SUO: Re: Lifecycle Integration Schema
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o
LIS. Discussion Note 69
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o
JA = Jon Awbrey
MW = Matthew West
Matthew,
Returning to a few old issues that have arisen anew,
collating an array of related questions, and onward!
MW: Please treat my thoughts as an initial
interpretation/categorisation to open the
discussion about what you meant when you use
these terms. You may have noticed I favour
a direct approach.
The word "direct" has many meanings, of course, and this will require us
approach its semantics with no small measures of caution, circumspection,
and reflection, which is to say, in a less "direct" or "literal" manner.
But let us not bandy words -- anybody who has a place for <abstract_objects>
in his scheme of things, as I know that you do from reading your scheme of
things, and who would have that scheme of things integrated into grander
schemes of things, much less the likes of an IFF or a LOT, has already
lost inncocence on that score, and cannot hope to reclaim it in quite
so ingenuous a manner.
So let me try to say it another way.
We are faced with this pheonomenon, that people use words
like "ego", "number", "quark", "unicorn", and all the rest
of the words they use, in many different ways. Unless you
now propose to recant all that you have been saying so far,
and return to the fold of the 1-lithic faith, this obvious,
persistent, recalcitrant, stubborn fact cannot be denied.
And this evident phenomenon presents us with an array of
problems that simply cannot be wished away, though many
have tried, and will most likely keep on trying to do.
One of these problems can be called the "explanatory" problem:
If it's really as bad as all that, just how do people manage to
communicate at all, which they just as evidentally do sometimes,
to some degree and extent, though of course, far from perfectly?
And, of course, part of the task of intercommunication, between
people, directly, or between their robotic proxies, indirectly,
is to address the current levels of adequacy in this capacity.
So I don't think that there's any way for serious people
to go back on their words here, and once again dismiss
the issue of many meanings for many significant words.
As I observed, most of the ontological schemes that we have the task
to integrate here, in an IFF or a LOT or whatever fits them, contain
a primal division between Abstract Sorts and Other Sorts of some sort.
But the very use of abstract objects involves a measure of indirectness,
not to say indirection, since it interposes a number of abstract objects --
algebras, arithmetics, categories, geometries, lattices, measures, models,
numbers, sets, spaces of every variety, and theoretical structures of all
kinds -- between the world of signs and the world of concrete, embodied,
physical, possible, space-time objects, or whatever descriptive feature
one wishes to attach to this latter order of existence.
JA: From what I have seen so far, the system that you are
proposing is undermined in its first distinction by an
"absolute fallacy", that misidentifies subject matters
that need to be expressed in relative terms with subject
matters that can be expressed in absolute terms.
JA: That is, there is no case being made for the
unexamined assumption that abstractness is
an absolute property of things, rather
than a relative property of things.
JA: The preponderance of the evidence that has developed
in these discussions has been that it may even be
a purely accidental or arbitrary attribution.
JA: The only way forward that I can now see is to backtrack to
the unconsidered distinction between absolute and relative
properties of things.
MW: This I do not follow.
JA: A practical test of whether a property of a thing
is a relative property of a thing is that one needs
additional information, beyond that which identifies
the thing, in order to make a decision about whether
the thing in question has the property in question.
MW: For example?
JA: Let me absorb the "absolute fallacy" under the more general
description of a certain type of "reductive error", whereby
one attempts to use a data model with too few dimensions,
or what amounts to the same thing, a relational model of
too low an arity to preserve the appearances or resolve
the anomalies of the phenomenon or problem in question.
JA: We are presently immersed in a poignant example, namely,
the acute situation of uncertainty that devolves from a
disagreement about the classification of terms, that is,
in effect, a disagreement about the application of some
terms to other terms within a discursive universe, mere
strings of syntax as well they may be, or, if it's safe
to say that the terms in the second set denote commonly
identifiable objects in our universe of discourse, then
about the classification of the things that are denoted.
JA: Table 1. Situation Of Uncertainty
o---------o----------------------o----------------------o
| <thing> | <abstract_object_JA> | <abstract_object_MW> |
o---------o----------------------o----------------------o
| ego | 1 | 0 |
| number | 1 | 1 |
| quark | 1 | 0 |
| unicorn | 1 | 0 |
o---------o----------------------o----------------------o
JA: One hypothesis about what has gotten us into difficulties here is that
we are taking it far too much for granted that there just has to be an
objective 2-adic relation, indeed, the type of objective function that
we know as a classification, that is the only acceptable model for the
underlying reality of the situation. Taking this for granted, whether
wittingly or unwittingly, we may fail to give due consideration to the
possibility that no such 2-adic model will fit the reality of
the case.
MW: Well naturally we have different concepts that
we happen to have given the same label to, and
now you have developed distinguished labels.
This is just normal. I really don't see
the big deal.
JA: Let !A! be a discursive universe, for now just a set of words, that
includes the terms "ego", "number", "quark", "unicorn", and let us
assume that each of these general terms has a non-empty extension
in some universe of discourse X, that is, a collection of things.
Furthermore, for the sake of simplicity, let us observe the rule
that all of the predicates that are listed in the column heads
of the Table apply equally well to all of the things that are
denoted by the terms that lead off the rows. That is to say,
any instance of an ego, number, quark, or unicorn represents
its class impartially with respect to the application of the
predicates abstract_object_JA, abstract_object_MW : X -> B.
JA: Now, if by some sporadic confluence of the stars, our two random observers
chanced to compute the very same function under the stem-and-leaf variants
abstract_object_JA, abstract_object_MW : X -> B, then we'd be justified in
reducing the data of the situation to a single classification, proposition,
or indicator function, convening the common stem, abstract_object : X -> B.
But until that happens, we have no choice but to save what's given, namely,
the full 3-adic relational data, or a model that preserves its information.
MW: And. There is nothing to stop you having your concept if you wish.
You can even give it the same name if you wish. You can even say
that it is your usage of the term, as distinct from the LIS usage
of the term. So just what is the problem?
"What's in a name?" you say? Shall we do the experiment of trying to
market software and hardware under the names of "MS_JA" and "PC_MW"?
I think that we already know the outcome of that gedankenexperiment.
Shall we ask our candidate ontologies to subscribe to the minimal
provisions of "truth in labeling" by always using subscript terms
like "abstract_object_LIS, "organization_OpenCyc", or "Set_SUMO"?
Been there, done that. By way of an explanation for the diehard --
no relation to any trademark of Sears & Roebuck, Inc. -- battery
and assault of silence that I've always received for suggesting
wild notions like accurate advertising, I can only guess that
it just plain goes against human nature and market strategy
even to entertain the idea. So let's not mince words here,
as nobody's really buying that pie. A standard ontology
of any brand is either a recommendation for general use,
or it is nothing.
Jon Awbrey
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o