can conceive of a Holmes who dislikes music and doesn’t smoke tobacco, but not of one who failed to solve problems that others too had found insoluble, nor one who is married with three children. Such a character would simply have ceased to be Sherlock Holmes.

Similar considerations apply outside of fiction. How far can any of us have been different and yet still claim the same essential identity? T.S. Eliot, for example, might have died in his teens, might have been an absinthe-drinker, might never have written The Waste Land, might have worn a gold lamé suit. He would still have been T.S. Eliot for all that – if a very different one from the Eliot with whom we are familiar. But he would not have been T.S. Eliot if he had been born of different parents and in another century; there are some characteristics that are necessary if we are to still be ourselves, others that are merely possible, whether also actualised or not.

Leibniz, that universal genius, in speaking of possible worlds, was a distant progenitor of what is now known as many-worlds semantics or modal logic, dealing with the modes of necessity and possibility across and within worlds. But how are we to understand this talk of possible worlds? For Saul Kripke, ‘a possible world isn’t a distant country that we are coming across, or viewing through a telescope’ – rather, it is a logical construct. He gives the example of a throw of two dice which results in their displaying two numbers face up. We can calculate that there are thirty-six possible states of the two dice, but only one is here instantiated, say, a three and a one. All the other thirty-five states, including the double six we hoped for, remain unsubstantiated, unfulfilled possibilities. The three and the one are instantiated in this world, what we think of as the actual world: all the remaining values belong to other possible worlds. There is our world which is not only possible but also actual; there are an infinity – or infinities – of other worlds which are possible but not actual, existing only as logical constructs, or as fictions.

In other possible worlds there exist counterparts of ourselves. We can identify them as our counterparts by reasons of our origin – being born of the same parents – but our lives and life-experiences may significantly diverge across worlds. I am currently sitting in a room, writing an essay for PN Review. In countless other possible worlds, other Roger Caldwells are doing the same. In one of them he gives up in despair. In another he is interrupted by a phone call, and forgets what he wanted to say. In this world, however, he goes on to complete the essay. There is a certain pathos in this: an infinity of near-identical figures doing more or less the same, and who are unaware of each other, cannot see or speak to one another, cannot ever meet. Our counterparts are the numerous possibilities we conceive for ourselves, but fail to actualise. Borges in 1934 planned in exquisite detail his own suicide, to be carried out in a room in the Hotel Delicias. As we know, he never executed the plan. In this world he lived to enjoy old age, but in another possible world the plan was executed exactly as conceived, with the result that in that world Borges killed himself in 1934.

But in speaking like this we are unwittingly falling into the trap against which Kripke warned us, of envisaging these possible worlds as somehow actual rather than merely possible, rather in the way we respond to characters in novels or dramas when we give assent to fictions. We see them as somewhere instantiated rather than as mere logical possibilities. However, according to David Lewis, in doing so we are seeing correctly. In his magisterial work On the Plurality of Worlds (1986), he argues eloquently for what he calls modal realism, that is, that every possibility is in some world actualised. This is the equivalent of Leibniz’s God, who, having considered the infinity of possible worlds, decides on a sudden whim to actualise not just the one world, but all of them. For Lewis, God does not come into the picture (though there are presumably worlds in which God is actualised, unless he is a contradiction in terms), but he turns the tables neatly on those who would protest at this superfluity of worlds: ‘What a remarkable bit of luck for us if the very world we are part of is the one that is absolutely actual! Out of all the people there are in all the worlds, the great majority are doomed to live in worlds that lack absolute actuality, but we are the select few. What reason could we ever have to think it was so?’ It should be noted that these are not entirely new concerns. The Greeks, as usual, got there first. Democritus and Epicurus both calmly accepted that there was an infinity of worlds. The Stoics, by contrast, believing that the same world was endlessly repeated down to the last detail after each world-conflagration or ecpyrosis, speculated as to whether in each reconstituted world its inhabitants were identical with their predecessors or were different, thus anticipating present-day talk of counterparts.

For Lewis, when we speak of something’s being actual what is meant is only that it is actual in this world – for example, that it is something we can point to. But in other worlds exactly the same situation applies. I, Roger Caldwell, actualised in this world, see my counterpart in another world as existing only in the realm of possibility. But for that Roger Caldwell in the other world, he is the real Roger Caldwell, and I in turn am to him only his shadowy counterpart, a possibility of existence, not actually actual, not quite real, a sort of fiction. For Lewis, then, what is seen to be actual simply depends upon what world you inhabit. Sherlock Holmes and Mrs Dalloway are fictions in this world: in others, since on Lewis’s account all possibilities are somewhere actualised, Sherlock Holmes does indeed live at 221b Baker Street and Mrs Dalloway is indeed preparing for another party.

But, of course, this returns us to the extravagant ontologies that the likes of Russell and Kripke resist in the name of common sense. In these worlds exist not only Sherlock Holmes and Mrs Dalloway but all the characters in every novel or drama ever written, or that will come to be written, or that could be written, and also all conceivable creatures from talking donkeys to centaurs and basilisks. This ontological overload seems too much a mad logician’s dream and the sane person’s nightmare. It is even more extravagant than the solution to the paradoxes of quantum physics espoused by such multiverse theorists as David Deutsch, although Lewis deals with worlds that are complete in themselves, whereas Deutsch deals with branching worlds. In the sub-atomic world with which quantum physics is concerned what we think of as particles exist in a superimposition of states – until they are measured, or observed, in which case we have what is known as the collapse of the wave function: that is, the particle acquires a particular position or velocity. The question is: what happens to all the other values? (The

20

PN Review 202 issue is logically akin to Kripke’s throw of the two dice, mentioned above.) Deutsch’s solution is that these too are instantiated, but in worlds that branch off from this one every time there is a collapse of the wave function. Here Sherlock Holmes and Mrs Dalloway fail to make their appearance, since they were never actualised, but we ourselves do, and again, we are multiply instantiated, existing in numerous mutually inaccessible split-off worlds.

Certainly, these are dizzying prospects, but surely we have moved from fiction to science-fiction. David Lewis’s book is one of the most densely argued works in modern philosophy – I have only scratched the surface of it here – but it has to be said that his brand of modal realism finds few full-blooded takers. David Deutsch’s ‘realist’ multiverse approach to quantum physics remains a minority view among scientists – while many quantum physicists are content if the mathematics works out and simply refuse to be embroiled in ontological speculation, all the more so if it leads to such counterintuitive conclusions as those of Deutsch.

For most of us, however, we may question what difference it makes in practice whether, as with Kripke, these other possible worlds remain as logical constructs or, with Lewis (and Deutsch), they are also actualised. Is this a difference that makes no difference? After all, we cannot have access to these worlds. What my counterpart does in another world is no concern of mine – I can’t reach across to him, nor he to me. There are those, however, who argue that the difference has ethical consequences. If my counterpart is merely a logical fiction it is certainly true that he is of no concern of mine. If, however, he is actualised in other worlds, he is alive in exactly the same way I am, he has had some of the same life-experiences that I have; not only this, but he exists in countless versions. At the very least this diminishes (drastically) our sense of our own uniqueness – and, potentially at least, our sense of personal responsibility. Why should I try to be good if there are already countless versions of myself that are better and countless versions that are worse? Why should I strive for perfection if elsewhere I have already achieved it, if there are countless counterparts of myself who will do the things that I ought to have done? The T.S. Eliot who lives in this world need not bother to write The Waste Land (in other worlds it is already written; in some of them it has been written by Ezra Pound); instead, knowing he was a great poet in other worlds, he might concentrate on a life of dissolution and frivolity, giving himself up to absinthe, and amassing a collection of gold lamé suits.

Needless to say, Lewis has answers to these objections, though this is not the place to adjudicate on them. I want now to move on to the obvious fact that, although in an empirical sense, we live in the one world, in another sense we inhabit numerous worlds. Sir Thomas Browne, in Religio Medici, famously saw man as ‘that great and true Amphibium whose nature is disposed to live not onely like other creatures in divers elements, but in divided and distinguished worlds’. Unlike other animals, we see ourselves and others in terms of possibilities – of what might be, and of what might have been. We tell stories about ourselves and to ourselves about possible futures and possible pasts – some of which may be true. We inhabit pictorial, acoustic, poetic, fictional worlds, worlds made for us by artists: this is the subject-matter of Nelson Goodman’s by now classic text, Ways of Worldmaking (1978). Of course, artists do not create complete worlds – but they are complete enough for their purpose. Further, what is possible in one artistic world is impossible in another – they operate according to different rules. The sound-world of Bach is distinct from that of Debussy as it is from that of Beethoven – though admittedly there are Bachian echoes in, say, the latter’s Opus 110 Piano Sonata, it is nonetheless Bach transformed into Beethoven. There is a Dickensian world and a Tolstoyan one – a Pecksniff would not be at home in War and Peace, and it is impossible to imagine Anna Karenina in Bleak House. What is possible in Mallarmé’s world is impossible in that of A.E. Housman: there can be no luckless lads in the former, no fauns and nymphs in the latter. We cannot live simultaneously in both Mallarmé’s and Housman’s worlds, but can only inhabit them successively. Further, all artistic worlds are compatible with living in this world. In this sense we live temporarily in innumerable different possible worlds when we flick through the pages of a poetry anthology. These worlds are not remote and inaccessible like the possible worlds of Kripke and Lewis, but ones we can inhabit: we continue to live in Emma Bovary’s world even as we put the book aside and search for our ticket when arriving at Liverpool Street Station in this world.

Goodman resourcefully explores some of the ways in which artists make these new worlds. But when he goes on to note ‘the very continuity and unity, the very affinity, of art and science and perception as branches of worldmaking’ we must take pause. He also tells us that ‘science and art proceed in much the same way with their searching and building’. But the truth is that they don’t: scientists are not worldmakers in the way that artists are. Neither is there parity between living in a Newtonian or Einsteinian world on the one hand and a Tolstoyan or Dostoevskyan one on the other. Nor is it simply true, for that matter, that the Newtonian picture is false and the Einsteinian one true: simply the latter one explains more than the other about this world. Art is not progressive in this sense: Virgil does not supersede Homer, Mahler doesn’t make Haydn redundant. Nor are we forced to choose between them; we can happily have both. Artists, as Nelson Goodman tells us, create new worlds – or partial worlds, or alternative sequences of this world (as Leibniz has it). But scientists don’t. Their business lies entirely with this world. Instead it is the artists who create for us possible worlds – thereby immeasurably expanding the imaginative space we live in – and making it possible for us to further reflect on this world from the vantage-point of another.

Roger Caldwell: ‘The Present King of France is Bald’: On Possible Worlds

21