedly accept 'xehe',
and continue to argue for 'xihi'.
---
And
From: Chris Handley <chandley@OTAGO.AC.NZ>
Subject: How do we say this?
Hi all,
A friend posted this to me:
=One of our guys is working on a Test Specification document which
=will be used for the formal testing and acceptance of a software
=project soon to be installed. In the document he says things like
=
= "test B is dependent on test A"
=
=ie test B will only be carried out if test A passes. B could be
=classed as one of A's dependants, but he wants to know what A's relationship
=to B is.
=
=(B is subordinate to A, A is superordinate to B? Mmmmm...doesnt have
=a good ring to it, does it)
=
=PS: Your children are your dependants. You are their what?
=PPS: Father is not the answer I want!
I would have used 'contingent' rather than 'dependent', but no matter.
Point is, can we express the inverse relationship easily in English (I
can't yet), or in Lojban?
Chris Handley
======================================================================
Chris Handley chandley@otago.ac.nz
Dept of Computer Science Ph (+64) 3-479-8499
University of Otago Fax (+64) 3-479-8529
Dunedin, NZ
______________________________________________________________________
"One needs to know a lot more to remain silent than to keep talking"
Fynn, Anna and the Black Knight
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: TECH: "any" & quantification
Jorge cusku di'e:
>> By the way, is "xa'a" as I've defined it the same as asserting existence?
>
>Does {mi viska lo pavyseljirna} assert existance, or the fact that there
>are no unicorns makes the sentence false?
How can "I can look at a unicorn" possibly be true statement, unless I
exist, and at least one unicorn exists? That's why I think useful
transparent statements with the hypothetical "xa'a" have to implicitly
assert existence.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Chris Bogart <cbogart@quetzal.com>
Subject: Thoughts on "any"
There's something special about the arguments of djica, nitcu, claxu, sisku,
and certain commands (ko cpacu lo tanxe), that is akin to a negative. There
seems to be a negation associated with most of these concepts (mi nitcu lo
tanxe at least suggests that mi na ponse lo tanxe; ko cpacu lo tanxe is the
same way)
I think the transparency/opacity distinction paralells the two possible
orderings of the negation and the existence assertion. In the following
examples I've changed the ordering in the "need" sentences as well, which
seems significant since I claim "nitce" is "sorta negative". But obviously
there's a problem with making the order of a certain small set of gismu
(djica, nitce, claxu, sisku, etc.) significant...
[I use xe'e to flag the opaque case, and xa'a to flag the transparent case,
but if you allow for a moment that argument order matters with nitcu, the
flags are unnecessary]
OPAQUE:
mi nitcu [xe'e] lo tanxe
I need a box (box describes my need, not any particular box)
SUGGESTS:
mi na ponse da poi lo tanxe
It is not true that (there exists a box such that I have it)
I don't have any boxes at all
BUT TRANSPARENT:
[xa'a] lo tanxe se nitcu mi
There is a box I need (I have a relationship of "need"
with some box)
SUGGESTS:
da poi lo tanxe na se ponse mi
There exists a box such that (I don't have it)
There is a box I don't have
In other words, the two possible interpretations caused by the combination
of the claim of existence ("da poi tanxe") and the negation ("na") in "mi na
ponse da poi lo tanxe", are disambiguated by word order, either in the main
sentence or out in the prenex. But I think there is something parallel
going on with "nitcu" and "claxu" as well, but we don't have a way of
disambiguating it, since they aren't explicitly negative. But something
like negation is going on, I think.
I don't know what suggestion to make based on that observation, but
hopefully someone else here will see something useful in it...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: general response on needing books
>Maybe my definitions are not very good, but they mainly say that {nitcu}
>and {djica} are forms of {claxu} with more properties for x1.
>Does {claxu} suffer from illicit raising as well?
I think claxu has the same capacity for the transparent/opaque problem that
nitcu does. But one could go further with your definitions and define
"claxu" as "na ponse" or something similar, and (re my previous posting) we
*have* a solution to the transparent/opaque distinction in the case of
negatives (at least when "da" is used: mi na ponse da vs. da na se ponse mi)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: remarks on gismu lexicology
> A heterogeneous assembly of comments & queries
> mostly concerning gismu lexicology.
Very interesting, too.
> (1) x2 of nitcu (need), djica (want) and cpedu
> (request), and x3 of pikci (beg) should be an
> event abstraction. "Need/want/ask/beg to have"
> should be rendered "ponse zei nitcu/djica/cpedu/
> pikci" (with x2 of the lujvo being x2 of ponse).
If we really are consistent, I wouldn't complain.
Let's see if people agree that the x2 of cpedu should
be an event.
> (2) Facki (find, discover) is redundant.
> Facki is equivalent to cilre (learn) (except that
> cilre has an x4 place for method, which is anyway
> not inappropriate for discovering.)
You could say that cilre is facki + assimilation
> (3) Sisku (seek) is redundant and too vague.
And yet so useful. (I mean, it would be useful if the x2
was an object, not a property.)
> "Mi sisku do"
> is either "Mi troci lenu mi { penmi [=encounter] / ponse
> [=possess] / jitro [=control] } do" or "I try to learn where
> you are" (don't know how to say that, but it involves
> "mi troci lenu mi cilre" (-- can someone please remind me how
> Lojban handles subordinate interrogatives, as in "I wonder
> what you're reading"?)).
mi kucli le du'u do tcidu makau
I wonder what you're reading.
mi troci le nu facki le du'u do zvati makau
I'm trying to find out where you are.
> These could be condensed into
> penmi/ponse/jitro zei troci (with x2 of lujvo being x2 of
> penmi/ponse/jitro).
Some redundancy is not necessarily bad. It is good that a word
can be defined in terms of others, otherwise we'll never have
a Lojban-Lojban dictionary.
That said, faktoi (facki troci) is a nice synonym for sisku.
clitoi (cilre troci) is more like tadni (study) than like sisku.
> (4) How to say "I search the pockets"? "mi zukte fe
> le nu catlu le daski kei fi le nu mi penmi/ponse/jitro/kavbu"?
> (Or with lujvo, "mi catlu zei zukte le daski le nu
> mi penmi/ponse/jitro/kavbu".)
Since {sisku} is not going away, sisycta (sisku catlu) may be
more clear than ctazu'e (catlu zukte). I don't see in the last
one where you get a place for the looked for object.
> I assume "catlu" means not "look" but "inspect, examine".
Sounds good. Then {catlu le daski} is enough for "search the pockets"
if you are not searching for something in them.
> (5) How to say "watch, heed, pay attention to"? 'Zgana'
> doesn't seem right.
I'v seen {kurji} used in this sense, but I don't like it.
How do you take care of an event?
> (6) simlu: x1 seems/appears to have property(ies) x2 to
> observer x3 under conditions x4.
> So "I seem blue" is "mi simlu le ka blanu"?
> Then how to say "It seems to be raining, it seems that it
> is raining"? I think we should be able to say "simlu fa
> le duhu carvi" - that is, x1 of simlu is a duhu abstraction
> and x2 is scrapped.
If you change du'u to nu, I agree. Nick mentioned this a short
while ago, too.
>"I seem blue" would be "simlu fa le duhu
> mi blanu".
> simlu: x1 (duhu) seems-to-be-the-case to observer x2
> under conditions x3.
> or, perhaps more usefully:
> simlu: to observer x1 x2 (duhu) seems-to-be-the-case
> under conditions x3.
> (This latter order avoids need for 'fa' to postpose the
> duhu clause, & lends itself as a translation of "it seems
> to me that...".)
No, I prefer the first ordering (with nu instead of du'u).
Otherwise, it turns too much into {jinvi}.
> (7) galfi: x1 (event) modifies/alters/changes x2 into x3
> stika: x1 (event) adjusts/changes x2 (ka/ni) in amount/degree x3
> I think the x1 place of these should be abolished.
> Galfi then becomes redundant with binxo:
> binxo: x1 becomes/changes into x2 under conditions x3
Then why change it?
> And I think binxo should have an extra place:
> binxo: x1 changes from belonging to category (ka) x2
> into belonging to category (ka) x3
The more lojbanic order would be into x2 from x3.
> I think some new but related meaning should be found for galfi,
> such as:
> galfi: x1 evolves from (ka) x2 into (ka) x3 under conditions/
> constraints x4 [e.g. natural selection]
That's {farvi}, easy to remember because it rhymes with Darwin :)
> (8) panci: x1 is an odor/fragrance/scent/smell emitted by x2
> and detected by observer/sensor x3
> sumne: x1 (experiencer) smells [transitive verb] x2;
> x2 smells/has odor to observer x1
> ganse: x1 [observer] senses/detects/notices/is aware of
> stimulus x2 by means x3 under conditions x4
> vrusi: x1 is a taste/flavor of x2
> (a) Given sumne, why does panci have this x3 place? Suppose
> I want to describe the smell of an unsmelt rose.
I agree. An odo(u)r needs a smeller as much as a colo(u)r needs a seer.
> (b) Why does sumne lack a place for the odour?
For the same reason that {viska} doesn't have a place for the colo(u)r
of the seen object, I guess.
> (c) Why does 'vrusi' have no 'transitive' counterpart? I suppose
> we could have:
> vrusi zei ganse: x1 tastes taste x2 of x3
> But in this case, why bother with having sumne?
Good question. I would add it, if it were possible.
> (d) Unless I've misunderstood, I suggest dropping the x3 of panci,
> and dropping sumne altogether, using panci zei ganse instead.
I agree with the first, but I don't think anything can be just dropped
altogether at this point.
> (9) Is there an agreed expression for look/appearance/
> countance/visual stimulus, without there being an implied
> perceiver?
I'd say {jvinu}.
Let's see: jvinu sance panci vrusi tengu
viska tirna sumne ????? pencu
Definitely, there's something missing.
> (10) tirna: x1 hears x2 against background/noise x3
> Could x3 be abolished, please? Otherwise, when there's
> no background noise we'll have to remember to use "xohe"
> (or whatever the sumti-abolishing cmavo is).
Hear, hear!
> (11) Is there a standard expression for 'saliva'?
The list gives molselpu'u, but I guess you don't like it. I can't
think of anything better.
> ---
> And
>
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: general response on needing books
> Claxu doesn't suffer from illicit raising, but it does lead to scope
> ambiguities.
> mi claxu lo cukta
> perhaps can mean either:
> Ex book(x) & not have(me,x) "There's a book that I lack"
> not Ex book(x) & have(me,x) "I don't have any books"
> Hopefully this second is not a possible meaning of the Lojban
> sentence.
Yes. But be prepared to see it often used like that.
> Given your not unreasonable interpretation of nitcu and djica as
> being kinds of claxu, you are quite right that they don't involve
> sumti raising. However I strongly urge that your interpretations
> should be rejected. First, in English I can say "I want/need this
> book" even if I already have it - I mean that I prefer having the
> book over not having it, & am not asserting that I don't have it.
> So for English, wanting/needing doesn't entail lacking.
> Second, if the x2 of nitcu/djica is an event, we can translate
> "I want/need you to go", which mean, roughly, I prefer your going
> over your not going. Third, "want/need to have" can then be
> rendered "ponse zei djica/nitcu". Fourth, wanting/needing +
> lacking can be rendered "djica/nitcu zei claxu".
Ok, you've convinced me. It should be an event, then.
> > > > I think you can't have specific/opaque.)
> > > To evaluate the truth of "Le cukta cu blanu" you identify
> > > the referent of "le cukta" and then check whether it's blue.
> >
> > Exactly.
> >
> > > "There is a certain book that I need to have":
> >
> > That "a certain book" sounds non-specific to me.
>
> I am not trying to translate the English sentence; the English
> sentence is merely an attempt to indicate the meaning I'm trying
> to describe. I can't use 'that/this/my', because reference of
> these is assigned at sentence level, not only in an inner bridi.
> There is no way I can think of to say in English what I want to say.
>
> If you will agree to define specificity as I did above (you do
> say "exactly"),
You shouldn't take my word so literally :)
Besides, I didn't think that was a definition of specificity. Just a
way of understanding how to evaluate the truth value of a sentence
containing a specific reference.
> then would you agree that there is a difference
> in meaning according to whether you have to identify the referent
> in the local bridi or in the outermost bridi.
I don't believe there is anything like "local specificity", or
specificity inside an abstraction. Specificity concerns the speaker
and the audience, and the only way you have more than one level of
these is with quotations.
> My point is precisely
> that there is a difference, & one worth making expressible in
> Lojban.
Then please describe a situation where such a use would make sense,
and write a Lojban sentence using your proposed xi'i with {le}.
> > You mean that for {lo} the xihi-less quantification would be outside
> > the abstraction? I think that goes against current usage.
>
> This is what I mean. I suppose it could be the other way around,
> but this would be needed less often.
I would have thought that local quantification was the norm, but that's
only an impresion.
> Alternatively there could
> be 2 cmavo, one for 'outermost' and one for 'local', with it
> understood that if neither is used there is a potential
> ambiguity.
I think I don't like it, but I have to think more about it.
> > > "xuhu" - "xuhu PA X cu blanu" indicates that PA things selected randomly
> > > from the set containing only every X are blue, but no claim
> > > is made about whether any additional X is blue.
> >
> > This is one of the many meanings of "any". Do you think it is the most
> > useful? I think that if you change it to "only PA things" then you can
> > recover your meaning with {xu'u su'oPA}, and it would be close to what
> > I meant by {xe'e} (I think).
>
> Ok, if you promise that "xehe suhore le cukta cu blanu" means
> "at least (any) two of the books are blue" then I'll go along
> with you.
I promise.
> OK: Forget 'xuhu'. I now wholeheartedly accept 'xehe',
> and continue to argue for 'xihi'.
>
> ---
> And
>
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Transparency / Opaqueness
la veion cusku di'e
> (a) I'll read two books. [any two]
> (b) I'll read two books, this and that.
>
> ((a) is the opaque case and (b) the transparent one)
Comment: Future tense is tricky for this. The truth value is
timeless, and therefore to evaluate it you have the advantage
of seeing all time at once. Even though (a) looks opaque, it
really isn't, you just have to look in the future and see if
the claim holds for two books, that are not "any" from this
timeless perspective.
The truly opaque "I read any two books" (tense is irrelevant) is
mindboggling.
> Now I'm wondering what would happen, if we said in
> general that {le} is used for transparency and {lo}
> for opaqueness.
It doesn't work. For most predicates, the "opaque" claim is
pretty useless.
> I think that even generally it
> might be useful to define an outer quantifier of the {le}
> descriptor to include an elidable {lo ro}.
It's already like that.
Notice that {pa le cukta} is non-specific. You can't have a
specific reference with any quantification but ro.
Thus {le pa le cukta} is specific again.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Transparency / Opaqueness
And responding to Veijo:
> It has been established (to my satisfaction, at any rate)
> that LE/LO is +/-specific [Colin propounded this most lucidly].
> It only relates to definiteness in
> that only +specifics can be +/-definite.
Could you explain what is definiteness in this context, please.
> I think that we do need a new cmavo & that LO/LE isn't
> the same as transparent/opaque. You seem to miss the ambiguity
> of (a).
> I'll read any two books. - pick two items freely from the
> set of all books, & it is asserted that I'll read them.
Which is a pretty nonsensical claim. And verifiably false: just
wait until you're dead, and then it will be obvious that picking
any two books you would most likely not have read them, and
therefore the claim that in the future you would was false.
(Unless truth values involve somehow your intent at the time?)
> There are two books I'll read. - examine every book & if
> you find at least two that I'll read, the assertion is
> true.
A transparent claim, and the one {mi ba tcidu re cukta} means.
(The examination of all the books need not be done
contemporaneously with the uttering of the claim.)
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Thoughts on "any"
Chris Bogart:
> There's something special about the arguments of djica, nitcu, claxu, sisku,
> and certain commands (ko cpacu lo tanxe), that is akin to a negative. There
> seems to be a negation associated with most of these concepts (mi nitcu lo
> tanxe at least suggests that mi na ponse lo tanxe; ko cpacu lo tanxe is the
> same way)
Be careful with negation in Lojban, it is not always what it seems.
mi na ponse da poi tanxe
da poi tanxe na se ponse mi
Both mean the same thing:
It is false that: there exists a box such that I have it.
To get:
> There exists a box such that (I don't have it)
> There is a box I don't have
You can say:
da poi tanxe naku se ponse mi
which means the same as
roda poi tanxe na se ponse mi
It is false that I have every box.
> In other words, the two possible interpretations caused by the combination
> of the claim of existence ("da poi tanxe") and the negation ("na") in "mi na
> ponse da poi lo tanxe", are disambiguated by word order, either in the main
> sentence or out in the prenex.
Partly true, but {na} always negates the whole claim, unless you use {naku},
for which order becomes important.
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: TECH: Transparence / Opaqueness
la xorxes. cusku di'e
> Date: Thu, 22 Sep 1994 22:02:18 EDT
> From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
> Subject: Re: TECH: Transparency / Opaqueness
>
> la veion cusku di'e
>
>> (a) I'll read two books. [any two]
>> (b) I'll read two books, this and that.
>>
>> ((a) is the opaque case and (b) the transparent one)
>
> Comment: Future tense is tricky for this. The truth value is
> timeless, and therefore to evaluate it you have the advantage
> of seeing all time at once. Even though (a) looks opaque, it
> really isn't, you just have to look in the future and see if
> the claim holds for two books, that are not "any" from this
> timeless perspective.
>
> The truly opaque "I read any two books" (tense is irrelevant) is
> mindboggling.
>
>> Now I'm wondering what would happen, if we said in
>> general that {le} is used for transparency and {lo}
>> for opaqueness.
>
> It doesn't work. For most predicates, the "opaque" claim is
> pretty useless.
I disagree.
If we consider those types of sentences where we have had
difficulties with quantification/transparency/opaqueness,
i.e. those involving {nitcu} & Co., and sentences like
mi pu'o catra ci le xa cinfo
I'm going to kill 3 of the 6 lions
ko dunda ci plise mi
Give me 3 apples!
we ought to observe that they have one feature in common:
the outcome of the sampling is not known at
the moment of utterance and we can draw only
a limited set of conclusions from the facts
given
We have a kind of linguistic Schroedinger's Cat which is
simultaneously alive and dead until the outcome of the
experiment is observed - until then we have opaqueness.
The basic Lojban sentence is timeless (even with the
simple tenses pu/ca/ba), and if I say
mi citka re le pano plise
I eat two of the ten apples
the {re le pano plise} are in limbo until I can say
mi ba'o citka le re le pano plise
at which time the two are fixed, definite, transparent -
until then there is just a kind of Poisson distribution
of the possible outcome of the sampling, things are not
sharply in focus but behind an opaque window.
Even though I have had but a single mother, I can say
re lo remna cu mamta mi
given the definition of {mamta} and the nature of the
Lojban tense system. Until I'm dead, and it can be
definitely said that no second remna turned out to
mother me, this statement holds opaquely.
Lojban aspires to be a logical language, and just for
that reason we must recognize the grey areas where we
cannot build on the premises of binary logic - or even
multivalued logic. The physicists recognize the opaque/
probabilistic nature of many basic phenomena, why
shouldn't we. Lojban as it stands has the means of
making the dual nature of reality explicit - without
losing logicalness. (Opaqueness isn't too bad - it
takes much more to stomach the tenses :-)
> Jorge
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: TECH: Transparency / Opaqueness
In-Reply-To: (Your message of Thu, 22 Sep 94 22:18:59 EDT.)
Jorge:
> And responding to Veijo:
> > It has been established (to my satisfaction, at any rate)
> > that LE/LO is +/-specific [Colin propounded this most lucidly].
> > It only relates to definiteness in
> > that only +specifics can be +/-definite.
>
> Could you explain what is definiteness in this context, please.
Oversimplifying a bit, it means that the addressee is able to
identify the referent (without asking 'which?').
Compare:
I bought a book.
Which book?
- normal
I bought the book.
Which book?
- which is not normal, & implies a failure in communication.
> > I think that we do need a new cmavo & that LO/LE isn't
> > the same as transparent/opaque. You seem to miss the ambiguity
> > of (a).
> > I'll read any two books. - pick two items freely from the
> > set of all books, & it is asserted that I'll read them.
>
> Which is a pretty nonsensical claim. And verifiably false: just
> wait until you're dead, and then it will be obvious that picking
> any two books you would most likely not have read them, and
> therefore the claim that in the future you would was false.
> (Unless truth values involve somehow your intent at the time?)
It is sensical, but quite false, as you say. The English means
"pick any two books & I am willing to read them". Then it becomes
true (potentially).
This is the case where one would use "xehe".
> > There are two books I'll read. - examine every book & if
> > you find at least two that I'll read, the assertion is
> > true.
>
> A transparent claim, and the one {mi ba tcidu re cukta} means.
Right. ns.
Right.> Which is a pretty nonsensical claim. And verifiably false: just
> wait until you're dead, and then it will be obvious that picking
> any two books you would most likely not have read them, and
> therefore the claim that in the future you would was false.
> (Unless truth values involve somehow your intent at the time?)
It is sensical, but quite false, as you say. The English means
"pick any two books & I am willing to read them". Then it becomes
true (potentially).
This is the case where one would use "xehe".
> > There are two books I'll read. - examine every book & if
> > you find at least two that I'll read, the assertion is
> > true.
>
> A transparent claim, and the one {mi ba tcidu re cukta} means.
Right.
---
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: TECH: Transparence / Opaqueness
In-Reply-To: (Your message of Fri, 23 Sep 94 19:06:49 EDT.)
Xorxes to Veion:
> > ko dunda ci plise mi
> > Give me 3 apples!
> Imperatives by definition are neither true nor false in Lojban. That
> means: make {do dunda ci plise mi} true. Since the distinction between
> opaque and transparent rests on how the truth value of the statement
> is determined, no such distinction is possible for imperatives.
But contrast:
I hereby command you to make it true that there are 3 apples that
I'm given
There are three apples such that I hereby command you to give
me them
The transparent/opque distinction (I suggest, out of my depth) has
to do with scope of quantification with resepect to some 'irrealis'
element of meaning, & a command, even in the form of an imperative,
constitutes an irrealis element. (CF. "Give me a book - any book".)
----
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: query about 'ka'
I find myself rather confused about 'ka'. If we look at duhu or nu,
we find that the following pair mean the same:
le nu/duhu mi nelci do
le nu/duhu do se nelci me
i.e. the event of me liking you/you being liked by me, or the
proposition that I like you or that you are liked by me.
But it seems to me that "the property of me liking you" and
"the property of you being liked by me" don't make sense,
and instead I would want to express these different meanings:
the property of liking you (the property of being the liker of you)
the property of being liked by me
But to express these, it seems that 'ka' would have to mark
a sumti rather than a bridi, as in:
le ka nelci do
le ka se nelci mi
or
le ka le nelci do kei
le ka le se nelci mi kei
That is, to me it seems that whereas nu/duhu express aspects of
a bridi, ka expresses an aspect of a sumti.
Could someone please explain...
----
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: general response on needing books
And:
> Le prenu cu jinvi le duhu xihile cukta cu blanu
>
> The person thinks "the book is blue".
I'd translate that English sentence as:
le prenu cu jinvi la'e lu le cukta cu blanu li'u
> To find out which book, you'd
> have to ask the person, not me the speaker.
Why the person? Is there something special about the x1 position?
How about:
le ninmu le nanmu cu tugni le du'u xi'ile cukta cu blanu
Who'd you ask, the woman, the man, both?
> The referent of xihile
> cukta is 'in mind' but the mind is the person's, not the speaker's.
Only if within quotations. Out of quotations, there's no connection
between the specificity of {le} and the mind of any "character"
appearing in the sentence.
Besides, linking the specificity of {le} with the quantification
of {lo} seems to me that is a big part of the confusion we are having.
> You need local quantification or reference-assignment only when it
> is in a clause subordinate to an irrealis element of meaning (i.e.
> something whose argument is not necessarily the case).
I disagree. The quantification applies to realis elements as well.
(The reference assignment is a different issue, which I don't think
has anything to do with locality.)
For example:
mi djuno le du'u lo cukta cu blanu
I know that there is a book that is blue.
da poi cukta zo'u mi djuno le du'u da blanu
There is a book such that I know it is blue.
Different claims, both with realis subordinate clauses.
(In the second one I have to know which book, in the first one I
may or may not know.)
> And in most
> utterances there won't be such an element. However, when there is
> such an element you may indeed be right that local rather than
> outermost quantification is more often what is wanted.
The question then is whether the need to use outside quantification
is significant enough to warrant the introduction of xi'i. You can
always be explicit using a prenex, is there really a need for the
more compact form with xi'i? I'm not sure.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: remarks on gismu lexicology
And:
> In this case I guess simlu would mean "x1 seems-to-be-actual/real",
> in which case there;s no need to restrict it to event abstractions.
> For example "Lo cukta cu simlu mi" would mean "There seems to me
> to be a book". I'd be happy with that.
Well, "there is a book that seems to be real to me". Let's leave it
as an event before all the worms start crawling out.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Transparency / Opaqueness
And:
> > > It has been established (to my satisfaction, at any rate)
> > > that LE/LO is +/-specific [Colin propounded this most lucidly].
> > > It only relates to definiteness in
> > > that only +specifics can be +/-definite.
> >
> > Could you explain what is definiteness in this context, please.
>
> Oversimplifying a bit, it means that the addressee is able to
> identify the referent (without asking 'which?').
> Compare:
> I bought a book.
> Which book?
> - normal
> I bought the book.
> Which book?
> - which is not normal, & implies a failure in communication.
Great!
I would add that -specific (lo) is always -definite, and that
"which one?" always makes sense in such case (which is not to say
that the speaker has to know the answer).
So the meaning of "ko'a nitcu lo tanxe" is the one for which the
question "which one?" makes sense. (Unless nitcu accepts
abstractions only, in which case the problem doesn't arise.)
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Transparence / Opaqueness
And:
> > > ko dunda ci plise mi
> > > Give me 3 apples!
>
> But contrast:
> I hereby command you to make it true that there are 3 apples that
> I'm given
> There are three apples such that I hereby command you to give
> me them
>
> The transparent/opque distinction (I suggest, out of my depth) has
> to do with scope of quantification with resepect to some 'irrealis'
> element of meaning, & a command, even in the form of an imperative,
> constitutes an irrealis element. (CF. "Give me a book - any book".)
You're right, but then the definition of {ko} can't be "make this
sentence, replacing {ko} by {do}, to be true". At least in the presence
of {xe'e} the meaning of {ko} would be somewhat different.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: query about 'ka'
> le ka nelci do
> le ka se nelci mi
> That is, to me it seems that whereas nu/duhu express aspects of
> a bridi, ka expresses an aspect of a sumti.
>
> Could someone please explain...
>
> ----
> And
I can't explain, but there was some discussion during Logfest on
this or a very related topic. One proposed solution was to have
a cmavo in selmaho KOhA (say xa'e) so that you'd have {le ka
xa'e nelci do} or {le ka mi nelci xa'e} for the two meanings you
want. (Something to do with lambda variables, I believe.)
Jorge
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: specificity and metonymy
In-Reply-To: (Your message of Sat, 24 Sep 94 09:56:57 D.)
Bob C:
> > It has been established ... that LE/LO is +/-specific
> This is somewhat misleading. {le} may well be *less* specific to a
> *listener* than {lo}.
I meant 'specific' in the technical sense, which (I say clumsily
in my ignorance) is that you don't find the referent by
quantification; the referent is a constant.
I do agree that metalanguage taken from everyday language can
be misleading. Would anyone care to suggest a lojban term for
specificity (of reference)? That would help.
On the subject of metonymy: unless this is built in to the grammar,
it is a matter of pragmatics, and if it is a matter of pragmatics
it isn't, strictly speaking, pertinent to a debate on the semantic
component of the grammar.
---
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: query about 'ka'
In-Reply-To: (Your message of Sat, 24 Sep 94 14:36:01 EDT.)
Jorge:
> I can't explain, but there was some discussion during Logfest on
> this or a very related topic. One proposed solution was to have
> a cmavo in selmaho KOhA (say xa'e) so that you'd have {le ka
> xa'e nelci do} or {le ka mi nelci xa'e} for the two meanings you
> want.
This indeed neatly solves the problem. Is it official?
> (Something to do with lambda variables, I believe.)
There are basically two types of people: geniuses, who understand
lambda variables, and the rest of us. (It has been pointed out
to me that there is a secondary dichotomy concerning whether
one knows which way round to write one's lambdas.) I have long
aspired to join the former class, but have always been thwarted
in my attempts by a brain that works like an 8086 PC in these
areas.
----
And
From: "Robert J. Chassell" <bob@GNU.AI.MIT.EDU>
Subject: Re: metonymy
In-Reply-To: <9409241909.AA17817@albert.gnu.ai.mit.edu> (message from ucleaar
on Sat, 24 Sep 1994 20:07:02 +0100)
On the subject of metonymy: unless this is built in to the grammar,
it is a matter of pragmatics, and if it is a matter of pragmatics
it isn't, strictly speaking, pertinent to a debate on the semantic
component of the grammar.
Depends what you mean by `built in'. In as much as Lojban has {noi}
and {poi} to specify referents to sumti, most of what you say most of
the time in Lojban is a short form for what you could say more
specifically in some grammatically correct longer form. It is the
same in a natural language such as English, except the minimal
specifiers are different.
English uses a bunch of terms, `a', `the', `some', `any', plus plural
and singular, as its minimal specifiers.
I need a box.
I need the box.
Lojban also uses a bunch of such terms, but their meanings are
different from English. The categories for short forms are different.
Most peculiarly, Lojban has a term indicating truth, `that which
really is', as well as one meaning `that which I designate'.
mi nitcu lo tanxe
mi nitcu le tanxe
In neither English nor Lojban do the terms *necessarily* specify the
box, although they may.
The presumption in English is that a listener is helped if he or she
is told whether the needed box is `the' box, or `a' box.
The presumption in Lojban is that a listener is helped if he, she or it
is told whether the needed box is `for-real', or something else that I
might be designating as a box in my mind, or a mathematical set of
boxes.
The minimal specifiers of the two languages are different. We often
translate `lo' into English `a', and `le' into `the', but such
translations are not very accurate. Better to say `that which really
is' and `that which I designate as' for `lo' and `le'.
A future research topic might be: does a fluent thinker in Lojban
unconsciously find more things that `might be designated as boxes, but
are not really boxes' than a person who is fluent only in English?
--
Robert J. Chassell bob@gnu.ai.mit.edu
25 Rattlesnake Mountain Road bob@grackle.stockbridge.ma.us
Stockbridge, MA 01262-0693 USA (413) 298-4725
Date: Sun, 25 Sep 1994 04:21:59 -0400
From: Logical Language Group <lojbab>
Subject: Re: query about 'ka'
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
"ka" refers to the properties of a bridi relationship which make the
relationship true. In some cases, it has been recognized that we want to
refer to the subset of properties of a relationship that apply specifically
to one element of that relationship, e.g.
the properties of you such that you are liked by me, as distinguished from
the properties of me such that I find you likable. But both are among
the properties of "me liking you", and there are some properties that pertain
to the relationship ("a lot" in the statement "I like you a lot" indicates
a quantity of something, which is ka mi nelci do).
lojbab
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: any
I'm leaving the transparency/opaqueness terminology as I
realized (after some thinking) that I wasn't using the
terms in Quine's sense.
I have tried to digest a reasonable number of the
messages which have been exchanged around this problem
(any, transparency/opaqueness, metonomy) and now I try
to recapitulate (very concicely).
Everything started out as Jorge thought he found a
discrepancy between
(1) mi ponse pa tanxe
(2) mi nitcu pa tanxe
in that a Lojban bridi asserts a relationship between
a number of entities (the sumti), in this case apparently
between {me} and {pa tanxe}. (1) seemed to be OK, but
there seemed to be something wrong with (2) in the case
the statement was opaque in the Quinean sense, i.e. I
needed just some box (I'm avoiding the word 'any' as it
leeds to unnecessary problems).
It is possible to express (2) in a long form which solves
the problem but - in the name of metonomy - it would be
desirable to be able to avoid this. Jorge suggested a new
cmavo {xe'e} to solve the problem but I'm not sure it
would actually solve the problem with the relationships.
I tried out several possible models to solve the problem
but got the feeling that there was always something wrong.
Now, I started thinking, could it be that no differing
structure would be necessary, after all?
There is a relationship which is correctly expressed by
both (1) and (2), even if (1) is apparently transparent
and (2) apparently opaque. Do you see it? It is the
relationship between {mi} and {pa}, the number of boxes,
I either have one or need one.
When we have an external quantifier we are not so much
concerned about the identity/specificity as the number of
entities. Perhaps we could sidestep the whole issue by
defining that an external quantifier is a type of
combined quantity abstraction descriptor. With this
definition
(3) mi kalte lo xanto
would be an opaque claim in the Quinean sense (given the
implicit external quantifier {su'o}). The transparent
case (which involves a specific elephant(s)) would be
(4) mi kalte le xanto
The implicit external quantifier {ro} makes the transparency.
-----
I'm done with this thread.
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
Date: Sun, 25 Sep 1994 12:05:00 -0400
From: Logical Language Group <lojbab>
Subject: Re: Thoughts on "any"
Subject: Thoughts on "any"
Chris Bogart <cbogart@quetzal.com>:
>There's something special about the arguments of djica, nitcu, claxu,
>sisku, and certain commands (ko cpacu lo tanxe), that is akin to a
>negative. There seems to be a negation associated with most of these
>concepts (mi nitcu lo tanxe at least suggests that mi na ponse lo tanxe;
>ko cpacu lo tanxe is the same way)
The commonalty of all of these are that they are "intensional" - the
thing wanted/needed/lacked/sought/acquired exists in the world of
intension - i.e. the mind - which need not be the same as the real
world. There are several other words that are, or could be intensional.
The implicit negation you are seeing is the acknowledgement built into
many of these words of the possible non-reality of the object.
It is possible that any time that you have an intensional, or an
irrealis condition, that there is an implicit need to introduce a more
limited scope for any quantifiers within that intensional sumti. pc has
said in private email (if I understand him) that many non-abstract
values in such places are inherently sumti-raising because of the need
for a different level of 'universe' with its own quantification scope.
(I'll let him explain this if it isn't obvious, since I am anything but
qualified to talk clearly on the subject.
I brought in irrealis into the last paragraph, because it hasn't thus
far been mentioned in this discussion. Irrealis mode seems to me to be
what we are talking about whenever we aren't sure that something exists,
asnd there may be some solution to the problems of quantification in
'mental world' sumti by using the marker for irrealis mode (Colin Fine
has argued this as the general meaning of "da'i" in UI).
I don't pretend that this answers all issues that have been raised in
this discussion (which has gone over my head), but maybe it will help on
some of them. Nora is reading the discussions and may eventually
comment, but she isn't yet managing to keep up with the flow of
messages, much less getting caught up to the point where she feels ready
to contribute. I believe that pc is roughly in the same position, and
perhaps John Cowan as well. The intensity of the discussion is good,
and perhaps a new record for this mailing list especially in that it is
involving people from over a half-dozen countries, and is at the deepest
logical levels of the language. So in no way to I want to urge a halt
to the discussion, while warning that pronouncements from "on-high" will
not be coming in the short term, especially while the discussion continues
at this level.
My compliments to all of you for your continued contributions.
lojbab
Date: Sun, 25 Sep 1994 12:05:48 -0400
From: Logical Language Group <lojbab>
Subject: Re; And/Jorge on gismu lexicon
And Rosta >>>
Jorge Llambias >> (only quoted where I respond to his posts)
And again >
Summary: one BIG open issue that probably needs a non-place structure
solution, 1 new lujvo proposal, and a couple of English synonyms to be
added to the dictionary. A lot of words to explain why (and more often:
why not).
My tone is perfunctory, but this is not intended to be critical or
condescending. I think it good that people pose questions and
criticisms about word meanings, as long as they are accepting of the
fact that the window of opportunity to change definitions of words in
other-than-clarificational ways is rapidly closing.
>>>(1) x2 of nitcu (need), djica (want) and cpedu (request), and x3 of
>>>pikci (beg) should be an event abstraction. "Need/want/ask/beg to have"
>>>should be rendered "ponse zei nitcu/djica/cpedu/ pikci" (with x2 of the
>>>lujvo being x2 of ponse).
I think this is part of the ongoing discussion. Unfortunately this is
not likely to be a complete list. See my other posting regarding
intensionality. I am hoping no place structure changes are needed (i.e
that we can solve it discursively or even pragmatically), since (as I
show below) the list of possible intensional predicates ranges into a
lot of weird areas.
>>>(2) Facki (find, discover) is redundant. Facki is equivalent to cilre
>>>(learn) (except that cilre has an x4 place for method, which is anyway
>>>not inappropriate for discovering.)
Even if I agreed with the equivalence, we do not delete gismu for mere
redundancy. In this case as well as others you pose, the redundancy is
from a gismu to a lujvo in all probability. This means that all lujvo
that might be built out of facki would take an extra term, making lujvo
longer. Several gismu are in the list primarily because it is believed
that they will be useful in making Zipfeanly short lujvo.
The current lujvo list has 7 lujvo for facki, all in final term
position, and 3 for cilre in various positions.
>>>(3) Sisku (seek) is redundant and too vague. (x2 of sisku is in my list
>>>object/event/property. I believe that this has been changed to just
>>>property, but either way I don't understand why.) "Mi sisku do" is
>>>either "Mi troci lenu mi { penmi [=encounter] / ponse [=possess] / jitro
>>>[=control] } do" or "I try to learn where you are" (don't know how to
>>>say that, but it involves "mi troci lenu mi cilre" (-- can someone
>>>please remind me how Lojban handles subordinate interrogatives, as in "I
>>>wonder what you're reading"?)). These could be condensed into
>>>penmi/ponse/jitro zei troci (with x2 of lujvo being x2 of
>>>penmi/ponse/jitro).
See above re redundancy, probably usefulness in lujvo (although sisku is
only in 1 lujvo so far in the current list).
>>That said, faktoi (facki troci) is a nice synonym for sisku.
Only if we solve the intensional places problem per And's (1). troci and
sisku relate to in-mind things, while you cannot facki unless the se
facki is 'real'. (Meanwhile it is possible to cilre a falsehood, so it
also might be intensional/in mind, along with things that are se ctuca)
You can bring our "any" discussion into these places and raise spectres
of existence/quantification problems with "I am teaching about
unicorns", "I am learning (anything/anything I can) about unicorns". "I
dreamt about unicorns", and "I remember seeing a unicorn (our memories
are not necessarily reality). It becomes easy to find places where "da
poi" quantification is a problem, and in some of these, the possibly
sumti-raised "da poi" is an abstraction, and eliminating sumti-raising
might force us to explicit 2nd-order abstractions:
mi cilre ledu'u da zo'u da ka de zo'u de mela .iunikorn.
I learn something that is a property of some unicorn.
I learn anything about unicorns.
Yeccch!!!
>>>(4) How to say "I search the pockets"? "mi zukte fe le nu catlu le
>>>daski kei fi le nu mi penmi/ponse/jitro/kavbu"? (Or with lujvo, "mi
>>>catlu zei zukte le daski le nu mi penmi/ponse/jitro/kavbu".) I assume
>>>"catlu" means not "look" but "inspect, examine".
The gismu list gives several synonyms, including those. Certainly catlu
is not limited to visual examination, but then neither is English
"look".
>>Since {sisku} is not going away, sisycta (sisku catlu) may be more clear
>>than ctazu'e (catlu zukte). I don't see in the last one where you get a
>>place for the looked for object.
>The looked for object would either be a "lenu ponse/cpacu" as
>x3 analogous to x3 of zukte, or it could be a stipulated x3
>for the lujvo.
It is possible to search pockets without having a specific (or general)
goal in mind for the object/property you are looking for. In fact, you
can search a pocket without expecting to find anything: "I search the
pocket to see if it is empty", which actually goes well with the current
'property' place structure.
"Stipulated x3" sounds like you are saying 'pull it out of thin air'.
In the era of Nick's analytical lujvo-making, this will be unlikely to
fly. (And it certainly won't vofli %^)
>>>(5) How to say "watch, heed, pay attention to"? 'Zgana' doesn't seem
>>>right.
If you grep the gismu list for "attention" you get exactly 1 word, which
is "jundi", which is probably the word you want. "heed" is not in the
list and I will add it as a synonym for tinbe as well as jundi. "watch"
greps in catlu and zgana - the former implying that the watching is
intentional.
>>I'v seen {kurji} used in this sense, but I don't like it. How do you
>>take care of an event?
I hope not from me %^) jundi works for events as well as objects
>I don't like it either.
Me too! Me too!
>>>(6) simlu: x1 seems/appears to have property(ies) x2 to observer x3
>>>under conditions x4. So "I seem blue" is "mi simlu le ka blanu"? Then
>>>how to say "It seems to be raining, it seems that it is raining"? I
>>>think we should be able to say "simlu fa le duhu carvi" - that is, x1 of
>>>simlu is a duhu abstraction and x2 is scrapped. "I seem blue" would be
>>>"simlu fa le duhu mi blanu".
>>> simlu: x1 (duhu) seems-to-be-the-case to observer x2
>>> under conditions x3.
>>>or, perhaps more usefully:
>>> simlu: to observer x1 x2 (duhu) seems-to-be-the-case
>>> under conditions x3.
>>>
>>>(This latter order avoids need for 'fa' to postpose the duhu clause, &
>>>lends itself as a translation of "it seems to me that...".)
It seems to be raining.
leka carvi cu se simsa
I seem blue
leka mi blanu cu se simsa
It is possible that x1 is redundant if we hammer down the method of
focussing on one sumti in a ka abstraction, which is an open issue. It
will still not be deleted, since simlu is used in 8 lujvo, all of which
use simlu1, and not all of which use simlu2. Objects tend to be what we
focus on in seemings, whereas what they seem like is often what the
other part of the lujvo is trying to talk about.
>> If you change du'u to nu, I agree. Nick mentioned this a short
>> while ago, too.
>
>In this case I guess simlu would mean "x1 seems-to-be-actual/real",
>in which case there;s no need to restrict it to event abstractions.
>For example "Lo cukta cu simlu mi" would mean "There seems to me
>to be a book". I'd be happy with that.
Intensionality again. There seems to me to be "da poi mela iunikorn"
fails if there are no unicorns. I think this one was changed as part of
the "sisku" change, for the same reason, except that it is useful to
have the object extractable from the ka abstraction to the main bridi
level for pragmatic reasons that are somewhat weaker for sisku.
>>>(7) galfi: x1 (event) modifies/alters/changes x2 into x3
>>> stika: x1 (event) adjusts/changes x2 (ka/ni) in amount/degree x3
>>>I think the x1 place of these should be abolished.
>>>Galfi then becomes redundant with binxo:
>>> binxo: x1 becomes/changes into x2 under conditions x3
>>>And I think binxo should have an extra place:
>>> binxo: x1 changes from belonging to category (ka) x2
>>> into belonging to category (ka) x3
Abolish a place so that the gismu can be deleted??? Whatever for? If
you don't ever need/use the place of galfi/stika, just use binxo/cenba
instead. But you either will need the former two, or will just recoin a
lujvo from binxo-rinka and/or cenba-rinka for the times when such is
useful. (I will admit that these seemed more useful before sumti
raising forced these into event clause places, thus requiring a
gasnu/zukte instead/in addition to give the Lojban for "The witch
changed the prince into a frog." and "The operator adjusted the speed of
the engine."
Again, Zipf demands shorter lujvo when possible, even when this causes
'redundancy'. 14 galfi lujvo, 29 binxo lujvo, 7 cenba lujvo, 5 stika
lujvo thus totalling 2% of our lujvo base (it was even higher in TLI
Loglan, and probably eventually will be for us).
>>>I think some new but related meaning should be found for galfi,
>>>such as:
>>> galfi: x1 evolves from (ka) x2 into (ka) x3 under conditions/
>>> constraints x4 [e.g. natural selection]
too much like farvi, and having nothing to do with agentive modification
unless you mean that your x4 is the current x1 of galfi (which might
actually be true - in which case your complaint about the current x1 is
really strange since you see it useful and meaningful when placed in a
different order and reworded slightly.
>>>(8) panci: x1 is an odor/fragrance/scent/smell emitted by x2
>>> and detected by observer/sensor x3
>>> sumne: x1 (experiencer) smells [transitive verb] x2;
>>> x2 smells/has odor to observer x1
>>> ganse: x1 [observer] senses/detects/notices/is aware of
>>> stimulus x2 by means x3 under conditions x4
>>> vrusi: x1 is a taste/flavor of x2
>>>(a) Given sumne, why does panci have this x3 place? Suppose
>>>I want to describe the smell of an unsmelt rose.
Like colors and flavors, there is considerable evidence of differences
among observers as to what is sensed by a smeller.
An odor that is unsensed is not a panci, and 'odor' to me implicitly
implies a detector. In fact, I am not sure what it is if there is no
(potential) detector.
vrusi is not necessarily limited to things sensed by the tongue (no
sensor place), and hence could be used to describe the attar of a rose.
I think we did this because of considerable scientific evidence that the
senses of taste and smell are no independent.
>>>(b) Why does sumne lack a place for the odour?
Because it would then be redundant to panci, wouldn't it?
People talk about smelling things, don't they?
>>>(c) Why does 'vrusi' have no 'transitive' counterpart? I suppose
>>>we could have:
>>> vrusi zei ganse: x1 tastes taste x2 of x3
Because it doesn't. "vrusi ganse" is sufficient for those cases where
you want to talk about the detector (though as I note above, vrusi
doesn't necessarily omit smell-related properties). I can see the
parallel to panci/sunme, and indeed may have argued for a parallel gismu
at one time, but was outvoted by the gismu minimalists. There are
useful lujvo for vrusi (3 currently) and panci (1), but none yet for
sunme, and I can't think that anything based on vrusi would be more
useful based on an observer/detector place. Generally, once we get past
the basic existence of the sensory apparatus, we tend to talk about
these things as if everyone tasted the same thing.
>>>But in this case, why bother with having sumne?
Historical reasons. JCB had a similar word (sutme) meaning panci, and I
was adding just the sort of parallels that Jorge suggests in his reponse
to you. It turned out that his word-making probably used the 'wrong'
meaning of smell, the detecting, when he used the word for the detected.
>>>(d) Unless I've misunderstood, I suggest dropping the x3 of panci,
>>>and dropping sumne altogether, using panci zei ganse instead.
We don't delete gismu merely because they aren't apparently useful.
(Repeat after me: "We don't..." %^)
>>>(9) Is there an agreed expression for look/appearance/ countance/visual
>>>stimulus, without there being an implied perceiver?
>> I'd say {jvinu}.
>
>Yes - I hadn't realized. (Gismu keywords are chosen to cunningly
>disguise the extensive patterns among gismu.)
Well, gismu keywords were chosen for use in LogFlash, where uniqueness
was more important than accuracy or pattern-recognition (even if we had
fully recognized the patterns back then). Of course we expected a
dictionary to replace the keywords as a way to look up based on English
words a LONG time ago. (I'm really close now!!! Just don't go making me
change any place structures at the last minute...)
jvinu gives something like this, but I am not sure what it means to have
a visual stimulus (or any other kind of stimulus) unless there is
something to be stimulated. Sounds and light and textures exist whether
they are perceived - they are physical phenomena. So do flavors. You
seem to be trying to talk about perceived properties while implying that
those properties are observer-independent. (I sense a simsa connection
in here somewhere).
>>Let's see: jvinu sance panci vrusi tengu
>> viska tirna sumne ????? pencu
>>
>>Definitely, there's something missing.
Let's try
observer ind. | jvinu sance vrusi tengu
| gusni
observer dep. panci molpanci?
sense viska tirna sumne molsumne? pencu
Note that pencu doesn't tell what texture is felt, viska is worded
ambiguously as to whether the x2 is the emitter of the light or the
light itself, and tirna likewise for the sound vs. sound maker.
I'm not sure this is a problem, and am unlikely to bother trying to solve
it at this late date. At worst, I think we have a metonymy situation,
which is one degree less complicated than sumti-raising.
>>>(10) tirna: x1 hears x2 against background/noise x3 Could x3 be
>>>abolished, please? Otherwise, when there's no background noise we'll
>>>have to remember to use "xohe" (or whatever the sumti-abolishing cmavo
>>>is).
That is "background/noise", not "background noise". A silent background
is not the same as no background. The background can affect what is
heard. The parallel statements are true for colors - there is always a
background, and it can affect what is perceived.
>>>(11) Is there a standard expression for 'saliva'?
>>The list gives molselpu'u, but I guess you don't like it. I can't
>>think of anything better.
>
>I thought I might have overlooked a gismu. A better lujvo is
>molselcigle.
molselcigla, and I can put it to Nick for possible addition (if he gets
time) to the lujvo list. He has several hundred others in his queue as
well, with more added everytime you guys write in Lojban. (.i'ecai doi
la goran. .e la ken. .e la xorxes.)
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any
Some comments on Veijo's very good summary.
(I don't agree with everything, but at least I think we
are starting to agree on what is the question.)
> (1) mi ponse pa tanxe
> (2) mi nitcu pa tanxe
> There is a relationship which is correctly expressed by
> both (1) and (2), even if (1) is apparently transparent
> and (2) apparently opaque. Do you see it? It is the
> relationship between {mi} and {pa}, the number of boxes,
> I either have one or need one.
I think I can see it, but then are we giving up the notion
that bridi describe relationships between the referents
of sumti?
> When we have an external quantifier we are not so much
> concerned about the identity/specificity as the number of
> entities. Perhaps we could sidestep the whole issue by
> defining that an external quantifier is a type of
> combined quantity abstraction descriptor. With this
> definition
>
> (3) mi kalte lo xanto
>
> would be an opaque claim in the Quinean sense (given the
> implicit external quantifier {su'o}). The transparent
> case (which involves a specific elephant(s)) would be
>
> (4) mi kalte le xanto
>
> The implicit external quantifier {ro} makes the transparency.
I think we agree that if the quantifier is {ro}, the claim is
transparent. (Be it {ro le} or {ro lo}.)
I'm not sure if you are proposing that in the case of
quantifiers other than {ro} the claim should be always opaque,
or either opaque or transparent, i.e. an ambiguous claim.
For example, suppose that I'm hunting a specific elephant
which we both agree to call {le xanto}. Then (4) is true.
Is (3) true in that case?
If (3) is false because {lo} is always opaque, then claims
with {lo} become mostly useless. It is only for a few predicates
that opaque claims are the most common ones.
If (3) is true, then (3) by itself doesn't tell us much,
because we don't know if you are claiming that there is such
an elephant being hunted, or that I am elephant-hunting,
no matter what all existing elephants are doing.
This gives me an idea. Why not {mi xanto kalte} for the opaque
claim? Similarly {mi tanxe nitcu}. Of course, if someone comes
asking me {do xanto kalte ma}, I wouldn't know what to respond.
Jorge
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: Thoughts on "any"
In-Reply-To: (Your message of Sun, 25 Sep 94 12:05:00 D.)
Date: Sun, 25 Sep 94 20:55:04 +0100
> I brought in irrealis into the last paragraph, because it hasn't thus
> far been mentioned in this discussion. Irrealis mode seems to me to be
> what we are talking about whenever we aren't sure that something exists,
> asnd there may be some solution to the problems of quantification in
> 'mental world' sumti by using the marker for irrealis mode (Colin Fine
> has argued this as the general meaning of "da'i" in UI).
I have been bringing in irrealis in the way you describe, but
doubtless with a great lack of perspicuity.
I feel that the discussion on 'any' has gone as far as it can
without expert contribution from, say, pc. Jorge & I seem to
be in agreement now, & I'm not finding that any of the recent
postings from anyone else are shedding new light on matters.
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Re; And/Jorge on gismu lexicon
> >>That said, faktoi (facki troci) is a nice synonym for sisku.
>
> Only if we solve the intensional places problem per And's (1). troci and
> sisku relate to in-mind things, while you cannot facki unless the se
> facki is 'real'.
How can I say "I'm looking for my umbrella"?
mi sisku le mi santa
is not right if the x2 of sisku is a property. So what would it be?
mi sisku le ka le mi santa cu zvati
I seek the property of my umbrella is somewhere.
Is that how sisku is supposed to be used?
I could also say:
mi troci le nu mi facki le du'u le mi santa zvati makau
I try to find out where is my umbrella.
or:
mi faktoi le du'u le mi santa zvati makau
I investigate where is my umbrella.
or:
mi zvafaktoi le mi santa
I search for my umbrella.
The lujvo list I have gives {stufa'i} for "to find something" (find
its location). That works for things that have a location, like
cities, but not for umbrellas, for which I think {zvafa'i} is better.
> (Meanwhile it is possible to cilre a falsehood, so it
> also might be intensional/in mind, along with things that are se ctuca)
There's no problem with the du'u places. There may be problems with the
"subject" places:
> You can bring our "any" discussion into these places and raise spectres
> of existence/quantification problems with "I am teaching about
> unicorns", "I am learning (anything/anything I can) about unicorns".
But you can always fill them with {ro lo} or {piro loi}
mi cilre fi piro loi pavyseljirna
I learn about the whole mass of unicorns
The subject is the whole mass. (Or it could be part of the mass, but a
well defined one, e.g. white unicorns.) For this reason, "subject" places
are relatively safe from the opaqueness problem.
> "I dreamt about unicorns",
This one is not a problem with {senva}, which doesn't have a "subject"
place.
It is strange that a {se senva} can be either an event or a fact, though.
Should we say:
le se senva cu ca'a fasnu
The dream (the dreamed event) really happens.
or:
le se senva cu jetnu
The dream (the dreamed predication) is a true predication.
I think x2 of {senva} should be events, not predications.
> and "I remember seeing a unicorn (our memories
> are not necessarily reality).
For that sentence, there's no problem:
mi morji le du'u mi viska lo pavyseljirna
Even with:
mi morji fi lo pavyseljirna
I remember (something) about a unicorn.
there's no problem. Which unicorn? The one I remember. It at least exists
in my mind. The box of "I need a box" doesn't exist even in my mind, it
is a "representative box", unlike the unicorn that you saw, which is
merely an imaginary one.
> It becomes easy to find places where "da
> poi" quantification is a problem, and in some of these, the possibly
> sumti-raised "da poi" is an abstraction, and eliminating sumti-raising
> might force us to explicit 2nd-order abstractions:
>
> mi cilre ledu'u da zo'u da ka de zo'u de mela .iunikorn.
> I learn something that is a property of some unicorn.
> I learn anything about unicorns.
>
> Yeccch!!!
The only problem with that English sentence that I can see is the "anything".
If you change to "something", then:
mi cilre da loi pavyseljirna
I learn something about unicorns.
> >>I'v seen {kurji} used in this sense, but I don't like it. How do you
> >>take care of an event?
>
> I hope not from me %^) jundi works for events as well as objects
kurji kuj ku'i take care of 'care'
x1 takes-care-of/looks after/attends to/provides for/is caretaker
for x2 (object/event/person)
How do you take care of an event?
Jorge
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: Re; And/Jorge on gismu lexicon
In-Reply-To: (Your message of Sun, 25 Sep 94 12:05:48 D.)
Date: Sun, 25 Sep 94 21:18:50 +0100
I take virtually all your points. I'm still not satisfied
with simlu, but I shan't go on about it.
I agree that sumti raising problems and intensionality
problems are going to recur and recur & some adaptable
solution needs to be found.
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: general response on needing books
In-Reply-To: (Your message of Sat, 24 Sep 94 13:42:22 EDT.)
> And:
> > Le prenu cu jinvi le duhu xihile cukta cu blanu
> >
> > The person thinks "the book is blue".
>
> I'd translate that English sentence as:
>
> le prenu cu jinvi la'e lu le cukta cu blanu li'u
I accept this as a solution (avoiding need for 'xihi' with LE)
so long as lu...lihu may merely represent a bridi, and not
necessarily an utterance.
> > To find out which book, you'd
> > have to ask the person, not me the speaker.
>
> Why the person? Is there something special about the x1 position?
In the pragmatic context it would be implied that it would
be the person that knows the referent of 'xihile cukta'. All
'xihi' would indicate grammatically is that the speaker doesn't
necessarily know the referent.
I think on the whole you are right & I am wrong: it is better
not to use 'xihile' and instead have a way of representing
other people's thoughts.
> > You need local quantification or reference-assignment only when it
> > is in a clause subordinate to an irrealis element of meaning (i.e.
> > something whose argument is not necessarily the case).
>
> I disagree. The quantification applies to realis elements as well.
> (The reference assignment is a different issue, which I don't think
> has anything to do with locality.)
>
> For example:
>
> mi djuno le du'u lo cukta cu blanu
> I know that there is a book that is blue.
>
> da poi cukta zo'u mi djuno le du'u da blanu
> There is a book such that I know it is blue.
>
> Different claims, both with realis subordinate clauses.
> (In the second one I have to know which book, in the first one I
> may or may not know.)
Are you *sure*? I agree there isn't an irrealis element
(assuming djuno is like 'know' rather than 'believe')
but your two examples (in both Eng. & Loj) seem to me
to mean the same thing.
> > And in most
> > utterances there won't be such an element. However, when there is
> > such an element you may indeed be right that local rather than
> > outermost quantification is more often what is wanted.
>
> The question then is whether the need to use outside quantification
> is significant enough to warrant the introduction of xi'i. You can
> always be explicit using a prenex, is there really a need for the
> more compact form with xi'i? I'm not sure.
What if people forget or don't bother to use a prenex, even
when their intended meaning warrants it?
Still, I think I will drop the 'xihi' proposal, in the hope
that combinations of xehe, lahelu, and use of prenexes will
do the job.
---
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: general response on needing books
> > mi djuno le du'u lo cukta cu blanu
> > I know that there is a book that is blue.
> >
> > da poi cukta zo'u mi djuno le du'u da blanu
> > There is a book such that I know it is blue.
> >
> > Different claims, both with realis subordinate clauses.
> > (In the second one I have to know which book, in the first one I
> > may or may not know.)
>
> Are you *sure*? I agree there isn't an irrealis element
> (assuming djuno is like 'know' rather than 'believe')
> but your two examples (in both Eng. & Loj) seem to me
> to mean the same thing.
I have 5 books on my table. I tell you that one of them is blue.
Assuming you believe me, then
do djuno le du'u pa le cukta cu blanu
You know that one of the books is blue.
but not
pa le cukta zo'u do djuno le du'u cy blanu
For one of the books: you know that it is blue.
> > The question then is whether the need to use outside quantification
> > is significant enough to warrant the introduction of xi'i. You can
> > always be explicit using a prenex, is there really a need for the
> > more compact form with xi'i? I'm not sure.
>
> What if people forget or don't bother to use a prenex, even
> when their intended meaning warrants it?
They would also forget xi'i... :(
> Still, I think I will drop the 'xihi' proposal, in the hope
> that combinations of xehe, lahelu, and use of prenexes will
> do the job.
I think that if nitcu et al. are changed to event-only, then xe'e
won't be of much use at all either.
>
> ---
> And
>
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: any
la xorxes. cusku di'e
> Date: Sun, 25 Sep 1994 13:56:54 EDT
> From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
> Subject: Re: any
>> (1) mi ponse pa tanxe
>> (2) mi nitcu pa tanxe
>
>> There is a relationship which is correctly expressed by
>> both (1) and (2), even if (1) is apparently transparent
>> and (2) apparently opaque. Do you see it? It is the
>> relationship between {mi} and {pa}, the number of boxes,
>> I either have one or need one.
>
> I think I can see it, but then are we giving up the notion
> that bridi describe relationships between the referents
> of sumti?
Separating {pa} from {pa tanxe} was just a gimmick to
help to see the situation. Let's concentrate on (1) to
avoid the problems with {nitcu}. (1) describes a relationship
between the referent of {mi} and the referent of {pa tanxe}
([pa tanxe]). [pa tanxe] is, however, something intangible,
an abstraction. Until I fix it with {le}, it can be just
any box, there is no difference what so ever to me needing
a box (some box, any representative of boxes). There isn't
even a way of getting at the identity without fixing it
with {le} ( {xu do ponse le vi tanxe} ).
Outer quantifiers are weird (let's at first keep {lo} out
of this - it's weird, too). Let's say I'm a co-owner of a
herd of cattle. I own 10 out the total of 50.
mi ponse pano le muno bakni
When I make a statement like this, it means, in a way, that
I own ALL the 50 as long as I don't start separating the
animals into two groups. Upto the count of 10, if you point
any animal and ask whether I own it, I can say 'yes'. So
there is a conditional ownership to the whole.
It is the same with all selbri - how ever concrete. We don't
need to have {nitcu} or {djica}. If you kill (with no selection
based on some criteria) 10 animals out of the 50, you kill
any/every animal up to the count of 10.
Say, we have a game of cards. You mix the deck of cards, take one
card and put it face down on the table. Until you turn the card
you have just ANY ONE of the cards. There is no way I can say
you don't have an ace.
When you say
mi nitcu ci [lo] tanxe
you need every box which you come across until you have collected
the 3. But there is NO difference if you say
mi ponse ci tanxe
If you express no restrictions (and no box is identified as belonging
to someone else) then any box up to the count of 3 can be said to
belong to you. There is no fundamental difference between {nitcu} and
{ponse} or any other selbri when we have an outer quantifier. If you
say otherwise, you must somehow be importing an implicit {le} to
turn the outer quantifier into an inner one.
So I think that {mi ponse ci tanxe} expresses a relationship between
{mi} and the referent of a two-part abstraction {ci}{lo ro tanxe}.
I have this {ponse} relationship to all the really are boxes as long
as my possessions aren't separated from the rest - then the number is
limited to 3 and {mi ponse le ci tanxe}. As long as you don't
individuate your possessions they are just abstractions - like most
of the money today, just numbers. When you are giving someone 'five
dollars' you are giving him an abstraction - 'a five-dollar bill'
is somewhat more concrete, 'the five-dollar bill' is the real stuff.
It's just sloppiness of usage if I say {mi ponse ci tanxe} when I
actually mean something like {mi ponse le ci vi tanxe}.
>> When we have an external quantifier we are not so much
>> concerned about the identity/specificity as the number of
>> entities. Perhaps we could sidestep the whole issue by
>> defining that an external quantifier is a type of
>> combined quantity abstraction descriptor. With this
>> definition
>>
>> (3) mi kalte lo xanto
>>
>> would be an opaque claim in the Quinean sense (given the
>> implicit external quantifier {su'o}). The transparent
>> case (which involves a specific elephant(s)) would be
>>
>> (4) mi kalte le xanto
>>
>> The implicit external quantifier {ro} makes the transparency.
>
> I think we agree that if the quantifier is {ro}, the claim is
> transparent. (Be it {ro le} or {ro lo}.)
>
> I'm not sure if you are proposing that in the case of
> quantifiers other than {ro} the claim should be always opaque,
> or either opaque or transparent, i.e. an ambiguous claim.
>
> For example, suppose that I'm hunting a specific elephant
> which we both agree to call {le xanto}. Then (4) is true.
> Is (3) true in that case?
>
> If (3) is false because {lo} is always opaque, then claims
> with {lo} become mostly useless. It is only for a few predicates
> that opaque claims are the most common ones.
>
> If (3) is true, then (3) by itself doesn't tell us much,
> because we don't know if you are claiming that there is such
> an elephant being hunted, or that I am elephant-hunting,
> no matter what all existing elephants are doing.
I think there would be very few situations where I'd use
(3) if I had a specific elephant in mind. If I'd like
to emphasize that a real elephant is in question, I'd say
mi kalte le pa lo xanto
(Let's keep Quine out of this mess. This time I brought him
and the elephants in just to show that they might fit within
the wider framework I was outlining.)
When you are out to get Clarence, technically it is quite
correct to state {do kalte lo xanto} as you ARE hunting one
of the really are elephants, but pragmatically it's somewhat
dubious.
The way I see it now, the Lojban implicit outer/inner quantifiers
solve certain problems quite elegantly - once we understand the
workings of the combined quantifiers correctly.
> This gives me an idea. Why not {mi xanto kalte} for the opaque
> claim? Similarly {mi tanxe nitcu}. Of course, if someone comes
> asking me {do xanto kalte ma}, I wouldn't know what to respond.
I thought of this, too. It get's ugly when you start inserting
numbers - {mi tanxe nitcu la'u re tanxe}.
> Jorge
I'm not absolutely sure whether my present approach is correct
or not, but I feel it just might be. I'll leave the matter to
the rest of us to ponder. Existence, specificity and what ever
are other cans of worms. This was just 'any' :-) / :-(
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
Date: Mon, 26 Sep 1994 05:15:48 -0400
From: Logical Language Group <lojbab>
Subject: dictionary files
I've left the dictionary files compressed in our directory to save you a few
seconds downloading. I will delete them Mon night, since they are easy
enough to get again from the ftp site.
I am obviously now waiting on the bogus list file and/or will have to do it
myself. Any hope?
I notified netcom that a major listserv node seems to not be able to
connect to them, but no response. I kinda hate to kick off 3 people, 1
of whom is active, especially when I know that I can email them directly.
But I will leave nomail decisions to you (I assume we can just send them
direct email telling them to try resubscribing). If you have a better
idea than mine regarding the French guy who just responded, please speak
up.
Self-congratulations are in order!!!
And Avgust just now for the first time woke up in the middle of the night to
empty his bladder. The alarm training is working very quickly! Only 2
weeks so far, and his bedwetting is down 75%. No more sheet changing every
day.
Hope things are improving for you.
lojbab
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: general response on needing books
In-Reply-To: (Your message of Sun, 25 Sep 94 18:04:28 EDT.)
Jorge:
> > > mi djuno le du'u lo cukta cu blanu
> > > I know that there is a book that is blue.
> > >
> > > da poi cukta zo'u mi djuno le du'u da blanu
> > > There is a book such that I know it is blue.
> > >
> > > Different claims, both with realis subordinate clauses.
> > > (In the second one I have to know which book, in the first one I
> > > may or may not know.)
> > Are you *sure*? I agree there isn't an irrealis element
> > (assuming djuno is like 'know' rather than 'believe')
> > but your two examples (in both Eng. & Loj) seem to me
> > to mean the same thing.
> I have 5 books on my table. I tell you that one of them is blue.
> Assuming you believe me, then
> do djuno le du'u pa le cukta cu blanu
> You know that one of the books is blue.
> but not
> pa le cukta zo'u do djuno le du'u cy blanu
> For one of the books: you know that it is blue.
I still don't see it. 'Know' means that that its duhu complement
is true (according to the speaker). You seem to be using 'djuno'
to mean 'believe', whose duhu complement is not necessearily
true - i.e. it is irrealis - and only on this interpretation can I
understand your examples to differ in meaning.
> > Still, I think I will drop the 'xihi' proposal, in the hope
> > that combinations of xehe, lahelu, and use of prenexes will
> > do the job.
>
> I think that if nitcu et al. are changed to event-only, then xe'e
> won't be of much use at all either.
I think xehe is still needed for "xehe PA" constructions: "I am
willing to read any three books", "any two people can sit on
the sofa". Can you do these without xehe?
----
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: any
In-Reply-To: (Your message of Mon, 26 Sep 94 09:06:38 O.)
Veijo:
> It is the same with all selbri - how ever concrete. We don't
> need to have {nitcu} or {djica}. If you kill (with no selection
> based on some criteria) 10 animals out of the 50, you kill
> any/every animal up to the count of 10.
I don't think this & your other similar examples are right, for
realis contexts. If you have killed 10 out of 50 animals, I
cannot just pick any ten of them & truthfully assert that
you killed them (unless by a freak of chance I happen to
pick the right 10). Your context would work for, say,
*intending* to kill 10 out of 50 animals.
---
And
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: TECH: Transparence / Opaqueness
>Imperatives by definition are neither true nor false in Lojban. That
>means: make {do dunda ci plise mi} true. Since the distinction between
>opaque and transparent rests on how the truth value of the statement
>is determined, no such distinction is possible for imperatives.
Is that really true? Aren't both interpretations possible of "ko cpacu le
tanxe"? It could mean "go get me a box (any box)" or "go get me a box
(whose location I'm about to reveal...)"
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: general response on needing books
And:
> I still don't see it. 'Know' means that that its duhu complement
> is true (according to the speaker).
Exactly. But a quantification inside the du'u complement refers to
that statement (the one that the speaker knows is true). A
quantification outside refers to the statement that says that the
speaker knows something is true.
> You seem to be using 'djuno'
> to mean 'believe', whose duhu complement is not necessearily
> true - i.e. it is irrealis - and only on this interpretation can I
> understand your examples to differ in meaning.
Suppose you find a handwritten letter without a signature.
>From your knowledge of how things get written,
do djuno le du'u da zo'u da ciska le xatra
You know that for some x, x wrote the letter.
However,
da zo'u do na djuno le du'u da ciska le xatra
It is false that for some x, you know that x wrote the letter.
or rephrasing,
roda naku zo'u do djuno le du'u da ciska le xatra
For each x, it is false that you know that x wrote the letter.
> > I think that if nitcu et al. are changed to event-only, then xe'e
> > won't be of much use at all either.
>
> I think xehe is still needed for "xehe PA" constructions: "I am
> willing to read any three books", "any two people can sit on
> the sofa". Can you do these without xehe?
Yes, for the same reason you can do nitcu: they are inside abstractions.
mi djica le nu mi tcidu ci selcku
I am willing to read three books.
kakne le nu re prenu cu zutse le sfofa
Can that two people sit on the sofa.
Who is it that can, in the last one, is tricky.
But yes, there has to be a way at least to emphasize the anyness, like
the English "whatsoever".
Jorge
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: any
>This gives me an idea. Why not {mi xanto kalte} for the opaque
>claim? Similarly {mi tanxe nitcu}. Of course, if someone comes
>asking me {do xanto kalte ma}, I wouldn't know what to respond.
I thought of this, too, and I started to post the suggestion but I deleted
my message because I talked myself out of it, because "mi nitcu xe'e lo
tanxe" seems to make some sort of a logical statement, while "mi tanxe
nitcu" is vague by definition.
But I don't really like "xe'e", and maybe an explicit notion of opacity
isn't really necessary for the language to be complete, if you can say what
you need to say without using it.
There are two problems which jump to mind though:
Is it OK to allow the implied transparent "zo'e" to sit there in the
x2 place when we're trying to be opaque? Or can the "obvious value" of
zo'e happen to be "ne'e"?
How do you say "I need two boxes?" I like "mi me re tanxe me'u nitcu"
semantically, but its not very zipfy. "mi remei tanxe nitcu" is a
little zipfier and vaguer. "mi nitce co me re tanxe"
>Of course, if someone comes
>asking me {do xanto kalte ma}, I wouldn't know what to respond.
I think that's OK, even though opaque and transparent don't occur together
in English.
mi xanto kalte I'm hunting elephants!
.i go'i ma? Which one(s) in particular?
.i go'i ne'e None in particular.
OR
.i go'i la dambos. I'm hunting elephants, specifically Dumbo
.i .o'onai do palci Oooh, you're evil!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: any
> Date: Mon, 26 Sep 1994 20:04:35 EDT
> From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
> Subject: Re: any
> la veion cusku di'e
[...]
> You seem to be proposing that non-specific quantification (any outer
> quantification other than {ro}) be always treated as opaque. This
> is possible in theory, but then you can't use Lojban to make logical
> statements the way we usually understand them.
That's right. They have to. Otherwise we are in a mess. The logics
will have to be worked out (probably has been done in the maths;
I seem to recall, from about 30 years back, that the preneces used to
have 3 kinds of elements in them: {for all x / ro da}, {there exists
x / su'o da} and {for n = ... / ny goi ...}; definite numbers do
complicate things)
I have this feeling. We may have to make this change in interpretation
in order to float. Any mathematicians around?
>> mi ponse ci tanxe
>>
>> If you express no restrictions (and no box is identified as belonging
>> to someone else) then any box up to the count of 3 can be said to
>> belong to you.
>
>> It's just sloppiness of usage if I say {mi ponse ci tanxe} when I
>> actually mean something like {mi ponse le ci vi tanxe}.
> I agree the two mean different things. In the second claim you are
> identifying the boxes. But even though you are not identifying them
> in the first claim, I think the claim says that the relationship holds
> for three of them that can (ka'e) be identified.
For identification you'll have to express some kind of restriction.
Our natural language habits (though languages disagree with each
other in the usages) make us feel that we must be able to make an
(inherently) opaque claim in a transparent situation. The problem
may be harder for people who's native tongue has definite articles.
They might find it awkward to insert a {le} into the claim (Finnish
has no articles and quite often I am either at a total loss whether
or which kind of an article to use in English, or I find it awkward
to insert one in situations where a Finnish usage seems to reveal
the underlying claim better.) Pragmatics (Zipf) makes us use language
(at least a natural one) inaccurately. It is, however, reasonably
easy to see the basic opaqueness of many types of statements - once
you stop to think about it. It took my fifteen year old daughter
about two seconds to understand the inherent meaning of a {mi ponse
re cukta}-type claim - actually I didn't have time to finish my
statement of the problem. But even though someone realises the
inaccuracy of a usage he may have no reasonable way of avoiding it
- in a natural language with petrified usages. This is causing an
immense amount of trouble in the maths (a prenex has to be build
with utmost care to avoid making false claims). We do not have to
carry over these usages into Lojban (which source language would we
use - not English if I do have a choice :-)
I think we must make the language definition watertight - what
ever the consequences would be if people used the language
strictly to the letter. This way we can handle formal things
with ease - that is after all one of the fundaments of the
Loglan Project.
In everyday language usage we can make concessions. There are
always numerous unexpressed restrictions. You cannot always state
all the premises, so you might, e.g., use an opaque expression
instead of a transparent one - though I personally see no pressing
need for that. And when you want to say that you really mean what
you are saying, you just insert {sa'e}.
> Jorge
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: any
> Date: Mon, 26 Sep 1994 19:41:34 +0100
> From: ucleaar <ucleaar@UCL.AC.UK>
> Subject: Re: any
>
> Veijo:
>> It is the same with all selbri - how ever concrete. We don't
>> need to have {nitcu} or {djica}. If you kill (with no selection
>> based on some criteria) 10 animals out of the 50, you kill
>> any/every animal up to the count of 10.
>
> I don't think this & your other similar examples are right, for
> realis contexts. If you have killed 10 out of 50 animals, I
> cannot just pick any ten of them & truthfully assert that
> you killed them (unless by a freak of chance I happen to
> pick the right 10). Your context would work for, say,
> *intending* to kill 10 out of 50 animals.
No. You miss the point.
If I go out, kill 10 animals of the 50 and come to tell you about
my deed, then until you go out and identify the animals (if you
have some need for the identification) I can have killed any
single one of them up to the count of 10. If you refer to THE 10
animals I killed, you cannot refer to them without a {le}, they
are {le poi mi ba'o catra ke'a ku'o pano bakni} - you DO NOT
do a resampling unless you want, e.g., to have a subset of the
killed animals. However, I think that in a situation like this
I'd already use {le} when first telling you
mi [ba'o] catra le bi'u pano le muno bakni
I killed these ten of the 50 cows
I think I have come across this kind of idiomatic usage of
{this/these} to avoid the problems of {-/a/the} ("I saw this
man. He was standing..."). {le bi'u} is a good equivalent.
I think that sometimes people are slightly uneasy with
inherently opaque expressions in natlangs and resort to
idioms which circumvent the problem of a strict indefinite/
definite duality.
> ---
> And
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: general response on needing books
In-Reply-To: (Your message of Mon, 26 Sep 94 20:37:28 EDT.)
Jorge:
> Exactly. But a quantification inside the du'u complement refers to
> that statement (the one that the speaker knows is true). A
> quantification outside refers to the statement that says that the
> speaker knows something is true.
Ok, you & John have got me to see it.
> > > I think that if nitcu et al. are changed to event-only, then xe'e
> > > won't be of much use at all either.
> >
> > I think xehe is still needed for "xehe PA" constructions: "I am
> > willing to read any three books", "any two people can sit on
> > the sofa". Can you do these without xehe?
>
> Yes, for the same reason you can do nitcu: they are inside abstractions.
>
> mi djica le nu mi tcidu ci selcku
> I am willing to read three books.
>
> kakne le nu re prenu cu zutse le sfofa
> Can that two people sit on the sofa.
I think that this is false only if there is no event of me reading
3 books that I consent to, and if there is no event of two people
sitting on the sofa. What I want is it to be false if before
you've finished selecting the three books & two people you
pick a book & it turns out that I don't consent to read it
or you pick a person and the event of them sitting on the
sofa isn't possible.
Given the form "I consent to an event such that there are
3 books and I read them", then you could offer me a book
& I could reject it without falsifying the statement. To
avoid this, we need "xehe".
----
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: any
In-Reply-To: (Your message of Tue, 27 Sep 94 11:11:59 O.)
Veijo:
> If I go out, kill 10 animals of the 50 and come to tell you about
> my deed, then until you go out and identify the animals (if you
> have some need for the identification) I can have killed any
> single one of them up to the count of 10.
This is the bit I don't understand. Suppose there are three animals,
Alfie, Boris and Candy, and you kill Alfie. Thereafter you have
killed Alfie, but I cannot see how (i) it is true that you can
have killed any single one of them (since you can't have killed
Boris or Candy), and (ii) why my subsequent identification of
the animals has any effect on which was the one you so casually
slaughtered for the sake of explaining semantics to us.
----
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Analogy
la djan cusku di'e
> 1) ro da poi broda cu brode
> 2) ro da poi broda zo'u da brode
> 3) ro broda cu brode
>
> are strictly equivalent in meaning, and imply nothing about the existence
> of widgets.
I agree that those are equivalent. I don't see any difference between
{lo broda} and {da poi broda} either, and they can be shown to be
equivalent using the equivalence of (1) and (3):
lo broda cu brode
<==> naku naku lo broda cu brode
<==> naku ro broda naku cu brode
<==> naku ro da poi broda naku cu brode
<==> naku naku da poi broda cu brode
<==> da poi broda cu brode
If there is no broda that is in relationship brode, then the claim
should be false, whether it is made using {lo broda} or {da poi broda}.
And the absence of a {lo broda} may or may not be because it is an
empty class. Why are they supposedly different?
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: any
> Date: Tue, 27 Sep 1994 20:11:20 +0100
> From: ucleaar <ucleaar@UCL.AC.UK>
> Subject: Re: any
> Veijo:
>> If I go out, kill 10 animals of the 50 and come to tell you about
>> my deed, then until you go out and identify the animals (if you
>> have some need for the identification) I can have killed any
>> single one of them up to the count of 10.
>
> This is the bit I don't understand. Suppose there are three animals,
> Alfie, Boris and Candy, and you kill Alfie. Thereafter you have
> killed Alfie, but I cannot see how (i) it is true that you can
> have killed any single one of them (since you can't have killed
> Boris or Candy), and (ii) why my subsequent identification of
> the animals has any effect on which was the one you so casually
> slaughtered for the sake of explaining semantics to us.
I ought to have continued: 'as far as you are concerned.'
(The same goes for me if I had no means of identifying them
in the first place.)
(The casual slaughtering was somehow related to Quine's
elephant-hunting - it keeps haunting me :-)
> ----
> And
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
Date: Tue, 27 Sep 1994 23:45:19 -0400
From: Logical Language Group <lojbab>
Subject: misc
1. Do you have some files in Transit that were supposed to have autodeleted?
We're at around 6 Meg again.
2. Doug is getting nasty about multiple users sharing accounts on digex.general.
WE may want to start thinking about getting you a separate account, which will
help the disk space problem (at a good cost, alas) at the same time, and
keep us on the up-and-up. If we do it before 30 Sept. we can get you a
discount.
If they are being aggressivce about it, the fact that you are logging in in
a different city probably will make it pretty clear that we are using
multiple users. And after over a year, it is hard to argue that this is
'short term temporary' anymore.
I consider that the LogFest approval of pc's account alos justifies paying for
one for you. I think it was discussed as such.
We should consider making us a 'group', however that is done, so we
can read each other's files when appropriate.
Because of the timing, let me know ASAP (or call Digex yourself).
lojbab
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: Analogy
In-Reply-To: (Your message of Tue, 27 Sep 94 16:09:53 EDT.)
Jorge:
> > 1) ro da poi broda cu brode
> > 2) ro da poi broda zo'u da brode
> > 3) ro broda cu brode
> >
> > are strictly equivalent in meaning, and imply nothing about the existence
> > of widgets.
>
> I agree that those are equivalent. I don't see any difference between
> {lo broda} and {da poi broda} either
Is there (I ask) a difference along the lines of (i-ii)?
(i) Ex A(x) & B(x) [There is an x such that x is an A and x is a B]
(ii) Ex A(x) -> B(x) [There is an x such that if x is an A then x is a B]
The second one doesn't claim there is an A.
---
And
From: Desmond Fearnley-Sander <dfs@HILBERT.MATHS.UTAS.EDU.AU>
Subject: Re: any
This may not help.
At 8:11 PM 27/9/94 +0100, ucleaar wrote:
Veijo:
>> If I go out, kill 10 animals of the 50 and come to tell you about
>> my deed, then until you go out and identify the animals (if you
>> have some need for the identification) I can have killed any
>> single one of them up to the count of 10.
>
And:
>This is the bit I don't understand. Suppose there are three animals,
>Alfie, Boris and Candy, and you kill Alfie. Thereafter you have
>killed Alfie, but I cannot see how (i) it is true that you can
>have killed any single one of them (since you can't have killed
>Boris or Candy), and (ii) why my subsequent identification of
>the animals has any effect on which was the one you so casually
>slaughtered for the sake of explaining semantics to us.
>
"And has slaughtered one of his three pigs."
"Oh! I hope it wasn't Alfie."
"No. It was Boris."
"That's a pity, Candy is the fattest."
"I was only kidding. It was really Candy."
.......
This discourse is about information, not facts. It makes sense
independently of the facts, and indeed, since it is all invented, there are
no associated facts. (In case someone is concerned, I'm afraid I don't
even know which pig was killed. If any.)
Even if there were facts, the discourse would still be about information.
The truth of the matter is external to the discourse. You may choose to
believe the first speaker in the dialogue, or not, but to be sure you will
have to visit the pigpen.
lojbab:
>Staements about non-0existent things are meaningless (they may be defined
>as false because they are meaningless, but this is definitional).
>
"Well if it was Candy, we're in for a feast."
Meaningless? False?
Desmond
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Outer quantifiers (was: any & al)
After strugling a couple of hours unsuccessfully with And's posting
> Date: Tue, 27 Sep 1994 19:49:16 +0100
> From: ucleaar <ucleaar@UCL.AC.UK>
> Subject: Re: general response on needing books
>> mi djica le nu mi tcidu ci selcku
>> I am willing to read three books.
>>
>> kakne le nu re prenu cu zutse le sfofa
>> Can that two people sit on the sofa.
>
> I think that this is false only if there is no event of me reading
> 3 books that I consent to, and if there is no event of two people
> sitting on the sofa. What I want is it to be false if before
> you've finished selecting the three books & two people you
> pick a book & it turns out that I don't consent to read it
> or you pick a person and the event of them sitting on the
> sofa isn't possible.
> Given the form "I consent to an event such that there are
> 3 books and I read them", then you could offer me a book
> & I could reject it without falsifying the statement. To
> avoid this, we need "xehe".
and ruminating all the talk about 'any' I have come to the conclusion
that the problem isn't the case with 'any'.
I decided to approach the problem from a different angle:
Given a sentence AND ONLY THE INFORMATION CONTAINED THEREIN write
a program which performs the requested action.
Now, of course, it turns out that it is impossible to derive any but
the opaque case from a blanket statement like 'mi nitcu re tanxe'
(I'm excluding the abstraction as it really doesn't change anything).
To get anything else but 'any two boxes' requires you to supply
information about any mental reservations you have or any preconditions
the boxes have to satisfy.
So the whole question about outer quantifiers boils down to
do we want an outer quantifier to be a 'pure' quantifier
(opaque, 'any') or do we allow it to be 'hazy' in the
sense that there always maybe some hidden reservation?
In the first case we have either to mark the transparent case or
supply at least part of the reservations/conditions, in the
second case we have to mark the opaque case with {xe'e}.
The consequences in each case would be
(1) opaque interpretation: logically sound but in everyday
usage either ambiguous (does he really mean 'any' or
just 'use your common sense'?) or requiring marking/
supplying the conditions explicitly.
(2) transparent interpretation: convenient (though ambiguos
in the NL way) in everyday usage but beyond logics
unless marked, and probably even then.
This is a policy question. I'd choose (1) and require that the
existence of the hidden conditions is at least acknowledged
with a marker or a vague restriction, e.g.
mi djica le bi'u re tanxe
I need these two boxes (the requirements for which
I have in mind but I'm not telling you)
mi djica re tanxe poi mapti
I need two suitable (in an unspecified way) boxes
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Colin Fine <C.J.Fine@bradford.ac.uk>
Subject: Re: coi doi goran
coi xorxes jo'u goran jo'u kris li'o
..i mi selpu'a le nu so'o da ca'o skamymri bau lo jbobau .i .oi ru'e temxaksu l
e
barda fa lei nu tcidu je jimpe le selmri
> .i mi bacru di'u ki'u mi'a ba'o jai gau xlali vau zo'o
..i di'u na te gerna .i .e'u pei setca lu le nu li'u le lidne be zo mi'a
>
> .izo'onai da'inai le glibau nunctuca ponjo cu na'e jetnu xlali
..i mi se spaji je se cfipu tu'a lu glibau nunctuca ponjo li'u .i ku'i mapti
..i'a ki'u le nu le nunctuca cu je'a ponjo
..i ku'i de'u se smuni le du'u py cu xlali gi'e nai jetnu xlali to jitfa
xlali pei toi
> .iku'ibo mi cilre
> le glibau ca lenu pu verba vi la kalyrados.
..i zo bo cu na'e sarcu
..i mi na birti le du'u ma kau sidbo do se ra'a lu mi pacna le jei cakla kei
li pinopa li'u .i mi smadi tu'a le xe fanva .i ku'i do na'e snada pe'i cusku
..i le remoi terbi be zo pacna cu ckaji lo kamsucta ju'o .e lo kamfasnu
..e nai bo lo niljetnu
co'o mi'e kolin
Colin Fine c.j.fine@bradford.ac.uk
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Transparence / Opaqueness
la kris cusku di'e
> Is that really true? Aren't both interpretations possible of "ko cpacu le
> tanxe"? It could mean "go get me a box (any box)" or "go get me a box
> (whose location I'm about to reveal...)"
It depends on how {ko} is defined. If {ko cpacu lo tanxe} means "make {do
cpacu lo tanxe} true", then you don't get to choose which box, because the
sentence will be true whichever box {do} chooses to get. If imperatives are
defined some other way, then I agree that the opaque/transparent distinction
might make sense for them too.
> Is it OK to allow the implied transparent "zo'e" to sit there in
> the x2 place when we're trying to be opaque? Or can the "obvious
> value" of zo'e happen to be "ne'e"?
What's {ne'e}? I thought it was undefined, and it is the actual cmavo I had
in mind for {xe'e}.
> >Of course, if someone comes
> >asking me {do xanto kalte ma}, I wouldn't know what to respond.
>
> I think that's OK, even though opaque and transparent don't occur together
> in English.
>
> mi xanto kalte I'm hunting elephants!
> .i go'i ma? Which one(s) in particular?
> .i go'i ne'e None in particular.
> OR
> .i go'i la dambos. I'm hunting elephants, specifically Dumbo
> .i .o'onai do palci Oooh, you're evil!
If you replace {ne'e} with {xe'eda}, I think I agree.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any
la veion cusku di'e
> Our natural language habits (though languages disagree with each
> other in the usages) make us feel that we must be able to make an
> (inherently) opaque claim in a transparent situation.
I think you are equating 'opaque' with 'non-specific', otherwise I
don't understand your assertion. Most claims are not opaque.
> Pragmatics (Zipf) makes us use language
> (at least a natural one) inaccurately.
I was in a physics seminar the other day, when to my joy and surprise
the speaker mentioned Zipf's law. Unfortunately, it turns out that it
had nothing to do with what I had learned about it from lojbanic lore.
According to this guy, Zipf discovered that if you plot on a log-log
graph the frequency of a word vs its position in order of frequency,
you get a straight line (of slope 1, I think). This is a very
interesting result, and apparently it works for any language, but has
nothing at all to do with lengths of words. Of course, I'm not
advocating that we lower Our Lord Zipf from his pedestal, but I
thought I'd just mention it.
> I decided to approach the problem from a different angle:
>
> Given a sentence AND ONLY THE INFORMATION CONTAINED THEREIN write
> a program which performs the requested action.
A sentence is not a requested action, unless it contains {ko}.
A sentence is a claim about the world (be it the real world, or any other
that we take as real for the purposes of evaluating the truth value).
To determine the truth value of the claim you examine that world, and
see whether the claim holds. In most cases, us poor inhabitants of that
world are unable to determine the truth value of most statements, but
fortunately, we can talk about those truth values.
> Now, of course, it turns out that it is impossible to derive any but
> the opaque case from a blanket statement like 'mi nitcu re tanxe'
To derive? We examine the world (from an omniscient perspective). We know
who {mi} is. We look at all tanxe. Are there exactly two in relationship
{nitcu} with {mi}? If yes, the statement is true. If not, it is false.
Of course, the purpose of language is to transmit information, not to
provide sentences for truth value evaluation. From the point of view of
the listener of the sentence, the sentence is assumed true, and the
information obtained is that there exist two tanxe in relationship {nitcu}
with {mi}.
> To get anything else but 'any two boxes' requires you to supply
> information about any mental reservations you have or any preconditions
> the boxes have to satisfy.
You can add as much information as you want, to help the listener
identify the two boxes, but this doesn't mean that the claim would
be true for any two boxes you pick.
> So the whole question about outer quantifiers boils down to
>
> do we want an outer quantifier to be a 'pure' quantifier
> (opaque, 'any') or do we allow it to be 'hazy' in the
> sense that there always maybe some hidden reservation?
Why is the 'pure' quantifier the opaque one? The normal logical quantifier
is the transparent one, and there's nothing hazy about it. If you say in
mathematics: "For all y, there exists an x, such that x + y = 0", then you
don't mean "any x". You don't give any clue as to which x you are refering
to, but it certainly is not any whatsoever.
In Lojban, informally this would be:
li no sumji roda de
0 is the sum of each y with some x
and you clearly don't mean any x. You don't specify which x, you don't
have to specify x for the claim to be true. All you need is that x exists.
> In the first case we have either to mark the transparent case or
> supply at least part of the reservations/conditions, in the
> second case we have to mark the opaque case with {xe'e}.
>
> The consequences in each case would be
>
> (1) opaque interpretation: logically sound but in everyday
> usage either ambiguous (does he really mean 'any' or
> just 'use your common sense'?) or requiring marking/
> supplying the conditions explicitly.
This one is logically UNsound, with the usual interpretation of logics.
The everyday usage would be what we expect for nitcu and family (where
the opaque form is the norm in English), but crazy claims for most other
predicates.
> (2) transparent interpretation: convenient (though ambiguos
> in the NL way) in everyday usage but beyond logics
> unless marked, and probably even then.
On the contrary, this is the one that logics requires. It is not ambiguous
in the NL way if the "any" case is required to be marked.
> This is a policy question. I'd choose (1) and require that the
> existence of the hidden conditions is at least acknowledged
> with a marker or a vague restriction,
Are you aware that with interpretation (1), {li no sumji roda de} means
"Zero is the sum of each x and any y (the first you can find)"?
How do you say "Zero is the sum of each x with some y" using
interpretation (1)?
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: Analogy
I agreed with John that:
> > > 1) ro da poi broda cu brode
> > > 3) ro broda cu brode
are equvalent, and therefore, so are:
da poi broda cu brode
lo broda cu brode
And asks:
> Is there (I ask) a difference along the lines of (i-ii)?
>
> (i) Ex A(x) & B(x) [There is an x such that x is an A and x is a B]
> (ii) Ex A(x) -> B(x) [There is an x such that if x is an A then x is a B]
>
> The second one doesn't claim there is an A.
I don't think so. The only claim being made is the relationship {brode}.
If you can't find a broda which is brodeing (either because no broda
actually exists, or because they're all busy doing something else) then
the claim is false. The claim is not about the existence of brodas.
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: doi xorxes. do ponse xo tanxe
xy. zo'u la xorxes. ponse vei xy tanxe .i li xy. du li xo
E(x) : Jorge has x boxes. x = ?
BTW. is it possible to say just {do nitcu xo tanxe}?
I find something like {do nitcu le nu do ponse xo tanxe}
somewhat odd. And, if {do nitcu xo tanxe} is OK,
then why not {mi nitcu re tanxe}?
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: doi xorxes. do ponse xo tanxe
la veion cusku di'e
> xy. zo'u la xorxes. ponse vei xy tanxe .i li xy. du li xo
> E(x) : Jorge has x boxes. x = ?
mi ponse so'i tanxe
> BTW. is it possible to say just {do nitcu xo tanxe}?
I think it is. But I don't think it means the same as what the English
question "How many boxes do you need?" usually means.
> I find something like {do nitcu le nu do ponse xo tanxe}
> somewhat odd.
That's exactly why I want {xe'e}, to avoid having to use the cumbersome
expression in the opaque case.
> And, if {do nitcu xo tanxe} is OK,
> then why not {mi nitcu re tanxe}?
It is OK. It just doesn't mean "I need (any) two boxes".
(All this is assuming that nitcu accepts an object. If it only accepts
abstractions, then there's no way around it other than {mi nitcu tu'a
re tanxe}, {do nitcu tu'a xo tanxe}, etc.)
Jorge
PS: Did my post with the Zipf anecdote get through? I didn't get my
copy, so maybe I should write it again.
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: any
Desmond:
> "And has slaughtered one of his three pigs."
> "Oh! I hope it wasn't Alfie."
> "No. It was Boris."
> "That's a pity, Candy is the fattest."
> "I was only kidding. It was really Candy."
> .......
- la and pu catra pa le ri ci xarju
- uu a'o na go'i la alfis
- na go'i i ja'a go'i la boris
- uicu'i i la kandis plana traji
- zo'o je'unai i ja'a ca'a catra la kandis
> This discourse is about information, not facts.
Or about pigs, rather. It involves exchange of information, for which
it is necessary to know how to deal with facts.
> It makes sense
> independently of the facts, and indeed, since it is all invented, there are
> no associated facts.
There are facts associated. That they are facts only in the imaginary world
where the discourse takes place is a different matter. That we don't know
which are the true facts is also another matter.
The first statement gives information. For the statement to be informative,
the listener has to assume that it is true. What does the statement mean
assuming that it is true? If we don't know the answer to that, then we
don't get any information from it.
> (In case someone is concerned, I'm afraid I don't
> even know which pig was killed. If any.)
I hope it was Candy!
> Even if there were facts, the discourse would still be about information.
> The truth of the matter is external to the discourse. You may choose to
> believe the first speaker in the dialogue, or not, but to be sure you will
> have to visit the pigpen.
I agree, but what do you mean by believing the first speaker? What do you
believe when you believe it? You need to know what the sentence means when
it is assumed to be true.
> lojbab:
> >Staements about non-0existent things are meaningless (they may be defined
> >as false because they are meaningless, but this is definitional).
> >
> "Well if it was Candy, we're in for a feast."
>
> Meaningless? False?
I don't agree that statements about non-existing things are meaningless,
but I don't think your example is what lojbab meant. In the universe of
that discourse, Candy is an existing pig.
- i ja'a go'i la kandis inaja ma'a pu'o citka salci la kandis
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Transparence / Opaqueness
la kris cusku di'e
> >> mi xanto kalte I'm hunting elephants!
> >> .i go'i ma? Which one(s) in particular?
> >> .i go'i ne'e None in particular.
> >> OR
> >> .i go'i la dambos. I'm hunting elephants, specifically Dumbo
> >> .i .o'onai do palci Oooh, you're evil!
> >
> >If you replace {ne'e} with {xe'eda}, I think I agree.
>
> I thought we were discussing using tanru *instead of* xe'e. Anyway, try
> replacing "ne'e" with "zi'o" and tell me what you think.
Yes, it makes some sense. I wonder if we can interpret zi'o in general as
"none in particular".
> All the proposals I've read and understood here, from assuming all sumti are
> opaque, to flagging some with xe'e, reject the notion that bridi should
> describe relationships among things referred to by sumti (since sumti, in
> either scheme, no longer necessarily have referrents).
I think you are right. There is one other case of sumti without proper
referents: le'e and lo'e. I think {mi nitcu lo'e tanxe} might also be
a possible solution (and this time there is no problem with needing
two boxes).
> At least with xe'e
> the anomaly is marked, but it seems like a weird solution since xe'e only
> seems to be needed in certain limited contexts (namely, the x2 place of half
> a dozen gismu)
In those contexts is where it would be most common, but I think it can be
used more generally. In any case, we should find some solution for those
contexts.
> Specifically, I don't think the transparency/opaqueness distinction exists
> elsewhere. I can't think what distinction the word "xe'e" could be making
> in the sentence: "xe'e lo blanu cu nelci la djan.".
Can I change it to "la djan nelci xe'e lo blanu"?
That would be "John likes anything that is blue", without asserting that
John likes everything that is blue. I know, it's weird, but again it is
very similar to {la djan nelci lo'e blanu} = "John likes the typical blue
thing". You can't conclude from that that John likes every blue thing,
either.
Maybe {lo'e} is the opaque gadri and we had it there all the time without
realizing it.
It even works for And's sofa:
re lo'e prenu cu kakne le nu zutse le sfofa
two (any/typical) persons can seat on the sofa.
mi nitcu lo'e tanxe
I need a (any/typical) box.
mi pu'o tcidu re lo'e selcku
I'm going to read two (any/typical) books.
Opinions?
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: TECH: Transparence / Opaqueness
la xorxes. cusku di'e
> Date: Sat, 1 Oct 1994 16:13:02 EDT
> From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
> Subject: Re: TECH: Transparence / Opaqueness
> la kris cusku di'e
[...]
>> All the proposals I've read and understood here, from assuming all sumti are
>> opaque, to flagging some with xe'e, reject the notion that bridi should
>> describe relationships among things referred to by sumti (since sumti, in
>> either scheme, no longer necessarily have referrents).
>
> I think you are right. There is one other case of sumti without proper
> referents: le'e and lo'e. I think {mi nitcu lo'e tanxe} might also be
> a possible solution (and this time there is no problem with needing
> two boxes).
[...]
>> Specifically, I don't think the transparency/opaqueness distinction exists
>> elsewhere. I can't think what distinction the word "xe'e" could be making
>> in the sentence: "xe'e lo blanu cu nelci la djan.".
>
> Can I change it to "la djan nelci xe'e lo blanu"?
>
> That would be "John likes anything that is blue", without asserting that
> John likes everything that is blue. I know, it's weird, but again it is
> very similar to {la djan nelci lo'e blanu} = "John likes the typical blue
> thing". You can't conclude from that that John likes every blue thing,
> either.
>
> Maybe {lo'e} is the opaque gadri and we had it there all the time without
> realizing it.
>
> It even works for And's sofa:
>
> re lo'e prenu cu kakne le nu zutse le sfofa
> two (any/typical) persons can seat on the sofa.
>
> mi nitcu lo'e tanxe
> I need a (any/typical) box.
>
> mi pu'o tcidu re lo'e selcku
> I'm going to read two (any/typical) books.
>
> Opinions?
I have been working along similar lines. I started from the hypothesis
that a {xe'e} would be needed, and tried assigning it to various
selma'o. LE seemed the best candidate, so I had a look at the
existing LE. {lo'e} seemed to have the best qualities. Of course,
it doesn't mean exactly 'any', you'd need a new LE ~= 'the arbitrary',
but I think that in most circumstances where you use 'any' (in this
meaning) in English you don't actually mean ABSOLUTELY any but just
any typical one(s). So, I might as well agree with you (at least,
until I get {lo} worked out :-)
> Jorge
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: TECH: Transparence / Opaqueness
>Maybe {lo'e} is the opaque gadri and we had it there all the time without
>realizing it.
>
>Opinions?
.ua pluka .i'e
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: TECH: Transparence / Opaqueness
Something still bothers me. If we accept that a structure like
re lo'e remna
two of all the typical human beings
means 'any two [typical] human beings' then
re lo remna
two of all the really are human beings
ought to mean 'any two human beings' as {lo'e} is more restrictive
than {lo}. {lo} requires that the described entity exhibits a
minimal set of characteristics to be considered {the really is},
{lo'e} on the other hand requires that the described entity
exhibits enough of the common characteristics to be considered
{the typical}. I think we need to be able to express {the typical},
so either we accept the above or we have to assign a new cmavo
for {the arbitrary}. Even then this type of structure may be
somewhat dubious.
It is possible to work out quite many things starting from the
above assumptions and remembering that the lojban descriptors
need not map 1:1 to, say, definite/indefinite/transparent/opaque.
But then there will always be the odd case which must be handled
differently.
I have also tried to assign {xe'e} to selma'o PA, but this seemed
to lead to even more trouble in some cases, and might perhaps
require a couple of additional cmavo to get a clean design.
I'm not so sure, anymore, that we need to be able to express
just anything with a simple sentence without resorting to the use of
a prenex or, in the case of real Quinean opaqueness, an explicit
abstraction. I always thought that using a prenex naturally in
everyday language was a feature of Lojban :-)
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: TECH: Transparence / Opaqueness
la veion cusku di'e
> Something still bothers me. If we accept that a structure like
>
> re lo'e remna
> two of all the typical human beings
No, this is confusing. There is no such thing as "all the typical
human beings" as a group from which to select. {lo'e remna} doesn't
have a proper referent. You can't point to someone and say that
they are a {lo'e remna}. (You can say that they are a typical
human, but that is {lo fadni remna}, which should not be confused
with {lo'e remna}). If you are making a claim about {re lo'e remna},
you are NOT claiming that there are two humans for which the claim
is true.
> means 'any two [typical] human beings' then
>
> re lo remna
> two of all the really are human beings
>
> ought to mean 'any two human beings' as {lo'e} is more restrictive
> than {lo}.
{lo'e} is not more restrictive than {lo}. {lo remna} has actual humans
as referents, {lo'e remna} has a platonic ideal as a referent.
> {lo} requires that the described entity exhibits a
> minimal set of characteristics to be considered {the really is},
Agreed.
> {lo'e} on the other hand requires that the described entity
> exhibits enough of the common characteristics to be considered
> {the typical}.
But the described entity doesn't exist in the same realm as the
entities described by {lo}. The {lo'e broda} exhibits the minimal
set of characteristics, and no others. Because it doesn't exist
in reality, it is not forced to have individual characteristics
like any everyday {lo broda}, which might make it in some way
"atypical".
> I think we need to be able to express {the typical},
> so either we accept the above or we have to assign a new cmavo
> for {the arbitrary}. Even then this type of structure may be
> somewhat dubious.
I'm not sure I see what is the problem you are describing. Perhaps
some examples would be useful. I don't think {lo'e broda} should
be used in the sense of {so'e broda}.
Jorge
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: admirers of 50% of symphonies
I have been trying & failing to render into Jbobau three different
interpretations of "(the) admirers of 50% of Brahms's symphonies"
[a nonspecific pair of his 4 symponies].
(1) Each admirer admires some but not necessarily each of this 50%
of symphonies
(2) Each admirer admires each of the same 50%
(3) Each admirer admires each of a possibly different 50%
In the sentence "I met all the admirers of 50% of Brahms's symphs"
these interpreations give us, respectively:
(1) There are two of B's S's such that I met all their admirers (everyone
who admires either of them).
(2) There are two of B's S's such that I met everyone who admires
both of them.
(3) I met everyone who admires two of B's S's.
Any of our didactic friends care to take this on for me?
---
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: admirers of 50% of symphonies
> I have been trying & failing to render into Jbobau three different
> interpretations of "(the) admirers of 50% of Brahms's symphonies"
> [a nonspecific pair of his 4 symponies].
Using {neltce} for admire, and {zgikrsimfoni,a} for Symphony:
> (1) Each admirer admires some but not necessarily each of this 50%
> of symphonies
lei neltce be su'o le re le zgikrsimfoni,a
The admirers of at least one of the two of the symphonies.
> (2) Each admirer admires each of the same 50%
lei neltce be le re le zgikrsimfoni,a
The admirers of the two of the symphonies.
> (3) Each admirer admires each of a possibly different 50%
lei neltce be re le zgikrsimfoni,a
The admirers of two of the symphonies.
> In the sentence "I met all the admirers of 50% of Brahms's symphs"
> these interpreations give us, respectively:
>
> (1) There are two of B's S's such that I met all their admirers (everyone
> who admires either of them).
> (2) There are two of B's S's such that I met everyone who admires
> both of them.
> (3) I met everyone who admires two of B's S's.
In (1) and (2), the people you met may or may not also admire the other
two symphonies.
> Any of our didactic friends care to take this on for me?
Didactic? I'm not didactic.
Jorge
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: admirers of 50% of symphonies
>> (1) Each admirer admires some but not necessarily each of this 50%
>> of symphonies
>
>lei neltce be su'o le re le zgikrsimfoni,a
>The admirers of at least one of the two of the symphonies.
What would it mean if you changed the first "le" to "lo"? Would that make
the count (2) of symphonies veridicial but still allow for loose use of the
word "symphony"? Does the "le" in your original sentence allow for, say
"2000" to be meant by "re", if it's understood in context?
>Didactic? I'm not didactic.
No need to be defensive -- everyone here is polydactic. zo'o
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: admirers of 50% of symphonies
In-Reply-To: (Your message of Tue, 04 Oct 94 20:04:34 EDT.)
> > (1) Each admirer admires some but not necessarily each of this 50%
> > of symphonies
>
> lei neltce be su'o le re le zgikrsimfoni,a
> The admirers of at least one of the two of the symphonies.
>
> > (2) Each admirer admires each of the same 50%
>
> lei neltce be le re le zgikrsimfoni,a
> The admirers of the two of the symphonies.
>
> > (3) Each admirer admires each of a possibly different 50%
>
> lei neltce be re le zgikrsimfoni,a
> The admirers of two of the symphonies.
How does this give us "50% of", or "0.5 of", or "one in two of"?
I want to be able to say this without knowing the total number
of symphonies. And in one of the cases do I have in mind a specific
subset of symphonies in mind.
---
And
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: admirers of 50% of symphonies
> >> (1) Each admirer admires some but not necessarily each of this 50%
> >> of symphonies
> >
> >lei neltce be su'o le re le zgikrsimfoni,a
> >The admirers of at least one of the two of the symphonies.
>
> What would it mean if you changed the first "le" to "lo"?
lei neltce be lo re le zgikrsimfoni,a
The admirers of at least one of two of the symphonies.
You are not being specific as to which two of the symphonies you're talking
about.
> Would that make
> the count (2) of symphonies veridicial but still allow for loose use of the
> word "symphony"?
The veridicality of {lo} is not its main feature, and I think it is a mistake
to insist so much with it. {lo} has to be veridical because it is non-specific.
If you don't require veridicality, you lose all meaning because of the
non-specificity.
But you can't use veridicality to disringuish the meanings of {le} and {lo}.
99.9 % of the time {le} will also be veridical. Because of its specificity,
you can allow for not quite veridicality and still convey meaning, but this
IMHO is a secondary feature of {le}.
The difference between {le} and {lo} is specificity. With {lo} you don't
give any indication as to what are the referents that make the claim true.
You only say that there are such referents. With {le}, you have those
referents in mind, and if you are kind with your audience you will make
sure it is clear to them too what are the referents. With {lo} you are
making a general claim, with {le} you are talking about a particular
situation.
> Does the "le" in your original sentence allow for, say
> "2000" to be meant by "re", if it's understood in context?
Maybe, but that's not the main difference between the two claims.
> >Didactic? I'm not didactic.
>
> No need to be defensive -- everyone here is polydactic. zo'o
Or polydactyl maybe.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: admirers of 50% of symphonies
> How does this give us "50% of", or "0.5 of", or "one in two of"?
> I want to be able to say this without knowing the total number
> of symphonies.
You should have said so before :). If you tell me they are four then 50%
is two. Now, let me try without knowing the total number:
> > > (1) Each admirer admires some but not necessarily each of this 50%
> > > of symphonies
lei neltce be pisu'o le pimu lei zgikrsimfoni,a
The admirers of at least part of the half of the symphonies.
> > > (2) Each admirer admires each of the same 50%
lei neltce be le pimu lei zgikrsimfoni,a
The admirers of the half of the symphonies.
> > > (3) Each admirer admires each of a possibly different 50%
lei neltce be pimu lei zgikrsimfoni,a
The admirers of half of the symphonies.
> And in one of the cases do I have in mind a specific
> subset of symphonies in mind.
(a) {lei zgikrsimfoni,a} is the whole mass of symphonies under discussion.
(b) {pimu lei zgikrsimfoni,a} is a non-specific half of them.
(c) {le pimu lei zgikrsimfoni,a} is the specific half/halves you have in mind.
(d) {lo pimu lei zgikrsimfoni,a} is one or more halves (non specific).
(e) {pisu'o le pimu lei zgikrsimfoni,a} is a non specific part of the
specific half you have in mind.
(f) {pisu'o lo pimu lei zgikrsimfoni,a} is a non specific part of a non
specific half.
And so on...
(I'm making all this up as I go, so there is the slight possibility that
I got it all wrong :)
Now:
{neltce lei zgikrsimfoni,as} means "admires the symphonies".
So:
{lo neltce be lei zgikrsimfoni,a} is at least one of those who fit
that predicate, so at least one of the admirers of (all) the symphonies.
Similarly:
{lo neltce be pimu lei zgikrsimfoni,a} is at least one of those who
admire some half of the symphonies. If you have more than one, they
don't have to admire the same half (I think). If you want them all
admiring the whole of one non-specific half, you'd have
{lo neltce be piro lo pimu lei zgikrsimfoni,a}
I think it all makes sense, but I wouldn't swear by it.
Jorge
Date: Thu, 6 Oct 1994 01:17:40 -0400
From: Logical Language Group <lojbab>
Subject: Re: A couple of questions
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
CS>1) Is there any difference between "lo broda cu brode" and "da poi broda cu
CS> brode"? If they are the same, the statement "lo [unicorn] cu brode"
CS> should be false, since noda cu [unicorn].
lo broda is not the same as da poi broda, and this is specifically one of
the differences - there is no claim that the referent exists in the unoverse
of discourse. Beyond that, however, is where the current debate seems to be
hanging - what else does "lo" mean? I am going to hang back from opining
further for now.
CS>2) Has it occured to anyone that "needing" something might not imply
CS>"needing
CS> to have it"? I might say in English, "I need a clean environment," yet
CS> I don't want to have (i.e., to possess) the clean environment. Perhaps
CS> this is yet another argument for constraining the x2 place of {nitcu}
CS> to abstractions.
I first note that there is "sarcu" which may apply to some of these cases.
Otherwise, you could well be right that an abstraction is in order. The
problem I see is that there are a LOT of places where intensionality is an
option, and it is currently seeming that intensionality requires an abstraction
place. But for optional intensionality situations (I do not see that "nitcu"
must ALWAYS be intensional), things are very murky.
CS>3) Does the sentence
CS>
CS> mi djuno ledu'u do djuno ledu'u makau blanu
CS> mean "I know you know what is blue" or "I know what you know to be
CS>blue"?
CS> Instinctively, the former should be correct, and the latter meaning can
CS> be expressed by
CS>
CS> mi djuno ledu'u do djuno ledu'u makauxire blanu
CS>
CS> Am I right?
Since "kau" is a discursive, it cannot be subscripted, so your solution is
rather vague in meaning - you have really subscripted the "ma".
I would do the second as
mi djuno tu'a makau poi do djuno ledu'u ke'a blanu
lojbab
From: "Mark E. Shoulson" <shoulson@CS.COLUMBIA.EDU>
Subject: Re: A couple of questions
In-Reply-To: <199410060453.AAA23937@cs.columbia.edu> (message from Chung-chieh
Shan on Thu, 6 Oct 1994 12:49:42 CST)
>Date: Thu, 6 Oct 1994 12:49:42 CST
>From: Chung-chieh Shan <ken%MATH.NTU.EDU.TW@CUVMB.CC.COLUMBIA.EDU>
>X-To: lojban@cuvmb.cc.columbia.edu
>2) Has it occured to anyone that "needing" something might not imply "needing
> to have it"? I might say in English, "I need a clean environment," yet
> I don't want to have (i.e., to possess) the clean environment. Perhaps
> this is yet another argument for constraining the x2 place of {nitcu}
> to abstractions.
I sort of thought that "nitcu" didn't imply "ponse"; that you'd need a
tanru/lujvo for that in any case (mi nitcu lenu mi ponse.../mi ponse
nitcu... or assorted sumti-raising: mi nitcu tu'a ...). So your example
would probably be simplest stated as "mi nitcu lenu le vanbi be mi cu
jinsa".
>----------
>Chung-chieh (Ken) Shan ken@cauchy.math.ntu.edu.tw
>"Ay, fashion you may call it. Go to, go to." -- Hamlet
~mark
From: i.alexander.bra0125@oasis.icl.co.uk
Subject: Re: A couple of questions
> CS>1) Is there any difference between "lo broda cu brode" and "da poi broda cu
> CS> brode"? If they are the same, the statement "lo [unicorn] cu brode"
> CS> should be false, since noda cu [unicorn].
cu'u la lojbab.
> lo broda is not the same as da poi broda, and this is specifically one of
> the differences - there is no claim that the referent exists in the unoverse
> of discourse.
...
I still can't help feeling that this is BOGUS.
I never did understand this concept, and I think the recent discussions
have helped clarify the situation.
Let me try to change your mind.
You seem to be saying that {lo broda cu brode}, e.g
(1) {lo mlatu je nanmu cu blanu}
could be true, even if there is
no such thing as a cat-man ***in the universe of discourse***
(far less the real world). This doesn't make any sense to
me whatsoever - I can't think of any interpretation of (1)
which doesn't imply existence.
I believe that all the situations where you might think you
needed such a concept are better handled in other ways.
Intensional contexts:
e.g. {nitcu}, {djica}, where there is an abstraction which
the typical NL elides, but we strongly encourage to be
acknowledged in Lojban.
Veijo's "You may choose two books" is similar. The abstraction
is compressed rather than totally elided in English -
{curmi lenu do cuxna re cukta}. (There are some other issues
in this translation which I'll skip over for the time being.)
Typical objects:
e.g. "I like an apple", which is {mi nelci lo'e plise}.
Some of these might even use {ro}.
Hypothetical objects:
e.g. the long-sought unicorn, which is {lo [da'i] pavyseljirna}.
The {da'i} is recommended for clarity, but we often live without it.
I may have forgotten some of the problem cases, but my instinct is
that they will all be soluble.
I strongly recommend that {lo broda} be defined as equivalent to
{da poi broda}, at least as far as the formal definition goes.
If we feel the need to allow some sort of laxity in informal
jbosku, that's a different matter - we all live with various
kinds of indiscretion in the rough and tumble of live usage -
but I don't like muddying the basics.
> CS>3) Does the sentence
> CS>
> CS> mi djuno ledu'u do djuno ledu'u makau blanu
> CS> mean "I know you know what is blue" or "I know what you know to be
> CS>blue"?
> CS> Instinctively, the former should be correct, and the latter meaning can
> CS> be expressed by
> CS>
> CS> mi djuno ledu'u do djuno ledu'u makauxire blanu
> CS>
> CS> Am I right?
> Since "kau" is a discursive, it cannot be subscripted, so your solution is
> rather vague in meaning - you have really subscripted the "ma".
Eh, what? I thought he'd got it 100% right. Last time I heard,
that WAS the way we distinguish nested constructs, including {kau}.
Is the disposition of the subscript a serious problem?
> I would do the second as
> mi djuno tu'a makau poi do djuno ledu'u ke'a blanu
I'm not sure either way about this as it stands, but it
certainly doesn't extend well to deeper nesting.
mu'o mi'e .i,n.
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: A couple of questions
In-Reply-To: (Your message of Thu, 06 Oct 94 12:49:42 CST.)
Ken:
> 1) Is there any difference between "lo broda cu brode" and "da poi broda cu
> brode"? If they are the same, the statement "lo [unicorn] cu brode"
> should be false, since noda cu [unicorn].
No difference.
> 2) Has it occured to anyone that "needing" something might not imply "needing
> to have it"? I might say in English, "I need a clean environment," yet
> I don't want to have (i.e., to possess) the clean environment. Perhaps
> this is yet another argument for constraining the x2 place of {nitcu}
> to abstractions.
Yes it has occurred to someone.
> 3) Does the sentence
>
> mi djuno ledu'u do djuno ledu'u makau blanu
>
> mean "I know you know what is blue" or "I know what you know to be blue"?
> Instinctively, the former should be correct,
Yes,
> and the latter meaning can be expressed by
>
> mi djuno ledu'u do djuno ledu'u makauxire blanu
>
> Am I right?
I think you mean "I know what x is such that you know x is blue".
(Rather than "I know x such that you know x is blue".)
Clever question. I wonder leduhu xukau you are right.
---
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: A couple of questions
In-Reply-To: (Your message of Thu, 06 Oct 94 17:24:44 A.)
Iain (replying to Lojbab):
> You seem to be saying that {lo broda cu brode}, e.g
> (1) {lo mlatu je nanmu cu blanu}
> could be true, even if there is
> no such thing as a cat-man ***in the universe of discourse***
> (far less the real world). This doesn't make any sense to
> me whatsoever - I can't think of any interpretation of (1)
> which doesn't imply existence.
Without wanting to disagree with you, I can think of a possibly
valid interpretation that doesn't imply existence.
"ro mlatu je nanmu cu blanu" doesn't imply existence, & we
can gloss this as "100% of catmen are blue".
We might therefore take "lo mlatu je nanmu cu blanu" to mean
"more than 0% of catmen are blue", again not implying existence.
(Just to confuse matters, English universal quantifiers usually
imply existence, unlike Jbobau's or logic's.)
---
And
From: Chung-chieh Shan <ken@MATH.NTU.EDU.TW>
Subject: A couple of questions
As the subject promised... a couple of questions.
1) Is there any difference between "lo broda cu brode" and "da poi broda cu
brode"? If they are the same, the statement "lo [unicorn] cu brode"
should be false, since noda cu [unicorn].
2) Has it occured to anyone that "needing" something might not imply "needing
to have it"? I might say in English, "I need a clean environment," yet
I don't want to have (i.e., to possess) the clean environment. Perhaps
this is yet another argument for constraining the x2 place of {nitcu}
to abstractions.
3) Does the sentence
mi djuno ledu'u do djuno ledu'u makau blanu
mean "I know you know what is blue" or "I know what you know to be blue"?
Instinctively, the former should be correct, and the latter meaning can
be expressed by
mi djuno ledu'u do djuno ledu'u makauxire blanu
Am I right?
----------
Chung-chieh (Ken) Shan ken@cauchy.math.ntu.edu.tw
"Ay, fashion you may call it. Go to, go to." -- Hamlet
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: admirers of 50% of symphonies
In-Reply-To: (Your message of Wed, 05 Oct 94 19:54:21 EDT.)
I can try to do my 3 meanings in rough pred.calc. form:
"I met every admirer of 50% of B's symphonies"
(3) Ax [Ey [y is a set containing 50% of B's symphonies]
& [Az [z is a member of y] -> x admires z]] -> I met x
(2) Ey [y is a set containing 50% of B's symphonies]
& [Ax [Az [z is a member of y] -> x admires z] -> I met x]
(1) Ey [y is a set containing 50% of B's symphonies]
& [Az [z is a member of y] Ax [x admires z] -> I met z]
Ignore my previous explanations of what I wanted to say, and try these.
> > > > (1) Each admirer admires some but not necessarily each of this 50%
> > > > of symphonies
>
> lei neltce be pisu'o le pimu lei zgikrsimfoni,a
> The admirers of at least part of the half of the symphonies.
I meant "be pisuho **lo** pimu lei Z".
But this doesn't work. I want the additional meaning that 50% of B's
symphonies are such that I met lei admirers of either of them. A context
in which I met admirers of #2 & admirers of #2 fits what I mean, but
not a context in which I met only admirers of #2.
> > > > (2) Each admirer admires each of the same 50%
>
> lei neltce be le pimu lei zgikrsimfoni,a
> The admirers of the half of the symphonies.
I don't have a specific half in mind. I want "There are entities
constituting 50% of B's symphonies such that I met lei people
who admire each of them".
You offer this, for this meaning:
{lo neltce be piro lo pimu lei zgikrsimfoni,a}
But to me this means that each admirer admires eaho osbydfeet5%-ie nepeain()
>>>> 3 ahamrramrsec fapsil ifrn 0
e eteb iulizirifn,
h dieso afo h ypois
O.S 3 stemaigyugtfr"e eteb opm e "
o edt nwhwt e htImatfr()&()
>>Adi n ft cssdIhvenmn scfc ^<notnt yo en i oeo h ae">>sbe fsmhne nmn.
--Ad
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: A couple of questions
> 1) Is there any difference between "lo broda cu brode" and "da poi broda cu
> brode"?
I agree with Iain and And, and disagree with lojbab. They should mean the same
thing.
> If they are the same, the statement "lo [unicorn] cu brode"
> should be false, since noda cu [unicorn].
If no unicorns exist in the world where the statement is used, then the
statement is false in that world, yes.
> 3) Does the sentence
>
> mi djuno ledu'u do djuno ledu'u makau blanu
>
> mean "I know you know what is blue" or "I know what you know to be blue"?
> Instinctively, the former should be correct, and the latter meaning can
> be expressed by
>
> mi djuno ledu'u do djuno ledu'u makauxire blanu
>
> Am I right?
pe'i go'i, you're right.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: A couple of questions
And:
> Iain (replying to Lojbab):
> > (1) {lo mlatu je nanmu cu blanu}
> > - I can't think of any interpretation of (1)
> > which doesn't imply existence.
>
> Without wanting to disagree with you, I can think of a possibly
> valid interpretation that doesn't imply existence.
> "ro mlatu je nanmu cu blanu" doesn't imply existence, & we
> can gloss this as "100% of catmen are blue".
Agreed.
> We might therefore take "lo mlatu je nanmu cu blanu" to mean
> "more than 0% of catmen are blue", again not implying existence.
But {lo} has an "at least one" quantifier, not "at least some %".
Otherwise, su'o doesn't work as the negation of ro.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: admirers of 50% of symphonies
> I can try to do my 3 meanings in rough pred.calc. form:
> "I met every admirer of 50% of B's symphonies"
>
> (3) Ax [Ey [y is a set containing 50% of B's symphonies]
> & [Az [z is a member of y] -> x admires z]] -> I met x
> (2) Ey [y is a set containing 50% of B's symphonies]
> & [Ax [Az [z is a member of y] -> x admires z] -> I met x]
> (1) Ey [y is a set containing 50% of B's symphonies]
> & [Az [z is a member of y] Ax [x admires z] -> I met z]
>
> Ignore my previous explanations of what I wanted to say, and try these.
>
> > > > > (1) Each admirer admires some but not necessarily each of this 50%
> > > > > of symphonies
> >
> > lei neltce be pisu'o le pimu lei zgikrsimfoni,a
> > The admirers of at least part of the half of the symphonies.
>
> I meant "be pisuho **lo** pimu lei Z".
> But this doesn't work. I want the additional meaning that 50% of B's
> symphonies are such that I met lei admirers of either of them. A context
> in which I met admirers of #2 & admirers of #2 fits what I mean, but
> not a context in which I met only admirers of #2.
This description doesn't match your pred.calc. form. If there are no
admirers of one of the symphonies, the context you want to exclude is
allowed there and in the Lojban form. If none of the symphonies have
zero admirers, then the Lojban form doesn't allow that context either,
just like your pred.calc.
> > > > > (2) Each admirer admires each of the same 50%
> >
> > lei neltce be le pimu lei zgikrsimfoni,a
> > The admirers of the half of the symphonies.
>
> I don't have a specific half in mind. I want "There are entities
> constituting 50% of B's symphonies such that I met lei people
> who admire each of them".
>
> You offer this, for this meaning:
> {lo neltce be piro lo pimu lei zgikrsimfoni,a}
> But to me this means that each admirer admires eaho osbydfeet5%-ie nepeain()
> >>>> 3 ahamrramrsec fapsil ifrn 0
>
> e eteb iulizirifn,
> h dieso afo h ypois
> O.S 3 stemaigyugtfr"e eteb opm e "
> o edt nwhwt e htImatfr()&()
> >>Adi n ft cssdIhvenmn scfc ^<notnt yo en i oeo h ae">>sbe fsmhne nmn.
> --Ad
You became incoherent in the middle of a sentence (or at least so it appears
from my terminal...)
I can guess at your objection. I think we can define {piro lo pimu lei zy}
to be different from {pimu lei zy}. The first one has one level less of
non-specificity, so that admirers of that would all have to admire the same
thing, while admirers of the second are allowed to admire different things.
What do you think?
Jorge
Date: Fri, 7 Oct 1994 03:19:46 -0400
From: Logical Language Group <lojbab>
Subject: Re: A couple of questions
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
IA>Hypothetical objects:
IA> e.g. the long-sought unicorn, which is {lo [da'i] pavyseljirna}.
IA>The {da'i} is recommended for clarity, but we often live without it.
IA>
IA>I may have forgotten some of the problem cases, but my instinct is
IA>that they will all be soluble.
But if this is "da poi pavyseljirna" you are not being hypothetical at all
you are CLAIMING existence: da zo'u da
pavyseljirna.
I can accept discursive marking with "da'i" in non-logical discussions, but
da'i seems incompatible with 'the present universe of discourse', it
specifically implies to me that we are moving OUT of said universe.
At which point "da poi pavyseljirna" is highly questionable to me.
lojbab
Date: Fri, 7 Oct 1994 04:48:21 -0400
From: lojbab
Subject: Re: A couple of questions
Cc: lojbab@access.digex.net
JL>> If they are the same, the statement "lo [unicorn] cu brode"
JL>> should be false, since noda cu [unicorn].
JL>
JL>If no unicorns exist in the world where the statement is used, then the
JL>statement is false in that world, yes.
1. Therefore the statement "Elves have pointed ears" is false since there is
no such thing as an elf. Likewise definitional statements "Elves are
humanoid" is also false even if definitional. How can you describe the
properties of a hypothetical but non-existent object if any statement about
such an object is false.
2. If statements about non-existent objects are false, then their negation
is true. We can possibly weasel around this with "na" negation (and I think
I did in the negation paper), but I am not sure.
3> And then there is the argument that all statements about non-existent
objects being equivalent to each other, since all are statements about the
members of the empty set.
I don't pretend to know the answers, but this is one of those questions that
comes up again and again and I never am satisfied enough with whatever
explanation is proposed to internalize it.
But the status quo remains, as far as I know, that "lo [unicorn] cu brode"
is not the same as da poi [unicorn] cu broda. Cowan or pc are welcome to
correct me, since they supposedly reolved this once before.
lojbab
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: admirers of 50% of symphonies
In-Reply-To: (Your message of Thu, 06 Oct 94 21:53:02 EDT.)
Jorge:
> > (1) Ey [y is a set containing 50% of B's symphonies]
> > & [Az [z is a member of y] Ax [x admires z] -> I met z]
> >
> > I meant "be pisuho **lo** pimu lei Z".
> > But this doesn't work. I want the additional meaning that 50% of B's
> > symphonies are such that I met lei admirers of either of them. A context
> > in which I met admirers of #2 & admirers of #2 fits what I mean, but
read: #2 #4
[I really fucked up that message]
> > not a context in which I met only admirers of #2.
>
> This description doesn't match your pred.calc. form. If there are no
> admirers of one of the symphonies, the context you want to exclude is
> allowed there and in the Lojban form. If none of the symphonies have
> zero admirers, then the Lojban form doesn't allow that context either,
> just like your pred.calc.
Hmm. You're right, I think. I can't even do my meaning in pred calc.
let alone Lojban. I want "I met every admirer of some symphonies,
and these admired symphonies constitute 50% of B's total".
> I can guess at your objection. I think we can define {piro lo pimu lei zy}
> to be different from {pimu lei zy}. The first one has one level less of
> non-specificity, so that admirers of that would all have to admire the same
> thing, while admirers of the second are allowed to admire different things.
> What do you think?
I think that it may be possible to stipulate such a rule for Lojban,
but that it doesn't follow from the present semantics as I (obviously
in a most confused way) understand them.
---
And
From: David Matuszek <dave@VFL.PARAMAX.COM>
Subject: Re: A couple of questions
In-Reply-To: Logical Language Group's message of Fri,
7 Oct 1994 08:40:18 -0400
<9410071257.AA28341@arbor.VFL.Paramax.COM>
lojbab writes:
> 1. Therefore the statement "Elves have pointed ears" is false since
> there is no such thing as an elf. Likewise definitional statements
> "Elves are humanoid" is also false even if definitional. How can you
> describe the properties of a hypothetical but non-existent object if any
> statement about such an object is false.
Whether universal quantification has existential import has been
argued extensively by philosophers and logicians, a lot of it during
the Middle Ages when Aristotle reigned supreme. There is no "correct"
answer. The standard interpretation in modern formal logic is that it
does not have existential import; only the explicit existential
quantifier has that. Again this is not "correct," merely what
logicians have decided is convenient.
However, if definitional statements can be false, then it's going to
be damned hard to ever do mathematics in Lojban. There will be no way
to prove something doesn't exist if you cannot define it in the first
place because the definition itself is false.
I think the question is whether you want to reinvent logic, or just
accept it as it exists today and work from there.
-- dave@vfl.paramax.com -- If my header says otherwise, it lies.
In memoriam: The Space Age, 1969-1972.
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: lo'e / lo / da / 50%
lo'e / lo / da / And's odd couple / 50%
I think I'm converging towards a coherent interpretation
which hopefully doesn't have (m)any major holes in it :-)
--------------------
I hope I can muster enough willpower to keep out of the
discussion after this. This takes just too much time and
has started to haunt me even in my dreams (the other day
I forgot to step out of the commuter train at my customary
station :-) I rather ought to concentrate on memorizing
cmavo and WRITING in Lojban - the Cafe is there waiting
for new stuff.
--------------------
A. practical considerations (mostly)
I've read and re-read the part of the draft reference grammar
which deals with {lo'e}. Clearly the aim was to be able to
state things like
lo'e cinfo cu xabju le fi'ortu'a
The [typical] lion lives in Africa
This in no way implies that ANY lion what-so-ever lives in
Africa - the construct was specifically designed to avoid that
kind of interpretation.
Much of the discussion concerning {any} was muddied (by ME and
others) with unnecessary complications related to Quinean
opaqueness, abstraction and sumti raising.
The English word {any} is used mainly in sentences containing
implied abstractions/sumti raising, intensionality, ability
or potentiality. I push the associated problems aside and try
to express indefiniteness in a way which, as far as I can see,
allows about the same expression to be used both for actualized
and for potential events.
If I have eaten/eat an apple I can say
(1) mi citka pa plise
I ate/eat one [which could have been/can be any one]
of the really are apples
Speaking about ability/potentiality I can say
(2) mi ka'e citka pa plise
I can eat one apple [any one of the really...]
This statement says nothing about existence. If I, OTOH,
want to say that there is an apple I can eat, I say
(3) mi ka'e citka pa da poi plise
There is one apple such that I can eat it.
and going towards definiteness
(4) mi ka'e citka le bi'u pa plise
I can eat this one apple.
(NB. 'this' is not used in the sense of {le pa vi}
but like 'there was this...')
(5) mi ka'e citka le pa plise
I can eat the one apple.
The difference between
(6) mi citka lo plise
I eat an apple. (could be just any apple)
I eat [one or more which can be just any of
all the really are] apples.
and
(7) mi citka da poi plise
There is an apple such that I eat it.
is that (7) claims the existence of the apple and
simultaneously makes a predication concerning me and the
said apple, whereas (6) makes a predication the truth
value of which is conditional to the said apple(s)
existing - but it doesn't claim the existence.
(NB. I discuss this difference in more depth in the
second part of this posting)
The difference between
(8) mi terpa tu'a lo'e cinfo
I'm afraid of the lion
and
(9) mi terpa tu'a lo cinfo
I'm afraid of lions
I'm afraid of any lion
is that (8) states that I'm afraid of the hypothetical
typical idealized {cilce cinfo poi xabju le fi'ortu'a}
but not necessarily of the odd household pet lion living
nextdoor, (9) OTOH states that I'm terrified by the really
are lions in general.
The difference between (8) and
(10) mi terpa tu'a le/lo fadni cinfo
I'm afraid of ordinary lions
is that (8) speaks about an abstract entity, whereas (10)
is concerned with real, flesh-and-blood, run-of-the-mill
lions, ordinary in some specific though unspecified way.
The differences between 8/9/10 are real and not even
very subtle - and we need to be able to express them.
With {fadni} we have some concrete quality(-ies) which
define(s) the reference set, with {lo'e} we have a more
or less vague collection of qualities and finally with
{lo} we just expect the referent really to be a lion.
Speaking about And's two people sitting in a chair, I
think we might preferably speak about pairs of people
instead of two individual people. So we would have either
{ [so'e] da poi remna remei} for [at least] one existing
pair who can sit in the chair or { lo remna remei } for
just any pair in general, with no reference to existence.
{ lo'e remna remei } is an abstract, Platonian, ideal pair
of people possessing qualities that are fit for a pair of
human beings - it isn't just any odd couple you might bump
into in the streets.
{lo cukta} isn't opaque in Quinean sense, just indefinite,
but it can be used to express {any} in appropriate
circumstances, i.e. within abstractions, with {ka'e} and
the like, with aspects and reasonably often in quite fadni
bridi.
(11) mi tcidu re lo cukta
I read two books (just two books, could be any)
(the emphasis is on the quantity)
(12) mi tcidu re da poi cukta
There are two books such that I read them
I read some two books.
(the emphasis is on the existence)
(13) mi tcidu lo'e cukta
I read books (Books, in abstract)
I am a book reader
(the emphasis is on the activity)
(mi cukta tcidu)
Even though the implicit quantifiers are defined for {lo'e}
I think using explicit ones (or even considering the
implicit ones) wouldn't be appropriate in most normal
circumstances. You just don't quantify ideals.
A few more examples
(14) ko'a kalte lo'e xanto
He hunts the typical elephant.
He is an elephant hunter.
(semi-opaque, idealized)
(15) ko'a kalte lo xanto
He hunts an elephant.
He hunts elephants (could be any ones, even
escaped circus elephants).
He hunts any elephant.
(indefinite, veridical, no a priori claim of
existence)
(NB. you can hunt a NONEXISTING lo xanto.
A couple of years back we had a case, here in
Finland, of tens of people hunting for a non-
existing lo cinfo. They were hunting one
which was supposed to have escaped from a
circus which was supposed to have been there!
It turned out there had been this practical
joker who had used a spoon to prepare a
reasonable facsimile of lion footprints in
soft mud somewhere. Well, perhaps we ought to
call it {le cinfo} :-)
ko'a kalte re lo xanto
He hunts for (just any) two elephants
(16) ko'a kalte da poi xanto
There are elephants such that he hunts them.
He hunts some elephant(s).
(existential, semidefinite, veridical)
ko'a kalte re da poi xanto
There are two elephants such that he hunts them.
(17) ko'a kalte le'e xanto
He hunts Elephants (the ones with the great
ivory tusks, you know :-)
He hunts what I consider really to be elephants.
(semi-opaque, subjectivized)
(18) ko'a kalte le bi'u xanto
He hunts this elephant (the one which I haven't
mentioned so far)
He hunts an elephant.
(semidefinite, existential, non-veridical)
ko'a kalte le bi'u re xanto
He hunts these two elephants
ko'a kalte re le bi'u xanto
He hunts for two of these elephants.
(19) ko'a kalte le xanto
He hunts the elephant(s)
He hunts something I call an elephant.
(definite, existential, non-veridical)
ko'a kalte le re xanto
He hunts the two elephants.
ko'a kalte re le xanto
He hunts for (any) two of the elephants
ko'a kalte le re le xanto
He hunts for the specific two of the elephants
ko'a kalte re da pe le xanto
There are two of the elephants such that he hunts
them
B. about the difference between {da poi broda} and {lo broda}
There is a fundamental difference between
(a) mi broda da poi brode
and
(b) mi broda lo brode
{da poi brode} is a variable definition imported from the prenex.
The normal form would be
(c) da poi brode zo'u mi broda da
This is an existential claim which says that there is at least
one such {da poi brode} that {mi broda da} is true. The claim
(d) mi broda re da
is equivalent to the claim
(e) da de zo'u mi broda da .e de
On the other hand (b) makes no existential claims as far as
{lo brode} is concerned. The variable in the statement is the
elided outer quantifier which has the value {su'o} in this case.
The expanded form of (b) is
(f) ny. goi su'o zo'u mi broda vei ny. lo brode
What we are claiming primarily is the quantity, {lo brode} is
in a way just a unit of measure. The actual relationship is
between {mi} and {piro lo'i vei ny. lo brode}, between me and
a set of entities taken, in principle, at random from the base
set {lo'i ro brode}. The statement can be true if such set exists
in the universe of discourse (the {lo cinfo} we were hunting in
Finland existed in theuniverse of discourse though not in the
material universe). The point to note is that the members of the
subset {could have been | can potentially be} ANY members of the
base set but this doesn't mean that we claim {ro da poi brode zo'u
mi broda da}, we just fix is the quantity. This makes the difference
between
(g) mi ba tcidu re cukta
ny. goi re zo'u mi ba tcidu vei ny. cukta
E(n = 2): I'll read n books
mi ba tcidu piro lo'i re lo cukta
ny. goi re zo'u mi ba tcidu piro lo'i vei ny. lo cukta
E(n = 2): I'll read the whole of a set of n books
and
(h) mi ba tcidu re da poi cukta
da poi cukta ku'o de poi cukta ku'o zo'u mi tcidu da .e de
E(x = a book), E(y = a book): I'll read x and y
This also gives us the two different types of asking 'how many?'
(i) do ba tcidu xo cukta
ny. goi xu zo'u do ba tcidu vei ny. cukta
HOW MANY books are you going to read?
(j) do ba tcidu xo da poi cukta
How many BOOKS there are such that you are going to
read them?
The first question ask for the number of volumes in the sense
that the same book might count as several over the years to come,
the second asks for the number of individuals.
C. How to say '50% of a number of entities' ?
This is quite simple
mi ba tcidu pimu lo'i/le'i pano cukta
I'll read .5 of the set of 10 books
i.e. .5 of a set of n broda instead of .5 of n broda
------
If you won't buy the above, I'll give up and agree with the rest
of you guys. Maybe I'm out of my depth, once more :-)
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: lo'e / lo / da / 50%
In-Reply-To: (Your message of Sat, 08 Oct 94 11:57:20 O.)
Veion:
> (the other day I forgot to step out of the commuter train at my customary
> station :-)
Hey - me too, & I was proccupied with exactly the same matter! Maybe
all over the world Lojbanists are forgetting to get off trains.
> C. How to say '50% of a number of entities' ?
>
> This is quite simple
>
> mi ba tcidu pimu lo'i/le'i pano cukta
> I'll read .5 of the set of 10 books
>
> i.e. .5 of a set of n broda instead of .5 of n broda
I think you'd need "luha pimu" - you read the members of half
a set, not half a set itself.
But my point was that for non-specific reference to less than
100% of brodas, there is a 3-way scope ambiguity. (e.g. in
"I met every admirer of {50% / 2} of Brahms's symphonies".)
For universal quant. & for specifics, there is 2-way scope
ambiguity, if reference is to more than 1 thing. For example,
"lei admirer of {ro/lere} symphony": (1) each admirer admires
each of ro/lere symph; (2) each admirer doesn't necessarily
admire each of ro/lere symph, but the totality of symphonies
admired by lei admirers amounts to ro/lere symphonies.
I've been wondering how these readings (for the 3-way & 2-way
ambiguities) are expressed in Lojban.
................
As for your arguments that (roughly) "lo = any", this solves
the problem, sort of, but I strongly suspect it won't hold
water. But I won't try to argue the point, since there are
dozens of subscribers to this list who are more competent
than me to pronounce on the matter (but haven't so far because
of the indigestible volume of recent traffic).
----
And
Date: Sun, 9 Oct 1994 15:33:07 -0700 (PDT)
From: "John E. Clifford" <pcliffje@crl.com>
Subject: Re: Mystery message
Cc: ucleaar@ucl.ac.uk, veion@xiron.pc.helsinki.fi,
Bob LeChevalier <lojbab@access.digex.net>
In-Reply-To: <9410091402.AA09372@kropotkin.gnu.ai.mit.edu>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
Sorry about that! The book says that ascii messages transmit unchanged
or are modified to fit the system. The book lies royally, apparently,
since my 8k ascii message has become an 11k hotchpotch. Suggestions
(other than learning to write within the internet feeder)?
pc>|83
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: 'any' is a discursive ?!
In the following I'll ignore the cases where 'any' is just 'all'
in disguise.
If we lived in an ideal universe and all the apples were ideal,
the following 2 statements would be equivalent
I'll eat one apple
I'll eat any one apple
However, we don't live in an ideal world, and some apples are rotten,
so, if someone states that he'll eat one apple, we'll understand
that he really means
I'll eat one apple which satisfies certain conditions which
I'm not telling you
mi ba citka pa plise poi ke'a mansa mi zo'e
The word 'any' is an emphatic discursive which is used to indicate that
a statement without explicit restrictions is to be taken at face value,
i.e. as really having no implicit restrictions, perhaps contrary to
common sense.
This means that logically there is no need for 'any', the need is
dictated by praxis. So we need a discursive which means 'Yes, there
really are no restrictions' - but within the bounds of the explicit
statement it has no logical content, it simply denies the existence
of any elided implicit restrictions, affirms the nonexistence of
unexpressed restrictions.
If you consider the statement 'I'll eat any one apple', it doesn't
mean that I'll eat all apples, it just means that I don't apply
any restrictions, it is the statement 'I'll eat one apple, period'.
So, IMHO, 'any' is, in a way, an implicit part of all indefinite,
unrestricted descriptions. When you think you have a case where
this doesn't apply, ask yourself whether the 'any' ought, in fact,
read 'all'.
In Lojban we have 3 choices at this point
1) we invent a new emphatic discursive {xe'e}
or 2) we twist an exiting one to cover this role
or 3) we negate the implicit {any} with a suitable restriction
(3) would be the most logical but not very Zipfian solution.
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: bob@GNU.AI.MIT.EDU
Subject: velju'o/epistemology
faupel@trshp.trs.ntc.nokia.com cuska di'e
....why not solve the problem of not being able to talk about
Elves and the such like by inventing a tense that would get us to
the (imaginary) location in which they do exist.
This has been done using a discursive. Suppose I say:
mi prami satre lo pa jirna xirma
mi pamsa'e lo pavjirnyxirma
I with love stroke/rub that which is truly a unicorn.
You say
ki'a
<confusion!>
I then say
da'i mi pamsa'e lo pavjirnyxirma
The {da'i} means the realm of discourse is hypothetical, I am
*supposing* that I am petting that which is truly a unicorn.
Alternatively, I could claim that unicorns really, truly exist, under
a particular epistemology:
mi pamsa'e lo pavjirnyxirma be vedu'o lo ranmi
I pet that which is truly a unicorn under epistemology myth.
In the Middle ages, people bought and sold narwhale tusks that they
thought were unicorn horns:
su'o lo selgu'e be fi lo ropno
[se gugde]
Some of the people of the land of Europe
puzu'u
for a long interval sometime in the past
vecnu lo jirna
sold that which is truly a horn
be lo pavjirnyxirma be vedu'o lo jitfa
of that which is truly a unicorn by epistemology false
As far as I can see, the problems regarding the veridicality of {lo
pavjirnyxirma} only occur when speaker and listener both agree the
unicorn does not exist in the epistemology of the conversation. But
in that case, it is bad grammar, as well as false, to use {lo} instead
of {le}.
Robert J. Chassell bob@grackle.stockbridge.ma.us
25 Rattlesnake Mountain Road bob@gnu.ai.mit.edu
Stockbridge, MA 01262-0693 USA (413) 298-4725
From: ucleaar <ucleaar@ucl.ac.uk>
Cc: lojban@cuvmb.cc.columbia.edu
Subject: Re: lo [nonexistent]
In-Reply-To: (Your message of Mon, 10 Oct 94 23:35:08 D.) <199410110335.AA20703@access4.digex.net>
Date: Tue, 11 Oct 94 19:28:16 +0100
> UC>> I can't say for sure about "lo'e", but this does not work for "ro elf".
> UC>> If the statement "ro [elf] cu [has pointed ears]" is true, then so is
> UC>> "ro [elf] cu [has unpointed ears]" and "ro [elf] na [has pointed ears]".
> UC>
> UC>I don't see this. But anyway, what matters is what would be true
> UC>if elves exist.
> Well, in English:
> Assume that there are no elves.
> All elves are green is a true statement, because you cannot disprove it
> by showing me a non-green elf.
By what epistemology? So you reckon "If your [=Lojbab's] name is Eric
then mine is Telemachus" is true? I suppose I can supply the answer:
by an epistemology in which any proposition that is not false is
necessarily true.
> You are talking about all members of the empty set
If I am, I'm talking about the empty set that would contain elves if there
were any elves, not the empty set that would contain drinkable instant
coffee if there were any drinkable instant coffee.
Maybe I'm being disingenuous: I recognize that what you say is true
for a respectable epistemology, but I don't think it suffices for
use in human language (e.g. Lojban).
I think we should lay this red herring to bed & recognize that
what does or doesn't exist out there in thw world is irrelevant
to language. Anyone who cares to can sprinkle in the odd "dahi",
according to taste (as Matthew's recent message proposed).
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: 'any' is a discursive ?!
In-Reply-To: (Your message of Tue,
11 Oct 94 08:55:45 O.) <m0qub7h-00005YC@xiron.pc.helsinki.fi>
coi Veion
> > Date: Mon, 10 Oct 1994 21:11:13 +0200
> > From: ucleaar <ucleaar@ucl.ac.uk>
> > Subject: Re: 'any' is a discursive ?!
> >
> > I can't understand how you get the difference between:
> > There are two poems by Rilke that I am willing to read.
> > I am willing to read any two poems by Rilke.
> > Is the former your "da poi" and the latter your "lo"?
>
> Yes. In the first case there exist 2 poems by Rilke such
> that I'm willing to read them, I'm implicitly excluding
> the others. In the second I'm willing to read (any) 2 out
> of all the poems by Rilke, no conditions, just give me 2.
> In a normal conversation I think I might use {le bi'u re
> pemci} as an alternative to {re da poi pemci} unless I'd
> want to emphasize the exclusion. This way I'd have a
> three step gradation at my disposal, from wide open to
> strictly limited.
It would make a big difference if you used "le bihu re pemci".
"Zabna fa le bihu re pemci" would be false only if the two
pemci you had in mind (Rilke's first and Rilke's last, say)
weren't zabna, whereas "Zabna fa re da poi pemci" would be
false only if each member of every pair of poems by Rilke
weren't zabna. [Cf. Jorge's recent laudable polemic/exposition
re LE/LO, and Colin's of a year or two ago.]
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: A couple of questions
In-Reply-To: (Your message of Tue, 11 Oct 94 09:36:26 D.)
John:
> > The lojbanic solution in such cases is usually to invent ways to
> > express both meanings (& to make both expressions "Zipfean" - i.e.
> > verbose in proportion to their infrequency). So I conclude that
> > we need:
> > (1) all, not implying existence
> > (2) all, implying existence
> > (3) some-but-not-necessarily-all, not implying existence
> > [This is the ">0%" I've advocated.]
> > (4) some-but-not-necessarily-all, implying existence
> > (1) is "ro" & (4) is "lo" & "da". It would be nice to have a convenient
> > expression for (2) & (3).
>
> I believe that by the current interpretations "lo" is #3. #2 can be handled
> by something like "rosu'o", "all of the at-least-one".
So is (4) handled by "da poi ..." then?
Are you sure "rosuho" would work for (2)? We want "all (& there is at
least one)".
E.g. since Brahms's symphonies number 4, "all (& there is at least one)
of B's symphs" refers to all four, while "all of at least one of B's
S's" could refer to (the members of) any non-empty subset of the set of
all of B's S's, e.g. to only 2 of B's S. Which one does "rosuho" do?
---
And
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: lo [nonexistent]
In-Reply-To: (Your message of Tue, 11 Oct 94 10:41:02 D.)
John:
> > I don't think lohe makes claims about the world;
>
> I agree.
>
> > it makes claims for
> > default properties of categories in our minds. So it works.
>
> On your view, then, the difference between "le'e" and "lo'e" is that
> the former refers to the speaker's idiosyncratic categories or properties
> thereof, whereas the latter refers to consensus categories or properties?
I haven't got access to my cmavo list, so I hope I am right in thinking
"lehe" is "the stereotypical".
I would say, to start off with, that it ought always to be possible to reach
a consensus about "lohe broda", if we know enough about brodas, but it
needn't always be possible to reach a consensus about "lehe broda"
(though this doesn't mean there must be an absence of such a consensus).
So, as you suggest, there is some objective basis for "lohe" but not
for "lehe", though we may in fact agree about "lehe broda".
In my view (but not necessarily on my view), properties attributed to
"lohe broda" should be truly the default for our state of knowledge,
so that if, given "lohe broda cu brode", we know that some broda isn't
brode, then we know that this broda is therefore exceptional.
With "lehe", on the other hand, properties we attribute to "lehe broda"
needn't be true by default (for our state of knowledge) of brodas -
it's as if "lehe broda" was appointed as representative of some group
without actually being representative [=adjective] of that group.
For example, suppose there is a race of people called Brodas, and
that virtually every Broda I know has been unfailingly kind to me, a fact
of which I am aware, but I happen to be a racist. In that case,
if I was rational, I would say "lohe mela brodas. cu zabna", and
it would be irrational of me to say "lohe mela brodas. cu mabla",
but it would be foul-minded rather than irrational of me to say
"lehe mela brodas cu mabla".
This is purely me trying to make sense of the distinction. I don't
know if it is the official line of whoever decided in the first
place that it was worthwhile making the distinction.
---
And
From: Chung-chieh Shan <ken@MATH.NTU.EDU.TW>
Subject: Re: Talking about non-existant objects
> I haven't been reading my Lojban mail that closely recently, partly because
> there's been so much of it
Well now I know I'm not alone. :)
> I then further thought, why not solve the problem of not being able to talk
> about Elves and the such like by inventing a tense that would get us to the
> (imaginary) location in which they do exist. Let's say this tense is "xa'o"
> (c.f. ka'o :-) then if I say:
>
> lo ricre'a cu crino
>
> you (believing the non-existance of elves) would infer:
>
> lo ricre'a xa'o crino
I like this proposal!
Maybe vedu'o works, or maybe we need a new modal/tense cmavo such as xa'o.
But this sounds most natural to me -- when I hear "elves are..." my mind
automatically enters a "hypothetical mode". Of course, we should all be
grateful to the implicit tense system of Lojban. Why hasn't anyone else
commented on this one?
This accepted-by-all-Lojbanists proposal xa'o ba add a new meaning to the
tense paper title "Imaginary Journeys". :)
----------
Chung-chieh (Ken) Shan ken@cauchy.math.ntu.edu.tw
"Ay, fashion you may call it. Go to, go to." -- Hamlet
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: "any"
PC said:
Veion suggests that 'any', when not all in disguise, is a
discursive that means "no hidden conditions apply". His example, "I will
eat any apple," however, seems to fail on both counts.
First, the sentence as he uses it is pretty clearly not a
prediction, but an offer. As such, it sets up an intentional (so also
intensional and thus opaque) context. "Any' then functions as usual as a
context-leaping universal, the whole being approximately, "for all x, if
x is an apple, then I am willing that I eat x" -- "all" in disguise
again, but outside the opaque context and binding into it, so covering
real apples only (with no guarantee that here are any, as is usual with
'any').
Using the often illuminating dialog exposition of quantifiers,
this offer would amount to the speaker saying "You get to pick the apple
but I am willing to eat whatever you pick." But I, the hearer am pretty
clearly not unrestricted in my choice of apples. In the first place, I
only get one pick (well, certainly the speaker can withdraw his offer
after some number, he is not committed to eating -- or even to being
willing to eat -- every apple there is). This is a feature of the
intentional part, though, not of the 'any.' But, further, my choice is
restricted in very inexplicit ways: I surely cannot expect him to take
the apple the queen has prepared for Snow white nor the one Eve gave
Adam nor probably even those soft brown ones in the bottom of the
barrel. "Any' is, after all, just the preferred bearer of such
conditions as "within reason" (the usual formulation of the hidden
clause). To insist that the speaker has agreed to eat a manifestly yucky
apple is on a par to denying that all wild ducks in North America fly
South for the winter because pet ducks, crippled ducks and ducks in city
parks do not. The objections may be technically correct but
conversationally irrelevant and inappropriate -- moderately good logic
but abominable language.
Veion's idea is a good one, IF he can find a case. But IMHO
'any' ain't gonna provide any.
pc>|83
GK continues:
This makes it clear just how pleiomorphic and ambiguous this
little word of English can be. I think if we ever get it
working properly in lojban it will have to take many different
forms. Just as the connective "and" did when lojbanized. I am
still opposed to trying to capture all the English meaning and
behavior of "any" in one word.
Looking at the above analysis it appears that the meaning can be
broken down into three elements.
1. exactly one apple is under discussion.
2. It is a typical apple. No outliers are under
consideration.
3. It is a randomly selected apple.
4. (2) and (3) are connected by the logical &.
So we have one typical and random apple:
In lojban this goes:
pa lomci je cunso plise (or possibly):
lo paboi lomci plise ije cunso plise?
Maybe we need a word for precisely this. It could be a start on "any"
Would it parse? Are quantifiers proliferating to excess?
XE'E, XE'E, XE'E, XE'E, XE'E ...........I heard that laugh, jorge.
djer
From: CJ FINE <C.J.Fine@bradford.ac.uk>
Subject: Re: "any"
In-Reply-To: (null)
On Mon, 17 Oct 1994, Gerald Koenig responded to pc:
> This makes it clear just how pleiomorphic and ambiguous this
> little word of English can be. I think if we ever get it
> working properly in lojban it will have to take many different
> forms. Just as the connective "and" did when lojbanized. I am
> still opposed to trying to capture all the English meaning and
> behavior of "any" in one word.
Absolutely. But is anybody trying to do so?
>
> Looking at the above analysis it appears that the meaning can be
> broken down into three elements.
> 1. exactly one apple is under discussion.
I'm not even convinced of this, but let it pass for the moment.
> 2. It is a typical apple. No outliers are under
> consideration.
Roughly, but I don't think you mean 'typical', which is after all an
idealisation. I suspect you mean 'ordinary' - 'fadni' in Lojban.
3. It is a randomly selected apple.
What has 'random' got to do with it? If I say I will eat any apple, you
may then with great care pick out an apple to give me - this is not
random, it is merely that I am explicitly forgoing any participation in
selecting the apple.
> 4. (2) and (3) are connected by the logical &.
>
> So we have one typical and random apple:
> In lojban this goes:
> pa lomci je cunso plise (or possibly):
> lo paboi lomci plise ije cunso plise?
I don't know what you mean by 'lomci' - it looks to me as if you have
tried to coin a brivla in back-formation from the rafsi 'lom'.
So building on your idea, I think it is something like
lo fadni je se cuxna be no mi be'o plise
which would function OK as a dictionary definition of 'any apple' in this
sense but is not terribly useful in everyday use. But maybe if we need to
express th8s precisely, this is how to do it.
Colin
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: "any"
In-Reply-To: (Your message of Mon, 17 Oct 94 01:59:20 MST.)
La Djer:
> Looking at the above analysis it appears that the meaning can be
> broken down into three elements.
> 1. exactly one apple is under discussion.
> 2. It is a typical apple. No outliers are under
> consideration.
> 3. It is a randomly selected apple.
> 4. (2) and (3) are connected by the logical &.
>
> Maybe we need a word for precisely this. It could be a start on "any"
> Would it parse? Are quantifiers proliferating to excess?
>
> XE'E, XE'E, XE'E, XE'E, XE'E ...........I heard that laugh, jorge.
Xehe indeed. This is pretty much my understanding of what Jorge proposed
& I seconded, with the following differences:
(1) "one" is only a default, and any number can be specified. (As in
"any five books")
(2) 'Typicality' in the sense of 'average, unexceptional' can I think
be left to pragmatics. That is, "I am willing to eat xehe one
apple" *should* entail that I am willing to eat a rotten apple
shat on by a skunk, and the fact that this is not what I intend
you to infer can be left to normal processes of communication.
(3) It is not *randomly* selected but *arbitrarily* selected.
So, in summary, "xehe broda" would mean "PA (and only PA) things
arbitrarily selected from the set containing every broda".
-----
And
From: Sylvia Rutiser <sru>
Subject: Re: lo [nonexistent]
In-Reply-To: <199410121000.AA07861@nfs1.digex.net>
coi doi ro
I would like to follow up on Bob Michael's comment.
IMHO, one of the uses for language is the discussion of whether or not
some object, relationship, or state exists. For example, the existence
or non-existence of deities is a frequent source of discussion and/or
quarrel. Also, ether, pholstogon, and songs sung by Milli Vanilli.
I would like to suggest that the distinction to be made is not whether
one or more examples of what you are talking about actually exists, but
whether the speaker is making a claim, or just a 'suppose' kind of
statement.
I think that there are UI selma'o cmavo to express the distinction.
mi'e silvian
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: "any"
The Webster definition of "any" is:
1. one or some indiscriminately of whatever kind.
a. one or another taken at random
b. every--used to indicate one selected without restriction
2. one, some, or all indiscriminately of whatever quantity.
a. one or more
b. all
3. a unmeasured or unlimited in amount,number, or extent.
b. appreciably large or extended.
Previously I suggested that "any x" should translate to: (quantifier)
typical random (x). This post is an attempt to embody some of the
dictionary meanings above, and to deal with the multifarious problems
pointed out by John Clifford. I have as yet no idea what these
suggestions would mean for resolving the many scope ambiguities of
"any" in Englsh. Nor do I know if the grammar here explored would
work.
Predicate calculus quantifiers lojban comments
All (x) ro true quantifier
Some E(x) su'o true quantifier
Iota I(x) lo pa * the (one) x (term)
S_any Alpha(x) * l'alfa * singular "any"(term)
P_any Sigma(x) * s'ma * plural"any"quantifier
True quantifiers should have grammar of PA4. Terms should have grammar
of LA. Sigma(x) should have grammar of PA4. The stars indicate new
ideas for discussion. Pardon my ungrammatical new word spellings.
Examples:
ro da zo'u tu'e da plise inaja mi cidja da
I eat all the apples. (They may or may not exist).
su'o da zo'u tu'e da plise inaja mi cidja da
I eat some apples. (They exist).
lo pa da zo'u tu'e da plise inaja mi cidja da
I eat the apple. (only one specific apple, it could be named Munchkin.)
l'alfa da zo'u tu'e da plise inaja mi cidja da
I eat any apple. (only one random apple, it could be named Crunchkin)
l'sma da zo'u tu'e da plise inaja me cidja da
I eat any apples. (subject to the built-in restrictions on "any")
These are of course too stilted for practical use. But maybe the
following forms would work:
mi cidja ro lo plise
mi cidja su'o lo plise
mi cidja lo pa plise
mi cidja l'alfa lo plise
mi cidja l'sma lo plise
These are all meant to have the same meaning as the corresponding
examples above.
If there is any value in these ideas, they should be harmonized with the
ideas put forth by .And on the question of existence in quantifiers. I'm
not sure I dealt with "any" as "all" either but that's it for today.
djer
From: i.alexander.bra0125@oasis.icl.co.uk
Subject: Re: 'any' as discursive
cu'u pycy
> Veion's idea is a good one, IF he can find a case. But IMHO
> 'any' ain't gonna provide any.
"I will eat any apple" may not be the right example, but I had this
idea as well, and I still think it's right.
The examples that come to mind are imperatives, e.g. "Pick a card,
any card" - "Let there be a card such that you choose it" -
"Make it such that there is a card which you have chosen".
These are also effectively opaque, but there doesn't seem to me
to be any universal quantification going on.
I'm not entirely sure that there is a universal quantification
in Veijo's example either. There is, I agree, such a quantification
implied by some instances of "any", but "I will eat any
apple you choose" certainly doesn't mean-to-me "For any x,
x an apple, I will eat x", which would be "I will eat
all the apples you choose". It's more like "I permit
you to choose an apple, which I will then eat" -
curmi le nu pa plise poi se cuxna do cu se citka mi
curmi le nu do cuxna gi'e mi citka vau pa plise
(This isn't quite right yet - more later.)
This doesn't tell the whole story, of course. We frequently
find hidden assumptions in everyday discourse. This says
what is permitted, but not what is forbidden.
cu'u la djer.
> 2. It is a typical apple. No outliers are under
> consideration.
This isn't so much part of the meaning of the single word
"any", as a much broader-based part of the extra-linguistic
context. You have no right to require me to perform any
particular action, such as eating a particular apple, but
we don't say this explicitly. There is in general no
economical way to make such things explicit - they have
to be left to common sense. I explicitly allow you to
choose an apple, within certain common-sense limitations,
which I then offer to eat - I may or may not be amenable
to eating more than one chosen-by-you apples, and you must
use your common sense to imagine what the limits might be
on the number as well as the quality.
This intensional (or do I mean intentional) stuff is tricky.
I'm not sure my earlier Lojban example is entirely accurate.
curmi le nu do cuxna pa plise, noi mi citka
That non-restrictive qualification feels right, although
I'm not entirely what the distinction means in Predicate
Calculus terms.
curmi le nu mi citka pa plise poi do cuxna
is wrong - it means I'll eat one of the apples you choose.
curmi le nu do cuxna pa plise poi mi citka
is better - I may eat other apples as well, but this only allows
you to choose one of them.
ca banzu
mu'o mi'e .i,n.
From: "John E. Clifford" <pcliffje@CRL.COM>
Adding to my last posting, mehr Licht?
Another feature of opaque contexts, which I forgot to men-
tion, is that Leibnitz's Law does not work in them: from a=b and
[a] we cannot infer [b]. This is certainly the case if one term
of the identity is a description, since what fits (or is intended
to fit) a description can vary with context. It is less certain
for proper names, since some logicians hold that names are rigid
designators, referring to the same thing in all contexts (at
least all those in which the thing named exists). But proper
names in natural languages do not seem to meet this requirement
-- more than one thing can have the same name, for example, even
in a single context. Thus, ordinary proper names seem to behave
pretty much like descriptions under this rule (as the Lojban
placement -- and official reading -- of _la_ suggests) and had
best be thought not to be replaceable under identity.
I found in the avalanche of the last couple of weeks a note
from lojbab that mentions that the mark for raised subjects is
_tu'a _, which is then the opacity marker under the suggestion in
that posting (_xe'e_ in its non-experimental form).
OTOH some expressions in English contain event-referring
expressions but seem to be transparent: " I saw someone shooting
pool," for example. This pretty clearly does imply that there is
someone that I saw shooting pool. Indeed, if it could be shown
that there was no one shooting pool in my visual range, I would
have withdraw my original claim, falling back to "I thought I
saw..." or "It looked like ..." or whatever. But notice that
the basic claim is not exactly the classic English form of an
event-referring expression, a "that"-clause or an infinitive. "I
saw that someone was shooting pool" does seem to be opaque,
approximately equivalent to "came to know ... by seeing" and so
inheriting the opacity of "know," failing to export when I could
not identify the player -- presumably by sight. Thus, we might
argue that the object of "see" in the transparent case is not the
event but rather just the subject, to which the event-reference
is somehow attached. That is, it may be that the analysis of "I
saw someone playing pool" is not "I saw (someone playing pool)"
but "I saw someone (playing pool)", which would both account for
the transparency and fit in with our general notion that the
object of seeing is an object not (generally) an event. That
leaves the question of how the "someone" and the "playing pool"
are to be linked together, for it is not just that the someone
was in fact playing pool but that I saw him doing it, so there is
an event-referring expression here after all, though perhaps
subordinately. None of the obvious suggestions in Lojban (e.g.
_poi_ or _noi_) seems quite right. Comments and suggestions
welcomed eagerly.
pc>|83
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: "any"
la djer cusku di'e
> su'o da zo'u tu'e da plise inaja mi cidja da
> I eat some apples. (They exist).
This is not what the Lojban sentence claims. In Lojban, you wrote:
"there is some x, such that if it is an apple, I eat it". This statement
is always true. It suffices to select some non-apple for da, and since
in that case {da plise} is false, the whole statement is true for at least
that da whether I eat it or not, and therefore the statement is true.
I eat some apples is
mi citka lo plise
I eat some apples
or equivalently:
su'o da poi plise zo'u mi citka da
For at least one x which is an apple, I eat it.
[Note: citka=eat, cidja=food]
> lo pa da zo'u tu'e da plise inaja mi cidja da
> I eat the apple. (only one specific apple, it could be named Munchkin.)
Again the same problem. You are claiming that for the one thing, if it is
an apple then you eat it. You don't claim that it is an apple, so the
claim is again trivially true.
The normal way to say "I eat the apple" is {mi citka le pa plise}.
You can say {mi citka lo pa plise}, but then you really mean that only
one apple exists.
> l'alfa da zo'u tu'e da plise inaja mi cidja da
> I eat any apple. (only one random apple, it could be named Crunchkin)
>
> l'sma da zo'u tu'e da plise inaja me cidja da
> I eat any apples. (subject to the built-in restrictions on "any")
>
> These are of course too stilted for practical use.
And they have a different meaning than the one you want.
> But maybe the
> following forms would work:
>
> mi cidja ro lo plise
> mi cidja su'o lo plise
> mi cidja lo pa plise
> mi cidja l'alfa lo plise
> mi cidja l'sma lo plise
>
> These are all meant to have the same meaning as the corresponding
> examples above.
I think by "l'alfa" and "l'sma" you mean the same I wanted to get with
{pa xe'e} and {su'o xe'e}. I don't understand why you say that one should
be a quantifier and the other an article.
Jorge
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject: Re: 'any' as discursive
la i,n cusku di'e
> This intensional (or do I mean intentional) stuff is tricky.
I think you mean intentional. What does intensional mean? With
great intensity? :)
> I'm not sure my earlier Lojban example is entirely accurate.
>
> curmi le nu do cuxna pa plise, noi mi citka
>
> That non-restrictive qualification feels right, although
> I'm not entirely what the distinction means in Predicate
> Calculus terms.
I agree with the non-restrictive. I disagree with the other two:
> curmi le nu mi citka pa plise poi do cuxna
>
> is wrong - it means I'll eat one of the apples you choose.
It means I'm allowed to eat one of the apples you choose.
> curmi le nu do cuxna pa plise poi mi citka
>
> is better - I may eat other apples as well, but this only allows
> you to choose one of them.
That means that you are allowed to choose one of the apples that
I eat. No guarantee that I will eat any one you choose. I like
the one with {noi}.
Jorge
From: Colin Fine <C.J.Fine@bradford.ac.uk>
Subject: Apposed participials
pc says:
> OTOH some expressions in English contain event-referring
> expressions but seem to be transparent: " I saw someone shooting
> pool," for example. This pretty clearly does imply that there is
> someone that I saw shooting pool. Indeed, if it could be shown
> that there was no one shooting pool in my visual range, I would
> have withdraw my original claim, falling back to "I thought I
> saw..." or "It looked like ..." or whatever. But notice that
> the basic claim is not exactly the classic English form of an
> event-referring expression, a "that"-clause or an infinitive. "I
> saw that someone was shooting pool" does seem to be opaque,
> approximately equivalent to "came to know ... by seeing" and so
> inheriting the opacity of "know," failing to export when I could
> not identify the player -- presumably by sight. Thus, we might
> argue that the object of "see" in the transparent case is not the
> event but rather just the subject, to which the event-reference
> is somehow attached. That is, it may be that the analysis of "I
> saw someone playing pool" is not "I saw (someone playing pool)"
> but "I saw someone (playing pool)", which would both account for
> the transparency and fit in with our general notion that the
> object of seeing is an object not (generally) an event. That
> leaves the question of how the "someone" and the "playing pool"
> are to be linked together, for it is not just that the someone
> was in fact playing pool but that I saw him doing it, so there is
> an event-referring expression here after all, though perhaps
> subordinately. None of the obvious suggestions in Lojban (e.g.
> _poi_ or _noi_) seems quite right. Comments and suggestions
> welcomed eagerly.
> pc>|83
>
I have a very strong intuition that it is not the event which is the object
of the seeing here. But as pc says the event is still part of the object.
In many cases we can actually express it as
mi viska da ca le nu da kelcrpuli
'I saw x at the time x plays-pool'
and I have a suspicion that expressions of this sort will always work,
but I'm not sure (they may not necessarily be temporal, but deciding
whether a temporal, spatial or other relation is appropriate will be
an example of the familiar process of being more precise when we translate
into Lojban. And we can always leave it vague with "va'o" or even "do'e".
pei
Colin
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: Transparancy
PC continues:
Adding to my last posting, mehr Licht?
Another feature of opaque contexts, which I forgot to men-
tion, is that Leibnitz's Law does not work in them: from a=b and
[a] we cannot infer [b]. This is certainly the case if one term
of the identity is a description, since what fits (or is intended
to fit) a description can vary with context.
GK> [Speaking of the sentence, "I saw someone shooting pool."]
GK> Perhaps in this case it is so clear what the "someone" refers to,
that the opaque, event formulation is not opaque. The "someone" seems
to be clearly the same person whether inside or outside the
parentheses. This could be a consequence of an implied pointing to an
individual. To see someone normally implies a light ray pointing to the
person. An indefinite description normally would specify a range of
individuals. This could be an instance of a definite description which
is traditionally equivalent to a proper name. I hear you when you say
that the traditional interpretion may not be practical. I understand
that The President=Bill is different from Bill=The President, where
there are thousands of Bills in the phone book. But I think there will
be a price to be paid if the equivalence of a definite description to a
name is abandoned.
PC> It is less certain for proper names, since some logicians
hold that names are rigid designators, referring to the same
thing in all contexts (at least all those in which the thing
named exists). But proper names in natural languages do not
seem to meet this requirement -- more than one thing can have
the same name, for example, even in a single context. Thus,
ordinary proper names seem to behave pretty much like
descriptions under this rule (as the Lojban placement -- and
official reading -- of _la_ suggests) and had best be thought
not to be replaceable under identity.
I found in the avalanche of the last couple of weeks a
note from lojbab that mentions that the mark for raised
subjects is _tu'a _, which is then the opacity marker under the
suggestion in that posting (_xe'e_ in its non-experimental
form).
OTOH some expressions in English contain event-referring
expressions but seem to be transparent: " I saw someone
shooting pool," for example. This pretty clearly does imply
that there is someone that I saw shooting pool. Indeed, if it
could be shown that there was no one shooting pool in my visual
range, I would have withdraw my original claim, falling back to
"I thought I saw..." or "It looked like ..." or whatever. But
notice that the basic claim is not exactly the classic English
form of an event-referring expression, a "that"-clause or an
infinitive. "I saw that someone was shooting pool" does seem to
be opaque, approximately equivalent to "came to know ... by
seeing" and so inheriting the opacity of "know," failing to
export when I could not identify the player -- presumably by
sight. Thus, we might argue that the object of "see" in the
transparent case is not the event but rather just the subject,
to which the event-reference is somehow attached. That is, it
may be that the analysis of "I saw someone playing pool" is not
"I saw (someone playing pool)" but "I saw someone (playing
pool)", which would both account for the transparency and fit
in with our general notion that the object of seeing is an
object not (generally) an event. That leaves the question of
how the "someone" and the "playing pool" are to be linked
together, for it is not just that the someone was in fact
playing pool but that I saw him doing it, so there is an
event-referring expression here after all, though perhaps
subordinately. None of the obvious suggestions in Lojban
(e.g. _poi_ or _noi_) seems quite right. Comments and
suggestions welcomed eagerly. pc>|83
GK> It seems to me that in both cases "someone" and "playing pool" are
related as subject and predicate, or sumti and selbri.
"I saw someone shooting pool" has two interpretions for me.
1. I saw a person, and that person was playing pool.
2. I saw an event, and that event was a person playing pool.
For the lojban I am going to truncate playing_pool to playing [zo'e].
I don't have a lot of confidence in bolga'a jubme.
1'. mi pu viska paboi prenu ije ri kelci
2'. mi pu viska lo nu paboi prenu cu kelci.
The selbri viska takes an object or an event indifferently as the X2.
I wish someone could explain the semantic difference in these sentences.
They mean pretty much the same thing to me.
If I understand you correctly, you are saying that the 'prenu' in
sentence 2' is opaque. Or rather that it should be opaque but isn't.
Here again I feel it would not be opaque because it is a very clear
identification. Identification means uniqueness, and hence transparency.
I hope this makes sense to you, I am still somewhat diaphanous about
transparency and opaqueness.
djer
From: Colin Fine <C.J.Fine@bradford.ac.uk>
Subject: Re: Transparancy
Gerald answers pc:
> GK> It seems to me that in both cases "someone" and "playing pool" are
> related as subject and predicate, or sumti and selbri.
That's one interpretation, but you can equally well say they are related
as term and attribue (eg smuti and seltanru)
> "I saw someone shooting pool" has two interpretions for me.
>
> 1. I saw a person, and that person was playing pool.
> 2. I saw an event, and that event was a person playing pool.
>
> For the lojban I am going to truncate playing_pool to playing [zo'e].
> I don't have a lot of confidence in bolga'a jubme.
>
> 1'. mi pu viska paboi prenu ije ri kelci
> 2'. mi pu viska lo nu paboi prenu cu kelci.
>
> The selbri viska takes an object or an event indifferently as the X2.
> I wish someone could explain the semantic difference in these sentences.
> They mean pretty much the same thing to me.
For one obvious difference, if I look through a window and see a pool
table and a ball being cued, but cannot see anything else in the room, I
suggest that it is true that
mi viska lo nu lo prenu cu kelcrpuli
but untrue that
mi viska lo prenu ije re kelci
(I am ignoring the 'pa' as irrelevant to this argument)
The English pretty clearly means the first, not the second.
> If I understand you correctly, you are saying that the 'prenu' in
> sentence 2' is opaque. Or rather that it should be opaque but isn't.
> Here again I feel it would not be opaque because it is a very clear
> identification. Identification means uniqueness, and hence transparency.
>
> I hope this makes sense to you, I am still somewhat diaphanous about
> transparency and opaqueness.
Me too, but I don't think it's anything to do with identification or
uniqueness.
1. mi pu viska paboi prenu ije ri kelci
is true iff I saw one person and that person plays. From it I can validly
deduce
2. pa prenu cu kelci
Now,
3. mi pu viska le nu paboi prenu cu kelci
is true iff I saw happening the event of one person playing pool.
I have indicated above why 3. does not entail 1. The question of transparency
is whether 3. entails 2. or not. I incline to the view that 'viska' does
not create an opaque context and therefore that it does entail 2 (but
'jinvi' certainly would not, and I conjecture that nor would 'sanji'). But I
believe that this is what we are arguing about.
Colin Fine
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: Apposed participials
la kolin. cusku di'e
> Date: Wed, 19 Oct 1994 17:11:10 BST
> From: Colin Fine <C.J.Fine@BRADFORD.AC.UK>
> Subject: Apposed participials
...
> I have a very strong intuition that it is not the event which is the object
> of the seeing here. But as pc says the event is still part of the object.
>
> In many cases we can actually express it as
>
> mi viska da ca le nu da kelcrpuli
> 'I saw x at the time x plays-pool'
>
> and I have a suspicion that expressions of this sort will always work,
> but I'm not sure (they may not necessarily be temporal, but deciding
> whether a temporal, spatial or other relation is appropriate will be
> an example of the familiar process of being more precise when we translate
> into Lojban. And we can always leave it vague with "va'o" or even "do'e".
I thought of several similar expressions and came up with the following
where it is quite difficult to say what is the actual object of the
sensing
'I sensed someone had been in the room'
The nearest translation I could think of was
mi ganse leka le kumfa ba'o se zvati da
but there may still be some quality of the original missing.
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: "John E. Clifford" <pcliffje@CRL.COM>
Subject: "any"
I,n asks where is the quantification in "Pick a card, any card." It must
be where a truth-functional quantifier can meet an non-truth-functional
directive sentence: at the satisfaction set, the set of all the
situations (sentences, for convenience) any one of which being true
satisfies the directive (request in this case?). That set is roughly {x
is a card in the deck presented : "I pick x"}, that is , for ever x which
is a card in the deck presented (the "hidden condition" in this "any"),
if I pick x, I satisfy the request. This is a very different set from
the one for "Pick every card", which has only "I pick every card" or the
conjunction of the sentences "I pick ..." for each card in it. The "any"
set is, however, exactly the set for "Pick a card", "if there is a card
in the deck that I pick, then I satisfy the request". This is what logic
says should happen in this case (assuming that the same deck is used
inside and out of the intentional context).
That same satisfaction set plays a related role in the original
question , about "need" and "any" functions again to leap out of that
context to home ground. It does it as well in "I will eat any apple you
choose", if you remember that "will" here is intentional, offering to do
something, not "just" a tense marker, so it makes the offer before the
choice of apples but the offer is still only good for a limited number of
apples.
pc>|83
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: pick any card
i.an's sentence, "Pick a card, any card", could be expressed as
ko cuxna l'alpha karda
using the experimental singular quantifier l'alpha= "any" I defined
in a previous post.
djer
Date: Fri, 21 Oct 1994 14:31:42 -0700 (PDT)
From: "John E. Clifford" <pcliffje@crl.com>
Subject: Re: "any"
In-Reply-To: <199410210815.AA06835@access4.digex.net>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
IMHO we have lambda more than covered for serious purposes with all those
event and property and whatever abstractions. Athide from thome thtrange
programming languageth, the main uthe of lambda calculuth ith to tholve a
technical problem in formalized logic: what egthactly can be thubthituted
for predicate letterth or function thymbolth in a tautology to enthure
that what you get will altho be a tautology. Lambda was a brilliant
solution in the 30's but we can do it more easily now. Of course, it is
also useful in higher order logics to talk about situations described at
lower levels, but that is back toevent descriptors again: we just leave
the spaces blank rather than binding them up.
I think.
pc>|83
From: Chris Bogart <cbogart@quetzal.com>
Subject: mi na nu'o catra ko'a
It's my understanding that when we say "mi catra ko'a", the tense is left
vague or to context, so it could in fact mean only "mi nu'o catra ko'a" --
which could be an interesting way to refute an earlier confession in a
lojbanic court! Do I have this right? It's a little weird, but necessary
so that "mi catra" can mean "I'm a killer", even if I haven't killed anyone
yet, and never get around to killing anyone.
Even worse, suppose I claimed "mi na catra ko'a" -- could that be
interpreted in some circumstances as "mi na nu'o catra ko'a"? Since "nu'o"
means "can but has not", does the "na" deny that the action was innately
possible, or that it didn't in fact happen? Or are both claims made by
"nu'o", so we have to use De Morgan's law ( NOT(A AND B) => NOT A OR NOT B )
to figure out what's going on:
mi na nu'o catra =>
mi na (na ca'a je ka'e) catra =>
mi (ca'a ja na ka'e) catra =>
mi ca'a catra .ijanai mi ka'e catra
"If I was capable of it, then I killed him" Easier to understand in the
future tense: "I'll kill him if I can". Given my train of reasoning, this
is a possible interpretation of "mi na catra ko'a", just as much as "I will
kill him" or "I was about to kill him", right? (BTW I'm not arguing against
this, but if my interpretation is correct, I'm warning that it's a weird
area to be careful in. If this is logical-but-counterintuitive and hard for
English speakers to grasp, but fluent Lojbanis manage to figure it out,
lojban will have demonstrated something very interesting, IMHO)
Well, I've convinced myself of this, so I guess I don't have a question
anymore, except "is this right and/or intelligible?"
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: jorge@PHYAST.PITT.EDU
Subject: Re: Apposed participials
la kolin cusku di'e
> I have a very strong intuition that it is not the event which is the object
> of the seeing here. But as pc says the event is still part of the object.
If both the event and the person are seen, then shouldn't the claim be:
mi viska da e le nu da kelcrpuli
I saw x and I saw x playing pool.
> In many cases we can actually express it as
>
> mi viska da ca le nu da kelcrpuli
> 'I saw x at the time x plays-pool'
>
> and I have a suspicion that expressions of this sort will always work,
That doesn't claim that I saw the event of playing, only that I saw x
at that time.
In any case, I would translate "I saw someone playing pool" as
{mi viska le nu da kelcrpuli}. The English sentence may have other
connotations, but that will always happen when translating from one
language to another.
It would be reasonable to assume, if I saw the event of someone playing
pool, that I saw that someone, even if I don't exactly say so. That is
a matter of context. (The same for the {ca} case, it would be natural
to assume that the playing was seen, if the person was seen at the time
of playing.)
Another way out is:
mi viska lo kelcrpuli
I saw someone playing pool
(Well, the English translation is tendentious, but the Lojban can mean that.)
I'm not sure what this has to do with the transparent/opaque distinction,
as both claims, either with the object or with the event being seen, seem
transparent to me.
la kolin di'e spusku la djer
> > I hope this makes sense to you, I am still somewhat diaphanous about
> > transparency and opaqueness.
>
> Me too, but I don't think it's anything to do with identification or
> uniqueness.
>
> 1. mi pu viska paboi prenu ije ri kelci
>
> is true iff I saw one person and that person plays. From it I can validly
> deduce
> 2. pa prenu cu kelci
>
> Now,
> 3. mi pu viska le nu paboi prenu cu kelci
>
> is true iff I saw happening the event of one person playing pool.
>
> I have indicated above why 3. does not entail 1.
We agree up to here.
> The question of transparency
> is whether 3. entails 2. or not.
I disagree with this. To me the question of transparency does not even
arise in this case, because the sumti in question is inside an abstraction.
Maybe my definition of opaque is wrong, but what I understand is that an
opaque claim, e.g. "I need a box" can be made transparent as "I need that
I have a box". The second claim is transparent in general, because the
verb "have" usually does not take opaque objects. (We could make up some
context like owning any one of the ten horses, but that is not a common
expression.)
Jorge
From: jorge@PHYAST.PITT.EDU
Subject: Re: "any"
la pycy cusku di'e
> I,n asks where is the quantification in "Pick a card, any card."
[...]
> The "any"
> set is, however, exactly the set for "Pick a card", "if there is a card
> in the deck that I pick, then I satisfy the request".
[...]
> That same satisfaction set plays a related role in the original
> question , about "need" and "any" functions again to leap out of that
> context to home ground.
I agree, but the question is how to implement it in Lojban.
In the case of commands/directives/requests there is no problem, the
way things are defined,
(1) ko cuxna lo karda
means: "Make true the statement {do cuxna lo karda}". And the statement
will be made true for any card that is picked.
Sentence (1) does NOT mean "there exists at least one card such that
I am requesting that you pick it", and therefore we don't run into the
problem we have with "need".
(2) mi nitcu lo tanxe
on the other hand, DOES mean "there exists at least one box such that
I need it". This is how we want it to work for most predicates, but for
{nitcu} it gives us problems.
I am happy to solve this using {mi nitcu lo'e tanxe}, which roughly would
mean "the archetypal box is such that I need it", or "the relationship
{nitcu} holds between {mi} and the archetypal box". This is why I don't
like {lo'e} being defined as "typical", which in any case is a strange
article to have.
> It does it as well in "I will eat any apple you
> choose", if you remember that "will" here is intentional, offering to do
> something, not "just" a tense marker,
Yes, so {ba} is the wrong word to translate "I am willing to".
I would say
(3) ai mi citka pa plise noi do ba cuxna ke'a
I will to eat one apple, which will be chosen by you.
Here the problem is solved by {ai}, which supposedly wrecks havoc with
truth values. We can interpret it as "I am willing that (I eat an apple,
which you will choose)". Then the same effect is achieved of having
it inside an abstraction.
I think from the point of view of logic, "any" can be handled one way
or another, but we can't translate the emphatic "any": "anyone whatsoever",
because there isn't a word to emphasize.
Jorge
From: jorge@PHYAST.PITT.EDU
Subject: Re: mi na nu'o catra ko'a
la kris cusku di'e
> It's my understanding that when we say "mi catra ko'a", the tense is left
> vague or to context,
I always interpret a tensless sentence as "ca'a", and if the time is needed
for it to make sense as "ca". I know that the usual claim is that any tense
can be meant, but as you point out, allowing it to mean "ka'e" and company
doesn't make much sense.
It's funny what {na nu'o} ends up meaning. I agree that {nu'o} is equivalent
to {ka'e jenai puca'a} and similarly, {pu'i} is {ka'e je puca'a}. I don't
really understand why they are part of the language. They seem to be possible
translations of "could", but they don't seem that useful to me to have a
special word.
Jorge
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Re: "any"
In-Reply-To: (Your message of Sat, 22 Oct 94 19:16:02 EDT.)
Jorge:
> In the case of commands/directives/requests there is no problem, the
> way things are defined,
>
> (1) ko cuxna lo karda
>
> means: "Make true the statement {do cuxna lo karda}". And the statement
> will be made true for any card that is picked.
>
> Sentence (1) does NOT mean "there exists at least one card such that
> I am requesting that you pick it"
This doesn't look like "no problem" to me. What if I want to say
"There is a card; pick it", or "Pick a card (& there is a card)"?
----
And
From: jorge@PHYAST.PITT.EDU
Subject: Re: "any"
> > (1) ko cuxna lo karda
> >
> > Sentence (1) does NOT mean "there exists at least one card such that
> > I am requesting that you pick it"
>
> This doesn't look like "no problem" to me. What if I want to say
> "There is a card; pick it", or "Pick a card (& there is a card)"?
>
> ----
> And
You can do the same thing you are doing in English, paraphrase:
ko'a karda i ko cuxna ko'a
It is a card; pick it.
or
ko cuxna lo karda (to ije lo karda cu zasti toi)
Pick a card (& a card exists).
I say in this case there is no problem because the simple expression
{ko cuxna lo karda} already has the opaque meaning, due to the way
that {ko} is defined, and furthemore the opaque meaning is the one
we usually want for this type of sentence. If you mean something else,
you have to be more wordy.
I just realized that it is strange that we have a special word for
the "command mode", but not for other similar things like the
"intentional mode", or the "volitional mode", etc, which are handled
with UIs, but could equally well have been something like {ko}.
For example, say {xi'u} was the "intentional {mi}", then we'd have
xi'u klama lo zarci ~ ai mi klama lo zarci
just like
ko klama lo zarci ~ e'o do klama lo zarci
I don't see why the imperative is somehow more fundamental than the
intentional, volitional, and all the others.
It would be interesting to make a list of the attitudinals that change
the sentence to opaque mode, like {ai} and {e'o}.
This is assuming I'm right that {ai mi klama lo zarci} means
"I intend that there be a store such that I go to it" and not
"there is a store such that I intend to go to it".
Jorge
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: mi na nu'o catra ko'a
la xorxes. cusku di'e:
>It's funny what {na nu'o} ends up meaning. I agree that {nu'o} is equivalent
>to {ka'e jenai puca'a} and similarly, {pu'i} is {ka'e je puca'a}. I don't
>really understand why they are part of the language. They seem to be possible
>translations of "could", but they don't seem that useful to me to have a
>special word.
For that matter, what's the pragmatic difference between "ca'a" and "pu'i"?
Both mean the thing happened, but only one states that the thing was
possible. But the reality of an action logically entails the possibility of
the action, doesn't it?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Tue, 25 Oct 1994 11:29:54 -0400
From: Logical Language Group <lojbab>
Subject: Re: mi na nu'o catra ko'a
Cc: lojbab@access.digex.net
Be aware Chris that you were auto-deleted due to a "user unknown" bounce
form Lojban List about 3 hours before this message was posted.
"pu'i" is more or less a past perfective, whereas ca'a (unmodified)
means that it actually is happening right now.
lojbab
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: "any"
I've been trying to do more lojban writing lately, and I have found that
opaqueness comes up more often than I would have thought, and again and
again I've found myself wondering whether to use Jorge's xe'e, or a tanru,
or what to do, and I've ended up using "lo" because it "felt right", despite
my previous opinions on the matter (I more or less agreed with Jorge).
I've been avoiding diving into the any/opaque debate, because it's
complicated and I'm not sure I'm keeping up with it properly. Specifically,
I've forgotten some of the arguments that convinced me that allowing "lo" to
be opaque was a bad idea.
But anyway, here goes: as you may have noticed I just LOVE making little
tables to help me understand things:
T=transparent, O=opaque, V=veridicial, NV=non-veridical
Jorge's system Lojbab's system
-------------- ---------------
T/V lo broda da poi broda
T/NV le broda le broda
O/V xe'e lo broda OR (.ai/ko + lo) lo broda
O/NV xe'e le broda OR (.ai/ko + le) ??
I haven't heard yet if Lojbab agrees with Jorge about the opacity of ko,
.ai, and some other attitudinals, and I haven't actually seen Jorge use
"xe'e le" but I'm extrapolating.
Now that I lay it out this way, it doesn't look like there's that much
difference except for notation, although I'm sure I'm glossing over some
subtle effects on the overall interpretation of lojban bridi based on the
different understandings of "lo broda".
I guess I'm leaning towards Lojbab's system because 1) opaqueness crops up a
lot and so Lojbab's is more Zipfy, and 2) in my personal usage veridiciality
seems to correlate with opacity.
But how does Lojbab handle an opaque non-veridicial reference? Could such a
thing actually be useful?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: jorge@PHYAST.PITT.EDU
Subject: Re: "any"
la kris cusku di'e
> I've forgotten some of the arguments that convinced me that allowing "lo" to
> be opaque was a bad idea.
An example: mi klama lo zarci
Does it mean "There is at least one store such that I go to it" (transparent)
or "I go to a store (but there isn't any one store that has the property that
I am going to it)" (opaque) , something like "I go shopping".
> But anyway, here goes: as you may have noticed I just LOVE making little
> tables to help me understand things:
>
> T=transparent, O=opaque, V=veridicial, NV=non-veridical
>
> Jorge's system Lojbab's system
> -------------- ---------------
> T/V lo broda da poi broda
> T/NV le broda le broda
> O/V xe'e lo broda OR (.ai/ko + lo) lo broda
> O/NV xe'e le broda OR (.ai/ko + le) ??
That's not my system!!! :)
(I don't think non-veridicality is the defining property of {le}, so I will
switch to S=specific, NS=non-specific.)
Jorge's real system
-------------------
S le broda
NS/T lo broda
NS/O xe'e lo broda - lo'e broda
The dichotomy transparent/opaque can only occur in the non-specific case.
In the specific case, the quantifier is always {ro}, and I can't give any
interpretation to an opaque {ro}.
NOTE 1: {le} with any quantifier other than its default {ro} becomes
non-specific. {re le broda} means "two of the broda", but it is not
specified which two. {le re le broda} is specific again ("the two of the
broda") and its quantifier is of course {ro}.
NOTE 2: {lo} with quantifier {ro} becomes specific, at least for all
practical purposes, since a claim made about every possible broda leaves
no doubt about to which specific broda the claim applies.
[Also, I'm not sure, but I think the distinction Lojbab makes between
{da poi broda} and {lo broda} is not one of transparent vs. opaque but one
of existence-claimed vs. not existence-claimed. That would mean that the
quantifier of {lo broda} changes to something other than {su'o} when no
broda exists, but {lo broda} means the same as {da poi broda} when at
least one broda exists.]
> I haven't heard yet if Lojbab agrees with Jorge about the opacity of ko,
> .ai, and some other attitudinals,
I'd like to know, too.
> and I haven't actually seen Jorge use
> "xe'e le" but I'm extrapolating.
I would define {xe'e le} as {su'o xe'e le}, because as I said {ro xe'e}
makes no sense to me.
> I guess I'm leaning towards Lojbab's system because 1) opaqueness crops up a
> lot and so Lojbab's is more Zipfy,
So to say "I go to a store", you'd say {mi klama da poi zarci}? I don't think
making opaqueness the default is more Zipfy (but I don't think that's what
Lojbab proposes either).
> and 2) in my personal usage veridiciality
> seems to correlate with opacity.
Examples?
> But how does Lojbab handle an opaque non-veridicial reference? Could such a
> thing actually be useful?
If you mean opaque specific, I can't think of anything like that. Opaqueness
is either a subclass of non-specificity, or it is a third category by itself.
Jorge
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: any & every
The English words, all, any, every, and each are all compressed
into the universal quantifier when expressed in predicate calculus.
Any subtleties they may have can only be expressed in predicate
calculus by sequencing the universal quantifier, or altering its scope.
I am sharing some examples from a logic text below which deal with the
question of "any" vs. "every". I have changed the metaphor to a pool
game as first used by PC, but the logical form of examples (1-2') is
from a logic text.
1). No ball entered every pocket.
2). No ball entered any pocket.
It is pretty clear from these that "every" \= "any"
Here is the textbook translation of these into predicate calculus:
1') -E(x){ball(x) & All(y)[pocket(y) => entered(x,y)]}
-------------------------
2') All(y){pocket(y) => -E(x)[ball(x) & entered(x,y)]}
-------------------------------------------
The lines indicate the scope of the universal quantifier. It is longer
for the "any" example. Apparently it has something to do with the fact
that these statements are negated, but I can't say that I understand
this.
Because lojban grammar is based on predicate calculus it is a fairly
easy matter to translate these into lojban, but I am not going to do it
here as I doubt that anyone would use these forms. It is like expressing
the number 5. as s(s(s(s(s(0))))).
Shifting the metaphor to the one raised by Jorge, one could say:
3). No person needs any box.
3'). All(y){box(y) => -E(x)[person(x) & needs(x,y)]}
Now, suppose that 3 was negated by putting "It is not the case that"
in front of it. I read this as saying , " a person needs any box."
Or, suppose that the -E(x) etc. were simply changed to E(x)etc in 3'
above. Does that say: some person needs any box? Or can "any" only be
expressed in the negative with predicate calculus and hence lojban?
Why don't we just use xe'e for "any" and be done with it? Because
"any" has the meanings of: one indiscriminatly taken; of some; of all;
and of (one, some, or all). Negation seems to contort it further.
Context determines which is meant and hence the word is not parseable.
In short this is one of those places where we have an opportunity to
vastly improve English, if we can just sort it all out.
As I posted previously, there are at least three "anys", I now believe
there are the 4 mentioned above. I call them alpha, sigma, zeta, and
rho. They are all quantifiers. Does anyone want to go for 5?
mi nitcu rho danfu
djer
From: ucleaar <ucleaar@ucl.ac.uk>
Subject: Why haven't we adopted xehe yet?
Jorge has defined his "xehe" very clearly, and it has been amply
established that it is needed, since noone has come up with a
satisfactory alternative, and noone has demonstrated that Jorge
is deluded in thinking "xehe" to be necessary.
Therefore, how about shifting the debate from the meaning of "any"
to reasons for not adopting "xehe"? If noone comes up with a good
argument against it (apart from inertia of status quo) - and I
see no evidence that anyone will - we can conclude that the matter
has been intensively debated and the case for "xehe" is irresistable.
---
And
From: jorge@PHYAST.PITT.EDU
Subject: Re: any & every
la djer cusku di'e
> 1). No ball entered every pocket.
> 2). No ball entered any pocket.
>
> 1') -E(x){ball(x) & All(y)[pocket(y) => entered(x,y)]}
> -------------------------
> 2') All(y){pocket(y) => -E(x)[ball(x) & entered(x,y)]}
> -------------------------------------------
>
> Because lojban grammar is based on predicate calculus it is a fairly
> easy matter to translate these into lojban, but I am not going to do it
> here as I doubt that anyone would use these forms. It is like expressing
> the number 5. as s(s(s(s(s(0))))).
You can express both simple forms in Lojban:
1'') no bolci pu nerkla ro kevna
No ball entered every pocket.
2'') ro kevna pu se nerkla no bolci
Every pocket was entered by zero balls.
The distinction every/any here allows you to reverse the order of quantifiers
in English, without having to reverse the order in which you say the
arguments. In Lojban you have no choice but to reverse the order of the
arguments (or use quantifiers in the prenex). {xe'e} doesn't help you here,
because it is not the right word to translate the "any" of (2).
This is not the problem in the case of the opaque "any". In the opaque
case, rearranging the arguments doesn't solve the problem.
> Shifting the metaphor to the one raised by Jorge, one could say:
>
> 3). No person needs any box.
> 3'). All(y){box(y) => -E(x)[person(x) & needs(x,y)]}
>
> Now, suppose that 3 was negated by putting "It is not the case that"
> in front of it. I read this as saying , " a person needs any box."
No, the negation of (3) is "at least one person needs at least one box".
Again, the opaque case doesn't appear here. In Lojban, you'd have:
3'') ro tanxe cu se nitcu no prenu
Every box is needed by zero persons
Which can be re-expressed as:
ro da poi tanxe no de poi prenu zo'u de nitcu da
ro da poi tanxe na su'o de poi prenu zo'u de nitcu da
na su'o da poi tanxe su'o de poi prenu zo'u de nitcu da
lo prenu na nitcu lo tanxe
And its negation is:
lo prenu cu nitcu lo tanxe
At least one person needs at least one box.
But this is the transparent {lo prenu cu nitcu lo tanxe}. Opaque arguments
are not strictly arguments in the logical sense, but rather they modify the
relationship, so it is not simply a matter of order of quantification.
> Or, suppose that the -E(x) etc. were simply changed to E(x)etc in 3'
> above. Does that say: some person needs any box? Or can "any" only be
> expressed in the negative with predicate calculus and hence lojban?
The meaning of "any" in a negative sentence is different form its meaning
in an affirmative sentence in English. The meaning of the negative sentence
is clear in predicate calculus. The meaning in affirmative sentences is not
always straightforward.
Actually, even in negative sentences you can have opaque meanings. Compare
I don't need any box. = I need no box. (transparent)
mi na nitcu lo tanxe
It is not the case that there is a box that I need.
I don't need just any box. (opaque)
mi na nitcu xe'e lo tanxe
It is not the case that I need any box whatsoever.
> Why don't we just use xe'e for "any" and be done with it? Because
> "any" has the meanings of: one indiscriminatly taken; of some; of all;
> and of (one, some, or all). Negation seems to contort it further.
Nobody is asking for a word to cover all the meanings of "any". Many of
those meanings are already covered in Lojban.
Jorge
From: Chung-chieh Shan <ken@MATH.NTU.EDU.TW>
Subject: xehe
Would someone (Jorge perhaps) reiterate the definition/meaning of
xehe, preferably with some examples. I'm sorry that I haven't been
able to follow up on all those discussion on xehe and related topics,
but I guess I'm not alone. :)
Thanks in advance...
----------
Chung-chieh (Ken) Shan ken@cauchy.math.ntu.edu.tw
"Ay, fashion you may call it. Go to, go to." -- Hamlet
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: any
Jorge quotes me(djer):
la djer cusku di'e
> su'o da zo'u tu'e da plise inaja mi cidja da
> I eat some apples. (They exist).
This is not what the Lojban sentence claims. In Lojban, you wrote:
"there is some x, such that if it is an apple, I eat it". This statement
is always true. It suffices to select some non-apple for da, and since
in that case {da plise} is false, the whole statement is true for at least
that da whether I eat it or not, and therefore the statement is true.
GK>
What the sentence claims is:
IF there is at least one object x such that x is an apple,THEN I
eat x.
I believe you have misplaced your IF.
For my sentence to be always true it would have to be a
tautology. It has the logical form of a material implication,
which is by definition not a tautology. Here is a truth table
for inaja and for a tautology, P=P.
P Q P=>Q P=P
TT T T
TF F T
FT T T
FF T T
It is possible to produce a false implication, for example, by
saying: There is an apple, and I do not eat it. My statement is
not trivially always true.
Jorge>
The normal way to say "I eat the apple" is {mi citka le pa plise}.
You can say {mi citka lo pa plise}, but then you really mean that only
one apple exists.
GK> Agreed. My purpose was to express the quantifier Iota(x). This
quantifier asserts the existence of a unique object.
Jorge>
I think by "l'alfa" and "l'sma" you mean the same I wanted to get with
{pa xe'e} and {su'o xe'e}.
GK> Not exactly. Alpha-any means " one taken at random, or
indiscriminately". Pa-xe'e can mean one taken according to an
order or plan. Sigma-any means " more than one and less than
all." Su'o-any means "at least one". I didn't pull these
definitions out of a hat. You can read them in a dictionary.
Not that I ever would have even noticed the "any" problem if
you hadn't brought it up. But I think you have found a
fundamental deficiency in lojban in its inability to easily
express the nuances of "any". Lojban has the same problem as
predicate calculus in this regard. This is because it has the
same quantificational scheme. I find it strange that I have
wound up arguing for the change, and you seem to be opposed.
Jorge>
I don't understand why you say that one should be a quantifier and the
other an article.
GK> It had to do with the semantic equivalence of a definite
description and a certain combination of quantifiers with
existence. It is a side issue and I changed my position as you
can read in my any & every post.
Jorge
djer
X-Sender: cbogart@teal.csn.org
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 27 Oct 1994 12:17:54 -0600
From: cbogart@quetzal.com (Chris Bogart)
Subject: Re: mi na nu'o catra ko'a
X-Mailer: <PC Eudora Version 1.4b17>
>Since they aren't defined in terms of the event contours, and are grammatically
>distinct from them, we cannot list them with the contours. We might have
>to do some careful thinking to explain how/why they are distinct, but I am
>sure that the distinction is real if only BECAUSE of the grammatical
>difference.
I can see that; I guess I was exaggerating when I said they ought to be
listed with the contours. The real problem is a confusion of "ba" and "pu"
at the end of Section 19 of the Imaginary Journeys paper which makes it hard
to understand the discussion of what "pu'i" and "nu'o" mean in the future
tense. (as a newcomer those papers and the textbook are my only windows
into the language, so they're my bible.)
>I won't comment on opaque non veridical, since I am not sure I can think of
>such a beast.
"I need an elephant", when referring, in context, to any one of a stack of
boxes with a picture of an elephant on it. Maybe "mi nitcu lo me le xanto"?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: jorge@PHYAST.PITT.EDU
Subject: Re: any
djer:
> > su'o da zo'u tu'e da plise inaja mi cidja da
> > I eat some apples. (They exist).
xorxes:
> This is not what the Lojban sentence claims. In Lojban, you wrote:
> "there is some x, such that if it is an apple, I eat it". This statement
> is always true. It suffices to select some non-apple for da, and since
> in that case {da plise} is false, the whole statement is true for at least
> that da whether I eat it or not, and therefore the statement is true.
djer:
> What the sentence claims is:
> IF there is at least one object x such that x is an apple,THEN I
> eat x.
> I believe you have misplaced your IF.
What the sentence claims is:
There is at least one x such that: IF x is an apple THEN I eat x.
Notice that the quantification is in the prenex before the {tu'e}, and
therefore does not apply only to the first part but to the whole thing.
> For my sentence to be always true it would have to be a
> tautology.
No, it is not a tautology. It is always true because {da naku plise} is always
true. At least one thing is not an apple. The statement would not be always
true only in a universe where everything is an apple.
> Jorge>
> I think by "l'alfa" and "l'sma" you mean the same I wanted to get with
> {pa xe'e} and {su'o xe'e}.
>
> GK> Not exactly. Alpha-any means " one taken at random, or
> indiscriminately". Pa-xe'e can mean one taken according to an
> order or plan.
Not my {xe'e}. The way I'd define it is something like: "Potentially everyone
but only one and no other, and none fits the predication in actuality". It is,
as you say, not definable within predicate calculus, as I understand it.
> Sigma-any means " more than one and less than all."
That sounds like {so'i}, or some other of the series.
> Su'o-any means "at least one".
{su'o} by itself already is "at least one". {su'o xe'e} would be: "Potentially
every one and it must be one or more (potentially), but none in actuality".
> I find it strange that I have
> wound up arguing for the change, and you seem to be opposed.
It wouldn't be the first time that I argue for opposing sides, but no,
I'm still in favour of {xe'e}, as long as it is not used without thinking
for every meaning of "any".
Jorge
From: Goran Topic <topic@math.hr>
Subject: preti zo goi .e zo nai
xu gendra je xamgu fa lu la broda goi ko'a li'u ji
lu la broda ku goi ko'a li'u ji
lu la goi ko'a broda li'u ji
le na se cusku be vi dei
.i le pamoi ka'e se jijnu .ijeku'i pe'i zo goi lasna fi zo broda .enai
lu la broda li'u .i le remai drani gi'enaiku'i xamgu ca'i la zipf. .i
tu'a le cimai se smadi mi
.i lenu ji'a le zo bai selma'o zo nai lidne cu cumki xu
.i mupli fa lo'u cu'unai ko'a le'u
.i ri gerna nagi'a se smuni
lu ko'a na cusku da li'u ji
lu na'e ko'a cusku da li'u ji
lu ko'a na'e cusku da li'u ji
lu ko'a cusku na'e da li'u
.i to le re fanmo cu na'e lakne toi
.i ki'e re'i co'o mi'e. goran.
--
Learn languages! The more langs you know, the more incomprehensible you can get
e'udoCILreleiBANgu.izo'ozo'onairoBANguteDJUnobedocubanRI'a.ailekadonaka'eSELjmi
From: Goran Topic <topic@math.hr>
Subject: any, opaque, transparent, xe'e...
la xorxes. cusku di'e:
> The normal way to say "I eat the apple" is {mi citka le pa plise}.
> You can say {mi citka lo pa plise}, but then you really mean that only
> one apple exists.
Just a moment, and let me check if my understanding is correct...
Doesn't it say that onlspacetime interval? Like, {mi citka lo pa plise be vi le
vi jumbe pagbu}?
And one other thing... About the opaque/transparent thing... I have been
listening to the discussion for about a month (or more?), and I basically
know what is wrong with "any", but I can't figure out what in fact do
"opaque" andate, but my mind
usually just skips over these words, not being able to parse them.
And, some definition of {xe'e} a bit more clear than
> Not my {xe'e}. The way I'd define it is something like: "Potentially everyone
> but only one and no other, and none fits the predication in actuality".
I think I got that it means one and just one arbitrarily chosen from whole
set (if used with default {lo} quantifiers), but I can't fathom the meaning
of "none fits the predication in actuality".
Thanks, and
co'o mi'e. goran.
--
Learn languages! The more langs you know, the more incomprehensible you can get
e'udoCILreleiBANgu.izo'ozo'onairoBANguteDJUnobedocubanRI'a.ailekadonaka'eSELjmi
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: any & every
Jorge quotes me (djer):
la djer cusku di'e
> 1). No ball entered every pocket.
> 2). No ball entered any pocket.
>
> 1') -E(x){ball(x) & All(y)[pocket(y) => entered(x,y)]}
> -------------------------
> 2') All(y){pocket(y) => -E(x)[ball(x) & entered(x,y)]}
> -------------------------------------------
>
> Because lojban grammar is based on predicate calculus it is a fairly
> easy matter to translate these into lojban, but I am not going to do it
> here as I doubt that anyone would use these forms. It is like expressing
> the number 5. as s(s(s(s(s(0))))).
You can express both simple forms in Lojban:
1'') no bolci pu nerkla ro kevna
No ball entered every pocket.
2'') ro kevna pu se nerkla no bolci
Every pocket was entered by zero balls.
The distinction every/any here allows you to reverse the order of quantifiers
in English, without having to reverse the order in which you say the
arguments. In Lojban you have no choice but to reverse the order of the
arguments (or use quantifiers in the prenex). {xe'e} doesn't help you here,
because it is not the right word to translate the "any" of (2).
-----------------------------------------------------
GK> Its not so easy as you think. Consider this scenario: The white
ball is marked with an X. During the course of a game it happens to get
hit into pockets one through six. So we can say as above in 1.), except
for the negation:
E(x)( ball(x) & All(y)( pocket(y) => entered(x,y)))
Notice that the other balls went into the pockets. When this sentence
is negated as in 1), it just denies the existence of the white ball
with the X on it, or any that behaved similarly. It doesn't say all the
pockets are empty.
Your sentence, 1'' says that "0 balls entered every pocket." You can't
translate the lojban word "no", which means the number 0, into the
English word "no" which is a logical connective, and make sense.
If you are inclined to argue about the meaning of the above sentence, I
suggest you look at page 215 of the book Logic and Prolog, Cambridge
University Press, which is where I learned about it.
I havn't digested your other comments to my any & every post, I
hope others will comment.
djer
From: jorge@PHYAST.PITT.EDU
Subject: Re: preti zo goi .e zo nai
la goran cusku di'e
> xu gendra je xamgu fa lu la broda goi ko'a li'u ji
> lu la broda ku goi ko'a li'u ji
> lu la goi ko'a broda li'u ji
> le na se cusku be vi dei
(to e'u ko na pilno ne'i le pa preti zo xu e zo ji toi)
i pe'i le ci selsku cu smuni le pa da
i ju'o le ci da gendra je xamgu
i zo'o mi na djuno le du'u xukau le na se cusku be di'u cu gendra je xamgu
i ki'a le na se cusku
> .i le pamoi ka'e se jijnu .ijeku'i pe'i zo goi lasna fi zo broda .enai
> lu la broda li'u
i ko na krici le du'u la parser cusku makau
i zo goi roroi lasna lo sumti lo sumti
>.i le remai drani gi'enaiku'i xamgu ca'i la zipf.
i ie lei famyma'u cu tolmelbi
>.i tu'a le cimai se smadi mi
i drani smadi
> .i lenu ji'a le zo bai selma'o zo nai lidne cu cumki xu
i go'i
> .i mupli fa lo'u cu'unai ko'a le'u
> .i ri gerna nagi'a se smuni
> lu ko'a na cusku da li'u ji
> lu na'e ko'a cusku da li'u ji
> lu ko'a na'e cusku da li'u ji
> lu ko'a cusku na'e da li'u
> .i to le re fanmo cu na'e lakne toi
i pe'i lu cu'unai ko'a li'u se smuni lu da naku se cusku ko'a li'u
i ri frica lu ko'a na cusku da li'u
(to lo'u na'e ko'a le'u na gendra i lu na'e bo ko'a li'u toi)
> .i ki'e re'i co'o mi'e. goran.
je'e co'o mi'e xorxes
From: jorge@PHYAST.PITT.EDU
Subject: Re: any & every
I translated this two sentences;
> > 1). No ball entered every pocket.
> > 2). No ball entered any pocket.
with result:
> 1'') no bolci pu nerkla ro kevna
> No ball entered every pocket.
>
> 2'') ro kevna pu se nerkla no bolci
> Every pocket was entered by zero balls.
This means that the effect of reversing the universal and negated existencial
quantifiers, that is achieved in English by changing from every to any, can
only be achieved in Lojban by actually reversing the arguments.
(This "any" is not {xe'e}.)
djer disagrees:
> GK> Its not so easy as you think. Consider this scenario: The white
> ball is marked with an X. During the course of a game it happens to get
> hit into pockets one through six.
Then you can say: su'opa bolci pu nerkla ro kevna
At least one ball (to wit, the white one)
entered every pocket
> So we can say as above in 1.), except
> for the negation:
>
> E(x)( ball(x) & All(y)( pocket(y) => entered(x,y)))
Yes.
> Notice that the other balls went into the pockets. When this sentence
> is negated as in 1), it just denies the existence of the white ball
> with the X on it, or any that behaved similarly. It doesn't say all the
> pockets are empty.
Of course not. 1'') doesn't say that either.
2) = 2'') both say that all pockets are empty.
> Your sentence, 1'' says that "0 balls entered every pocket."
Maybe the English translation is confusing, but 1'') does not say that
every pocket is empty. It simply says that the number of balls that
entered all of the six pockets is zero. If each pocket was entered by
one different ball, it is still true that {no bolci pu nerkla ro kevna}.
If you say {pa bolci pu nerkla ro kevna} you claim that one ball entered
each and every pocket, the same ball. If each pocket has one ball, but
not the same one, then it is false that {pa bolci pu nerkla ro kevna},
but true that {ro kevna pu se nerkla pa bolci}.
> You can't
> translate the lojban word "no", which means the number 0, into the
> English word "no" which is a logical connective, and make sense.
Yes you can, in most cases. {noda} can always be replaced by {naku su'oda}.
If we do this in this case:
no bolci pu nerkla ro kevna
naku su'o bolci pu nerkla ro kevna
It is false that at least one ball entered every single pocket.
No ball entered every pocket.
> If you are inclined to argue about the meaning of the above sentence, I
> suggest you look at page 215 of the book Logic and Prolog, Cambridge
> University Press, which is where I learned about it.
I don't understand your comment. Sentences 1) and 2) mean two different
things. I don't see that we have any disagreement on what the English
sentences mean, so what can I check in that book? What we seem to disagree
on is the meaning of the Lojban sentence 1'').
Jorge
From: jorge@PHYAST.PITT.EDU
Subject: Re: any, opaque, transparent, xe'e...
la goran cusku di'e
> > The normal way to say "I eat the apple" is {mi citka le pa plise}.
> > You can say {mi citka lo pa plise}, but then you really mean that only
> > one apple exists.
>
> Just a moment, and let me check if my understanding is correct...
> Doesn't it say that onlspacetime interval? Like, {mi citka lo pa plise
> be vi le vi jumbe pagbu}?
Good point. I'd like to hear John Cowan's comment on that.
> And one other thing... About the opaque/transparent thing... I have been
> listening to the discussion for about a month (or more?), and I basically
> know what is wrong with "any", but I can't figure out what in fact do
> "opaque" andate, but my mind
> usually just skips over these words, not being able to parse them.
Well, consider "I need a box".
That can mean (at least) two things.
In the transparent sense, it means "there is a box such that I need it".
If you go over every single thing that is a box in your universe of
discourse, you can find one with the property that I need it.
That is what {mi nitcu lo tanxe} means. When you go over all things that
tanxe, I'm saying that at least one of them is needed by me.
But "I need a box" can have an opaque meaning as well. In this case, I'm
not claiming a property of any of the things that are boxes. I'm only
claiming something about me. Boxes could even not exist and the claim
would still be true. Or every single box could serve for my purposes,
and I still would need only "a box".
This opaque meaning is not possible in {mi nitcu lo tanxe} if we want
to keep Lojban logical.
For that I proposed {mi nitcu xe'e tanxe}. I also think that {mi nitcu
lo'e tanxe} works for this.
> And, some definition of {xe'e} a bit more clear than
>
> > The way I'd define it is something like: "Potentially everyone
> > but only one and no other, and none fits the predication in actuality".
>
> I think I got that it means one and just one arbitrarily chosen from whole
> set (if used with default {lo} quantifiers), but I can't fathom the meaning
> of "none fits the predication in actuality".
Well, because you are not really choosing one. You can't choose one and say
that the claim is true for that one, no matter how arbitrarily you choose it.
The claim only is valid before the selection. There is no actual referent
that fits the relationship. That is what I tried to say by making the potential
vs. actual distinction. I know it's not a clear definition, but I can't do
better for the moment.
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: any & every
la xorxes cusku di'e
> Date: Fri, 28 Oct 1994 20:16:13 EDT
> From: jorge@PHYAST.PITT.EDU
> Subject: Re: any & every
[...]
> If you say {pa bolci pu nerkla ro kevna} you claim that one ball entered
> each and every pocket, the same ball. If each pocket has one ball, but
> not the same one, then it is false that {pa bolci pu nerkla ro kevna},
> but true that {ro kevna pu se nerkla pa bolci}.
Here I must disagree. Sumti conversion DOESN'T change the LOGICAL meaning
of a sentence, only the EMPHASIS. IMHO the claim that one and the same
ball ( = a ball ) entered each and every pocket would be in Lojban
le [bi'u] [pa] bolci pu nerkla ro le kevna
{pa bolci pu nerkla ro le kevna} and all its conversions just claim that the
number of balls entering was one. I may be in minority, but I still claim
that a sumti with an external quantifier basically claims the number,
just the way {a ball} and {one ball} differ in English. {A ball entered
all the the pockets} claims that one and the same ball entered, {one ball
entered..} claims just the number of balls entering each pocket.
a ball = le [bi'u] [pa] bolci
one ball = pa bolci
{Two balls entered all pockets} is, at least slightly, ambiguos in English
but can be disambiguated in Lojban using the above device. I think {two
of the balls entered all pockets} (which again isn't totally unambiguos
but most often implies 'the same two') ought to be rendered as
le [bi'u] re le bolci nerkla ro le kevna
i.e. with an inner quantifier to fix the identities of the balls to be
the same at each pocket.
re le bolci nerkla ro le kevna
doesn't, IMHO, claim that the same two balls entered all the pockets
(neither does it exclude the possibility).
(My command of English isn't perfect so I hope to be forgiven if I
always don't get the shades of meaning quite right.)
> Jorge
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: The problem with external quantifiers
I may again be out of my depth but here goes.
In Lojban we have basically considered a bridi as a function
with as many arguments as there are sumti. Everything works
out alright as long as we don't have external quantifiers,
i.e. the ability to answer the question 'How many?'.
When we, however, introduce external quantifiers this model
breaks down. We would actually need twice as many arguments
as there are sumti because quantification is predication
about predicates.
Those of you who know Prolog might try to work this out :-)
I think many problems concerning external quantification,
existence, opaqueness/transparency etc. can only be solved
by resorting to second-order predicates. I'll leave that
to those who are better qualified.
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: jorge@PHYAST.PITT.EDU
Subject: Re: any & every
la veion di'e mi spusku
> > If each pocket has one ball, but
> > not the same one, then it is false that {pa bolci pu nerkla ro kevna},
> > but true that {ro kevna pu se nerkla pa bolci}.
>
> Here I must disagree. Sumti conversion DOESN'T change the LOGICAL meaning
> of a sentence, only the EMPHASIS.
I think sumti conversion DOES change the logical meaning, except in special
(but quite common) cases.
The meaning DOES change: When a sumti with quantifier {ro} changes
position with a sumti with any other quantifier. (Only exception: ro=pa).
This follows directly from the fact that the universal and existential
logical quantifiers DON'T commute.
Example 1: ro da prami de
Everybody loves someone.
For each x, there is a y such that x loves y
de se prami ro da
Someone is loved by everybody.
For some y, every x is such that x loves y
Example 2: le ci nixli cu citka re plise
Each of the three girls eats two apples.
For each of the two girls, there are two apples, such that
the girl eats them.
re plise se citka le ci nixli
Two apples are eaten by the three girls.
There are two apples, such that each of the three
girls eats them.
You will find that any other example where {ro} changes position with a
non-{ro}, the logical meaning changes.
> IMHO the claim that one and the same
> ball ( = a ball ) entered each and every pocket would be in Lojban
>
> le [bi'u] [pa] bolci pu nerkla ro le kevna
Yes. In this case, you could even reverse it and get the same meaning, since
both are quantified with {ro}:
ro le kevna pu se nerkla le bolci
Each of the pockets was entered by the ball.
However, the non-specific claim:
pa bolci pu nerkla ro le kevna
There is one ball that entered each of the pockets
claims that one and the same ball (and only one) entered all the pockets.
> {pa bolci pu nerkla ro le kevna} and all its conversions just claim that the
> number of balls entering was one.
For each of them the number was one, or the number was one of the balls that
entered all of them?
> I may be in minority, but I still claim
> that a sumti with an external quantifier basically claims the number,
> just the way {a ball} and {one ball} differ in English. {A ball entered
> all the the pockets} claims that one and the same ball entered, {one ball
> entered..} claims just the number of balls entering each pocket.
>
> a ball = le [bi'u] [pa] bolci
> one ball = pa bolci
>
> {Two balls entered all pockets} is, at least slightly, ambiguos in English
> but can be disambiguated in Lojban using the above device. I think {two
> of the balls entered all pockets} (which again isn't totally unambiguos
> but most often implies 'the same two') ought to be rendered as
>
> le [bi'u] re le bolci nerkla ro le kevna
>
> i.e. with an inner quantifier to fix the identities of the balls to be
> the same at each pocket.
>
> re le bolci nerkla ro le kevna
>
> doesn't, IMHO, claim that the same two balls entered all the pockets
> (neither does it exclude the possibility).
I think you are saying that {ro} quantifiers should always be understood
to come first, no matter the order in which they appear in the sentence.
I don't think that is the standard interpretation.
Jorge
From: jorge@PHYAST.PITT.EDU
Subject: Re: The problem with external quantifiers
la veion cusku di'e
> In Lojban we have basically considered a bridi as a function
> with as many arguments as there are sumti. Everything works
> out alright as long as we don't have external quantifiers,
But we ALWAYS have external quantifiers! Sometimes they are not
explicited, but they are always there.
> i.e. the ability to answer the question 'How many?'.
To me, the more interesting question that quantifiers answer,
(or don't answer) is 'Which one(s)?'.
> When we, however, introduce external quantifiers this model
> breaks down.
No it doesn't! The model doesn't work without the quantifiers.
We'd be limited to make only the vaguest of claims.
> We would actually need twice as many arguments
> as there are sumti because quantification is predication
> about predicates.
I don't understand that.
> Those of you who know Prolog might try to work this out :-)
I don't know Prolog, maybe that's why I don't understand :)
> I think many problems concerning external quantification,
> existence, opaqueness/transparency etc. can only be solved
> by resorting to second-order predicates.
I don't think there is any problem with external quantifiers
and existence. It is not always obvious what a claim means,
because we are used to thinking in English or other languages
that do not always follow predicate logic in the most aseptic
way, but the claims in Lojban are always unambiguous in that
respect.
As for opaqueness/transparency, I agree that by using embedded
abstractions (if that's what you mean by second-order predicates)
you can always dispose of the opaqueness. {xe'e} is only a way
to make it explicit without resorting to that.
> I'll leave that to those who are better qualified.
Oops... Well, I already wrote my answer, so I'll send it anyway.
Jorge
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: any, opaque, transparent, xe'e...
>And one other thing... About the opaque/transparent thing... I have been
>listening to the discussion for about a month (or more?), and I basically
>know what is wrong with "any", but I can't figure out what in fact do
>"opaque" andate, but my mind
>usually just skips over these words, not being able to parse them.
A transparent reference is a linguistic reference to something; an opaque
reference is what *appears* to be a reference, but it doesn't *in fact*
reference any particular thing. It's *kind of* like an uninstantiated variable.
"I need that red box" is transparent, because, in context, there is some red
box that the speaker and listener both know is being referred to specifically.
"I need a red box" (in its usual interpretation) is opaque, because "a red
box" doesn't refer to any particular box, really it describes the kind of
need I have. We've been using "any" as shorthand for this situation, but
that's not perfectly correct: consider this spooky Hallowe'en dialog:
Santa Claus: "I need a red box to wrap Tommy's toy train."
Elf: "Here, use this one."
Santa: "No, I need a bigger one"
Obviously not just "any" red box was good enough for Santa. It's also been
suggested that "a red box" here actually does refer to something: one
unidentified member of the set of red boxes which exist. But that's not
strictly true either; consider:
Santa: "I need a red box to wrap Tommy's toy train!"
Elf: "I'll get some tape and red cardboard and make one for you"
or even:
Santa: "I need a red unicorn"
Elf: "Tough luck. The unicorn is a mythical beast."
You can't pin down what "a red box" is referring to because it doesn't refer
to anything. It merely describes (some of) the needed thing's required
characteristics.
Notice that that only happens in certain contexts, where there's a sort of
implied negative in the semantics of the verb, that isn't handled by the
usual negation mechanisms: "need", "want", "lack", are grammatically
positive but semantically negative. If you say "I smell a red box", or "I
see a red box", or "I sit on a red box", there's a particular box or boxes
being talked about, and the reference isn't opaque.
I have a suspicion that the *ideal* logical solution to all this, if we were
starting from scratch, is to eliminate words like nitcu, djica, and claxu
from the lexicon, and find a way of communicating those concepts that is
explicitly marked as negative. But I don't think that's very practical at
this point. Realistically I think we have to use either Lojbab's or Jorge's
marking schemes, or forbid opaque references altogether, using other
mechanisms like "mi tanxe nictu" to get the point across.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: jorge@PHYAST.PITT.EDU
Subject: Re: any, opaque, transparent, xe'e...
Thank you Chris!!! Excellent explanation of what opaque means.
> I have a suspicion that the *ideal* logical solution to all this, if we were
> starting from scratch, is to eliminate words like nitcu, djica, and claxu
> from the lexicon, and find a way of communicating those concepts that is
> explicitly marked as negative.
Or allow those words to only accept events, like was done for {djica} and
{sisku}, but not yet for {nitcu}, {claxu}, {kalte} and maybe some others.
I don't think that is the best solution, though.
> But I don't think that's very practical at
> this point. Realistically I think we have to use either Lojbab's or Jorge's
> marking schemes, or forbid opaque references altogether, using other
> mechanisms like "mi tanxe nictu" to get the point across.
What is it that you are calling Lojbab's scheme? Use of {tu'a}, or use of
{lo tanxe} for these predicates with a different meaning than for most
other predicates?
Jorge
From: Gerald Koenig <jlk@NETCOM.COM>
Subject: any & every & naku
Jorge said:
I translated this two sentences;
> > 1). No ball entered every pocket.
> > 2). No ball entered any pocket.
with result:
> 1'') no bolci pu nerkla ro kevna
> No ball entered every pocket.
>
> 2'') ro kevna pu se nerkla no bolci
> Every pocket was entered by zero balls.
This means that the effect of reversing the universal and negated existencial
quantifiers, that is achieved in English by changing from every to any, can
only be achieved in Lojban by actually reversing the arguments.
Then you can say: su'opa bolci pu nerkla ro kevna
At least one ball (to wit, the white one)
entered every pocket
GK> Your sentence, 1'' says that "0 balls entered every pocket."
Maybe the English translation is confusing, but 1'') does not say that
every pocket is empty. It simply says that the number of balls that
entered all of the six pockets is zero. If each pocket was entered by
one different ball, it is still true that {no bolci pu nerkla ro kevna}.
GK> You can't
> translate the lojban word "no", which means the number 0, into the
> English word "no" which is a logical connective, and make sense.
JL>Yes you can, in most cases. {noda} can always be replaced by {naku su'oda}.
If we do this in this case:
no bolci pu nerkla ro kevna
naku su'o bolci pu nerkla ro kevna
It is false that at least one ball entered every single pocket.
No ball entered every pocket.
What we seem to disagree on is the meaning of the Lojban sentence 1'').
Jorge
-------------------------------------------------------------------
GK> (djer) continues: I do continue to find the translation
1'' "no bolci pu nerkla ro kevna
very questionable. But I think your alternative form that you believe
equivalent,
1''' naku su'o bolci pu nerkla ro kevna
is very, very good and does offer a valid short form alternative to the
full logical expansion from predicate calculus without altering the
meaning.
When I tried to replace the "no" in your second translation,
2'' ro kevna pu se nerkla no bolci
with what you say is equivalent I got:
ro kevna pu se nerkla naku su'o bolci
which looks again very questionable. I now suggest:
2''' ro kevna pu na se nerkla su'o bolci
Each pocket was not entered by one or more balls; which carries the
meaning of "no ball entered any pocket" in a short form. I hope I got
that negation right. I just read the paper.
Your system of following the order of the predicate
calculus formulation of 2. and then converting the selbri gives a
really compact expression. So now we have:
No ball entered every pocket.
naku su'o bolci pu nerkla ro kevna
No ball entered any pocket.
ro kevna pu na se nerkla su'o bolci
The neat symmetry of your "no" formulation is lost. But finally the
meaning of"any" has been expressed in a compact form without actually
using the word. Think about this:
naku su'o bolci pu nerkla ro kevna
It is not the case that at least one ball entered every pocket.
No ball entered every pocket.
naku so'o bolci pu nerkla "zeta-any" kevna
It is not the case that at at least one ball entered (one, some, or
all) pockets.
No ball entered any pocket.
I previously defined zeta-any to mean: (one, some, or all). Or use xe'e
if it works here. I'm not sure how you defined it. The use of a word for
"any" requires a lot less mental gymnastics and maybe that's why it
exists.
Do you agree with this post?
djer
From: "John E. Clifford" <pcliffje@CRL.COM>
Subject: mi na nu'o catra ko'a
Given the _loj_ in _lojban_, which usually means at least that logically
relevant structure is explicit, it is probably a mistake to think that
_nu'o_, which is not obviously complex, is in fact complex, whatever the
gloss may say. Apparently, Lojban divides up the the spectrum of reality
and possibility in a different way from English, with several fundamental
items where we might use only a couple and complexes. Thus, denying one
Lojban item gives a disjunction of the other, not some simpler compound
of them and negations (though we might eventually prove a theorem of
equipollence)
But, does _na_nu'o_ in fact deny _nu'o_? _na_ has basic sentence scope.
Presumably _nu'o_, like the tenses does too -- or perhaps even broader,
as the tenses often do: ko'a na ba klama lo zarci rarely means "He will
never go to the store" but usually "He will not go to the store" on the
occasion we are interested in. Logically, the tense is outside the
negation here, F not Kxz rather than not F Kxz. So may it be with
_nu'o_. But that still means that _na_nu'o_ means something strange,
roughly "I could have avoided it but didn't" = "I didn't resist", which
is probably not what was meant.
pc>|83 Trying to be the conservative logical brake on the wheel of change.
From: "John E. Clifford" <pcliffje@CRL.COM>
Subject: some definitions
I am having trouble following parts of the discussions here because I do
not understand some words the ways they are being used. Since the issues
we are concerned with are logical, I find it handy to get definitions in
terms of the basics of logic: truth conditions or validated inferences.
Anyhow, here are a few terms as I use them (standard stuff, I think) and
a few requests for help with some other terms.
OPAQUE (referentially opaque context) Terms which occur in opaque
contexts cannot be exported to surrounding transparent contexts, cannot
be generalized by nor used to instantiate external quantifiers, doe not
replace under external identitities, if bound need not instantiate to
external objects. (I admit that this collapses several notions into one,
but most logicians hold that they go together.)
EXTENSIONAL The truth of every complex is a direct function of the truth
of its constituents. Hence, referring expressions refer to and
quantifiers range over the established domain only. Extensional contexts
are, therefore, transparent (not opaque).
INTENSIONAL Not extensional, for short: the truth of some complexes are
not determmined by the truth of their components. Intensional contexts
may be opaque and, so, most (I would say "all") must be treated as such.
(The exact line between extensional and intensional is controversial and
system-bound. For example, modal and tense systems can be worked either
way, with no significant _formal_ difference).
INTENTIONAL. Having to do with what an agent intends: purpose, goal,
motive, etc. This somewhat elastic, since objective needs are sometimes
included along with subjective wants. These are always intensional.
I hope that helps. Now to help me, could someone please explain what
truth conditions or inferences are involved in "opaque" as a property of
sumti(? descriptors? it is not clear what is being classified here). In
the same discussion -- and maybe more critically -- could someone explain
"specific" and "definite". These two words are used interchangably by
some groups and by others to make any number of distinctions (often one
group uses one to make the distinction another makes with the other),
what exactly is going on here.
The discussion is not helped by the fact that some of the examples of
apparently non-controversial cases seem just wrong: _lo_brida_ keeps
appearing as a general term rather than a singular one, if translation
and inferences are any guide.
pc>|83 Trying to be the logical brake on the wheel of change
From: jorge@PHYAST.PITT.EDU
Subject: Re: any & every & naku
> > > 1). No ball entered every pocket.
> > > 2). No ball entered any pocket.
>
> > 1'') no bolci pu nerkla ro kevna
> > No ball entered every pocket.
> >
> > 2'') ro kevna pu se nerkla no bolci
> > Every pocket was entered by zero balls.
>
> -------------------------------------------------------------------
> GK> (djer) continues: I do continue to find the translation
>
> 1'' "no bolci pu nerkla ro kevna
>
> very questionable. But I think your alternative form that you believe
> equivalent,
(This equivalence is mentioned in one of the grammar papers, BTW)
> 1''' naku su'o bolci pu nerkla ro kevna
>
> is very, very good and does offer a valid short form alternative to the
> full logical expansion from predicate calculus without altering the
> meaning.
Notice that 1''' can also be written
su'o bolci na pu nerkla ro kevna
lo bolci na pu nerkla ro kevna
It is false that: at least one ball entered every pocket.
> When I tried to replace the "no" in your second translation,
>
> 2'' ro kevna pu se nerkla no bolci
>
> with what you say is equivalent I got:
>
> ro kevna pu se nerkla naku su'o bolci
>
> which looks again very questionable.
Looks can deceive :)
The sentence indeed means the same thing as 2''.
> I now suggest:
>
> 2''' ro kevna pu na se nerkla su'o bolci
This says:
It is false that: For every pocket, at least one ball entered it.
This is NOT equivalent to 2) = 2'')
> Each pocket was not entered by one or more balls;
No, be careful. The negation applies to the whole claim, not only to the
part after the universal quantifier.
I think that what you want is:
2'''') ro kevna naku pu se nerkla su'o bolci
which is equivalent to 2 and 2''.
> which carries the
> meaning of "no ball entered any pocket" in a short form. I hope I got
> that negation right. I just read the paper.
That explains your confusion. :) That paper is not one of my favourites.
> Your system of following the order of the predicate
> calculus formulation of 2. and then converting the selbri gives a
> really compact expression. So now we have:
>
> No ball entered every pocket.
> naku su'o bolci pu nerkla ro kevna
>
> No ball entered any pocket.
> ro kevna pu na se nerkla su'o bolci
The last one is wrong, but you can fix it by changing to {naku}.
> The neat symmetry of your "no" formulation is lost. But finally the
> meaning of"any" has been expressed in a compact form without actually
> using the word.
I still think the expressions with {no} mean what I said, but you are
welcome not to use them if you don't like them.
> Think about this:
>
> naku su'o bolci pu nerkla ro kevna
> It is not the case that at least one ball entered every pocket.
> No ball entered every pocket.
Agreed.
> naku so'o bolci pu nerkla "zeta-any" kevna
> It is not the case that at at least one ball entered (one, some, or
> all) pockets.
> No ball entered any pocket.
(one, some, all) is su'o = "at least one".
And that's what you arrive at if you keep manipulating. Start with 2'''':
ro kevna naku pu se nerkla su'o bolci
Now, if you pass a negation by a {ro} you change it to a {su'o}, so:
naku su'o kevna pu se nerkla su'o bolci
And finally, existential quantifiers commute, so:
naku su'o bolci pu nerkla su'o kevna
which is your "zeta-any" expression.
And you can even simplify it. {su'o bolci} is the same as {lo bolci}, and
a {naku} at the beginning of the sentence is the same as a {na} in front
of the selbri, so we end up with:
lo bolci na pu nerkla lo kevna
It is false that at least one ball entered at least one pocket.
Or more colloquially: No ball entered any pocket. We've come full circle.
> I previously defined zeta-any to mean: (one, some, or all). Or use xe'e
> if it works here.
No, it doesn't. {xe'e} is not a logical quantifier. Zeta-any is su'o.
> I'm not sure how you defined it. The use of a word for
> "any" requires a lot less mental gymnastics and maybe that's why it
> exists.
Yes, but "any" has many different meanings in English.
> Do you agree with this post?
With some parts. :)
Jorge
From: "John E. Clifford" <pcliffje@CRL.COM>
Subject: any
Aside from the problem with _lo_, _ko_cuxna_lo_karda_ finally convinced
me that we do need something more to d in Lojban what we do in English
with "any". The problem is that imperatives set up opaque contexts
without a word that sets the context. Thus, we cannot show the
quantifier outside the intensional context by putting it before the
context-forming word -- the classic way of dealing with a context leaper
in logic. Even fronting the quantifier leaves it in imperative mode: ko
cuxna ro karda = ro da poi karda gu ko cuxna da and both say "Pick every
card" (I suspect I need something stronger than just "gu" there) which is
as wrong as the forcer-deck _lo_ for what we want. We can stick with the
safe ko cuxna pa karda or su'o karda or da poi karda, but those, being in
an opaque context, need not be restricted to (or even include) the cards
actually presented.
Using the subject-raising cmavo won't help here, since we want exactly to
get out of the opaque context, not warn that we are still in it
(although, we might make it a toggle -- what is the happyface for "Yuck,
ptui"?)
I was glad to see that it seems to be accepted that opaque contexts arise
from event descriptions. It is not clear whether it is accepted
that that is the only source, as I was trying to make the case. But, in
any case, if this is to be an adequate explanation, we do need to do
something about "I saw someone playing pool", for, if that is an event
description it ruins the explanation, since it is transparent and thus it
is some factor or than event-descriptions which cause opacity. ("Obtains"
is also a problem but so strange that it can probably be handled by the
fact that "that p obtains" is in every way equipollent to just p, not
merely that the complex implies the simple.)
pc>|83 Trying to be the logical brake on the wheel of change
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: About outer quantification
I have been thinking quite a lot about outer quantification and
would like to present my present understanding. I have presented
some of these ideas already previously but at a 'IMHO' level.
Now I think I can present some justification for my case.
Outer quantification can be defined in a way which is
1) unambiguous in all cases
2) invariant in regard to sumti conversion
3) logically sound and comprehensible
Whether this way contradicts the 'official line' or not, I'll
leave to others to decide. I know it contradicts Jorge's view
of the matter :-)
If we have a number of balls and and a number of pockets
into which the balls may enter and re-enter, we have the
following subcases (e.g.)
a) there are 2 of the balls such that each of the 2
enters 3 pockets (which need not be the same)
b) there are 3 of the balls such that each of the 3
enter the same 2 pockets
c) there are 2 pockets such that 2 balls (which need
not be the same in each case) enter each of the 2
pockets
d) some 4 balls enter some 5 pockets
The Lojban equivalents are (? ; or could be)
a) [ro] le re le bolci nerkla ci le kevna
= ci le kevna se nerkla le re le bolci
the number of pockets each of the 2 balls enters is 3
b) [ro] le ci le bolci nerkla [ro] le re le kevna
= le re le kevna se nerkla le ci le bolci
each of the 3 balls enters each of the 2 pockets
c) re le bolci nerkla [ro] le re le kevna
= le re le kevna se nerkla re le bolci
the number of balls entering each of the 2 pockets is 2
d) vo le bolci nerkla ci le kevna
= ci le kevna se nerkla vo le bolci
the number of balls entering a pocket/pockets is 4 and
the number of pockets being entered by a ball/balls is 3
This means that an outer quantifier expresses the number of
the quantified sumti for which the bridi WITHOUT outer quantifiers
is true. In example (d) we have two INDEPENDENT quantifications.
A first-order predicate has NO (outer) quantifiers. Outer
quantification requires second-order predication where we have
a quantification function for each first-order argument we want
to quantify (implicitly for each sumti in a bridi).
>From this standpoint it is also clear that sumti conversion doesn't
interact with outer quantification. Outer quantification is outside
the scope of the first-order predication expressed by the unquantified
basic bridi in Lojban.
The various descriptors make it possible to express DEPENDENT
quantification and also to move quantification inside the first-order
bridi where the logic so dictates.
This definition of outer quantification requires that we use inner
quantification in a number of cases which at first may seem strange
from a NL point of view. After getting rid of all connotations of
NL articles, however, the inclusion of {le} doesn't, after a while,
feel so strange - and with the default outer quatifier {ro} it
actually reflects the logical situation quite accurately.
The default outer quantifiers
a) exist to make the system orthogonal while allowing elision
of unnecessary second-order arguments {su'o|ro}
b) are defined to have logically convenient values
{su'o|ro} as outer quantifiers are borderline cases as they also belong
to the first-order logic ( su'o ~ E(), ro ~ A() ).
In this system we can unambiguously make the following questions
and correspondigly get equally unambiguous answers:
1) [ro] le xo le bolci nerkla ci le kevna
2) [ro] le re le bolci nerkla xo le kevna
3) [ro] le xo le bolci nerkla xo le kevna
4) [ro] le xo le bolci nerkla [ro] le re le kevna
5) [ro] le ci le bolci nerkla [ro] le xo le kevna
6) [ro] le xo le bolci nerkla [ro] le xo le kevna
7) xo le bolci nerkla [ro] le re le kevna
8) re le bolci nerkla [ro] le xo le kevna
9) xo le bolci nerkla [ro] le xo le kevna
10) xo le bolci nerkla mu le kevna
11) vo le bolci nerkla xo le kevna
12) xo le bolci nerkla xo le kevna
To recapitulate:
The Lojban system of outer quantification overlays the basic
first-order structure with a number of second-order counter
functions indicating the number of instances of the associated
sumti for which the basic bridi is true. This means that actually
the truth value of the whole bridi depends on the counter functions
having the indicated values (or the default values in case the
quantifier is elided) and the basic first-order bridi is hidden
behind one level of abstraction (the counting). All sumti with an
outer quantifier EFFECTIVELY different from {ro} lose their identity
(are indefinite).
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: About outer quantification
While entering my posting I seem to have omitted one essential
clarification:
{ro} as an outer quantifier is a special case also in the sense
that the rest of the bridi WITH all the quantifiers intact must
hold for each and every instance of the sumti quantified by {ro}.
This gives us the ability to use an inner quantified {le}
description to define a specific number of instances for which
the rest of the bridi must hold.
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: any, opaque, transparent, xe'e...
>What is it that you are calling Lojbab's scheme? Use of {tu'a}, or use of
>{lo tanxe} for these predicates with a different meaning than for most
>other predicates?
I think a post of mine got lost somehow, which I'll quote from:
>T=transparent, O=opaque, V=veridicial, NV=non-veridical
>
> Jorge's system Lojbab's system
> -------------- ---------------
>T/V lo broda da poi broda
>T/NV le broda le broda
>O/V xe'e lo broda OR (.ai/ko + lo) lo broda
>O/NV xe'e le broda OR (.ai/ko + le) ??
I went on to point out that in my lojban writing I was noticing a
correlation between veridicial and opaque, which was causing me to lean more
towards Lojbab's system, for Zipfean reasons.
Lojbab wrote back and said that he couldn't fill in the question marks on
his side of the table because he couldn't imagine an opaque non-veridicial.
I think I can imagine one, but I don't think it's terribly useful.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: About outer quantification
I found out that a further clarification may be required.
When deriving the truth value of a bridi it is necessary
to omit the explicit outer quantifiers only in a sub-bridi
which contains non-{ro} outer quantifiers ({ro}s being
handled differently). The omissions amount to replacing
the explicit quantifiers with {su'o} (even before a {le}).
E.g. in the bridi
le re nanmu cu tavla re ninmu ci cukta
for each of the two nanmu (ro le re) it is necessary to consider
a bridi like
ko'a cu tavla re ninmu ci cukta
and count separately
a = the number of ninmu to whom ko'a speaks about su'o cukta
( ko'a tavla xo ninmu su'o cukta )
b = the number of cukta about which ko'a speaks to su'o ninmu
( ko'a tavla su'o ninmu xo cukta )
The original bridi is true if for each nanmu a = 2 and b = 3.
The total number of ninmu can range from 2 to 4 and the total
number of cukta from 3 to 6, i.e. there may or may not be an
overlap between the nanmu.
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: jorge@PHYAST.PITT.EDU
Subject: Re: About outer quantification
la veion cusku di'e
> Outer quantification can be defined in a way which is
>
> 1) unambiguous in all cases
>
> 2) invariant in regard to sumti conversion
>
> 3) logically sound and comprehensible
No doubt about it. It CAN be defined that way, but I believe it hasn't.
I think your system can be expressed thus: "All universal quantifiers
come before all existential quantifiers, no matter the order in which
they appear in the sentence."
Lojban's system (as defined in the grammar papers) is: "All quantifiers
are taken in the order in which they appear in the sentence".
Both systems are unambiguous, logically sound and comprehensible.
Lojban's is more flexible in one sense, yours is more flexible in
another sense.
> Whether this way contradicts the 'official line' or not, I'll
> leave to others to decide. I know it contradicts Jorge's view
> of the matter :-)
It's not just my view. Here's a quote from the negation paper:
>> (We must note that there is an exception to the blanket statement that
>> changing the order of a Lojban bridi has no effect on meaning. When there
>> are multiple quantified variables in a sentence and some are existential
>> ("da" or "su'oda") and some are universal ("rode"), this statement is true
>> only when these variables are quantified in the prenex. If the prenex is
>> omitted, the variables may be quantified "on the spot" in the sentence, but
>> the order of the quantifiers first appearance in the sentence cannot be
>> changed by a reordering of the sentence, without changing the meaning. This
>> special circumstance is the rough equivalent of the two English sentences
>> "Everybody loves someone." and "Someone is loved by everyone.", which are
>> significantly different in their claim. This special case has no particular
>> special effects on negation, since Lojban rules for transforming negations
>> always involve moving all quantified variables to the prenex, whence the
>> problem does not occur. See also Section 9.)
I think that your system (all universals first) usually mimics natural
language better, but Lojban's system is more like predicate logic.
There is one other difference:
> d) vo le bolci nerkla ci le kevna
> = ci le kevna se nerkla vo le bolci
With the standard interpretation as I understand it, the statement
(in both forms, since there are no universal quantifiers here, so they
are equivalent) claims 12 events of ball entering pocket.
> the number of balls entering a pocket/pockets is 4 and
> the number of pockets being entered by a ball/balls is 3
The ambiguous interpretation that you give, can only be obtained with:
lei vo le bolci cu nerkla lei ci le kevna
The four balls (as a mass) enter the three pockets (as a mass).
(Or piro loi, instead of lei, if you prefer.)
In any case, I don't find your interpretation unsound, only that it
differs from the one I was taught.
Jorge
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: About outer quantification
la xorxes cusku di'e
> Date: Sun, 30 Oct 1994 13:52:31 EST
> From: jorge@PHYAST.PITT.EDU
> Subject: Re: About outer quantification
> Lojban's system (as defined in the grammar papers) is: "All quantifiers
> are taken in the order in which they appear in the sentence".
>
> Both systems are unambiguous, logically sound and comprehensible.
> Lojban's is more flexible in one sense, yours is more flexible in
> another sense.
>It's not just my view. Here's a quote from the negation paper:
:-) That was the one paper I had given a none too thorough reading :-)
I had been been wondering were amongst the hundreds of pages I had
seen the thing spelled out.
> In any case, I don't find your interpretation unsound, only that it
> differs from the one I was taught.
I have no need of trying to replace a working system. I'm content
of having got mine specified in a reasonable way. Now I'll file
it under 'Exercises completed' (unless someone wants to discuss
the details of a hypothetical alternate system). But I really
liked the sumti conversion invariance - it would have made life
easier while rearranging complex sentences with relative clauses
etc. to a more readable/comprehensible form :-)
> Jorge
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: jorge@PHYAST.PITT.EDU
Subject: Re: some definitions
la pycy cusku di'e
> OPAQUE (referentially opaque context) Terms which occur in opaque
> contexts cannot be exported to surrounding transparent contexts, cannot
> be generalized by nor used to instantiate external quantifiers, doe not
> replace under external identitities, if bound need not instantiate to
> external objects. (I admit that this collapses several notions into one,
> but most logicians hold that they go together.)
This seems to be the way I was using "opaque". The key point is that an
opaque sumti does not instantiate to external objects, while I think
{lo broda} in {lo broda cu brode} should always instantiate, no matter
what {brode} is. That is a difference from English, where "a box" is
transparent in "I see a box" but opaque in "I need a box". I don't think
in Lojban the semantics of the selbri should influence the opaqueness of
its sumti. {mi nitcu lo tanxe} should behave like {mi viska lo tanxe}
as far as the instantiation properties of {lo tanxe} is concerned.
> INTENTIONAL. Having to do with what an agent intends: purpose, goal,
> motive, etc. This somewhat elastic, since objective needs are sometimes
> included along with subjective wants. These are always intensional.
Intentional has also been used to describe a property of {le}, which does
not have to do with agents, but with the speaker of the utterance.
This intentionality of {le} does not make the utterance intensional, it
only serves to specify the referent of the sumti.
> I hope that helps.
Sure does. Thanks for the definitions.
> In
> the same discussion -- and maybe more critically -- could someone explain
> "specific" and "definite". These two words are used interchangably by
> some groups and by others to make any number of distinctions (often one
> group uses one to make the distinction another makes with the other),
> what exactly is going on here.
I use "non-specific" to mean that the question "which one(s)?" has not been
in principle answered, but the question is relevant.
So, in {mi klama lo zarci}, {lo zarci} is non-specific, because I only
said that I go to "a market", no information was given to specify which
one, but the question "which one?" must have an answer, i.e. there _is_
a market such that I go to it.
The speaker need not know the answer to "which one?" for the claim to be
true, but there must be one.
In {mi klama le zarci}, {le zarci} is specific. The relevant question
if "the market" cannot be identified by the audience is not "which market?",
that has already been answered as "the market", the relevant question is
"what do you mean by 'the market'?", or "What is the market?".
In English, the dialogue: "I'm going to the market." "Which market?"
does make sense, but only because the assumption by the speaker that
the audience understands the referent of "the market" fails, not because
the speaker left it non-specific, as in "I'm going to a market", in which
case the audience has not been told the referent of "a market".
That seems suitably muddled.
> The discussion is not helped by the fact that some of the examples of
> apparently non-controversial cases seem just wrong: _lo_brida_ keeps
> appearing as a general term rather than a singular one, if translation
> and inferences are any guide.
I don't understand that. {lo broda} means (I hope) "at least one of
all things that broda". What would be general/singular in this case?
(Those weren't in the list of definitions :)
Jorge
From: jorge@PHYAST.PITT.EDU
Subject: Re: mi na nu'o catra ko'a
pc says:
> But, does _na_nu'o_ in fact deny _nu'o_? _na_ has basic sentence scope.
> Presumably _nu'o_, like the tenses does too -- or perhaps even broader,
> as the tenses often do: ko'a na ba klama lo zarci rarely means "He will
> never go to the store" but usually "He will not go to the store" on the
> occasion we are interested in. Logically, the tense is outside the
> negation here, F not Kxz rather than not F Kxz.
I thought truth values were atemporal, so it wouldn't make sense that
some claim is true now but not true in the future. Effectively, this
means that F and not must commute, which means that F can be neither
"at some time in the future" (existential quantification) nor "at every
time in the future" (universal quantification). It rather has to be
"at the one time in the future that I'm talking about", which commutes
with "not".
If the tenses are quantified with anything but unitary (is that the
right word?) quantification, then we will have ambiguities with sentences
with tense.
Jorge
From: Randall Holmes <holmes@CATSEYE.IDBSU.EDU>
Subject: The "problem" with external quantifiers
Veijo Vilva is out of his depth.
A predicate (bridi in Lojban) can be considered as a function
(with n arguments if the predicate has n arguments) taking values
in {true, false}. This does not change when quantifiers are introduced.
Nor are second-order predicates needed to handle these concepts
(universal quantification and existence). The use of second-order
or higher-order predicates is another idiom for the introduction
of sets, and nothing about first-order logic (the logic of
connectives ("and", "or", "if...then...", etc.) and quantifiers
(for all... there exists a...) requires the introduction of
second-order predicates.
The problems of opaqueness and transparency are not handled by
introducing second order predicates, either; they are handled
best by a careful distinction between use and mention (between
representations of things and concepts and the things and concepts
themselves).
In particular, it is certainly not the case that the number
of arguments of a predicate doubles when it is quantified, nor
is it the case that a quantifier has to be interpreted as a predicate
of predicates (it _can_ be interpreted as a predicate of predicates
in a system with higher-order predicates, but it does not have to
be; in first-order logic quantifiers are not explained in this way).
Pardon the lack of Lojban examples, but I'm with the competing brand!
--Randall Holmes. logician with TLI
P.S. I don't see the point of the appeal to Prolog.
From: Chris Bogart <cbogart@quetzal.com>
Subject: Re: any, opaque, transparent, xe'e...
>> I think a post of mine got lost somehow, which I'll quote from:
>
>No, it wasn't lost. Maybe my answer to it was lost.
The list server mysteriously dumped me off for a few days right about then
-- that must be why I didn't see it.
>In my answer I asked you for examples. I can't think of common examples
>not involving {nitcu}, {djica} and the four or five like them, where
>it would be more convenient to consider {lo broda} opaque.
I went through looking for words in that category and I found:
cpedu, cuxna, denpa?, facki, nitcu, djica, claxu, friti, drata?
but I didn't go through the whole gi'uste, so there are probably more.
Also some tenses can make other gismu have opaque places:
ka'e: *mi mutce xagji .oi .i mi ka'e citka xe'e lo xirma.
di'i: *mi di'i citka xe'e lo xirma
-roi: *mi ciroi citka xe'e lo xirma
It's interesting that in all these examples it's the x2 place that causes
the problem.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject: Re: The "problem" with external quantifiers
> Date: Mon, 31 Oct 1994 10:22:57 -0700
> From: Randall Holmes <holmes@CATSEYE.IDBSU.EDU>
> Subject: The "problem" with external quantifiers
> Veijo Vilva is out of his depth.
I agree. Got carried away. I better keep to things I can handle :-)
> --Randall Holmes. logician with TLI
--
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
From: "Mark E. Shoulson" <shoulson@CS.COLUMBIA.EDU>
Subject: Re: any, opaque, transparent, xe'e...
In-Reply-To: <199410291909.PAA17455@cs.columbia.edu> (message from Chris
Bogart on Sat, 29 Oct 1994 12:46:18 -0600)
>Date: Sat, 29 Oct 1994 12:46:18 -0600
>From: Chris Bogart <cbogart%QUETZAL.COM@CUVMB.CC.COLUMBIA.EDU>
>A transparent reference is a linguistic reference to something; an opaque
>reference is what *appears* to be a reference, but it doesn't *in fact*
>reference any particular thing. It's *kind of* like an uninstantiated
variable.
>"I need that red box" is transparent, because, in context, there is some red
>box that the speaker and listener both know is being referred to specifically.
>"I need a red box" (in its usual interpretation) is opaque, because "a red
>box" doesn't refer to any particular box, really it describes the kind of
>need I have. We've been using "any" as shorthand for this situation, but
>that's not perfectly correct: consider this spooky Hallowe'en dialog:
Hey, I think I'm starting to understand this opaque/transparent thing now!
But it sounds (as someone already said) that "I need a red box" therefore
means "I need the *existence* of red box (possibly of other qualities), and
probably that said box be accessible by me." So isn't that something like
"mi nitcu lei nu [da] xunre tanxe"?
~mark
From: RRICCI@VAXRMA.INFN.IT
Subject: Re: any, opaque, transparent, xe'e...
I was reluctant to intervene in the discussion about "any" which has been
monopolizing the mailing list in the last weeks, since I haven't had the
possibility to read in due depth all the postings. So my opinion could well
be a mere rephrasing of ideas already stated by somebody else or, more
probably, simply miss the point. In this case, just ignore this posting.
It seems to me that the main difference between
"I need a box" (or "Give me a box")
and
I need any box (or "Give me any box")
is simply a subjective or attitudinal one: in other words, I'm not referring
to different domains of discourse, nor am I asserting anything about the
existence or non-existence of the referent of my statement or request - I
could equally well say that I need (or want to be shown) "a unicorn" or
"any unicorn", the same "ambiguity" resulting in this case too. I feel that the
real peculiarity of using "any" instead of "a" relies on a different
attitude towards the thing spoken about: using "any" - in most
"non pathological" cases at least - simply conveys the idea that I am
explicitly stating my *indifference* (for ideological or practical reasons)
on what specific item should be substituted for the thing I'm referring to.
So my modest suggestion for rendering this kind of "any" in Lojban
is to continue using normal descriptors ("le" or "lo"
depending on speaker's intention to be or not to be veridical) together with
some sort of attitude cmavo denoting indifference or "suspension of judge"
(I don't know whether it's already at hand, or if "xe'e" is intended to
mean just this).
Does it sound too naive to your sophisticated philosophically-inclined ears? :-)
---Roberto Ricci co'o mi'e rob.
From: Chris Bogart <cbogart@quetzal.com>
Subject: "any" FAQ needed?
> So my opinion could well
>be a mere rephrasing of ideas already stated by somebody else or, more
>probably, simply miss the point.
It is a lot to read through. What if we created a FAQ that described our
collective thinking on the question, and our discussion could then focus on
changing details of the document, until we reach consensus on it. Someone
could post it every few weeks with up-to-date changes. That would make it a
little easier to focus, and allow other people to participate. John
Clifford's definitions would be a good start, along with some less-technical
amplification.
What we put together here might be a good resource when the logic paper is
written for the reference grammar. I realize that sounds like a lot of work
for a topic we would all probably like to see resolved once and for all, but
the question will come up again and again if new people don't have a single
cohesive place they can read our reasoning and conclusions. Imagine the
chore of going back and reading through all the old postings on any,
opaqueness, etc!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Chris Bogart
cbogart@quetzal.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: "John E. Clifford" <pcliffje@CRL.COM>
Subject: some definitions
la xorxes cusku di'e
> la pycy cusku di'e
But I like "pycyn" I've been away but don't names still have to end in
consonants?
> > OPAQUE (referentially opaque context) Terms which occur in
> >opaque contexts cannot be exported to surrounding transpar
> >ent contexts, cannot be generalized by nor used to instan
> >tiate external quantifiers, does not replace under external
> >identities, if bound need not instantiate to external
> > objects. (I admit that this collapses several notions
> > into one,> but most logicians hold that they go together.)
> This seems to be the way I was using "opaque". The key point
> is that an opaque sumti does not instantiate to external
> objects, while I think {lo broda} in {lo broda cu brode}
> should always instantiate, no matter what {brode} is. That
> is a difference from English, where "a box" is transparent
> in "I see a box" but opaque in "I need a box". I don't
> think in Lojban the semantics of the selbri should influence
> the opaqueness of its sumti. {mi nitcu lo tanxe} should
> behave like {mi viska lo tanxe} as far as the instantiation
> properties of {lo tanxe} is concerned.
I can't understand the expression "opaque sumti" except as "sumti in an
opaque context" (see the explanation). The _lo_tanxe_ works the same in
both cases, but the context is different, the real word with _viska_, some
intensional (and probably intentional) one with _nitcu_. It is not the
term that changes but where it is. And we cannot do away with that, since
we have to have words like "need" that take event description arguments.
We could do away with subject raising, perhaps, but at the cost of a lot
of wasted words (do we really care what he needs the box for?). But then,
I probably do not understand what "instantiate" means here. In the logic
racket, instantiation is the move from a quantified sentence to the result
of replacing all occurrences of the variable bound by the dominant
quantifier by occurrences of some one singular term, going from "all x Fx"
to "Fa" (or from "Some x Fx") Now the "a box" in "I need a box" surely is
a quantifier expression and the one in "I see a box" may be, but what they
can instantiate to need not be the same, since they are in two different
contexts. Each instantiates perfectly well within its context, however.
But the Lojban does not work, since _lo_tanxe_ is, by its form, not a
quantifier expression at all, but a singular term -- a candidate for
replacing, not for being replaced in instantiation. The two _lo_tanxe_
might refer to different boxes, since they are in different contexts. The
problem with opacity seems to be just that we do not notice it and so make
wrong inferences and the solution is to mark it so that we won't. If the
theory I have been peddling works out, all cases will be cases of event
descriptions and the troublesome cases will be just subject raising from
these, so the subject-raising tag will be warning enough.
> > INTENTIONAL. Having to do with what an agent intends:
> >purpose, goal, motive, etc. This somewhat elastic, since
> >objective needs are sometimes included along with subjec
> >tive wants. These are always intensional.
> Intentional has also been used to describe a property of
> {le}, which does not have to do with agents, but with the
> speaker of the utterance. This intentionality of {le} does
> not make the utterance intensional, it only serves to speci
> fy the referent of the sumti.
Speaking is not an act nor a speaker an agent? To be sure, "intentional"
is not being used in the above sense here, a sense which applies to
contexts again. _le_ displays some aspect of the speaker's intention but
does not describe it.
> >In the same discussion -- and maybe more critically --
> >could someone explain "specific" and "definite". These two
> >words are used interchangeably by some groups and by others
> >to make any number of distinctions (often one group uses
> >one to make the distinction another makes with the other),
> > what exactly is going on here?
> I use "non-specific" to mean that the question "which
> one(s)?" has not been in principle answered, but the ques
> tion is relevant.
> So, in {mi klama lo zarci}, {lo zarci} is non-specific,
> because I only said that I go to "a market", no information
> was given to specify which one, but the question "which
> one?" must have an answer, i.e. there _is_ a market such
> that I go to it.
> The speaker need not know the answer to "which one?" for the
> claim to be true, but there must be one.
> In {mi klama le zarci}, {le zarci} is specific. The relevant
> question if "the market" cannot be identified by the audi
> ence is not "which market?", that has already been answered
> as "the market", the relevant question is "what do you mean
> by 'the market'?", or "What is the market?".
> In English, the dialogue: "I'm going to the market." "Which
> market?" does make sense, but only because the assumption by
> the speaker that the audience understands the referent of
> "the market" fails, not because the speaker left it non-
> specific, as in "I'm going to a market", in which case the
> audience has not been told the referent of "a market".
> That seems suitably muddled.
To me too :).
> > The discussion is not helped by the fact that some of the
> >examples of apparently non-controversial cases seem just
> >wrong: _lo_brida_ keeps appearing as a general term rather
> >than a singular one, if translation and inferences are any
> >guide.
> I don't understand that. {lo broda} means (I hope) "at least
> one of all things that broda". What would be general/singu
> lar in this case? (Those weren't in the list of definitions)
Sorry about leaving them out; it is hard to remember what parts of jargon
are common coin. Anyhow, your specific/non-specific distinction sounds a
lot like the general/singular one: quantifiers are inherently general or
non-specific, the question of "Which one" has not been answered in
principle; names and de- scriptions, singular expressions (though in
Lojban the can refer to more than one thing), are inherently specific: the
question has been settled in princi- ple, even though we -- speaker and
hearer -- may not know the answer.
But then _lo_broda_ clearly lies on the specific side. It looks like a
de- scription and that counts for something in a logical language. It is
historically just a correction to _le_broda_ to emphasize the correctness
of the description (as _le_broda_ and the later "dthat" of logic were
corrections to the Russell or Frege description that did not account for
context). Both have the inferences of a name, passing through negations
and the like unchanged (indeed, what would be the dual of _lo_, related to
it as "all" is to "some"?). And, of course, _lo_ binds no variable, even
implicitly, as a quantifier does.
I suspect the problem is that someone once said that _lo_ is like English
"a," as if that pleiomorphic (nice word, thanx djer - wasn't it) kin of
"any" meant just one (or one dozen or...) things. Often "a" is a
quantifier, as in "A whale is a mammal" or "Take a teaspoon of flour and
blend into the melted butter". But often it is a descriptor, forming (for
want of a better word) indefinite descriptions: "A man came into a bar..."
(or -- but keep the exis- tential import out of it a moment -- "A unicorn
came into a barn..."). In that use, the continuation of the tale always
picks up the "a" description with "the". It is that "a/the" use of "a"
that Lojban _lo_ was meant to get: specific (the identity is fixed before
-- or at least as -- the description is uttered, supralapsarian election
as we used to say at Kirk) so like _le_, but the identity not necessarily
manifested, unlike the definite _le_, where the first user at least must
know the referent, even if he doesn't describe it too well. Quantifiers,
by contrast, are infralapsarian, they do not really refer at all and so
what they are talking about is found established only when it is
discovered, after the utterance, in an act of instantiation (Well, this is
one story -- the one Lojban bought. It descends from Hilbert via Hermes.
The other story, mainly from Karttunen (veion waves the flag!), is that
"a" is always a quantifier and that the "the" (our _lo_ again, mind you)
is the instantiation of it to continue the story. Note that _lo_ still is
determined before uttered. Karttunen's story solves djer's problem of
wanting a particular quantifier whose scope is the antecedent of the
conditional but which governs a term in the term in the consequent: "I
some boy comes, all the girls will dance with him" is just "If some boy
comes, all the girls will dance with the boy (who came). We do not
actually have to choose between these stories, since Hermes showed that
all the standard logic of quantifiers could be reconstructed with the
indefinite descriptor. His demonstration is not quite a word for word or
even sentence for sentence translation and it does make use of the "most
likely F" theorem -- there is an x such that if something is an F, then x
is an F -- but it serves our purpose and we can work either way.)
> I thought truth values were atemporal, so it wouldn't make
> sense that some claim is true now but not true in the future.
> Effectively, this means that F and not must commute, which means
> that F can be neither "at some time in the future" (existential
> quantification) nor "at every time in the future" (universal
> quantification). It rather has to be "at the one time in the
> future that I'm talking about", which commutes with "not".
> If the tenses are quantified with anything but unitary (is
> that the right word?) quantification, then we will have
> ambiguities with sentences with tense.
You can't have atemporal truth values and tenses in the same system, since
it is the nature of a tensed sentence to be true sometimes - when in the
appropriate relation to the event described, and false otherwise. "It
will rain" is true, if at all, only up until the raining starts, then "It
is raining" comes true -- as it was not before. It, in turns, ceases to
be true when "It has rained" and "It rained" come into their truth (well,
"It has rained" can overlap "It is raining"). We can occasionally manage
to construct a sentence that is about a particular event and is always
true, but that takes specifying the event in all its gory details as far
as date and time and place and maybe metaphysical or storical context goes
-- all the things we can fiddle with in the tense slot. This won't get us
into any (unsolvable) problems if we have a fixed notion of where pieces
go. I used to be sure that in Lojban the implicit order of "tenses" was
modal temporal spatial negation. Now I am not sure, nor do I remember
what the device is for shifting them -- _ku_ maybe. In any case, it all
takes place at prenex level and so the shifts in the heart of the sentence
are not affected. (We can -- and Karttunen did -- deal with the
quantifier version of tense with the appropriate variant of the "a-the"
shift.) In short, we may not be sure what we said, but it is not ambiguous
(specific even if not definite).
pc>|83 Trying to brake etc.
From: "Mark E. Shoulson" <shoulson@CS.COLUMBIA.EDU>
Subject: Re: some definitions
In-Reply-To: <199411012023.PAA24719@cs.columbia.edu>
(pcliffje%CRL.COM@CUVMB.CC.COLUMBIA.EDU)
>Date: Tue, 1 Nov 1994 12:23:20 -0800
>From: "John E. Clifford" <pcliffje%CRL.COM@CUVMB.CC.COLUMBIA.EDU>
>X-To: lojban list <lojban@cuvmb.cc.columbia.edu>
>Content-Type: text
>Content-Length: 12011
>la xorxes cusku di'e
> > la pycy cusku di'e
>But I like "pycyn" I've been away but don't names still have to end in
>consonants?
Um, let me see if I still have it right... Wouldn't "la pycy. cusku" mean
"that-named 'expressor' restrictively associated with the referent of the
pronoun "P.C.""? No... that would have to be py.cy. But otherwise that's
right? Just like "le ko'a tanxe", replacing "le" with "la" (allowed by
grammar) and "ko'a" with "py.cy." (both pro-sumti)... right?
Just checking..
~mark
From: Goran Topic <topic@math.hr>
Subject: Re: preti zo goi .e zo nai
> la goran cusku di'e
>
> > xu gendra je xamgu fa lu la broda goi ko'a li'u ji
> > lu la broda ku goi ko'a li'u ji
> > lu la goi ko'a broda li'u ji
> > le na se cusku be vi dei
>
> (to e'u ko na pilno ne'i le pa preti zo xu e zo ji toi)
.i nu'e baze'e na go'i vau zo'o
> i pe'i le ci selsku cu smuni le pa da
> i ju'o le ci da gendra je xamgu
>
> i zo'o mi na djuno le du'u xukau le na se cusku be di'u cu gendra je xamgu
> i ki'a le na se cusku
.i xu se curmi fa lenu pilno zo na vi le skiselsku
.i go'i .inaja gliban cusku zoi gy. something I didn't say here .gy.
> > .i le pamoi ka'e se jijnu .ijeku'i pe'i zo goi lasna fi zo broda .enai
> > lu la broda li'u
>
> i ko na krici le du'u la parser cusku makau
> i zo goi roroi lasna lo sumti lo sumti
.i go'i .i mi ruble morji ledu'u zo cei zasti .i le preti cu se spuda
> >.i le remai drani gi'enaiku'i xamgu ca'i la zipf.
>
> i ie lei famyma'u cu tolmelbi
.i famyma'u ki'a
> > .i lenu ji'a le zo bai selma'o zo nai lidne cu cumki xu
>
> i go'i
>
> > .i mupli fa lo'u cu'unai ko'a le'u
> > .i ri gerna nagi'a se smuni
> > lu ko'a na cusku da li'u ji
> > lu na'e ko'a cusku da li'u ji
> > lu ko'a na'e cusku da li'u ji
> > lu ko'a cusku na'e da li'u
> > .i to le re fanmo cu na'e lakne toi
>
> i pe'i lu cu'unai ko'a li'u se smuni lu da naku se cusku ko'a li'u
> i ri frica lu ko'a na cusku da li'u
>
> (to lo'u na'e ko'a le'u na gendra i lu na'e bo ko'a li'u toi)
.i ki'anai
> je'e co'o mi'e xorxes
co'o mi'e. goran.
--
Learn languages! The more langs you know, the more incomprehensible you can get
e'udoCILreleiBANgu.izo'ozo'onairoBANguteDJUnobedocubanRI'a.ailekadonaka'eSELjmi
From: jorge@PHYAST.PITT.EDU
Subject: Re: some definitions
la pycyn cusku di'e
> But the Lojban does not work, since _lo_tanxe_ is, by its form, not a
> quantifier expression at all, but a singular term -- a candidate for
> replacing, not for being replaced in instantiation.
Here is the crux of our disagreement!
If I understand correctly your terms, I always thought {lo tanxe} was
a quantifier expression, not at all a singular term. Otherwise, its
default quantifier {su'o} = "at least one", makes no sense.
I don't see why we would want to have both {le} and {lo} as singular
terms, and no simple quantifier expression.
> Anyhow, your specific/non-specific distinction sounds a
> lot like the general/singular one: quantifiers are inherently general or
> non-specific, the question of "Which one" has not been answered in
> principle; names and descriptions, singular expressions (though in
> Lojban the can refer to more than one thing), are inherently specific: the
> question has been settled in principle, even though we -- speaker and
> hearer -- may not know the answer.
Great, at least we now agree on the terms.
> But then _lo_broda_ clearly lies on the specific side.
How do you explain its default quantifier, in that case?
> It is
> historically just a correction to _le_broda_ to emphasize the correctness
> of the description ().
> Both have the inferences of a name, passing through negations
> and the like unchanged
But {lo} doesn't pass unchanged through negations!
> (indeed, what would be the dual of _lo_, related to
> it as "all" is to "some"?).
{naku lo broda} goes to {ro broda naku}. That is what the negation paper says.
{le broda} doesn't in general pass unchanged either, because its quantifier
is {ro}, but in 99.99% of the cases it is singular, {ro le pa broda}, and
so passes unchanged, because {su'o le pa broda}={ro le pa broda}.
> You can't have atemporal truth values and tenses in the same system, since
> it is the nature of a tensed sentence to be true sometimes - when in the
> appropriate relation to the event described, and false otherwise.
That's a matter of definitions. I say truth values are atemporal, but
utterances are not. You are saying that the utterance is atemporal, but
its truth value isn't.
> "It
> will rain" is true, if at all, only up until the raining starts, then "It
> is raining" comes true -- as it was not before.
Or, in my language: "It will rain" is true if at the future of the time
of the utterance it rains. I prefer to think that it is the utterance
that is associated with a time. The same words "It will rain" constitute
a new utterance if said again when it starts to rain.
> This won't get us
> into any (unsolvable) problems if we have a fixed notion of where pieces
> go. I used to be sure that in Lojban the implicit order of "tenses" was
> modal temporal spatial negation. Now I am not sure, nor do I remember
> what the device is for shifting them -- _ku_ maybe.
Yes, it can be done with {ku}, but not when they are directly attached to
the selbri. But are tenses singular terms, or quantified? I thought that
by themselves they were singular terms.
So, what does {mi na ba klama} mean?
1- It is false that at some future time I will go.
2- At some time in the future, it will be false that I go.
3- (My candidate) It is false that at the time in the future that
I'm talking about I will go.
4- (=3) At the time in the future that I'm talking about, it will
be false that I go.
3 and 4 correspond to taking tenses as singular, which seems the best
choice. In that case, they commute with negation. If they don't commute
with negation, we need some convention to interpret their order.
Jorge
From: bob@GNU.AI.MIT.EDU
Subject: context in Lojban
This opaque/transparent discussion often contains confusing cases.
For example, some people consistently refer to
mi nitcu lo tanxe
as
I need a box
suggesting that the box is unspecified or opaque. Some of the Lojban
introductory materials encourage `a' for {lo}. However, by default,
this translation is wrong for Lojban, although the interpretation is
correct in Logic and English.
Lojban is a dialog, not monologue, based language, as Lojbab pointed
out many years ago. Context is always understood. If context is not
understood, then someone should say {ki'a}.
mi nitcu lo tanxe
best translates as
I require that which is really a box in the context understood by
you and me (and whoever else is part of this conversation).
Suppose you and I have the top of a box and a real box in front of us;
the top of the box can look like a real box if you look at it from the
top, but is only a top. In this context, {lo tanxe} is transparent,
is specific, and is *more* specific than {le tanxe}.
It is a bad habit to use `a' for {lo} and `the' for {le}. When
contexts are known, {lo} is often, perhaps mostly more specific than
{le}.
Please express examples with appropriate context. Unfortunately,
the Santa and the Elf example of a few days ago did not tell what
would have been evident to the conversationalists, namely the number
and reality of the various boxes and things that might be designated
boxes in front of the conversationalists. The Santa and Elf case made
sense to Logicians and English speakers because people in these
languages expect low context monologues; but the situation is very
unlike what Lojban is supposed to be.
Robert J. Chassell bob@gnu.ai.mit.edu
25 Rattlesnake Mountain Road bob@grackle.stockbridge.ma.us
Stockbridge, MA 01262-0693 USA (413) 298-4725
From: bob@GNU.AI.MIT.EDU
Subject: Re: singular and plural
In-Reply-To: <9411020530.AA01042@albert.gnu.ai.mit.edu> (jorge@phyast.pitt.edu)
jorge@phyast.pitt.edu cuska di'e
I don't see why we would want to have both {le} and {lo} as singular
terms, and no simple quantifier expression. ...
Because Lojban does *not* consider singular and plural to be primary.
In English and in most people's interpretations of logical
expressions, number is primary. Lojban is different.
The primary distinction in Lojban is between `that which I designate
in my head' and `that which is real according to the epistimology and
context of the conversation'.
lo tanxe
is at least one box because it is real and things which do not exist
are not real, at least not in the context of this and most other
Lojban list dialogs.
le tanxe
is likely to be (but is not necessarily) at least one something
because I tend not to designate a thing as a box unless the thing
designated is real, although that thing may not be a box, again in the
epistemology and context of the dialog.
Robert J. Chassell bob@gnu.ai.mit.edu
25 Rattlesnake Mountain Road bob@grackle.stockbridge.ma.us
Stockbridge, MA 01262-0693 USA (413) 298-4725