The incoherence problem: Bryant contra Harman

One of my favorite moments in the classroom as an undergraduate was getting to the end of Wittgenstein’s Tractatus—the part about how you should shut up already if you can’t talk about something. Even when quietly intoned with Cambridgey words like “whereof” and “thereof,” the teenage mind easily identifies (and identifies with) the underlying sentiment of rage at self-contradictory types who want things both ways, at tragic stories with unprepared happy twist endings, at authorities who would legislate how you spend your free time, at hubristic cosmologists who claim to know what aliens are like

(they seem friendly to me)

That desire to draw strict world-defining boundaries comes to mind when I think of the evolving dispute between Levi Bryant and Graham Harman. Of course many smart people get into disagreements that reach an impasse, despite years of correspondence and ongoing personal and professional respect. What is fascinating in this case is how close these two seemed at one point to agreeing about fundamental issues (albeit with different terminology and backgrounds). Only then did their intersection reverse into a still-growing gap. It can’t be seen as reciprocal, since Bryant is the moving target. For this post I want to look at one element that he has made into a major theme of dispute. The familiar sentiment at the center of his current conviction against Harman’s philosophy is “you can’t have it both ways.”


Bryant recently published an article, “The Interior of Things,” based on a talk he gave earlier last year with the same title and published on his website. The full text of the officially published version is not yet available (for me, anyway), and he may have changed key paragraphs since delivering his paper in Milwaukee. But judging from the abstract and citations, it looks like the argument hasn’t changed. [Addendum 3/2017: Nope, no changes. All quotations below are accurate!] Here is how he characterizes Harman’s philosophy and his own relationship to it:

That contemporary vector of ontological thought of which I was once something of a proponent is very much a revival of Aristotelian substance-ontology. We must remember that…[Harman’s] position is not identical with speculative realism, and also that object-oriented philosophy is not identical with object-oriented ontology. Speculative realism and object-oriented ontology are both broader than Harman’s object-oriented philosophy.

I think the timeline goes something like this. Harman develops his philosophy around the turn of the century and calls it object-oriented. Bryant doesn’t like the idea of it, but then he reads some and likes it, either noticing some similarities with his own thinking or finding himself moving in its direction. He proposes object-oriented ontology (OOO) as an umbrella term for what the two of them and some others are doing. Harman likes the updated moniker. It captures his own project (ontology), allows for a diverse and productive group of bickering allies, and doesn’t lead to the “oops” joke of OOP or to confusion with a popular programming language. He thinks: okay, I also do triple-O. In the background, the other speculative realists disavow their own title or vary it (speculative materialism anybody?), leaving it largely associated with Harman and OOO. The result of their ambivalence: “Harman can have the term.”

Several years later, as the quoted passage above indicates, Bryant seems wary about even associating object-oriented ontology with Harman’s philosophy. He avoids phrasing that might imply Harman has produced just one variant of OOO, a twist on a set of shared basic views. Instead of writing something like, “Harman’s position is not the only object-oriented one,” he suggests that Harman’s ideas are qualitatively different than the broad tenets of OOO. He also assigns Harman’s project a new nickname to contrast it with speculative realism: “I am advancing a speculative realist critique of Harman’s object-oriented philosophy or neo-Aristotelianism.”

What to call it, then? In most other contexts “Harman’s OOO” would be appropriate, yet Bryant in his essay invites a more drastic choice. Personally, I find the idea of a strict difference between OOP and OOO kind of annoying. (“Philosophy” is the broader term anyway.) From a practical standpoint, it would make more sense for Bryant himself to either let the terminological issue go or stop using object-oriented. In any case, his own views have changed and there’s a strong historical association between Harman and object-oriented thought. So I will refer to Harman’s work as OOO. He can have the term.

The main element in OOO that Bryant does not like is that it responds to the problem of non-relational objects with a doctrine of indirect or vicarious causation. (Ultimately, this means Bryant does not like the specter of non-relation…) If I wanted to portray the first step in the starkest possible way, I’d choose this passage from Tool-Being (295) or something like it:

The world has been said to contain no relations—nothing other than entities. But entities are always primarily withdrawn tool-beings, and as such, they are sealed away in a vacuum devoid of all relation. If this is true, then the world is packed with noncommunicating vacuous zones, ontological bubbles, none of them able to transmit energy or influence to the others. There are neither windows nor doors to be found. Any contact between distinct entities would seem to be impossible; for the same reason, any sort of alteration in the universe would also seem impossible.

My decision to quote the passage outside of its rhetorical context sharpens the sense that objects can never touch, on pain of logical contradiction. In the immediately preceding sentence, however, Harman does seem aware that the picture of vacuous zones just painted raises a question about how relations are possible:

I have already mentioned that a challenging difficulty results from all of this…

And immediately after as well:

Is there any way to avoid these consequences by pointing to a medium through which tool-beings might genuinely interact? How can one vacuum impart its secrets to another?

The precise nature of the apparent paradox gets clarified in Guerrilla Metaphysics, when Harman again claims that his doctrine of non-relation raises major questions (97):

Heidegger began with a negative approach to his own question of being, with the profound initial insight that being cannot be thought of as presence-at-hand. I would give it an even more general twist, and say that being cannot be conceived in terms of any relations at all. This is a useful negative starting point, but still tells us too little about what vacuous tool-being really is.

Not only do relations need to be accounted for, but tool-being itself has not been sufficiently examined. In other words, tool-being isn’t exhausted by the doctrine of non-relation! And that obliges the critical reader to grapple with the reason Harman takes his version of vacuum-sealed objects to open up the topic of causation for further study rather than simply foreclose it. But let’s consider what Bryant has to say. First, he gives his own account of Harman’s non-relational bubbles, as well as the following step of vicarious causation:

The outcome of Harman’s object-oriented philosophy is that real objects cannot relate because they are so thoroughly withdrawn they never touch in any way. Again, they are, as he likes to say “vacuum sealed.” As he repeats throughout his works, this places him in a position similar to the occasionalists. The occasionalists argued that no object directly relates to another…Unlike occasionalists like al-Ghazali, however, Harman does not evoke God to sustain objects in their existence and relate them to one another. Rather, he develops a concept he refers to as “vicarious causation.” While the squirrel that climbs a tree does not relate the tree as a real tree because, insofar as beings are withdrawn from one another, the squirrel as one real object does not relate to the tree as another real object, the squirrel does nonetheless relate to the tree as a “sensuous object.” The sensuous object is an object on the interior of a real object; in this case, an object on the interior of the squirrel. As Harman remarks, “…we never touch real objects, we always touch sensual objects. Sensual objects would not even exist if they did not exist for me, or for some other agent that expends its energy in taking them seriously.”

Bryant produces an important ambiguity when he defines the sensual object as something on the interior of a real object. But before getting to that, let’s finish up the argument that follows. It’s a long paragraph, yet worth viewing in one chunk:

I confess that, along with many others, I have a very difficult time understanding Harman’s account of vicarious causation or how it solves the quandary surrounding relations that his doctrine of withdrawal places him in. First, at the most basic level, his account of vicarious causation seems incoherent. Claiming that the squirrel relates to a sensuous tree rather than a real tree gets us no further because we still don’t understand how the sensuous tree can possibly relate to the real tree. It would seem that you can’t have it both ways. If all objects are absolutely withdrawn from one another, if they are all vacuum sealed, they don’t relate in any way. Here the squirrel is not relating to the tree, but to itself. It’s as if Harman were attempting to say that the squirrel both does and does not relate to the tree, but that move has already been foreclosed by his thesis of withdrawal. As an aside, we can also see how Harman’s ontology of withdrawal also undermines the entire ecological dimension of being insofar as it conceives beings as fundamentally self-enclosed and unrelated. Second, and I owe this insight to David Roden, the doctrine of withdrawal makes it difficult to see how the squirrel can be related to a tree at all, for where objects are withdrawn beings can never know what it is that they are related to. As Roden puts it, “it’s difficult to see how a proponent of [Harman’s object-oriented philosophy] can be ontologically committed to crystals (or clowns) given that we never have access to them.” Where objects are completely withdrawn from one another it becomes impossible to say what is because we have no access to these objects whatsoever. What we instead get is a sort of pan-solipsism where, at best, we can say that there are sensuous trees, squirrels, and clowns for us, without ever being able to determine whether there are real trees, squirrels, and clowns.

Bryant’s first and most emphatic point is in bold: “If all objects are absolutely withdrawn from one another…they don’t relate in any way.” He holds that a real object cannot, by definition, interact with anything. The second axis thus only results in solipsism, with lots of sensual objects for me and nothing more. (Pan-solipsism is a false concession, since according to this critique there’s no reason to suppose the existence of a plurality of objects.) Bryant also claims that the return to the occasionalist problem in OOO forecloses ecological thought because objects are cut off from each other.

Now both of the above points come down to just one: Harman cannot allow for relations. In Bryant’s view (which he says he owes to David Roden), any attempt to do so is condemned from the start. Given the understanding that real objects are withdrawn from all contact, any talk about their relating is incoherent. This is very different from Harman’s own perspective at the end of Tool-Being. Once we arrive at the occasionalist dilemma, he writes, “we must ask through what medium [objects] affect one another at all” (296). So which one is it: a cul-de-sac into obligatory silence, or a launch pad?


Harman would probably claim that the underlying assumption in Bryant’s critique is that contact must be one/zero, either/or, all/nothing. Indeed, Harman disputes this very notion. He claims that if contact between real things cannot come down to an immediate influence between entities, access must instead occur indirectly. It would also be wonky to argue something like, “if things can’t touch anything, then it’s contradictory to say they touch anything.” Built into Harman’s earliest accounts of objects is the claim that what we do touch is always a distortion. Translation comes with the package of withdrawal.

Yet hidden in the bones of Bryant’s critique is a very interesting point. Namely, “we still don’t understand how the sensuous tree can possibly relate to the real tree.” This statement is worth exploring a bit, because it connects with one of the weirder features of Harman’s ontology.

Starting out, let’s grant (contrary to Roden and Bryant) that anyone’s occasionalism, including Harman’s, requires things to relate indirectly through some medium. In order to avoid rehashing all of Tool-Being, let’s also grant that the mediator or “vicar” in Harman’s philosophy is not a single canvas shared by all objects, not even one that can be folded into different shapes. Instead, something about the specific form or “style” of a given entity makes it relate to other things in its own particular way.

Bryant’s paraphrase at this stage is not strictly accurate. He claims that the sensual object exists “on the interior of a real object.” The tree encountered by a squirrel is on the interior of the squirrel. But that’s incorrect. The sensual object in OOO is not actually on the interior of the entity that brushes against it. Rather, it is on the interior of the relation between the two objects, which itself exceeds the execution of the squirrel’s being. That is, in the attachment squirrel-tree, or RO-SO, Harman would not say that SO lies on the interior of RO. Instead, the tree SO exists on the interior of a compound entity RO[2], a distinct thing altogether from the reality of the squirrel RO[1] alone.

By the same token, it is misleading to claim that withdrawal “undermines the entire ecological dimension of being insofar as it conceives beings as fundamentally self-enclosed and unrelated.” Beings are made of relations for OOO, after all. The tree exists on the interior of what we might call “The Squirrel-Tree Encounter.” (For a concise version of this claim, see Quadruple Object, ch. 8 section B.) To repeat a point made above, the concept of “sensual object” or translated object refers to one thing relating to others in its own particular way, which is surely different than the idea that real trees project images of themselves to squirrels or that squirrels are trapped in their own heads. The role of creators and intellectuals of all stripes is thus to investigate the dynamics of translation, and for OOO this is only possible by digging through levels of objects.

It is one thing to produce a new object through indirect contact with another (the squirrel-tree encounter), but it is another for change to come to the first real object. Harman has tried to account for such retroactive change in his recent writings on aesthetic and social theory. As I mentioned in my post last month about Dante’s Broken Hammer, he claims that change happens aesthetically. The object conforms to a sensual object like a character actor playing a raging magician or a golf club wrapping around a tree; it fuses with the surprising qualities of its relation through the pretext of superficial similarities. The actor may be affected in a real way, even if we can only tell by certain symptoms that pass us by or fascinate us in turn: he hunches over, speaks with a strange accent, and madly waves his golf club like he’s stirring the forest into a frenzy.

One issue that might come up is what it means for something like a virus to change the human as a whole. Maybe the virus directly contacts certain things (e.g., cell membranes or the operational structure of the cell), and these then translate diffusely into an impact on the host, who feels like crap. Thus, it might be said, one can agree with Harman that an object is affected “vicariously” by means of its parts or shared relations without embracing the doctrine of withdrawal. But this would avoid the question that Harman wishes to address in his ontology: how does the virus contact the cell membrane? In order to maintain a notion of direct impact, the explanation keeps pushing to quantized units (chemical components, atoms, etc.), each one functioning as a kind of plastic material for the level above it. From there, we can only fight the dualism between matter and form by means of asserting their constant combination in a single entity, such as a self-shaping surface. Harman is the one in this situation who would claim that you can’t have it both ways. You can’t have both direct contact and object-oriented translation.

Another issue is that terms like “vicar,” “buffer,” and “mediator” can be taken to imply a neutral space between two or more things, a shared line of contact such as the iconic noodle in Disney’s Lady and the Tramp. Harman himself sometimes describes objects as linking up like magnets via opposite poles: real-sensual-real-…

But that kind of symmetry around a center (A-B-A) is anathema to OOO. It is surely convenient to say that the noodle in the restaurant scene references a single event with multiple points of contact: a noodle shared by two dogs, like a bridge crossed by squirrels, dust, warlords, and diplomats. But for the characters, anyway, it participates in two different life-changing experiences by the accident of bringing each one in for a kiss. Quick ‘n dirty, let’s say one of them focuses Lady’s skittish wonderment at the large world through her attraction to the footloose tour guide who made it possible; the other one gives shape to Tramp’s careless chaos through the ego-stroking enjoyment of his companion. (The filmmakers are aware of this asymmetry, too—only one character does a cartoon dog impression of a shy blush.) Aside from the question of the noodle’s role in the larger story or the single fascinating moment it facilitates for some viewers, the noodle qua sensual medium serves as two different vicars for two different realities.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s