https://mw.lojban.org/index.php?title=%27%27%27lo%27%27%27_down&feed=atom&action=history'''lo''' down - Revision history2024-03-28T13:51:01ZRevision history for this page on the wikiMediaWiki 1.38.4https://mw.lojban.org/index.php?title=%27%27%27lo%27%27%27_down&diff=111414&oldid=prevGleki: Created page with "<div>Posted by [http://www.lojban.org/tiki/user11 pycyn] on Tue 27 of Sep., 2005 20:53 GMT posts: 2388</div><div> A primary occurrence of a sumti is one not in the scope of a..."2015-01-26T10:49:44Z<p>Created page with "<div>Posted by [http://www.lojban.org/tiki/user11 pycyn] on Tue 27 of Sep., 2005 20:53 GMT posts: 2388</div><div> A primary occurrence of a sumti is one not in the scope of a..."</p>
<p><b>New page</b></p><div><div>Posted by [http://www.lojban.org/tiki/user11 pycyn] on Tue 27 of Sep., 2005 20:53 GMT posts: 2388</div><div><br />
A primary occurrence of a sumti is one not in the <br />
scope of a negation, abstraction, modal or <br />
non-assertive speech act. In general, primary <br />
occurrence sumti can serve as the premise of a <br />
particular generalization: from {[http://www.lojban.org/tiki/sumti sumti] broda} <br />
to {da broda}. The usual upshot of this is that <br />
sumti refer to things that exist in the world for <br />
which the sentence is being evaluated.<br />
<br />
In xorlo, some primary occurrences of sumti which <br />
do not refer to things existing in the evaluation <br />
world are nonetheless true: certainly {lo <br />
pavyseljirna cu pavyseljirna}, probably {lo <br />
pavyseljirna cu simsa lo xirma}, maybe even {lo <br />
pavyseljirna cu blanu}. <br />
These are all strictly false in prelo, the <br />
previous interpretation of {lo}. However, there <br />
are a variety of workarounds. For xorlo we could <br />
change the domain of descriptions to, say, one in <br />
which every well-formed sumti had a referent and <br />
insofar as possible given existence constraints <br />
where that sumti indicated what sort of thing <br />
the referent should be, it was that sort of thing <br />
({lo pavyseljirna co se darxi be mi} would be a <br />
unicorn but not in this world — one hit by me, <br />
since what is hit has to exist in the world of <br />
the hitting). The quantified variables would <br />
then range over this domain. Thus, the first <br />
inference would go through, but the second (from <br />
there is a to there exists a) would not, in <br />
any case. There are some inconveniences with <br />
having things this way, but some with not having <br />
it as well. And the inconveniences with having it <br />
are practically less than theory suggests, since <br />
so many interesting predicates require that <br />
significant places be filled by reference to <br />
existents, not just to beings.<br />
<br />
For prelo, the standard workaround has always <br />
been that if we are talking about unicorns then <br />
we are in a world where unicorns exist. But that <br />
only works sometimes; clearly if someone rushes <br />
in saying I just saw a unicorn and we reply <br />
Not likely, since they dont exist we have not <br />
made the shift. More satisfying for the certain <br />
cases is the notion that tautologies, like <br />
unicorns are unicorns are true even with <br />
vacuous terms: even brumpfs are brumpfs is true <br />
whether or not we have any idea what brumpfs are. <br />
Strictly speaking this takes advantage of the <br />
fact that anything can be omitted in Lojban (if <br />
it is clear from the context) and modals and <br />
the like are particularly apt to fall under this, <br />
so here we have omitted an obvious {cae} <br />
(assuming that is the right thing for <br />
definitional claims and closely related items <br />
like tautologies). The second claim above, that <br />
unicorns are like horses, can similarly be <br />
justified as a hidden {kau}. The third, that <br />
unicorns are blue, is harder, since this goes <br />
against cultural norms (unicorns are typically <br />
white and, at worst, run through horse colors in <br />
the cultural understanding). But the claim is <br />
also, for those very reasons probably false, even <br />
though it is logically possible that there be <br />
blue unicorns (and on the broad domain view there <br />
certainly are since {lo blanu pavyseljirna} is a <br />
well-formed sumti and nothing prevents its <br />
referent from being blue any more than from being <br />
a unicorn). <br />
That is, the practical difference between xorlo <br />
and prelo in this area is vanishingly small and <br />
comes down eventually to slightly different <br />
interpretations, differences that will appear <br />
within each of the positions standing alone <br />
(i.e., the difference between the two systems is <br />
no greater than between different instances <br />
within a given system).<br />
<br />
Such easy relief does not appear for the other <br />
case of difference. To say that I want a unicorn <br />
in prelo requires {tua}: {mi djica tua lo <br />
pavyseljirna}, while xorlo can say just {mi djica <br />
lo pavyseljirna}, what used to be a malglico <br />
solecism. In this case, the fact that there are <br />
no unicorns in the evaluation world is not <br />
crucial; the same problem arises with {mi djica <br />
lo mikce}. Nor does expanding domains help any <br />
here: even in the widest domain, the move from <br />
{mi djica lo broda} to {da (poi broda) zou mi <br />
djica da} does not work in general, for it can be <br />
shown for every broda, fub, even in the extended <br />
domain that {la fub zou mi djica fy} is false <br />
even when the original is true — because I would <br />
have been as satisfied, my desire met, by any <br />
other broda every bit as well as this one.<br />
<br />
The problem is, of course, that of intensions. <br />
For expressing them there are generally two <br />
solutions: intensional places or intensional <br />
expressions. English does a mix; Lojban aims at <br />
doing the second, though even in unchallenged <br />
areas there are exceptions. For a logical <br />
language to use places would seem to require that <br />
those places be overtly marked to prevent by <br />
formal interdiction the objectionable inferences <br />
(and to allow unmarking where appropriate). <br />
Merely learning a list (even if guaranteed <br />
exhaustive) does not seem sufficient. So, for <br />
Lojban, the suggestion that {djica2} is <br />
intensional is at variance with the program, to <br />
be taken up only as a last resort (it also does <br />
not allow marking cases where the inference goes <br />
through).<br />
<br />
But there is the other choice, namely to take <br />
expressions in {djica2} as referring to <br />
intensional objects. One doesnt want to do this <br />
across the board, since that would again preclude <br />
a marking for the generalizable cases and, more <br />
to the point, some expression pretty much have to <br />
refer to particular, identifiable things (even <br />
though occasionally nonexistent ones) where the <br />
inference usually goes through with only the <br />
existence problem (for which solutions are <br />
available). That is, most kinds of sumti refer <br />
unambiguously to extensional objects (or <br />
abstracta considered extensionally). In fact, the <br />
severe problems arise only with {lo}, which, <br />
because of its generality, is particularly liable <br />
to the sort of fallacy sketched above. Thus, <br />
much of the problem presented is solved if {lo <br />
broda} is taken to refer to an intensional <br />
entity, say the broda species or brodahood or <br />
brodaness (what happens in these case differs <br />
somewhat but the overall pattern is pretty much <br />
the same; in what follows we will stick to <br />
brodaness, the property of being a broda). That <br />
is, {mi djica lo broda} describes a relation <br />
between me and brodaness, a relation involving a <br />
tension that would be resolved just in case I <br />
come into some (unspecified, but covered by <br />
having) relation with something that has <br />
brodahood, suo broda.<br />
<br />
Of course, {lo broda} does not refer to brodahood <br />
only in {djica2}(and other intensional places, as <br />
it were), for that would recreate the problem of <br />
unmarked opaque places, which we are trying to <br />
avoid. Thus, {lo broda} refers to brodahood in <br />
all places where there is no mark that it does <br />
not (assuming we want to have some such places). <br />
In particular, since it is brodahood in {djica2} <br />
it must be brodahood in all places of all <br />
predicates (else we would have unmarked places). <br />
On the other hand, for example, {lo broda} as <br />
part of a more complex description: {Q lo broda} <br />
can be given an extensional reading, if that is <br />
desired (in any case the whole expression is <br />
extensional).<br />
<br />
Now ordinarily, we say that {[http://www.lojban.org/tiki/sumti sumti] brode} is <br />
true just in case the referent(s) (in the <br />
evaluating world) of [http://www.lojban.org/tiki/sumti sumti] is/are in the set <br />
of things assigned the predicate {brode}, that <br />
is, have the property brodeness in the evaluating <br />
world. This will not work for the intended <br />
purpose of {lo broda cu brode} to say that <br />
brodas have the property — however, since <br />
generally brodaness the referent of {lo broda} <br />
-- will not have the property brodeness, even if <br />
all the brodas do (or may have it even when all <br />
the brodas lack it). So, for xorlo to work, we <br />
need a different rule the truth of {lo broda cu <br />
brode}. Since we want this to sometimes be true <br />
even when there are no brodas, this new rule <br />
cannot appeal directly to things which have <br />
brodaness or the like. It must rather be <br />
written in terms of the relation between <br />
brodaness and brodeness directly, that there is a <br />
semantic or conceptual overlap. Now this <br />
overlap, to be effective in the way intended can <br />
come about in either of two ways: by conceptual <br />
inclusion (as in {lo pavyseljirna cu <br />
pavyseljirna} and perhaps {lo pavyseljirna cu <br />
blabi}) or by factual overlap (as in effect <br />
happens in the usual way of doing way of <br />
evaluating truth things that fall under one <br />
property fall also under the other).<br />
<br />
Having introduced this intensional definition of <br />
truth for {lo} expressions, we need either to <br />
find a way to apply it to all other sumti <br />
expressions or to recognize that there are two <br />
radically different definitions of truth involved <br />
here, where there was only one with old {lo}. It <br />
seems clear that we cannot make the intensional <br />
definition work with many sumti expressions or <br />
can do so only with the introduction of <br />
considerable ad hoc complexity. Variable by <br />
their nature do not fall under some concept to be <br />
used. Neither do names (unless you have named <br />
so-and-so as a concept, which seems to be <br />
straining the notion of concept a bit). And <br />
{le} expressions are specified exactly by their <br />
referents, not by any concept (unless again is <br />
called such-and-such is a concept and even that <br />
has content only by pointing to the intended <br />
individuals). So, it seems we must recognize <br />
that we have a more complicated and doubled <br />
conceptual apparatus with xorlo. Of course, this <br />
makes no practical difference in using the <br />
language, only in theory; so, if it brings about <br />
some simplification in use, it is probably <br />
acceptable.<br />
<br />
As noted earlier, one change that is achieved is <br />
only a theoretical advance: the various things <br />
that can be said about nonexistents work as well <br />
in prelo as in xorlo and, at least in some cases, <br />
the same hidden devices are needed: elided modals <br />
or the like. As for the intensional cases, we <br />
still need {tua} for those desires that are <br />
referred to by other expressions than {lo}s: {mi <br />
djica le mikce} or {mi djica la djinis} or even <br />
{mi djica da} all immediately have as they do <br />
in prelo the reading equivalent to the fronted <br />
form: {le mikce zou mi djica my}, {la djinis <br />
zou mi djica dy}, {da zou mi djica da}, even <br />
when this is not meant (Jeanie may be only in my <br />
dreams as may be the doctor that I want and I can <br />
obviously want something without there being <br />
something I want). And, lacking {tua} with {lo} <br />
we are deprived of the possibility of quantifying <br />
out that is available for other sumti: I cannot <br />
say directly that the thing I want is something <br />
recognizable even separate from my wants. {mi <br />
djica lo broda} = {lo broda zou mi djica by} in <br />
prelo but not in xorlo, since the special reading <br />
of application in terms of property overlap <br />
applies only to {lo} expressions and {by} is not <br />
one, but rather is merely coreferential with one <br />
and so has me wanting a property. Since, this <br />
problem will arise with every anaphora of a {lo} <br />
expression} we can expand our rule of <br />
interpretation to cover such anaphora as well, <br />
although this requires that we always know what <br />
sort of an expression is being anaphorized <br />
(perhaps, to keep this all formally correct, we <br />
should require different anaphora for {lo} <br />
expressions so we can always tell). In a similar <br />
way, we need special variables for generalization <br />
of {lo} expressions. Otherwise the <br />
generalization from {lo pavyseljirna cu <br />
pavyseljirna} would be {da pavyseljirna} which <br />
says there are unicorns as things, not just about <br />
the overlap of properties. This is all to <br />
complicate the language and either put strain on <br />
the cmavo space or complicate the structure of <br />
pronouns and variables. Or, of course, we could <br />
ignore these problems and regularly use <br />
expressions that are ambiguous at a fundamental <br />
logical (and ontological) level.<br />
<br />
So far, I have seen no concrete suggestions about <br />
how either to deal with these problems or to <br />
circumvent them. On the whole then, I think it <br />
would make sense to do away with xorlo and return <br />
to prelo. The one change that this makes at a <br />
practical level (aside from not needing the <br />
duplication of pronouns and variables) is the use <br />
of {tua} in all cases except those where the <br />
existence of a particular object is stressed and <br />
this was already a part of Lojban. And the <br />
theory is greatly simplified, with all sumti <br />
functioning in the same way the way we <br />
naturally think they function anyhow.<br />
</div></div>Gleki