Re: Judith's challenge #1 - repost of JJK comment, with indexing

[John Kineman <***>]

Everyone,
JM pointed out that my net ettiquite is bad because I'm not labeling
who's saying what. I apologize, I didn't realize that because in my
email viewer I get all the html indents, etc. here's a re-post in better
style of the last comment. I will try to maintain this in the future.

HP, speaking of whether or not Bob's criticism makes sense:

Howard Pattee wrote:

> ...Bob’s statements like, “Von Neumann confounds simulation (computation) with construction.” Von Neumann was known for clearly formulating this and similar epistemic distinctions and there is no evidence that he ever confounded them.

JK: How about ontologically? could he have confounded them there? See below:

> ...
>
> HP: Here are only three of Bob’s many quotes over the years. You may judge if the charge is repetitive or not. “Indeed, in addition to invalidating von Neumann’s specific argument, we learn that great care must be exercised in general when arguing from purely logical models (i.e., from models pertaining to efficient cause) to any kind of physical realization, such as developmental or evolutionary biology (which, as noted, pertain to material cause.)” (Anticipatory Systems, p. 419).
JK: This says applying a mechanistic model to natural ontologies won't work.

HP:

> “John von Neumann’s well-known discussion of the self reproducing automaton is based on a similar equivocation. Specifically, he argued that Turing’s demonstration of the existence of a universal computer (a mathematical machine) implied the existence of a “universal constructor.” He thought of the latter, apparently, as a material device; his argument was that construction (i.e., the execution of a blueprint) was no less an algorithmic process than computation (i.e., the execution of a program); hence whatever was true of the latter was equally true of the former.” (Causal structures in brains and machines, Int. J. General Systems, 12, 107-126, 1986)

JK:

Again, construction involves ontology - creating something. If that
ontology is not imbedded in the system, then one is applying a
mechanistic model as the explanation of origins. The only way RR's
statement here would be false if if VN believed that nature itself was
complex down to the core - i.e., a complementarity between realization
and formalization. Did he believe that??? If so, then RR was mistaken
about VN and both were implying that the basis for life exists at the
origin of the universe and in its most fundamental elements. In that
case, physically defined material systems "emerged" from this living
essence - we call it complexity so as not to get confused with life form
-- and physics is thus a sub-set of theoretical biology. Is that what VN
believed???

HP:

> “I remark parenthetically that the confounding of simulation (computation) with construction, which lies at the heart of, e.g., von Neumann’s well-known discussion of “self-reproducing automata” arises precisely here and rests entirely on the equivocal and inconsistent harware/software distinctions to which I have just called attention.” (Life Itself,  p. 234)

JK: Again, let me repeat the definition of an "equivocation error." It is
changing the sementics of a term in the middle of a logical argument.
For example (from the Web): "Odd things arouse human suspicion. But
seventeen is an odd number. Therefore, seventeen arouses human
suspicion." It is using a word or phrase with more than one meaning.

So, I suggest that VN was NOT confounding computation and construction
within a strictly epistemological adaptation of the modeling relation,
i.e., within his own model of self-reproducing automata, but it becomes
equivocation if he then claims that that is how reality works (i.e.,
equivocating between epistemological use of the modeling relation and
ontological use of it). In the epistemological use, software is imagined
to be separate from hardware, but is not really so. Hence the results of
reasoning on the basis of that separation is not applicable to actual
complex nature unless you have a way for software to be developed from
semantic inputs that are not part of the code. This is the nesting I
referred to earlier.

HP:

> All I am claiming, and what I will show by direct quotes, is that Bob’s statements about von Neumann’s not distinguishing simulation and construction (or software and hardware, physical laws and syntactic rules, matter and symbols, the two sides of the epistemic cut, etc.)  were not fair. I repeat, this is not a judgment about von Neumann’s models, it is about correctly reporting what a man says.

JK:
I think the misunderstanding is in the vein I have suggested. Recall my
response about Platonism. The traditional use of the epistemic cut is
Platonic, distinguishing between laws and nature. Those laws are "out
there" to discover, and we get them wrong so the modeling relation
doesn't commute and you get an epistemic form of complexity in the
business of science. That's as far as the agreement between RR and VN
seems to go (based on what I've seen quoted). RR's work continues to
distinguish laws from realization but also embeds that distinction in
nature itself. Once that is done the Platonic realm is dispensed with
and the laws themselves can be seen to be emergent from complexity along
with their realizations. This is not explicitly stated, but I believe is
implied by the whole work which leads to this conclusion.

The epistemic cut is not thus dispensed with, but instead fractioned
into components of the system model. That works conceptually, but itself
cannot be reproduced in a computer without simplifying it, because you
can't provide the semantic inputs at each level except by simulation.
Hence, the criticism of the ontological view would be that we still
can't represent the model without destroying its complexity - there are
no complex models, as such. However, it turns out to be an excellent
design principle for building adaptive modeling systems that are
complex, and for preventing the waste of gobs of money on other
approaches that can't work for specific purposes, for example, in
ecosystem management.

In fact my current model is an automata/agent-based model simulation.
But as such it isn't predictive. We thus treat it as a scenario
generator and produce many scenarios to imagine future possibilities. We
may also embed it in a modeling system, with subjective human inputs,
that can be used by humans adaptively with management. The living
component makes it truly complex, but still only capable of describing
nature in scenarios.

Physicists will always, and should always, attempt to get around this
epistemic cut rather than thinking of it as natural. That was the whole
issue in the Copenhagen convention and the discipline tolerated Bohr's
view that uncertainty/complexity is natural only for a brief time. Then
other approaches with hidden dimensions and other worlds surfaced in a
new attempt to axiomatize all of nature. I doubt that will succeed, but
as a scientific discipline physics is obligated to try. In ecology we
are obligated to work with complexity. For that we need to think of the
complexity as fully real and unavoidable, so an embedded complex
ontology is more appropriate there.