Re: place structure of lujvo

[Logical Language Group <cbmvax!uunet!pucc.princeton.edu!lojbab>]

This is my response to Dave's private message to me on this subject:
_______
Re your discussion > with Nick >> on his fable:
>Regarding your fable, I said of this sentence,
>
>> ... le maxri noi se cimri'a ku'o lei manti goi ko'a cu se sudri'a [vau]
>
>>To me, this has said "the wheat was in a dry fashion caused by the ants."

The wheat which was moist-caused, by the ants, was dry-caused.  What the
lujvo mean is subject to negotiation.  Generally, with -ri'a causals
x1 has been a cause and x2 has been a result.

>As I read rinka, x2 is the thing-caused.  If it's the event of wheat
>being dry that was caused, shouldn't it be around in x2, as a sumti?

Thus I agree with you ... almost.

What we really have here is an application of sumti raising.  This is
because rinka talks about events causing other events.  Thus, if Nick
had used tanru instead of lujvo, he would have been wrong not to use
tu'a on both x1 and x2 (The ants doing something caused the wheat to
become dry)

>And I offered "lei manti cu rinka le nu le maxri cu sudga [kei] [vau]"
>or something like that.

tu'a lei manti cu rinka lenu lei maxri cu binxo lo sudga
or
tu'a lei manti cu rinka tu'a lei maxri sudga

It has been commonly accepted that since such things like sumti raising
are such pains, that lujvo can be defined to foreshorten them. lujvo are
NOT as predictable as tanru as to the place structure.  You choose what
is most useful - a lujvo must compress out some structure, and the sumti
raising seems as much a candidate as anything.  I argue this based on
Zipf's Law, not on dikyjvo.

Carter's dikyjvo are NOT part of Lojban in themselves.  What I have been
arguing form is that they CAN be used as a means to predict 'most
useful' interpretations of lujvo, and thus are not wholly useless.

Your version used no lujvo, hence was longer.  Nick did a Zipfean
compression using a lujvo, which incidentally matched the dikyjvo
conventions.

Nick has been very aggressive at using lujvo where the concept he is
trying for seems like a natural lujvo-sized thing.  This is good!  There
will be some confusion, because you don't necessarily know what the Zipf
compression was.  I usually interpret based on context, and was in this
case, able to figure it out, once I realized Nick was using SOV order.

I did this because ants are not events, nor is wheat, in most usages.
Therefore rinka in a final-spot of a lujvo implies that the ants doing
something cause the wheat to do something.

Your interpretation isn't plausible because in real life ants don't
cause wheat (or maybe they do - this was LE maxri - perhaps they
dryingly caused dry wheat).  If I had wanted causation of a thing, I
would use zbasu if made from something else, or finti if made from
nothing, whole cloth.  Thus the caused event is not likely the mere
existence of the wheat because there are better, clearer, ways to say
it.  The next plausible interpretation to consider is the lenu sudga
expansion that you used.

Some other things that Nick might have considered are sudga-galfi and
sudga-stika as lujvo.  Would these have been clearer to you?

Whether Nick is using a system like dikyjvo, or not, is up to him.  The
more regular the rules he uses, the more likely the reader will be able
to predict the result.  This is the advantage of dikyjvo.  The
disadvantage is that dikyjvo put Lojban into a straightjacket too early,
creating systematic rules that might be malglico, that might not
consider all of the options that should be considered, etc., because the
rules are being devised by English speakers who are admittedly still
thinking in English, and wothout participation by the community.

I suggested to jimc, and implicitly to Nick, that they could freely use
dikyjvo as a method for lujvo-making WITHOUT marking it or using it to
justify their lujvo choices, and see whether people 'naturally' figured
out what the lujvo meant.  Nick did, and you did not, thereby showing
that dikyjvo are not as natural to Lojban as jimc thinks.  Since you,
Dave, have demonstrated as much command of Lojban as Jimc has
(well-formed individual sentences - as long as they are simple - Jim has
tried longer and more complex things, but has generally been incorrect
in some way on virtually every sentence therein), and indeed you have
had and read the draft lessons and he has not, your skill/opinion about
what seems 'right' for Lojban certainly has as much weight as his.

Nick has bought my position that dikyjvo are merely good guidelines,
perhaps as crutches, for English speakers who are afraid to make lujvo
more instinctively.

>To which you responded...
>>Ever heard of jimc's dikyjvo?  Quite simply, you already know the x2 of
>>rinka - it's lenu sudga.  The x1 of sudga becomes the x2 of sudri'a.
>>Compare:  xy. citka y'y .i zy. rinka lenu xy. citka zy. i zy ctiri'a xy.
>>zy.  X eats Y. Z causes that X eats Y. Z feeds X Y. (This example is a
>>bit deceitful, as it hides behind the English dative; normally, it'd be
>>feeds Y to X.)  This means of simplifying sumti is elegant, nicely
>>analogous to Esperanto's -ig- suffix. sudga means is-dry; sudri'a is a
>>single word meaning to dry:  x1 dries x2 of x3 (xyxipa rinka lenu xyxire
>>sudga xyxici).  Transitives are very very useful, and it's boring to say
>>{rinka lenu} all day, so I'll defend my usage, which also maps more
>>comfortably with NL factitives.  But any usage you suggest (and have
>>suggested) is perfectly legit.

And this last sentence is where he admits that your interpretation and
preference is valid, though he notes that his lujvo choice is plausible
and matches the pattern of English and Esp-o.

>Lemme see if I have this straight.  For any two gismu (let's keep it to
>gismu for now), symbolically
>
>        g1 G g2...
>
>and
>
>        h1 H h2...
>
>then if you make a lujvo GH, its argument pattern is
>
>        h1 GH g1 ...?
>
>what's the rest of the pattern?  I would presume that g2-gn are
>discarded and it's "h1 GH g1 h2...hn" ?

The default convention (and these are all merely proposed conventions)
is that a lujvo with no contextual aid is

         h1 GH h2 ... hn g1 ... gn

but since this leads to indefinitely long place structures, you try to
find reasons to compress redundant or irrelevant places, and even to
switch some places in order if it makes sense.

h1 causes h2 under conditions h3
g1 is dry of liquid g2

h1 dry-causes h2=g1 under conditions h3 to be dry of liquid g2

is approximately the sequence I use.

But this compression is not systematic:

g1 is a pot containing g2, of material g3
h1 cooks h2 by recipe/method h3 with heat h4

h1 cooks h2=(g1 or g2?) ...

You have to rely on real-world knowledge that you cook contents of pots,
not the pots themselves.

>Is this jimc's proposal?  I have not bothered trying to read his very
>dense postings, so can't tell if you are using dikyjvo or contrasting to
>that scheme.  I can see how this would be very useful for H==rinka and a
>few others, but am dubious about the general case.  At any rate, it
>feels a bit brain-twisting that, when meeting a lujvo, I have to call to
>mind not only the argument list of H but that of G also, or part of it.

Yes it is brain twisting, but whether using dikyjvo or not, this is the
case.

>If your fables ARE using dikyjvo, which if I understand correctly is not
>actually part of Lojban?  I think you should have attached a notice to
>that effect.

That would violate the experiment I suggested to them and would also
lend greater authority/credibility to dikyjvo, which I'd rather see
forgotten or left in low profile for 20 years.  Why?  Because to refer
to dikyjvo everytime they use them, means that the average newcomer will
ask what are dikyjvo, and we get the jimc lecture over and over again
until people give up in disgust, or maybe break through to understanding
and take the system whole cloth.  Using it without comment, people will
try to figure it out knowing the actual English and the translation.  If
the writer has credibility that his Lojban is reasonably good, the use
of a pattern will serve as authority for the validity of that pattern.
Thus people will learn patterns and not rules, and learn them by
analyzing them probably by a much more subsconscious method than any of
the above methods, especially including dikyjvo.  I would contend that
these subconscious methods are those we use in learning any new
language, 1st or 2nd.

At this point however, only Cowan and I seem to have that kind of
authority, and Nick and others in the next skill tier have no qualms
about questioning OUR efforts as well.

We need to get more people trying to use lujvo, making them by whatever
method seems natural, and learning by the misunderstandings that result.
I would much rather, as a result of your reaction, that Nick not see
"ri'a" as the equivalent of an Esp-o transitivizing suffix, but as a
word symbol of its own right that must be freshly analyzed when used in
lujvo each time, until the process is internalized.  The result may be
the same, but the method is the key to handling the more obscure cases
that don't fit Jimc's neat patterns (or fit more than one).

Finally I want to note that Nick is imprecise in the first place by
using rinka merely because the x1 of rinka should be an event - there is
nothing in dikyjvo that tells you that the x1 of the compound is the
tu'a agent of the causing event.  If he wants agentive causation, I
suggest to him using "gasnu" "-gau" which currently has the official
place structure x1 is the agent of event x2 occuring.  The lujvo could
be sudri'agau, but I think this will be seen to naturally collapse to
sudgau if the x1 of rinka is not a fully expressed event in one of the
places.  And if it is a fully expressed event, sudri'a IS correct.  Thus
you use sudri'a for events causing drying, and sudgau for agents causing
drying:

lenu le solri dirce glare le terdi sudri'a lei maxri
The Sun radiatingly-warming the Earth dried the wheat.

lei manti sudgau lei maxri
The ant(s) (agentively doing something) dried the wheat.

I encourage you to edit this with appropriate leader material for Lojban
List.  The answer to your question is of general interest, since it
wasn;t obvious before, and the choice of -gau vs.  -ri'a is something
that has evolved mostly within the last year, indeed after the
sumti-raising discussion.

lojbab