It’s a sunny spring day and the birds are busy. Some are flying about, perhaps foraging. I notice one staring at me from beneath a bush. What does it want? Suddenly, the bird returns with a little stick in its beak, flies away, and disappears into a tree. Ah—it’s probably building a nest.
I have chosen birds, but any organism could have played the protagonist in this opening scene. There are fish in the Pacific that crack open clams by tossing them repeatedly against hard surfaces. In the microscopic world, some amoebas carry bacteria when moving to a new location—in order to farm them. I could go on indefinitely. My point is simple: we distinguish what is alive from what is inert by its intentional behaviour. That is the essence of biology—the study of life.
This may sound naive, anthropomorphic, or unscientific. ‘Surely,’ you object, ‘organisms don’t really behave intentionally. They only appear to. Their behaviour results from natural selection, not from some life-spirit.’ Indeed, many take the essence of biology to be evolution. I do not mean to dismiss Darwin’s insight—God forbid! But biologists do not study rocks or molecules; that is the domain of physicists, geologists, and chemists. Biologists study material entities that move and interact as if with purpose.
So what do we mean by intention, if not this peculiar capacity for end-directed behaviour? And where did our mental intentions come from, if not from mindless animals without brains? Could we, in some sense, have mindless intentions too?
Reevaluating the Notion of Intention in Biology
Modern science is characterised by unabashed materialism, and yet intentions are conceptualised as mental states—alongside beliefs and desires. In this view, if a bird cannot think something like ‘I am going to build a nest,‘ it cannot have intentions. It is said to merely react to stimuli—like a robot, albeit a complex one.
This view descends from an old philosophical and theological tradition, which filters biological observation through the lens of our linguistic subjectivity. It actually runs counter to the Darwinian revolution, which was meant to bring humans into continuity with animals—not elevate them with metaphysical exceptions. It also reveals that scientists lack a biological definition of language, relying instead on philosophical notions that treat words and minds as something magical.
But biology, properly understood, need not take sides in metaphysical debates about matter and spirit. Biologists need not attribute minds to organisms, but they have every reason to attribute intentions. Indeed, the idea that living beings are evolved machines is a philosophical assumption, not a biological one. To make this clear, let me turn to a fundamental feature of life: communication.
Defining Signals in Animal Behavior: A Biological Perspective
Let’s go back to the beginning of this piece, where it’s a sunny spring day and I hear the birds singing. The purpose of these acoustic signals is not as obvious as the purpose of flying about or picking up little sticks. It takes a little more observation and inference to conclude what such a display is for.
According to biological observation, bird song is for attracting females and driving away competitors. Similarly, a bower bird’s flashy nest is not actually for laying eggs; rather, careful observation suggests that this structure is for attracting females. With the present text, I, as an animal, must be intending something similar. While I do not deny that my human intentions are more elaborate, I think that they are better explained through their relation to those of other animals.
Hence, in my view, bird song, the present text and other such signals reflect an intention to affect other organisms, to change their behaviour. Of course, animals act and develop in many ways that end up affecting others. However, signals are characterised by their dependence on the receivers’ intention to respond: The male bird sings or builds a flashy nest only because there are females around looking to maybe respond by choosing it as a mate. These females may be occupied with food or a predator. They must also decide which signals they will respond to, which signaller they will choose. Receivers cannot and won’t always respond to signals, however, they have an intention and interest in doing so; otherwise, there would not be any point in signalling.
This conceptualisation distinguishes signals from two other types of acts: coercion and cues. In coercion, an animal tries to change the behaviour of others who do not intend to behave that way. For example, the males of many species will harass females until they allow copulation; alternatively, they will punish them for refusing to mate with them, which increases the chances that they will accept the coercer. This behaviour is clearly not one of signalling, even though it succeeds at changing the other’s behaviour. In cues, the opposite happens. The female bird’s nest, for example, is not aimed at anyone; however, it is a cue for the cuckoo, who certainly seeks to lay eggs inside.
In sum, signals are acts or structures through which an organism (the signaller) intends to change other organisms’ behaviour, and which are effective because those organisms (the receivers) intend to respond to them. This (my) definition seems consistent with many observations of animal behaviour. It allows us to recognise signals in the same way that we recognise 1) acts and structures not aimed at others, such as nest building and nests, 2) acts and structures aimed at others who do not intend to respond (coercion), and 3) acts and structures that others react to but were not intended for them (cues).
In contrast, Maynard Smith and Harper (2003) define a signal as ‘any act or structure which alters the behaviour of other organisms, which evolved because of that effect, and which is effective because the receiver’s response has also evolved.’ Notice how this definition replaces intention – a concept describing observable phenomena – for the abstract notion of evolution, which refers to the ultimate origin of such phenomena. But before looking at the reasons for this terminological preference, let’s see how it impairs the definition of signal.
The Pitfalls of a Traditional Definition of Signals in Biology
Consider the way in which lions cooperate to hunt down big prey. Each organism reacts to the actions of others in a coordinated fashion, for example, one lion moves towards the back because another has moved to the front. Smith and Harper’s definition leads us to conclude that the first action is a signal: 1) the move has altered the other’s behaviour, 2) it has evolved because of that effect, without which the hunt would not be successful, and 3) the move is effective because the receiver’s reaction (its ‘response’) has similarly evolved. By contrast, in my definition, the lions’ actions cannot be signals because their aim is the prey. This is why I wrote that the lions react to each other’s behaviour in hunting, whereas they respond to each other’s signals.
The fact that responses (to signals) are intentionally cooperative, yet directed or aimed at receivers, sets them apart from other similar behaviours, such as group hunting. The same applies to conflicting interactions: males who retreat after hearing a mighty roar clearly intend to retreat, even though they will sometimes choose to fight.
But Smith and Harper do not see it this way: ‘If one stag pushes another stag backwards, that is not a signal but coercion. If it roars and the other stag retreats, it is a signal because the response depends on the evolved properties of the brain and sense organs of the receiver’ (Maynard Smith & Harper, 2003, emphasis mine). Here, the authors refer to a red deer contest in which males roar at each other. If the roars are not effective, the stags resort to pushing each other until the loser retreats. However, retreating is not a mechanical effect of being pushed backwards; it is a decision to abandon the fight, i.e. an evolved reaction that ‘depends on the evolved properties of the brain.’ The same is the case when females finally yield to a male’s aggression in sexual coercion, as I mentioned above. However, this makes coercion the same as a signal in the authors’ definition.
Conclusion
This post will continue. But for now, let me say this: despite decades of research, biology lacks a proper definition of signalling—one of its most basic and crucial concepts. This failure stems from a philosophical obsession with ultimate causes, rather than a commitment to empirical clarity. To be truly scientific, biology must set aside metaphysical anxieties about the origin of life. It must return to its proper task: observing and explaining the behaviour of living things.
References
Maanmieli, J. (2021). Prescription: A biological definition of language. Alethes.net, 1(2).
Maynard Smith, J. & Harper, D.G.C. 2003. Animal Signals. Oxford University Press, Oxford.