Message-Id: <m0sdUPs-0000ZHC@xiron.pc.helsinki.fi>
Date:         Tue, 1 Aug 1995 20:03:43 -0700
Reply-To:     "John E. Clifford" <pcliffje@CRL.COM>
Sender:       Lojban list <LOJBAN@CUVMB.BITNET>
From:         "John E. Clifford" <pcliffje@CRL.COM>
Subject:      lu'a
To:           Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Content-Length: 3056
Lines: 55

        Last week's note from lojbab (with Cowan the only people whose
grammar I trust more than xorxes') tends to confirm my reconstruction of
the descriptor set.  In LAhE there appear to be a set of operators which
move -- I would say -- from one set-related type to another.  In the terms
of my reconstruction these would be roughly

lu'i strips away the modifications to get to just the set (le, lei, le'e
to le'i, Ax e {broda}, Mas({broda}), Avg({broda}) to {broda})

lu'a takes the underlying set distributively (lei, le'i, le'e to le, etc.)

lu'o takes the underlying set to its mass (le, le'i, le'e to lei)

        I miss the converter to Avg({broda}), which seems important to me,
given the confusion between the two when talking about individuals: the
mass of lions inhabits North America (inter alia) but the typical one
inhabits Africa. I wonder if the ones in the list convert from the average
to the other objects; I have assumed they do.

        The sequencer is new and rather nice; I assume the sequence is
always of length k({broda}) so that all and only the set members get in
and each exactly once. The status of this is a bit unclear, however.  It
does not give a unique sequence, so presumably it gives some member of the
set of all sequences, a kind of lo.  I would like to see it in use a bit
before I felt I knew what it meant practically.

        I now begin to see wherein was my problem with xorxes' exposition
in the "quantifier" thread.  I took his

>{lo'i broda} and {loi broda} are short forms of {lu'i ro lo broda} and
>{lu'o ro lo broda}, and because of the {ro} there is only one of each.

as definitions for the more
fundamental notions, rather than as equivalences.  I also did not notice
(nor does lojbab mention) the important point that these converters can be
used to form subsets and related items by being applied to descriptions
quantified by other than ro (presumably this works for sequences as well).
Xorxes extols the virtues of this capability later in the other thread.

        Given all of this, what can be made of the discussion that
occupies this thread (I missed the beginning of it, so part of this may be
only guesswork -- including, alas, identification of the speakers in some
of the pieces lojbab excerpted.  I guess at th em being himself and xorxes
and make a stab at assigning the views aright)?  I take it that lu'a le
selcku is a repetitive redundancy (or else illegal), since it convert a
distributive reference to a set into a distributive reference to that set;
lu'a ci le selcku might make sense, bringing us down to a new set
(assuming that I was calling more than three things selcku originally).
But even starting with lei selcku, lu'a lei selcku does not get us bits
and pieces of selckus (whatever they are, books?) but rather just back to
le selcku again (barring other quantifiers again).  It's the same set
massified, identified, distributed or averaged.  To get book parts out,
you gotta put book parts in. Massification is a logical operation, not a
Waring blender.
pc>|83