Dramaturgy and information control

One starting point for this inquiry is the dramaturgical sociology of Erving Goffman, which offers a general perspective on the social world based on the metaphor of the theatre. The term ‘information control’ is drawn from Goffman, who uses this concept in his account of the self, developed initially in The Presentation of Self in Everyday Life and further in Asylums and Stigma.Footnote ⁹ What is distinctive about Goffman's perspective is its focus on the dialectic of revelation and concealment through which social actors – or ‘performers’ – present themselves to their ‘audiences’. In a variety of contexts Goffman attends to the artful staging of social life, examining the theatrical techniques actors use to present their identities and exploring the slippages that often disrupt the show. One can grasp how the dramaturgical perspective operates by considering Goffman's concept of ‘region behaviour’ – activity intended to separate two regions – the front stage (which is publicly presented) and the backstage (which is not). In traditional theatre and in everyday life, performers partition physical space, and structure their speech and bodily comportment, in ways intended to shape what audiences perceive.Footnote ¹⁰ Through the strategic use of words, gestures, clothing and many other devices, people create virtual ‘regions’, attempting to highlight their creditable attributes while hiding stigmatizing ones. In other words, what Peter Galison aptly terms ‘anti-epistemology’ – the art of obscuring knowledge and making it difficult to secure – is not restricted to the national security community but is a mundane feature of interpersonal encounters.Footnote ¹¹

Goffman's point is not simply that the analyst can treat such phenomena using the metaphor of theatre; he also argues that social actors themselves possess a profound theatrical self-consciousness, experiencing themselves as attempting to shape how they appear. Nor is his claim that all performers cynically attempt to mislead their audiences (although some clearly do), for he stresses that delivering sincere performances requires the strategic management of impressions to avoid fostering unintended messages. But no matter how cleverly performers seek to shape the impressions that they foster, their ability to control information is inevitably imperfect, and discrediting information often leaks out. Body language or tone of voice, for example, can convey unintended messages that call into question the authenticity of a performance. At times, audiences tactfully ignore these unintended messages, allowing the inept performer the benefit of the doubt. At other times, they warily seek signs that the performance is misleading, striving to avoid being taken in.

Theatrical considerations are reflected in social organization through the formation of what Goffman calls ‘teams’ – groups of people who work together to stage a performance. In a real theatre, teams of actors, writers, set designers, directors and stage crews work to deliver an effective show.Footnote ¹² Goffman argues that analogous teams are found in many settings. Team membership is defined not by formal organizational structures but by shared engagement in ‘dramaturgical cooperation’ – collective work to manage the impressions that audiences receive in the interest of pulling off a compelling performance.

Goffman's central interest was the study of face-to-face interaction, and his approach can be immediately applied to such scientific encounters as conversations, lectures and professional meetings. But it also can be extended to consider collective modes of information control expressed in the production and selective circulation of texts and other entities that contain knowledge, which also can be regarded as performances prepared with audiences in mind. This approach has been used to address the question of how credibility is performatively constructed by scientific advisory organizations.Footnote ¹³ It is also relevant to the practices of the military and intelligence agencies.Footnote ¹⁴ Such contexts feature ongoing collective management of what to include in – and omit from – texts, and with whom to share them and when. The dramaturgical perspective highlights the construction of enclosed regions where knowledge can be strategically packaged in preparation for its (possibly partial) presentation to wider audiences. It also highlights the gatekeeping activities that precede the release of texts (and other entities that contain information). It highlights slippages and leakage of information from the world of the backstage, calling attention to processes through which regimes of information control break down. And it highlights the construction of contained (sometimes virtual) spaces within which groups, defined in part by their access to concealed knowledge, enjoy privileged epistemic positions and identities.Footnote ¹⁵ Most fundamentally, this perspective frames information control as entailing dynamic interaction and suggests that strategic information management must be ubiquitous in science, found – as it is in social life more generally – in any physical or virtual space where performers and audiences meet.

Containers and materiality

The familiar image of a white-coated figure gazing intently into a test tube metaphorically captures a second aspect of this perspective. To develop an analysis of information control adapted to the laboratory, one needs to extend the concept of ‘information’ to include a wide range of material entities. Thus, in considering this iconic image, my focus is not on the scientist, or the gaze, or the white coat, despite the rich resonances of each. Instead, I want to foreground the lowly test tube as an exemplar of an important dimension of scientific practice: the procedures and machinery used to contain objects of technoscientific significance and choreograph their circulation through specific channels. If the white coat can symbolize the cultural authority of science, then the test tube can be a metaphor for the social and technical machinery intended to enclose and protect entities of technoscientific interest, to regulate their disclosure and package them for shipping, to structure the channels through which they circulate, to effect control over their uses and futures. The metaphor of containment and controlled circulation is meant to highlight the materiality of both the means of containment and the entities – or ‘technoscientific things’ – that are packaged therein. The term ‘technoscientific things’ is used broadly here to refer to entities that relevant actors believe to be of potential epistemic, technological and/or strategic significance in ongoing, emerging or imagined technoscientific activity. Technoscientific things come in many forms, including such entities as biomaterials, datasets, software, texts, speech, visual displays, specialized instrumentation, skilled personnel and assemblages of these items.Footnote ¹⁶

Spaces containing knowledge are formed – and, on occasion, breached – using a variety of means, and practices aimed at regulating the flow of technoscientific things play a central role in science. These practices are enacted using a diversity of machinery, including the relatively tangible (like the snap-cap tubes common in the life sciences) and the relatively virtual (like the password-protected databases and auditing software of a contemporary laboratory information system). They also include a variety of transformations that materials and information undergo during laboratory work. These transformations take place continually within laboratories for a variety of technological and epistemic reasons, including preparing things for circulation by, say, packaging biomaterials, wrapping electronic data in metadata or editing texts to meet length constraints.Footnote ¹⁷ Such transformations not only enable circulation, but also provide opportunities for building varying kinds of constraint and freedom of action into the very things that circulate. For example, during the extraction of genomic DNA from a tissue culture, the cells are destroyed and cellular matter is discarded so that the isolated DNA can subsequently be analysed, free of contaminants. The cells, which no longer exist, cannot be studied or reproduced, so a researcher who receives live cells from a colleague enjoys more freedom of action than one who receives an aliquot of purified DNA. As in the case of the trapdoor functions that underlie public-key cryptography, materials and data in laboratories often pass through trapdoor transformations – one-way operations that render their outputs unusable for particular purposes. These trapdoor transformations are often routine parts of scientific practice, as mundane as performing statistical operations that cannot be reversed to yield the original data. They also include changes introduced for the express purpose of constraining downstream users, such as redacting a document. In either case, trapdoor transformations build encumbrances directly and materially into the things themselves, thus allowing things to circulate while constraining their users.Footnote ¹⁸

Trapdoor transformations are only one mechanism for establishing constraints intended to control future uses of circulating things. Other mechanisms include those that rely on trust (like a colleague's promise) or sanctions (like the threat of criminal penalties) or sealed packages (like embedding a microchip in epoxy). Through such mechanisms, alone or in various combinations, actors can inscribe qualitatively different forms of freedom and constraint into circulating things. In this way, the (in)capacities of technoscientific things and their users are co-produced.

Regimes of closure

The dynamic interactions alluded to above do not take place on an open field, unconstrained by history, identity and institutions. For this reason, we need another concept: the regimes of closure that constitute generic agents of certain types, along with their rights, duties, privileges and concepts of normal courses of action as they pertain to other agents and to control over technoscientific things. These regimes, which often take legal or quasi-legal forms, are embedded in institutionalized discourses and practices (e.g. scientific publication, military or trade secrecy, patent law, privacy and even the epistemic cultures of particular scientific fields).Footnote ¹⁹ By ‘generic agents’ I mean well-institutionalized discursive constructs – such as the ‘author’, the ‘inventor’ or the ‘human research subject’ – that are widely used as framing devices to describe abstract situations or to conceptually package concrete ones (in which specific agents are implicated). Regimes also involve generic spaces – such as ‘the individual's body’, the ‘scientist's laboratory’ or ‘the published literature’ – in which technoscientific things are found. In addition, regimes of closure involve generic technoscientific things, such as ‘scientific papers’ or ‘anonymized DNA samples’.

In many situations, the relevant regime provides some mixture of informal guidance and legal force about the appropriate relations among these agents, spaces and things as they pertain to regulating the distribution of control. No researcher, for example, has the right to pierce the human subject's skin with a needle and extract technoscientifically interesting DNA from her body without permission. The material is neatly contained, at least in the imaginary of bioethics, not only inside her body but also within a regime of human-subject protections and, more broadly, of individual rights. Yet freedom of action for the researcher can be contractually constructed. Using the legal instrumentation of informed consent and contract law, the right to control the future use of the material can be extracted from the subject and transferred to the scientist along with the blood sample. Residual encumbrances, such as a duty to keep the identity of the source confidential, may also be legally constituted, attached to the material and transferred along with the material to future users. As this example suggests, encumbrances not only are sometimes removed but also may be added using legal devices (for example, material transfer agreements, non-disclosure agreements and military classification), allowing targeted, partial or limited circulation. Patents, which require unconstrained disclosure of descriptions of inventions while granting monopolistic encumbrances on their use, are another example.

Some of the regimes operative in the laboratory are specific to particular fields of science and even particular research projects. In academic molecular biology, the professor who is ‘head’ of the laboratory enjoys extensive authority over both its internal operation and its external interactions. Subordinates are typically granted some rights (such as authorship on multi-authored papers) in connection with specific projects deemed to be ‘their’ work, but the laboratory head retains ultimate authority over everything done in ‘his’ or ‘her’ lab. This regime frames expectations about how rights in laboratory products will be allocated as projects develop or personnel join and leave research teams, but disputes occur with some regularity. Rather different patterns are found in other fields, such as high-energy physics.Footnote ²⁰ Quasi-formal regimes pertaining to the disposition of technoscientific things also are sometimes imposed by funding agencies, or used to help constitute multi-laboratory collaborations.Footnote ²¹

As this discussion suggests, manoeuvring about the control of technoscientific things takes place in a context of multiple regimes. To understand their operation, it is important not to focus on any one regime in narrow isolation from those that surround or, more accurately, are entangled with it, for doing so will paint a misleadingly one-dimensional picture.Footnote ²² The regime of scientific publication provides an instructive example. People who consider publication in isolation from other regimes often appear to imagine that it constitutes a world of unfettered openness. And in a sense, it does. The scientific literature steadily accumulating in libraries represents a massive storehouse of information that approaches open-to-any-and-all, free-of-charge, no-questions-asked access.Footnote ²³ Moreover, taking an idea (which economists would call a non-rival good) from the literature does not reduce its availability to others. The duties of those who use such facilities are few, consisting mainly of following proper citation practices and respecting copyrights. Indeed, the image of the open scientific literature as the ultimate repository of scientific knowledge is arguably the root metaphor on which the commonplace understanding of science as ‘naturally open’ is based. But before we stop the analysis with the self-congratulatory conclusion that the open literature is the endpoint where information control stops and wide-open access begins, we might consider what happens to the facts, data, ideas, papers and technoscientific things that a researcher finds in the literature, copies and transfers to his or her laboratory. There, in the enclosed space of the laboratory, a regime of privacy governs: the scientist is under no obligation to share his or her work, but enjoys the privilege of deciding whether to reveal it, when, and to whom. Things taken from the literature and transported into the laboratory enter a private space: the fact of their presence in the laboratory, the uses to which they are put, the new things they are used to create – all these fall under the exclusive control of the laboratory. As this example suggests, analysis of regimes of closure needs to consider how closure, disclosure and re-enclosure – a phenomenon whose many permutations we might gloss as ‘dis(en)closure’ – operates at points where regimes interact.

My point is not that the openness of the literature is a sham. It is not. Indeed, in my opinion, public libraries – for their institutional vision as well as for their contents – rank among the most impressive achievements of human civilization. The point is rather that openness in science is in some ways transient even where it is most complete. For a library to be completely open, the things within it must be completely unencumbered. And for those things to be unencumbered, the users of those things must be allowed freedom of action to use them as they see fit within the enclosed spaces of their private thoughts and actions. Openness is in part constituted by its other.

The dialectic of revelation and concealment

The first encounter illustrates several aspects of the dialectic of revelation and concealment, especially as it is found in face-to-face interaction:

This sort of opened up the topic of advances in electrophoresis, and Clyde started asking what was going on with gel mathematics. He seemed to know that some of the people involved in the Company One/Unnamed University nexus had been studying what happens in this type of sequencing gel. The issue that he considered important to understand is exactly where the molecules are in real time as they are moving down the gel. Not all of the DNA molecules move at the same rate even if they are the same size, because the different fluorescent tags on them slow them down to different degrees. A mathematical understanding of the behaviour of the molecules in the gels is necessary for sequencing machines; the base calling and ordering algorithms must be adjusted to take account of these differences. So Bill and Clyde went back and forth on this for a little bit, with Clyde probing Bill with technical questions. Bill finally said, ‘Yep, the gel mathematics, they are now understood.’ Clyde said, ‘I never saw the reference.’ Bill said, ‘Company One did the study and Company Two has also done the study.’ Clyde said, ‘Oh, that's why.’ No one needed to say what was obvious to everyone present, namely that the companies had decided not to publish papers on the topic, owing to their desire to retain a competitive edge in the race to improve sequencing machines. It was clear that this was all that Bill was going to say, and no one pushed for more.

This ethnographic vignette illustrates the dialectic of revelation and concealment in conversation, as information is concealed, revealed and concealed as it is revealed. The participants appear to self-consciously recognize their efforts to control the flow of information. In particular, Bill seems to be artfully releasing some information (for example that the company has studied and now understands gel mathematics) while withholding other information (such as even the briefest gloss on what was found).Footnote ²⁸ At the same time, Andy and Clyde seem interested in extracting information from Bill. The artful construction of regions – by Bill and (apparently) by both companies – is a salient aspect of the interaction. In addition, everyone seems to accept that for the company to impose a regime of secrecy in this context is well within the normal range of ordinary action.

The conversation also points to the selective enclosure and disclosure of different kinds of information. The unpublished findings of research on ‘gel mathematics’ that the two companies apparently conducted seem to be instances of what the national security community would call ‘objective secrets’, a kind of knowledge deemed to describe stable and universal facts.Footnote ²⁹ The conversation also involves what that community would term ‘subjective secrets’ – information of the ‘loose lips sink ships’ variety that is deemed transient and particular.Footnote ³⁰ For example, Bill's comportment and speech seemed to self-consciously point to a boundary marking off the confidential terrain of information about Company One's activities, although (in what may have been another layer of self-conscious information control) he took no steps to clarify whether he was edging up against that boundary or breaching it while maintaining deniability. It is also possible that to deter competition or encourage exuberance among stock analysts and their technical consultants (Clyde might be such a person), Company One wants relevant people to ‘know’ that it understands the gel mathematics, yet does not want to tell outsiders the details. Bill invited the other participants to know that the Company knows without permitting them to know what the Company knows (or does not know).Footnote ³¹

As this discussion suggests, this conversation, though brief, seems to involve information-control practices pertaining to both objective secrets (about gels) and subjective secrets (about company activities). More deeply, however, the conversation points to ambiguities in the distinction between objective and subjective secrets. Company One's unpublished findings about gel mathematics (which might be imagined to be objective secrets) remain, at least for outsiders not privy to the work, incompletely vetted and consequently of at least questionable epistemic status. Company insiders, on the other hand, might have more extensive grounds for being convinced, but outsiders cannot be sure. Bill seems to believe that the phenomenon is ‘now understood’ but this leaves important questions open: should one share his confidence that the company has achieved the correct (or the deepest or the most useful) understanding? Can one be certain that Bill is not bluffing or overstating his certainty? Should Bill's assertion that Company Two ‘has also done the study’ be taken to indicate that the two companies did ‘the [same] study’ (something Bill did not exactly say) and found the same thing? Can we assume that the two (possibly) secret studies ‘replicated’ each other? Company outsiders who implicitly or explicitly answer ‘no’ to such questions will not perceive the company's knowledge of gel mathematics to be classifiable as an objective secret. By adjusting the level of scepticism, the outsider can slide back and forth from ‘we know that the company knows the objective secrets of gel mathematics but will not disclose them’ to ‘we know only that one scientist in the Unnamed University/Company One nexus disclosed the subjective and possibly secret claim that it has studied and understands gel mathematics but hasn't published what that person says the company knows’. Information-control practices thus not only shape the social distribution of objective and subjective information; they also shape the social distribution of the capacity to be confident about which information can be regarded as subjective versus objective. Far from only being about access, as consequential as that is, these practices also shape the epistemic qualities of knowledge.Footnote ³²

The boundaries of groups

Above, I used the terms ‘insider’ and ‘outsider’ without exploring how the division between insiders and outsiders is constructed. Information control plays an important role in the construction of group boundaries. This comes out clearly in our second vignette:

This conversation underlines the role of ‘sharing’ technoscientific things in efforts to constitute collaborations, which genome researchers usually imagine as transient alliances among individuals and/or research groups. These alliances are typically entered into by laboratory heads whose ‘ownership’ of resources and authority over ‘their’ laboratory underwrites their ability to make commitments. Eileen seeks information about whether she should consider providing some potentially valuable biomaterials to a well-known researcher. Such interlaboratory exchanges of resources are common in molecular biology, since no single laboratory can produce everything it needs. However, as Karin Knorr Cetina notes, cooperation is problematic in the individualistic culture of molecular biology.Footnote ³³ Although some resources are transferred with no strings attached, others are deemed too valuable to provide without a plan for joint publication or some similar shared ‘ownership’ of research results. The terms of such exchanges are incompletely standardized and frequently become points of tension. Eileen's inquiries are a mode of intelligence gathering aimed at learning who can be trusted to participate in (what she considers) ‘fair’ exchanges. The promise of the materials and her control over their disposition gives her much power to decide with whom to explore their potential. Similarly, she enjoys strong control over when she transfers the materials. Finally, she is in a strong position, at least initially, to specify the rights that she and her collaborators will hold in any outputs of their work. (A few months later, I interviewed Eileen privately and she explained her method of selecting collaborators in strategic terms. Her account prominently mentioned reputation in ways that support Steven Shapin's argument about the importance of character in the small, familiar worlds of technoscientific practice.Footnote ³⁴) But let us return to the conversation:

George also started telling stories about how John was trying to appropriate their data and ideas. He provided an example of a particular idea that John had forbidden him to pursue. He said that he had thought of that idea himself, that it was obvious, that there was nothing proprietary about it – all of the necessary materials were publicly available. But John's view, according to George, was somehow that this idea was his and that it would be inappropriate for George to do it. I asked George, ‘What kind of argument did he make to try to justify that?’ He said that John presented the issue as, ‘You used to work for me, and there's a hierarchy. It's up to you to follow instructions. Anything you do at the time you're working for me, I own.’ It was not done in terms of an argument such as, ‘On 12 April we discussed this idea.’ John did not claim that previous discussions had led to the development of that particular idea.

Here the conversation offers a one-sided account of a dispute about whether a particular set of technoscientific things falls under John's authority as a laboratory head. George and Frank contest his claim to jurisdiction on temporal grounds and by asserting that they independently developed the relevant ideas, whereas John reportedly asserts that the ideas were part of what they did while working for him. This ethnographic vignette shows how information control is entangled in practices that shape group boundaries, both during the formation of groups, as illustrated by Eileen's inquiries about potential collaborators, and during their dissolution, as shown by the dispute over who ‘owned’ work undertaken ‘within’ a contested period that included (or, to highlight a point of contention, preceded or followed) the departure of personnel.