“if what the computer scientist says about computers in theory does not agree with behavior, he or she can always change the computer to match the theory”1
This perception that computers are artifactual and not susceptible to theorizing arises and persists because there has not been a conceptual link made between natural forms of computation and the way humans build computers in the same way as, for instance, the conceptual correspondences between a birds wing and how humans build airplane wings.
The consequence is a belief that there is no primitive essence to pursue; that any conceptual model need go no further than what a human can understand and build. This is evident in the Turing machine model comprising a state machine and a tape control mechanism and a paper tape. Humans can clearly understand and build a working state machine and tape mechanism so there is no need to further characterize the state machine or the memory.
Lets pursue a little primitivity. A sequential process is realized with a sequence controller. There must be a most primitive sequential process with its sequential controller which controller cannot itself be a sequential process with a sequential controller but must be some form of process, not sequence controlled, that behaves on its own terms. It must posses its own liveness, logic and progress behavior. We have discovered a necessary primitivity to explore but, wait, a human can build it with a clock and Boolean logic or with a steam engine and gears and levers or with tinkertoys. We have not escaped artifactness.
We find ourselves in a self referential artifact labyrinth with no exit. Whichever way you turn you encounter an impassable wall of human ingenuity. We glimpse a possible way out of the labyrinth and a pesky human slams it shut. There is no entrance to discover with an Ariadne thread because the computer was born inside the labyrinth. Yet, there is an outside. It is generally accepted that nature computes: life, brains… So there must be some connection, some portal out of the labyrinth to non-artifactual computing.
But we have created our labyrinth with the very notion of artifactness. Just because humans build something without initially consulting nature does not mean that nature cannot also build a very similar thing. The portal out of the labyrinth is recognizing the similarity. Illuminating this portal, revealing the similarities and discovering primitivity is the task that CSR addresses.
1. P. Ceruzzi, Electronics technology and computer science, 1940-1975: A coevolution, Anals of the History of Computing 10 (4,1989): 257-275, p 267.