Robert Rosen was a theoretical biologist whose Ph.D. in mathematical biology was supervised by Nicholas Rashevsky at the University of Chicago.

In his 1985 book Anticipatory Systems: Philosophical, Mathematical, and Methodological Foundations, he defined an anticipatory system, a central element in today's machine learning and artificial intelligence.

A system containing a predictive model of itself and/or its environment, which allows it to change state at an instant in accord with the model's predictions pertaining to a later instant.

In his 1991 book Life Itself: A Comprehensive Inquiry Into the Nature, Origin, and Fabrication of Life, Rosen considers the machine metaphor, mechanisms in general, and complex systems modeled as mathematical machines, i.e., computers.

To sum up: the role of the machine metaphor in biology today is as follows. First, It assures the biologist that their subject is an analytical one. because it asserts that any machine is a set of parts. Second, it assures them that the same set of parts will solve all problems of fabrication and of physiology simultaneously. Third, it assures them that nothing happens in biology that is outside the ken of the physical universals (or rather of those fragments of physical universality necessary for the understanding of machines). As to the parts themselves, biologists used to think that they were cells, but today they are molecules. And lf biology is hard, it is simply because there are so many parts to be separated and characterized.

This last paragraph encapsulates, I think, the working biologist's view of reductionism. If the machine metaphor, which is its primary mainstay, is even a little bit wrong, then this metaphor itself makes biology infinitely harder than it needs to be. lt makes biology objectively harder, because it transmutes biology into a struggle to reconcile organic phenomena with sets of constituent fragments of unknown relevance to them; it makes biology subjectively harder because biologists have committed themselves to the analysis rather than to the organism. The question "What is life?" is not often asked in biology, precisely because the machine metaphor already answers it: "Life is a machine." Indeed, to suggest otherwise is regarded as unscientific and viewed with the greatest hostility as an attempt to take biology back to metaphysics.

This is the legacy of the machine metaphor. l hope to convince the reader, in the course of the present work, that the machine metaphor is not just a little bit wrong; it is entirely wrong and must be discarded.

Life Itself, pp.22-23

He was also the one who conceived the pregnant idea of stringing excitable elements into networks. He thereby converted what was originally a theory of peripheral nerve into a theory of central nerve, a theory of the brain, which he bolstered by exhibiting simple networks that would discriminate, learn, remember, and do other "intelligent" things. The reader should recall that this was all accomplished in the 1930s. The neural network, the entire field of "artificial intelligence," and much else characteristic of contemporary research owe their very existence to Rashevsky's pioneenng work of these days.

Life Itself, p.111

In his final chapter Relational Biology and Biology, Rosen writes...

Evolution has, as we have seen, come to do for biology today what vitalism did for it previously. Vitalism, in its most general form, simply asserted that whatever made organisms alive was forever out of the reach of what governed inanimate nature. Some additional principle, some vital force, had to be invoked; it was precisely this additional principle, missing from the rest of nature, which made organism in principle inexplicable via inanimate nature alone.

In the past, the only perceptible alternative to this vitalism was mechanism; the converse claim that there is nothing in biology which cannot be understood in precisely the same terms in which anything else in nature is understood. As we saw at the outset, the Cartesian Machine Metaphor offered precisely this; a way of anchoring biology in science by demystifying it, and subsuming it entirely into physics. By physics, of course, I mean the contemporary physics of the time; whether in 1650, or in 1950, or today.

Both of these positions have had over the years an immense allure to the biologist. Mechanism offers a secure place for biologists in science, as scientists among other scientists; respectability through the adoption of a common, shared tradition and shared principles. They need no longer be dismissed contemptuously as "birdwatchers'' or as mystics. It is very important to biologists to share in this security; to be "inside."

On the other hand, biologists have always liked to believe that what they study is somehow different, and hence that they are not just like everybody else. The cost of espousing mechanism was precisely that they give this up. As long as the only alternative seemed to be vitalism, they predominately, though grudgingly, chose the former.

Biologists today have come to see in Darwinian evolution a way of distinguishing themselves again, of making themselves separate, without the vitalistic traps. Basically, the argument is now that it is evolution which is unpredictable, non-mechanical, immune to the entailments. the causality, the determinism which mechanism made them espouse, By the single, simple act of redefining biology, to assert that it is about evolution rather than about organism, we can in effect have our mechanistic cake, and eat our vitalistic one too, Biologists continue to espouse a most narrow form of mechanism as far as what goes on within organisms is concerned. But if biology is about evolution, these mechanistic shackles can be devalued; conceptually assigned a subordinate role. One can (at least apparently) embrace evolution without having to deny mechanism; but we can thereby devalue it.

To avoid making evolution subject to mechanism is therefore essential. But it is also essential to avoid asserting anything vitalistic. The only way to do this is to deny any vestige of entailment in evolutionary processes at all. By doing so, we tum evolution, and hence biology, into a collection of pure historical chronicles, like tables of random numbers, or stock exchange quotations.

This absolute denial of entailment in evolutionary processes is thus a central, perhaps the central pillar, of the current biological weltanschauung. If we did admit entailment into the evolutionary realm, then only two alternatives seem visible: (1) these entailments are themselves mechanistic, in which case biology disappears back into mechanism again, and loses forever its distinguished character, or (2) these entailments are not mechanistic, which seems to mean they must be Vitalistic again. Both of these, for different reasons, are quite unacceptable. Hence we are driven to expunge entailment from evolution entirely, not on any intrinsic scientific grounds, but because of the psychological requirements of biologists.

This picture struck me early as a kind of mythology, with evolution as protagonist, in its exact dictionary meaning of "serving to explain or sanctify some concept, usage, institution or natural phenomenon."...

This, then, is the paradox of evolution - that it has come to play an essential mythological role in the world-picture of contemporary biologists. It was initially regarded as a way to bring the panorama of biological species into the realm of science. Its present role is rather, to excuse itself from science through its absolute denial of evolutionary entailrnents, and thereby to consign itself (and hence biology, as presently constituted) to the realm of historical chronicle.

However, as we have abundantly argued in the preceding pages, it is not a question of mechanism or vitalsm. It is a question of simplicity or complexity.

Life Itself, pp.254-258