Re: lo lunra selgusni ninmu

["Bob LeChevalier (lojbab)" <lojbab@lojban.org>]

>From: "=?us-ascii?Q?Jorge_J._Llamb=EDas?=" <jorge@intermedia.com.ar>
>>By the text and examples on pg 216, 13.5) on pg 234 is equivalent to
>>puku mi baku klama le zarci.
>
>Example 13.5) suggests that, I agree.
>
>>Introducing zu or co'a or co'u should not change how this works,
>
>I agree about {zu}, but introducing {co'u} does necessarily change
>how it works, because {puco'uba} is not a permissible Lojban tense.

Well it raises the questioned dealt with below.

>> though the
>>backwardsness of some of the tense/modals when used as sumti tcita might
>>cause them to need to be reversed in some cases.
>
>This is certainly a problem, but I think it doesn't affect this discussion.
>I suppose that {mi ba'oku klama le zarci} is the same as
>{mi ba'o klama le zarci} and not {mi ba'o zo'e klama le zarci}, right?

I hesitate to say, because John and I have answered the question before,
and it might even be in the Book.  My interpretation barring John saying
otherwise, however, would be that the ku form presumes the ellipsized
sumti.  The paradigm that had us add puku for example was originally that
of ellipsized sumti, and not as a semantics-free transformational grammar
maneuver.  It just was convenient and logical to make puku adjacent to the
selbri be equivalent to pu in the selbri.  But I think that
transformability need not be so for ba'o.

>>"Noncompoundable" is a grammatical issue arising solely from what John had
>>to do to make the grammar work under YACC and stay simple.
>
>That surely must be wrong. The simplest grammar would have been
>to make all tense cmavo part of the same selmaho and allow any

>and all combinations. That would certainly work under YACC.

The goal of the tense system grammar was not the simplest grammar, but one
which described the tense system.  That system went through 3 major
evolutionary stages during development, and all specified the grammar in
considerable detail based on PC's presentation to me and later John of the
logic of tense.  The current system is simplified to remove constraints,
but we don't want random strings of tense-related cmavo to be legal - we
want things that we can interpret via grouping rules etc.  (Originally PA
also had sub-grammar so that not just any old string of numbers and
semi-numbers could be legal, but we found that too many of the possibility
had usable meanings and chose to stop trying to specify them all, which was
impossible using LALR1 by that point.)

I know that in support of the compounding interpretation, there were some
things that could not be said with a single tenseconstruct because
ungrammatical, which John said would be expressed using two consecutive
tenses.  For example,

mi baki ne'iki klama 

parses by inserting a ku after the baki but 

mi ba ne'iki klama  
and
mi ba ne'i klama

parse with both together in a compound selbri tense

Clearly the baki[ku] ne'iki is the same construct, but using the ku to
avoid the grammaticality problem, which means in turn that it should be the
same as 
baki [ku] ne'iki [ku] because I can always stick a third ki tense in there.

Now this might override what I said above about the desirable
interpretation of ba'oku, since the grammatical generation of ba'oki ku is
inevitable at some point.  Which is why I will defer to John if he decides
one or the other should take precedence.  He has dominated the tense
grammar since he came up with the imaginary journey metaphor and wrote the
definitive tense paper - I only invented the grammar concept %^)


>If the complex tense grammar has any reason of being is
>precisely to _prevent_ some combinations from happening.

Indeed that is the intent, but especially to prevent ambiguous groupings.
However the necessities of LALR1 made very complex tenses easier to deal
with as multiple tenses especially with the imaginary jounreys metaphor and
the related storytime convention that preceded it and thereby gave a basis
for interpreting compounds.

Remember that for a long time you could not have both orders
space time
and
time space

The reason was that I had never figured out a way to get YACC to accept
either unambiguously and hence required the illegal one to be stated using
a ku separator (that this was indeed the solution du jour for that problem
is why I am sure that successive ku tenses are treated as sequential
elements in a hyper compound)

It was a late modification that John made that allowed both orders to be
possible without a ku.

>> I think it has
>>been clearly stated on the List, if not explicitly in the Book, that two
>>adjacent "noncompoundable" or "compoundable" for that matter tenses should
>>be treated as if they were compounded.
>
>That doesn't make sense. If they're noncompoundable they can't be treated
>as if they were compounded, by definition.

Grammatically noncompoundable.  Semantically, we can do whatever seems
necessary, and "as if compunded" seems like the simplest interpretation
(and I am not sure you have presented an alternative one).

>> I am only unsure whether this was
>>stated for particular kinds of noncompoundables or as a general case.
>
>I don't remember it ever being discussed. Maybe it was before my
>joining the list.

I am sure it was afterwards, since you joined while the imaginary journey
was still young and the reversability of space and time tenses did not yet
exist.  That changed, as I recall vaguely, after a discussion you started,
possibly even with a comment on the paper.

>>>>>{puco'aku baco'uku}
>>
>>Take an imaginary journey to the past and we have an initiation of an
>event.
>>That event is the future (relative to the pu offset already stated)
>>conclusion of X.
>
>I'm afraid I can't make sense of that. Let's make it more concrete. Let's
>say I have been painting my house, and painting the door will be the

>conclusion of the larger event of painting the whole house. Then I
>might say:
>
>                mi pu co'a co'u cintypu'i le zdani ca le cerni
>                I started the conclusion of painting the house this morning.
>
>meaning that I started to paint that last door this morning. What could
>the additional {ba} possibly mean? That the painting of the door was
>in the future of its start? How could it be in the future, since it has to
>start there? Does it mean that a part of it was in the future? And to make
>it more confusing still, what about something like {puco'aku puco'uku}?
>Starting of the conclusion that happened earlier?

You are using the ZAhO tenses that are points to make things more difficult
than they might otherwise be, resulting in semantic nonsense.

Given enough torturous thinking, I may be able to come up with an
interpretation that makes sense, but not online when I am trying thus far
unsuccessfully to be brief today so I can prepare some orders.

>>It is not clear whether or not that conclusion is in the past or future of
>>the space time reference.  puzuco'aku bazico'uku would be in the past of
>>the reference whereas puzico'aku bazuco'uku would be inthe future of the
>>reference.
>
>In any case, the conclusion would have to start in the past, wouldn't it?

I think that the use of the imaginary journey metaphor is such that once
you have moved away from the reference on the journey, it ceases to be a
very useful reference for that bridi.

>>>I don't see the need to force it when you have not
>>>one compond tense but several distinct tenses.

>>
>>I'm not sure if any other interpretation makes sense, so I don't see how it
>>is "forced".  Certainly not an implicit logical connective, since that can
>>so easily be made explicit with a multiple compound tense.
>
>Certainly not. I explained in my first answer why logical connection is
>different. For example: {mi pu je ba citka lo plise} means that I ate
>an apple in the past, and I will eat an apple in the future. Probably not
>the same one. {mi puku baku citka lo plise} would mean that my eating
>an apple was taking place in the past and will be taking place in the
>future, the same event, and thus the same at least one apple.

Remember that it is not necessarily the case that logical connectives
expand into separate bridi. I am not sure what has been said about tense
logical connection.  But nonlogical connection in any event is not
expandible, so pu joi ba should work.

>>  It's just
>>carrying a logical pattern to a rather extreme conclusion that probably
>>will never be useful (but pc could probably come up with an example if Nora
>>or John couldn't).
>
>I'm afraid I don't understand what you mean here. An example of what?

An example parallel to your house painting stumper above that makes sense
and justifies my interpretation as useful without going into time travel
scenarios (which were BTW envisioned as the ultimate interpretation of
nonsensical imaginary journeys in time).

lojbab