Thursday, April 16, 2020

Metaphysics and the Problem of Consciousness

In a recent post I talked about different frameworks for addressing the subjective dimension of consciousness. One path used ideas from philosophy of mind, the other looked to evolutionary biology. Of course, many who ponder solving this and related aspects of the mind-body problem take a more overtly metaphysical turn. Here I’ll briefly discuss why I don’t think these efforts are likely to get it right.

Against “vertical” metaphysical relations

My first post in this recent series was prompted by reading Philip Goff’s book presenting his panpsychist approach to the problem of consciousness.1 In the sections where he addresses the combination problem, Goff considers alternative strategies for situating a macro-size conscious subject in the world: several of these involve appeals to “grounding”. To sketch, grounding (in its application to ontology) is a kind of non-causal explanatory metaphysical relation between entities, with things at a more fundamental “level” of reality typically providing a ground for something at a higher level. For example, a metaphysician fleshing out the notion of a physicalist view of reality might appeal to a grounding relationship between, say, fundamental simple micro-physical entities and bigger, more complex macro-size objects. It’s a way of working out the idea that the former account for the latter, or the latter exist in virtue of the former. There are a variety of ways to explicate this kind of idea.2 Goff presents a version called constitutive grounding. He thinks this faces difficulties in the case of accounting for macro-sized conscious subjects in terms of micro-sized ones, and discusses an alternative approach where the more fundamental thing is at the higher level: he endorses a view where the most fundamental conscious entity is, in fact, the entire cosmos (“cosmopsychism”). In this scenario, human and animal concsciousness can be accounted for via a relation to the cosmos called grounding by subsumption. Goff motivates these various notions of grounding with examples that appeal to how certain of our concepts seem to be linked together, or to how our visual experiences appear to be composed.

Please read the book for the details.3 Here, I want to comment on why I don’t find an approach like this to be very illuminating. It is actually a part of a more general methodological concern I have developed over time. Certainly, trying to uncover the metaphysical truth about things is always a somewhat quixotic endeavor! But I think it is extremely likely to go wrong when done via excavation of our intuitions in the absence of close engagement with the relevant sciences.4 To make a long story short, I’ll just say that here I concur with much of Ladyman and Ross’s infamous critique of analytic metaphysics.5 But to get more specific, I have a deep skepticism in particular about the whole notion of synchronic (“vertical”) metaphysical relations. Not only panpsychist discussions but a great many philosophy of mind debates are structured around the idea that ontological elements at different “levels” are connected by such relations as part-whole, supervenience, or grounding. Positing these vertical relations, in turn, has contributed to confusion in debates about notions of (ontological) reduction and emergence. The causal exclusion problem, I believe, is misguided to the extent it is premised in part on the existence of these vertical relations.

I see no evidence that there are any such synchronic relations in the actual world investigated by the natural sciences (although they may characterize some of our idealized models). At arbitrary infinitesimal moments of time there exist no relata to connect: there are no such things as organisms, brains, cells, or even molecules. All these phenomena are temporally extended dynamic processes. Any static conception we employ is an artifact of our cognitive apparatus or our representational schemes. Reifying these static conceptions and then drawing vertical lines between entities at different scales is a mistake. My view is that all relations of composition in nature are diachronic.

Solve the problem with a new metaphysics of causation?

Given this, I think questions about how phenomena at different scales relate to each other involve a causal form of composition. So, one might ask whether thinking about the nature of causation help can with the problem of consciousness. Even before doing my own deep dive into research on the topic, I was drawn to those panpsychist approaches that explored this avenue. As mentioned in the earlier post, Russell’s account takes a causal approach to the structuring of subjects, although he himself doesn’t go on to offer a detailed theory.6 I think Whitehead’s speculative metaphysics can be characterized, at least in part, as an attempt to use a rich metaphysics of causation to account for the integration of mind and world. In more recent times, Gregg Rosenberg developed an account that found a home for consciousness in the nature of causation.7

Over time, however, I have also become skeptical of these more expansive causal theories. This is in spite of my view of the central role causation should play in any account of the composition of natural systems. Here, the problem is that these approaches go too far by baking in the answer to the mind-body problem from the beginning. Methodologically, I believe we should resist the urge to invent a causal theory that is so enriched with specific dualistic features that it directly addresses the challenge. For example, in Whitehead’s system every causal event (“actual occasion”) already has in place both a subjective and an objective “pole.” For Rosenberg, two kinds of properties (“effective” and “receptive”) are involved in each causal event, and this ultimately underpins the apparent dualism of the physical and mental. In contrast to these speculative solutions, we should be more conservative and pursue a causal theory that makes sense of our successful scientific explanations of natural phenomena, and then see how that effort might shed light on the mind. I’ll discuss my view on this in a future post.

1 Consciousness and Fundamental Reality. 2017. Oxford: Oxford University Press.
2 Here’s the SEP article on grounding.
3 Also, check out Daniel Stoljar’s review.
4 A quite different way metaphysics can go wrong is when those who are truly and deeply engaged with science (specifically physics) succumb to the tendency to (more or less) read ontology off of the mathematical formalism.  But that is a discussion for another time.
5 Everything Must Go: Metaphysics Naturalized. 2007. James Ladyman & Don Ross. Oxford: Oxford Univerisity Press. See. Ch 1.
6 At least this is true of The Analysis of Matter (1927), where the view now known as Russellian Monism was most fully developed.  In his later Human Knowledge: Its Scope and Limits (1948), he presents a bit of a theory via his account of “causal lines:” specifically, this comes in the context of an argument that such a conception of causation is needed to account for successful scientific inferences (part VI, chapter V).  As an aside: by this time, Russell seemed to come quite a long way toward a reversal of the arguments presented in his (much more cited) “On the Notion of Cause” from 1913. There, Russell argued that the prevailing philosophical view of cause and effect does not play a role in advanced sciences. Someone looking to harmonize the early and late Russell might argue that the disagreement between the two positions is limited: one could say the later Russell is developing causal notions that better suit the practice of science as compared to the more traditional concept that is the focus of criticism in the earlier article. However, I think it is clear that the later book’s perspective is quite a sea change from the earlier paper’s generally dismissive approach to the importance of causation to science.
7 A Place for Consciousness. 2004. Oxford: Oxford University Press. I have some older posts about the book.

Tuesday, March 31, 2020

A Panpsychist Surprise Ending

While I have been catching up on some reading relating to panpsychism lately (see recent posts here and here), I had turned back to thinking about another topic, the philosophy of quantum theory, when following up on some references led me to a surprise endorsement of panpsychism.

Specifically, I have long been interested in relational quantum mechanics (RQM), an interpretation first introduced by Carlo Rovelli in the 1990’s (good SEP article here).  I now suspect it is the interpretation that best fits with the theory of causation I am attracted to for independent reasons – but I will talk about that another time.  A philosopher whose views I find interesting and compatible with my own approach to thinking about quantum theory is Mauro Dorato.  He has a fine article discussing some of the philosophical issues raised by RQM (pre-print here; note there are some differences from the final article).1

Dorato, when discussing briefly the general stance that relations might be a fundamental metaphysical category, referenced a paper I had not read (though I imagine it is familiar to many who are interested in structuralism): “The Mathematical Structure of the World: The World as Graph,” published in 1997 (link) by Randall R. Dipert.2   Dipert, who passed away just last year (see here), was a scholar focused on a number of areas, including the thought of C.S. Peirce. The paper offers reasons to take (symmetric) relations forming (asymmetric) graphs as metaphysical bedrock (not expressed via logic or set theory, but as fundamental components in and of themselves).  In any case, while it is a wide-ranging and fascinating paper, I was surprised by the turn taken in the last paragraph:
There might at first seem to be no place in these cold graphs for minds, consciousness, and other mental phenomena unless, that is, everything is mental. Although within the dialectic of this essay it is wild and possibly irresponsible speculation, we should perhaps consider seriously the possibility that something like the pan-psychism of Spinoza, Leibniz, or Peirce is true, and that vertices are pure feelings (Peircean "firstnesses"), constituting a distinct thought or object only when connected to other such entities (358).
I take it Peirce was a panpsychist of sorts (see here), and evidently Dipert was as well.

1. Dorato, Mauro. 2016. Rovelli’s Relational Quantum Mechanics, Anti-Monism, and Quantum Becoming. In The Metaphysics of Relations, ed. Anna Marmodoro and David Yates. Oxford: Oxford University Press.
 2. Dipert, Randall R. 1997. The Mathematical Structure of the World: The World as Graph. The Journal of Philosophy, (94) 7, pp. 329-358

Wednesday, March 25, 2020

Different Approaches to Subjectivity

In the last post, I endorsed a Russellian approach to the mind-body problem (specifically the view labeled "panqualityism"), noting that one of the important tasks this framework leaves us with is explaining the subjective dimension of consciousness. This problem arguably requires less of a deep dive into metaphysical waters, but rather a consideration of the ways a naturalistic approach can tackle the phenomenon, once isolated from the other aspects of mind.

From Philip Goff’s book, Consciousness and Fundamental Reality, I learned that versions of panqualityism has been recently defended by philosophers Sam Coleman and Tom McClelland.  When it comes to addressing subjectivity, both draw upon what one might call the philosophy of mind toolkit.  Philosophers of mind have long been trying to understand conceptually how different aspects of our mental faculties might be understood.  The relationship of this activity to work in the relevant sciences (neurobiology, cognitive science) varies: a common approach for the philosopher is to try to keep up with the sciences, avoiding inconsistency, but then to proceed to theorize in advance of what is known (from the proverbial armchair) in pursuit of possible solutions.

In a recent article defending his approach, Coleman invokes higher-order thought (HOT) theory as a tool to understand the subjective character of conscious experience.1 HOT theory was put forward by David M. Rosenthal: roughly the main idea is that a mental state is conscious when we also have a thought about that state (the higher-order thought).  While HOT theory has faced many objections as an overall approach to consciousness, perhaps it can be applied to solve the relatively stripped-down question of what makes a mental state subjective. McClelland, on the other hand, invokes a self-representation model of subjectivity.2  Here (referencing work by Uriah Kriegel), the idea is that a state is subjective when, in addition to representing something else (say, an aspect of the external world), it also represents itself (unlike HOT theory, there is only one state involved rather than two).  The discussion of by the authors about how both ideas work out in conjunction with a Russellian framework is interesting and worth more consideration.

For now, I want to just make a meta-philosophical point by contrasting these approaches to subjectivity with a very different one. In a recent paper, Peter Godfrey-Smith examines the evolution of subjectivity as a biological phenomenon.3  One of the benefits of having a philosopher of science/biology like Godfrey-Smith working on the problem of mind is that it expands the territory of the possible solution space being considered.  This paper is very rich, and includes a great discussion of different biological features that appear relevant to the concept of subjectivity and their role in different kinds of organisms.  Given our common intuition that consciousness extends beyond humans at least to some extent, this kind of work can make a critical contribution for philosophy of mind.  I think it is especially relevant from the perspective of a divide and conquer strategy like panqualityism: a position that implies subjectivity is an aspect of phenomenal consciousness that should be relatively tractable to scientific explanation.

1. “Panpsychism and Neutral Monism: How to make up One's Mind,” In Brüntrup & Jaskolla  (eds.), Panpsychism: Contemporary Perspectives. Oxford University Press (2016).  A preprint can be accessed here.
2. In “The Neo-Russellian Ignorance Hypothesis: A Hybrid Account of Phenomenal Consciousness”, Journal of Consciousness Studies, 20, No. 3–4, 2013, pp. 125–51.
3. See (open access) “Evolving Across the Explanatory Gap,” Philosophy, Theory and Practice in Biology (2019) 11:1. See also this paper: "Varieties of Subjectivity."

Friday, March 20, 2020


I have been enjoying following some debates about the problem of consciousness on twitter and in blog posts.  In particular, philosopher Philip Goff has been tirelessly advocating the merits of panpsychism.1   As usual, this meets with a mixture of principled objections and more ungenerous responses.  I thought I would revive my blog to make a few comments, beginning with some reflections about my past thinking.  As a caveat, while I earned a PhD in philosophy recently (it's never too late!), it is not on this topic.2  My remarks are informal and tentative, and I invite corrections and reading suggestions.

My Long History with Panpsychism

I have always enjoyed puzzling over the mind-body problem.  It is one of the things that got me interested in philosophy early on.  In the 1990’s in particular, the explosion in attention the subject received (recast as the problem of phenomenal consciousness/the explanatory gap/the “hard” problem) had me riveted.

When I started my philosophy-oriented blog in 2004, it was one of my main topics, and I was a fan of panpsychist approaches. I reasoned that we know our own conscious experience exists (as well as or better than we know anything), and it doesn’t appear to be the kind of thing that can be scientifically explained using wholly non-experiential ingredients. I embraced the idea that in order to explain the character of our first-person conscious experience we must go beyond the relevant science to philosophically re-assess our assumptions about the fabric of the world.


While panpsychism is the common name for the idea that consciousness is a ubiquitous part of nature, on my blog I originally preferred to use “panexperientialism”.3  So how does panexperientialism differ from panpsychism?  The difference might seem to be just marketing.  “Panpsychism” implies that some inanimate things have minds, and that is a big affront to our usual intuitions: conscious minds seem to be something that only humans and some animals possess. Perhaps we can more easily picture “experience” as a phenomenon that might extend in attenuated fashion into the less complex building blocks of nature. A defender of panpsychism would reply by noting that the notion of experience certainly seems to require a subject of experience. And positing the ubiquity of experiencing subjects brings us right back to panpsychism.

But it seemed to me that there could be a principled difference lurking here.  While, despite some efforts at reading Whitehead, I was not well-versed in process philosophy myself, I took to heart the importance of resisting a static view of nature.  “Mind” and “subject” have static, object or substance-like connotations, while “experience” conveys the notion that consciousness is an ongoing activity. Panpsychists face a big hurdle (the much discussed “combination problem”) in explaining how small parts of nature that possess minds (particles, neurons) could combine to form a larger mind like ours. But if we view the natural world as a dynamic evolving web of processes, perhaps we can make sense of how a subject of experience could come about by degrees. The question, of course, is whether this suggestion can be fleshed out. In particular, it seems to put an emphasis on the need to understand the relationship of experience to causation (and causal forms of constitution) in the natural world.

Turning to Russell

Sometime in the mid-2000’s I went back and took a more careful look at some of Bertrand Russell’s later philosophical work on the problem, which was receiving more and more attention in the contemporary debates.  (This was in contrast to the relative lack of attention given to his old collaborator Whitehead’s process metaphysics, which most find very difficult to penetrate and assess).

Russell set out to show how careful attention to the way physical theories are constructed can reveal a very general common framework connecting what we think of as the physical and mental realms.4  First, he argues that the subject matter of physics can be interpreted as the abstract description of events and their linkage in causal relations. He then argues that the mental realm can likewise be described in terms of events, and that, given a causal theory of perception, we can view perceptual events (“percepts”) as connecting with physical events. And a key point is that our knowledge of the physical events (which is inferred and ultimately derived from observation) includes nothing which is known to be inconsistent with the mental. Russell speaks of events (or groups of events) as having “intrinsic qualities” or “intrinsic character”. Qualitative character is known to be an aspect of percepts, but is not part of physics, given its abstract structure. Russell doesn’t assert that events described by physics must have qualities like those of percepts – he is agnostic -- but he argues there’s no reason they couldn’t.5

In contemporary debates, these ideas are typically recast in the following way (although I think a bit is lost in this translation). The idea is that physical models seek to represent the relational, extrinsic, or dispositional properties of natural systems, but leave untouched the non-relational, intrinsic or categorical/qualitative properties. This latter sort of property is a fundamental aspect of nature that provides a suitable building block for underpinning consciousness in a way that can’t be done with the resources of a traditional physicalist metaphysics.

In his recent book defending panpsychism (Consciousness and Fundamental Reality6), Goff takes the Russellian framework to be a more promising alternative to physicalism, then describes how it can be further fleshed out in a number of different ways on the way to offering his preferred solution. These ways include panpsychist theories (the building blocks of nature are conscious subjects), and panprotopsychist theories (the intrinsic properties of the building blocks do not directly involve consciousness but somehow help give rise to consciousness).

Another view Goff discusses (but does not endorse) is panqualityism: this is the view closest to that of Russell himself.  This approach begins by noting that phenomenal consciousness actually has two distinctive problem-causing aspects: its qualitative character and its subjective character.  As Russell stressed, physical models, being formal and abstract depictions of a causal “skeleton”, do not capture the intrinsic qualities inherent in natural systems.  But we are acquainted with this aspect of nature via consciousness, so we conclude it is ubiquitous. As for the fact that consciousness is something subjective (that is, it is not public or third-person, but rather private or first-person), Russell posits that subjects are constructions from causally linked networks of elementary events (in the same way our familiar macro-objects are constructed).  He does not offer a satisfying fleshed-out account of this idea (and in fact gives no theory of causation – a concept he had once famously derided7 ), but a deflationary or reductionist approach to the subject is clearly intended. Again, we are pointed to thinking more about the nature of causation and its role in complex natural systems when thinking about the problem of consciousness.

Panpsychism is Half Right

Right now, my thinking about the mind-body problem continues to be very much in line with Russell’s.  My study of philosophy of science, especially on the nature and use of scientific models, has tended only to reinforce my view that there is something right about the central Russellian insight: a view of nature derived solely from the content of physical theories will fall short of accommodating consciousness because it doesn’t encompass qualities.

On the other hand, I suspect that some of the recalcitrant intuitions we have about the nature of conscious subjects (e.g. as irreducible things) can be defeased.  Questions about which natural systems are subjects and how this works can succumb to scientific inquiry, assisted by an understanding of how successful causal explanations (including constitutive causal explanations) work in the natural sciences.  I’ll look to follow up on this topic in a future post.

  1. See @Philip_Goff on twitter and the links on his website. Goff has a new book for a popular audience, called Galileo’s Error: Foundations for a New Science of Consciousness. By coincidence the copy I ordered arrived as I was writing this post. I shall read it shortly.
  2. My research is in philosophy of science, with a focus on causation and scientific explanation.
  3. The term is due to the Whiteheadian process philosopher David Ray Griffin.
  4.  The main source here is The Analysis of Matter (1927), Nottingham: Spokesman. See the SEP article on Russellian Monism here.
  5. I wrote in a bit more detail on Russell in the draft paper attached to this prior post.
  6. Consciousness and Fundamental Reality (2017). Oxford: Oxford University Press. See Chapter six.  Contemporary philosophers mentioned by Goff who pursue a panqualityist approach are Sam Coleman and Tom McClelland.
  7. In “On the Notion of Cause” (1912-13/1918). In B. Russell, Mysticism and Logic (pp. 142-165). London: Longmans, Green & Co.