Virtual History: Alternatives and Counterfactuals (12 page)

But uncertainty has outlived Einstein; and it has no less disconcerting implications for historical determinism. By analogy, historians should never lose sight of their own ‘uncertainty principle’ - that any observation of historical evidence inevitably distorts its significance by the very fact of its selection through the prism of hindsight.

Another modern scientific concept with important historical implications is the so-called ‘anthropic’ principle, which in its ‘strong’ version states that ‘there are many different universes or regions of a single universe each with its own initial configuration and perhaps with its own set of laws of science ... [but] only in the few universes that are like ours would intelligent beings develop’.
¹⁶¹
This naturally raises obvious problems: it is not clear what significance we should attach to the other ‘histories’ in which we do not exist. According to Hawking, ‘our universe is not just one of the possible histories, but one of the most probable ones ... there is a particular family of histories that are much more probable than the others’.
¹⁶²
This idea of multiple universes (and dimensions) has been taken further by physicists like Michio Kaku. The historian does not, it seems to me, need to take too literally some of Kaku’s more fantastic notions. Because of the immense amounts of energy which would be required, it seems doubtful if time travel through ‘transversible worm holes’ in space-time can be described as even ‘theoretically’ possible. (Apart from anything else, as has often been said, if time travel were possible we would already have been inundated with ‘tourists’ from the future - those, that is, who had elected not to travel further back in time to avert Lincoln’s death or to strangle the new-born Adolf Hitler.)
¹⁶³
Nevertheless, the idea of an infinite number of universes can serve an important heuristic purpose. The idea that - as one physicist has put it - there are other worlds where Cleopatra had an off-putting wart at the tip of her celebrated nose sounds, and is, fanciful. But it provides a vivid reminder of the indeterminate nature of the past.

The biological sciences have made similar moves away from determinism in recent years. Although Richard Dawkins’s work, for example, has a deterministic thrust to it, with its definition of individual organisms, including humans, as mere ‘survival machines built by short-lived confederations of long-lived genes’, he states explicitly in
The Selfish Gene
that genes ‘determine behaviour only in a statistical sense ... [they] do not control their creations’.
¹⁶⁴
His Darwinian theory of evolution is ‘blind to the future’ - Nature has no predestinarian blueprint. Indeed, the whole point about evolution is that replicator molecules (such as DNA) make and reproduce mistakes, so that ‘apparently trivial tiny influences can have a major impact on evolution’. ‘Genes have no foresight, they do not plan ahead’. Only in one sense is Dawkins a determinist, in that he rules out the role of ‘bad luck’ in natural selection: ‘By definition, luck strikes at random, and a gene that is consistently on the losing side is not unlucky; it is a bad gene.’ Thus those organisms which survive the slings and arrows of fortune are those best designed to do so: ‘Genes have to perform a task analogous to prediction ... [But] prediction in a complex world is a chancy business. Every decision that a survival machine takes is a gamble ... Those individuals whose genes build brains in such a way that they tend to gamble correctly are as a direct result more likely to survive, and therefore to propagate those same genes. Hence the premium on the basic stimuli of pain and pleasure, and the abilities to remember mistakes, to simulate options and to communicate with other “survival machines”.’
¹⁶⁵

Other evolutionists, however, take issue with this line of argument, with its still deterministic implication that the race goes to the strong individual organism (or ‘meme’ or ‘phenotype’, Dawkins’s other forms of replicator). As Stephen Jay Gould shows in his
Wonderful Life
, certain chance events - major environmental catastrophes like the one which apparently happened after the so-called ‘Cambrian explosion’ - do disrupt the process of natural selection.
¹⁶⁶
By completely changing long-standing ecological conditions, they render valueless overnight attributes honed over millennia in response to those conditions. The survivors survive not because their genes have designed and built superior ‘survival machines’ but often because vestigial attributes suddenly turn up trumps. In short, there is no getting away from the role of contingency in prehistory. The traditional chains and cones of evolutionary theory, as Gould shows, are simply rendered obsolete by the diversity of anatomical designs revealed in the 530-million-year-old Burgess Shale in British Columbia. No Darwinian law of natural selection determined which of the organisms preserved in the Burgess Shale survived the great crisis which beset the earth 225 million years ago. They were just the lucky winners of a cataclysmic ‘lottery’ . Had the cataclysm taken a different form, therefore, life on earth would have evolved in quite different, and unpredictable, ways.
¹⁶⁷

Once again, it is easy to scoff at Gould’s alternative worlds inhabited by ‘grazing marine herbivores’ and ‘marine predators with grasping limbs up front and jaws like nutcrackers’ - but not by
Homo sapiens
(‘If little penis worms ruled the sea, I have no confidence that Australopithecus would ever have walked erect on the savannas of Africa’).
¹⁶⁸
But Gould’s comments on the role of contingency in history are far from absurd. In the absence of the scientific procedure of verification by repetition, the historian of evolution can only construct a narrative - replay an imaginary tape, in his phrase - and then speculate as to what would have happened had the initial conditions or some event in the sequence been different. This applies not just to the fortuitous triumph of the polychaetes over the priapolids after the Burgess period, or the triumph of mammals over giant birds in the Eocene period. It applies to that brief eighteen-thousandth of the planet’s history when it has been inhabited by man.

True, Gould’s argument depends heavily on the role of major upheavals - like those caused by the impact of extraterrestrial bodies. Yet this is not the only way in which contingency enters the historical process. For, as the proponents of ‘chaos theory’ have demonstrated, the natural world is unpredictable enough - even when there are no falling meteors - to make the task of accurate prediction well-nigh impossible.

In its modern usage by mathematicians, meteorologists and others, ‘chaos’ does
not
mean anarchy. It does
not
mean that there are no laws in the natural world. It means simply that those laws are so complex that it is virtually impossible for us to make accurate predictions, so that much of what happens around us
seems
to be random or chaotic. Thus, as Ian Stewart has said, ‘God can play dice and create a universe of complete law and order in the same breath,’ since ‘even simple equations [can] generate motion so complex, so sensitive to measurement, that it appears to be random’.
¹⁶⁹
To be precise, the theory of chaos is concerned with stochastic (that is, seemingly random) behaviour occurring in deterministic systems.

This was originally a phenomenon of interest only to disciples of the pioneering French mathematician Henri Poincaré. Poincaré had maintained that periodicity must ultimately arise if a transformation were repeatedly applied in a mathematical system; but, as Stephen Smale and others came to realise, some dynamical systems in multiple dimensions did not settle down to the four sorts of steady state identified by Poincaré for two dimensions. Using Poincaré’s topological system of mapping, it was possible to identify a number of ‘strange attractors’ (such as the Cantor set) to which such systems tended. The ‘strangeness’ of these systems lay in the extreme difficulty of predicting their behaviour. Because of their extreme sensitivity to initial conditions, it was necessary to have an impossibly accurate knowledge of their starting points to make accurate forecasts.
¹⁷⁰
In other words, apparently random behaviour turns out not to be completely random - just non-linear : ‘Even when our theory is deterministic, not all of its predictions lead to repeatable experiments. Only those that are robust under small changes of initial conditions.’ Theoretically, we could predict the outcome when we toss a coin if we knew exactly its vertical velocity and rotations per second. In practice, it’s too difficult - and the same applies a
fortiori
in more complex processes. So although the universe is notionally deterministic after all, ‘all deterministic bets are off. The best we can do is [
sic
] probabilities ... [because] we’re too stupid to see the pattern.’
¹⁷¹

The applications (and derivatives) of chaos theory are numerous. One of the first was in the classic physics problem of ‘three bodies’ - the unpredictable gravitational effects of two equally sized planets on a grain of dust - which astronomers have seen in practice in the apparently random orbit of Hyperion around Saturn. Chaos applies to turbulence in liquids and gases too: this was Mitchell Feigenbaum’s main area of interest. Benoit Mandelbrot discovered other chaotic patterns in his work
The Fractal Geometry of Nature
: a fractal, as he defined it, ‘continued to exhibit detailed structure over a large range of scales’ - just as the Feigenbaum ‘fig tree’ does. Edward Lorenz’s research on convection and the weather provides one of the most striking examples of chaos in action: he used the phrase ‘Butterfly Effect’ to characterise the climate’s sensitive dependence on initial conditions (meaning that the flapping of a single butterfly’s wing today could notionally determine whether or not a hurricane would hit southern England next week). In other words, tiny fluctuations in the state of the atmosphere could have big consequences - hence the impossibility of even roughly accurate weather forecasting (even with the biggest available computer) for more than four days to come. Chaotic patterns have also been found by Robert May and others in the fluctuations of insect and animal populations. In a sense, chaos theory finally confirms what Marcus Aurelius and Alexander Pope long ago instinctively knew: even if the world appears to be ‘the effect of Chance’, it still has a ‘regular and beautiful’ - if unintelligible - structure. ‘All Nature is but art unknown to thee; / All Chance, direction, which thou canst not see’.

Clearly, chaos theory has important implications for the social sciences. For economists, chaos theory helps to explain why predictions and forecasts based on the linear equations which are the basis of most economic models are so often wrong.
¹⁷²
The same principle ‘that simple systems do not necessarily possess simple dynamic properties’ can presumably be applied to the world of politics as well.
¹⁷³
It is, if nothing else, a warning to all pundits to avoid simple theories about the determinants of elections. The most we can do with our understanding of chaotic systems, as Roger Penrose has suggested, is to‘ simulate
typical
outcomes. The predicted weather may well not be the weather that actually occurs, but it is perfectly plausible as a weather.’
¹⁷⁴
The same applies to economic and political predictions. The best the long-range forecaster can do is give us a number of plausible scenarios, and to admit that the choice between them can only be a guess, not a prophecy.

But what are the implications of chaos for historians, who are concerned not with predicting the future, but with understanding the past? It is not enough simply to say that man, like all creatures, is subject to the chaotic behaviour of the natural world, though it is certainly true that, right up until the late nineteenth century, the weather probably was the principal determinant of most people’s well-being. In modern history, however, the acts of other people have come to play an increasingly important role in this regard. In the twentieth century, more people have had their lives shortened by other people - as opposed to nature - than ever before.

The philosophical significance of chaos theory is that it reconciles the notions of causation and contingency. It rescues us not only from the nonsensical world of the idealists like Oakeshott, where there is no such thing as a cause or an effect and the equally nonsensical world of the determinists, in which there is only a chain of preordained causation based on laws. Chaos - stochastic behaviour in deterministic systems - means unpredictable outcomes even when successive events are causally linked.

In fact, this middle position was already implicit in much that had been said by philosophers of history about causation in the 1940s and 1950s - before the advent of chaos theory. The fundamental determinist idea that causal statements could only be predicated on laws can, as we have seen, be traced back to Hume. In his
Treatise of Human Nature
, Hume had argued that a causal link between two phenomena X
₁
and Y, could only be posited if
series
of cases in which events X
₁
, X
₂
, X
₃
, X
₄
... had been followed by Y
₁
, Y
₂
, Y
₃
, Y
₄
... had been observed - a series sufficiently long to justify the inference that Xs are always (or very likely to be) followed by Ys. As refined by Hempel, this became known as the ‘covering law’ model of causation, which states that any statement of a causal nature is predicated on a law (or ‘explicit statement of the [presupposed] general regularities) ’derived from repeated observation.
¹⁷⁵

However, Karl Popper cast doubt on the possibility of establishing such laws of historical change, if by ‘law’ was meant a predictive statement analogous to the classical laws of physics. Popper’s point was simply that scientific methodology - the systematic testing of hypotheses by experimentation - could not be applied to the study of the past. Yet Popper’s rejection of determinism - what he rather confusingly called ‘historicism’ - did not imply a rejection of the notion of causation altogether, in the way that Oakeshott’s had.
¹⁷⁶
Popper accepted that events or trends really were caused by ‘initial conditions’. The critical point was that it was possible to have a causal explanation in history which did not depend on such a general statement or deductive certainty. Collingwood had already distinguished between the Hempelian (or nomological) type of causal explanation and the ‘practical science’ type of explanation, in which a cause is ‘an event or state of things by producing
or preventing
which we can produce or prevent that whose cause it is said to be’.
¹⁷⁷
Here the best criterion for establishing a causal relationship was not the Hempelian covering law, but the so-called ‘but for’ or
sine qua non
test, applying the principle that ‘the effect cannot happen or exist unless the cause happens or exists’. Popper made the same point: ‘There are countless possible conditions; and in order to be able to examine these possibilities in our search for the true conditions of a trend, we have all the time to try
to imagine conditions under which the trend in question would disappear
.’
¹⁷⁸
Indeed, Popper’s most telling charge against ‘historicists’ was their inability to ask such questions - ‘to imagine a change in the conditions of change’ (something of which idealists like Oakeshott, as we have seen, had been just as guilty).