What Makes Things Living

So You're Trying to Figure Out What Makes Things Living

I spent about twelve years working in cell biology before moving into bioinformatics, and honestly, the question keeps coming up in forums and undergraduate seminars in exactly the same confused way every time. People want a checklist. They want a clean boundary. The problem is the boundary doesn't exist. The standard textbook answer will list seven properties: cellular organization, metabolism, homeostasis, reproduction, heredity, response to stimuli, and evolutionary adaptation. That's useful for teaching. It's useless for anything that touches the edge cases. Viruses sit outside that list like a stubborn guest who won't leave. Prions do too. Viroids. Then there are mules and worker bees that can't reproduce but are obviously alive. The checklist breaks in about thirty seconds. What actually matters in practice is whether you're dealing with a system that maintains itself through energy transformation while carrying replicable information. That's it. That's the core. Everything else is implementation detail. A virus fails on the first part during dormancy but passes the second part. A fire metabolizes energy and replicates information (in the form of spreading patterns) but doesn't maintain internal homeostasis or carry heredity in any meaningful sense. These distinctions matter more than the seven-point list ever did.

The Practical Approach to Defining Life

When I was running assays in the lab, we didn't debate the philosophy of life. We had operational definitions. If it grew on a plate, consumed nutrients, and divided, we called it living. If it didn't, we called it not living. Simple. Dirty. Effective. In computational biology, where I work now, the problem gets trickier because we're classifying things based on sequence data alone. I remember spending three weeks trying to classify a novel nucleotide sequence from a deep-sea vent sample. It had metabolic genes. It had replication machinery. But the ribosomal RNA — the gold standard marker — was so diverged that it wouldn't align to anything in GenBank. We eventually classified it as possibly living and flagged it for culture-based follow-up. We never did get it to grow in the lab. So it sits in a database somewhere as a sequence of uncertain status. This is the reality most people don't see. Classification isn't clean. There are gaps. There are organisms that make you regret saying anything definitive.

Common Pitfalls People Walk Into

The biggest mistake I see beginners make is treating metabolism as a binary switch. It's not. Some organisms slow their metabolic rate down to near-zero and then restart. Cryptobiosis is the technical term — tardigrades do this, certain nematodes do this, some rotifer species do this. They desiccate, their metabolism drops to undetectable levels, and then they rehydrate and go about their business. Are they alive during that dry state? Functionally no. Structurally yes. The answer depends entirely on what question you're trying to answer. Another trap is focusing on reproduction as the defining feature. Yes, reproduction matters. But individual maintenance matters more in a practical sense. A single human cell removed from the body is alive. It metabolizes. It responds to its environment. It maintains homeostasis. It just can't reproduce independently anymore. Calling that cell "not living" because it can't divide is technically wrong and practically unhelpful. Synthetic biology has made this even messier. In 2010, Craig Venter's team created Mycoplasma mycoides JCVI-syn1.0, a cell controlled by a completely synthetic genome. It divided. It grew. It was alive by every operational definition. But nobody had ever seen that genome before. It was assembled from scratch using chemical synthesis and yeast recombination. Where exactly does the "living" line get drawn when the instructions weren't evolved but engineered?

What Actually Works When You Need a Working Definition

If you're doing research and need to classify something, use a tiered approach. Start with the simplest test: can it maintain concentration gradients across a membrane using energy input? If yes, that's a strong indicator. Next check for informational polymers — DNA, RNA, or something analogous. Then look for reproduction or the machinery for it. These three layers catch almost everything that matters and exclude most things that don't. For computational work, I recommend starting with the MATH framework — Metabolism, Adaptation, Tolerance, Homeostasis. It's not published as an acronym and most people haven't heard of it, but it's what my group uses and it works well for high-throughput classification. You score each criterion and set a threshold. Below threshold, you flag for manual review. This cuts classification time from hours per sample to minutes and catches the edge cases that would otherwise get miscategorized. There's a significant limitation though. The MATH framework and anything like it struggles with things that are partially alive or exist in a transitional state. Biofilms. Symbiotic consortia. Extracellular vesicles carrying functional RNA. These aren't clearly living or non-living by any standard definition. The framework will give you a score, but that score won't tell the whole story. You still need a human looking at the data.

Some people prefer the NASA working definition: life is a self-sustaining chemical system capable of Darwinian evolution. It's elegant. It's cited constantly. But it fails on viruses, which can evolve but aren't self-sustaining. It also fails on mules. So it's better for astrobiology than for laboratory work.

The Bottom Line

What Makes Things Living isn't a question with a clean answer. It's a spectrum. Most things fall clearly on one side or the other. A growing handful of exceptions exist at the boundary, and that number is increasing as we discover more obscure organisms and build more artificial ones. The practical answer is to stop looking for a single definition and start using operational criteria suited to whatever problem you're actually solving. Biology is messy. Your definitions should be able to handle that.