Advantageous superstitions

1. Henry Fuseli

The Swiss artist Henry Fuseli (Johann Heinrich Füssli, 1741-1825) made this remark:

"We are more impressed by Gothic than by Greek mythology, because the bands are not yet rent which tie us to its magic: he has a powerful hold of us, who holds us by our superstition or by a theory of honour" (Knowles, The Life and Writings of Henry Fuseli, volume 3, aphorism 105, page 102).

The word is "superstition" in the singular. This points to our disposition to believe or fear, rather than any particular beliefs or fears. There is something dubious within us, a mode of thought that it is easy to dismiss as irrational. But it is individual superstitions that may directly steer our thoughts and actions by virtue of the hold they have on us. The disposition only matters to the extent that there are individual superstitions, the hold of which it facilitates. But then it matters a great deal, as the reason why they can have a hold.

We may think that any rational person's mentality should exclude the disposition, and that we should automatically dismiss the idea that there was anything to black cats, horseshoes, or astrology.

Here we shall ask whether this is really so. We shall propose that while the only sensible response to such traditional superstitions is indeed to dismiss them out of hand, there is more to be said about beliefs that we all accept but that come to seem optional when we look closely. They are not traditional superstitions, but they share the feature that we lack adequate grounds to think that the world really is as they indicate. Moreover, it can be advantageous for them to have a hold on us, and therefore advantageous that we have a disposition that allows such a hold.

2. Traditional superstitions

There are two perfectly good reasons to dismiss traditional superstitions out of hand.

The first reason is that, given our scientific understanding of the world, there is no plausible mechanism by which they could operate. How could a black cat affect one's fortune? How could the presence or absence of a horseshoe on the wall, or which way up it was (a point on which superstitions differ), make any difference to the course of one's life? How could the position of the Sun relative to patterns of distant stars which only appear in those patterns when viewed from the vicinity of Earth make any difference to a baby's future character? And so on. A claim that there is a mechanism we have not yet understood is a reasonable speculation when connections within the same general field are proposed, for example a claim (made before tectonic plates were understood) that there could be some geological mechanism to connect the present positions of continents with earlier positions in which they would have had a jigsaw-like fit. It is a wholly unreasonable speculation when a connection between unrelated phenomena is proposed, a claim such as one that connects a passing black cat or a pattern in the stars with the immediate or long-term course of a life.

The second reason to dismiss traditional superstitions out of hand is that they have not been tested properly, and we have a strong suspicion that if they were tested, they would fail. In order to test them, we would first define the relevant variables carefully. For example, we would define what counted as good luck or bad luck. Then we would take a large sample of people and note whether a black cat did or did not cross each person's path on a particular day. We would measure whether each person had good, bad, or indifferent luck in the week following that day. Once we took into account the negatives (no black cat) and all types of luck, we might very well find that the supposed connection between black cats and luck vanished. We cannot be sure that this would be the result, but given the first reason, the lack of plausible mechanism, it would not be surprising. And given that first reason, the onus must be on proponents of the superstition to get such a test performed if they want the rest of us to take any notice of their claim.

This does not mean that we should entirely ignore traditional superstitions. We should be aware of their psychological hold on people, even while regretting that hold. We also need to understand them and their hold in order to understand much of our art and literature. This includes modern work that draws on the hold that long-standing superstitions have, for example the writings of Angela Carter and the paintings of Leonora Carrington. But that is all.

3. Advantageous superstitions

In this section we shall give some examples of superstitions which are so useful that it is advantageous that they should have a hold on us, at least in everyday life. Their advantages will be set out in later sections.

3.1 Scientific superstitions

3.1.1 The idea of a scientific superstition

The idea of a scientific superstition may seem bizarre. Over recent millennia, science has gradually displaced superstition. And the last survivors are exposed to the challenge noted above that from a scientific point of view, there is no plausible mechanism by which portents and actions could make the results that superstitions predict any more likely than they would have been in the absence of those portents or actions.

We can however find candidates if we broaden our notion of superstition to cover beliefs that underpin our ways of grasping the world, so that they have a powerful hold on us, but that are not forced upon us by the evidence as it would appear if not viewed through the lens of those beliefs. This is not to say that contrary beliefs would be permitted by the evidence, but that the beliefs we do have would, when the evidence was viewed without their influence, seem to be optional.

It is important that the superstitions should on current evidence be seen as optional, rather than their being seen as ruled out. Traditional superstitions are ruled out by evidence, or at least probably would be if we bothered to formulate and administer rigorous tests. For example, we would very likely find that people with horseshoes fixed to their walls had lives no better than people without them, or that the disposition of constellations at the times of people's birth could not be correlated with either their characteristics or the courses of their lives. The scientific superstitions we have in mind are not so vulnerable. Scientific progress might eventually lead us to regard them as false rather than optional, but we are not there yet.

3.1.2 Causation

Our first example is causation. We observe causes and effects, and rely on causal predictions, every day. Yet as David Hume pointed out, this may be just a habit of the mind. We repeatedly see event C followed by event E, and take it that C is the cause of E. But we do not see the causal power that necessitates E once C has occurred (Enquiry Concerning Human Understanding, section 7). Bertrand Russell, whose objections were wider-ranging and on the whole more technical than Hume's, wrote that "The law of causality, I believe, like much that passes muster among philosophers, is a relic of a bygone age, surviving, like the monarchy, only because it is erroneously supposed to do no harm" ("On the Notion of Cause", page 180). Belief in the reality of causation appears to be optional, and indeed its presence in fundamental physics is at best a shadow of its presence in everyday life. It seems as though we could get by with regularities along with confidence that they were underwritten by something or other. (They would have to be underwritten by something for the Universe to be sufficiently well-behaved for us to have evolved.) And neither Hume nor Russell objected to our noting regularities and acting on the basis that they would probably recur. But we take the option of identifying causation as that which underwrites the regularities. For this reason, causation is a superstition in our extended sense.

We do here speak of causation in the sense in which most people would understand it, including some sort of necessitation that distinguishes it from accidental regularity. Philosophers have developed other theories of causation, such as conterfactualism, dispositionalism, and process theories, and some have turned the emphasis from what causation really is to what we mean when we make causal claims, while others have been ambitiously anti-realist about causation. But such theories do not make inroads into the consciousness of most of us when we recognise individual instances of causation. We naturally think in terms of causes pushing parts of the world so that the usual effects occur. Moreover, the equations that the natural sciences offer us, equations that set out how systems evolve over time, do not in themselves satisfy our natural sense of parts of the world pushing other parts, even though one might think of the direction of time as the direction of causality.

It is belief in causation that is optional, not use of the concept. In most of the sciences and in everyday life, we have to identify events as causes of other events. It is only when we are challenged that we bring our assumption of causation to consciousness and turn it into a belief. It is then that we are driven to reflect on whether our talk of causes is legitimate. Indeed it is at that point that we may reflect on and choose between the various available theories of causation. Our regular avoidance of such debates is facilitated by the fact that we tend to focus on specific events, rather than on causation as a general notion. We say "this caused that", rather than "here is an instance of causation".

3.1.3 Free will

Our second example is free will. For most of us, it is essential to our images of ourselves and our understanding of people and social relations that we assume free will in an uncomplicated sense, without getting into the formulations that compatibilists offer in response to the challenge that the physical world offers only determinism and randomness and that the mental supervenes on the physical. We have to take it that our inner sense of freedom at the moment of decision corresponds to the reality of the world. There are some people who study the neurology and the philosophy and then say "I am a determinist, free will is an illusion, and compatibilist guidance control does not even come close to amounting to the free will that I deny". But this attitude is unlikely to be easy to sustain in daily life. Most of us are much happier retaining our uncomplicated assumption of free will.

We shall therefore regard belief in free will, the belief that arises when we bring our assumption to consciousness, as a superstition in our extended sense. It is a popular belief. It also has a powerful hold, in the sense that if a philosopher tells us we are mistaken we will cling to it as long and as firmly as we can. But given the controversy, free will cannot be regarded as established fact. And the belief lies somewhere between belief in causation and traditional superstitions in its vulnerability to turning out not even to be optional but having to be discarded. Our knowledge of neurons, and the natural assumption that the mental supervenes on the physical, for lack of anything else on which it might supervene or any plausible mechanism that would allow it not to supervene on anything, together place belief in free will at specific and substantial risk.

As with causation, it is only belief in free will that is optional. Its assumption in daily life, an assumption that does not generally involve explicit assent to its existence, is all but inevitable.

3.2 Ethical superstition

We tend to regard ethical rules as at least provisionally binding, and not as mere suggestions that may be ignored whenever that would happen to suit our preferences. We could say that the rules that commanded general support were the ones that were binding on us individually, with the preferences of the majority being the ground of their binding nature. But that might not do. It would not allow for a whole society to be collectively ethically misguided. We feel most comfortable with a largely unspoken assumption that ethical rules are grounded in some facts about the world that hold independently of what we currently happen to think. That is, we incline to some form of ethical objectivism. (This is not the same as objectivism in the sense associated with Ayn Rand, although we might see Rand's objectivism as an example of the more general objectivism that we have in mind.)

It is when we are challenged on our assumption of ethical objectivism that the assumption turns into a conscious belief. Then we find that there is an unresolved debate. There are plenty of philosophers who regard the belief as correct, although it takes different forms in the views of different thinkers. But there are also plenty of philosophers who think the belief is incorrect, and who offer a wide range of objections.

The existence of serious challenges to ethical objectivism means that we cannot regard it as established fact. There might not be any ethical rules that were binding for reasons other than our individual or shared preferences. And yet the idea that there are such rules has a powerful hold on us. We do not stop to think that there might not be any objective rules when we routinely praise some conduct and condemn other conduct. So the belief that there are such rules is a superstition in our extended sense. Quasi-realists might say that it was a superstition that was indeed mistaken, rather than one that might be mistaken, but they would in their own way endorse recognition of the strong hold that the superstition had on us.

As to vulnerability, this superstition is too far in its nature from the natural sciences to be exposed to scientific testing of the sort that might force its abandonment. It cannot be located on the scale of vulnerability that runs from belief in causation, through belief in free will, to traditional superstitions. Nonetheless, we cannot be confident that those who reject ethical objectivism are mistaken.

It is also not easy to say that while belief in ethical objectivism is optional, its assumption in the daily life in which it is not made explicit is all but inevitable. There does seem to be real scope for a day-to-day attitude that ethical rules are to be followed without thinking of them as objective. We could be conscious prescriptivists without disturbing our pattern of life.

4. Limits to optionality

4.1 Scientific superstitions

The world appears to us as if causation and free will are real, and we act accordingly. We have to dig some way into our scientific understanding to raise doubts. And even then, we are not forced to deny their reality. The arguments against their reality may be strong, but they are not logically conclusive, and we could always try the escape route of saying that our scientific understanding might be a misunderstanding.

The upshot is that we cannot regard causation and free will as definitely real. Belief in them is optional in that while the world appears as it would if they were real, that appearance might arise without their being real.

This is however a limited form of optionality. We are not driven to belief in the reality of causation and free will. We are however driven to acceptance of their reality in everyday life, and for causation in a great deal of science, and for free will in the routine psychological understanding of people. There is no suggestion that it would be practical to get back to the underlying detail of reality and work with that in order to eliminate use of the concept of causation, or use of broad-brush psychological concepts like the concept of voluntary decision.

4.2 Ethical superstition

Most of us live as if ethical rules had more objectivity than our individual or collective preferences, taking that objectivist attitude both to guide our own conduct (or at least to feel bad about conduct that breaks the rules) and to comment on other people's conduct. If we dive into philosophical arguments, we become less sure of this. Yet we can see that even if we were to reject objectivism we could go on conducting ourselves in the same ways, so long as we did not at moments of decision think that the rules only represented preferences and were therefore trivial to ignore. Thus the belief in objectivity is optional, but ridding ourselves of the unconscious assumption of objectivity would be mentally inconvenient. It would also be risky unless an unconscious dedication to prescriptivism were substituted.

5. The respectability of belief

5.1 Scientific superstitions

Belief in the reality of causation and free will comes across as perfectly respectable. These beliefs are very widely shared, either explicitly when people consider them or implicitly when they are simply assumed, and their assumption has turned out to be very useful in navigating the natural and the human world. There is no reason to worry that systematic testing at the level at which they are put to work, as distinct from a deep exploration of physics or neurology, would instantly demolish the beliefs. In relation to free will, the Libet experiments and their successors come closest to such demolition, but they relate to very circumscribed choices like pushing buttons, and their interpretation is contested.

This is not to say that these beliefs should go unchallenged. Pressure to examine them closely can be brought to bear by philosophers as well as scientists. There are differing views of the nature of causation, and differing views as to what is actually claimed if one asserts its reality. Turning to free will, debates over the extent to which people are responsible for their actions lead us to ask what kind of free will can really be attributed to human beings. But most of the time, we get by perfectly well with straightforward and unsophisticated assumption of the reality of causation and free will.

5.2 Ethical superstition

Similar things could be said about belief in the objectivity of ethical rules. The implicit assumption of that belief is very widely shared, and it plays an important role in everyday life. The belief is also not vulnerable unless we conduct serious philosophical argument. Even then, it is only placed in doubt, not refuted. And most of the time it is taken for granted, not expressed in any way that would put it up as a target for debate. The belief can and sometimes should be debated, but we can get by perfectly well without having such debates all the time.

6. The virtue of belief

6.1 Scientific superstitions

It is not merely convenient to believe in or assume the reality of causation and free will. Doing so can be a condition of certain enquiries into the nature of the world, enquiries that we can expect to be fruitful. Someone who believes in or assumes the reality of causation will ask of some phenomenon what causes it, and may find an answer that he or she would not have found otherwise. Someone who thinks that people make their own choices will be motivated to investigate the psychology of choice, and will look for humanistic rather than mechanistic answers. This will not only make the answers more comprehensible. It will also greatly improve the prospects for getting answers that are of real interest and use. If we insisted on thinking in the mechanistic terms of neurons, the prospects for getting fully explanatory answers in relation to particular decisions would be remote.

What we have just said may make it seem that the scientific superstitions that concern us are no more than heuristic tools, rendering their superstitious nature harmless because it will not infect substantive claims that are consciously made. But this is not quite so. If one starts by saying "Let us pretend that there is causation", or "Let us pretend that there is free will", one will not be able to say that one has established the cause of the phenomenon of interest, or the psychology of free human choice. If research is built on a pretence, the fact that it is a pretence will undermine the status of results that also reflect the pretence. If a researcher when presenting results speaks of causes or free choices, any sense that the words used reflect pretences will cut the ground from under the results. This problem is averted when results are presented in terms that do not reflect the pretence inherent in some heuristic tool, but with causes and free choices, the problem is likely to arise. So the researcher must believe rather than pretend, or at least put awareness of pretence out of mind.

If the assumption of scientific superstitions is a high road to results, a willingness to assume them may be regarded as an epistemic virtue of a responsibilist kind. It may sit alongside virtues such as creativity and inquisitiveness when they are taken as virtues that drive enquiry forward, virtues in the context of discovery, rather than virtues that lead one to subject claims to rigorous evaluation, virtues in the context of justification. So we can celebrate our disposition to rely on superstitions that are not forced upon us by the evidence, but that are immensely useful in advancing our understanding of the world.

We can however only celebrate this disposition so long as it is kept under control. If some scientific superstition turned out to be ruled out by a combination of our evidence and theories that were well-supported, it would have to be discarded. Any results that were formulated in terms that assumed its correctness would then have to be reformulated. Less drastically, if a scientific superstition turned out to be less useful than one might have hoped, perhaps encouraging research that took us down dead ends, it would have to be called into question given that it was not in itself forced upon us by evidence.

We must also emphasise that the virtue in question here is virtue in the pursuit of knowledge, not virtue in the giving of psychological comfort. The cheering effect of wishful thinking, or a belief that a lucky charm or a treatment that happens to be a placebo will aid recovery from a medical condition, might be thought to make such states of mind virtuous, or make the encouragement of them in other people virtuous. But that is not what we have in mind here, quite apart from the fact that it is prima facie unlikely that virtue would really be found in such states of mind rather than in a robust respect for truth.

6.2 Ethical superstition

The virtue of belief in or the unconscious assumption of ethical objectivism is clear. We are encouraged to take our obligations seriously, so that we get along together. Constant doubt about the status of ethical rules would make us less reliable neighbours and colleagues than we actually are.

7. The grounds of usefulness

7.1 Scientific superstitions

7.1.1 Broad-brush work

Scientific superstitions most obviously gain their usefulness from the fact that it would be counterproductive to get down to the microscopic detail of the world, even assuming (which may not be the case) that there would be a microscopic level that would capture all the basic facts upon which everything else supervened and that if we had arbitrarily great computational power it would be possible to investigate all the phenomena of interest in the world by working at that level.

That is, superstitions are a way to work at a broad-brush level. They are not the only way to do so. One could work with concepts that bore a close computational relationship to what would be seen at the microscopic level. This is for example what is done in statistical mechanics. Scientific superstitions, by contrast, are set out and put to work using new concepts that are not inherent in the microscopic detail. Their applicability is at closest emergent from what would be seen at the microscopic level, and perhaps not even that.

(We note, but shall not here pursue, the thought that if it were not even possible to identify emergence, it would be arguable that the relevant scientific superstitions were not optional, but were straightforwardly correct theories of the relevant aspects of the world. The starting point for the argument would be that they were not eliminable, even hypothetically on an assumption of arbitrarily great ability to collect data and arbitrarily great computational power.)

7.1.2 Structuring reality

The concepts used to formulate superstitions allow us to structure the world at a high level of abstraction. This is an aspect of working at a broad-brush level, but the aspect of structure deserves particular attention.

The classic notion of the use of concepts to structure reality is that of Immanuel Kant, for whom the world would be unintelligible if we did not use the obvious tools of spatial and temporal ordering, the identification of causal relationships, and so on.

Kant's approach was at the heart of a complex metaphysic that put the world as it was in itself wholly and permanently outside our grasp. Even if he had seen the depths to which modern physics penetrated reality, he would have maintained that this was so. Theories that identified fundamental particles and fields, and that made computations of the structure of spacetime, would still only show us the world as it appeared to us, not the world as it was in itself.

We do not need to follow Kant in order to understand the power of scientific superstition, but our attitude to scientific superstitions does have something of the flavour of his approach.

We can see scientific superstitions as giving more structure to our image of the world than it would otherwise have. Events are related by causation. Circumstances and human actions are put in clear patterns in relation to one another by use of the concept of free choice. And when we have found structuring devices that work, such as these superstitions, our belief in their correctness is encouraged.

They work not merely by making sense of what we currently observe, but also by facilitating progress. Once we have a stable picture of the world, we can re-apply the superstitions to formulate new hypotheses and then test them. We look for what caused what, or for how a subject thought through the options before making some free choice.

7.2 Ethical superstition

A belief in or assumption of the objectivity of ethical rules has the great benefit that it will support compliance without too much consideration of alternative rules, thereby sustaining a harmonious and pleasant society.

This psychological and sociological result is enough to make the superstition useful. The ultimate ground of usefulness is that people are responsive to what they perceive as binding rules, and are happy to have their interactions with others given some structure and predictability.

On the other hand, it is a short step from belief in or assumption of the objectivity of ethical rules to refusal to countenance changes to them. We sometimes need to ease up on a devotion to ethical objectivism in order to make progress. Historically it has been those who identify a supposed conscious source of objectivity, by asserting the existence of a divine lawgiver, who have been the chief obstacles to ethical progress.

8. Maintaining belief

There is no mystery about how we come to believe in or assume causation, or free will, or the objectivity of ethical rules. These beliefs or assumptions are held by people generally, and we just absorb them in the course of growing up.

What is mysterious is that the beliefs can survive recognition that while they might be correct, they might equally well be incorrect, and indeed that there are strong although not wholly persuasive arguments against them. Causation is vestigial in serious physics. Free will is at risk from the presumably mechanical operation of neurons together with the highly plausible view that the mental supervenes on the physical. And when we start to ask how ethical rules might be objective, and what strange kinds of facts might support ethical rules, we see that a claim to their objectivity would need far more substantial defences than a mere habit of belief or assumption.

Despite these risks, it may be perfectly acceptable to resolve to think and act as though the superstitions were correct. They have after all not been ruled out, and if inference to the best explanation can include in its evaluation of explanations their success in everyday life, they have a lot going for them. But as already noted, anything that smacked of pretence to correctness, of an attitude that it was merely as if they were correct, would risk undermining conclusions that were worded in the same terms as the superstitions.

So on what basis could superstitions that might be correct but the correctness of which was easy to doubt be upheld in a way that was strong enough not to undermine conclusions expressed in their terms? Let us consider two options.

The first option would be some form of doxastic voluntarism. We might simply choose to accept the superstitions in a sufficiently strong form of acceptance, because they were so useful. Voluntarism is however contentious. Belief is supposed to aim at truth, so it should be driven by the state of the world and not by our desires, not even noble desires to advance our knowledge and to live in harmony.

The second option would simply to resign ourselves to the psychological inevitability of the superstitions. Their hold on us is powerful not just because we have been brought up with them, but because they play large roles in our ways of life. As David Hume pointed out, philosophical doubts are only to be sustained in one's study, and are quickly abandoned once one enters into everyday society (A Treatise of Human Nature, book 1, part 4, section 7, page 269).

In relation to the ethical superstition of objectivism, the option of doxastic voluntarism would be the one we would be likely to select when placed on the defensive, and might even seem respectable. In the context of ethics, there is a degree of willingness to see an act of voluntary commitment as intellectually respectable. There is however a delicate balance to be maintained. Objectivism requires facts, and while it is certainly possible (and indeed admirable) to make a voluntary commitment to accept the facts whatever they may turn out to be, it is dubious to make a voluntary commitment to specific supposed facts in advance of receiving sufficient evidence. It is precisely the objectivist nature of the ethical superstition that should make us doubt whether voluntary commitment would be respectable.

Turning to scientific superstitions, the option of resignation to psychological inevitability would be the one we would be more likely to select when placed on the defensive. The very idea of doxastic voluntarism is manifestly antithetical to the project of finding out about the world.

Having thus resigned ourselves, we would also resign ourselves to not doubting the respectability of conclusions that we reached and that were worded in the same terms as were used in scientific superstitions, such as causal terms or the terms of free choice. Doubts that should perhaps have been transmitted from the superstitions to our expression of the conclusions would simply find no foothold in our consciousness.

This would however not suffice to make the conclusions as expressed in ways that took scientific superstitions for granted undoubtedly respectable. They would remain under the same shadow of doubt as the superstitions. But if we are to make progress in understanding the world with our limited capacity to handle microscopic detail, we must put up with that because we have to use concepts that are not forced upon us by the microscopic detail of the world and do not capture that detail. As noted above, this is not merely a matter of averaging out detail, but of conceiving the world in different ways. And we should not assume that if our brains (or our artificial intelligence systems) had arbitrarily great computational power, this need to move away from microscopic detail would end. That need might reflect limitations in principle, rather than merely in practice, of microscopic mathematical modelling or of its power to capture the nature of the world.

So the hold of advantageous scientific superstitions on us may very well be inevitable both now and in the future, either for ever or until some fundamentally new way of grasping the nature of the world is introduced. We are undoubtedly better graspers of the nature of the world for having abandoned traditional superstitions. But we are not perfect.

References

Hume, David. A Treatise of Human Nature, second edition, edited by L. A. Selby-Bigge and revised by P. H. Nidditch. Oxford, Clarendon Press, 1978.

Hume, David. Enquiries Concerning Human Understanding and Concerning the Principles of Morals, third edition, edited by L. A. Selby-Bigge and revised by P. H. Nidditch. Oxford, Clarendon Press, 1975.

Knowles, John. The Life and Writings of Henry Fuseli, 3 volumes. London, Henry Colburn and Richard Bentley, 1831.

Russell, Bertrand. "On the Notion of Cause". Chapter 9 of Russell, Mysticism and Logic and Other Essays. London, George Allen and Unwin, 1959.

Atheism, religious faith, and political discourse

1. Introduction

Our concern here is with the roles of religious faith in political movements. "Faith" will cover both the kind of faith that involves affirmation of religious claims of a factual nature, and other ways of being religious. "Political movements" will include everything from loose groupings of people who think in the same way and identify as having that much in common up to fully organised political parties.

There is a view within some some political movements that religious faith has a central role to play in politics. Some members of these movements speak of family, flag, and faith. When this slogan is used, Christian faith is usually meant. Other religious beliefs could be given a comparable political role, but for the purposes of this discussion we shall concentrate on Christianity.

Here we shall explore whether atheists (such as the present author) could comfortably align themselves with movements within which faith had a significant political role.

Would atheists be as easily convinced as religious people that the specific political positions taken were adequately supported against alternative positions?

More generally, would atheists be entirely comfortable working within movements in which faith was important? There is no suggestion that atheists would be deliberately excluded. Atheism, in which we include the agnosticism that has no definite idea of a deity and might as well be atheism for all practical and most theoretical purposes, is now widespread. It would be an unusual political group that would exclude such a large proportion of the population and thereby lose a lot of potential members and supporters. Moreover, many people would probably not wish to be associated with a group that took such a censorious attitude based on religious difference. But atheists, even if accepted, might still feel ill at ease.

2. Support for political positions

Religious faith may play a role in supporting political positions as against alternative positions. Arguments for one position as against another might be arguments with other people, or arguments within the minds of individuals who were aware of alternatives but who sought to be fortified in their convictions.

We must consider whether an atheist could see faith as providing any such support, and then if not, whether that would create difficulties for atheists.

When we speak of faith, we mean actual religious faith (although, as we shall see, not necessarily centred on propositional content). Some speak of faith in Christian values, which one might read simply as commitment to a range of values traditionally associated with Christianity without any need for distinctively religious faith. But that would reduce the relevant political debates to arguments about the merits and the applicability of different values, to be conducted in the ordinary way without any specifically religious element. We shall in sections 3.2 to 3.4 consider ways of understanding religious faith that edge toward that sort of position, but they still keep one foot in religion and thereby prevent such an easy reduction.

3. Could an atheist see faith as providing support?

Religious faith may be regarded as a matter of belief in certain propositions, or as something else. We shall consider the first option in section 3.1, and the second one in sections 3.2 to 3.4. We shall use the term "religious belief" to denote the propositions that may be believed, and the term "people of faith" to encompass both people who do believe them and people who understand their religion in some other way.

3.1 Faith as belief

From an atheist starting point, religious belief as it is found in Christianity or any other monotheistic religion would require belief that God existed, taken as a factual claim. There are variants. A Jesuit priest once told the present author that the wording should be "God is", rather than "God exists", but that variant would not help in the discussion that follows here. Nor would other sophisticated theological approaches help, because the claims made would still be rejected by atheists as false or regarded by them as meaningless. Sophisticated theology may appeal to those who are already believers. They do not need to find a way round any objection that there is no good reason for those not already committed to a religion to believe the factual claim that there is a God. It has no appeal to the atheist as a way round such an objection. And since atheists are our concern here, their need for straightforward facts must be respected if they are to be convinced.

Atheists simply cannot allow a theistic claim. So they cannot accord religious belief an argumentative role in supporting political positions against alternatives. An understanding of religion could still be important in political debate, helping the comprehension of positions by reference to their historical and social contexts. But that would be all. Suppose that some religious claim is designated "p", and some political claim is designated “q”. "If p, then q" can provide no encouragement either to accept or to reject q if one thinks that p is false, or indeed that it is meaningless.

There would also be no good way to repair the damage that would be done to the political force of faith as belief by the rejection of theistic claims. Given that rejection, and without adopting some other model of faith, Christianity would be reduced to exhortation backed up by stories that might be historically accurate in their naming and description of people, places and events, but in the reading of which literal readings of all references to divine aspects would have to be avoided in order to avoid factual inaccuracy. Then Christianity would be too weak to be a source of imperatives with sufficient force behind them to override competing imperatives.

There would be an additional issue for atheists. They would want the polity to be one that was for them just as much as it was a polity for people of faith. If religious belief were to be essential to being a full member of the polity, they would find that unacceptable. And the larger the role that was given to religious belief, the greater the danger that belief would be essential to full membership. Even if pretended belief sufficed, that would require the great discomfort of continual dishonesty.

Finally, any choice or rejection of religious beliefs to suit politics would imply that the beliefs were not regarded as factual matters. Where facts are at stake, we do not get to choose the facts. Indeed, if one were to start from political preferences and then choose religious beliefs, any support for the politics from those beliefs would be founded on circular reasoning.

3.2 Faith without belief

If we take faith as belief, the beliefs involved would be ones that atheists would reject. Faith in that form could not for them do anything to support political positions. We should however consider forms of religious faith that, while rooted in straightforwardly theistic traditions, do not require factual claims that atheists would reject. Could any such forms play roles in supporting political positions against alternatives?

Such forms of faith may be classed under two headings, although any individual might take elements from both or even fully combine them. The division we envisage is one between the primacy of an imperative to live rightly and the primacy of a frame of mind.

3.3 An imperative to live rightly

There is an approach to religion that goes under the name of orthopraxy, or correct conduct. While this encompasses both liturgical conduct and ethical conduct in everyday life, there is a decided rooting in a religious tradition in which orthodoxy is central, leaving little or no scope for atheists to participate in orthopraxy in its full form.

Atheists could however take on the notion of orthopraxy in the sense of ethical conduct, and they often do. The notion is not merely the idea of conduct of a particular nature. It incorporates what is captured by the prefix "ortho". This indicates a link to regulation by some external source of direction, so that mere private habits would not qualify even if they happened to lead to the same conduct. In relation to ethical conduct in everyday life, atheists could recognise the significance of having an external source of direction just as much as people of religious faith could recognise it, even though their external source would be likely to be an impersonal body of ethical thought rather than a personality or a sacred text.

The specific conduct that interests us here is the promotion of selected political positions because their promotion seems to be the right way to act in debate. We must ask whether the way in which atheists might consider themselves required to live rightly would in that context be a good fit with the way in which people of faith might rely on their religion to justify promoting selected positions.

3.3.1 Thinking in parallel

People of faith within the political movements that interest us might well take the view that in the political context religion was primarily about doing the right things, not by accident but because of religious motivation, and that even if factual theistic belief was also important to them, it did not matter so much when deciding which political positions to promote or oppose.

With that focus on promoting or opposing political positions, atheists might not have any difficulty in working closely with people of faith who saw the application of religion in that context as not being dependent on factual religious belief. The immediate question for atheists in politics would be, "Are these the policies to promote, and those the ones to oppose?". They might easily give the same positive and negative verdicts on particular policies as people of faith would. They would do so on grounds that did not include even implicit reference to the divine, but that to them would be just as imperative as the divine would be to people of faith.

In order to parallel in their minds the model of religion as a matter of living rightly that we envisage here, atheists would have to be motivated by considerations that made their choices of policies mandatory. They could however supply such considerations from philosophical positions that were not theistic in nature. Examples would include utilitarianism and virtue ethics. There would still be a difference from religion in that the source of considerations would be purely secular, but that would be a tolerable difference. It would however be important in supporting political positions against alternatives that they did have some such grounds. The approach to politics that we envisage without some supporting orthodoxy or other external motivation would not count for much in political argument.

3.3.2 Current disagreement

Atheists might sometimes give different verdicts on policies from people of faith, but that would not be a serious obstacle to working comfortably together so long as differences did not arise in respect of prescriptions that were important enough to make acceptance of them practically a condition of belonging to the political movement, or important enough to individuals to make them think it unethical to work with anyone who advocated a contrary position. After all, different people of faith can also have different views on the same issues.

Having said that, the role of religion might make it likely that there would be conflict in respect of prescriptions of great importance, a contemporary example being a prescription actively to oppose (or to support) abortion rights. We should expect religion sometimes to create conflict in respect of prescriptions of great importance, because religions typically have specific things to say about fundamental aspect of human life.

When such a conflict arose, people on one side of the debate might not want to work closely with people on the other side in the promotion of a party that sought a place in government, for fear that the party would then legislate for the position on that issue that they abhorred.

3.3.3 Stability and future disagreement

For a person of faith for whom orthopraxy was in the forefront, orthodoxy would still be in the background, both as an optional extra that the person of faith could choose to add to his or her orthopraxy and by virtue of orthopraxy's roots in traditions of orthodoxy. That would give a certain stability to the detailed positions adopted, at least so long as the individual's faith lasted. There would always be scope to interpret a religion differently, as shown by the fact that people of different political persuasions manage to claim the same religion as central to their own positions. But the way in which a specific individual interpreted his or her religion for political purposes would probably be reasonably stable, especially if he or she was in a political movement in which many others were of the same religion.

For an atheist, there would not be the same kind of source of stability. General ethical positions, such as utilitarianism or some virtue ethic, might provide reasonable stability, along with some predictability as to how novel issues would be handled. But such positions are at sufficient distance from the specifics of life for this to be far from guaranteed. Religions, on the other hand, having grown up over centuries in actual societies where there was a need to achieve agreement on laws and conduct, tend in their principles to be closer to the specifics of life.

The result of this difference, together with the fact that religious principles sometimes promote results at variance with what schools of secular ethics can recommend, would be potential for disagreement over questions of what should be done that might arise in future. There might even be a current fear that such future differences could lead to schism within the relevant political movement. Moreover, the different grounds of commitment to specific positions, religious and secular grounds, would encourage different forms of political argument and rhetoric in the present, making the difference in grounds visible to the public and thereby perhaps reducing support because the risk of a political party's taking an unexpected direction would deter voters. For reasons such as these, disquiet at being in the same political movement as people on the other side of the divide between people of faith and atheists might persist.

3.4 A frame of mind

It is possible to regard religious faith as a matter of seeing the world in a particular way. This does not involve assent to propositions that would be specific enough to be assessed for truth or falsity in the way that theists and atheists might assess a claim that God existed. Rather, it amounts to commitment to a view that, while it could expressed in propositional terms, would have to be expressed in propositions so general that they would not be testable. Examples would be the proposition that reality was all a single substance, in a Spinozist kind of way, or the proposition that everything had mental qualities, in a panpsychist kind of way.

Such propositions would be accompanied by an imperative to look at the world in their light. Thus someone who adopted a Spinozist approach would see the world as both a divine thought and a physical reality, while a panpsychist might see each event as explicable by reference to the purposes of the participant entities. One consequence might be to respond to the world as a person of orthodox faith might well do so, at least at the level of broad values such as concern for people and the environment and respect for all. There might also be an inclination to attend religious services, not in order to affirm the contents of creeds but because one would feel at home there.

Any such framing of the world would be flexible in its effects on the selection of particular prescriptions for views or conduct. This flexibility might be reduced by a background of orthodoxy. Such a background might have an influence even if an individual did not subscribe to any such orthodoxy, because the place of the orthodoxy in the relevant cultural history would have shaped the range of options likely to occur to the individual and the ways in which he or she would interpret the chosen option. But there would still be some flexibility, and some potential to derive alternative and perhaps opposing prescriptions.

It is therefore unlikely that such a framing could suffice to support political positions against alternative positions. Even if a given framing were agreed, it would be possible to derive support for different positions from it. And it is unlikely that a framing would be agreed precisely, given that it would not even be formulated with any great precision. So atheists who rejected such framings, seeing them as mystical, would not be any less able than people who adopted such framings to find support for political positions against alternatives.

A framing might however still suffice to resolve a debate within the mind of a single individual, with the outcome fortifying the individual in his or her convictions.

4. Working with people who rely on faith

No political movement that could only attract people who had religious faith, or could only attract people who did not, could expect to make great progress in a modern democracy. So we must consider any difficulties that might arise when atheists and people of faith sought to work together.

4.1 Support for positions

It would be difficult for an atheist to see faith as providing support for political positions. Faith as belief could not do so, because an implication from a religious belief to some political position would have what the atheist regarded as a false antecedent. And faith without belief, while it might encourage people with the faith to sustain their political positions, would be seen by the atheist as providing merely psychological support, the existence of which had no probative value.

This would not in itself need to present any difficulty, so long as atheists could find adequate reasons of their own to support the positions that were important to the movement, or alternatively if any failures to support certain positions could be tolerated.

There might however be a felt loss of coherence of the overall stance of the movement. If the main positions were mostly natural consequences of the relevant religious faith, that would for the faithful justify holding the positions as a package. Without that central core of faith, the different positions might come across as a disparate bunch, justified individually but not collectively.

Moreover, future loyalty to the movement would be more uncertain for the atheist than for the faithful. There is a view that this would be a good thing. It is rational to consider each political position on its own merits, and to be willing to change one's views if appropriate. On the other hand, the faithful might be uncomfortable with the constant risk that their atheist colleagues would start to deviate from the programme, and atheists might feel uncomfortable at the expectation of continued loyalty to a programme that could not remain static but would have to evolve as new developments in the world demanded new responses.

4.2 Conflicts over positions

If religious faith were central to an approach to political issues, it is quite likely that at least a few positions would flow naturally from the relevant faith but would conflict with the values of some of the people without faith. So there would be occasions when atheists would want to reject some positions within a broad faith-inspired programme.

How far this would be likely to happen would depend on the core values that went along with the relevant faith. We shall now take a look at what might be called Judaeo-Christian values, and then consider some specific conflicts.

4.2.1 Christian and Judaeo-Christian values

The phrases "Christian values" and "Judaeo-Christian values" are widely used, but they are easy to interpret in several different ways. If pressed, those who use them within the political movements that interest us here will cite some specific values, such as justice, mutual respect, doing as you would be done by, charity, the sanctity of life, and avoidance of behaviour traditionally regarded as sinful. But those values too are not especially specific in their implications for contemporary policy. It can however be clearer what they forbid than what they require, as when references to the sanctity of life are used to argue against abortion rights, or an imperative to avoid sin is used to justify forbidding certain personal relationships.

Clarity on prohibitions reflects the Old Testament inheritance, with its regulations and ideas of structures of authority. Those regulations and ideas of authority are more muted in the New Testament. Their strongest recurrence is in the Pauline Epistles, but there they come across as balanced on the single point of faith in Jesus, lacking the force that comes from being rooted in the profound history of a people that is recounted in the Old Testament.

It is therefore not surprising that the concept of Judaeo-Christian values, including the Old Testament part, should have considerable appeal to members of a political movement who advocate family and flag. But it also prepares the ground for conflicts on some specific issues between members with religious faith and atheist members.

4.2.2 Specific conflicts

Atheists are likely to object to the promotion of any particular religion, or of religion generally, in schools or in the life of the state. They regard the beliefs promoted as simply false, and therefore as not to be advocated. This opposition may bring them into direct conflict with those who advocate a political role for faith. Those who think that faith has a valuable role in supporting their political choices are likely to want those choices to be supported by as many people as possible, in order to ensure their implementation in a democracy.

A related point is that an atheist of integrity would not be able to say that while he or she did not believe the propositions of religion, it would be good if the population at large did. That would be to say that it was good for people to be misled in certain ways, a position inconsistent with taking truth seriously.

On the issues of abortion and euthanasia, atheists are more likely than people of faith to be in favour of the freedom of the mother and the terminally ill respectively. There is no guarantee that any given atheist or person of faith would take the position one would expect. There will be anti-abortion and anti-euthanasia atheists, and pro-abortion rights and pro-euthanasia people of faith. But a political movement will tend to want a single coherent position on each major issue, and the more people of faith dominate, the more likely it is that this position will be anti-abortion and anti-euthanasia. That may leave a substantial proportion of atheist supporters uncomfortable, and even schismatic.

Another area of potential conflict is personal lifestyle. Atheists are more likely than people with religious faith to take the position that individuals should be free to live as they choose so long as their lifestyles do no harm to others. And someone else's horror that anyone should be living in a particular way would not count as harm. Without a notion of divine disapproval, the regulation of conduct should be limited to conduct that can be characterised as wrong on purely secular grounds. People of faith, on the other hand, may well have a notion of divine disapproval and see it as part of the role of the state to forbid conduct that would incur that disapproval.

Here, the notion of freedom that is sometimes joined to family, flag, and faith is at stake. The libertarian position would be that freedom included leaving harmless people alone. A more conservative and quite possibly faith-inspired position would be that it required allowing such freedoms as might be suitable while in some respects constraining people for their own good or for the good of society. As with other issues, there is no guarantee that any given atheist or person of faith would take the position one would predict in this way. But again, there is a tendency for a political movement to want a single coherent position on each major issue. And whichever way such a decision was made, one might expect some disaffection, either predominantly on the atheist side or predominantly on the side of people of faith.

One reason why conflicts over issues like abortion, euthanasia, and personal lifestyle may be hard to resolve by discussion within a political movement is that the grounds for positions may be different. For the atheist, the grounds may be given in arguments against a background of secular values that are themselves open to challenge by argument which may lead to adjustment of those values. For the person of faith, the grounds may be given in arguments against a background of values that come from the relevant religious tradition and that are not themselves to be challenged by argument, but only rejected by abandoning the faith in question.

5. Avoiding reliance on religious tradition

There would be advantages in a political movement that had hitherto given an important role to religious tradition moving away from any reliance on that tradition, whether the tradition was still linked to actual faith or was seen as merely historical.

5.1 The risks of reliance

When a tradition was linked to actual faith, there would be a risk that people, including some of the keenest supporters, would lose their faith. Then if faith was important to the political movement, those people might very well leave the movement. Alternatively, if they were a large enough group, they might seek radical change within the movement, and schism could easily result.

Even when a tradition was seen as merely of historical significance, and as a way to influence current choices of policy on the basis that the tradition was central to the relevant country's culture and that it would be sensible to choose policies that fitted well with that culture, the argument for respecting the tradition would weaken as faith diminished in the population. To sustain respect, there would need to be a reasonable proportion of the population who thought that the relevant religious beliefs were true. In a generally atheistic population there would be considerable doubt that the tradition deserved continuing respect, simply because the beliefs on which it was founded were seen as mistaken.

5.2 Alternative support for positions

If faith were discarded, and moreover the weight of a tradition viewed merely historically was reduced because the beliefs on which it was founded were seen as mistaken, what alternatives might be available to support the positions typically favoured by the political movements that interest us?

Fortunately, there are rich traditions of ethical thought on which to draw, some developed in Europe from Ancient Greece onward, and some developed elsewhere in the world. These traditions have deep roots in culture, placing them on a par with established religions in that respect. Abstract ethical theories when considered independently of history would come across as lacking such roots, leading to a fear that they would be flimsier than religions, but in fact there is history to hand.

European traditions of thought that have endured were, in years AD, influences on and influenced by Christianity, but they are still identifiable as traditions that could stand independently of theistic religion. To see that, one only has to wind the clock back to the time before Christianity became widespread, a time for which there is also not much sign of dependence on Jewish thought, and then imagine winding the clock forward again on a path without Christianity. The vaunting of truth, beauty, and in political arrangements courage and cooperation with others can all be found in such traditions. And subsequent ethical thought, such as Locke on government, Kant on duty and justice, or Mill on liberty, could easily have arisen and stood independently of theistic traditions. (Locke and Kant would require some editing, but their edifices would not crumble.) 

We do not here reach directly for the work of those who characterise themselves as modern-day humanists, although their promotion of values is explicitly non-theistic. This is not to scorn their work. But the values for which they tend to argue are not ones that are distinctive of the political movements that interest us. Rather, they are values of scientific progress, liberty, and consideration for others that reach across the political spectrum. We may however make indirect use of the modern-day humanists by pointing out that their publications show how values can perfectly well be supported without any kind of religion.

5.3 Consequences of reliance on alternative support

Support from religious belief can easily be support for a complete package of political positions. Quite a lot of interpretation and argument to conclusions may be needed to reach political positions appropriate to the modern world from an inherited corpus of texts and doctrines, but support for anything in the package can be seen as derived from that corpus.

Correspondingly, the exclusion of rival political positions may be supported by interpreting and drawing conclusions from the same corpus. Rivals may be able to cite the same corpus in support of their positions. But at least the debate can be shifted from "Your position is mistaken" to "You are interpreting the corpus and drawing conclusions in the wrong way". Then it is easier to be confident in one's own positions, because one's confidence is based on a view that one is epistemically superior in general terms, a view that need not be defended by reference to any particular interpretation of the corpus or by reference to any other body of information at a comparable level of specificity. This is not to say that a claim of epistemic superiority would be justified. One might invent the claim that opponents were misusing the corpus simply because one did not like their conclusions, and then hunt around for moves in their thinking that could be claimed to be mistakes. But at least one would have found a move to make.

All this changes when one moves away from support for positions that is based on religious belief. Support for a complete package would be most unlikely without a single foundational corpus. The most extensive foundational corpus would be the set of established results in the natural sciences, but while such results would sometimes be necessary to make a good secular case for a political position, they would rarely if ever be sufficient. So one would be left having to justify positions one by one, rather than as a package. There might well be constraints on positions by reference to other positions. The final package would have to be coherent. The adoption of some positions might require or exclude the adoption of others. But imposition of the need for coherence would come after the search for support for specific positions against rivals. And while coherence might be promoted by adoption of a single overall approach to ethics, such as utilitarianism or virtue ethics, such approaches would be somewhat too distant from specific political positions to give much of a guarantee of coherence, given that many policy choices could not easily be ranked by their exhibition of virtues or generation of happiness. Only some extreme choices would be likely to be easy to exclude by reference to such criteria.

One consequence is that specific positions would come to seem more open to change than when they were based on religious belief. Any position could at any time be seen to have lost its secular support, without that requiring the large mental shift that would be involved in abandoning a religious tradition. There would be no guarantee underwritten by a tradition that we would keep our political positions stable.

On the other hand, a consequence that might be found more pleasing would be that political positions might maintain their support even if religious belief were lost, or respect for a religious tradition faded. Religious belief has declined among the population as a whole, and with it any automatic entitlement to respect that religious traditions may have enjoyed in the past. Even within a movement in which most members were committed to religious belief or at least to deep respect for a tradition, the fact that belief and respect had declined among the wider population should serve as a warning that they could just as easily decline among the membership of the movement.

It follows that support from religious belief or from respect for a religious tradition would be fragile, however long the belief or the respect might have lasted so far. Reliance on secular arguments for positions might be safer, even if the consequence was that the overall programme of the movement might come across as piecemeal.

Members of a political movement that had historically given a central role to religious belief or to respect for a religious tradition might however still feel nervous about a shift to secular support for their positions, even if such support was reasonably easy to find. As the development of modern-day humanism has shown, it is entirely possible that values somewhat out of keeping with the traditions of movements of the kind that interest us would come to get equally good support and would be adopted.

Vinteuil's sonata

1. Proust and the musical phrase

1.1 The text

In Marcel Proust's novel Du côté de chez Swann, we find these reflections on a phrase in a sonata by the fictional composer Vinteuil:

Swann n’avait donc pas tort de croire que la phrase de la sonate existât réellement. Certes, humaine à ce point de vue, elle appartenait pourtant à un ordre de créatures surnaturelles et que nous n’avons jamais vues, mais que malgré cela nous reconnaissons avec ravissement quand quelque explorateur de l’invisible arrive à en capter une, à l’amener, du monde divin où il a accès, briller quelques instants au-dessus du nôtre. ... Et une preuve que Swann ne se trompait pas quand il croyait à l’existence réelle de cette phrase, c’est que tout amateur un peu fin se fût tout de suite aperçu de l’imposture, si Vinteuil ayant eu moins de puissance pour en voir et en rendre les formes, avait cherché à dissimuler, en ajoutant çà et là des traits de son cru, les lacunes de sa vision ou les défaillances de sa main.

(Proust, Du côté de chez Swann, pages 413-414.)

Here is C. K. Scott Moncrieff's translation:

So Swann was not mistaken in believing that the phrase of the sonata did, really, exist. Human as it was from this point of view, it belonged, none the less, to an order of supernatural creatures whom we have never seen, but whom, in spite of that, we recognise and acclaim with rapture when some explorer of the unseen contrives to coax one forth, to bring it down from that divine world to which he has access to shine for a brief moment in the firmament of ours. ... And one proof that Swann was not mistaken when he believed in the real existence of this phrase, was that anyone with an ear at all delicate for music would at once have detected the imposture had Vinteuil, endowed with less power to see and to render its forms, sought to dissemble (by adding a line, here and there, of his own invention) the dimness of his vision or the feebleness of his hand.

(Proust, Swann's Way volume II, pages 184-185)

1.2 Reality and constraint

There are several philosophical themes one could draw out of this paragraph and the surrounding text. Our concern will be with the claim that the reality of the phrase was shown by the fact that it was constrained to be exactly as it was. Any change would have been substantially to its detriment. It had to be as it was in order to be superlative, so it must have pre-existed to allow it to be discovered rather than merely invented. Only the pre-existing real has to have a particular nature.

In section 2, we shall consider how a piece of music might be constrained to be exactly as it was.

In section 3, we shall consider what sort of reality might be possessed by pieces of music.

The phrase in the sonata has also been discussed by Maurice Merleau-Ponty (Le visible et l'invisble, pages 195-204: Vinteuil is not mentioned by name, and Merleau-Ponty simply refers to the "petite phrase"). We shall however not pursue Merleau-Ponty's line of thought here. That would require a discussion in a different intellectual context from our preferred one, the context of mainstream analytic philosophy.

2. Constraint

2.1 Constraints and the superlative

The constraints that interest us are constraints on a piece of music's being superlative, not on its being music. There is no suggestion in Proust's words that the piece of music to which he refers would be constrained in the sense that alterations would lead to its ceasing to be music. It would merely cease to be superlative.

We shall focus on pieces being regarded by human beings as superlative. We shall take no interest in the idea that a piece of music might be superlative without human beings having the ability to recognise its quality. This is not to say that musical quality would be decided by majority vote, or by universal agreement. It is merely that we shall not suppose a quality that human beings could not discern. We do leave open the possibility that a piece's quality might only be well-discerned by people with particular expertise. But we take care to say "well-discerned". We do not say "rightly discerned", because that would hint at there being a standard independent of human judgement such that even the best human judges might not be reliable judges.

2.2 Sources of constraint

Constraint might arise from there being a quality that was possessed by only a small proportion of possible pieces and that was not to be derived from other principles, along the lines of goodness as a simple non-natural quality in G. E. Moore's book Principia Ethica (chapter 1, sections 5-14). But the quality could still be discerned by human beings. 

Constraint might alternatively arise from principles of musical quality that were not fully to be derived from features of the world outside the sphere of aesthetics. The principles might be wholly of non-natural origin, or they might be partly explained by reference to natural features of human beings, in particular their perceptual organs and the workings of their brains. And the principles could be discerned by human beings.

Finally, constraint might arise entirely from natural features of human beings. The natures of our ears and our neural systems might suffice to determine which pieces of music would qualify as superlative.

2.3 Evidence for a simple quality or for principles

Constraint on a piece of music to be exactly as it was in order to be superlative would amount to evidence that there was some way to classify pieces of music as superlative or not which did not amount to allocating pieces to one class or the other at random.

It would amount to such evidence on the basis of an inference to the best explanation. A non-random way to classify pieces, whether by reference to a simple non-natural quality or by reference to principles, could easily forbid even very small changes. A random allocation of pieces would be more likely to allow into the class of superlative pieces some pieces that were only slight variations on a given superlative piece.

Thus the presence of constraint on pieces to be exactly as they were in order to be superlative would indicate that one or other of two alternatives held. The first alternative would be that there was such a thing as an intuitive grasp of musical quality, an ability reliably to recognise some simple non-natural quality that put pieces of music in the superlative class. The second alternative would be that there were principles of classification that could be discovered. And any such indication, whether of an intuitable quality or of principles, would be particularly strong if we could discern features that a good number of superlative pieces of music had in common but that lesser pieces lacked. That would give reassurance that there was something of systematic effect behind the identification of some pieces as superlative. 

3. The reality of pieces of music

3.1 Ways to be real

Now we can consider whether it would make sense to say that superlative pieces of music that could not be altered without losing their status were real.

There is a straightforward sense in which they would be real. Sequences of notes and indicators of style of performance could all be written down, and recordings could be made. Pages of sheet music and drives that hold audio files are not ethereal, they are physical. And their informational contents, along with the informational contents of advice on how to perform pieces, are as real as other abstract objects. Moreover, we routinely refer to particular pieces of music in the way that we refer to other everyday objects.

This would however be a trivial sense of reality, which would extend to all pieces of music. It would not be any sense of reality that Proust could have had in mind. He was thinking of the pre-existing reality of a given superlative piece, waiting for a composer to discover it. The pre-existence in question would of course not involve the actual auditory production of the piece or the physical existence of sheet music. Rather, it would be abstract pre-existence of the kind that mathematical objects can easily be argued to have, the sense in which groups, rings and fields arguably existed long before the first developments in abstract algebra during the late eighteenth century.

One way to make sense of the supposed pre-existing reality of pieces of music, despite their naturally being taken to be human creations entirely dependent on their composers, would be baldly to assert their existence in a Platonic realm. Another way would be to draw on recent thought in the philosophy of science that ties reality to the task of making sense of the world. We shall explore each in turn, but first say a bit more about the idea of discovery rather than invention.

3.2 Discovery rather than invention

If a composer experiments with various options for motifs, sequences, key changes, harmonisations, and so on, and eventually concludes that there is only one way to turn an initial idea into a superlative piece, it seems right both to acknowledge that the initial idea was invented and to say that the finished piece was discovered. This may seem ambitious. Pieces of music seem obviously to be invented. But a case for discovery can be made.

The finished piece was not deducible from the initial idea, because it does not in general follow from an initial idea that a piece of music (superlative or not) must be as the finished piece actually is. Rather, there was a search for a piece that could be seen as perfectly embodying the initial idea. The position is analogous to that of a biochemist who has the idea that some molecule or other of a given general nature must exist to perform some important function in organisms, and then sets out to discover what the molecule actually is.

We are however left with the question of whether discovery would imply pre-existing reality. And the ontologically parsimonious could say that for music it did not. Rather, they could say that discovery in the sense in which it covered both molecules and music was simply a particular type of mental process of the discoverer, one that involved working out what was needed within tight constraints. And while that kind of process was required when the things discovered were pre-existing (as molecules certainly are), it could also be required when the targets of the mental process were not pre-existing (as pieces of music might be argued not to be).

We must therefore turn to approaches that would move us beyond what the mere idea of discovery can offer.

3.3 A Platonic realm

It would be possible simply to assert that a superlative piece of music pre-existed in some Platonic realm, waiting to be discovered. Pieces of music might be in the set of superlative pieces as a brute fact, and be possessors of a simple intuitable quality of superlativeness, or they might be in the set when they complied with certain constraints, constraints that arose out of certain principles. We shall consider the second alternative before finding ourselves driven to the first one.

Suppose that compliance or non-compliance with constraints founded on principles was a reliable guide to pieces being or not being superlative. In order adequately to motivate speaking of a Platonic realm, rather than taking the option of merely reducing pre-existence to compliance with principles and constraints that themselves pre-existed, the principles and constraints would need to be discerned by first contemplating pieces within the set of superlative pieces. They would need to be regarded as principles and constraints that indicated, rather than led to, membership or non-membership of the set. The direction of explanation would need to be top-down from the Platonic realm, rather than bottom-up from the principles and constraints. Without such motivation, assertion of a Platonic realm would be a mere ad hoc convenience to satisfy Proust.

We say that the direction of explanation would need to be top-down because if membership of the set of superlative pieces were to follow from general musical principles, the fact that alterations would spoil a piece would be adequately explained by the fact that those principles would then be violated. There would be no need to cite the piece's membership of the set, and an altered version's failure to belong to the set, as the explanation. To cite membership as the explanation, we would need membership to be a brute fact, on a par with the brute fact of the existence of a particular concrete object, not explained by general features of the world.

This does however give rise to a challenge to the Platonic approach. We would very often expect it to be explicable by reference to general principles why a piece of music had to be as it was, and not be any different, in order to be superlative. Even if being superlative amounted to possession of a non-natural quality, similar to G. E. Moore's quality of goodness in that it was not to be defined by reference to natural qualities, it might depart from Moore's approach in that it might not be a simple quality. Instead it might be defined in terms of other non-natural qualities, the instantiation of which was required by principles which were not grounded entirely in natural facts about human beings and the world. So it would seem that unless we could insist on Moore's feature of simplicity as well as that of non-naturalness, membership of the set of superlative pieces would rarely if ever do enough work to justify hypothesising a Platonic realm as a manner of pre-existence.

Moreover, working out the details of a Platonic realm of superlative pieces of music would not be straightforward. In order to be selective and not be open to all pieces of music, the realm would have to be populated by specific pieces, rather than by the form of musical pieces in general.

On the bright side we could say that pieces of music, at least when wordless, were form rather than content in the sense that they could be characterised in ways that did not require reference to physical objects. To that extent pieces of music would be well-suited to existence within a Platonic realm. One could explore such questions further by considering Schopenhauer's reflections on music in Die Welt als Wille und Vorstellung, volume 1, chapter 52, although we shall not do so here.

Despite this reassuring thought about form and content, we cannot conclude that asserting existence within a Platonic realm is a well-motivated way to claim pre-existence for pieces of music. We must therefore leave Proust with less than is suggested by the second sentence of the paragraph we quoted, with its reference to supernatural creatures. We may however still be able to justify seeing superlative pieces of music as pre-existing in some less poetic sense. We shall now turn to an approach that is an alternative to the assertion of a Platonic realm.

3.4 Making sense of the musical realm

Hasok Chang, in his book Realism for Realistic People, makes a case for what he calls active realism. He ties reality to the activity of investigating and manipulating the world. As he puts it, "an entity is real to the extent that there are operationally coherent activities that can be performed by relying significantly on its existence and its properties" (page 121). The requirement for coherence in activities rules out attributions of reality ad hoc to entities that it might temporarily be convenient to regard as real, and attributions that would serve no purpose or would lead to incoherence in our activities. In Chang's words, "in order to try to make something happen, the agent has to coordinate carefully various movements and thoughts with each other and with external circumstances, towards the achievement of an aim. ... I propose to use the term operational coherence to refer to such a state of aim-oriented coordination" (page 24).

Chang's work is a contribution to the philosophy of science. And he gives us a concept of reality that can be a matter of degree (page 142). So we cannot expect simply to borrow his idea and use it to generate Platonic reality (an absolute, not a matter of degree) in music (a significantly different field when our concern is the aesthetic quality of pieces, rather than any physiological or neurological form of musicology). There is however some use to be made of Chang's approach.

We can make use of the approach by focusing on the question of whether there is good reason to regard superlative pieces of music, but not other pieces, as pre-existing, even though all pieces look as though they are simply invented by composers. This question of good reason is what matters within the context of a pragmatist approach such as Chang's. Under pragmatism, there is no thought that there might be a truly Platonic realm of things that existed independently of human thought. The question of existence reduces to a question of whether we should regard certain things as existing. Is it useful to regard superlative pieces, but not others, as pre-existing? Here use means use in the conduct of operationally coherent activities, not just temporary convenience.

We can identify a coherent activity that is aided by regarding superlative pieces as pre-existing. This is to account for the impossibility of varying superlative pieces without demoting them to everyday pieces. If superlative pieces pre-exist, they will do so in their precise form. Neighbouring pieces that incorporate small variations will not be within the realm of pre-existing superlative pieces. The composer of superlative music must discover the precise piece. Lesser pieces, on the other hand, can be seen as invented in that they could safely be varied, the choice of one variant over another being part of the act of invention.

This is not to say that the composition of superlative pieces is purely an act of discovery. The composer will engage in invention when he or she decides that what is needed is a piece along such and such general lines, and perhaps narrows that characterisation. But the final step will be an act of discovery of exactly how the piece should be. (We might compare the explorer who has a general idea that a river must have a source, perhaps a spring or perhaps a lake, and possibly a few thousand kilometres upstream, but who has to discover the actual location and nature of the source.) The composer of a lesser piece, on the other hand, need make no such final step of discovery. His or her work can be invention all the way, given both the freedom to make large and small adjustments, and the lack of a single correct solution out there to be discovered.

It is also not to say that this supposition of pre-existence is the only way in which we could account for the impossibility of varying superlative pieces without demoting them. Their composition could be all invention, with no need for a component of discovery. But it is a good way to account for the impossibility of variation. It gives us a picture of a landscape of possible pieces, with altitude corresponding to quality, in which there are some isolated peaks of quality with vertical drops to the surrounding terrain, rather than gentle slopes of variation on which one could take a position of high but not exceptional quality.

We should at this point note another theme in Hasok Chang's book, that while nature may be pre-existing it is not pre-figured (sections 2.1 to 2.4). Nature is as it is, so our theories can be found satisfactory or unsatisfactory. But that is not to say that the concepts we use to understand nature are already to be found in any pre-existing structure of nature. We have to bring our concepts to bear. As Chang puts it, reality is mind-framed (section 2.1). If use of the concepts we currently have permits coherent activity, that is all to the good. If it does not, we must try other concepts.

Turning to music, there are two layers to identify in the build-up to superlative pieces. The first layer comprises natural facts about human beings which mean that they will find some combinations and sequences of sounds more attractive than others. Those facts are set out using concepts that we have invented, and that allow us to conduct the coherent activity of investigating auditory perception. The second layer is a musical tradition, which includes the principles that govern the quality of music within that tradition. The tradition has been created by us to allow the coherent activity of composing good music. It is a discovery that the concepts used in formulating the tradition, and the tradition itself, are productive given the first layer. But the first layer does not contain either the musical concepts or the tradition as a whole. To apply Chang's picture, relative to the second layer (but not absolutely), the first layer gives us nature in itself and the musical concepts and tradition turn out to be a way to work productively given how things are in the first layer.

Finally, we can fit superlative pieces into the picture. The identification of them as superlative is accomplished without introducing any new concepts. Here there is a break with Chang's vision of the natural sciences. Given the musical tradition, it is not an act of conceptualisation that makes certain pieces superlative. They simply are superlative.

Here is permission, although not in itself sufficient reason, to think of superlative pieces as pre-existing. They are not fashioned out of a musical tradition by introducing new general concepts. Their individual existence is directly latent in the tradition. To move on to sufficient reason, it makes sense to think of them as pre-existing so as to account for the fact that small variations would demote them to pieces of much lower stature.

We have not achieved Platonic reality. But we have at least made out a case for an admittedly ontologically daring way to make sense of the feature of a superlative piece that so struck Swann, the impossibility of making small variations.

References

Chang, Hasok. Realism for Realistic People: A New Pragmatist Philosophy of Science. Cambridge, Cambridge University Press, 2022.

Merleau-Ponty, Maurice. Le visible et l'invisible. Paris, Éditions Gallimard, 1964.

Moore, G. E. Principia Ethica, revised edition edited by Thomas Baldwin. Cambridge, Cambridge University Press, 1993.

Proust. Marcel. Du côté de chez Swann. Paris, Éditions Gallimard, 1954.

Proust, Marcel. Swann's Way, translated by C. K. Scott Moncrieff, volume II. London, Chatto and Windus, 1934.

Schopenhauer, Arthur. Die Welt als Wille und Vorstellung. Several editions are available.

What is it like to be a ghost?

This is the topic for our Cambridge philosophy café this Halloween week. Here are the notes that have been circulated, with small changes. They comprise only questions. Answers are left as an exercise for the reader.

The ghosts of Halloween are the ghosts of folklore. They may be scary, but they vanish in the sunrise of sobriety. The ghosts of philosophy are something else. Studying them may teach us something about being human.

We find them in Jean-Paul Sartre's screenplay Les jeux sont faits, filmed in 1947. The deceased are ghosts who can walk around the city, can see and be concerned about what goes on, but can only interact with other ghosts. The living cannot see or hear them, and they can do nothing in the world of the living unless they are briefly returned to life - as happens to the two main characters.

Detachment from the world, fully aware of what was going on but unable to help, could be annoying. But more than that, is it essential to our humanity that we can do things? Does your sense of who you are depend on your being in a world that changes in response to your actions? Are you only who you are because you are at a specific location, a point from which you not only see certain things you would not see from elsewhere but can do things you could not do if you were elsewhere? (Lucy O'Brien's book Self-Knowing Agents is particularly relevant here.)

Another element of the sense of self that ghosts would lack would be a sense of achievement. It is not just that we do things. We find that the world resists. Doing things takes effort, and sometimes things do not go to plan. We can say "I did that" with pride. How important is a sense of achievement to our sense of self? Sartre's ghosts would lack any such sense.

Sometimes we must say "I did that" with shame. We make choices that we later see as bad ones. We take responsibility for our actions - or at least, we do if we are good existentialists. Is that possibility vital to our humanity? Again, ghosts would be safe from any such sense because they would not act in the world.

One of our greatest challenges is time, or the lack of it. In the screenplay we are told that the living seem to be in a hurry, while the dead just stroll around. Would we be fully human if we did not feel the pressure of time?

Situations demanding action, our responses, our successes and our failures make us different people. But if we were ghosts, without the pressures and possibilities of life, would each of us still be unique? Perhaps after a while all the ghosts Sartre imagined would be the same, dressed in the costumes of the times when they died but otherwise indistinguishable.

If you would like to join our philosophy café, information is here:

https://www.meetup.com/think-and-drink-cambridge/

Behemoth the Cat

1. Introduction

In Mikhail Bulgakov's novel The Master and Margarita, the most endearing character is Behemoth the Cat (Кот Бегемот). He stands on his hind legs, reaching over a metre in height according to his presentation in the 2005 Russian television adaptation. He is black and plump, has fine whiskers, speaks perfectly good Russian, eats with a fork, and plays chess. He also delights in creating chaos in the human world. He is immune to bullets or any other human assault, save that he does not like it when in chapter 23 Margarita digs her nails into his ear. He is one of the entourage of Woland, that is to say Satan. In his undisciplined way he supports Woland's project of disrupting Moscow life and exposing the follies of the government and the people.

If one gets truly engrossed in the novel, or indeed in the 2005 adaptation (and maybe in other adaptations), he becomes entirely convincing. He does so to a greater extent than the other members of the entourage, Koroviev, Azazello and Hella, all of whom are in human form, and perhaps to a greater extent than Woland himself. This may be because we are familiar with the inner mental lives of human beings and the natural sense that people have of their own limitations, so the human or quasi-human characters with their extraordinary powers are obviously not quite right. There is no plausible coherent mental life of a sane adult human being to match their actions. But we are not familiar with the inner mental lives of cats, so we could impute a coherent mental life to Behemoth without being able to say that it was implausible. Indeed, the mental life of an exceptionally mischievous human child who was also precociously knowledgeable and self-controlled could be attributed to him.

The convincing nature of Behemoth reaches the point that if one were to finish the novel, go out into the street, and be met by a large bipedal talking cat, one would not feel forced to conclude that one was drunk or dreaming. The encounter might seem odd, but one could take seriously the idea that this really was a talking cat.

The strange encounter would however still demand an explanation. It might be, but would not have to be, an actual talking cat. What might the explanation be? How could we come to select the best explanation, or at least identify explanations to rule out?

In section 2 we shall set out some possible explanations. In section 3 we shall move on to the problem of selection. Then in sections 4 and 5, we shall look at a way to choose explanations that would normally be ruled out and at what we might learn by exploiting that way.

2. Possible explanations

2.1 Altered states of mind

The explanation of the cat and the other strange goings on that is offered by the authorities in the epilogue to the novel is mass hypnosis. Correspondingly, if one had appeared to meet a talking cat, one might later conclude that one had been under the influence of some hypnotist. The fact that one could not recall any hypnosis session would not count against this, because that lack of awareness would itself have been engineered by the hypnotist.

Another possibility would be that one had been dreaming. That might however be ruled out by the coherence of experience over a period of some hours, and the clear awareness that we can have when awake that we are in fact awake, a particular type of awareness of being awake that seems to be absent from dreams.

A third possibility would be that one had been under the influence of alcohol or hallucinogens. But again, extended coherence of experience and a sensation of sobriety could rule this out.

2.2 Fake cats

We might be taken in by a human being in a costume, posing as a very large cat. We might also be taken in by a robot cat. In both cases, it would help that the large size and ability to talk would undermine our expectation that this was an ordinary cat. Thus it would not be vital for a human being or a robot accurately to mimic every normal characteristic of cats.

2.3 An actual talking cat

The most obvious but on sober reflection least plausible explanation would be that there had in fact been a talking cat. This could not be ruled out on logical or metaphysical grounds, or even on the ground of physics. It would take biology to discard the explanation as wholly unreasonable, even though we might without scientific thought be confident that it was not in fact correct. We shall keep this explanation in play.

3. Choosing an explanation

3.1 Discarding explanations

We are familiar with the idea that observations may appear to have a particular explanation, but on reflection that explanation turns out to be no good.

An explanation may be discarded either when further information becomes available, or when we think about what would have to be the case for the explanation to be plausible and realise that it would not be sensible to expect those further requirements to be fulfilled.

Either route to discarding an explanation could be relevant to the case of an apparent talking cat.

3.1.1 Further information

Further information could include learning that one had been under the influence of a hypnotist, of alcohol or of hallucinogens, or that one had in fact been asleep despite a clear sense that one had been awake. Then explanations other than hypnotism, alcohol, hallucinogens or sleep would probably be discarded as unnecessary (a human being in a cat suit or a mechanical cat) or implausible (an actual talking cat).

Further information could also include noticing something about the cat's movements, or sounds it emitted, that indicated either that it was a human being in a cat suit or that it was a mechanical cat. Again, other explanations would probably be discarded as unnecessary (hypnotism, alcohol, hallucinogens or sleep) or implausible (an actual talking cat).

Further information might also lead one to discard explanations directly, rather than by rendering them unnecessary. Perhaps unaltered perception of the world around oneself, or an ability to read a book easily or to follow a mathematical argument, would count against hypnotism, alcohol, hallucinogens or sleep. And an X-ray or MRI scan of the cat might show no sign of any human or mechanical presence under the fur, counting directly against such explanations.

3.1.2 What would have to be the case

Sometimes explanations can be discarded without obtaining further information, simply by reflecting on their implications against a background of existing knowledge. 

The one explanation of an apparent talking cat that might easily be discarded in this way would be that there really had been a talking cat. The first point of relevant background knowledge would be supplied by zoology: talking cats have never been observed. That would not suffice to render them impossible, but we could then move on to more general biological considerations. We could cite the need for certain brain and vocal structures in order to talk, structures that have never been observed in cats, and the need for large bipedal cats to have evolved and to have developed a human language (presumably by interaction with human beings) without anyone having noticed them before now.

The conclusion would be that talking cats were extremely unlikely to exist. At the level of biology we could firmly exclude the possibility. But we would need to invoke special sciences like biology in order to rule them out. Physics would not be enough, even though physical considerations would be invoked when thinking about evolutionary processes, necessary vocal structures, and the like. It might be argued that biology could be reduced to physics so that physical considerations encompassed biological ones, but that would be to overstate the power of physics to rule out particular organisms. The physics of the universe accommodates the actual biology of this planet and explains why certain biological combinations would be impossible, for example certain combinations of vocal structures and linguistic abilities, but the same physical laws could equally well accommodate other biologies so long as they did not require physical impossibilities. We would need to move up to the level of actual biology to argue against the possibility of talking cats, and to add the need for integration of evolution and of food chains to rule out the existence of some hitherto unnoticed parallel biology on Earth that physics on its own would also permit.

3.1.3 Saving explanations from their implications

If an explanation had to be discarded because its implications would clash with our wider knowledge, it might be tempting to narrow our focus and not take account of all potentially relevant knowledge or not draw all potentially relevant inferences.

At first glance, that move would be flagrantly illegitimate. In our academic pursuits, whether in the natural sciences, the social sciences or the humanities, we bring all potentially relevant knowledge to bear when assessing putative explanations. (We shall call this general approach of all respectable disciplines the academic approach, an approach characterised by a demand to do the best we can to discover the truth and not to flinch from drawing inconvenient inferences.) A hypothesis that an apparent talking cat is an actual talking cat is not ruled out by physics alone, but it is ruled out when we broaden our view to take biology into account. Such broad views of our knowledge are needed to stop us believing nonsense such as magic or astrology.

We do however have another option, to switch from the academic approach to what we shall call the poetic approach. This is the approach that is used in fiction, whether prose or poetry, when one wishes to say things that our academic knowledge would exclude, but it may also have application in real life. Under the poetic approach, we avoid constraints that would be imposed if we took a broad, academic, view of our knowledge. We may do so in order to save explanations such as the existence of an actual talking cat from being discarded. We shall see how the poetic approach can be put into effect, and some consequences of adopting it, in sections 4 and 5.

3.2 Retaining explanations

We can now consider would would be needed for an explanation to be retained.

It is not merely that an explanation should not have implications that would clash with existing knowledge (except to the extent that adoption of the poetic approach allows us to avoid this constraint). An explanation should also have considerable positive plausibility, even though (as Karl Popper pointed out) we never establish an explanation beyond all doubt. The best we can do is to test it severely and fail to falsify it.

Thus if hypnosis were the explanation under consideration, hypnosis in general would have to be shown to be able to be strong enough. Dreams would have to be able to be coherent and long-lasting enough to fool the dreamer into thinking that he or she was awake. And alcohol and hallucinogens would have to be able to have the required effects on people. Likewise, if one were to explain an apparent talking cat as a fake cat, it would not suffice that people could fit into catsuits or that robot cats could be built. Fooling people with a fake cat would have to be a plausible thing for someone to do, unless the need for plausibility fell by the wayside because the fake cat was captured and the creature was scanned or the suit removed to reveal what was underneath.

Turning to the explanation of a real talking cat, which only remains in play so long as one is being poetic rather than academic and can therefore avoid constraints imposed by existing knowledge, the locus of the test of plausibility shifts from the world considered academically to the content of some poetic work or to a poetic description of an episode in real life. Does the work or the description find a natural place for a talking cat? If it does, the explanation that the talking cat is real may be retained. Moreover, one should take no interest in adopting other explanations (altered states of mind or fake cats), although reference to them might form part of a work of fiction (as mass hypnosis features in The Master and Margarita, in chapter 12 and in the epilogue) or part of an account of an episode in real life (as in a statement that any such prosaic explanation would leave the account unable to capture the impact of the experience).

4. The poetic approach

When we take the poetic approach, we avoid some of the constraints imposed by our existing knowledge in order to make room for desired readings of stories or for appealing explanations of experiences. We may do this in order to immerse ourselves fully in a good story. In The Master and Margarita, Behemoth has to be an actual talking cat (disregarding his transformation in chapter 32). We may also do this in order to capture the impact of an experience in real life. It might be that only a prima facie bizarre explanation would suffice to capture the impact that the experience had.

4.1 What the poetic approach involves

Admitting the existence of an actual talking cat would be a radical step, plainly foolish from an academic point of view. It would mean overstepping the limits set by academic knowledge. It would indeed be a desire to overstep those limits that had driven us to the poetic approach. Poetry and dramatic fiction written within the bounds of academic knowledge could also be satisfying, but within those bounds we would not be allowed to speak literally of talking cats.

Strong reasons would be needed to make it legitimate to adopt the poetic approach in the context of explaining experiences in real life. (There is no such demand for strong reasons in the context of reading fiction, because it is clearly understood to be fiction.) We shall look at reasons shortly, but first we need a better idea of what the approach involves.

What goes on when we adopt the poetic approach in order to overstep limits set by academic knowledge? There are two broad possibilities.

The first possibility would be to modify our understanding of the world in general, scientific laws, general ways people behaved, and so on, in order to accommodate the views of the contents of stories, or the explanations of experiences, we desired.

There are two ways in which this first possibility might be put into effect. The first way would be to overlook specified inconvenient elements in our academic knowledge. The second way would be to transfer our entire imagination to an alternative world view, where whatever facts about the world in general there might be would be envisaged to be such as to permit whatever views of the contents of stories, or explanations of experiences, we desired. It is not that gaps would be put in knowledge, as in the first way. Rather, it would be supposed that the world in general was in a complete state that was as permissive as required, although the details of that state would be supposed to be whatever would do the job rather than their being specified.

The second way to put this first possibility into effect would be the more plausible way. When being poetic, we are not likely to specify general facts about the world that would be needed or would need to be ignored. Moreover, overlooking specified inconvenient elements in our academic knowledge would make us conscious that we were engaging in make-believe, pretending that there could be talking cats. Supposing an appropriate but unspecified complete state of the world in general, by contrast, could be done by transferring our entire imagination into the supposed world, losing most of our awareness that we were violating the norms of academic knowledge.

The second possibility would be to preserve our academic understanding of the world in general, but to avoid discomfort by failing to draw inferences from our desired views of the contents of stories or from our chosen explanations of experiences.

This second possibility might require less intellectual effort than supposing that the world was in an unspecified but suitably accommodating state. On the other hand, it would create the highly specific intellectual discomfort of knowing that one was closing one's eyes to constraints imposed by some important facts. That could be harder to bear than the more diffuse discomfort of knowing that one was supposing the world in general to be accommodating. Indeed, it would lead to a sensation of engaging in make-believe much like that which would be engendered by overlooking specific inconvenient parts of our academic knowledge.

For such reasons, we favour the second way in which the first possibility might be put into effect, that is, supposing the world in general to be in an appropriate but unspecified state, as a conception of how we are poetic beyond the bounds of academic knowledge.

4.2 Reasons to adopt the poetic approach

What kind of justification for the poetic approach could be given? How could someone legitimately say "I am being poetic, and in so doing am breaking free from the constraints of academic knowledge, but we should still take seriously what I say"?

If one were being poetic for its own sake, as for example in the writing or the reading of fiction, no further justification would be needed. Not all intellectual activity has to be in the pursuit of knowledge. And who knows, we might happen upon inspiration that would lead to academic knowledge. The human brain can progress to new ideas that have academic value by strange paths, and not only by inferential progression from a combination of existing knowledge and new empirical data in an academically respectable form.

As noted above, one might also need to be poetic in order to give an experience in real life an explanation that would represent the experience in such a way as to capture its full impact, even though the explanation would prima facie be bizarre.

Such justifications for being poetic would however not give that activity any status as an independent source of knowledge. Fiction is not rejected for failure to comply with real-world theories and data, but that is because it is recognised as fiction, not as fact. And a lively account of an experience must be stripped of those elements that depend for their legitimacy on use of the poetic approach before it can be regarded as the literal truth.

This difficulty in seeing the poetic approach as a direct route to the literal truth should not surprise us. Setting the mind free from potentially relevant knowledge is a bad start in the search for truth.

There is however another way for the poetic approach to be of cognitive value. It may offer a short cut to an appreciation of humanity that could only much more laboriously be obtained by academic means, and that might not even be open to being stated in academic terms. Fiction in particular can show us how people may feel and act in unusual situations, and thereby disclose to us the human condition and the workings of the human mind. And a lively description of an actual experience can show both the experiencer and his or her audience what is significant to people in such experiences and how people are inclined to respond to them.

The epistemic status of what was learnt would however be somewhat unclear. It would not be on a par with the status of academic knowledge, unless it were practical to make academic checks on its content. And the making of such checks would require statement of the content in a suitably academic form, when it might not be easy or even possible to state the content without reliance on metaphor.

Our discussion has become rather general. It is time to get back to Behemoth. In so doing, we may get a better idea of what sort of thing might be learnt through adopting the poetic approach.

5. Back to Behemoth

5.1 Teaching us about human beings

In The Master and Margarita, Behemoth is a supporting character in the team that exposes various human follies. The exposure of follies teaches a valuable lesson about human beings that survives our reversion to an academic approach. And while the story is made highly effective by being told in a way that happens to rely on the reader's acceptance that the possibilities of nature are not constrained by our academic knowledge, it could have been told without that reliance, using less engaging characters. So here we see the poetic approach allowing a highly effective route to academic knowledge of human folly.

There is also scope for a more specific benefit of adopting the poetic approach than the benefit of permitting engaging characters. Seeing a talking cat as real is a good way to flex the concept of humanity. The creature exhibits qualities, such as talking and then acting in accordance with what he says, that we normally regard as specific to human beings, while exhibiting other qualities, such as feline form, that we regard as never exhibited by human beings. To say that a human being is cat-like can be a way to convey information about that person. We attribute to him or her some of the qualities we attribute to cats, such as being demanding, cunning, or aloof. (Many such qualities were highlighted by the characters in the musical Cats and, less dramatically, in the inspiration for that musical, T. S. Eliot's collection of poems entitled Old Possum's Book of Practical Cats.) We might think of ourselves as merely drawing analogies between cats and human beings, but when Behemoth comes on the scene, bipedal and talking, the gap between cats and human beings that prompts us to think we only have analogies closes. We can directly see properties of cats as being attributable to some human beings. Behemoth is cat and human being in one, felinity yielding the properties and humanity allowing their direct inclusion in the range of potential properties of people.

We might then progress to some lateral thinking. "Just suppose that cats could talk and join in human society. Then what?" We would be prompted to think about whom to regard as a moral agent or a moral client, adjustments to laws and social security to accommodate the different needs of cats while preserving our own safety and security, whether cats should be allowed roles in political processes, and so on.

One lesson for us back in a world without talking cats would be that people are different in ways that may test the boundaries of legal and social arrangements, so that we must work out how to adjust and strike balances. We would already have known that much, but only in a theoretical way if we had so far lived in a homogeneous society. And it is back in the real world that we would test the lessons learnt, by contemplating the natures of actual people, the tensions that heterogeneity can create, and the need to respond to those tensions.

5.2 Experiences that change us

Sometimes, reading literature is a significant experience in itself. It can change us. It can be what in German would be called an Erfahrung. This concept emphasises the long-term change in the experiencer. It may be contrasted with the concept of an Erlebnis, a concept that emphasises simply the experience itself.

When we change our approach to one that allows for talking cats as part of the natural world, how we see and think about the world under that shift of approach can leave mental changes that endure after we shift back to an academic approach. We do not need to have in mind a specified alternative set of natural laws so as to work out that they would allow for talking cats. Rather, we need to go with the flow and welcome talking cats and other strange creatures as part of how life is. Rather than saying that certain laws of nature do not apply, we wholly imagine that nature is such as to make the creatures unsurprising.

In doing that, we both allow ourselves to contemplate alternative ways the world might have been, leading us to notice features of the actual world that we might otherwise have missed, and give new flexibility to our minds. We also leave ourselves wide open to believing nonsense far beyond what is in the text we are reading, but that is not a problem because we shall in due course revert to an academic approach under which the nonsense will be eliminated. Talking cats will be eliminated at the same time, but that is no loss because by then we will have had the Erfahrung, and its effects are what will stay with us.

Turning to real-life experiences, a description of one that is only legitimate under the poetic approach, such as a description that takes a talking cat to be real, can articulate how the experience was an Erfahrung. Giving such a description may indeed be instrumental in elevating an Erlebnis to an Erfahrung, as it leads the experiencer to appreciate the impact on him or her.

So we should not think of an Erfahrung of either the literary or the real-life kind as a direct source of new knowledge. Rather, we should think of it as developing mental capabilities, self-understanding, and an understanding of humanity. By experiencing an acceptance of talking cats, we can develop our ability to break free of the constraints of existing knowledge and thereby make new discoveries, albeit mostly of a non-academic kind. The sleep of academic reason can give rise to the most delightful and instructive monsters. We just need to make sure that academic reason reawakens after the experience.