A pedantic analysis of the academic criteria of life: the most boring title I could think of

I read a lot of sci-fi and one of the ideas of science fiction that has held the most allure as a plausible but not yet realized technology is the Von Neumann machine (VNM). A VNM is simple to define but difficult to build, it is just a machine that can make another version of itself without outside intervention. The golden age of sci-fi ran with this idea, imagining autonomous spacecraft that can travel between solar systems and utilize local resources to reproduce exponentially. My favorite example of this is the '“Bobiverse” series of books by Dennis E. Taylor, definitely worth a read. In addition to reading, I also have a day job of teaching science classes. Every biology class starts out with a definition of life that is vague and usually falls back on requiring several criteria to be fulfilled for a thing to be considered alive, each necessary but not sufficient on their own. Depending on the text, decade, website, or whims of the author, the lists of criteria a thing must meet to be considered alive varies. I’d like to rebut a generalized version of this list, so if you were taught an additional criteria in your high school biology class that I’ve left off, well then take that up with the philosophers.

The first criteria on the list usually seems to be that a thing must be able to grow to be alive. I don’t know why this usually comes first, maybe it is some subtle form of plagiarizing earlier lists of criteria. Now I’ll remind you that all of these criteria are admittedly necessary but not sufficient, so the challenge isn’t to find something that grows that isn’t alive, but rather to find something that is alive that doesn’t grow. To grow is to gain mass, and presumably complexity. In other words, to be assembled. A bacterium that splits in half to make two new bacteria, necessarily makes daughter bacteria that are half as big, so if that species didn’t include the ability to grow back to full size, each successive generation would be smaller and smaller. Growth does seem necessary in that instance. Humans similarly start out small and get bigger, at least for a while. Jokes aside about growing around the midsection, most adults don’t grow. Does that mean they aren’t alive? Should we have to amend the criteria to a more specific form ‘for something to be alive it must grow at some point in its life”? Maybe, but adding caveats smacks of a poorly chosen rule to begin with. If this is supposed to be a universal truth, the least it could do is be elegantly simple. Is that too much to ask? So what about a VNM? Does assembly count as growth? I’d say that a machine assembling another machine isn’t a huge philosophical leap from a mother developing an embryo in her womb. There is a deeper implication that might bely a bias of how we see life. Our lives start as single cells, accrue material and assemble it into larger, more complex parts. Could we conceive of a counter example? Instead of building from the ground up, starting large and calving small living things from it? Like an iceberg from a glacier. Perhaps a god that formed beings out of clay, or a Star-Trekian energy could that could make humanoid monsters for Kirk to fight. Our mythologies and fictions are filled with examples of living things that were created spontaneously from larger, more complex and more powerful things, but just because we haven’t found any real world versions we have decided to exclude that from our definitions of life.

The next criteria is typically that a thing must be able to reproduce to be considered alive. The VNM clearly passes this test, but a lot of my friends don’t. My wife has verifiably given birth to three children, so presumably that means I meet the criteria for being a living thing. I can feel my self esteem rising already. However, just like adults who aren’t growing anymore, there are tons of people who haven’t or can’t reproduce. The caveat that is added onto this criteria to address this obvious shortcoming is even worse than the last one. The legalese that is often tacked onto the criteria of reproduction is along the lines of: “a thing must be a member of a species or population that is capable of reproduction to be considered alive”. Even if we grant this caveat, this criteria easily breaks down in ad absurdum examples: there are only two Northern White Rhinos left in existence and they are both female, ergo, neither belongs to a population or a species that is capable of reproduction. Does that mean they shouldn’t be considered living things? The bananas you buy and eat are seedless and can’t reproduce. But that doesn’t make sense because there are new bananas every year. Well, banana plants grow multiple stems, which can be cut apart and grown as two separate plants. Do this often enough and you can have huge groves of identical ‘clones’, which are essentially chopped up bits of one huge super banana plant. So does that mean that banana plants aren’t alive? Seems that this criteria is having trouble holding water too.

I’ll address one last criteria that is often split as two: metabolism and homeostasis. A thing must use energy to create and maintain an internal conditions that are distinct from the environment around it. This is the most interesting criteria of them all. The first problem I’d like to crash headlong into is the dual verbs of create and maintain. The implication is that the thing itself is what is using the energy to create and maintain its conditions, as opposed to a watchmaker who carefully assembles a machine that can run for a while but is incapable of self-repair. It turns out, I’m getting worse and worse at self repair as I get older. Some moths don’t eat but instead rely entirely on energy that was obtained when they were a caterpillar, during these last few weeks of their lives they are degrading rapidly. The caterpillar was the watchmaker and the moth is the watch, slowly winding down, not maintaining homeostasis. A VNM that makes another VNM and is also capable of self repair would be maintaining homeostasis, but those critical internal conditions that are different from the environment around it were created by a previous robot, not by itself. This could also be said to be true of ourselves, as we start out as single cells that weren’t made by us but were made by other people. Look back at the first criteria and we are obviously growing that homeostasis, but we didn’t create it. The verb maintain is problematic as well. Must you be actively maintaining homeostasis to be alive? What about a VNM with no broken parts to repair? Must you have the capability to maintain? What about a terminally ill human, incapable of sustaining life but certainly alive at the moment? The only part of this criteria (or of all three) that seems philosophically sound is that living things are in some state of homeostasis that is distinct from their environment. Life is a momentary lack of entropy in a small area of space. There are clearly a lot of things that fit that definition that aren’t alive, but again the game is to find something that is alive that doesn’t fit that criteria.

After years of telling thousands of young people what defines life, I’ve only eroded away my certainty in any of the definitions. Our intuitions about if a thing is alive or not is still a better guide than the most well crafted checklist of criteria. I’ll borrow a bit of wisdom from Supreme Court Justice Potter Stewart when he struggled to define something obvious: “I know it when I see it”.