Subject: Moor From: "William J. Rapaport" Date: Mon, 22 Mar 2010 11:45:00 -0400 (EDT) Here's a question from one of you prompted by Moor's article; my reply (which is almost as long as the question:-) is online at: http://www.cse.buffalo.edu/~rapaport/584/S10/EMAIL/question20100321.pdf -------------------------- I'm still troubled regarding Moor's article. In particular, his account for computer vs computer program. I'll refer back to the LISP example. To run a lisp program, we first need a LISP interpreter. Eventually at some point we go through interpreters until we reach a point of Physicality and no longer need something which can 'understand' our program. Once we hit that point, we are said to be at the computer. This to me is problematic, and I don't particularly like the results. It would seem the best case scenario is that there is only one large computer(the universe) and we are placed into an uncomfortable predicament with our language. The worst case scenario is that we have an infinite regress and there are no computers. Suppose we have a LISP program running on your off the shelf x86 platform. At what point does the interpretation start being computed by the computer? Supposedly it's once the instructions reach the CPU and at that point no longer need an interpreter. However I will argue that even the CPU(and any other device for that matter) is in fact in need of an interpreter. Namely, if we embrace the modern physics perspective, the CPU is reliant upon a lower layer of say molecules. This molecular layer decomposes into various elements, atoms, protons/neutrons/electrons, quarks, boson's mesons, etc. In order for the CPU to execute, there must be a lower laying executing the CPU. Just as there needs to be a lower layer to implement any high level programming language. Now let us suppose there is some truly fundamental particle. It follows that in reality the universe then is one big computer composed of said particle. All of our concrete particulars in said universe are just interpreted programs. In reality there would be only one 'computer', and when we speak of our x86 computers we are really speaking about programs. This is the less harmful, but nonetheless uncomfortable scenario. The more harmful scenario runs as follows. Inductively it appears to me that with any modern scientific atomism comes an infinitely recurring sequence of smaller and smaller pieces. Given infinite time for discovery, we will commit our selves to an infinite number of 'basic' particles. If you accept my sloppy inductive proof, it would seem there is no end to the interpretive process. As a result, we would commit ourselves to the existence of only computer programs. Further more, we would commit our selves to the non-existence of computers. As a result, we have only computer programs without a computer to run them on. How can it be the case that this is possible? If you haven't figured out by now, I'm arguing it isn't. It may be argued that we need not hit the lower most layer of physicality to declare ourselves in the computer. However, should this be the case, why not push our primitive higher up the ontology? It seems to me that any computer program must itself be physically represented in some form(baring the existence of a program in the mind and mind's being non-physical). There would appear to be no reasoning as to why it must be placed near the CPU/Registers/etc. I would argue that A CPU is a grouping of transistors, those transistors are really just a grouping of elements, those elements atoms... A computer program may just be a series of bits on a hard drive, those bits a magnetic orientation, that orientation an arrangement of electrons/protons and on and on. Or if you want to account for programs on paper, a program may exist as a series of symbols, those symbols then made of ink on paper, and on and on. One possible objection I do see, is to maintain that some how programs exist in some sort of other worldly plane of existence. Really a program exists there, and the physical computers just model it. However, I find myself opposed to these sort of super natural and apriori states of affairs. Also, I see no reasoning why this ought to be adopted except for the ability to save the prior account of a computer/computer program. However, if that should be our only reason for introducing this new plane of existence, I should think we would be better off finding a different account for computer/computer program. It would seem that introducing such a realm provides us with no greater explanatory power.