murder by structure and latour

I’ve been making my way through Latour’s Reassembling the Social in preparation for teaching it to my graduate theory class on the advice of scatterbrains. (Note to self: consider reading books before assigning them?) Meanwhile, I’m also working through my large pile of back journals and came across Andrew Papachristos’ “Murder by Structure” in AJS 115:1.

The contrast is interesting, at least if I have it right.

Papachristos uses a really creative dataset and approach–essentially, network data about the “exchange” of murders among gangs for three year-long periods in Chicago–to interpret murder, particularly for “expressive” reasons, as a gift exchange. Murders are “traded” between gangs intent upon maintaining symbolic and territorial monopolies over the use of force. The article is great IMHO, for its use of data, its mixed method (extensive fieldwork and in-depth interviews alongside the police network data), and its theoretical innovation. It’s particularly cheeky–and therefore elegant–to argue that a gift is “reciprocated” when the gift is lethal, i.e., the recipient is by definition dead before s/he can reciprocate! Thus the reciprocation cannot be individual in any sense–it must be social, the product of the collectivities that frame action in this context. It’s a very nice article, if morbid. I’m sure I’ll teach it sometime.

Meanwhile: if I understand Latour correctly, he begins with a strong claim that action is something done by actors, which may be individual people, or may be artifacts. Action is carried out through networks connecting these actors. Crucially, though, Latour rejects the idea that collectivities can act (I’m writing this away from my copy of the book, so can’t offer textual support. If I’m wrong please correct me. Did I mention that I’m teaching this book later in the semester? Or that I encouraged the students in the class to read scatterplot?). This constitutes a major break with the French tradition of Durkheim, Foucault, Bourdieu, all of whom in one way or another argue for collectivities as ontologically real, albeit constituted and acting in different ways.

I don’t mean to suggest that there’s a contradiction here, or that Papachristos proves Latour wrong, or anything so grand as all that. I just noticed an interesting contrast between two interesting works I’m reading right now. But Papachristos’ work strikes me as demonstrating that the most parsimonious way to think about the data is as collectivities acting. You could re-draw the networks of gangs as networks of individuals nested within gangs, but that would add complexity that doesn’t actually explain anything, since the murders are always exchanged between gangs, not between individuals (for obvious reasons). For those of you who consider Latour a revelation: am I right that Papachristos’ approach is a-Latour-ic? Is that a problem?

Author: andrewperrin

Johns Hopkins University - Sociology and SNF Agora Institute

20 thoughts on “murder by structure and latour”

  1. I’m pretty interested, but unconvinced, at your characterization of Latour as rejecting “the idea that collectivities can act.” I’m away from my books too (let’s be honest–I probably wouldn’t have the time to scan texts on the second day of the semester) so I may be mislead here…but when I think of the way that Latour describes the “delegation” of authority to door hinges and “porters” in his “Missing Masses” essay, I imagine he must mean “we” (people) collectively act (we build doors to turn tombs into rooms, then put in doors and hinges and later porters), and pass knowledge of that action–that delegation–to others, who imitate us, and expect the door to close after we pass through it. No?


  2. Jenn, I’m very much not an expert on Latour, and I have not read the door piece, but it looks great and I hope to read it soon. I’m getting the sense that Latour rejects the idea of collectivities doing the acting from, e.g., page 5 of Reassembling the Social: “`social’ is not some glue that could fix everything including what the other glues cannot fix; it is what is glued together by many other types of connectors.” … “social does not designate a thing among other things, like a black sheep among other white sheep, but a type of connection between things that are not themselves social.”


  3. Both of you seem to assume that Latour has been consistent throughout his writing (the door piece is from 1992). Am I the only one finding myself asking “What was the argument again?” when reading stuff published after circa 2000?


  4. For what it’s worth, I recommended something else!

    That said, I’m teaching a graduate seminar this Fall on Bourdieu, Foucault, and Latour. Any profound insights into Latour are welcome. I’m the least secure with him. But like you, I guess I’ll learn!


  5. I think the notion that Latour breaks with collectivities in the Durkheimean sense is right in that he sees himself as working more in the Tardean tradition, if there is such a thing, given Durkheim’s obliteration of Tarde. But you might say that for Latour (again taking into account the changes in his views over time, as phnk points out) the networks / assemblages of human and non-human objects are collectivities of a kind that ‘act’ because they are linked together in ways that allow them to sustain challenges, or ‘trials of strength’ and so on. For me what’s been more difficult to grasp is not so much his understanding of the social, collectivities than his conception of what it means to ‘act’ in the first place.


  6. Regarding the Papachristos piece, I would point out that Bourdieu wrote of revenge-acts as the functional equivalent of gift-acts. Both revenge-killings and birthday presents are exchanged via receiprocity. Neither are really direct “products” of the social. They both derive from the “sense of honor” inculcated in individuals, especially individuals within “pre-modern” social structures. They are the direct product of individuals, though the ultimate producer is society, you could say. This anyway is PB’s argument (in Outline of a Theory of Practice).

    Regarding Latour, I’ve never been able to pin down where he stands on the question of society. Once an assemblage gets really big–enough actants are enrolled–it seems to take on a life of its own. But more in a taken-for-granted way (black-boxed or reified, a la Berger and Luckmann) than a sui-generis way.


  7. In my opinion, network analysis makes it possible to understand collective action, even assuming – and maintaining – ad individualistic perspective: in a word, it is often proposed as a possible solution to the so called “micro-macro link problem”.
    Even if groups cannot materially act, there is no doubt that the “feud” mechanism can be maintained only if the offence is conceived in a collective way, and a member of the group is obliged to the revenge to (re)gain honour to the group as a whole. At the same time, in Italy, there were specific norms that obliged specific “individuals” (a brother, a son etc.) to revenge.
    Networks may be of use in empirically analysing this kind of social mechanisms.
    Thank you for the interesting post.


  8. AV, I’m skeptical. Network analysis, as far as I can tell, offers tools for describing and visualizing these processes, but the theoretical work is generally subsumed in the content of the ties between nodes. In other words, the presence or absence of a network tie does not explain action (or the lack thereof) by either node; it’s the content of the tie between them that may contain the explanation. This doesn’t so much solve the micro-macro link as document it!

    Furthermore, if the operative question is: can collectivities act even outside the volition of a given individual within the collectivity, then the network-analytic tool of just assuming that it’s nodes who act makes the whole question a tautology.


  9. Network analysis should not be reduced to a complicated method to describe a group. Only sociologists do so: the purpose should be understanding how groups or communities “act”.
    It is clear that I assume as a matter of fact that collectivities do not act, because they simply cannot do it.
    Collectivities act by means of individuals, signs, symbols and institutional objects. This is the “Durkheimian” side of many individualisms (Latour as well as Boudon, but even social interactionism or phenomenology).
    The point here is understanding why sometimes (often, always or never)people act complying collective norms and values.
    On the contrary, assuming as a matter of fact that collectivities do act, network analysis appears to be unable to explain what really matter, i.e. collective actions.
    To be honest, this is not completely true, since network analysis may reveal emergent properties of collectivities, not derived from individuals’ actions.
    In any case, in my previous comment I just tried to say that there is no tautology nor contradiction in assuming an individualistic approach, and attempting to resolve empirical problems that naturally imply collectivities, such as a theory of “feud”, revenge or gift.
    There are several good reasons to be skeptic on both network analysis and individualism, of course, but the attempt seems consistent with the premises.


  10. I assume as a matter of fact that collectivities do not act, because they simply cannot do it.

    …which is precisely why the approach you advocate cannot evaluate whether, in fact, collectivities can act. I’m certainly not denigrating network analysis in general, but if you assume away the possibility of action at a super- or supra-individual level, you can’t test whether such action is possible.

    “All sorts of processions, dances, and songs had been underway
    by torchlight since nightfall, and the general effervescence was
    increasingly intense. At a certain moment, twelve of those present
    each took in hand a large lighted torch; and, holding his own torch
    like a bayonette, one of them charged a group of natives. The blows
    ere parried with staves and lances. A general melee followed. Men
    jumped, kicked, reared, and let out wild screams the torches blazed
    and crackled as they hit heads and bodies, showering sparks in all
    directions. “The smoke, the flaming torches, the rain of sparks, the
    mass of men dancing and screaming—all that created a scene whose
    wildness cannot be conveyed in words.”

    It is not difficult to imagine that a man in such a state of
    exaltation should no longer know himself.”


    1. I think that it is not evaluable, because it is a postulate, not an hypothesis. As well as my (semi-)individualistic assumption, of course. It is a typical issue of social ontology.


      1. It’s a claim. My point early in the post was that, in a way, Papachristos has found an instance where it must be the collective that “acted,” because the gift-recipient is now DEAD and therefore, most would agree, not able to act. So it’s not evaluable, perhaps, using a strictly scientific approach, but it’s nevertheless amenable to evidentiary argument, but only insofar as you don’t assume it away before commencing that argument!


  11. I agree on this point, definitely. That is why I’ve found your post so interesting.
    My point in my first comment was that this is not contradictory with an individualistic approach.
    Asserting that only individuals “act” does not imply that collectivities do not exist, and do not produce effects.
    I only do not know if the word/concept “action” – as used in social theory – is appropriate to describe what collectivities do.
    Just a question, to clear this point. What kind of actions are you referring to, in general terms? What do collectivities do?


  12. Here is my best shot at what I think Latour would
    say about this. In Reassambling the Social he states
    “it is clear… that the notion of local interaction
    has just as little reality as global structure (203)”
    So I don’t think he would claim that these murders had
    much to do with the individuals who committed them.

    Further back on page 166 he says: “Interactions do
    not resemble a picnic where all the food is gathered
    on the spot by the participants, but rather a
    reception given by some unknown sponsors who have
    staged everything down the last detail.”

    So I think what Latour would say is that here the
    gangs have sponsored these murders. And that
    Papachristo, being a good ANT, followed the traces
    (murders) back to their actors, and let those actors
    articulate the setting in which they were located
    (the gang). As he states on page 205: “especially
    important is that which allows the actors to interpret the setting in which they are located.”

    The key, according to what I think Latour is saying,
    is that the setting or context comes last and that “the social” or “the gang” or whatever is an actor, but as he says somewhere “only one actor”. We should
    not begin with “local interaction” or “macro theory”
    but start with the traces and let the actors tell us
    what the context is. If what I’ve said about Latour
    is true, then I think that, on the contrary,
    Papachristo’s study is an exemplary of ANT. (I should mention here that I haven’t read this paper of Papachristo’s, I just relied on Perrin’s characterization of it.)


  13. Charles: you seem to have quite a good grip of what is going on here. Therefore, would you mind writing down the one-line summary of Papachristo’s study

    1. from a macro-theoretical perspective
    2. from a local interactionist perspective
    3. from a Latour/ANT perspective

    And then try to answer the triple question: Which approach is the most…

    Q1. easy to articulate in writing [legibility test]
    Q2. easy to follow in the field [practicality test]
    Q3. direct path to the explanans [parsimony test]

    (I wrote the test in counter-lexicographic preferential order, which is just a complicated to say I value parsimony over practicality, and value both over legibility. Which must be why I like Bourdieu.)


  14. (Disclaimer of confusion: the difference between local interaction and ANT is, to me, equivalent to the difference between a chicken and a chicken wearing a chicken costume. More feathers, but makes the same sound in the end.)


  15. I’m not sure I buy either Charles’ or phnk’s approach here. Charles, I don’t think Latour is willing to let “the gang” be an actor, since it’s an assembly in itself, and he’s so into disentangling the traces of these assemblies.

    phnk, can you say more about what “local interaction” is? It’s not a term I’m familiar with. IMHO neither legibility nor practicality is a particularly important test of a theoretical approach, since the point is to accurately represent and/or explain, not to do so easily (necessarily).


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: