I am reading Graham Harman's Object-Oriented Ontology: A New Theory of Everything. I'm finding it interesting and a lot of the ideas resonate, although I'm quite sure I don't completely understand it.

The second chapter, on Aesthetics, has an extensive discussion of metaphor, and this seems to be a foundational concept (as it is referred to a lot later). This includes five "lessons", and I'm struggling with several of them. (That is — not that I reject the ideas, but that I do not understand the concept or argument as presented.)

The first, that art (and metaphor in specific) gives us (however obliquely) access to "real objects". I this take as basically an axiom of the system rather than something that is proven. The second, though, has me lost. It is this claim:

[M]etaphors are non-reciprocal, since one of the two terms is inevitably in the subject position and the other in the object position (in the grammatical sense of subject and object).

I don't understand — or disagree with? — the argument for this, but perhaps more importantly I don't understand why this lesson is important for the overall theory. This has me concerned for my overall understanding (as a child cannot really understand multiplication if they don't get the concept of addition).

Harman bases his explanation on an essay from José Ortega y Gasset, in which there is an example metaphor from a poet (López Pico): the cypress is like the ghost of a dead flame. Harman likes this, but says:

We begin with the topic of asymmetry. Ortega's mistake comes in the following sentence, which details the fusion of the cypress and the flame: 'we are to see the image of cypress through the image of a flame; we see it as a flame, and vice versa.' But why 'vice versa'? Here, he is too hasty, and fails to think through the consequences of his claim. For if there were really a 'vice versa' between the cypress and the flame in the metaphor, then we ought to be able to flip the order of the metaphor without this leading to any change.

He proceeds to offer "a flame is like the ghost of a dead cypress" as an inversion of the line. But... this makes no sense to me, as it treats "ghost" and "dead" as part of the template, or filler, or something. But, I don't think it's something that can really be ignored in impact of the metaphor in the first place, and therefore must participate in the inversion.

Isn't the inversion simply: "the ghost of a dead flame is like a cypress"?

He continues:

In the first case, the metaphorical object is a cypress with flame-qualities; in the second, it is a a flame with cypress-qualities. Hence, the assumption of a 'vice versa' at work in metaphor is a fatal error.

But, as I read it, the metaphor does work both ways. That is, when I read the original, I have associations brought from my memories and experience of dead and dying fires and my knowledge and experience of cypress trees, and together they make a new combined idea which doesn't seem to be directional.

There might be some grammatical argument centered around "well, actually, this is a simile..." — not for sheer pedantry but because by definition simile makes a more limited comparison — but I don't think that's really relevant. Anyway...

Later in the chapter, Harman gives another example, from Hemingway: 'The sail was patched with flour sacks and, furled, it looked like the flag of permanent defeat.' For this, he imagines a literal flag of defeat, and gives this inversion: 'It was patched with duct tape and, furled, it looked like the flag of a decrepit, elderly fisherman.'

He says "Both metaphors work fairly well, but they are obviously not the same."

But, isn't the inversion, rather: "A flag of permanent defeat looks like a sail patched with flour sacks and furled"? Or, with more exposition as given in Harman's version: "A flag of permanent defeat looks like the tattered, furled sail of an elderly fisherman"?

Again, leaving aside grammatical concern, I don't see a meaningful directionality.

But, Harman says this is critical to object-oriented ontology:

What does Ortega lose by missing out on the asymmtry between cypress and flame in his example? He misses one of the pillars of OOO: the deep divide or tension between an object and its qualities.

There is then a long explanation of this point, which makes plenty of sense to me on its own. Then, further expansion into an argument against objects as only "bundles of qualities" — which I also think I follow — and then continues to discussion of real objects, sensual objects, and sensual qualities. All of this seems fine, until:

Whereas Husserl allows us to distinguish between the the sensual object snowmobile its sensual qualities, Ortega gives us a different distinction (or would have, if he had not missed the asymmetry of metaphor) between the real object 'snowmobile' and its sensual qualities. The reason the qualities are still sensual in Ortega's case rather than real is because while the object 'cypress' in the metaphor is deeper and more mysterious than can be summarized by listing flame-qualities, these flame-qualities are by no means hidden from the one who experiences the metaphor. We know what they are, even if sub-verbally, and simply of difficulty attaching them to a cypress.

(Italics and quotes are in the original — I'm not sure why snowmobile is in quotes when related to real object and not previously.)

Harman continues, but now just in summarizing what he is just said now using SO, SQ, RO, and RQ as shorthand for sensual and real objects and qualities.

I believe I understand exactly what he's saying about the way the sensual qualities sub-verbally relate to the real object — I'm not sure a logically-convincing argument is presented here, but as I said, this is what he later summarizes as the first feature of metaphor, which I'm willing to take as a given of the framework. That is, metaphors are about the RO-SQ relation.

But, I'm totally, completely missing why the point about asymmetry matters. As I see it, the cypress-flame metaphor brings together the objects 'cypress' and 'flame' through subjective qualities of each, resulting in a new object which mixes the two. This idea of symbiosis comes up in the next chapter, and here's where I'm concern that my lack of understanding of this building block is hampering my further understanding, as Harman says:

Third and finally OOO stresses the nonsymmetrical nature of symbioses. As was seen in the case of metaphor, symbiosis is not a literal case of two objects exchanging common features and benefits, but of one object stripping away the qualities of another.

So, it's clearly important! But...

  1. What am I missing in Harman's argument about inversion?
  2. Is there a different way I could understand this asymmetry?
  3. Why is asymmetry so important to the RO-SQ tension?
  • 1
    Cypress has the quality of flame but not vice versa, this is simply like inheritance relation in ooo where the child could be said to strip away its parent qualities, not unlike an awk shell had the qualities of a sed… Jun 28 at 23:56
  • 1
    @DoubleKnot: Linux analogies is a bit niche..
    – CriglCragl
    Jul 26 at 20:29

1 Answer 1


It makes me think of aboutness, that one thing is being used from something, towards another thing.

I look at "the cypress is like the ghost of a dead flame", and it strikes me the ghost-of-a-dead-flame is not a thing, but a sensation of having an afterimage from an extinguished flame, and in moonlight the silhouette of the tree has that shape, and the sense of unreal ghostliness. I presume this metaphor, despite the poem it's from apparently not being available in English that I can find, has been chosen because it is challenging and elusive, it evokes a sensation. Poetically, in terms of the purpose of the metaphor, inversion is then non-sensical, a cypress tree cannot add to your understanding of the flame.

I also think of Witthenstein's saying vs showing distinction. Metaphors, and poets, stretch what can be said. Objects, the shown, can give rise to language, but transcend it, the description of the sound of an instrument say is a very poor stand-in for someone who has never heard it. We can point at sensations, and through our intersubjectivity we share experiences - such that we often forget that words cannot contain all of our experiences, we need cues and context and to have the thing we are discussing 'in front of us' or in our shared way of experiencing (which he called forms of life) while we learn it's labels and how to use them (eg 'game' is easy to show positive examples, but hard to delimit - we develop a sense of relevance, rather than a precise deginition).

I am all for Bundle Theory, which I see as aligning with Mahayana Buddhist ontology. Objects are always changing, so they don't have an essence seperate to the properties, their interactions; it is just useful for us to 'chunk' their phenomena together, and we have a cognitive bias towards narrating events happening to subjects. But Kant for instance held to 'noumena', the thing-in-itself that transcends specific knowledge and interactions; I guess that has relevance for considering how an object has unrealised potentials, hidden qualities, and a certain degree of uniqueness. For me the dispute comes down to how you consider the status of mathematics, and whether abstractions are less real or more real than objects, with Kant saying things are 'out there' as inferred logical objects, like math is.

Symbiosis, is interesting. Many symbiotic relationships evolve from parasitism, through mutualism. When obligate symbiosis has occurred, the organisms provide essential services to each other such that they cannot survive without them. I'm not convinced it's useful to see one side of a symbiosis as always stripping away qualities of another, or one side as always being bigger, or having more control, to say so is open to challenge from the vast panoply of evolution, and doesn't get at something critical to symbiosis. It's more that if symbiosis is more successful than living seperately there will be faculties which become a waste of energy, harming a symbiote that keeps them vs one that saves the energy. The cell nucleus and mitochondria are thought to have begun as parasites, but began entirely new paths of evolution through what could not be done without them.

You will likely find 'Surfaces and Essences: Analogy as the Fuel and Fire of Thinking' by Hofstadter & Sander, an interesting alternative approach to the topic. There are a lot of examples of how our language is really structured by metaphors, to such an extent we tend to forget how much we use them.

  • To be clear: OOO is opposed to bundle theory. In short, as I understand it: 1. Sensual qualities may change but we still have a sense of a consistent thing beyond those qualities (take a bite of an apple, still the same apple), and 2. It explicitly accepts Kant's "noumena" -- but these are not abstractions like math. (Harman calls reducing things to their component bits "undermining", and reducing them to abstract generalizations "overmining.)
    – mattdm
    Jun 26 at 14:22
  • Anyway, thanks for your comments. Right now, I'm not looking for wider approaches to the general topics brought up here, but rather specifically trying to understand Harman's meaning in the context of his own theory.
    – mattdm
    Jun 26 at 14:23
  • Kant's noumena are uncontactable, we can only derive approximations from phenomena, the noumena must be logically inferred, never directly experienced. It's not so easy to dismiss bundle-theory, see Anatta, in Buddhist thought, or the Ship of Theseus, or Sorites paradox. Or just Heraclitus' 'We can never step in the same river twice'. You might find this answer interesting, on the problems with our intuitions about isolated objects for how we think about causality: philosophy.stackexchange.com/questions/70930/…
    – CriglCragl
    Jun 26 at 15:53
  • Sure, I'm aware of that. However, I'm trying to understand the theory as presented.
    – mattdm
    Jun 26 at 17:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .