Message-Id: <199310040701.AA17521@eli.CS.YALE.EDU>
Date:         Mon, 4 Oct 1993 02:59:26 -0400
Reply-To: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Subject:      TECH: local unit gismu, hierarchical place structures
To: Erik Rauch <rauch>
Status: RO

>Subject: Re: TECH: Submultiples of local units
>Date: Mon, 27 Sep 1993 13:34:15 -0400 (EDT)
>In-Reply-To: "Logical Language Group" at Sep 24, 93 03:40:47 am
>
>mi'e djan. kau,n

mi'e je'abo la lojbab.

>Jorge asks about what "minli" means when the local distance unit is
>already kilometers.  Probably it means nothing, although the local units
>are more persistent than you'd think.  In France we still have the
>pound, now defined as 500 grams; in Italy it is 300g.  (A kilogram worth
>of food seems to be a bit too large for easy use.)  Likewise in Germany
>there is the "Klafter" (I think -- I only heard it used once, and I've
>never seen it written), which is 2m and corresponds roughly to the
>English "fathom".

In any case, these are, like the culture words, intended to be
lujvo-ized so that you can make it clear what kind of distance unit you
were talking about.

>I agree that the x3 place defaults to 0, but I think that the x4 place
>is required to know what the x3 place really means.  Some may consider
>the furlong to be the "first subunit" for the mile; others may have
>other choices.  I also don't see how there can be an arbitrary number of
>subunit places following x3, since x4 is nailed down as the standard.

On the basis of this argument, I would accept that any subunit places
should follow the standard place, moving the latter to x3.  There
certainly need not be something special about two levels of unit/subunit
(though it does seem that at least in English measurements we seldom
specify more than two even when there exist more.

If people want this change, I will change all five local unit gismu
identically.

>Perhaps the question of x3 in these unit gismu should be reconsidered.
>It seems to me that subunits could be handled within x2 using "pi'e":
>
>        le minli be li papi'ere
>        that-which is-measured-in-miles-as the-number 1;2
>        something that is 1 mile 2 furlongs (or whatever) long.

I would accept that this is a valid expression under the status quo,
without even requiring a change.

But...

you know no more about the semantics of the subunit place under this
scheme than you do under the status quo.  All you know is that there are
some kind of units, and some kind of subunits, and the pi'e suggests
that the number of subunits in one unit is not the current base
(typically 10).  You would be better to argue for the subunit side of
the pi'e to be expressed as a fraction with denominator equal to the
conversion factor.  Then at least you would recapture some missing but
useful semantics.  Of course, people don't do that in everyday language,
but it would be clearer.

I think the norm in modern English (or at least American) usage given
miles and subunits is for teh subunits to be either yards or feet, not
furlongs.  So you stand the risk under either method of having the
number interpreted as 1 mile and 2 feet.

We DO have the option of creating lujvo for the various combinations
that are most used and useful, specifying EXACTLY which subunits are
applicable.  A merkyxirmyjivnyminli (American-horse-compete-"mile")
would be miles and furlongs, while some other lujvo might be defined as
miles and feet.

>Lojbab refers to "jutsi" and "du" as the paradigm case for these
>indefinite-length place structures.  "du" is unproblematic, since all
>the places have the same semantics ("x1, x2, x3, ... are identical"),
>but I have always been troubled by "jutsi".  The difficulty is that some
>levels in Linnean classification are mandatory (species, genus, family,
>order, class, phylum, kingdom), whereas others are optional.  We might
>say
>
>1)      la'o .ly. Homo sapiens .ly. cu jutsi
>                la'o .ly. Homo .ly.
>                la'o .ly. Hominidae .ly.
>                la'o .ly. Primates .ly.
>                la'o .ly. Mammalia .ly.
>                la'o .ly. Chordata .ly.
>                la'o .ly. Animalia .ly.
>
>There are, however, several other hypernyms of "Homo sapiens":
>superfamily Hominoidea and suborder Anthropoidea, which fall between
>Hominidae and Primates; superclass Eutheria, which falls between
>Primates and Mammalia, and subphylum Vertebrata, which falls between
>Mammalia and Chordata.  Not all orders or phyla are subdivided, and not
>all families and classes are grouped.  Therefore, the meaning of "ve
>jutsi" or "xe jutsi" or "sexipano jutsi" are awfully vague.  Is this
>really tolerable?

They are vague but you know one thing of import:  that each succeeding
place is a higher level in the hierarchy.  It would NOT be valid to say

1a)      la'o .ly. Homo sapiens .ly. cu jutsi
                la'o .ly. Animalia .ly.
                la'o .ly. Homo .ly.
                la'o .ly. Mammalia .ly.
                la'o .ly. Chordata .ly.
                la'o .ly. Hominidae .ly.
                la'o .ly. Primates .ly.

since that doesn't matche the hierarchical order of the categories.  I
will agree that le xe jutsi is relatively meaningless without all more
detailed branches of the tree specified (i.e. se and te and ve).

Note also that in this case you don't have the option of "pi'e".  I'm
not sure there's a practical alternative even if we agreed that there
was a problem to be solved.  To express it all in one "name" doesn't
even give an indication of which portion of the name is the more
detailed part of the tree (unless maybe the morphology of the words
tells you, and I doubt that most scientists, much less the general
public, knows the morphological forms needed to make this informative.

Normal scientific usage of course, is backwards from our scheme, with
the genus followed by species followed by subspecies.  But this leaves
no room for omitting the genus and including the subspecies, or talking
at higher levels of the hierarchy than the genus level.  Thus we end up
looking like time-of-day (tcika) or dates (detri) which have similar
need to work at differing levels but without the capability to use
numbers separated by pi'e.  So you end up having to use sumti joined by
joi, or ce'o with the whole surrounded by mass brackets.  Yecch!

How do you know that a time "03:54" is hours and minutes, or minutes and
seconds?  Convention coupled with context.  How will you know what the
places are in a species name?  Convention coupled with context.  I like
the separate places version better than the single place version since
it has more flexibility, and the capability doesn't exclude the use of
the single place technique.  You CAN say "minli be li cipi'evo as easily
as minli be li ci bei li vo (no Zipfean advantage either way - 2
syllables however you do it)).

lojbab

lojbab