Replies

Well it wouldn’t be if you two would speak something other than Swahili :)

LOLOL!

I’ll try to “sum it up” this way…

There are four different kinds of “causation.” To use an example, the formal cause would be the blueprint for your house. The material cause would be the lumber, nails, etc. The efficient cause would be the construction workers who build it. And the final cause would be the house itself, the reason for the previous three causes.

Since the days of Newton, science has ignored formal and final cause with the assumption that the everything in the universe is a machine that can be understood by material and efficient causes.

Among other things, this allowed them to insist philosophers and theologians stay away to let them do their work.

And their presupposition has been wildly successful for centuries because, with the notable exception of living things, the rest of the universe can be understood as a machine.

Evidently, the scientists always considered biology to be a “special case” – minor in comparison to the rest of the universe – and not really worth their time. The machine presupposition works well in physics and chemistry, so it’s just a matter of time before they can explain life as a machine, too.

The biologists meanwhile didn’t care either. The machine way of looking at things works well enough in the laboratory until people ask inconvenient questions – and besides they can always claim that life is evolution, the historical record itself. Which is to say, it is because here we are (see Anthropic principle.)

Well, enter the mathematical biologists (Rosen and his predecessors) and mathematicians/physicists who dared to ask (vonNeumann, Pattee, Yockey, Chaitin, Wolfram et al) and it becomes glaringly apparent that life is not simply a machine after all.

From Rosen’s outstanding arguments we see there is no (efficient) cause outside of the organism doing the maintenance, repair, metabolizing and building. It’s doing it on its own. And so he has developed a relational biology, a mathematical model looking at the organization itself. And thus Rosen declares that "a material system is an organism if, and only if, it is closed to efficient causation."

That is how he answers the question “What is life?”

His model is not static, the organism doesn’t just sit there dead as a doornail. There is a flow in the organizational model from one element to the next. And that flow involves both encoding and decoding. That is “chasing” in the model. His model is not concerned with time but with the ordering, the flow, the chasing.

The same is true of Shannon’s mathematical model of communications. It is all about the chasing. Information is defined by Shannon as the reduction of uncertainty in the receiver (an element to Rosen’s model) – the chasing, the flow – not the message itself.

My only complaint so far about Rosen’s book is that he did not give enough credit to Shannon even though his theory relies on Shannon’s work.

To compare the two, think of Shannon as a discrete single chase through Rosen’s organization, e.g. it starts with a sender, a message which is encoded and sent through a channel subject to noise whereupon it is decoded and thereby reduces the uncertainty of the receiver. Shannon's has a beginning and an end. It is discrete.

Rosen's is not a discrete instance, his goes endlessly one to another, turning it into a circular model. One flow (input>process>output> to another (input>process>output) seamlessly.

And so, if anyone asks me “What is life?” I will answer them with both.

Under Shannon, that which successfully communicates in nature is alive. If it cannot, it is either dead or non-life. Shannon’s model doesn’t care whether the elements of the model are biological, radios, tvs, computers, non-physical, etc. Thus the Shannon definition applies to biological organisms (nature), alien life forms (cosmos), artificial intelligence (man-made), spiritual beings, etc.

Under Rosen, expanding his above definition beyond the material (nature) - a thing is alive if it is closed to efficient causation. Which is to say, the thing doesn’t need an outsider to do maintenance, repair, etc.

Because of this, Rosen’s definition rejects artificial intelligence and thus has been criticized by some in that camp. It also arguably would only recognize God as having Spiritual life in Himself (as the Scriptures say.)

The two models are not mutually exclusive. Which one I emphasize in a debate will probably depend on the subject matter.

The Shannon model has a track record in pharmaceutical and cancer research. The Rosen model is just now getting some attention and its application is also reaching to physical cosmology (Fineman et al.)

Did that help?

I think that Rosen was stating that living systems aren’t to be compared to mechanical or software due to ‘chasing’ which occures when the programs search for the best answers (and protect the ‘species’ or program by isolating it from errors that would otherwise affect living systems??) is this correct? If so, that’s about the bottom line of things as GA’s certainly can’t mimic living systems, and end up artificially protecting the ‘evolving’ systems, and artificially introducing elements not seen in nature such as far greater rapid development minus the negative effects seen in nature.

I think what Rosen is saying is that living organisms can't meaningfully be compared to mechanisms or machines (though all are material systems in nature) because the latter are "simple" systems (i.e., essentially reducible to their algorithms) while the former are "complex" systems (i.e., not ever so reducible — the causal structure of such systems simply isn't "algorithmic" or "computable").

Rosen's idea of causality comprehends a system of entailments, which relate the events and phenomena occurring within.

The "chasing" behavior A-G pointed out in the context of the machine model is perhaps just Rosen's way of describing the moving input–output character of machine processing. Rosen expresses such notions with relational diagrams that show the system of causal entailment that models a given particular system. The mechanistic models and the machine models he shows all demonstrate situations of "paucity" of causal entailment. That is, there is insufficient causal basis in such systems to account for such real phenomena as life and mind. That's putting it in a (very abstract) nutshell! :^)

Another way to put it is to say all material systems in nature are studied by science (as presently constituted) on the assumption that they can exhaustively be explained in terms of three of the Aristotelian causes: formal, material, and efficient. It is postulated: There can be no final cause!!!

Yet the relational diagrams of living systems are rich with "final causes." The very idea of biological function is related to the idea of final cause. So Rosen argues, science must put the final cause back into its method if it wants to deal with issues of life and consciousness (mind).

In particular, he believes this would be essential to any understanding of the dualistic genotype–phenotype relation, which essentially involves a "cause–effect" relation between two incommensurable phenomena.

The restoration of final cause to science is a suggestion most strenuously resisted nowadays, by physicists and biologists alike.

The reasons given for abolishing formal causes from science: (1) they are not "objective"; (2) they are not computable (i.e., reducible to an algorithm); (3) worst of all, they clearly point to a type of cause which is "anticipatory" in some sense, and this is forbidden by the Newtonian Paradigm, which demands that causation must always flow from past to future.

The book Life Itself largely deals with the inadequacy of contemporary physics to deal with the questions, "What is life?" and "What is mind?" Indeed, you pointed to one of the main strategies — mimesis — WRT your comment regarding "GAs." (I agree with your conclusions there.) The other main strategy (and the more common) is reductionism. Here's a funny story Rosen tells (in "Mind as Phenotype," Essays on Life Itself) about the reductionist strategy:

Many years ago, I heard a routine of Woody Allen that bears on exactly this point. As he told it, he acquired a Rolls-Royce while in England and wanted to return with it to the States. On the other hand, he didn't want to pay the duty on it. So he hit upon the idea of disassembling it, packing the parts into many suitcases, and describing them to the customs inspectors as modern sculpture, not dutiable as art. He was successful, got his many suitcases home, and proceeded to try to reassemble his car. In his first attempt, the parts yielded 200 bicycles. On the second attempt, he got many lawn mowers. And so it went; he never could retrieve the car.

I won't further belabor reductionism here, except to say that what is "lost" in any such reduction is precisely information about how to reconstruct the system to restore it to its original form. No study of the "parts" can give one any notion of this. Analogically speaking, you have the elements of the genotype (i.e., its parts, the genes); but you have no clue how to reconstitute the phenotype (the original phenomenal system and its behavior).

Anyhoot, we see reductionism most clearly today in the way biochemistry and microbiology are usually conducted.

Regarding mimesis [the basic strategy of artificial intelligence and artificial life studies]: What is involved here is to replace the system of interest by some behavioral or phenotypic mimic, then to study the mimic to see what you can learn about the original system.

Case in point: the typical assumption that a mimic capable of demonstrating a sufficient number of "thinking behaviors" is actually, in fact, "thinking." What we learn about its "thinking" is assumed to shed light on how human beings actually think. Or at the very least, the mimetic model is imputed to human thought as the best descriptor of it.

There seems to be a huge "leap of faith" involved in this strategy. The mimic is a (comparatively small) collection of behaviors manifested by the original system. What behaviors one puts into the mimic is largely arbitrary, and no attention is given to the causal underpinnings of such behaviors. Then, as Rosen points out, "something like an Occam's Razor is invoked to argue that explaining these behaviors in the mimic is adequate for explaining them in the original system."

As Rosen points out, both these strategies involve replacing the actual system of interest by some kind of surrogate, and then studying the surrogate.

What is clear to me is that both strategies drain all life and mind aspects from living organisms simply by employing the methods that characterize the strategies themselves. The irony is these strategies are attractive to scientists because they meet the scientific criterion of "objectivity." And yet each strategy, as a model, is relentlessly subjective, a "choice" of how one wants to look at material reality, and an assumption that one's model is up to the task.

Neither reduction nor mimesis looks to be a "wining strategy" if what we want is to answer the question, "What is life?"

Don't know if any of this helps, CottShop, to indicate what Rosen is up to in his work. His insights are rigorous, penetrating, profound, and — among other things — often go straight to the very foundations of mathematics.