Abstract
Thomas S. Kuhn’s (1970 [1962]) Structure of Scientific Revolutions (SSR) and Bruno Latour’s (1987) Science in Action (SiA) were amazingly successful vectors for spreading the authors’ ideas through expansive networks. Perhaps I should add that the ‘ideas’ that these books provoked were not necessarily those of the authors, as both Kuhn and Latour endeavored for many years afterwards to qualify and disavow some of the ideas for which they were credited and blamed. There also are some key differences between the journeys the two books have traveled. Kuhn’s book is one of the most widely read and cited academic books of the past half-century, and its success has had no rival in history, philosophy, and social studies of science. It also was, and is, widely read and cited in fields that are quite remote from history and philosophy of science. Degraded versions of Kuhn’s use of the terms ‘paradigm’ and ‘paradigm shift’ are now part of the vernacular, and are frequently used to mark epochal changes of all kinds. Latour’s SiA launched actor-network theory (ANT), and provided (and still provides) an introduction to that theory for many students. Interest in the book is by no means limited to a single, specialized academic community, but the networks through which it circulates are still far more localized than those of Kuhn’s SSR.
It is impossible to take stock of either of the two books separately from their uptake. While it is clear that SSR was by far Kuhn’s most significant book, it is not necessarily the one he would most liked to have been remembered for writing. Unlike Kuhn’s other major works on the Copernican revolution (1957) and the history of quantum theory (1978), SSR is a synthetic interpretation of the Harvard Case Histories in history of science, which provided a concise interpretation using a range of current ideas in philosophy and the social sciences. SiA is one of many books by Latour, and perhaps not the most significant one for securing his reputation within and beyond STS. SiA also is a synthesis – a textbook of sorts – that draws heavily from other science studies research, especially the laboratory ethnography presented in Latour and Steve Woolgar’s (1979) breakthrough book Laboratory Life. Moreover, while SiA ranks with the most widely read and cited books in STS, it may be less significant for the broader field of history, philosophy, and social studies of science than Steven Shapin and Simon Schaffer’s (1985) Leviathan and the Air Pump. Nevertheless, for this journal and the STS field, SiA now represents a youthful successor, exactly half the age of the venerable SSR, which has largely claimed the space once occupied by ‘Kuhnian’ sociology of scientific knowledge.
In this brief comment, I suggest that Kuhn’s and Latour’s philosophies are self-exemplifying. Though based on historical case studies and, in Latour’s case, ethnographies and popular sources as well as historical studies, both present general philosophies. Although Kuhn’s philosophy can be classified as neo-Kantian and Latour’s as stridently post-Kantian, both have effected transformations that roughly follow the trajectories the authors trace for the cases they describe. It would be worth reflecting upon the extent to which these transformations have not attained (and perhaps cannot attain) the robust stability and technical specificity that Kuhn and Latour, in their own ways, attribute to the sciences.
Self-exemplifying revolutions
Kuhn famously counseled against present-centered histories, but sometimes hindsight can be a source of insight. Viewed in retrospect, Kuhn and Latour succeeded marvelously with SSR and SiA. Indeed, it does not take too much interpretive license to treat the success of each book as being exemplary of the respective author’s comprehensive philosophy. Kuhn’s book effected a revolution in history and philosophy of science, and set the stage for a ‘paradigm shift’ in the sociology of scientific knowledge during the 1970s which is widely regarded as a crucial turn in the establishment of STS (a convenient conjunction of letters that covers two names: Science and Technology Studies, and Science, Technology, and Society). Latour, in turn, followed the path laid out in his insightful and entertaining history of Louis Pasteur and the Pasteurians (Latour, 1983, 1986). Though the Latourization of STS hasn’t quite attained the scope of the Pasteurization of France, Latour’s skilful exercise of enrolment, interessement, and the other organizational achievements he ascribes to Pasteur is at least as impressive as his theorizing about them. As both Kuhn and Latour teach us, however, such success is contingent: many would-be revolutions fail to catch on, and many efforts to enroll allies are ignored or rebuffed. The conceptual rubrics both authors deployed did not predict success, though in both cases they turned out to be highly successful in the extent to which they were taken up and used by others.
Both authors most likely were surprised by how well their particular books caught on, but there is no mistaking the ambitions they expressed from the very outset. Both Kuhn and Latour deploy revolutionary rhetoric. For Kuhn, transformative revolutions are both a topic (of historical research) and an aim (for historiography). The very first sentence of his introduction boldly proclaims an aim to ‘produce a decisive transformation in the image of science by which we are now possessed’ (1970 [1962]: 1). The means for doing so is ‘a historiographical revolution in the history of science’ – one that, he says, already was underway, although the historians contributing to it may not (yet) have realized that they were effecting such a transformation (p. 3). Kuhn presented this historiographical transformation as an instance of a gestalt switch – an expository device he used for framing shifts between incommensurable paradigms in physical science and natural philosophy. Kuhn proposed to overthrow the reigning paradigm in history of science, which viewed history through the lens of present-day textbook accounts of established theories, laws, and methods, and which inquired into how those contents were discovered, established and resisted. His new paradigm framed past, now defunct or discredited, sciences in their own conceptual terms and in their own historical contexts. This is a form of historical realism (Sharrock and Read, 2002), though Kuhn is often taken to be promoting a version of relativism.
Latour also portrays two diametrically opposed pictures of science in SiA. However, instead of using the alternating figure of the duck-rabbit, he presents two simultaneously visible faces of a Janus figure, one of which voices dictums about ‘ready-made science’ and the other of which makes opposing statements about ‘science-in-action’. Rather than forcing a choice between incommensurable propositions, Latour goes on to integrate them within a scheme that begins with science-in-action and produces ready-made science as a contingent outcome of successful research. Through the magic of historical inversion, ready-made science then becomes the stable antecedent for the concretely situated actions that continually generate it. For Latour (although he’s not always consistent on this point), ready-made science is not (as in Kuhn) a misleading picture of science to be superseded by the new ‘paradigm’ of science-in-action. Instead, Latour’s Janus figure suggests that the ready-made version’s talking head is eternally fused with the science-in-action talking head, in the manner of a Siamese twin whose skulls and brains are fused at the back. Accordingly, to accept ready-made science at face value is misleading only when it is treated as an accurate image of how research proceeds in action. For Latour, ready-made science becomes a product of contingent trials of strength that may someday turn out differently. However, once the dust is settled, the venerable image of science does not lose its hold on well-established scientific findings, at least not until further notice.
SiA is written as a uniquely organized textbook. It digests a large body of work into a coherently presented set of topics, illustrative vignettes, idiosyncratic diagrams, and inverted propositions and principles. (An example of an ‘inverted proposition’ is Latour’s (1987: 99) Third Rule of Method, in which antecedent and consequent positions in a backward-looking realist pronouncement – ‘Nature is the cause that allowed controversies to be settled’ – are reversed and recast in the future tense: ‘Nature will be the consequence of the settlement’.) Latour’s text is unique, because it has none of the leaden style of a conventional textbook. It is written with tongue in cheek. Latour (p. 7) parodies the inscription at the threshold of Dante’s Hell, when counseling his novice reader to ‘abandon all knowledge of knowledge’ when embarking on the investigative journey. Not only that, he invites us to abandon an entire array of vernacular terms and distinctions that frame the way we commonly talk about things, people, actions, politics, and causes and effects (ultimately disowning the term ‘knowledge’ itself). The abundant diagrams (evidently drawn with an early Apple Macintosh program) are quirky and challenging to decipher, and a lively narrative threads itself through a series of stories drawn from academic and popular case studies of scientific and technological innovation.
SiA certainly works as an introduction to ANT, if not to STS as a whole. The main problem with it is that it has been taken much too seriously by too many people in the past 25 years. It is now 13 years past the publication date of a volume that announced the aftermath of ANT (Law and Hassard, 1999), but that volume and much of what followed it further expanded the ANT colony into novel territories. There is much to celebrate in the wake of such expansion and, like many in STS, I have learned from (and have learned to draw selectively from) some key works by Latour, Michel Callon, and others. But, speaking as an editor who has read upwards of 200 new submissions per year over the past decade, I have been led to the sad conclusion that the volume of BADANT (Banal and Derivative Actor Network Theory) greatly exceeds the well-researched, original, and broadly informative written work that rides under the ANT banner. (ANT is by no means the only theoretical perspective that has suffered this fate.)
Given the half-century that has passed since its publication, it is not surprising that Kuhn’s SSR is not often treated anymore as a source of novel insight into the history and social study of science. SSR is mentioned much more than it is used, but the uses and abuses of Kuhn’s ideas in the past half-century have been legion. As noted earlier, SSR has traveled further and more widely than SiA, but in both cases we can inquire into how and how well they have traveled. Have they given rise to stable ‘normal sciences’, robust technical systems, or expansive disciplines? Should they? Would we expect or want them to? Or, have the innovative and transformative themes and conceptual rubrics in both books become more diffuse as they have diffused?
Actor-networks and disciplinary matrices
One point of intersection, if not commonality, between SSR and SiA has to do with what Kuhn (in his 1970 ‘Postscript’ to the second edition of SSR) calls a ‘disciplinary matrix’ and Latour (in SiA) calls an ‘actor-network’. Kuhn begins with a discussion of what he calls scientific ‘communities’ focused around particular lines of research. He cites sociological research (specifically, bibliometric studies of ‘invisible colleges’) as providing evidence of the outlines of such communities, but unlike the social scientists he cites, he goes on to elaborate upon the socio-technical origins of the ‘disciplinary matrix’: symbolic generalizations such as concepts, constants, and formulae; metaphysical commitments of a specialized sort (heuristic models and preconceptions that underpin research activities and their sense); communal values that define proper work and well-formed results; and, finally and most importantly, ‘exemplars’ (models and pedagogical problems that provide taken-for-granted starting points for further work). For Kuhn, the elements of a disciplinary matrix are deeply integrated with a specialized community: without the specific symbolic, material, and moral elements of the matrix, there would be no community for the biblio-metrologists to map and reorganize. It is not a giant step from Kuhn’s disciplinary matrix to Latour’s actor-network. Both are heterogeneous, and both fold social, semiotic, and material elements into an overall nexus that is at once technical and communal. There is one large difference, and that is the conception of agency with which Latour perfuses the network, but I will pass over that difference here.
At the risk of reinforcing familiar lines of demarcation between ‘science’ and ‘metaphysics’, I will briefly suggest that the philosophies enunciated in SSR and SiA have traveled somewhat different pathways from those of the sciences that Kuhn and Latour describe. Kuhn and Latour, in their different ways, both stress that the linkages that integrate and hold together disciplinary matrices and actor-networks are not limited to semiotic links between words or citations. The outlines of research ‘communities’ may be indicated by common symbolic generalizations and citation networks, but the networks are integrated by material practices and infrastructures. Universities and research institutes provide sites and facilities, but as Latour so strongly emphasizes, a particular laboratory (or the field station, or some other concentration of people, instruments, techniques, specimens, and archives) is the indispensible ‘passage point’ for specific research ventures. Although Kuhn sometimes writes as though a paradigm is akin to a theory or worldview, a disciplinary matrix is not simply a conduit for ideas and symbols. It includes practical and communicative skills cultivated through extensive practice with instruments and materials; skills that underpin (and sometimes undermine) the semiotic linkages provided by common vocabularies, themes, citations, and other verbal and literary devices.
The dissemination of ‘Kuhn’ and the Latourization of STS has mainly been achieved through the sale and reproduction of texts, as well as through bibliographic citations and verbal recitations. These semiotic materials circulate through journals, conferences, and university seminars, and occasionally break out into more popular media. The mode of travel permits (indeed, relies upon) drift and slippage in meaning from one node to another, with relatively weak and decentralized policing. Kuhn’s historiographic paradigm and Latour’s own actor-network have become integrated with other agendas, for better and worse. There is no privileged center or material matrix with which to nail down ‘meaning’ and control its use. When I say this, I do not mean to express science envy, but to mark a relative difference between the sites and liberties of interpretation that integrate the networks Kuhn and Latour describe and those that provide the milieu in which their texts continue to live.
Both Kuhn and Latour tried to press beyond the expository method of documented narrative with which they organized their best-known writings. However, their efforts to ‘get technical’ never amounted to very much. In one of his follow-up essays to SSR, Kuhn (1977: 310) mentioned that he and some colleagues were attempting to develop a computer program that would implement his model of acquired similarity relations – an explanation of how individuals in a collectivity learn to apprehend ‘stimuli’ as ‘sensations’ in accordance with a shared cognitive system. However, as he later acknowledged (1977: 17), he did not complete the project. Latour and some of his colleagues have used various computer programs to analyze networks of words, as a way to trace and map associations among semiotic ‘actants’ in texts. Although, unlike Kuhn’s program, their programs have been featured in publications (see, for example, Latour et al., 1992), those publications are far from the most prominent items in the Latourian corpus. More importantly, neither Kuhn’s nor Latour’s efforts to get technical established coherent material practices that aspiring students and followers were obliged to acquire as entry points to STS, or that their critics were obliged to master as prerequisites for producing credible refutations. Instead, students and critics continue to work with Kuhn’s and Latour’s texts and develop interpretations of them in a virtual disciplinary matrix.
Conclusion
Latour and Kuhn both tread a fine line between effecting a broad transformation and attempting to curb the enthusiasms with which others appropriated their texts. Kuhn (1991) famously denounced the Edinburgh School’s Strong Programme by linking it to ‘deconstruction’. This remark seemed to be designed less as a characterization of the research done at the Science Studies Unit in Edinburgh, and more as an attempt to deter his critics in philosophy and history of science from linking his philosophy to relativism and subjectivism. Latour deftly affiliated his theorizing with trends in information science and cultural studies, but also distanced himself from critical social and cultural theory (Latour, 2004). Neither Kuhn nor Latour succeeded in controlling the drift of their key texts through charted and uncharted waters.
In my view, the relatively free diffusion of Kuhn’s and Latour’s writings is not a mark of failure, but is instead a condition of their academic success. The authors sometimes seemed to hanker for something more, however: a historiographic and cognitive foundation for philosophy of science, in Kuhn’s case; a novel ontology with an explanatory technics of its own design, in Latour’s. When SSR and SiA are read with hindsight, it becomes possible to say that the limited conditions of their success are forecast in the texts themselves. Kuhn’s book is often read (and, arguably, allows itself to be read) in a way that conflates paradigms with grand theories, thus reducing the contingent communicative and material production of science to a coherent intellectual history. Latour’s adoption of semiotics as a primary source for his analytical vocabulary eases the job of appropriating his writings for high-flown literary characterizations of ‘science’, ‘materiality’, and ‘modernity’. However, if what I have suggested in this commentary is anything close to being right, both Kuhn and Latour benefited greatly from the very misuses of their key texts from which they distanced themselves.