three types of causality

"Each of the entities I am dealing with possesses the three types of
causality that [Bloor] keeps so cleanly separated: each of them is
self-referential -- like society in his scheme -- each of them is
causal -- as in what he calls non-social nature -- and yet each of them
underdetermines the next one in the series -- as in his conception of
the gap between sensory data and interpretation. Instead of
distributing his three types of causality according to different
ontological domains (one for non-social nature, one for social nature,
and one for their connection), I attribute each of them to all of the
entities. This is what has allowed me to tackle anew the question of
what society and nature are made of. I claim that this question has
become empirically studiable only since this methodological move has
been taken."

- Bruno Latour, "For David Bloor... and Beyond: A Reply to David
Bloor's 'Anti-Latour'"