Dear Hans,
Thanks for your reply. Please see my
responses embedded into your email below,
-Rich
Sincerely,
Rich Cooper
EnglishLogicKernel.com
Rich AT EnglishLogicKernel DOT com
9 4 9 \ 5 2 5 - 5 7 1 2
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Hans Polzer
Sent: Friday, March 09, 2012 12:40
PM
To: '[ontolog-forum]
'
Subject: Re: [ontolog-forum] Constructs, primitives, terms
Rich,
I think it would be
better not to use terms like “semantic baggage”, which suggest some
lack of objectivity on the part of whoever defined C.
I can’t take credit
for that term; David was the first to use it, but I heartily agree with the
term because I firmly believe that the originator(s) of any ontology
necessarily exercised their subjective cache of beliefs to create the
ontology. That is really my point in this discussion. People are
always subjective, even when pretending toward the objective view. We see
our world through the glasses of a subjective experience from birth to
now. Objectivity is what we call it when people don’t disagree much
on a concept. That is why people can agree on very simple ontologies such
as Dublin Core, but not on more complex ontologies.
For example, the new health
coding system for doctors to use in seeking reimbursement has 140,000 codes,
each with an English description. Here is an article describing one point
of view on how that change from the present 18,000 codes to the new 140,000
codes that distinguish more detail on health conditions. For example,
according to this author, there are 36 different codes for treating a snake
bite, depending on the type of snake, its geographical region, and whether the
incident was accidental, intentional self-harm, assault, or undetermined. The
new codes also thoroughly differentiate between nine different types of
hang-gliding injuries, four different types of alligator attacks, and the
important difference between injuries sustained by walking into a wall and
those resulting from walking into a lamppost:
http://www.amazon.com/forum/politics/ref=cm_cd_dp_rft_tft_tp?_encoding=UTF8&cdForum=Fx1S3QSZRUL93V8&cdThread=TxBWPF3HLJGNLL
At the risk of
getting into a discussion of Plato, the key point is that every definition of
C, C’, and C”, are based on some context (often assumed and
implicit), some frame(s) of reference for describing entities/concepts within
that context, and with specific (if often implicit) scope, and from some
perspective upon that context. Until we have a shared language for describing
context, frames of reference, their scope, and the perspective from which the
context is described, we will always have variations in definitions of C,
C’. and C”. Indeed, there will be as many variations of C as there
are context dimensions and scope values for those dimensions as might have a
material influence on the definition of C.
What is the cost of
developing these shared codes, i.e., training every physician to use the “proper”
code for each condition of the 140,000 distinct codes? The context is
captured in more detail than ever before, but will the codes be “properly”
assigned? I doubt it. What is the motivation for the physician to
distinguish states of mind, such as the difference between assault with a
snake, versus self injury with a snake, versus accidental discovery of a snake
engaged in biting the patient?
Let’s call this ontology
of 140,000 codes, each code j corresponding to one context C[j], the health
code ontology. But what physicians will actually memorize, discriminate
and record health conditions correctly within this 140,000 health code ontology
forest? The subjectivity of the reporting physician will be superimposed
on this ontology, and the data that is actually recorded will not truly match
the code. That mismatch is the “semantic baggage” mentioned
above.
Furthermore, the original
developers (probably a committee) who created the 140,000 code health ontology must
have debated and reconsidered their codes many times to reach the complexity of
140,000 codes. But why stop there? I am sure there are other, more
specialized contexts which could be coded, and also more general, aggregated
contexts that could be matched against a vector of those codes. The
choices they made to reach the specific ontology of 140,000 codes were due simply
to their collectively subjective judgments leading to a result by the project
deadline.
Which brings up
another important point, namely that of purpose of the definition, or of the
concept/entity being defined, modulo the above discussion. The purpose of the
definition is what determines whether a context dimension is material or not.
If the differences in definition of C and C’ do not alter the
intended/desired outcome for some purpose (or set of purposes over some context
dimension scope ranges), then they are functionally equivalent definitions in
that context “space”.
A “purpose”
is by definition subjective. I suspect that the committee making up the
140,000 codes in the health ontology considered some attributes of the health
care situation, though even at that rich level of predication, The considered
attributes couldn’t possibly describe every situation into which a patient
can find herself disposed. Yet I doubt very much whether the committee members
all truly agreed with the attributions made on health contexts. More
likely, the chairman, or manager, or director, or pick some other title for the
alpha leader, overruled some which she considered outlandish, added some on
which she alone insisted, and broke ties among committee members to reach a
politically acceptable consensus for her own context of working on the project
to get results which satisfy her and her bankers.
So I still believe that
there is a C- context, not just a C, C’ and C” context, which has
to be considered. Large ontologies such as the health ontology above
absolutely require politically acceptable contexts in which to operate.
That context is the C- (if it could only be described by a perfectly objective
uninvolved and unconflicted observer who would probably fall asleep designing
the ontology since there would be no motivations for such an objective,
uninvolved, unimpacted and unaffected observer. That is why those
observers don’t exist.
This is the
pragmatic aspect of “common” semantics, which many on this forum
have brought up in the past. Commonality is a meaningful concept only if one
specifies the context “space” (i.e., the range of context
dimensions and scope attribute value ranges for each dimension in that
“n”-space) over which the concept or entity definition is
functionally equivalent among the actors intending to use that definition for
some set of purposes.
Again, those aren’t
“actors”, those are subjectively motivated and peripherally
inspired politically directed participants seeking the implementation of their
own aspects of the ontology which they individually feel are important to them
and perhaps to the people they represent. The rest of the ontology, each “actor”
must feel, can do whatever they want with it. That doesn’t make all
committee members motivated to apply the 140,000 codes in toto, just in the
areas they want to measure.
The NCOIC SCOPE
model is an attempt to define such a context space and scope dimensional
“scales” so that two or more systems can determine whether they can
interoperate correctly for their intended purposes. Note that semantic
interoperability is only a portion of the SCOPE model dimension
set. Conversely, the SCOPE model is explicitly limited in scope to interactions
that are possible over a network connection. It does not address physical interoperability,
for example.
It doesn’t seem to
me that network communications are as significant as representational
divergence. The 140,000 codes will not be interpreted in the same way by
all physicians, most of whom will only worry about what has to be reported so
they can get reimbursed. Physicians are already over managed and
overregulated; they don’t even have time to talk to patients much any
more. At most, fifteen minutes goes toward listening to the patient and
giving a prescription or a referral.
Analysis of the data
force fitted into this 140,000 code ontology will be based on what little
familiarity each physician has to gain about the codes in his specialty
area. Yet all kinds of statistical analysis, classifications, inferences
and abductions will be drawn from databases containing signs entered into
databases purportedly in compliance with the ontology.
Hans
-Rich
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Rich Cooper
Sent: Friday, March 09, 2012 1:41
PM
To: '[ontolog-forum]
'
Subject: Re: [ontolog-forum] Constructs, primitives, terms
Dear David,
You wrote:
… In this
example, the terms as used in C' and C'' are effectively specializations (via
added constraints) of the term in C. To transmit a C' or C'' thing as a C
thing is a fair substitution; but to receive a C thing as a C' or C'' thing
does an implicit narrowing that is not necessarily valid.
…
In practice, though, such
an understanding of the differences (or that there are differences) among similar terms as used in C, C' and
C'' often comes out only after a failure has occurred. In real-world use of any
sort of language that does not have mechanical, closed-world semantics, that
potentially invalid narrowing is not only unpreventable, but is often the
"least worst" translation that can be made into the receiver's
conceptualization. Every organization and every person applies their own
semantic baggage (added constraints) to supposedly common terms; said
"local modifications" are discovered, defined and communicated only after a problem arises.
Your analysis seems promising, but I
suggest there is at least one more complication; the description of C must also
have been loaded with the “semantic baggage” of the person who
defined it, just as C’ and C” and therefore C seems likely to also
be a specialization of some even more abstract concept C- which may not have
contained the baggage of C, C’ or C”.
There is no pure abstraction C- in most of
the descriptions for concepts so far as I have seen in our discussions.
Every concept seems to have been modulated by the proposer’s semantic
baggage. Since it is always a PERSON who produces the conceptualization C
in the first place, it isn’t possible to be that abstract.
-Rich
Sincerely,
Rich Cooper
EnglishLogicKernel.com
Rich AT EnglishLogicKernel DOT com
9 4 9 \ 5 2 5 - 5 7 1 2
On 3/5/2012 9:08 AM, John F. Sowa wrote:
Base vocabulary V: A collection of terms defined precisely at a level
of detail sufficient for interpreting messages that use those terms
in a general context C.
System A: A computational system that imports vocabulary V and uses
the definitions designated by the URIs. But it uses the terms in
a context C' that adds further information that is consistent with C.
That info may be implicit in declarative or procedural statements.
System B: Another computational system that imports and uses terms
in V. B was developed independently of A. It may use terms in V
in a context C'' that is consistent with the general context C,
but possibly inconsistent with the context C' of System A.
Problem: During operations, Systems A and B send messages from
one to the other that use only the vocabulary defined in V.
But the "same" message, which is consistent with the general
context C, may have inconsistent implications in the more
specialized contexts C' and C''.
My thinking began similar to what Patrick Cassidy wrote. In this example,
the terms as used in C' and C'' are effectively specializations (via added
constraints) of the term in C. To transmit a C' or C'' thing as a C thing
is a fair substitution; but to receive a C thing as a C' or C'' thing does an
implicit narrowing that is not necessarily valid.
In practice, though, such an understanding of the differences (or that there are differences) among similar terms as
used in C, C' and C'' often comes out only after a failure has occurred.
In real-world use of any sort of language that does not have mechanical,
closed-world semantics, that potentially invalid narrowing is not only
unpreventable, but is often the "least worst" translation that can be
made into the receiver's conceptualization. Every organization and every
person applies their own semantic baggage (added constraints) to supposedly
common terms; said "local modifications" are discovered, defined and
communicated only after a problem
arises.
Should we then blame the common model (ontology, lexicon, schema, exchange
format, whatever) for having been incomplete or wrong for the task at
hand? Nobody wants to complicate the model with the infinite number of
properties/attributes that don't matter. You just need to model exactly
the set of properties/attributes that are necessary and sufficient to prevent
all future catastrophes under all integration scenarios that will actually
happen, and none of those that won't happen. Easy! if you can predict the
future.
In digest mode,
--
David Flater, National Institute of Standards and Technology,
U.S.A.