Leaving aside the idea that ‘difference differs’ (which suggests to me that you can’t actually identify a difference because identification is re-identification, cognition is recognition, and something that is always differing isn’t stable enough to be identified) I am claiming that there’s a paradox here, but it’s not exactly the one you outline. To put the matter simply, ‘difference’ is an insufficient notion for developing an ontology and this insuffiency is exhibited in the lack of context and context-dependence that differentially construed objects exhibit.
The idea seems to be that if everything differs then we don’t get enough stability within being for entities to exist, much less be theorized. It seems to me that this criticism moves too quickly, failing to explore the resources that difference offers in responding to these sorts of issues. What this criticism fails to take into account are differences in scale and duration characteristic of the world. A blooming flower is constantly undergoing change at each moment of its existence, both at the cellular level of the processes or operations being undergone by each cell and at the the level of the petals, opening and turning towards the sun. On the one hand, I am unable to perceive the cellular activity of the flower because of the scale at which I exist. Likewise, I seldom notice the process the flower undergoes as it blooms, instead noting only the result or final product of the now open flower. If this is the case, then it is because of the relatively slow duration that characterizes the movement of the flower with respect to the duration that characterizes my being.
Following Graham and Latour, it can be said that the flower is a black-box or has undergone “punctualization” with respect to me. That is, I treat the flower as an identical and unchanging object, oblivious to the heterogeneity that characterizes its being. This phenomenon of punctualization is not simply an epistemic phenomenon, but rather is ontological. The tree relates to the other trees in its vicinities not as a heterogeneity, but in punctualized form, as an obstacle to its light. A rock falling down a mountain does not encounter the mountain as a constantly changing processes at both the molecular and molar level, but as a series of obstacles informing its trajectory down the side of the mountain.
Moreover, in contemporary cognitive neuroscience we get a conception of cognition that is, precisely, differential in character, where abilities such as re-identification and re-cognition are the result of constant operations, not unlike the way cells must perpetually make themselves, and where each repetition introduces a difference in the neurological system as a whole. Re-identification and re-cognition are results and effects of these differential processes, products, not conditions.
The real question strikes me not as the question of how recognition and identification are possible, but rather as the question of when differences become appreciable in relations among assemblages. At each moment my eye contracts millions of photons of light. The physicists tell me, in turn, that these photons are themselves constantly differing, being both waves and particles, and that each of these packets of energy or photons are themselves undergoing all sorts of processes and variations. Yet when I look at this table before me, all I see is the color brown. At what point, through what process, does this brown shift from being the deep, rich brown that it now presents itself as to being a nearly blinding, shiny silver color as occurs when I shift my head slightly?
If there were no differences of scale and duration, Alexei would be right. However it is precisely these differences in scale and duration that allow for the phenomenon of punctualization or “black-boxing” that enable contexts or plateaus to be formed. This is also the importance of the notion of singularity as a difference that constitutes an appreciable difference. Singularities can only be spoken of in terms of inter-assemblic relations because what counts as a singularity will be a function of the assemblages entering into relations with one another.
Thus, for example, while it is ontologically true that a subatomic particle, as a being, makes a difference, it is generally unlikely that a subatomic particle constitutes a singularity for a wall. This is because, depending on the sort of particle it is, it is able to pass right through the wall as if it didn’t even exist. Likewise, the wall, for such a particle, offers little in the way of singularities in relation to the particle. Singularities, as I argued in “Margaret’s Pepper Principle” and “Towards a New Transcendental Aesthetic“, are points of density, condensation, friction, or resistance, thereby playing the role of appreciable differences in the genesis of a form or quality. However, when referring to singularities, it is always necessary to ask “with respect to what assemblages?” “With respect to what scale?” “With respect to what duration?”