Duty, Virtue and Consequences

Tit for Tat

Mecircle To what extent is it fair to say that all behaviour is selfish? Can one have a viable ethical system based around individuals pursuing their own self-interest above all other factors, and if so, what would such a system be like? And given that human society is founded upon the assumption of co-operation between its members, how is such trust justified?

These issues relate to a particular doctrine known as egoism, and to explore this subject thoroughly we must look at a famous dilemma, a particular strategy involved in its resolution, and the surprising concept that co-operation is not just a good idea, it’s in everyone’s best interests.



The idea that co-operation might be an inherent behaviour in nature was not taken seriously until very recently. Indeed, despite long standing and largely undisputed philosophical objections, many people hold the naïve view that all human action (and presumably all animal action too) is selfish – a view we can almost instantly demolish.

To begin with, we must accept that a specific connotation of the word ‘selfish’ is to act in one’s own benefit to the detriment of others. To see this clearly, consider the examples of going to the doctor when feeling ill, or brushing one’s teeth. It is simply not consistent with how we use the word ‘selfish’ to describe either of these two actions as selfish, although they are self-interested. Conversely, consider the example of parking in a handicapped space when one has perfect personal mobility – this is selfish behaviour, as it takes something away from others (a handicapped space) for one’s own benefit (such spaces are usually closer to the store). 

It is readily apparent, therefore, that the issue is not whether all behaviour is selfish (which by any conventional rendering of this word it is not) but whether all behaviour is self-interested. This position is of significant philosophical import, and is known as psychological egoism – which leads in turn to the theory of ethical egoism, as espoused by Friedrich Nietzsche and Max Stirner (and relating to views expressed earlier by Thomas Hobbes and David Gautier) which proposes that what a moral agent ought to do is what is indicated by self-interest (a view which is commonly in conflict with more common religious views, such as that suggested by the Golden Rule).

Those who believe in psychological egoism espouse a viewpoint in which all behaviour is reduced to self-interest, even altruistic behaviour with great cost to the person making the relevant sacrifices. Often, the viewpoint is supported by presupposing a certain good feeling in the individual which is the self-interested motivation for acting altruistically (such as pride, the satisfaction of a desire to comply with a moral code, or the expectation of reciprocal benefit). Even so, it is hard to render the sacrifice of a soldier who throws themself onto a grenade to save their squad in terms of that soldier’s self-interest without courting considerable criticism and argument. If such a sacrifice of one’s life can be considered self-interested, the very notion of self-interest comes under scrutiny. 

The trouble with psychological egoism is that it presents an overly rational image of behaviour, which is not borne out in observation. Suppose a certain person has a terrible toothache, but an intense fear of dentists. It would be in their self-interest to attend the dentist and have this problem resolved, but they do not because of their phobia, and thus experience further pain, entirely against their self-interest. The same argument can be seen in the example of a smoker who despite knowing the risk of cancer proceeds to smoke anyway. Here we see a conflict between the pursuit of pleasure (hedonism) and self-interest.

Arguments such as these were presented by the English Bishop Joseph Butler in 1726, who remarked “The thing to be lamented is, not that men have so great regard to their own good or interest in the present world, for they have not enough.” Butler also observed the circularity of the central argument behind psychological egoism: if a person willingly performs an action, they must derive personal enjoyment from it, therefore people only perform acts that give them personal enjoyment – this is circular logic, as the conclusion is identical to the hypothesis.

In a rare case of David Hume siding with a religious perspective, Appendix II of Enquiries Concerning the Principles of Morals (written in 1751) goes to great lengths to oppose the view that self-love can be seen as the motivating force behind benevolence, friendship and public spirit. Hume concludes his comments on this subject by observing that vengeance can lead to a person sacrificing their interests in pursuit of revenge, and notes: “what a malignant philosophy must it be that will not allow to humanity and friendship the same privileges which are undisputably granted to the darker passions of enmity and resentment.” 

None of this means that an individual cannot choose ethical egoism as their moral system (free will means we each get to choose our own ethics), it simply calls into doubt the principles of psychological egoism upon which this position is founded.


The Prisoner Dilemma 

Morgan_jail_2 The classic Prisoner Dilemma has its roots in game theory, but for our purposes we shall put aside a mathematical interpretation and focus on Albert Tucker’s version of the dilemma in terms of prison sentences. It is Tucker’s account of the earlier work by Merrill Flood and Melvin Dresher in 1950 which gives us the name ‘The Prisoner Dilemma’ which has stuck. Here is the basis of the quandary:

You and an accomplice have been arrested for robbing a bank, and you both care more about your own freedom than that of your accomplice. The district attorney makes you the following offer: “You can either confess, or remain silent. If you confess, and your accomplice remains silent, I will drop all charges against you and see that your partner is put away for some serious time. If they confess and you don’t, they go free and you do the time. If you both confess, you’ll both be convicted, but I’ll see to it that you get an early parole. If neither of you confess, I’ll prosecute you both for firearms possession, and you’ll get small sentences.” You cannot communicate with your accomplice by any means: you must make your decision alone. 

The horns of the dilemma, so to speak, lie in the fact that irrespective of what the other prisoner chooses, each is better off confessing than remaining silent. If the other prisoner confesses, your best choice is to confess – otherwise you go away for a long time. If the other prisoner remains silent, your best choice is to confess – as that way you go free.

But the outcome this implies – that both confess – is far worse than if both remain silent. If both confess, both go to jail for a while (albeit they might get out earlier than if sold out by their accomplice). But if both remain silent, they each are charged with only a minor offence. Worse than going free, of course, but far better than the jail time implied if both confess. 

The Prisoner Dilemma (often abbreviated PD) is of considerable interest in game theory, and also in moral philosophy. It is sometimes asserted that because the outcome one receives is better in the case of confession – irrespective of what the other prisoner chooses – that confession is the rational choice. Of course, this viewpoint brings into question what we mean by the term ‘rational’.

At the heart of the issues raised are questions of trust – if both prisoners trust the other party to take their mutual interests into account, remaining silent becomes a viable strategy. In general terms, the question of co-operation always revolves around trust. 


Tit for Tat

In the early 1980s, the Political scientist Robert Axelrod ran a tournament based around the iterative Prisoner’s Dilemma. In this version of the problem, there are many participants, each of whom encounter each other multiple times in a version of the basic dilemma (usually rendered economically – in terms of different degrees of gain, rather than different degrees of penalty). The choices remain the same: to co-operate, or to defect (betray), with the payoff being greatest when the player defects while the other party co-operates, but with mutual gain if both co-operate in excess of the small gain if both defect. 

Axelrod asked fellow academics to submit computer programs as players to this tournament, and many different approaches were submitted with great varieties of complexity and strategy. However, the winning approach was the simplest submitted, a strategy named “Tit for Tat” by its creator Anatol Rapaport. It consists of the following two rules:

When you first meet another player, co-operate.
Thereafter, choose the response that the other player chose when last encountered. 

The extraordinary success of the Tit for Tat strategy has prompted considerable interest in its potential application as an explanatory principle in evolutionary biology. Before pursuing this, it is worth noting that the Tit for Tat strategy is further improved by adding an element of forgiveness: in the absence of this, Tit for Tat can enter into a vengeful pact with another player from which it never recovers (you defect, so I defect, so you defect and so on). In “Tit for Tat with forgiveness”, a small chance is given for the program to choose to co-operate after the opponent defects – allowing a chance of escape from a cycle of defections.

Additionally, there is an alternative successful strategy “Tit for two Tats” – which allows the other player one defection before retaliating. The success of this strategy depends upon the nature of the other strategies encountered: against very aggressive opponents, Tit for Tat will outperform Tit for two Tats, but against a mixed crowd it performs approximately as well, and occasionally better. 

Axelrod concluded from his experiments that a successful strategy at the iterated Prisoner’s Dilemma had to be nice (never the first to defect), retaliating (willing to defect), forgiving (willing to attempt to regain trust by breaking a defection cycle) and non-envious (not specifically attempting to outscore individual opponents). This leads to the conclusion that even a selfish individual finds their self-interest in being nice, forgiving and non-envious – not to mention that contrary to the aphorism, nice guys can finish first.


Tit for Tat in Nature 

Since Axelrod’s work, there have been many observations of Tit for Tat-like behaviour in nature, leading some to label this approach an evolutionary stable strategy – that is, a pattern of behaviour which effectively resists ‘invasion’ by alternative strategies. (It works so well, it’s hard for anyone to consistently take advantage of it).

Various primate species, including vervet monkeys and olive baboons, have been observed coming to the assistance of animals that have helped them in the past, which is often linked to a Tit for Tat-like strategies, but even animals considered to be less social show signs of related behaviour. For example, tree swallows of the species Tachycineta bicolour live in groups, but not all the birds in each group raise chicks. Even though non-breeders occasionally kill the chicks of parents and usurp a nest, parents tolerate the non-breeders (there are many hypothetical benefits to doing so, including the increased strength of numbers to resist predators, but we should be careful about drawing such conclusions). Co-operation is the norm in this species, and in many others. 

Threespine_stickleback A bizarre experiment with sticklebacks (small river fish) by Manfred Milinski also demonstrated Tit for Tat like behaviour. The fish are preyed upon by larger predators such as pike and perch, but when such a predator is nearby, a small number of fish will break from the shoal for a “predator inspection visit” – presumably to gauge the degree of threat. Working together, the small fish can get closer, and scatter if attacked, which can confuse the predatory fish.

In the experiment, Milinski used an adjustible mirror to give the impression to a lone fish that it was accompanied by a partner. The position of the mirror determined whether their reflection seemed to be approaching with them, or falling behind. The behaviour of the fish in the experiment was exactly what was expected by Tit for Tat – if the mirror was set to ‘co-operate’, the fish would complete an inspection, but if the mirror was set to ‘defect’ they would rapidly withdraw. Forgiveness was also observed – with a ‘defecting’ mirror, the fish would repeatedly try to advance, only to withdraw again when it realised it would not be accompanied all the way. 

It transpires that co-operation in nature is common – in manners that go beyond what could strictly be interpreted as Tit for Tat strategies. For example, hermit crabs are known to conduct trading. Hermit crabs live in empty gastropod shells, and intact shells are in short supply (as most predators damage the shell when they eat the gastropod). If a large hermit crab living in a small shell encounters a small hermit crab living in a large shell, they exchange the shells – initiating the exchange by tapping or shaking the other party in a specific manner. The exchange only occurs when there is mutual benefit.

Entire coral reef ecosystems rely on co-operation at various different scales, from the symbiotic alliance of polyps and algae at the base of the ecology to the clownfish’s relationship with its anemone home. Most striking is the behaviour of cleaner wrasse (and other similar fish) which set up ‘cleaning stations’ on the reef, where other fish come to have dead skin and parasitic organisms picked off. Both parties benefit: the cleaner fish takes the skin and parasites as nutrition, and the visitor fish is healthier for the service. Even predatory fish that would normally eat fish the size of the cleaner fish leave the cleaners alone. This remarkable example of a naturally occurring “service industry” goes further – there are cleaner shrimp which provide the same service to the cleaner fish! The creatures that provide these services are apparently exempt from predation. 



The success of the Tit for Tat strategy (both in nature and experiment) suggests that trust is compatible with self-interested behaviour. The classical image of nature as “the war of all against all” (as characterised by Thomas Hobbes) does not bear up to scrutiny. While predation is an inescapable fact of nature, co-operation is widespread and essential to the health and well-being of many communities and ecosystems. Trust occurs within and between species that we would not normally consider capable of cognising ‘trust’ as a concept – it is a natural and beneficial behaviour.

Ironically, even if one believes in psychological egoism (despite the objections raised above), and chooses ethical egoism as one’s moral system, co-operation is still necessitated as self-interested. Even the egoist must co-operate if they are to behave in a rationally self-interested fashion. 

Some individuals may argue from the success of Tit for Tat to psychological egoism, taking the reductionistic view that since altruism (at least in the guise of co-operation) is necessitated by rational self-interest, then self-interest is sufficient to explain altruism, and we do not need to invoke altruism at all. Here, the fallacy of reductionism as an inherently superior view is exposed, as this is to suggest that identifying constituent components is more instructive than observing whole behaviours – but nothing in our understanding of quarks, the theoretical components of all matter, tells us anything about atomic forces, chemistry, biology or psychology, and similarly the recognition that self-interest is compatible with co-operation cannot reduce co-operation to self-interest by any mechanism other than the explicit choice to view things this way.

We will not eliminate ethical egoism – there will always be some individuals who prefer this perspective – but the wonderful thing about the success of the Tit for Tat strategy is the capacity to turn selfishness into co-operation. Since ethical egoism dictates individuals ought to act in their own self-interest, Tit for Tat demonstrates that such egoists ought to co-operate with the people around them (at least under the conditions dictated by the Prisoner Dilemma). This is a considerably weaker stance than that espoused by the Golden Rule, but at least it means that in the minimal ethical case, we can all work together towards common goals. 

Some examples of animal co-operation and Tit for Tat in nature used in this piece were taken from this webpage. 


Feed You can follow this conversation by subscribing to the comment feed for this post.

Something's always bothered me about the Prisoner's Dilemma: it models a conflict in which resources aren't scarce. There's enough freedom for everybody. Should that be a concern when looking at ethical systems?

c.f. The Sharer's Dilemma: link

If you say, "a specific connotation of the word ‘selfish’ is to act in one’s own benefit to the detriment of others" then maybe "going to the doctor when feeling ill" IS arguably selfish - as you're depriving other individuals of the doctors' time.

In (a more serious) response to the article as a whole - perhaps those who act benevolently at the true expense of themselves are doing so because that's the kind of world they wish to live in. Maybe they, subconsciously, believe that others will follow in their example. Tit for tat, with adjustment for possibly inspiring a greater number than the immediate 'neighbours' and so skewed drastically in favour of 'tolerance'.

caller #6: it is the case that the Prisoner's Dilemma ignores issues of scarcity of resources; I don't think this is necessarily problematic as a general case observation, however. In our modern Western societies, we live in absurd plenty and questions of scarcity are rarely enormously relevant. (However, there are odd special cases - such as the especially scarce commodity of 'silence', which can often only be generated through co-operation).

Furthermore, many ethical issues do not relate to limited resources. Consider, as a trivial example, road usage. Impolite driving gains the selfish individual a few scant seconds or minutes of time on their journey, but this time is not a limited resource we are all competing for, per se, as it is universally available. No matter how much time the selfish driver gains, there is still plenty of time left for everyone else! :) Here, the Prisoner's Dilemma is apt - as if all drivers co-operate to use the road, there are fewer delays and everyone benefits.

Issues relating to scarce resources quite frequently turn into questions of politics, which I'm hoping to skirt around for now.

As for the role of freedom in ethics, this is certainly an issue. We'll definitely look at this issue in the context of Kant shortly.


Bezman: if you are suggesting that a motivating factor in benevolent behaviour is the desire to encourage further benevolent behaviour, I would tend to agree. Benevolence (to invoke a common ethical watchword) is an easier behaviour to universalize than selfishness - as the selfish person does not usually desire to encourage other people to act selfishly. :)


Thank you both for the comments!

I was suggesting exactly that.

Once again, your succinctness and clarity surpasses mine.

"In our modern Western societies, we live in absurd plenty and questions of scarcity are rarely enormously relevant."

Where they are relevant, they tend to be relevant on a large scale. Consider the UK National Health Service. How many suicides could be avoided by increasing funding to early-intervention depression counselling services for each cancer death avoided by early-stage treatment with Herceptin? Or should UK plc simply push more resources into the service to do both? Or less?

Bezman: I merely precised your thoughts; it was you who raised the idea. ;)

Peter: your comment here gives me pause for thought. When I wrote of 'plenty' I meant in respect of the basics of life, such as food and water; the issue of health care is far trickier - as here there is potentially no upper limit for expenditure in our modern cultures. Counselling services are a good example, as there is not enough to go around, and waiting lists make such things rather redundant... How much economic force can be diverted to such issues?

I feel that in this particular instance the community should be able to offer support to individuals. If we rely on the health services to provide such things, we set precedents that may be hard to maintain. One of the great tragedies of the twentieth century is how the local communities of Western nations have become eroded in favour of national or regional identities. There is some benefit in this, I'm sure, but there are also terrible costs.

But this whole issue moves beyond the ethical and into the political. :)

Best wishes!

Ethical egoism has never precluded cooperation and 'benevolance'. In fact, as you mentionned, has even often been presented as its sole motivation.

In a situation where selfishness is not exerted by a mindless machine, we can expect any semi-intelligent entity (we can hope that includes humans) to minimally seek to achieve a solution that's Pareto optimal (to keep the discussion in the paradigm of game theory), if only to elicit reciprocity, should the tables be turned.

Where I'm dubitative about what your article is when you use quarks versus biochemistry make a point against ethical egoism as true reason behind cooperation. This is what Dennette would probably likens an 'intuition pump', it summons up an intuition but it demonstrate nothing.

In fact, it is generally agreed (in the scientific community at least) that identifying constituants, causes, and rules explaining a phenomenon is generally 'superior' (more instructive to use your words), than mere observation of the said phenomenon.

I personnally have no conviction as to the 'cause' of altruism, but after reading a good lot on the subject, I've yet to find a somewhat convincing demonstration that egoist cannot be it.

As you pointed out however, this is of little practical value since there is no abiguity that in a non-zero sum environment (such as life), cooperation will pretty much always be a desirable behavior regarless of the motivation behind it... So there is hope =)

Oliver: firstly, many thanks for your kind words in your other comment! I'm always glad when this ragtag fugitive fleet of intellectual malcontents and weirdos acquires a new recruit! ;)

"In fact, it is generally agreed (in the scientific community at least) that identifying constituants, causes, and rules explaining a phenomenon is generally 'superior' (more instructive to use your words), than mere observation of the said phenomenon."

I'm not certain about this claim. In psychology, for instance, observation (case studies) are generally valued as much or more than theoretical work - case studies are permanently valuable, while theoretical work is only valuable in so much as it continues to contribute to the development of useful theoretical frameworks. For instance, what value any of the research into phrenology now that this field has been abandoned? While direct observational work conducted in the 19th century remains valuable because it is not tied to specific theoretical frameworks.

The point I am making is not that theories have no value, it is that reductionistic thinking has no a priori claim to superior explanatory value.

The point about quarks is that the theory of quarks and leptons is valuable for explanation at an atomic scale, but has no value for explanation at a chemical scale, much less at a biological or psychological scale: looking for "smallest components" is not a guaranteed way of getting the best handle on a situation.

Scientists get into a reductionistic habit and forget sometimes that "smaller causes" are not always the best explanation. We forget, for instance, that descent with modification - Darwin's idea - is not reductionistic in nature. It only became so in "the new synthesis" when it was assumed that genetics explains everything in this process (a claim I find premature).

In the context here, my position is that animal behaviour in general, and human behaviour in specific, is a holistic outcome of many different forces, and is not best explained by reduction to single factors - although reductionistic models can shed light on the elements that contribute to the overall behaviour.

One must take into account belief systems (resulting from our neural architecture), emotions (resulting from our endocrine system) and instincts (resulting from our nervous system) as contributory factors to behaviour - no one thing stands out as the dominant component.

At least, such is my view. :)

Thanks for the comments!

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)