Perplexed in Peoria <
[email protected]> wrote:
> "John Wilkins" <
[email protected]> wrote in message
>
news:[email protected]...
> > Perplexed in Peoria <
[email protected]> wrote:
> >
> > > "John Wilkins" <
[email protected]> wrote in
> > > message
news:[email protected]...
> > > > Perplexed in Peoria <
[email protected]>
> > > > wrote:
>
> > > > > [Regarding the definition of "analog" and
> > > > > "digital"] [The] main thrust of my response to
> > > > > John was that any discussion on these topics has
> > > > > to take place in the context of Shannon's
> > > > > communication theory. That means:
> > > > >
> > > > > 1. That information theory inevitably carries a
> > > > > dose of teleology - it is an engineering
> > > > > discipline, not a branch of pure descriptive
> > > > > science. Of course, Nature (natural selection)
> > > > > is also an engineer. A central part of this is
> > > > > Shannon's idea of the active channel.
> > > >
> > > > Shannon began his classic paper by observing that it
> > > > didn't matter what the meaning of the message was in
> > > > his theory, so long as the stream received was the
> > > > stream sent. I don't think this is in any way a
> > > > teleological theory, and I believe you are
> > > > overinterpreting. It applies nicely to a
> > > > nonteleological system (for example, cell--cell
> > > > signalling).
> > > >
> > > As it happens, my lost post anticipated and dealt with
> > > this response. You need to imagine three levels here.
> > > Meaning reduces to information, which in turn reduces
> > > to physical state. Shannon deals only with the lower
> > > two levels, but the information level retains a
> > > modicum of teleology. Here is the quote from Shannon
> > > that you referred to, with my EMPHASIS added:
> > >
> > > The fundamental PROBLEM of communication is that of
> > > reproducing at one point either exactly OR
> > > APPROXIMATELY a message selected at another point.
> > > Frequently the messages have meaning; that is they
> > > refer to or are correlated according to some system
> > > with certain physical or conceptual entities. These
> > > semantic aspects of communication are irrelevant to
> > > the ENGINEERING PROBLEM. The significant aspect is
> > > that the message is one selected from a set of
> > > possible messages. The system MUST BE DESIGNED to
> > > operate for each possible selection ...
> >
> > That there is a design problem in Shannon systems is a
> > fact about the context in which Shannon developed the
> > maths. It has no more teleology in it that game theory
> > needs rational egoists to make sense. In a *telegraphic*
> > system there is an element of intentional design,
> > because that is how telegraphic systems get made and
> > why. But application of Shannon theory to a biological
> > system does not involve teleology, any more than the
> > application of game theory to genetics or evolution
> > requires reflective self-interest-maximising agents.
>
> I am clearly using "teleology" in a broader sense than you
> would prefer. More on this below. But in response to this
> specific point, I would claim that any analysis of an
> information theoretic nature must take into account that
> there is some goal, purpose, or function involved in the
> preservation or transmission of the information.
> Otherwise, the "information" is not really information, it
> is just an interesting pattern of noise.
And this is *entirely* a matter of how it is described by
the observer, so is it any wonder that there is some
intentional aspect of the description? But physically, any
system that can be described using (in this case) Shannon
maths is an "information-transmitting" system, no matter if
there is intention in it or not.
The mistake is to infer from our use of intentional
metaphors that anything we can describe in that way must be
intentional (which is what we have when we infer from the
intentionality of a Shannon-designed system to any other
natural system that any Shannonesque system must be
intentional also). It is simply projection.
"Noise" in one formalisation can be a "signal" in another -
consider using a different frequency of a TV signal to
carry, say, the stock reports. The picture may be degraded
by that interference, and be noise relative to the TV
signal. What is noise depends on the uses made by humans, or
on descriptions of processes made by humans.
>
> Similarly, any application of game theory involves some
> kind of objective function or goal of the agents - that is
> why the key parameter is called a "payoff". The same math
> can be used on other kinds of systems; ones that don't
> involve goal-oriented (I'm trying to avoid "directed" or
> "seeking") agents. But then it is no longer "game theory".
The maths is identical. Game theory is the maths. You
surprise me, Jim; I thought you would not make that
sort of claim.
>
> > > However, it must be that you use the word teleological
> > > differently than I do. Cell-to-cell signalling is
> > > quite teleological, as I use the word, assuming of
> > > course that this signalling is an adaptation created
> > > by natural selection to serve some function. Your
> > > reductionist tendencies (or perhaps tendencies toward
> > > Cartesian dualism) may be worse than I thought.
> > > Natural selection EXPLAINS the appearance of design -
> > > it does not dispell it. To my mind, NS restores the
> > > respectability of teleology in biology - it doesn't
> > > demand that teleological language and forms of
> > > explanation be suppressed.
> >
> > I think you need to read up on the way teleology has
> > been dealt with since Mayr and Pittenrigh <sp?> worked
> > out the notion of a teleonomic system.
>
> I thank you for directing my attention in that direction.
> "Teleomatic", and "teleonomic" seem to be useful terms.
> However, in the course of researching this, I find that
> "teleological" is used in two senses. In the broad sense,
> it is a subject area that encompasses all of the "teleo-"
> words. In the narrow sense, it means "goal-directed" and
> seems to carry a connotation of consciousness to some
> commentators. More on the narrow sense below, but here is
> a link to a broad sense use:
>
http://www.hku.hk/philodep/courses/ac/2011telogic.html
I find that biologists often do not quite get teleology,
which is why they so often return to it. But you are being
misled by words, that's all.
I wrote this once before, if I may quote myself here:
"Quite generally in biology, teleology appears to have been
decisively rejected, despite the resurgence of so-called
"systems" theory in popular environmentalism. Mayr makes the
point that goal-directed thinking in biology is rarely
teleological in the sense of Teilhardian Omega Points or
inevitable progress, but rather is functionalist. O'Grady
and Brooks (1986), following Mayr's initial (1982)
discussion, distinguish between three types of "end-
attaining" activity: teleomatic ("end-resulting", referred
to in the Figure as "lawlike"), teleonomic ("end-directed")
and teleological ("goal-seeking"). Generally, philosophers,
under Aristotle's influence, assume the teleological
approach. They nest the three kinds into each other (Figure
right side). The idealist (Kantian) conception of purpose
according to Mayr's characterisation is exactly the inverse
of the Darwinian, and as a result, evolutionary change since
Lamarck has often been seen as necessarily directed to an
end, since there is an over-riding purpose to all facets of
nature. Thus, purpose is the more general set, including the
other two in reverse order. Still, the only process that is
undeniably teleological, in the sense of being actively goal
seeking, is the cognitive process of higher animals. Other
processes may be apparently or arguably goal seeking, but as
with many neo-Kantian doctrines, this one often puts the
cart before the horse. Darwin overturned that cart, and
denied that because we need to see purpose in processes,
that purpose was required to explain those processes. Under
a darwinian model, according to Mayr, purposive behaviour is
a special case of general functional processes (Figure left
side). This satisfies my own biases and intuitions.
Darwinian conception Idealist conception
teleomatic---------lawlike processes----+ +-----teleological
| |
includes | | includes
| |
teleonomic---------functional processes----------
teleonomic
| |
includes | | includes
| |
teleological-------cognitive processes-----+ +--------
teleomatic
The significance of this for a general theory of evolution
is that it resolves many of the confusions that arise when
biologists discuss functional systems - where it is more
than merely convenient to refer to the processes of a system
as goal seeking. By being able to refer to the state-
maintaining processes as teleomatic, and where necessary as
teleonomic, without implications of agency or purpose,
allows us to make clear distinctions between different kinds
functional systems without anthropomorphic confusion such as
trailed the publication of Dawkins' 1977."
>
>
> > Teleonomy is goal-seeking behavior, teleology is goal-
> > directed behavior. Shannon building a telegraphy system
> > for Bell is teleological, because Shannon is a gaol-
> > directed system.
>
> Surely all of Shannon is not gaol directed - only the
> heavy drinking, football watching portion of the
> population. ;-)
Claude Shannon was a population? Are you going all
Minskyan on me?
>
> > Cell--cell signaling is a teleonomic system because it
> > seeks a particular outcome, but it does so because of
> > decidedly nonteleological processes.
>
> By Jove, I believe that we have Dr. Wilkins committing the
> fallacy of composition, led in that direction by Ernst
> Mayr! Certainly, the ontogenic development of signaling
> apparatus by a cell is teleonomic, rather than narrowly
> teleological. In Mayr's terms, the cell is following a
> "program". Same goes for the use and operation of that
> apparatus. An extreme reductionist might even call these
> events "teleomatic" because we are simply witnessing the
> working out of physical laws.
It doesn't take an *extreme* reductionist to do this.
Ordinary science will be able to call processes teleomatic
even when cells are following (and I note the necessary
scare quotes) "programs". Which is another metaphor...
>
> But what about the *design* of the signaling apparatus?
> That is, what explanation do we give for the fact that
> almost all cells of that type in that species construct
> the same apparatus and apparently use it for the same
> function? In my world-view, a designer can be identified -
> it is "Nature", also known as "Natural Selection".
The use of the term "design" is yet another metaphor.
Evolution doesn't design things - it makes things look to
systems that do recognise design
- us - as if they were designed. Dawkins calls this
designoids, I call it "quasidesign" (where "qua si" means
"as if" in Latin). Selection doesn't design - it
quasidesigns. And the recognition of that design lies
totally in our pattern recognition capacities. And *we*
evolved to both design and recognise design.
>
> Is NS acting teleonomically? Mayr and Wilkins apparently
> think so, but they are reasoning fallaciously by
> composition. Each cell is following a program, but is the
> "species" of such cell types following a program? Is NS
> following a program? I will argue below that it is not.
>
> Of course, our extreme reductionist could claim that the
> whole process is teleomatic at all levels. Such a
> viewpoint is irrefutable, but not particularly useful
> (except perhaps to someone who spends too much time on
> talk.origins, and is motivated to refute the argument from
> design.)
I have larger ambitions than that. I think that intentional
language ought to be restricted to systems that are actually
capable of having intentions. So I was to drop "function" as
well as "goal", "purpose" and "design" from descriptions of
the natural world, and realise that these intentional
aspects of things actually reflect the nature of our models
rather than the nature of the modeled things (unless we are
dealing with intentional systems, like us).
>
>
> > Think of it like this:
> >
> > "Blind" systems include teleonomic systems include
> > teleological systems. The pre-evolutionary view was the
> > exact reverse.
>
> In the course of my research, I was directed to a
> particular FAQ archive containing the following essay:
>
http://www.talkorigins.org/faqs/evolphil/teleology.html
> which seems to be the only source for the idea that
> teleomatic systems include teleonomic systems, which
> include strictly teleological systems. I'm not sure which
> direction of inclusion is more useful, so for the
> remainder of this discussion, I am going to assume they
> are disjoint. (Hence the "strictly", below.)
See above. Besides, the author was just a tyro at the time
he wrote that.
>
> I would define:
>
> A (strictly) TELEOMATIC system is one whose behavior is
> "hardwired". It simply follows physical law. While its
> behavior may be constrained by its surroundings, it is in
> no sense "informed" by the surroundings. Examples might
> include (1) the behavior of a dropped rock, which only
> seems to seek the center of the earth, (2) a mixture of
> chemicals, which only seems to be directed towards
> equilibrium and the maximization of the world's entropy.
>
> A (strictly) TELEONOMIC system is one whose behavior is
> "programmed". It "senses" some aspects of its
> surroundings, and adjusts its behavior in response in a
> way that optimizes some objective function, such that it
> can be said to be goal-seeking. Since such systems very
> likely include non-linearities in their feedback cycles,
> they may have a kind of "memory" and can be said to
> contain a rudimentary model of their surroundings. A
> slightly more traditional word for "telonomic" is
> "cybernetic". Examples of teleonomic systems might include
> (1) the famous rotating pressure regulator for steam
> engines,
> (2) homeostasis, or cell-to-cell signaling, in the context
> of a single organism.
>
> A strictly TELEOLOGICAL system is one whose behavior is
> "intelligent" or even "rational" in a broad sense. Such
> systems go beyond simply maintaining a model of the
> surroundings. They actively construct such models, and
> perform "thought experiments" or physical experiments to
> answer "What if?" questions. They are goal-directed, and
> they achieve those goals, typically, by designing and
> constructing teleonomic/teleomatic systems that will
> achieve those goals. Examples might include (1) Claude
> Shannon, (2) the US NASA, though it is not very good at
> it, (3) an AI system, if such a thing is ever constructed,
> and (4) more controversially, Natural Selection.
>
> Only someone who insists that the "consciousness"
> (whatever THAT is) of the agent and the agent's
> "understanding" of its goal must be part of the definition
> of "teleological" could disagree. (And, to someone of my
> world view, such restrictions would make "teleological" a
> useless word, having no real-world entities to describe.)
> NS is a very hands-on kind of thinker. She models in the
> same medium in which she constructs. Her model of a
> species is a subpopulation of a species. She finds the
> answers to her "What if?" questions by carrying out
> physical experiments on a restricted scale. If her
> experimental designs are successful, then they are rolled
> into full scale production.
>
> So, to summarize, and to return to the original question
> of "analog" vs "digital": these words are terms of art,
> and are properly used only in the case of systems that are
> at least teleonomic. They ought to be used to describe
> strategies used to store or transmit information, though
> they are sometimes used more loosely to say whether the
> message has been selected from a discrete or a continuum
> set. They are almost certainly not used appropriately in
> the analysis of teleomatic systems, which have no use for
> "information". You pretty much agreed with this on another
> thread, unless I misread you.
>
> But we seem to have a new issue - whether Nature is acting
> strictly teleologically when she designs teleonomic
> systems. I am willing to distinguish two different kinds
> of strict teleology, and to relegate Nature's kind of teleo-
> to a lower level that that of you or I. But I don't think
> that Nature is acting strictly teleonomically. In making
> this claim, I am partly drawing on my intuition that NS
> has more "foresight" than she is usually given credit for.
> But even without this bit of heresy on my part, my case
> seems pretty good that NS is not merely teleonomic.
True. It is purely teleomatic...
--
John S Wilkins PhD - www.wilkins.id.au a little emptier, a
little spent as always by that quiver in the self,
subjugated, yes, and obedient. -- Seamus Heaney