Message-Id: <199501281609.AA06074@nfs2.digex.net>
Date:         Sat, 28 Jan 1995 08:08:08 -0800
Reply-To: "John E. Clifford" <pcliffje@CRL.COM>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: "John E. Clifford" <pcliffje@CRL.COM>
Subject:      ago
To: Bob LeChevalier <lojbab@access.digex.net>
Status: RO

        The "ago" thread and some of its subfibres illustrate the paradox
of the whole notion of a Logical Language.  The logic part pulls us
towards evr greater precision (surprisingly, given the imprecision - not
to say plain wrongness - of so much that is said in logic books).  The
language part pulls us toward simpler modes of expression until we get
the right size for what we have to say.  (BTW the right size in Lojban
will always be a little longer than in English, since the lojban units
are a little longer - so "the right size" is relative to the base of the
language.)
        The happiest upshot of this tension will be a right sized
expression which does not misfit the situation being described, even
though it is not a perfect fit.  It will be "suitably imprecise," so that,
while it is true of the situation being described, it is also true of
other situations which are (occasionally significantly) different from the
one aimed at.  (In mathematics, Peano's postulates are almost always taken
as describing the natural numbers, although they fit all manner of other
critters -- Robinson arithmetic, for example -- that are grossly
different.) The hope is that the conventions espoused in pragmatics will
get us to the right target the first time and then rote idiom learning
will fix this expression in usage.
        Sadly Griceans (and Austinians et al) have been great exegetes
but less impressive prophets. Given that utterance u uttered in situation
s carries message m (which is different from the literal, contextless
message of u), pragmaticists of this ilk can give good (often blindingly
brilliant) explanations of how and why it works that way.  But before the
fact, they have had less success with predicting what message u will bear
in s (where it is in violation of some conventions - itself not always an
easy fact to spot) and virtually no (that I know of -- and I admit to
being a Maj-Gen Stanley, plucky and adventury but sometimes a decade
behind) luck in figuring out what u will convey a given m in in a given
s.  Thus, we cannot turn to them (alas -- and may I be stood corrected)
to tell us where to look nor when we have found the right expression for
a task.  EXCEPT (Praise Jesus!) that we want one that satisfies the test
they can do: we declare that our u does carry m in s and they find a
good explanation of how and why, so that it is believable that we will
regularly pick out the right situation out of all the ones that the
expression fits.
        By way of suggesting a couple of devices which may work for "ago"
(I assume that _zai_ in my ancient cmavo list is defunct or, at least,
does not work as it seems to from the description there), here are some
things that are about the right size and are true in the situation in
question.  I give them in English because of the unreliability of my
cmavo lexing, which follow.  For "three years ago"
        "before the past year triad" _pu lopu nirne cimei_
(this can be adapted for the accuracy of the "three years" down to "at the
beginning of the past year triad" _ca lepu<INIT>_ or "just before"_puzi_)
        "during the fourth past year" _ze'e lopu vomoi nirne_ (_nirne co vomoi
be la Cac_ to be on the safe side)
        Both of these assume that the years are taken as discrete and
adjacent (not overlapping and not a scattered collection) and in natural
order.  This set of assumptions -- presuppositions? implicatures? --
might even extend to justifying
        "Before three past years" _pu cipu nirne_, an expression close
to English and so probably just right for Lojban in length.
pc>|83