There is no single, universally accepted scientific definition of life. Biologists, chemists, physicists, and philosophers have proposed well over a hundred definitions, and none has won consensus. The closest thing to an official standard is a working definition used by NASA: life is “a self-sustained chemical system capable of Darwinian evolution.”1PubMed. The Origin, Extension, and Future of the “NASA Definition” of Life That phrase is popular in the astrobiology community, but even its supporters acknowledge it leaves out things most people would call alive and arguably includes things they would not. The difficulty is not a failure of scientific effort; it reflects the fact that life, as we know it, is not a clean category with sharp edges.
The Checklist Approach
The most intuitive way to define life is to list what living things do. Biology textbooks have traditionally taught a version of this: living organisms metabolize, maintain internal stability, organize themselves, grow, adapt, process information, and reproduce.2PubMed Central. Biomolecular Basis of Life If something does all of these, it is alive. If it does none of them, it is not. The approach is straightforward, and it works well for everyday purposes. A cat clearly checks every box. A rock clearly does not.
The trouble with checklists is the borderlands. A mule metabolizes, maintains homeostasis, and grows, but it cannot reproduce. Fire consumes fuel, grows, and responds to its environment, yet nobody calls it alive. Crystals organize themselves into orderly structures and can even “seed” the growth of new crystals, which superficially resembles reproduction. A checklist either has to be made so specific that it becomes a description of Earth biology rather than a general principle, or it stays general enough that non-living systems keep sneaking in. These edge cases are not trivial academic puzzles. When scientists design instruments to look for life on Mars or Europa, they need criteria that will work even if the life they find does not look like anything on Earth.
The NASA Definition and Why It Dominates
In the early 1990s, a committee of scientists working with NASA settled on a more compact approach: life is “a self-sustaining chemical system capable of Darwinian evolution.”3PubMed Central. Defining life The phrase packs two requirements into one sentence. First, the system has to sustain itself chemically, meaning it takes in energy and matter from its surroundings and uses them to maintain its own existence. Second, it has to be capable of evolving through natural selection, which means it can copy itself imperfectly, introduce variation, and allow that variation to be filtered by the environment over time.
The appeal is its generality. It does not mention DNA, cells, carbon, or water. In principle, it could apply to a biochemistry completely unlike our own. That makes it especially useful for astrobiology, where you want a definition broad enough to catch alien life without assuming it mirrors terrestrial biology. But the definition has real weaknesses. A single rabbit is alive, yet one rabbit alone cannot undergo Darwinian evolution; it needs a population. A sterile organism cannot reproduce at all, let alone evolve, but calling a mule “not alive” feels absurd. And some computer programs can already self-replicate and evolve, which raises the awkward question of whether they count.
Life as Self-Making
An alternative framework focuses not on evolution but on self-production. The concept of autopoiesis, proposed in the 1970s, defines a living system as one that continuously produces the components that make up its own boundary and internal structure.4PubMed. Autopoiesis: a review and a reappraisal A cell, for example, synthesizes its own membrane, its own enzymes, and the molecules it needs to keep doing all of this. The network of chemical reactions inside the cell is both the producer and the product. Under this view, the defining feature of life is not replication or evolution but autonomous self-maintenance.
Autopoiesis offers a concrete, pragmatic way to think about what separates the simplest living cell from a complex but non-living chemical reaction. It has been refined over the decades into a criterion for minimal life that connects to related ideas about biological autonomy and self-organization.5PubMed. Autopoiesis 40 years later. A review and a reformulation. Its limitation is roughly the inverse of the NASA definition’s: it captures individual organisms well but says little about populations, evolution, or how life changes over time. A definition of life that ignores evolution is leaving out a lot of what makes biology different from mere chemistry.
The Thermodynamic View
In 1944, the physicist Erwin Schrödinger asked “What is life?” in a short book that became enormously influential. His answer was thermodynamic: living things feed on “negative entropy,” importing order from their surroundings to maintain their own internal organization. The idea has been explored in detail for microbial growth, where careful analysis shows that organisms can rid themselves of internally generated entropy either by producing high-entropy waste products or by releasing heat, or both.6PubMed. Does microbial life always feed on negative entropy? Thermodynamic analysis of microbial growth
This line of thinking connects to the broader concept of dissipative structures, a term coined by the chemist Ilya Prigogine. These are organized patterns that arise and persist precisely because energy flows through them and gets dissipated as entropy. A hurricane is a dissipative structure, and so is a living cell.7PubMed Central. Dissipative Structures, Organisms and Evolution The thermodynamic perspective reframes life not as a special substance but as a particular kind of process: a far-from-equilibrium system that channels energy to sustain and reproduce its own complex organization. It is a powerful insight, but by itself it casts too wide a net. Flames, convection cells, and tornadoes are all dissipative structures, yet they are not alive by any ordinary understanding.
Life as Information Processing
Another school of thought emphasizes information rather than chemistry. On this view, the fundamental characteristic of life is that it processes information at every level of organization, from molecular interactions up to entire ecosystems. Memory is maintained not just in DNA but in protein states, ecological relationships, and developmental patterns. The “purpose” of all this information processing, to the extent you can use that word, is self-perpetuation: the system processes information in order to keep processing information.8PubMed. Living is information processing: from molecules to global systems
This framing has gained traction partly because it bridges biology and computation. It also aligns with the observation that even organisms without nervous systems engage in sophisticated sensing, decision-making, and adaptation. Research on basal cognition has shown that bacteria, single-celled eukaryotes, and plants all exhibit capacities usually associated with “higher” organisms: perception, memory, valence, learning, and communication.9PubMed Central. Reframing cognition: getting down to biological basics If even the simplest life forms are processing information and making context-dependent decisions, then information processing starts to look like a genuinely universal feature of living systems rather than an emergent property of brains.
Viruses and the Boundary Problem
No discussion of “what is life” gets very far before someone asks about viruses. They have genetic material, they evolve rapidly, and they hijack cellular machinery to make copies of themselves. But they lack metabolism, they do not maintain homeostasis on their own, and outside a host cell they are inert particles. Are they alive?
The honest answer is that the question depends entirely on which definition you pick, making it more of a definitional dispute than a biological one. Within a framework that places all biological replicators on a continuum from completely selfish to fully cooperative, typical lytic viruses sit at the selfish extreme, while temperate viruses and mobile genetic elements occupy the middle ground.10PubMed Central. Are viruses alive? The replicator paradigm sheds decisive light on an old but misguided question. In other words, rather than drawing a hard line between “alive” and “not alive,” you can think of viruses as occupying one end of a biological spectrum.
Giant viruses have made this question even harder to resolve. Over the past couple of decades, several families of enormous viruses have been discovered with genome sizes, gene contents, and physical dimensions that overlap with those of small cellular organisms.11FEMS Microbiology Reviews. The rapidly expanding universe of giant viruses: Mimivirus, Pandoravirus, Pithovirus and Mollivirus Some of these giant viruses encode proteins once thought to be exclusive to cells, including components of the translation machinery that reads genetic code to build proteins.12Proceedings of the National Academy of Sciences. Why did some viruses evolve to be giants while others did not? They still need a host to replicate, but genetically and physically, they look more cell-like than anyone expected. Giant viruses have become one of the strongest arguments that life and non-life are not separated by a bright line.
Prions and Self-Replication Without Genes
Viruses at least have nucleic acid genomes. Prions push the boundary further. These are misfolded proteins that propagate by converting normally folded copies of the same protein into their own aberrant shape.13PubMed Central. Prion disease and the ‘protein-only hypothesis’ No DNA, no RNA, no genetic code in the traditional sense. Unlike nucleic-acid-based replication, which requires new polymers to be synthesized, prion propagation works through structural transitions in existing protein molecules.14PubMed Central. Prion propagation: the role of protein dynamics
Nobody seriously argues that a prion is alive. But prions do replicate, and in some yeast systems they even show a rudimentary form of heritability and variation. They are a reminder that self-replication, often treated as the hallmark of life, can be achieved by remarkably simple mechanisms that fall well short of what we would call living. Any definition of life that rests heavily on self-replication needs to explain why prions do not qualify.
The Third State Between Life and Death
If viruses challenge the definition from the “maybe not alive enough” direction, some organisms challenge it by appearing to shut down life’s machinery entirely. Tardigrades, the microscopic creatures sometimes called water bears, can enter a state called cryptobiosis when their environment becomes hostile. During cryptobiosis, measurable metabolism drops to undetectable levels, sometimes described as a reversible standstill.15PubMed. New insights into survival strategies of tardigrades This state has been called a third condition between life and death: the organism is not metabolizing, not growing, not responding to stimuli, yet it is not dead, because when favorable conditions return, it resumes all of those functions.16PubMed. Cryptobiosis: a new theoretical perspective
Cryptobiosis is not unique to tardigrades. Certain bacteria, fungi, brine shrimp, and plant seeds enter similar states of suspended animation. These cases are inconvenient for any definition that requires ongoing metabolism or active self-maintenance. A desiccated tardigrade is not metabolizing. An autopoietic definition would struggle to call it alive in that moment. Yet it clearly is not the same as a dead tardigrade, because a dead one will never revive. The simplest resolution is to say that life is better understood as a property of lineages and histories than of instantaneous states: the tardigrade is alive because it is part of a continuous living process, even though that process is temporarily paused.
What Is the Minimum Requirement for a Living Cell?
Synthetic biology has tried to answer the definition question from the engineering side: instead of debating what life is, build the simplest possible version and see what you need. In 2016, researchers designed and synthesized a minimal bacterial genome containing just 473 genes, packed into about 531,000 base pairs of DNA.17Science. Design and synthesis of a minimal bacterial genome The resulting organism, called JCVI-syn3.0, was a functioning, self-replicating cell, but even its creators could not assign a clear function to roughly a third of those 473 genes. We know they are essential because removing them kills the cell, but we do not know what they do.
This is a humbling finding. Even the simplest possible cell we can build contains deep mysteries about what is actually needed to be alive. It also provides a kind of lower bound: you probably cannot get a free-living cell to work with substantially fewer than a few hundred genes. Below that threshold, the system cannot maintain itself and reproduce. Whether some entirely different chemistry could achieve the same result with less remains an open question.
Digital Organisms and Artificial Life
Computer scientists have built self-replicating programs that evolve, compete, and form ecological networks resembling those of biological organisms. In platforms designed for this purpose, digital organisms start from a single ancestor and, through mutation and selection, diversify into webs of interacting species that experience competition, predation, parasitism, and mutualism.18PubMed Central. Evolving digital ecological networks These programs evolve genuine adaptations: under high mutation rates, for instance, digital organisms restructure their genomes to become more robust, tolerating a larger fraction of neutral mutations by making their instructions more independent of one another.19Artificial Life. Evolution of Robustness in Digital Organisms
Do these programs satisfy the NASA definition? They self-replicate, they mutate, and they undergo Darwinian evolution. They are not chemical systems, though, so a strict reading of “self-sustained chemical system” would exclude them. Whether that exclusion reflects something meaningful about life or just a limitation of the definition depends on your philosophy. The existence of digital organisms is one reason some researchers argue that definitions of life should be process-based rather than substance-based: it may be more productive to ask what life does than what life is made of.
Searching for Life We Would Not Recognize
Every definition of life carries an implicit risk: if it is too closely tied to Earth biology, it might cause us to miss genuinely alien forms of life. This worry has led astrobiologists to develop “agnostic biosignatures,” which are signs of biology based on molecular complexity rather than on the specific molecules found in Earth organisms.20Annual Review of Earth and Planetary Sciences. Agnostic Biosignatures: Expanding the Search for Life in the Solar System The idea is that any living system will generate molecules with a degree of chemical complexity and pattern that is statistically unlikely to arise from non-biological processes alone, regardless of whether those molecules happen to be amino acids and nucleotides.
Silicon-based life is the classic thought experiment here, and it has been explored rigorously. Silicon sits just below carbon on the periodic table and shares some chemical properties, making it an obvious candidate for an alternative biochemistry. However, detailed analysis across a range of planetary solvents paints a discouraging picture: in water, silicon’s chemistry is severely limited by its tendency to form inert silica; in extremely cold solvents like liquid nitrogen, essentially nothing dissolves at all. Sulfuric acid turns out to support a surprisingly wider range of organosilicon chemistry than water does, but even there, a biochemistry primarily built around silicon does not look feasible.21PubMed Central. On the Potential of Silicon as a Building Block for Life Carbon’s dominance in biology appears to be more than a historical accident; it reflects genuine chemical advantages that would be hard to replicate with other elements.
Where Life May Have First Emerged
The question of what life is intertwines with how it started. One influential hypothesis places life’s origin at alkaline hydrothermal vents on the ocean floor, where warm, mineral-rich, alkaline fluid meets cooler, acidic, carbon-dioxide-rich seawater. The natural chemical gradient across the porous walls of such vents could have provided a ready-made energy source, essentially a geological version of the proton gradients that living cells use today to generate chemical energy.22PubMed Central. On the origin of biochemistry at an alkaline hydrothermal vent The argument is that the ability to tap into a pre-existing gradient is older than the ability to generate one from scratch.23PubMed. Hydrothermal focusing of chemical and chemiosmotic energy, supported by delivery of catalytic Fe, Ni, Mo/W, Co, S and Se, forced life to emerge.
Under this scenario, the first proto-cells harnessed geochemical energy at the vent–ocean interface before they developed genetically encoded metabolic pathways. Only later did they learn to generate their own proton gradients, a transition that was necessary for life to spread beyond the immediate vicinity of the vent.24BioEssays. How did LUCA make a living? Chemiosmosis in the origin of life If this model is correct, the earliest life was barely distinguishable from the geology that sustained it, and the line between living chemistry and non-living chemistry was even blurrier at the start than it is now.
A separate hypothesis, the RNA World, focuses on the emergence of a self-replicating molecule that could both copy itself and mutate, and therefore evolve toward more efficient copying.25Cold Spring Harbor Symposia on Quantitative Biology. 9 Introns and the RNA World These two ideas are not necessarily in conflict: the vent hypothesis describes the energy setting, while the RNA World describes the informational chemistry. But they do emphasize different things as the key threshold of life. The vent model says energy flow came first; the RNA model says self-replication did.
Why Some Scientists Think Defining Life Is a Mistake
Given the persistent lack of consensus, a number of researchers have argued that trying to define life is a fundamentally misguided project. Their reasoning often comes down to the idea that “life” may not be a natural kind with a fixed essence. Just as the color “red” does not have a precise wavelength boundary, life may not have a clean set of necessary and sufficient conditions. One philosophical response to this objection is to treat life as a non-essentialist natural kind: a real pattern in the world, but one whose boundaries are fuzzy and whose members share family resemblances rather than a single defining feature.26Quaderns de Filosofia. Defining Life as a Non-Essentialist Natural Kind
Historically, these conceptual struggles have a long pedigree. Before the chemical revolution showed that organic and inorganic matter obey the same physical laws, vitalism held that living things were animated by a special vital force absent from non-living matter. The demolition of vitalism, and later Pasteur’s overthrow of the ancient idea that life arises spontaneously from non-life, established that biology is continuous with chemistry and physics.27Frontiers in Synthetic Biology. Historical paradigm shifts in defining life: from spontaneous generation and vitalism to the Pasteurian Wall and the quest for artificial creation That very continuity is part of what makes a clean definition so hard: if life follows the same rules as everything else, there is no separate “life force” to point to.
Life as a Planetary Phenomenon
One more scale worth considering is whether life can be meaningfully defined at the level of an entire biosphere rather than individual organisms. The original Gaia hypothesis suggested that Earth’s biosphere acts as a kind of self-regulating system, maintaining environmental conditions favorable for life over geological timescales. Critics pointed out that natural selection requires populations with variation, reproduction, and heredity, and a single biosphere is not part of such a population. More recent defenses have argued that selection can act on persistence alone, without reproduction, and that a single lineage can undergo sequential selection over time.28Proceedings of the National Academy of Sciences. Function, chance and purpose in the biosphere: a critical examination of the Darwinized Gaia hypothesis Whether you find this convincing or not, the debate highlights that the scale at which you ask “what is alive” matters enormously. Individual cells, organisms, species, ecosystems, and entire biospheres each show different combinations of the properties we associate with life, and the definition that works at one scale may fail at another.