Any attempt to formulate an absolute definition that distinguishes between life and non-life represents a similar false dichotomy. The first cell did not just appear, fully formed. Rather, life must have arisen through a sequence of emergent events - diverse processes of organic synthesis followed by molecular selection, concentration, encapsulation and organisation into various molecular structures. The emergence of self-replicating molecules of increasing complexity and mutability led to molecular evolution through the process of natural selection, driven by competition for limited raw materials.I appreciate Hazen's distinction between top-down and bottom-up attempts to work out a definition, and agree that our current conception of life is too advanced for us to likely ever know what came before.
What today appears as a yawning divide between non-life and life obscures the fact that the chemical evolution of life occurred in this stepwise sequence of successively more complex stages. When cells emerged, they quickly consumed virtually all traces of the earlier stages of chemical evolution. "Protolife", a rich source of food, was wiped clean by voracious cellular life.
Our challenge, then, rather than to define life in absolute terms, is to establish a progressive hierarchy of steps leading from a prebiotic Earth enriched in organic molecules to cellular life. The nature and sequence of these steps may vary in different environments, and we may never know the exact sequence - or sequences - that occurred on Earth. Yet many of us suspect that the chemical path has a similar, inexorable direction on any habitable planet or moon.
Elsewhere (subscribe already!) Hazen notes that the short definition is pragmatic and defensible, since it would exclude "artificial" (read: digital or virtual) lifeforms. However, here's where blurs begin again. What happens when computers, a la The Terminator, take control of material resources and begin replicating? It's only a matter of time.