No single definition of life has gained universal acceptance in biology, despite more than a century of serious attempts. Scientists have proposed over a hundred definitions, and the most common approach is a checklist of properties: organization into cells, metabolism, growth, reproduction, response to stimuli, and evolution. But every checklist runs into something that meets most of the criteria and fails one or two, or something clearly nonliving that sneaks through. The difficulty is not a failure of effort. It reflects something genuinely slippery about the boundary between living and nonliving matter, a boundary that biology, chemistry, philosophy, and even space science are all still trying to draw.
The Checklist Approach and Why It Persists
Most biology textbooks teach a version of the same list. To count as alive, something should be made of cells, maintain an internal environment, metabolize energy, grow and develop, reproduce, respond to its surroundings, and adapt over generations through evolution. This “list definition” framework remains the most widely taught because it captures what most organisms obviously do. A bacterium, an oak tree, and a human all check every box.
The appeal is practical. If you are a student in a biology course or a field researcher cataloguing organisms, these criteria work well enough for everyday purposes. The trouble starts at the edges. A mule is alive but cannot reproduce. A crystal grows but is not alive. Fire consumes fuel, responds to its environment, and spreads, yet nobody seriously argues fire is living. Each criterion, taken alone, fails to separate living from nonliving. The hope has always been that the full cluster of properties, taken together, would do the job. But attempts to define life through a strict set of necessary and sufficient conditions have repeatedly failed, leading some researchers to argue that life is better understood not as a sharply bounded category but as a cluster of properties that tend to travel together without any single one being absolutely required.1Biological Theory. Life as a Homeostatic Property Cluster
That framing helps explain why the checklist persists despite its flaws. If life is a cluster rather than a neat definition, then organisms near the center of the cluster (a bacterium, a fern) are unambiguously alive, while entities near the edges (a virus, a dormant spore) will always generate argument. The checklist is not wrong so much as incomplete, a map that covers the heartland but leaves the borders blurry.
The NASA Definition and Darwinian Evolution
When the question shifts from “what is alive on Earth” to “how would we recognize life anywhere in the universe,” a different definition takes center stage. NASA’s working definition, widely used in astrobiology, describes life as “a self-sustained chemical system capable of Darwinian evolution.”2PubMed. The Origin, Extension, and Future of the “NASA Definition” of Life This definition strips away Earth-specific details like cells, DNA, and carbon and focuses on two abstract requirements: the system sustains itself chemically, and it can evolve through variation and selection.
The strength here is generality. If life on another planet used a completely different chemical basis, the NASA definition would still apply as long as that chemistry could sustain itself and undergo open-ended evolution. The weakness is that the definition is hard to apply to a single organism. A lone bacterium is alive, but one bacterium cannot undergo Darwinian evolution by itself. Evolution is a population-level phenomenon. So the definition works well for identifying living systems but awkwardly for deciding whether any one thing is alive. It also arguably excludes things most people would consider alive, like a sterile animal that will never contribute to evolution.3PubMed Central. Defining life
Viruses and the Gray Zone
Viruses are the most famous boundary case, and the debate over whether they are alive has gone on for decades with no resolution. On the “not alive” side: viruses have no metabolism, cannot reproduce on their own, and are inert outside a host cell. On the “alive” side: they evolve rapidly, carry genetic information, and hijack cellular machinery with extraordinary sophistication. One analysis put it bluntly: the question of whether viruses are alive is effectively without substance, because the answer depends entirely on which definition of life you adopt, and that choice is bound to be arbitrary.4PubMed Central. Are viruses alive? The replicator paradigm sheds decisive light on an old but misguided question
What makes the virus question genuinely interesting rather than just semantic is that viruses are not all alike. Giant DNA viruses, such as mimivirus, have genomes larger than some bacteria and carry genes once thought exclusive to cellular life. Recent work has found that mimivirus encodes a complete, functional translation initiation complex, the molecular machinery needed to start building proteins, which it uses to replace the host’s own machinery during infection.5PubMed. Cap in hand: giant viruses, stolen translation, and a road to endosymbiosis? That kind of capability blurs the line between viruses and cells in ways that simple “alive or not” framing cannot capture. These giant viruses sit in a place where the traditional checklists start to break down: they carry out some functions associated with life while still depending on a host for others.6PubMed Central. Giant DNA viruses encode a hallmark translation initiation complex of eukaryotic life
Even stranger are viroids, tiny loops of RNA found in plants that are far simpler than viruses. A viroid is just a short stretch of RNA, typically 246 to 401 nucleotides long, that encodes no proteins at all. It has no coat, no enzymes, nothing. It relies almost entirely on the host cell’s own machinery to copy itself. Viroids are essentially parasitic information, naked genetic instructions that exploit a cell’s pre-existing tools. If viruses are a gray zone, viroids are a deeper shade of gray, showing that something can behave like a biological entity with almost none of the physical trappings we associate with life.7PubMed Central. Derailing the host machinery to achieve replication: how viroid and viroid-like RNAs successfully copy their genomes in hostile territory
Suspended Animation and Dormancy
Living things do not always look alive. Tardigrades, the microscopic animals sometimes called water bears, can enter a desiccated state of suspended animation in which all detectable metabolic activity stops.8PubMed Central. Time-series transcriptomic screening of factors contributing to the cross-tolerance to UV radiation and anhydrobiosis in tardigrades In that dried-out state, a tardigrade does not metabolize, grow, respond to stimuli, or reproduce. By every observable criterion except structural integrity, it resembles a dead speck of dust. Add water, and it revives. Seeds can remain viable for centuries. Bacterial endospores have been revived after thousands of years in permafrost.
These cases pose a real problem for any definition that relies on active processes. If you define life by what something is doing right now, a dormant tardigrade is not alive. If you define it by potential, you have to explain why a bag of amino acids does not also count, since it has the chemical potential to participate in living processes under the right conditions. Most biologists handle this by treating dormancy as a temporary state within a life cycle rather than a challenge to the definition itself, but it is a reminder that “alive” and “actively living” are not always the same thing.
What Is the Bare Minimum for a Cell?
One way to pin down what makes something alive is to strip a cell down to the smallest possible version that still works. In 2016, researchers synthesized a near-minimal bacterial cell by building a genome containing only the essential and quasi-essential genes of a simple bacterium. The resulting organism could replicate DNA, transcribe RNA, build proteins, and divide, and that was about it.9PubMed Central. Minimal Cells-Real and Imagined It had just 473 genes, the smallest genome of any free-living organism, yet roughly a third of those genes had unknown or only generic functions.10Cell. A 75-gene genome constraint for mammalian design and cell division That is a humbling finding: even at the absolute minimum of life, we do not fully understand what every necessary part does.
Synthetic biology has also approached the question from the other direction, building upward from nonliving components. Researchers have engineered artificial vesicles, tiny membrane bubbles, that can carry out transcription and translation from a synthetic genome, replicate their own DNA, and even synthesize new membrane material.11bioRxiv. A synthetic cell with integrated DNA self-replication and membrane biosynthesis These synthetic protocells are not alive by anyone’s standard, not yet. But they demonstrate that individual functions associated with life, like information processing, self-replication, and boundary maintenance, can be built from scratch and integrated inside a single compartment. The gap between these constructs and a genuinely living cell remains large, but it is getting smaller.
The Transition From Chemistry to Biology
If we struggle to define life even among things we can study today, the question gets even harder when we look backward in time. At some point roughly four billion years ago, chemistry on early Earth became biology. How? The honest answer is that nobody knows the full mechanism, but research on protocells offers clues. Protocells are simple membrane-enclosed compartments that form spontaneously from certain molecules in water. They cannot replicate themselves or evolve, but they can concentrate chemicals, create reaction environments, and exhibit behaviors that look like crude precursors to cellular functions.12PubMed Central. Investigating Prebiotic Protocells for A Comprehensive Understanding of the Origins of Life: A Prebiotic Systems Chemistry Perspective
Recent experiments have shown that protocell-like structures can form at the same time as the organic building blocks of life under conditions plausible for early Earth. When researchers simulated lightning in an atmosphere of gases thought to resemble early Earth’s, they found that hollow, vesicle-like structures formed alongside amino acids and other prebiotic molecules, not as separate steps but together.13PubMed Central. Concomitant formation of protocells and prebiotic compounds under a plausible early Earth atmosphere This suggests that compartmentalization and chemical complexity may have co-evolved from the beginning, rather than one preceding the other. The transition from nonliving to living was probably not a single event but a gradual accumulation of capabilities, with networks of chemical reactions becoming self-sustaining before anything resembling a modern cell existed.14PubMed Central. Autocatalytic Networks at the Basis of Life’s Origin and Organization
This gradual picture is precisely why a sharp definition is so hard to formulate. If the transition from nonliving to living was continuous, then any line we draw between the two is, in some sense, imposed by us rather than discovered in nature.
Detecting Life Without Knowing What It Looks Like
All of Earth’s known life shares DNA, RNA, proteins, and a common set of metabolic pathways. But if you are designing an instrument to detect life on Mars or Europa, you cannot assume alien life will use the same chemistry. You need a way to recognize life that does not depend on recognizing specific molecules. This is where assembly theory comes in, a relatively new framework that asks a different question entirely: not “what is this made of?” but “how hard is this to make?”
The idea is that certain molecules are so complex that they are extremely unlikely to form through random chemical processes. The more steps required to assemble a molecule from simple building blocks, the higher its “molecular assembly” number. Above a certain threshold, the argument goes, the molecule’s existence implies that some selective, information-driven process, something life-like, produced it.15Nature Communications. Identifying molecules as biosignatures with assembly theory and mass spectrometry The approach has been tested using mass spectrometry on Earth samples and shows promise as a biosignature that works regardless of what elements or biochemistry a living system uses.16PubMed Central. Molecular assembly as a universal biosignature measurable by mass spectrometry
Assembly theory does not define life directly. Instead, it sidesteps the definition problem by focusing on the detectable products of living processes. If a rock on another planet contains molecules too complex to have formed by accident, something must have made them, and that something is what we would want to investigate further. It is a pragmatic workaround for a philosophical impasse.
Expanding the Genetic Toolkit
All known life on Earth uses the same four-letter genetic alphabet in DNA. But that is not necessarily a requirement of life itself. Researchers have created organisms that incorporate unnatural base pairs into their DNA, expanding the genetic alphabet beyond the familiar four. These semi-synthetic organisms can store and retrieve genetic information using letters that evolution never invented.17PubMed Central. Expansion of the Genetic Alphabet: A Chemist’s Approach to Synthetic Biology The organisms are alive by any reasonable standard: they grow, replicate, and maintain themselves. But their existence raises the question of whether life on another planet might use a completely different informational system. If life is defined by its chemistry, these engineered organisms stretch that definition. If life is defined by its functions, they fit right in.
Some researchers have even speculated about the possibility of a “shadow biosphere” on Earth itself: microbial life that uses fundamentally different biochemistry and has gone undetected because our methods are tuned to find only the familiar kind.18PubMed. Epistemological issues in the study of microbial life: alternative terran biospheres? No such life has been found. But the idea highlights how deeply our concept of “living” is shaped by the one example of life we know, and how a broader definition might be needed if that example turns out to be less universal than we assume.
Where Does One Living Thing End and Another Begin?
Even among organisms everyone agrees are alive, the boundaries of a living individual are murkier than they appear. You carry trillions of microbial cells that are not genetically “you” but are functionally essential. Many metabolic pathways in animals and plants depend on symbiotic microorganisms to work properly, to the point where the host organism alone is physiologically incomplete.19PubMed. A symbiotic view of life: we have never been individuals This has led to the concept of the “holobiont,” treating the host plus its persistent microbial partners as a single biological unit that develops, functions, and in some cases evolves together.20PubMed Central. What Is a Hologenomic Adaptation? Emergent Individuality and Inter-Identity in Multispecies Systems
This matters for the “what is living” question because it challenges the intuition that a living thing is a discrete, bounded entity. If your life depends on thousands of species of bacteria living in your gut, are those bacteria part of you or separate organisms? The answer affects how we think about biological individuality, immunity, and even identity. A coral reef is a community of animals, algae, and microbes so tightly integrated that removing any component can kill the whole structure. Is the reef alive, or only its parts? These are not trick questions. They reveal that “living” may be a property of systems and relationships, not just of individual bounded objects.
When Life Doesn’t Fully Stop
Death, the obvious opposite of life, turns out to be its own gray zone. After an organism dies, its organs, tissues, and individual cells do not all stop functioning at the same time. This is the principle behind organ donation: a heart or kidney can remain viable and functional after the organism it belonged to has been declared dead. Some cells continue to express genes and carry out metabolic processes for hours or even days after organismal death.21PubMed. Perspective on Death: A Gateway to a New Biology
This residual activity raises questions about what exactly dies when an organism dies. If individual cells are still metabolizing, responding to signals, and even dividing, are those cells alive? By the standard checklist criteria, yes. The organism as a whole is dead, but life at the cellular level persists for a time. It is another case where the scale at which you ask the question changes the answer. A dead human contains living cells, just as a living human contains dead cells (the outer layer of your skin, for example, is made of dead keratinocytes). Life and death, like living and nonliving, are less like a light switch and more like a dimmer.
Cognition Without Brains
A more recent strand of research asks whether some form of cognition, the ability to sense, learn, and make decisions, might be a fundamental feature of all living things rather than a special property of animals with nervous systems. Bacteria exhibit behaviors that look like memory, learning, decision-making, and anticipation, capacities that researchers in “basal cognition” treat as part of an adaptive toolkit shaped by evolution and present across the tree of life.22PubMed Central. On the prospects of basal cognition research becoming fully evolutionary: promising avenues and cautionary notes If this view holds up, information processing and rudimentary decision-making may be as central to what makes something alive as metabolism or reproduction. That would add another dimension to definitions of life, one that current checklists mostly ignore: not just what living things are made of or what they do chemically, but how they process and act on information about their environment.
Whether or not this framing becomes mainstream, it captures something that anyone who has watched a slime mold navigate a maze or a single-celled organism flee a toxin can appreciate. Living things do not just exist; they behave as if they have stakes in the outcome. That quality, difficult to formalize but easy to recognize, might be as close to a universal marker of life as anything on the checklist.
Virtual Life and the Question of Medium
Computer scientists and artificial-life researchers have pushed the question even further from familiar biology. If you simulate an evolving, self-maintaining system inside a computer, is it alive? Some researchers argue that the specific physical medium does not matter: if a system exhibits self-maintenance, agency, and open-ended adaptation, it meets the abstract requirements for life regardless of whether it is made of carbon, silicon, or code running on a server.23Philosophical Transactions of the Royal Society B. Towards origins of virtual artificial life: an overview
This position is controversial, and most biologists would reject the idea that a software simulation is genuinely alive. But it forces a useful clarification. If you insist that life must be chemical, you are making a claim about the medium. If you insist that life is defined by what it does, you open the door to non-chemical life. Neither position is obviously wrong, and the disagreement reveals how much of the definition problem comes down to whether we think life is a kind of stuff or a kind of process. Earth gives us only one example of life, and that example happens to be chemical. Whether that is a deep truth or an accident of our sample size is, for now, an open question.