Re: do djica loi ckafi je'i tcati

[Logical Language Group <lojbab>]

(Administrivia: I'm back, suffering the usual Monday morning blues, but
otherwise all right.)

la kris. bogart. cusku di'e

> I was thinking more about the discussion of whether the answer ".a" to the
> question:
> 
> 1)      "do djica loi ckafi ji loi tcati"
> 
> is evasive or not.  When the question is answered ".a" that means:
> 
> 2)      "mi djica loi ckafi .a loi tcati"
> 
> which expands to:
> 
> 3)      "mi djica loi ckafi .ija mi djica loi tcati"
> 
> Now suppose I want tea, and I don't want coffee; and suppose I answer ".a"
> to the question. I would be telling the truth, because "mi djica loi ckafi"
> would be false, "mi djica loi tcati" would be true, and the ".ija" of the
> two statements would therefore be true.  So saying ".a" doesn't really tell
> the questioner what I want to drink!

This analysis, originally proposed by Jorge, is perfectly correct, and I
will remove the offending examples from the next version of my paper on
the subject.  (BTW, I have numbered all your examples for ease of reference.)

> But instead of adding new connectives or something, I think the real problem
> is that the question was misphrased.  The correct translation of "would you
> like coffee or tea?" should be:
> 
> 4)      "do djica loi ckafi je'i tcati ku"
> 
> ("je'i" being the tanru connective question-word) -- and the response could
> be "ja" or better yet "jonai" to mean "either"; or "jenai" to mean you want
> tea, or "na'i" to mean you're not thirsty.
> 
> I suggest using a tanru connective not because the semantics of the tanru
> connectives are hazier, but because it allows you to put the connective
> within the scope of "loi", which (I think) is a barrier preventing
> conversion of the sentence (answered by "ja") into two sentences connected
> by ".ija".  So if I don't want coffee I can't get away with evasively saying
> "ja", since the possibility of
> 
>         "mi djica loi tcati" being true, and
>         "mi djica loi ckafi" being false,
> 
> is not really implied or even addressed by the statement:
> 
> 5)      "mi djica loi ckafi ja tcati".

You are correct in your conclusions, but not (I think) in your reasoning.
It is indeed the special property of tanru connectives, namely that they
don't split into full sentences, that makes this solution usable.
For example:

5a)	ti nanmu ja ninmu vimkumfa
	This is-a-(man or woman)-type-of excretion-room.

does not split into:

5b)	ti nanmu vimkumfa .ija ti ninmu vimkumfa
	This is-a-man excretion room, or this is-a-woman excretion-room.

because the "ti" of Example 5a refers to a unisex toilet/bathroom, which
is precisely neither a "men's room" nor a "women's room".

> Alternately you could find some way of abstracting the connective without
> resorting to tanru connection, like:
> 
> 6)      "do djica lenu mi dunda do loi ckafi ji loi tcati".
> 
> I think ".a" here as an answer would generalize to:
> 
> 7)      "mi djica lenu do dunda mi loi ckafi
>                      gi'a dunda mi loi tcati kei"
> 
> but *not* to:
> 
> 8)      "mi djica lenu do dunda mi loi ckafi
>     .ija mi djica lenu do dunda mi loi tcati".

Correct, and this is because of a general principle that a logical
connective splits only the bridi it is in, not any superordinate bridi.
This works in English too:

8a)	I believe that there is life on Jupiter or there isn't.

is true, since the content of the belief is a tautology, and I believe all
tautologies provided I am aware of them; but

8b)	I believe that there is life on Jupiter.

is false, and so is

8c)	I believe that there isn't life on Jupiter.

since I have no evidence one way or the other.

> 9)      "do djica tu'a loi ckafi ji loi tcati lu'u"
> 
> (can "tu'a" enclose two connected sumti like that?  Or is there an implied
> "lu'u" before "ji"?)

This, I think (and so does Jorge, I believe) is the right general solution.
"tu'a" (and members of LAhE generally) are long scope, and include the whole
logically connected sumti.  (Historically, LAhE was short scope and LUhI
was long scope: when they were merged, part of the reason was to allow LAhE
to have long scope when needed.)

-- 
John Cowan		sharing account <lojbab@access.digex.net> for now
		e'osai ko sarji la lojban.