Neo-rationalism
Jean-Pierre Caron, a philosopher and an artist, was a guest speaker at one of GAIA’s meetings, “Artificial Intelligence and Arts Group”, held at the Innovation Center of the University of São Paulo (USP).
For that meeting, he suggested we’d discussed “What is Philosophy?”, an article by Iranian philosopher Reza Negarestani.
That article was my introduction to Reza’s work and, mainly, to the so-called Neo-rationalism.
Getting to know the project, or rather, the concept of, which advances the possibilities of building knowledge and evolving thoughts, has really enlightened me.
I invited J.-P Caron to present the edges of Neo-rationalism.
I am also launching an interview-based version of News From the Future, in which a guest will answer three questions.
1.Could you try to synthesize the neo-rationalist thought and comment on how it influences the conception/understanding of the human?
I don’t really think there is a unified neo-rationalist thought. So-called neo-rationalism is basically a label that is used to name a certain set of thinkers (which includes at least Ray Brassier, Reza Negarestani, Pete Wolfendale, Daniel Sacilotto and the Laboria Cuboniks collective), that were influenced to a greater or lesser degree from a subset of rationalist concerns that were emerging at the time of Speculative Realist wave. Of course, there are more specific descriptors of what neo-rationalism could be beyond this initial genealogy, but I can get into that more easily by first revisiting it (1).
Speculative Realism (from now on, SR) started in 2007 at the Goldsmiths conference of the same name in London, which was attended by one of the thinkers I just mentioned — Ray Brassier who is the English translator of Quentin Meillassoux’s Après la Finitude, the book that is credited as the inspiration for the so called “movement”. I use inverted commas to talk of it as a movement, since it is highly disputed if there was in fact something like a unified Speculative Realism (2). But I am using the moniker here as a convenient way to refer to a certain intellectual milieu more than to a set of recognizable doctrinal points.
Going back to the Meillassoux book, rehearsing all the arguments in that book is beyond the point here, but basically what Meillassoux proposed, counteracting the hegemonic tendency in the European Continental Philosophy, was a form of realism that would support the reality of primary, mathematizable properties. The book proposed a way out of what it baptized correlationism, which was conceptualized by Meillassoux as the idea of a thought sealed within in its own conditions, not being able to go beyond what is transcendentally guaranteed. Thought never advances beyond what is in Kantian philosophy described as what is yielded through the transcendental conditions of experience- which determines the difference between phenomenon, that which appears to us; and noumenon, the thing in itself beyond the correlation of thinking and being. Meillassoux’s critical hypothesis is that this was somewhat replicated ever since in different guises in the history of Western thought: be it the transcendental conditions of experience (Kant), the thematization of acts of conscience by which objects are constituted (Husserl), or the language by which we organize our experience (Wittgenstein). Correlationism, therefore, is the name that was given to the hegemonic tendency since Kant to think of ourselves as beholden to our own conditions of access to the world, losing the world (according to Meillassoux, “the great Outside”) itself of sight.
In trying to get out of the correlationist strictures, Meillassoux drew the consequences of an argument: It is a scientific fact that we can date events prior to the presence of any living being therefore prior to the actuality of any transcendental structures. It is the arche-fossil, or information that comes to us from a time prior to the time of the emergence of life — the accretion of the Earth, for example. For Meillassoux, if this datum is thinkable, and it is by its very nature not contained in the access conditions of any living being, it would mean that the Real is thinkable beyond the correlation itself. From there he went to derive the capacity of mathematics to pierce through being itself, with surprising novel consequences.
The rationalist element in Meillassoux’s approach was an important ingredient in Brassier’s philosophy, first exposed in his book Nihil Unbound. In it, not the past arché-fossil, but the future extinction of the universe for all intents and purposes considered in our scientific horizon today as inevitable, and thus the ability of thought to thematize its own disappearance, is taken as a demonstration of the independence between the interests of thought and interests of life. Rational thinking would be capable of sustaining the idea of its own disappearance, and the book seeks to trace the consequences of this mutual indifference between thought and the world.
That argumentation was important insofar as the battle against correlationism, then prevalent involved trying to extract thought from the correlational circle. And, in doing so, Brassier deepened the idea of the world’s disenchantment, as the conceptual representation would therefore not be one about a world pre-formatted for it, akin to a reservoir of transcendentally guaranteed meanings for the living, but instead a product of a process of confrontation with something that is fundamentally indifferent to that which confronts it. What is key to retain in order to understand the so-called neo-rationalism that appears later is the idea of the autonomy of rational thought and its “Promethean” capacity (“Prometheanism” is used as another name for neo-rationalism) to overcome finite conditions, first of thought itself, then of the human, on one hand and, on the other hand, the disconnection of reason and the horizon of meanings taken as “given” — which untethers reason from the “reasonable”, a theme that we will explore later on.
The idea of Prometheanism anticipates the post-human topic that finds its most developed form until now within the neo-rationalist constellation in the works of Reza Negarestani. In “The labor of the inhuman” (3), Negarestani launches his rational inhumanism, which is the result of sustaining two ideas contained in the addition of the prefix “in” to humanism: the idea of the non-human, a superseding of the present image of the human, but also the idea of an “inside” of the human, of “traversing through” the human. Inhumanism, then, is the immanent development of humanism in such a way as to overcome it from within, following the inferential threads that hide beneath the very concept of the human. In a very simplified way, Negarestani mobilizes an apparatus of philosophy of language — Robert Brandom’s inferentialism — that gives priority in the identification of the meaning of a concept to its inferential relations with other concepts. This means that, unlike more traditional philosophical approaches to semantics, the explanatory order of priority is not derived from reference — the object that falls under the concept– but from the determination of concepts by the inferences they enable and in which they partake, and which are inspectable by the assertoric sentences that we can justifiably produce with these same concepts. And these inferences are also a function of the discoveries that can be made about the object of the concept. Thus, a discovery of a new property of a certain chemical element — for example, “gold dissolves in aqua regia” — implies the addition and/or revision of the previous inferences in which the concept of that chemical element participated (“dissolving in aqua regia” now participates in the conditions of use of the concept “gold”).
But more interesting than this role of empirical discovery in determining concepts is their properly logical unfolding. One of the most interesting consequences of inferentialism is that we use concepts without strictly knowing all that they entail — thus, all that constitutes them. If the concept is a node in a network of commitments organized by the inferences in which it partakes, we can always discover new inferences, which implies an amplifying and revisable process embedded in the use — which is not external but is constitutive of the conceptual content. Under that light inferentialism also weaves closely the semantic and the pragmatic dimensions of language- an assertion is a rule-bound doing that amounts to the undertaking of the commitment expressed by its content, which in turn results from the inferential commitments embedded in the contents of each of its constitutive concepts. As a commitment, once undertaken, one ought to also undertake the consequences that it entails.
Something like this reasoning is applied by Negarestani in “The Labor of the inhuman” to the very concept of the human — as concept that trigger a revisionary process that leads to a transformation of the concept of Human”. If the index of the autonomy of reason in 2007’s Brassier was that it was able to uphold a world in which it was rational to sustain that life itself would disappear- as we have seen, untethering the interests of life and the interests of thought- and trying to collect the rational conditions and consequences to support this thesis; in 2014’s Negarestani, the index of the autonomy of reason is not extinction, but the transformation of the Human through the action of the Concept. That idea is the inception of Negarestani’s concern with artificial intelligence as a philosophical way to inspect the cognitive composition of the human, which appears most clearly in his 2018 book Intelligence and Spirit.
In a certain sense, the underlying problem changes from 2007 onward, also because it is no longer a matter of guaranteeing a form of realism, but of working immanently within the processes inherent to rationality. At this point, some form of correlationism is also recovered — not an ontological correlationism, but an epistemic one — that is, the process of knowledge depends on the cognitive apparatus: it is not by negating the role of mediations in the knowledge of the outside that one gains traction on the world, but by the patient uncovering of the mediations that are constitutive of thought, their limits and affordances. The authors of the so-called neo-rationalism basically develop different approaches to these immanent processes that constitute reason itself, in a process of knowledge and intellection of its world(s).
From this very brief exposition, I think one can draw a proto-definition of neo-rationalism as a philosophical research platform that tries 1- to recover the centrality of reason (and not of the reasonable — its parochial and limited version) for the conditions of intelligibility of thought in general; 2- tracking and sustaining the most radical possible consequences of this centrality for the human figure and for the reconstruction of the world, from which political consequences can be drawn; 3- the treatment of this reason in compositional terms, which in some versions of neorationalism considers cognition as a form of computational processing (4), seeking to identify and explore the space of restrictions and affordances that produces an economy immanent to thought, intelligent behavior and representational capacity. Here we find an inversion of what we had said in the previous paragraph: if artificial intelligence appeared in “The labor of the inhuman” as a consequence and product of the autonomy (and automation) of reason, here in turn artificial intelligence, through computationalism, would provide us with an external model of an agent, through which we can inspect and reverse engineer the rational constitution of the human being as such — it becomes the means of construction and not just the product of inferential unfolding. This is the function of the Toy Model in Negarestani’s book Intelligence and spirit.
The three ideas proposed above can be summarized and complemented by this definition proposed by Wolfendale: “To reject all rational intuition in the name of reason, in order to insist that not only there is no intuitive faculty of rational knowledge, but that there is no intuitive purchase of reason’s own structure, possibilities, and limits. Reason is not what you think it is. Reason is not rationalization. Reason is not reasonable.(5)”
Thus, the neo-rationalist project would therefore relate not only to the centrality of reason, but to its expansion and self-transformation, in such a way that reason and reasonableness are not coextensive, but in some aspects opposite: the reasonable is consistent with common sense and an established understanding of how to deal with life and the real; the rational may include the reasonable, but it ultimately surpasses and redefines it, as it is not subject to a given, “intuitive” image of the possibilities of reason (6), but the process by which reason is defined is also the process by which it is elaborated beyond the operative coordinates of the currently rational. This concern with the rational self-transformation of the human also appears in political projects related to the emergence of neo-rationalism, and it is not a surprise that it has intersection points with Srnicek and Williams’ work on Left Accelerationism (“Manifesto” for an Accelerationist Politics” (7), 2013) and the collective Laboria Cuboniks’ on Xenofeminism (“Xenofeminism — a politics for alienation” (8), 2015), which would continue to be reflected in the individual works of some of the members of that collective, such as artist/ writer Patricia Reed, who has produced several important essays since then (9).
2. How does neo-rationalism share a concern with meeting the needs of non-cognitive, non-conscious, and non-representative processes?
Taking up the cue from the end of the previous answer, I can develop an answer to this question in two more or less divergent lines. The first would be simply the fact that an approach to the structure of mind would necessarily have to include non- or pre-conscious elements that constitute and/or condition the cognitive, conscious, and representational capacities in rational agents. Neo-rationalism was heavily influenced by the American philosopher Wilfrid Sellars (although there’s been recently a drift towards other intellectual influences) who was very active in the late 50s and 60s and was precisely concerned with the functional and the material structure that conditions rationality itself. As Sellars himself says, quoted by Brassier, “the task of ‘transcendental logic’ is to explain the concept of a mind that obtains knowledge about a reality of which it is a part. The acquisition of knowledge by such a mind involves it being the object of the action of, or ‘affected’ by, the objects it knows.(10)” But, more radically, this includes not only being affected but being constituted by the material processes of which, in fact, one seeks to obtain knowledge.
The problem for Sellars is that you would have no way of putting yourself “outside” your own cognitive abilities to describe your precognitive or material constraints. The 3rd person point of view have to be constituted within the resources of our 1st person point of view. This maps onto the distinction that Sellars proposes between a “Scientific Image” and a “Manifest Image” of the human being in the world, respectively. The manifest image roughly corresponds to the self-understanding we have of ourselves as rational and self-determined individuals, and the commonsense ontology that accompanies this image, composed of molar objects, animate and inanimate beings, and people. The scientific image corresponds to the clumping of the forms of description that science has historically produced — culminating in a “particulate” image of reality Sellars describes in “Philosophy and the Scientific Image of Man”. One of the difficulties that Sellars sees in this is bringing both images together in a “stereoscopic vision” since nothing like a self-determined rational individual who obeys rational norms of inquiry can appear in the scientific image at this stage. But if the scientific image proves to have greater explanatory and predictive power, thus seeming to be true, how can we solve the paradox that this scientific image is itself the result of a thought the structural features of which are describable with the resources of the manifest image, and for which the normativity proper to reason is indispensable? This difficulty gives rise to a bifurcation between philosophers influenced by Sellars — namely, those who develop the rational infrastructure of the manifest image, among whom are Brandom and his inferentialism; and those who develop the consequences of the scientific image, including eliminativists who would advocate the overcoming of the manifest image by the scientific image through the replacing of the animic vocabulary of emotions and mental states of the manifest image by neurophysiological vocabulary, such as Patricia and Paul Churchland envisage.
Finding the Archimedean point between the two images is a way of posing the problem of the precognitive, pre-representational, and preconscious constitution of cognition, representation, and consciousness without taking the dogmatic step of presupposing these conditions. These must be derived within the cognitive, conscious, and representational processes themselves, even though their objects are causally prior to and external to these spheres. In other words: the inferential-cognitive capacities one sees in the Manifest Image would have to be accounted for by the resources of the Scientific Image, which in turn arises from the framework of the Manifest Image. This is the critical and transcendental restriction to not presuppose dogmatically the world-constitution without engaging in the way the mind works out these constraints.
In Brassier’s particular case, and here I can place the second line of answer, there has been an engagement with the thought coming from the Marxist tradition in recent years. While it is debatable whether the engagement with Marxism is still thinkable within the confines of neorationalism, here, too, an element that would externally constrain cognition arises — namely, the abstraction of value, its constraints, and impingements upon thought. Brassier’s research has been trying for several years to combine a Sellarsian approach to representation with Marxist approaches to social form, complexifying the constitutive elements of reason to encompass real abstractions that are historical products of the capitalist mode of production. So, in addition to material conditions of individual cognition, the material conditions embodied by social practice appear as elements that constrain and determine reason, which at the same time are retroactively describable by reason within certain historically determined frameworks — in a way analogous to retroactive description, contained in the manifest image, of the a-subjective processes that give rise to reason, within the scientific image, through the resources of reason itself.
3. In your opinion, what is the impact of artificial intelligence and neuroscience on the evolution of neo-rationalist thought? Do they interfere, contribute, if they contribute in what way?
I think this question was partially answered above, but it might be worth developing a bit further. We have already touched on two ways in which AI appears, for example, in the work of Reza Negarestani: 1. As a result of the artificialization of the human under the thrust of the rational elaboration of its concept. 2. As a part of this elaboration itself, as a model of a rational agent that’s external to ourselves, with which we can test possible cognitive structures, not only to reproduce our own — which makes the artificial agent an instrument for the understanding of ourselves — but to build novel structures.
This touches on a point I have emphasized little so far. It seems to me that neo-rationalism is not only interested in adequately describing the mediations that constrain and constitute reason in all ideal, material, social spheres; but it is also interested in enhancing these cognitive abilities. If Marx said that philosophers have hitherto been trying to understand the world and the task would be to transform it, Brassier in “Concepts and objects” tells us that the inability to transform the world may be related to the inability to understand it. Thus, the project of understanding the conditions for the autonomy of reason would intersect with the project of liberating this same reason, and our capacities for intelligibility in general — which goes through Artificial General Intelligence, but also through research in logic, construction of worlds (“worldmaking” as stated by Nelson Goodman), analysis of capitalist sociality, and political organization.
Finally, neuroscience and the philosophy of mind provide explanatory hypotheses for this same mediation between mind and material support, to be critically evaluated and assembled by neo-rationalism, even though this takes them beyond the proper scope of those disciplines, joining the way in which computer science has addressed the same and related themes. The choice for functionalism, for computationalism, and for the debate with eliminativism (and not necessarily the adoption of its positive thesis) delimits a specific scope within neuroscience research that has been more interesting to neo-rationalism than other areas, for presenting more acutely the problems that concerns it: the idea of multiple realizability of the mental over the material and therefore an untethering of mindedness from an understanding in terms of a thinking “substance” (of functionalism), the explicitly algorithmic approach that seeks to trace the multiscalar relationships between lower-level operations, meso-level and higher-level functions in terms of types of computations (something that has a complex history and different models, from Turing-machines to the interactive paradigm in computationalism), (11) and the naturalization hypotheses to be accepted or rejected, which is provoked by the debates with eliminativism.
Regarding the adoption of functionalism and computationalism, it is important to stress that “(i)n reality, neither functionalism nor computationalism entails one another. But if they are taken as implicitly or explicitly related, that is, if the functional organization (with functions having causal or logical roles) is regarded as computational either intrinsically or algorithmically, then the result is computational functionalism.”
What this recipe yields is the possibility of a fine-grained multi-level analysis of the composition of a mind in terms of its multiple functionalities in their variegated relationships to computational operations and types. It is this compositionality that is crucial for Negarestani, without which there is no analysis of the mind in terms of its realizability out of its “natural habitat”- which entails the transformation of its meaning and substantial conditions of realization. It is in that sense that AGI is interesting for neorationalism- as an outside model of cognition by which we can gain knowledge of our own cognition, opening the black box of experience and motivating its radical reformatting.
______________________
- One other important historical thread leading to neorationalism would be the philosophy of the Cybernetic Culture.Research Unit in Warwick in the 90s, that was an important formative influence for both Brassier and Negarestani. But that would make us drift even farther from the main subject.
- See Brassier “Postscript: Speculative Autopsy” in Wolfendale, P. Object-oriented philosophy: the noumenon’s new clothes, London, Urbanomic, 2014, pp. 409–421
- Part 1 https://www.e-flux.com/journal/52/59920/the-labor-of-the-inhuman-part-i-human/ Part 2 https://www.e-flux.com/journal/53/59893/the-labor-of-the-inhuman-part-ii-the-inhuman/ In PT-BR https://zazie.com.br/wp-content/uploads/2021/05/REZANEGARESTANI_ZAZIEEDICOES_2021.pdf
- As presented in this text by Wolfendale: https://deontologistics.co/2021/05/20/tfe-what-kind-of-computational-process-is-a-mind/
- https://deontologistics.co/2018/02/11/on-neorationalism/
- Mark Fisher, in his period that was closest to what would later come to be neo-rationalism — namely his Cold Rationalism from 2004:
The anomalous is not irrational.
Reason is not common sense.
Rationality does not reveal a world that fits the human operating system’s scan pattern. That’s why Cold Rationalists are psychotic
7. https://criticallegalthinking.com/2013/05/14/accelerate-manifesto-for-an-accelerationist-politics/
8. https://laboriacuboniks.net/manifesto/xenofeminism-a-politics-for-alienation/
9. For instance : https://www.e-flux.com/architecture/artificial-labor/140674/xenophily-and-computational-denaturalization/
In portuguese: https://aestheticmanagement.com/writing/xenolia-e-desnaturalizacao-computacional/
10. Sellars apud Brassier https://www.glass-bead.org/article/transcendental-logic-and-true-representings/
11.The recipe that combines functionalism and computationalism is well approached in Negarestani, R. “Revolution Backwards: Functional realization and computational implementation” In: Pasquinelli, M. Alleys of your mind: augmented intelligence and its traumas. Meson Press, 2015. Pp 139–156. https://meson.press/wp-content/uploads/2015/11/978-3-95796-066-5_Alleys_of_Your_Mind.pdf
It is critically assessed by Daniel Sacilotto in “A thought disincarnate: what does it mean to think?” https://www.glass-bead.org/research-platform/a-thought-disincarnate/
________________________
J.-P. Caron is a philosopher and artist based in Rio de Janeiro, Brazil. He is an associate professor of philosophy at the UFRJ (Universidade Federal do Rio de Janeiro) and an instructor for The New Centre for Research and Practice. He’s also been practicing noise and experimental music for more than 15 years and manages with friends his own label, Seminal Records. His recent work deals with problems of rationality, scale-sensitivity in several sectors of experience, and social ontology.
Sanannda Acácia is a Brazilian artist that works in the medium of both sound and image. In recent years, she has performed with her solo projects Quasicrystal and Insignificanto, and in collaboration with other artists in the projects MUTA, zEros and several others. She is a member of the experimental music label Seminal Records. For the past 9 years, she has been performing at festivals and events in the Brazilian underground, as well as having worked and collaborated with art spaces and collectives such as Fosso.
