Ambivalence,equivocation and the politics of experimental knowledge: A transdisciplinary neuroscience encounter

Introduction

The desire to work between disciplines is an emerging feature of the contemporary academy. The trend towards the dissolution of boundaries between different ways of producing knowledge can be seen in a variety of sites, including novel degree programmes (Worton, 2013), genre-busting research-management strategies (European Science Foundation (ESF), 2013) and centres and objects of research (Hadorn et al., 2008). For scholars within science and technology studies (STS), such hybrid research projects have particular salience. Not only has STS scholarship developed tacit expertise that crosses different disciplinary knowledges (Collins and Evans, 2002), but also the very production and sustenance of disciplined knowledge-objects has itself been a long-standing interest within these literatures (Gieryn, 1983; Star and Griesemer, 1989). Moreover, having been trained to identify the gaps and continuities between others’ intellectual practices, STS scholars are well placed to actually work within and through interdisciplinary research projects (Jasanoff, 2011). Unsurprisingly then, as interdisciplinarity has emerged as a potent rhetoric of contemporary knowledge production and as a way of enacting a hybridized STS knowledge, sustained attention to the collaborative logic of interdisciplinarity has appeared within STS (Gorman, 2010; Maasen, 2000; Rabinow and Bennett, 2012) and cognate literatures (Frodeman et al., 2010; Schmidt, 2008).

This article reflects on the progress of one collaborative and transdisciplinary project, a novel experiment in neuroscientific lie detection, in which the authors of this article, who are from different disciplinary backgrounds – STS, literary studies, sociology, anthropology, clinical psychology and cognitive neuroscience – were entangled. There are still relatively few accounts (Rabinow and Bennett, 2012, discussed just below, is an exception) of what it is actually like, in the most basic sense, to participate in such collaborations, and even fewer analyses of the broader logics of collaboration that root themselves in such experiences, or in the feelings engendered through them. This is potentially significant. The humanities and social science authors of this article, for example, entered the collaboration with ideal-type accounts of the logic of cross-disciplinary labour, based on notions of trade (Galison, 1997), interaction (Collins and Evans, 2002) and integration (Huutoniemi et al., 2010). And yet these accounts did not serve the group at all well when they encountered the more mundane realities of actually producing knowledge within a collaborative and transdisciplinary research space. In preparing this reflective article, which tries to refocus attention away from the ideal and the abstract and towards the mundane and the tangible, we have come to the realization that the mundane pragmatics of this collaboration were not so much guided by logics of collaboration, shared interest, and mutual enrichment; they took place, in fact, within a rather less transparent, rather less unified and rather less propitious sphere of interaction and exchange – a field that was dominated, instead, by some more subterranean feelings of ambivalence, critique, reserve and even dishonesty. This article is thus an attempt to unfold the consequences of a realization that collaboration is sometimes not so much about dialogue or trade, that it can also be a much more ambiguous intertwinement of knowledge, affect and power.

In their book Designing Human Practices, an account of a similarly collaborative effort within the emerging field of synthetic biology, the ethnographer-collaborators Paul Rabinow and Gaymon Bennett (2012) describe their interaction with molecular scientists as ‘a productive experiment’ (p. 1). And yet, as with our collaboration, such productivity does not mean that Rabinow and Bennett’s experimental experience was an unending joy. Indeed, reflecting on an experience that produced some acrimony, the ethnographers go on to explore, at some length, their scientific collaborators’ ‘wide-ranging lack of curiosity outside of their specialties’ and the ways in which agreed understandings were, in practice, ‘rarely addressed seriously and … easily neglected, when it [came] to the inclusions of the human sciences in bioscientific enterprises’ (pp. 4, 7). They describe how their social-science questions were frequently met with ‘perplexity, indifference, and/or hostility’, from their bioscientist colleagues, and they show how they sometimes even experienced ‘active resistance’ to their research priorities (pp. 8, 20). Amid this ‘hierarchy of power and privilege’, they point out, it was taken as natural that Rabinow and Bennett ‘were conversant with the molecular biology and eager to learn more’ – whereas, on the part of the molecular biologists themselves, ‘no reciprocity emerged, nor was it encouraged’ (p. 29). As we will show below (but we will depart markedly from Rabinow and Bennett’s conclusion) this sense of discomfiture resonates with our experience.

In an account of ethnography and collaboration at Xerox Park, and drawing on Barry et al.’s (2008) account of a more ‘agonistic’ interdisciplinarity, Lucy Suchman (2013) asks, ‘what could it mean … to treat [such] resonances and tensions as productive?’ (p. 156). This article provides one answer. Its core argument is that attending to these experiences opens up a more expansive and compelling space for theorizing interdisciplinary work; more specifically, by reflecting on the mundane everyday experience of collaboration, and on the feelings and senses of ressentiment ¹ entangled in it, this article argues that we might more fruitfully explore the awkward forms of ‘experimental politics’ that sometimes animate collaborations like this one. Thus, in the first section, we describe the experiment in question and our reasons and methods for approaching it. In the second, we reflect on what it was like to participate in this experiment, focusing on the presence of three unanticipated collaborative registers in our project: ambivalence, critique and dishonesty. The third and fourth sections argue that these feelings might be interpreted as effects of the epistemological politics of collaboration, and here we attempt a more generalized account of the relationship between politics and experiment in such collaborations. In conclusion, reflecting on the ambiguous, shifting and perhaps intractable nature of such a politics, we do not call for more genial or transparent modes of collaboration but for more attention to equivocation as a way to live with, and to work through, such modes of collaborative exchange.

It should be noted that when we reflect on our experience, the feelings and intentions reflected upon are not always admirable, nor are they flattering to the authors. But our commitment here is not to judge the moral content of our experience. The article’s aim, instead, is to unfold the roles played by some less savoury feelings and motivations in our collaboration and to bring these into understanding within a wider attention to interdisciplinary work. Setting aside the desire for a normative or practice-based account of this experience, we explore how such feelings might actually underpin a collaboration, indeed, one that might otherwise be understood as ‘successful’, or even ‘good’.

The experiment in question

The experiment at the heart of this article came from a ‘NeuroSchool’ on ‘social neuroscience and neuroimaging’, run by the European Neuroscience and Society Network (ENSN). ² Despite its name, the NeuroSchool was not designed as a straightforward pedagogical institution. Its goal was to look for more complex interactions between the neural and social sciences, not only by promoting cross-disciplinary knowledge-sharing, but also by requiring attendees to think with and through the perspective of another science. As such, the programme declared,

Training in the neurosciences is very often ahistorical and asocial … [whereas] students trained in the social studies of neuroscience do not always have a chance to be directly exposed to how rationales and questions in neuroscience experimentation are formulated. (ENSN, 2009)

The goal was not simply to pool neuroscientific and social–scientific expertise, but was to generate a space in which neuroscientists and social scientists might work together to conceptualize a neuroscience embedded in its own culture and history, as well as a social world worked through the structures and functions of the human brain.

The ENSN attempted to generate an infrastructure for scholars to think and collaborate in the space between a burgeoning neuro-discourse and the wide-open field of related ethical and social issues. In an era that did not lack for neuroscientists thinking about the societal consequences of their work (Iacoboni, 2008) or for social scientists keen to talk about the cultural a prioris of neurobiological knowledge (Martin, 2004), the ENSN (2007) was distinguished by its attempt to locate a specifically transdisciplinary space between neuro-biological and sociological interests:

‘transdisciplinarity’, unlike interdisciplinarity, does not simply mean laying two or more disciplines next to each other. Rather, it means to set about a question simultaneously taking into account visions and methods on the same topic from seemingly different perspectives.

Participants at the NeuroSchool not only exchanged ideas on social neuroscience and the role of neuroimaging technology, they also formed teams to collaborate on the design of neuroimaging experiments that were expected to grant intellectual parity to the role of social context. The team that designed the most engaging proposal was invited to carry out its experiment at the Centre for Integrative Neuroscience in Aarhus, Denmark. The design that became the experiment at the heart of this article was the winning proposal. ³

What united many of the participants at the NeuroSchool – and certainly the authors of this article – was an inchoate sense that there was ground to be claimed by pushing at the edges of an emergent ‘social neuroscience’ (e.g. Lieberman, 2006). The transdisciplinary goal of both the workshop and the present authors was to ask whether experimenting with a ‘more social’ social neuroscience could help us, in turn, to reimagine forms of engagement across the biological and social sciences. Participants worked together, as in an interdisciplinary model, but in addition they tried to reimagine (however partially) some of the fundamental questions and assumptions of the ‘other’ disciplines (see Klein (2010) for a careful parsing of forms of inter- and transdisciplinarity). ⁴ Thus, the ostensible and publicly shared normative commitments both of the members of this collaboration and of the NeuroSchool in general were not so much to the critique of neuroscience as such, but to the emergence of a more expansive, transdisciplinary (especially) social neuroscience, one that maintained a commitment to the rigours of experimental practice but that was also mindful of the plurality of disciplinary perspectives on, and approaches to, ‘the social’. In addition to its ostensible goal, therefore, the ‘winning’ proposal would also be an experiment in collaboration itself. How would a truly transdisciplinary engagement between sociological and biological knowledge actually play out in practice? This was a question that animated the NeuroSchool and our collaborative group.

It is worth noting that the methodology of this larger experiment in transdisciplinarity was ad hoc and informal. The groups were formed seemingly only with an eye to disciplinary spread; there were no obvious processes for ongoing assessment or mediation (beyond the competition), nor did there seem to be an accounting for the quality and process of collaboration. We note this not as a critique of the ENSN, but to stress that virtually all of our group’s design energy went into the neuroimaging experiment, rather than to the formation of our own interaction. This meant that we lacked an explicit method or forum for addressing and recounting the kinds of feelings that we describe in this article or for associating them with particular moments or issues in the collaboration. Our collective experience is that informality and ad hoc procedures are more common than not in interdisciplinary assemblages; thus, there was a serendipitous ‘ecological validity’ both to the ENSN’s procedure as well as to the experiment that followed. We note also that this collaboration worked. Despite some apparently unusual features, and the ambivalence that at least partly resulted from them, we produced a finding worthy of submitting to a peer-reviewed journal. We will reflect more on this relationship between the ‘success’ of the experiment and the lack of specificity in our procedures in the ‘Conclusion’ section.

Our group’s experimental design centred on neuroimaging studies of lie detection (Langleben et al., 2005; Wolpe et al., 2010). While attempts to produce a scientific basis of lie detection have a long history, several scholars from within the neurosciences have recently wondered whether the brain-imaging method, fMRI (functional magnetic resonance imaging), might not finally provide a method for reliably locating valid acts of deception within the body and brain of an individual. However, historical and cultural critics of lie detection – including Littlefield (2009, 2011) – had drawn attention to the way in which such an endeavour recapitulated some of the most basic and problematic aspects of previous attempts at scientific lie detection. These included the assumption that ‘truth’ and ‘lie’ are, if not natural categories, robust kinds with solid biological correlates; they also included the tacit suggestion that deception, as an embodied phenomenon, might be understood as a departure from a biological norm, and that truth, therefore, could be positioned as the natural condition of the body at rest. The experiment attempted to transform this fundamentally sociological and historical critique into a meaningful collaboration, and thus to rearticulate a well-established socio-critique through the very methods and assumptions of neuroimaging itself. Could we collaborate across epistemic cultures (Knorr Cetina, 1999) while designing a more reflexive experiment in fMRI lie detection, one that would have regard for the rigour and robustness of neuroscientific experiment, but that would also reflexively integrate social and cultural questions into its basic paradigm?

There is not space here for a formal account (see Littlefield et al., submitted), but the experiment drew on an expansive historical and literary tradition to argue that truth and lie might not be so distinctive as the neuroimaging literature assumes (we alighted on lie detection in particular because it was a literature with which the group, via Littlefield, had some prior expertise). The core goal of the experimental design was to generate an ‘ecological’ situation in which a ‘socially-stressful’ truth would be operationalized. In the final experiment, we recruited 27 participants from choirs in Aarhus, told them we were running a brain-imaging study on team evaluation in the context of a competitive environment and invited them to a day-long series of choir-based team-building activities, culminating in a team singing competition. ⁵ On arrival, participants were randomly split into two teams, which were to individually bond, compete with one another and practise singing together as separate teams, until the final singing competition. During the day, each participant was individually led to a separate room, and asked to record a short video, singing solo to a camera. They were told that when they each came for their subsequent individual fMRI scan, they would evaluate four randomly selected videos. It was stressed that the ‘accuracy’ of their responses would contribute to an overall team score and would be revealed at an unspecified future date when everyone would get together again and all results be revealed.

What the participants did not know, however, was that each team contained two actors, one female and one male, who were working with the experimenters. The actors had two core instructions: (1) to play the role of a very likeable, outgoing member of their team, and (2) to sing very, very badly. Of course, when the ‘real’ participants were taken to record their solo videos, nothing was recorded; only the actors made (deliberately awful) videos. And when participants came for their brain-imaging appointment, we fixed it so that everyone could only select the names of the four actors. The point of all this subterfuge was to get the participants to tell an awkward truth about the terrible performance of a member of their own team when that person was particularly likeable, when they had just spent a day bonding, when this would be made public to the teammate in the future ⁶ – and also when the accuracy of everyone’s judgement was very much at stake. We defined this ‘socially stressful truth’ as a truth elicited from participants in the fMRI scanner under conditions that render the teller unsure, or slightly ambivalent, or particularly attentive, or self-aware. We hypothesized that the cognitive activity involved in the socially stressful truth would belie the use of truth as a baseline condition in deception studies, showing truth-telling as a complex, costly and sometimes awkward activity in its own right. And indeed we found that truth, no less than lying, showed activity in areas associated with mentalizing, empathy, attention, decision-making and so on. Thus, we tried to use humanistic knowledge to expand and problematize this area of neuroscience by using the language and methods of neuroimaging itself.

As novices in this kind of project, we were well versed in ‘ideal-type’ accounts of cross-disciplinary interaction and collaboration (Aboelela et al., 2007; Collins and Evans, 2002; ESF, 2013; Nissani, 1997), as well as long-standing STS theories of boundary-working and boundary-crossing in scientific spaces (Gieryn, 1983; Star and Griesemer, 1989). This is a valuable corpus on how collaboration can or should be done–that is, with transparency, clarity and a productive outcome; it sets out the instrumental justifications for pursuing these forms of collaboration and the gains to be made through them, and it begins to characterize some of the forms of knowledge and action (both tacit and explicit) that might actually allow researchers to talk across borders.

But the truth is that we simply do not recognize our own collaboration in such descriptions. And we are increasingly convinced that these accounts are too conciliatory and too instrumental. In short, they are too distant from our own more contrary experience of working across similar boundaries. Our goal in what follows is to expand these accounts of collaboration and to call for more attention to the intensely ambivalent, transgressive and affective qualities of epistemic boundary-crossing. Our gambit is twofold: (1) that the ressentiment that characterizes much of the experience described here is not unique, or even terribly unusual, despite its low prominence in a literature that tilts towards encouragement, and (2) that if we wish to understand possible registers of collaboration more widely, we need some account of the presence of negative feelings in these spaces, and, perhaps more importantly, an understanding of the relationship of such feelings to experimental outcomes. The remainder of this article is an attempt to work through this initial gambit. In the next section, we reflect on the unspoken tensions and lurking resentments that may haunt the space of some contemporary collaboration, but whose role, up to now, has not been fully elaborated. In the following section, we elaborate on the politics of experiment demanded by this reflection, and we ask what such an attention might tell us about collaboration at the broader scale.

Ambivalence, critique and dishonesty in collaborative practice

The politics of experimental collaboration

It is not our goal to unmask the transdisciplinary experiment. But we are interested in expanding the conversation about what is actually at stake in the daily experience of collaborative labour. Resisting both bureaucratic proclamations of why interdisciplinary is a good thing and technocratic attention to how it should be performed, our attention to the feeling of participation in these spaces brings another perspective into focus. In particular, our account of ambivalence and ressentiment leads us away from a straightforward epistemic or practice-based account of our experience, and towards an interpretation of the politics of experimental practice.

Clearly, a macro-politics of knowledge structures interactions between the neural and psychological sciences, on the one hand, and the humanities and social sciences, on the other. Scholars from the humanities and social sciences are frequently exhorted to seek connection with ‘science’ ⁸ but often with little reference to the political economies that have made humanistic and interpretive interests increasingly unsustainable. ‘[T]he humanities are being driven into defensive positions’, wrote the vice-provost of University College London recently: ‘Despite isolated counter-actions, they experience marginalisation as martyrdom and tend to look inwards rather than outwards to new possibilities, such as recovering their status and influence through interdisciplinary working’ (Worton, 2013). In the background of this advice, of course, is the fact that humanities scholars are well advised to seek collaboration with the sciences because of many states’ radically different financial and rhetorical commitment to the two areas (Browne et al., 2010; cf. Holmwood, 2010; McGettigan, 2013). This politics of knowledge was an implicit part of the design of our experiment, to the extent that, in effect, it recapitulated biologically a point that we already knew historically. This recapitulation was founded precisely on the recognition that a neurobiological claim is more rhetorically and politically potent than a historical one even where the fundamental argument remains the same. It seems inescapable that much collaboration is similarly underwritten by cross-disciplinary differences in institutional power and epistemic prestige.

Equally, in the ENSN’s NeuroSchool, despite clear desires for mutuality, the ‘neural’ was often unconsciously positioned as the thing to be understood, and the ‘social’ as a mildly querulous constraint upon it. Social scientists were introduced to magnetic resonance and the operation of the scanner, but there was no assumption built into the collaboration that social neuroscientists should learn something about science fiction, the machinations of power/knowledge or anything similar. Littlefield, leader and originator of the study, was often asked about the history of lie detection’s experimental designs, and how these might be reimagined in a novel fMRI paradigm. But the flow of the experimental situation left little space for the numerous socio-cultural critiques that she levelled against the pursuit of lie detection through mechanical intervention. When the group collaborated around the pragmatics of the experiment, knowledges and tools got aligned in very specific ways. The white board, for example, was filled with ‘2×2’ factorial designs, ‘x and y’ axes and ‘vectors’ of various kinds; it never contained sociological or humanistic theories. There are, of course, straightforward organizational lessons to be drawn here (the humanists may well have simply disrupted such alignments). But we still draw attention to the fact that even in avowedly transdisciplinary collaborations like ours, some knowledges have to interject and insist on their own usefulness; others have the privilege of taking their universal utility for granted.

Such dynamics of exchange are clearly governed by a larger-scale epistemological politics that renders methodological and conceptual differences between the social, natural and humanistic sciences as a hierarchy of intellectual prestige. Although this politics is well described elsewhere (Benneworth and Jongbloed, 2010; Philpott et al., 2011), it is strangely absent from the more formal and instrumental descriptions of – and sometimes exhortations to – cross-disciplinary collaborations considered above (Collins and Evans, 2002; Huutoniemi et al., 2010). By contrast, the reconsideration of this kind of scholarship at the level of mundane experience, through registers of feeling and in the everyday pragmatics of transdisciplinary labour, brings these contests, and the politics that governs them, inescapably to the fore.

Another politics of experiment is possible ⁹

But there is more to be said about the relationship between experiment and politics here. Experiment is exhausted neither by the ‘nitty-gritty’ of laboratory life nor by its recognition as a ‘very peculiar human practice’; we can also think of experiments as aesthetic practices, as ‘trying out … novel forms of intervention’ (Roepstorff and Frith, 2012: 105). We draw on this analysis of the experiment as a novel aesthetic gesture, in order to reposition the politics of experimental knowledge as it emerged within the ambivalent space of our collaboration. If the politics of our experiment derives from hidden machinations of disciplinary power and prestige in transdisciplinary collaboration, we are also in pursuit of some more generative ‘experimental politics’, understood as an ethic and a method that allows us to resist a straightforward account of disciplinary victimhood and to restate why we think it worth entering these spaces in the first place.

This experiment was not only an attempt at collaboration that proved (sometimes) surprisingly difficult. It was also a commitment to build on the work of other scholars already attempting to muddle through the sometimes troubling and awkward, but nonetheless ultimately productive, work of drawing an anthropological, sociological, literary–historical and neurobiological interest through one another (Roepstorff et al., 2010; Rose and Abi-Rached, 2013; Singh, 2012; Wilson, 2004). On the other side of the difficult relations of power and prestige that structure these kinds of collaborations, this experiment came from a gambit that there might nonetheless be an experimental space worth claiming, one in which novel political and epistemological frontiers would, at least, come into view. By insisting on such an ‘experimental politics’ in our collaboration, we try to grasp some of the ways in which this project was also an attempt to think with the experimental, in both its laboratory and aesthetic senses, as a mode of knowledge, one with particular kinds of constraints, effects and possibilities (Fitzgerald and Callard, submitted). Drawing on feminist STS and allied areas, we root such a politics in a methodological refusal of sharp distinction between the objects that are given to, or the questions that can be asked by, scientific and nonscientific research-practices (Haraway, 1997; Latour, 1999). Precisely through such a refusal, we have pursued a transdisciplinary mode of intervention in which the neurobiological legibility of ‘truth’, for example, is not simply affirmed scientifically or criticized sociologically; instead, this legibility is expanded and complicated through more risky and generous imaginaries of cross-disciplinary connection, a method that works to figure, in this case, the traffic between lies, bodies, feelings, theories, situations and laboratories.

Such an experimental politics relies on a sympathetic view of what neuroimaging either is or might be. In her Psychosomatic, Elizabeth Wilson (2004) reminds us that ‘neurological material is more confident, flexible, resilient, and assertive than many critics have yet acknowledged’ (p. 22). Such material may even prove itself a ‘resource for theoretical endeavour, rather than the dangerous and inert substance against which criticism launches itself’ (Wilson, 2004: 29). By attending, here, to the flows of ambiguity, reserve and even critique in our collaboration, and to their indivisibility from analyses of the experiment as such, we have similarly tried to show how a neuroimaging experiment can be a more pliable, resourceful and self-aware mode of knowledge-production than many of its critics have yet realized (Choudhury et al., 2009; Martin, 2013). Several of us entered this experiment committed to a view that, beyond well-worn critiques of ‘neuro-reductionism’ and ‘neuro-colonialism’ (Cromby et al., 2011; Martin, 2004), the neuroimaging experiment may harbour an untapped potential, one that might even claim the neuroimaging laboratory as a site for forging novel alliances between biological, sociological and humanistic knowledges. We now interpret our awkward, and yet also somehow ‘successful’, experience as a sign of how we have learned to live with such alliances – even where they are difficult, or unhappy – to assist more marginal modes of knowledge as they seek to become both say-able and witness-able.

This pursuit of novelty, the forging of alliance and the commitment to aesthetically experimental technique, is the frame, ultimately, through which we have come to understand the logics of ambivalence and reserve that flowed through our collaboration. Situating the linkage between humanistic and neurological knowledge as neither inevitable nor impossible, this experiment sought a space in which such connection might be actively configured. Our focus on feelings of discomfort and dishonesty illuminates both the way in which such an experimental politics may require a rethinking of the logic of interdisciplinary collaboration and the modes of sometimes temporary and uneasy circulation through which it gets practised. We attempted to mobilize, in the experimental mode, something out of different bits of history, sociology, anthropology and neuroscience. And we are reminded that if such a composition is a difficult and tenuous achievement, it is an achievement all the same (Latour, 2010). Transdisciplinary awkwardness is neither simply a subterranean logic of collaboration (although it is partly that), nor is it only a reason for despair (although it is sometimes that too). But the reservations and ambiguities of our collaboration help us to reimagine what experimental politics makes possible in a collaborative mode. We believe there is scope for other inhabitants, from STS and elsewhere, to draw on, and to share, their own experiences of both ambiguity and possibility in collaborative spaces.

Conclusion: practising equivocal speech

Writing this article brought our sense of ambivalence and reserve into the open. But even here we have not placed everything on the table. This article itself, as an open process, was produced through a hesitant, electronically mediated, iterative procedure in which one author, voicing his or her own reflections and feelings, asks others to comment, respond, add or delete and so on. Such a process, which is not unique to us, is also governed by an ethic of reticent politesse. So we do not claim to have resolved the macro- and micro-politics of knowledge animated by that reticence. In this sense, the solidity of the connection between these reflections, and the deep-rooted thoughts, affects and memories of the authors as such, must remain an open question. Conclusion, here, is not always closure.

There are also other ways to account for our experience, not least to wonder whether our ambivalence is not simply a result of poor organization, and whether this article should be interpreted as a call for more attention to method in the composition of projects like this one. We tackle some of these more organizational issues in an accompanying publication (Littlefield et al., 2014). But our experiment was also successful. We made a transdisciplinary hypothesis that drew on a literary–historical insight in order to both trouble and expand a neurobiological literature; the experiment that we designed produced a positive publishable finding, in line with our hypothesis; we wrote that finding up together and we submitted it to a journal. So we do not want to lament the feelings that accompanied our progress, nor are we minded to recommend ways to avoid our fate. Quite the opposite: our interest in dwelling on these feelings has to do with the fact that, unlike many others, our transdisciplinary experiment more or less worked out. It is in this sense that we have suggested that good collaboration might be less a question of fair and transparent commerce and much more a practice of learning to live with feelings of ambivalence and reserve.

Let us return to Rabinow and Bennett. They should have known, they conclude, that ‘the price to be paid for the power and instrumental mastery of modern science was the abandonment of hermeneutical meaning, general cultural significance, enhanced moral practice, and political or ethical spirituality. We underestimated the existential price to be paid’ (Rabinow and Bennett, 2012: 173). The scales and stakes are rather different between our collaborative endeavours and those of Rabinow and Bennett. In both cases, however, social scientists and natural scientists were trying to do an actual, real, nonfigurative ‘experiment’; collaborators found that one of their most significant results was a disappointing preponderance of disciplinary hierarchy. Rabinow and Bennett (2012) suggest a very particular kind of response; in the final pages of their monograph, they argue that their distinctively collaborative and critical orientation produces insights that need to be ‘put into play in a serious and consequential manner’, lest ‘their salutary effects on the practice of thinking … be deflated or distorted’ (p. 178). The ethnographers urge other would-be collaborators from the human sciences ‘to speak the truth frankly … practicing frank speech in consequential situations actually makes one more capable of seeking the truth’ (2012: 179).

We want to conclude by suggesting a different kind of response. If our collaboration was not a great deal more comfortable than that of Rabinow and Bennett, it was still ‘successful’ – at least to the extent that it ended with a more or less publishable shared result, within the loose remit of the ambitiously transdisciplinary framework it had set up for itself. Our proposal is that this success might have come precisely because we did not speak frankly; we did not seek the truth; we totally failed to acknowledge – let alone discuss – the consequences of our experimental situation. What we did, instead, was to try to work and live within a zone that was just about ambiguous enough to keep everything together – that was sufficiently averse to frank-speaking to keep the worst of the resentments at bay. We suggest that collaborators attend instead to an ethic of ‘equivocal speech’ – a mode that is attentive to the things that are better left unsaid, to the feelings that are as well off not articulated and to the senses of awkwardness and ignorance that probably will not help anything if openly acknowledged. Against Rabinow and Bennett’s solution, and its imagination of a collaborative subjectivity founded on openness and speech, and precisely against the clarity and transparency demanded by frankness, our solution is rooted in the more nuanced equivocations of feeling. What we have tried to describe, here, is the capacity for collaborating scholars, especially from the social sciences and humanities, to feel ambiguity, resentment and subversion, but also, in particular, their willingness to go on feeling them, their experimental desire to keep collaborating through them; and thus the commitment to, if not the enjoyment of, living with them anyway.