[Next] [Previous] [Top] [Contents]

The Possible Irreducibility of Artificial Software Life - Bruce Edmonds

1 Introduction

Sometimes it is assumed that simulating (or even producing) life as software on a computer would be tantamount to reducing it. In fact, it seems that many of the arguments that real artificial life is impossible seem to be driven by the fear that this would necessarily mean that our life is similarly reducible.

This implicit argument can be summarised like this:

  1. Something is reducible if there exists a Turing machine that can accurately model it.

  2. All software can be computed on a Turing machine.

  3. Life is irreducible.

Therefore

  1. Artificial Software Life would be reducible.

And thus

  1. Artificial Software Life is impossible.

The assumption most frequently objected to is (3). This is not a question I aim to address here*1. I also leave the truth of (5) open. All I aim to show is that if artificial software life ever arose then it might be as irreducible as natural living systems.

The assumption I will argue against is (1). That merely having a program of something is not the same as reducing it and that it is possible that evolved software life would be as irreducible as its more tactile cousins.

The Possible Irreducibility of Artificial Software Life - Bruce Edmonds - 20 MAY 97
[Next] [Previous] [Top] [Contents]

Generated with CERN WebMaker