In considering the arguments for the final aspect of Harman’s system, namely vicarious causation, we are again put in a difficult position. Though Harman devotes a considerable amount of space to elaborating his account of allure,1 and presents some additional reasons why we should want such an account of causation, the principal motivation for the account is provided by the arguments we have already considered and rejected. Harman issues the following challenge to those who would assess his account of causation in Guerrilla Metaphysics:
Once it was conceded that the world is made up of withdrawn objects, utterly sealed in private vacuums but also unleashing forces upon one another, all the other problems follow in quick succession. Let anyone who does not agree with the strategies of guerrilla metaphysics specify clearly which of its initial steps is invalid.2
This is precisely what I have done. None of these initial steps has proved valid, let alone all of them. This seems to rule out vicarious causation by default. Still, there are some more probative reasons that Harman presents for his account of causation. He provides a further historical narrative regarding the tradition of occasionalist accounts of causation, which is meant to suggest that the problem his theory responds to emerges from a broader range of concerns than his own. He also suggests that the scientific account of causation demands supplementation by a metaphysical theory of causation of precisely the kind he provides. I will now address both of these claims, but in between them I will try to reconstruct the core of Harman’s argument for vicarity on the basis of the independence of objects from one another. This will provide a proper contrast with the motivations of the occasionalists, as well as contextualize the demand for supplementation.
According to Harman, the problem of how distinct things can causally interact has a long lineage.3 On the one hand, he sees it being raised in explicitly metaphysical terms in the Islamic occasionalism of the Ash’arite school, the modern occasionalism of Descartes, Malebranche, and Leibniz, and in the more contemporary occasionalism of Whitehead. All of these thinkers invoke God as a mediator capable of overcoming what they see as the causal gap between entities, whether as the source of all causal power (the Ash’arites), the source of the connection between different kinds of substance (Descartes), or the medium through which entities are able to encounter one another (Malebranche, Leibniz, and Whitehead). On the other, he sees it being raised implicitly in the epistemological scepticism/critique of Hume and Kant. He reads these thinkers as invoking the mind as a mediator which provides the causal connections between appearances, whether through mere habit (Hume) or through transcendental necessity (Kant). Harman criticises both of these trends for advocating a ‘global occasionalism’, insofar as they require all causal relations to be mediated by the same thing—God in the former, the mind in the latter—and proposes, along with Latour, a ‘local occasionalism’ in which causal relations between entities are mediated by further entities.
Now, although this strikes me as presenting a somewhat perverse reading of Kant and Hume, insofar as it reads their epistemological concerns in metaphysical terms they would abjure, there are definite continuities here. There are overlapping themes that seem to motivate a similar account of causation, insofar as they all demand some form of causal mediation. However, this demand does not arise from a single problem held in common by the various sub-traditions that make up this narrative. For instance, not only did Islamic occasionalism provide a theological solution, it was motivated by a theological problem about the power of God. This is remarkably different from Descartes’s problem concerning the split between thought and extension, Leibniz’s problem concerning compossibility, and is light years from the concerns with the nature of explanation that motivate Latour’s occasionalism. If we do not share any of these diverse concerns, then this problem has no hold on us. Harman hardly takes the theological concerns of the Ash’arites to be pressing, so he cannot lean upon them to motivate his own theory of causation. In short, we still need some good reasons, above and beyond this narrative, to accept the problematic status of unmediated causal relations.
Harman’s own reasons for taking unmediated causal relations to be impossible all stem from his claims about the independence of objects from their relations to one another. These turn up at various different points in the three arguments for withdrawal we have considered, but they are never motivated independently of claims about the excess of objects over our grasp of them, whether they are explicitly connected or implicitly conflated. This should be unsurprising given the dominance of phenomenological themes throughout these arguments, even when they are illicitly intertwined with metaphysical ones. My aim now is to make this tangle of claims about epistemic access and causal interaction a bit clearer, not by reconstructing a further argument, but by unearthing a non sequitur underlying the other arguments. This amounts to a final attempt at cutting the Gordian knot of methodological issues underlying Harman’s project, prior to considering his ideas about the relationship between philosophy and science.
I think the key here is Harman’s offhand remark that ‘despite its various degrees of efficacy, [physical causation] must ultimately either work or fail to work’.4 This remark is made in the context of an exposition of the parallels between causation and allure, which he similarly takes to either succeed or fail in this binary fashion. This adds an extra layer of depth to the picture of vicarious causation presented above: not only is sincerity insufficient for causal interaction, allure is sometimes insufficient too. Successful causation requires successful allure. But what is really interesting is the claim that causal interaction should be understood in terms of success at all. If the problem of how one object can affect another is actually the problem of how one object can successfully affect another, then this tells us something more about the implicit motivations of the problem. Knowledge can be understood in terms of representational success. If one conflates representation and causation by treating causation in intentional terms, then one can seemingly infer the impossibility of successful causation (causal independence) from the impossibility of knowledge (epistemic excess). This conflation can only be held together by the sort of functional language that Harman refuses to abandon at the end of the argument from execution, as it lets us treat things as striving for ends. We can say that things try to affect one another, even if they always fail.
Of course, there still must be some way in which causation can succeed. The absolute ban upon causal contact is thus qualified using the notion of directness: all direct access fails, therefore all direct causation fails. The hope of an indirect form of access (if no longer strictly epistemic in character) thus holds open the hope of an indirect form of causation. This hope is answered in both cases by allure. The latter provides a supposedly non-representational way for us to access the real, and in doing so provides a way for the real to affect us. However, the fact that these relations proceed in opposite directions should give us pause for thought. The object that tries to affect is the object hiding behind the sensual object, whereas the object that tries to access is the object encountering this facade. What is going on here?
The crucial question may be put as follows: In precisely what way can allure be said to succeed where representation fails? It is the equivocation between the standards of representational success and causal success that allows us to convert epistemic excess into causal independence. If there is no sense in which allure is held to the former standard, or to some deeper standard that it shares with representation, then there is no good sense in which it can overcome causal independence. The problem is that the only concrete standards of success that Harman ever deploys in his discussions of allure concern how the allure affects the one who experiences it.5 Does the joke make me laugh? Does my mistake embarrass me? Does the metaphor make me think? The fact that these are the questions that determine the success of allure indicates why successful allure is a model for successful causation. These allusions can only succeed or fail insofar as there is some effect they are supposed to produce upon us. They are thus more like access to narcotics than access to information. It doesn’t seem to matter that there is no substantive comparison with representational success, only because allure is already understood in causal terms. The non sequitur is hidden by blatant circularity. Harman’s aesthetics is an introspective theory of emotional affection.
Finally, we come to Harman’s defence of the importance of his theory of vicarious causation by way of his thoughts on the relationship between philosophy and science. Let’s jump straight in at the deep end:
For several centuries, philosophy has been on the defensive against the natural sciences, and now occupies a point of lower social prestige and, surprisingly, narrower subject matter. A brief glance at history shows that this was not always the case. To resume the offensive, we need only reverse the long-standing trends of renouncing all speculation on objects and volunteering for curfew in an ever-tinier ghetto of solely human realities: language, texts, political power. Vicarious causation frees us from such imprisonment by returning us to the heart of the inanimate world, whether natural or artificial. The uniqueness of philosophy is secured, not by walling off a zone of precious human reality that science cannot touch, but by dealing with the same world as the various sciences but in a different manner.6
Harman thus sees his metaphysical system as an attempt to return philosophy to its rightful subject matter. He defends philosophy’s right to tackle the same topics as the sciences by claiming that it can approach them through other means. Given the difficulties we have encountered in determining Harman’s methodology thus far, we are entitled to some curiosity regarding just what these means are, and how they are supposed to differ from those of the sciences. This is where the theory of vicarious causation is supposed to shine, by providing us with an exemplar of the divergence between the scientific and philosophical approaches:
From the naturalistic standpoint, ignoring for now whatever complications one might wish to infer from the quantum theory, causation is essentially a physical problem of two material masses slamming into each other or mutually affected through fields. One object becomes directly present to the other, whether through physical contact or some other form of intimacy. But there is also a metaphysical problem of causation.7
The initial problem with this is that all of the contrasts Harman makes between the supposed scientific understanding of causality and his own metaphysical one present an incredibly crude version of the sciences.8 Although he pays lip service to the implications of quantum mechanics, he entirely ignores the advanced mathematical techniques (e.g., phase space modelling, statistical analysis, information theory, etc.) that the sciences have developed to model phenomena since Hume talked about billiard-ball dynamics, along with the intricate theoretical questions regarding the nature of causation that these have spawned, both in the sciences and the philosophy of science (e.g., emergent capacities, statistical causality, information transmission, etc.).9 However, on second thought, the real problem is that Harman’s approach precludes him from paying any attention to these things anyway. As far as he is concerned, the sciences don’t tell us anything about reality. They only talk about it as it seems, whereas philosophy can talk about it as it is. This isn’t to say science is useless, but simply that the truth is entirely inaccessible to it. Maybe this truth will be relevant to the sciences, maybe it won’t, but there’s no real debate to be had here, even if there might be mutual inspiration.
There is a tremendous irony in this: the strange methodological hybrid of phenomenological description and metaphysical argument that Harman adopts amounts to the practice of introspective metaphysics. It is important to understand that this is different from what is often called ‘armchair metaphysics’ insofar as it has nothing to do with the a priori as traditionally understood. It is not a matter of retreating from observation to contemplate and reason about the fundamental concepts that underpin observation, but a matter of seeking out a special kind of intuition unknown to the sciences. Harman claims to get at the reality that the sciences can never describe by closely describing the structure of seeming. Far from challenging the retreat of philosophers from the world into the bastion of consciousness, he has simply extended the domain of consciousness into the world. On this basis, he provides us with an introspective theory of causation modelled upon emotional intensity. This theory is independent of the sciences insofar as it is based on a form of evidence entirely alien to them, but it strikes me as equally alien to the proper practice of philosophy. The phenomenological trappings in which Harman’s metaphysical introspection are clothed are at best a bad disguise, as if an unusually pensive crook were to don a rubber Husserl mask to preserve his anonymity during a hold-up. What they hide is a series of questionable assumptions and sometimes outright misunderstandings regarding important epistemological and metaphysical issues. Our next task must be to peel back this mask and bring these assumptions into the open, in order to better understand why one might be tempted to endorse OOP despite the convoluted and deeply flawed arguments presented for it.
1. Cf. Guerrilla Metaphysics §§8–12; ‘On Vicarious Causation’.
2. Guerrilla Metaphysics, 97.
3. Cf. ‘On Vicarious Causation’, 188, 202, 218–19; Prince of Networks, §5C.
4. Guerrilla Metaphysics, 176.
5. Cf. Guerrilla Metaphysics, §§8–9, 211–13.
6. ‘On Vicarious Causation’, 190.
7. Guerrilla Metaphysics, 18.
8. Cf. Tool-Being, 19, 209; Guerrilla Metaphysics, 79.
9. This is evident in the way he approaches the work of Ladyman and Ross in ‘I am also of the opinion that materialism must be destroyed’, 772–90, where he all but explicitly refuses to consider the scientific issues that motivate many of their crucial metaphysical choices.