Chapter 7 Chapter 5 Aggressive Behavior: Stability and Selfish Machines

This chapter deals primarily with the largely misunderstood topic of aggressive behaviour.We shall continue to speak of the individual as a selfish machine programmed to do whatever is best for all of its genes as a whole.This way of speaking is for simplicity of description.At the end of this chapter we will return to the single gene. For a survival machine, another survival machine (not its own offspring, nor another close relative) is as much a part of its environment as a rock, a river, or a loaf of bread. This other survival machine can cause trouble, but it can also be exploited.One important difference between it and a rock or a river is that it tends to fight back.Because it is also a machine, with immortal genes on which its future rests, and in order to preserve these genes, it will go through fire and water.Natural selection favors genes that control their survival machines to make the most of their environment, including other survival machines of the same species and of different species.

At times, survival machines seem to have little effect on each other's lives.Moles and blackbirds, for example, do not eat each other, mate, or compete for territory.Even so, we can't think that they have nothing to do with each other.They may be competing for something, perhaps for earthworms.That doesn't mean you'll see moles and blackbirds vying for an earthworm; in fact, a blackbird may never see a mole in its life.But if you wiped out the mole population, there might be a noticeable effect on the blackbird, though I dare not speculate about the details of the effect, or by what indirect route of tortuous encounters.

Survival machines of different species interact with each other in a variety of ways.They may be predators or prey, parasites or hosts, or rivals for some rare resource.They can be exploited in various special ways, for example, flowers use bees as pollinators. Survival machines belonging to the same species tend to affect each other's lives more directly.This happens for many reasons.One reason is that half the members of one's own species are likely to be potential mates, and industrious and useful parents to their offspring; Both are machines that preserve genes in the same kind of place, and have the same way of life, so they are more direct competitors for all the resources necessary for life.The mole may be a rival to the blackbird, but it is not nearly as important as the other blackbird.Moles and blackbirds may compete for earthworms, but blackbirds and blackbirds compete with each other not only for earthworms but for everything else.They may also compete for mates if they are of the same sex.Usually male animals compete with each other for female mates, for reasons we shall see.This suggests that a male may be doing his own genes a favor if he does harm to another male he is competing with.

Thus, the logical strategy for a survival machine seems to be to kill its competitors and then, preferably, eat them.Although slaughter and cannibalism do occur in nature, to think that it is common is a naive understanding of the theory of the selfish gene.In fact, Lorenz emphasized the restrained and gentlemanly nature of animal fights in his book On Aggression.In his opinion, there is one thing worth noting about animal fights: their fight is a normal competitive activity, like boxing or fencing, it is carried out according to rules.Animal combat is a fight fought with a blunt sword or with gloves on.Threats and bravado have replaced real threats and bravado.The victor respects the gesture of submission, which does not mortally blow or bite the surrender, as our naive theories may assert.

Interpreting animal aggression as restrained and regulated behavior can be controversial.Especially to say that poor historic humans are the only species to slaughter their own kind, the sole heirs of the mark of Cain and all the lurid accusations of the like.Whether a naturalist emphasizes the violent or restrained side of animal aggression depends partly on the kinds of animals he usually observes, and partly on his biases in evolution. "people.Even if descriptions of how animals fight are somewhat exaggerated, there is at least some truth to the idea that animal civilizations fight.On the surface, this phenomenon appears to be a form of altruism.The theory of the selfish gene has to take on the difficult task of explaining this phenomenon.Why don't animals go out of their way to kill rivals of their own species at every possible opportunity?The general answer to this question is that there are benefits as well as costs to that kind of doom and gloom, and not just the obvious loss of time and energy.For example, suppose both B and C are my competitors, and I happen to meet B.As a selfish individual, I should try to kill B.But take it easy and listen to me. C is both my opponent and B's opponent.If I kill B, I get rid of an opponent for C, and I do a good thing for C invisibly.It may be better if I let B live, because then B may compete or fight with C, and I can reap the benefits.There is no apparent benefit to killing an opponent indiscriminately, and that is the moral of this hypothetical simple example.

In a large and complex competitive system, getting rid of a competitor is not necessarily a good thing, and other competitors are likely to get more benefits from it than you.Those officials responsible for controlling pests were taught such harsh lessons.You have a serious agricultural pest, you discover a good way to get rid of it, and you happily do it.Little do they know that the eradication of this kind of pest benefits another kind of pest, and its degree even exceeds the benefits to human agriculture.As a result, your situation is worse than before. On the other hand, it seems like a good idea to discriminately kill, or at least wrestle, certain specific competitors.If B is an elephant seal (elephant seal) and has a large group of "harem" (harem), and I am also an elephant seal, and I can get his "harem" by killing him, then I will do this Might be wise.But even in selective wrestling there are losses and risks.It is in B's interest to fight back to defend his valuable property.Had I provoked a fight, I would have ended as it had, probably in death.Maybe it is even more likely that I will die and it will not die.I want to wrestle it because it holds a valuable resource.But why does it have such resources?It may have been won in battle.It may have beaten off other challengers before it fought me.It could be a valiant fighter.Even if I win the fight and get the "wives and concubines", I may be seriously injured in the fight and not be able to enjoy the benefits.Also, wrestling drains time and energy.It may be better to temporarily save time and energy.If I focus on eating and not causing trouble for a while, I grow bigger and stronger.

Eventually I'll fight it for the wives, but I may have a better chance of winning if I wait instead of rushing in now. The above self-monologue is purely illustrative: Before deciding whether to fight or not, it is best to make a complex, if unconscious, but complex trade-off of gains and losses.While there are undoubtedly certain advantages to be gained from wrestling, it is not all advantages and disadvantages.Likewise, during the course of a fight, every tactical decision involving escalation or de-escalation has its pros and cons, and these pros and cons can in principle be analysed.Individual ecologists have known this for a long time, albeit not quite clearly, but only Smith, who is not usually considered an ecologist, has made it so forcefully and clearly .He collaborated with Price and Parker on a branch of mathematics called Game Theory.Their original insights can be expressed in words rather than mathematical symbols, though with some loss of precision.

The evolutionarily stable strategy (hereinafter referred to as ESS) is the basic concept proposed by Smith.He traced back to the source and found that the first to have this idea was Hamilton (WDHamilton) and MacArthur (RHMacArthur). A "policy" is a programmed behavior policy.For example, "Attack your opponent; if he flees, chase him; if he fights back, flee" is one strategy.It is important to understand that the strategies we are talking about are not consciously formulated by the individual.Don't forget that we're picturing animals as robotic-like survival machines whose muscles are controlled by a preprogrammed computer.Writing a strategy into a simple set of instructions in words is just for our convenience.Animal behavior, produced by some mechanism that is difficult to specify, seems to be based on such instructions.

Most members of the population adopt a certain strategy, and the benefits of this strategy are not comparable to other strategies, this strategy is evolutionarily stable strategy or ESS.The concept is both subtle and important.In other words, the best strategy for an individual depends on what most members of the population are doing.Since the rest of the population is also made up of individuals, each of which seeks to maximize its individual achievement, what will persist will be a strategy that, once established, is impossible for any individual behaving erratically. Compare with it.After a major change in the environment, there may be a short period of evolutionary instability in the population, and there may even be fluctuations.But once an ESS is established, it becomes stable: deviations from the ESS will be punished by natural selection.

To apply this idea to explaining aggression, let's consider one of the simplest examples that Smith postulates.Suppose there is a particular species called eagle and dove. There are only two fighting strategies in this population.In our hypothetical population, all individuals are either hawks or doves.The hawk always fights with all his might, desperately, and never backs down unless he is seriously wounded; but the dove only threatens and intimidates in the customary and graceful manner, never harming other animals.If the hawk fights the dove, the dove runs away quickly, so the dove is not hurt.If hawks fight hawks, they will fight until one of them is seriously injured or dies.If pigeon meets dove, no one is hurt; they engage in a confrontation for a long time, until one of them gets tired, or gets bored and decides not to continue the confrontation, and thus makes a concession. .Let us assume for the moment that an individual cannot know in advance whether its opponent is a hawk or a dove.It only knows when it's wrestling with it, and it can't remember which individuals it has wrestled with in the past, so it can't learn from it.

Now, as a purely arbitrary rule of the game, we stipulate the "score" standard for competitors as follows: 50 points for a win, 0 points for a loss, -100 points for serious injuries, and -10 points for those who prolong the game and waste time point.We can think of these scores as a currency that translates directly into genetic survival.Individuals with high scores and high average "profits" will leave many genes in the gene pool.On a broad scale, the actual numbers don't mean much for analysis, but they help us think about it. It is not of interest to us whether the hawk has a tendency to beat the dove when the hawk fights the dove, which is important.We already know the answer to that question: Eagles always win.What we want to know is whether the hawk or the dove is the evolutionarily stable strategic type.If one of them is an ESS type and the other is not, then we think that the ESS type will evolve.Theoretically, it is possible that there are two ESS types.Whatever strategy happens to be adopted by the majority of the population—whether it is a hawk or a dove—if the best strategy for any individual is to follow the herd, then there are two ESSs type is possible.In this case, the population generally remains in the one of its two stable states which it reaches first.However, as we shall see, both of these strategies, hawk or dove, are in fact unlikely to be evolutionarily stable on their own, so we should not expect either of them to evolve.To illustrate this, we have to calculate the average profit. Suppose there is a population consisting entirely of pigeons.No matter when they fight, no one is hurt.Such contests are long, ritualized contests, perhaps eye-to-eye confrontations, which end only when one opponent backs down.The winner then gets 50 points for acquiring the resource in question, but gets a -10 penalty for wasting time by staring at each other for a long time, so a net score of 40 points.The loser was also penalized -10 points for wasting time.On average, each pigeon can be expected to win and lose 50/50.The average profit per game is therefore the average of +40 points and -10 points, which is +15 points.So, every pigeon in the pigeon population seems to be doing well. But now suppose a mutant eagle appears in the population.Since he was the only hawk around, every fight he fought was with the dove.The hawk is always unbeaten against the dove, so he nets +50 points per fight, and this number is his average profit.Since the dove's profit is only +15 points, the eagle enjoys a huge advantage.As a result, hawk genes spread rapidly through the population.But the hawk can no longer count on the doves for all its opponents.To take another extreme example, if the successful spread of the hawk gene causes the entire population to become dominated by hawks, then all fights become hawk-to-hawk fights.This time the situation is completely different.When eagle meets eagle, one of them is seriously injured and gets -100 points, while the winner gets +50 points.Each hawk in a hawk population is expected to win a fight half and half.Its average expected profit per fight is therefore half of +50 points and -100 points, or -25 points.Now let us imagine the situation of a natural pigeon living in a population of hawks.There is no doubt that it loses every fight.But on the other hand it never hurts.Therefore, its average payoff in the hawk population is 0, whereas the average hawk payoff in the hawk population is -25.Pigeon genes therefore have a tendency to spread out in the population. According to the way I describe it, it seems that there is a continuous swing state in the population.The hawk genes skyrocket to rapid dominance; the result of the hawk majority is that the dove genes necessarily benefit, and the numbers increase until the hawk genes start multiplying again, and so on.It doesn't have to be that volatile, however.There is a steady ratio of hawks to doves.You just do the math on whatever prescribed scoring system we use, and it turns out that the stable ratio of doves to hawks is 5/12:7/12.Up to this steady ratio, the average payoff of hawks and doves is exactly the same.Therefore, natural selection does not favor A over B, but treats them equally.If the number of hawks in the population starts to rise and the ratio is no longer 7/12, the doves start gaining an extra advantage and the ratio returns to a steady state again.As we shall see the stable ratio of sexes is 50:50, so in this hypothetical example the stable ratio of hawks to doves is 7:5.In both of the above ratios, if a swing from the stable point occurs, it need not be very large. This situation sounds a bit like group selection at first glance, but it actually has nothing in common with group selection.The reason this sounds like group selection is that it reminds us of a population in a stable equilibrium that tends to gradually recover whenever the equilibrium is disturbed.But ESS is a far more subtle concept than group selection.It has nothing to do with the fact that some groups are more successful than others.This is well illustrated by applying the arbitrary scoring system in our hypothetical example.In a stable population consisting of 7/12 hawks and 5/12 doves, the average individual payoff turns out to be 61/4 points.This is true whether the individual is a hawk or a dove. 6 1/4 points is much less than the average profit (15 points) per pigeon in a population of pigeons.As long as everyone agrees to be a pigeon, each individual benefits.According to pure group selection, any group whose individuals all agree to be doves will achieve much more than a competing group that stays at the ESS ratio. (Actually, a group consisting purely of pigeons is not necessarily the most successful group. A group consisting of 1/6 hawks and 5/6 doves has an average profit of 16 2/3 points per race .The group formed according to this ratio is the group most likely to be successful. But as far as the present topic is concerned, we don’t need to consider this situation. For each individual, it is relatively simple to form a group consisting entirely of pigeons, because each The average profit of an individual is 15 points, which is much superior to ESS.) Therefore, group selection theory believes that evolution to a group consisting entirely of pigeons is a development trend, because eagles account for 7/12 of the group's probability of success be smaller.The problem is that even those groups that offer benefits to each of its members in the long run are bound to have bad apples.It is true that every single pigeon in an all-pigeon colony is better off than a pigeon in an ESS colony.Unfortunately, in the group of pigeons, a single-handed hawk can do unparalleled deeds, and no force can stop the evolution of hawks.Therefore, the group could not escape the doom of disintegration due to internal betrayal. The stability of the ESS population is not because it is particularly beneficial to the individuals in it, but only because it has no hidden danger of internal betrayal. Human beings are able to form various alliances or groups, even if these alliances or groups are not stable in the sense of ESS, they are beneficial to each individual.This situation is possible only because each individual can consciously use his foresight ability, so that he understands that it is in his own long-term interest to abide by the provisions of the covenant.The temptation for some individuals to violate the covenant for the possibility of large short-term benefits can become overwhelming.This danger is always present even in the covenants made by human beings.Monopoly prices are perhaps the most telling example.It is in the long-term interest of all gas station owners to set the flat price of gasoline at some artificially high level.Those price-manipulating groups can exist for a considerable period of time because they make conscious estimates and judgments about the highest long-term interests.But every now and then an individual is tempted to lower the price by the temptation to make a quick profit.The peers near this kind of person will immediately follow suit, and the wave of lower prices will spread to the whole country.To our regret, the conscious foresight of the gas-station owners now resumes its role and concludes a new pact to monopolize prices.So, even in a species endowed with the gift of conscious foresight, human beings, pacts or groups based on the highest long-term interests are often on the verge of falling apart due to internal rebellion.Even less often, group interests or group strategies can develop in wild animals, where they are controlled by competing genes.What we can see must be: Evolutionarily stable strategies are everywhere. In the above example, we simply assumed that each individual was either a hawk or a dove.We end up with an evolutionarily stable ratio of hawks to doves.In fact, that is to say that the genes for the hawk and the genes for the dove achieve a stable ratio in the gene pool.This phenomenon is called stable polymorphism (polymorphism) in the term of genetics.Mathematically speaking, a fully equal ESS without polymorphism can be achieved in the following way.If every individual can behave either like a hawk or a dove in each particular competition, an ESS can be achieved in which all individuals have exactly equal probability of behaving like a hawk.In our concrete example this probability is 7/12.In fact, this situation shows that each individual, participating in each competition, has decided arbitrarily in advance whether to act like a hawk or a dove in this competition; although the decision is made at will, it is always Consider the ratio of hawks to 7 doves to 5.While these decisions are in favor of the hawks, they must be arbitrary, in the sense that it is crucial that an opponent cannot guess in advance what the other will do in any particular contest.For example, it is absolutely not advisable to play the role of the hawk for seven fights in a row, then the dove for five fights in a row, and so on.If any individual employs such a simple fight sequence, its opponents will quickly see through the tactic and take advantage of it.Against this strategist who uses simple fight sequences, you are only in a good position to fight with hawk moves knowing that he is acting as a dove in the fight. Of course, the story of the hawk and the dove is a bit childishly simple.This is a "pattern" that, while it doesn't actually happen in nature, can help us understand what actually happens in nature.Patterns can be very simple, as in our hypothetical pattern, and still be useful for understanding an argument or developing a concept.Simple patterns can be enriched and extended to gradually form more complex patterns.If all goes well, as the patterns become more complex, they will become more like the real world.One way to develop the hawk-dove model is to introduce more strategies.Hawks and doves are not the only possibilities.A more complex strategy introduced by Smith and Price is called the Retaliator. The retaliator behaves like a dove at the beginning of each fight, that is to say, it is not like a hawk, which is desperate and fierce at the beginning of the attack, but puts on the usual threatening confrontation posture, but once the opponent attacks it, It fights back.In other words, a retaliator behaves like a hawk when attacked by a hawk, behaves like a dove when confronted by a dove, and behaves like a dove when confronted by another retaliator .A riposte strategist is a conditional strategist.Its behavior depends on the behavior of the other party. Another conditional strategist is called a bully.It behaves like a hawk in every way, but when it is retaliated, it flees immediately.Another kind of conditional strategist is the prober-retaliator.It basically acts like a riposte tactician, but also tentatively sometimes escalates the contest briefly.If the opponent does not fight back, it persists in acting like a hawk; on the other hand, if the opponent fights back, it reverts to the usual threatening posture of the dove.If attacked, it retaliates just like a normal riposte tactician. If all five of the strategies I mentioned were pitted against each other in a computer simulation, only one of them, the riposte strategy, was evolutionarily stable.Tentative retaliatory tactics are nearly stable.The dove strategy is unstable because hawks and bullies invade the dove population.The hawk strategy is also unstable since the hawk population is subject to attack by doves and bullies.The bully strategy is also unstable due to the fact that the bully population is subject to encroachment by hawks.In a population of retaliators, since no other strategy does better than retaliator itself, it will not be violated by any other strategy. However, the pigeon strategy can also perform equally well in a population composed of pure retaliators.This means that, other things being equal, the number of pigeons will slowly and gradually rise.If the number of doves rises to a considerable extent, the tentative retaliator strategy (and along with the hawk and the bully) starts to gain the upper hand, because they perform better against the dove than the retaliator strategy.The tentative-retaliator strategy itself, unlike the hawk and the bully strategies, has only one other strategy in the population of tentative-retaliator strategies, the riposte strategy, that performs better, and only slightly better. .In that sense, it's almost an ESS.We can therefore imagine that a mixed strategy of riposte and tentative riposte might tend to dominate, perhaps even with a modest swing between the two strategies, while the very small proportion of pigeons also differs in number. increase or decrease.We no longer have to think about problems based on polymorphism, because according to polymorphism, each individual will never adopt this strategy, or adopt another strategy.Each individual can in fact employ a complex mix of riposte, heuristic riposte, and pigeon strategy. The conclusions of this theory are not far from the actual situation of most wild animals.In a sense, we have addressed the "civilized" side of animal aggression.As for the details, of course it depends on the actual "score" of wins, injuries and wasted time etc.For the elephant seal, the winning prize may be the near exclusive right to a large group of wives.Therefore, the profit of this kind of winning should be said to be very high.It's no wonder that fighting is so vicious and the possibility of serious injury is so high.Compared with the cost of being injured in a fight and the benefits of winning, the cost of wasting time should be said to be small.But on the other hand, for a small bird in a cold climate, the cost of wasting time can be enormous.Great tits feeding chicks need to catch an average of one prey every thirty seconds.Every second of the day is precious.In hawk-to-hawk combat, the time wasted is relatively brief, but perhaps it should be considered a more serious matter than the risk of injury to them.Unfortunately, we know too little about the losses and benefits of various activities in nature to come up with actual figures.We cannot easily draw conclusions purely from our own arbitrarily chosen numbers. These general conclusions are important that ESSs tend to evolve; that ESSs are not the same as the best that can be achieved by any mass group; and that common sense can lead people astray. Another type of war game that Smith contemplates is called a "war of attrition."It may be argued that this "war of attrition" occurs in a species that never engages in dangerous combat, perhaps a well-armored species that is less likely to be injured.All disputes among this species are settled by posturing in the traditional manner.Contests always end with concessions by the party participating in the contest.If you want to win, all you have to do is keep your eyes on the opponent and hold on until the opponent finally runs away.It is obvious that no animal can be threatened indefinitely; for there are important things to do elsewhere.The resources it competes for are valuable, but not infinite in value.It was worth only so much time, and as at auction, everyone was prepared to pay only so much.Time is the currency used in this auction with only two bidders. We assume that all of these individuals estimate precisely in advance how long a particular resource (such as female animals) is worth spending.Then a mutant individual who intends to give it a little more time is always the winner. Therefore, a strategy with a fixed bid limit is unstable.Even if the value of the resource can be estimated with great precision and all individuals bid appropriately, this strategy is unstable.If any two individuals bid according to the limit strategy, they will stop bidding at the same instant, and no one will get this resource as a result!In this case, it is better to simply abstain from the beginning than to waste time in the competition.The important difference between a war of attrition and an actual auction is that in a war of attrition, after all, both parties in the contest pay the price, but only one gets the item.Therefore, in a population of extreme bidders, a strategy of abstaining from the start of the contest will succeed and spread through the population.The inevitable consequence is that for those individuals who do not abstain right away, but wait a few seconds before abstaining, some of the benefits they might have accrued begin to accrue.This is an advantageous strategy used against individuals who have already dominated the population and retreated without a fight.Thus natural selection favors holding out for a period of time before abstaining, gradually extending the period until again as far as the real economic value of the resource in question will allow. In talking, we unknowingly describe the phenomenon of swaying in populations.Yet again, mathematical analysis shows that this wobble is not inevitable.There is an evolutionarily stable strategy, which can be expressed not only in mathematical formulas, but also in words: each individual confronts for a period of time that cannot be estimated in advance, that is, it is difficult to estimate in advance in any specific situation , but according to the actual value of resources, a binary number can be obtained.For example, suppose the actual value of the resource is five minutes of stamina.In an evolutionarily stable strategy, any individual may last more than five minutes, or less than five minutes, or exactly five minutes.The important thing is that the other party has no way of knowing how long it is going to last on this particular occasion. In a war of attrition it is obviously of the utmost importance that the individual should not give any hint as to how long it intends to last.The mere twitch of a whisker at the mere thought of throwing in the towel puts any individual at a disadvantage at once.If a twitch of the whiskers is a sure sign of retreat in less than a minute, a very simple strategy for winning is: "If your opponent's whiskers twitch, no matter how long you're prepared to last, you'll fight again." Wait an extra minute. If your opponent's whiskers have not shaken, and it is less than a minute before you are ready to throw in the towel, abstain without wasting any more time. Never shake your own beard." Thus, whisker shaking, or any similar exposure that heralds future behavior, is quickly punished by natural selection.A calm facial expression will be developed. Why keep a straight face instead of openly lying?The reason is again that the act of lying is unstable.Assuming that it is the case that in a war of attrition most individuals only bristle the nape when they really want to fight for a long time, then what would be able to develop is the obvious opposite strategy: when the opponent puffs up the nape Immediately throw in the towel.But this is when the ranks of liars may begin to build up.Individuals that really had no intention of fighting for a long time bristled the scruff of their necks in every confrontation, and the fruits of victory were easily reaped. The liar gene thus spreads.When liars become the majority, natural selection favors those individuals who can force the liars to a showdown.Thus the number of liars will decrease again.Neither lying nor telling the truth is an evolutionarily stable strategy in a war of attrition.The nonchalant facial expression is an evolutionarily stable strategy.Even if the final surrender, it will be sudden and unpredictable. So far we have only discussed what Smith calls the "symmetric" race.This means that we assume that the contestants are equal in all respects except fighting strategy.我们把鹰和鸽子假定为力量强弱相同，具有的武器和防护器官相同，而且可能赢得的胜利果实也相同。对于假设一种模式来说，这是简便的，但并不太真实。帕克和史密斯也曾对"不对称"的竞赛进行了探讨。举例说，如果个体在身材大小和搏斗能力方面各不相同，而每一个体也能够对自己的和对手的身材大小进行比较并作出估计的话，这对形成的ESS是否有影响？肯定是有影响的。 不对称现象似乎主要有三类。第一类就是我们刚才提到的那种情况：个体在身材大小或搏斗装备方面可能不同；第二类是，个体可能因胜利果实的多寡而有所区别。比如说，衰老的雄性动物，由于其余生不会很长，如果受伤，它的损失较之来日方长的、精力充沛的年轻雄性动物可能要少。 第三类，纯属随意假定而且明显互不相干的不对称现象能够产生一种ESS，因为这种不对称现象能够使竞赛很快见分晓，这是这种理论的一种异乎寻常的推论。 比如说，通常会发生这样的情况，即两个竞争者中的一个比另一个早到达竞赛地点。我们就分别称它们为"留驻者"（resident）和"闯入者"（intruder）。 为了便于论证起见，我是这样进行假定的，不论是留驻者还是闯入者都不因此而具有任何附加的有利条件。我们将会看到，这一假定在实际生活中可能与事实不符，但这点并不是问题的关键。问题的关键在于，纵令留驻者具有优于闯入者的有利条件这种假定无理可据，基于不对称现象本身的ESS也很可能会得以形成。 简单地讲，这和人类抛掷钱币，并根据钱币的正反面来迅速而毫不用争执地解决争论的情况有类似之处。 "如果你是留驻者，进攻；如果你是闯入者，退却，"这种有条件的策略能够成为ESS。由于不对称现象是任意假定的，因此，"如果是留驻者，退却；如果是闯入者，进攻"这种相反的策略也有可能是稳定的。具体种群中到底采取这两种ESS中的哪一种，这要取决于其中的哪一种ESS首先达到多数。个体的大多数一旦运用这两种有条件的策略的某一种，所有脱离群众的行为皆受到惩罚，这种策略就因之称为ESS。 譬如说，假定所有个体都实行"留驻者赢，闯入者逃"的策略。就是说它们所进行的搏斗将会是输赢各半。它们绝不会受伤，也绝不会浪费时间，因为一切争端都按任意作出的惯例迅速得到解决。现在让我们设想出现一个新的突变型叛逆者。假定它实行的是纯粹的鹰的策略，永远进攻，从不退却，那么它的对手是闯入者时，它就会赢；而当它的对手是留驻者时，它就要冒受伤的很大风险。平均来说，它比那些按ESS的任意规定的准则进行比赛的个体，得分要低些。如果叛逆者不顾惯常的策略而试图反其道而行之，采取"如身为留驻者就逃；如身为闯入者就进攻"的策略，那么它的下场会更糟。它不仅时常受伤，而且也极少有机会赢得一场竞赛。然而，假定由于某些偶然的变化，采用同惯例相反的策略的个体竟然成了多数，这样它们的这种策略就会成为一种准则，偏离它就要受到惩罚。可以想见，我们如果连续观察一个种群好几代，我们就能看到一系列偶然发生的从一种稳定状态跳到另一种稳定状态的现象。 但是实际生活中可能并不存在真正的任意不对称现象。如留驻者实际上可能比闯入者享有更有利的条件，因为它们对当地的地形更熟悉。闯入者也许更可能是气喘吁吁的，因为它必须赶到战斗现场，而留驻者却是一直待在那里的。两种稳定状态中，"留驻者赢，闯入者退"这种状态存在于自然界的可能性更大，其所以如此的理由是比较深奥的。这是因为"闯入者赢，留驻者退"这种相反的策略有一种固有的自我毁灭倾向，史密斯把这种策略称为自相矛盾的策略。处于这种自相矛盾中的ESS状态的任何种群中，所有个体总是极力设法避免处于留驻者的地位：无论何时与对手相遇，它们总是千方百计地充当闯入者。为了做到这一点，它们只有不停地四处流窜，居无定所。这是毫无意义的。这种进化趋势，除无疑会招致时间和精力上的损失之外，其本身往往导致"留驻者"这一类型的消亡。在处于另一种稳定状态，即"留驻者赢，闯入者退"的种群中，自然选择有利于努力成为留驻者的个体。对每一个体来说，就是要坚守一块具体地盘，尽可能少离开，而且摆出"保卫"它的架势。这种行为如大家所知，在自然界中到处可见，大家把这种行为称为"领土保卫"。 就我所知，伟大的个体生态学家廷伯根（Niko Tinbergen）所作的异常巧妙和一目了然的试验，再精彩不过地展示了这种行为上的不对称性。他有一只鱼缸，其中放了两条雄性刺鱼。它们在鱼缸的两端各自做了巢，并各自"保卫"其巢穴附近的水域。廷伯根将这两条刺鱼分别放人两个大的玻璃试管中，再把两个试管并排放一起，只见它们隔着玻璃管试图相互搏斗。于是产生了十分有趣的结果。 当他将两个试营移到刺鱼A的巢穴附近时，A就摆出进攻的架势，而刺鱼B 就试图退却；但当他将两个试管移到刺鱼B 的水域时，因主客易地而形势倒转。廷伯根只要将两个试管从鱼缸的一端移向另一端，他就能指挥哪条刺鱼进攻，哪条退却。很显然，两条刺鱼实行的都是简单的有条件策略："凡是留驻者，进攻；凡是闯入者，退却。"这种领土行为有什么生物学上的"好处"？这是生物学家时常要问的问题，生物学家提出了许多论点，其中有些论点稍后我们将会提及。但是我们现在就可以看出，提出这样的问题可能本来就是不必要的。这种领土"保卫"行为可能仅仅是由于抵达时间的不对称性而形成的一种ESS，而抵达时间的不对称性通常就是两个个体同一块地盘之间关系的一种特点。 体积的大小和一般的搏斗能力，据认为是非任意性不对称现象中最重要的形式。 体积大不一定就是赢得搏斗不可或缺的最重要特性，但可能是特性之一。如果两个个体搏斗时比较大的一个总是赢的话，如果每一个体都能确切知道自己比对手大还是小，只有一种策略是明智的："如果你的对手比你体积大，赶快逃跑。同比你体积小的人进行搏斗。"假使体积的重要性并不那么肯定，情况也就随之更复杂些。如果体积大还是具有一点优越性的话，我刚才讲的策略就仍旧是稳定的。如果受伤的风险很大的话，还可能有一种"似非而是的策略"，即"专挑比你大的人进行搏斗，见到比你小的就逃"！称之为似非而是的原因是不言而喻的。因为这种策略似乎完全违背常识。它之所以能够稳定，其原因在于：在全部由似非而是的策略者组成的种群中，绝不会有人受伤，因为每场竞赛中，逃走的总是参加竞赛的较大的一个。一个大小适中的突变体如实行的是"合理"的策略，即专挑比自己体积小的对手，他就要同他所遇见的人中的一半进行逐步加剧的严重搏斗。因为，如果他遇到比自己小的个体，他就进攻；而较小的个体拼命还击，因为后者实行的是似非而是策略；尽管合理策略的实行者比似非而是策略的实行者赢得胜利的可能性更大一些，但他仍旧冒着失败和严重受伤的实际风险。由于种群中的大部分个体实行似非而是的策略，因而一个合理策略的实行者比任何一个似非而是策略的实行者受伤的可能性都大。 即使似非而是的策略可能是稳定的，但它大概只具有学术上的意义。似非而是策略的搏斗者只有在数量上大大超过合理策略的搏斗者的情况下才能获得较高的平均盈利。首先，这样的状况如何能够出现实属难以想象。即使出现这种情况，合理策略者对似非而是策略者的比率也只要略微向合理策略者一边移动一点，便达到另一种ESS--合理的策略--的"引力区域"（zone of attraction）。所谓引力区域即种群的一组比率，在这个例子里，合理策略者处于这组比率的范围内时是有利的：种群一旦到达这一区域，就不可避免地被引向合理的稳定点。 要是在自然界能够找到一个似非而是的ESS实例会是一件令人兴奋的事情，但我怀疑我们能否抱这样的侈望[我话说得太早了。在我写完了上面这句活之后，史密斯教授提醒我注意伯吉斯（Burgess）关于墨西哥群居蜘蛛Oecobius civitas（拟壁钱属）的行为所作的下述描绘"如果一只蜘蛛被惊动并被赶出其隐避的地方，它就急冲冲地爬过岩石，如岩石上面无隙缝可藏身，就可能到同一物种的其他蜘蛛的隐蔽地点去避难。如果闯入者进来时，这个蜘蛛正在家里，它并不进攻，而是急冲冲爬出去再为自己去另寻新的避难所。因此，一旦第一个蜘蛛被惊动，从一个蜘蛛网到另一个蜘蛛网的一系列替换过程要持续几秒钟，这种情况往往会使聚居区的大部分蜘蛛从它们本来的隐蔽所迁徙，到另一只蜘蛛的隐蔽所"（群居蜘蛛，《科学美国人》，1976年3月号）。这就是第109页上所讲的那种意义上的似非而是的现象］。 假如个体对以往搏斗的结果保留某些记忆，情况又会是怎样呢？这要看这种记忆是具体的还是一般的。蟋蟀对以往搏斗的情况具有一般的记忆。一只蟋蟀如果在最近多次搏斗中获胜，它就会变得更具有鹰的特点；而一只最近连遭败北的蟋蟀，其特点会更接近鸽子。亚历山大（RD Alexander）很巧妙地证实了这种情况，他利用一个模型蟋蟀痛击真正的蟋蟀。吃过这种苦头的蟋蟀再同其他真正的蟋蟀搏斗时多数要失败。我们可以说，每个蟋蟀在同其种群中有平均搏斗能力的成员作比较的同时，对自己的搏斗能力不断作出新的估计。如果把对以往的搏斗情况具有一般记忆的动物，如蟋蟀，集中在一起组成一个与外界不相往来的群体，过一段时间之后，很可能会形成某种类型的统治集团。观察者能够把这些个体按级别高低的顺序排列。在这一顺序中级别低的个体通常要屈从于级别高的个体。这倒没有必要认为这些个体相互能够辨认。习惯于赢的个体就越是会赢，习惯于失败的个体就越是要失败。实际情况就是如此。即使开始时个体的胜利或失败完全是偶然的，它们会自动归类形成等级。这种情况附带产生了一个效果：群体中激烈的搏斗逐渐减少。 我不得不用"某种类型的统治集团"这样一个名称，因为许多人只把"统治集团"（dominance hierarchy）这个术语用于个体具有相互辨认能力的情况。在这类例子中，对于以往搏斗的记忆是具体的而不是一般的。作为个体来说，蟋蟀相互辨认不出，但母鸡和猴子都能相互辨认。如果你是一个猴子的话，一个过去曾经打败过你的猴子，今后还可能要打败你。对个体来说，最好的策略是，对待先前曾打败过它的个体应采取相对的带有鸽派味道的态度。如果我们把一群过去相互从未见过的母鸡放在一起，通常会引起许多搏斗。一段时间之后，搏斗越来越少，但其原因同蟋蟀的情况不同。对母鸡来说，搏斗减少是因为在个体的相互关系中，每一个体都能"安分守己"。这对整个群体来说也带来好处，下面的情况足资证明：有人注意到，在已确立的母鸡群体中，很少发生凶猛搏斗的情况，蛋的产量就比较高；相比之下，在其成员不断更换因而搏斗更加频繁的母鸡群体中，蛋产量就比较低。生物学家常常把这种"统治集团"在生物学上的优越性或"功能"说成是在于减少群体中明显的进犯行为。然而这种讲法是错误的。 不能说统治集团本身在进化的意义上具有"功能"，因为它是群体而不是个体的一种特性。通过统治集团的形式表现出来的个体行为模式，从群体水平的观点上来看，可以说是具有功能的。然而，如果我们根本不提"功能"这个词，而是按照存在有个体辨认能力和记忆的不对称竞赛中的各种ESS来考虑这个问题，这样甚至会更好些。 迄今我们所考虑的竞争都是指同一物种的成员间的竞争。物种间的竞争情况又是如何呢？我们上面已经谈过，不同物种的成员之间的竞争，不象同一物种的成员之间那样直接。基于这一理由，我们应该设想它们有关资源的争端是比较少的，我们的预料已得到证实。例如，知更鸟保卫地盘不准其他知更鸟侵犯，但对大山雀却并不戒备。我们可以画一幅不同个体知更鸟在树林中分别占有领地的地图，然后在上面叠上一幅个体大山雀领地地图，可以看到两个物种的领地部分重叠，完全不相互排斥，它们简直象生活在不同的星球上。 但不同物种的个体之间也要发生尖锐的利害冲突，不过其表现形式不同而已。例如，狮子想吃羚羊的躯休，而羚羊对于自己的躯体却另有截然不同的打算。虽然这种情况不是通常所认为的那种争夺资源的竞争，但从逻辑上说，不算竞争资源，道理上难以讲通。在这里，有争议的资源是肉。狮子的基因"想要"肉供其生存机器食用，而羚羊的基因是想把肉作为其生存机器进行工作的肌肉和器官。 肉的这两种用途是互不相容的，因此就发生了利害冲突。 同一物种的成员也是肉做的，但为什么同类相食的情况相对来说这样少呢？这种情况我们在黑头鸥中见到过，成年鸥有时要吃自己物种的幼鸥。但我们从未见到成年的肉食动物为吞食自己物种的其他成年动物而主动去追逐它们。为什么没有这种现象呢？我们仍旧习惯于按照"物种利益"的进化观点去思考问题，以致我们时常忘记摆出这样完全有道理的问题："为什么狮子不去追捕其他狮子？"还有一个人们很少提出的其实是很好的问题："羚羊为什么见到狮子就逃，而不进行回击呢？"狮子之所以不追捕狮子是因为那样做对它们来说不是一种ESS。同类相食的策略是不稳定的，其原因和前面所举例子中的鹰策略相同。遭到反击的危险性太大了。而在不同物种的成员之间的竞争中，这种反击的可能性要小些，这也就是那么多的被捕食的动物要逃走而不反击的道理。这种现象可能源出于这样的事实：在不同物种的两只动物的相互作用中存在一种固有的不对称现象，而且其不对称的程度要比同一物种的成员之间大。竞争中的不对称现象凡是强烈的，ESS一般是以不对称现象为依据的有条件的策略。"如果你比对手小，就逃走；如果你比对手大，就进攻"，这种类型的策略很可能在不同物种成员之间的竞争中得到发展，因为可以利用的不对称现象非常之多。狮于和羚羊通过进化上的趋异过程而形成了一种稳定性，而竞争中本来就有的不对称现象也因此变得日益加强。追逐和逃跑分别变成它们各自的高超技巧。一只突变型羚羊如果采取了"对峙并搏斗"的策略来对付狮子，它的命运同那些消失在地平线上的其他羚羊相比，可能要不妙得多。 我总是有一种预感，我们可能最终会承认ESS概念的发明，是自达尔文以来进化理论上最重要的发展之一。凡是有利害冲突的地方，它都适用，这就是说几乎在一切地方都适用。一些研究动物行为的学者沾染了侈谈"社会组织"的习惯。他们动辄把一个物种的社会组织看作是一个具备作为实体的条件的单位，它享有生物学上的"有利条件"。我所举的"统治集团"就是一例。我相信，混迹于生物学家有关社会组织的大量论述中的那些隐蔽的群体选择主义的各种假定，是能够辨认出来的。史密斯的ESS概念使我们第一次能够清楚地看到，一个由许多独立的自私实体所构成的集合体，如何最终变得象一个有组织的整体。我认为，这不仅对物种内的社会组织是正确的，而对于由许多物种所构成的"生态系统"以及"群落"也是正确的。从长远观点来看，我预期ESS概念将会使生态学发生彻底的变革。 我们也可以把这一概念运用于曾在第三章搁置下来的一个问题，即船上的桨手（代表体内的基因）需要很好的集体精神这一类比。基因被选择，不是因为它在孤立状态下的"好"，而是由于在基因库中的其他基因这一背景下工作得好。好的基因应能够和它必须与之长期共同生活于一系列个体内的其余基因和谐共存，相互补充。磨嚼植物的牙齿的基因在草食物种的基因库中是好基因，但在肉食物种的基因库中就是不好的基因。 我们可以设想一个不矛盾的基因组合，它是作为一个单位被选择在一起的。在第三章蝴蝶模拟的例子中，情况似乎就是如此。但现在ESS概念使我们能够看到，自然选择纯粹在独立基因的水平上如何能够得到相同的结果，这就是ESS概念的力量所在。这些基因并不一定是在同一条染色体上连接在一起的。 其实，划船的类比还没达到说明这一概念的程度。它最多只能说明一个近似的概念。我们假定，一个赛艇的全体船员要能真正获得成功，重要的是奖手必须用言语协调其动作。我们再进一步假定，在桨手库中，教练能够选用的桨手，有些只会讲英语，有些只会讲德语。操英语的奖手并不始终比操德语的桨手好些，也不总是比操德语的桨手差些。但由于通话的重要性，混合组成的桨手队得胜的机会要少些，而纯粹讲英语的或纯粹讲德语的所组成的桨手队得胜的机会要多些。 教练没有认识到这点，他只是任意地调配他的桨手，认为得胜的船上的个体都是好的，认为失败的船上的个体都是差的。如果在教练的桨手库中，英国人碰巧占压倒优势，那么，船上只要有一个德国人，很可能就会使这条船输掉，因为无法进行通话；反之，如果在桨手库中凑巧德国人占绝对优势，船上只要有一个英国人，也会使这条船失败。因此，最理想的一队船员应处于两种稳定状态中任何一种，即要么全部是英国人，要么全部是德国人，而绝不是混合阵容。表面上看起来，教练似乎选择清一色的语言小组作为单位，其实不然，他是根据个体桨手赢得竞赛的明显能力来进行选择的。而个体赢得竞赛的趋向则要取决于候选桨手库中现有的其他个体。属于少数的候选桨手会自动受到惩罚，这倒并非因为他们是不好的桨手，而仅仅是由于他们是少数而已。同样，基因因能相互和谐共存而被选择在一起，这并不一定说明我们必须象看待蝴蝶的情况那样，把基因群体也看成是作为单位来进行选择的。在单个基因低水平上的选择能给人以在某种更高水平上选择的印象。 在这一例子中，自然选择有利于简单的行为一致性。更为有趣的是，基因之被选择可能由于它们的相辅相成的行为。以类比法来说明问题，我们可以假定由四个右手划桨手和四个左手划奖手组成的赛艇队是力量匀称的理想队；我们再假定教练不懂得这个道理，他根据"功绩"盲目进行挑选。那么如果在候选桨手库中碰巧右手划桨手占压倒优势的话，任何个别的左手划桨手往往会成为一种有利因素：他有可能使他所在的任何一条船取得胜利，他因此就显得是一个好桨手。反之，在左手划桨手占绝对多数的桨手库中，右手划桨手就是一个有利因素。这种情况就同一只鹰在鸽子种群中取得良好成绩，以及一只鸽子在鹰种群中取得良好成绩的情况相似。所不同的是，在那里我们讲的是关于个体--自私的机器--之间的相互作用；而这里我们用类比法谈论的是关于体内基因之间的相互作用。 教练盲目挑选"好"桨手的最终结果必然是由四个左手划奖手和四个右手划桨手组成的一个理想的桨手队。表面看起来他好象把这些桨手作为一个完整的、力量匀称的单位选在一起的。我觉得说他在较低的水平上，即在单独的候选桨手水平上进行选择更加简便省事。四个左手划桨手和四个右手划桨手加在一起的这种进化上稳定状态（"策略"一词在这里会引起误解）的形成，只不过是以表面功绩为基础在低水平上进行选择的必然结果。 基因库是基因的长期环境。"好的"基因是作为在基因库中存活下来的基因盲目地选择出来的。这不是一种理论，甚至也不是一种观察到的事实，它不过是一个概念无数次的重复。什么东西使基因成为好基因才是人们感兴趣的问题。我曾讲过，建造高效能的生存机器--躯体--的能力是基因之成为好基因的标准，这是一种初步的近似讲法。现在我们必须对这种讲法加以修正。基因库是由一组进化上稳定的基因所形成，这组基因成为一个不受任何新基因侵犯的基因库。大部分因突变、重新组合或来自外部而出现的基因很快就受到自然选择的惩罚：这组进化上稳定的基因重新得到恢复。新基因侵入一组稳定的基因偶尔也会获得成功，即成功地在基因库中散布开来。然后出现一个不稳定的过渡阶段，最终又形成新的一组进化上稳定的基因--发生了某种细微程度的进化。按进犯策略类推，一个种群可能有不止一个可选择的稳定点，还可能偶尔从一个稳定点跳向另一个稳定点。渐进的进化过程与其说是一个稳步向上爬的进程，倒不如说是一系列的从一个稳定台阶走上另一个稳定台阶的不连续的步伐。作为一个整体，种群的行为就好象是一个自动进行调节的单位。而这种幻觉是由在单个基因水平上进行的选择所造成。基因是根据其"成绩"被选择的，但对成遗的判断是以基因在一组进化上稳定的基因（即现存基因库）的背景下的表现为基础的。 史密斯集中地论述了一些完整个体之间的进犯性相互作用，从而把问题阐明。鹰的躯体和鸽子躯体之间的稳定比率易于想象，因为躯体是我们能够看得见的大物体。但寄居于不同躯体中的基因之间的这种相互作用犹如冰山的尖顶。而在一组进化上稳定的基因--基因库--中，基因之间绝大部分的重要相互作用，是在个体的躯体内进行的。这些相互作用很难看见，因为它们是在细胞内，主要是在发育中的胚胎细胞内发生的。完整的浑然一体的躯体之所以存在，正是因为它们是一组进化上稳定的自私基因的产物。 但我必须回到完整动物之间的相互作用的水平上来，因为这是本书的主题。把个体动物视为独立的自私机器便于理解进犯行为。如果有关个体是近亲--兄弟姐妹，堂兄弟姐妹，双亲和子女--这一模式也就失去效用。这是因为近亲体内有很大一部分基因是共同的。因此，每一个自私的基因却同时须忠于不同的个体。 这一问题留待下一章再加阐明。