my preference in terms of sorting out descriptions/techniques would be to take on case-studies and show (rather than abstractly say) what is gained and lost (to the degree that we can capture such things, obviously any view-point/lens by foregrounding certain things leaves others out of focus, and than many things just exceed our grasps) by the applications of the differing approaches/styles.Case studies are indeed essential. The central intuitions of ANT have to be maintained, one of which is: we can only learn through trials in which the researcher makes herself open to transformation by her participants; she makes herself plastic, quite literally impressionable.
However, we cannot go into a case study 'unequipped'; without 'tooling up,' so to speak, beforehand! This brings me back to the modes project and the relationship between fieldwork and metaphysics in general.
Can't we take metaphysical arguments as 'wagers' that practically precede but do not empirically predetermine fieldwork? We might never be sure when turning up for work whether we'll need a hammer or a scalpel but if we familiarise ourselves with both tools and bring them along with us then we will be both sensitive to and prepared to deal with the issues that arise.
Arriving with only one tool, one mode - that of networks - brings to my mind a joke: give a man a hammer and every problem starts to look like a nail...
Having a whole set of tools (not just one or two) requires us to make a decision, it requires us to think about the kind of problem we are facing. It may even prompt us to realise that none of our tools fit and we have to go 'back to the drawing board' - that is, back to the space of abstraction, of planning, of architecture, of philosophy.
A plurality of metaphysics makes us more impressionable because it makes us more uncertain, more hesitant as to our next step. It makes us a little bit more 'lost' in our moment but, at the same time, increases the number of routes open to us.