What are friends for and how can a friendship be tested? By behaving altruistically, would be the most common answer and by sacrificing one's interests in favour of one's friends. Friendship implies the converse of egoism, both psychologically and ethically. But then we say that the dog is "man's best friend". After all, it is characterized by unconditional love, by unselfish behaviour, by sacrifice, when necessary. Isn't this the epitome of friendship? Apparently not. On the one hand, the dog's friendship seems to be unaffected by long term calculations of personal benefit. But that is not to say that it is not affected by calculations of a short-term nature. The owner, after all, looks after the dog and is the source of its subsistence and security. People and dogs have been known to have sacrificed their lives for less. The dog is selfish it clings and protects what it regards to be its territory and its property (including and especially so - the owner). Thus, the first condition, seemingly not satisfied by canine attachment is that it be reasonably unselfish.

There are, however, more important conditions:

• For a real friendship to exist at least one of the friends must be a conscious and intelligent entity, possessed of mental states. It can be an individual, or a collective of individuals, but in both cases this requirement will similarly apply.

• There must be a minimal level of identical mental states between the terms of the equation of friendship. A human being cannot be friends with a tree (at least not in the fullest sense of the word).

• The behaviour must not be deterministic, lest it be interpreted as instinct driven. A conscious choice must be involved. This is a very surprising conclusion: the more "reliable", the more "predictable" the less appreciated. Someone who reacts identically to similar situations, without dedicating a first, let alone a second thought to it his acts would be depreciated as "automatic responses".

For a pattern of behaviour to be described as "friendship", these four conditions must be met: diminished egoism, conscious and intelligent agents, identical mental states (allowing for the communication of the friendship) and non-deterministic behaviour, the result of constant decision making.

A friendship can be and often is tested in view of these criteria. There is a paradox underlying the very notion of testing a friendship. A real friend would never test his friend's commitment and allegiance. Anyone who puts his friend to a test (deliberately) would hardly qualify as a friend himself. But circumstances can put ALL the members of a friendship, all the individuals (two or more) in the "collective" to a test of friendship. Financial hardship encountered by someone would surely oblige his friends to assist him even if he himself did not take the initiative and explicitly asked them to do so. It is life that tests the resilience and strength and depth of true friendships not the friends themselves.

In all the discussions of egoism versus altruism confusion between self-interest and self-welfare prevails. A person may be urged on to act by his self-interest, which might be detrimental to his (long-term) self-welfare. Some behaviours and actions can satisfy short-term desires, urges, wishes (in short: self-interest) and yet be self- destructive or otherwise adversely effect the individual's future welfare. (Psychological) Egoism should, therefore, be re-defined as the active pursuit of self- welfare, not of self-interest. Only when the person caters, in a balanced manner, to both his present (self-interest) and his future (self-welfare) interests can we call him an egoist. Otherwise, if he caters only to his immediate self-interest, seeks to fulfil his desires and disregards the future costs of his behaviour he is an animal, not an egoist.

Joseph Butler separated the main (motivating) desire from the desire that is self- interest. The latter cannot exist without the former. A person is hungry and this is his desire. His self-interest is, therefore, to eat. But the hunger is directed at eating not at fulfilling self-interests. Thus, hunger generates self-interest (to eat) but its object is eating. Self-interest is a second order desire that aims to satisfy first order desires (which can also motivate us directly).

This subtle distinction can be applied to disinterested behaviours, acts, which seem to lack a clear self-interest or even a first order desire. Consider why do people contribute to humanitarian causes? There is no self-interest here, even if we account for the global picture (with every possible future event in the life of the contributor). No rich American is likely to find himself starving in Somalia, the target of one such humanitarian aid mission.

But even here the Butler model can be validated. The first order desire of the donator is to avoid anxiety feelings generated by a cognitive dissonance. In the process of socialization we are all exposed to altruistic messages. They are internalized by us (some even to the extent of forming part of the almighty superego, the conscience). In parallel, we assimilate the punishment inflicted upon members of society who are not "social" enough, unwilling to contribute beyond that which is required to satisfy their self interest, selfish or egoistic, non-conformist, "too" individualistic, "too" idiosyncratic or eccentric, etc. Completely not being altruistic is "bad" and as such calls for "punishment". This no longer is an outside judgement, on a case by case basis, with the penalty inflicted by an external moral authority. This comes from the inside: the opprobrium and reproach, the guilt, the punishment (read Kafka). Such impending punishment generates anxiety whenever the person judges himself not to have been altruistically "sufficient". It is to avoid this anxiety or to quell it that a person engages in altruistic acts, the result of his social conditioning. To use the Butler scheme: the first-degree desire is to avoid the agonies of cognitive dissonance and the resulting anxiety. This can be achieved by committing acts of altruism. The second-degree desire is the self-interest to commit altruistic acts in order to satisfy the first-degree desire. No one engages in contributing to the poor because he wants them to be less poor or in famine relief because he does not want others to starve. People do these apparently selfless activities because they do not want to experience that tormenting inner voice and to suffer the acute anxiety, which accompanies it. Altruism is the name that we give to successful indoctrination. The stronger the process of socialization, the stricter the education, the more severely brought up the individual, the grimmer and more constraining his superego the more of an altruist he is likely to be. Independent people who really feel comfortable with their selves are less likely to exhibit these behaviours.

This is the self-interest of society: altruism enhances the overall level of welfare. It redistributes resources more equitably, it tackles market failures more or less efficiently (progressive tax systems are altruistic), it reduces social pressures and stabilizes both individuals and society. Clearly, the self-interest of society is to make its members limit the pursuit of their own self-interest? There are many opinions and theories. They can be grouped into:

• Those who see an inverse relation between the two: the more satisfied the self interests of the individuals comprising a society the worse off that society will end up. What is meant by "better off" is a different issue but at least the commonsense, intuitive, meaning is clear and begs no explanation. Many religions and strands of moral absolutism espouse this view.

• Those who believe that the more satisfied the self-interests of the individuals comprising a society the better off this society will end up. These are the "hidden hand" theories. Individuals, which strive merely to maximize their utility, their happiness, their returns (profits) find themselves inadvertently engaged in a colossal endeavour to better their society. This is mostly achieved through the dual mechanisms of market and price. Adam Smith is an example (and other schools of the dismal science).

• Those who believe that a delicate balance must exist between the two types of self-interest: the private and the public. While most individuals will be unable to obtain the full satisfaction of their self-interest it is still conceivable that they will attain most of it. On the other hand, society must not fully tread on individuals' rights to self-fulfilment, wealth accumulation and the pursuit of happiness. So, it must accept less than maximum satisfaction of its self-interest. The optimal mix exists and is, probably, of the minimax type. This is not a zero sum game and society and the individuals comprising it can maximize their worst outcomes.

The French have a saying: "Good bookkeeping makes for a good friendship". Self-interest, altruism and the interest of society at large are not necessarily incompatible.