home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!caen!ingles
- From: ingles@engin.umich.edu (Ray Ingles)
- Newsgroups: alt.atheism
- Subject: Re: contradictory nature of free will and omni-anything
- Date: 23 Jan 1993 01:19:22 GMT
- Organization: University of Michigan Engineering, Ann Arbor
- Lines: 62
- Distribution: world
- Message-ID: <1jq6eqINNm3f@srvr1.engin.umich.edu>
- References: <16B5C12448.I3150101@dbstu1.rz.tu-bs.de> <1jl1rfINNc91@srvr1.engin.umich.edu> <1993Jan22.170557.4387@st-andrews.ac.uk>
- NNTP-Posting-Host: syndicoot.engin.umich.edu
-
- In article <1993Jan22.170557.4387@st-andrews.ac.uk> nrp@st-andrews.ac.uk (Norman R. Paterson) writes:
-
- [deletions]
-
- >I think free-will is an illusion. There's also a difference between being
- >determined and being predictable, as I will try to show. In the following
- >I use the term "computer" to mean "hardware + software".
- >
- >A computer's behaviour is entirely determined (ignoring quantum effects for
- >the moment). It's progression from one state to the next is hard-wired.
- >
- >However, it is impossible to _predict_ what a given computer will do. If
- >you want to find out, you have to let it run, and wait for it to finish.
- >(This is the basis of the halting problem: you can't even predict whether
- >any given computer will halt or not.)
-
- Well, in *practice*, yes, things may not be predictable. But in principle
- they need not be unpredictable. As an example, take the process:
-
- Given an integer:
- If even, divide by two.
- If odd, multiply by three and add one.
- Repeat.
-
- As far as anyone knows, any integer you do this to will eventually go
- to one. But no one has proved it; there may be a 'loop' that doesn't
- include the number 'one'.
- I could make a turing machine to run this procedure for all integers,
- stopping if it finds a sequence that loops. Now, as far as anyone knows,
- it will never stop.
- But, there is nothing that says that we might not find a proof that it
- will or won't stop. Some clever individual may find a proof that all
- integers go to one under that process. *Then* we would know.
- Or consider a simpler case; a Turing machine with one command: "Halt."
- I can know whether or not that one will stop.
- The halting problem does not say that there is no way to know if a
- program will halt or not; it says that there is no *algorithm* that will
- tell for *all* programs.
-
- Also, what if we run the same program on two computers, but one is faster
- than the other? If the faster one stops, we can know that the slower one
- will stop.
-
- Now, in practice, it may prove too complex to figure out what a computer
- will do. Even ignoring the possibility of cosmic rays messing with the
- hardware, some programs are just too big. And, of course, for many systems,
- the least uncertainty in the data will make prediction beyond a certain
- point impossible.
- But this is in *practice*, notin *principle* as you state above.
-
- >So I think we are determined but unpredictable.
-
- I tend to agree, but for different reasons. More precisely, I suspect
- we are determined but we can't ever know it for sure, becuse we can't
- test by predicting what someone will do.
-
- Sincerely,
-
- Ray Ingles ingles@engin.umich.edu
-
- "The meek can *have* the Earth. The rest of us are going to the
- stars!" - Robert A. Heinlein
-