Contents

Chapter 1: Interactionism..............................................................................................1

1.1 Introduction..........................................................................................................1

1.2 How Monism Wins Popularity..........................................................................7

1.2.1 Four Dualist Positions..................................................................................7

1.2.2 Causal Closure...............................................................................................9

1.2.3 Epiphenomenalism.....................................................................................12

1.3 Epiphenomenalism Is an Unattractive Position............................................14

1.4 Burden of Proof..................................................................................................17

1.5 Proportionality....................................................................................................19

Chapter 2: The Paradox of Justified Phenomenal Judgement...............................25

2.1 The Paradox........................................................................................................25

2.1.1 Interactionist Responses.............................................................................27

2.1.2 Chalmers’ Response....................................................................................30

2.1.3 Criticism of Chalmers’ Response..............................................................31

2.1.4 Interactionism and Certainty.....................................................................36

2.2 A Revised Account.............................................................................................42

2.2.1 Exposition.....................................................................................................42

2.2.2 Criticism........................................................................................................51

Chapter 3: Two Objections..........................................................................................61

3.1 Introduction........................................................................................................61

3.2 A Conceptual Problem......................................................................................61

3.2.1 The Argument..............................................................................................62

3.2.2 A Counter‐Argument.................................................................................63

3.2.3 Further Considerations...............................................................................64

3.3 Defining Interactionism.....................................................................................81

Chapter 4: Conservation of Energy...........................................................................87

4.1 Introduction........................................................................................................87

4.2 The Status of CoE...............................................................................................88

4.3 What Is a Closed System?.................................................................................90

4.4 The Relevance of CoE to Interactionist Dualism...........................................93

4.4.1 Purported Irrelevance.................................................................................94

4.4.2 The Prima Facie Relevance of CoE.............................................................95

4.4.3 Diagnosis......................................................................................................97

4.4.4 Relevance Reasserted................................................................................100

4.5 Taxonomy of the Defences..............................................................................103

4.5.1 A Tripartite Taxonomy.............................................................................104

4.5.2 An Alternative Taxonomy.......................................................................104

4.6 Popper’s Defence..............................................................................................105

4.6.1 Context........................................................................................................105

4.6.2 Interpretation.............................................................................................107

4.6.3 Evaluation..................................................................................................110

4.7 Compensation Defences..................................................................................114

4.7.1 Two ‘Levelling Out’ Views......................................................................114

4.8 Defences that Allege Question‐Begging.......................................................120

4.8.1 Question‐Begging I: Averill and Keating..............................................121

4.8.2 Question‐Begging II: Rejecting 1*...........................................................124

4.9 Conclusion.........................................................................................................127

Chapter 5: Other Arguments....................................................................................129

5.1 Introduction......................................................................................................129

5.2 Papineau............................................................................................................130

5.3 Meta‐Inductions on the History of Science..................................................132

5.3.1 The Arguments..........................................................................................132

5.3.2 The Response.............................................................................................133

5.3.3 Other Varieties...........................................................................................136

5.4 Unity of Science................................................................................................138

5.5 Lack of ‘Wiggle Room’....................................................................................145

Conclusion...................................................................................................................147

Appendix A.................................................................................................................151

Abstract....................................................................................................................151

A.1 Introduction.....................................................................................................151

A.2 Dualism by Degrees........................................................................................153

A.3 Units of Measurement....................................................................................157

A.4 Action at a Distance........................................................................................159

A.5 Who Were Right?............................................................................................161

A.6 Contrasts...........................................................................................................162

A.7 Present‐Day Physics.......................................................................................168

A.8 Avoiding Vacuity............................................................................................171

A.9 Concluding Remarks......................................................................................173

References....................................................................................................................175

Chapter 1: Interactionism

1.1 Introduction

Mind‐body dualism is, on the face of things, an attractive position. Before philosophical inquiry begins, we have on the one hand interacted with a multitude of different physical objects, and on the other enjoyed a rich ‘inner life’, abounding with thoughts, emotions, experiences and so forth. That the two seem very different is rarely disputed.1 What is disputed is how much weight this appearance should be given when we try to find out what the world is like.

Appearance of difference does not always track real difference in the world. Wind appears differently to us than do most physical objects of our experience. We cannot see wind, but we can see the effects it has, for sometimes leaves, branches, sand or fog move in the wind. We are used to seeing the causes of physical changes in our environments, so in this respect (and others) wind is unusual. We do not, however, take this to indicate that wind is deeply different from physical objects of our everyday experience. Wind is air that is moving. We know how air is composed, and although we cannot make accurate small‐scale predictions, we can with reasonable accuracy predict what the general direction and strength of wind will be on most days. Moreover, even in those few instances — such as tornadoes — where even large‐scale predictions elude us, there is widespread consensus that it is only the large number of variables and the complexity of their interactions that stop us. Wind, although invisible, is no longer deeply mysterious to us.

Mind‐body dualism therefore requires considerations to motivate the belief that the appearance of difference between mind and body runs deeper than the appearance of difference between wind and everyday physical objects such as rocks and books. There is, in other words, a need to show that we are now not merely in a position vis‐à‐vis the mind similar to that which people were in vis‐à‐vis wind, when wind was still mysterious to them.

1 Rarely, but sometimes. In ‘The Headless Woman Illusion and the Defence of Materialism’ Armstrong (1968) can be understood as suggesting that rather than being aware of apparent difference, we are unaware of apparent similarity, between the two. “It can now be suggested by the Materialist that we tend to pass from something that is true: ‘I am not introspectively aware that mental images are brain-processes’ to something that is false: ‘I am introspectively aware that mental images are not brain-processes’.” (pp. 48-49). I have added single quotation marks for clarity.

Mind‐body dualists are usually taken to believe that the difference between mind and body is deep, or even fundamental. One consideration that may underpin such a belief is the following. A pre‐scientific encounter with wind might have been characterised by a sense of a mysterious and unseen force that moves branches and leaves. That force has now been explained by, as it were, constitution; we have an account of what ‘constitutes’ the ‘force’ that moves the branches, and of why it is that we cannot see it. With the mind, however, many dualists hold that no analogous explanation could be forthcoming, because any attempt at giving an explanation in terms of (say) atoms and how they move would amount to changing the subject, in the strong sense of changing what we are talking about, not just how we talk about it.

We know a fair bit about the physical requirements for our inner life. The indications that our inner life is strongly connected with the brain are very clear; we can for example ‘turn off’ the conscious part of a person’s inner life by giving them a solid knock to the head. Evidence to show that damage to a specific part of the brain is likely to cause specific changes in the mental life also abounds. Yet, mind‐body dualists think that the appearance of difference is not compromised by this knowledge of the physical requirements for our inner life. Even though we know a lot about the working of the brain, we would be wrong, they claim, to say that that is what we are talking about when we report on our inner life. It is not as if we are inaccurate when we say e.g. ‘I feel sad’, and really should have stated something about the neural activity in the brain. What is referred to by expressions such as ‘I feel sad’ cannot turn out to refer to brain states, or so the dualist is said to think.

This report, ‘I feel sad’, is carefully chosen, for it is most common today to be a dualist about only a limited section of our inner life. ‘I feel sad’ is taken to report on a state within that section. What most dualists believe today is that there are some properties of experience that are not physical. There are many names for these properties, phenomenal experience is one, qualia is another, what it is like to be (someone who has a certain experience) is yet another. The thought is that there is something it is like to have an experience, there is a something it feels like to see a colour, hear a sound, smell something. This is something it is very hard to argue for. If you were inclined to deny that there really is something it is like to have an experience, there is not much I could do to convince you to change your mind. For the purposes of this thesis it is assumed that phenomenal experience really is a property of experience. That is not too controversial.

It will furthermore be assumed that this property of experience, the ‘what it is like’ to have that experience, is not physical. It will be assumed, in other words, that any attempt to replace reports like “I feel sad” with a report like “My brain is in such‐and‐such a state” has to fail; the two reports would not be co‐referential. Someone would have changed the subject.2 That is very controversial. In fact, a significant proportion of recent philosophy of mind has been occupied by argument over this point. Nevertheless, I think it is worthwhile to make that assumption, and I will try to explain why.

In philosophy of mind monism — the view that mind and body are fundamentally similar — is the received view. Jackson and Braddon‐Mitchell exaggerate when they claim that “dualism is almost as unpopular as idealism” (1996, p. 4); a publication such as Chalmers’ The Conscious Mind is either a symptom or a cause of a resurgence in the popularity of dualism; most likely a bit of both. Furthermore, even among monists it is surely true that dualism, in one form or another, has some pull on them some of the time. The intuition that mind and body are very different from one another, perhaps fundamentally different, is ubiquitous and compelling. So dualism is not nearly as unpopular as idealism. However, though it does not rule the scene as unopposed as Jackson and Braddon‐Mitchell claim, they are right in saying that monism is orthodoxy.

2 Again, this is meant in a strong sense. On some accounts the use of a different predicate guarantees in itself that a different property is being predicated. On such accounts it comes out as trivially true that changing the ‘mode of referring’ from “I feel sad” to “My brain is in such-and-such a state” changes the predicate. I assume that changing the sense does not necessarily suffice to change the reference, and thus that the claim of having changed the subject in this example is non-trivial.

Against this backdrop the decision to assume that dualism is true is perhaps surprising. Less so, hopefully, once it is realised that the dominance of the monist position does not rest solely on the merits of that position itself. Rather, it depends, to a not inconsiderable extent, on the elimination of dualist positions as unviable. Many philosophers believe that no credible dualist position is available. If none were, it would be unreasonable to blame someone for retreating to a (presumably) less problematic monist position instead. The purpose of this thesis is not to attack the merits of monist standpoints in the philosophy of mind, but rather to question whether the dismissal of a particular dualist position — interactionism — can be justified in the way that it is widely supposed that it can.

Assuming the truth of dualism for the purposes of this thesis is justified, because the widespread belief that interactionism is unviable rests to a very large extent on arguments that are inapplicable to other dualist positions. The assumption does not, therefore, lend interactionism any unfair dialectical advantage. It serves merely to focus the discussion in the following way: it opens for an evaluation of whether interactionism can be defended from the attacks that it is actually subjected to. It remains the case, of course, that any general or principled criticism that can be levelled against any dualist position applies just as forcefully to interactionism. This is, however, as a matter of fact not how the criticism of interactionism usually proceeds. Interactionism is criticised for claims that are specific to that position, and crossed off the list over available dualist alternatives. Only when other dualist alternatives are crossed off as well is the conclusion reached that monism must be true. Interactionism is, in other words, to a very large extent criticised qua interactionism, and not qua dualism. Assuming that dualism is true for the purposes of evaluating the viability of interactionism takes this fact seriously.

In addition to leaving any general or principled criticism of interactionism qua dualist position intact, defending interactionism from an argument widely used to discredit it does not, in itself, contribute to motivate belief in the position. A comprehensive argument for interactionism would contain either arguments that lead directly to interactionism (if such could be devised) or arguments that lead first to dualism and then, in the second step, to interactionism as the best dualist position. The present thesis comes closer to employing the second of these strategies, since epiphenomenalism — an alternative dualist position — is criticised in some detail herein. But the other half of the strategy is conspicuous in its absence; I make no independent attempt to motivate dualism here.

Interesting questions can be raised about the viability of the strategy of motivating interactionist dualism in two steps. It is interesting, in particular, to ask whether the arguments that these days are used to motivate dualism could constitute the first step of such a two step strategy.3 The question is whether these arguments would survive the move to interactionism in the second step; there may be some reason to doubt that they would. For example, one might doubt if the knowledge argument could go through exactly as before.

That argument — the important features of which were present already in Dunne’s An Experiment with Time (1927/1958), but which was since put into its oft‐cited form by Jackson (1982) — describes a brilliant scientist who is confined to a monochrome environment, but who nevertheless studies, among other things, colour perception. The intuition is elicited that when she escapes her confinement she learns something new, namely what it is like to have a colour experience, and that this is something she could not have known beforehand. However, if phenomenal experience is causally efficacious, as interactionism holds that it is, then Mary would know about phenomenal experiences while she was still confined, at least in the sense that she could ‘pick them out’. She would be able to refer to them using unique identity conditions, for she would know their place in the causal web of things. She could not pick them out in the same way that most others pick out colour experiences, of course, and there is still a powerful intuition that there would be something she would not know. How important that something would be, and whether it would be enough for the argument to go through, is uncertain. Perhaps similar worries could be raised about other arguments for dualism.

3 I am grateful to Leon Leontyev for relentlessly driving this point home to me.

I set this question to one side, however, for whether a two‐step argument for interactionist dualism is possible is tangential to the current project. The aim of this thesis is not to complete the ambitious project of building a comprehensive case in favour of interactionist dualism. For the purposes of this thesis it is more than sufficient to note that a two‐step argument for interactionist dualism does not obviously fail, and that interactionism in any case has considerable intuitive attractiveness.

The aims of this thesis are quite modest. The overall aim of the project is just to show that interactionist dualism is a ‘live option’ in the philosophy of mind, one that deserves more attention and investigation than it is, at present, getting. First it is argued that epiphenomenalism — the currently dominating dualist position — lacks important and attractive features that interactionism has. This motivates the claim that if both positions are viable we have good reason to accept interactionism over epiphenomenalism. Secondly I attempt to show that interactionism is a viable position by showing that the objections most commonly raised against the position hold little force.

In conjunction, these two steps go some way toward showing that interactionist dualism is the most attractive dualist position. Since the starting point is the assumption that dualism is true they can obviously at most contribute toward half the job of showing that interactionist dualism is the most attractive position in the philosophy of mind overall. However, if the claim that monism wins a not inconsiderable proportion of its support by means of elimination of dualist alternatives should be accepted, it is hoped that the merits of showing that one such alternative is not unviable (or at least not unviable for the reasons most commonly given) will be seen as well. If interactionist dualism comes to be seen as a real alternative position, it seems that this may have the potential to impact the debate in philosophy of mind quite significantly.

1.2 How Monism Wins Popularity

It is not easy to argue directly for the claim that the popularity of monism is owed largely to a lack of alternatives, without taking an opinion poll. A good case can be made for the claim in indirect ways, however, for the process that leads to the ‘crossing out’ of one dualist position after another — and thus eventually to monism — is not too hard to reconstruct.

1.2.1 FOUR DUALIST POSITIONS

In respect to causality, the logical space governing the relation between the mental and the physical contains but four types of dualist positions: 4

i. There are no causal relations between the mental and the physical.

ii. There are causal relations between the mental and the physical such that mental states of affairs bring about physical states of affairs.

iii. There are causal relations between the mental and the physical such that physical states of affairs bring about mental states of affairs.

iv. There are causal relations between the mental and the physical going both ways. That is, there are causal relations between the mental and the physical such that mental states of affairs bring about physical states of affairs, and such that physical states of affairs bring about mental states of affairs.

4 This thesis does not discuss what Chalmers (2002) calls ‘Type-F monism’. Aside from the standard excuse that one cannot do everything at once, the most important reason for this stems from the objection that the position seems to fail to cater for the core intuition that the mind makes a difference to the physical in a direct sense. It seems that the ‘subtle’ causal relevance that Type-F monism (1996, pp. 153-56) allots to phenomenal properties is too subtle to account for our intuitions. Chalmers notes the problem: “It remains the case that natural supervenience feels epiphenomenalistic” (p. 156).

The first of these positions shoulders a very significant explanatory burden, namely that of explaining why it constantly seems to us that the there are causal relations between the mental and physical. This position was famously defended by Leibniz. He argued that God had established a harmony between the mental and the physical from the outset, so that the two domains run in parallel. Whenever there is a wish to, say, raise one’s arm, there is also, according to parallelism, operations of the mechanisms in the physical world such that the arm is indeed raised. All of this takes place without any interaction between the two domains. It is safe to say that pre‐established harmony strikes most contemporary philosophers as a very implausible doctrine, and I know of no remaining defenders of that or any other form of parallelism.

The second position is not to be confused with idealism. Idealism is a form of monism, but unlike the materialist or physicalist monism now popular, idealism holds that everything is mental. In contrast, the position under consideration is a dualist position, one that holds that physical states of (say) human bodies are brought about by mental states, but never the other way around. The position should further be differentiated from occasionalism, one form of which holds that all states of affairs are brought about by God through continuing creation. Occasionalism is similar to the position presently under consideration in that causation between mind and matter is limited to the one direction from mind to matter. However the view under consideration is not a view about God, but about finite creatures that have both bodies and minds, and the causal relations between those.

This position has to my knowledge never been defended, although it appears to be a possible stance. One might hold that the (immaterial) soul causes the body to grow and develop, and that the soul controls the body throughout the life of a person.

Considerations mentioned above can be brought to bear against this position. There is abundant prima facie evidence for the existence of causal relations from the physical to the mental in our everyday life, so it is very hard to believe that causal relations should be limited to the direction from the mental to the physical. It seems safe to conclude that neither parallelism nor the position just discussed are viable dualist positions. They can safely be set to one side.

1.2.2 CAUSAL CLOSURE

After this first brief round of exclusions we are left with only the last two dualist alternatives. The fourth position is interactionist dualism. The minimal interactionist position is that the mind either sometimes brings about a physical event that would not otherwise have occurred or brings it about that an event occurs differently than it otherwise would have. A single, very influential objection is generally taken to put this alternative out of the running for good. The objection relies on what is known as the principle of causal closure of the physical world. Even though we are far from completing the scientific project, we know, it is claimed, enough about the world to know that all physical phenomena are fully explainable in terms of other physical phenomena; they have only physical causes. That the principle of causal closure does not rely on determinism is widely recognised; 5 allowing for indeterminacy you might say that inasmuch as a physical event can be said to be an effect, inasmuch as the event has causes, all those causes are physical.

It is not hard to see how the principle of causal closure leads to the denial of interactionism. There are two domains, one is physical and the other is not. The principle says that the physical domain is causally closed; that nothing outside of that domain has any causal influence on anything within that domain. Then it follows straightforwardly that the other domain cannot have any causal efficacy on the physical domain. That some aspect of the mental has such causal efficacy is precisely the minimal claim an interactionist must make, so the principle of causal closure leads to the rejection of interactionism.

The belief that the objection from causal closure is fatal to interactionism has been extremely influential. Thus, for example, Papineau argues that “[a]ll physical effects are fully caused by purely physical prior histories” (2002, p. 17), Jackson and Braddon‐Mitchell state that “[a]ll the evidence suggests that there is nothing non‐physical ... that steps in and somehow regulates the physical world’s goings‐on” (1996, p. 9), Chalmers writes that “[t]he best evidence of contemporary science tells us that the physical world is more or less causally closed: for every physical event, there is a physical sufficient cause” (1996, p. 125) and Lewis endorses what he calls the “materialistic working hypothesis”, namely that “physical phenomena have none but purely physical explanations” (1966, p. 17). With the exclusion of Chalmers, who has since changed his mind (2002, n. 26), these theorists believe that interactionism is excluded by causal closure.

It is worth noting that there is a significant difference between the formulations just mentioned. Lewis’ and Jackson and Braddon‐Mitchell’s formulations expressly exclude non‐physical causes for physical phenomena, there are no entities such that they are both causes of physical phenomena and themselves non‐physical. Chalmers’ formulation here does not; it states merely that for all physical effects there are sufficient causes that themselves are physical, and Papineau’s formulation is also restricted to an existential claim.

5 E.g. by Papineau (2002, p. 17, n. 2), Jackson and Braddon-Mitchell (1996, p. 13) and Chalmers (1996, p. 150).

Formulations that assert the existence of sufficient physical causes for all physical effects, rather than deny the existence of non‐physical causes for physical effects, open the door to the thesis of causal overdetermination. This is the thesis that some events, and in particular those which we tend to believe have mental causes, may have two sets of causes, each in itself sufficient to bring about the effect. If that were so it looks like interactionism might not be incompatible with the causal closure thesis defined as an existential claim, for the assertion that there are sufficient physical causes for all physical effects is compatible with there being sufficient non‐physical causes for some physical events, as well.

It will be argued in this thesis that it is important to direct close attention to the investigation of interactionism as an alternative position on the mind‐body issue, and in particular to the scrutiny of the causal closure thesis. It may seem to follow that further research into causal overdetermination should also be recommended. That is not the recommendation, for two reasons. First, while the investigation of causal overdetermination is not a large research project it is fair to say that it has received much more attention (recently) than the project of investigating the viability of interactionism. But even more importantly, causal overdetermination does not seem to deliver enough ‘oomph’ to the mental for it to account for our intuitions about the causal efficacy of phenomenal experience. It does not offer a significant improvement over epiphenomenalism. As Robinson points out, the central intuition is “that the mental makes a difference to the physical, i.e. that it leads to behaviour that would not have happened in absence of the mental” (2003), and that intuition is not appeased by causal overdetermination. Consequently, in what follows, all mention of the causal closure thesis is to be understood as making reference to a thesis that is incompatible with causal overdetermination, to a thesis, that is, that expressly denies the existence of non‐physical causes for physical events.

1.2.3 EPIPHENOMENALISM

The exclusion of interactionism on the grounds that it conflicts with the principle of causal closure thus leaves only one dualist position: epiphenomenalism. According to epiphenomenalism the phenomenal properties of experience — what it feels like to have an experience — have no causal efficacy upon our actions (or upon anything else that is physical). On this view, our non‐physical mental life arises from the physical material in the brain, and the changes and events that take place there, so there are causal relations going from the physical to the mental. The epiphenomenalist holds, however, that it is those physical changes and events themselves that account for all our bodily behaviour; they do not countenance causal relations going from the mental to the physical. Epiphenomenalism has sometimes been marketed as a ‘safe’ way to be a dualist. An epiphenomenalist do not have to engage in bothersome speculations about how something non‐physical can cause a change in something physical — speculations that lead Descartes to the much ridiculed conclusion that the ‘cross over’ point is in the pituitary gland.6 One can be a dualist without having to sound like someone who believes in fairies, as Jackson once put it (1982, p. 128).

In this thesis it will be argued that epiphenomenalism is an extremely unattractive position. If that is right it appears that the dualist has no escape; there is no attractive and tenable position left. In reference to the point of how monism wins popularity it may be noted that even if it is not true that there is no tenable and attractive dualist position left for a dualist to hold, the widespread belief that this is so goes a very long way toward explaining why monism dominates to the extent that it does. Or so I claim.

6 In The Passions of the Soul he wrote about the pituitary gland: “[T]he activity of the soul consists entirely in the fact that simply by willing something it brings it about that the little gland to which it is closely joined moves in the manner required to produce the effect corresponding to this volition” (1649/1985, p. 343).

Given first and foremost that dualism enjoys intuitive plausibility, and secondly that it has enjoyed renewed popularity of late, considerations that leave the dualist with no tenable position serve to highlight the importance of two projects of inquiry. First, it naturally becomes very important to investigate whether or not epiphenomenalism is as untenable as it may seem. Perhaps the intuitions against epiphenomenalism cannot be translated into rigorous arguments against that position. If so, perhaps the mere fact that epiphenomenalism is a counterintuitive position should not lead us to abandon it. This project has already received significant attention. When there appears to be no tenable dualist alternatives left, that situation should also, however, alert us to a second important project for further research, one that has received far less attention than it deserves.

That project is, of course, the project of determining whether interactionism really is an untenable position for the dualist to hold. Do we really have good reasons to conclude that interactionism is incompatible with current scientific knowledge? Are there other solid arguments against the position? Above it was claimed that, as a matter of fact, the most influential objection against interactionism has been that it is incompatible with the principle of causal closure. If that is true it follows that an important project is to investigate whether or not we have good reason to believe that the principle of causal closure is true. Or, perhaps even more to the point, it should be investigated whether we have as good reason to believe in the principle as the majority of contemporary philosophers of mind seem to believe or assume that we have.

1.3 Epiphenomenalism Is an Unattractive Position

The importance of investigating the viability of interactionism — and thus also the importance of investigating the credibility of the principle of causal closure — depends on the claim made above, namely that epiphenomenalism is an unattractive position. The most accessible reason to reject epiphenomenalism is also the most compelling. It is simply that epiphenomenalism is completely at odds with what most people are convinced is going on all the time in their everyday lives; it clashes thoroughly with our commonsense view of the world.

How so? If you were to ask someone (perhaps, to be safe, someone unpolluted by philosophy) why they did something — why they did a bungee jump or jumped out of a plane with a parachute or had a piece of cake or worked out or got their hair done or whatever — then their eventual answer would, in a very large proportion of the cases, be that they did it because it, or something it facilitates, feels good somehow. It feels good to have a big adrenalin rush, chocolate tastes good, it feels good to do physical training (or at least to have done it), it feels good to be satisfied with your appearance. Similarly with avoidance behaviour; to ask for an explanation of why we avoid humiliating experiences amounts to admitting that you do not understand what ‘humiliating’ means. It feels patently bad to be humiliated, so we try to avoid it.7

This is not to say, of course, that the motivation for all our behaviour can be explained this simply. One of the marks of maturity is the ability to forego an immediate or very proximate desire‐satisfaction in favour of one that is a more distant but greater. This ability obviously affects our behaviour in profound ways. We keep working, for example, even when our immediate preference is to go to the beach, and we sometimes refrain from eating tasty foods in order to maintain our health. Nearly as obviously, this ability affects the way we speak. Envisage the following (rather caricatured) conversation with a young person:

7 There are exceptions to this, stemming inter alia from certain sexual fetishes. It is no less likely in those cases, however, that phenomenal experience — a certain very arousing phenomenal experience, probably — plays a crucial role in motivating the (unusual) behaviour of seeking out humiliating experiences.

You: I am sorry, we cannot start that game now; I have to leave very soon.

Child: Why?

You: Because I have to be at the university by one o’clock.

Child: Why?

You: Because I want to go to a lecture then.

Child: Why?

You: Because it is important to be there.

Child: Why?

You: Because it will help me get a good grade in this subject.

Child: So?

You: I want a good grade in this subject.

Child: Why?

You: Because it will help me get a good job.

Child: So?

You: Getting a good job is important.

Child: Why?

You: A good job will help me care for my family.

Child: So?

You: Well, if I cannot care for my family, I will be very sad, and they will be very sad too, because they will not get the things they need.

Child: Oh, OK.

Here the account of the motivation is gradually being ‘pushed backwards’, until finally a point is reached where the interlocutor understands why; a motivational factor the child shares has been reached. Speaking with young people is a good thing, and speaking with people who hold very different opinions is not bad either. When we speak with people who share our ambitions and values we are not usually challenged to justify our actions, and this can lead to complacency about the value of what we are doing. Sometimes the ‘unpacking’ of motivations is a complicated and time consuming process. That, however, does not change the fact that at the bottom of very nearly all motivation is phenomenal experience.8

Thus interactionist dualism is arguably very well aligned with common sense; both hold that phenomenal experience plays a causal role upon our behaviour. When it is taken into account how successful a theory commonsense is, this weighs very heavily in favour of interactionist dualism.9

That phenomenal experience has causal efficacy on our behaviour is precisely what epiphenomenalism denies. According to epiphenomenalism, the fact that the chocolate tastes good and the fact that it feels bad to be scared or humiliated has no causal influence on your behaviour. In particular, it does not explain why you buy chocolate or avoid heights and belittling people. That is incredible. Inasmuch as epiphenomenalist dualism is committed to this claim, it is an incredible position, too.10

One standard reply from the epiphenomenalist to this challenge needs to be mentioned briefly. According to this reply certain brain states have dual effect; they cause both a phenomenal experience and an action (the latter of which may be stipulated to occur further downstream, after the intermediate occurrence of further brain states, neural impulses, muscular contractions and so on). Because the phenomenal experience is reliably correlated with the kind of brain state that causes physical action downstream, it is not surprising that we should come to believe that phenomenal experience is causally efficacious on our actions. Nevertheless, so the argument goes, it is fully plausible that the causal relation is an illusion.11

8 Of course, we sometimes do things for others. In those cases perhaps we are motivated by the phenomenal experience we intend to elicit in our beneficiaries, or perhaps we are still motivated by a phenomenal experience in ourselves (the latter claim tempts some to deny the existence of true altruism).

9 For eloquent praise of the “extraordinary predictive power” of commonsense psychology, see Psychosemantics (Fodor 1987, pp. 1-10).

10 Similar objections to epiphenomenalism are widespread in the literature. Kneale argues that “the great paradox of epiphenomenalism” is “the suggestion that we are necessarily mistaken in all our ordinary thought about human action” (1959, p. 453); and that “the proposition ... that mental events are sometimes causes of physical events, is one which belongs to the hard core of common sense” (p. 454); Taylor characterises the thesis that “all bodily behavior is caused by bodily processes alone” as “quite impossible to believe” (1963, p. 23); Shaffer argues that “[i]nsofar as Epiphenomenalism asserts that mental events are never causes, it does seem to be in flat contradiction to our ordinary descriptions of familiar experiences” (1965, p. 101); Lewis argues that if the phenomenal difference between two tastes is epiphenomenal, that “makes it very queer, and repugnant to good sense” (1988/1999, p. 285); and Jackson and Braddon-Mitchell argue that, not only do we have good reason to believe that phenomenal experience, as a matter of fact, causally influences our behaviour, but a propensity to cause certain actions is part and parcel of many of our concepts for phenomenal experience (1996, p. 6), a thought which Lewis also entertains (1995, p. 141). (Kneale, Taylor and Shaffer are expressing opposition to epiphenomenalism as a thesis about a broader range of mental states than the thesis I criticise in the text advocates.)

The persuasive force of this argument varies very significantly. Some philosophers seem willing to accept the conclusion that phenomenal experience only seems to matter for action, and that this appearance is illusion. It does not, however, seem at all plausible to me that the causal connection I perceive between the certain experiences and my subsequent behaviour is illusory. Furthermore, it does not seem any more reasonable to suppose that the actions of the people I relate to in my daily life are not similarly influenced by their phenomenal experience than it does to suppose that they have no phenomenal experience at all. My belief that there is such a connection is, one might say, as entrenched as my belief in other minds.

For the present purposes, however, whether a solid argument can be mounted to replace these vague musings is not quite to the point. What matters more is that the vast majority of the ‘folk’ — as well as a fair few philosophers — are entirely convinced that the causal connection is real and not illusory. That does not mean that that belief cannot be mistaken. It does mean, however, that denying that belief is a significant theoretical burden, and one that certainly counts heavily against the epiphenomenalist position.

1.4 Burden of Proof

This leads to the question of what the burden of proof for a theoretical position about the mind‐body issue should be. In the current dialectic climate one might easily get the impression that interactionist dualism is so outlandish and crazy that someone wishing to raise the mere possibility that it might be true must present arguments truly out of the ordinary. One cannot help but feeling, sometimes, that this impression is deliberately cultivated. For example, the title of Jackson’s review article of The Self and Its Brain (Popper and Eccles 1977) was ‘Interactionism Revived?’ (1980), perhaps a somewhat tendentious choice of words. Similarly, to say that “dualism is almost as unpopular as idealism” and that “[t]o many, dualism is as discredited as vialism” pulls in the same direction, this time for dualism as a whole (Jackson and Braddon‐Mitchell 1996, p. 4).

11 See e.g. ‘Epiphenomenal Qualia’ (Jackson 1982, p. 133), The Conscious Mind (Chalmers 1996, pp. 181-82), ‘Epiphenomenalism’ (Robinson 2003).

It would of course be silly to think that anything but an honest appraisal of the standing of the debate is being expressed here. Nevertheless, expressions such as these do make it appear that a Herculean argumentative effort is needed to raise the possibility that dualism is a live option for contemporary philosophers, and all the more so for interactionist dualism. To borrow a now common legalese expression, one gets the impression that dualism, to be considered, would have to be supported by arguments that establish the case ‘beyond reasonable doubt’, and further still in the case of interactionist dualism.

That, however, is an unreasonable standard, one we should not impose. Sticking to legalese we might say that evidence that proves the case on the balance of probabilities is all we can reasonably demand. We should expect a case for the claim that dualism in general — or interactionist dualism in particular — is more likely to be true than are its rival theories, and nothing more than that. That this is the only reasonable standard to impose is a truth that has been somewhat obscured by aspects of the current climate in philosophy of mind, but it is a truth, nevertheless.

Against the backdrop assumption that dualism is true, the considerations given above quickly show that the main rival theory for interactionism is epiphenomenalist dualism. One argument against that position has already been offered. Another will discussed in the next chapter: the argument from the paradox of justified phenomenal judgement.

It is not here claimed that these two arguments constitute a knock down case against epiphenomenalism. However, if, as I have argued, a proof on the balance of probabilities is what we should be looking for, the two arguments against interactionism jointly go a long way toward making the case that the viability of interactionist dualism needs to be more carefully investigated. To make interactionism appear more likely than epiphenomenalism, it is not necessary to show that epiphenomenalism cannot be true. It may be sufficient to show that it shoulders very significant theoretical burdens. It has been argued that epiphenomenalism runs contrary to widely held and deeply entrenched intuitions about motivation, and that is a significant theoretical burden. Added to that is the slightly more technical argument in the next chapter to the effect that epiphenomenalism cannot satisfactorily account for the justification of beliefs and judgements about phenomenal experience. The hope is that together they will at least pique an interest in interactionism.

Piquing an interest in interactionism is necessary in the current climate, but it is not sufficient. If it were true, as is widely held, that interactionist dualism is flatly disproved by current scientific knowledge then the position would hardly merit a second look. The latter two chapters of this thesis argue that this is not the case. In chapter four it is argued that interactionist dualism is compatible with the conservation of energy law in physics and in chapter five interactionism is defended against some common, less convincing, arguments.

1.5 Proportionality

In the prevailing climate in recent analytic philosophy of mind, interactionism has been considered dead and gone; in need of ‘resuscitation’ by arguments beyond reasonable doubt. Investigation of which position is more likely on the balance of probabilities is, in contrast, a context conducive to the close appraisal of some arguments that have thus far largely escaped detailed scrutiny. This thesis contends that the argument against interactionism from the causal closure of the physical, in particular, stands in need of much closer scrutiny than it has thus far been subjected to. If causal overdetermination is excluded, the validity of the argument is rarely doubted: if the physical is causally closed, then the mental is either itself physical, contrary to the dualist part of interactionist dualism, or it is causally inefficacious, contrary to the interactionist part. What is required is further investigation into the truth‐value of the crucial premise: that the physical is causally closed.

Consider a simplistic analysis of the circumstances under which one might recommend that a thesis be accepted for the sake of argument. Intuitively, there are two conditions under which that can do little harm, namely when either (i) not much hinges on the acceptance of the thesis, or (ii) we are superbly confident that the thesis is true. If either of these conditions is satisfied to a very high degree it is reasonable to ask that a thesis be accepted for the sake of argument.

It is rarely the case that we know with near‐certainty that a thesis we are being asked to accept is true, and it is, of course, also rare that we are asked to accept a thesis on whose acceptance nothing of importance hinges. That does not mean that we are almost never warranted in accepting theses. What we require is a suitable mixture of the two conditions, and we have that if the degree of confidence in the thesis is proportionate to how much hinges on its being true or false. Call this simplistic analysis the proportionality constraint on accepting a thesis.

This work contends that acceptance the thesis of causal closure of the physical violates the proportionality constraint, because the importance of what hinges on its acceptance is disproportionate to the reasons we have for accepting it. We should not yet accept that thesis of causal closure is true.

Above it was argued that monism wins a not inconsiderable proportion of its support through elimination of the available dualist alternatives. Then an objection to epiphenomenalism was presented, alleging that epiphenomenalism is wildly at odds with widespread and firmly entrenched intuitions about the causal efficacy of phenomenal experiences. Conjoining that argument with the argument in the next chapter yields a solid case for the claim that epiphenomenalism is an implausible position. If that is right — and if the brief considerations above against positions of type i. and ii. are effective — interactionism emerges as the only plausible dualist position. Since the thesis of causal closure leads to what is widely taken to be a decisive argument against interactionist dualism, accepting the thesis of causal closure will force a decision on the question of whether to accept dualism or monism, for once interactionism is excluded, no plausible dualist position is left. In this context it is safe to say that a lot hinges on the acceptance of the thesis of causal closure.

We might still respect the proportionality constraint if we have reason to be supremely confident that the thesis of causal closure is true. That, no doubt, is what a number of contemporary philosophers believe to be the case. However, as the last two chapters of this thesis show, the arguments in favour of the thesis of causal closure fail to adequately support belief in that thesis.

All of this is not to say, of course, that we should stop considering what would be the case if the thesis were true; working out the consequences of physicalism is as important as before. It would be well to remember, however, that the results are conditional. It would be well, too, to consider the consequences of the thesis being false. The arguments in favour of the thesis of causal closure fall well short of what would be required to justify allowing the acceptance of the thesis of causal closure to, as it were, ‘become stable’. That is, to an extent, what has happened; recent analytic philosophy of mind in the English‐speaking world has been dominated by the view that interactionism is dead and gone. This thesis aims to show that that is a conclusion which, by that tradition’s own standards, lacks adequate support. Analytic philosophy of mind would profit from accepting interactionism as what it is: a live option, a reasonable position.

Before seeing interactionism as such, there is no reason to demand a demonstration that the thesis of causal closure is not at all supported by the arguments. Once it is accepted that much hinges on the acceptance of the causal closure thesis it becomes clear that, to justify discounting interactionism to the degree which that has been done, what is required is not just some support for the causal closure thesis, but very strong support. This thesis attempts to show that that kind of support is not forthcoming.

The causal closure thesis may still play legitimate roles in philosophical arguments, for example in arguments for the truth of physicalism. It cannot, however, legitimately play such roles without arguments justifying belief in the thesis itself, or with only brief and superficial arguments. What is suggested here is not that the thesis of causal closure should not be accepted if a solid argument is put forth in its favour, but that it should not be accepted in the absence of such an argument.

Regardless of the merits of the simplistic analysis of when one should accept a thesis, it is certainly a striking fact that arguments for belief in the causal closure thesis are very rarely presented. Of the philosophers mentioned above — Jackson and Braddon‐Mitchell, Chalmers, Lewis and Papineau — only the latter presents much of an argument for the causal closure thesis, and yet their arguments all explicitly rely on the thesis. Moreover, it is reasonable to think, with Papineau, that many materialist arguments that do not mention the thesis explicitly nevertheless depend implicitly on the truth of the thesis (2002, pp. 233‐34). If the arguments in chapter four and five of this thesis are successful, this predicts difficulty for anyone wishing to exclude interactionism on the basis of an argument from the thesis of causal closure. It does not, of course, close the issue. To close the issue is, however, not the aim of the current thesis. The aim of the current thesis is to open it up.

Chapter 2: The Paradox of Justified Phenomenal Judgement

2.1 The Paradox

The state of affairs a judgement is about usually figures in an explanation of what does, or does not, justify that judgement. Were I to make the judgement that the sky is overcast today, a part of the explanation of why this is a justified judgement would be played by the meteorological state of affairs in the region where I am. It seems obvious that this is not an accident, and that the state of affairs the judgement is about is not merely a useful pedagogical tool. Being appropriately related to a state of affairs that is the way the judgement claims that it is seems certain to be what makes the judgement justified. This does not, I claim, come under threat from any lack of knowledge about the precise nature of the appropriate relation, nor from lack of clarity about what it means for a judgement to claim that the world is a certain way.

Sometimes we make judgements about our phenomenal experience. One might say (or say to oneself, or write, or sing, or ...) that one’s visual experience is dominated by a certain colour, or that one is having a particularly pleasant taste‐experience at that time. Most of the judgements thus expressed are no doubt justified.1 It will be argued below, however, that not only might you be wrong about your phenomenal experiences and thus make erroneous judgements about them but that you are furthermore capable of making unjustified judgements about your own occurrent phenomenal experience. This is by no means uncontroversial but the examples of unjustified phenomenal judgements I give below will hopefully go some way toward defending the claim. (Indeed, that we can be aware of properties of experiences as such is also sometimes disputed, but I do not defend that claim here.) Given unjustified phenomenal judgements, an explanation of what distinguishes justified judgements from judgements that are not justified is required.

1 There is an implicit restriction to judgements about intrinsic properties of experiences. If that restriction is not imposed, there is no difficulty in supposing either that judgements about phenomenal experiences can be erroneous or that they can be unjustified.

According to epiphenomenalist dualism, judgements are physical events.2 Additionally they often cause other physical events downstream, for example when the judgement is expressed. Epiphenomenalism holds that phenomenal experience forms no part of the causal history of any physical event. One natural way of cashing out what it is to be ‘appropriately related’ to a state of affairs — of explaining in virtue of what most of our judgements about phenomenal experience are justified — is therefore unavailable to epiphenomenalists: they cannot give a causal theory of justification. On their account, there is no causal link between the experiences and the judgements about those experiences.

Whether epiphenomenalism is forced to accept that the phenomenal experience judgements are about has nothing to do with the judgements’ status as justified is a further question. On the face of things, however, that seems to be a likely outcome, for once a causal link is excluded it is not clear what role the phenomenal experience a judgement is about could play in accounting for the justification of a judgement about it. Thus we have, at least prima facie, a paradoxical situation, in which the intentional object of a judgement is unrelated to the status of that judgement as justified or not. One might call that result the paradox of justified phenomenal judgement.3

2 See e.g. Chalmers (1996, pp. 173-74).

3 Chalmers gives vivid expression to a closely related paradox in The Conscious Mind (1996, chapter 5). The paradox Chalmers discusses in his book is couched in terms of explanatory irrelevance: given that judgements are taken to be reducible events, there should be a purely physical explanation of how they come to occur. But phenomenal experience, which is not reducible, will not figure in that explanation, so phenomenal experience will be explanatorily irrelevant to judgements about phenomenal experience (p. 177). Thus the discussion here is pitched in terms of a slightly different paradox than that which Chalmers discusses. This is innocent, however, for although Chalmers does not formulate the paradox of justified phenomenal judgements, he well could have, for he explicitly recognises the problem of how judgements about experience can be justified in the absence of a causal theory of justification (p. 193). I take his attempted solution to that problem as proposed solutions to what I have called the paradox of justified phenomenal judgement, and perpetrate no injustice that I can see in so doing. (See also Kneale (1959, p. 454), Shaffer (1965, p. 100 ff.), Elitzur (1989, pp. 8-9) and Penrose (1987, p. 116; 1989, pp. 408-09) for earlier formulations of closely related paradoxical results.)

One way of making the paradox vivid is by imagining a zombie twin world. There, everything is just like it is in our world, except that all the humanoid inhabitants of that world are phenomenal zombies: they lack phenomenal experience altogether. Any inhabitant of the zombie twin world is ex hypothesi just as likely as its twin on our world to make judgements about phenomenal experience. The difference is, of course, that in our world the vast majority of our phenomenal judgements are justified, but in the zombie world a justified phenomenal judgement is almost never made.4 How do we account for the difference?

2.1.1 INTERACTIONIST RESPONSES

The paradox of justified phenomenal judgement does not arise for an interactionist dualist account of the mind. This is not surprising, for material in allowing the paradox to arise for the epiphenomenalist is the causal inefficacy of phenomenal experience. The interactionist, in contrast, can subscribe to a widely accepted and intuitively very plausible theory of justification: the causal theory.5 According to the causal theory of justification it is a condition of a belief being justified that the object of the belief figures appropriately in the causal history of that belief. An interactionist can maintain that a causal theory of justification is appropriate in general, and in particular that it applies to judgements about phenomenal experience. Then our judgements about phenomenal experience are justified because they are at least partly caused by the experiences they are about, and the paradox fails to arise. In chapter one it was argued that interactionism is better aligned with common sense because it takes phenomenal experience to be causally efficacious. Given the intuitive plausibility of the causal account of justification, the solution to the paradox outlined here constitutes a natural choice for the interactionist, a suitable continuation of the alignment with intuition and common sense.6

4 Almost never, but perhaps sometimes, as eliminativists about consciousness would have counterparts in the zombie world, too. Their judgements would represent the limiting case. The judgement ‘I have no phenomenal experience’ might perhaps be said to be a degenerate case of a judgement ‘about’ phenomenal experience. Since it corresponds to the state of affairs it is ‘about’, perhaps an account of justification which made that judgement come out justified could be developed.

The zombie case, by the way, is an illustration here and the claim that justified judgements are almost never made in the zombie world is not required for the argument. All the argument requires is that there be some unjustified judgements in our world, a claim I defend below. I think that the claim about the zombie world is plausible, but I do not defend it here.

5 For arguments in favour of the intuitive plausibility of this theory see section IV of Benacerraf’s ‘Mathematical Truth’ (1973, pp. 671-73).

The paradox of justified phenomenal judgements is a very memorable problem for epiphenomenalism. Furthermore, the fact that interactionism has a solution to the paradox is made more noteworthy still by the fact that interactionism can solve other problems that face epiphenomenalism as well. Epiphenomenalism was criticised above for clashing radically with ubiquitous beliefs about motivation. Interactionism has no such problem. It is open to interactionism to claim that it is (at least partly) because you know that chocolate tastes good that you purchase a chocolate bar, that you exercise (at least partly) because you wish to feel good about your physical condition (and you believe that exercise will further this goal), that it is (at least partly) because being humiliated feels bad that most people avoid humiliating situations etc. Interactionism aligns well with common sense here.

Other problems faced by epiphenomenalism on account of the causal inefficacy of phenomenal experiences are avoided by interactionism in similar ways. The evolutionary objection complains that epiphenomenalism leaves us without an account of how phenomenal experience has evolved in humans. A causally inefficacious feature of our constitution cannot contribute to our reproductive fitness, so, a fortiori, cannot have been among the traits picked out by evolution. This poses a challenge for the dualist to explain how phenomenal experience came to evolve. The early Jackson’s solution to this problem — he later characterises it as ‘hardly attractive’ — is that phenomenal experience is an unavoidable by‐product of traits that are adaptive (Jackson 1982, pp. 133‐34; Jackson and Braddon‐Mitchell 1996, p. 7). Interactionism, in contrast, offers a rather more elegant solution; phenomenal experience was picked out by evolution because it is adaptive.

6 What is needed is differential treatment of justified and unjustified phenomenal judgements, and that is what it will be argued that epiphenomenalism fails to provide. It is assumed here that the causal theory is an excellent candidate for supplying such differential treatment, and that a systematic difference between the causal histories of justified and unjustified phenomenal judgements is discoverable; more specifically the difference of containing and not containing, in the appropriate place, the right phenomenal experience. This, of course, remains a somewhat substantial assumption until much more is known about this issue.

Interactionism will admittedly face an isomorphic problem when it comes to accounting for phenomenal experiences that are manifestly maladaptive. Panic angst is an obvious example. Panic angst is all‐consuming and debilitating, and many of the situations where it sets in would be far better dealt with by action than by the inaction that it causes, so surely the capacity for panic angst is maladaptive if anything is. Unsurprisingly, the answer to this challenge is also isomorphic; the interactionist should claim that phenomenal experience comes as a ‘package deal’, and that it is, overall, more adaptive to have the potential for the whole spectrum of phenomenal experiences than to not have any. This may make it seem like the advantage interactionism can claim in favour of epiphenomenalism here is very modest. Modest or not, it is not insignificant. Resorting to the package deal solution for a small subset of phenomenal experience is much less problematic than it is to thus explain the entirety of a phenomenon as ubiquitous and all‐encompassing as phenomenal experience is. The solution, in other words, is acceptable in very limited cases, but its use should be minimised. That is what interactionism does. Similarly, what Jackson and Braddon‐Mitchell call ‘the epistemological’ objection to epiphenomenalism (p. 7), that of how phenomenal experience can be remembered, does not cause problems for interactionism, for causally efficacious phenomenal experiences can leave traces in the world and thus can be known and remembered.

These considerations reinforce the view argued above, that interactionism is a position that fits very well with our commonsense view of the world. This is not to say that its plausibility cannot be undermined by argument, levelled both specifically at interactionism and generally at dualism. It does, however, put strain on any argument summoned to perform this duty, and, to reiterate, particularly if it is true, as was argued in the first chapter, that the plausibility of the dominant alternative position — monism — depends to a not inconsiderable extent on the elimination of tenable dualist alternatives.

2.1.2 CHALMERS’ RESPONSE

We now turn to the solution Chalmers proposes to the paradox in The Conscious Mind. As mentioned, an obvious candidate solution is to adopt a causal theory of justification; this stops the paradox from arising in the first place. Chalmers rejects this solution. The problem with such an account is, he argues, that causal connections are contingent — “a causal connection that holds might not have held” — so any knowledge that relies on such a connection cannot be certain either (p. 195). This, he argues, conflicts with the certainty of our knowledge of phenomenal experience.

We take ourselves to know much about, for example, medium sized objects that surround us. Nevertheless, it is widely held that for all we know we could be brains in vats; all our sensory input could be the result of an elaborate scam and not of the objects we believe occupy our surroundings.7 In contrast, according to a similarly widely shared intuition, it is not possible to construct a sceptical scenario about being conscious analogous to the sceptical brain‐in‐vat scenario about being embedded in a physical world. Sceptical scenarios rely on the possibility that everything might seem just the same to the subject while the situation is radically different from what the subject believes it to be. Therefore, given that “[t]here is no situation in which everything seems just the same to us but in which we are not conscious, as our conscious experience is (at least partly) constitutive of the way things seem”, Chalmers argues that a sceptical scenario about being conscious is impossible to construct (p. 195). This shows, he argues, that our knowledge about conscious experience is certain, and consequently that any account that allows for uncertainty about that knowledge (and the causal theory of justification is among them) is inadequate.

7 Chalmers denies elsewhere (2003b) that brains–in–vats would be massively deluded, but the contrast to the sceptical scenario and its uncertainty is really only illustrational here. What matters is that Chalmers takes the causal account to be unable to deliver the required degree of certainty, a certainty he thinks that our knowledge that we are conscious demands: “There will always be a skeptical scenario in which everything seems just the same to the subject, but in which the causal connection is absent and in which [the purportedly known object] does not exist; so the subject cannot know for certain about [that object]. But we do know for certain that we are conscious; so a causal account of this knowledge is inappropriate” (1996, p. 195).

According to Chalmers in The Conscious Mind, judgements about phenomenal experience are justified simply by “having the experiences” (p. 196):

[T]here is something intrinsically epistemic about experience. To have an experience is automatically to stand in some sort of intimate epistemic relation to the experience—a relation that we might call “acquaintance.” There is not even a conceptual possibility that a subject could have a red experience like this one without having any epistemic contact with it: to have the experience is to be related to it in this way (pp. 196‐97).

2.1.3 CRITICISM OF CHALMERS’ RESPONSE

There is something right about Chalmers’ picture here, but, I shall argue, something that is wrong about it as well. It is clearly true that it is impossible to have an experience without standing in any relation to it. The problem, however, is determining the nature of this relation with as much precision as possible. The claim that the relation is epistemic does not on its own take us any closer to an explanation or understanding of the relation, and neither does the assertion that experience itself has an ‘intrinsically epistemic’ aspect. Both statements are surely true on some reading of them. However, like the claim that simply having an experience justifies judgements about the experience, they do not elucidate what requires elucidation; they restate what needs to be explained in a different way.

Is having phenomenal experience a necessary condition for being able to make justified judgements about phenomenal experience? Can a judgement about phenomenal experience be justified if there is no such experience? It is hard to think of a case where this would be so, but perhaps not impossible.8 Intuitions may vary, however, and not much hinges on this point.

2.1.3.1 HAVING EXPERIENCE IS NOT SUFFICIENT

What is much more important is that just having phenomenal experience is not a sufficient condition for making justified judgements about phenomenal experience. We can imagine someone having one set of conscious experiences but making judgements as if they had a different set. It appears that Chalmers would have to agree, for this is just a description of the zombie case. However, contrary to Chalmers, I think the claim holds not just in the limiting case when the set of experiences that obtains is the null set, as is the case with phenomenal zombies. We can coherently imagine someone who has the right ‘amount’ of phenomenal experience, but who nevertheless makes a large number of mistakes in their judgements about phenomenal experience.

That this is possible follows from the possibility of making single erroneous judgements about phenomenal experiences, which Chalmers does not contest. If it is possible to make one erroneous judgement about phenomenal experience, it is also possible to make another mistake immediately following the first one, adding up to two consecutive mistaken judgements. Then it is also possible to have a long series of nothing but mistakes. The larger the number of consecutive mistakes the more improbable the scenario, but given a large enough sample to draw on even the very improbable will obtain. (Put enough monkeys to work on enough typewriters, and you will — later rather than sooner, probably — end up with the complete works of Shakespeare.) There may be an upper limit to the amount of mistakes we can coherently attribute to another person.9 But those limits allow enough mistakes to make it reasonable to hold that some proportion of such a person’s judgements would have to be classified as unjustified; lest the concept of justification lose all content.

8 Imagine subjects repeatedly administered a drug with a dual effect: first to cause strong pleasant taste experiences, and secondly to interfere with the mechanism by which gustatory experiences are judged, causing it to ‘go haywire’. In the first few trials the subjects might, while still under the influence of the drug, intermittently report very pleasant and very awful taste experiences. After the drug wore off they might be expected to report that, in fact, their taste experiences had been very pleasant throughout.

After repeated trials, and so long as no memory impairment is supposed, it is reasonable to expect the reports during the time under the drug’s influence to become gradually more and more positive. What is going on, we might stipulate, is that the subjects remember that the drug has a dual effect, and overrides their initial inclination to judge that they are having dreadful taste experiences, because they know that they are influenced by the very same drug as before, and know that, in the previous instances, when the drug wore off, they confidently asserted that their taste experience had been pleasant all along.

If unbeknown to the patients the drug was now changed to one whose first effect was to suppress all gustatory experience while the second was the same as before, it seems likely that the subjects would still judge that they had pleasant gustatory experiences, though perhaps with decreased frequency and somewhat lessened confidence. In such a case I think it would be best to say that the judgements were justified. The causal theory can account for this. The distinctive feature of the situation is that the subjects learns to base their judgements about experiences immediately prior to the judgements on previous judgements, judgements that in turn are causally linked to experiences. The effect is that the subject’s latest judgements become causally dependent on experiences much further removed from the judgements than what is normally the case. Nevertheless, there is a causal connection between experiences and judgements about those experiences that explains why we should judge them to be justified.

Perhaps some would protest, however, and argue that to admit that we might make mistaken judgements about phenomenal experiences is not yet to admit that those judgements are unjustified, regardless of the number of mistakes we suppose are made.10 How should we evaluate such a claim? We might start with the observation that while infallibility may or may not imply justification, the failure of infallibility certainly does not imply failure of justification.11 The question of whether someone can fail to be justified in beliefs about their phenomenal experience is therefore quite correctly claimed to be a further, distinct question from that of whether one can be mistaken about one’s phenomenal experience. Moreover, the two questions appear to be independent of one another, in the following sense: it seems possible to make mistaken but justified judgements about phenomenal experience (cf. n. 8 above), and it seems possible to make unjustified but correct judgements about phenomenal experience (as when a judgement is passed in excessive haste, but happens to be correct; more on this in a moment). Given this, some might claim that all our judgements about phenomenal experience, erroneous or not, are justified. For the paradox to arise in the form discussed here (cf. n. 3 above) it must be possible to make unjustified phenomenal judgements, so the question about the possibility of making unjustified phenomenal judgements is relevant here.

9 See e.g. ‘On the Very Idea of a Conceptual Scheme’ (Davidson 1974, especially p. 19).

10 I take it that such a person would be defending about phenomenal experience the view that Alston calls ‘self-warrant’: “For any proposition, S, of type R, it is logically impossible that P should believe that S and not be justified in believing that S” (1971, p. 235).

11 Alston claims that infallibility implies justification: “[O]ne could hardly have a stronger (epistemic) justification for holding a certain belief than the logical impossibility of the belief’s being mistaken” (1971, p. 229). I think this is less than obvious, for, since they are necessarily true, true mathematical beliefs are logically guaranteed to be correct; nevertheless there certainly seems to be examples of people holding true mathematical beliefs without being justified in so doing; viz. the first person to ever entertain the belief that there is no largest prime number. However, I remain agnostic on this issue here.

2.1.3.2 SOME UNJUSTIFIED PHENOMENAL JUDGEMENTS

Are all phenomenal judgements justified? I think not.12 Examples of erroneous phenomenal judgements where it seems fully possible for the subject to avoid the erroneous judgement — for instance because the mistake arises from haste or carelessness — seem amply available. Furthermore, it seems reasonable to say that when it was fully within a subject’s capabilities to avoid making an erroneous judgement, then she was unjustified in passing that judgement. Chalmers mentions as an example of an erroneous phenomenal judgement someone misclassifying the phenomenal experience of cold for heat when passing a quick judgement in the context of expecting to be burnt (2003a, p. 241). Given that it is likely that at least some such judgements could have been avoided with relative ease the example works just as well as an example of unjustified judgement.

Here is another example, from an exhibition I once enjoyed at the tech