We have seen that it is characteristic of a certain very popular style of contemporary moral philosophy to hold that there are no questions at all that philosophy should not ask. Those who take this line tend to see their stance as one of heroically bold and open-minded free thought. They apparently regard the contrary idea, which I have been defending by describing the virtue of innocence, as timid, no-debate obscurantism.

So far in this book I have raised lots of questions about friendship. I have said something about the nature of friendship. I have considered some of the most famous philosophical theorists of friendship, and looked at just a few of the many works of philosophy, art, and literature where friendships are important or central. I have said something about the way in which a key part of the delight of friendship, much of the time, is how unemphatic, unselfconscious, tacit, innocent, spontaneous, and unmoralistic it is.

So we should begin the philosophy of friendship not with a formal, logically watertight definition, but rather with a rough definition of friendship, such as my own characterisation of friendship as benevolent companionship over time, or such as the dictionary definition that I quote in Chapter 3. We can also base our reflections on the question ‘What is opposite to friendship?’, and on an explicit and conscious list of examples of friendship. And it’s all the better if this list is explicit and conscious, because our thoughts about friendship almost certainly will be guided by our own experiences of friendship whether we realise it or not, so we will be using some list of examples either consciously or unconsciously; consciously is preferable.

At the most basic level of natural history, some creatures simply do not get on well living solitarily, and do much better when they are part of a larger group of some sort. There are species of mammal for which it is natural and normal to live alone (except for breeding and rearing purposes): for example snow leopards, and polar bears, and platypuses, and moose. But such species as lions, and sheep, and gorillas, and sparrows, and dolphins are plainly social species. Elephants and chimps are well known for the clear distress that they show at the loss of a member of their group. Every dog owner knows how a pet dog will go on looking for another pet dog, or for a member of its human family, after they have died. Anyone who wants to keep rabbits is advised by all the experts to make sure that it is indeed rabbits, plural, and not just a single rabbit. Even cats, though they often seem to project a sleek air of solitary superiority, are in fact nearly always busy seeking out the company both of humans (especially ones with food and warmth to share) and of other cats (especially ones that they can either fight or fornicate with). For all these kinds of animals, it is obviously and demonstrably true that living together is better than living in isolation. And this seems to be such a general, physiology-level fact that it is almost superfluous to ask why they do better in company: they just do, in a whole range of very basic ways.

There isn’t a logically watertight formal definition. And the fact that we can’t formally define friendship teaches us some important lessons about the nature of philosophical inquiry – lessons that we have (or should have) been learning ever since Plato, and that Wittgenstein helps to remind us of. But loosely and roughly, and without any promise of counter-example-proofness: friendship is benevolent companionship over time. Here what I mean by ‘companionship’ involves mutual second-personality: it means that I see someone else as a person who sees me as a person.

Perhaps the most influential and familiar taxonomy of the varieties of love – a quick bit of googling shows its striking prevalence even today, nearly sixty-five years after publication – comes from C. S. Lewis’s 1960 classic of popular theology, The Four Loves.

In the Nicomachean Ethics Books VIII–IX, Aristotle is the author of an immensely influential study of philia – friendship, as it is usually translated, though I have already suggested that this is a misleading translation, and shall be saying more about the point in a moment. Aristotle famously divides philia into three kinds. The first and ‘highest’ of these three kinds is the focal and paradigmatic case of philia; the other two kinds are ‘lower’ and derivative cases, which may sometimes perhaps only count as philia in an extended or metaphorical sense (kat’ analogian, in Aristotle’s terminology). The focal case is ‘the friendship of virtue’, a relationship of intimate companionship and shared moral/intellectual endeavour between educated and aristocratic men. The two derivative cases of philia that most interest Aristotle are ‘the friendship of honour’, typified by partnerships formed between free adult men in order to achieve the ends of honour such as political office or military or sporting success, and ‘the friendship of advantage’, of which friendships based on amusement or pleasure, and on money-making, would both apparently count for Aristotle as examples.

The award-winning children’s film Shrek came out at just the right time (in 2001) to be a favourite with our children when they were small, for family viewing (in fact, for multi-repeat family viewing). Shrek is a fairy-tale animation about a sad, lonely, angry, alienated ogre in a world that rejects him for being scary, ugly, and different. With Donkey’s help, Shrek has just rescued the beautiful Princess Fiona from a dragon. Since this is a fairy-tale, Shrek has of course fallen in love with Fiona, but since he is an ogre and she is (apparently) a princess, he is sure his love is in vain. He has always kept up a tough and cynical exterior, but now he has stormed off from Princess Fiona because he has overheard her saying ‘Who could love a hideous, ugly beast?’ – and mistakenly assumed that she meant him.

You could go up into the heavens, you could contemplate the whole nature of the world and the beauty of the stars from up there – but if you had no one to tell about it, your revelation would bring you no pleasure. Whereas if you did have someone, it would be the most wonderful thing ever.

Aristotle’s and C. S. Lewis’s systematising theories of friendship both face some of the same objections, and two in particular. First, they treat what is really only local to their own social and historical contexts as if it were universal and natural. Secondly and connectedly, their own prejudices are all too visible in what they choose to centre, and what they choose to marginalise or ignore.

The phrase ‘tacit knowledge’ is normally connected with the work of the Hungarian-British scientist and philosopher Michael Polanyi (1888–1976). Polanyi’s (1958) slogan in philosophy was ‘We know more than we can tell’; ‘tacit knowledge’ was Polanyi’s name for all the things that we know but cannot tell.

Chapter 1 has raised a whole range of questions about friendship. Of course many of them had already occurred to me anyway, and no doubt to you too. But a lot of these questions came up, or came up again, when I did my preliminary reading around for this book in the extant philosophical literature on friendship. This goes back a very long way historically, and even today it is still growing alarmingly rapidly.

I have been saying things about the philosophy of love at least some of which it wouldn’t really do to say about the philosophy of friendship – even though some of them are true. Why wouldn’t it do? Well, we come back here to a problem that I noted in the literature of the philosophy of friendship as far back as Chapter 2: as I called it there, po-facedness.

In general, then, moralism in ethical philosophy means obsessively reducing everything we care about to ought and duty, right and wrong, permission and obligation. So moralism about friendship means doing this obsessive reduction in the specific case of friendship.

Alongside the involvement of friendship with spontaneity, unselfconsciousness, and humility, friendship also involves at least one other virtue that necessarily involves tacit knowledge. I shall call this virtue innocence. Innocence is an interesting and difficult concept. In rather the same way as humility, spontaneity, and unselfconsciousness, its difficulties have to do with the familiar idea of ‘thinking the unthinkable’. Innocence, in brief, is the virtue of not thinking the unthinkable – of having the tacit knowledge that enables you to avoid even thinking about what it is better not even to think about.