Re: Still a few thoughts about ZAhOs

[VILVA@VIIKKI21.HELSINKI.FI]

This posting contains responses to several postings

> Date:         Mon, 16 Aug 1993 12:29:32 EDT
> From:         Jorge LLambias <jorge%PHYAST.PITT.EDU@FINHUTC.hut.fi>
> Subject:      Re: Still a few thoughts about ZAhOs

> Veijon, this discussion is proving to be very helpful, at least to me.
> I think we are converging, slowly but surely, to the same understanding.

  me too.

> You haven't addressed the issue of the mix up between ba'o and pu'o,
> what are your thoughts on that matter? (I make some comments on it
> answering one of your points below.)

  I'll handle this later on.

> > This means that "ZAhO broda" defines a second-order relation.

> With respect to "broda", yes. The same can be said of "PU broda" or
> the other tenses.

  Here I don't quite agree. Due to the aoristic nature of the PU tenses
  they have no special effect on the underlying relation.

         da PU broda de di

  is roughly

         da broda de di PU le *cabna

  where *cabna refers to the time of utterance. The utterance merely
  states that the base bridi was true at least at the indicated part
  of the time axis relative to the time of utterance but may well be
  true also at any other time.

> > There is NO transformation which given the ZAhO
> > and the broda would give the corresponding relation ZAhO_broda. The
> > meaning must be inferred from extratextual knowledge pertaining to the
> > nature of the broda.

> Here I'm not quite sure of what you are saying. The meaning of the
> relation has its bones in the meaning of "ZAhO broda", and the context
> provides, of course, the flesh, but the meaning of the "ZAhO broda"
> should be clear. If by NO transformation, you mean that you cannot
> express the sentence in a different way having _exactly_ the same
> meaning, then I agree. But it is not necessary to do that in order to
> understand the basic meaning of the relationship.

  What I meant was that given 2 different relations broda and brode
  there is no general function f such that

      f(ZAhO,broda) = ZAhO_broda

  and simultaneously

      f(ZAhO,brode) = ZAhO_brode

  i.e. that there is no general rule which would define how to derive
  (preferably programatically) ZAhO_broda from any given ZAhO and broda.

> > At Lojban level this doesn't present any great difficulties once you
> > learn to regard events as sequences of phases/states and ZAhO bridis
> > as mental snapshots of these phases -- there is actually no need to
> > analyze a phase into the basic ingredients.

> This works for non-ZAhO bridis as well, doesn't it?

  Yes, of course. What I tried to say was that it isn't necessary
  to actually KNOW the ZAhO_broda.

> I agree that the ZAhO are extremely important, that's why it is a pity
> that ba'o and pu'o got mixed up.

> But if on top of time being a nuisance, the similarities
> ba-ba'o and pu-pu'o lead you to make the wrong connections, it becomes
> much harder to grasp their true meaning. Every time I read "ba'o", I have
> to tell myself: remember that the future has nothing to do here". The
> intrusion of the past in "ba'o" would not hurt at all, and it may even
> help, since what happened in the past is relevant to the state described
> by the ba'o bridi.

  see later.

> The simple replacement of "ba'o" by "...is in the aftermath of..." is
> misleading, since it is not necessarily true that "x1 is in the aftermath
> of brodaing"

  Well, it seems to me we haven't quite yet converged :-)

  For me stating

             mi ca ba'o zvati la paris.

  certainly implies something about me and Paris at the present.
  'Me and Paris are in the aftermath of my having been to Paris.',
  neither of us is quite the same :-)

  As to the general case, statements

  (A) mi ba citka pu'o le nu do ba klama mi
      I'll eat before you come to me.

    and

  (B) mi ba ba'o citka pu'o le nu do ba klama mi
      I'll be in the state of having eaten before you come to me
      I'll be in the aftermath of having eaten
        before you come to me
      (*I'll have eaten before you come to me)

 tell each something different. In (A) the present definition says
 that the event "le nu mi ba citka" will be finished before you come
 (see later). If the aftermath wouldn't carry any weight, (B) would
 mean exactly the same. However, (B) is DEFINITELY a statement about
 "mi" or rather the state of "mi" in the aftermath.

 "la paulus. ba'o klama la damaskus." means that Paul is in the
 aftermath of having gone to Damascus. This doesn't merely state
 that "la paulus. pu klama la damaskus." is true. It says that
 Paul will ever since have the quality of having gone to Damascus.

> > At a moment like
> > that you most certainly don't think "ko'a has loved me" but "ko'a and
> > I are past ko'a loving me".
>
> "past" is the key word here. ;)  :(

 (-: Here the English is slightly mixed up as the meaning of the
     preposition "past" refering to time is 'beyond in time; after',
     e.g. past noon, half past six.
     Similarly for locations: He went past the gate. :-)


> Date:         Mon, 16 Aug 1993 12:18:58 -0400
> From:         John Cowan
<snark!cowan%GVLS1.VFL.PARAMAX.COM@FINHUTC.hut.fi>
> Subject:      Re: Still a few thoughts about ZAhOs

> > This gives us (approximately, ignoring perhaps some finer points
> > relating to the ZAhO in question)
> >
> >         da *ZAhO_broda de di ZAhO le nu da broda de di
> >
> > e.g.
> >
> >         da ba'o klama de di
> >     =>  da *ba'o_klama de di ba'o le nu da klama de di
> >
> > where "*ba'o_klama" very clearly cannot equal "klama" as the relation
> > between da, de and di IS NOT "da klama de di" anymore in the AFTERMATH
> > of the coming -- da is already at de, not coming to de anymore.

> However, I cannot agree with the above conclusions, because of the
> lack of tense on "le nu da klama de di".
>
> You treat "da klama de di" as if it meant "da caca'oca'a klama de di", but
> IT DOES NOT.  It is open as to tense proper, aspect, and actuality.
> So "ba'o_klama" is not a different relationship from "klama" proper, but
> simply a subtype of it: it is klama seen from the aftermath perspective.
>
> We are prone to believe that "caca'oca'a", that which is actually
continuing
> now, is the most important part of the event, and can always be assumed as
> the default, but it cannot.

  As I noted I WAS ignoring some fine points.

  I was using "le nu da klama de di" in the usual NL sense to refer
  to "le nu da ca'oca'a klama de di" i.e. the actual realized action
  part of the event in question. The main purpose was to derive an
  internal time reference using a ZAhO-tagged sumti. (BTW I have doubts
  about inserting a PU tense into the event descriptor as PUs are relative
  to the time of utterance which here isn't relevant as we have a definite
  event which we are refering to.)

  I have nothing against calling "ba'o_klama" a subtype of "klama" --
  actually it makes sense because, what ever the relation is, it must
  contain traces of "klama".

> "klama" is a most expansive relation, and
> just as well associates the coming of Paul to Damascus, an event which
> long ago entered its "ba'o" stage, as my going home tonight, which is
> hardly even in its "pu'o" stage yet.

  Combining this, Bob's comment on this and what I was trying to do
  leads to a mess. Things do get kind of hairy and it's quite easy
  to end up with circular references like 'the aftermath of this event
  is in the aftermath of this event'. I'll try to reformulate.

  My starting point was that, at least externally, "la paulus. ca ba'o
  klama la damaskus" is a statement about Paul and Damascus and NOT
  a statement about the event of Paul's going to Damascus. This I
  felt was a very important distinction if we are to maintain that
  Lojban is a logical language.

  The explicit PU tense doesn't really matter as it merely shifts
  the whole mess relative to the narrator. Then I tried to clarify
  things by replacing the original relation with a new one which
  would hold only on a clearly demarcated part of the time axis.

  Firstly I wanted to show that the relation had to be different at
  different phases of the event. Secondly I wanted to show that
  using the cmavo "pu'o" for the ZAhO tense refering to the aftermath
  in no way contradicted the use of "pu'o" as a sumti tcita.

  Let's pass the first by noting that these relations are just
  subtypes of the klama relation appropriate to the phases of the
  event.

  Now the question about the assignment of the names of the cmavo
  remains to be solved.

  As I have stated and implied several times before I see NO
  contradiction what so ever between "pu'o" and "pu", either
  as tenses or as sumti tcita.

  Let's start from the basics.

  (1) a PU tense defines the position of the action of the base
      bridi (the ca'o phase) relative to the time of utterance
      (ignoring the aoristic nature of the PUs)

  (2) a ZAhO tense defines the position of the indicated phase
      of the process relative to the process proper (action, ca'o).

  Both of these are a matter of definition. In (2) we could choose
  to regard the action from the standpoint of the sumti in which
  case the action would be first in the future and then in the past.
  However, ZAhO tenses handle the event contour or the phases of
  a process. We are already used to talk about preprocessing and
  postprocessing, pre-war, post-war.

  ZAhOs are about being not doing. We might say

       He is in pre-action state      ko'a ca pu'o zukte
       He is in action                ko'a ca ca'o zukte
       He is in post-action state     ko'a ca ba'o zukte

  as contrasted to

       He is going to act             ko'a ba zukte
       He is acting                   ko'a ca zukte
       He has acted                   ko'a pu zukte

  As I have stated before, being in post-action state in no way
  implies existence, the state may as well be the state of non-
  existence -- a state is a state.

  Is the fact that e.g. "ba'o" can be just moved to link a nucleus
  event descriptor to the basic bridi in a way that CORRECTLY reflects
  the temporal position of the part relative to the nucleus of the whole
  a sufficient justification for the present assignment?

  We presently have the following parallel transformations (again
  ignoring a lot of details but don't let's get into that anymore,
  I know (roughly) what I'm eliding.)

  ko'a ba klama       => ko'a klama ba le *cabna
  ko'a ca ba'o klama  => ko'a ca *klama/b ba'o le nu ko'a co'aca'a klama

  (Having the explicit "ca" already greatly diminishes the pervasiveness
  of associating "ba'o" with a future tense.)


> Date:         Mon, 16 Aug 1993 14:21:23 EDT
> From:         Jorge LLambias <jorge%PHYAST.PITT.EDU@FINHUTC.hut.fi>
> Subject:      Re: Still a few thoughts about ZAhOs

> la djan. cusku di'e
> > la veion. cusku di'e
> >
> > > A bridi like
> > >
> > >         da ZAhO broda de di
[...]
> > I believe that most of your thoughts on ZAhOs are entirely sound.
> > However, I cannot agree with the above conclusions, because of the
> > lack of tense on "le nu da klama de di".

> This gets interesting! Then the ba'o as sumti tcita does _not_ define
> a contour for the internal event of the sumti? This is what I would like,
> but not what I've been told up to now, assuming that I understood what
> I've been told.
>
> To be consistent with the other tenses, the contour should apply to the
> main bridi, but it doesn't. In fact, it seems that as sumti tcita, the
> ZAhOs acquire the ability to give a contour to both the main bridi and
> the bridi inside the sumti.
>
> So, although I don't like the present interpretation, it seems to indicate
> that Veijo is right. (Accepting as he does that he is ignoring some finer
> points.)


  The Imaginary Journeys paper states about ZAhO used as sumti tcita that
  "The event described in the sumti is viewed as a process, and the action
  of the main bridi constitutes the part of the process which the ZAhO
  specifies."

  If I understand this and the given examples correctly it implies that
  the starting and ending points of the process defined by the outer bridi
  are CONTAINED within the ZAhO part of the sumti process. A PU tcita on
  the other hand requires only, if I understand things correctly, that
  there is an OVERLAP of the ca'o of the outer process and the part of the
  time axis indicated by the PU tcita.

  My PRESENT position regarding the event contours is

  (1) as far as the outer bridi is concerned, in the absence of an
      explicit ZAhO tense the ZAhO tcita sets OUTER limits to the
      action, i.e. the ca'o phase. This defines the event contour
      in a rather vague way (depending on the actual ZAhO used) but
      more precisely than the use of a PU tcita would. More precision
      can be obtained in some cases by using a ZAhO tense in the outer
      bridi.

  (2) the ZAhO tcita has usually a subsidiary role so that we are
      only interested in the event contour as a kind of scale which
      makes it possible to temporally locate the main bridi (A).
      There are, however, times when we may want to add emphasis to
      the contour (B).

         (A) mi klama ba'o le nu do pu klama
         (B) mi klama ca'o le nu do pu ba'o klama

      I think that in (A) the event of the attached sumti is seen
      as a whole (regardless of the ZAhO used) which is used solely
      to define a reference interval/point, while in (B) the internal
      contour is important:

          (A) I came after you
          (B) I came while you already were there

      There are situation where more complex contouring is necessary:

          mi (ba) klama do pu'o le nu do ba ba'o citka

----------------

  About the range of pu'o and ba'o

  I have stated before that I don't think "pu'o" means "all the time
  until the event" and "ba'o" "all the time after the event". On the
  other hand I have indicated that these phases may have indefinite
  ranges practically covering all eternity. This seems to be a
  contradiction. This is partly explained by the fact that my thoughts
  have been developing, partly by the fact that I consider both the
  leader and the trailer phases to have a kind of density function.

  I think "pu'o" and "ba'o" are used for two separate kinds of
  reference:

      (1) to refer to an indefinite interval somehow connected
          with the event

      (2) to refer to the immediate inchoative/perfective phases
          of an event, e.g. "I'm about to ...", "I am just
          cleaning up after the act of ..."

  These are both necessary and there are cases when separating them
  isn't desirable. I think that excluding one type of use would
  create the necessity of adding two new contours to the existing
  set.

  This is, however, the kind of vagueness we are used to deal with
  in human communication. At strictly definitional level I tend to
  agree with Bob though I have used and, most likely, will use
  these cmavo in the sense of (2).


  co'o mi'e veion

------------------------------------------------------------------

 Veijo Vilva       vilva@viikki21.helsinki.fi