Theories of power in communication A critical assessment 1

Full text

Theories of power in communication
A critical assessment1
Fredrik Engelstad
Language and communication as a precondition for power have attracted increasing attention over
the last thirty years or so. The introduction of concepts such as discursive power (Foucault 1970)
and symbolic power (Bourdieu 1991) turned attention away from traditional thinking taking
resources and positions as the main sources of power (Hindess 1996, Engelstad 1999). But the two
approaches are in no way incompatible. Even if resources and positions are retained as the basic
elements, the communicative aspect is an integral part of virtually any exercise of power, given that
power holders have to make their intentions or wishes to subordinates understood by direct or
indirect communication in order to influence their behavior.
Admittedly, such a general statement is not without significant exceptions: power may be exercised
by the use of brute force, or by directly affecting actors‘ resources or options without their
knowledge. But even in such cases indirect communication is often part of the power game. In
military operations a crucial part of the strategy is misleading the adversary about one’s own plans.
In more peaceful contexts power holders frequently do not communicate their real intentions and
long term plans, as every student of Machiavelli is aware of. Nevertheless, they have to make clear
what they want others to do, regardless of whether these others understand the further implications
of their actions.
The meaning of the concept of power is not unaffected by the change of perspective from force and
resources to communication. However, this does not necessitate forging a completely new concept
of power, but rather the enrichment of traditional notions. The core of the concept is still captured in
the very simple definition of power as “getting an actor to do something that she or he otherwise
would not have done”. This is in line with definitions by such diverse analysts as Max Weber,
Robert Dahl and Michel Foucault (Engelstad 1999). Understood this way, power has an intentional
element. The assumption of intentionality is controversial, and is challenged from two sides.
Intentional acts have unintended consequences, which in given circumstances may serve the
interests of power holders. Thus, the outcome alone is not a sufficient indication of the existence of
a power relation. In most cases this could be classified as luck, not exercise of power (Dawding
1997). More complicated, however, is the fact that communication is based on preconditions which
to a large degree are unreflected and taken for granted. These constitute aspects of what is often,
rather haphazardly, termed “structural power”. The hard question is how these preconditions enter
into actual power relations. A basic assumption in the present discussion is that this must be
understood in the light of transformations from micro level to macro level models.
Power in communication may be analyzed under three different angles. First, aesthetic means may
be used to make the communication maximally efficient. Typical examples are aesthetic means of
persuasion, such as increasing sexual attraction by elaborate clothing, tailoring advertising to elicit
desires in consumers, or political rhetoric persuading voters to support a given party. Aesthetic
means may also be used to enhance the legitimacy of personal power holders, given offices or their
incumbents. Glittering parties, impressive parades, sumptuous clothing and great buildings inspires
awe in spectators, and supports the feeling that the people who command all these resources deserve
to exercise power the way they do (Engelstad 2001, Daloz 2003).
These means–ends oriented forms of communication are important, but normally quite easily
understood. More complicated are the forms of power that are inherent in the communicative
process itself. This is the case with the second and third types. The second is what is often
characterized as the “language is power” approach (Blakar 1972), or what I would call the standard
1 Thanks to Marianne Gullestad, Marianne Nordli Hansen, Håkon Leiulfsrud, Erling Sandmo and Aagoth Storvik for
valuable comments to an earlier draft.

2
theory of power in communication: actor A influences actor B by affecting her conceptions of the
world. The power of actor A resides in getting B to accept a given description. One source of
inspiration to the standard theory was the political movements of the 1960s. Opponents to the war
in Vietnam pointed to the serious distortions of reality affected by official accounts, even if they
might be ”true” in a very narrow sense of the word (Chomsky 1973). Others demonstrated how
grave problems of social inequality were defined out of the political debate (Edelman 1977).
Likewise, the feminist movement revealed linguistic conventions reflecting male dominance, and in
many cases actively supporting female subordination (Blakar 1972).
Underlying these cases is the assumption that the description of any object is prone to some form of
perspectivism, simply because we cannot grasp the world from a single, objective, fixed point. In
describing a state of the world we (i) select those objects which are to be regarded as relevant, as
well as those intrinsic aspects which should command our attention. (ii) Objects and actions are
framed and classified, and thereby (iii) ascribed a specific value. (iv) When events and states are
conceived as produced in a given way, the actions generating them may be exposed in the passive
mode, thus hiding some crucial characteristics of relevant actors. (v) Descriptions of single events
are chained together and given a narrative and dramatized structure.
The acknowledgement of these sources of perspectivism need not entail a general relativism, but it
points to the fact that we never escape interpretation of the world, even if some truth claims are
more persuasive and well-founded than others. The necessity of interpretation leads to the formation
of groups with competing world views striving for hegemony in the public discourse, such as left
versus right in politics, monetarists versus Keynesians in economics, traditionalists versus
modernists in literary debate .
In the third type, which is the theme of the present paper, the focus is not on characteristics
of the message, but on the implicit structuring of the social relations among those who participate in
communicative interaction. In the following I shall take speech act theory as the point of departure,
and discuss the fruitfulness of using it as a frame of reference in comparing the work of Erving
Goffman, Michel Foucault and Pierre Bourdieu. The discussion starts at the micro level. I then
proceed to the macro level, contrasting the macro references of Foucault and Bourdieu to the
institutional conception of John Searle, based in speech act theory.
Speech act theory
Communication is never only communication, i.e. a message to be interpreted. The basic
assumption in the theory of speech acts is that communication also has an action side; it expresses
that the sender wants to achieve something. The sender relates to one or more receivers with the
expectation of a response of a certain type. This response depends not only on the interpretation of
the meaning of the message, but on its mode, what the sender wants from the other. The sender of
the message may transmit a statement, declaration, question, promise, threat or command (the so-
called ‘illocutionary force’ of the utterance). The message elicits a response in the receivers, who
obey the command, answer the question or assume that the description is correct. Or to the contrary
they refuse to accept the message.
The set of core questions in speech act theory concerns the description of the felicity conditions of
these various acts, for example, what it is that constitutes such linguistic constructs as a ‘question’,
a ‘request’ or a ‘statement’ (Austin 1961, Searle 1969). Less discussed, but more central to the
present discussion, is the point that the power of speakers to influence listeners is what speech act
theory is about. In other words, that communication inherently has a meaning side and a power side.
Erving Goffman and the definition of the situation

3
Closely parallel to speech act theory, Erving Goffman raises a similar question in his analysis of
what constitutes social interaction. Goffman did not present himself as a theorist of power, and is
hardly ever cited in discussions of power and communication (but see Rogers 1977, 1979). To
Goffman, power was only one among many relevant aspects of everyday interaction. Nevertheless,
it is striking how his work shows itself to be an important frame of reference to other theories of
power and communication.
Building on the work of W. I. Thomas, Goffman developed his conception of the “definition of the
situation” in The Presentation of Self in Everyday Life (1959). At the start of an interaction
sequence actors have a strong propensity to acquire a more or less conscious idea of the kind of
situation they enter into. The central questions in this respect is: “What kind of themes are deemed
appropriate or inappropriate in this conversation?”, with the implication that any definition will
have a moral content (Goffman 1959 [1969:24]). This leads up to the question: “Who are the
relevant partners?”, “What are their relevant social positions?” To a large degree these questions are
not raised explicitly, but the answer is inferred from the behavior of the partners in the process by
reference to former patterns of definition as well as through observation of gestures and attitudes
and inductive inference from the ongoing conversation.
In these analyses, Goffman’s main theme is not power, but above all the understanding of the
conditions for continuing the relationship in an undisturbed way. Nevertheless, his point of
departure is individuals’ need to protect and strengthen their self in interaction, and consequently
their striving for control of the impression they make vis-à-vis others. This point is underlined by
Goffman’s explicit reference to an unpublished (at that time) paper by Jay Haley, probably
elaborated in Strategies of Psychotherapy (1963), where Haley emphasizes the salience of power
relations in therapeutic processes (Goffman 1969:15).
Moreover, the examples given by Goffman demonstrate the salience of power in processes of
definition of the situation, because the outcome is the answer not only to the question “What is
going on here?” but quite often also “Who gets the upper hand in this interaction?” Typical
examples are a teacher who enters a new class, and immediately demonstrates who will be in charge
during the time she will spend together with the students, or nurses at a psychiatric ward pursuing
the same strategy in their encounter with new patients (Goffman 1969:23).
In a later work Goffman develops a related idea into ‘frame analysis’ (Goffman 1974). Here the
focus is not so much on the definitions of the situation between actors as on the interpretive side of
communication. At the same time he makes explicit references to speech act theory, emphasizing
that “[t]he point is to try to apply to all social behavior something of what linguists and logicians
have considered in regard to statements” (Goffman 1974:44). The main thrust of Goffman’s
exposition is the demonstration of the enormous flexibility of framing found in everyday life. We
continuously enter into and go out of frames of interpretation, distinguishing between spontaneous
behavior, irony, theatrical acting, deceit, etc. within a time span of a few minutes, and even seconds.
In the context of power analysis, the concept of the frame is important in two ways. First, it is
immediately clear that the ability to manipulate other peoples’ frames of interpretation will enhance
the power of a given actor. On the other hand, it should also be seen that the flexibility of framing
allows for a reframing of messages sent by power-holders to subordinates. Allusive jokes, irony,
and reinterpretation setting the fate of the subordinate in a different light, may turn commands and
statements from power-holders into subversive communication.
Although the idea of definition of the situation is fairly elementary, it serves as a matrix of models
of the exercise of power in various settings, both in public and in more restricted settings. Models of
non-decisions (Bachrach and Baratz 1961), and agenda setting (Coombs and MacKuen 1981), the
tactics in channeling a decision making process into the most favorable arena (Hellevik 1972), or
the “garbage can” theory of decision making (Cohen, March and Olsen 1976) all contain situational
definitions as a crucial element. They describe various ways that actors may influence the common
understanding of what is going on and what are the rights and rules of appropriate behavior in the
given setting.

4
Michel Foucault and power in discourse
Among theorists of power, Michel Foucault is probably the most cited in the last two decades.
Nevertheless, it is a matter of dispute whether it is possible to extract one specific theory of power
from his heterogeneous work (Kelly 1994, Sandmo 1999). A sympathetic analyst makes the
observation that “Foucault’s work is at roots ad hoc, fragmentary and incomplete” (Gutting
1994:2).2 Three aspects of power may be distinguished in Foucault’s work, linked to discourse, to
discipline, and to self-reflection. These are partly overlapping, discourse being most salient in the
work from the 1960s, and self-reflection the dominant concern in the early 1980s.
The sources of Foucault’s concept of power that have gained most attention are his studies from the
1970s, Discipline and Punish (1977), and the first volume of History of Sexuality (1979). In both
books his concept of power is extremely instrumental, the former containing analyses mostly
reminiscent of Frederick Taylor’s Scientific Management, while the latter, more theoretical work, is
permeated by such expressions as “technologies of power”, “strategies of power”, and where he
even compares social relations to military operations. The analyses in Discipline and Punish focus
primarily on the manipulation of bodies, i.e. the installation of material constraints and continuous
supervision of the behavior of subordinates. In the next round, however, a strongly programmed
internalization of norms is presupposed to take place. Despite the fact that in these books Foucault
discusses situations that presuppose communication, he makes few if any attempts at depicting
power mechanisms with a distinct hermeneutic basis. A possible exception is the prominence
accorded to the gaze, demonstrating superior competence on the part of the teacher or the prison
officer.
Thus, in the present context it is Foucault’s concept of discourse, elaborated in his Archeology of
Knowledge (1969 [1972]) and reiterated in his inaugural lecture at the Collège de France in 1970,
”The Order of Discourse” (Foucault 1972), that is relevant. While the former is a treatise on
historical methodology, which has little or nothing to say on power, “The Order of Discourse”
outlines several dimensions of power in communication. In his treatment of discourse Foucault first
points to a set of procedures of exclusion: various types of prohibitions concerning content and
performance of speech; delineation against madness; and the division between truth and falseness.
Secondly, he considers principles constructing order in the discourse. Order is achieved by the
assumption of an authorial character, by the distinction between source texts and commentaries, and
by the organization of disciplines of knowledge. And third, he discusses rules delineating who are
competent speakers. These rules are institutionalized in rituals, membership in learned societies,
doctrinal groups, and educational systems.
This somewhat pedantic summary makes it clear that a number of parallels exist between Foucault
and Goffman, even if Foucault centers on academic discourse, and Goffman on exchanges in
everyday life. Above all, they have in common a parallel to speech act theory: communication is the
enactment of more than the communicative message; it is constituted by a set of underlying rules,
and it has several social implications. Even if Foucault’s discussion of discourse does not imply a
“definition of the situation” in direct relationship to participants in the interaction, it does imply a
definition of the message and the more general communicative situation surrounding the learned
discourse. Concerning the difference between everyday life and exchange in academic circles, the
main difference is found in the second dimension, that of unity of discourse. Everyday
communication is much more dispersed and free floating, and concepts such as ‘author’ and
‘commentary’ have less relevance. The main power aspects, however, are not found here, but in the
2 Cf. also the following (half ironic?) statement from the introduction to The Archeology of Knowledge: ”Do not ask
who I am and do not ask me to remain the same: leave it to our bureaucrats and our police to see that our papers are in
order.” (Foucault 1972:17)

5
exclusionary processes, whether they are connected to the selection of themes or the set of relevant
speakers. In these respects, the two types of communication are basically similar.
Symbolic power and symbolic violence in Pierre Bourdieu
A common characteristic in Michel Foucault and Pierre Bourdieu is a desire to avoid the traditional
social psychological idea of ‘internalization’ of norms, understood as directives of action, and
concentrate on institutionalized practices and processes of power and control in ongoing interaction
processes. In Bourdieu’s work this is combined with symbolic elements in several ways, more or
less closely associated to the main theme here.
First, Bordieu has coined a concept of symbolic capital, which is only marginally connected to
symbols and communication in the traditional meaning of the words. ‘Symbolic capital’ denotes the
reinforcement of a person’s position based on former performance. The artist, the stock-broker or
the politician is judged not only on the basis of an assessment of his talent, but also his successes in
former transactions. Another notion covering the same phenomenon might be ‘rumor’ or
‘credibility’ (Danielsen and Nordli Hansen 1999). By contrast, symbolic power is directly connected
to processes of interpretation. It is constituted by actors’ ability to impose their own interpretation
of reality on others (Bourdieu 1977:165), in line with what was called above the standard theory of
power in communication. Thus, what Bourdieu has in mind is not restricted to a definition of the
situation, but refers to a wider definition of society, or of the world via a corpus of ideology.
This leads to the more specific notion of symbolic violence, where the focus is set on the receiver. It
may be summarized in a model with five components: (a) a differential mode of interpretation
(‘habitus’), based in variation in institutionalized practices in different groups, (b) strategic emission
of signals based, for example on dress codes, language codes, art, or other symbolic systems, (c)
variation in the ability to interpret signals, with the consequence that signals are understood by
some groups and not others, (d) lacking ability of interpretation commonly leads to a feeling of
shame and self contempt, (e) suppression of their reactions by receivers through a process of
misrecognition, which gives legitimacy to prestige signals or codes. Thus, an oppressive mode of
communication is taken for granted and regarded as “natural”.
Much of Bourdieu’s writing is variations on this theme. The institution of a standard written
language is a simple example (Bourdieu 1991:43 ff). Standard French rests on linguistic codes that
are not shared by the majority of the population but are elaborated and refined by a Paris-based
cultural elite. At the same time, the media, the publishing industry, and above all the school system,
hold up the official norm as the only natural medium of expression for the French. People from the
Béarn district for example – where Bourdieu grew up – using their local dialect in communication
with people connected to the official norm, are confronted with their own feeling of insufficiency.
Accepting this, Béarn people cannot avoid defining themselves as culturally inferior to the Parisian
elite. They are particularly exposed to such symbolic attacks when people with high prestige are
using both the low and the high prestige code, as when the mayor uses both standard French and the
local dialect (Bourdieu 1991:68), or the boss acts as if he were on a par with his employees of low
rank.
In this theory, Bourdieu exposes another variant of the definition of the situation. In Goffman’s
version, the question of social dignity was central, but was seen at least partly as a product of the
negotiation between the parties in the interaction. In Bourdieu’s conception, the definition of the
situation is more strongly associated to the modes of communication. Actors are accorded prestige,
and thus varying degrees of self-confidence as a product of the general social prestige of the codes
they are using.
The difference to the Goffmanian conception may also be described from the perspective of speech
act theory. Given his general orientation it might be expected that Bourdieu would embrace speech
act theory. But he shows himself quite critical in this respect. The reason is that Bourdieu denies

6
that the utterance itself can have any separate “force”. His main point is that any authority in the
communicative situation is derived directly from socially legitimated authority (Bourdieu
1991:107). It is difficult to read this otherwise than as a statement that symbolic power after all is an
epiphenomenon. The message itself has only negligible power effects; it is the match or lacking
match of codes in senders and receivers that is at stake. If Foucault points to the exclusion of given
topics and speakers, Bourdieu focuses on the definition of social positions inherent in linguistic
systems. Despite the label ‘symbolic violence’, symbols are seen as almost direct extensions of
social relations rather than having effects of their own.
From micro to macro
Rightly understood, all the theories discussed contribute significantly to the analysis of power
relations on the micro level. At the same time they lead up to a general point: if it is true that power
is inherent in communication, it follows that power is found in all kinds of social relationships. The
salient issue in the analysis of power, however, is not the description of innumerable social events,
but how these events are translated and aggregated into relatively stable and systematic social
structures, networks and institutions. Based on the communicative approach, there has been a
continuous stream of assertions over the last thirty years or so, that power is ‘floating all over
society’, more or less invisible but nevertheless omnipresent. However, the problem with most of
these assumptions is that they do not account for the way that this translation into macro patterns is
taking place. The macro level is simply ascribed a given structure, without further empirical
investigation.
Three different approaches to the translation from micro to macro may be singled out. First, a
relationship of homology, which under certain conditions is also a metaphorical relationship. Quite
commonly, one institution is treated as a metaphor for society as a whole (e.g. the prison as an
image of society).3 Alternatively, homologies may be ascribed to basic types of social structures.
An interesting example is found in the debate between Talcott Parsons (1963) and James Coleman
(1963, 1970) in the early 1960s. Despite vast differences in their general approach, both took as
their point of departure the idea that power and money have a similar structure. In this respect, both
adhere to the homology assumption.
Second, models of structural determination. In this case, an underlying “systems logic” is assumed
to be at work. Empirical observations are assumed to be shaped by a set of more or less constant
factors which cannot be observed directly but have to be inferred. They may be compared to a
“grammar” or a deep structure not unlike linguistic rules, governing central parts of social life.
A problem with both these approaches is that the macro structure is accorded a life of its own. This
is not to deny that structural constraints do exist. Likewise, there is no reason to discard causal
chains where central elements cannot be reduced to intentional action. Nevertheless, the effects of
social structures is not something that can be taken as axiomatic, but must be assessed empirically.
The alternative to homologies and structural determination are historical theories that are less
satisfactory in the sense that they are more open and less parsimonious, but closer to empirical
reality. An example is Charles Lindblom’s discussion of money and power as “media of
communication and exchange” (Lindblom 1977). In contrast to Parsons and Coleman, Lindblom
does not assume that money and power are homologous, but treats them as complementary. One of
the most important differences pointed out by Lindblom is that money to a large extent is
convertible across institutions, whereas power more often is confined to specific positions in
distinct institutional arenas. An implication of this view is that it is not meaningful to speak of the
allocation of power in society in the same way as the allocation of money.
A further elaboration along the same line was made by Michael Mann (1986) in his conception of
four levels of power in macro. Mann’s main concern is an attempt to avoid reification in our
3 It is unclear whether Foucault (1977) himself makes this connection, but undoubtedly many of his interpreters do.

7
interpretation of society. He suggests that in societies power functions on four different levels, each
with their specific modus operandi: military, political, economic and ideological. These levels are
only partially overlapping. Cultural patterns may have an extension quite different from that of the
military resources, which again differ from the structure of markets and commercial transactions.
The combination of possibilities for control in each sphere created by technology and organizational
devices, gives each social formation its distinct features.
On this background I shall proceed to the discussion of the theories of Foucault and Bourdieu on the
macro level. Since Goffman hardly went outside the micro relations in his analyses, comparisons of
Foucault and Bourdieu with Goffman will not be made systematically on the macro level. However,
towards the end I shall take up a macro theory of power and communication based on speech act
theory, which is compatible with Goffman’s thinking, even though there is no reason to ascribe
such a theory to him.
Foucault and knowledge regimes
On the macro level, the concepts linked to Foucauldian discourse are those of ‘discursive
formation’ and ‘episteme’. Taken together they may be characterized somewhat loosely as a
‘knowledge regime’. As defined in The Order of Things, an episteme is not a specific form of
knowledge; it is the condition of possibility of knowledge in a given epoch – that which makes
it possible to organize ideas and observations into bodies of knowledge (Foucault 1966:171),
whereas in The Archeology of Knowledge it is defined as the total set of relations that unite the
discursive practices of a given period (Foucault 1972:191). A discourse formation is
characterized as a rule differentiating specific discourses. If discourses manifest power in
micro, knowledge regimes may be seen as deep structures that determine their modus
operandi. In this respect, they might be seen as deep-seated power structures in macro.
The question is immediately raised about the relationship between such knowledge regimes and
Thomas Kuhn’s concept of ’paradigm’ (Kuhn 1961). In the 1950s, Foucault encountered Kuhn
(Sandmo 2000:x), and the obvious similarities in their approaches to historical discontinuities make
it plausible that Kuhn influenced Foucault’s thinking for a period. Kuhn describes scientific
activities as going on within a given theoretical paradigm. This is what he calls ‘normal science’. If
new evidence emerges, which is difficult to assess within the frames of the existing paradigm,
scientists will initially overlook – or try to redefine – such disturbing findings. But as more and
more findings are accumulated over time they eventually lead to a breakdown of the existing
paradigm and the establishment of a new one.
The interesting point in this connection is not the question of influence, but the debates over Kuhn’s
work may shed some light on Foucault’s concepts as well. A first problem is simply that of
empirical confirmation. If they are to have any meaning, the notions of episteme and discourse
formation must have an empirical reference. An answer near at hand is that these phenomena are
latent and unobservable. But this is begging the question. Given that it is unobservable, it becomes
very difficult to give a clear delineation of a phenomenon, and it may well be that it is wholly
imaginary. This is not a denial of the possible existence of deep structures but a caution that a
realistic assessment of it is very demanding. In this respect there is a parallel in the problems
pointed out by Masterman concerning Kuhn’s use of paradigm. Masterman (1970) showed that in
Kuhn’s own texts, the notion of ‘paradigm’ was given 21 different interpretations. A similar
inconsistency is found in Foucault’s assumptions about episteme. In The Order of Things
(1966:171) he implied that there could be only one episteme in an epoch, in The Archeology of
Knowledge (1972:176) and other texts he conceded that various disciplines have different epistemic
status at the same time (see also Flynn 1994:33). In the former he defines episteme as a
precondition of knowledge, in the latter as the totality of relations of knowledge.
A second problem resides in the question of change. The main thrust of Foucauldian thinking is the
transformation of knowledge regimes, or epistemic shifts, basically between the Renaissance,
Enlightenment and post revolutionary epochs. This raises the question of how such epistemic shifts

8
are enacted, but Foucault is not very clear on this point. Within the frames of his empirical research
on changes in medical diagnoses, treatment of madness, or disciple devices in schools, prisons and
armies, this may be unproblematic. However, his theoretical account of changes in knowledge
regimes does little more than acknowledge that changes do occur, pointing to their basis in a change
of rules of discourse formation (Foucault 1972:166 ff.). In other words, changes in modes of
thinking are assumed to be determined by a deep structure of which we have no knowledge except
vague notions, neither of its static nor its dynamic features.
In themselves, these problems suffice to create a general skepticism to the construct of episteme.
How can we know that such a thing exists at all? Undeniably, the idea of deep structure, something
being the condition of ordering of knowledge, has an immediate appeal. Grammar is an obvious
example. However, even if native speakers of any language have learnt its grammar in an implicit
way, it is by no means unobservable. All written languages have codified grammars, which may be
the object of discussion and deliberative change.
The notion of deep structure should be held up, however, against the Goffmanian concept of frame.
What is most significant in this context, is Goffman’s underlining of the flexibility of framing in
everyday life. Within a time scope of a few minutes any competent speaker will change frames of
reference several times, for example, in having an ironic chat with a friend on the phone, writing out
a check, planning dinner with a family member, watching a TV ad (Goffman 1974). In itself this
flexibility calls forth doubts as to the necessity of a unified deep structure governing communication
and interpretation.
To the extent that this argument is accepted, we can do without the concept of episteme, and simply
situate discourse and the aspects of power that it entails, within academic institutions in the same
way as power structures in schools, factories, asylums and prisons are situated basically within
these institutions. It seems that Foucault arrived at a similar conclusion in the mid 1970s, when he
stated in an interview that the idea of discursive regime in his earlier work was “confused … too
much with systematicity, theoretical form, or something like a paradigm” at the cost of
understanding “the effects of power peculiar to the play of statements.” (Foucault 1980:113). I take
this to mean that Foucault had given up the idea of macro structures of communicative power,
beyond discourse itself.
Bourdieu, institutions and hyper-integration
If Foucault gave up a general account of micro-macro relations, the same cannot be said of
Bourdieu. On the contrary, for a long time he insisted on having solved the micro-macro problem
by building a comprehensive theory embracing both levels. The core concept in this endeavor is that
of habitus. In Bourdieu’s work ‘habitus’ denotes habit, institutionalized practice, but does not so
much refer to internalized norms as to modes of interpretation.4 It has the position as the mediating
link between social structure and individual behavior. Thus, in Bourdieu, the concept of habitus has
much of the same position as that which episteme had in Foucault. The main difference is that
habitus is not a common organizing principle across society as a whole, but that the various classes
are ascribed different habituses. Taken together, however, the various habituses form a social
totality.
In order to account for a general pattern in social structure, habitus must be sufficiently well ordered
and clear to secure uniform reactions across situations. Likewise, the events enacted by habitus
must have quite foreseeable effects. If there is a substantive lack of clarity and constant threats of
misinterpretation, and constant presence of serious unintended consequences of actions, habituses
simply will not suffice to uphold stable social structures. It is probably sufficient to refer to traits in
any communicative process to make it clear that such stability is a wishful dream. Ambiguities,
4 Bourdieu expressly maintains his aim to account for social regularities without the notions of rule-oriented behavior
(Bourdieu, after Schwartz 1997:95).

9
misunderstandings and meta-communication are a constant feature of any communicative field.
This is true not only of messages, but applies to the interpretation of criteria of legitimacy as well
(cf. Thompson 1984:59). Thus, there arises a general need for discussion of competing
interpretations of the world.
Of course, it may be countered, this point was already made by Bourdieu himself, in his exposition
of competitive positions in fields of action as, for example, within the field of literary interpretation
(Bourdieu 1996). But how is a constant positional struggle in public discourse compatible with a
general conception of social reproduction through habitus? This is possible on one condition,
namely that cultural power struggles are not primarily fought by individuals defending their own,
idiosyncratic positions. Instead, actors are assumed to be representatives of competing social
interests, and articulate the “objective strategies” based on their interests. This, however, is hardly a
fair description of a field such as literature where the pressure for innovation is one of the strongest
forces. On the other hand, if cultural power struggles are considered as more open, and only
partially reducible to constellations of interests, the model of social reproduction breaks down.
The stability of social hierarchy is challenged from another angle as well, which may be
summarized in the idea of the ‘productivity of power’ exposed by Parsons as well as by Foucault.
Bourdieu seems to imply that the standard culture solely serves the reinforcement of power in the
dominant classes. But that is an arbitrary assumption. Take once more the simple example of
standard language. Undoubtedly, under certain conditions one of its consequences may be
oppression, but it also serves to empower those who learn it, for example by making it possible to
organize common interests across local groups, or establishing arenas of political debate on the
national level.
Hence, communicative power is bound to the actual functioning of social institutions. In his
empirical studies, Bourdieu outlined the power of schools and institutions of higher education in an
admirable way. But paradoxically, he seems to imply that institutions have one given power
structure, which is kept constant even if the institution itself undergoes changes in line with the
proverb “Plus ça change, plus c’est la même chose”. But as a consequence of institutional changes,
not only does the form of hierarchies in schools change, but the so does the efficiency of these
hierarchies. Consequently, the analysis of power in macro is dependent on understanding these
institutional changes, and not on some vaguely conceived power matrix assumed to be constantly
present in society.
A speech act theory of power and macro phenomena
So far the discussion has pointed to the commonalities of Foucault and Bourdieu’s conceptions of
power in communication with the theory of speech acts. Their attempts at elaborating macro
theories have shown themselves less convincing, however. An alternative macro theory of power in
communication was introduced by the philosopher of language John Searle in his The Construction
of Social Reality (1995). The question, then, is whether the contributions by Foucault and Bourdieu
may be connected to this theory.
Searle’s main work has been within speech act theory and the philosophy of mind, themes at a far
distance from macro sociology. Thus, even if he draws up some important aspects of power in
communication, his contribution necessarily remains sketchy. The innovative point in Searle, is the
way he connects speech acts to institutions. Institutions are not only a framework for
communication, but are themselves made up by communication. Institutions are conceived by
Searle as generated by speech acts, albeit through a complex iteration of speech acts, and they
basically have the same structure. In this respect there is a close connection between Searle’s theory
and Goffman’s work. On a par with promises, threats or statements, social institutions may be
ascribed ‘felicity conditions’ in the sense that the binding force of the institution is based on the

10
rules connected to their roles. Authority is accorded to given roles within that institution when
social members implicitly acknowledge that “We accept that X has the power to do Y in C”.
Thus, the power in the institution may be said to rest on two factors: The legitimacy accorded to
instituted positions and the illocutionary force of the speech acts making up that institution. To take
a simple example: When Paul and Mary marry, the pastor asks “Will you take Mary/Paul as your
wife/husband?” and the two answer “Yes”. By this performative speech act marriage is completed.
It rests on two separate conditions: that the pastor actually is empowered to pose the question, and
that the future spouses make the promise. The pastor’s position is based on a preexisting social
order, but this order could not function if the promise made by the spouses did not have a normative
binding character.
Institutions have a force in society not only because they stabilize relationships, but also because
they permit the maintenance of these relationships in various flexible ways. The couple remain
married even if they live separately, marriage concerns several sets of rights not directly related to
their living together, such as legislation regulating their property, or their duties toward their
children, to take a few examples. Thus, performative speech acts may function as essential nodes in
a larger network of institutional order. Moreover, underlining the communicative element in social
institutions emphasizes the necessity of legitimate acceptance of power structures, be it in the form
of active loyalty or passive consent from the members of society. In the end, power based mainly on
force or violence alone will become paralyzed.
As mentioned above, Bourdieu (1991:74, 107 ff.) criticizes Austin on the assumption that a speech
act only has an effect if delegated by positions with social power. In the wedding example that is
true for the pastor, but Bourdieu misses the point by focusing exclusively on the position of the
pastor, while disregarding the significance of the promise made by the spouses (see also Thompson
1984:68). Thereby the communicative basis for power is underlined.
At the same time Bourdieu has a point, even though he states it somewhat crudely. It is obviously
mistaken to assert that the force of speech acts always rests on social power, but it is equally naïve
to assume that purely linguistic effects are at work. Even if Searle strongly emphasizes the interplay
of power and communication, it has no place in his earlier work on speech acts in micro.
Commands are simply assumed to be issued and followed, and statements are simply taken as
statements about the world. In this respect, the insights into the inherent power in communication
elaborated by Foucault and Bourdieu are a valuable supplement to a more general theory on speech
acts and power.
In accordance with the present discussion, Searle’s theory is ‘general’ in the sense that power is part
of communication both in micro and macro. However, it is ‘local’ in the sense that speech acts and
discourse are shaped by the specific institutions within which they take place, and no a priori
assumptions are made concerning the relationship between social institutions. In most respects this
is analogous to the conception advocated by Foucault in The History of Sexuality (1979). But there
is more to social life than institutions. Before concluding, I shall give a few examples of power
mediated by performative speech acts that are not institutional in the sense outlined here.
Collective action and varieties of public discourse
As a complement to this general conception, two examples of power linked to the public discourse
will be sketched. The point here is the nature of the social institution at work. In the case of public
discourse, the most salient institutions are those related to communication itself, such as modern
mass media and comparable channels of exchange. Public discourse consists to a large extent of
struggles over hegemony in defining social problems and their solutions. The core of these struggles
is credibility of argumentation. But credibility has a broader range than the quality of argumentation
itself, something that enhances indirect modes of power.

11
The first example is that of spreading rumors. In groups and societies with tight networks, spreading
rumors may be a very efficient way of exercising power, for example in order to exclude
unwelcome competitors from the scene. The source of the rumor may be very difficult to identify,
and the person assailed has no chance to falsify the rumor, as long as he or she does not have direct
access to every receiver.
In cases where mass media are the main channel, this situation becomes more ambiguous. On the
one hand, rumors may be spread more efficiently. On the other, even if the original source may still
be anonymous, the newspaper transferring the news is not. Communication is structured in such a
way that it becomes easier to challenge the rumor right on by a public rejection of the accusation.
Despite widespread cynicism, the medium in charge usually needs to protect its credibility in the
public arena, something that may contribute to the tempering of flows of rumors. A further twist
occurs when rumors are not reported directly, but when actors publicly take a stand on them, such
as X denies that he is a muslim, or has friendly relations to terrorists. In most cases this is a counter
productive strategy seen from the position of X, as the denial of the rumor implies the confirmation
of its existence, well founded or not.
The structure of mass media has also brought forth another source of power connected to
performative speech acts. This rests on a structure related to the spread of rumors, found in
collective action problems where actors cannot get in contact with each other in order to verify or
falsify a bit of information. Mass media is an exceptional source of credibility for actors holding
public positions. If the press declares that “we” do not have confidence in a given politician, a
director of the central bank, or a talk show host, this is both a description and the creation of a social
fact. “We”, the readers, have no choice but to accept the statement, thus making it true. In modern
societies, this makes campaign journalism an immense source of social power, giving the press the
force of driving large parts of public officials out of their positions, partly on quite shaky grounds.
Conclusion
The relationship between power and communication may be analyzed on three distinct levels: (i)
Communication as a means to the exercise of non-communicative action (ii) power following from
descriptions of the world, whether they refer to specific states or to more general conceptions of the
social world; (iii) the intermingling of power and communication due to the preconditions
surrounding the communicative message. Here the main focus has been on the latter point. Taking
speech act theory as reference, three different constellations of power and communication have
been pointed out. In Goffman, power is used as a means for defining the situation and creating the
condition for interaction. In Foucault the interactional aspect is lacking. Instead, Foucault focuses
on long term stylized discourse patterns where power resides in pre-established exclusion of themes
and interlocutors. Bourdieu takes a position between these as he retains Foucault’s conception of
pre-established structures while focusing mainly on the recipient and his or her reactions when
confronted with a superior position. Somewhat simplified, the three positions might be
characterized as “definition of the situation”, “definition of the theme”, and “definition of the
recipient”.
Turning to the macro level, it has been discussed whether the concepts of episteme and habitus
might form the basis for general theories of power in communication. The answer has been
negative. Even if a unitary basis for societal power in communication was assumed to be possible,
these concepts lack the dynamic character that would be necessary to explain micro-macro
relations. In the absence of plausible alternatives, a more adequate strategy is to concentrate on an
institutional frame of reference – incidentally the same as that used by Foucault in his historical
analyses of power. A broader notion of societal power is then dependent on the actual constellation
of institutions in a given social formation. The attempt by John Searle to use speech act theory as a
basis for the understanding of social institutions, seems promising in this respect. Finally, an

12
example from the institution of mass media illustrates the power of the media, not only in their
descriptions of the world, but also in their construction of social facts through performative speech
acts.
References
Austin, J.L. (1962), How to Do Things with Words. Cambridge: Harvard University Press
Bachrach, P. and Baratz, M. (1962), “The two faces of power”. American Political Science Review,
57:947-952.
Blakar, R. M. (1972), Språk er makt. Oslo: Pax.
Bourdieu, P. (1977), Outline of a Theory of Practice. Cambridge: Cambridge University Press.
Bourdieu, P. (1991), Language and Symbolic Power. Cambridge: Polity Press.
Bourdieu, P. (1996), Rules of art. Cambridge: Polity Press.
Chomsky, N. (1973), For reasons of state. New York: Pantheon
March, J and Johan P. Olsen (eds.), Ambiguity and Choice in Organizations. Oslo:
Universitetsforlaget
Coleman, James S. (1963), “Comment on ‘On the Concept of Influence’”. Public Opinion
Quarterly, 27:63-82.
Coleman, James S. (1970), “Political Money”. American Political Science Review, 64:1074-1087.
Coombs, S. and MacKuen, M. (1981), More Than News: Media Power in Public Affairs. Beverly
Hills: Sage.
Daloz, Jean-P. (2003), ”Ostentation in Comparative Perspective: Culture and Elite Legitimation”.
Comparative Social Research, 21: 29-62.
Danielsen, A. and Marianne Nordli Hansen (1999), “ In: F. Engelstad (ed.), Om makt. Teori og
kritikk. Oslo: AdNotam Gyldendal.
Dawding, K. (1996), Power. Buckingham: Open University Press.
Edelman, M. (1977), Political Language. New York: Academic Press.
Engelstad, F. (1999); “Maktbegrepet etter Max Weber”. In: F. Engelstad (ed.), Om makt. Teori og
kritikk. Oslo: AdNotam Gyldendal.
Engelstad, F. (2001), “Expressive forms as generators, transmitters and transformers of power.”
Forthcoming in International Encyclopedia of the Social and Behavioral Sciences. Oxford:
Elsevier.
Flynn, T. (1994), “Foucault’s Mapping of History”. In: G. Gutting, ed., The Cambridge Companion
to Foucault. Cambridge: Cambridge University Press.

13
Foucault, M. (1966), Les mots et les choses. Paris: Gallimard.
Foucault, M. (1972), The Archeology of Knowledge and The Discourse on Language. New York:
Pantheon.
Foucault, M. (1977), Discipline and Punish. London: Allen Lane.
Foucault, M. (1979), History of Sexuality. Vol. I: The Will to Truth. New York: Vintage
Foucault, M. (1980), Power/Knowledge. New York: Pantheon
Goffman, E. (1959), The Presentation of Self in Everyday Life. Garden City, NY: Doubleday.
Goffman, E. (1974), Frame Analysis. Boston: Northeastern University Press.
Gutting, G. (1994), “Introduction”. In: G. Gutting, ed., The Cambridge Companion to Foucault.
Cambridge: Cambridge University Press.
Haley, J. (1963), Strategies of Psychotherapy. New York: Wiley.
Hellevik, O (1972), Kan Bratteli gjøre som han vil? Oslo: Pax
Kelly, M., ed. (1994), Critique and Power. Recasting the Foucault/Habermas Debate. Cambridge:
MIT Press.
Kuhn, T. (1962), The Structure of Scientific Revolutions. Chicago: University of Chicago Press.
Lindblom, C. (1977), Politics and Markets. New York: Basic Books.
Mann, M. (1986), The Sources of Social Power. Vol. I: A History of power from the Beginning to
A.D. 1760. Cambridge: Cambridge University Press.
Masterman, M. (1970), “ The meaning of a paradigm“. In: Imre Lakatos and Alan Musgrave, eds.,
Criticism and the Growth of Krowledge. Cambridge: Cambridge University Press.
Parsons, T. (1963), ”On the Concept of Influence”. Public Opinion Quarterly, 27:37-62.
Rogers, M. (1977), “Goffman and Power”. American Sociologist, 12:88-95.
Rogers, M. (1979); “Goffman on Power, Hierarchy and Status”. In: J. Ditton (ed.), The View from
Goffman. London: Macmillan.
Sandmo, E. (1999), ”Michel Foucault som maktteoretiker.” In: F. Engelstad, ed., Om makt. Teori og
kritikk. Oslo: AdNotam Gyldendal.
Sandmo, E. (2000), ”Innledende essay”. In M. Foucault, Galskapens historie. Oslo: Bokklubben
Dagens Bøker.
Schwartz, D. (1997), Culture & Power. The Sociology of Pierre Bourdieu. Chicago: University of
Chicago Press.

14
Searle, John (1969), Speech Acts. Cambridge: Cambridge University Press.
Searle, John (1995), The Construction of Social Reality. New York: Wiley
Thompson, John B., (1984). Studies in the Theory of Ideology. Cambridge: Polity Press.