p24

Many philosophers have argued that Presentism, the view that only present things exist, entails that our (basic) experiences must be instantiated at particular times, rather than over extended temporal intervals. As Dainton (2018) writes, “If reality is confined to a momentary present in the way that Presentists maintain, then it is difficult to see how [Positive Duration] can be true. Our immediate experience cannot extend through time if time itself has no extension.26 Time-Slice Internalism is therefore neutral on broader metaphysical questions about the nature of time, whereas a view like Minimal Positive Duration is not.27 Apart from these more metaphysical concerns, one could also motivate Time-Slice Internalism on more intuitive grounds. For example, insofar as one finds it natural to think of persisting physical objects (such as brains) as being ultimately composed of successive time-slices, perhaps it is similarly natural to think of persisting streams of consciousness as being ultimately 25 There’s an interesting question of how Time-Slice Internalism might relate to approaches to quantum gravity in which time and space are discrete, in which case fundamental physical properties would be instantiated over finite, discrete “units” of time rather than “instants” of time. While our formulation of Time-Slice Internalism does presuppose that time is continuous, it turns out that the very same arguments in favor of Time-Slice Internalism can also be run in a setting where time is fundamentally discrete. Those arguments would then support (say) Time-Unit Internalism, where Time-Unit Internalism is the view that phenomenal properties are instantiated at a single minimal “unit” of time, which (by The Temporal Inclusion Principle) would have physical correlates present at the very same unit of time. Here is how the argument would go. Let Extended Duration (analogous to Minimal Positive Duration), say that phenomenal properties must be instantiated over at least two units of time. Let Not Wholly Future Dependent* be the thesis that if someone is conscious at a unit of time, they must be conscious at that unit of time in virtue of instantiating a phenomenal property that does not extend into any times that are strictly later than that time (which is supported in the same way by the Two Rooms thought experiment). Then, Extended Duration and Not Wholly Future Dependent* would imply the Inductive Premise*, which says that if someone is conscious at some unit of time, then they are conscious at the immediately prior unit of time. This then implies that we have all lived forever as conscious beings. Just as we rejected Minimal Positive Duration in favor of Time-Slice Internalism, we can then reject Extended Duration in favor of Time-Unit Internalism. So, the arguments in this paper really support the disjunction of Time-Slice Internalism or Time-Unit Internalism, depending on whether time is continuous or discrete. 26 Dainton (2018) further supports this conclusion by the particular “Extensionalist” model of Positive Duration that he supports, in saying that “[I]f earlier and later stream-phases are experienced together, in the way that Extensional models require, then it seems very plausible to suppose that these phases must both exist. (Or to put it another way: an experience which no longer figures in the sum total of reality is not in a position to be part of the same unified state of consciousness as an experience which does so figure, any more than a non-existent brick can help hold up a wall.)” Frischhut (2017) similarly argues that Presentism is incompatible with Positive Duration. 27 See Markosian (2004), Bourne (2006), Zimmerman (2011), Ingram (2019), Emery (2019, 2020) and Builes and Impagnatiello (forthcoming) for defenses of Presentism. Perhaps the strongest objection to Presentism is due to the absence of absolute simultaneity in modern physics, but see Monton (2006), Barbour (2012), and Maudlin (2018) for different ways that absolute simultaneity might be supported by modern physics. Also see Builes and Teitel (2022) for a recent philosophical argument in favor of absolute simultaneity. If there turns out not to be a relation of absolute simultaneity in physics, then debates about the “temporal” structure of conscious experience will have to be significantly reconceived (e.g. see Lee 2007).15 composed of successive time-slices. One might also have phenomenological intuitions that support Time-Slice Internalism. For example, suppose your visual experience was gradually changing its phenomenal color between two times. It is natural to think that, for any time within that period, your visual experience would have some particular phenomenal color purely in virtue of how your visual experience is at that very time. At least in our more naïve moments, many of us are inclined to think of our conscious experiences as some kind of “internal movie”, playing directly before us. Just as ordinary movies have intrinsic states at particular times (just hit pause!), it is natural to think that our conscious experiences do as well. For all these reasons, we think Time-Slice Internalism is worth taking seriously. However, let us now turn to consider some of the major objections that it faces. 8. Time-Slice Internalism and the Experience of Change One major objection to Time-Slice Internalism is that (i) our most basic experiences have temporally extended contents (e.g. they represent positive temporal durations) and (ii) no instantaneous experience can have a temporally extended content. However, because both (i) and (ii) are very controversial philosophical assumptions, we don’t think that this is an especially strong objection to Time-Slice Internalism. The most popular way to resist this argument is by resisting (ii): the mere fact that an experience itself is not temporally extended does not imply that the experience cannot represent a temporally extended state of affairs. This kind of response, sometimes called the “retentionalist” model of temporal experience, has been defended by many different philosophers, including Broad (1925), Horwich (1987), Tye (2003), Grush (2006), and Pelczar (2010a). For skepticism about whether our basic experiences have temporally extended contents in the first place, see Chuard (2011) and Arstila (2018). Since we don’t have anything to add to the already very large literature criticizing (i) and (ii), here we will simply note that rejecting Time-Slice Internalism on the basis of (i) and (ii) is highly controversial. 9. Time-Slice Internalism and Possible States of Consciousness Another way to resist Time-Slice Internalism is to be skeptical of the possibility of a subject of experience that lasts for a single durationless instant. Prima facie, it is hard to make sense of someone having (say) a painful experience that doesn’t last for any positive duration at all. In order to address this kind of worry, consider the following case: Instant Twin: Suppose you are in pain during some temporal interval. Consider some time t within that time interval. Suppose further that a perfect physical duplicate of you at time16 t popped into existence at t and only existed for a single instant. Was your Instant Twin in pain at t? Of course, the assumption that you could have an “Instant Twin” like this is highly unrealistic. However, insofar as Time-Slice Internalism seems to imply that such a case is at least metaphysically possible, and that your Instant Twin would be in pain, this might seem like an implausible consequence of Time-Slice Internalism. A similar seemingly implausible consequence of Time-Slice Internalism can be illustrated with the following case: Frozen Twin: Suppose again that you are in pain during some temporal interval. Consider a time-slice physical duplicate of you at time t that is “frozen”. That is, consider a physical duplicate of you that is completely immobilized, forever (i.e.. every time-slice of the duplicate is exactly similar to every other time-slice). Is your frozen twin in eternal pain?28 Prima facie, it seems like Time-Slice Internalism implies that such a case is metaphysically possible, and that your frozen twin would eternally be in pain (since they have the relevant physical correlate for pain at every time), which again seems implausible. After all, feelings of pain are supposed to be associated with certain corresponding mental functions and behavioral dispositions (e.g. aversive behavior), which are all absent in the case of your Frozen Twin. In response, we do not think that Time-Slice Internalism is committed to these two implausible verdicts. To see this, we need to distinguish two different versions of Time-Slice Internalism. According to a “Humean” version, the relevant physical correlates of a phenomenal property at a time do not have any intrinsic causal powers, i.e. they do not necessitate any corresponding physical effects simply in virtue of their intrinsic properties. Instead, the relevant physical correlates can be freely modally recombined with any other physical states in their immediate past and/or future. Such a Humean version of Time-Slice Internalism does have the implausible consequences that your Instant Twin and Frozen Twin are (i) metaphysically possible and (ii) in pain. However, a Non-Humean version of Time-Slice Internalism does not have this consequence. According to this version of the view, the relevant physical correlates of any given phenomenal property at a time must be Non-Humean causal powers, which necessitate their corresponding effects wholly in virtue of their intrinsic properties. According to this view, if you made a perfect 28 This case is also discussed in Builes (2020). There, Builes argues that Phenomenal Internalists can avoid Frozen Twin if they endorse intrinsic rates of change. While this is true, below we will argue that Frozen Twin can also be avoided given a Non-Humean theory of causal powers. For more on intrinsic rates of change, see Arntzenius (2000), Lange (2005), and Builes and Teitel (2020).17 physical duplicate of a normal human being in pain, then that physical duplicate would have the very same Non-Humean causal powers of the original human being. These causal powers would then necessitate that the duplicate human being evolves through time in similar ways as the original human being in its immediate future, contrary both to Instant Twin and Frozen Twin.29 So, on the assumption that the physical correlates of consciousness have Non-Humean causal powers, it is dubious whether Instant Twin and Frozen Twin are metaphysically possible in the first place.30 In response to this point, however, one might worry that combining Time-Slice Internalism with the relevant version of Non-Humeanism decreases the overall plausibility of Time-Slice Internalism, given that the relevant version of Non-Humeanism is controversial.31 However, it turns out that there are entirely independent reasons why Time-Slice Internalism should not be combined with a Humean metaphysics. As has been argued by Hawthorne (2004) and Pallies (2019), anyone who accepts Phenomenal Internalism (which is implied by Time-Slice Internalism) must reject Humeanism. This is because every plausible account of the physical correlates of phenomenal properties implies that such physical correlates must have certain causal or dispositional properties associated with them, corresponding to (say) various mental 29 Because the duplicate might exist in very different environmental conditions (e.g. the duplicate might pop into existence in the empty void of space), it might be that these environmental differences make the duplicate quickly evolve in a different way than the original human being. However, because the physical effects of the environment on the duplicate will not be instantaneous (because causal signals cannot travel faster than the speed of light), the immediate evolution of the duplicate will be arbitrarily similar to the immediate evolution of the original. Consequently, because it is plausible to think that arbitrarily similar physical states correspond to arbitrarily similar phenomenal states, it is plausible that the duplicate will remain in pain for some non-zero interval of time. 30 We should address two worries about this proposal. First, one could worry that various macro-level powers or dispositions often times do not manifest their corresponding effects for various reasons (e.g. a fragile glass may not break when struck because it is covered in bubble wrap, a poisonous drink may not kill a victim when ingested because of a later antidote, etc.). However, non-Humean dispositional essentialists such as Bird (2007: 43-65) use these possible ways in which a disposition might not manifest its corresponding effects, because of the presence of “finks” or “antidotes”, to explain ceteris paribus laws that admit of exceptions. In the case of strict, exceptionless physical laws, dispositional essentialists should think that there can be no finks or antidotes that block the manifestation of fundamental physical dispositions. See Bird (2007: 60-65) for independent reasons to think that finks and antidotes are not present at the level of fundamental physics. For our purposes, we do not need to worry about these complications since we are considering a perfect physical duplicate, which will precisely duplicate all of the fundamental (and exceptionless) causal powers found in the original human being. A second worry is that, if the laws of physics turn out to be indeterministic, then it might be that the relevant Non-Humean causal powers don’t necessitate any particular effect, but rather they simply necessitate different chances for different effects. Our first response to this point is to simply note that, while it is widely assumed that quantum mechanical theories are indeterministic, in fact two of the three currently most popular interpretations of quantum theory are fully deterministic. While spontaneous collapse theories are indeterministic, Bohmian (‘hidden variable’) interpretations and Everettian (‘many worlds’) interpretations are fully deterministic. Our second response is that plausible indeterministic interpretations of quantum mechanics will still entail that Instant Twin and Frozen Twin have a chance of zero of occurring. For example, the objective chances in the GRW spontaneous collapse interpretation of quantum mechanics govern the particular positions of particles, but it is still nomically impossible (and therefore metaphysically impossible given the relevant version of Non-Humeanism) on such a theory that the collection of particles that make up Instant Twin spontaneously disappear. See Lewis (2016), Norsen (2017), Maudlin (2019) for introductions to different interpretations of quantum mechanics. 31 However, for defenses of Non-Humean causal powers, see Shoemaker (1980), Bird (2007), Strawson (2008), Heil (2010), Jacobs (2011), Demarest (2017), and Builes (2022, forthcoming).18 functions and behavioral dispositions. However, according to Humeanism, such causal properties are not intrinsic in the required way (i.e. according to Humeanism, perfect physical physical duplicates might have different causal and dispositional properties). So, in sum, we think that certain Non-Humean versions of Time-Slice Internalism can avoid the implausible implications suggested by Instant Twin and Frozen Twin, and because anyone who accepts Phenomenal Internalism already has antecedent reasons to reject Humeanism, we don’t think that appealing to Non-Humeanism in this way significantly reduces the plausibility of Time-Slice Internalism. 10. Time-Slice Internalism and Physical Correlates The final objection to Time-Slice Internalism rests on the worry that (i) the physical correlates of conscious experience are temporally extended and (ii) due to Phenomenal Internalism (or the Temporal Identity Principle), we should think that our conscious experiences are similarly temporally extended. By design, Time-Slice Internalism complies with Phenomenal Internalism, so this worry boils down to the empirical claim that the physical correlates of conscious experiences are temporally extended. In responding to this objection, we want to begin by acknowledging up front that this is perhaps the strongest objection to Time-Slice Internalism. Nonetheless, we still think that there are plausible ways that this objection can be resisted. The first point to make is that Time-Slice Internalism is only a claim about basic experiences, so it is perfectly compatible with there being temporally extended physical correlates of (non-basic and coarse-grained) temporally extended conscious experiences. So, Time-Slice Internalism does not necessarily imply that orthodox accounts of the physical correlates of consciousness are false so much as incomplete. In other words, even if our most basic conscious experiences (and their physical correlates) are more temporally fine-grained than contemporary accounts suggest, that does not mean that there aren’t important approximate temporally coarse-grained physical correlates of consciousness that contemporary neuroscience is uncovering. To consider an analogy with physics, the fact that Newtons’ theory of gravity was superseded by General Relativity is perfectly consistent with Newton’s theory being approximately accurate in a wide range of cases. Defenders of Time-Slice Internalism should think that our current understanding of the physical correlates of consciousness is in some ways analogous to Newtonian Mechanics: current empirical theories are giving us useful coarse-grained and approximate information about which physical processes go along with which conscious processes, but we do not yet have good19 empirical grounds to be confident about the physical correlates of our most basic phenomenal experiences. Second, there are principled methodological reasons for why we should expect our current science of consciousness to provide a temporally coarse-grained account of consciousness, even if it is ultimately temporally fine-grained. So, the fact that the current search for neural correlates of consciousness is leading to temporally coarse-grained theories shouldn’t count as much evidence against Time-Slice Internalism. The main reason for this is that the primary source of experimental evidence we have for phenomenal consciousness comes by way of our cognitive access to those states, and it is uncontroversial that our cognitive access to our conscious states, or “access consciousness”, must be temporally coarse-grained. For example, it is no surprise that we are unable to notice (and hence unable to make verbal reports about) features of our phenomenal consciousness that lie below certain time-scales. By itself, this empirical datum is entirely neutral on the question about the temporal grain of phenomenal consciousness. It is only when this empirical datum is combined with certain controversial philosophical assumptions, such as the assumption that phenomenal consciousness cannot be more rich than (or “overflow”) access consciousness, that we arrive at an argument against Time-Slice Internalism.32 In the absence of these philosophical assumptions, it is difficult to see how there could be any wholly empirical case against Time-Slice Internalism. Although in recent times there have been a number of methodological proposals for scientifically investigating aspects of our phenomenal consciousness that may (or may not) overflow cognitive access, these proposals remain highly controversial. As Phillips (2018) argues, “given our present data and methods, not only do we not know whether consciousness requires [cognitive access], we do not know how to find out. Until that changes, we must adopt an attitude of humility towards the phenomenal” (7). Thirdly, there are independent reasons for thinking that current theories of the neural correlates of consciousness must be incomplete anyway, as defenders of Time-Slice Internalism should believe. Earlier, we described how many philosophers are inclined to think that phenomenal consciousness cannot be vague. However, our current best theories of the physical correlates of consciousness inevitably appeal to vague concepts, such as “global workspace”, “higher-order thought”, “representation”, “neuron”, etc. Therefore, anyone who wishes to defend the view that phenomenal consciousness is not vague already has to think that contemporary accounts of the physical correlates of consciousness are only coarse-grained approximations. For all these reasons, we think there is ample reason to be skeptical that the empirical search for the physical correlates of consciousness has demonstrated that the physical correlates of our most basic conscious experiences are temporally extended. 32 For discussion on both sides of the debate, see Block (2011), Cohen and Dennett (2011), Brown (2012), Phillips (2016), Gross and Flombaum (2017), and Carruthers (2017).11. Conclusion 20 Having finished arguing against Minimal Positive Duration and in favor of Time-Slice Internalism, we would like to close by highlighting an important feature of the arguments we have been defending. One can very roughly categorize approaches to the metaphysics of consciousness as “inflationary” or “deflationary”. According to inflationary views, consciousness is usually taken to be fundamental, precise, and intrinsic. Examples of inflationary views include dualist, pan(proto)psychist, and idealist approaches to the metaphysics of consciousness (and perhaps versions of physicalism on which consciousness is precise). According to deflationary views, consciousness is usually taken to be a vague, reducible, and higher-level phenomenon, much like other biological phenomena such as photosynthesis or digestion. At a number of points in our defense of Time-Slice Internalism, we have appealed to various “inflationary” views about the metaphysics of consciousness. For this reason, we think that the best way forward for those who wish to uphold Minimal Positive Duration is to adopt a deflationary view about the metaphysics of consciousness. We hope this point illustrates our broader goal of showing how questions about the relationship between experience and time are inextricably linked to a variety of other central philosophical concerns.33