Been digesting further and further Latour's process philosophy and looking at ways in which there is no genuine correlationism there. But feel more and more attracted to some version of actualism like the one process philosophy and my ontology of fragments affords. Placed in a proper network of alliances I can play the zither and speak Javanese. Talk about capacities and skills is talk about soft biases in the following sense: they are all ceteris paribus claims, they are claims about the current network of alliances I am inserted. Alliances can be made with the whole universe. A given network could have skills but it depends on being kept as such by the rest of the world. (A zither player can undergo, say, a traumatic experience of any sort that would wipe off his skills.) Then I feel more sympathetic towards conditional analyses: to be fragile is no more than to be broken in neighbouring possible worlds. Finkish dispositions? There are loads of antidotes that can stop a possible course of events to take place, even for good. But there is no potentiality harboured inside anything in a brute state (see Deleuze's take on coupling for he posits no dispositional properties).
I just claimed that alliances can be made with the whole universe and I made use of possible worlds - how do I understand this modal talk? I think talk of possible worlds is fine, they imply individual actants (or fragments) and their possible alliances. In other possible worlds there are no counterparts but also there is no individuation that is done once and for all. Nothing is forever either an entity or a network - only tests of strength can decide that, Latour says in Irreductions. Only with respect to an actant - a fragment - we can conceive of a possible world. Therefore, no concretism. Naming: Socrates is Socrates if the network of alliances that constitute it is close enough to the actual one. If Socrates became wind or fire or ashes thrown in the Aegean sea, the name would barely have a holder.
More on this soon.