Undoing Power
Klaus Krippendorff
The Annenberg School for Communication
University of Pennsylvania, Philadelphia
From: Critical Studies in Mass Communication 12, 2: 101-132, 1995
"Liberation is a way of talking about power."
--- Bruce A. Ackerman (1980, p.6)
"Philosophers merely interpreted the world ... the point is to change it."
--- Karl Marx, 1845 (1978, p.7)


This essay speaks into power -- not about what the powerless lack or what the powerful have too much of and in terms of which empowerment would mean appropriating it from Others. It explores some of the entailments of talking and conceiving our world in these rather common monological terms. In the course of this examination, an understanding of power arises that is tied to being in dialogue with Others and involves the bodily experiences of languaging. Here, empowerment means something altogether different.

As the deliberately ambiguous title, "undoing power," suggests, this essay not only explores the undoability of power, it also shows, to the extend that a written paper can, the power to "undo" such phenomena, whether as critical social scientists or in everyday life.

In pursuit of these aims, I will start with an experiment in perception. It may not appear obvious how this relates to "power." However, it demonstrates, at least to me, our ability to see our seeing differently which creates an opening for us to enter into different uni-verses of being. From these, we will be able to look back to traditional ways of writing about power.

Doing Ordinary Languaging

Frankly, I much prefer talking to writing. Listening to the voices of Others and responding to what shows up while being with people can hardly be approximated by writing. Without the possibility of dialogue with you, the reader, I have to imagine your reading and in my internal dialogue, I am offering to walk with you through a common but far too little practiced experience.

Imagine glancing over the advertisement pages of a newspaper, like the New York Times, where this portrait of an Indian (Ernst, 1986, p.28) may be found.

I invite you to ignore the fact that this "recontextualized" image now occupies a rather prominent position on this page of a scholarly journal. Chances are that, without my pointing to this figure, you might have never noticed it among the dazzling displays of an ad page competing for attention. If this is so, I claim, your having it in focus is not literally caused by a physical stimulus, by what the ad page and the image "is". You did not start by sensing all details on this page before interpreting them, "figuring" out what each meant and ending up talking about what you selected for attention. Rather, the "causality" of seeing came from active reading, from your choosing to accept my invitation to look for a face, from our joint participation in a discourse. Your accepting my invitation coordinated your attention relative to mine. It primed you to look for something affording what you understood me to say it is. In other words, your reading brought forth what you are now seeing.

If this gedankenexperiment seems unconvincing, perhaps because you undoubtedly know or are telling yourself like everyone else presumably does that 'what you see in front of your eyes objectively exists there for everyone else to see exactly as it is,' much like a camera takes pictures of what it faces, I am inviting you to look again, but this time at the eskimo. At the eskimo? I'm not kidding! Having settled on seeing a human face, seeing something else will most certainly be difficult.

I know that my request to see B instead of A troubles a whole cluster of foundational and highly rationalized epistemological commitments: our ability to distinguish between reality and illusion, our belief in being able to rely on observations to settle ontological disputes, our customary reliance on the already mentioned optical metaphors to explain human perception which supports such ideas as cognitive representation, information processing, the communicational notion of coding and many more. Such rational/technical explanations -- after all a lens is a well understood technical device -- make my request to see an eskimo in place of a human face difficult to accept, at least on logical grounds. Remarkably, it is these explanations, not our seeing, that get in the way of seeing our own seeing, and impede our understanding of our own linguistic involvement in the world we see. Seriously; please stop reading here and take a minute of your time until you no longer look for an eskimo, but see it (her or him).

If you haven't gotten it yet, here is my suggestion: Don't look at the Indian's eye; forget his nose and mouth, and focus instead on the lower part of the drawing. Look for someone's shoes pointing away from you. When you recognize them, then move your attention upward again and you will find they belong to someone facing away from you, the dark background, and wearing a thick winter coat, like eskimos supposedly do or did.

Having accomplished this perceptual feat, hopefully, try to see something even more outrageous without my coaching. If this proves difficult, at least put yourself in an artist's shoes and explore how it could have been drawn, or try to see it like a printer might who is looking for production defects.

+ - + - + - + - +

However far you were willing to go into such re-visions, I want to draw four conclusions from the experiences I imagine you have had.

First, we tend to see the for each of us obvious, the expected and the familiar. Going through the above experiences, your reading primed you to see a human face and you most likely did. At that moment, you allowed yourself to be locked into one way of seeing, into one single uni-verse, and this excluded or blinded you from seeing any other uni-verse. Without the effort on your part to break out of this self-confinement, you would have neither experienced nor missed alternative ways of seeing, much less imagined that equally sane people could simultaneously be in altogether different uni-verses, without ever knowing this of each other. This is a quite remarkable fact that most common sense accounts of perception, of communication and, as we shall see, of power conveniently ignore.

This "way of seeing," I might add here, appears to us as a particular relational experience: Although it is constitutively impossible for us to distinguish between what something "really" is and how it appears to us -- for both take place within our very own process of seeing -- we certainly can experience our own actions. For example, we can move something in and out of our visual field, cover it with our hand or our retrace it with a pen on paper, and thereby become cognizant that our seeing is related to what we are or someone else is doing. We might conceptualize seeing as a process of "being-with" something that is, however, constitutively unknowable outside its participation in this relationship.

Second, your ability to shift from one way of being-with to another way of being-with, your experience that what you see in one universe overrides, extinguishes, and excludes what you had seen in any other universe, completely absolves the unchanging physical properties of the image from being the primary cause of your seeing. This experience pulls the rug out from under any objectivist metaphysics which assumes objects exist independent of our seeing and derives from the above mentioned optical=causal=observer/spectator metaphor of seeing. To be fair, we do not deny the so-called physics of the above image. However, this is nothing other than the account of yet another way of seeing, one offered by a standard physicist, perhaps aided by in-themselves-blind instruments, but not therefore the only one description (of the universe) that counts. And it would be difficult to imagine, at least for me, how any one way of seeing or being-in-a-world could literally cause another way of seeing or being-in-another-world. While our experiment speaks against objectivist reality claims and in favor of postmodern conceptions, it does not license solipsism either. In fact, the non-arbitrariness or "reality" of our seeing shows up in the very difficulty we experience in trying to get out of one universe and into another. Some such shifts require more efforts, some less and some are simply not afforded in any kind of being-with. One could say, the "real" does not lie in what is, but in the constraints on bringing forth alternative ways of seeing and these are partly our own.

Third, our ability to remember the re-visions we experienced, to re-cognize the universes we passed through and could be visiting again, leads us to conclude that we live in a multi-verse that is constituted in the possible reality constructions it affords. Sadly, instead of celebrating this liberating insight, we tend to privilege one universe, one version of the world, at the expense of all others, for example, by submitting to the exclusionary view of physics, and dismissing all our experiences that are inconsistent with this one exclusionary view as imperfect, subjective, unreliable, biased, illusionary, unreal, mere interpretations, and so forth. This focus on ontology at the expense of epistemology or hermeneutics and languaging also leads us to disown our experiences and inscribe them onto external stimuli. This is evident when we take our image, for example, as a container of meanings (having content), as ambiguous (leading to confusing perceptions), as an illusionary artifact (intended to deceive us), as a mere representation (that can be either true or false -- the "real uni-verse" in contrast always being what it is), and so forth. The privileging of one universe responds to what Richard J. Bernstein (1992, p.17) calls our "Cartesian anxiety" or to what Humberto Maturana and Francisco Varela (1987, p.16) call our "temptation of certainty." Attitudes like these have little if anything to do with what exists but much with how we live. I recommend vigorous questioning of these deep seated epistemological fallacies whatever their motivation and wherever they show up. Belief in a singular uni-verse, in an one ontology for everyone to live in, essentially denies us our own experiences. Our experiment demonstrates the possibility of multi-verses in the domain of seeing. We shall draw on this in the undoing of power.

Fourth, and perhaps most importantly, our ability to shift from one universe to another also is a joint accomplishment (Shotter, 1993, pp.38ff). We (you and I) virtually (because we are not co-present) collaborate in creating these experiences through your reading of my writing, through languaging, albeit restricted here to writing. As already suggested, had you come to this image accidentally, you would have seen it in whichever way it would have appeared to you. It is your reading, my writing, and our longstanding dialogical involvement in language that made the difference. Our experiment demonstrates that languaging -- the process of people co-ordinating their voicing and listening, their writing and reading, their feelings and doings and their being-with each other -- virtually "says" things into being, "brings forth" ways of seeing.

Languaging is a social process in which we jointly construct realities for each of us to see, occupy and to talk into. Evidently, languaging can open our "eyes" to alternative ways of seeing and create new ways of being-with Others. Talking or writing of power is just one example which we shall now explore.

Some Discourses on Power

"Power" is first of all a word. This is probably the only statement everyone can agree on. Words are public, occur in dialogue and, as we have seen above, accomplish something in this world. The use of the word "power" is no exception, yet extremely varied. To give some examples, let me quote from the writing of Robin Tolmach Lakoff who, as a linguist, gives us a good sense of some common uses of "power."

Power ... is vibrant, the very word conjuring up images of strength, force, action. Whether positive or negative, those images are strong. Power is physical: it changes reality, it gets things done or undoes what exists. It creates effects that can be seen, felt, and measured. Power is the engine of the 747 that lifts the behemoth off the runway; it is the "plow that broke the plains"; the firing squad; the nuclear bomb under discussion at the conference table; the parent who can give or withhold the keys to the car; the boss who can hire or fire at whim. All of these operate to change reality for better or worse. We may admire power or resent it, but we can see its operations and feel its physical effects on us (1990, p.12). Let me also add a couple of passages written by Michel Foucault who succeeded in introducing a whole new vocabulary into the discourse on power. He dismissed the notion of power as a simple, isolatable relationship between master and slave or between oppressor and oppressed. Suggesting it to be manifest instead in complex networks of relationships, including of institutions, of knowledge, and so forth, he saw power as double sided: What makes power hold good, what makes it accepted, is simply the fact that it doesn't only weigh on us as a force that says "no", but that it traverses and produces things, it induces pleasure, forms knowledge, produces discourse. It needs to be considered as a productive network which runs through the whole social body, much more than a negative instance whose function is repression. (1980, p.119). He also recognized its pervasiveness: Power is everywhere; not because it embraces everything but because it comes from everywhere. ... Power comes from below. ... There is no binary and all-encompassing opposition between rules and ruled at the root of power relations, and serving as a general matrix (1978, pp.92-94). ... power is "always already there," ... one is never "outside" it, ... there are no "margins" for those who break with the system to gamble in (1980, pp.141). Roland Barthes virtually echoes Foucault's voice: ... power is present in the most delicate mechanisms of social exchange: not only in the State, in classes, in groups, but even in fashion, public opinion, entertainment, sports, news, family and private relations, and even in the liberating impulses which attempt to counteract it (1982, p.460). Beyond this collage of different uses of "power," what can we say about this everywhere present, awe-invoking and otherwise quite contradictory "something"? Are there commonalities? Should there be any? What fuels the fascination with the concept? Let me make two points that only partially answer these questions. The first is constructive and concerned with power as an undoable social construction, that is, one whose very nature is that it can also be undone. The second is critical (or deconstructive?) and concerned with its predominantly disembodied and disabling formulation.

Power as Undoable

Having lead you through the above re-visions, it should not be difficult to see why I am lead to oppose the above cited notions. I suggest power to be not "natural" in the sense of being unavoidable, omnipresent, affecting everyone albeit in unequal measures and regardless of how it occurs in language. The fact that traditional social scientists have theorized power as independent of observation and as outside human agency can hardly serve as reason not to conceive power as essentially erasable, extinguishable, voidable, overcomeable or undoable,. Let me offer a four point definition of undoability and then discuss each with reference to the above uses of power.

(1) Efforts to undo a particular phenomena are preceded if not motivated by feelings of being locked into a particular way of languaging with Others, having come to a dead end, or seeing Others as being so involved.

(2) Undoable phenomena must be variously articulable, mutable and can thus show up in several forms.

(3) Undoable phenomena are contestable in principle, but surface as such only in actual contestations, when their present articulations are questioned, challenged, appropriated, denied or re-articulated in dialogue with Others.

(4) Undoable concepts may be replaced by re-articulations.

Accordingly and re (1): Power is not an entity, a thing or a resource. It arises as an experience that is brought forth and clarified in dialogue. In Foucault's (1979, p.16) terms, it "is exercised rather than possessed" to which I insist on adding, by someone and in "words." This addition is important because, to me, power is most profitably seen as embodied in the lives of people with very real bodies saying things to each other, in their actual languaging, which includes uttering explanations, commands, dismissals, threats, promises as well as giving indications of acceptance, obeyance, compliance, submission or agreement. What people say to each other can not be separated from the emotions accompanying the process of their being with each other, from what it means to them, without losing track of the phenomenon in question. In particular, I am proposing that power is felt in being locked into a burdensome way of languaging, into a worrisome dialogical practice, into an oppressive Wittgensteinian language game, much like being locked into a way of seeing, from which, however -- and quite unlike our forgoing experiment -- there seems to be no readily apparent escape (Krippendorff, 1989b; Stolzenberg, 1978). The words accounting for the feelings of being burdened, powerless, confined, troubled, oppressed, marginalized, disadvantaged or worried, for example, distinguish among the emotions that could concur with such dialogical entrapments. Although this notion of power seems to emphasize one side of a relationship, at least superficially, those in positions of authority are hardly immune to feelings of entrapment either, for example, when their authority is challenged without the apparent possibility to reassert it. This is true for popular heroes who are at the mercy of their fans as much as it is for tyrants who can not exist without rewarding their subjects or henchman for executing their commands. The fear of losing such authority indicates entrapment as well. The feeling of being without burden is not distinguishable from the feeling of simply going on with life, effortlessly and seamlessly, whether it involves using one's mind, employing a technology or relying on Others. In suggesting that we understand power as dialogically embodied, I am turning away from mere theoretical abstractions, grammatical categories, purely verbal accounts of behavioral events and from individualistic and mentalistic constructions. To me, concepts are both felt and articulated in view of Others' feelings. Thus, I am using the word "power" not to represent anything mysterious, but to gain access to the kind of languaging in which it arises.

Re (2): The articulability of undoable concepts. The quotations in the previous section (Lakoff, Foucault, Barthes) make it abundantly clear that we do not face a singular phenomenon called "power." The power of an atom bomb is certainly different from the power wielded by a CEO, the power exercised by a judge in court, the power of disqualifying Others from talking about a subject matter, or the power of reading a text that shifts one's perceptions to something heretofore unseen. Although many writers on power use the word without definition -- as if its meaning were self-evident and obvious -- short of concluding that it is a catch-all-and-say-nothing concept, the examples mentioned so far exhibit, at best, Wittgensteinian family resemblances (Wittgenstein, 1958, p.32). They do not share a single set of features but are connected by pair-wise commonalities.

To illustrate, let me quote two of Lakoff's examples of dialogue fragments in which power is not articulated, for they indicate nothing of what is felt:

He: Wanna go to the movies?
She: Oh, I don't know ... do you? (Lakoff, 1990, p.18).

Here is one articulation: This is an episode of a currently widespread male/female game. He is making a suggestion phrased as an inquiry into her wants. She can accept his suggestion, propose an alternative or talk of her own likes (including leaving the scene). In either case, she would be taking responsibilities for the consequences of her responses. Particularly in making a counter proposal, she would present herself as equal to him. However, should such a proposal fail, she would feel guilty for not maintaining the relationship she is part of. Fearful of the outcome, she hedges, leaves the decision to him, never gets where she wants to be and ends up unhappy and powerless with him in charge of her feelings. Lakoff (1990, p.18) also offers a complement to this example:

She: Tell me how to fix this Xerox machine.
He: Oh, don't you worry about this, honey. Leave it to me.

Here she takes the initiative in asserting what she wants. On the surface he is caring, the "honey" suggesting intimacy, and his offer to save her time and energy seems generous. But, look again. In requesting help, she also grants him possession of knowledge she is lacking. Under the guise of being polite and considerate, he seizes on the opportunity to keep his knowledge to himself. He, thus, enacts the stereotype that fixing a machine is a man's job, and that women lack this ability. Such games played over and over again and in different empirical domains -- their legitimation and objectification -- are a way of generalizing her feeling of incompetence and his feeling at being so capable. This accounts for why the two episodes are so common.

Note that neither of these two dialogue fragments contain the word "power," nor any of its synonyms much less any expression of what he and she actually feels. Taking these two-liners literally and by themselves, without access to either their history of interactions or the possibility of exploring the participants' feelings in the presence of each other, one could not possibly say much about power in this dialogue. It is Lakoff who, as an expert outsider, articulates these episodes in terms of power, and there surely could be other articulations as well. Her linguistic observer role prevents her access to the feelings concurrent with these dialogues. Her articulations are hers and any projecting them onto others remains problematic.

Respecting the ownership of these articulations, we can not be concerned with which if these articulations or any others one may advance are correct. What I wish to point out is: (a) In these (Lakoff's or my) articulations of the two episodes, we can already distinguish at least two entirely different notions of power at work. In the first example, the dialogical inequality is invited by her yielding -- not to his initiative, but to her fear of tampering with the relationship. In the second example, her initiative is undermined by his diverting her request into a game in which he can keep his knowledge -- a valued resource for his superiority -- to himself. And (b), although these articulations make sense, at least to me, if they would enter the very languaging they articulate, they may be accepted but they could also meet rejection or become radically re-articulated. A single disembodied concept of power covering all of these phenomena will not only conceal their dialogical differences, but outside observers may not have a clue as to what is going on here unless they get interactively involved.

Going back into these articulation, the asymmetry in both examples -- the unfairness women must feel all too often, the complaints this situation must fuel, the powerlessness that accompanies their involvement -- reveals to us the very dialogical entrapment that prevents taking emancipatory actions or formulating liberating re-visions of the power involved. It is these articulations that reveal power relationships and can encourage or discourage contestations. If they do not enter a dialogue, concepts of power can become abstract, unwittingly projected onto Others, and remain without empirical consequences.

Re (3): I believe it was Wittgenstein's family resemblances that led W. B. Gallie (1962, p.125) to introduce into the philosophical literature the notion of an essentially contested concept which chimes in with my undoability notion. His concept squarely challenges the assumption of agreements as a prerequisite of scientific understanding. For him, an essentially contested concept (a) signifies a valued achievement of (b) an internally complex whole that (c) is variously describable and (d) admits considerable modification in the light of changed circumstances. Finally (e), each party recognizes the fact that its own use of the concept is contested by the other parties. Gallie's examples include works of art and the concept of democracy, and he relies on the metaphor of competitive sports to illustrate how contestants have a stake in seeing to it that the concept in question stays being challenged. Power clearly is a valued achievement, which would satisfy Gallie's point (a), whether it means getting something done or getting Others to do what they would not do otherwise. The idea of family resemblances is expressed in power as being describable in multiple ways, (c), and as being modifiable (d), which conforms to my notion of various articulability (2). There is, thus, considerable similarity between essentially contested and undoable concepts, except for Gallie's fifth point (e), the issue of whether it makes sense to regard power as essentially contested.

In the literature on power, Gallie's idea has met understandable controversy. Steven Lukes (1974, p.9) presents his "conceptual analysis of power" claiming that power is "ineradicably evaluative and 'essentially contested'." Later, William E. Connolly (1983, p.126) claims that "attention to the grammar of 'power' makes it readily understandable why it is an essentially contested concept." In contrast, Thomas E. Wartenberg (1990, p.13) vehemently rejects the thesis of the essentially contested nature of power because, he believes, it can not explain the divergences among theories of power. Undoability can hardly be an issue for those who conceive their theories as descriptive of objective facts and motivate reformulations of them by claiming them either to improve predictability or to embrace other concepts, both in the belief of furthering consensus among diverse scholars. These traditionally behaviorist and positivist goals of science rest on the strict separation between the nature of power and theories thereof. They treat the former as the objects -- excluding the linguistic involvement by its human participants -- of the latter's generalizations and neither as articulations in their own right. This separation contradicts my point (2) and provides no place for appreciating the re-articulability of concepts. Here, I side, but for somewhat different reasons, with Anthony Giddens (1979, p.89) who argued -- with generalizations in mind -- that the thesis of essential contestedness "is mistaken if the implication is that some notions in the social sciences are essentially contested while others are not."

The real problem seems to be this: Gallie's definitional requirement (e) does not address the possibility that the very decision of whether or not a concept is essentially contested could come under the influence of "powerful" institutions that have an interest in keeping their "power" and thus discourage the concept from being challenged. In fact, if power is tied to an institutional identity, it becomes the most obvious target for such interests. For example, the belief in the "power" of the mass media to influence the public not only justifies the public attention entailed by this belief but also motivates the flow of its revenues. Because this belief has become constitutive of the mass media as a public institution it has become difficult to contest its power from within the very public sphere the mass media have created to be influential. Its power derives from living (doing research and participating) in a self-sustaining definition. Much communication research is actually dedicated to show that this power does exist and thus provides the ground of its continued sustenance. Naturally, all scientific generalizations of theories of power are popular and appreciated in circles that thrive on its "reality." Denying the constructed nature of power is a way to keeping that power effective and in the very place its detached theory creates for it. To me, efforts to hold on to such objectivist notions -- constructions that make alternative constructions difficult to see or implausible -- virtually cry out for articulations of the privileges its proponents enjoy or seek to achieve within society. My point (3) asserts that all parts of a discourse are contestable, at least in principle, which means they may not be contested in practice. Contestedness can not therefore be the issue because the absence of contestations -- as in our short dialogical episodes -- does not speak to their possibility.

As already said, when power is not articulated within a dialogue, we just can not say much about its reality. Where its articulations do occur but remain uncontested, willing compliance is a possibility. It is where they are subject of complaints, opposed, contradicted, questioned, and challenged, that everyday entrapments reveal their burdens, that vested interests show their colors, that conflicts can become violent and thus bring the actual strength of power to the fore.

Suppose the second of the above two dialogue fragments would continue as follows:

She: Tell me how to fix this Xerox machine.
He: Oh, don't worry about that, honey. Leave it to me.
She: I'm not your honey, all I am asking is to tell me how.
He: Wooow! -- Women can't do this, you know!
She: Don't give me this patronizing talk. Will you tell me or not?
He: Okay, okay, I'll tell you. Let's be friends.

To us as readers of this transcript, the third line signals "business" to be no longer "as usual." She rejects the kind of help Others in her position of relatedness may have gladly taken. She disrupts the usual game of male superiority and female submission and thus contests his role as an overtly generous helper. Repeating a request, "tell me how," almost always calls for literal interpretations, not metaphorical ones, and here it means "don't interpret my request (from line 1) as an expression of helplessness. I really mean what I requested of you." So, in this third line, the "rules" of this common game are in fact contested. In the fourth line, he is not only trying to hold on to the comfortable position the familiar language game would grant him, but bluntly reveals what seems to be his assumption in playing it: the stereotypical analogy -- male is to competence as female is to incompetence -- regarding fixing things. His construction of reality is challenged by her unanticipated "aggressiveness." It is at the moment of such contestation that power surfaces, becomes public and discursively accessible to us as outsiders, not before. In the fifth line, she does not merely respond to his previous assertion, but to the whole sequence of exchanges she had experienced so far. "Patronizing" signifies a complex coordination of sayings and feelings. Its dialogical meaning resides precisely within this and possibly similar past episodes, surfaces in processes of actual contestation and would not be apparent without it.

While this makes sense in the example, we must ask ourselves, why do we have to reject outside observers' claims of power being evident where contestations or expressions of entrapment are absent? Although linguists and objectivist conversational analysts may not like my answers, I am continuing to suggest that the two line episodes do not contain (internal) evidence of power or of what might be driving the dialogue. "He" and "she" could be reminiscing about how they met, playing a therapeutic game with reversed roles, or reading a script. We just don't know anything about their bodily involvement. Our reading (including Lakoff's interpretation and my re-interpretation of hers) puts these fragments into the context of our own experiences of languaging with Others, not into the context of their dialogical lives, and re-articulating their articulations is really all we can do without becoming interactively involved. However, claiming power to be at work where no one questions, interrogates, challenges or contests what is going on, is analogous to claiming intellectual capabilities superior to those being observed, listened to or read. Obviously, this objectivist position, shared by many discourse analysts, is very similar to that of the "he" in our episode, proclaiming an ability of fixing machinery women do not have. My point (3) would speak against such monological constructions of power whose truths are predicated on claiming privileged access to facts and on dismissing the articulations of dialogically involved Others as being unable to see their own oppression, as possessing false consciousness, as being ideologically biased or as denying the truths only superior analysts can see. The consequence of exploring a notion of power whose strength is revealed in contestations is that we have to become more actively involved and enter the dialogue in which these contestations take place. Here, we can do this only vicariously.

Finally, re (4): It should be obvious by now that all re-articulations -- much like those of our dialogue example -- do create new contexts, new frames, new meanings and are constitutive of new ways of languaging. Where re-articulations are woven back into the very fabric they articulate, they can no longer be considered neutral. Family members listening to the comments made by a family therapist can create new dialogical realities. Even observing one's habits on a television screen can encourage not only self-awareness but also behavioral changes. At a minimum, re-articulations add another voice and to have emancipatory effects listening to them is important. For instance: Re-articulations can be understood to complain about "the unfairness of it all" and, coupled with a belief in not being able to do anything about it, encourage and justify rage. Re-articulations can "blame others" for one's troubles and reinforce the entrapment evident in the original articulation. Re-articulations can make us aware of things otherwise unnoticed by telling us "this is where I always seem to fail." But re-articulations can also distinguish between the dialogical trap one finds oneself in and what could be outside of it. They can transform rules, overcome constraints on heretofore unimaginable options, create new ways of being and effect shifts in the languaging of Others as well. The two interlocutors in our episodes could not possibly have continued their games had they talked to Robin Lakoff or read this paper. In view of these possibilities, power can be seen as undoable.

The word "undoability" connotes first of all an action, the possibility of untying something complex and, second, of creating a new "fabric" in its place, perhaps with some of the same "strings." This might not always entail radical replacements. For once, there is individual memory which allows us to imagine what no longer exists. For another, the undoing of power always involves Others who may happen to re-vision and weave a different fabric. There is also the possibility of these Others feeling no burden at all, not playing into the contestations or not collaborating in the efforts to undo what one party feels burdensome. By having to involve Others, power may thus be undone only gradually, diminish or erode their strength in time, especially when participants resist, fight back or can escape. There are no guarantees.

The undoability of power is not a hypothesis that can be tested in terms of the frequencies of its occurrence. As we have seen, it is a possibility that is open to those capable of languaging, capable of re-visions and hence capable of freeing themselves from the need to react to past dialogical involvements. I would even go so far as to say that the ability to eliminate old ways of being with Others and construct new forms of languaging in their place is the essence of human agency. But this is another topic. Formulating power as undoable in this possibilistic sense calls for accounts (articulations or theories) of power that respect the linguistic agency of those implicated in its ontology. And creating novel accounts might be all that can be done to prevent theories of power from becoming the parasites that prey on human agency. Undoability is the ultimate consequence of the adage that power becomes slippery when reflected upon.

Mythical Regimes

In my critical comment on the discourses on power I want to take seriously the above contention that power arises in languaging, can be contested in languaging, can be overcome in languaging and is, thus, embodied in the languaging among real people. Given this point of departure, most theories of power -- from Max Weber to Talcott Parsons, Robert A. Dahl, Michel Foucault, Roland Barthes, to name only a few -- fail in two respects: One is their disembodiment and the other is their disabling formulation.

Both of these deficiencies have been recognized before, although in the context of criticizing another and classical epistemological pathology. In German Ideology, Karl Marx and Friedrich Engels (1970; 1978), writing one hundred fifty years ago, described what they called the "trick in three acts" that theorists, ideologists and philosophers -- especially Hegel -- performed in proving the hegemony of the human spirit in history or the importance of ideas as ruling an epoch. Marx and Engels were clear that detaching these ideas from those who hold them was not only a mistake, they called it "an illusion," but played into the hands -- one might even say into the heads -- of the ruling classes who distinguished themselves from those ruled by believing in just such ideas. Marx and Engels described this "trick" in the following three moves:

No. 1. One has to separate the ideas from the class of the [for a variety of reasons actually] ruling individuals and take them to be the ruling ideas. ...

No. 2. One has to bring an order into these ruling ideas, show [empirical] connections... [between them] which is accomplished by providing a [complete and] self-contained account of these... (this is possible because these ideas really are connected through their empirical grounding and produced by distinctions intrinsic to these ideas).

No. 3. To remove the hypothetical appearance of this self-contained account [of the order, history, system or logic of ideas], one attributes agency to it, refers to it as a person -- "the self consciousness" -- or, to appear thoroughly objective, as a sequence of persons [-- the historical consciousness --] who represent these ideas in history. The thinkers, philosophers or ideologists then come to be its "counsel of guardians"... Herewith, all material elements have been removed from history, and speculation [mythology] is given free rein (Marx and Engels, 1978, p.49; my translation; square brackets my insertions).

In this essay, however, I am not concerned, as Marx and Engels were, with idealism, how it provided the ruling classes with self-serving illusions and what caused their inability to listen to the emerging revolutionary voices. We have other issues today. One could also apply their criticism to their own theorizing and show their historical determinism to exhibit a similarly self-sealing and therefore no longer contestable autonomism. But this is not my issue either. What does concern me is that we seem to continue to perform the same or similar tricks on ourselves and create disembodied knowledge and regimes of power, whose formulation makes their understanding and contestation increasingly difficult. Let us see how this "scientific" trick in three moves -- abstraction, systematization, and mythologization -- appears in dialogical terms and with respect to power:

First move, abstraction: (1) Single out and abstract, relevant and objectively describable units or events from the (often dismissively regarded as "messy") languaging of real people listening to each others' voices and coordinating their lives relative to each other.

There exists a considerable consensus that power is transacted through language. Barthes (1982, pp.460-461) says "power is inscribed, for all human eternity, (in) language, or to be more precise, (in) its necessary expression: the language we speak and write... . Once uttered, ... speech enters the service of power." "I call the discourse of power any discourse that engenders blame, hence guilt, in its recipients." (Barthes, 1982, p.495). Foucault (1972, p.140) suggests that "the fundamental operation of power (should) be thought of as ... a speech act: enunciation of law, discourse of prohibition. The manifestation of power takes on the pure form of 'thou shalt not'." And in Lakoff's (1990, p.13) words: "Language is the initiator and interpreter of power relations." How this consensus is translated into accounts of power is a wholly different issue.

Traditionally described by its syntax and semantics, language already is an abstraction from processes of languaging as embodied in people, an abstraction that has been shaped by linguists in pursuit of an easily formalizable object for research, by bureaucratic and technological needs for standardized communications, and by educational demands for reproducible performances of students. Rules of grammar, thesauri, dictionaries, text books, codes of conduct and organizational manuals are some of the products of their common institutional interests (Harris, 1981, p.12). Building upon the phonemes, words and sentences of our conventional notion of language -- logically extended to the domains of speech acts, messages, texts, literatures -- conveniently excludes all references to their users, to the bodily life of languaging beings and to the history of their interweaving voices. On top of that, the above-mentioned speech acts are units of speech that derive from a particular theory of language, one that sees speech as a succession of distinct acts. Thus, taking speech acts as the smallest meaningful units ties power to already pretty abstract linguistic forms that may or may not be recognizable by those actually speaking them -- for example, by the "he" and "she" of our dialogue fragments.

But, the pursuit of descriptive economy and socially significant units of analysis -- the preference for governing a large range of phenomena with the fewest number of theoretical prepositions or for generalizations, ideally from a god's eye perspective -- tends to drive social theorists further up the abstraction ladder. If this is what is cherished, why would one be concerned with "insignificant" details, with local perspectives? Just as keys are more easy to describe than what they open, so do signifiers provide units that are more powerful than what they signify and so do texts become preferable to how people actually use them. This is what invites theorists of power to theorize about huge bodies of texts, discourse in general, reproduced representations, leading technologies, enduring institutions or such super-individual abstractions as social classes, the state or culture for the great many phenomena they presumably command. These "bodies" of abstractions are not only far removed from everyday life, their unitization is justified by the preferred theories and covering laws whose claimed generality or descriptive governance already is a form of power, thus creating units that support what power theorists are predisposed to see and thereafter claim to "discover."

Second move, systematization: (2) Give an exhaustive and coherent account of the connections between these units, including, over time -- invent and formulate relationships, rules of conduct, grammars of discourse, systems of signifiers, (inter-)textualities, information economies, communication networks, regimes of différance, knowledge structures, structuring structures and so forth -- that simulate or explain and in the end obviate the need to examine how these units are embodied.

Power relations that can be constructed to hold between abstract units necessarily are abstract as well. Not that they are therefore vacuous, their empirical justification derives from these units' "roots" in the dialogical nature of languaging, ultimately, from what "he" and "she" say to each other and how they feel burdened in their relationship. In Dahl's (1957, pp.202-203) definition, "A has power over B to the extent that he can get B to do something that B would not otherwise do;" the relation between A and B can still be put to the test by asking people, A and B, what they say to each other, how this directs their action and why. But in the hands of most theorists of power, this root is cut off to the point of no longer being visible. For example, while Foucault always stresses the relational character of power, the relationships he constructs do not account for how speakers are being-with each other, what they feel they can or would not say, but they account for how one action determines another, leaving the human actor without agency, without voice and without a place. In his (1982, p.220) terms, "What defines a relationship of power is that it is a mode of action which does not act directly and immediately on others. Instead it acts upon their actions." Acknowledging the complex and dynamic character of social life, in Foucault's writing, power relations naturally turn into "networks of power," "technologies of power," "economies of power," "systems of power," "regimes of power," "hegemonies," and into virtually autonomous power fields:

Power must be understood in the first instance as the multiplicity of force relations imminent in the sphere in which they operate and in which they constitute their own organization; ... as the support which these force relations find in one another, thus forming a chain or a system;... and lastly as the strategies in which they take effect, whose general design or institutional crystallization is embodied in the state apparatus, in the formulation of the law, in the various social hegemonies (1978, p.92-24; my emphasis). Here, the unquestioned complexity and dynamism (two attributes that merely acknowledge descriptive difficulties) of everyday life is being mapped, as it were, into a perhaps equally complex construction of power, rendering their embodiment invisible.

Just as in Marx and Engels' criticism of idealism, such systems unquestionably (a) are the inventions of experts (sociologists, political economists, for example), (b) describe the empirical connections between abstract units -- for example, "networks" of "force relations" -- that can substitute for these units' embodiment in the languaging among real people because of their exhaustiveness and coherence, and (c) exist in the discourse of these experts' community, not in the dialogical practices they claim to describe. Computer programs, which must be exhaustive in detail and coherent in form, are the celebrated exemplars of such theories. Both Saussurian and Chomskyan linguistics incorporate these two moves into their methodology -- Saussure (1960) in seeking to account for langue as opposed to parole and Chomsky (1965) in being interested in competence rather than performance. Both account for language as an autonomous and ideally computable system -- and so does systems theory, especially in sociology, structuralism, the recent cognitivism, and other scientific practices aimed at mathematical modelling, for example.

Third move, mythologization: (3) Find a metaphor that not only asserts the system's autonomy, self-organization or self-determination but also (a) dissociates it from its creators (absolve their inventors from all responsibilities for having created them, for example, by taking their accounts as objective representations) and (b) severs it from its embodiments -- in the case of power, from the feeling of being locked into a dialogical trap.

Marx and Engels criticized their scholarly opponents for attributing personhood (we now call it agency), self-consciousness, historical consciousness, or will to "regimes of ideas" and accused them of practicing a kind of animism. I am suggesting that "regimes of power" arise from similar attributions but are now to an even larger measure fueled by a physical metaphor.

In classical physics, "power" denotes a non-dimensional quantity that expresses the rate at which energy is transformed into mechanical work. Thermodynamic laws depict energy as flowing one way, from a source (of high levels of energy) to a sink (of low levels of energy). What appears in its path slows down the process, but creates nothing of its own. In physics, power and resistance are always equal in magnitude, but opposite in direction. Power is active. Resistance is passive. Physics has formulated its calculus of power to be rationally deployable. The design of mechanical systems fueled by energy, technologies of control, are logical consequences of such world constructions.

It is truly amazing to observe how social scientists, despite their overt commitments to using literal language, have uncritically embraced this physical metaphor. John Searle describes how speech acts affect what people do to each other by attributing illocutionary or perlocutionary forces to them. Since physical force becomes evident only where resisted, the metaphor makes it natural to confuse power and force and speak of power and resistance in one breath. For example, for Foucault (1980, p.142) "there are no relations of power without resistances" and, explaining these "power relationships" in considerable length, he concludes that "(t)heir existence depends on a multiplicity of points of resistance" (1978, p.95). Max Weber ([1922],1957, p.152), far less abstracting than Foucault, defines "power (a)s the probability that one actor within a social relationship will be in a position to carry out his own will, despite resistance, regardless of the basis on which this probability rests." There are numerous examples of this stereotypical association.

Note in the above that the word "basis" invites elaborations. And in the probably unconscious effort to preserve metaphorical coherence, it is then quite natural for theorists to look for -- or better said "invent" -- energy analogues as the social resources that are then promptly made responsible for creating this power. For example, Ellen Langer (1983, p.15) defines

Power (a)s the ability to affect tangible outcomes whether that power is achieved through accumulation of money, property, muscles, beauty, or knowledge and whether the effect is for self or others. As those commodities increase, one's power proportionally increase. People vary in their desire and need for power... . To be fair to other conceptions of power, several social theorists, such as Talcott Parsons, Karl Deutsch, and Niklas Luhmann describe power as a kind of currency that circulates within an entire society, accumulates unevenly, can be selectively transformed into work and has, hence, transformative capacity. Such a zero-sum notion is also found in Barthes (1982, p.460) when he describes power as being everywhere and "perpetual in historical time. Exhausted, defeated here it appears there; it never disappears. Make a revolution to destroy it, power will immediately revive and flourish again." Here one can see economic metaphors entering the discourse and constructing equally abstract and autonomous systems that are then claimed to be governing everyone's life.

Either metaphor not only disconnects the abstractly constructed "networks of force relations," "laws," "structures," or "systems" from everyday languaging, but also renders them as "autonomous mechanisms," as "regimes of power," as super-individual "social hegemonies" that have "their own organization" and "operate" purposively or "strategically" without any participation of people. These quite remarkable mythologies are constructed so that nobody, not even the theorist, can be held responsibility for them. Moreover, these regimes, being considered natural or physical, and hence outside human control -- Barthes (1982, p.460) speaks of "trans-social organism(s)" -- are overwhelmingly powerful and allow no escape. Referring back to power as the feeling of being dialogically entrapped, it is these theories of power, when taken seriously, that do the entrapping and create the power they innocently claim to account for. Myth, according to Barthes (1973, pp.109-158) is a second-order type speech that naturalizes its object and renders it autonomous and uncontestable. Clearly, it is the physical metaphor that energizes the mythical regimes of power described in the above discourses on power.

Re-Embodying "Power"

What can we do to overcome the phenomenon of power as described above? I am seriously suggesting, perhaps even as a general strategy of continuous liberation, to undo or reverse the moves that I suggest has brought power to its obscure reign. In our case, first, by distinguishing between physical power and its metaphor and treating each in a manner appropriate to their embodiments, the former in nature (as described in physics), the latter in languaging (as practiced in everyday life). Second, by de-abstracting, deconstructing the abstract and generalized notion of systems of power, allowing already existing everyday vocabulary to draw finer and individually more meaningful distinctions without imposing a systematicity other than what our situated language affords, honoring loose family resemblances at the most. And third, by getting closer to and ideally participating in the actual languaging in which these vocabularies have made their homes. This, of course is literally impossible for any one-way writing and the whole project is also far too ambitious to be undertaken here. So, I will have to be sketchy here.

(i) The first counter move has to be de-mythologization. Here, de-mythologizing means distinguishing clearly between physical power, I propose to call it "force," and its metaphorical use. The need for this distinction rests on the contention that the confusion between the two is largely responsible for the mythical character power has acquired.

In physics, force and resistance are simultaneous and occur without any human intervention, for example, when two cars collide, when one person beats a defenseless Other, or when people are on a roller coaster ride. In these examples, force appears as a strictly causal phenomena and any agency that those affected might possess is irrelevant to how the process unfolds. Indeed, once released, a bullet proceeds along its trajectory until it hits something or someone, regardless of anyone's intentions. When an agent A is said to exercise force over agent B, from the perspective of physics, any understanding B might have about A's intentions -- about the physics of A's actions or about the social relationship between A and B that prompts A's use of forces -- can play absolutely no role in what happens. According to this notion, force can neither be circumvented nor make agents do anything in particular, but it could impair or even destroy their future agency.

Based on this clarification, I am proposing the obvious: that we restrict physicalistic explanations to occurrences of physical force and open up the remainder -- let me provisionally call this "power" -- to social, relational, contextual, and local articulations. Accordingly, force would then designate a strictly causal phenomenon, one that becomes evident when energy flows meet resistances and one that is explainable in terms of thermodynamics or mechanics. By contrast, power depends on distinctions drawn in language, on its conception or on the constructions enacted by those affected. The feeling of being locked into a burdensome way of dialoguing is not describable in terms of physics. Neither are the contestations of dialogical practices that outsiders may observe in the articulations of burdens. Physical metaphors of force are, first of all, metaphors and occur -- not in an unattended nature -- but in the dialogical process of languaging. This would distinguish between terrorist threats, for example, and what happens when a bomb explodes, or between fear or its articulation and actual death or bodily injury. Thus, power and force operate in two incommensurate discourses. Power is relational, social and mostly negotiable, force is neither. People can liberate themselves from power, at least in principle, but not from force.

Why is this distinction so important? To give a simple answer, let me reiterate: An understanding of language does not enter the kind of world physicists conceive of, describe and presume to live in. The entailments of physical metaphors reduce power to force, often quite unbeknownst to their users, which gives the compelling impression that power is not the product of its articulation nor that it could be affected by contestations and be undone. Physicalistic explanations keep their users locked in a single uni-verse devoid of language. Within the empirical domain of physics, undoable concepts are meaningless and impossible to justify. One is reminded on the laws of preservation of matter and energy.

In contrast, metaphors surely are discursive forms. (To be clear, physics too is essentially a discursive phenomenon, but it defines itself by denying this to its empirical domain). As such, the appropriateness of metaphors is contestable and the entailments of these metaphors are examinable and hence undoable. Yet, as we have seen in the previous paragraph, physical metaphors lure their users into the very domain of physical explanations that are predicated on not reflecting their linguistic nature. It follows that physical metaphors conceal their own workings, prevent their users from becoming aware of their own dialogical involvements, protect their entailments from contestations and thus keep the mythical nature of power alive. It is because of this thorough blinding of naturalist social theoreticians that "power" remains so powerful.

Hannah Arendt (1970) draws a similar distinction between violence and power, characterizing violence as a consequence of non-consensual political acts, such as torture and imprisonment. While torture clearly involves force, there are non-violent effects of force as well, in sports for example. There also are symbolic uses of force, ethnic cleansing, for example. The virtue of her distinction lies in her ability to identify misuses of power, misuse being given the a priori meaning of "without consent." C. Wright Mills (1956; 1959, p.100) applies a definition of power whose virtue is to hold "men of power" responsible for any specific courses of events they initiate. Both conceptions are definitionally tied to acts as originating in agents, leaving their targets as passive respondents, (in Arendt's case reduced to the granting or not granting of consent). The virtue of my distinction is that it puts power in the dialogical processes of languaging, where articulations are created, contested or acquiesced, where responsibilities are assumed, rejected or assigned and where undoing but also reification can take place. Liberation is a way of talking into (not about) power. It is not an application of force.

(ii) The second counter move consists of de-systematizing, recovering ordinary language distinctions. Having separated power from force, and thereby opened the possibility of treating the physical metaphor as a metaphor (that is, no longer confusing it with its ontological entailments), we now can remove the abstract and externally imposed system, logic or network of relationships in terms of which power tends to be described and revert back to the language in which the distinctions have been drawn. Even our only provisionally adopted conception of power, while important to redirect our attention to actual contestations of dialogical entrapments, may still be too general to be of use when it comes to understanding what real people do and feel toward each other. Indeed, we must ask ourselves, why should we impoverish our social science discourse just for the sake of pursuing abstract generalities when even the most ordinary folks -- the people we are involved with or observe in their effort to oppose unjust authorities, confront illegitimate regimes or expose the oppressive language of their employers -- have rich vocabularies available to distinguish the many ways relationships can go sour? Let me mention a few familiar examples. Each of these have their own histories, their own logic, and require, hence, their own undoing strategies. The point of the following accounts is not to be exhaustive, but to give a sense of the variety of phenomena that no general theory of power can summarize without at the same time denying their undoability.

There are numerous terms attributing to someone an ability or capacity to get Others to do something for them, even those in positions of authority: might, vitality, vigor, strength, enthusiasm and so forth. I already cited Langer's list of resources for this ability to which one could add other appreciable characteristics such as supportiveness, political connections, values and so forth, most of which have polar opposites describing an incapacity or lack. Even helplessness, vulnerability or sickness can move people to act -- none of which "exerts a force."

There are many common metaphors of painful imprisonment, being constrained, confined, forced against ones will, incapacitated, held back, shut in, or facing insurmountable barriers or obstacles due to circumstances outside one's control. I relied on this metaphorical complex to access the feeling of being locked into a way of seeing or into a way of dialoguing to characterize power. Let me offer some diverse manifestations of these.

There is coercion, to oblige someone to do something by threatening hurt, punishment, or by predicting something harmful to happen, to command compliance by playing on fears, or to enforce a rule by an arbitrary authority -- blackmail, for example.

There is persuasion, encouraging someone to do something by reason, coaxing, urging or to influence someone's belief in something by referring to models, offering self esteem, promising association or group membership, appealing to morals or values.

There is also a whole vocabulary for allocating blame, humiliation, embarrassment, denunciation, ridicule, intimidation, degradation, ostracizing, accusing, scapegoating, criminalizing and so forth. These take their effect by being asserted in front of a community of witnesses, in public, through the mass media, for example. Related to this are induced feelings of shame (causing one to wish one could disappear from view) or of guilt (the feeling of having done something wrong).

There is the structural silencing of voices, the marginalizing if not ignoring whole classes of otherwise available phenomena, such as people, ideas or ways of life.

There is the vocabulary of social diminution which creates dialogical inequalities: superiority/inferiority, competence/incompetence, worthiness/worthlessness, leader/follower, being qualified/not being good enough, winner/loser, beautiful/ugly, attractive/not loveable and so forth. These polarities can be taught and are learned, but they could also be opposed and neutralized.

There are the notions of alienation, exclusion and isolation from a community: the feeling of being distraught, dispossessed, disqualified, disenfranchised, disaffected, dislocated, being shunned and so forth, whose positive complements are connectedness, being-with, synergy, solidarity, sharing, participation. Barry Barnes's (1988) notion of power focuses on power in the public sphere, the strength derived from belonging to a group and the weakness of non-belongingness.

There is domination, akin to the German "Macht" as distinct from "Kraft" which is roughly strength or force. The English "power" lumps these two together. Domination requires of people first of all to grant select Others the role of leaders, of popular heroes, of authorities, of being more important or more capable than they feel they are. This not only legitimates these Others' actions on account of the "positions" they thereby come to hold but also entails, secondly, submission by those that have put them into these positions.

There is hegemony, which is a domination, not by individuals, but by collectivities, majorities, institutions, traditions, the state or one state within a federation of states. All of these derive their reality through collective (widely shared and valued) commitments to rather abstract conceptualizations of systems: ideologies, conventions, social organizations, governments. All of these become manifest in certain valued and often routinely executed interactive practices. Abstractions have no powers per se, but believers typically are entrapped in them.

And then there is the nourishing guidance by parents, mentors, coaches and teachers, that make the development of competence of one party a joint accomplishment. This is the kind of power that feminist theory has recently focused on and identified as more typical of women (Ruddick, 1989; Code, 1991).

Power can also be paradoxical, as in "commanding disobedience," "forcing someone to be free," "denying one's ability to contest," "encouraging the very power one complains about," "causing someone to see possibilities" -- which underlies our re-vision demonstration and last but not least, "the power to undo power" which gave this essay its title. Paradoxes are unsettling vacillations between mutually exclusive understandings and known to invite re-visions.

Obviously, this collage of concepts of burdensome dialogical situations is just the tip of an iceberg that is kept afloat by astonishingly rich vocabularies of less visible forms, unhappy relationships or power plays, all of which can draw one into a way of being-with others which are difficult to leave (Stolzenberg, 1978). Each of these relational concepts/feelings deserve their own theory and call for different dialogical involvements. They can be considered different language games. Our albeit cursory understanding of these common sense concepts questions the value of social theories that put such things as alienation, terrorist threats, acquired incompetence and humiliation in one bag labeled "power." The only thing that this vocabulary of power may have in common is the feeling of awe, a peculiarly paradoxical emotion combining fascination and fear. Taking this emotion as causing ones reactions or as the stimulus for ones responses is precisely what prevents its undoing because its causal=monological construction constitutes a trap that is not easy to see from its inside.

(iii) The third counter move calls for de-abstracting which means locating (the objectively removed) concepts in dialogically embodied practices and thus preparing to re-enter dialogue. Beyond recovering meaningful distinctions, I want to take just four kinds of "power" from the above collection and show the dialogical "grammar" or the interactive "logic" of their respective workings and explore the languaging that could actually undo each one of them.

As a first example, I refer to Lakoff's male/female dialogue we already discussed. With physical metaphors in mind, we would seek to explain why he appears to get his way by asking what he has that she is lacking. If she takes power to be an entity, one can see how she must blame him for possessing something that she must also deny having. But if one asks what it in fact is that flows from here to there one would be lead to quite absurd answers which shows the ridiculousness of such constructions. But this might not matter much when the feeling of being entrapped is confused with the entrapping explanation.

From a relational stance, one might say that his controlling and her compliance complement each other. Observing the process rather that its outcome raises the question of punctuation, whether anyone, and if so who, initiated the sequence. However, probably the fairest answer is that punctuation is arbitrary yet drives each interlocutors' behavior and forms a dialogical trap from within which she may see herself as letting him set the parameters for her being. Clearly, submission need not be seen as an adaptive response to the exertion of power. It may well be regarded as what invites power to arise (Krippendorff, 1989b). Because we are so primed to blame "powerful" Others for all our mishaps, this reconceptualization might look like "blaming the victim." However, the issue is not to fix blame but to construct a lever by which one can bootstrap oneself out of the feeling of entrapment, to assume a position from which one can see possibilities for contestation, to shape the relationship one finds burdensome. Undoing starts by recognizing one's complicity in a relationship of power. Denying this possibility is precisely what constitutes the feeling of being locked into a disabling situation, into confusing power with force, into defining one's situation as objectively unalterable, into having been and (therefore) continuing to be the target of Others' powers.

By examining the dialogical role of her submission, she can be led to her own way of languaging, her using a vocabulary which constructs a universe that enables him to assume the very role she finds objectionable. In Lakoff's first dialogical example, it was her deference that invited him to take charge. Complaining about such situations without commitment to act expresses the very helplessness that seals this dialogical trap. This kind of helplessness is surely acquired -- not, however, by accommodating to objectively existing conditions, but by preserving one's conception of relatedness or one's institutional involvement, not wanting to risk changes. This condition can be re-visioned, re-articulated and overcome, but not without realizing ones own conceptualization as the reason for the burdensome relationship and conceiving the reconstructability of ones relation to Others and oneself in it. Surely, this form of power with its numerous variations arises in language and can be undone by languaging oneself out of the entrapment in which it resides.

Let me take coercion, as a second example. Its logic is understood far better than the previous one. Suppose a mother tells her child "if you do this again you are grounded for a week." With this simple threat, the mother (a) commits herself to respond to a child's possible action. She assumes that (b) the child fears the consequences of her response, that (c) this fear entails a constraint on the actions otherwise available to the child and that (d) the child desires the actions in question. The mother can control the child's behavior only if the child accepts the mother's reality construction in all of these four respects.

Suppose the child knows that the mother has no interest in even having the child around for the week, that she can not monitor whether she stays at home or not, or has never made good on similar injunctions before, then the threat has no effect. Suppose the child wants to stay home for a week, then the punishment turns into a plus and the threat no longer exists. Suppose the child is no longer interested in doing what the mother wants her not to do, then the mother may think her threat has succeeded, but the child was not affected at all and the mother's power has in fact been rendered quite illusionary on account of the child's reality.

The point of going through these conditions is to show that even simple threats always come at the end of a history of languaging. Isolated speech acts mean little by themselves. And for threats to be effective they require that the coerced party has constructed and holds on to a reality in which the particular statement becomes a threat. This construction must also be known to or assumed by the coercing party. Obviously, a considerable conceptual collaboration is required for something to be a threat -- not to belabor the possibility that someone may be threatened by something not intended to coerce. Force is inevitably causal, but threats can never be explained that way. Undoing a threat boils down to constructing and shifting to a universe in which the conditions of the coercer are not met. Someone who demonstrates not to understand what the threat means, for example, by joking about its seriousness can not be made to do anything under a threat. When Crocodile Dandy, in the movie by this name, engages the thugs that tried to rob him in a conversation about delivering their threat more effectively, he manages to shift to a different dialogical universe in which the threat no longer exists. People who happen to see the supposed punishment as a virtue, like martyrs who conceive being rewarded after dying for a larger cause, can hardly be threatened, not even by the prospect of losing their life. Someone who can claim not to have any control over his or her actions, like the pilot who convinced the hijacker that his airplane was controlled from the ground, can not be made to do anything different from what he or she already does. Someone able to create options more favorable than those a threat intends to limit has immunized him or herself to that threat. For example, an employee who has attractive job offers in his pocket hardly fears being fired. Terrorists, kamikaze pilots, and assassins are a problem for society largely because their idiosyncratic rationality immunizes them against coercive efforts. Thus, this simple speech act belies the complexities involved but is completely undoable by constructing as variable a reality that the threatening party assumes fixed.

I take domination as my third example. In extension of the above, domination can be said to result from being burdened by an authority that is seen as both illegitimate and unquestionable. To be sure, not all authority is burdensome. To have access to a good expert, mentor, coach, teacher or parent is a blessing, not a curse. The images of domination that come to mind are tyrannical rulers, ruthless bosses, mean gang leaders, unfeeling parents. It is remarkable that we first think of people occupying a superior position, people we can blame for our feelings, all the while ignoring that authority always resides in a complementary relationship. Authorities do not exist without loyal or obedient subjects. If this is so, disloyalty and disobedience will erase their ascribed capabilities.

We know of three justified reasons for disloyalty and disobedience: when a higher authority leads a subject to act differently, when a subject is commanded to do something outside the legitimate range of the commanding authority, and when the history of acquiring the commanding authority is no longer considered legitimate or acceptable (which includes being an unjust burden). The Nürnberg laws, by which Nazi crimes against humanity were judged, appealed to an authority higher than that governing Germany at the time and affirmed the first reason. Convicting public officials who demand favors for services or bosses who sexually harass their employees affirms the second reason. Annulling a fraudulent election or persecuting a practicing physician without a MD affirms the third. While legal proceedings can justify disobedience to illegitimate authority, I do not want to belittle the difficulty of undoing domination in everyday life. The existence of child abuse, mafia style protection rackets or even government lobbying exemplify that domination is difficult to get rid of precisely when it is confounded with (a history of) benefits, loyalties and solidarities. Gregory Bateson's (1972, pp.242-249) double-bind shows how this works in families. Nevertheless, since authority is created and legitimated in dialogue, usually before it turns illegitimate, unjust or burdensome, it is never immune to contestations. The mass media are particularly good at making or breaking public heroes whether by celebrating accomplishments or interpreting personal histories.

Undoing domination also starts with recognizing one's complicity in the unquestionableness of a burdensome authority, by daring to articulate the previously taken for granted support and to thus reconstruct the reality in which it grew. Revolutionaries that adopt physical metaphors for domination are naturally led to the conclusion that unjust authorities must be toppled by force greater than they command. However, history tells us that domination may also be eroded by withdrawing the kind of loyalties, obedience and admiration on which such authorities depend by either invoking higher authorities, referring to moral principles, for example, or conceptually side stepping the resources along which obedience was commanded. The non-violent movements of Mahatma Ghandi and Martin Luther King are two well known examples of undoing domination. Strategies for coming out of relational entrapments in everyday life are not much different. All authorities, from national leaders to playing the boss in the family, face the possibility that their followers, supporters, subjects withdraw their support and move their concerns to areas in which a dominant authority has no answer or no rights to make demands. As we can see, undoing domination follows a logic very different from defying threats or overcoming male-female power relations.

Finally, let me talk into the "power of language" which is recognizably woven into all manifestations of power. Some writers see natural language as a constraint and, consequently, as a collective "instrument of oppression" from which no speaker of that language can escape. I contend that this harsh judgement derives from conceiving language as a disembodied system that is abstracted from processes of languaging. For example, for Barthes,

(l)anguage is legislation, speech is its code. We do not see the power which is in speech because we forget that all speech is a classification, and that all classifications are oppressive: ... In French ... I am obliged to posit myself first as subject before stating the action which will henceforth be no more than my attribute: what I do is merely the consequence and consecution of what I am. ... I must always choose between masculine and feminine, for the neuter and the dual are forbidden me. Further, I must indicate my relation to the other person by resorting to either tu or vous; social or affective suspension is denied me. Thus, by its very structure my language implies an inevitable relation of alienation. ... (L)anguage -- the performance of a language system -- is neither reactionary nor progressive; it is quite simply fascist; for fascism does not prevent speech, it compels speech (1982, p.461). His conclusion that language is fascist is quite extraordinary and has created much discussion in France and alarmed semioticians like Umberto Eco (1986, pp.239-255).

Barthes' statement logically follows from the popular understanding of language as constituted by unalterable rules, as being autonomous, as having a super-individual structure, and from a notion of communication as the "performance of a system" that users need to comprehend. This conclusion is based neither on observations, nor on experiences and it is certainly not derived in view of the possibility of undoing a phenomenon. To be fair, Barthes does recognize the possibility of linguistic change but describes it as a "cheating of language" which confirms his Cartesian conception of language as an autonomous system.

Obviously, we are born into already ongoing processes of languaging with Others, but it is only when we conceive of (or submit to) language as a fixed system of rules, as codified, formally taught and enforced by appropriate institutions that we limit our creativity, including the possibility of constructing ourselves anew. Through our belief in its normative stability we see language as making us male or female. In our using common subject-predicate-object constructions we render us as actors that manipulate things. Through the convenience of using binary opposites we seduce each other into dualist world constructions. By relying on a conception of language whose facility of pointing to and of representing things has been so thoroughly institutionalized by linguists and formalized by logicians we silence the alternatives, rule out of existence the illogical if not dialogical, discriminate against what can not be stated clearly, the emotional, the fantastic, the construction and the undoing of phenomena. The convenience of prepositional categories directs us to distinguish between "power-of, -over, -with and -to," attribute power-of to objects, power-over and power-with to Others, and power-to to instrumental acts -- the latter being often equated with "empowerment." I suggested in the above that scientific theories of power unwittingly ride the wave of physical metaphors which is a good example of the unawareness of languaging, the blindness to dialogical processes and the inability to regard language as embodied, as a process of constructing perceptions among others. If one wants to pursue something like a power of language, it seems to arise when we let us being spoken by language, not recognizing the language game we are playing. This is to say, that we willingly confine ourselves to the linguists' language game that Barthes (1982, pp.460-461) might indeed be enacting when saying "language is legislation...a generalized reaction...(it) has no exterior: there is no exit." But this is just a construction by linguists motivated solely by the need to create an institutionalizable object of scientific research. It is produced by what Marx and Engels identified as a "trick" through which we lose the connection with ordinary languaging. This power of language can be undone when we take the responsibility for creating us in dialogue.

Much of this essay addressed the possibility of creating many language games, of re-articulating one's being-with Others, of re-visioning perceptions and of constructing and moving between alternative uni-verses, even in everyday languaging. Gambattista Vico saw the creation of new forms through metaphors. Non-establishment (avant-garde) artists have always taken the liberty to ignore old forms of expression, forms that have become obsolete, intolerable or oppressive, by simply creating new ones. And Richard Rorty (1989) considers languaging as a process in which new forms continuously kill off old ones, assigning "strong poets" the task of creating new vocabularies and of re-articulating new realities. We can not blame speech and writing as the source of our problems with power, as Barthes does (just as we can not blame the figure for causing our seeing) when it is we who adopt a disembodied system called language that linguists, administrators and educators have constructed for their purposes, when it is we who in speaking have the choice of taking liberties with its structure or going by the books.

Undoing power, getting out of dialogical entrapments is a challenge that individuals have to create for themselves. Artists know it. Therapees come to realize it. Revolutionaries do it. Avoiding certain constructions that apparently limit us -- like physical metaphors for power -- is the seed that could alter languaging in the long run. Dialogical possibilities always are far greater than we can realize in our individual articulations. Unfortunately, we can experience these possibilities only through another language, in retrospect or through the eyes of Others we are being-with, hence the imperative to always seek conditions for these possibilities to arise and to explore them in practice.

Summary and Conclusion for Critical Theory

In sharp contrast to our common and variously institutionalized understanding of language as a medium for representing something detached from it, as having form and content and a super-individual nature separate from our lives, we started our journey by looking at an image. Languaging with it, your reading of my writing, showed you the possibility of being guided in language out of the otherwise unnoticeable experience of being locked into an exclusive way of seeing. It enabled you to experience the possibility of quite radical re-visions. It made you aware of how your seeing is brought forth in language and your ability to shift to different uni-verses of being-with. Evidently, such re-visions are not unusual in ordinary languaging. They involve very real and bodily experiences and demonstrate the reality of multi-verses (not merely different versions of or perspectives on one) in which we can make ourselves appear in rather different roles.

It is our ontological commitment to a single uni-verse that blinds us to experience them or dismiss these experiences as illusionary.

I then presented to you a collection of uses of the word "power," mostly from academic writing. The variety of contexts of its use is astonishing indeed. The only commonality might be the feeling of awe: fascination and fear. Since this ambiguous emotion may not be a good starting point for attending to the phenomenon of power, I suggested the notion of undoability from which emerged a concept of power that, on the one hand, is embodied in feelings of being entrapped and, on the other hand, surfaces dialogically in actual contestations. This removes the concept of power from the domain of physics and brings it into that of languaging and involves speakers and listeners as constituents of the process. Since power can not arise outside the language that makes it threatening or enabling, I argue, it can also be undone in language and through appropriate re-visions or re-articulations.

In contrast, and without insisting on sweeping generalities, contemporary articulations of power, especially social science theories thereof, seem to move in the opposite direction, into an abstract and monological objectivism, governed by physical metaphors. Its theoreticians seem blatantly unaware of how their own language virtually creates the abstract, detached, mechanistic, and mythically empowered objects of their research. The analogy of being locked into one way of seeing is apparent and this applies to the ontology of social research as well. Marx and Engels analysis of another epistemological pathology aided this conclusion regarding contemporary social theory.

Reversing the steps of this analysis enabled us, (i), to de-mythologize power as a self-sustaining system that commands submission; (ii), to de-systematize the abstract if not obscure regimes of power relations into the multiplicity of concepts that actually do occur in everyday articulations; and (iii), to de-abstract these concepts by re-entering them, here only in writing, into the very dialogue in which they arise and making it thus possible to re-articulate there the entrapments that previously could only be felt, judged and reacted to. This kind of re-embodiment would speak against general theories of power and, instead, for rather specific and locally practicable emancipatory or therapeutic articulations.

Undoing power relationships, I suggest, can not be caused from outside their embodiments, for example by usurping the powers Others seem to command or by pleading with them to relinquish their position of influence. After all, feelings of dialogical entrapment happens in being-with Others. Feelings are neither mysterious, systematic nor abstract. However, the lessons of our examples go beyond what the three counter moves can accomplish, that is, bringing power to where it is felt and exercised. Let me summarize these lessons in the form of four additional moves that, together with the above, may provide critical scholars "a blueprint for processes of liberation:"

(iv) Cognizing/admitting participation. Emancipatory efforts presuppose that one recognizes ones dialogical involvement in burdensome relationships. We always are in complicity with our dialogical entrapments, usually far more than we seem to be willing to see, at least initially.

(v) Contesting such entrapments brings power into the very languaging responsible for it, to the surface, as we said. Merely complaining is at best a ritual form of denying (iv), which makes us into subjects, villains or martyrs and draws us deeper into the very entrapment we complain of. Typically, this self-disablement has benefits, not only costs, and these go into the very construction of a trap.

(vi) Re-articulating and re-visioning relationships or participating in the construction of new ways of being-with Others opens the possibility to stepping out of a contested relationship and into new dialogical involvements. This is the creative component of the blueprint.

(vii) Enabling Others. In emancipatory dialogue people are neither alone nor can they be in charge. Re-articulations of individual re-visions will always be meaningful to Others but not necessarily as intended. Since all forms of power imply togetherness, are coordinations of actions, re-articulations have to offer dialogical possibilities for Others as well. Since power arises in social relationships, emancipation from it must not deny the possibility of Others to go on. It has to acknowledge the being-with Others.

What have we learned along this journey? How should we as communication scholars read, write, listen to and speak social science discourse? Clearly, "objective" and "accurate" accounts of existing power relations, using corroborating evidence in the form of statistical inequalities, whether by means of content-, discourse-, or conversation-analyses, can hardly be our aim. It would merely objectify and stabilizes what may have to be questioned and undone. Nor would it be ethical to pursue a hermeneutics of suspicion that proofs the reality of unwanted sources of constraints or hidden connections only we can see in the taken-for-granted world of everyday life. Had I so construed my task, I would have ended up, as many critical scholars do, complaining, but thereby also reifying the very prominent physical metaphor that reproduces the burdensome nature of power. It would have demonstrated my being locked into a self-concealing construction without an escape route. My deviation from these traditional paths of social inquiry has, I hope, demonstrated that it is simply not enough -- or, said more strongly -- positively self-defeating for us to criticize institutions for their evil practices and to thereby solidify the very powers we criticize. Equally questionable is the practice of critical theorists who invent utopias or ideal (speech) situations and then proceed to "reveal" the hidden obstacles to achieving them. The absence of dialogical involvements in these strategies of criticism keeps social theorists detached from experiencing power and prevents their theories from being contested within where they are embodied or could be realized in practice. Good intentions are far too easily and unwittingly undermined or limited by buying into modernist validity criteria of truth (celebrating accurate descriptions, predictability and control) or by seeking widespread consensus (cherishing inter-subjective agreement among experts) that marginalizes those who suffer from its "facts."

Criticizing is one thing and undoing is something quite different. I am suggesting that criticism be deliberately constructive of new realities in which people could live and conceivably find each other. Given the possibilities of re-visions and re-articulations languaging enables, given the multi-verses new vocabularies can bring forth, given that critical social theorizing too is a way of languaging, critical theory would do best to go beyond criticizing existing conditions from a self serving monological perspective and propose alternative versions of languaging that are both contestable nor undoable. We, as critical scholars, should simply set ourselves the task to:

(1) continuously invent new possibilities of being-with Others in language or in dialogue,

(2) critically examine these inventions for their realism, for their practicability, acknowledging our own involvement and

(3) remain accountable to those who populate our constructions or who could find each other in them. This is what I hope this essay illustrated.

With (1) I am proposing that critical scholarship commit itself to the continuous creation of new vocabularies capable of drawing distinctions heretofore unavailable, to the invention of new ways of languaging that can realize social relationships heretofore inconceivable, to keep alive the undoability of (not necessarily the undoing of) the very institutions that modernist vocabularies implicitly conserve, enable and predict (with theories that preclude human involvement), and to oppose temptations to construct universalizing theories whose inherent imperialism discourages local understanding and diversity.

To acknowledge the self-reflexive and socially constructive process of social science discourse, with (3) I am proposing that critical scholarship be accountable to those to whose lives it addresses by which I mean that its claims must make sense, be re-articulable and potentially contestable by its constituents, particularly by those talked-for, talked-of or affected-by this scholarship.

To keep these possibilities alive in practice, with (2) I am proposing that critical scholarship consider itself validated to the extent it actually does open new (as opposed to closing existing) possibilities for Others to dialogically construct their own worlds and feel at home in them. Adopting "possibilities for Others" as a new validity criterion would keep critical scholarship socially responsive, dialogically enabling, individually emancipatory and embodied in social processes.

This proposal could give the social sciences a new life and critical scholarship a new justification for being.


Ackerman, Bruce A. (1980). Social justice in the liberal state. New Haven, CT: Yale University Press.

Arendt, Hannah (1970). On violence. New York: Harcourt Brace Jovanovich.

Barthes, Roland (1982). Inaugural lecture, College de France. In Susan Sontag (Ed.), A Barthes reader (pp.457-478). New York: Hill and Wang.

Barthes, Roland (1987). Mythologies. London: Granada (Paladin Books).

Bateson, Gregory (1972). Steps to an ecology of mind. New York: Chandler.

Bernstein, Richard J. (1992). The new constellation. Cambridge, MA: MIT Press.

Chomsky, Noam (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.

Code, Lorraine (1991). What can she know? Ithaca, NY: Cornell University Press.

Connolly, William E. (1983). The terms of political discourse. Princeton, NJ: Princeton University Press.

Dahl, Robert A. (1957). On the concept of power. Behavioral Science, 2:201-215.

Dreyfus, Hubert L. (1991). Being-in-the-world. Cambridge MA: MIT Press.

Eco, Umberto (1986). Language, power, force. In Umberto Eco, Travels in hyper reality (pp.239-255). New York: Harcourt Brace Jovanovich

Ernst, Bruno (1986). Het begoochelde Oog; onmogelijke en meerzinnige Figuren. Amsterdam: Bruno Ernst and Meulenhoff Nederland bv.

Foucault, Michel (1978). The history of sexuality. Vol 1: An introduction. New York: Pantheon.

Foucault, Michel (1980). Power/knowledge. Colin Gordon (Ed.). New York: Pantheon.

Foucault, Michel (1982). The afterword: The subject and power. In Hubert L. Dreyfus and Paul Rabinow (Eds.), Michel Foucault, beyond structuralism and hermeneutics (2nd ed.)(pp.208-226). Chicago, IL: University of Chicago Press.

Gallie, W. B. (1962). Essentially contested concepts. In Max Black (Ed.), The importance of language (pp.121-146). Englewood Cliffs, NJ: Prentice Hall [Originally (1955-56), Proceedings of the Aristotelian Society,56,167-190].

Giddens, Anthony (1979). Central problems in social theory: Action, structure, and contradiction in social analysis. Berkeley and Los Angeles: University of California Press.

Harris, Roy (1981). The language myth. London: Duckworth.

Krippendorff, Klaus (1989a). On the ethics of constructing communication. In Brenda Dervin, Larry Grossberg, Barbara J. O'Keefe, and Ellen Wartella (Eds.), Rethinking communication; Vol. 1, Paradigm issues (pp.66-96). Newbury Park, CA: Sage Publications.

Krippendorff, Klaus (1989b). The power of communication and the communication of power: Toward an emancipatory theory of communication. Communication, 12(3):175-196.

Krippendorff, Klaus (1993). Conversation or intellectual imperialism in comparing communication (theories). Communication Theory, 3(3):252-266.

Lakoff, Robin Tolmach (1990). Talking power. New York: Basic Books.
Langer, Ellen J. (1983). The psychology of control. Beverly Hills, CA: Sage Publications.
Lukes, Steven (1947). Power: A radical view. London: Macmillan.
Marx, Karl, & Frederick Engels (1970). The German ideology, part one. C. J. Arthur (Ed.). New York: International Publishers.
Marx, Karl, & Friedrich Engels (1978, [1845-46]). Werke (Band 3). Berlin: Dietz Verlag.
Maturana, Humberto R. and Francisco J. Varela (1987). The tree of knowledge. Boston: Shambhala.
Mills, C. Wright (1956). The power elite. New York: Oxford University Press.
Mills, C. Wright (1959). The causes of world war three. London: Secker and Warburg.
Rorty, Richard (1989). Contingency, irony and solidarity. New York: Cambridge University Press.
Ruddick, Sara (1989). Maternal thinking: toward a politics of peace. New York: Ballantine.
Saussure, Ferdinand de, (1960). Course in general linguistics. C. Bally & A. Sechehaye (Eds.). London: Routledge.
Shotter, John (1993). Conversational Realities; Constructing Life through Language. London: Sage Publications.
Stolzenberg, Gabriel (1978). Can an inquiry into the foundations of mathematics tell us anything interesting about mind? In George A. Miller &
     Elizabeth Lenneberg (Eds.), Psychology and biology of language and thought (pp.221-269). New York: Academic Press.
Wartenberg, Thomas E. (1990). The forms of power. Philadelphia, PA: Temple University Press.
Weber, Max (1957). The theory of social and economic organization. Talcott Parsons (Ed.). Glencoe, IL: Free Press [1922].
Wittgenstein, Ludwig (1958). Philosophical investigations. New York: Macmillan.
Wrong, Dennis H. (1988). Power: Its forms, bases, and uses (2nd ed.). Chicago, IL: University of Chicago Press.


1 I am aware, of course, of the stereotype depicted here. American Indians rarely run around dresses with feathers except in the stories white
   Americans tell each other about "them." The ethnic designation of the Figure is Bruno Ernst's.

2 If you happen to have started seeing the fur coated eskimo before you saw the human face, this does not distract from the demonstration
   being made here.

3 From Martin Heidegger, Basic Problems of Phenomenology. Bloomington: Indiana University Press, 1982, as cited in Dreyfus (1991).

4 Literally, "undoing" is reversing something done, an action, metaphorically, untying a knot or a woven fabric. I like its connotation of involving many strands that could be put together differently. Most of its synonyms have less desirable entailments: "Erasing" was my favorite for it undoes something said and creates a space for re-writing. But it is also tied to writing and drawing on paper. "Canceling" suggests countering an action by one of equal force. "Diffusing" means spreading something thinner and over a larger area. "Obliterating" implies covering up or smearing over something that does not disappear but is no longer recognizable for what it was. "Avoiding" amounts to circumventing or walking around without questioning its existence. "Voiding" leaves a space behind but also suggests missing it thereafter. Undoing leaves no void but the possibility of creating something new.

5 I am grateful to John Shotter who brought this passage to my attention.

6 For earlier steps see Krippendorff, 1989a, 1989b, 1993.


Undoing Power

by Klaus Krippendorff
The Annenberg School for Communication, University of Pennsylvania, Philadelphia

 This essay examines the possibility for critical scholarship to re-articulate power so as to aid its undoing in social and, hence, dialogical practices. Its reading might appear cumbersome because it can not simply speak about power as usual, as if it existed independent of people feeling its impact or using it to their advantage. Such objectivist accounts would draw the writer as well as the reader into the very trap in which power tends to be constructed as a necessary evil, as something that nobody can escape from, as something that can not be undone. Writing enablingly is difficult and perhaps not entirely possible.

The essay starts with an experiment in perception designed to make readers aware of how language is implicated in bringing forth the reality we see and of the possibility of its re-articulations.

It presents several well known articulations of power, largely from academic writing, that exhibit the awe-invoking and omnipresent nature of power, leaving nothing to be done.

It then proposes four defining conditions for undoable phenomena and applies them to power, examining the foundation of its inevitability and looking for a lever to undermine its constructions. Surprisingly, the beginnings of both are found in an old analysis of the pitfalls of idealism, which is extended here to modern objectivist account of power. The latter turns out to be the product of an epistemological "trick in three acts" that one seemingly allows being performed on oneself by remaining unaware of ones use of language in dialogical constructions reality and, hence, unable to see the mythical ontogeny of the supposedly scientific concepts of power. Reversing this "trick" is proposed as a way to recover the lost access to the multitude of manifestations of power in dialogue. And with the help of four more procedural steps it is suggested how power may be contested, re-articulated and undone.

Key to the debilitating notions of power is their reliance on physical metaphors whose entailments make any undoing inconceivable. This gives rise to the distinction between power and force. The undoing of power is demonstrated on three examples: simple threats, domination and the power of a concept of language that can keep social scientists stuck in a debilitating way of languaging.

The essay concludes with a recommendation for what critical theory could and should do: maintaining the possibility of emancipatory dialogue with Others.