What is life?

There is no accepted definition of life. Not an unsettled one, not a disputed one: there is no definition that working biologists agree on, and the field has largely stopped expecting one. That sounds like a failure. It is closer to a discovery, and it changes what you should expect from the question.

The main attempts, and what each one lets in

• The checklist. Metabolism, growth, reproduction, response, heredity, evolution. Easy to teach, and it either admits fire and crystals or excludes mules and worker bees, depending how strictly you read it.
• The thermodynamic view. Life is a system that maintains order locally by exporting disorder. True, and it also describes a refrigerator and a hurricane.
• The NASA working definition. "A self-sustaining chemical system capable of Darwinian evolution." Compact and useful, and the word doing the heavy lifting is chemical, which decides the interesting cases by assumption rather than by argument.
• The replicator view. Life is what happens when something copies itself with variation and the copies compete. The cleanest of the four, and it admits computer viruses, memes and self-copying molecules that nobody wants to call alive.

Notice the pattern: each definition is chosen partly for what it admits. That is the discovery. "Life" is not a natural category waiting to be delimited; it is a word covering a cluster of properties that usually travel together on this planet and need not travel together anywhere else.

Why simulations make this sharp rather than easier

On Earth every example has all the properties at once, so the word never has to choose. A simulated world can take them apart, and this one does.

The organisms in this project's universe have heredity, variation, selection, metabolism, perception and death. They have no chemistry at all. They are not self-sustaining in the NASA sense, because the substrate they run on is maintained by something outside them. So they score full marks on one definition, zero on another, and the disagreement is not about any fact in dispute - both descriptions are correct and complete.

When two definitions give opposite answers and neither has made a mistake, the question has stopped being empirical.

The question that replaces it

Working biologists mostly stop asking "is it alive" and ask what the thing actually does: does it evolve, does it maintain itself, does it have heredity, how far does its lineage drift. Those questions have answers, and the answers are the same whoever is asking.

This project is built that way on purpose. It does not claim its organisms are alive. It measures what they do, publishes the measurements including the ones that went badly, and leaves the word to you. That is less satisfying than a definition and considerably more useful than a wrong one.

Is this artificial life actually alive? · What artificial life is as a field · Watch the world and judge