In 1976, two Viking landers scooped Martian soil and fed it nutrients tagged with radioactive carbon. Something in the soil released gas, much as living microbes would. Yet another instrument found no organic molecules at all. Nearly fifty years later, scientists still debate what Viking actually detected.
The ambiguity was more than technical. It exposed a deeper problem: biologists study life in extraordinary detail, yet they have never agreed on what life is. NASA's working definition calls it a self-sustaining chemical system capable of Darwinian evolution. It sounds precise, but viruses, mules and even fire expose its seams. Why has defining life proved harder than studying it, and does the search for life elsewhere depend on getting the definition right?
The Borderline Cases
Most attempts to define life work by listing criteria: metabolism, growth, reproduction, response to the environment, evolution. Each seems reasonable on its own. The trouble appears when we test these criteria against real cases. Every one of them either lets in things we would not call alive or shuts out things we clearly would.
Consider fire. It consumes fuel, grows, spreads to new places and reacts to its surroundings. By a metabolic checklist, it qualifies. Crystals grow and pass their structure on to new layers. Prions, the misfolded proteins behind diseases like mad cow disease, spread by converting normal proteins into copies of themselves, and some strains can outcompete others over time. Viruses evolve rapidly, yet have no metabolism of their own and reproduce only inside host cells.
The reproduction criterion fails in the opposite direction. A mule cannot reproduce, and neither can a worker bee, yet nobody doubts that either is alive. NASA's definition avoids this by applying evolution to populations rather than individuals. But that raises a new question: is life a property of organisms, of lineages, or of whole systems? The pattern is revealing. We classify clear cases with confidence, but we cannot state the rule we are using to sort them.
TakeawayWhen every proposed rule fails at the edges, the solution may not be a better rule. The concept itself may rest on something we have not yet understood.
Definitions or Theories
Philosopher Carol Cleland argues that the search for a definition of life is misguided, at least for now. Her analogy comes from the history of chemistry. Before modern chemistry, people identified water by its observable properties: it was wet, clear, tasteless and dissolved many substances. These features worked most of the time, but they also misled. Some clear liquids were not water, while muddy river water still was.
What settled the matter was not a cleverer definition but a theory. Once molecular chemistry explained what substances are made of, water could be identified as H2O. That theory explained why water usually has its familiar properties, and why some samples look unusual. In Cleland's view, a definition built only on surface features is premature. Understanding the underlying nature of a thing has to come first.
Life, she suggests, is where water was before chemistry. We list its observable features because we lack a general theory of what living systems are. The obstacle is that all known life on Earth descends from a common ancestor and shares the same basic chemistry of DNA and proteins. We have, in effect, a sample of one. Generalising from it is like trying to understand all mammals by studying only zebras.
TakeawayDefinitions often come at the end of inquiry rather than the beginning. A precise definition is something a mature theory earns, not something investigation needs before it can start.
Searching Without a Definition
If we cannot say exactly what life is, how do we design instruments to find it? Astrobiologists have largely shifted from looking for life itself to looking for biosignatures: chemical, structural or atmospheric traces that are hard to explain without biology. Oxygen alongside methane in a planet's atmosphere, for instance, suggests something is constantly replenishing gases that should react with each other and disappear.
The weakness is that many biosignatures assume Earth-like biochemistry. To avoid that bias, some researchers pursue more agnostic strategies. One looks for molecular complexity: molecules so intricate, and so abundant, that ordinary chemistry is unlikely to produce them. Another, favoured by Cleland herself, looks for anomalies, patterns that resist explanation by known physical and chemical processes, and treats them as targets for closer investigation.
This shift has a Popperian flavour. Rather than checking a signal against a definition, scientists propose that it is biological and then try hard to explain it away. Viking's results show why this matters: without independent tests, a suggestive signal can remain ambiguous for decades. Life detection becomes less a matter of ticking boxes and more a process of ruling out alternatives until biology is the best remaining explanation.
TakeawayYou can search productively for something you cannot fully define, provided you are clear about what would count as evidence and what would rule it out.
The difficulty of defining life is not a failure of biology. It reflects how scientific concepts often develop: we recognise clear cases long before we understand what unites them. Viruses, mules and fire are not embarrassments but clues, marking where current understanding runs out.
Philosophical analysis clarifies what the search for alien life can achieve. Finding life with a separate origin would do more than tell us we are not alone. It could provide the second example we need to understand what life is.