[UPDATED - after an illuminating Twitter discussion with Pete Wolfendale and Shane Denson, I've updated the post to include their necessary and helpful contributions. Warning - its 3130 words or so.]
So I’ve been pretty rubbish at blogging lately – you can blame fatigue / Speculations work / extra research for that and a whole bunch o’ personal stuff – more on that soon.
Anyway – its nice to finally have a spare evening in which I can write about some recent thinking, that probably won’t go anywhere. So I have a couple of essays coming out on Speculative Realism and aesthetics, in which I discuss and interrogate, what I see as the major divide between the “original four” of the 2007 conference – what I call (borrowing Meillassoux’s distinction in that conference) the difference between Demonstration and Description. I am of the view that in order to properly apply SR to other domains of interest (and for me that includes aesthetics and computer science especially) one has to locate this divide both historically and present within them. In my essay for And Another Thing (coming out soon) I briefly show how this division applies in aesthetic terms, and how they potentially divide Kant’s historical legacy, both aesthetic and philosophical in half. In Dublin, I also briefly showed how this may apply to Computer Science too (for me, the two histories are both intertwined philosophically);
In the first instance there is ‘Demonstration’ the primacy of epistemological ‘fact’ or ‘knowledge‘: that a passive, inert material reality can be epistemologically demonstrated and known through the formal, inferential properties of thought and an extrinsic principle of the fact, so that thought becomes radically divorced from a non-anthropomorphic being. This position is shared both by Meillassoux and Ray Brassier through their joint commitment to explain the truth of reality rationally using the skeptical tools of correlated thought. For Meillassoux it is absolutely clear that reality must only be thought, not taken as given (otherwise it is simply correlated). Instead of appealing to the givenness of laws or predictability of reasonable conjecture, correlative thought must be demonstrated to achieve knowledge of the absolute contingency of everything, where the only law is absolute lawlessness. Similarly as much Brassier would like to cut himself off from Speculative Realism, he too is implicated into Demonstration just as much as Meillassoux. Brassier’s recent development of Wilfrid Sellars‘ work, is at once helpful and a startling extension to his work in Nihil Unbound; he adopts most of Sellars’ rejections including his Myth of the Given, in which knowledge is to be gained from self-analysis of the meta-lingustic functioning of language games. From this, Brassier follows Sellars’ gung ho synoptic project into fusing the inorganic manifest intentional image-function with the visionless, natural non-anthropomorphic scientific image-function. This, it seems, gives Brassier the leverage to develop Sellars’ absolute processes inherent in cognitive behaviour, as a reconciliation of both images. The crucial, ‘demonstrative’ point, is that this can only be done by rigorously analysing the normative structure (rule-behaviour) of cognitive order, developed from (and through) its material achievement and development. For Brassier, pragmatic conceptual function rules. Whether through function or fact, both Meillassoux and Brassier believe that reality can only be demonstrated.
Yet, in complete opposition, there is ‘Description’: the primacy of ontological realities in their own right. For this side reality is composed of fundamental entities, objects, things, forces and powers which are ontologically no different in kind than the epistemological limits of cognition and, moreover, they exist in their own right. This is an intrinsic principle of the thing. The correlated relationship between thinking and being is radicalised into entity-specific things, with substantial essence on the side of individual objects (Harman, Levi Bryant), a second empiricism of networked actants (Bruno Latour) or the panpsychist, vitalist dynamism of active matter on the other (Iain Hamilton Grant, Steven Shaviro). For Harman, as we know – the in-itself is not a primordial lump, nor a product of human understanding, but a mezzanine of entities, which can only be described. But this description is not all there is, for other entities are defined by their own logics of descriptions and this is the OOO gamble, as it were – how do other entities, including humans describe? Grant on the other hand, does not locate his metaphysics in bodies, or individuals but in Nature – and moreover the irreducible products of nature as its own history. Both Harman and Grant have noted (in The Speculative Turn and elsewhere) that the inanimate world is the crucial orientation which they both share – not direct access through epistemological knowledge, but (in Grant) through a thinking of a single ground upon which thinking is a product (knowledge for Grant is for the most part, a possible contingent product of Nature itself). Both Harman and Grant agree on at least showing that the ontological primacy of reality, in its own right, is metaphysically prior to the theories of the knowledge of the natural sciences, and thus not handmaiden to it. Reality can only be described.
We all know this division of course, even if its only second-hand. What is important here is the utter incompatible nature of both orientations, precisely insofar as correlationism was a pre-synthesis of both. No middle way is possible, because correlationism was that middle way. Like a Hegelian dialectic stuck in reverse gear, Speculative Realism ‘fractures’ the correlate into these two halves and only these two. Fuse both Demonstration and Description together and you arrive back at correlationism; that the world cannot be known directly, and only be internally related by human thought. Speculative Realism, then, simply is this incompatible splintering, this fracture, and its existence emerges from this rejection. Demonstration argues that the correlate can fruitfully prove or deduce knowledge of itself and the world, whilst Description argues that the human correlate never fully deduces anything, and is no different in kind from anything else. Once the correlate is rejected, there is no middle way to stake a claim apart from these two broad orientations of ‘what is’; either reality must be epistemologically demonstrated or reality must be described in terms of real ontological variance. Once once side is chosen, the other recedes from view.
However to simply leave this distinction as it is, only delivers half of a development. For a while now, I’ve been trying to understand and develop a second axis which can account for the differences between Meillassoux and Brassier’s Demonstrative side but equally account for the differences between Harman and Grant’s Descriptive side in equal measure and application. Such a division would go some way into understanding how such a fourfold could arise and mould into four very different orientations. I sincerely doubt that what I’ve come with with, ‘sticks’ so to speak, yet I cannot apply anything else which equally matches. So, feel free to shoot me down, should the model fail to preach what needs to be preached.
So critically the first axis, which one could sufficiently call the ‘Kantian’ or the ‘decorrelationist axis’ is relatively simple: either a Speculative Realist chooses the primacy of knowledge to deduce inert material reality, or one simply chooses the primacy of inanimate ontological reality, given to us but also metaphysically present to start off with. This division is located horizontally, with Meillassoux and Brassier on the top and Harman and Grant on the bottom. Is there another axis which reveals a different pairing no matter how tenuous and faint? I believe there is.
This second axis could be called the
‘context of application’ axis ’Modal’ Axis- although I have no finalised name for it at present. It’s basically using two well known divisions in Analytic Philosophy of language and philosophies of logic: the division between Extentionality and Intensionality – but one could easily apply a distinction between Facticity and Ground, and the basic split still holds.
A crucial caveat is needed though: their application here only stands to illuminate or pick out various philosophical regularities I see in the philosopher’s respective work – what I’m not attempting to do is say that ‘x’ philosopher is making an intensional truth-value, or functional context, a commitment to semantic use, a logic proposition, or whatever. I’m only using the general differences between the two principles so as to pick out tensions between the SR fourfold as I see it. So I’m well aware in advance that hardly any of the SR four explicitly define themselves as extensional or intensional language philosophers – and to say so, would negate the purpose of establishing a first axis anyway. Rather I’m adding the principles of extensional and intensional principles to the division of Demonstration and Description, so as to give us a stable system in which one can pair off and divide different types of arguments and claims. What I am not doing is to fit or locate extensionality and intensionality as either Demonstration and Description, as this would be appealing to their respective Analytic uses once more (extension is typically used to refer to objects as referents: i.e. description – whilst intension is used to gain knowledge about the necessary reasons for meanings i.e. Demonstration) – to apply them in this narrowly analytic way is to arrive back at the realism/antirealism distinction again. What is crucial is that this axis is perpendicular to the Kantian one and not the same, even if it has been historically treated as the same. This way is more fun too.
Moving on quickly, I believe that what Meillassoux and Harman share in their philosophies is a joint development of extentionality. What is extenstionality? It basically refers to a set of principles upon which all objects or entities are equal if they have the same external properties. If you can describe one object by a set of external principles, i.e. extend them, (Robert is the only person on the street who is 29 years of age), then if you can apply this set of qualities to another set of objects (‘Robert’ and ‘the only person on the street who is 29 years old’) then each object is equal by definition. When one says ‘wallet’, the extensional principle simply requires that a history of all wallets satisfy this definition. The main point here is that extensionality is defined by applying a set of principles upon which a given object or entity can satisfy those principles, as in when and how they are applied – and the consistency of the philosophy is this application. Also both share a primacy of the fact as application towards a reality prior to reasonable ground: Meillassoux starts from contingent facts to arrive at the demonstration of absolute knowledge, as opposed (and similar in kind) to Harman who starts from describing actual contingent objects without locating a prior ground for their own existence (i.e. it could be said that individual discrete objects are their own reason, without a prior ground – which again is extensional).
By default, Meillassoux gets his extentionality by Badiou’s axiomatic endorsement of set theory (ZFC axiomatic set theory is defined by extension tout court), but there are other portions of his philosophy which encompass extensionality – for instance the application of possible worlds beyond this one, on account of absolute contingency applying to them, even if they nonexistent at present. This would be a demonstrative extensionality, (or Extensional Demonstration) driven by what can be known – and insofar as this is epistemological, this is about whether infinite worlds can be known or realized. Extensionality also plays a major part in Meillassoux’s anti-constructive, ex nihilo, anti-reason, ‘post-metaphysics – metaphysics’ – because like Badiou, there is no reason for how groupings of sets or worlds are brought together in certain principles, and thus they are always subject to change. In Meillassoux’s strange but intriguing cosmic history of contingent ‘jumps’ – there is no reason for matter, nor life, nor thought, nor God – only extensions of the absolute.
On first glance, Harman’s object oriented philosophy doesn’t correspond to this at all. Harman’s philosophy is weird on this basis. If it is an Extensional Description, then how would it differ? The key here would perhaps be how Harman uses the tools of extension to pinpoint the description of objects, speculating on their hidden reality outside of knowledge. First off, we can clearly see that extension is used to denote that objects are equal and must be accounted for. Objects, in themselves aren’t extensional sets of course, but the crucial point is how objects are realised at all in this metaphysics, and how such a system accounts for them. One of the classic OOO methods – the Latourian litany – establishes an enumerative procedure which lists various objects one by one in a list, so as to show how each object extends the OOO principles of discreteness, equality and irreducibility. Each object is understood as fitting the definitions (but of course objects are not simply ‘just’ their definitions – they exist in reality whether one pays attention to them or not). Another key aspect is the finiteness of extension for how OOO describes things that exist – all examples are good enough, and entirely equal in application (Although not part of OOO, Garcia is just as happy to apply the extensional nature of his flat ontology to absolutely everything “no matter what”, either existing or non-existing, contradictory or non-contradictory, because it simply exists according to a principle of, “no matter what”). For Harman of course, there must be something deeper in the thing that withdraws from all extensional description, even if we only have access to this particular method – we don’t have any access to a deeper knowledge of the essence of the thing, even if it is necessary (yet the thing is also contingent). Extension in this regard, is the plural equality of access in such a realism: and so all finite extensions are good enough to fail the grasping of discrete objects.
On the flipside though, its becoming increasingly clear that Brassier and Grant also share a similar joint development of Intensionality within the similar split of Demonstration and Description. So what is intenstionality? (not ever to be confused with intentionality) – this is a little harder to explain, but I’ll give it a go. Intensional principles (or nonextenstional principles as they are sometimes referred to) denote similar sets or objects so defined, but the difference here is that such principles require an explanation of the inner, necessary and sufficient reasons for those things, including meanings, properties and reasons. Intensionality refers to the necessary ‘rules’ of the structure which are coterminous with the thing itself. These principles must be sufficient enough to account for all manifestations of the rule (infinite or otherwise), as they are defined and applied; so if an extensional principle of a Bachelor is a list of all bachelors who have ever existed, an intensional principle of a Bachelor would simply be ‘an unmarried man’ – precisely insofar as ‘an unmarried man’ is a sufficient and necessary condition of knowledge that requires both to be the same or operating under the same rules.
So Brassier is clearly a philosopher who can be located in the Intensional Demonstration category without much fuss. For him there must be necessary and sufficient reasons for cognitive structure and representation – and moreover we can demonstrate these structural reasons, using the tools of objective science and nominalist theory, but utilising normative principles. The bankrolling of intensional reasons is what separates Brassier from other post-Hegelian philosophers, as he can escape what would be (in his eyes) the extenstional inadequacy of locating thought as an ex nihilo ”voided animal’, whose genesis is left up to radical contingency or vitalised matter. The birth of manifest representation needs sufficient explanation. Sellars can provide a nice and sturdy, meta linguistic, post-Hegelian table to stand on for this category -which requires using meta-lingustic function as leverage to a non-linguistic reality in thought. Equally Pete Wolfendale is equally home with the Brandom angle, and the procedural rationalism he endorses (or thats how I read his position(s) anyway). What is possible is only what is normatively possible.
Grant on the other hand has little time for appealing to the genesis of thought as its primary problem, but in equal measure, applies an Intensional Description into metaphysics. Nature is the functional necessary and sufficient reason in-itself outside of human demonstration. What Grant is after is the sufficient conditions upon which such all objects (including thought) are possible from Nature (as I understand it). Such individuals or objects are produced intrinsically from the ground of conditions, which in turn produce that ground (and ground is its production). The necessary sufficient powers that craft such objects (even all aspects of phenomena) are not separable from what they do, but quite simply ‘are’ what they do, as necessary production.
What both Grant and Brassier share, that Meillassoux and Harman respectfully do not, is an intensional commitment to function and process, but the Demonstration and Description distinction is crucial to highlight the differences; for Brassier, Sellarsian absolute processes are not processes which can be located ontologically outside of thought, but an ontological discovery about reality through concepts and intrinsic to the conceptual manifest image itself. Grant on the other hand, locates function and process on metaphysical grounds, where Nature and powers simply are, whether manifest or not. If Brassier is wanting to ground the epistemological reasons for facts, then Grant is equally similar (and equally opposing) in that he wants to ground the ontological grounds for existents, which are only described second hand.
I’ve gone on long enough for a blog post – but I’ll sign off with one last outcome which arises from this fourfold. From the structure as I’ve described it – we have found four possible connections and thus four equal possible tensions; (EDis and EDes (Harman and Meillassoux): IDis and IDes (Grant and Brassier): EDes and IDes (Meillassoux and Brassier): EDis and IDis (Harman and Grant). It follows then that there are two – and only two – critically incompatible tensions where no connection is possible whatsoever, if one accepts the dual axis, so defined: (that is EDes and IDis (Meillassoux and Grant) and EDis and IDes (Harman and Brassier)) – and this makes total sense in my eyes. Both sets of confrontations are so opposing that its unclear how any reconciliation is possible (I mean take one look at my awful drawing above: its clear that the philosophies of Wolfendale and Garcia, should never be made to dance). Likewise it explains the incompatibility of Meillassoux and Grant’s positions, the tension of which is never really covered in any great depth: Meillassoux seeks to remove the necessity of sufficient laws by simple epistemological deduction alone, where even matter and the probable possibility of matter, is rid of any intrinsic power to change or produce change. This then, is the complete opposite to Grant who seeks to locate the necessary and sufficient metaphysical reasons for the ontological production and possibility of change.