Version classiqueVersion mobile

Behaviour, Development and Evolution

Patrick Bateson

1. Appearance of Design1

Texte intégral

  • 1 This chapter is taken in part from Bateson, P. & Martin, P. (1999), Design for a Life London: Jonat (...)
  • 2 Paley, W. (1802), Natural Theology. London: Faulder.
  • 3 Paley illustrated his idea of relations by considering the various features of the mole :
    ‘The stron

1In everyday life design implies a beneficial means to an end. The idea of design has been central to much discussion in biology. Bishop William Paley in the early nineteenth century wrote about the reactions of a person discovering a watch on a mountainside, pondering on how it was made.2 Paley wrote: ‘It is the suitableness of these parts to one another; first, in the succession and order in which they act; and, secondly, with a view to the effect finally produced’. Paley emphasized how different parts of an animal’s body relate to each other and contribute to the whole.3 He regarded the design he saw everywhere in nature as proof of the existence of God. These days few biologists would try to pin their religious faith, if they have any, on biological evidence, and the apparent design to which Paley referred would be attributed instead to the evolutionary mechanism which Charles Darwin called natural selection.

Charles Darwin in old age. Line drawing adapted from a photograph from Life and a Letters of Charles Darwin (1891). Wikimedia,​wiki/​File:Charlesdarwin-portrait-sitting-on-chair-sketch.png, Public Domain.

2Darwin’s theory of evolution by natural selection is universally accepted among serious biologists (except for a few so-called creationist scientists), even if arguments continue over the details. Darwin proposed a three-stage cycle that starts with variation in the form and behaviour of individuals. In any given set of environmental conditions some individuals are better able to survive and reproduce than others because of their distinctive characteristics. The historical process of becoming adapted notches forward a step if the factors that gave rise to those distinctive characteristics are inherited in the course of reproduction. Suppose, for example, that an individual bacterium happens to have heritable characteristics that make it resistant to an antibiotic. While all the others are killed by the antibiotic, this one will survive and multiply rapidly. Before long, the world is full of antibiotic-resistant bacteria. Darwinian evolution requires no unconscious motives for propagation — let alone conscious ones.

3Biologists should not write evolutionary accounts in which the past is seen as leading purposefully towards the goal of the present blissful state of perfection. A clear distinction is necessarily and wisely drawn between the present-day utility (or function) of a biological process, structure or behaviour pattern, and its historical, evolutionary origins. Darwin noted, for example, that while the bony plates of the mammalian skull allow the young mammal an easier passage through the mother’s birth canal, these same plates are also present in the mammals’ egglaying reptilian ancestors. Their original biological function clearly must have been different from their current function.

4The distinction between current function and historical evolution is all the more necessary because current adaptations may result from the experience of the individual during its lifetime. Human hands form calluses to protect against mechanical wear, and muscles develop in response to the specific loads placed upon them during exercise. Behaviour, in particular, becomes adapted to local conditions during the course of an individual’s development, whether through learning by trial and error or through copying others. These are all examples of adaptations that are acquired during the lifetime of the individual, and they are clearly distinct from adaptations that are inherited.

  • 4 Tinbergen, N. (1963), On aims and methods of ethology. Zeitschrift für Tierpsychollogie, 20, 410–43 (...)

5An important advance in thinking was made by the Nobel Prize winner, Niko Tinbergen.4 He pointed out that a number of fundamentally different types of question may be asked when studying behaviour. ‘How does it work?’ ‘How did it develop?’ ‘What is it for?’ and ‘How did it evolve?’ In the case of fully-formed behaviour, questions to do with control and function are current, whereas questions to do with evolution and development are historical.

6Tinbergen’s distinctions can be illustrated with a commonplace example. Suppose drivers are asked why they stop their cars at red traffic lights. One answer would be that a specific visual stimulus—the red light—is perceived, processed in the central nervous system and reliably elicits a specific response (easing off on the accelerator, applying the brake and so on). This would be an explanation in terms of the way in which the traffic light controls the behaviour of drivers. A different but equally valid answer is that individual drivers have learnt this rule by past observation and instruction. This is an explanation in terms of development. A functional explanation is that drivers who do not stop at red traffic lights are liable to have an accident or, at least, be stopped by the police. Finally, an ‘evolutionary’ explanation would deal with the historical processes whereby a red light came to be used as a universal signal for stopping traffic at road junctions. All four answers are equally correct, but they reflect four distinct levels of enquiry about the same phenomenon. To use the traffic lights example once again, their efficient regulation of the behaviour of drivers suggest that they have been designed by human agency—undoubtedly correct in this case.

7The perception that behaviour is designed springs from the relations between the behaviour, the circumstances in which it is expressed and the resulting consequences. The closeness of the perceived match between the tool and the job for which it is required is relative. In human design, the best that one person can do will be exceeded by somebody with superior technology. If you were on a picnic with a bottle of wine stoppered with a cork but had no corkscrew, one of your companions might use a strong stick to push the cork into the bottle. If you had never seen this done before, you might be impressed by the choice of a rigid tool small enough to get inside the neck of the bottle. The tool would be an adaptation of a kind. Tools that are better adapted to the job of removing corks from wine bottles are available, of course, and an astonishing array of devices have been invented. One ingenious solution involved a pump and a hollow needle with a hole near the pointed end; the needle was pushed through the cork and air was pumped into the bottle, forcing the cork out. Sometimes, however, the bottle exploded and this tool quickly became extinct. As with human tools, what is perceived as good biological design may be superseded by an even better design, or the same solution may be achieved in different ways.

8The proposition that living organisms’ bodies, brains and behaviour were adapted over the course of evolution and by their suitability to the conditions in which they live is familiar to most non-biologists. An adaptation is a characteristic of an organism that makes the organism better suited to survive and reproduce in a particular environment—better suited, that is, than if it lacked the crucial feature.

Design of machines

9Within an individual, as well as between individuals, different systems of behaviour are variable both in their development and in their organisation. Some insight into why this should be may be obtained by looking at machines. Tailoring a system to a specified job while building in flexibility is a problem that human designers of machines must face again and again. Robots with even simple forms of regulation do things that look remarkably life-like. Similarly, in a game like chess simple rules can generate games of great complexity.

  • 5 Bateson, P. & Martin, P. (1999), Design for a Life. London: Jonathan Cape, pp. 96–97.

10The difficult challenge for the designers of chess-playing computers is to beat the creativity, flair and imagination of a chess Grandmaster. IBM rose to the chess challenge and started its Deep Blue project in 1989.5 The Deep Blue computer relied on massive parallel arrays with dedicated hardware and software. It had 256 chess-specific processor chips operating in tandem, each capable of analysing up to three million chess moves every second. The whole array could process 50–100 billion moves in the three minutes allotted for each move. It was also equipped with an enormous database of Grandmaster games played in the previous century.

11In the initial stages of the project no attempt was made to mimic human thought. Without any ‘psychology’ to mess things up, the machine would never get tired or make a silly mistake. It would instead depend for its success on raw computing power and its enormous memory. In one second Deep Blue could search ahead through several hundred million possible moves, while its human opponent, the Russian Grandmaster and one-time World Champion, Gary Kasparov, could analyse only one or two. Kasparov himself admitted that quantity sometimes becomes quality. But he had the compensatory benefits of intuition, judgment and experience.

The World Chess Champion Garry Kasparov played successfully against IBM’s Deep Blue computer but was beaten by the next version, Deeper Blue. Photo by Jürg Vollmer (2009), Flickr,​photos/​maiakinfo/​3858951927, CCBY-SA2.0.

12Compared to computers, humans calculate slowly, but are good at recognising patterns. Chess Grandmasters are much better than novice players at remembering patterns of pieces from real games, but no better at remembering arbitrary patterns. Experience helps them to remember patterns that have meaning and link these with the sequences of moves that have the best pay-off in the long run. The surprising consequence is that humans see traps that lie beyond the search horizon of even an exceedingly fast computer.

13In 1996 Kasparov played Deep Blue in a six-game match. Kasparov lost the first game, but then put his human skill to good effect and went on to win the match. He was able to do this because he could adapt his strategy in response to what he discovered were weaknesses in his machine opponent. Deep Blue, on the other hand, could not respond to the overview of its human opponent. IBM rose to the new challenge. Deeper Blue, their 1997 successor to Deep Blue, was faster and smarter.

14In particular, it could modify its basic strategy between tournaments in response to the playing style of its human adversary. This time the machine managed, albeit with some difficulty, to win the next match against Kasparov.

15These chess matches emphasised how important adaptability is in such competitions. An interesting development has been the cooperation between machine and humans.6 Amateur chess players coached their laptop computers to explore deeply specific positions using human pattern recognition together with their computers’ computational power. The resulting combination overcame in competition the superior chess knowledge of grandmasters and the superior computational power of big computers. Average players with average machines beat the best players and the best machines.

16In more practical uses, such as the control of traffic flow by co-ordinating the switching of traffic lights or regulating speed limits, a capacity to adapt to new situations is desirable. Faced with novelty, such systems must not change everything at once. If they did, they would quickly collapse into chaotic malfunctioning. So, as with animals, buffering some aspects of the computer’s capacities from change is crucial. These essential capacities must continue to function in the same way despite radical changes in input. Life brings many requirements.

17Impressive though IBM’s Deeper Blue computer was, it was dedicated to one complicated but narrowly defined task—playing chess. Gary Kasparov may have met his match on the chess board, but he was able to do a great many other complicated things of which Deeper Blue was incapable. He could make decisions about chess matches and holidays that he would take years into the future. He could run a complicated social life and allocate time to his main biological appetites, none of which were shared with Deeper Blue. He could feel moved by patriotism or spiritual feelings. He could write books and enjoy music. From time to time, he doubtless reflected on his life and his character.

Conflicts in motivation

18Kasparov like every other human and every other animal, had many strands to his life. The systems that are involved in running each of these aspects sometimes seem to be semi-autonomous, usually functioning smoothly together but occasionally coming into conflict. Humans feel the conflict most strongly, perhaps, in times of war, when their craving for leadership and their identification with their own group, tribe or nation conflicts with their peace-time commitments and pleasures and, indeed, perception of their own self-preservation. But everybody feels the pull, on most days of even the most routine life, between incompatible activities. You can’t eat and sleep at the same time; you can’t have a warm shower and take a walk simultaneously—except perhaps during a cloud burst in the tropics.

19Much of animal and human behaviour and physiology operates on the basis that considerable autonomy has seemingly been designed into each behavioural system or organ. Interaction necessarily occurs between them to prevent total breakdown when the different parts pull in different directions. A problem of great interest to engineers has been how far machines should emulate biology, using specialised modules like those in the brain that are dedicated to particular jobs such as recognising faces. How far should the modules be built into separately organised systems, each competing for time when they cannot operate simultaneously? Should a ‘boss’ allocate priority where it is impossible for two activities to occur at the same time? Or should a decision to express a particular form of behaviour depend on weighting the needs of competing systems?

20Designers of intelligent machines often opt for distributed control, known as heterarchy7 (as opposed to hierarchy), because of the efficiency it brings. The solutions to the problems of running a smart machine, or an individual life, are also found in the management of human organisations. In contrast with traditional hierarchical bureaucracies, modern public and commercial institutions increasingly tend to have structures and organisational cultures that focus on tasks or projects. The emphasis is on getting the job done efficiently, and this is achieved by bringing together groups of people with the right knowledge and skills. Expertise and teamwork are what counts, rather than formal status. The organizational structure tends to be a matrix of project teams rather than a traditional top-down hierarchy. Such management relies on great flexibility and considerable autonomy for each part of the organisation, with exchange of information and competition occurring when the well-being of the whole demands it. The central control over day-to-day work is minimal and the ways in which each team is set up depends on the need.

21How do animals achieve comparable solutions in the development and integration of their behaviour? The ultimate arbiter of priority in organising their own behaviour is reproductive success. The consequences of giving priority to this aspect of their biology are sometimes astonishing—at least when judged from a human perspective. The male emperor penguin brooding his mate’s egg over the Antarctic winter cannot be relieved by his mate because the growth of the ice shelf puts the sea and food beyond reach. So, in the interests of producing an offspring, he fasts for months—a feat any human would find impossible. Other potential solutions to this problem, such as shorter stints of brooding and trekking repeatedly across the ice shelf during the winter, presumably proved to be less successful. The penguins that fasted all winter were the ones whose ancestors had best survived with this adaptation. Examples like this emphasise how dependent is the organisation of behaviour on the ecology of the species. Differences between individuals in the processes of development are to be expected.


22The developmental progression from a single cell to an integrated body of billions of cells, combining to produce coherent behaviour, is astonishingly orderly. Just as animals grow kidneys with a specialised biological function, adapted to the conditions in which they live, so they perform elaborate and adaptive behaviour patterns without any previous opportunities for learning or practice. Particular behaviour patterns are like body organs in serving particular biological functions; their structure is often likely to have been adapted to its present use by Darwinian evolution and by their adaptability. It depends on the ecology of the animal. Their structure and behaviour develop in a highly coordinated and systematic way.

23From the standpoint of design, systems of behaviour that serve different biological functions, such as cleaning the body or finding food, should not be expected to develop in the same way. In particular, the role of experience is likely to vary considerably from one type of behaviour to another. In predatory species, such as cats, cleaning the body is not generally something that needs special skills tailored to local conditions, whereas capturing fast-moving prey requires considerable learning and practice to be successful. The osprey snatching trout from the water does not develop that ability overnight. Animals that rely upon highly sophisticated predatory skills, such as birds of prey, suffer high mortality when young and those that survive are often unable to breed for several years because they have to hone their skills before they can capture enough prey to feed offspring. In such cases, a combination of different developmental processes is required in order to generate the highly tuned skills seen in the adult.

24Retaining the concept of design brings with it insights that biologists might well not have had without it. Even so, its use generates an unforeseen problem in the current world. The pre-Darwinian ideas about intelligent design have been taken up by the creationists in their attempts to disguise their beliefs as a form of science. As a result all sorts of unpalatable associations are brought up in the minds of biologists when they hear the word ‘design’. In attempts to make accessible complicated processes intelligible, various devices are used—like attributing metaphorical intentions to genes (see Chapter 8) or to the weather. These linguistic devices are easily misunderstood. The take home message is, then, that when using a term like design, which means different things to different people, great efforts must be made to ensure the language is not taken too literally. The design in biology is only apparent, but shorn of its religious connotations, understanding the relations between the parts of an organism remains as useful as ever.


1 This chapter is taken in part from Bateson, P. & Martin, P. (1999), Design for a Life London: Jonathan Cape, pp. 95-101.

2 Paley, W. (1802), Natural Theology. London: Faulder.

3 Paley illustrated his idea of relations by considering the various features of the mole :
‘The strong short legs of that animal, the palmated feet armed with sharp nails, the pig-like nose, the teeth, the velvet coat, the small external ear, the sagacious smell, the sunk protected eye, all conduce to the utilities or to the safety of its underground life. … The mole did not want to look about it; nor would a large advanced eye have been easily defended from the annoyance to which the life of the animal must constantly expose it. How indeed was the mole, working its way under ground, to guard its eyes at all? In order to meet this difficulty, the eyes are made scarcely larger than the head of a corking-pin; and these minute globules are sunk so deeply in the skull, and lie so sheltered within the velvet of its covering, as that any contraction of what may be called the eye-brows, not only closes up the apertures which lead to the eyes, but presents a cushion, as it were, to any sharp or protruding substance which might push against them. This aperture, even in its ordinary state, is like a pin-hole in a piece of velvet, scarcely pervious to loose paricles of earth. Observe then, in this structure, that which we call relation. There is no natural connection between a small sunk eye and shovel palmated foot. Palmated feet might have been joined with goggle eyes; or small eyes might have been joined with feet of any other form. What was it therefore which brought them together in the mole? That which brought together the barrel, the chain, and the cogs, in a watch—design; and design, in both cases, inferred from the relation which the parts bear to one another in the prosecution of a common purpose. …
In a word; the feet of the mole are made for digging; the neck, nose, eyes, ears, and skin are peculiarly adapted to an underground life; and this is what I call relation’.

4 Tinbergen, N. (1963), On aims and methods of ethology. Zeitschrift für Tierpsychollogie, 20, 410–433. An appreciation and an update of Tinbergen’s thinking half a century later is given in Bateson, P. & Laland, K. (2013), Tinbergen’s four questions: an appreciation and an update. Trends in Ecology & Evolution 28, 712–718.

5 Bateson, P. & Martin, P. (1999), Design for a Life. London: Jonathan Cape, pp. 96–97.

6 See Shyam Sankar, The rise of human-computer cooperation, TED talk (June 2012),


Table des illustrations

Légende Charles Darwin in old age. Line drawing adapted from a photograph from Life and a Letters of Charles Darwin (1891). Wikimedia,​wiki/​File:Charlesdarwin-portrait-sitting-on-chair-sketch.png, Public Domain.
Fichier image/jpeg, 36k
Légende The World Chess Champion Garry Kasparov played successfully against IBM’s Deep Blue computer but was beaten by the next version, Deeper Blue. Photo by Jürg Vollmer (2009), Flickr,​photos/​maiakinfo/​3858951927, CCBY-SA2.0.
Fichier image/jpeg, 38k


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search