Determinism, Free Will and Some Insights of Modern Science
Mike Arnautov, August 2007
The notion of free will still seems to be causing philosophical trouble even amongst professional philosophers, despite having been debated at great length over some centuries. Having recently run into several more bouts of highly confused discussions of this subject, I finally decided to put down my take on it in writing. I realise that most philosophical aspects of the free will question have already being exhaustively written about by people smarter than I am, and there would be little point in rehearsing the arguments yet again, just for the sake of it. My intent in revisiting this well trodden ground is to try dispelling some persistent confusion, and to bring to the reader's attention some aspects of the modern scientific understanding of the world which appear to have some bearing on the issues. A detailed examination of these matters (and of the much broader background of the relevant philosophical and scientific theories) is of necessity beyond the scope of a brief paper. It simply is not feasible to go into the detailed investigation combinatorics, Chaos, Complexity and Quantum Mechanics, let alone to summarise everything written on the subject by philosophers. The best I can hope to do is to try to keep things simple, avoid technical jargon in so far as possible, and show where and how these subjects are thought to interact with the problem of free will, and leave it to the reader to pursue such connections to a greater depth, if he is so inclined.
The only coherent interpretation of free will is to see it as our evolved ability to make complex choices in an uncertain world. Our experience of it has nothing to do with non-causality.
In a deterministic setting, free will and inevitability are two sides of the same coin. That which we freely will, inevitably happens. That which inevitably happens, we freely will. The two statements are equivalent, but since we are not some independent observers standing apart from our universe, only the former statement makes any sense from our point of view, because it describes our actual experience.
An admixture of randomness (which is indistinguishable from real non-causality) does not in any way affect the only coherent conclusion: we may or may not be fully determined, but this has no impact our freedom to make choices – on our meaningful experience of possessing a free will.
I. The incoherence of naïve conceptions of free will
In popular conceptions, free will is usually taken to mean an aspect of human behaviour which is not determined (i.e. completely and unavoidably caused) by circumstances. While it is not entirely clear what exactly are or are not to be counted as determining circumstances, such conceptions of free will agree on one point: there is a component of human actions (the "size" of this component being unclear) which is not determined by "circumstances". This component is free and spontaneous. It is un-caused. It is what makes us free agents, rather than mechanistic automatons, and gives us control over our behaviour and hence our lives.
William James pointed out the incoherence of this view in an admirably concise way:
If a "free" act be a sheer novelty, that comes not from me, the previous me, but ex nihilo, and simply tacks itself on to me, how can I, the previous I, be responsible? How can I have any permanent character that will stand still long enough for praise or blame to be awarded?1
One traditional response to this criticism is to assume that a whole complex, dynamic component of one's personality (generally identified with the soul) exists beyond the confines of the mundane material world. However, if actions of this soul do not flow causally from other factors in its non-corporeal world, then William James' questions apply to it with the same force. If on the other hand the soul's actions are subject to causation within its non-corporeal world, then this gambit merely removes some causal sources of our behaviour from the sphere of the potentially knowable factors, without in any way freeing us from the overall constraints of causality.
Another escape route is to assume that the requisite un-caused influence is an aspect of God, but this would negate the individual human aspect of free will, subsuming it into the universal will of God. On the other hand, if the divine irruption is considered to be merely opening a space for a non-causal element of decision making, this manoeuvre fails to avoid the difficulty spelled out by the above quote from James. It is, therefore not surprising that theologians and philosophers tread carefully around notions of true non-causality, largely relegating it to the concept of the Prime Cause, the unique source of all causal chains.
It is sometimes claimed that Quantum Mechanics2 offers a third way out, by introducing an irreducibly random element into the very fabric of our world.3 But is randomness on its own a sufficient non-causal factor, to satisfy the popular conceptions of free will? A simple thought experiment shows that this is not the case.
Consider game-playing programs – Chess, draughts, 3D noughts and crosses or any other formal game of sufficient complexity to require a reasonably complex evaluation of positions in searching for the optimal move (i.e. a move most likely to lead to a victory). Not even the most enthusiastic proponents of Artificial Intelligence would claim such programs to have any semblance of free will. While deterministic in nature, game playing programs do usually have an element of randomness built into them. When they comes across a plurality of different moves, all of which appear to be equally advantageous with no further way of differentiating between them, the usual tactic is to choose one alternative at random. Standard pseudo-random computing methods are quite sufficient to ensure that a program takes different choices in otherwise identical situations, even though its behaviour remains completely deterministic. Suppose, however, the choice were truly random, based on QM effects. Would the program then have free will?
Such free will would be a paltry thing, and as one would expect, the suggestion is firmly rejected by proponents of an non-causal free will4. In any case, William James' objection is relevant here as well: in what way is a truly random choice in some way my choice?
In short, the notion of free will as an non-causal component of human behaviour remains incoherent, whichever way we hold it up against the light of rational analysis,
II. The experience of free will
Despite all of the above, we do have a direct personal intuition of possessing and exercising free will, and this experience needs to be accounted for, rather than being dismissed as meaningless.
Unfortunately, dismissing personal experience when it does not square with scientific theories is a mistake scientific thinkers are somewhat prone to, when turning their attention to philosophical matters. The extreme behaviourist stance, which flourished briefly in mid-20th century (and caused damage to cognitive research for decades afterwards), is an example of this attitude.5 A much more productive approach is exemplified by physicists squaring the common perception of the solidity of everyday objects, with the realisation that even the most solid of such objects are 99.9999...% "empty space"6.
We do feel that we have the ability to make free, reasonably unconstrained choices, and our intuition tells us that when we make a choice, it is our choice, not swayed by the random outcome of some QM "coin tossing". We could have chosen otherwise of our own free will. Note, however, that this last assertion is unprovable, since it is impossible to repeat the choice under exactly identical conditions. This is a well-rehearsed argument, but its full depth is not always appreciated.
Firstly, the combinatorial complexity of a human nervous system is so unimaginably huge7, that any experimental verification cannot rely on different individuals being presented with "identical" choices. Using the same individual for repeated "identical" choices is also not feasible. In addition to the obvious observation that memory (conscious or unconscious) of earlier repeats itself becomes a causative factor ("let's do something else this time!"), there are the far more fundamental difficulties which flow from the Chaos Theory and QM. To put it quite simply, the more we know, the more we are in a position to appreciate the full strengths of the Heracleitian dictum that one cannot "step twice into the same river", and one consequence of that is to preclude any chance of an objective verification of the existence of free will.
There is an obverse side to these difficulties: what we see as apparently identical situations, are in fact extremely unlikely to be truly identical, and hence it is quite possible that they would result in us making different choices. Hence such experiments would not demonstrate the existence of free will any more than they could refute determinism. All that can be inferred from any such experiments is the uncontroversial fact that in very similar situations (rather than in an absolutely precise specific case) we do as a rule have a whole array of options available to us. The greater this repertoire of options, the better we are in responding to the world around us in a nuanced way, whether or not our behaviour has a free will aspect to it.
If we accept that free will is something we feel to have, while also accepting that this cannot be experimentally verified, it makes sense to enquire as to the circumstances which trigger this feeling. As is often the case with such matters, it is in fact easier to start by looking for the opposite – for kinds of choices in which we do not feel free will to be (necessarily) involved.
In some situations, where we do not greatly care about the outcome of a decision, or where all likely outcomes are seen as being roughly equivalent, we often summarise our decision making process as tossing a mental coin (or sometimes even tossing a real one). As we have already seen, random choices are not perceived as expressions of free will, and indeed this kind of decision making is readily accepted as having no deep philosophical implications. It could, of course, be argued that the primary decision to use a mental coin-tossing approach is itself a free will choice, but the fact remains that "it's-all-the-same-to-me" choices can be readily attributed to purely random decision making, with no injury to our intuition of having a free will.
At the other extreme, we encounter situations where the preferred choice is entirely obvious. In such cases we often say that we "have no choice", even though this is clearly not literally the case. The expression simply signifies the lack of any involvement of the free will faculty.
While seemingly opposite to each other, both of these cases can be subsumed in a single category. In both cases we know enough about the available options and their potential outcomes. We can clearly see that there a set of "best" (or perhaps just least-worst) choices (which may include all the choices available), all of which (if there is more than one) are equally good/bad from our point of view. In other words, such situations can be characterised by the fact that we feel we have enough data for the set of "best" choices to be quite clear to us. These are "no-brainer" choices, which in themselves do not require free will. It is of course still possible for a "perverse", non-optimal choice to be taken, but this would be clearly attributable to causal psychological factors (feeling "bolshy", being tired of repetition, not wishing to be predictable, etc...).
Where we do have full information, we have no difficulty in accepting that the outcome of our choice may be fully determined. By contrast, the situations in which we feel our choice to be in some way "free" are those where we find that instead of dealing with hard facts, we are balancing likelihoods, or more precisely our perceptions and intuitions of likelihoods, against the values we tentatively ascribe to various outcomes. Yet there is nothing inherently indeterministic about such decision-making. We may say that we need to "sleep on it", or "search our hearts", that we proceed on "gut feel" and on hunches etc... Even then we often attempt to rationalise retrospectively our eventual decision.
This experience of making complex, incomplete-data choices in itself offers no indication as to the decision-making process being purely deterministic, or including any kind of non-causal element. Given the well established fact that a significant part of our cogitation is unconscious (or subconscious) and hence not open to our introspection, there is in any case no way for us to differentiate between these possibilities. What we experience is weighing alternatives on the basis of the available data, and coming to a conclusion which may or may not apparently follow from the factors we have consciously examined. It is, of course, quite possible that all our decisions are in the final analysis still "no-brainers" at a sufficiently high level of discrimination, which just happens to be beyond the limits of our awareness.
Nevertheless, it would seem that the more choices we have, with potentially very different (and hard to evaluate) outcomes, the more we are inclined to invoke free will as a deciding factor in our choice, despite there being no internal psychological experience, no qualia specific to exercising one's free will, which can be distinguished from complex decision making without any non-causal elements.
Does that mean that we may be fundamentally no different from a game-playing program making its deterministic (or purely random) choices? Well, "yes and no"8. In a point-by-point comparison, it is indeed hard to see any significant difference. However, the Marxist dictum that "quantity transforms into quality" is surprisingly often correct.9 The sheer scope and complexity of human decision-making, of its influencing factors and of the dynamic interplay between the two, make the comparison with the game-playing program quite absurd. In the language of Complexity Theory, our perception of free will is an emergent property of an irreducibly complex process.10 Or of you prefer a more traditional, physics-inspired metaphor, somewhere along our evolutionary path our capacity to make decisions underwent a phase transition, making comparisons with decision making by, e.g., bacteria or present day robots, about as meaningful as comparisons between the properties of water and ice.
To summarise, phenomenologically, our experience of free will is indistinguishable from our capacity for making meaningful choices on the basis of incomplete data.
None of the above analysis is very new, but the belief in free will persists even amongst well-informed people. Since it is in principle impossible to prove its existence experimentally, and since (as argued above) free will does not generate any subjective qualia to distinguish it from a sufficiently complex deterministic decision making in the absence of sufficient data, we must conclude that the belief persists for other reasons. The most obvious candidate is the fear, sometimes explicitly expressed, that without free will there can be no personal responsibility and no moral judgement, and that all action becomes meaningless.
In other words, the final argument in favour of free will is that if it doesn't exist, it is necessary to invent it. Daniel Dennett persuasively argues that evolution has done just that on our behalf.11 In his view it is a mistake to concentrate on the absolute specifics of any given moment, because the freedom of choice, which we do have, is conferred by the very large repertoire of options available to us when faced with general situations, sharing some salient feature necessitating our choice. If in playing golf, we miss a simple putt on the green, this does not mean that we could not have made the putt if we tried again – it only means that in the absolutely specific circumstances of that moment, for whatever reason, we did miss. Nothing more can be read into that fact, since it is impossible to revisit/alter that past moment or any past moments leading up to it, regardless of the world being deterministic or not.
III. But what of the moral consequences?
Let us assume, for the sake of the argument, that we live in a completely deterministic universe. We know that the reality is rather more complex than that, but it is instructive to deal with the context of "hard" determinism by thinking through the consequences of free will conceptualised as an ability to exercise informed judgement in the absence of complete information.
Perhaps the most fundamental fear is simply that if our behaviour is completely deterministic, then all our choices are predictable in advance and hence that all choice is simply meaningless. Why agonise over a complex choice, if the outcome can be predicted anyway?12 That would be a good question were it not a non sequitur. Determinism does not equate to predictability, and the existence of fundamental reasons why this is so was one of the more surprising discoveries of the 20th century.
There are three obstacles to predictability of our behaviour in a deterministic world: complexity, chaos and emergence.
We have already dealt with some aspects of the complexity issue when considering the impossibility of an experimental verification of free will. There we were thinking of an artificially simplified situation, but for predicting our choices in ordinary life, one must also consider the entire physical and social environment in which we exist. If that weren't bad enough, there is another, novel and radical complicating factor. Human minds, as well as human societies, are very complex examples of "iterative systems" – i.e. systems whose "outputs" are also a part of their "inputs". To express it in a less technical (if less precise) language, they constitute a substantial part of their own environment. As we have learned in the last few decades, even very simple iterative deterministic processes can have some very unexpected properties.13
Chaos Theory teaches us that for many iterative processes – weather systems being a well known example – it is necessary to know the current state of a process with absolute (infinite!) precision, in order to make reliable predictions of its behaviour. Arbitrarily small discrepancies between the actual state of the system and of our description of it, can very quickly lead to arbitrarily large discrepancies between our predictions and the system's actual behaviour.
Even more fundamentally, we know from studies sometimes lumped together under the grand name of Complexity Theory that even non-chaotic iterative processes can exhibit an irreducible complexity – a feature sometime inaccurately equated with the wider concept of "emergent behaviour". For such a process there are no short-cuts for predicting their behaviour – the quickest way to discover what it is going to do, is simply to observe it doing it.14 This does not preclude the existence of some regularities amenable to prediction, but there we are inevitably moving away from behaviour of individuals, to the statistics of behaviour of large numbers of individuals.15
While the applicability of all of this to human behaviour is hypothetical, there are reasons to suppose that both kinds of this radical unpredictability may well apply.16 It is very likely that as a matter of principle, human behaviour cannot be reliably predicted even under the conditions of "hard" determinism. Such a conclusion is in tune with our intuition, our experience and our hopes.
What about other fears? Does deterministic free will invite fatalism? A frequently expressed concern is that "if it is all preordained, there is no point of doing anything". But that simply misses the point – preordained or not, what we do is what actually happens, and it will not happen without us doing it. An individual may certainly conclude that there is no point of doing anything at all17, but that is still a a decision he takes, and inaction is also an action.
There is a fundamental misconception lurking here, and the same misconception underlies all other morally-based objections to determinism. They implicitly assume one's self to be some helpless, tragic puppet, a philosophical homunculus, swept along by the implacable machinery of the deterministic universe, who observes what is happening to it without being able to influence anything. But there is no such homunculus. We, and our decision making are an integral part of the "machinery" of the world, not some alienated outsiders.
Does determinism absolve us from moral responsibility? It does not, but to see this, we must look deeper into the notion of "being responsible" for something. Once again, the best approach is to consider situations in which we feel that "being responsible" does not apply.
Young children are not held responsible for their actions because they cannot be assumed to have a clear grasp of the distinguish between what is right and what is wrong. We do of course treat them as if they were held responsible, but that is merely an educational stance, which would not apply in a court of law. Having, or at least being capable of having a moral code is therefore one of the prerequisites for being held responsible for one's actions.
We are also absolved from responsibility for consequences of our actions where there are reasons to think that those consequences could not have been reasonably foreseen. As an extreme example, consider turning on a light switch which unbeknown to us has been wired up to cause a devastating explosion. Clearly no blame attaches to the person who unknowingly triggers the explosion.
There are further considerations, which have less of an on/off character and are felt to reduce responsibility to a greater or lesser extent, without necessarily eliminating it. Even if negative consequences of our actions are foreseeable, our responsibility for them is moderated if it can be seen that those consequences were not intended by us. If something dreadful happens because we absent-mindedly forgot to do something, that is very different from us actually intending it to happen. The responsibility is felt to be even further diminished if the omission is not due to mere absent-mindedness, but was caused by a significant external distraction.
The degree of awareness or lack of it is another factor. A sleep-walker causing an accident is not generally considered responsible, assuming it is possible to establish that the sleep-walker was in fact sleep-walking. Other conditions which reduce or eliminate our awareness of our actions also have the effect of reducing or removing blame (or praise!), depending on the degree to which we are responsible for bringing about such a condition (e.g. by excessive drinking).
To summarise this very brief analysis: responsibility requires (1) a potential for having a moral code, (2) foresight, (3) intention and (4) awareness. It would be clearly possible to have an extended discussion concerning the nature and scope of these requirements, but for the purpose of this paper it is sufficient to observe that none of them are incompatible with a deterministic universe. Specifically, in so far as we judge humans fulfilling the above four requirements, we can and should hold them responsible for their actions, regardless of our assumptions concerning the (in)deterministic nature of the world.
The world being completely deterministic or otherwise makes no difference to our internal experience of making choices. They remain our choices. We choose the way we choose because of who we are, for reasons which are ours, based on our unique experience, striving to achieve our individual goals. How then can we not be responsible for our actions? Whether or not we would always have chosen in exactly the same manner in exactly the same circumstances, does not in any way affect our ability to make choices, nor the importance to our life of making good choices.18
Does determinism make moral judgement irrelevant? Not at all. Our interactions with others are causative influences on their future behaviour. It would be illogical and immoral to let a criminal escape punishment (in its widest sense of an attempt at behaviour correction) on the grounds of the crime having been "inevitable". The crime was (to use Dennett's expression) perfectly "evitable" if only the criminal and/or some of the circumstances of his life were somewhat different – and it behoves a civilised society to realise this "evitability" in practice by taking steps to correct or at least to prevent such behaviour in the future.
The common perception that a determinist view somehow absolves a criminal from all blame for his crime, again hinges on the unspoken assumption that he is just a helpless puppet, trapped in the relentless machinery of the workings of his mind and life. But there is no such puppet to be excused – such a dualist body/soul view has no place in a deterministic universe. The criminal is not in any way separate from the workings of his mind and cannot be absolved from blame.
There is a knack to thinking about freedom of choice in a deterministic setting. It is not unlike the experience of looking at one of those ambiguous pictures, which can be viewed in two different ways – e.g. as a black silhouette of a vase on a white background or as silhouettes of two human profiles facing each other against a black background. Our mind flips between the two interpretations, and it requires a considerable effort and practice to see (to really see) the gestalt of the picture itself.
Similarly, we tend to flip between our subjective viewpoint of being "free" choice-makers, and the objective view of a deterministic universe, and this causes us to contaminate the deterministic view with a lurking image of the inner "free" homunculus somewhere inside – a helpless passenger of the runaway juggernaut of the world. It takes a conscious effort to bear in mind that the two viewpoints are mutually complimentary rather than contradictory, without letting some kind of "soul" spill into the picture.
What of deterministic universes with some randomness thrown in? We have already dealt with this question when considering the example of a game-playing program. Randomness does not offer anything that would alter any of our above conclusions. Even in the absence of any weightier reasons, this is obvious from the fact that in ordinary life we are unable to distinguish true randomness from pseudo-random processes such as coin throwing.
While it is true that QM springs a surprise in the shape of its Many Worlds interpretation, from the point of view of a singular world-line, even this oddity result in nothing more than pure randomness. Philosophical libertarians invoke QM in vain – it offers no help.
IV. Changing the future
A fairly common misconception concerning determinism and its interaction with free will is exemplified by the question "If the future is determined, wouldn't it happen whatever we do?" Clearly, this is based on the wrong assumption that our actions have no causative effect on the future. The future will be as it will be because of whatever we do, not despite it.19 When hammering a nail, the immediate future I would experience after hitting my thumb would be very different from the one caused by my hitting the nail as intended – and neither would happen if I don't pick up the hammer in the first place.
A related question would seem to be a more sensible one: "If determinism is true, wouldn't it mean that we cannot really change the future?". Intuitively, the meaning of this query appears fairly obvious, but perhaps surprisingly, the more one thinks about it the less obvious it becomes.
What exactly could we mean by "changing the future"? Changing it from what? The future has not happened yet, so what are we supposed to be changing it from? From what it is determined to be? But we do not know what it is determined to be, and as argued above, there are excellent reasons to think that we cannot know it as a matter of principle. The only way we have of knowing for sure what is really going to happen, is to let it happen, by which time it has happened and is no longer the future. On the other hand, if we could somehow observe the future before it slips into the past, the resulting knowledge would merely become yet another causative factor influencing our behaviour, leading to the merry fun of unstable time-loops and time-travel paradoxes, beloved of science-fiction writers.20
Yes, some eminent physicists postulate the existence of a "block universe", in which our past, present and future all objectively exist on an equal footing in the four-dimensional block of space-time, and the passage of time is merely an illusion of an unknown origin and significance.21 However, the only way we could observe such pre-existing future would be by stepping outside the putative space-time block in order to examine its arbitrary sections. This route to omniscience is, of course, precluded by the hypothesis itself – the block universe necessarily consists of all space and time, making it in principle impossible to step outside it.
In short, while one may postulate an omniscient God point of view to which such a concept might be meaningful, from our space-time bound perspective, the notion of "changing" the future is simply a linguistic construct with no ascertainable meaning. The only sensible answer to the question whether we can "really" change the future, is the one already given: we affect the future by all our actions and inactions, and that is all that can be meaningfully stated.
Contrary to a common misconception, even "hard" determinism does not preclude the subjective experience of having a free will, without such an experience being reduced to a meaningless illusion. The key to this insight lies in the realisation that we are active participants in the workings of the world, rather than some detached, passive observers. In a deterministic setting, that which we freely will, inevitably happens, and equivalently, we freely will that which inevitably happens. However, the latter statement only makes sense from an omniscient viewpoint completely outside of the space-time of our universe. Embedded as we necessarily are within the workings of the universe, the experience of actively contributing to the temporal unfolding of events is the only one sensibly applicable from our point of view.
- o O o -
If you feel like leaving a comment, please feel free to do so!
1 William James, "Pragmatism", Lecture III. "Some Metaphysical Problems Pragmatically Considered".
2 Quantum Mechanics is bound to raise its ugly/beautiful head in any serious examination of these matters sooner rather than later. From here on we'll use the traditional abbreviation QM instead of spelling it out in full.
3 As mathematicians know only too well, randomness is a surprisingly slippery concept. Most processes which are generally considered to be random (e.g. tossing coins, throwing dice etc...) are anything but. Such processes are pseudo-random, their apparent randomness being due to the vast complexity of factors affecting their outcome – the complexity of which is well beyond our ability to take fully into the account. Only QM gives us what currently appears to be a true source of randomness, and practical devices can be (and have been) constructed to exploit QM phenomena in order to generate truly random effects.
4 The eminent mathematician Roger Penrose speculates that our intelligence does have its roots in QM phenomena, but the primary purpose of his speculations is to tackle the issue of alleged non-computability of mind, rather than those of free will. (See his "The Emperor's New Mind" and "Shadows of the Mind".)
5 Behaviourism simply denied any validity to internal experiences, declaring them to be meaningless epiphenomena of the nervous system, unworthy of any scientific attention. The very idea of examining the phenomenon or the contents of conciousness, let alone searching for its origins, would be palpably absurd to a serious behaviourist.
6 Amusingly, albeit alarmingly, it was reported a few years ago that some military "thinkers" at the Pentagon were interpreting this scientific fact as a justification for attempts to walk through solid walls. In reality, the apparent contradiction is resolved by taking into the account interactions of force-fields binding together the subatomic particles, atoms and molecules which which together generate the observed solidity of objects.
7 The number of possible states of the nervous system of a human being probably exceeds the number of atoms in the observable universe. This counter-intuitive state of affairs is caused by what mathematicians and computer scientists call the combinatorial explosion – the propensity for the number of possible relationships between elements of a system to grow explosively with the number of elements in the system.
8 Gordon Lack in conversations, passim.
9 Adding hydrogen to a (large) volume of space is a good illustration. At first this results in just a cloud of gas. Adding more, eventually leads to a qualitative change – a gas planet is formed. As more hydrogen is added, the planet grows until it ignites into a star. But keep on pumping, (which will be now be very hard to do, but this is a thought experiment!) and the star will collapse into a black hole (unless it explodes as a nova first). No magic ingredient has been added anywhere along the way. We just kept pumping hydrogen.
10 We shall return to this theme later.
11 See Daniel Dennett's "Freedom Evolves". Dennett's conception of free will is evolutionary. It embraces both an individual and his social milieu and thus escapes the existentialist trap of assuming the individual to be his sole "prime mover", thereby also escaping existentialist alienation and despair.
12 Some highly unpleasant consequences of such a state of affairs have been explored in a number of books of speculative fiction – e.g. in Philip K Dick's "Minority Report".
13 The great French mathematician Poincare had a glimpse of these things at the end of the 19th century, but generally speaking, science was not yet ready for exploring the strange landscape he discovered.
14 This is what Stewart and Cohen term "the ant country" in their "Figments of Reality", the reference being to Langodon's Ant – a very simple deterministic system exhibiting this kind of behaviour.
15 In 1898, Ladislaus Bortkiewicz, an army surgeon, observed that the distribution of rare events of a particular kind, such as the number of deaths in the Prussian army resulting from a horse or a mule kick, obeyed a simple mathematical formula. This, one might say, kick-started the study of statistics – a branch of mathematics, which can say nothing about an individual, while being capable of reliable predictions when dealing with large populations of individuals. Statistical regularities have no bearing on individual free will.
16 On the simplest level, in the "arms race" between predators and prey, unpredictability offers a clear advantage. It is therefore natural to suppose that evolution, not having direct access to quantum randomness, would utilise the much simpler devices of chaos and irreducible complexity, particularly since these traits come naturally to iterative systems, of which a brain is certainly an example. More generally, Stuart Kaufman argues in his "Investigations" that life inevitably evolves towards "the edge of chaos".
17 And die of hunger and thirst soon afterwards, thereby demonstrating that it was not a well-informed choice. Or he may decide to get on with his life instead. Nobody, including himself, can decide beforehand which of these options he will choose, without going through the process of choosing. In any case, such extreme fatalism is not a survival trait, which probably accounts for the lack of really extreme fatalists in the world. Even Turgenev's fictional titular "hero" Oblomov, proverbial for his passivity, ate and drank well.
18 Which is where evolution comes in, according to Daniel Dennett, in refining and developing our capacity to make meaningful choices in an uncertain world, which is the only coherent conception of free will.
19 Surprisingly enough, even professional philosophers can stumble on this one. For instance, here's John Searle's attempt to demonstrate the supposed absurdity of determinism from the 1st person point of view: "You cannot say to the waiter, 'Look, I am a determinist. I will just wait and see what I order because I know that my order is determined.'" ("Freedom and Neurobiology", p.11) To be fair to Searle, he is not claiming that this proves the existence of free will, but the argument does not work even in the weaker context of the 1st person perspective.
20 Isaac Asimov's novel "The End of Eternity" is still one of the best explorations of the theme of time-travel paradoxes.
21 See for example David Deutsch's "The Fabric of Reality".