Interviewer: Lynn Chiu

Lynn: David, I'm really excited about your new paper "Mind the matter: Active matter, soft robotics, and the making of bio-inspired artificial intelligence." It touches upon a lot of exciting topics in philosophy of mind and biology, with highly topical implications. Before diving into what I consider a core philosophical contribution of the paper-- a "multiple realization 2.0"-- in a few words, what is the paper about?

David:

The paper touches on a cluster of issues in the philosophy of life and mind sciences and, despite being a primarily theoretical project—more specifically, what we call a metatheoretical project—it is targeted towards a more ‘praxis’ oriented audience. My co-authors and I target a conceptual space in which critical ideas are configured, with our specific emphasis coming down on the nature of ‘materiality” and “matter.” In targeting this contentious area of philosophy, we are able to incorporate exciting developments in the fields of active matter physics, materials sciences, and basal cognition—areas that have received less philosophical attention than we believe they deserve.

Lynn: Why is the nature of “materiality” a philosophically contentious issue?

David:

A long and entrenched view of matter positions it as a predominantly passive feature, with only mind and mindedness being depicted as an active, plastic, and responsive feature of the mind-body interaction (here, the body is crudely associated with matter). A corollary of this view is that a very sophisticated cognitive architecture is required for ascriptions of cognition to be justified. The kind of architecture under consideration here is, of course, the highly centralised nervous systems found in humans and the ‘higher mammals.’ However, in pulling from the above adumbrated disciplines, we try to show that the materiality and corporeality of minded creatures displays a more sophisticated set of capacities and behavioural dispositions than is often appreciated. Specifically, we look at work in basal cognition which highlights how comparatively ‘simpler’ creatures, lacking the architectures often associated with cognition, still exhibit several distinctively ‘cognitive’ capacities: such as memory, learning, goal-directedness, and so on.

The distinctive move in our paper comes from focusing on the interactions and interconnections between these capacities and the material substrate on which they depend. Stated differently, these systems do not exhibit obvious dissociations between their behaviour and being the bioenergetic systems they are. Among other things, this is meant to emphasise a more essential link between ‘matter’ and ‘mind’ than is commonly supposed. This is ultimately why we found an engagement with the multiple realisability (MR) literature so central to our argument: the idea that the ‘matter matters’ sits entirely in opposition to this version of MR.

Lynn: Let’s talk about this. In the paper, you propose a “multiple realization 2.0.” What is it?

David:

To understand the main motivations and implications behind our argument, and therefore to understand Multiple Realizability 2.0, it is necessary to first say a bit about multiple realisability more generally. Basically, MR is the theoretical thesis—and sometimes hypothesis—that cognition is a ‘substrate independent’ phenomena: a kind of software that, like the name suggests, can be instituted and realised on diverse sets of hardware. However, just how diverse this range of substrates can be is a matter of great debate and discussion. On the more liberal view, there are no constraints on the types of materials that can manifest cognitive capacities of interest: “we could be made of Swiss cheese and it wouldn’t matter”, as Hilary Putnam once famously put it.

Lynn: This is Multiple Realizability 1.0, right? What’s wrong with it?

David:

Because MR, as originally presented, fails to apply the appropriate constraints, the literature is replete with fantastical scenarios that are meant to show the dissociability between mind and matter. Not just Swiss cheese, but constellations of beer bottles and other pell-mell materials can be considered as having minds on this account! This is sometimes called the liberality problem. We need to put some constraints on MR, but we also need to avoid committing so-called ‘biochauvinistic’ errors: the idea that only biological materials are capable of entertaining mindedness. It is precisely by incorporating advancements in soft robotics and materials sciences that we hope to show how minds are distributed in the world.

Multiple Realizability 2.0 is an account that tries to allow for a rather wide distribution of mindedness and its possible recreation (in ML or AI) while still placing constraints on the kinds of systems under examination. We incorporate insights from soft robotics and active matter physics to show the interdependencies imposed by being a soft matter system and having a mind. This comes down on what is called “existential needs.” One thing we want to encourage in this paper is that there are certain demands placed on being a system with existential needs and the mindedness associated with it.

Lynn: How would a more nuanced understanding of mind and matter help the multiple disciplines working on robotics and AI?

David:

One of the insights we take from active matter physics is that this comes down on being the kind of material system that can maintain far-from-equilibrium thermodynamic states. I believe this renewed investigation into active matter and soft robots benefits the field by placing constraints on the space of possible minds in a way that is amenable to our wider, blue-sky goals of a kind of ‘artificial general intelligence’.

Lynn: Thank you so much, David, for unpacking the central thesis of the paper. For our final question, I’ll turn to the researcher behind the paper. What are the inspirations and motivations that drive your work?

David:

This piece is very personal to me (even if I can already see so many in which its exposition and delivery could have been improved!), mostly because it represents a culmination of ideas I have been thinking about since well before I began a Master’s program dedicated to this topic. I have always been interested in what distinguishes (paradigmatically) living systems from machines—why in the former you find such an interminable (or seemingly interminable) will to live and ‘care’ that you just don’t find in non-living systems.

One of the initial inspirations for exploring this was through the angle of phenomenology, specifically Heidegger’s existential analysis of ‘care’ as a constitutive feature of human subjects. But why stop at the human? Heidegger had scant interest in the interiority and mental lives found in non-human animals (and ultimately the restriction to the human subject is ultimately what erects a barrier for me in taking these ideas further), but the contention that the existentially fundamentally different feature of living systems is a kind of ‘centre of concern’ is captured in philosopher John Haugeland’s quip that the problem with AI is that ‘it doesn’t give a damn’. Thus, crudely and colloquially, I was interested in what made organisms generally—and humans specifically, you and I—‘give a damn’.

These questions broadly motivated me to pursue postgraduate studies at the University of Edinburgh. To my surprise, the inner go of the brain was of only passing discussion in my various (philosophy of) cognitive sciences courses. Instead, what I saw was highly formal and abstract depictions of mentation that seemed to not have an eye towards its underlying substrate. The 20^th century German philosopher Dilthey once commented on Kant’s transcendental philosophy as being devoid of life and living activity: “no real blood flows in the veins of the knowing subject constructed by Locke, Hume, and Kant”. At best, living activity was treated as a kind of ‘behind-the-scenes set-up’, as Michael Levin puts it, with little relevance for understanding cognition and mind.

Lynn: If we zoom out to take a bird's-eye view of your projects and near-future plans, how would you situate this paper in your greater scheme of things?  

David:

Zooming out this piece is situated within two wider theoretical trends: one more pedigreed and one more new-fangled. The former is the life-mind continuity thesis most famously associated with Maturana & Varela’s theory of autopoiesis, though it can be found in thinkers like Whitehead, Bergson, and Schelling. My biggest inspiration here probably comes from Schelling, though elaborating on that would probably exhaust the remainder of our time!

The latter wider theoretical project would be what is called ‘new materialism’, which puts matter and materiality as an active participant in the social process of knowledge making and subject making. However, it has since been extrapolated to encompass a wider physicalist ontology for understanding emergence, top-causation, and relationality: concepts that are critical for understanding the emergence of subjectivity, agency, and mind in nature. Our paper tries to push the conversation further in this direction, and I believe it helps to further emphasise, along with the above theoretical pools, the importance of our corporeality in understanding the nature of mindedness. What happens when we insert the living blood into the knowing subject constructed by, e.g., Locke, Hume, and Kant? In some sense, this question is a beacon for our wider investigations.

Lynn: Thank you for the fantastic insights, David!

Publication: