[Date Prev][Date Next][Thread Prev][Thread Next]   [Date Index] [Thread Index] [Author Index] 

Re: machines vs. living systems

Tim Gwinn wrote: 
The book, Life Itself, has as one of its primary themes to identify clearly
what the Cartesian machine metaphor means and to show that organisms are not
machines: "I hope to convince the reader, in the course of this the present
work, that the machine metaphor is not just a little bit wrong; it is
entirely wrong and must be discarded." [LI 23] This theme leads up to the
"Central Argument: The Limitations of Entailment in Machines and
Mechanisms".[sec. 9F] It is the limitations therein which lead to the
proposed definition of life in chapter 10 ("a material system is an organism
if, and only if, it is closed to efficient causation"), which definitively
distinguishes organisms from machines.

But, as you said, "Because the term 'machine' is used in a very specific and narrow sense here,..." If the use of the term, here, in this conversation with you, is so specific and different from the normal English sense of the term 'machine', then how does it help distinguish between the objects we normally, in English, refer to as "machines" and the objects we normally, in English, refer to as organisms? If the term is so specific, how is helping in that everyday context?

When you asked, "What I don't understand is this distinction between
"machine" and ... well, "non-machine".  What do you mean by "life is no
machine"?", it sounded as if you must not have had a chance yet to read Life
Itself, since the entire book follows just that theme. If you had read it,
then this theme has either eluded you or not convinced you, including the
reasoning for the definition of 'machine' and what that definition entails,
which crosses three chapters.

Well, first off, I was not talking to Robert Rosen. I did not ask Robert what _he_ meant by "life is no machine." [grin] And given that, it's odd that you would respond this way.

The theme of LI has convinced me that: formal systems can be defined (in a somewhat minimal way) so as to capture what seem to be some essential properties of life.

In arguing that certain formal systems can capture some essential properties of life, RR built upon some preliminary definitions of formal systems he identified with the tokens "mechanism" and "machine". His use of those terms seemed intended to hook to our qualitative intuitions of what those terms meant to map to the objects in the real world that we call "machines". But, at no place that I remember, did he claim that his mathematics actually distinguished actual machines from actual organisms. The distinction is made between the formal systems (which, admittedly, were intended to map to the actual systems). (I read the book about 8 years ago... so bear with me, here.)

When you say: "My original claim was that in order for a formal system that
describes car engines to be practically useful, it must be capable of
expressing complex causality.  In that sense, complexity is _not_ an
adequate classifier for distinguishing machine from organism.", this also
suggests to me that you misunderstand Life Itself, since complexity alone is
not what distinguishes an organism from a non-organismic complex system.
Instead it is based on the proposed definition in ch. 10, which happens to
necessarily entail that an organism is a complex system.

Again, I wasn't conversing with RR. I believe I was talking to Judith at the time I made that statement. Forgive me if I distinguish between the two. [grin] At the time, Judith was making an argument that complexity was a central part of the distinction between machines and organisms. Hence, I responded by saying that complexity is inadequate to the task; so, even if it is _part_ of the distinction, it is not the whole distinction.

And since I was talking to Judith and the other list members, it does not indicate a lack of understanding of Life Itself. It only indicates a willingness to hold a conversation without continuously thumping a book.

In the modeling relation, all that sits on the natural system side is
phenomena. All our models for a natural system sit on the formal side. When
you say:
============
"RR's machines (particular formal systems with particular properties) and
RR's organisms (particular formal systems with particular properties)
are different and I believe I understand that difference."

My question was more about the comments on the list about Venter's work,
which is not about formal systems.  In fact, Venter's work is totally
unrelated to formals systems.  "My question was more about the comments on
the list about Venter's work, which is not about formal systems.  In fact,
Venter's work is totally unrelated to formals systems."
============
this suggest to me that you are unclear about the modeling relation. To even
identify a particular set of phenomena in the natural world as a natural
system means to build a modeling relation between those phenomena and formal
counterparts to them.

All it means (and all I intended it to mean) is that we, as humans, are embedded in the ambient muck. We are not separate from natural systems in _any_ way. When I rub my eyes, press on my accelerator, pick a flower, or make a logical inference, I am _acting_ within and upon the ambient muck.

Dr. Venter, when he creates his new artifacts out of pre-existing artifacts, his creation will be an extant thing out in the ambient muck. Venter is _constructing_. Venter (and Venter's _mind_ including any models that reside within that mind) are all part of the physical, biological, and social process that leads to (indeed, causally entails) the new artifact.

What you are doing, is trying to separate (for good reasons that I wholly support) the stuff that happens "out there" and the stuff that happens "in here". RR explicitly treats this issue as well; but, only to a small extent.

When Venter begins poking and juxtaposing his building blocks, he is _not_ engaged in inferential entailment alone. Yes, he is thinking about the materials he's manipulating and, in such thought, is engaged in inference. But, that whole thinking process is thoroughly, deeply, inextricably, embedded in the causal entailment of the materials and systems he's manipulating.

That larger process is his main concern. Venter would, I suppose, be completely willing to toss out any thought or inferential system at a whim, as long as the actual systems he produces work and survive in some sense. I.e. the _primary_ (and canalizing) driver for his work is _not_ formal systems. It is extent systems.

And that is wholly unrelated to any mental gymnastics you, Venter, or the fly on the wall is doing during the course of Venter's constructive dynamical process.

So, to revisit, my comments do not suggest that I don't understand Life Itself, particularly as a work of literature. (It's true that I may and probably do fail to understand the work... I'm aware of my limitations. [grin] But, the comments you cite do not indicate that.)

When Venter calls some sub-collection of that
phenomena "DNA", this is again relating model(s) on the formal side with
this particular phenomena. Epistemologically, in the Rosennean paradigm, the
scientific investigation of any natural system entails a modeling relation
and the process of the building of formal models of that natural system.
Likewise, engineering of alterations in a natural system (Venter's work
included)  mean a modeling relation between formal models and the the
natural system.

I admit that Venter engages in the act of modeling (at least I hope he does... though it's important to note that he does not _have_ to in order to manipulate an extant system...).

But, the article did not say "Venter is going to present us with a formal system that adequately models bacteria." The article said: "Dr. Venter hopes to become the first to whip up a made-to-order bacterium."

That means his primary purpose is to manipulate actual systems. I suspect he is willing to adopt and discard formal systems like he changes socks, as long as the actual outcome is achieved.

This means that Venter's work is _not_ about formal systems.

RR's work is about formal systems. And although RR does a nice job showing us how his work might be mapped to an engineering effort, he does not do so in Life Itself. I suspect there is absolutely _nothing_ in Life Itself that will help Venter complete his task. (Though I'm sure the mind-expanding ideas in Life Itself could have the side benefit of sparking creative ideas or cautionary tales, helping Venter refine his thoughts and models.)

Rosen's organism's are not "particular formal systems with particular
properties". Organisms are - like every natural system is in the Rosennean
paradigm - defined by their models. For Rosen, the class of natural systems
which qualify as organisms are those which possess a model which meets the
criteria in ch. 10.

I can believe that he (or any reasonable person) would say that a natural system is _described_ by their models. And I believe that RR establishes a distinction between formal systems that model organisms and formal systems that model machines.

But, the distinction is still between formal systems. The only thing that firmly establishes that a model adequately captures its referrent is validation. RR treats this explicitly in LI: "The formal system F is then called a model of the natural system N if we always get the same answer, whether we follow path (1), or whether we follow the path (2)+(3)+(4)." ("always get the same answer" is the relevant part of that sentence.)

If there is a place, where I _disagree_ with RR, it lies in the assumption that we can follow path (1), which means "follow the dynamical process that leads a natural system from one situation to a subsequent one". I doubt that we can actually follow path (1), directly. But, as I understand it, RR wasn't really suggesting this. He was just using convenient language to make his point. He later goes on to describe the actual, indirect, hypothetical process we use.

Regardless, all of his technique is based on formal systems. And any distinction he makes between "machines" and "organisms" lies on the formal systems side. Hence, all he gives us is a classifier for formal systems. "... it is for us to discover the keys, the encodings and decodings, by which they [miracles like prediction] can be brought to pass."

I believe he goes much farther towards the practical usage of such ideas in "Fundamentals..."; but, you didn't accuse me of misunderstanding that book. [grin] So, I'll leave that for later.

All that aside, it is certainly possible for you to present us with some
other definition of 'machine', one which differs from Rosen's, and then draw
some other conclusions accordingly, such that an organism will be a member
of that class of machines as so defined. The preciseness of that definition
of 'machine' will of course be key - how precise (imprecise) the definition
is will determine how meaningful (vacuous) will be the classification of
'organisms as machines'.

I do not have an alternate definition. In fact, I don't even understand how to apply RR's definition to the machines I bump into in my everyday life. And that is why I started this conversation.

I also don't believe imprecision implies vacuousness. RR _could_ have chosen a different word to refer to "machines". He might have called these formal systems (which can be modeled by a system that can only be modeled by simulable systems) "widgets" (or any other neutral term). But, he didn't. He used the term "machine", in part, to harness the precision (or lack thereof) of the English word "machine". Spoken languages have the ambiguity they have because that ambiguity is _useful_. Spoken language tokens are not vacuous, yet they are often very imprecise.

So, again, harkening back to what I said when I started off on this tangent, many people on this list seem to use the English word "machine" when they speak of distinguishing between machines and living systems...
And I was just asking about the source of their vernacular distinction and the confidence they have in their distinctions. I fully accept that the distinction they make is based upon RR's work in mathematics. ([grin] based on the name and description of this email list) But, all the subsequent conversation has conflated the ambient muck "out there" with the mental gymnastics "in here".

--
glen e. p. ropella              =><=                Hail Eris!
H: 503-630-4505                       http://ropella.net/~gepr
M: 503-971-3846                        http://tempusdictum.com