Challenging the ‘New Descriptivism’ – Rod Benson’s talk from QualPolComm preconference
by Rasmus Kleis Nielsen
Rod Benson (NYU) served as a respondent at our conference on Qualitative Political Communication Research in May at ICA in Seattle.
His response created quite a stir (it has become known in some circles as his “ANT-takedown”) and was clearly aimed at more than the four particular papers on the panel.
We are happy he has agreed to share the text of his talk, posted below. It’s very good, provocative, and we hope it will spur debate.
Rodney Benson
“Challenging the ‘New Descriptivism’: Restoring Explanation, Evaluation, and Theoretical Dialogue to Communication Research.”
Remarks at the Qualitative Political Communication Pre-Conference,
International Communication Association
Seattle, May 22, 2014
Given time constraints, I will cut to the chase. I’ve been asked to stir things up a bit, so I’ll see what I can do.
We’ve just heard two excellent ANT (actor-network theory) papers (Joshua Braun; Burcu Baykurt) and two excellent systems/institutional theory papers (Hannah Middendorf; Lucas Graves and Magda Konieczna).
This seems like a good opportunity to put these different approaches in dialogue. And to do so, I’d like to raise four big meta-questions:
First, where is the explanation?
Second, why should we care – what is at stake?
Third, how should we see society – as settled or unsettled?
And fourth, is it possible for ANT and institutional theories to work productively together – or are we faced only with the choices of mutual indifference or all-out-war?
(1) Let’s start with the first question. Where is the explanation?
In other words, as we saw with some of the papers in this session, the research starts and ends with a descriptive account. We wait in vain for the explanatory analysis. Why? Why these findings and not others? The answers never arrive.
I call this the “new descriptivism.”
I can’t prove it with quantitative evidence, but I’ve begun to notice a widespread tendency of more and more research offering descriptions and nothing else: very detailed, very sophisticated, very interesting descriptions, but at the end of the day, just descriptions. What has happened to explanation? (And, related, as I’ll ask in a moment: what has happened to critique?)
I can think of several reasons, not necessarily good reasons, but reasons nonetheless for the lack of explanatory analysis in so much of our contemporary research:
First, people are overwhelmed with “big data” or they’re geeking out on the new computer tools they can use to map and track this data. In this context, it makes eminent sense to begin with description and get around to explanation later. But there is so much to describe – and so much to turn into nifty visual graphs and charts – that the moment for explanation never seems to arrive.
Second, there is a lack of variation incorporated into research design. To explain, you need variation. When we explain, usually what we’re doing is saying that more or less of x contributes to more or less of y. And if you don’t have variation built into your research design, you can’t do that.
Third, people are working with theories that are antithetical to explanation, understood as I’ve just defined it, such as:
-Holistic, totalizing theories, like Foucault’s neo-liberal governmentality – in this case, the description and explanation are basically the same; or
-Theories of contingency and chaos and flux, heterogeneity (to use the ANT term), and unpredictability – in which case, everything is so unstable and unpredictable that nothing can be explained; or
-Epistemologically relativistic theories, such as ANT: in other words, to explain, some accounts must be privileged over others, and this is precisely what ANT not only will not but cannot do, given its epistemology.
Let me expand on this last point about epistemology, specifically in relation to actor-network theory. As Joshua Braun mentioned in his talk, there is a strong link between ANT and STS “controversy studies.”
What does ANT mean by controversy studies? Well, it has a close cousin in sociology – and that cousin is what is called the social construction of “social problems” or “social constructionism” (or sometimes social “constructivism”) for short.
There are two broad strains of social constructionism:
1) Strict constructionism (represented most notably by Stephen Woolgar, who not coincidentally has co-authored with Bruno Latour): social reality in media and other public forums is constructed by discourses or in turn by interpretive communities; given that we know that all knowledge is constructed, no construction can be privileged over another; hence, the role of the analyst is to try to faithfully catalogue the array of constructions.
and
2) Contextual constructionism (represented most notably by Joel Best): social construction of public problems in media and other public forums has to be situated in a context. For example, the social construction of the immigration problem, as with research for my book Shaping Immigration News: A French-American Comparison, should be understood in relation to the economy, political events, unemployment, immigration flows, etc. Of course, these kinds of official and expert discourses are also constructions, and contested ones, but nevertheless we privilege some of them (with all due skepticism) as part of the explanation of why some constructions win out over others in the public debate.
Now there are many varieties of contextual constructionism. There is the almost ad hoc kind favored by Best, the value-added model of William Gamson (social movement strategies, media logics, cultural resonances, objective factors), and there is the “structural constructionism” represented by institutional and field theories. What all of these “contextual” accounts have in common is that they do not “level” the social world and treat all accounts equally; they privilege some accounts over others in order to arrive at explanations.
So to bring this back around to ANT and institutional/field theories.
Clearly, ANT is a version of strict constructionism. ANT does not privilege one account over another. Latour, and John Law, are very explicit about this. It’s part of their credo of keeping the world flat. And this is one of the key reasons why ANT is so unsatisfying to anyone who reads it looking for anything in the way of explanation. It’s just not there and it’s not there by design.
On the other hand, institutional/field theories are versions of contextual constructionism – with the virtue of offering the kind of systematic “systemic” context needed for effective cross-national research.
ANT has the virtue of raising epistemological questions directly and of offering a consistent epistemology. ANT shifts the burden of justification to those who do want to explain, who do want to judge. If we want to explain and judge, we have to make the case for how any particular disposition from a position – how any socially-produced form of knowledge – is superior to someone else’s. Bourdieu makes the case with his notions of autonomy and reflexivity. Most of us simply bracket or ignore the question and carry on with our research.
For the moment, ANT-style relativism seems to be in vogue. In terms of epistemological consistency, again, there is an argument for such relativism. But, let’s be clear, there is a cost:
The cost is: First, no explanations – that would privilege some accounts over others. And second, no evaluations – nothing is better or worse than anything else.
(2) Which leads me to my second big question, very briefly. Why should we care – what’s at stake?
ANT studies don’t – won’t – answer this question. In return, one might respond: Why should we care about ANT?
(There might still be some reasons – I’ll return to this at the end.)
(3) Moving on quickly, third big question: What is our basic understanding of society?
Is society’s default position settled or unsettled?
Most institutional and field approaches presume settledness. What has to be explained is how things ever get unsettled.
There are periodic moments of crisis and contestation. Somebody’s solution wins out. The solution is anchored in institutions over time through path dependency; it gets naturalized so no one can imagine an alternative – until the next crisis appears and then there is the small possibility of change. Fligstein and McAdam (A Theory of Fields) are very explicit about this.
We see this presumption in the research question posed by Lucas Graves and Magda Konieczna’s paper: Under what conditions is it possible to “unsettle previously stable arrangements”?
Alternatively, ANT presumes unsettledness: society is fundamentally unsettled and the production of order is the miracle we have to explain.
Braun, quoting John Law: “How is it that some kinds of interactions appear to succeed in stabilizing and reproducing themselves”?
So is the glass half-full or half-empty? The two approaches differ in which half of the glass they want to call attention to.
This would be okay if there were occasional nods to the other half, but in fact, the accounts tend to be one-sided.
(4) Which leads me to my fourth and final question: Can ANT and institutional/field theories play nice – or not?
Yes, I think they can play nice. We’ve played nice this afternoon and all of these papers in this panel stress the need to bring multiple strains of theory together.
I want to close by listing four possible ways in which ANT and institutional/field theories can be complementary to each other: 1) temporal, 2) spatial, 3) topical, and 4) analytical.
The first is a temporal division of labor. The argument is that historical conditions make one approach more appropriate than another. We hear this argument often made for ANT but actually I think it’s the weakest possible argument.
In effect, when times are settled, we need institutional theories; when times are unsettled, as they are now, we need an unsettled theory like ANT.
For instance, Josh Braun argues that ANT is the best model for studying certain kinds of media work that now seem on the ascendance, the kind described by Mark Deuze as “project-based” liquid journalism.
Again, I’m not convinced by this argument. Things seem unsettled – at least to some scholars and journalists – but are they really that unsettled, overall, for other people, for most people?
Several studies have shown in fact that journalistic norms and practices have not changed that much in the aggregate.
So we need to keep this question of the unsettledness of this historical moment “open” – and not let our theories prejudge the answers.
In fact, far from historical conditions demanding a certain theory, I would caution that the seeming appropriateness of some new theory might mask “ideological effects.” ANT may actually be the least appropriate theory to analyze new developments in media, precisely because it valorizes and legitimizes a view of the world – of rapid change, innovation, agency – consistent with the worldview and interests of powerful interests in Silicon Valley, Wall Street, and Washington, D.C.
Maybe, either way, we just find what we’re looking for – settledness or unsettledness. But going in, if possible, rather than presuming the correctness of one or the other theory, we should see ANT and institutional/field theories as competing hypotheses and then try to let the evidence speak for itself.
Now, second, to the spatial dimension – and I find this a much more compelling way to parse out distinct roles for ANT and institutional/field theories. Gil Eyal makes this argument in his essay “Fields and the Spaces Between.” He argues that society is composed of settled fields and unsettled actor-networks that are not fields or just not yet fields. I find this argument much more convincing, but we need more research to try to prove that there are actually such differentiated social spaces and that they operate according to different logics.
Third, we have the topical dimension. Fred Turner, Nick Couldry, Chris Anderson and Daniel Kreiss, and others argue that we need a theory like ANT because it calls attention to phenomena ignored by other theories, namely technical artifacts and objects, which have their own form of “agency.” I’m not entirely sure about the idea of non-human, object-based agency. But I am persuaded that ANT has played an important role in focusing attention on the power of new technologies, and in ways potentially quite useful and distinct from McLuhanist technological determinism.
Fourth, and finally, as I’ve already alluded to, there can be an analytical division of labor. We can use different theories for different purposes: descriptive, explanatory, and normative analysis. ANT may only be descriptive, but as long as we are not relying only on ANT, this isn’t a fatal flaw. We see this use of multiple types of theories in Chris Anderson’s Rebuilding the News. He draws on ANT in the first part of his book, which has a more descriptive focus in order to call attention to the ways in which new technologies are transforming the ecology of media. Then, in the last half of the book, he uses institutional theory for explanation (explanation of persistence and lack of change in many cases) and democratic normative theory to make judgments about whether these developments are good or bad.
All such combinations of course presume heterodox rather than orthodox interpretations and uses of theory. I had a French colleague who was once introduced at some French conference as a “Bourdieusienne heterodox,” the implication being that there was something almost wrong or weird about this. In America, maybe, we can more easily get away with being heterodox. Still, we need to be mindful of the kind of heterodoxy we pursue.
The danger with heterodoxy is taking theoretical concepts out of context and totally distorting their meaning.
The potential benefit of heterodoxy is putting theory to work so that it can open up rather than close down a dialogue with other scholars. To me, that makes a lot more sense than just staying inside our churches and speaking only to our co-religionists.
Thank you.
Suggested citation
Benson, Rodney. 2014. “Challenging the ‘New Descriptivism’: Restoring Explanation, Evaluation, and Theoretical Dialogue to Communication Research.” Presentation at Qualitative Political Communication Pre-Conference, International Communication Association, Seattle, May 22, 2014,
[…] at Scatterplot, Andy Perrin has a nice post pointing to a recent talk by Rodney Benson on actor-network theory and what Benson calls “the new descriptivism” in political […]
I very much enjoyed Rod’s challenge to the field, and specifically his critique of the “New Descriptivism.” In the end, I suspect that we end up in same heterodoxical place, which has always been my own way of working.
That said, I did want to respond to this provocative essay on a couple of points.
First, a key aspect of ANT, and work that Latour cites such as Thomas Hughes’s Networks of Power, is that there is an underlying explanatory theory that temporality matters. Latour has explicitly linked ANT to historical methods more broadly in arguing for description being explanation.
That temporality matters is a theoretical claim in itself – Thomas Sewell calls this a “theory of temporality.” I wrote about this on this blog a few months back, but like Michael Schudson’s work at its heart there is a claim that all explanatory accounts have to proceed from locating things in time. To retread my earlier post, Sewell compellingly argues that historians see heterogeneous time, where certain events are highly consequential in shaping the course of what comes after. A related point is that sequences matter; the order of events and the processes that constitute those events shape outcomes. In Sewell’s view, an “eventful history,” “recognizes the power of events in history”; it “is one that takes into account the transformation of structures by events” (100). There are three aspects of eventful histories in Sewell’s account: path dependence, causal heterogeneity, and contingency. Path dependence means, generally, that earlier happenings affect the following sequence of events and outcomes. Causal heterogeneity means that causal mechanisms are not uniform and can be altered by events: “because the causalities that operate in social relations depend at least in part on the contents and relations of cultural categories, events have the power to transform social causality” (101). Contingency means that outcomes cannot be deduced from general laws; they are dependent upon everything that came before and therefore are inherently unpredictable. For historians much of social life is contingent – whether a cause has an effect, and what that effect is, depends on the context within which it is introduced. Another way of saying this is that the causes are not inherent in the agent itself. There is what Sewell calls “causal heterogeneity”: “the consequences of a given act are not intrinsic in the act but rather will depend on the nature of the social world within which it takes place” (10).
The point is that description, particularly historical description of the kind that Latour has called for and that historians routinely do, is premised on an underlying theory of what matters in terms of developing explanatory accounts of the social – and that entails not only expanding the range of actors we consider in our accounts of the world, but also tracing their work over time.
Second, I do not think that ANT approaches necessarily favor instability over stability. Across Latour’s corpus of work, he shows the processes through which things are made to become stable or durable, or alternately to become undone. The Janus face of science-in-the-making and ready-made science in Science in Action illustrates this well. Latour’s point is that things/facts in science only become stable after many long and arduous trials, but once stabilized are really hard to destabilize. C.W. Anderson and I recently showed how historical studies of electoral politics and journalism can reveal how certain social arrangements become stable (such as the way we vote) and unstable (such as journalistic production).
Third, I think there is a difference between relativism and relationality. Latour himself makes this point, and so do many other scholars. At least as I read ANT and scholars such as Stephen Fuchs who have taken up Latour’s work, the point is not so much that we can never provide accounts of the world, it is that we all inhabit different locations as observers and, in the social sciences at least, this entails wielding different sets of theories, methods, practices, technologies, and languages for making knowledge claims. The point of a book such as Science in Action is precisely to show how claims win out within certain observer communities through massive trials that pit scientists, research labs, instruments, and nature against one another. Explaining why something becomes settled becomes the analyst’s task.
But even more, Latour argues that we should respect the outcomes of these trials, even if they are never enough to forever prevent a ‘fact’ from being reopened (he uses the example of the ongoing debates over the structure of DNA). So, I see nothing that suggests that ANT requires us to treat all knowledge claims equally; it’s just the opposite, to produce scientific knowledge that counts in the observer community you want to speak to you have to go through massive trials to be taken seriously and the analyst should respect the outcomes of other fields’ trials (and seek to explain them).
So, in sum, I disagree with the claims that ANT permits no explanation and no possibility for normative evaluation. Even more, I think ANT makes no a priori claim for whether the world is settled or unsettled; the point is to show how things become settled or unsettled in a rich explanatory and empirical way.
Of course, this does not mean I whole heartedly endorse an ANT-ian framework. ANT has little capacity to explain the social meanings that Isaac Reed so beautifully shows needs to be at the heart of historically-grounded causal explanation. As C.W. Anderson and I conclude our own use of ANT’s conceptual tool-kit:
“It is not clear, for instance, that ANT has a ready-to-hand conceptual vocabulary that accounts for, in Schudson’s (2001, 423) terms, the “symbols, meanings, and enacted rituals” that make up politics as culture, or the “knowledge (including cultural presuppositions)” that underlies practice. Or, to borrow the clear and concise statement of famous sociologist and field researcher, Howard Becker (1982/2008,4), the “rationales” according to which “activities make sense and are worth doing.”
Thank you for this! For me, every line of argument here is spot on except for the last three paragraph that I think would be better clear as “…I think ANT makes no a priori claim for whether the world is settled or unsettled; the point is to show how things become settled, [or sustained in order to show their temporariness and continuous struggle for stability] or unsettled in a rich [descriptive and empirical way that are in themselves explanatory]”.
I really enjoy the challenges highlighted by this essay. A quick comment, though: the same questions are actually voiced from within ANT, that has evolved a lot since, let’s say, “Laboratory Life”. For instance, in his latest book, Latour acknowledges that there is something bigger out there (“explanations”, as Rod Benson puts it in this essay) and that arguing that everything is heterogenous is a dead-end. The book opens with that exact same argument:
[the first part of the book, Latour stages the journey of a young anthropologist who embraces ANT] “And yet, to her great confusion, as she studies segments from Law, Science, The Economy, or Religion she begins to feel that she is saying almost the same thing about all of them: namely, that they are “composed in a heterogeneous fashion of unexpected elements revealed by the investigation.” To be sure, she is indeed moving, like her informants, from one surprise to another, but, somewhat to her surprise, this stops being surprising, in a way, as each element becomes surprising in the same way.” (An inquiry into modes of existence, p. 33)
Later on, there’s even a rather blunt mea culpa on the limitations of ANT:
“This theory played a critical role in dissolving overly narrow notions of institution, in making it possible to follow the liaisons between humans and nonhumans, and especially in transforming the notion of “the social” and society into a general principle of free association, rather than being an ingredient distinct from the others. Thanks to this theory, society is no longer made of a particular material, the social—as opposed, for example, to the organic, the material, the economic, or the psychological; rather, it consists in a movement of connections that are ever more extensive and surprising in each case.
And yet, we understand this now, this method has retained some of the limitations of critical thought: the vocabulary it offers is liberating, but too limited to distinguish the values to which the informants cling so doggedly. It is thus not entirely without justification that this theory is accused of being Machiavellian: everything can be associated with everything, without any way to know how to define what may succeed and what may fail. A tool in the war against the distinction between force and reason, it risked succumbing in turn to the unification of all associations under the sole reign of the number of links established by those who have, as it were, “succeeded.”” (An inquiry into modes of existence, p. 63)
Finding explanations: that’s the whole point of the “Modes of Existence” project — personally I’m not convinced by all the “explanations” that Latour then proceeds to formulate, but at least the questions are there.
Reblogged this on ANTHEM and commented:
“ANT-takedown” via http://installingorder.org/
First off, its fantastic to see such an engagement at any academic conference and I hope to keep the conversation going.
Benson’s comments help clarify both the threads of actor-network theory useful for the investigation of power as well as this project of New Descriptivism. I’ll begin with comments on this term New Descriptivism then move to some comments on selectively reading Latour before offering some suggestions for a descriptive project.
Benson combines a number of threads to create what he calls New Descriptivism or research that emphasizes description over explanation. His criticism echoes back to seminal debates over the philosophy of the social sciences, particularly its lack of a lone Kuhnian paradigm. I find this distinction between explanation and description reminiscent of the understanding/explanation divide proposed by Martin Finch in his book The Philosophy of Social Science: An Introduction. Finch at once tries to map (as many have) the fields of thought in the social sciences and posit dialogs between various quadrants. Given the divide that Benson draws, I would be interested in understanding how his definitions between the two. Why in particular does Foucaultian analysis suffers from conflating the explanation with the description? I wonder whether Benson might be attempting to describe a fault line in the social science. What is this divide? How might it be bridged?
The symptoms of New Descriptivism combine what-I-see as an exciting revitalization of Qualitative Methods in Political Communication (especially in the United States) with more troublesome trends in big data and formalized research design. I wonder whether this combination might be a bit unfair? The necessity of returning to more descriptive research (articulated by W. Lance Bennett and Shanto Iyengar) was a reaction, in part, against the formalization of the field’s methods. Since their publication, political communication has merged big data with formal methods to intensify this problem. We have to be wary of any method formalized in code especially given the trend toward large-scale studies of Twitter (a platform whose brevity seems custom-build for computational analysis). I would point to Dave Karpf’s call for kludginess in digital methods in contrast to this big data reseach. I tend to agree with Benson that the idea that big data does not solve the problem of explanation. Big data also intensifies inequities in critical research by restricting who has access to the data (see danah boyd and Kate Crawford, 2012). Neither of these criticisms, however, apply to kind of qualitative research seen at the pre-conference.
Now, we can be a bit more precise with New Descriptivism. In my mind, it refers to work: 1) aiming toward Finch’s sense of understanding and 2) drawing on the broad research labelled as actor-network theory. I have given some comments that might help describe the former (or at least start more debate) and I would like to give some comments addressed at the later, namely the shortcomings of actor-network theory.
I took a break from Latour at what-might-be-called the height of his New Descriptivism. Nested within his latest theoretical iteration Re-assembling the Social (now superseded by An Inquiry in Modes of Existence), Latour offered a “(somewhat) Socratic”, (likely) farcical dialog between a PhD student and a professor whose lone advice seems to be to just keep describing. The Professor suggests, “As a rule, context stinks. It’s simply a way of stopping the description when you are tired or too lazy to go on”. For a PhD student (like myself at the time), Latour seemed to have reduced this empirical contribution to a call for long descriptions. How else can you present all the sides as a flat ontology? It was a troubling time for my relationship with ANT.
Some time and perspective later, I realize that Latour (like all good theory) offers more of a toolbox of concepts. The latest book is not necessarily the best. I think that two concepts from Latour do help flush out the contributions of ANT: contingency and stabilization.
Benson emphasizes the binary between settled and unsettled, but that link does not resonate as much work my sense of ANT that focuses on contingency. Or how things settled might be unsettling. In the figure/ground debate that Benson alludes to, contingency reminds researchers to prepare for unexpected, unpredicted outcomes. That claim, I think, differs sharply from valorizing “rapid change, innovation, agency” but emphasizes the complex articulations of power and influence in modern society.
ANT does address power and the temporal through its concept of stablization. Latour provides a beautiful account of the doomed Aramis transit project. The book chronicles the transition from an idea to a project. It uses the concept of stabilization to how an actor-network grows and becomes more stable by bringing in new connections and parts. Stabilization then involves the drawing of longer and denser networks. I’ve always wondered if size might be the underlooked means to consider power in actor-network theory. If flat ontology is like a map, could we compare the size of territories occupied by factions. ANT then offers a means to understand the history of these networks through its document analysis and interviews.
I personally would like to see greater collaboration between ANT and articulation theory developed by Laclau and Mouffe. They might balance out the flat ontology of ANT. The challenge of a flat ontology resurfaced when I taught a course on Approaches to Media and Technology where I noticed the emphasis of Latour on the concept of the network in his famous article on the door opener. The network as a form emphasizes positives connections and missing what Ulises Mejias calls the paranodal. Laclau and Mouffe go to the other extreme to draw on exclusion in discursive formations. The inverse of faults — Latour being too positive and technological, Laclau and Mouffe being too negative and linguistic — suggest that the new New Descriptivisms might help uncover the complex articulations of power and discourse at the center of critical inquiry.
As a final statement, I’d suggest that its vital for the social sciences to embrace its hybridity. We’re always drawing on diverse sources. Andrew Chadwick has made a compelling case for hybridity in political communication. He also draws upon both field theory and actor-network theory in the book. The social sciences face the same problem outlined in the parable of the elephant. All perspectives offer some truth. I think that we should always keep that in mind.
Lots to chew on here, and I take all of these responses very seriously.
I had a chance to revisit this issue and these comments recently in a talk I gave at Concordia University in Montreal. Herewith some additional thoughts, by no means the last word:
Several of the comments made here note that ANT is not in fact wholly descriptive, that it does offer explanations. It’s just that these explanations are rooted in specific historical conditions unique to each contingent case. This would be the defense of much of history or ethnography, as well, in that thick historical or ethnographic description offers a kind of explantion. The explanation of something lies in recounting all that preceded it; the explanation lies in temporality, and that it is always case specific. This is basically what Latour argues explicitly in “Technology is society made durable.”
In some case, perhaps, that may be perhaps the only kind of explanation or the most appropriate kind of explanation we can offer.
However, what I had in mind as explanation involved a bit more parsimony and a bit more ambition. Without trying to reduce complex phenomena to a single factor, there are good reasons why we would want to identify the three or four most important factors and why we might want to try to generalize beyond the individual case: for one, to facilitate understanding to facilitate action, whether activist or governmental.
I find George Steinmetz’s “critical realist” understanding of theory and the use of cases to advance theory far more satisfying and compelling than Latour’s.
In his 2004 essay Steinmetz makes the case for a theoretically-driven, comparative research program that tries to steers a path between large-scale quantitative research oblivious to context and the small scale qualitative research that insists that all cases are ultimately incommensurable.
See: G. Steinmetz, “Odious Comparisons: Incommensurability, the Case Study, and “Small N’s in Sociology.” Sociological Theory, Vol. 22, number 3, pp. 371-400.
And for my formulation of a similar type of project, specifically for media sociology, see:
R. Benson, “Strategy Follows Structure: A Media Sociology Manifesto.”
In Silvio Waisbord, ed., Media Sociology: A Reappraisal (Cambridge, UK: Polity): 25-45.
As Steinmetz (ibid., p. 394) rightly insists, “the production of sociological knowledge involves movement among case studies, comparisons among case studies, and theory.” In other words, this kind of sociology continues to generalize – carefully, yes – but yes, we can say the word: “generalize” beyond a single case. It does not allow context, complexity, and contingency to completely disable this legitimate sociological aspiration.
The other issue where I still await clarification is the political and normative underpinnings of ANT. What’s at stake here? I read through so-many ANT-inspired studies and I get to the end and think: so what? Sorry to be so blunt, but the findings so often seem so trivial. Obscure highly technical objects we hadn’t thought about much before matter … but how much really? To answer such questions, we need more comparative design: cases where the object was part of the equation and cases where it wasn’t, but maybe there is simply a dearth of situations like this to test.
And finally, what are the implications of the finding that this or that object made a difference? Concretely? In a way that might generalize beyond the specific case?
Rod, thanks for returning this. An important and interesting discussion. As you note, the kind of explanations ANT seems more inclined to offer are “explanations are rooted in specific historical conditions unique to each contingent case”. I would place it–along with theoretical cousins like various varieties of post-structuralism–as in the more humanistic end of the social sciences, bearing in mind the old idea of ideographic humanities and nomothetic natural sciences with the social sciences uneasily perched between–with law-seeking nomothetic ambitions but largely ideographic, humanstic roots. As you point out, this is how much history-writing and ethnography works (whether employing a fully articulated and sometimes convoluted theoretical vocabulary or not), largely embracing the search for specific, contingent, situated explanations, but also ultimately, however indirectly, taking up the challenge that you point to, of generalizing findings. We all in some way fail at this, as I see it (i.e. we rarely produce truly general results), and I suppose then it is in part a matter of intellectual taste whether one wants to fail from the bottom up (ideographic explanations falling short of general applicability) or from the top down (nomothetic explanations not really identifying the general laws or regularities that they seek)?
[…] Rod has now returned to his original post and added a response to the many comments, with more to come. Check out the discussion here. […]