In my older (2003-2004) terminology polyagents were anything that had polyagency (which simply was a generic term for human and non-human agency). Back then I was also influenced by Bachelard’s discontinuous view of time (as opposed to the continuous view of time of Bergson and Deleuze which I later felt to be more convincing). After my dissertation period I also realized that I should drop the polyagency part all together. The idea of material agency vs. human agency largely relies on a dialectic world view (of which Bachelard adhered to). Bergson and Deleuze take us away from that view. In any case, I continued to use the polyagency idea in my dissertation but it is a bit awkward there since it tries to squeeze the dialectic view into a non-dialectic view…
Anyway, these polyagents (artifacts, causeways, pyramids, potholes, etc.) have an internal and external world. It is in the encounter with other polyagents that polyagents first get disorganised before the situation become organized. The boundary between polyagents is therefore a set of ways to deal with the otherness. Boundary crossing, as when a polyagent interacts with another polyagent, lead to a discontinuity in spatial and temporal relations. These discontinuities provide shelters for more organized spatial and temporal settings. The time shelters are “persistent forms of event-discontinuity in the world” (Wood 2000:226). Thus, the interaction between polyagents form time shelters, but polyagency itself is not temporally or spatially located, it lies in-between.
Ballgame rituals, causeway processions, earthquakes, daybreak, etc., were or are such time shelters. These time-shelters all have a semiautonomous temporal organization, which mean that they are not just organized in time but of time. I therefore argued that we need to suspend the idea of a unity of time. All times do not line up in a single series that have the same form or that are serial. The unity of time is a principle of boundary permeability. Time shelters are what we believe is real (Wood 2000:228ff).
In my dissertation I briefly used the time-shelter idea. Here the time shelter became the actualization that forms a frame, a territory and an internal boundary. The time shelter shelters an internal duration that remains united as an actualization, until it breaks up when it is destratified/deterritorialized in another direction. Boundary crossing occurs when an actualization interacts with another actualization, which leads to a discontinuity. These discontinuities provide new time shelters for spatial and temporally folded settings.
Causeways are such actualized time shelters that shelter various foldings of duration (the duration of limestone in bedrock, duration of the quarried limestone blocks, sascab and plaster). A causeway is a time shelter in relation to its various phases of actualizations. Within the world there is a constant boundary crossing between causeways and other polyagents that deterritorializes and reterritorializes different strata. Thus, existing actualizations are the territorial surfaces from which human goals emerge and from which the social world is formed. This means that existing actualizations affect our goals (Normark 2006).