In this talk I will try recommend the idea of memory assemblages, a central notion in my current project around specters and addition. It is an ultrametaphysical (or postmetaphysical) notion and, in this sense, it is perhaps not even properly an idea – it contrasts both with ousia (or presence, or substance) and with physis (or truth-as-unveiling, or real object). By ultrametaphysics I mean the project of thinking away from the usual ontotheological assumptions of metaphysics inaugurated by Heidegger and carried on in different veins after him by philosophers like Derrida and Malabou, of whom I borrow the phrase ‘ultrametaphysics’. Memory assemblages are also distinct from Heidegger’s turn towards the event (Ereignis) which tend to be still oriented towards the present, if not towards presence. Memory assemblages are in contrast with presences because they involve an anachronic coupling of what is placed in retention and what is eventually retrieved. Things are not given or fully contentful because whatever they keep in store depends on eventual retrievals. The correlationist circle, to use Meillassoux’s term, is a memory assemblage where, say, concepts are employed to retrieve whatever is stored in a parts of the world. But the correlation is only one among the memory assemblages, they are, I argue, ubiquitous.

This ultrametaphysical panmnemism (or pan-archivism) holds that the past – even what is made available in perception – needs to be remembered and therefore depends on what will be added in the future. Memory assemblages couple what is at store – not fully present, not fully contentful – with what is insistently added to what there is. Whatever has happened – and that includes any event - leaves traces in retention and cannot be taken to be fully forgotten (or fully remembered) because the addends bring in new possibilities of retrieval (think of archaeological research but also to frozen germs becoming effective with climate change or the megafauna coming back to life through genetic, and epigenetic, research). There is a sense in which this vindicates some writings of Parmenides concerning the impossibility of lapsing into inexistence – Severino himself, who perhaps promoted the greatest revival of Parmenidean notions in recent years, takes what seems like lapsing into nothingness as a storage, except his storage is thought of as independent of any specific retrieval. If retention always needs retrieval, one can find in Benjamin’s criticism of historical truths – more than in Kantian ideas - the embryo of a generalized suspicion towards metaphysical knowledge. 

In a world full of memory assemblages where what is kept is always hostage to what will come from the exterior, the doctrine that memory assemblages are ubiquitous is also itself an memory assemblage. This makes the notion of memory assemblage not only ultrametaphysical but also ultraspeculative for it doesn’t lead to a broad thesis concerning how things are – as speculative gestures are suppose to lead – but rather to some way of storing things that is itself beholden to future retrievals.