We forget where we parked. We misplace our keys. We misread instructions. We lose track of the time. We call people by the wrong name. ‘To err is human,’ as the English poet Alexander Pope wrote in his Essay on Criticism (1711). But it is not exclusively human. All animals do things that prevent them from surviving, reproducing, being safe, or being happy. All animals get things wrong. Think of a fish that takes the bait and accidentally bites into a metal hook. Think of dogs that forget where they have buried their bones, or frogs that aim their tongues in the wrong direction. Birds build flimsy nests. Whales beach themselves. Domestic hens try to hatch golf balls.
But not everything in the Universe can make mistakes. While living things navigate a world filled with biological errors, the fundamental building blocks of the cosmos adhere to the laws of physics with unwavering consistency. No one ever caught an electron erring, let alone an atom, a sodium ion, a lump of gold, a water droplet or a supernova. The objects that physicists study, the pure objects of physics, do not make mistakes. Instead, they follow ineluctable laws.
And this is where a problem emerges. Mistake-making organisms, like everything else in the Universe, are made from law-abiding atoms and molecules. So where does mistake-making begin and end in living things? How deep does it go? Can the parts and subsystems of organisms, like immune systems or the platelets in blood, make mistakes too? And, if they do, is there something that connects human mistakes to those made by biological subsystems?
The answers to these questions have profound implications for how we think about life. If things go wrong only when physics becomes biology, biology might truly be irreducible to physics and chemistry, despite centuries of reductionism saying otherwise. It might also mean that organisms really have ‘correct’ goals and purposes that they can mistakenly deviate from – they really are teleological, despite a long history of mechanistic arguments claiming otherwise. And if life’s errors really are as ubiquitous as they appear, it might mean we need a ‘grand’ framework to explain what happens when things go wrong: a theory of biological mistakes.
As a philosopher, I have spent much of my life studying the puzzles of metaphysics and ethics. I have explored the nature of reality, the concept of being, and the moral implications of human action. But, in recent years, I have been working on the problem of mistakes with a team of researchers at the University of Reading in the UK. What drew us to this topic was our puzzlement about a major gap in the history of biological thought. Surprisingly, mistakes have mostly been ignored by researchers, even among biologists and philosophers of biology, and traditional definitions of life have largely overlooked the role of mistakes, focusing instead on successes, adaptations and beneficial mutations. That is why, in the late 2010s, our team began investigating how a more rigorous look at mistake-making could generate novel scientific hypotheses. How could mistakes be understood in a more systematic and interdisciplinary way, we wondered?
During the past few centuries, scholars and scientists have tended to focus on what goes right rather than what goes wrong. The idea of rightness in living things has taken many forms. In the 17th century, during the early days of the scientific revolution, René Descartes characterised animals as automata: ‘machines’ made of tissue that obeyed mechanical laws, like the movements of a clock. ‘No movement can take place,’ Descartes wrote, ‘either in the bodies of beasts, or even in our own, if these bodies have not in themselves all the organs and instruments by means of which the very same movements would be accomplished in a machine.’ The idea of automata implies that animals can only malfunction or break down rather than make mistakes – understanding the internal circuitry tells you all you need to know about the ‘right’ way for an animal to behave.
Two centuries later, a different view of biological rightness and wrongness emerged through the work of Charles Darwin. From a Darwinian perspective, whether something counts as a mistake can be assessed only in the cold light of evolutionary time, after a species has either perpetuated its lineage or died out. Organisms, according to the ‘standard’ view of evolution, are simply the product of blind natural selection working through the success of random genetic variation. In this case, the ‘right’ variation will lead to a species being more adapted to its environment, and more likely to survive, reproduce and continue evolving.
No matter how lofty or powerful a species may be in its environment, they all get things wrong
To understand how things can go wrong for animals outside of evolutionary time, the animal behaviourists of the 19th and 20th centuries placed renewed emphasis on the study of individual organisms. I am thinking of behaviourists such as B F Skinner, but also of ethologists such as Charles Otis Whitman, Oskar Heinroth, Konrad Lorenz and Nikolaas Tinbergen. Their writings contain examples of mistakes made by animals, such as gulls misidentifying eggs and ducklings attaching themselves to inanimate objects. Biologists influenced by Lorenz’s and Tinbergen’s seminal works now routinely investigate mistake-making in various forms. However, there is still no grand conceptual framework, no theory of mistakes, that might create an interface between philosophy and biology.
Thinking about mistakes gives us the right kind of orientation to understand ourselves and other organisms. It focuses our attention on the fact that living systems, from paramecia to people, are subject to normative standards of right and wrong. This can be explained simply: when living things operate in some ways, they do well; when they operate in other ways, things go badly.
Life is strewn with attempts to avoid, correct or minimise mistakes. Living things employ all kinds of strategies to keep themselves on the normative straight and narrow. It is no surprise that recent work by researchers such as Daniel Kahneman and Amos Tversky into human mistake-making has been so important and influential. We humans use ‘heuristics’ – mental shortcuts or rules of thumb – to judge situations, rank preferences, assess people and so on. Very often these heuristics serve us well (sometimes you can judge a book by its cover), but other times they lead us astray. No matter how lofty or powerful a species may be in its environment, they all get things wrong.
This is why our team has sought to develop a rigorous conceptual framework for thinking about mistakes and normativity. Such a framework will, we hope, help generate new, testable hypotheses for experimental biologists, and might shed light on many of the mistakes to which we humans are prone. But developing our framework has taken us in unexpected directions. We believe that mistake-making might illuminate the nature of life itself. It might show, once and for all, that biology is irreducible to the laws of physics and chemistry – atoms, remember, don’t make errors.
If organisms are bundles of atoms that obey fundamental physical laws, how do mistakes emerge? Alongside everything else in the Universe, we are also influenced by physical laws like gravity, but laws are not all that influence what organisms do. There is something else going on when bundles of atoms become living beings. It is called ‘biological normativity’.
Organisms are governed by norms of correct behaviour and, when they depart from these norms, they can get sick, fail to adapt, suffer, die or disintegrate. To avoid such fates, they mostly need to do what is right for them: they need to act at the right time and place, in the right circumstances, in the right way. The predator must get its timing correct, strike accurately, expend enough energy to subdue its prey without exhausting itself.
Physical laws alone cannot explain what is right or wrong for an organism because, in physics, all sequences of events are on a par – they are all treated equally. Consider the transfer of electrons from one molecule to another, known as the electron transport chain. This transfer is crucial for generating energy in most, if not all, living things – it keeps organisms alive and healthy. And, from a purely physicochemical perspective, the process of electron transport is always the same. What the historian Arnold J Toynbee said of history can be said of physics: ‘Just one damned thing after another.’ However, the process of electron transfer can go very wrong. A molecule that fails to transport electrons in the right way will cause mitochondrial dysfunction, leading to a diseased organism. So, not all instances of electron transport are equal. When it comes to life, some sequences are simply better than others because they promote health, integrity, survival. They promote flourishing. Different courses of action work for or against the organism in its environment. The wrong course of action is a mistake.
It is crude to identify the future state with the goal pure and simple. The goal cannot be a future state
This may sound obvious, but the ideas embedded in this definition are complicated and controversial. For some, claiming that the ‘wrong course of action is a mistake’ may reek of teleology, a concept that was virtually banned for much of the 20th century. The word comes from the Greek telos (meaning ‘end’ or ‘purpose’), which is the classical term for what is now more commonly called ‘goal-directedness’ or ‘purposiveness’. Invoking teleology was a serious problem for 20th-century biologists. In 1988, the German American evolutionary biologist Ernst Mayr took issue with the concept because he believed it involved the positing of mysterious backwards causes. How can future goals direct the present behaviour of organisms? As the biologist Colin Pittendrigh put it: ‘Biologists for a while were prepared to say a turtle came ashore and laid its eggs, but they refused to say it came ashore to lay its eggs.’ Saying the turtle came ashore with the goal of laying its eggs would suggest that, even while it was out in the ocean, the turtle was directed by a future state that pointed it toward the beach by mysteriously working backwards in time to influence its behaviour in the here and now. Backwards causation (obscure physics aside) is hard for most philosophers to accept – along with the rest of us. Surely, whatever explains the turtle’s egg-laying behaviour must be wholly in the here and now, and a product of past evolutionary processes.
Goals or purposes do indeed refer to future states an organism aims to be in, such as reproducing, surviving, adapting to the environment, being healthy or living in a well-functioning social group. However, it is crude to identify the future state with the goal pure and simple. The goal cannot be a future state.
I can have the goal of climbing Everest long before I ever set foot on the mountain. The same goes for all other organisms. Having a goal means making something real. It means pursuing something – whether that involves seeking food, shelter or a mate – rather than being pushed around in the present by a future state. The goals I am talking about here are those that are hardwired into organisms as drives, tendencies and dispositions to act in certain ways, such as being healthy or surviving. Goal-directedness must be present in the here and now while it aims an organism toward future states.
And none of this implies any requirement for awareness, let alone something as complex as self-consciousness.
In recent decades, the ‘ban’ on teleology has been lifted and some philosophers have been willing to take the concept seriously. But many, especially those influenced by the philosopher of science Ernest Nagel, will still insist that goal-directedness comes down to physics and chemistry. To these thinkers, there are no sui generis biological explanations for why organisms get things wrong. This is a reductive view that misunderstands how mistakes are made. To make this kind of error, an organism must depart from standards of correctness. It must do something wrong. And this normativity doesn’t come from physics or chemistry.
For reductionists, notions of ‘good’ and ‘bad’ are easily explained away by evolution. To these sceptics, normativity is no more than a numbers game: ‘badness’ simply appears when a species fails to produce enough offspring to adapt and perpetuate itself. ‘Goodness’ is the converse, appearing when a species successfully reproduces enough to generate the genetic variations needed to adapt and survive. Understanding ‘normativity’, then, only requires an understanding of how an organism contributes to its species’ fitness. Organisms either help their species adapt to their environment by successfully creating offspring, or they contribute to their species’ extinction by failing to reproduce. For the sceptic, who thinks the only meaningful ‘mistakes’ an organism can make are related to fitness, there are no ‘good’ or ‘bad’ actions – biological normativity doesn’t exist.
I don’t believe that this adequately explains mistake-making. Flourishing is not just about successful reproduction. It also involves doing things like catching prey or finding food better than the competition. It is because a bird builds the right kind of nest out of the right kind of materials in the right location that it can successfully raise offspring. Building a flimsy nest would be a mistake.
Normativity can exist, even if we have a poor understanding of what is good or bad for an organism
Would physical and chemical information be enough to predict what counts as a mistake for a given organism, like a nest-building bird? Even the French scientist Pierre-Simon Laplace’s vision of an all-knowing demon – an omniscient observer who knows, moment by moment, everything about the physical state of the Universe – would not be able to make an accurate prediction. Perfectly understanding an organism’s physical structures, body movements, sound emissions, nest-building skills or other features doesn’t allow us to predict which of its actions are correct and which are mistakes. Not everything can be reduced to physics and chemistry. Survival is more than a numbers game. Instead, we need to know how all this physics relates to action in the environment: we need to understand how an organism experiences the world. Is it flourishing? Is it healthy? Is it mentally and physically integrated? Is it literally happy with its situation (maybe not true for fungi or worms, but certainly so for dogs and zebras)?
There is, however, one lingering issue with taking mistakes and biological normativity seriously: values. It’s one thing to say that humans can act in a ‘good’ or ‘bad’ way, but can we really use those value-laden concepts to describe the behaviour of frogs or bacteria? Those who are sceptical of normativity in biology would say: ‘No.’ The philosopher of biology Justin Garson, for example, says that normativity has nothing to do with ‘values or goals, oughts and shoulds, prescriptions or commands, the good or the just’. If we take Garson’s argument seriously, then the malfunctioning of a dog’s heart is in no way literally bad for the dog, even though it may end up sick or dead. But is this correct? After all, things do not tend to go well for a dog with a bad heart. Things look pretty good, however, for a dog that has plenty of nutritious food, fresh air and fellow dogs to play with.
So, what does this have to do with things going biologically wrong? Well, we can happily use the term ‘value’ in the context of mistakes if we understand that something can be good or bad for an animal even if it doesn’t consciously value that thing. And that thing can be good or bad even if we don’t value it, either. Normativity can exist, even if we have a poor understanding of what is good or bad for an organism.
This is why mistakes cannot be banished from the conceptual toolbox of biology. And today, few biologists would ever seek to do so, as opposed to philosophers of biology in thrall to reductionism (or the idea of a total discontinuity between humans and other living beings). As we will see, biological mistakes open the door to a new and refreshing way of understanding living beings. Viewing living things through their mistakes is powerful because it provides a broad canvas within which to explore and scientifically study organisms. It also vindicates the special nature of biology.
However, once we accept the possibility of biological mistakes, a set of complex problems emerge: what distinguishes mistakes from other kinds of problems? And how do we locate and identify mistakes? Until now, we have discussed only familiar organisms, like birds, dogs and people. Normativity and mistake-making, however, seem to play a much more fundamental role in life on Earth.
Though the theory of biological mistakes involves several technical definitions of what it means to make such an error, the outlines are relatively simple: an organism makes a mistake when it does something that, if not mitigated in some way, will undermine its flourishing. We say that ‘mistakes happen’. But that’s not true. Mistakes are always made by individuals at specific times and places. This means that mistakes are not simply failures or malfunctions.
A failure is something that happens to you, not something you do. Being hit by lightning is not a mistake unless you ignore the weather warning and go for a stroll in the park during a thunderstorm. Horses and buffaloes can’t understand weather reports, so whenever they get struck it’s a mere failure. The same is true if they get attacked by a parasite that makes them sick or kills them – it’s just bad luck. A malfunction is similar. It is something that goes wrong with an organism’s biological functioning, like disease or deformity, but not something the organisms does.
Different kinds of biological mistakes are similar to each other in that they are all made, but that does not mean they are all exactly alike. One way that biological errors differ is in terms of their preventability: some mistakes are avoidable, others are unavoidable. Broody domestic hens, for example, will try to hatch golf balls or other egg-like objects that are left in their coop. They don’t do this because of some failure or malfunction but simply because they don’t have the perceptual equipment to distinguish eggs from things that look a fair bit like eggs. Their mistake is unavoidable because there’s nothing wrong with these hens. Avoidable mistakes, on the other hands, occur when an organism can act in ways that would help it thrive in a specific situation but fails to do so. Consider a buffalo, alert for predators. If an approaching lion is visible, but the buffalo is distracted, an attack would be an avoidable mistake.
An antibody misidentifying a pathogen is like you mistaking someone else’s mobile phone for your own
Whether avoidable or unavoidable, mistakes are always made. But who or what exactly can make these biological mistakes? As our research team looked closer, we found mistakes that were not limited to single organisms. A collection of organisms can also make a mistake – think of a flock of birds that flies into a skyscraper, or a beached pod of whales. Mistakes can also be made by parts of living things. Some of the most well-known examples involve DNA. Various mistakes can occur in the process of genetic transcription, translation and regulation, leading to cancer, genetic disorders, developmental issues or other problems. Another example is antibodies. Sometimes we get sick because our antibodies are fooled by deceptive pathogens that pretend to be part of our bodies. For example, the meningitis bacterium Neisseria meningitidis can mimic the appearance of the body’s cells and induce a particular part of the immune system to erroneously refrain from activating against it.
One puzzle for our research team is whether features of mistake-making are shared across all living things. On the surface, there is a chasm of difference between the mistakes of antibodies and of people, but might there be similarities? Consider two errors: an antibody misidentifying a pathogen, and you accidentally taking someone else’s mobile phone, mistaking it for your own. To do their work, antibodies respond to what mistake theorists call ‘markers’, which are prompts for action, like ‘cues’ but without the psychological connotation. Markers can take the form of receptors or shapes on the surface of the pathogen, which deceive the antibodies. But we also use markers in our daily lives. When you mistakenly take the wrong phone, you are responding to the colour, shape, size or position of someone else’s phone that may mimic your own. People and antibodies rely on markers to act because neither has the time nor energy to carefully inspect the entire target. This is an important but understudied area of mistakes. We still don’t fully understand these markers for action, but through them we might begin to classify the shared features of biological mistakes.
This is where mistake theory begins to make its boldest and most surprising claims: mistakes are made wherever there are living systems. They are a universal feature of biology. Our research team suspects that mistakes may even appear among the parts and subsystems of organisms. Consider one ‘part’ that our research team has been thinking about a lot: the haemostatic (blood-clotting) system. Blood clotting is a complex pathway of molecular activations involving tiny, disc-shaped cell fragments in our blood called platelets. And it seems to be a highly normative affair. If the process begins too late, an injured organism can bleed to death. If it begins too early, the organism can suffer a debilitating thrombosis as blood clots block veins or arteries. The process must occur in the right location, the site of injury. Clotting must also end at the right time for the same reasons. Platelets play a crucial role in this normative process.
When blood vessels are damaged, the collagen within them is exposed. Blood platelets activate when they are exposed to this collagen at an injury site. However, platelets can sometimes be activated by collagen that appears without the presence of an injured blood vessel. This can lead to thrombosis, with potentially deadly consequences for an organism. And there are many other ways that platelets can get things wrong: the clot they produce must be of the right size and shape to function correctly. Though platelets can potentially get many things wrong, is it possible for blood-clotting systems to make mistakes?
We know that platelets are activated by specific amino-acid sequences within collagen called ‘GPO triplets’. For the mistake theorist, this immediately raises the question as to whether GPO triplets are present in other proteins, or whether other protein sequences or post-translational modification might produce markers very similar to the GPO in collagen. Could blood platelets misidentify collagen? Could they even be activated by a collagen mimic? This could result in mistaken platelet activation – an activation caused by the wrong protein – with potentially disastrous consequences. Could platelets be fooled? We don’t yet know the answer. And there are further unknowns.
Spikes in the dopamine of zebra finches correlate in real time with fluctuations in song quality
Another example, which shows the uses of mistake theory and the possible depth of biological mistakes, is bird song. Each male zebra finch has a specific song they sing to court potential mates, and they teach this song to their male offspring. There is margin for fluctuation in the learned song – it needs to be a faithful reproduction, not a perfect copy. That means a true mistake would occur only if the learned song departed too much from the correct one. But how much is too much? How do zebra finches learn to sing the right song?
Research indicates that dopamine is released during zebra finch singing to keep their song at the correct pitch. With this knowledge, mistake theory can offer some testable hypotheses. According to our definition, a mistake is made only when the ‘departure’ undermines a finch’s flourishing. In this case, flourishing relates to the attraction of mates, which involves attracting enough of the right mates at the right time and so on, possibly over generations. (Flourishing is not merely a numbers game but, for most organisms, reproductive fitness and success is part of what it means for their lives to go well.) In experiments, spikes in the dopamine of zebra finches correlate in real time with fluctuations in song quality, suggesting a kind of evaluation is being made. The birds, without any awareness, seem to judge or calibrate the performance of their song based on changes in dopamine levels. They are responding to the correctness or incorrectness of their song. The bird will use auditory feedback as it adjusts its singing, but there seems to be something else going on: an evaluative function performed by the dopaminergic neurons themselves.
Perhaps the dopaminergic system has a representation of the correct song against which the actual song is compared, which would leave open the potential for getting things wrong. In this case, mistakes appear even among systems of neurochemicals. This is at the edge of what we know, but mistake theory can stimulate organised investigation into such phenomena.
The theory of biological mistakes appears to be a universal feature of biology, which demarcates the living from the realms of physics and chemistry, thereby rendering it irreducible to either. Despite this, mistakes are not yet subject to systematic investigation by biologists. Mistake theory is a framework within which to generate novel, testable hypotheses. And there are so many questions in need of systematic investigation: how can things go wrong in relation to timing, location, measurement, evaluation of quality and identification? How do organisms attempt to avoid mistakes? Which mistakes are unavoidable? How are they corrected? How does an organism monitor, in real time, whether it is deviating onto a pathway that will threaten its flourishing?
And then there are questions about the contradictory cases in which mistakes paradoxically help an organism in the long term despite threatening flourishing in the short term. This relates to the role of exploration or play in life. Organisms generally need to explore their environments, whether in search of food, or a mate, or shelter, and so on. However, too much exploration is wasteful and dangerous. It would be a mistake to allow too many mistakes, but some are required for us to flourish in our environments. Indeed, mistakes in DNA copying, for example, produce the variation that drives life’s diversity. But if these mistakes vary too much, systems fall apart. Interrogating these errors experimentally may give us a window into the phenomenon of biological normativity, helping us understand how organisms act correctly, or badly, in their environments.
Mistake-making is neither limited to organisms nor bound by scale. Mistakes can be made by the tiniest bacteria as well as the largest animals – even whole populations. They can also be made by non-organisms, such as platelets, antibodies and cells belonging to organisms. It is the ubiquity of mistake-making, as well as its potential, that demands an equally broad theory to organise investigation into the phenomenon.
Life is often defined by what we get right. It is explained by growth, replication and adaptation to the environment. But mistakes are everywhere. A theory of mistakes will help us understand, in a systematic and experimentally driven way, behaviour that threatens the flourishing of living beings. It will also help us appreciate the normativity that runs through life. While some still view ‘teleology’ with scepticism, mistake theory may well be the antidote that challenges conventional wisdom about the goals of living things. In the intricate biological dance of right and wrong, we might just find the key to understanding the deeper purposes that drive life on Earth.
This essay is based on Mistakes in Living Systems: A New Conceptual Framework for the Study of Purpose in Biology, a project supported by the global John Templeton Foundation research programme Agency, Directionality, and Function (grant no 62220). David S Oderberg was Principal Investigator, and team members include Jonathan Hill, Ingo Bojak, Jon Gibbins, François Cinotti and Christopher Austin. The opinions expressed in this article are those of the author and not those of the John Templeton Foundation.
