I’ve been seeing a lot on Harman and capitalism and his model of causation as “nonsense” and whatnot and thought I’d try my hand at an explanation. For one, I don’t see why Harman’s model of causation is so hard to grasp but maybe its because I have a different background than most of those involved in the theory-corner of the blogosphere. I also want to stress that I’m not an object-oriented philosopher. I have serious misgivings about OOP which will be evident from my paper for Speculations. In fact, my paper will be on the subject of change and causality. That doesn’t mean however that I don’t think highly of the theory or that Harman should be insulted or attacked. Disagreements happen, we’re all adults here.
There are essentially two modes to understanding Vicarious Causation. The first is Aristotelian, the second is Kantian. It should be noted that both of these give us different versions of Occasionalism, that is, a mediated model of causality. I think the main problem people have with Harman’s theory is that they approach it strictly from the perspective of Heidegger’s tool-analysis, which while foundational for Harman’s thought has been overshadowed by a newer model of OOP over the past year. I think this this clear from lectures he’s given recently where the tool-analysis is explained but not foundational. He’s found new, better ways to ground the theory which makes it much more historically relevant and probably much easier to grasp by those without the Heideggerian or even phenomenological baggage.
1.
In Aristotle’s Categories he distinguishes between subjects and predicates. The Greek word for “subject” is hypokeimenon (ὑποκείμενον) meaning “underlying thing.” Essentially, it is that which is predicated but remains beneath the layers of predicates. We can also understand this through substance and accidents. The substance of the thing is that which the accidents adhere to without itself becoming anything fundamentally new. My car is still a car even if I have it painted a new colour for instance. The predicate “silver” does not alter the substance “car” in any substantial way. So there are substances and there are accidents. Great. The chief occasionalist insight to be made here is through the chain of causality. The position is one that says substances don’t touch each other. Let’s use an example. When I have a relationship with a person, there is more to that person than our interactions. Let us assume it is a romantic relationship between lover and beloved. Does this relation exhaust the other’s being? Is it not the case that there is far more to the person than their relation to me? While we would likely share much of our lives with each other, there remains a fundamental gap between the two of us. Don’t we interact on the level of accidents and not substance? When I talk to or touch my girlfriend, there is always more to her than these interactions. This is also the case for my interactions with non-human objects, for instance the relationship I have to the laptop I am writing this on. There are infinite possibilities for relations within a thing, it can interact with practically anything else in the universe in any number of ways, none of which could exhaust its possibilities. This is the point of the fire and cotton example. Cotton can do a lot more than burn, and the fire only engages the cotton on that level and not on the part of the cotton (to use improper language) that could become denim or a Q-Tip. While the fire destroys the cotton, this does not mean it has exhausted those potentialities, it has simply destroyed them.
We can see then that substances don’t interact directly, but are mediated by accidents. It is not the table-being that keeps my drink from spilling but the particular arrangement and strength of the material all of which is accidental to the essence of table-being. The same is true of my use of a tool. It is never the whole being of the tool that is used but only those limited set of properties that are useful for whatever task I have planned; a screwdriver has many uses other than driving screws and could potentially do any number of things outside of the human-tool relation.
2.
Harman remains a Kantian, though one who walks on their head. I think it was Shaviro who originally used the term “Kantianism for non-human objects” (though I could be wrong there), and this fairly accurately sums up Harman’s position. When one is taught the history of philosophy, there is usually a course on the Moderns which is presented in epistemological terms. That’s how it was during my undergraduate degree and also how it is here in Newfoundland as well. The great debates of Modern philosophy are reduced to theories of knowledge. What Harman tells us in his treatment of this period however is that what is central is not knowledge but causality. The Rationalists are all Occasionalists, where there remains at least some aspect of necessary mediation. For Descartes, it is no problem to explain how matter impacts matter, but some serious legwork has to be done to explain how mind can affect anything else. There is a fundamental split between the order of thinking and the other of material things. I have a Cartesian friend who claims Descartes can get out of this without God but I don’t buy it. God remains a necessary causal mediator. This is different for Leibniz however for whom substances remain auto-affective, that is, able to alter or change itself.
On the other hand, the Hume’s skepticism claims that causality is merely convention, a bad habit that we can’t break with for practical purposes. Kant’s great move in the history of philosophy is to unite both skepticism and occasionalism, taking from the latter the idea of causality through mediation and from the former gaining the insight that it is the human mind that does this rather than God. What is interesting to me is that both Meillassoux and Harman play the role of Kant here, uniting these same traditions that Kant did in different ways. While Meillassoux accents the Humean Principle of Factiality (any cause can have any effect, meaning anything can be other than what it is or, put another way, anything is possible), Harman accents the occasionalist point of mediation, which, when taken strictly as an epistemological point becomes the Principle of Correlation. What Harman does however is unleash it as a metaphysical principle that says all interactions are mediated in some way because all things are distinct substances unable to interact with each other.
3.
This is where the concept of vicarious causality becomes important. What unites both the occasionalists and Kantians is the power of a single entity when it comes to mediating causality, either God or Mind. What Harman claims is that there is not logical reason for there to be simply one entity doing this, and that it is not a deficit of the human mind causing the appearance of a gap. If we unite modes 1 and 2 as presented here, the mediation of substances through accidents, along with the occasionalist model of causality, we are left with a “local” model of occasional cause. That is to say, it is not an epistemological fact that causality is mediated, but a metaphysical one. Rather than attempting to understand this through the eyes of a 20th Century Continental philosopher, we should think of it in terms of Aristotelian substance-theory mediated through the Moderns.
In Guerrilla Metaphysics and “On Vicarious Causality” Harman uses phenomenological language, drawing on the withdrawal of objects found in Heidegger and the epistemology of Husserl, but I don’t think this is necessary at all. I don’t think you need to accept phenomenology in order to accept vicarious causality (I don’t accept either, for the record). I am quite confident that an Aristotelian or a Scholastic could not only understand but accept such a theory of causality. When Harman talks of “Real Objects” versus “Sensual Objects” we can replace such language easily with substance and accidents. The sensual object I interact with is not the being of the thing, but the predicates and properties of it, all of which are accidental. When I pet my cat lying next to me, I don’t touch her cat-being, but nor do I simply touch a bundle of qualities. Rather, I touch the former through the latter, reaching the substance via the accidents. While real objects cannot touch, sensual objects can, and the latter can touch the former. That is, when I perceive a tree, I don’t see the tree-being, but a tree at a particular time of day in a particular place under particular circumstances. And yet there remains a tree underneath all of these qualities, something that is not exhausted by these circumstances and yet related to them. They are predicates of the thing without themselves being the thing.
As for all of the “head-scratching,” I think it is precisely this last point. Harman’s model of causality is decidedly pre-modern, which makes sense from the perspective of Heidegger scholarship when one takes into account his theological background, specifically his training in Scholastic philosophy. Levi’s philosophy, at least what I know if it, is rooted far more in structuralism, assuming we take structuralism as loosely as possible in its claim that what a thing means is not any positive claim, but is only a difference. A thing is not this or that. While he is a realist, it is a different kind of realism, with Harman being connected with the Aristotelian-substance tradition and Levi being connect with the Structuralism-difference tradition. Both posit a fundamental split, but they are different in kind. It only makes sense then that there would be head-scratching since they are reaching the same point (reality is made up of objects) through different claims (the heart of an object is a vacuous substance versus an object is that which is different from other objects).
Complete Lies. 
Pingback: Austin speaks « Object-Oriented Philosophy
Nice post — I was just thinking through Harman’s causation essay from Collapse, and trying to separate out how essential Heideggerian withdrawal was to the whole thing. I was already cool with the idea that a causal connection doesn’t “exhaust” an object.
For me, the head-scratching comes from a different place. We seem to have an occasionalist model that is “local”, which is totally fine. But what’s difficult to figure out is the intentional “space” within which a causal interaction takes place. What creates that space? And how does the mediation take place? If there’s a real me and a sensual tree, what makes the real tree change? If it’s the sensual tree, then there’s a causality that takes place outside of the intentional space that needs to be explained. If the real tree doesn’t change at all (only its accidents change), then how do its accidents adhere to it? Or how could it be destroyed?
I think the real difficulty is the spatial metaphor of the intentional space. It’s difficult to tell how that space is itself caused and why anything is on the inside of it.
Pingback: Harman’s Objects and Vicarious Causation « Larval Subjects .
It seems to me that the question “what makes the real tree change” is itself representative of the problem people have with the theory. Let us assume that Harman is right, that the tree-being is inexhaustable by any possible relations, that there is always an ‘indivisible remainder’ and that we only have access to the sensual object. Can I do anything to the tree that touches its tree-being? The only option is to destroy it. If I set fire to the tree, I don’t exhaust its relational possibilities, I do not exhaust its being, but I do destroy them. Great, but where does that leave us? Essentially, the tree-being is infinite, not spatially but relationally. If I remove the leaves or branches the tree remains a tree. If I transport it, the location changes but this does not alter its substance. Is there any way I ever do anything to the real tree?
I can think of three possible relations I can have to the (real) object: 1. I can destroy its being, 2. I can conceal its being, and finally, 3. I can reveal its being. I do not in any way “change” the being of the thing unless I eliminate it. In the case of concealment, whenever I caricature a thing, that is interact with it at all, I veil its being in some way. When I use an apple as a weapon to attack a person, I am concealing possibilities, not allowing it to do things it could do in other possible relations. On the other hand, this also reveals something about the apple, namely that it really hurts when it hits the back of your head. The same is true if I cook the apple, transforming it into a delicious pie. This reveals new things about the apple-being that are concealed as it hangs from a tree. This doesn’t mean that I “change” the apple-being, but when I interact with it, I in some sense conceal its being while in another sense reveal its being. This is all done however in a mediated manner where I do not interact with these potentialities but work with my own image of the thing, drawing out new possibilities.
Michael – Thanks — that answers some questions and raises others.
If I can destroy the real tree, then that is a special case, since otherwise, destruction is on a continuum with dismantlement or damage. So it would need to be explained what makes total destruction different. If I *can’t* destroy the real tree, then it needs to be explained how the real tree still exists when it has been incinerated and its ashes scattered over the Pacific Ocean.
But for me, that’s not the central head-scratcher. The central head-sratcher is where the intentional space comes from, why we have the metaphorical spatial relationship with it that we do, from where it derives its causal powers, and what allows it, rather than anything else, to effect causal change. What I see in Harman’s essay is a structure or model in which causation occurs, but in which an explanation of the actual causation does not seem to be present.
http://www.notbored.org/space.html
+/-
http://thinkingculture.blogspot.com/2004/12/production-of-space-by-h-lefebvre-aloy.html
Michael,
It sounds like our positions are actually very close to one another in a number of respects. My objects are structured or organized dynamisms of change. The tree is, for me, a substance but a substance is not a fixed or static thing but rather what I call an “endo-consistency”. That is, it is a pattern that persists through time. These objects could just as easily be thought of as processes or events. Where Latour and Whitehead think of events as instantaneous, however, I think of events as extended across time. For example, a football is an event but it has a duration or extension in time. The evental nature of the object is the processes by which the object perpetually reproduces its pattern across time.
Like Harman, the qualities of an object do not, for me, get at the proper being of the object or its substantiality. Rather the endo-consistency of the object consists of the internal relations that define its patterned being along with singularities or attractors. Attractors are states towards which an entity tends. Accompanying attractors is a phase space or set of points through which attractors can be actualized qualitatively. Points in the phase space of these attractors are actualized through a combination of internal states of the endo-consistency of the object (its ongoing internal dynamics) and relations it enters into with other objects and entities. For example, sunlight causes the actualization of the green of a leaf as one possible point in the phase space of the endo-consistency of the leaves.
I suppose that one way you could think about my objects is on the Deleuzian model of the virtual and the actual. The virtual, for me, is the substantiality of the object. Unlike Deleuze, however, I don’t think the virtual is a one-all or a whole, but is composed of discrete virtualities or multiplicities. The actual would be actualizations of attractor states within the phase space of the object.
Regarding your question of change, I think there are a variety of different forms of change. First, as a dynamic system objects are perpetually changing while maintaining their endo-consistency. Second, objects evolve or develop as in the case of trees growing. This is presided over through a combination of the endo-consistency of the object and the networked relations it enters into with other objects. Third, through the actualization of various points in phase space, objects can hit bifurcation points where the object becomes an entirely new object with a very different endo-consistency. Finally, through entering into networked relations new objects can emerge with different endo-consistencies. Hopefully this makes some sense and also puts some flesh on my point that my objects are not simply differences from other objects (exactly the opposite of what I’m arguing).
Isn’t the substance/accident distinction somehow mirrored in our puzzlement about how it is possible that we are able to identify objects which are subject to change ( both in form than in exposition )?
The standard answer is that a neural network is a device for clustering data. A huge set of data is mapped onto a few data points using approximations. It is possibly much greater than our simulations are able to yield but not much different conceptually.
The nice thing about identification as clustering is that pre-defined categories don’t have to be assumed. They emerge in the clustering process adaptively. On the other hand features have to be distinct and get integrated. In the model there are feature vectors. Together with a metrics defined on those vectors perception can be put to work.
Oh, the cat is more than the sum of its parts? I wonder why people are thinking so low about sums?
It would be o.k. to say that the cat is not just a user interface for its environment and in particular ourselves but the distinction is not so absolute as it seems. It is a useful abstraction and those abstractions are plentiful and are of course not restricted to human beings but are created also by cats and trees. Here are a few examples of what trees are doing when they build abstractions of their environment – this time without neurons:
http://tinyurl.com/y9uuke4
I know this is probably just a way we talk these days, but this sentence strikes me as odd:
“My objects are structured or organized dynamisms of change. The tree is, for me, a substance but a substance is not a fixed or static thing but rather what I call an “endo-consistency.”
I thought the whole point of realism is that there are no “my” or “your” objects – I mean I understand that you are saying “my understanding of objects is…” but still there’s this subjective perspective. It seems that if I wanted to be a consistent realist I would say that “objects are this and that way – I see them as such and I am either right or wrong, if they are the way I see and understand them, everyone else who disagrees with me must be wrong.” I don’t see how we can have various versions of object-oriented ontology floating around without a major fight for the one true version. We can never get to the truth, of course, but I find this peaceful coexistence of various theories to be somewhat puzzling.
Mikhail,
I take it that my use of words like “my” and “I” is an expression of modesty, not a reference to subjectivity. All I am doing when I write like this is acknowledging that there are other theories of objects and the real. Harman theorizes objects differently than I do. Ian Hamilton Grant develops a critique of “somatism” (objects) throughout his work. Latour and Whitehead propose a different ontology. The point isn’t that all of these theories are “true”, but that there are debates out there as to what objects and the real is. This is no difference than disputes between Fichte, Schelling, and Hegel. Each of these thinkers claims to provide a theory of what reality is and have disputes with one another over their philosophical positions.
Michael, hi. You write:
As I always do when talking about OOP/SR, I will qualify this with “I may be missing something,” and I may be missing something, but this description seems to me to actually cause Harman’s view of substance to lead to a refutation of his view of causality, or vice versa. That is, his account of causality, which follows from his theory of substances, actually cancels his theory of substances, or his theory of substance cancels his theory of causation, or perhaps both.
To see how I think this happens, take a hypothetical example. Imagine a scientist learns to gradually replace the DNA of a tree over time, so that it doesn’t die, but eventually all of its DNA is replaced and it no longer has the DNA of a tree but that of a cactus. If this scenario were to play out, we’d have to ask, is the scientist, in mucking with the tree’s DNA, contacting the tree’s substance? It doesn’t seem that we can say that he’s merely destroying it, unless any change in the DNA is seen as a destruction of the tree (which would lead to all sorts of strange conclusions: e.g., genetic diseases which don’t kill the tree but alter some of its DNA have destroyed the tree as it was and created in its place a new substance, an entirely new tree, which looks and behaves exactly like the old tree in the same spot and so on). So the scientist is either contacting the tree’s substance, or he’s working with the accidental. This latter explanation leads two two problems. One, it doesn’t allow for an explanation of when (or even if) the tree becomes a cactus. And two, the bigger problem, the theory-cancelling problem, is that it means there is no treeness (or even individual treeness) to the substance of the tree.
Now, you may object that the tree’s DNA is not its substance, an objection with which I’m fine, but it misses my point. What I mean by the second problem above is that if there is nothing that we can do to the tree, no way that we can possibly exhaust its possibilities, and therefore no way that we can touch its substance (short of destroying it, and thereby destroying its possibilities), then its possibilities are both infinite and coextensive with the possibilities of all other beings (since they are also infinite). In other words, there are not multiple substances, but instead only one substance, which manifests itself differently (that is, it has different accidental properties, or attributes, as some Dutch guy who made a similar point a few centuries ago, in response to a similar theory that there were multiple non-interacting substances, called them in some Propositions, maybe I-XXIII) in trees and cacti and shale and people and so on.
So we’re left with a sort of double whammy to vicarious causation: either we accept individual substances and argue that, somehow, they can causally interact with each other substantially, or there is only one substance, in which case vicarious causation is nonsensical. Either way, vicarious causation has to go (and with it, a particular view of substance).
Again though, I could be missing something.
Chris,
I don’t think you’ve shown that we must accept either the theory of substance or the theory of causality and toss the other. In fact I think you’ve brought out the proximity Harman’s theory has to Leibniz.
You said: “In other words, there are not multiple substances, but instead only one substance, which manifests itself differently”
We could go either with the Spinozist response or the Leibnizian one. Sticking with the example of the changing tree, we could say that the tree-being doesn’t change at all since the tree-being is infinite and inexhaustible. By turning it slowly in to a cactus, we are simply unveiling new possibilities within this being. I understand your inclination to say that this means they are all the same and so are all one, however, if we follow Leibniz rather than Spinoza, we have a model whereby all objects (monads) are identical in one sense (all are infinities) and yet don’t touch, but remain distinct.
The real objects still never touch, but can only ever be mediated. Your turning a tree in to a cactus does not change the fact that that being remains vacuum-sealed and distinct from all others. You veil and unveil the thing through changes to its accidents, not to its substance which cannot be changed (for how could an infinity change into anything else?). It is simply a matter of different accidents showing different aspects of existence.
Rather than one object we end up with an infinity of objects. Monads are not entirely self-identical but hold differing positions, reflecting all others from different vantage points. This is why for Leibniz causality is autoaffective, one monad affects itself and since all monads reflect all others on to infinity this affects the whole of reality. For Harman, shifts in accidents reveal new and different qualities of existent substances which exist as a sort of self-contained cosmos.
Levi,
I understand that point, I was going a bit further with my question, but I suppose I didn’t make it clear: I agree, of course, that we can have debates about, say, what constitutes objects and so on (“in my view, objects are X”), but in a realist view there must be one true theory (things exist independently of us and they are this/that way), right? I mean if there is one definite way that objects are, then shouldn’t we put all our efforts into figuring out what that way is instead of creating and polishing our own perspectives on the matter? I mean I’ve been following these conversations for what feels like many years, sometimes more actively, sometimes more passively, and I am continuously being reminded of Kant’s description of the scandal of philosophy in the preface (I forget which one) to the first critique, namely, that we are still not able to have one good theory of what reality is like. Isn’t there a sense of this scandal still present when we talk about various ontologIES?
Hmm… I wonder if holding different positions doesn’t limit their possibilities, but maybe I’m taking the spatial metaphor too seriously. I’d also wonder about divisibility of an infinite substance (a point Spinoza raises, of course), whether Harman adheres to some sort of Leibnizean universal harmony (which the Leibnizean position seems to require), as well as all the other commonly raised objections to monads or entelechies or whatever we’re calling individual substances these days.
But more than that, I’d wonder whether the choice isn’t an arbitrary one (an old objection, I know, and is one of the few to actually bring out a sense of humor in Kant). That is, if we accept that substances can’t be altertered (at least not from the outside, to keep with the spatial metaphor), then why should we choose between one substance and many substances? What does choosing many substancdes buy us? I suppose since their’s a Kantian element to the Leibnizean elements in Harman’s thought, we could ask both why Leibniz instead of Spinoza and why Kant instead of Schopenhauer? I haven’t yet read Guerrilla Metaphysics (it seems to be perpetually checked out at my library, which I suppose means people are interested), so maybe there are in depths arguments for why this instead of that in that book?
I must admit, this is a problem I’ve had in trying to follow OOP in general: why this and not the other? I don’t mean to imply that they don’t provide arguments (I know that’s a common charge on the intertubes), merely that it’s not clear what their choices ultimately buy them, and why they are any better than their contraries (e.g., why many substances and not one? why objects instead of events? why realism instead of anti-realism? etc.).
Mikhail,
I’m really not sure what to do with your comment here. It seems to me that you’re conflating theory with the world. An ontology is not the world but a theory of the world. It is possible for that theory to be mistaken. Noting that there are different ontologies is merely pointing out that there are different competing theories of what being is. Each of these theories are trying to get at the truth and each of these theories critiques other positions and offers arguments in favor of its position. I am not “polishing my perspective”, but presenting a position or theory of the being of beings. In other words, I am trying to get at reality or the being of objects. This point seems like it should go without saying.
Levi, I might be conflating theory with the world since I don’t think there’s a pre-theoretical world and that I can ever compare a theory with a world it theorizes and therefore see how it does or does not fit. So for me there is no one true theory of the world, since that presupposes that I have an access to this world and so on.
You did answer my question though, so I think this is one productive interaction between us that just needs to stand here for everyone to enjoy.
Michael,
Can you e-mail me your Speculation piece? I’d love to read it in total. Just send it to me at joncogburn at yahoo.com if that’s O.K.
Best,
Jon Cogburn
Jon,
You can certainly read it… when it’s done.
Chris,
You write:
I think a number of these questions have been answered in both my own work and Harman’s published work. Since you’re having trouble getting hold of Guerrilla Metaphysics, you might check out Prince of Networks which is available for free online. You can find it here (warning pdf):
Click to access OA_Version_780980544060_Prince_of_Networks.pdf
Harman does provide arguments as to why we require many substances, not just one substance. You’ll find these arguments, in particular, in his critiques of what he calls “relationism”. These are arguments I share with Harman.
In my own work, I believe my recent posts on transcendental realism give a sense of why epistemically I believe that realism is a superior position to anti-realism. These are arguments that will be developed in greater detail in the future. And I certainly appreciate critique of these arguments.
In my view the binary choice between objects and events is a false dilemma. Within the framework of the ontology I am proposing, objects are events. Consequently, I have no qualms with someone substituting the word “event” whenever they encounter the word object in what I write. Part of the argument for this reason is a posteriori. The more we’ve learned about the natural world, the more we’ve learned that objects are not fixed and enduring substances lying beneath accidents or properties, but that objects are dynamic and ongoing processes. Second, I believe we need an “evental” conception of objects to capture the sense in which objects act as they interact with other objects about them. Here, I think, is perhaps a major difference between the ontology I am proposing and the ontology Harman proposes. Where Harman is a strict actualist holding that objects are completely concrete, I argue that objects have potentialities and capacities that differ from the actualized acts.
Pingback: Paradoxes of Realism and Self-Reflexivity « Larval Subjects .
Pingback: Links to links to links to links « CineMadison
Pingback: Against Experience « Deontologistics
Pingback: on a comment by Levi « Object-Oriented Philosophy
Pingback: Translation and Information « Larval Subjects .
Pingback: Levi Apparently Has Never Read Harman’s Theory of Causation « Frames /sing
Pingback: Empirical Realism and Ontological Realism « Larval Subjects .
Pingback: On Vicarious Head-Scratching | Complete Lies. A good explication – The AntManBee