John Levi Martin



Note:  This is a response to Neil Gross's thoughtful review in Contemporary Sociology.  Neil has been offered the chance to reply here at whatever length he desires.... watch for his counter punch!

One has to sympathize with Neil Gross. A brilliant and promising theorist, he was for some reason assigned to review the book Social Structures, even though he does not particularly like reading about social structures. Still, he dutifully slogged through the book, understood it quite well, but couldn’t help missing the things that make sociology interesting, like stories of individuals and their lives—sort of like the critique of the New York Times as having "too much news, not enough human interest." Now I admit it is my fault for beginning with Simmel; it easily can give the impression that I am somehow joining in the efforts of those executors of his intellectual will to complete his project (whatever that might be). Thus if I misled anyone into thinking that this should be judged by the standards of Simmelian orthodoxy (whatever that might be) I will repent of that. (Interestingly, the one Simmelian method I do use – armchair reflection as to subjectivity – Gross disapproves of; perhaps it is because I argue that this only has heuristic value and not the phenomenological validity that Simmel apparently assumed. Of course, assuming does not make it so.) The key issue is what it means to "go inside" the actors, for it is this quasi-novelistic aspect, the absence of which so troubles Gross.

As some may know, this is an issue that has been of interest to me for a bit of time, and so what I say here may be incorrect, but it is not off the cuff. Social Structures does not deal with this inside-the-head, make-it-come-to-life, phenomenological aspect of social interaction for two reasons, both extremely good. The first is that the book had its own focus and was able, for better or worse, to pursue that independently. That is not to say that information about the subjective correlates of structure would not be extremely pertinent. But it is not necessary for the project of understanding the formation and development of small social structures. It would be desirable, certainly, but (and this is the second reason for the omission) it is impossible. Simmel simply made stuff up. We of course do not tolerate this sort of thing in this day and age. Instead, we talk to informants and have them make it up for us. (This argument, and the implications for research practice, will be defended at greater length in the forthcoming Explanation of Social Action, so I drop the discussion at this point.)

So what I could have done, and what I think Gross would have preferred, would be to (1) find the regularities in the structures that emerge by surveying a range of social structures [which is done in Social Structures]; (2) illustrate these with specific cases not briefly [as is done in Social Structures] but in greater depth [not done]; (3a) make detailed interpolations as to the subjectivities of the actors involved [which is not done in Social Structures] instead of (3b) trying to deduce purely on formal grounds the simplest possible rules of action compatible with the observed structures [which is done in Social Structures].

Carrying out 3a might make for a richer and more persuasive account, but probably a worse book. Take perhaps the simplest heuristic I discuss, that of the popularity tournament—"go with the winner." I could use novels, interview studies, personal reflections, logic, ethnographies, and even films, to try to explain why popular people are liked more than unpopular people (though as I point out in a recent article, this isn't always true). The thing is…I have examined this issue enough to be extremely doubtful of all of these as valid sources regarding the cognitive processes that might be involved in friendship choice or even regarding the subjective experiences that accompany said choice. If I included them, the argument would seem more plausible, and it would certainly make for a more pleasant read, but it would be less likely to correct. There is nothing wrong with preferring Simmel’s account to mine (or the nineteenth century to the twenty-first, for that matter), but I believe sociology must return to the issues regarding the place of subjectivity in its explanations, and make a distinction between the enjoyable and the defensible.

Science is not always fun. And as Isaiah Berlin said, there is no guarantee that when you find it, the truth will prove to be very interesting.

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Just added! (At the time I am writing this, which could be years ago):  A response to Paul DiMaggio's very kind review in the American Journal of Sociology:

In my haste to get to the substantive matter, I did not give a sufficient explanation of the analytic method, leading DiMaggio (2011, 1669), like others, into the understandable error of imagining that the relations surveyed were ones that I picked on some sort of whim, and that the results would be quite different were other ones the focus of each chapter.  That is, there are a large number of symmetric relationships besides friendship—rivalry (raised by DiMaggio) is one, taking-the-same-bus-to-work is another, and so on and so forth.  Why choose one as opposed to another?  There were three principles, good ones, guiding my approach. 

(1) My choice was based on the fact that my goal was to determine what we know about how certain relationships have tendencies to spontaneously take on structural forms – I was looking only at “simple” structures that could at least be analytically “grown” outside an organizational or institutional framework.  Taking-the-same-bus-to-work is not a relationship that can be seen as simple as opposed to scaffolded by another structure; further, no one has ever made the claim that large social structures come from the compilation of relationships of taking-the-same-bus-to-work, nor will they nor should they.  Such claims and investigations have been made about friendship and alliance, and this is the sort of claim that should be taken seriously. 

(2) I was looking for relationships in which it was not merely that the content was formally symmetric in structural terms (such as taking-the-same-bus-to-work) but where the content was, to the greatest extent possible, nothing other than this structural symmetry.  That is, is there a relationship in which the action-profile of both parties is “assume the structural position of the other person”?  As I recount, alliance is clearly and recognizably such a relationship; friendship is widely understood to be a close second (and this leads to the four Spinozan rules of balance).  It turns out that it is not quite that way, and I show why.  But this is the correct place to begin for the investigation.

(3) Finally, with a proper selection based on principle (2), it should turn out, and it does turn out, that (contrary to DiMaggio’s supposition), the results are robust to swapping in different relationships.  Thus DiMaggio suggests “rivalry.”  If we put aside the question of whether to count any person as her own rival, and simply eliminate this degenerate aspect of the structure (as I state in Social Structures, reflexivity in these cases is purely a matter of convenience and has no substantive weight), we can pose the following questions:  a.  Is rivalry symmetric?  Yes it is.  b.  Is rivalry transitive?  Yes it is; that is, just as with friendship, we will understand an inherent tendency for it to be so, and any cases in which it is not are more complex and involve more complications from other structures than those in which it is.  Hence rivalry would form cliques the same as friendship.  However, rivalry is not as good a case to begin with because (i) in no case in which I am aware has a person ever claimed that rivalry structures could be important for coordinating large scale action; (ii) in no case in which I am aware has a simple rivalry structure been identified (that is, one in which the rivalry is not embedded in another set of relations, such as organizational or kin relations); (iii) in no case in which I am aware has anything like any rivalry structure at all been identified as something that has the possibility of stability.  There is only one example of which I am aware in which cliques of hostile rivals existed in some sort of stable formation for a long time, and this was the University of Chicago social sciences division,* and I am happy to say it finally collapsed under its own ugly weight.

 For these reasons we find that DiMaggio’s other offered suggestion, substituting exchange for dominance, also is not a helpful one.  First, exchange is not a simple relationship, but a compound.  If you give me butter and I give you bread, this is two relationships, in that I could give you bread without you giving me butter and although there would perhaps be a different sense to the bread giving, it would still be a bread-giving.  (This is not true, or so I argued, for “advice-giving” inside an anti-symmetric patronage relationship and that outside – they actually feel different, as if bread bought was toast and bread given wasn’t.  Whether that argument is accepted is unimportant for the point here.)  In contrast, dominance is precisely that relationship that is not only a fundamentally asymmetric content, but a necessarily fused asymmetry in which the content is itself that asymmetry. 

* My source here are stories of a presentation James Davis made perhaps around 1980? where he displayed sociograms of positive and negative ties in the UC SSD.  Note to self:  go down hall and ask him if this is true before posting this.  Reply to self: oh my, have you internalized a peer reviewer in your head?  You know you aren't going to check this; it's now or never for fixing this page.  And anyway, it's still a good story.