Should dualists locate the physical basis of experience in the head?

Abstract

Dualism holds that experiences are non-physical states that exist alongside physical states. Dualism leads to the postulation of psychophysical laws that generate experiences by operating on certain sorts of physical states. What sorts of physical states? To the limited extent that dualists have addressed this question, they have tended to favor a brain-based approach that locates the physical basis of experience in the head. In contrast, this paper develops an argument for a form of dualism on which experience has a physical basis that reaches out into the world. Specifically, the argument uses a puzzle about spatial experience to support a theory I call tracking dualism. On tracking dualism, a psychophysical law generates experiences by operating on complex physical states that involve both brain states and tracking relations that they bear to features of the environment. To develop the argument, I discuss tracking dualism alongside its more familiar physicalist counterpart, the tracking intentionalist theory associated with Dretske, Lycan, and Tye. I conclude by outlining a tracking dualist research agenda.

1 Introduction

Is the physical basis of (sensory) experience in the head?¹

This question has sown division in physicalist quarters. According to physicalism, experiences just are certain physical states.² Some physicalists embrace metaphysical internalism, the thesis that experiences are narrow physical states, i.e. physical states that are constitutively independent of their bearers’ environments.³ Among metaphysically internalist physicalists, it is uncontroversial that the physical basis of human experience lies in the brain. Other physicalists instead defend metaphysical externalism, the thesis that experiences are wide (i.e. non-narrow) physical states, physical states that reach into the world.⁴ However, the question does not presuppose physicalism. It also arises for dualists.

Dualism holds that experiences are non-physical states that exist alongside physical states. Since experiences vary systematically with interventions on our sensory systems, physicalists and dualists should agree that experiences are systematically correlated with physical states. These phenomenal-physical correlations cry out for explanation. On dualism, these cries are to be answered by positing fundamental psychophysical laws that generate experiences as a systematic function of physical states. Whether the physical basis of human experience lies in the head then turns on whether the physical states these laws take as inputs are narrow or wide. Put in terms of Kripke’s (1980) metaphor, the question becomes: upon selecting the psychophysical laws and an entity’s narrow physical states, did God thereby settle what experiences that entity has?

While this issue has received little discussion in the literature, nearly all⁵ dualists who have weighed in seem to favor nomic internalism on which experiences are nomically necessitated by narrow physical states. Given nomic internalism, narrow physical duplication preserves phenomenology within the space of worlds that have the same laws as ours. However, there is reason to think that dualism should instead be combined with nomic externalism, which holds that experiences are nomically necessitated by wide physical states but not by narrow physical states.⁶ In what follows, I will argue for a nomically externalist theory: tracking dualism.

To a first approximation, tracking dualism goes beyond dualism by positing a specific fundamental psychophysical law that generates experiences as a function of “tracking” states, roughly complex physical states that consist in the holding of certain causal covariation relations between subjects and features of their environments. As we will see, tracking dualism can be usefully understood as a dualistic analog of the tracking physicalist theory associated with Dretske, Lycan, Tye, and others.⁷ On tracking physicalism, experiences are identified with certain representational states which are in turn identified with certain tracking states.

Introducing tracking dualism and explaining its appeal for dualists is the paper’s main aim. I will develop what I regard as the best argument from physicalism for tracking physicalism and use it as a springboard for mounting a roughly parallel argument from dualism to tracking dualism.⁸ These arguments depart from familiar motivations for tracking physicalism: they do not rely on the contentious transparency thesis that when we try to attend to features of experience we can attend only to external phenomena; nor do they rely on an analogy between experiences and other mental states that supposedly require externalist treatment.⁹ Instead, they support tracking theories about spatial experience by contending that they provide the best explanation of a kind of puzzle case. The arguments then contend that tracking theories should be generalized to cover all experiences.¹⁰ The ambitions of my argument for tracking dualism are doubly modest. First, the argument is conditional on dualism, which I’ll assume rather than argue for here.¹¹ Second, even conditional on dualism, that support is non-demonstrative. For reasons given at the end of the paper, evaluating tracking dualism in light of the known relevant considerations is a large research project.

I will proceed as follows. §2 introduces tracking theories in more detail. §3 develops the noted argument for tracking physicalism. §4 presents the case for tracking dualism. §5 outlines a tracking dualist research agenda.

2 Tracking theories

I take it as a background assumption that we have experiences. What it’s like to have a given experience is its phenomenal character. To explain tracking theories in more detail, I will work with the following example: suppose that you have a visual experience (E) while looking at a kale salad. The phenomenal character (C) of E is one that we would naturally describe in terms of the ostensive presentation of various colors and shapes before you. Why do you have E?

Tracking physicalism, as I will understand it, answers as follows. First, tracking physicalists offer an identity intentionalist account of phenomenal character: they identify C with a physical property represented by E.¹² More specifically, C is held to be a complex physical property whose elements are various properties that E presents to you.¹³ For tracking physicalists, these presented properties include ‘sensible’ colors and shapes.¹⁴ Next, tracking physicalists hold that your having E consists in your being in a certain sort of physical state that stands in a certain representation relation to C. In particular, tracking physicalists hold that having E just is being in a physical state that both tracks C and plays the e-role.¹⁵ These notions require unpacking.

Here, ‘tracks’ denotes the wide physical relation¹⁶ (if such there be) that fixes contents and characters of experiences. Whichever relation tracking is, it is representational in the minimal sense that it can relate subjects to properties even when those properties are not instantiated before the subject, as in cases of illusions and non-veridical hallucinations.

The main candidates for tracking are the physicalistic representation relations from the psychosemantics literature, such as asymmetric dependence, functional indication, and causal covariation under optimal conditions.¹⁷ For convenience, we can understand tracking as causal covariation under optimal conditions. This notion allows for misrepresentation: in non-optimal conditions a state may track a property that is not instantiated before the subject of the state by being disposed to causally covary with the property under optimal conditions. The e-role is whatever functional role distinguishes tracking states that yield experiences from those that do not. Why should tracking theorists posit such a role? To avoid attributing consciousness where it isn’t—for example, in thermometers. For convenience, we can think of a state as occupying the e-role as a matter of being poised for cognitive use.¹⁸ So, on tracking physicalism, upon looking at the kale salad, you are conscious (rather than unconscious) because you are in some physical state that occupies the e-role while tracking a physical property. And you have E, rather than an experience as of different sensible shapes and colors, because you are in an e-role occupying physical state that tracks C, rather than some other physical property with different sensible qualities as elements.

Officially, tracking physicalism is defined by its generalization of this to all experiences. That is, tracking physicalism claims:

Every experience e has a phenomenal character c such that (i) c is identical with a physical property and (ii) to have e is to be in a physical state that tracks c and plays the e-role.

Tracking dualists forgo this systematic identification scheme in favor of a systematic psychophysical law. Roughly, the law says that whenever subjects are in an e-role occupying physical state that tracks a property, they have a corresponding non-physical experience as of that property.

To state the law more precisely, I need to introduce the notion of phenomenal representation. Here is a rough characterization that suffices for my purposes: an experience phenomenally represents whatever it must, of metaphysical necessity, represent in virtue of its phenomenal character.¹⁹ This notion of phenomenal representation leaves open whether phenomenally represented contents—and, in particular, sensible qualities—constitute phenomenal characters or are instead distinct entities. For instance, your kale salad experience phenomenally represents various shapes, whether or not shapes help constitute its phenomenal character.²⁰ When the physical properties tracked by a state are the same as those phenomenally represented by an experience, we can say that that state and experience match.

With this terminology, we can state the law to which tracking dualism appeals.

Tracking Law

If a subject is in a physical state that occupies the e-role while tracking a physical property, then she has an experience that matches that state.

Officially, tracking dualism claims that dualism is true and that Tracking Law is the fundamental psychophysical law that fixes the distribution of experiences.²¹ Tracking dualism is neutral on a number of issues. It does not require phenomenal representation to be identified with tracking: while tracking dualists maintain that tracking is a kind of representation that constrains the distribution of experience, they are not committed to tracking being the only kind of representation. For example, it is open to tracking dualists to hold that phenomenal representation is a primitive (non-tracking) form of representation.²² Like tracking physicalism, tracking dualism is neutral both on exactly which wide physical relation the tracking relation is and on exactly which functional role the e-role is.²³ Aside from construing experiences as non-physical states, tracking dualism also leaves open the nature of experience. Tracking dualists could join tracking physicalists in embracing identity intentionalism. But, unlike tracking physicalists, tracking dualists could instead opt for (metaphysical) adverbialism or the sense data theory to account for the nature of experiences.

To further elucidate tracking dualism, let’s consider its explanation of why you have E when you look at the kale salad. You are in a physical state that plays the e-role while tracking a physical property P. There is a match between that physical state and E, since P is what E phenomenally represents. On tracking dualism, Tracking Law dictates that whenever a subject is in such a physical state, she has such an experience. That, according to tracking dualism, is why you have E.

As for why you are conscious rather than unconscious, tracking dualists will say that you are conscious because Tracking Law holds and you are in some physical state that plays the e-role while tracking a physical property. And, they will say, you have E, rather than an experience with a different phenomenal character, because that physical state tracks P rather than any other physical property P* that would have instead yielded a match with an experience with a different phenomenal character.

3 A path from physicalism to tracking physicalism

I will now present a bolstered variation of what I regard as the best argument for tracking physicalism: Pautz’s (2019) argument alluded to in §1. It will serve as a useful contrast for developing the argument for tracking dualism in the next section. I will develop his argument in two stages, assuming physicalism throughout. The first stage uses a puzzle for metaphysically internalist physicalists to support tracking physicalism about spatial experience. The second generalizes from tracking physicalism about spatial experience to tracking physicalism simpliciter.

Here’s the scenario that generates the puzzle. Consider Swampy, a brain who in Swampman²⁴ fashion pops into existence as the result of an unlikely but nomically possible quantum fluctuation. Unfortunately, Swampy materializes in the absence of a body. But, on the bright side, the quantum fluctuation creates not only Swampy, but also a computerized vat that interfaces with Swampy. As it turns out, the computerized vat renders Swampy a narrow physical duplicate of your brain as you have an experience of triangularity. Now, let us stipulatively impoverish Swampy’s external physical connections to triangularity to the greatest extent allowable by the laws of nature that prevail in our world. For instance, (let us stipulate that, to the extent that the laws allow) Swampy does not stand in any significant discriminating physical relations to triangles. Swampy is not a triangular swamp brain. It has no history of interacting with triangles. Swampy is the only member of its kind, and so not a member of a kind that has a history of interacting with triangles. It admits of various forms of embodiment and exhibits great variation in behavior across the nearest worlds in which it is embodied. As a result, it’s not the case that Swampy would, if embodied, behave in a triangle-appropriate manner.²⁵

What is it like to be Swampy? Given the dearth of a priori connections between narrow physical states and experience, this question cannot be settled just by intuiting a verdict about the case.²⁶ But metaphysically internalist physicalists can answer the question without appealing to intuition. On their view, narrow physical duplication preserves phenomenology. So, since your brain and Swampy’s are narrow physical duplicates, metaphysically internalist physicalists will predict that they are phenomenal duplicates.²⁷ And since your experience phenomenally represents triangularity, they must say, so too does Swampy’s.

This is a puzzling result. After all, by construction, Swampy has been stripped of interesting physical connections to triangularity. Evidently, then, Swampy is not in a state that stands in any physical relation to triangularity with which the phenomenal representation relation might be identical. Yet metaphysically internalist physicalism entails that Swampy has an experience that phenomenally represents triangularity. Metaphysically internalist physicalists must explain how, despite being physically isolated from triangularity, Swampy manages to have an experience that phenomenally represents triangularity.²⁸ Given the dearth of physical relations that connect Swampy to that property, the only available explanation of how it has an experience that phenomenally represents triangularity is that it stands in a non-physical relation to triangularity. However, this explanation is unavailable to physicalists.²⁹

Can tracking physicalists do better? Yes. On tracking physicalism, Swampy can have an experience that phenomenally represents triangularity only by being in a physical state that tracks triangularity. But that is precluded by Swampy’s isolation from triangularity. Therefore, tracking physicalism predicts, Swampy does not have an experience that phenomenally represents triangularity. A fortiori, tracking physicalism is not obliged to explain how Swampy manages to have an experience that achieves such representation. Of course, there is nothing special about triangularity. This kind of puzzle case arises for the phenomenal representation of spatial properties more generally. Tracking physicalism provides a general treatment of such cases: the experience of spatial properties is essentially configured by the tracking of those properties. So, where swamp brains fail to be in states that track spatial properties, they fail to have experiences that phenomenally represent them. Hence, on tracking physicalism, there is no (unresolved) puzzle as to how swamp brains have such experiences. At least when it comes to spatial experience, then, there are grounds for accepting tracking physicalism over metaphysically internalist physicalism.

This concludes the first stage of the argument. To complete the argument, we need only extend tracking physicalism about spatial experience to other sorts of experience. The alternative would be to opt for a fragmented view on which tracking physicalism holds of spatial experience and some other theory holds of other experiences. Here is how Pautz implements this stage:

… The spatial argument indirectly supports externalism generally [via] considerations of uniformity. On the resulting view, sensible redness, like roundness, is an objective, mind-independent feature... In one natural version, it is a reflectance property… Further, we phenomenally represent sensible redness… in the same way we represent roundness, namely, by having a brain state that tracks it... This theory explains how, even in hallucination, we can ostensibly experience sensible colors in various locations and as conjoined with other spatial properties. Unless we are willing to accept sense data in a private mental space, or the mysterious “visual field regions” of Peacocke (2008), how else might we explain this? True, there are traditional arguments against the view that sensible colors are objective properties of external objects, concerning perceptual variation, spectrum inversion, and so on. But externalists (Dretske, Tye, others) have tried to answer those arguments. (2019: pp. 383-4; emphasis suppressed)

Let me bolster this stage of the argument by adding some points in support of this generalization move.

I start with a defensive point. The worry that tracking physicalism requires an implausible, objectivist account of sensible qualities can be further allayed by noticing that tracking physicalism is compatible with a range of options about the sensible qualities—its fate does not turn on whether sensible redness can be identified with a reflectance property. On the adopted broad understanding of ‘physical’, tracking physicalism is compatible with the view that sensible qualities are physical properties that are as they appear but which elude scientific detection. For instance, it is compatible with a non-reductive account of sensible qualities according to which sensible redness is an irreducible physical property grounded in certain reflectance properties.³⁰ Tracking physicalism can also be combined with a Russellian monist view of secondary qualities on which sensible redness is the categorical, quiddity-involving property underlying the dispositional reflectance profile that correlates with sensible redness.³¹

In fact, tracking physicalism is even compatible with a more traditional “colored-brain” form of Russellian monism that identifies sensible colors with quiddity-involving properties of the brain that we erroneously attribute to external objects. On this theoretical package, experiences as of triangularity would involve tracking an external shape property, while experiences as of blueness would instead involve tracking a property of your brain.

It might seem that the colored-brain form of Russellian monism would be at odds with using uniformity considerations to support the proposed generalization. For the resulting view would be disjunctive in that it would take some experiences to constitutively involve external properties while it would take others to constitutively involve properties that are not external. However, tracking physicalists who opt for this package could respond that experiences are all essentially of the same form: being in a physical state that plays the e-role while tracking property X. Many values of X (e.g. sensible redness) have their external status accidentally and just so happen not to be external in our world. But, according to this response, that is a “Cambridge” disjunction of a sort that generally infects natural classes, not a disjunction of the sort that uniformity considerations tell against.

To support the generalization, Pautz appeals to uniformity considerations. Only by flouting the canons of induction could one deny that such considerations support the move from tracking physicalism about spatial experience to tracking physicalism simpliciter. But two other considerations also support the generalization.

First, accepting tracking physicalism about spatial experience while declining to extend it is a recipe for metaphysical monstrosities.³² To illustrate, consider an experience as of a blue triangle. If tracking physicalism holds only for spatial experience, it must be that this experience is somehow a combination of a tracking state that accounts for its being an experience as of a triangle and some other state such as an adverbial modification that accounts for its being as of blueness. It is a point in favor of the generalization that it avoids the concoction of such chimeras. For once generalized to all experience, tracking physicalism can elegantly account for experiences as of blue triangles as follows: one has (say) an experience as of a blue triangle by entering a physical state that plays the e-role and tracks the property of being blue and triangular.³³

Second, the injunction to minimize basic, arbitrary-seeming posits lends support to the generalization.³⁴ To see this, notice that tracking physicalists can avail themselves of the following principle to systematically capture phenomenal-physical correlations. The principle is that E is an experience with phenomenal character C just in case E is a physical state that plays the e-role while tracking C. This principle may seem arbitrary. But this is just a reflection of the explanatory gap that all physicalists face. Once the principle is accepted, it can dispel the apparent arbitrariness of the specific physical-phenomenal identities it entails.

Other forms of physicalism typically fail to provide such a principle. As a result, they are left with a host of arbitrary-seeming physical-phenomenal identities.³⁵ Physicalists who opt for tracking physicalism about spatial experience but go in for another view about other experiences risk incurring a commitment to numerous identities of this objectionable sort. In the best case scenario, such physicalists will end up with multiple arbitrary-seeming principles that dispel the apparent arbitrariness of physical-phenomenal identities within their scope. Even then, since tracking physicalism gets by with just one principle of this sort, the injunction to minimize arbitrariness tells in favor of tracking physicalism.

4 A path from dualism to tracking dualism

The next item on the agenda is to think through a dualistic variation of the just considered argument. In parallel fashion, I will develop the argument in two stages, assuming dualism throughout. The first uses the puzzle about spatial experience as it arises for nomically internalist dualists to support tracking dualism about spatial experience. The second stage generalizes to arrive at tracking dualism.

To start, return to the Swampy scenario. As before, Swampy is a narrow physical duplicate of your brain as you have an experience that phenomenally represents triangularity. By stipulation, Swampy’s physical connections with triangularity are severed to the greatest extent allowable by law. Again, we can ask: what is it like to be Swampy?

Dualists are no more in a position to settle this question by intuition than physicalists are. But just as metaphysically internalist physicalism yields a prediction about the case, so too does nomically internalist dualism. For, on nomically internalist dualism, narrow physical duplication within the space of nomically possible worlds preserves phenomenology. Since your brain and Swampy’s are narrow physical duplicates and the same laws prevail in your worlds, nomically internalist dualists will predict that your brain and Swampy’s are phenomenal duplicates. And since your experience phenomenally represents triangularity, such theorists will say that Swampy’s experience does too.

At this point in the previous argument, metaphysically internalist physicalists faced the puzzle of how Swampy manages to have an experience that phenomenally represents triangularity, despite its physical isolation from triangularity. It may seem that nomically internalist dualists do not face any analogous puzzle here. For, as we saw in the previous puzzle, there is a non-physicalist solution to the puzzle: explain how Swampy manages to glom onto triangularity by positing a non-physical phenomenal representation relation that Swampy’s experience bears to that property. While no such move is available to physicalists, nomically internalist dualists, being unburdened by physicalistic scruples, can help themselves to this explanation.

One might worry that this explanation just pushes the explanatory bump under the rug. For we can ask how Swampy manages to have an experience that stands in a non-physical phenomenal representation relation to triangularity. But nomically internalist dualists can reply that psychophysical laws generate Swampy’s experience, and phenomenal representation of triangularity automatically comes with that sort of experience. (For instance, they might say that there is a law that just operates on the brain state you’re in when having the triangle experience which dictates that whenever something enters that brain state it has that sort of experience.) Apparently, since dualism requires some explanation of this form to explain how experiences phenomenally represent triangularity, invoking this explanation does not put nomically internalist dualists at a disadvantage relative to rival forms of dualism.

This appearance is illusory. There is a distinctive puzzle here for nomically internalist dualists. The puzzle traces to what I’ll call the systematicity constraint on psychophysical laws. Chalmers nicely captures this constraint in the following passage:

… The cornerstone of a theory of consciousness will be a set of psychophysical laws governing the relationship between consciousness and physical processes. These laws will tell us just what sort of experience will be associated with different sorts of physical process… An ultimate theory will not leave the connection at the level of “brain state X produces conscious state Y” for a vast collection of complex physical states and associated experiences. Instead, it will systematize this connection via an underlying explanatory framework, specifying simple underlying laws in virtue of which the connection holds… Ultimately, we will wish for a set of fundamental laws. Just as physicists seek a set of basic laws simple enough that one might write them on the front of a T-shirt, we should expect the same for a theory of consciousness. (1996: pp. 213-4)

I’ll understand the systematicity constraint as dictating that basic psychophysical laws be systematic, i.e. simple and powerful enough to yield a rich array of experiences on the order of what we find in our world. For nomically internalist dualists, this constraint entails that the psychophysical laws take the form of a systematic algorithm going from narrow physical states to experiences and hence to phenomenally represented properties. However, by construction, Swampy is physically isolated from triangularity. Swampy therefore does not stand in any physical relation to triangularity that might serve as an input to a systematic psychophysical law which generates its triangle experience. Admittedly, as we have seen, there are conceivable psychophysical laws that would generate Swampy’s triangle experience just by specifying that whatever has the relevant brain state has a triangle experience. However, it is hard to see how such laws could take anything but an unsystematic form, if they are to yield the vast and diverse array of spatial experiences that nomically internalist dualism predicts across the many Swampy isolation scenarios we can specify. Thus, nomically internalist dualists seem unable to explain Swampy’s spatial experience without running afoul of the systematicity constraint.

Given minimizing arbitrariness as a desideratum for psychophysical laws, an additional vice of nomic internalism is that it requires arbitrary psychophysical laws. This can be seen through variations of the Swampy scenario in which Swampy is a narrow physical duplicate of you as you have an experience as of an external property (such as that of being left-glove shaped) that is related by symmetry to another external property that you are not having an experience as of. Given the symmetry between these external properties, any narrow basis for representing one rather than the other would be arbitrary.³⁶

Tracking dualists can do better. On their view, Swampy has an experience that phenomenally represents triangularity only if it is in a physical state that tracks triangularity. Since Swampy is physically isolated from that property, tracking dualism predicts that Swampy doesn’t have an experience that phenomenally represents triangularity. The same goes for cases in which Swampy is a narrow physical duplicate of a subject who has an experience as of other spatial properties. A fortiori, tracking dualism does not lead to an explanation of Swampy’s spatial experience that runs afoul of the systematicity constraint. Moreover, tracking dualism about spatial experience satisfies the systematicity constraint with respect to spatial experience. For Tracking Law offers a simple yet powerful algorithm going from physical states to spatial experiences—this is what nomic internalist dualism needs but fails to provide. Tracking Law also yields a non-arbitrary basis for spatial experiences: when a subject experiences one asymmetric property rather than its mirror-counterpart, that is because the subject tracks the former but not the latter. The upshot is that, at least when it comes to spatial experience, tracking dualism is explanatorily superior to nomically internalist dualism.

So much for the first stage of the argument. To reach our target conclusion, we need to generalize from tracking dualism about spatial experience to tracking dualism simpliciter. As before, while the generalizing move is straightforwardly supported by uniformity considerations, it also enjoys several further sources of support.

First, Tracking Law is topic-neutral in that it yields spatial experiences without explicit reference to spatial experience. To restrict tracking dualism to spatial experience, evidently we would need to replace Tracking Law with a more complicated law that explicitly mentions spatial experience. Given that simpler laws are to be preferred, a fully general tracking dualism would therefore enjoy an advantage in accounting for spatial experience with respect to simplicity over a form of tracking dualism that is restricted to spatial experience.

Second, declining to generalize raises a binding problem for tracking dualists about spatial experience. To illustrate, suppose you have an experience as of a blue triangle, red square, and pink chiliagon. A theory of experience should be able to explain not only how you have experiences as of these shapes and experiences as of these colors, but also how you experience these shapes and colors as co-instantiated in the way that you do. Tracking dualism can meet this need by holding that you have that experience because Tracking Law holds and you are in a physical state that occupies the e-role while tracking the property of being an array with a blue triangle, red square, and pink chiliagon as elements. In contrast, it is hard to see how a fragmented view that adopts tracking dualism for spatial experience and some other view for experiences as of other sensible qualities such as colors might explain this.³⁷

Third, there is a dearth of dualist proposals for satisfying the systematicity constraint. This poses a serious problem for dualism that has been neglected by dualists and anti-dualists alike. The systematicity problem is that of at least sketching a dualist theory that satisfies the systematicity constraint. Without at least a sketch of this sort, it is doubtful whether psychophysical laws of the sort dualists quantify over could take anything but an unsystematic (e.g. list-like) form. We have already seen that tracking dualism about spatial experience satisfies the systematicity constraint as it applies to spatial experience. Correspondingly, once that restriction is lifted, tracking dualism satisfies that constraint in its full generality, as Tracking Law provides a simple algorithm that yields the distribution of experiences as a function of physical states. In contrast, it is not clear how tracking dualists about spatial experience who decline to generalize might go about solving the systematicity problem.³⁸ At best, they may find a tracking law for spatial experiences along with non-tracking psychophysical laws that systematically yields experiences as of other sensible qualities (and perhaps a binding law) that together yield the total distribution of experience. Even this sort of psychophysical legislation would be less parsimonious and less explanatorily powerful than one that gets by with Tracking Law alone.

Fourth, Tracking Law is also minimally arbitrary as far as psychophysical laws go. For it induces a mapping between physical states and experiences by exploiting a commonality between them, namely representation of the same properties. Restricting tracking dualism to spatial experience would sacrifice this virtue. For the restriction to spatial experience would be arbitrary. In addition, even if a non-tracking psychophysical law systematically generated experiences as of sensible qualities (perhaps by exploiting an isomorphism between a class of physical states and a class of experiences) they would not be guaranteed to minimize arbitrariness (e.g. fundamental “constants” might be needed to yield experiences as of one class of sensible qualities rather than a distinct yet isomorphic class of sensible qualities). Thus, in the absence of another way to minimize arbitrariness in psychophysical laws, there is pressure to extend tracking dualism about spatial experience to experience in general.

5 A tracking dualist research agenda

I have developed an argument that supports tracking dualism over nomically internalist forms of dualism. For dualists, the argument yields a presumption against locating the physical basis of experience in the head. The argument also suggests a potentially fruitful research agenda for dualists. To conclude, I highlight some items that deserve a place on such an agenda.

First, like tracking physicalism, tracking dualism is programmatic. Thus, there is the task of developing tracking dualism into a more comprehensive theory. The space of tracking dualist theories exhibits at least seven important dimensions of variation. They correspond to the following questions:

(1) What is the nature of the e-role?

(2) What is the nature of the tracking relation?

(3) What are the sensible qualities and how are they distributed?

(4) What is the (non-physical) nature of experience in our world? For instance, are experiences narrow states that lie in the head, or do they extend into the world like their physical bases?

(5) What is the nature of the phenomenal-physical nomic connection? For example, does the Tracking Law hold in virtue of brute physical-phenomenal regularities or is it instead an unexplained explainer that enables physical states to cause experiences?

(6) What, if anything, do experiences cause?

(7) What are the combination rules for experience? For example, can experiences as of tracked properties combine to yield experiences of properties that are untracked but which consist of tracked components?

If we take answering these questions as desiderata for developing tracking dualism, then each question can be used to constrain answers to others. For instance, what a given version of tracking dualism tells us about the nature of the tracking relation should cohere with what it says about the nature and distribution of sensible qualities. Similarly, what a theory says about the nature of experience should cohere with what it tells us about what experiences cause. For instance, a form of tracking dualism on which experiences themselves are wide would need to either embrace epiphenomenalism or explain how experiences manage to contribute to behavior without being excluded by narrow physical causes.

Second, there is also the task of assessing the viability of tracking dualism. Of course, this partly turns on whether dualism is viable. But for dualists, a more pressing issue is whether tracking dualism is susceptible to problems of the sort that have been raised for tracking physicalism and, if so, whether they can be overcome. Central among these are:

(a) “mismatch” cases—such as Inverted Earth, Swampman, coincidental variation, and e-role tampering scenarios³⁹—about which tracking theories are alleged to implausibly predict that experience comes apart from behavior or judgment in certain bizarre ways,

(b) objections to objectivist views of sensible qualities,⁴⁰ and

(c) objections from indeterminacy.⁴¹

It should not be assumed in advance that whether tracking dualists can solve such problems stands or falls with whether tracking physicalists can. The two theories incur different commitments and command different resources. For instance, as noted in §3.1, tracking physicalism inherits the counterintuitive consequences of identity intentionalism. Not so for tracking dualism, as it is not committed to identity intentionalism. And whereas it is open to tracking dualists to posit basic interactionist laws that render mismatch cases nomically impossible, no such move is available to tracking physicalists. In addition, tracking dualists could handle cases of indeterminacy with a stochastic rendering of Tracking Law: they could hold that the degree of determinacy with which the tracking condition for a given type of experience holds fixes a corresponding objective chance that that type of experience will be generated.⁴² In contrast, tracking physicalists who countenance such indeterminacy are hard pressed to avoid counterintuitive states at the borderline between experiences and non-experiences. This is not to say none of the same moves are available to both types of tracking theorists. Take, for instance, the familiar worry about an objectivist view of sensible qualities that identifies them with physically reducible properties (e.g. reflectance profiles), namely that those properties differ in similarity structure from the sensible qualities. It is open to tracking dualists and tracking physicalists alike to avoid this objection by opting for a different objectivist view of sensible qualities, one that construes them as irreducible or quiddity-involving properties with the required similarity structure.⁴³

Third, unless tracking dualism is shown to be untenable, there will also be the task of adjudicating between it and rival views. I have contrasted tracking dualism with tracking physicalism and with nomically internalist dualism. However, a developed form of tracking dualism will face competition from various quarters. This is because tracking dualism shares foundational assumptions of several major research programs in philosophy of mind. Tracking dualism’s commitment to explaining experience in terms of representation relations (namely, tracking and phenomenal representation) places it within the broad confines of the representationalist research program. But, as we have seen, tracking dualism is compatible with the rejection of some of the more ambitious views within that program. Per its commitment to explaining experience in terms of tracking, tracking dualism aligns with the program of naturalizing the mind. This alignment is, however, only partial, as the physicalistic aspirations of the naturalization project are incompatible with tracking dualism’s dualistic commitments. Those commitments lend to the hypothesis that experiences are basic phenomena which earn their keep in our theories by explaining other phenomena. In this respect, tracking dualism is aligned with the phenomenal intentionality program, which seeks to explain intentionality in terms of experience. However, tracking dualism’s rejection of nomic internalism and its commitment to explaining experience in terms of tracking are at odds with typical implementations of that program. Finally, tracking dualism perhaps finds its most natural home within the relatively underdeveloped research program of constructing a satisfactory dualist theory of experience and its place in nature.