[lojban] Re: A Proposed Explanation of {gunma}

[John E Clifford <lojban-out@lojban.org>]

--- Jorge Llambías <jjllambias@gmail.com> wrote:

> Here are some of my thoughts on loi/joi and
> gunma/cmima/pagbu.

Gee, it's nice to agree with xorxes on a
potentially controversial issue.
> 
> 1- I will avoid using the term "mass" or any
> other similar noun
> to talk about how loi/joi work. This is in part
> to avoid confusion
> with English "mass nouns", but mainly to avoid
> a reifying
> interpretation of loi/joi which any noun
> automatically brings
> in with it. This is not to say, of course, that
> we can't talk in
> Lojban about groups, collections, assortments,
> bunches,
> sets, crowds, packs or any other creatures with
> constituent
> members. Those things are perfectly valid
> things to talk about,
> but in my opinion loi/joi do not by themselves
> bring in any
> reference to any of them. If we talk about a
> mass of some
> things, then we are immediately talking about a
> new entity
> with properties of its own, and then we have to
> discuss how
> properties are inherited back and forth between
> this new thing
> and its constituents, and that is really not
> needed for loi/joi.

Well, while {joi} (and perhaps {loi}) don't call
for new entities, they don't prohibit them
either.  And if you are a traditionalist in
requiring singular reference/quantification, then
such references are required.  The nice thing
about bunches is that they work either way and
the inheritance is limited to what is already
needed in terms of distributive and collective
predication.

> 2- The ONLY thing loi/joi do is block a
> distributive reading
> for its referents. {joi} always involves at
> least two things,
> one on either side. {loi} need not in principle
> involve at least
> two, but it makes little sense to block a
> distributive reading
> for one single thing, so at least pragmatically
> it tends to
> involve at least two things.
> 
> 3- loi/joi contrast with lo/jo'u which do not
> block a distributive
> reading (but neither do they force one). For
> example, it is
> perfectly acceptable for me to say something
> like:
> 
>    la djan jo'u la meris pu jgari lo tutci gi'e
> co'a zbasu lo zdani
>    John and Mary grabbed the tools and started
> to build a house.
> 
> where "grabbed the tools" can be distributive
> (each one of them
> grabbed a different tool) and "started to build
> a house" collective
> (they started to build the same house
> together). It is possible
> to be more precise if desired:
> 
>    la djan .e la meris pu jgari lo tutci ije la
> djan joi la meris
>    co'a zbasu lo zdani
> 
> where {.e} in the first sentence is
> distributive and {joi} in the
> second is non-distributive. This precision is
> often unnecessary,
> but it is available when needed. {lo} and
> {jo'u} are useful because
> we often want to apply both distributive and
> non-distributive
> properties to the same referents at the same
> time.
> 
> {ko'a .e ko'e} is distributive
> {ko'a joi ko'e} is non-distributive
> {ko'a jo'u ko'e} is silent on distributivity.
> 
> {ro broda} is distributive
> {loi broda} is non-distributive
> {lo broda} is silent on distributivity.

It should be noted -- pragmatically again -- that
in contexts where the distributivity/collectivity
is at issue, {lo} becomes the marker for
distributivity just by not being {loi}.  But such
situations are are rare and so {lo} usually
functions simply as the designator for the things
involved, without indication of how they are they
are predicated of.  On the other hand, the gadri
and quantifier methods of indicating predication
are technically inadequate since they do not say
what kind of predication is involved in the
description itself: "Those who (together)
surround the building are red-haired" might need
to be distinguished from "Those who (each)
surround the bulding are very long walls" for
example.  The gadri indiciation also complicates
collapsing cases of two kinds of predication with
a common core: "The wall (by itself) and the
students (together) surround the bulding" and as
above "John and Mary (separately) picked up tools
and (together) started to build a house." 
Context does not always solve the question and
accuracy -- using gadri or ,e, connectives --
does not allow the collapse in Lojban (which
currently has no fall-back device).  Finally, {ro
broda} for distributive predication would only
work if it is firmly in mind that it is short for
{ro lo broda}, since this would no longer be the
same as {ro da poi broda}, another natural
reading.

> 4- Using {loi} for substances and generics is
> in my opinion pragmatically
> wrong, because in order to block distributivity
> {loi} has to first
> bring the issue
> up, and with substances and generics the issue
> should not even arise. I find
> {le kabri cu vasru lo djacu} much better than
> {le kabri cu vasru loi djacu}
> for "the cup contains water", even if both are
> theoretically possible.
> {loi djacu} brings in the totally irelevant
> possibility of the distribution
> of quantities of water, just in order to block
> it.

I am not sure what this means exactly.  The point
of using {loi} to construct generic or substance
expressions (mass nouns in the usal sense)is
exactly that, in Lojban, such notions are just 
individuals taken (predicated of) collectively
and so inherently different from distributive
cases.  the possibility of distributing over
these individuals is not irrelevant; it is simply
not what is going on in these cases. It would be
irrelevant in Urgoo interpretations of
substances, but that is not Lojban's way.

> 5- I think {gunma} should mean:
> 
> "x1 is a
>
mass/group/bunch/aggregation/collection/assortment
> consisting of constituents x2"
> 
> where x1 is a single entity and x2 are many
> entities that together
> (non-distributively, obviously!) constitute x1.
> So for example:
> 
>   le kamni cu gunma la djan joi la meris
>   The committee is formed by John and Mary.
> 
>   la djan .e la meris cu cmima le kamni
>   Both John and Mary are members of the
> committee.

One is tempted to say things like {loi broda cu
gunma lo broda}, even though it is not *required*
that {loi broda} and {lo broda} refer to the same
objects taken together or not.

> 6- {pagbu} differs from {cmima} in that {pagbu}
> is a transitive
> relationship and {cmima} is not. If A is a part
> of B and B is a
> part of C, then A is a part of C, but if A is a
> member of B and
> B is a member of C, then generally A need not
> be a member
> of C.

{pagbu} will not always work for "part" if
transitivity is maintained, since I am a part of
society and my liver is a part of me, but my
liver is not a part of society.  Presumably the
first of these should be {cmima} or even some
more specific word. The "in" relation for bunches
("among" for plurals) is transitive, but {pagbu}
feels wrong for it -- probably refoecting the
aura of the English word again.
 
> {pagbu} is not much related to joi/loi:
> 
>    lo xislu cu pagbu lo karce
>    A wheel is part of a car.
> 
> mu'o mi'e xorxes
> 
> 
> To unsubscribe from this list, send mail to
> lojban-list-request@lojban.org
> with the subject unsubscribe, or go to
> http://www.lojban.org/lsg2/, or if
> you're really stuck, send mail to
> secretary@lojban.org for help.
> 
> 



To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.