Re: SUO: Definition of life
Eric --
I do think there are some important distinctions to be made here. I
believe they are uppermost, from an ontological perspective. If that means
they are "upper" in the sense of SUO, is another question.
My thinking over the last couple of years has gravitated toward the notion
of system as perhaps the most fundamental concept in my personal view of
the world. On other occasions I have posted a list of close to 100
variants on the concept of system, but for the moment I am happy with Ross
Ashby's definition, which is "... any set of variables that is selected
...". This is a good working definition because it focuses on the other
fundamental concept in my own uppermost ontology, which is purpose.
What I am thinking is that a system is any of the many variations as noted
elsewhere, having as characteristics identifiable parts that are
interrelated and subject to some form of behavior. Parts, relatedness, and
behavior are all intrinsic to any system. Based on the Ashby definition,
it is axiomatic that the system itself is identified by an observer, based
on some purpose of the observer. The purpose of the system is then further
identified in relation to the purpose of the observer in calling it out
from the surrounding background.
Terms like "final causation" and "teleological explanation" have ring to
them, that in my mind conjurs notions of pre-ordained cosmic design. This
is a level of purpose that could be ontologized, and if it's lurking in
unspoken axioms, it would be good to bring it out explicitly. In fact, an
ontology of purpose would be valuable exercise. John's note below brings
up some further aspects of a full rendering of the meanings of purpose,
such as his discusssion of goals and subgoals in a chess match. His
attribution of goal and subgoal respectively to final causation and
efficient causation seems wrong to me, if efficient causation is also
defined as nonliving. Clearly all moves in chess are the results of living
entities, so there seem to be several issues that need to be sorted out
here.
By the way, you put an interesting twist on the ontology of purpose when
you say "useful for our purposes". Purpose of ontology?
And one more "by the way": neither system, nor purpose, are concepts
ontologized in SUMO.
Doug McDavid
IBM Business Innovation Services
Member, IBM Academy of Technology
mcdavid@us.ibm.com
"Imagine all the people ... living life in peace."
Erik Larson <elarson_78746@yahoo.com> on 06/01/2002 03:46:34 PM
To: Douglas McDavid/Boulder/IBM@IBMUS
cc: cg@cs.uah.edu, standard-upper-ontology@ieee.org
Subject: Re: SUO: Definition of life
Doug,
I might be misreading John, but I thought he was just committing himself to
a definition of life as a 'triadic' relation ('X does P to achieve R') as
opposed to the dyadic relations of instances of efficient causation. That
would seem agnostic as to the notion of "purpose" he had in mind--certainly
it wouldn't imply something like a theistic worldview, I would think.
One more thing: for the purposes of building ontologies I think we
can avoid worrying about whether our concepts square with a worldview,
e.g., whether a notion of teleology that yields a good definition of living
things might be in conflict with scientific naturalism, etc. People, after
all, disagree about these things, and we shouldn't worry about taking sides
(it shouldn't be a criteria for adoption in an ontology at any rate). For
instance, I think that the concepts of so-called "folk psychology"
--beliefs, intentions, desires, etc--will do just fine for
ontologizing aspects of the mind, even though if you are a die-hard
naturalist (or I should say an "eliminative materialist" along the lines
of Churchland, etc) you'll think all of these concepts are metaphysical
illusions. But as I say, that won't matter much in an ontology. After
all, most people do believe th! ese things (they are not eliminative
materialists), and obviously people qua inferencing systems have done quite
well (better than AI systems) in predicting what other people are up to and
why, etc. So if the point here is to get a def. of life to distinguish it
from non-life, if John did suggest your second interpretation of
"teleology" (and I'm not sure he did--but that is for John to clarify), and
it's useful for our purposes, then so be it.
Erik
Douglas McDavid <mcdavid@us.ibm.com> wrote:
John --
I think you're conflating two senses of "purpose" in what you say below.
One is the purposefulness of the amoeba in pursuing its survival-oriented
life processes. The other is purposefulness with respect to the very
existence of the amoeba, as in fulfilling some grand plan or design. The
second meaning is what should more properly be termed teleological, and
scientists are wise to avoid arguments based on ultimate purpose of some
omnipotent designer. My preferred definition of life is autopoiesis, as
defined by Maturana and Varela. Their definition explicitly rejects
teleology, as I'm using it. I'm afraid your two heroes, Peirce and
Whitehead, build on a shaky foundation in this regard, given the prominence
of the concept of "God" in their arguments.
Doug McDavid
Delivery Excellence Team
IBM Business Innovation Serv! ices
Member, IBM Academy of Technology
mcdavid@us.ibm.com
"Imagine all the people ... living life in peace."
"John F. Sowa" @majordomo.ieee.org on 06/01/2002 11:26:38
AM
Please respond to "John F. Sowa"
Sent by: owner-standard-upper-ontology@majordomo.ieee.org
To: cg@cs.uah.edu, standard-upper-ontology@ieee.org
cc:
Subject: SUO: Definition of life
One of the discussions that has erupted on these mailing lists from
time to time is the question of Peirce's Thirdness and its usefulness
as a distinctive feature for defining concepts in a formal ontology.
An example that I have used is Peirce's semiotic definition of life.
I don't have the quotation handy at the moment, but in any case,
a brief quotation is not convincing to people who are not already
thinking in Peircean categories. Therefore, I'll give an explanation
in my own words:
! 1. Nonliving causality is based on "efficient causation", which can
be formalized in terms of chains of implications: p1 causes p2,
which causes p3, which causes p4, etc. This chain of causation
can be represented by a collection of ordinary axioms or rules
expressed in terms of the usual material implication:
p1 -> p2, p2 -> p3, p3 -> p4, ...
2. Causality in living beings also involves efficient causation,
but the defining characteristic that distinguishes a living being
from a nonliving being is "final causation" or a "teleological
explanation": an entity X is living if and only if it performs
actions that must be described in terms of a final cause, goal,
or in Greek "telos".
3. Following is the criterion for distinguishing an efficient cause
from a final cause by some being X, which produces some result R.
In efficient causation, X does R.
In final causation, X does something P, which implies R.
What makes this chain of two implications different from a chain
that occurs in nonliving entities is the relative importance of
P and R for preserving and maintaining X's "livelihood". The
result R is something that helps to preserve X, but P is
something that is not directly relevant to preserving X.
4. A syntactic criterion for distinguishing final causation from
efficient causation is the kind of words used to describe it.
In AI terminology (as in game-playing programs), one would say
that R is the goal, and P is a subgoal for achieving R. In
ordinary English, one would say that X does P "in order to"
achieve R. The difference between a goal and a subgoal is
the relative importance: in chess, for example, checkmate is
the ultimate goal, and every other move is a subgoal.
To illustrate this distinction, I would like to cite the following
article, which discusses how an amoeba moves:
http://www.sciencedaily.com/! releases/2002/05/020531072413.htm
Following are three paragraphs excerpted from that article:
"How do cells determine which direction to go to find an attractant?
How do they sense the differences in concentration of the chemical,
alter their membranes and move forward?" asks Peter Devreotes, PhD,
professor and director of cell biology in the school's Institute for
Basic Biomedical Sciences. "It's a very complex puzzle, and we've
found another piece."
Reporting in the current issue of the journal Cell, the scientists
show that a protein called PTEN goes to the back of the cell when a
chemical attractant is sensed, allowing the cell to move purposefully
toward the attractant. Because PTEN "brings up the rear," the
molecules crucial for allowing the cell to reach out and move forward
are restricted to the front of the cell.
"PTEN is found only in the back of the cell in moving amoeba, and is
actually attached to the ! cell's membrane," says Devreotes. "Without
PTEN or without it attached properly, the amoeba couldn't determine
direction as well. Instead of moving in a straight line and adjusting
quickly if the source of the attractant is moved, cells without PTEN
have bigger 'fronts' that tugged them in a number of directions at
once, impeding their progress."
Notice the terminology used in the article: the reporter actually uses
the word "purposefully". The scientist (in the quoted portions) uses
words that are not as obviously teleological, but which have implicit
purposes buried in their definitions: "attractant", "determine",
"properly", "sense", "adjust", and "progress".
By their training, scientists have been "brainwashed" by David Hume and
his followers to avoid teleogical explanations. Unlike the reporter,
they are careful to avoid the word "purpose" or "goal", but they use
disguised synonyms like "attractant" and words like "progr! ess", whose
definition implies some goal, purpose, or "attractant".
One can, of course, explain everything that happens in the amoeba by
purely efficient causes, as the scientist does in explaining the motions
of the proteins and their interactions. Similar explanations can be
given for the chess-playing programs, which are merely executing the
instructions encoded by some human programmer.
Those explanations, however, do not explain *why* the amoeba or the
program has a structure that just happens to be explainable in
teleological terms. For the program, the answer is simple: some
human encoded his or her goals into the sequence of instructions.
For the amoeba (or the human, for that matter), the answer is
"evolution": organisms that perform intermediate steps like P
"in order to" achieve R can more easily preserve their "livelihood"
than organisms that use single-step efficient causation.
That may be true. But that i! s all that I (or Peirce) was trying to say:
The distinguishing characteristic of life in all its forms is the use
of "intermediate goals" to achieve some result that preserves its
"livelihood" or, if you prefer some word that does not use the term
"life", we could just say "that preserves it in a stable state".
That, in short, is characteristic of Thirdness: the essential use
of an intermediary that mediates between two other participants.
Efficient causation is a sequence of dyadic relations, but a
teleological explanation always involves an irreducible triad:
an agent, a goal, and a means used by the agent to achieve the goal.
The technical term, which scientists use to disguies their anti-Humean
assumptions, is "homeostasis", which simply means "preserving the same
state." But of course, that word "preserve" implies that there is a
purposeful (or at least preferred) state that is worth preserving.
Bottom line: There is! no way to define life or even to distinguish
a living process from a nonliving process without giving a definition
that implies a goal, a purpose, or an intermediate state whose only
reason is "homeostasis", which assumes a teleological explanation.
John Sowa
Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup