Emergent necessitation and contingency
For two reasons, to justify the Harmony Thesis O’Connor needs the “neces-sitation’ of emergent active power by the subvenient base. The best way to clarify ‘‘necessitation’’ is to characterize it in the context of presenting the first reason. To get at that reason, it will be useful to begin by reviewing insights from Frank Jackson that were presented in chapter one.61
Jackson contrasts serious metaphysics with a shopping list approach to ontology. According to the latter, metaphysicians adopt a pluralist approach to ontology and seek to provide big descriptive lists of the various kinds of things there are, resting content to add sui generis entity to sui generis entity. What is missing in such an approach is any attempt to provide a comprehensive, discriminatory account of what is real and how things came to be. According to Jackson, advocates of N should take naturalism to be a piece of serious metaphysics because in so doing, they pattern the epistemic justification of N on that of good scientific theories, and they provide grounds for preferring N to its rivals on the basis of N's superior explanatory power.
Jackson correctly observes that if one takes N to be an expression of serious metaphysics, then one must face the location problem. According to Jackson, given that naturalists are committed to a fairly widely-accepted physical story about how things came-to-be and what they are, the location problem is the task of locating or finding a place for some entity (for example, semantic contents, mind, agency) in that story. The mereological hierarchy is the ontology that results from serious metaphysics and, among other things, it requires one to locate emergent entities up the hierarchy in terms of the basic, subvenient entities at the level of physics.
For Jackson, location amounts to entry by entailment: some entity is located and, therefore, has a place in one's ontology if it is entailed by the basic account. Applied to mental and physical entities, any world that is a minimal physical duplicate of our world is a psychological duplicate of our world.
Because he is concerned to locate certain supervenient structural properties (e.g. broad mental contents), Jackson opts for global supervenience as a way of clarifying his view further: A world which is a minimal physical duplicate of our world is a world that (a) is exactly like our world in every physical respect and (b) contains nothing else (e.g. Cartesian souls) in the sense of nothing more by way of kinds or particulars than it must to satisfy (a).
In the context of O’Connor’s thought, since the concern is to locate emergent and not structural properties, a certain version of strong super-venience would be widely accepted by naturalists as adequate for the task. Expressed in terms of physical and psychological properties, N seems to require that in no possible world with the same physical properties and laws as the actual world is there a particular that shares its physical properties with a particular in our world but fails to share its psychological properties. In this way, the physical may be said to ‘‘necessitate’’ the psychological. It is important to keep this framework in mind for what follows.
Though he does not mention it explicitly, O’Connor seems concerned to take N as an expression of serious metaphysics, and he understands this to require the location of emergent properties, including mental properties such as active power, in terms of the understanding of ‘‘necessitation’’ just mentioned. Since he is concerned to show that those who accept N are not thereby given adequate grounds for rejecting AGC, O’Connor must be assuming that AGC may be adequately located in N and, moreover, that AGC does not provide evidence for a rival to N, say theism, along with substance dualism as a component of theism. As O’Connor admits, many—perhaps most—have seen AGC as evidence against N and reject the Harmony Thesis. Thus, O’Connor argues that if one is going to have a scientific understanding of an emergent property, one cannot merely accept a property as emergent without explaining its existence. Rather, one must require that an emergent property be causally grounded in its base properties if it is to be naturalistically explicable.62
Elsewhere, O’Connor claims that if an emergent property is depicted in such a way as to be contingently linked to the base properties causing it to emerge, then apart from an appeal to God’s contingent choice that things be so, and to God’s stable intention that they continue to be so, there will be no explanation for the link itself or its constancy.63 In short, if the link is contingent, the Harmony Thesis is false and AGC provides evidence for theism, and given theism, there is less need to preserve physicalism in the category of individual.
There is a second reason why O’Connor needs the “necessitation” of emergent active power by the subvenient base: the view of causation that forms the core of O'Connor's depiction of both N and AGC. Recall that for O'Connor, the causal powers of properties are essential aspects of those properties and, thus, belong to properties with an absolute, metaphysical necessity. The causal potentialities of a property are part of what constitutes the property's identity.64 O’Connor’s realist view of causation—event and agent—entails that a cause produces or brings about its effect in virtue of the properties of the cause, and properties are universals that have essentially their causal powers.65 Since most philosophers identify the supervenience relation with the causal relation in the case of emergent properties, it is in this causal sense that in the right circumstances, the instantiation of a subvenient property necessitates the instantiation of its associated emergent property.
Given this view of causal potentialities, a certain result seems to follow regarding emergent properties. Since an emergent property is simply the actualization of causal potentialities in the right circumstances, the emergent property would seem to be a part of its causal property’s identity as well. In this sense, an emergent property would seem to require its base property(s) to exist. Interestingly, in an earlier account of emergence, O’Connor accepted this robust claim about emergent properties.66 He held this precisely because he took an emergent property to be an expression of the very nature of the subvenient base causing it.
However, in Persons and Causes he says that the notion that an emergent property could not exist without its subvenient base is “possibly gra-tuitous.’’67 His concession seems to result from his desire to offer as minimalist an account of emergence as possible to increase its chances of being accepted by critics and, thus, he leaves open the sort of modality (metaphysical, nomological) required for a minimalist account of emergence. But O’Connor himself continues to accept the more robust account of causality, and this would seem to require that he also continue to accept the stronger notion of emergence.68
Unfortunately, while the Harmony Thesis requires the relevant physical circumstances to necessitate emergent mental properties, including active power, the link between mental properties and the relevant physical circumstances seems to be utterly contingent. Grounded in strong con-ceivability, thought experiments proliferate throughout the literature in philosophy of mind, which provide strong justification for this claim. For example, inverted qualia and Chinese Room scenarios seem to be coherent and entirely possible. No strictly physical proposition of N employing solely physical terms for particulars, properties, relations or laws renders these thought experiments broadly logically impossible, even in worlds that resemble ours in every physical respect.
Again, different forms of the well-known Knowledge Argument seem to be quite plausible. Since O’Connor himself accepts a property dualist interpretation of the argument, given this interpretation, no knowledge whatever of merely physical facts gives one any information at all about the presence, absence or nature of mental facts. If this is so, it is difficult to see how one could justify the claim that, say, φ entails Ψ. No amount of information about the former entails anything at all about the latter. φ is consistent with our world and with inverted qualia and zombie worlds that are minimal physical duplicates of our world. If this is so, then the physi-cal/mental link seems contingent indeed.
Further, the modal argument for substance dualism seems to be quite plausible. If so, then at least certain versions of the argument imply that physical entities are not necessary for the instantiation of mental properties. Indeed, theism itself presents (at least) one case in which active power is not dependent upon a physical base. Surely, the existence of God (and of angels) with libertarian power is metaphysically possible, and if so, it is just not clear why the property of active power is causally tied to a physical base.
These various thought experiments have been around a long time and there is no sign that they are going away. They provide evidence against the necessitation claim that is central to the Harmony Thesis. So far as I know, O’Connor does not consider the force of the modal argument. I would be especially interested to see how he would handle cases in which the agent cause is a pure spirit (God and angels, disembodied souls?). If he says that the presence of the relevant physical base necessitates the emergence of active power, but that the latter could obtain without the former, then this would seem to amount to a denial that an emergent property is an essential aspect of the subvenient property whose potentialities actualize it. At the very least, it would imply that the presence of active power is underdetermined by emergence and substance dualism such that there would be no adequate grounds for preferring the former.
Given a functionalist analysis of mental kinds, it may be that a type of mental state could be ‘‘realized’’ in spirits and brains and this fact is consistent with particular brain states in certain circumstances necessitating the realization of a mental state by being sufficient for such a realization. But for two reasons this admission would not provide O’Connor with a rejoinder to my argument from the instantiation of active power in spirits. Given that active power is a simple, intrinsically characterized property that is instantiated, and not a structural property that is realized, O’Connor depicts active power as a disposition of its metaphysical base as a matter of metaphysical necessity, and it is hard to see how this disposition could be actualized without its categorical base.
Further, most naturalists do not cash out emergent supervenience merely as the logical sufficiency of the subvenient base. Rather, they spell out emergence in terms of two other principles that, together with logical sufficiency, constitute minimal physicalism:
(1) The anti-Cartesian principle: There can be no purely mental beings (e. g. substantial souls) because nothing can have a mental property without having a physical property as well.
(2) Mind-body dependence: What mental properties an entity has depend on and are determined by its physical properties.69
Naturalists employ (1) and (2) in their analysis of emergence precisely because they want to ensure that emergent properties are located in the naturalist ontology by guaranteeing that such properties require, depend on and are causally determined by their entirely physical subvenient bases. If most naturalists are correct about this requirement for locating an emergent property in the ontology of N, then the actuality or even the metaphysical possibility of the instantiation of active power in a pure spirit is a problem for the Harmony Thesis. It is one thing to reject the existence of God and angels. It is another thing altogether to claim that God or angels are metaphysical impossibilities, even if the modal status of such a claim is limited to possible worlds with the same physical particulars, properties, relations and laws as the actual world.
O’Connor does address the Knowledge Argument and inverted qualia thought experiments. Regarding the former, he opts for a dualist interpretation of the argument and claims that at least two features of many mental phenomena are emergent properties causally necessitated by the appropriate physical bases: the phenomenal feature and subjectivity, which he interprets as the fact that one can come into contact with a conscious property only by having it.70 What about the apparent contingency of the mental/physical causal link? O’Connor simply denies that all causal necessity must be transparent. He says that there is no good reason to think that when we come to have a scientific understanding of some phenomenon, we will just be able to see that a causal effect had to follow from its cause. In the case of conscious properties, though their causal bases necessitate
them, we just cannot see the necessity of the causal connection. Regarding inverted qualia, O’Connor adopts the same dismissive strategy, claiming that inverted qualia thought experiments ‘‘implausibly drive a wedge between a phenomenal property’s qualitative features and its causal role.’’71
In my opinion, O’Connor’s rejoinder to these arguments for causal contingency sounds very much like a mere denial that there is a problem. But the intuitions of contingency that lie behind the various dualist arguments in focus are rooted deep within our pre-philosophical intuitions. Surely,
then, there is a burden of proof on O’Connor that is not met by his dismissive strategy.72
Besides that, there are four considerations whose cumulative effect undercuts O’Connor’s claim that conscious properties are emergent. First, O’Connor himself admits that ‘‘there are no widely accepted working theories that are committed to the existence of emergent properties,’’73 and ‘‘there is a lack of hard evidence in favor of emergence in areas that are well understood.’’74 He does not find this particularly troubling. Rather, he believes that our scientific knowledge is so incomplete that the absence of emergent properties is far from empirically established. However, surely the burden of proof lies in the other direction, and given the state of things, the proper conclusion to draw is that, currently, ‘‘the hypothesis of emergence’’ is yet to be justified.
Second, it is false to claim that ‘‘there is convincing evidence’’75 that (at least many) mental properties are emergent. For three reasons, it is difficult and may be impossible to justify their emergence empirically: (1) The emergent hypothesis and substance dualism are empirically equivalent models and there is no empirical evidence that can count in favor of one over the other.76 (2) In the attempt to correlate mental and physical properties as a first step towards justifying the hypothesis of emergence, one of the two correlates is not available for empirical inspection, and this makes straightforward empirical justification of emergence more difficult. (3) It is only in the case of fairly simple mental states, e.g. specific sorts of pains, that we have any hard evidence of specific mental/physical correlations. There is no evidence whatever that complex mental properties, such as the property thinking-about-the-history-of-skepticism, are correlated with specific base physical properties, much less emergent on them. Part of the problem here is the difficulty of providing criteria for individuating complex mental states in an empirically testable way, a problem that O’Connor himself acknowledges.77 On a fine-grained theory of properties, this may well be an impossible task, and not just a difficult one. This is why many strong physicalists adopt a course-grained view of mental properties as a response to inverted qualia arguments, but this move requires that mental properties be identified with functional roles, and it is not available to O’Connor.
Third, even if we grant that mental properties are, in some sense, emergent, that does not entail that they are emergent in O’Connor’s sense. Recall that for O’Connor, emergent properties have these three features:
(1) they are simple, intrinsically characterizable, new kinds of properties,
(2) they have their own ontologically basic type of causal influence, and (3) they are causally necessitated by their subvenient physical base.
Roughly, the first two features correspond to what John Searle calls emergent;! and emergent2, respectively.78 Now Searle is typical of those naturalists who accept emergent properties as merely emergent1 and not emergent2. Since we have seen reasons for this in chapter one (e.g. the claim that N requires the causal closure of the physical), I will not rehearse them here. But one point needs to be emphasized. O’Connor claims that mental properties are the best examples of emergent properties, since they exhibit subjectivity and a phenomenal nature, and he claims that we have ‘‘direct evidence’’ of emergence in the case of consciousness.79
I agree that we have direct access to and introspective knowledge by acquaintance of our own mental states, but naturalists such as Searle claim that this ‘‘direct evidence’’ merely justifies conscious properties as emer-gent1 and not emergent2. As I will argue below, the sort of introspective evidence that might be cited to support the claim that some mental properties, especially active power, have their own causal powers also supports substance dualism and, thus, that evidence provides a defeater for the claim that mental properties are emergent. At the very least, this additional introspective evidence goes beyond the sort of direct evidence O’Connor cites to justify consciousness as emergent2. At best, it merely justifies them as emergent;!.
However, even if this “direct evidence’’ justifies taking active power to be emergent in the first two senses, it utterly fails to justify the third sense.
The vast majority of people agree that in introspection they are completely unaware of anything physical. They have no introspective acquaintance with their brain or any other strictly physical object, or with any sub-venient physical properties. When philosophers argue that consciousness is a set of emergent properties, one thing seems clear: they do not appeal to first-person introspection to justify the claim. Indeed, no inspection of the brain or any other candidate for the subvenient physical base from a first or third-person perspective provides ‘‘direct evidence’’ for treating any conscious property as emergent in sense three.
This is an important conclusion that O’Connor apparently fails to see. In a publication subsequent to Persons & Causes, O’Connor acknowledges that ‘‘[t]he emergentist can and should allow that there is an epistemological presumption against emergentist hypotheses for systems of currently-untested complexity levels absent special reason to suspect that they are different from run of the mill cases.’’80 So far so good. But right after granting this concession, O’Connor attempts to refute a claim by Brian McLaughlin to the effect that, while emergence is a coherent concept, it is enormously implausible that there are any such properties, and least for those with ostensible scientific sobriety.
O’Connor’s response consists in two claims: (1) a person’s experiences and other conscious mental states are sui generis simple emergent properties and (2) claim (1) is defeasibly justified by direct first-person awareness of conscious states with an epistemic strength that precludes the a posteriori ascription to them of hidden micro-structure hidden to introspection. But O’Connor is simply mistaken about this. Direct first-person awareness completely fails to provide any justification whatsoever for his third characterization of emergent properties and this is the sense he needs to justify conscious properties as emergent in the sense needed for his Harmony Thesis.
Finally, given O’Connor’s admission that ‘‘direct evidence’’ is involved in justifying the claim that conscious properties are emergent ones, the epistemic grounds for this claim derive from first-person introspection and not from empirical research. As we have just seen, O’Connor insists on this. Given that this evidence provides accurate information about the intrinsic nature of mental properties (his sense one of emergence), and given that we have a fairly good idea of the nature of physical properties, most have seen their connection to be contingent, and that is why naturalists have had such a hard time ‘‘locating’’ them in light of the necessitation condition discussed earlier.
The contingency of the link between mental/physical properties stands in stark contrast to naturalist examples of paradigm cases of located macroproperties. For example, Jackson cites macro-solidity, understood as impenetrability, as something easily construed as necessitated by subvenient base traits, cashed out in terms of intermolecular forces, lattice structures, and so on.81 Jackson also points out that the pre-scientific notion of macrosolidity as being everywhere dense has been rejected by those who accept N. The reason for this rejection is clear. If real, the latter notion of solidity would be a macro-property only contingently connected to its microphysical base and, thus, it would not be located in N.
I think most naturalistic philosophers would hold that irreducible mental properties are like the pre-scientific notion of solidity. Since they cannot be located, our dualistic pre-scientific conception of them must be revised according to some strong physicalist strategy. If mental properties are new kinds of properties as O’Connor claims, they fail to resemble paradigm cases of located macro-properties (e.g. solidity as impenetrability), and O’Connor has failed to provide an adequate justification for assimilating them to the paradigm cases. Interestingly, he acknowledges that “[rjeductionism nowadays is much disparaged. Yet by our lights, the most plausible variety of physical-ism is reductionist, as it does not require one to make dubious moves in the underlying metaphysics of physical properties.’’82 I have been at pains to show that it is no accident that strong physicalism is and ought to be the ontology of naturalism precisely because it does not require one to make dubious moves in the underlying of naturalism. For self-reflective naturalists who claim explanatory superiority for their worldview, the Constitution Thesis is an essential component that fits naturalism like a hand in a glove.
In contending for his views, O’Connor makes implicit or explicit reference to particular epistemic features of his case both for AGC and the Harmony Thesis. I shall focus on two of these features and argue that, if applied consistently, they place a burden of proof on O’Connor’s defense of the Harmony Thesis—specifically, the harmony of AGC and a physical agent—that he has failed to meet: the role of pre-philosophical intuitions in his case, and his view of the nature of pre-philosophical intuitions about mental properties.
O’Connor and the role of pre-philosophical intuitions
In arguing for AGC, O’Connor commits himself to two important epis-temic requirements for his task:
(i) one’s view of agency should be guided by and justified in light of prephilosophical, common sense intuitions which place a burden of proof on any view that requires abandonment of them; and
(ii) these pre-philosophical intuitions are a source of justified beliefs about the nature of human action itself, and not merely about our concept of human action.83
O’Connor uses these intuitions to place a burden of proof on compati-bilists and on critics of the Harmony Thesis. Thus, his task in both areas of debate is to rebut and not refute his interlocutors.
Applied to agency, O’Connor claims that incompatibilism is prima facie justified by these intuitions, they ground a modal style argument for incompatibilism, and compatibilists fail to overturn the argument based on these prima facie justified intuitions. Applied to the Harmony Thesis, given N and the pre-philosophical intuitive justification of AGC, O’Connor says that the burden is on those who reject the Harmony Thesis and accept the Constitution Thesis. Since the latter is neither entailed by the Causal Unity Thesis nor empirically established, then we are not required to accept it. Failure to meet this burden, coupled with positive grounds for emergent properties to be described below, means that there is no good reason to reject the Harmony Thesis.
How does one know when one has solid pre-philosophical intuitions with sufficient justification to do the work required of them in O’Connor’s case? I suggest there are at least two features of such intuitions. First, such intuitions should be held widely and deeply by normal folk with no ideological axe to grind. Throughout the literature, friends and foes of incompatibilism acknowledge that it enjoys this sort of intuitive support, and O’Connor makes explicit use of this fact in his case.84 Second, both sides of a dispute employ concepts derived from or based on those intuitions. John Bishop is typical of many compatibilists when he explicitly employs a libertarian concept of agency to develop his own compatibilist model that falls under that concept ‘‘closely enough’ to be adequate.85 Bishop allows a libertarian conception of agency to guide the development of his own account, and to be the legitimate source both of counter arguments in the form of thought experiments and of the sense of adequacy for his responses to those counter arguments. Libertarian intuitions seem pervasive in debates about agency.
Now both of these characteristics seem present for intuitions on behalf of substance dualism and against physicalist views of the self. Friends and foes of dualism widely admit that it is the common sense view, and the vast majority of people throughout history have been dualists about the self in one form or another. Jaegwon Kim acknowledges that “We commonly think that we, as persons, have a mental and bodily dimension. ... Something like this dualism of personhood, I believe, is common lore shared across most cultures and religious traditions.’’86 Along similar lines, Frank Jackson says that ‘‘our folk conception of personal identity is Cartesian in character.’’87
Pre-philosophical intuitions in support of a substantial, immaterial self are widely and deeply held, and they ground the modal argument for substance dualism.88 Moreover, these intuitions seem expressed in the concepts and arguments used by dualists and physicalists. The intelligibility of NDE’s (Near Death Experiences), arguments from the unity of one’s conscious field, thought experiments about personal identity to the effect that the person is merely contingently related to his body or psychological traits, and responses to these thought experiments (e.g. various causal chain analyses of personal identity) seem to employ a substantial, immaterial conception of the self.
O’Connor could respond that in the case of substance dualism, grounds for N justify a rejection of these pre-philosophical intuitions, but in light of his own employment of similar pre-philosophical intuitions for AGC and the Harmony thesis, this response seems arbitrary. After all, most naturalists employ N to justify a rejection of the intuitions in support of AGC, a fact that O’Connor acknowledges. Most naturalists agree that pre-philosophical intuitions are on the side of AGC and substance dualism, but they adopt a consistent attitude—rejection—towards both sets of intuitions. While strictly consistent with the grounds for N, most naturalists believe that AGC and substance dualism are not as plausible as compatibilism (or non causal versions of incompatibilism) and physicalism in light of those grounds.
Moreover, just as the Causal Unity Thesis fails to entail the Constitution Thesis and the latter has not been empirically established, so the empirical grounds for N fail to entail or empirically establish a physical agent. If O’Connor thinks otherwise, he is invited to cite the empirical evidence that accomplishes this feat. In the absence of such evidence and in light of his own epistemic characterization of the requirements placed on those who would reject the Harmony Thesis, it is hard to see what O’Connor would say to the same claim made by substance dualists about the epistemic status of physicalism, given the presence of pre-philosophical intuitions for substance dualism.
O’Connor and the nature of pre-philosophical intuitions
In addition to the role of pre-philosophical intuitions in O’Connor’s case for AGC and the Harmony Thesis, the nature of those intuitions is also of crucial importance. Philosophers differ about the nature of intuitions, e.g. some hold that they are merely dispositions to believe certain things. However, the traditional view of intuitions takes them to be cases of (perhaps, defeasible) first-person direct awareness of a relevant intentional object. So understood, they are reported by way of the phenomenological use of “seems’’ or ‘‘appears.’’
O’Connor seems to agree with this view of intuitions. He claims that intuitions in support of AGC are the way things ‘‘seem’’ to people.89 In arguing that consciousness is an emergent property, he claims that people have ‘‘direct evidence’’ of the nature of conscious properties themselves. Here, he seems to take it that one has direct first-person access to one’s own mental states and, indeed, if this is so, such access seems to provide non-doxastic justification for pre-philosophical beliefs about/concepts of mental properties, including the nature of active power. He also claims to experience himself directly bringing about the formation of an intention.90 If one accepts this account of intuitions, then one has the resources to explain why certain beliefs are so widely and deeply held.
But dualists, regarding intuitions about the self, often make the same claim. For example, Stewart Goetz has argued that we are directly aware of ourselves (e.g. of our own simplicity) and, on this basis, we are justified in believing substance dualism.91 It is on the basis of such first-person selfawareness, that people have the pre-philosophical dualist beliefs they do, and this is why these beliefs (or, at least, dualist concepts) play such a regulative role in philosophical arguments about personal identity and related topics.
Of course, it is fashionable today to claim that people have direct access to their mental states but not to their selves. From the time of Hume, the major strategy employed to justify this assertion is the claim that, in fact, people just are never aware of themselves. I believe that dualists have provided adequate responses to this strategy, but that is beside the present point, because I do not believe that O’Connor can avail himself of this strategy. To see why, we need to examine his response to an epistemological objection raised against his version of AGC. The objection is that we could not, in principle, ever know whether any events are produced in the manner that AGC postulates since agent caused events would be indistinguishable from ones that were essentially random.92
O’Connor points out that this Humean-type objection would be equally telling against his realist version of event causation (event causes produce or bring about their effects). The Humean skeptic will say that all we have direct evidence for is the pattern of relations among types of events, not of the causal event bringing about its effect. In reply, O’Connor says that in at least some cases we seem to observe directly the causal connectedness between cause and effect. He illustrates this by pointing out that we do not merely observe the movement of the hammer followed by the movement of the nail; rather, we see the hammer moving the nail.
Now it is not clear how one can directly see the hammer moving the nail without directly seeing the hammer. Similarly, it is hard to see how one could directly be aware of one’s own self-producing an intention to act without being directly aware of one’s own self. Indeed, O’Connor acknowledges that ‘‘in the deliberate formation of an intention, the coming to be of my intention doesn’t seem to me merely to occur at the conclusion of my deliberation; I seem to experience myself directly bringing it about.’’93 This would seem to imply that people are able to be directly aware of their own selves. If so, and given that pre-philosophical intuitions are widely acknowledged to be of a substance dualist sort, the very nature of intuitions as first-person forms of direct access seems to offer defeasible justified beliefs of a substance dualist sort.
It may be that O’Connor has other reasons for rejecting the use of first-person direct awareness of the self as grounds for substance dualism. To my knowledge, he has not addressed the topic in writing, but I could be wrong about this. If he does, there would seem to be two requirements for any such response. First, without begging the question, he is going to have to provide sufficient grounds for rejecting first-person awareness of the self and the role such awareness plays in justifying substance dualism in such a way that he does not undermine his own use of first-person awarenesses as a source of justification for AGC. For example, he cannot simply assert that naturalism makes substance dualism implausible, so we must reject the force of this dualist argument, because the same thing is widely said about the epistemic impact of naturalism on the justification of AGC.
Second, he would need to offer an explanation of the origin and justification of the various dualist intuitions that are a part of O’Connor’s own characterization of the agent. I am not disagreeing with that characterization. I am simply asking where it came from and why we should believe it. I believe there is a good answer to these questions—first-person awareness of the self—but these questions would need to be answered in a way that avoids lending support to substance dualism. For example, it seems implausible to suggest that we have first-person awareness of ourselves as physical substances. If we are physical substances, yet we lack first-person awarenesses that this is so and, in fact, seem to have awarenesses that support substance dualism, we would need to know the source of and justification for dualist intuitions that form an essential part of the self qua agent.
The simple fact is that it does not seem to most folk that they are macrolevel objects. On the contrary, it seems to most people from the first-person perspective—the perspective upon which O’Connor draws to justify agent causation—that we are mental subjects who fail to be aware of exemplifying any physical properties. The issue then becomes whether there is any good reason to think we are physical objects, though we are not aware of being such. As far as I know, O’Connor never gives us any reason to think we are physical objects, and he must provide such an argument. When he does, he runs the danger of bringing forth considerations of a kind (e.g. from the third-person perspective) that, if persuasive, could also be brought forth to undermine our conviction that we have libertarian freedom. If he simply breaks rank with most people and says that he is, in fact, aware of being a material object by first-person introspection, then this would at best justify locating his view within panpsychism and not within naturalism.
In sum, I agree with most of O’Connor’s model of agent causation. But I do not believe he has provided sufficient grounds for accepting the Harmony Thesis and I have tried to say why I cannot follow him in this regard.