« A Capitalist Call For Socialized Medicine? | Gene Expression Front Page | Free is good »
February 22, 2004

No Pain, no Gain

In my post Changing the Subject I said I would come back to proposition D:

“Some subjective sensations, such as pleasures and pains, are adaptive traits of organisms.”

I don’t want to spend all day on definitions, but I must quickly say what I mean by subjective sensations. They are broadly what in recent philosophy have been called qualia, and include:

- the qualitative aspects of what we perceive by the traditional five senses, such as a patch of blue in the visual field, a buzzing sound, or the taste of a piece of lemon
- sensations appearing to be inside the body, such as pains, itches, tingles, etc
- mental images occurring in memory, imagination, or dreams
- sensations associated with emotional states, such as feelings of anger or

Of course, extreme Behaviorists would deny that some or all of the above exist, but I won’t waste time discussing that. The best reply to an extreme Behaviorist is like Dr Johnson’s refutation of Berkeley - except that you kick the Behaviorist, not a stone. [I can’t resist repeating an old joke: the Behaviorist says to his wife, after sex: ‘That was great for you, honey - how was it for me?’]

It can be plausibly argued that all of these categories of sensations have some effect on behaviour, and that they have evolved by natural selection for that purpose. For example, the feelings of anger help us win fights and resist aggression. However, I want to concentrate on the narrower category of pleasures and pains, which I interpret as including pleasant and unpleasant sense perceptions, such as delicious or disgusting tastes and smells.

The fundamental argument for the causal efficacy of pleasures and pains was well-expressed by William James (Principles of Psychology, Dover edn., vol. 1, p.143):

“It is a well-known fact that pleasures are generally associated with beneficial, pains with detrimental, experiences. All the fundamental vital processes illustrate this law”.

To take obvious examples, pain is usually associated with injury and illness. Extreme heat or cold are painful as well as biologically harmful. The unpleasant sensations of hunger, thirst, and suffocation are associated with deprivation of vital resources. Excrement and rotting food are bad to eat (for humans, but not for dung-flies) and also smell bad (to humans, but presumably not to dung-flies). Poisonous defensive chemicals produced by plants (or animals such as toads or millipedes) taste bitter or acrid. Conversely, moderate warmth feels pleasant, and nutritious foods such as meat, fruit, and sugars (beneficial during most of our evolutionary history) taste good. The most intense pleasure of all is associated with sex and therefore with reproductive fitness.

As William James recognised, there are apparent exceptions to the general rule. It is important to take account of these exceptions, and I discuss them more fully in the continuation. I conclude that there are very few genuine anomalies, and that most of the apparent exceptions positively support an evolutionary interpretation. It is not surprising, for example, that we do not feel pain from ionizing radiation, because it is not a hazard to which we have been exposed during our evolutionary history. It also seems clear that the correspondence of pleasures with biological benefits, and pains with biological harm, is too close to be explained by mere chance association.

How is this correspondence to be accounted for? The common-sense hypothesis is that the intrinsic pleasantness and unpleasantness of sensations causes us to seek out or avoid the circumstances giving rise to such sensations. For example, the pleasant taste of chocolate is one reason why we eat chocolate, and the pain of getting a finger caught in a door is one reason why we are careful in closing doors. (People who do not feel pain, such as lepers, frequently injure themselves).

So far as we know, there is no fundamental physical reason why pleasure is attached to some circumstances and pains to others. We might have evolved to find the sensation of having a finger caught in a door intensely pleasurable, just as dung-flies may well relish the smell of excrement. On the evolutionary hypothesis, the fact that pleasures are generally associated with beneficial, and pains with detrimental, experiences, is the result of natural selection among random variations: those individuals who happen to have an association of this kind have higher biological fitness than those who have no such association, or the reverse association.

This is obviously not the place to develop a full evolutionary theory of pain and pleasure. Unfortunately I don’t know of any satisfactory published treatment to refer to. I was very disappointed in the book Pain: the science of suffering by Patrick Wall, one of the world’s leading authorities on pain. It contains a great deal of fascinating information, but evolutionary considerations are conspicuous by their absence - another sign that in some respects Darwin’s time has still not come.

As I argued in my previous post, natural selection can only operate if it affects reproductive fitness. If proposition D is accepted, sensations must have causal efficacy in the physical world. But this is philosophically a very controversial position (which doesn’t worry me), and it may seem to conflict with physical laws such as the conservation of energy and momentum (which does worry me). So in my next post I will consider whether there are any viable alternatives to this hypothesis. I will also try to respond to comments.

Continuation: Exceptions and anomalies

If we classify sensations into the categories pleasant, unpleasant, and neutral, and the circumstances with which they are associated as beneficial, harmful, and neutral, then the following associations are appropriate:

pleasant sensation - beneficial circumstances
unpleasant sensation - harmful circumstances
neutral sensation - neutral circumstances.

All other combinations are inappropriate, and may be considered possible exceptions to William James’s general rule.

There are therefore six categories of possible exceptions to be considered.

1. Harmful circumstances associated with pleasant sensations

There are some obvious examples: notably addictive drugs such as alcohol, heroin and cocaine, which give pleasure but are harmful in the long run. (Tobacco addiction is slightly different, as tobacco seldom gives the addict much pleasure, but withdrawal is unpleasant.) Heavy consumption of fats and sugars is also pleasant but harmful.

These exceptions present no serious threat to the evolutionary theory, as addictive drugs were not generally available during most of our evolutionary history. In the case of alcohol, it is possible that different ethnic groups have evolved a degree of resistance proportionate to their length of exposure, so that long-civilised groups like the Chinese have high resistance. (This cannot be explained by a general immunity to addiction, since the Chinese are notoriously addicted to gambling.)

The attractiveness of fats and sugars is no mystery, as during most of our evolutionary history they were in short supply and highly beneficial. It is only within the last hundred years or so that they have become harmful.

2. Harmful circumstances associated with neutral sensations (or absence of sensation)

Again, there are many obvious examples. X-rays and other ionising radiation are harmful but not immediately painful. Many poisonous chemicals and gases are tasteless or odourless. These present no evolutionary puzzle, as the hazards are too recent to have produced an evolved response.

More interesting is the pattern of response to injury and disease. Serious injuries do not always cause pain. It is said that injuries to the brain itself are not painful, and I have seen it stated that most of the internal organs are not sensitive to cutting or burning, though they are sensitive to pressure. This makes good evolutionary sense: there would be no selective advantage in evolving a pain-response to such injuries, since in natural conditions they would almost always be fatal. By the time your guts are roasting over an open fire, there is no point in learning to avoid the experience in future, because you have no future.

It is also said that massive traumas, like being hit by a truck or a bullet, are usually painless. This may be because such traumas were rare and/or invariably fatal in our evolutionary past. Another possibility is that major traumas would have occurred mainly in attacks by wild animals or fights with rivals, and that in such circumstances the ‘fight or flight’ response, charged by adrenalin, takes over.

Another interesting example is cancer, which is usually painless until it is well-advanced, and the cancer begins to press on nerves or vital organs. From an evolutionary point of view this makes sense, as our ancestors could do nothing either to avoid or to treat cancer. In modern circumstances the absence of pain is disastrous, as cancer can often be cured in its early stages but not by the time it has become painful. (Patrick Wall, op. cit. p. 104, makes this point but characteristically overlooks its evolutionary significance.)

In general, the main evolutionary function of pain following serious injury or disease (as distinct from minor cuts, stings, etc) is probably to promote healing, rather than to deter the victim from repeating the event. E.g. the pain following a bone fracture promotes healing by preventing the victim from putting strain on the injured area. Unfortunately, a lot of pain experienced by individuals is probably ‘useless’, in the sense that it does not enhance their personal reproductive fitness. For a pain response to evolve, it is not necessary that it should always promote fitness, only that it should do so often enough to outweigh the cost, in extra nerve cells or whatever, of producing it.

3. Beneficial circumstances associated with unpleasant sensations

I can’t think of many examples, but there is one biggie: childbirth. Childbirth is obviously vital to reproductive fitness, but it is also very painful. One might expect this to deter women from having children. So why hasn’t natural selection eliminated the pain? Of course, it is not surprising that having a large object extruded from a narrow orifice is painful, but we might expect evolution to have provided a flood of natural painkillers (endorphins) to remove the pain.

I don’t claim to have a conclusive explanation. One argument might be that painful childbirth is a modern phenomenon, a product of our soft Western civilization. There is a widespread modern belief that in ‘primitive’ peasant or hunter-gatherer societies women feel little pain from childbirth; they just squat behind a bush, pop out the baby, and get on with their work. This belief has encouraged the ‘natural childbirth’ movement, which teaches that a fit, healthy and unstressed woman should be able to give birth without pain. Women who demand drugs or anaesthetics are made to feel guilty and inadequate. Unfortunately it is all bollocks. ‘Primitive’ women do feel pain, it is just that in some cultures they are expected to keep quiet about it (see Wall, op. cit., p. 84). Not for the first time, a large area of modern medical practice is based on a delusion. Demand the drugs, girls.

Another possibility is that the mother’s pain is beneficial to the baby, by encouraging her to push harder and get the damned thing out more quickly.
But there is no evidence for this: childbirth proceeds just as well under complete anaesthesia, as the Victorians discovered, led by Victoria herself.

I can only suggest that for much of human evolutionary history the pain of childbirth did not act as a deterrent to having children because women had no choice in the matter. They didn’t have any choice about having sex, and if they had sex, they had children. Moreover, the connection between having sex and having children may not have been understood until a few thousand years ago, perhaps when man started keeping and breeding livestock. According to Malinowski, even in modern times the Trobriand Islanders did not see the connection, except in the limited sense of recognizing that a virgin could not give birth. Hence it was necessary for girls to be deflowered at a very early age, to ensure they could have children. (I’m not sure I believe Malinowski’s account! See also the discussion in Westermarck’s History of Human Marriage.)

4. Beneficial circumstances associated with neutral sensations (or absence of sensation)

I can’t think of many examples, but it is interesting that air and water are necessary for life, but tasteless and odourless. In the case of air, this is probably because we are normally surrounded by air, and do not need to seek it out. But why doesn’t water taste pleasant? Well, drinking water when you are thirsty is very pleasant, even if the water is tasteless. And since saliva contains water I guess it might be inconvenient if water had a taste, which would lead to overproduction of saliva.

5. Neutral circumstances associated with pleasant sensations

I can’t think of many examples, but it is puzzling that flowers smell so good to us. Indeed, the scent of flowers is probably the nicest smell of all. This would be understandable if humans fed on nectar or pollen, like bees, but we do not, except in the form of honey. I can only suggest that some of our primate ancestors did, and we have inherited their pleasure in the scent of flowers. (Some primate species do eat flowers or suck nectar.) But it must be many thousands of years since flowers were a major food item for humans, so this is a thin explanation.

6. Neutral circumstances associated with unpleasant sensations

I really can’t think of any obvious examples. This is a strong point against any suggestion that the distribution of pleasures and pains is random with respect to biological utility.

Posted by David B at 05:17 AM