The Gordian Knot of Consciousness Meets the Sword of Darwinism

Consciousness is a topic on which anything, it seems, can be said. The question “what is consciousness?” is a question as old as philosophy itself, if one interprets ancient Egyptian and Hindu texts with minimal charity. And despite the achievements of our technoscientific age, one can be forgiven for thinking that consciousness does not cohere with a properly scientific worldview. Indeed, it has recently been argued (Frank et al., 2024) that we shall need to drastically revise and expand what counts as “a scientific worldview” if we are to cease ignoring the role of consciousness as making experience, and therefore science itself, possible.

One of the chief difficulties in trying to say anything sensible about consciousness is that there are simply too many possibilities. Someone venturing into the marketplace of ideas looking to buy a theory of consciousness is confronted with a surfeit of options. Are only humans conscious? What about other animals? Could all life be conscious or harbor a kind of primitive sentience (biopsychism)? But why stop there, when everything that exists could be conscious in some way or other (panpsychism)? Or, for that matter, what if nothing is conscious—at least not in the sense that has generated worries about “the hard problem”? (This latter position, called illusionism, is considerably less absurd than its detractors make it out to be; see Dennett, 2015 and Frankish, 2016). The problem is not just that there are too many different competing positions, but that we seem to lack any criterion for deciding between them. It is not for nothing that some philosophers of mind decide to take up something on which there is at least the illusion of progress to be made, such as intentionality.

But now the Gordian knot of consciousness met the sword of Darwinism as capably wielded by Walter Veit in his A Philosophy for the Science of Animal Consciousness. He is far from the first philosopher-scientist to take up a Darwinian approach to consciousness, but his approach is richly informed by contemporary cognitive ethology and paleontology. Importantly, he avoids becoming mired in interminable debates about whether or not consciousness is a natural phenomenon. Rather, his basic starting-point is to say that if consciousness is a natural, biological phenomenon at all, then it must have an evolutionary explanation for why it evolved. ¹ Consciousness is so widespread in the animal world that it is very unlikely to be a “spandrel”—a by-product of selection acting upon other features—and if it an emergent property of deeper biological laws, it is mysterious as to what those laws might be. In any event, we have nothing to lose by considering the strongest possible case (to date) for the thesis that consciousness is an adaptation, in the strict sense of increasing the overall (personal and inclusive) fitness of those species that have it.

Veit’s book should be read as a contribution to what Godfrey-Smith (2001) calls “philosophy of nature.” A philosophy of nature, Godfrey-Smith proposes, is a distinct kind of intellectual project:

When undertaking philosophy of nature in my sense, a writer comments on the overall picture of the natural world that science, and perhaps other types of inquiry, seems to be giving us. But this commentary does not have to use language in the same way that scientists find convenient for their own work. It can use its own concepts and categories. These concepts should be ones suited to the task of describing the world as accurately and as precisely as possible when a diverse range of scientific descriptions are to be taken into account and when a philosophical concern with the underlying structure of theories is appropriate. (2001a, p. 265)

Thus construed, Veit joins recent work by Godfrey-Smith (2016; 2020), Sterelny (2003; 2012), and Dennett (2017), all projects that advance a resolutely naturalistic worldview to address ostensibly core features of human existence. ² That is, Veit’s theory is a philosophy of nature about consciousness.

I shall focus on Veit’s brief discussion and criticism of autopoiesis theory. This may seem a very minor quibble, but I think that it can be illuminating to consider Veit’s project in relation to autopoiesis and the closely related tradition of enactive cognitive science or enactivism. In doing so, I hope to contribute more bridges between otherwise disparate intellectual traditions whose advocates too often end up speaking past one another. To begin with, however, I shall briefly describe the underlying logic of Veit’s account.

The crux of Veit’s account is what he calls “the pathological complexity thesis”: to understand the teleonomic functions of consciousness, we need to consider consciousness in terms of the life-history of the organism in question. As Veit puts in the précis for his book (this issue):

All organisms designed by natural selection have an optimal life history strategy in the pursuit of reproductive fitness. The species-specific life history challenges they will encounter during their lives is what I refer to as “pathological complexity.” This should not be confused with complexity related to pathogens, but rather the living complexity of any system in trying to implement their optimal strategy. … And this can and must be extended to behavior and cognitive processes that ultimately need to pay off for the organism. This is why the ethologists such as Konrad Lorenz emphasized the importance of distinguishing pathological variations of behavior in order to extend the Darwinian revolution to include behavior … and my goal in this book was to do the same for consciousness: extend the Darwinian revolution once more.

In other words, the pathological complexity thesis holds that the strategy to naturalistically explain consciousness is the same as the strategy for naturalistically explaining behavior: by explaining why the trade-offs at work in the evolution of consciousness are such that consciousness tends to promote the personal and/or inclusive fitness of those organisms that have it.

With this context in mind, let us turn to Veit’s criticism of autopoiesis, which is first introduced as “a direct competitor to the pathological complexity thesis” (Veit, 2023: 53). Given that the pathological complexity thesis is a thesis about the kinds of selective pressures that drove the early evolution of animal consciousness, what makes autopoiesis a competitor to it? Veit’s answer seems to be that autopoiesis is a competitor to the pathological complexity thesis, not by virtue of being an alternative account of the selective pressures on early animals, but by virtue of being a very different kind of account: one that is “internalist” rather than “externalist.” As Veit sees it, autopoiesis was motivated by a concern with

the organismal phenomena of “viability,” “self-maintenance,” “self-organization,” “self-production,” and “self-reproduction.” It is these phenomena that they saw being neglected in an alleged object-like view from the Darwinian perspective that neglected the autonomy and subject-like nature of biological organisms. Life, the autopoietic tradition emphasizes, must be bounded and self-maintaining in order to allow its continuing existence and these features should accordingly be the center of our attention. (ibid.)

While this is true to a degree, somewhat more nuance is required to understand the ways in which autopoiesis theory is (and is not) a rival to Darwinian explanations in general and the pathological complexity thesis in particular.

To understand why autopoiesis theory might be construed as a rival to Darwinian explanations, it is important to notice that autopoiesis theory has a history of its own. Its founders, Humberto Maturan and Francisco Varela, had different intellectual formations and different routes toward and away from autopoiesis theory as they developed it in the early 1980s. More importantly, Varela himself, quite close to the end of his career, discovered the German phenomenologist Hans Jonas (Weber & Varela, 2002). It is through Jonas that Varela realized that autopoiesis theory allowed for a naturalization of teleology (Thompson, 2009). By contrast, the original formulation of autopoiesis theory was staunchly opposed to teleological and teleonomic explanations. Accordingly, I will refer to the foundational work of Maturana and Varela as autopoiesis 1.0 and the Jonasian turn—including the later Varela, Thompson, Di Paolo, and others—as autopoiesis 2.0. This distinction matters because Veit takes aim at both, but without appreciating some important differences between them.

Autopoiesis 1.0 attempts to answer the question: what is essential to life that makes life distinct from non-life, such that there can be a natural science of life—biology—at all? This question arises from a dissatisfaction with a mere laundry list of characteristics that all living things have, since it aims at answering why those traits are all bundled together. Just as importantly, however, Maturana and Varela held that a scientific answer to that question, in order to count as genuinely scientific, must take the form of a mechanistic explanation. What is needed is an explication of the specific kind of mechanism that is distinct of all and only organisms. Given a background commitment to the importance of mechanistic explanation in modern science, then the status of biology as an objective science that is irreducible to chemistry or physics depends upon specifying the kind of mechanism that makes something an organism. ³

There are two distinct aspects of autopoiesis 1.0 that are relevant to Veit’s criticism: the claim that all life is cognitive and the rejection of teleonomic explanations. It is the conjunction of these two claims that make autopoiesis 1.0 appear to be a rival to a teleonomic explanation of consciousness, hence a rival to Veit’s pathological complexity thesis. However, I am less sure that autopoiesis 1.0 is a genuine rival to Veit once these claims are properly interpreted, although autopoiesis 2.0 might be.

The claim that all life is cognitive is importantly different from claiming that all life is sentient or aware. Instead, the claim is that all life is cognitive in the sense of performing inductive inference: what has happened to a cell is expected to happen again. Far from being a rival to Veit’s claims about the biological origin and function of consciousness, it is a claim about how organisms (both single-celled and multi-cellular) perform computations in response to how they are perturbed over the course of interactions with their aspects of their niche. This anticipates Lyon’s thesis of biogenic explanation of cognition (2005) and the cognitive cell (2015), but it has little direct bearing on evolutionary explanations of hedonic evaluation as a dimension of consciousness. ⁴

Veit correctly notes that autopoiesis 1.0 (as I am calling it) rejects teleonomic explanations. But to be more precise, autopoiesis 1.0 rejects teleonomic explanations for what makes something an organism. Maturana and Varela give two arguments against teleonomic explanations. Firstly, they hold that evolutionary theory cannot be the conceptual basis for biology as modern objective science. They construe evolution as the cumulative effect of organismal autonomy within populations across time. But for that very reason, evolution presupposes organismal autonomy and thus cannot account for it. ⁵ Secondly, they insist upon a strict distinction between the cognitive realm of observer, which includes the organism in its relations with its environment, and the cognitive realm of the organism itself. ⁶

While there is clearly something right about this distinction, Maturana and Varela assert that that the organism’s cognitive realm is never its environment, but only its own responses to that environment. The relationship between the organism and its environment is in the cognitive realm of the observer, not that of the organism. Hence that relation plays no role in their scientific ontology of what makes something an organism, which in turn is the conceptual foundation of biology. Whereas Dobzhansky famously said, “nothing in biology makes sense except in light of evolution,” Maturana and Varela would have said, “nothing in biology makes sense, including evolution, except in light of autopoiesis.”

It is here that I think Veit makes an important criticism of autopoiesis 1.0. One of Veit’s explanatory desiderata is to avoid “an unfortunate dilemma between internalist and externalist approaches to consciousness, that is, that consciousness is to be explained through recourse of properties either internal to the organism or external to it” (p. 52). We need to be, to use a somewhat abused term, “dialectical” in how we understand the interdependence between whatever organisms are doing and whatever is being done to them—an interdependence that is diachronic and takes place at various timescales, some of them unfolding over development, and others at the level of evolutionary change. Veit is correct that autopoiesis 1.0 is not a theory of the relationship between organisms and their environments; it is a wholly internalist explanation. For precisely that reason, it is not entirely clear as to how Maturana and Varela saw autopoiesis 1.0 as offering a conceptual foundation for evolution. For it is one thing to assert with Lewontin (1983) that we need to conceptualize the organism as a subject of evolution as well as an object of it; it is quite another to assert that the environment plays no constitutive role in biological agency.

Regardless of how Maturana and Varela would have addressed this issue in the early 1980s, autopoiesis theory itself continued to evolve. This brings us to what I am calling autopoiesis 2.0. Autopoiesis 2.0 differs from autopoiesis 1.0 in two crucial respects: (1) autopoiesis is now seen as explaining the reality of teleology, rather than giving us reasons to reject teleological discourse as merely the observer’s choice of descriptive locutions and (2) the emphasis on biology as essentially cognitive has been revised to a conception of all life as affective and sentient. In both respects autopoiesis 2.0 reflects the phenomenology of life and philosophy of nature developed by Hans Jonas, especially in his The Phenomenon of Life (1966). Mario Villalobos and David Ward call this development “the Jonasian turn” within autopoiesis (Villalobos & Ward, 2016). ⁷ It is this turn that makes autopoiesis 2.0 appear to be in potential conflict with the pathological complexity thesis as an explanation for the rise of hedonic evaluation within the phyla that emerged during the Cambrian Explosion.

I emphasize the Jonasian turn here because it is this development within autopoietic theory that motivates Thompson’s version of biopsychism. It is as a Jonasian biophenomenologist that Thompson is at least willing to entertain the possibility that all life is sentient (2022). But precisely because this position is grounded in Jonas’s phenomenology of life, it is not clear how best to adjudicate its potential conflict with a philosophy of nature of consciousness. A philosophy of nature of consciousness is a philosophical construction grounded in (but not limited to) scientific explanations; it does not accord any epistemic privilege to phenomenological descriptions of subjective experience. Thus, whereas Veit sees Thompson’s willingness to take biopsychism seriously as being in theoretical competition with Veit’s restricted zoopsychism (since not even all animals are conscious, but only chordates, mollusks, and arthropods), I would suggest that Veit and Thompson are not even engaged in the same kind of project. Thompson’s project is not a philosophy of nature in the Godfrey-Smith sense.

To be sure, we still need a clear distinction between a philosophy of nature of consciousness and the Jonasian turn within autopoiesis. It should be stressed that Veit’s book is A Philosophy for the Science of Animal Consciousness and not A Philosophy of Animal Consciousness. That is, it is not a phenomenological inquiry into our experience of how animals experience their worlds; it is a conceptual framework for the scientific explanation of how animals experience their worlds. It is precisely because of the difference between those projects that we can raise the question as how, if at all, those projects can be related. It should be noted that some recent enactivist philosophers have taken up the term “philosophy of nature” from Peter Godfrey-Smith and applied it to the Jonasian turn within autopoietic enactivism (Gallagher, 2017; Meyer & Brancazio, 2022). I concur that there is some important truth to the idea that Jonas does have a philosophy of nature. ⁸ However, it is not clear whether this should be counted as a philosophy of nature in Godfrey-Smith’s sense, whereas Veit’s project much more clearly is.

Insofar as autopoiesis 2.0 belongs to a different tradition of theoretical biology than does Veit’s philosophy of nature, there is an important project in showing how and why they must be related. I believe that autopoietic enactivists would benefit from reading Veit’s book. All too often, perhaps due to what enactivists have taken over from phenomenology, terms such as “sense-making” are used with a variety of ways with little care for differences in context. Veit’s precise focus on what makes consciousness or sentience distinct from cognition more generally, and his argument for the ontological priority of hedonic evaluation as a dimension of consciousness, deserves the full attention of biophenomenologists, neurophenomenologists, and enactivists—not only for the similarities but also for the productive conversations to be had exploring those differences (Dennett, 1991; Maturana, 1979, Maturana & Varela, 1979).