Seth provides a wonderful guide to the complex landscape of debate concerning the scientific explanation of consciousness. But in so doing, he makes an assumption that I would like to call into question. The assumption is that there is a meaningful difference between the study of consciousness and the study of intelligence and cognition more widely construed.
Seth begins by distinguishing consciousness from “intelligence” and by glossing intelligence as being about “what a system can do,” whereas consciousness is about what it feels like to be that system. At that point, he approvingly invokes Nagel’s (Reference Nagel1974) discussion of “what it is like-ness.” By way of contrast, consider the position known (somewhat misleadingly in my view) as illusionism (Clark, Reference Clark2019; Clark, Friston, & Wilkinson, Reference Clark, Friston and Wilkinson2019; Dennett, Reference Dennett2016; Frankish, Reference Frankish2012, Reference Frankish2019). It is a mistake, illusionists suggest, to think of conscious experience as involving the presence of special “intrinsic properties” where these (Dennett, Reference Dennett and Dennett1998a) are features of experience that outrun, or that could outrun, our own “reactive dispositions” – our capacities to discriminate, respond, react, sort, sift, judge, describe, and so on.
The picture of experience as involving such intrinsic properties (“qualia”) is, of course, at the heart of Chalmers’s (Reference Chalmers1996) easy problem/hard problem distinction. Notice that if you mean something else by “qualia” (or by “phenomenology” or “conscious experience”), then you need have no quarrel with illusionism at all – for a nice analysis, see Frankish (Reference Frankish2012). As Dennett (Reference Dennett and Dennett1998b) once wrote:
“I am not denying the reality of consciousness at all; I am just denying that consciousness, real consciousness, is like what they think it is like” (p. 139).
If the illusionist is right, then consciousness is not special. It is just a label for a large and complex set of reactive dispositions. There is no hard problem because there are no “classic qualia” – experiential features that might, in principle, be present, absent, or different without that difference showing up in some profile of responses and reactions. That profile (in the human case) includes our reports of puzzlement and of experiencing “ineffable qualitative properties,” and anything else that is actually testable by scientific means (see Dennett, Reference Dennett1991, on “heterophenomenology”).
Adopting the illusionist stance in this way rules out the total failure of functionalism. It is, however, consistent with the potential failure of what Seth labels “computational functionalism.” It might, for example, turn out that to get some specific functional profile you need to build your large-scale system out of smaller systems, each of which has its own “interests” and “goals,” echoing the nested structure of living organisations. Or it may be that certain other properties of living matter itself are vital. What matters, however, is to keep sight of the fact that the only way we can test such ideas is by looking for profiles of response and reaction and asking what substrates best support them.
The illusionist route resolves what seems to me to be a tension in Seth’s own picture of what “consciousness” is. On the one hand, Seth seems to identify “consciousness” with classical appeals to “what it is likeness.” But such appeals pick out something distinctive only on condition that the “what it is likeness” is not exhausted by a complex response profile of some kind (see Frankish, Reference Frankish2012). On the other hand, Seth later nudges us towards what he calls the “real problem of consciousness,” which, as I understand it, is solved whenever an account manages to get to grips with enough standard data points (patterns of behaviour, reaction, and report) to deliver improved capacities to predict, control, and intervene. This is an illusionist-friendly picture of the scientific study of consciousness.
Of course, nothing in this area comes for free. To carry through the illusionist project requires solving Chalmers’s (Reference Chalmers2018) “meta-problem” – the problem of explaining why our own conscious experience appears (to us) as so puzzling. What are the tricks that our own minds are playing on us that push us towards seeing conscious experience as so very weird and puzzling as to be apparently resistant to all known means of cognitive scientific attack? We do not yet have a full and convincing solution to the meta-problem, but for some attempts, see Clark (Reference Clark2019) and Clark et al. (Reference Clark, Friston and Wilkinson2019). In many ways, Dennett’s (Reference Dennett1991) work was really a full-blooded attempt to resolve the meta-problem too, by tracking the many mistakes we make when we think about experience as occurring in a kind of “Cartesian Theatre.”
Taking the illusionist stance radically alters the discursive landscape. It rules out scenarios in which qualia might be present, absent, or different without any clear means of telling, and thus keeps the study of consciousness on a firm scientific footing – exactly as Seth’s own “real problem” formulation suggests. The scientific study of consciousness, tackled in this way, targets a complex high-dimensional space of abilities and capacities. Within this space, there are many important divisions. For example, some of our human reactive patterns involve advanced meta-cognitive capacities while others – including many of those shared with non-human animals – do not. Given the complexity of this space, there will (as Dennett long argued) typically be no simple “fact of the matter” about the presence or absence of a given “qualitative” state. Instead, different animals will turn up as occupying different locations in a high-dimensional space, with many criss-crossing patterns of similarity and difference.
In sum, I suggest that Seth makes his project unnecessarily hard by endorsing a picture of consciousness which uneasily straddles the broadly behavioural (where that includes everything we could ever actually test or probe) and the elusive “what it is likeness” notion – a notion whose defining feature is nothing other than its built-in (zombie-breeding) resistance to the kinds of scientific approach that Seth ultimately champions. Even if deep properties of life turn out to be crucial, we should not set up the initial puzzle by identifying “what-it-is-like-ness” as a special kind of target, distinct from what systems do and can do.
Seth should join the (badly named) illusionists!
Seth provides a wonderful guide to the complex landscape of debate concerning the scientific explanation of consciousness. But in so doing, he makes an assumption that I would like to call into question. The assumption is that there is a meaningful difference between the study of consciousness and the study of intelligence and cognition more widely construed.
Seth begins by distinguishing consciousness from “intelligence” and by glossing intelligence as being about “what a system can do,” whereas consciousness is about what it feels like to be that system. At that point, he approvingly invokes Nagel’s (Reference Nagel1974) discussion of “what it is like-ness.” By way of contrast, consider the position known (somewhat misleadingly in my view) as illusionism (Clark, Reference Clark2019; Clark, Friston, & Wilkinson, Reference Clark, Friston and Wilkinson2019; Dennett, Reference Dennett2016; Frankish, Reference Frankish2012, Reference Frankish2019). It is a mistake, illusionists suggest, to think of conscious experience as involving the presence of special “intrinsic properties” where these (Dennett, Reference Dennett and Dennett1998a) are features of experience that outrun, or that could outrun, our own “reactive dispositions” – our capacities to discriminate, respond, react, sort, sift, judge, describe, and so on.
The picture of experience as involving such intrinsic properties (“qualia”) is, of course, at the heart of Chalmers’s (Reference Chalmers1996) easy problem/hard problem distinction. Notice that if you mean something else by “qualia” (or by “phenomenology” or “conscious experience”), then you need have no quarrel with illusionism at all – for a nice analysis, see Frankish (Reference Frankish2012). As Dennett (Reference Dennett and Dennett1998b) once wrote:
“I am not denying the reality of consciousness at all; I am just denying that consciousness, real consciousness, is like what they think it is like” (p. 139).
If the illusionist is right, then consciousness is not special. It is just a label for a large and complex set of reactive dispositions. There is no hard problem because there are no “classic qualia” – experiential features that might, in principle, be present, absent, or different without that difference showing up in some profile of responses and reactions. That profile (in the human case) includes our reports of puzzlement and of experiencing “ineffable qualitative properties,” and anything else that is actually testable by scientific means (see Dennett, Reference Dennett1991, on “heterophenomenology”).
Adopting the illusionist stance in this way rules out the total failure of functionalism. It is, however, consistent with the potential failure of what Seth labels “computational functionalism.” It might, for example, turn out that to get some specific functional profile you need to build your large-scale system out of smaller systems, each of which has its own “interests” and “goals,” echoing the nested structure of living organisations. Or it may be that certain other properties of living matter itself are vital. What matters, however, is to keep sight of the fact that the only way we can test such ideas is by looking for profiles of response and reaction and asking what substrates best support them.
The illusionist route resolves what seems to me to be a tension in Seth’s own picture of what “consciousness” is. On the one hand, Seth seems to identify “consciousness” with classical appeals to “what it is likeness.” But such appeals pick out something distinctive only on condition that the “what it is likeness” is not exhausted by a complex response profile of some kind (see Frankish, Reference Frankish2012). On the other hand, Seth later nudges us towards what he calls the “real problem of consciousness,” which, as I understand it, is solved whenever an account manages to get to grips with enough standard data points (patterns of behaviour, reaction, and report) to deliver improved capacities to predict, control, and intervene. This is an illusionist-friendly picture of the scientific study of consciousness.
Of course, nothing in this area comes for free. To carry through the illusionist project requires solving Chalmers’s (Reference Chalmers2018) “meta-problem” – the problem of explaining why our own conscious experience appears (to us) as so puzzling. What are the tricks that our own minds are playing on us that push us towards seeing conscious experience as so very weird and puzzling as to be apparently resistant to all known means of cognitive scientific attack? We do not yet have a full and convincing solution to the meta-problem, but for some attempts, see Clark (Reference Clark2019) and Clark et al. (Reference Clark, Friston and Wilkinson2019). In many ways, Dennett’s (Reference Dennett1991) work was really a full-blooded attempt to resolve the meta-problem too, by tracking the many mistakes we make when we think about experience as occurring in a kind of “Cartesian Theatre.”
Taking the illusionist stance radically alters the discursive landscape. It rules out scenarios in which qualia might be present, absent, or different without any clear means of telling, and thus keeps the study of consciousness on a firm scientific footing – exactly as Seth’s own “real problem” formulation suggests. The scientific study of consciousness, tackled in this way, targets a complex high-dimensional space of abilities and capacities. Within this space, there are many important divisions. For example, some of our human reactive patterns involve advanced meta-cognitive capacities while others – including many of those shared with non-human animals – do not. Given the complexity of this space, there will (as Dennett long argued) typically be no simple “fact of the matter” about the presence or absence of a given “qualitative” state. Instead, different animals will turn up as occupying different locations in a high-dimensional space, with many criss-crossing patterns of similarity and difference.
In sum, I suggest that Seth makes his project unnecessarily hard by endorsing a picture of consciousness which uneasily straddles the broadly behavioural (where that includes everything we could ever actually test or probe) and the elusive “what it is likeness” notion – a notion whose defining feature is nothing other than its built-in (zombie-breeding) resistance to the kinds of scientific approach that Seth ultimately champions. Even if deep properties of life turn out to be crucial, we should not set up the initial puzzle by identifying “what-it-is-like-ness” as a special kind of target, distinct from what systems do and can do.
Seth should join the (badly named) illusionists!
Financial support
This work was carried out with the support of European Research Council, Synergy Grant (XSCAPE) ERC-2020-SyG 951631.
Competing interests
None.