While it is not unusual to speak of “early” and “late” phases in a philosopher’s career, Bruno Latour is perhaps unique in having gone through both phases simultaneously.Harman has written words along these lines several times before. It's becoming something of a catchphrase for his commentary on Latour. The problem with this remark, however pithy (and I like pith), is that Latour is very clear in pretty much everything he's published on the subject in the last few years* that he was always working on his 'modes' project, he just happened to concentrate on the 'network' element for a long time. So, the only 'phases' are the phase when he was writing about one mode and not speaking openly about the rest and the present phase when he's speaking about all of them together.
Harman's quip (and I like quips) gives the impression that Latour has produced notably distinct and different philosophies at different times. I don't think this is at all accurate. If we take what Latour himself says seriously then the former work was simply one element of a larger project that is now coming to fruition. It's a subtle distinction but it matters.
I like Harman's first book on Latour, the 'Prince of Networks,' but it does prove that the author is dead. Harman's Latour doesn't always bear that much resemblance to the Latour I find in the his own books. There's nothing wrong with this as such -- I've written in the past about how Latour's own public persona is a fiction -- so it's not about authenticity, it's about being candid and clear about just what's going on.
Harman's Latour is a translation, a continuation-with-deformation that Harman fabricates largely for his own purposes. Harman's Latour, as presented in Prince of Networks, is a stepping stone, a foil for Harman's own philosophy. He says 'this is what Bruno said, this is what he got right, this is what he got wrong, this is how I'd improve on it...' But his reading isn't ever neutral, it's always oriented towards the latter part of the book, the 'how I'd improve on it' part, which is the real end -- the earlier parts are the means. He has to present Latour in a certain light in order to present his own thought as an improvement on his precursor's. In order to do this he gets bends some concepts, leaves some stuff out and gets some other things just plain wrong. Again, nothing wrong with that but it is what it is.
I look forward to reading the Prince of Modes and I don't object to Harman's simplifications as such, it just concerns me that many people will read Harman's Latour and think that this is Latour's Latour. Both are equally 'fictions,' conscious and specifically stylised presentations, but they do differ significantly and this isn't as widely recognised as it should be.
Going back to the original point, the whole idea of 'phases' is rather inadequate to describe Latour's many trajectories. I think the concept of a tree captures it better. There's a largely continuous if somewhat twisted trunk, a distinctive signature, running through everything, even back to Laboratory Life and beyond. However, there are many divergent branches and branches of branches of varying thickness and convolution. Many of them represent tangents have been abandoned or that have changed beyond recognition. (Okay, it's not a perfect metaphor but I think it makes my point.)
So, his thought has changed, certainly. But does it form anything like a pair of 'phases'? I don't see it at all. Nor, according to what he's written, does Latour himself. He's never been self-consistent enough to corral into two of anything and yet, at the same time, his key concepts haven't changed enough to warrant any kind of temporal demarcation. No befores and afters, even if they're happening at the same time.
It seems to me that the before and after myth that Harman is crafting lets him keep his critiques of Latour1 while bracketing off any exceptions or remainders that don't seem to apply to that figure by giving them a home in Latour2. By setting up two relatively purified figures he then has a duo of fixed targets, each of which perform different purposes in his own narrative. If he was forced to acknowledge that there's both more and fewer than 2 Latours his narrative of promotion, critique and succession wouldn't work. What if, for instance, he was forced to abandon the idea that Latour 'reduces things to relations' when the counterarguments to this are legion? He'd have nothing to succeed. Cutting up Latour into two phases means he can put discordant ideas wherever they don't inconvenience the narrative.**
This argument is partially speculative and perhaps a little premature because, of course, I don't know what Prince of Modes will say, but based on his commentary to date I think I'm on the right track.
* e.g. 'Coming out as a philosopher', his Gifford lectures, to name but two.
** See the example of 'plasma' for an example of where a concept is comprehensively redefined in order to suit the narrative.