[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lojban] Re: A (rather long) discussion of {all}

On 5/20/06, Maxim Katcharov <maxim.katcharov@gmail.com> wrote:


Context is anything that could aid you in determining what was meant
by an ambiguous description, so yes. It's just that "me" (I this one
speaker) is not ambiguous, and same for a word with a sticky tense
applied - it's not ambiguous.


There is a difference between ambiguity and vagueness. Lojban is very
good at avoiding ambiguity. Its syntax is unambiguous and it has no
homonyms, for example. Vagueness is an intrinsic property of
language, you can have more or less of it, but you can't eliminate it,
and too little vagueness is as bad as too much of it.

"Me" is not ambiguous, I agree. But it is not devoid of vagueness, you
still need context to undestand what it refers to. For example, in
"Look at these pictures, this is me in 1987 and this is me in 1995",
the referent of "me" changes from one use to the next.


I would not support having {L_ ro cribe} definitely mean {L_ ro cribe
poi nenri la ...}, derived solely through context.


Neither would I. Context can never give such definite answers, it can
only suggest, with greater or lesser force, that that is what is meant.


> {ro nai} is currently ungrammatical, though I think {nai} should be allowed
> after any word.

I think I agree with you, since it sounds a bit odd. Is {rome'i lo ro
cribe cu zasti} proper then?


{me'i} = {me'iro}, "less than all", but {rome'i} = {rome'iro}, "all,
less than all"
is either self-contradictory or the first {ro} is redundant, depending on how
you treat two adjacent quantifiers. (The wiki page for the BPFK section
on quantifiers has some proposal about that.)


> > >      mi pu viska lo pixra skina i lo ro cribe pe ri cu dasni lo mapku
> > >      "I saw a cartoon. All bears in it wore hats."
>
>       mi pu klama lo dalpanka i lo ro cribe pe ri cu dasni lo mapku
>       "I went to a zoo. All bears in it wore hats."
>
Yes, zasti-typeness is figured out from context. {pe} is, after all,
highly context-based, and so are, usually, ti/ta/tu and friends.

´
{pe ri} = {poi ke'a srana ri}. It is not more context-based than other
restrictive clauses. The point here is that depending on what the
restriction is, the unrestricted set is changed. In one case it
includes non-existent bears and in the other it doesn't.

We

have ways to say "in India", and "in the year 1970", but let's say
that we have a work of fiction based in India, 1970, and we want to
assert things regarding that story. Like "Alice came to India in June"
- that's surely not true, because this Alice is fictional/imaginary,
and so never came to India. But in the book, she did! (Just like the
bears never wore any hats, but in the cartoon, they did.) And we
surely don't mean that imaginary-Alice came to India? Though perhaps
we mean imaginary Alice came to imaginary India. And imagined by who?
What if I wrote of two different imaginary Indias? I'd have to say
which of these perceptual universes I mean. And if I share this story
with you, don't you partake in this other perceptual universe with me?
So really the imaginer doesn't matter. But when writing a story we'd
have to say something like "Alice (an imaginary person) was on the
(imaginary) ship as it approached the (imaginary) dock." Understand
that I don't mean {xanri bloti}, I mean {bloti poi xanri la [Alice In
India]} (the boat imagined by the book titled). Anyway, we have to do
this, or we aren't being very logical, are we? Not that we have to be
logical if we don't want to, even in Lojban. I'm just presenting it as
something to think about.


That's why context is so important. When we are reading a fictional work,
we know what the context is, we don't need to add "who is a fictional
character" every time we mention Alice, and indeed we shouldn't, because
within the context of the story she isn't.


> > > (2) A semi-formal situation where for example an instructor with a prepared
> > > speech explains the agreement to a group of people. They may start their
> > > talk by saying: "In what follows, when I say 'bear' I will mean the grizzly
> > > bears at GSA" and so on.
>
> (2a) Instead of saying "In what follows, when I say 'bear'..." the instructor
> opens with "I am going to speak to you about the grizzlies at GSA" or
> something of that sort, so he makes no explicit reservation of the word "bear".
> Would that change anything for the later interpretation of {lo ro cribe
> poi se catra lo prenu}, which may come perhaps twenty minutes later?
> If yes, does that mean that one is forced to keep in mind not just the sense
> of what the instructor is saying, which is basically the same in both cases,
> but also the exact wording used?

"I am going to speak to you about the grizzlies at GSA" implies "I am
going to speak to you about *only* the grizzlies at GSA", which seems
to be the same as "when I speak about bears, I mean only bears in
GSA".


Not logically, no. That I am going to talk about X does not preclude my
talking about Y too. It does suggest that when I say "all bears killed by
people" it will mean the grizzlies at GSA, but if I were to say "the bears
at GSA are not like bears in other places", then "bears in other places"
is obviously not restricted to bears at GSA.


 I don't really see a difference, it's like me saying "I define
bears as bears that are in GSA". It's not context, it's that you're
editing the language for the purposes of the contract.


Remember that in this case this is a talk, not a written contract where you
can go back and check how things were defined. You can't expect people
to keep an exact record in their memories of every word said. Just the
general sense of what has been said before can be used to interpret
later utterances.


> (2b) One of the attendees at the conference arrives late so they miss the
> instructor's warning that they will use 'bear' to reffer to grizzlies at GSA.
> When the instructor mentions that "all bears killed by people must be
> accounted for under the quota", however, the late arrivee has no trouble
> figuring out that it doesn't include bears that were killed by people two
> hundred years ago in another continent. I predict that the same would happen

Yes, the speaker can infer certain things, like that there's a mistake
somewhere, from context anyway, even with my{ro}. The late person can
ask what the instructor means by "all" - does he mean the entire
forest? and get an answer - no, just the GSA. I can assume that you
made a typo, and that you really mean ninmu and not nanmu. This
doesn't mean that it's ok for the speaker to exclude that statement at
the start of the lecture. He wouldn't be speaking the language
correctly if he did.

So looking at this from the perspective of the listener is probably
not the right way to do it. After all, the speaker knows what bears
he's referring to or making assertions about, and he knows that he
used the language to properly describe them, and that the language fit
his thoughts.


OK, I guess we'll just have to disagree about that. I don't think Lojban
will differ from other languages in this regard.


> in English or in Lojban. Would you rather have a fluent Lojban speaker be
> confused in that situation? (This situation is basically the same I faced
> when reading this sentence for the first time, I only saw the single sentence
> without knowing that it belonged to a formal agreement with definitions
> at the beginning, and yet I understood immediately that they were not
> talking about all bears ever. The very sense of the sentence itself made
> that quite clear.)

This to me shows that humans are capable of recognizing and working
around mistakes in communication. This seems to be a way of saying
"well, it's ok if we don't handle this concept of language, because
the listener will in most cases get our point anyway".


Only if you think it is a mistake. If I make a typo, and you can still figure
out what I mean, but when you point it out to me I agree that I made a typo,
that was a mistake. If I use inner {ro} in a way that does not refer
to everything
you expect it to refer (which I'm still not completely sure what it is), I may
accept that I was not precise enough for you, but not that I misused
the language. I will try to reword what I said, so that communication is
more successful, but I won't say "you're right, I said something I didn't mean,
what I really meant was...". Being vague, when precision is not called for,
is not the same as being sloppy in the use of the language. The degree
of vagueness that is acceptable and even necessary in any given situation
is a matter of pragmatics, and I don't have a full theory of pragmatics to
make explicit rules about when for example {lo ro broda} admits imaginary
brodas and when it doesn't. I just know that sometimes it does and some
times it doesn't.



> > > (3) An informal situation: "You're new here, right? OK, then you have to know
> > > this: all bears killed by people must be accounted for, no exceptions are
> > > admitted".
> >
> > In this case I wouldn't use {ro}/"all": "...know this: bears killed by
> > people...", or I'd probably say something like "all bears killed here
> > must be accounted for", to which they may reply "disambiguate 'here'",
> > to which I'd reply "lo Wasomething Settlement".
>
> But the "all" is important here, as emphasized by the "no exceptions are
> admitted". Indeed the original had some other type of emphasis, something

My blank inner assumes the most contextually sensible number of bears.
In this case, the most sensible referent is "all bears in this place
here" ("in this place here" being deduced from the preceding "you're
new here, right?"). It is informal speech, after all. If he wanted to
be a bit more precise, he'd speak less casually and provide the full
restriction.


What if he wanted to emphasize the {ro} but still remain informal. Something
like "*All* bears killed by people must be accounted for". How do you
emphasize a word that must be understood but cannot be pronounced?


My May 19th message asked:

Show me how to say "nothing" (described above) via your method in a
way that isn't convoluted.

I would like you to address that. If you're still thinking, or have a
way, please let me know. If you have no way to do it, then please
consider the implications that that has on our discussion.


As I said, I would use {no da} for almost any plain "nothing". I say
"almost" because maybe you can come up with a context where
something else might be better, but any nonconvoluted "nothing"
should be very well translated with {no da}. Rather than ask for a
translation of a single word, it is better to put it in a context. That
often makes translation easier.

mu'o mi'e xorxes