Hold me closer, tiny actor

You have indeed brought us to new territory, and in good Latourian fashion, by translating my little methodological concerns about how Latour relates to the rest of sociology into a much more interesting discussion about how Latour himself thinks. I really like your final formulation: "Bruno Latour is the thinker of scale: it's what makes him always on the edge of the qualitative and the qualitative … and also … able to pay minute attention to the long connections of processes that allow them to pass into each other." Everything that follows might be thought of as a gloss on that remark.

First of all, let me say that I understand better now what you were getting at by questioning my interest in interests. What you're intrigued by, if I understand you correctly, is that for Latour, interests are not just human/social — he rejects the sociological/scientistic vision that separates off our petty human manueverings, "the dynamics of social groups" as you put it, from what is really happening in nature. If everything has interests, in other words, then interest is no longer an anthropocentric concept: but there's still good reason to retain it, because it gets at how everything can interact, be associated or dissociated with each other. I think Latour can seem like a strange bedfellow for anti-correlationist philosophers like Meillassoux, because one of the immediately striking things about him is that he's always talking about things as if they were human: like he never gave up the animistic universe that supposedly got exterminated during the "disenchantment of the world." But Latour does this, I think you're implying, not to assert the primacy of the human but rather to assert the primacy of the social (which includes nonhumans) over the philosophical: as you put it so excellently in discussing BL's relationship to Derrida, "Latour is saying that nothing is singular (irreducible) because it always needs others … [T]he misunderstanding comes in to affirm the fact that a thing needs others."

So we come to measurement. Again, it could be confusing that Latour is so interested in this, because it seems like such a fundamentally human concept: only humans measure things. (Probably there are anthropologists out there who argue that chimpanzees or something have systems of measurement, but you know what I mean: strict, formalized measurements, the kind that requires counting and standardized tools.) But as with "translation," he wants to take "measurement" out of the realm of the human and turn it into an ontological principle, a way that objects relate to one another, and in fact even "need" each other (because one can't act without a sense of how big or small one is). This is the metaphysical part of Latour's project (which he and Harman have both called "empirical metaphysics"); the "empirical" part comes from his attention to the ways humans — and here he's a conventional sociologist — actually do scale and measure, which he sees as dependent on a prior measurement or adequation, and thus questionable. I hope what I'm saying will become clearer as you continue reading this post.
For Latour, interests arise out of actors measuring themselves against one another, scaling each other up or down. Here's how he puts it in Irreductions 1.3.1:

All entelechies may measure and be the measure of other entelechies … Nevertheless, certain forces constantly try to measure rather than be measured and to translate rather than be translated. They wish to act rather than be acted upon. They wish to be stronger than others.

I have said ‘certain’ rather than ‘all’ as in Nietzsche’s bellicose myth. Most actants are too far apart or too indifferent to rise to the challenge, too undisciplined or devious to follow for long those that speak in their name, and too happy and proud to take command of others. In this work I speak only of those weaknesses that want to increase their strength. The irreducible others have need of poets rather than philosophers. (167)

We can see that Latour is skittish about coming too close to Nietzsche's "will to power" here (and I notice that you also contrasted him with Nietzsche; care to elaborate on that?). Again, while we get this weirdly anthropocentric language ("they wish to act," "indifferent," "undisciplined," "devious," "happy," "proud"), what Latour is trying to say, I think, is that "will" is just one way of looking at what entelechies do. In fact, for Latour, only certain kinds of entelechies — scientific networks being, of course, the prime instance — possess this "will" to "measure rather than be measured" and "act rather than be acted upon." So this seemingly anthropocentric move, giving characters and physiognomies to nonhuman actants, works against another, more common anthropocentric move (associated with Nietzsche), of psychologizing the natural so that it has a "will to power" and a desire to destroy. (I have to say though, I don't know about Latour's chivalrous offer of the irreducible, strength-indifferent entities to the "poets": seems to me poets are as bellicose and power-willing as anyone. Cf. Harold Bloom, for starters.)
This brings us back to sociology, and the micro-macro issue you asked about, because there is another type of entelechy, besides science, that Latour sees as wanting to dominate and measure everything else, and that is what are sometimes referred to in sociology as "macro-actors": institutions, corporations, governments, etc. Contrary to your flattering assumption, I don't actually have a particularly good handle on the micro-macro question: I know it was a major disciplinary concern in the eighties, because sociologists of different traditions were worried that they couldn't even talk to each other, let alone synthesize their findings into some sort of grand synoptic picture of the social world. (Jeffrey Alexander et al's The Micro-Macro Link, published in 1987, would be the locus classicus for this debate, I believe.) As for whether "whether Latour indeed addresses it," he does so explicitly in a 1981 article coauthored by Michel Callon called "Unscrewing the Big Leviathan: How Actors Macro-Structure Reality and How Sociologists Help Them To Do So." There, he and Callon make an impassioned argument for a single methodology to describe micro- and macro-actors (which, they insist, should not be used synonymously with "individuals" and "institutions"). "Too often sociologists — just like politicians or the man in the street — change their framework of analysis depending on whether they are tackling a macro-actor or a micro-actor," Latour and Callon say. "…[N]o sociologists at present examine macro-actors and micro-actors using the same tools and the same arguments." In effect, they are reconceiving the micro-macro problem as a problem of measurement: it's not that there are different "levels" which demand different modes of analysis (as in physics), but that we need common "tools" and "arguments" in order to be able to pass from one kind of actor to another. And for these authors, the consequences of the assumption that there really are different levels of society in need of explanation are grievous:

It seems to us that sociologists are too often on the wrong foot. Either, believing that macro-actors really do exist, they anticipate the actors' strength by helping them to grow more vigorous [here, I believe, he's referring to Marxian, Weberian, and other "big systems" analysis]. Or else they deny their existence, once they really do exist, and will not even allow us the right to study them [referring to ethnomethodologists and symbolic interactionists]. These two alternate but symmetrical errors stem from the same presupposition: the acceptance as a given fact that actors can be of different or of equal size. As soon as we reject this presupposition, we are once again faced with Hobbes' paradox: no actor is bigger than another except by means of a transaction (a translation) which must be examined. (280-281)

For Latour and Callon, macro-actors are not at all things of an entirely different kind or size or even complexity than micro-actors: they are, rather, "micro-actors seated on top of many (leaky) black boxes" (286). In other words, they have made many assumptions and simplifications in order to attain dominance, and these can always be challenged. (They go on to illustrate this with a bravura, though rather depressing, example of how Renault undermined Electricity of France in the early 70s to keep them from developing a viable electric car.) The point is important methodologically but also politically, because it keeps us away from the despairing feeling (which Latour associates elsewhere — cf. Irreductions 1.4.6, 2.1.10, and Interlude VI — with Hegelianism and Marxism, and which I must admit I myself associate with Fredric Jameson) that there are vast systems and structures and forces of which we ourselves are only a tiny part: in other words, that there is a macro to our micro. But Callon and Latour are very emphatic here that the macro-actor is only a network of "devious" micro-actors, and that they are in many ways simpler and easier to study than micro-actors. In one of the most charming descriptions of technological modernity I've ever read, the authors state: "A tiny actor becomes a macro-actor, just like in the French nursery rhyme: 'The cat knocks over the pot, the pot knocks over the table, the table knocks over the room, the room knocks over the house, the house knocks over the street, the street knocks over Paris: Paris, Paris, Paris has fallen!'" (296). This rhyme is particularly apropos because, for Callon and Latour, it is the sociologist's role to step in early and say, "Wait a minute, how does a table knock over a room?" — thus arresting the whole process, or at least redirecting it.
I'm going to try to sum up, because this is getting too long. Entelechies can only scale each other up or down by enlisting an incredible number of micro-actors — and ANT seeks to follow these enlistments in order to question, and thus adjust, the scales. It's in this way that measurement is, for Latour, both an ontological fact — as he says in the passage you adduced last time: "What is neither reducible nor irreducible has to be tested, counted, and measured. There is no other way" (1.1.3) — and a revisable human practice.