TECH: Narrative connectives?
Veijo Vilva
Logically connected sentences
(1) mi nelci ti .ije mi nelci ta
mi citka .ije mi pinxe
can be conveniently shortened to
(2) mi nelci ti .e ta
mi citka je pinxe
However, there is no way to shorten a straigth narrative
(3) mi nelci ti .i mi nelci ta
mi citka .i mi pinxe
in a similar way. Sometimes it would be quite natural and
convenient to be able to say, e.g.
(4) *mi nelci ti gi ta
I like this ... and that
*mi citka gi pinxe
I eat ... and drink
Here {gi} is a new type of non-logical connective, a narrative
connective. It builds loose compounds which don't fit well
into the set of existing non-logical compound types.
({i} can be thought of as a narrative sentence connective.)
This structure could then be expanded to handle afterthought
comparisons (as I proposed in an earlier posting)
(5) mi nelci ti .isemaubo mi nelci ta
-> *mi nelci ti gisemaubo ta
(still using the {gi} which Jorge wouldn't like to have :_)
and tense relationships within tanru (not presently allowed)
(6) mi citka .ibabo mi pinxe
-> *mi citka gibabo pinxe
which would be especially useful in certain descriptions
(7) *le sipna gisemaubo cikna
*le zukte giba'obo pensi
Presently there doesn't seem to be any (easy) way to express
these!
co'o mi'e veion
---------------------------------
.i mi du la'o sy. Veijo Vilva sy.
---------------------------------
Veijo Vilva
> Date: Thu, 4 Aug 1994 17:22:00 EDT
> From: Jorge Llambias <jo...@PHYAST.PITT.EDU>
> la veion cusku di'e
>> I was just thinking...
> Always a dangerous activity :)
>> Logically connected sentences
>>
>> (1) mi nelci ti .ije mi nelci ta
>> mi citka .ije mi pinxe
>>
>> can be conveniently shortened to
>>
>> (2) mi nelci ti .e ta
>> mi citka je pinxe
>
> Actually, the second one should be
>
> mi citka gi'e pinxe
>
> because a tanru is not really automatically expandable to two sentences.
>
Right.
> Also, you could have had {mi citka ti .ije mi pinxe ta} which is
> {mi citka ti gi'e pinxe ta}, and you probably, in (1), didn't mean to say
> that you were eating and drinking the same thing.
Again an unfortunate choice of an example sentence :_)
> Your point is just as valid, though.
[...]
>> in a similar way. Sometimes it would be quite natural and
>> convenient to be able to say, e.g.
>>
>> (4) *mi nelci ti gi ta
>> I like this ... and that
>>
>> *mi citka gi pinxe
>> I eat ... and drink
>
> Unfortunately, I think it doesn't work. Consider:
>
> ge mi citka ti gi do pinxe ta
>
> Will the parser understand it as:
>
> ge (mi citka ti) gi (do pinxe ta)
>
> or as:
>
> ge mi citka (ti gi do)
>
> and then find an error with the next word?
Replace {gi} with, say, {xi'e} (I should have done it,
just to be sure :_(
[...]
>> and tense relationships within tanru (not presently allowed)
>>
>> (6) mi citka .ibabo mi pinxe
>> -> *mi citka gibabo pinxe
>
> I think this was permitted under Mad Proposal. It has to be bridi-tail
> connection though, not tanru connection.
Bridi-tails have been there... and checking from the selbri-
level of the grammar it seems tanru with mixed tense connection
( mi citka jebabo pinxe ) are allowed - so it is just a question
of connective type :_)
> Jorge
So the question boils down to: Is there a different enough connection
(from a logical AND and the various non-logical connections) involved
to warrant one or more additional cmavo? In the case of comparisons
there is a difference between
(a) mi nelci ti .esemaubo ta
and
(b) *mi nelci ti xi'esemaubo ta
in that (a) makes clearly all the 3 claims involved but (b) could
omit {mi nelci ta} (or even {mi nelci ti} ?). Of course, the same
argument could apply to
(c) mi zukte jeba'obo pensi
and
(d) *mi zukte xi'eba'obo pensi
i.e. in (c) I DO think afterwards but (d) could leave that open.
But then, this would differ from the straight narrative
(e) mi zukte .iba'obo mi pensi
which seems to make all the 3 claims. Eh? But how does
(f) mi zukte .ijeba'obo mi pensi
differ from (e) ? And could there be a difference between
(g) mi nelci ti .e ta
le zukte jeba'obo pensi
and (using {xi'i} for a narrative connective)
(h) *mi nelci ti xi'i ta
*le zukte xi'iba'obo pensi
Might we actually need TWO sets of new connectives, afterthought
mixed connectives to express just the tense/modal relationship and
narrative connectives along the lines of my original posting
(correcting the details :_) ?
The afterthought case can be handled in a rather non-Zipfian way,
of course:
(j) mi nelci ti .emaugi ti gi ta
(but that's even worse than (b) :__(
Logical Language Group
> Logically connected sentences
>
> (1) mi nelci ti .ije mi nelci ta
> mi citka .ije mi pinxe
>
> can be conveniently shortened to
>
> (2) mi nelci ti .e ta
> mi citka je pinxe
>
> However, there is no way to shorten a straigth narrative
>
> (3) mi nelci ti .i mi nelci ta
> mi citka .i mi pinxe
Well, actually there are ways, using different mechanisms. (I know you
know about these, but not everybody on this list will.)
3a) mi nelci ti .i go'i ti
mi citka .i pinxe
where the latter case elides x1 but is not a classic "observative".
> in a similar way. Sometimes it would be quite natural and
> convenient to be able to say, e.g.
>
> (4) *mi nelci ti gi ta
> I like this ... and that
>
> *mi citka gi pinxe
> I eat ... and drink
I believe that this proposed change (whatever its syntax) would violate
the general nature of non-logical connectives (which is really what it is,
as you say below: a vague non-logical connective). The essence of a non-
logical connective is that it doesn't transform into multiple sentences:
"mi pinxe loi ckafi ku joi loi tcati" is not the same as
"mi pinxe loi ckafi .ijoi go'i loi tcati", and it is questionable what
the latter might mean at all.
> Here {gi} is a new type of non-logical connective, a narrative
> connective. It builds loose compounds which don't fit well
> into the set of existing non-logical compound types.
> ({i} can be thought of as a narrative sentence connective.)
Your analysis is sound enough.
[material on afterthought comparisons deleted]
--
John Cowan sharing account <loj...@access.digex.net> for now
e'osai ko sarji la lojban.
Jorge Llambias
> I was just thinking...
Always a dangerous activity :)
> Logically connected sentences
>
> (1) mi nelci ti .ije mi nelci ta
> mi citka .ije mi pinxe
>
> can be conveniently shortened to
>
> (2) mi nelci ti .e ta
> mi citka je pinxe
Actually, the second one should be
mi citka gi'e pinxe
because a tanru is not really automatically expandable to two sentences.
Also, you could have had {mi citka ti .ije mi pinxe ta} which is
{mi citka ti gi'e pinxe ta}, and you probably, in (1), didn't mean to say
that you were eating and drinking the same thing.
Your point is just as valid, though.
> However, there is no way to shorten a straigth narrative
>
> (3) mi nelci ti .i mi nelci ta
> mi citka .i mi pinxe
>
> in a similar way. Sometimes it would be quite natural and
> convenient to be able to say, e.g.
>
> (4) *mi nelci ti gi ta
> I like this ... and that
>
> *mi citka gi pinxe
> I eat ... and drink
Unfortunately, I think it doesn't work. Consider:
ge mi citka ti gi do pinxe ta
Will the parser understand it as:
ge (mi citka ti) gi (do pinxe ta)
or as:
ge mi citka (ti gi do)
and then find an error with the next word?
I had thought along these lines at the time of the Mad Proposal. :)
> This structure could then be expanded to handle afterthought
> comparisons (as I proposed in an earlier posting)
>
> (5) mi nelci ti .isemaubo mi nelci ta
> -> *mi nelci ti gisemaubo ta
>
> (still using the {gi} which Jorge wouldn't like to have :_)
I don't mind it, if it can be made to work, but my way is more zipfy :)
(That's always a good argument to use in convincing Lojban Central :)
> and tense relationships within tanru (not presently allowed)
>
> (6) mi citka .ibabo mi pinxe
> -> *mi citka gibabo pinxe
I think this was permitted under Mad Proposal. It has to be bridi-tail
connection though, not tanru connection.
Jorge