What is life?
Simple question, thousands of years of human intellectual torture trying to answer it. The truth is that 'life' really does seem to defy easy definition. We can say that it's a natural phenomenon - yes, OK. Actually it might be better thought of as a number of deeply connected natural phenomenae, OK, that too. But what are the most critical pieces of these phenomenae that allow us to distinguish life from something that isn't quite life?
I don't think anyone has a good solution. And it may be that there's something a bit off with the question in the first place - our attempts to pin down a complex and sometimes extraordinarily variable set of characteristics with a few parameters may just not work well.
Nonetheless, for astrobiology it would be really useful to have some kind of working hypothesis. A practical set of rules that would help identify 'life', even if they were not always sufficient to describe it - these would at least let us eliminate false detections. What we could use is a universal life litmus test.
Self-organization, reduction and oxidation, entropy, self-replication and information storage, cellular structure, energy and metabolism, change and mutation, all characteristics that might form part of such test. But the devil is truly in the details.
Now listen to this excellent podcast from Sift, the brainchild of Aaron Bishop Sand, where he gracefully and intelligently torments a whole groups of scientists with just these questions. You will hear me on this, and if you listen to the middle you'll get to find out what I think chickens have to do with it all...
It's a tremendous piece, a veritable sound sculpture. Enjoy.