Here I will speak of the distinction between primary and secondary qualities, long viewed with scorn by cutting-edge philosophy, but back in the center of discussion today. John Locke is the most famous theorist of primary and secondary qualities, if not the first. Primary qualities are those that must belong to an entity whether or not they are perceived, while secondary qualities exist only insofar as they are perceived. Roses would continue to reflect the same wavelength of light (primary quality) even if all sentient beings were exterminated, but they could not in any way be experienced as red (secondary quality). The sensation of red requires the interaction between roses and eyes. Thus, roses could not be red if all eyes were destroyed, any more than if all roses were destroyed. Nor would the rose have a rosy smell if all noses were annihilated; the smell of a rose requires the existence of noses no less than that of the flowers that are smelled. Locke generalizes the point further:

What I have said concerning Colours and Smells, may be understood also of Tastes and Sounds, and other the like sensible Qualities; which, whatever reality we, by mistake, attribute to them, are in truth nothing in the Objects themselves, but Powers to produce various sensations in us, and depend on those primary Qualities, viz. Bulk, Figure, Texture, and Motion of parts; as I have said.;1

Nor did this problem die out with the seventeenth century thinkers. One of the most significant books of present-day philosophy opens with the same theme. Quentin Meillassoux writes as follows: “The theory of primary and secondary qualities seems to belong to an irremediably obsolete philosophical past. It is time it was rehabilitated.”2

Meillassoux’s proposed rehabilitation of the theory hinges on the following claim:

all those aspects of objects that can be formulated in mathematical terms can meaningfully be conceived as properties of the object in itself. All those aspects of the object that can give rise to a math- ematical thought (to a formula or to digitalization) rather than to a perception or sensation can be meaningfully turned into properties of the thing not only as it is with me, but also as it is without me.3 

In the brief space remaining, I will make three claims. First, Locke’s distinction between primary and secondary qualities must be maintained. Second, Meillassoux is wrong to identify primary qualities with those that can be mathematized. Third, Locke is partially wrong to hold that colors, smells, tastes, and sounds are secondary rather than primary qualities. Let’s get right to the point.

There are two possible ways to deny the existence of primary qualities. One is to deny the existence of anything primary at all. On this view, everything is secondary: nothing exists except insofar as it is perceived by or related to something else. The second way would be to concede the existence of a primary reality beyond all perception, while still denying that this primary reality possesses anything like “qualities.” Let’s deal briefly with both of these objections. The first denial is found in most extreme form in George Berkeley’s delightfully sarcastic idealist claim that “it is indeed an opinion strangely prevailing amongst men, that houses, mountains, rivers, and in a word all sensible objects, have an existence, natural or real, distinct from their being perceived by the understanding.”4 Here no objects or qualities are primary because, for Berkeley, nothing exists apart from being given to the human mind. But no human-centered idealism is required to deny the existence of primary qualities.

Bruno Latour, for instance, tries to place all entities (or “actors”) on the same footing, granting no privilege to human over inanimate actors. But while this differentiates him significantly from Berkeley, it does not prevent Latour from defining entities solely in terms of their effects on other entities: “there is no other way to define an actor but through its action, and there is no other way to define an action but by asking what other actors are modified, transformed, perturbed, or created by the character that is the focus of attention.”5

What Berkeley and Latour share is the assumption that realities exist only in relation to other realities. In this respect they both deny anything “primary” that would exist independently of anything else. The reason we cannot accept such a claim is that if reality existed only in relation to something else, it would be fully exhausted by that relation. There would be no surplus of reality not currently deployed or expressed in the world here and now, and hence no reason why the current relations between things would ever change. 

The second way of denying real qualities would be to agree that objects have a primary reality preceding their relations, but simply to deny that this primary reality has anything like qualities. The clearest exponent of such a position is my fellow object-oriented ontologist Levi R. Bryant.6

In Bryant’s philosophy, there is a distinction between the non-relational “virtual proper being” of an object and its “local manifestations,” and only the latter can be said to have qualities.7 Yet it is difficult to see how this restriction of “qualities” to the relational realm is anything more than a terminological decision on Bryant’s part. Presumably two dogs, a toothbrush, a skyscraper, and a clown all have different “virtual proper beings,” since otherwise these objects would be exactly the same.

Bryant does not want to ascribe qualities to these “virtual” entities, but prefers to distinguish them through “what they can do” as opposed to “what they are.” But there would be no distinction between what they can do unless there were already a difference in what they are, and I see no reason not to use the term “qualities” to describe the features that differentiate what they are. Obviously, since these qualities precede any relation insofar as they might generate a large number of possible relations, we must call them primary rather than secondary qualities.

That brings us to Meillassoux’s claim that primary qualities are those that can be mathematized. Aside from a recent half-finished attempt in an unpublished Berlin lecture from April 2012, his clearest statements on this topic are perhaps those found in his 2010 interview discussion with me.8 While conceding that he has not yet demonstrated “the capacity [of mathematics] to describe that which is independent of all thought,”9 Meillassoux insists that mathematics can do this despite being a human construction. He offers the analogy of archaeological work, in which

the ‘constructions’ (a complex of winches, sounding lines, scaffolding, spades, brushes, etc.) are not destined to produce an object, as in the case of architecture. On the contrary, they are made with a view to not interfering with the object at which they aim: that is to say, excavating the ruins without damaging them, in unearthing them ‘as is’, and not as modified or even destroyed by the impact of the excavation tools.10

However, the problem with mathematizing entities is not that mathematics is a human construction. The problem, instead, is that mathematics claims to present reality directly in a way that is utterly foreign to other disciplines. Archaeology never confuses sarcophagi with the winches and pulleys used to lift them. By contrast, mathematics does encourage the belief that its models of things are exact replicas of those models. The mathematical ontologist may well concede that the things have some material substrate in which its mathematizable features inhere. But the nature of this substrate remains untheorized in Meillassoux, who occasionally calls it “dead matter” but then fails to explain the difference between such dead matter and the mathematized version of it that is supposed to yield its primary qualities.

We have now argued as follows: (a) primary qualities must exist, (b) they cannot be mathematized, which is another way of saying that they cannot be thoroughly known. Belief in a reality outside the mind must not be confused with the belief that this reality can also be directly known, as Meillassoux demands. Instead, respect for reality requires a respect for its somewhat inscrutable character, not directly translatable into any form of knowledge, so that it can only be known indirectly or allusively, with a sideways glance.

Now, along with speaking of primary and secondary qualities, we must also speak of primary and secondary objects. For just as certain qualities belong to a thing without anyone seeing it (primary qualities), while oth- ers require the presence of a perceiver (secondary qualities), the same duality can be found in objects. For just as we presume the existence of genuine things apart from the mind and indeed apart from all relation whatsoever, so too there are objects existing only in relation to the mind, such as centaurs and golden mountains.11 This might also seem to be true of entities such as sounds: a musical tone, after all, seems to exist only for some hearer. We might assume that the “primary” object is merely a physical sound wave, while the musical tone is a merely “secondary” object, and as such, only the primary sound wave could have primary qualities, and the secondary musical tone could only have secondary qualities that exist for some hearer alone, and not in themselves.

Paradoxically enough, the world does not work in this simplistic a fashion. Let’s change the terminology of this essay from the more classical “primary” and “secondary” to the related terms more prominent in my previous works: “real” and “sensual.” For object-oriented philosophy, real objects are those that withdraw from relation and exist quite apart from any relations in which they might become involved. Sensual objects are 

those that exist only in relation to some other entity, and they must be distinguished from their sensual qualities, due to Husserl’s critique of David Hume’s influential “bundle theory” of experienced entities. Hume famously holds that we do not experience a unified object such as a zebra, but only a series of stripes, hairs, eyes, mouths, legs, and tail that seem to move together with such regularity that we form the habit of taking all of these together to be a single thing. But Husserl’s phenom- enology overturns this theory by showing that Hume has it backwards.

We do not encounter qualities or parts in isolation, but first encounter the object itself. What we actually encounter is not a dissolute set of partial objects, but the zebra as a whole. This zebra can be viewed from many different angles and distances and still be regarded as the same zebra. This zebra is a sensual object rather than a real one, since it is fully present in my consciousness, not withdrawn in the least, unlike the spooky subterranean entities that populate the tool-analysis of Husserl’s renegade pupil Martin Heidegger. And this sensual zebra has numerous and ever-shifting sensual qualities, which sparkle across its surface and pass away forever without affecting the integrity of the sensual zebra. And yet, this phantasmal zebra in the mind has other qualities that cannot pass away without consequence.

There are certain features that the zebra must always retain, under penalty of my no longer regarding it as the same thing, but instead as a different zebra, a warthog, or perhaps even a wagon, sunbeam, crutch, or puzzle. These real qualities that the zebra needs, in order to remain the very zebra that it is, can be called its real qualities. Husserl concedes that these real (or “eidetic”) qualities cannot be known through the senses, though he mistakenly assumes that they can be known through direct intuition by the mind, as if the difference between the sensual and the intelligible were such a massive rift.

We now have the paradox that a sensual object can have real qualities, or in quasi-Lockean terms, that a secondary object can have primary qualities. This allows for a strange intrusion by sensual, phantasmal, and purely fictional entities into the realm of autonomous reality. Even if we agree with the most hardheaded realists that Popeye, Don Quixote, centaurs, and hallucinations are not real objects in the same manner as neutrons or cliffs of limestone, Popeye and his cousins must still be acknowledged to have genuine, real qualities. 

Perhaps this should not come as a surprise, given that by an analogous paradox, real withdrawn objects signal into the realm of perception by means of sensual qualities, in a manner exploited most thoroughly by the arts. Even if the sounds heard by human and animal ears are not as autonomous as waves travelling through air, these sounds immediately generate qualities that are every bit as independent as physical disturbances. The wizard of the mixer who makes a sound scratchy or ghostly, who transposes it into a different key or gives it a different mood, is merely a profiteer reliant upon the durable real qualities of a sensual thing— as if a hallucinated apple had a more-than-imagined color, smell, taste, and even price. 



1. John Locke, An Essay Concerning Human Understanding, p.137. (Oxford : Oxford University Press, 1979.)
2. Quentin Meillassoux, After Finitude: Essay on the Necessity of Contingency, trans. R. Brassier, p. 1. (London : Continuum, 2008.)
3. Meillassoux, After Finitude, p. 3. Emphasis removed.
4. George Berkeley, A Treatise Concerning the Principles of Human Knowledge, page 24. (New York : Cosimo Classics, 2005.)
5. Bruno Latour, Pandora’s Hope: Essays on the Reality of Science Studies, page 122. (Cambridge, MA : Harvard University Press, 1999.)
6. See for example Bryant’s post of January 5, 2012, “More on Withdrawn Objects,”
7. See Levi R. Bryant, “The Ontic Principle: Outlines of an Object-Oriented Ontol- ogy,” pp. 261-278, in L. Bryant et al. (eds.), The Speculative Turn: Continental Materialism and Realism. (Melbourne:, 2011).
8. See “Interview with Quentin Meillassoux,” pp. 159-174 in Graham Harman, Quentin Meillassoux: Philosophy in the Making. (Edinburgh: Edinburgh University Press, 2011.)
9. Harman, Quentin Meillassoux, p. 167.

10. Harman, Quentin Meillassoux, p. 167. 11. I have often argued this point in connection with the philosophy of Edmund Husserl. See for example Graham Harman, The Quadruple Object, pp. 20-34. (Winchester, UK : Zero Books, 2011.) 

Commissioned for Not for Human Consumption

Graham Harman is Distinguished University Professor at the American University in Cairo.

He is the author of eleven books, most recently Weird Realism: Lovecraft and Philosophy (2012) and Bells and Whistles: More Speculative Realism (2013). He is the editor of the Speculative Realism book series at Edinburgh University Press, and (with Bruno Latour) co-editor of the New Metaphysics book series at Open Humanities Press.

Download PDF