Pyschiatrist Randolf Nesse is a gifted writer who I have followed for many years. I first learned of Nesse’s work when I read Why We Get Sick: The New Science of Darwinian Medicine. Nesse is one of the many respondents to this year’s annual question by Edge.org: “What will change everything?”
Nesse’s answer: RECOGNIZING THAT THE BODY IS NOT A MACHINE
As we improve our knowledge of bodies, they don’t fit very well within our venerable metaphor of the body as a “machine.” One of his points is that we can describe machines, whereas a satisfying description of bodies seems so elusive. The complexity of the body is, indeed, humbling:
We have yet to acknowledge that some evolved systems may be indescribably complex. Indescribable complexity implies nothing supernatural. Bodies and their origins are purely physical. It also has nothing to do with so-called irreducible complexity, that last bastion of creationists desperate to avoid the reality of unintelligent design. Indescribable complexity does, however, confront us with the inadequacy of models built to suit our human preferences for discrete categories, specific functions, and one directional causal arrows. Worse than merely inadequate, attempts to describe the body as a machine foster inaccurate oversimplifications. Some bodily systems cannot be described in terms simple enough to be satisfying; others may not be described adequately even by the most complex models we can imagine.
[Related DI post: The Brain is not a Computer]
Computer software evolves in a way similar to biological evolution, and can result in similar indescribable complexity. Programmers of different levels of skill will build on and modify existing software in a way similar to mutations–they don't always know what they are doing, or why, just that it works. This is especially true in systems that are decades old and in systems that are really massive. In each case, there are components that are essentially black boxes, that no one really understands. As a result, programmers try this, try that, try something else, test, test, test, and ultimately hope for the best. The result can be massive, complex systems like the internet that no one really understands at a detailed level. (There are better examples, but that's the most familiar one.)
Maybe the biologists and computer scientists can get together (maybe with some mathematicians and physicists) and try to figure out ways of modeling and describing the complexity. That'd be nice.
There are actual evolving forms of software. Programmers set up code kernels that mate with each other and swap code fragments to propagate in a multi-thread environment, to automatically breed optimum functions to meet a specification (selection criterion).
It works, and it doesn't. The resulting code (for example a satellite guidance routine) beats anything written by a programmer in tests. But it gets so complex that no human programmer can understand how it works. Therefore it is unmaintainable in captivity, should new conditions emerge.
Back in the early 1960's the prevalent metaphor was the "everything is like a factory" concept emphasize how all the parts contribute to a greater whole. Like the different machines in a factory, the different systems in the body worked together to contribute to the functioning of the body, and all parts are equally important. This metaphor was also applied to society, indicating the importance of each individual's contributions to the progress and fabric of society.
I think it was during the Reagan years that whole factory took a backseat to the individual machine, Society became secondary to the individual, the team became less important than the quarterback (or center for us B-ball types), a troop, which once described a group of soldiers, became a single soldier, and our leadership quit being a part of the common constituency and placed themselves above and outside of the standards and laws to which they held the rest of us accountable.
Part of the machine metaphor is that most machines are easily replaced by newer and more compliant models. Ones that better respond to what their operators require of them. Most of all machines don't think, Machines do as they are told.