Human Studies

, Volume 37, Issue 3, pp 369–392 | Cite as

Multistability and the Agency of Mundane Artifacts: from Speed Bumps to Subway Benches

  • Robert Rosenberger
Theoretical / Philosophical Paper


A central question in philosophical and sociological accounts of technology is how the agency of technologies should be conceived, that is, how to understand their constitutive roles in the actions performed by assemblages of humans and artifacts. To address this question, I build on the suggestion that a helpful perspective can be gained by amalgamating “actor-network theory” and “postphenomenological” accounts. The idea is that only a combined account can confront both the nuances of human experiential relationships with technology on which postphenomenology specializes, and also the chains of interactions between numerous technologies and humans that actor-network theory can address. To perform this amalgamation, however, several technical adjustments to these theories are required. The central change I develop here is to the postphenomenological notion of “multistability,” i.e., the claim that a technology can be used for multiple purposes through different contexts. I expand the postphenomenological framework through the development of a method called “variational cross-examination,” which involves critically contrasting the various stabilities of a multistable technology for the purpose of exploring how a particular stability has come to dominate. As a guiding example, I explore the case of the everyday public bench. The agency of this “mundane artifact,” as actor-network theorist Bruno Latour would call it, cannot be accounted for by either postphenomenology or actor-network theory alone.


Postphenomenology Actor-network theory Multistability Technological agency Technological mediation 


In accounts of technology in the fields of the philosophy of technology and science and technology studies (STS), a key issue is how to understand technological agency. Various theories attempt to come to grips with what role—if any—should be ascribed to technologies themselves within accounts of the larger social collectives of people and objects within which those technologies are used.1 Does the agency of a technology somehow reduce to the choices and actions of its user? Or does a technology instead somehow shape a user’s choices and actions? How should we conceive of the ways that a technology is at once both constructed by a collective of actors and at the same time maintains an influence on that collective? Let us refer to this issue as the “problem of technological agency”.

As a guiding example, consider the public bench. By this I refer simply to those pieces of outdoor furniture that provide a public place for people to take a seat, common to parks, bus stops, and subway platforms. At first, it may be difficult to think of a technology simpler and more seemingly innocent. But on further consideration we see that a bench is more than simply a place to sit for those roaming a park or waiting for public transit; benches contribute to the flow and control of movements of the population. What is interesting about these devices is that, despite their everydayness, they are sites of conflicts over how they should be used and who should be using them.

Things become more complicated when we remember that public benches are sometimes used for a purpose other than that for which they were designed. That is, benches are used by the homeless as a place to sleep. As a long and horizontal surface, they can be used as a makeshift bed. Consideration of these two uses of a single technology—sitting and sleeping—prompts questions regarding the problem of technological agency. What makes possible these specific assorted uses for this device? What social factors are involved? What material factors make possible some options for usage and not others?

There is even more to this simple story. We should also consider a specific and common form of public bench: those designed specifically to prohibit sleeping (e.g., Figs. 1, 2, 3). A common version is a bench fitted with vertical dividers between the places in which people sit. The dividers make it impossible to lay across the bench in the way otherwise possible in their absence. For lack of a better term, let us refer to this kind of bench as a sleep-prevention bench.
Fig. 1

Subway bench, Union Turnpike Station, New York City

Fig. 2

Wall-mounted benches, Laurier Station, Montreal

Fig. 3

Public bench, Ueno Park, Tokyo

With the example of the sleep-prevention bench, further questions are raised regarding the problem of technological agency. In what ways can the different possible uses of a technology be reconfigured? What social and material factors are involved in these reconfigurations?

Two contemporary theories that bear on the problem of technological agency are “postphenomenology” and “actor-network theory” (ANT). Postphenomenlogy refers to a school of thought that builds on ideas in the philosophical tradition of phenomenology, but adds certain interests and commitments. For example, there is a focus on human bodily relations to technology, and there is a commitment to the anti-foundationalism of the tradition of American pragmatism. Work in postphenomenology begins from the framework of concepts developed by Don Ihde, and engages in concrete case studies of particular human-technology relations. ANT is a paradigm within the field of STS, and below I focus on its account of technology as developed by Bruno Latour. ANT can be understood as sociological theory expanded to also include artifacts—rather than only humans—in accounts of collective actions. In this view, a technology is a participant in the chain of people and objects that work together to enact an agenda.

One way to investigate the problem of technological agency is to cross-analyze these two theories. Several thinkers have suggested that postphenomenology and actor-network theory can be usefully combined (Hildebrandt 2007; Smith 2003; Verbeek 2005). It is not difficult to see their motivation. The intuition is that these two perspectives each articulate something about technology that the other misses. Where postphenomenology is good at describing the nuances of the relationships developed between an individual user and a technology, it is suggested that it is not as readily prepared to talk about the effects of these relationships on the larger world. Conversely, actor-network theory is taken to be adept at describing the ways that collections of people and technologies together have effects on the world, but not so proficient at addressing the nuances of particular relationships between individual humans and technologies. That is, in the view of these amalgamators, postphenomenology and actor-network theory can productively supplement one another.

However, in my view, more needs to be done if these theories are to be amalgamated. It will not be enough to only point out why it would be advantageous to ascribe to both of these frameworks. More work is required to integrate the two in a philosophically technical way, addressing points of incompatibility, and advancing these ideas. I make some of these steps below.

The combination of perspectives that I outline below draws out aspects of the nature of technological agency that neither perspective captures alone. Most centrally, I examine and expand on the notion of “multistability,” a key concept in the postphenomenological framework. This refers to the ever-present potential for a technology to be used in multiple ways through multiple contexts. This idea is often wielded by postphenomenologists as a part of arguments against overly-deterministic accounts of human-technology relations. Below, I develop a method for appreciating not only a technology’s multistability, but also for identifying the factors that contribute to users’ inclinations to approach a technology in terms of one particular possibility despite the availability of alternatives. In what I call “variational cross-examination,” the factors under investigation include both a user’s individual intentions and habitual inclinations toward a technology, and also the particular ways that same technology has been materially customized by others for their own purposes.

My suggestion, returned to throughout this paper, is that in order to account for the dynamics of the sleep-prevention bench, and technologies like it, we require the insights from both ANT and postphenomenology.

In what follows, I begin with a review of the ANT account of technology, with focus on the view articulated in a series of articles written by Bruno Latour on the concept of mediation. In the second section, I review relevant features of the postphenomenological perspective. In the third section, I consider what it would mean to amalgamate ANT and postphenomenology. Section four expands the postphenomenological toolkit by developing the method of variational cross-examination. This sets the stage for a final set of reflections on the sleep-prevention bench specifically, and nature of technological agency more generally, when viewed through the dual lenses of phenomenology and social theory.

The Actor-Network Theory of Technology

While first developed as an account of laboratory practices in particular, and more generally as an account of the concrete work involved in establishing a claim as a widely-accepted fact, actor-network theory has also been expanded into a theory of technology. Actor-network accounts of technology investigate the social roles of “mundane artifacts,” i.e., the various technologies encountered in everyday life. Put into the terminology of this conceptual framework, technologies are nonhuman actors enrolled into networks which include both humans and nonhumans. Here, I would like to focus on the particular ANT account of technology developed through a series of articles by Bruno Latour (e.g., Latour 1991; Latour et al. 1992; Latour 1992, 1994, 1995, 1999).

Latour’s (1994: 31) analysis centers on the notion of mediation, which in his account refers to the way that humans and nonhumans work together through networks to enact an agenda, or a program of action, as he calls it. To understand such mediation, it is helpful to consider the experience of a person that runs up against a program enforced by a powerful network of actors. This is the case in Latour’s (1994: 38) iconic example of the speed bump. In this example, he reflects on the experience of a driver who is forced to slow down by the actions of others, such as the police. One option for the police is to station an officer along the roadside at the site where they want the driver to reduce speed. The driver thus faces the network enforcing the police’s program of action at this point, complete with its particular human actor, namely that officer with her or his ability to issue traffic citations to anyone caught violating the speed limit. Another option available to the police is to enroll a nonhuman actor, a speed bump. When the driver encounters the bump, she or he must reduce speed or risk damage to the vehicle.

Latour (1992: 232) uses the term delegation to describe this particular form of mediation, the kind in which duties are assigned to nonhumans. In the above example, the task of enforcing the speed limit is delegated by the police to the speed bump. Accordingly, the act of resisting a particular program requires the development of a competing network of actors, or an antiprogram. Even the program of enforcing a speed limit is itself a response to a competing program of action in which a driver intends to speed. Latour (1992: 247) explains, “If you study a complicated mechanism without seeing that it reinscribes contradictory specifications, you offer a dull description, but every piece of an artifact becomes fascinating when you see that every wheel and crank is the possible answer to an objection. The program of action is in practice the answer to an antiprogram against which the mechanism braces itself”.

According to Latour, agency should be attributed not to individual actors, but to networks overall. Playing off The National Rifle Association’s popular claim that it is people—not guns—that kill, Latour claims instead that, “It is neither people nor guns that kill. Responsibility for action must be shared among the various actants” (Latour 1994: 34). Thus, in this view, to consider the notion of agency in relation to technology, it is important to consider a device’s role in an overall program of action. Latour explains, “every time you want to know what a nonhuman does, simply imagine what other humans or other nonhumans would have to do were this character not present” (1992: 229). In the case of the speed bump, to understand its role we must consider the duties of the human police officer playing the same role. These are the assignments that have been delegated to the bump. Latour notes that in fact the term for speed bump in French means “sleeping police officer”.

The example of the sleep-prevention bench is in some ways offered here as a straightforward echo of Latour’s speed bump account. The typical bus stop bench is a nonhuman actor produced and installed as part of a program of action which includes providing riders a temporary place to sit as they wait for the next bus. When the homeless use these same benches as safely public places to sleep, this can be understood as an antiprogram. The homeless resist the intended usage of the first program, and enroll the nonhuman actor as a contributor to their alternative agenda. From here, just like in the speed bump account, proponents of the original program have the option of enlisting a human actor to enforce the original agenda, for example by employing a police officer to patrol the area around the bench, and to thus regulate the space in which the homeless are permitted to dwell. (Of course cities exercise this first option all the time, establishing anti-loitering laws which can be used to harass and/or arrest homeless people who spend time in public areas. One reason why the homeless often chose to sleep at bus stops in particular is that they are places in which people are permitted to spend long periods of time—ostensibly waiting for the bus.) Alternatively, the task of prohibiting the homeless from making use of the bench for sleeping can be delegated to the sleep-prevention design.2

This ANT interpretation of the sleep-prevention bench captures much of the agential character of this device. But it is not yet the whole story.

Postphenomenology and Human-Technology Relations

An account of human-technology relations which can be productively contrasted and in some ways integrated with ANT is the emerging philosophical perspective called “postphenomenology” (e.g., Ihde 2003; Verbeek 2005; Ihde 2009; Rosenberger 2009; Ihde 2010; Verbeek 2011; Rosenberger 2012; Hasse forthcoming; Friis and Crease, forthcoming; Rosenberger and Verbeek 2014).3 Phenomenology is a tradition of thought which addresses philosophical questions through the description and analysis of human experience. Postphenomenology builds from this tradition, but it stakes out a number of diverging commitments and it focuses on less conventional issues. This includes a commitment to the ontological positions of American pragmatism, and a central concern over the actional and perceptual bodily relationships humans develop with technologies. Postphenomenological work proceeds through the investigation of concrete case studies of technologies, and through the application and analysis of Don Ihde’s body of thought.

Like ANT, postphenomenology also focuses on the notion of technological mediation, but through a different conception of this central term. In this view, mediation refers to the way that technologies are not simply some aspect of the world upon which a user perceives or acts; a technology plays a mediating role, coming between a user and the world, transforming each participant and their relationship to one another in the process. Peter-Paul Verbeek (2005: 130) puts it most strongly, “What humans are and what their world is receive their form by artifactual mediation. Mediation does not simply take place between a subject and an object, but rather coshapes subjectivity and objectivity”. Mediation, in this view, is non-neutral and deeply context dependent. A mediating technology enables certain possibilities for a user, while perhaps also foreclosing others, all of this relative to the particular user, the particular device, and the particular use-context.

One kind of human-technology relationship that Ihde explores is what he calls an embodiment relation. In such a relation, a user integrates the device into her or his bodily encounter with the world; the technology becomes a means through which the user relates to the world in a transformed manner (Ihde 1990: 92). A standard example is a wearable technology such as a hearing aid or a pair of glasses. A user does not pay attention to the hearing aid or the glasses themselves, but instead it is a transformed world that is experienced through the mediation of these kinds of devices. Such technologies are embodied as they are used.

Transformed capacities are not the only result of human-technology relations. In this view, as a user becomes accustomed to a technological relation, her or his awareness of the relation itself may decrease. In the examples of wearable technology, a user who develops an everyday, taken-for-granted, deeply familiar relationship with the technology may in many moments be barely aware of that device at all. Ihde (1990: 75) refers to this phenomenon as transparency, and he diagnoses a general tendency, “I want the transformation that the technology allows, but I want it in such a way that I am basically unaware of its presence”. For example, despite the quite significant changes a technology such as a pair of eyeglasses makes to a user’s entire field of vision, despite the glasses frame within that field, and despite the fact that the device sits directly on the user’s face, the experience of the glasses themselves may fade into the background of awareness as the user perceives the world through the device. Indeed, for the accustomed user, a pair of glasses would be considered problematically distracting if it did not take on an almost complete transparency.

The level of transparency depends on a number of factors, including the user’s individual level of familiarity, accustomedness, expectation, and bodily habituation with regard to the device. In addition to asking what degree of transparency an individual human-technology relation may possess, we can also ask just how deeply, how automatically, and how stubbornly this characterization holds. Calling on a metaphor used throughout phenomenology, I use the term sedimentation to refer to the strength of the habituation that characterizes a particular human-technology relation. That is, any human-technology relation whose features are accompanied by a strong force of habit is one that is deeply sedimented.4 For example, for an accustomed glasses wearer, the pair of glasses themselves not only take on a high degree of transparency, but do so with an unhesitation, resiliency, and ease that reflects the depth of the sedimentation of this relationship.

Postphenomenology’s pragmatic commitment to anti-foundationalism and anti-essentialism is reflected in its conception of technologies as multistable. Multistability refers to a technology’s potential to support multiple relations; a single technology can be understood in multiple ways, taken up in many contexts, and employed for various purposes. For example, a hammer can be used for purposes other than its common usage. Ihde (1993: 37) says, the hammer “may be used in a number of ways. It could, and perhaps is dominantly used, for its designed purpose—to hammer. But it could be used as a paperweight, an objet d’art, a murder weapon, a pendulum weight, a door handle, etc. This ambiguity of uses, however, is not indefinitely extendable”.5 The notion of multistability thus simultaneously highlights two points: (1) multiple relations to a technology are always possible, and (2) this potential is at the same time limited by the technology’s materiality, i.e., the particularities if its physical composition. In this terminology, an individual stable relation to a technology is known as a stability and also as a variation (these two terms are deployed interchangeably in this literature.)6

A key methodology in postphenomenology is what Ihde calls variational analysis, the creative brainstorming of possible stabilities (or variations) for a given technology. As Ihde (2009: 12) puts it, through the variational analysis of many cases over the years, “what emerged or ‘showed itself’ was the complicated structure of multistability”. In the example of the hammer that could also be a murder weapon or an objet d’art, what is revealed through this creative brainstorming is the device’s status as itself something capable of participating in multiple stable human-technology relations. That is, according to Ihde, variational analysis reveals a technology’s status as multistable. It reveals a technology’s form to be essentially ambiguous, though ambiguous in way that is at the same time only finitely extendable; a technology can always be used for multiple purposes, but not simply any purpose.7

The method of variational analysis has most often been employed by Ihde and others to challenge overly-deterministic accounts by revealing the unacknowledged multistability of the technologies at issue. An example is Ihde’s critique of Heidegger’s dismissive comments about the typewriter. In Parmenides, Heidegger (1982) reflects on the way that language flows through one’s hand and through one’s pen in the act of handwriting. A situation which worried Heidegger was the increasing popularity of the typewriter. He claimed that, in contrast to the pen, the typewriter stifles the individual character of one’s writing through its mechanization. Ihde (2010: ch. 5) critiques Heidegger’s assessment on two fronts. First, he claims that Heidegger fails to recognize the way individual differences in familiarity with a writing implement inform writing style; for users deeply acquainted with typing, writing flows similarly to how it can with the pen. Next, Ihde notes that the printed word itself can be seen to be situated differently in the different histories in which it had developed across the globe. He says, “The point is that the same technology—printing—can be and is differently embedded and has multiple ‘histories’” (2010: 127). Like all technologies, the typewriter is encountered in terms of both the user’s individual experience and her or his cultural context.8

The public bench is an example of a multistable technology. The dominant stability for this technology is that usage for which it was designed and manufactured, i.e., sitting. And it would be possible to elaborate on the kind of sedimented and transparent embodiment relation an accustomed user will develop with a particular bench, say, one that the user encounters each morning while waiting for the bus. While we could brainstorm a multiplicity of alternative stable relations possible for this device, we have observed that an important alternative is actually often taken up in practice; it is possible to use a bench for sleeping, and we have observed the homeless to at times use benches as makeshift beds. This alternative usage includes a very different bodily relationship to the same device. In this way, a bench is a multistable technology with both bench-as-seat and bench-as-bed stabilities. The sleep-prevention bench can then be conceived as a multistable technology which has been redesigned to shut out one of its potential stabilities.

The perspectives of ANT and postphenomenology each reveal different aspects of the sleep-prevention bench technology. This raises several questions. Is it possible to patch together the insights of each into a combined account? If so, then what is further revealed about the bench technology in particular, and about technological agency in general? And what is revealed about the postphenomenological and ANT perspectives themselves?

Amalgamating ANT and Postphenomenology

Those who argue for amalgamating postphenomenology and ANT point out the ways each perspective can importantly supplement the other (Smith 2003; Verbeek 2005; Hildebrandt 2007). Postphenomenologists generally have long claimed that the reach of their descriptions extends to more than only accounts of individual experience. Nevertheless, they also recognize the ability of ANT to address the chains of interactions between various humans and technologies that extend beyond the transformations involved within an individual human-technology relation. These authors simultaneously note that ANT lacks nuance in its accounts with regard to the exact juncture within networks upon which postphenomenology specializes: individual human-technology relationships. As Aaron Smith (2003: 189) puts it, “Latour’s view, however, does not develop in nearly the same depth the direct personal relationships with artifacts that Ihde’s does. Instead, Latour’s project could be seen as picking up where Ihde’s ends because it emphasizes systems of relations”.9 Put another way, one place that the postphenomenological perspective advances a much more sophisticated account is at the point of interface between humans and technologies.

I agree with these amalgamators that benefits would come from the combination of the postphenomenological and ANT perspectives. But combining these perspectives requires more than an enumeration of the advantages that would result. Both postphenomenology and ANT require technical adjustments and expansions if these perspectives are to be integrated coherently.

For example, these two perspectives are not necessarily on the same page with regard to the issue of technological “symmetry”. It is no simple oversight that leaves Latour open to the criticism from postphenomenologists that his account of human-technology interface lacks nuance; it is the result of his active positioning. Within debates in STS, some have argued against Latour that any inclusion of nonhumans into the overall agency of collectives introduces an objectionable technological determinism.10 Latour instead of course argues that nonhumans are contributors to the agency of networks, going so far as to understand the contributions of humans and nonhumans to be symmetrical. He says, “The degree of attachment of an actant to a program of action varies from version to version. The terms ‘actant’ and ‘degree of attachment’ are symmetrical—that is, they apply indifferently to humans and nonhumans” (1991: 108). To approach networks in terms of the Latourian conception of symmetry, neither humans nor nonhumans should be regarded as the primary movers of the network; neither simply falls under the agency of the other. Latour writes, “In the symmetry between humans and nonhumans, I keep constant the series of competences, of properties, that agents are able to swap by overlapping with one another” (1999: 182). This is the case, for example, in the speedbump account in which the properties of the officer and the artifact are cast in transposable terms. This “swapping” reflects a kind of interchangeability of the actors that form a network, neither humans nor nonhumans possessing characteristics that could not be swapped out.11

Latour’s conception of the symmetry between humans and nonhumans is reflected even in the rhetorical connotations of his conceptual framework. For example, the work of nonhumans is cast in terms normally reserved for the description of human interactions. For instance, as we have seen, the notion of “delegation,” normally used to refer to the assignment of a task to one human by another, is repurposed by Latour for use in describing the enrollment of nonhumans into networks. Conversely, human roles within networks are labeled in terms typically reserved for the description of the interactions between objects. For example, as a user learns the interface of a device, she or he is “incorporated” into the network (Latour 1992: 231).

An adjustment to Latour’s conception of symmetry is required for the amalgamation of ANT and postphenomenology. I suggest that what is needed is an explicit recognition that humans and technologies may play structurally different roles within networks, and that such a recognition does not require abandoning a symmetrical approach. This recognition leaves in place the Latourian conception of relations between humans and nonhumans as one in which the agency of actors does not reduce from one to the other. However, Latour’s conception of humans and nonhumans as always possessing entirely interchangeable properties should be called into question; indeed this conception must be challenged if the notion of “human-technology relations” is to have any meaning that does not reduce simply to “actor–actor” relations.

I suggest that it is possible for an analysis to at once remain symmetrical, and to also remain open to the possibility that the work required to enroll a human into a network can be at times different from that required for the enrollment of a nonhuman. Recognition of the distinctions between human and nonhuman contributions to the overall agency of a network need not result in the reduction of responsibility to one or the other.12 Under this modified conception of symmetry, we can begin to amalgamate ANT’s accounts of chains of actors with postphenomenology’s accounts of human-technology relations. It becomes possible to consider the roles of user experience, and the role of technological multistability, within the overall agency possessed by networks of actors.13

Variational Cross-Examination

The amalgamation of ANT and postphenomenology requires more than only the adjustment of Latourian commitments. It requires an expansion of postphenomenological methodology. Here I would like to focus on the notions of multistability and variational analysis. I suggest that for both the project of integrating ANT and postphenomenology, and that of exploring the problem of technological agency, variational analysis must be conceived as only a first step within a larger methodological framework.

Variational analysis is most often cast as a method for demonstrating a technology’s multistability. By creatively brainstorming alternative possible meanings and uses for a technology, its status as multistable is revealed. However, I contend that if one’s objective is to do anything more than level a counterpoint against an overreaching account, then more is required of postphenomenology. It is possible to expand postphenomenological methodology so that it can also be used constructively rather than only polemically.14 To do so, multistablility must become a starting point, rather than an endpoint, of analysis.

I suggest that variational analysis ought to be conceived as a first step of a larger two-part approach. When beginning a postphenomeonological analysis of a human-technology relation, the first move should be to conduct variational analysis, establishing the technology’s multistability through the identification of concrete alternative stabilities.

Next, I suggest that postphenomenological analysis should continue with a second step: the critical contrast of the particular stabilities that have been identified. The features of these stabilities can be highlighted against one another. This process may reveal new information about these stabilities.15 As a sister method to variational analysis, I refer to this second step as variational cross-examination.

While it may be possible to use variational cross-examination to investigate any of the particular stabilities that have been identified for a technology, this method is especially useful for scrutinizing what could be called the dominant stability. This refers to the stability through which the technology is typically used (and most likely—though of course not always—the one for which it was designed or manufactured). Returning to Ihde’s (1993: 37) quote above regarding the hammer, he suggests that the device is “perhaps dominantly used, for its designed purpose—to hammer”. For the hammer, the dominant stability is reflected in its name; it is used to hammer nails into some other object. We have seen that Ihde performs variational analysis, identifying those alternative stabilities reviewed above (murder weapon, objet d’art, paperweight, pendulum weight, and door handle), establishing that the hammer is indeed multistable. To continue to the next step of the two-step method I propose, that is, to conduct variational cross-examination upon the example of the hammer, we should next contrast the features of these alternative stabilities with those of the dominant stability.

Thus far I have referred only abstractly to the “features” of stabilities that can be contrasted with one another. To articulate some of the features I have in mind, we must revisit a few of the central concepts of postphenomenology and ANT, including embodiment, sedimentation, delegation, and programs of action. Though the following list may not be exhaustive, I consider three general categories of features that characterize various stabilities: (1) comportments and habits, (2) role within a program, and (3) material tailoring.

Comportment and Habits

One category of features that can be cross-examined is the particular set of bodily comportments and habits involved in each stable relation to a technology. That is, in order to relate to a technology in terms of a particular stability, one must take up a particular bodily and perceptual approach, what could be called a relational strategy. 16 For example, to use a hammer in terms of its dominant stability, a user comports her or his body in a particular way with respect to the device, e.g., holding the handle rather than the head, swinging it in a particular manner. And more, if a user grows accustomed to using the hammer in a particular way, the device itself can take on a high degree of transparency; the user focuses more on the work being done with the hammer than on the device itself. This relationship can also become deeply sedimented. That is, a user can come to approach the technology with a level of automaticity, the device instantly appearing as part of a particular context, with certain aspects strongly and immediately taking on their usual transparency. It may be productive at times to contrast the comportments and habits associated with the dominant stability with those of alternatives.

The hammer-as-murder-weapon, for example, is an alternative stability that would involve a different relational strategy to the one associated with hammering nails. That is, using this device as a weapon would call upon a different set of bodily comportments—e.g., it may be wielded with a different swing—than those of typical nail hammering. If one were to become accustomed to using the hammer as a weapon, it is possible that the hammer could again take on a kind of transparency but this time in the bodily comportmenal terms of this alternative stability. (Of course the context in which this transparency becomes manifest for the weapon brandisher would be different from that of the carpenter!)

Ihde identifies the hammer-as-pendulum as another alternative stability. We can ask again what comportments would be involved in, say, threading a hammer-shaped pendulum weight onto a chain dangling from a mechanical clock. Or we can ask what sorts of perceptual preconceptions go into interpreting a grandfather clock, allowing, for example, the swinging pendulum to fade into the background as one focuses upon interpreting the clockface display or the hourly auditory chimes.

By contrasting what relational strategies are required for engaging the alternative stabilities possible for the hammer, aspects of typical hammering may be productively put into relief. It is precisely because the dominant stable relation is steeped in habit and transparency that aspects of this relation may be otherwise difficult to identify.

Role Within a Program

A second category of features upon which stabilities can be contrasted is the roles a technology could potentially play in various networks of associated actors. That is, one way to draw out the associations a technology maintains with other actors is to consider what contributions a technology could potentially make within alternative programs of action. For example, the alternative stabilities of the hammer-as-paperweight and the hammer-as-artwork each call forth sets of relations to other actors. The actors to which the hammer-as-artwork may relate (e.g., museums, curators) are different from those of the network of the dominant stability (e.g., nails, carpenters). The space of usage into which the hammer-as-paperweight fits would be the workdesk of papers rather than the workbench of tools.

As in ANT, in variational cross-examination the technology is understood as a nonhuman that has been enrolled into a program of actors. Also as in ANT, this nonhuman’s role in the program can be identified by considering what other nonhumans and humans would do in its place. What variational cross-examination adds, however, is consideration of how that same nonhuman—as multistable—could potentially play different roles within alternative programs. By considering the potential ways that a technology could fit in a stable manner into alternative networks of actors, we may gain further perspective on what has been involved in its actual enrollment into the dominant network. It can be difficult to identify in retrospect just what work had been done to translate an individual nonhuman into a substantial team player precisely because networks of actors can be marshaled together through the exertion of considerable power. It is on this point that variational cross-examination can be useful.17

Concrete Tailoring

A third category of features upon which stabilities may be contrasted regards the particular ways a technology may be physically altered in the process of making it useful toward a specific purpose. I use the term concrete tailoring to refer to the ways a technology has been converted specifically for use through a particular stability. For example, Ihde suggests the possibility of a hammer-as-door-handle stability. What physical changes would a typical hammer need to undergo in order for it to be fit into this alternative use-context? By considering the modifications that would be necessary for making a technology fit into the context of an alternative stability, we may discover things about the way that technology has already been customized in the process of making it fit into the context of the dominant stability.

A bump built into the road may effectively perform the officer’s duty of enforcing the speed limit, but that bump can itself be modified to better perform this duty. It can be painted with bright colors or stripes to better catch a driver’s eye. A sign reading “bump” can be placed nearby. These material adjustments concretely tailor the speed bump for more optimal fit into the specific program of the police, and into the larger governmental and economic system.18 Subjecting this modified speed bump to variational cross-examination can be useful for drawing out those concrete tailorings just listed. Even if we limit our thoughts here to other kinds of long bumps that could be built into a roadway, alternative stabilities can be brainstormed. A long bump in the road could be used to provide a roadside edge (i.e., a curb); to direct flows of rainwater into drains; or to provide a starting block for drag racing. In these cases, such bumps would be concretely tailored differently. We can ask whether these alternative stabilities would also call for those material tailorings associated with the dominant stability, such as the stripes or the sign reading “bump”. For a driver who approaches the speed bump tailored for the dominant stability, these stripes and signs may not be conspicuous. That is, for the accustomed driver, these tailorings would simply be encountered as already transparent and sedimented, set as “normal” and expected within the power of the culture, driving norms, and the state. By considering what sorts of tailorings may be appropriate for alternate possible stabilities, light may be productively shed on the particular ways that the dominant stability has been actually concretely tailored.19

It is important to note the deeply situated character of any given postphenomenological analysis, a character made explicit by the addition of variational cross-examination to the methodological toolkit of this perspective. The notion of a “dominant” stability, for example, highlights the situatedness of the person performing the analysis. Recognition of the primary way a technology is taken up by a community implies a subject doing the recognizing, and implies some relation to that community. Is she or he a member of the community within which the technology is used in its dominant stability? If not, then from whom has she or he learned about these practices? Is this investigator’s vantage point one of privilege, institutionalized recognition, marginalization, etc.? This situatedness is highlighted by the perspectival nature of the two-step postphenomenological method I advocate. If one stops after the first step, satisfied to only demonstrate that a technology is multistable, it may not matter which particular stabilities have been identified, only that more than one has been identified. But by continuing to the second step, the position of the person performing the postphenomenological analysis—including what alternative stabilities this person does and does not consider—has a bearing on what may be learned. For example, the cross-examination of even more stabilities, if they can be identified, could provide the opportunity to learn more.

A related point to note is that the notion of “concrete tailoring” may at first appear to introduce a problem for variational cross-examination: if we are not only considering the ways that a technology-as-is can be fit into multiple contexts, but are also considering altered versions of that technology, then it may not be immediately clear just how far such alterations should extend in our analysis. How extreme must a material change be for an alternative to no longer count as the same technology? Who makes this decision? My response is that for one to see this to be a problem peculiar to variational cross-examination, one must already be committed to the idea of an independent, non-embodied, non-situated subject, and also to the idea of an independent and innocent device. Instead, in this view, the performer of postphenomenological analysis is understood to be embodied and thus saddled with a socially and historically situated perspective from which things appear with meaning and relevance. It is not assumed that one’s viewpoint is automatically detached and disinterested—but instead just the opposite; it is embedded within the world, perspectival, fallible, limited, and susceptible to bias. Any particular postphenomenological analysis should keep this fact explicit and should remain open to scrutiny from other perspectives. The point of variational cross-examination, after all, is not for a disembodied subject to apprehend the abstract essence of an object; it is to understand what has gone into the tailoring of an actual device for a specific purpose, whose purpose that may be, and how automatically meaningful that device appears within the experience of an individual user.20

Cross-Examining the Public Bench

Let us return again to the example of the sleep-prevention bench. I suggest that it takes at least the combined insights of ANT and postphenomenology to account for the details of this mundane technology, and that the two-step postphenomenological method I have articulated above is well-suited for the task. Many features of the sleep-prevention bench can be made explicit through variational cross-examination, that is, through the critical contrast of the bench-as-seat and bench-as-bed stabilities.

First, consider the bench in terms of the bodily comportments and habits associated with the two stabilities identified. For a person accustomed to sitting on a particular bench in an everyday sense (say, waiting for the bus each morning), habits of conception and bodily approach develop. The details of the relationship an accustomed user shares with the typical (i.e., non-sleep-prevention) bench in its dominant stability can be drawn out by contrasting that stability with the bench-as-bed alternative. That is, for one that sits on a bench each day, the details of her or his “sitting” relationship with that bench can be drawn out by comparing it with what it would be like if this person were to instead lay across that same bench. These details include the particular manner in which one comports one’s body with respect to use in terms of the dominant stability, just how deeply transparent this relationship is, and just how stubbornly sedimented these aspects of one’s approach may be. If that same user were asked to consider lying down across the bench, she or he may not engage this alternative stability nearly as automatically or transparently, nor would the same aspects of the bodily interaction with the device become transparent were she or he to eventually become accustomed to this alternative usage.

In the case of the sleep-prevention bench, an everyday user of the dominant sitting stability typically will not have already performed any kind of varational analysis. This person will not be thinking about the bench as something that can be both sat upon and lain across. Thus, insofar as the bench is approached in terms of its dominant stability, she or he would most likely not even notice the vertical dividers since they do not impede usage in terms of the bench-as-seat stability. (However, of course we can imagine some instances in which a user could become aware, and of course we can also imagine some bench-as-seat users for which the dividers would present an impediment). The addition of the dividers—or other such sleep-preventing modification—both leaves the user free to sit as she or he would, and also at the same time renders the user unable to take up the device as a bed. And this modification remains hidden within the unnoticed background aspects of an everyday user’s seating experience, concealed by habitual routine.

Second, consider the sleep-prevention bench in terms of its role in the larger program of action into which it has been enlisted. As described above, the material configuration of the sleep-prevention bench provides clues about the particular agenda toward which it has been enrolled as an actor. Without the dividers, a bench can be made to fit into either of the purposes under discussion here, sitting and sleeping. With the dividers, however, of course a more nuanced agenda is revealed. The program to which the sleep-prevention bench belongs includes not only the goal of providing a place for users to sit; this program simultaneously includes the goal of discouraging homeless people from spending time in the area around the bench by ruling out the bench’s use as a bed. In this way, to the sleep-prevention bench has been delegated the task of controlling the whereabouts of the homeless population.

There is a complexity to the bench’s enrollment into the program of influencing the homeless that ANT cannot fully address alone. The dispute over the sleep-prevention bench does not reduce to a clash between two networks. To understand what role the sleep-prevention bench plays in the program of allowing sitting and prohibiting sleeping, it takes more than identifying what other humans or nonhumans would do in the bench’s place, as Latour would suggest. The story of the sleep-prevention bench includes more than merely an account of interchangeable actors (e.g., conceiving of the sleep-prevention bench as swappable with, say, the typical bench plus a police officer shooing away those looking for a safely public place to sleep). This story additionally requires recognition that this particular nonhuman, the bench, is a multistable technology open to a particular limited variety of uses. It is precisely because the typical bench can be approached in terms of a particular variety of stable uses that the issue of prohibiting certain uses arises in the first place. Sleep-prevention benches were never objects introduced into areas purely for the purpose of manipulating the homeless in the same way a speed bump is introduced into an area for the purpose of manipulating speeders. The bench-as-bed stability is instead an alternate usage that happens to be possible for a typically shaped bench. The social dispute is a side effect that first arises from the introduction of a bench into an area, one made possible by the fact that the bench is multistable and that bench-as-bed is one of its stabilities, and also of course by the fact that the need of the homeless for a safely public place to sleep exists already independently of any benches.

This brings us to the third feature to consider: exactly how the device has been concretely tailored to fit into a program. In the case of the sleep-prevention bench, of course the particular tailoring at issue is the addition of the vertical dividers or other such sleep-prohibiting modification. Unlike the example of a speed bump which may be concretely tailored with painted stripes that facilitate a better performance of its dominant speed-limiting duties, the sleep-prevention bench has been concretely tailored to foreclose an alternate stable usage. The vertical slats do not encourage sitting the way stripes further encourage speed reduction; the sleep-prevention modifications are a countermeasure against an alternative usage introduced by the multistability of typical bench design. Against the interests of a homeless population that would use the bench in an alternative manner, the sleep-prevention bench is tailored to restrict how it may be used. Explicitly contrasting the dominant bench-as-seat stability of the sleep-prevention bench against the bench-as-bed stability draws out the specific ways that the device has been concretely tailored for the purpose of excluding that specific alternative usage. Because the dominant bench-as-seat stability is not interrupted by this tailoring, and because the homeless population is disenfranchised, it could be easy for the functioning of this tailoring to go unnoticed by the bench-as-seat users.

Discussion: Zeroing in on Technological Agency

Above, I have attempted to amalgamate aspects of postphenomenology and actor-network theory around the notion of multistability. I have primarily performed this amalgamation through the development of “variational cross-examination,” a method for investigating technology that integrates insights from both perspectives. This procedure involves first brainstorming possible stabilities for the target technology, and then contrasting their features, including their relations to users’ bodies, their places within larger social assemblages, and their concrete material configurations. A notable quality of this method is that, in tune with the philosophical commitments of both actor-network theory and postphenomenology, its results remain nonfoundational; that is, they do not purport to reveal the essential nature of the technology. Instead the information yielded is about the relations themselves, situated and context specific, revealed through the critical comparison of those very relations.

As a final set of thoughts, I would like to consider the implications of this expanded version of postphenomenology, amalgamated with actor-network theory, for the problem of technological agency. As defined negatively above, the problem of technological agency refers to the task of conceptualizing technology’s agential role in the actions of human-technology assemblages in a way that reduces to neither that of users and social collectives, nor that of the artifacts themselves. While the amalgamated perspective I have developed here does not, in my view, provide a simple or complete solution to the problem, it does approach it in a way that neither postphenomenology nor actor-network theory have so far even attempted. That is, the amalgamated view outlined above realigns these theories such that they can at least begin to zero in upon a solution to the problem of technological agency.

Actor-network theory and postphenomenology both employ the same maneuver to displace the problem of technological agency, at least as I have defined it here. This maneuver, as we have seen above specifically in the works of Ihde, Latour, and Verbeek, is to maintain that the notion of agency should only ever be attributed to collections of humans and technologies together. In the case of Latour’s version of actor-network theory, the notion of agency should only be attributed to networks of associated actors, human and nonhuman, rather than to any individual human or nonhuman. According to the versions of postphenomenology advanced by Ihde and Verbeek, the notion of agency should only be attributed to human-technology relations, rather than to an individual user or to an individual technology. Even granting these maneuvers, I suggest that the problem of technological agency remains: what is the specific role played by a technology within the overall agency of the assemblage?

Technologies often sit at the intersection of user experience and larger social programs of usage. I suggest that neither side of this intersectionneither the technology’s role in user experience, nor its role in a larger social programcan be fully conceived without an understanding of its relation to the other side. This is clear in the example of the sleep-prevention benches. The story of how and why the sleep-prevention measures are added to a typical bench requires addressing several factors in relation to one another, including the relation between a user’s bodily habits and the different stabilities of the bench, the relation between users’ intentions and the sleep-prevention tailoring, and the relation between those stabilities and the agendas of different larger communities of actors. This point can be explicated by highlighting what each theory, postphenomenology and actor-network theory, independently overlooks even in its own terms.

Without actor-network theory, postphenomenology remains unable to account for the actions of others on the technology with which a user shares a relation. While postphenomenology may specialize in describing the relationships that can occur between a user and a technology, these relationships do not occur within a social vacuum. A larger community of actors may tailor that technology for their own purposes. These tailorings influence what kinds of relationships are possible between an individual user and a technology. That is, a postphenomenological account of an individual human-technology relation is incomplete—even in its own terms—without an understanding of how that technology has been shaped by larger communities of actors. As Latour (1994: 52) puts it, “Technical action is a form of delegation that allows us to mobilize, during interactions, moves made elsewhere, earlier, by other actants”. In the case of the sleep-prevention bench, the postphenomenological perspective by itself remains unable to account for the larger network of actors that provide the sleep-prevention modifications—modifications which refigure what kinds of stable relationships are possible between a user and the device.

Without postphenomenology, actor-network theory remains unable to account for the effects of both individual user experience, and also the multistability of technology, on the overall agency of a network. That is, (1) the particular way an individual human actor conceives of a technology (and embodies those conceptions through habituation), is relevant to the agency of a larger network. Put broadly, changes to human consciousness and embodiment are significant to the agency of chains of actors. And, (2) the multistability of a technology, i.e., exactly which uses are stable within user experience and which are not, is also relevant to the overall agency of a network. Again put broadly, the fact that a technology, as a material actor, makes certain things possible and not others for human users (and that customizing that technology can refigure these possibilities) is significant to the agency of a chain of actors.

The example of the sleep-prevention bench draws out these limitations. The actor-network perspective remains unable to account for the role of the bench’s multistability in its evolving history. That is, ANT cannot alone account for the ways that alternative specific possibilities for action emerge from one group’s design of a technology for a particular purpose. It is not only the case that two groups are in disagreement how a technology should be used. It is that the technology itself makes possible multiple-though-limited relationships with human bodies, and that social disagreements emerge because of specific possibilities enabled by that multistability. From there, the more powerful community addresses this phenomenon with a new material design that makes one bodily relation impossible while retaining the possibility of their preferred bodily relation.

The point is that a technology’s role in the overall agency attributed to an assemblage does not reduce to either individual bodily relations or to group dynamics, and that it is not possible to provide a complete account for either factor alone without engagement with the other. The development of a more complete conception of technology usage turns out to be more complex than even the richness of actor-network theory and postphenomenology imply. Any account of technological agency must contend with a technology’s status as a distinct multistable material participant that remains continuously implicated within both individual user experience and the agendas of larger communities.


  1. 1.

    For example, a foundational disagreement between the Sociology of Scientific Knowledge (SSK) perspective and actor-network theory (ANT) is over whether technologies should be considered “actors” at all (e.g., the “chicken debate” between Bruno Latour, Harry Collins, Steven Yearley, and others in Pickering 1994). Harry M. Collins’ 1994 review of Bruno Latour’s (1994). We Have Never Been Modern is emblematic of accounts which dismiss a conception of actors that includes anything more than humans. Collins writes (1994: 674), “The battle is between society and the individual, not humans and things. To invite inanimate objects into the debate adds nothing but confusion because inanimate objects cannot be social except when the term is used in such a recondite way as to avoid the crucial issues”. A sampling of works across the spectrum that wrestle with the issues of technological agency include (Akrich 1992; Latour 1992; Callon and Law 1995; Pickering 1995; Haraway 1997; Verbeek and Kockelkoren 1998; Latour 1999; Oudshoorn and Pinch 2003; Barad 2003; Verbeek 2005; Ihde 2009; Pinch 2010).

  2. 2.

    For more on the physical changes that technologies can undergo as they are enrolled into networks, see Madeleine Akrich’s (1992) notion of “inscription” specifically, and her script theory account of technology more generally.

  3. 3.

    See also the 2008, 31(1) issue of Human Studies; the 2011, 16(2–3) issue of Foundations of Science; and the book series “Postphenomenology and the Philosophy of Technology” with Rowman Littlefield Press/Lexington Books.

  4. 4.

    According to Maurice Merleau-Ponty (1962: 130), our previous experiences leave within us, “a ‘world of thoughts,’ or a sediment left by our mental processes, which enables us to rely on our concepts and acquired judgments as we might on the things there in front of us, presented globally, without there being any need for us to resynthesize them”. Elsewhere I have used the notions of transparency and sedimentation to describe user relationships with cellular phones, and to explore the ways phone usage causes driver distraction (e.g., Rosenberger 2012).

  5. 5.

    Ihde’s conception of embodiment relations is deeply indebted to Martin Heidegger’s (1927/1953) account of tool use in his work Being and Time. Ihde’s account of multistability in some ways builds from, and in some ways is a critique of, Heidegger’s account of the breakdown of tools. And Ihde’s example of the hammer is an ironic play on Heidegger’s iconic descriptions of the use of a hammer as it breaks. The reason that I do not rely on Heidegger centrally in this paper is that I do not wish my account to become implicated in the larger ontological framework that he develops.

  6. 6.

    Despite sharing the same meaning, these two terms—stability and variation—each appear to more strongly rhetorically emphasize one of the two ideas which the notion of multistaiblity simultaneously emphasizes. The term “variation” more strongly calls to attention the idea that a variety of relations are always possible between a user and a technology. In contrast, the term “stability” more strongly calls to attention the idea that only some relations between a human and a technology will be stable. That is, it calls attention to the fact that a technology cannot be used to do simply anything.

  7. 7.

    It is important to note that, though it is sometimes claimed that the postphenomenological notion of “multistability” is equivalent to the social construction of technology (SCOT) notion of “interpretive flexability,” the two are quite distinct concepts. Verbeek has implied this equivalency, for example, in the following quote, “the existence of multistability—a product’s ‘interpretative flexibility’ as Bijker (1995: 20) calls it—need not hamper designers in explicitly trying to anticipate the mediating role of products in their use context” (2005: 217).

    But the two terms point to different things. Where multistability refers to the way that a current technology can be interpreted in multiple ways, the notion of interpretative flexibility instead refers to the state of a community’s relationship to a technology in an early stage of the history of its social acceptance. Trevor Pinch and Wiebe Bijker write, “By this [interpretative flexibility] we mean, not only that there is flexibility in how people think of, or interpret, artefacts, but also that there is flexibility in how artefacts are designed” (1984: 421). That is, the notion of interpretative flexibility points to the early stage of the social acceptance of a technology in which the conception of the device remains general enough to be subject to multiple designs. This highlights an important difference in emphasis between SCOT and postphenomenology (one that parallels a difference that can also be seen between SCOT and ANT): where SCOT focuses on the history of the social conflict leading to the establishment of a technological design, postphenomenology instead focuses on the continuing potential for any technology to be fit into multiple contexts.

  8. 8.

    Interestingly, despite Latour’s consistent use of Heidegger’s later work as a key example of exactly what he is arguing against, a literature is developing which analyzes the similarities between the ideas of these two thinkers (Khong 2003; Riis 2008; Harman 2009; Kochan 2010; Paddock 2010).

    Ihde too is a consistent critic of Heidegger’s later work (see, especially, 2010). For counterarguments against the criticism of Ihde and others against Heidegger, see Scharff (2010).

  9. 9.

    Verbeek (2005: 164f.) puts it this way, “Actor-network theory is primarily interested in unraveling the networks of relations by virtue of which entities emerge into presence, while a postphenomenological approach, by contrast, seeks to understand the relations that humans have with those entities”.

    It should be noted that Latour has a history of making dismissive comments about phenomenology which I do not refute here (e.g., Latour 1993: 58; Latour 1999: 9). See Smith (2003) and Verbeek (2005) for counterarguments.

  10. 10.

    See the debates cited in footnote 1. Trevor Pinch (2010) raises these issues in critique of Latour and Verbeek’s use of the notion of delegation. Pinch claims that the idea that nonhumans influence human action, exemplified by the speed bump’s influence on drivers, is trivial; nonhumans are constantly influencing humans everywhere. He writes, “Latour and Verbeek, and others who advocate an ontological turn, need a means of sifting through all the myriad non-humans and their countless interactions with humans so that we can see the significant choices made by humans” (2010: 87). According to Pinch, the contemporary influence of particular technologies can only be understood by returning to the history of political disagreements between people (not people plus technologies), a history, in his view, made visible better by the tools of SCOT than by ANT or postphenomenology.

    I disagree. While Pinch is certainly correct that understanding the history of human disagreements over a technology is crucial to understanding that technology’s contemporary influence on humans, such a history alone is not always enough. The agential role of technology itself—that difficult ontological topic of philosophy—is also relevant, since the effects of that technology on humans may not reduce to the intentions of any of the human participants in the historical disagreement.

  11. 11.

    As Lee and Brown (1994: 782) summarize this conception, “The scorched-earth policy that Callon, Law, and Latour adopt, the leveling of the taken-for-granted natural, social, and technological worlds into a heterogeneous field of force, is a calculated risk taken to subvert the established world picture of nested systems”.

  12. 12.

    I claim that my modified conception of symmetry offered here is ultimately consistent with the provocative set of reflections on the nature of technological agency put forward by the actor-network theorists Callon and Law (1995). They argue persuasively against a traditional conception of agency that would limit its application to only actors capable of possessing intentions, i.e., humans. Postphenomenologists agree with this extension of agency across human-technology relations, and themselves also struggle to determine how best to conceive of technological agency (see especially Verbeek 2005). My contention is that Callon and Law’s arguments in favor of extending agency to nonhumans, and against a conception that limits its application to only intention-possessing humans, do not require that the roles played by humans and nonhumans within their overall networked agency be identical. Nor do they require that all properties of humans and nonhumans reduce to those which happen to be interchangeable. Indeed, by recognizing that humans have a special intention-possessing ability (in their argument that intention possession is not required for agency), Law and Callon remain open the possibility of taking up the modified kind of symmetry I propose.

  13. 13.

    A distinction can be made between my own adjustment to Latour’s conception of symmetry here, and Ihde’s criticism of Latour. For Ihde, full blown symmetry like the kind Latour defends (and I affirm here) is too strong a position to hold, and thus an asymmetry is preferable. He writes, “I would find it hard to say—at least without claiming a highly metaphorical attribution—that the speed bump (sleeping policeman) is filled with designers, administrators, and policemen! I can’t quite bring myself to the level of ‘socializing’ artifacts. They may be interactants, but they are not quite actants” (2003: 139). In contrast, I am willing to grant full symmetry to the analysis of the agency of humans and nonhumans, so long as, contra Latour, humans and nonhumans are able retain certain distinctive attributes that render them, at least at times, non-interchangeable.

  14. 14.

    I take the methodology developed in this paper to be in tune with the spirit of Latour’s series of papers lamenting the current trend of “critique” in academic discourse. In Latour’s view, contemporary critique tends only to tear down, rather than build constructively upon, previous work (see esp. Latour 2004).

  15. 15.

    A distinction can be made between Edmund Husserl’s method of eidetic reduction and the method of variational cross-examination that I outline here. For Husserl, it is possible to reach the “essence” of a target of investigation by considering it from various perspectives. By taking up those various perspectives it becomes clear which aspects of one’s observations are accidental to a particular perspective, and which instead are the essential features of that object (Husserl 1950). In contrast, in variational cross-examination the target of investigation is the stabilities themselves. Nothing is revealed about the target’s essence independent of perspective, independent of the context of the examination. This difference between the essence-targeted method of eidetic reduction, and the relations-targeted method of variational cross-examination, is an example of the non-foundational, anti-essentializing, pragmatic commitments of postphenomenology.

  16. 16.

    I first developed the notion of “relational strategies”—the specific understanding and bodily approach one brings to a technology to relate to it in terms of a particular stability—through the example of desktop computing. There I contrasted different ways to approach computing in the event that one’s internet connection suddenly and unexpectedly slows (Rosenberger 2009, 2013b). This notion has also proven useful in a recent debate over the ethical and pedagogical implications of computer-simulated frog dissection in the classroom (Friesen 2011; Rosenberger 2011b).

    The notion of relational strategies is also related to the notion of “hermeneutic strategies,” which refers to the interpretive approach that one brings to the perception of a multistable readable technology, such as a laboratory image (e.g., Rosenberger, 2011a, 2013a).

  17. 17.

    From here, a fruitful direction to develop these ideas further would be to integrate ANT notions of the strength and density of the nodes and connections that make up a network (e.g., Latour 1999). With these metaphors, ANT practitioners attempt to articulate just how powerfully an actor is fixed within a network. This is another aspect of stabilities subject to critical contrast through variational cross-examination.

  18. 18.

    It is on this point where philosopher of technology Andrew Feenberg picks up. He articulates the importance of understanding the larger political/economic systems within which programs of actions fit (e.g., Feenberg 1999).

  19. 19.

    Of course the notion of concrete tailoring has much overlap with the script analysis notion of “inscription” (Akrich 1992). However, I want to establish a distinct concept so that the kind of material manipulation of concern here does not get swallowed up by the linguistic and social connotations carried by the script terminology. For a critique of script theory in tune with my reservations, see Verbeek and Kockelkoren (1998).

  20. 20.

    This introduces epistemological questions into postphenomenological thought—a step forward, I think, for this account. How to productively combine and cross-analyze situated perspectives to draw out biases in the analysis of multistable technology is a question that I cannot address here. But it is clear upon whose work such a project should be built (e.g., Code 1991; Star 1991; Haraway 1991; Harding 1991; Alcoff 1997; Hartsock 1998). It seems possible to utilize the insights of feminist standpoint epistemology and feminist phenomenology to develop ways to “operationalize” this kind of analysis, as Harding would put it, for considering the situated biases of various actors.

    For a thoughtful analysis of the notion of multistability that also approaches the sorts of methodological concerns raised in this section of the paper, see Whyte (forthcoming).


  1. Akrich, M. (1992). The description of technical artifacts. In W. E. Bijker & J. Law (Eds.), Shaping technology/building society (pp. 205–224). Cambridge: MIT Press.Google Scholar
  2. Alcoff, L. M. (1997). The politics of postmodern feminism, revisited. Cultural Critique, 36, 5–27.CrossRefGoogle Scholar
  3. Barad, K. (2003). Posthumanist performativity: Toward an understanding of how matter comes to matter. Signs, 28(3), 801–829.CrossRefGoogle Scholar
  4. Bijker, W. E. (1995). Of bicycles, bakelites, and bulbs: toward a theory of sociotechnical change. Cambridge: MIT Press.Google Scholar
  5. Callon, M., & Law, J. (1995). Agency and the hybrid collectif. The South Atlantic Quarterly, 94(2), 480–507.Google Scholar
  6. Code, L. (1991). What can she know?. Ithaca: Cornell University Press.Google Scholar
  7. Collins, H. M. (1994). Review of we have never been modern. Isis, 85(4), 672–674.CrossRefGoogle Scholar
  8. Feenberg, A. (1999). Questioning technology. New York: Routledge.Google Scholar
  9. Friesen, N. (2011). Dissection and simulation: Brilliance and transparency, or encumbrance and disruption? Techné: Research in Philosophy and Technology, 15(3), 185–200.Google Scholar
  10. Friis, J. K. B. O., & Crease, R. P. (Eds.) (forthcoming). Technoscience & Postphenomenology: The Manhattan Papers. Lexington Books/Rowman Littlefield Press.Google Scholar
  11. Haraway, D. (1991). Situated knowledge’s: The science question in feminism and the privilege of partial perspective. In Simians, Cyborgs and Women, chap. 9. New York: Routledge.Google Scholar
  12. Haraway, D. (1997). Modest_Witness@Second_Millenium. New York: Routledge.Google Scholar
  13. Harding, S. (1991). Whose science? Whose knowledge?. Ithaca: Cornell University Press.Google Scholar
  14. Harman, G. (2009). Prince of networks: Bruno Latour and metaphysics. Melbourne: Scholar
  15. Hartsock, N. C. M. (1998). The feminist standpoint revisited and other essays. Boulder: Westview Press.Google Scholar
  16. Hasse, C. (forthcoming.) The Anthropology of Learning Organizational Culture. Springer.Google Scholar
  17. Heidegger, M. (1927/1953). Being and Time. (J. Stanbaugh, 1996 Trans.). Albany: SUNY Press.Google Scholar
  18. Heidegger, M. (1982). Parmenides. Frankfurt: Klostermann.Google Scholar
  19. Hildebrandt, M. (2007). Ambient intelligence, criminal liability, and democracy. Criminal Law and Philosophy, 2(2), 163–180.CrossRefGoogle Scholar
  20. Husserl, E. (1950). Cartesian Meditations. (D. Cairns, Trans.). Dordrecht: Kluwer Academic Publishers.Google Scholar
  21. Ihde, D. (1990). Technology and the lifeworld. Bloomington: Indiana University Press.Google Scholar
  22. Ihde, D. (1993). Postphenomenology: Essays in the postmodern context. Evanston: Northwestern University Press.Google Scholar
  23. Ihde, D. (2003). If phenomenology is an albatross, is post-phenomenology possible? In D. Ihde & E. Selinger (Eds.), Chasing technoscience (pp. 131–144). Bloomington: Indiana University Press.Google Scholar
  24. Ihde, D. (2009). Postphenomenology and technoscience: The Peking University Lectures. Albany: SUNY Press.Google Scholar
  25. Ihde, D. (2010). Heidegger’s technologies: Postphenomenological perspectives. New York: Fordham University Press.CrossRefGoogle Scholar
  26. Khong, L. (2003). Actants and enframing: Heidegger and latour on technology. Studies in History and Philosophy of Science, 34, 603–704.CrossRefGoogle Scholar
  27. Kochan, J. (2010). Latour’s Heidegger. Social Studies of Science, 40(4), 579–598.CrossRefGoogle Scholar
  28. Latour, B. (1991). Technology is society made durable. In J. Law (Ed.), A sociology of monsters (pp. 103–131). London: Routledge.Google Scholar
  29. Latour, B. (1992). Where are the missing masses? The sociology of a few mundane artifacts. In W. Bijker & J. Law (Eds.), Shaping technology/building society (pp. 226–258). London: MIT Press.Google Scholar
  30. Latour, B. (1993). We have never been modern. Cambridge: Harvard University Press.Google Scholar
  31. Latour, B. (1994). On technological mediation: Philosophy, sociology, genealogy. Common Knowledge, 3, 29–64.Google Scholar
  32. Latour, B. (1995). A door must be either open or shut: A little philosophy of techniques. In A. Feenberg & A. Hannay (Eds.), Technology and the politics of knowledge (pp. 272–281). Bloomington: Indiana University Press.Google Scholar
  33. Latour, B. (1999). Pandora’s hope: Essays on the reality of science studies. Cambridge: Harvard University Press.Google Scholar
  34. Latour, B. (2004). Why has critique run out of steam? From matters of fact to matters of concern. Critical Inquiry, 30, 225–248.CrossRefGoogle Scholar
  35. Latour, B., Mauguin, P., & Teil, G. (1992). A note on socio-technical graphs. Social Studies of Science, 22(1), 33–57.CrossRefGoogle Scholar
  36. Lee, N., & Brown, S. (1994). Otherness and the actor network: The undiscovered continent. American Behavioral Scientist, 37(6), 772–790.CrossRefGoogle Scholar
  37. Merleau-Ponty, M. (1962). Phenomenology of Perception. (C. Smith, Trans.). New York: Routledge.Google Scholar
  38. Oudshoorn, N., & Pinch, T. (Eds.). (2003). How users matter: The co-construction of users and technologies. Cambridge: MIT Press.Google Scholar
  39. Paddock, T. (2010). Bridges: Technology and the social. Environment, Space, Place, 2(2), 7–27.CrossRefGoogle Scholar
  40. Pickering, A. (Ed.). (1994). Science as practice and culture. Chicago: Chicago University Press.Google Scholar
  41. Pickering, A. (1995). The mangle of practice: Time, agency, and science. Chicago: University of Chicago Press.CrossRefGoogle Scholar
  42. Pinch, T. J. (2010). On making infrastructure visible: Putting the non-humans to rights. Cambridge Journal of Economics, 34, 77–89.CrossRefGoogle Scholar
  43. Pinch, T. J., & Bijker, W. (1984). The social construction of facts and artifacts: Or how the sociology of science and the sociology of technology might benefit each other. Social Studies of Science, 14(3), 399–441.CrossRefGoogle Scholar
  44. Riis, S. (2008). The symmetry between Latour and Heidegger: The technique of turning a police officer into a speed bump. Social Studies of Science, 38(2), 285–301.CrossRefGoogle Scholar
  45. Rosenberger, R. (2009). The sudden experience of the computer. AI & Society, 24, 173–180.CrossRefGoogle Scholar
  46. Rosenberger, R. (2011a). A case study in the applied philosophy of imaging: The synaptic vesicle debate. Science, Technology and Human Values, 36(1), 6–32.CrossRefGoogle Scholar
  47. Rosenberger, R. (2011b). A phenomenological defense of computer-simulated frog dissection. Techné: Research in Philosophy and Technology, 15(3), 215–228.Google Scholar
  48. Rosenberger, R. (2012). Embodied technology and the dangers of using the phone while driving. Phenomenology & the Cognitive Sciences, 11(1), 79–94.CrossRefGoogle Scholar
  49. Rosenberger, R. (2013a). Mediating mars: Perceptual experience and scientific imaging technologies. Foundations of Science, 18, 75–91.CrossRefGoogle Scholar
  50. Rosenberger, R. (2013b). The importance of generalized bodily habits for a future world of ubiquitous computing. AI & Society, 28, 289–296.CrossRefGoogle Scholar
  51. Rosenberger, R., & Verbeek, P.-P. (Eds.) (2014). Postphenomenological Investigations: Essays on Human-Technology Relations. Lexington Books/Rowman Littlefield Press.Google Scholar
  52. Scharff, R. C. (2010). Technoscience studies after Heidegger? Not yet. Philosophy Today, 54, 106–114.Google Scholar
  53. Smith, A. (2003). Do you believe in ethics? Latour and Ihde in the trenches of the science wars (Or: Watch out Latour, Ihde’s got a gun). In D. Ihde & E. Selinger (Eds.), Chasing technoscience (pp. 182–194). Bloomington: Indiana University Press.Google Scholar
  54. Star, S. L. (1991). Power, technologies and the phenomenology of conventions: On being alergic to onions. In J. Law (Ed.), A sociology of monsters? (pp. 26–56). London: Rutledge.Google Scholar
  55. Verbeek, P. -P. (2005). What things do: Philosophical reflections on technology, agency, and design. University Park: Penn State University Press.Google Scholar
  56. Verbeek, P. -P. (2011). Moralizing technology. Chicago: Chicago University Press.CrossRefGoogle Scholar
  57. Verbeek, P. -P., & Kockelkoren, P. (1998). The things that matter. Design Issues, 14(3), 28–42.CrossRefGoogle Scholar
  58. Whyte, K. P. (forthcoming). What is multistability? A theory of the keystone concept of postphenomenological research. In J. K. B. O. Friis & R. P. Crease (Eds.), Technoscience and Postphenomenology: The Manhattan Papers. Lexington Books/Rowman Littlefield Press.Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2014

Authors and Affiliations

  1. 1.School of Public PolicyGeorgia Institute of TechnologyAtlantaUSA

Personalised recommendations