Sociology Mind
2012. Vol.2, No.4, 355-361
Published Online October 2012 in SciRes (http://www.SciRP.org/journal/sm) http://dx.doi.org/10.4236/sm.2012.24047
Agency at Work: A Dynamic Interpretive Approach
to the Study of Action
Colin Campbell
Department of Sociology, Un iversity of York, York, UK
Received July 16th, 2012; r evised August 19th, 2012; acc e p t e d A u gust 29th, 2012
Roy’s 1950s paper “Banana Time” is used as the basis for an exploration of the nature and relationship of
agency and action. Roy’s activity in playing his “game of work” is shown to be a feature of individual
conduct that, despite possessing subjective meaning, is largely neglected by contemporary sociologists,
mainly because of its covert character. What an examination of this aspect of his conduct suggests is the
need to revise the conventional observational approach to the definition of the unit act by recognising that
there may well be an additional actor’s covert definition sitting within the accepted social definition and
that it is therefore necessary to use the criterion of attentionality to identify the unit act. An analysis of
Roy’s game of work also helps to shed light on the possible relationship between action and agency, re-
vealing that while the power of agency enables individuals to act, it is also frequently necessary for indi-
viduals to act in order to maintain or restore their power of agency. Finally, a consideration of the func-
tion fulfilled by Roy’s game of work shows that a behaviourist-style stimulus-response analysis of con-
duct is not at odds either with voluntarism or the adoption of the actor’s standpoint. This is because Roy
demonstrates how actors are themselves lay behaviourists, fully aware of how they need to manipulate
stimuli in order to produce desired responses in themselves.
Keywords: Agency; Definition of Action; Covert Action; Dynamic Interpretivism
Introduction
In 1959 David Roy published an article in the journal Human
Organization entitled “Banana Time”: Job Satisfaction and
Informal Interaction” that has subsequently come to acquire the
status of a classic in the field of industrial sociology (Roy,
1959). Although the focus of the article was on the social and
interactive functions performed by a ritual enacted by a group
of workers (this features the contents of a lunch box, hence the
significance of the title) it also includes an intriguing account of
how Roy, who enrolled as a worker in order to conduct his
research through participatory observation, struggled to adapt to
the manual role allotted to him. The job in question was that of
a clicker and involved operating a machine that used steel dies
to punch out pieces of leather or plastic. The work was ex-
tremely simple and very repetitive with the result that the te-
dium of the twelve-hour job quickly led Roy to experience “the
beast of monotony”, leading to boredom and fatigue1. These
feelings were so acute that he considered abandoning the job
and with it his projected programme of research. That he didn’t
in the end do so was because he discovered a way of coping
with the intrinsic boredom of the task allotted to him by turning
his actions into what he called his game of work. This involved
alternating the colours of the items worked together with
cleaning the die-head of the machine according to a set of rules
that he invented. The effect of this was to invest the extremely
simple movements that constituted the task in hand with a dis-
tinct meaning thereby providing for himself a system of re-
wards (albeit, as he admits, fairly minimal) linked to the
mini-goals so created. Roy himself describes the game of work
he created in this way as serving to create a “sequence of
short-range production goals with achievement rewards in the
form of activity change” (1959: p. 161). The effect of this was
that instead of being confronted by a seemingly endless series
of undifferentiated “happenings” Roy managed to create mean-
ingful experiences for himself, and in this way he found that he
could sustain an activity which previously he had experienced
as so excruciatingly boring, and consequently tiring, that he had
contemplated handing in his notice and abandoning the research.
This account of Roy’s experiences while undertaking research
would appear at first glance to have no particular significance
for sociology, except perhaps to serve as a warning of the dif-
ficulties that can be encountered when engaged in participatory
observation. It is certainly not immediately obvious how a con-
sideration of his experience might lead to significant insights
into the nature of action and agency let alone the development
of a novel perspective for their study. However these are pre-
cisely the claims that will be advanced here.
Action at Work
1Roy described his work situation as “standing all day in one spot beside
three old codgers in a dingy room looking out through barred windows at
the bare walls of a brick warehouse, leg movements restricted to the shifting
of body weight from one foot to the other, hand and arm movements con-
fined, for the most part, to a simple repetitive sequence of place the die,
punch the clicker, place the die, punch the clicker, and intellectual activity
reduced to c o mputing t h e hours to quitting time” (1959: p. 160).
The first point of note about the events described by Roy is
that they would not normally be recognised as constituting a
phenomenon worthy of sociological investigation. Since his
game of work is not part of an on-going social interaction, ori-
ented towards others or indeed visible to others, it would not
count as social action as this term is generally understood. But

C. CAMPBELL
then it is also unlikely to be recognised as action by those per-
spectives, currently popular in the discipline, that do focus on
individual, non-interactive conduct, such as rational choice or
rational action theory. For characteristically these perspectives
only employ abstract models of human behaviour, taking little
interest in the real subjective experience of actors, whilst also
focusing exclusively on decision-taking. Hence, while such
perspectives might be applied to an understanding of why Roy
decided to take the job of clipper in the first place (and possibly
even, had he been forced to abandon it, why he had done so),
how he managed to handle the strains and stresses of the job
itself would not be subject to investigation2.
Yet there would seem to be no good reason for excluding
what Roy recounts doing in his game of work from the subject
matter of the discipline. For it does fall into the category of
action as traditionally defined and understood. That is to say it
meets Weber’s requirement of being “human behaviour to
which the acting individual attaches a subjective meaning”
(1964: p. 88), while the fact that his behaviour is solitary in
nature and neither oriented to others or observed by them would
not appear to be a problem given that Weber’s most famous
example of action—that of the woodcutter chopping wood—is
also an instance of an individual acting alone and presumably
unobserved. Furthermore the fact that the conduct he describes
was apparently covert, or at least not actually noticeable to
others, would also not seem to be a problem given that, as We-
ber also asserts, action may be “either overt or purely inward or
subjective” (1964: p. 88). One is thus left to wonder why, in
this case, the kind of experience Roy recounts is typically not
the subject of sociological attention. But then, despite Weber’s
remark, it is actually the case that almost none of the theoretical
traditions that flourish in contemporary sociology concern
themselves with covert conduct, focusing exclusively on that
portion of an individual’s behaviour which is overt and hence
observable to others3. To what extent sociologists should be
censored for this omission is debatable given that Weber him-
self appears to overlook the significance of his own observation
when illustrating his interpretive method as he fails to provide
any real examples of covert actions; that is ones resembling that
provided by Roy4, apparently making the mistake of assuming
that the concept covered no more than the deliberate absence of
overt action5.
But then what is especially interesting about the behaviour
Roy describes is that it is neither a simple example of overt or
of covert action but rather a combination of the two, and hence
it is a surprise to realise that Weber seems to have made the
assumption—in which he was followed by Schutz (1967)
among others - that all actions could be divided into the two
simple categories of overt and covert rather than considering
the very real possibility that they might be a combination of the
two. Now although Roy does not make it clear in his account to
what extent it would be apparent to the casual observer that his
actions as a clicker-operator followed his own distinctly pat-
terned sequence, he does observe that he subsequently discov-
ered that many of his fellow-workers employed similar strate-
gies to his own game of work, something that he would pre-
sumably have discovered on his first day had this been an easily
observable feature of their conduct. So here we have an exam-
ple of a genuine action (or sequence of actions) which are
partly overt in so far as Roy’s fellow-workers, not to mention
his supervisor, can clearly see that he is operating a clicker
machine, whilst also being partly covert in the form of his game
of work, a combination that presents special difficulties as long
as one sticks to the conventional sociological approach to the
identification of action6.
Identifying Action
A long-standing and stubborn problem at the heart of action
theory has been that of how to successfully identify the unit act
and in identifying it describe it correctly. Weber himself as-
sumed that this could be achieved by a process of “direct ob-
servational understanding” (1964: p. 95) an assumption that is
still prevalent in sociology today. And yet, as Schutz has ob-
served, only the actor is really in a position to give an accurate
description of what he or she could be said to be “doing” (1973:
p. 22). Certainly if we applied Weber’s famous formula to
Roy’s conduct we would be forced to conclude that he was
“operating a clicker machine” much as Weber confidently de-
clared that the woodcutter was “chopping wood” (an assertion
that has a suspiciously tautological ring to it). But we know
from Roy himself that this is an inadequate or at least incom-
plete description of what he was doing. Where therefore a con-
sideration of Roy’s experience appears to lead is to a reconsid-
eration of the difficult problem of how to successfully identify
and describe action. Clearly the danger with the conventional
approach is the error of assuming that an individual’s action
consists of merely that portion which is overt and observable.
But of course not only may this not be equivalent to all that an
individual is “doing” it may not even be the portion that con-
tains “subjective meaning”, being performed habitually, as is
currently the case for example with my typing of this article;
for it is the covert portion—my thinking about what it is that I
am typing—that represents the section of my conduct that
really does possess “subjective meaning”. As long therefore as
sociologists consider it important to study action, as opposed
that is to mere behaviour, there will always be a crucial issue
that needs to be decided: notably, which portion of an individ-
2In principal phenomenological sociology, given its derivation from a phi-
losophical tradition that focused on the description of experiences, ought to
b
e a theor eti cal p ers pect iv e th at co ul d in deed reg ard Ro y’s g ame o f wo rk as
a phenomenon worthy of study. In reality however this strand of theorising,
as it has influenced sociology, has tended to focus on shared knowledge and
experience (see Berger & Luckmann, 1966).
3The decision-making that comprises the basic subject matter of rational
action theory could be said to be largely covert in character since it pre-
sumably consists of mental processes. However, as noted, this theoretical
perspective does not study real phenomena as such choosing to work with
models of human action.
4As suggested below Weber could be said to supply real examples of covert
action in what might be called his applied work, such as The Protestan
t
E
thic and Spirit of Capitalism, even if he doesn’t do this in his brief meth-
odological statement.
5Schutz also observes that action may be covert and he gives the example o
f
someone attempting to solve a scientific problem mentally. However he
seemsto forget about this example in his subsequent discussion where he
follows Weber’s lead in assuming that covert action equals inaction (1973:
p. xxxiv an d 20).
6In fact most true actions are likely to be partly covert in the manner o
f
Roy’s game. That is to say some crucial part of the action is likely to be
occurring intra-subjectively in the form of cognitions, feelings or imagin-
ings that are not apparent to an observer. One of the examples that Weber
gives to illustrate understanding an action in terms of motive illustrates this
p
oint. He writes that we can understand the action of an individual “w r it[ing
down the proposition twice two equals four” in the course of “balancing a
ledger” (1964: p. 95). But of course only part of the relevant action here is
overt: the actual writing in the ledger. The calculation itself is covert, since
it occurs in the accountant’s mind, not on the paper in front of him.
356

C. CAMPBELL
ual’s conduct it is that is actually possessed of subjective
meaning.
But then this does raise the question of quite how the term
“subjective meaning” itself should be interpreted given its key
role in serving to distinguish action from behaviour. Conven-
tionally this is understood to refer to the intentions (or possibly
goals or motives) that are presumed to be manifest in the action
in question, and indeed it is usually the nature of the intention
that serves to define the action, as in the claim that the wood-
cutter’s action consists of “chopping wood”. However, when
we turn to consider Roy’s activity we see that there are over-
lapping intentions underlying his conduct. His initial decision
to take the job of a clipper was because of his intention to en-
gage in participatory observation. On the other hand his game
of work was undertaken with the intention of combating bore-
dom and fatigue, while looking even more closely at his be-
haviour one could say that any one discrete action Roy per-
forms is undertaken with the intention of completing the current
game. Clearly one crucial difficulty with using intentionality as
the principal means of defining action is that it is unclear where
precisely the parameters should be drawn, given that conduct
commonly embodies a variety of what might be considered
overlapping intentions, with the consequence that it becomes
difficult to know precisely how the action should be defined.
Put succinctly one can say that a single act many embody many
intentions, just as a single intention may guide many different
actions. Hence in this case one could describe Roy’s actions as
a) playing the game of work; b) working as a clicker; c) under-
taking participant observation7.
Yet a consideration of Roy’s case can help resolve the prob-
lem of discriminating between multiple intentions. For although
Roy is clearly doing (or performing) the job of a clicker, some-
thing he has undertaken in order to be in a position to engage in
participatory observation, it would be more accurate to say that
he is “playing his game of work” for that is what he is actually
attending to. Hence attentionality can provide the key to dis-
criminating between intentions and therefore between alterna-
tive descriptions of action. What is more, employing attention-
ality in this way has the considerable advantage of facilitating
the process of distinguishing between action—viewed as con-
taining subjective meaning in the sense of being “meaningfully
oriented” (Weber, 1964: p. 116)—from habitual conduct. For
while these cannot be distinguished in terms of intentionality
(most habitual conduct manifests intention), they can be dis-
tinguished in terms of attentionality, since to attend to some
aspect of one’s conduct (that is to perform it “attentively”) is by
definition to imbue it with meaning, while to act inattentively is
to act “without giving any thought to what one is doing”. At the
same time attention is a faculty that is limited by definition (it
is difficult to “attend to” more than one thing at a time), whilst
there are no limits to the number of intentions that can be sub-
sumed by any action or programme of actions. Consequently
what Roy’s example shows is that, succinctly expressed, we
can say that what someone is “doing” at any one time is what
they are “attending to”8. This does not mean of course that the
actor is unaware of the other intentions that are also guiding, or
perhaps one should say framing, the action in question. Hence a
more precise definition of action would also include these. Thus,
in Roy’s case we could say that his action consisted of “playing
the game of work so that he could continue to perform the role
of working as a clicker”; or even “playing the game of work so
that he could continue to perform the role of working as a
clicker in order that he could carry out his research”; this par-
ticular nesting formulation of the nature of action being a more
truthful manner of describing the conduct of actors than the
more conventional, single-intention style definitions.
Essentially the problem of defining and delineating action
comes down to the issue of how the continuing flow of human
conduct is to be broken up into meaningful units, and there
would seem to be only two feasible ways of doing this. The
one—which currently dominates in the discipline—is to rely on
situational or conventional institutional guidance in order to
identify actions, and it is by employing this approach that it is
possible to identify given classes of actions, such as greetings
for example, or specific tasks, such as Roy’s “working as a
clicker”. But the drawback with this approach is that it may not
correspond to that which possesses subjective meaning for the
actor concerned. This is not to say that the actor does not rec-
ognise that her activity has this meaning (that is to say has a
conventional or socially agreed name or denotation), but that,
when viewed from the actor’s perspective, it lacks “meaning-
fulness”. And the only way to ensure that conduct is divided up
in a manner that accounts for this dimension is to employ the
criterion of attention. It is however important to note that such a
position implies recognition of the fact that which sections of
an actor’s conduct can be judged to constitute action and which
behaviour (setting aside involuntary respondent behaviours)
may change over time as attention is switched between differ-
ent aspects of an actor’s overall “doings” 9.
Agency at Work
Having noted how Roy’s conduct poses questions for the
conventional approach to defining action it is time to observe
that his game of work is actually more intimately related to the
issue of agency than action per se. Or at least it is to that par-
ticular conception of agency that defines it as “the capacity for
willed (voluntary) action” (Marshall, 1994: p. 7) or “the so-
cioculturally mediated capacity to act” (Ahearn, 2001: p. 112)
or what Campbell (2009) refers to as type 1 agency. For one of
the most interesting features of Roy’s game of work is the rea-
son why he invented it in the first place, which was in order to
maintain his power of agency; that is to say, his ability to initi-
ate and carry through a course of action. This observation
points to a crucial feature of human conduct that sociologists
rarely acknowledge: the fact that action can fail. This basic fact
of life is ignored by virtually all forms of action theory, with
the actor’s ability to perform actions treated as a taken-for-
granted premise; for, as Stephen Fuchs notes, “theorists assume
that persons ‘have agency’” (2001: p. 27). So it is important to
remember that actors routinely embark on programmes of ac-
tion—as Roy did when he signed on as a clicker—only to find
7Schutz makes this point arguing that Weber grossly oversimplified the task
of eliciting the intended meaning of an act given that this changes over time.
“One cannot... speak simply of the intended meaning attached to an action...
to become fully meaningful, it requires a date index specifying the moment
of the meaning-interpretation” (1967: p. 65).
8It may be of course that an individual is not “acting” at all. That is to say
they may not be ‘attending to’ anything (for example if they are day-
dreaming or simply on auto-
p
ilot), even though they can still be seen to be
“behaving”.
9It is also impor tant to reali ze that which sectio ns of an indiv idual ’s con duct
constitute action and which behaviour also change over time biographically,
processes normally referred to under such headings as “learning” or “ha-
bituation”.

C. CAMPBELL
that they are struggling to carry them out, given that there are
inherent inescapable limits to any actor’s power of agency
(Campbell, 2011). To that extent action is necessarily a contin-
gent phenomenon, dependent for its accomplishment on the
presence of prerequisites, the nature of which need to be exam-
ined. It is an interesting observation on the state of contempo-
rary sociology that the ability of individuals to interact suc-
cessfully is not take-for-granted but subject to a detailed an
extensive programme of research; so why, one wonders, is their
ability to ac t not subje ct to a similar programme?
Roy recognised after only a short period in his new job that
he might have to abandon it because boredom and fatigue were
sapping his will to carry on. Had he done so this would have
amounted to a failure of agency. He therefore took steps to
maintain, or repair, this faculty by inventing the game of work.
This, as he said, had the effect of imbuing his actions with
meaning, or perhaps more accurately with meaningfulness since
they already had a formal meaning in the sense of being a link
in the chain of actions that resulted in the creation of a manu-
factured product. What turning them into a game achieved was
to relate them directly to his own system of needs, providing
some personal satisfaction and hence the motivation to continue
with his allotted task. What is of special interest about this ex-
ample is that it illustrates the manner in which actors can use
their ability to manipulate meaning to control their behaviour.
As we can see in this instance the degree of freedom that indi-
viduals possess to use their meaning-manipulating skills to
define their immediate environment is not necessarily at odds
with the more formal social definitions of situations with which
sociologists are familiar. In that respect Roy was not defining
his situation differently from the way in which others defined it:
he was still working as a clicker. He had merely constructed a
purely personal action-related definition of his situation, one
that could as it were sit inside his formal social situation. It is
also important to note, in view of the prevailing tendency in
sociology to judge all cultural phenomena to be not merely
social but shared, that Roy’s game was his own invention, and
although no doubt he applied some highly general ideas about
the nature of games in order to construct it, it still remained his
game, one entirely invented by him.
This example is also revealing when considered from the
perspective of the other interpretation given to the concept of
agency, that which stresses its meaning as the “capacity of in-
dividuals to act independently of structural constraints” (Aber-
crombie et al., 1984: p. 6) or “independently of the determining
constraints of social structure” (Jary & Jary, 1991: p. 10).
Roy’s conduct as a clipper was clearly `constrained’ by social
structure in the form of the occupational role he occupied, set as
it was within the formal institutional context of a factory work-
place. And yet he managed to innovate in this context all the
same, not merely renegotiating the meaning of the activity that
he was required to undertake but actually re-arranging the order
in which its individual actions were performed. In other words
Roy succeeded in creating a space for himself, one in which he
could be creative and express something of his personality,
within the context of the highly structured role that he had no
choice but to act out. No choice that is once he had committed
himself to it. For of course the constraint that Roy experienced
in this connection was self-imposed, having entered into it vol-
untarily. He was not, as some of his fellow workers may have
been, effectively forced to work as a clipper in a factory by the
need to earn a living. However, having made the decision to
adopt this role, his freedom to act was then powerfully con-
strained. In other words it was as a consequence of the manner
in which he exercised his power of choice in the past that he
subsequently experienced considerable constraint on his free-
dom to act in the present. This is a significant observation be-
cause it shows how agency is an essentially self-limiting faculty
and that actors will always experience some constraint on their
freedom as a direct consequence of their very exercise of that
freedom. Indeed what Roy’s experience in this respect demon-
strates is the need for an adequate theory of action to go beyond
simply focusing on how individuals exercise their ability to
make choices, together with their reasons and motives for doing
so, to include how they manage to accomplish actions once
they have embarked on them. For it could be argued that for
many people the central problem they face in their lives is less
what choices to make than how to successfully accomplish
those programmes of action to which they are already commit-
ted.
Agency in Action
The above discussion helps to shed light on what has long
been recognised as something of a problem in sociology theory,
namely the relationship between agency and action. If we use
Roy’s experience as a guide we can say that while it is neces-
sary to possess the power of agency in order to be able to un-
dertake actions it is also sometimes necessary for an individual
to undertake actions with the express purpose of maintaining or
restoring the power of agency. What is distinctive about this
particular class of actions, which for convenience we might call
agency management or control actions, is that they tend to take
the form of creating goals or outcomes that in turn help to sup-
ply the motivation necessary to initiate or complete a pro-
gramme of action, and as such the normal process of explana-
tion is reversed. For conventionally one explains an action by
seeking the goal (or sometimes the motive) that underlies it, as
Weber does when he says that “we understand the chopping of
wood… if we know that the woodchopper is working for a wage
or chopping a supply of firewood for his own use” (1964: p. 95).
But understanding agency management actions requires that
one explains the goals actors have selected in terms of the ac-
tion that they serve to maintain, as for example when Roy says
that he “created a sequence of short-range production goals”
(1959: p. 161) for himself by inventing his game of work. In
other words agency management actions all have the same goal,
which is to maintain or increase the power of agency. Hence in
these cases one explains goals in terms of actions rather than
vice versa.
But then understanding agency management or control ac-
tions also requires that one abandons the logical or cognitive
bias that for so long has dominated the investigation of mean-
ing in action theory. Traditionally sociologists have interpreted
Weber’s recommendation that one should understand actions
by placing them in a more inclusive context of meaning as im-
plying that they should be considered as if they were connected
rationally, principally in a means-end manner, as in the quote
given above concerning the woodcutter’s actions. Yet in order
to make sense of Roy’s invention of the game of work contex-
tual understanding requires the application not of a means-end
framework but of a dynamic or behavioural one; specifically an
appreciation of the role that actions can have as stimuli that
produce a gratifying response. Indeed it is important to note
358

C. CAMPBELL
that there is no logical connection between Roy’s game and his
activity as a clicker (in the way that it is sometimes claimed that
beliefs and values are logically connected with actions). The
connection is a dynamic or behavioural one in which Roy’s
game has to be seen not as a means to an end but rather as an
end in itself. That is to say as an autotelic activity, one per-
formed for the intrinsic satisfaction it provides. Unfortunately
the application of such a behavioural model has, from the very
beginnings of action theory, been seen as at odds with both
voluntarism and the adoption of the actor’s frame of reference.
And indeed Weber himself refers to understanding action that is
a response to a stimulus as “affectually determined and thus in
a certain sense irrational” (1964: p. 95, italics added). Yet it is
clear from Roy’s example that this is not necessarily the case,
as in this instance a behavioural perspective is included in the
subjective standpoint employed by the actor. Consequently in
understanding the action in question as constituting a stimu-
lus-response pattern one is not stepping outside the actor’s
frame of reference. This is because Roy reveals himself to be
something of a lay behaviourist, knowing full well what kind of
stimuli he needed to invent in order to produce the desired re-
sponse in himself. Indeed it was because Roy understood how
his behaviour was controlled by stimuli that he was able to
exercise voluntary control over his actions. It follows from this
example that sociologists are wrong to assume that a behav-
iourist-style analysis is at odds with an interpretivist and vol-
untaristic approach to the study of action. The reason they are
not incompatible is because human beings are different from
animals. For one thing they can use knowledge and foresight to
re-arrange their environment so as to adjust the stimuli to which
they are exposed. But then crucially humans do not so much
respond to stimuli as respond to their meanings and because
individuals have some control over these meanings it is possi-
ble for them to use this control to influence their responses, as
Roy did in this case. The movements that he was required to
undertake in order to fulfil his role as a clipper are the stimuli in
question, together with the objects themselves—principally the
pieces of plastic—that he was required to manipulate. By using
his own capacity to create meaning to impose a pattern on these
movements as well as inventing a set sequence in which the
plastic pieces were processed Roy managed to change the na-
ture of these stimuli and hence his own response to the activity.
So we can see that Roy’s actions in inventing the game of work
followed first from his monitoring of his own behaviour, and
his clear recognition of the various behavioural cues that indi-
cated fatigue and boredom on his part, and then second, from
his understanding of the nature of the stimuli he needed to cre-
ate in order to alter his own behavioural condition in a manner
that would produce a re-energised state.
It can be seen from this that there is no contradiction between
an interpretive analysis of the meaning of an actor’s conduct
and a behavioural analysis in terms of cause and effect. To
suggest, a some sociologists have done, that sociology should
be concerned with “understanding action rather than observing
behaviour” (Silverman, 1970: p. 126) is to fail to recognise that
actors are observers of their own behaviour and, as lay behav-
iourists, endeavour to control their environment, and hence the
stimuli to which they are exposed, in order to control them-
selves. It is also to overlook the fact that while it may be true to
suggest that a full-blooded behavioural perspective has no room
for an interpretive analysis of meaning the reverse is not the
case, if only because action is actually a form of behaviour10.
But then given that what individuals do has always comprised
not simply a combination of action and behaviour but an inte-
grated pattern of conduct in which these two forms are inti-
mately interlinked it was always mistaken to imagine that one
could successfully be examined without reference to the other.
The value of a dynamic interpretive approach to the study of
human action therefore is that it involves recognising that ac-
tors regularly manipulate both stimuli and their meanings in
order to control their conduct.
Objections
There are two possible objections that might be raised
against the suggestion that a consideration of Roy’s account of
his experience as a clipper could form the basis for a revised
approach to the study of action and agency. One is that the
activity that he describes engaging in is relatively rare and
hence not significant enough to warrant its use as a justification
for a novel theoretical perspective, while the second is that the
phenomenon concerned, although of interest, really belongs to
the discipline of psychology rather than sociology. However, a
little reflection reveals that neither of these really has much
force. In the first place it is clear that Roy’s experience was not
in any way unique. Indeed as he notes he subsequently discov-
ered th at many of h is fellow workers h ad devise d simila r, if not
necessarily identical, strategies to his own game of work in
order that they too might be able to perform their allotted roles
successfully, suggesting that the phenomenon in question is not
isolated or unusual but on the contrary commonplace. That is to
say this form of behaviour is something that people have in
common even if it is not something that they could necessarily
be said to share. Indeed it would seem highly probable that
agency control actions of this kind occur whenever and wher-
ever individuals find that their ability to successfully initiate or
accomplish actions is in doubt11. Then of course there is the
interesting question of the cultural resources available for actors
to draw upon in order to re-define or re-negotiate both the
meaning of their own actions and the stimuli they encounter in
their environment. Where this material comes from and how it
is employed in the processes of action accomplishment is
clearly a matter of some interest to sociologists. But then what
also suggests that this should not be regarded as by any means a
purely psychological phenomenon is the fact that actors do not
simply use their ability to manipulate meaning in order to ac-
complish action in small-scale or short-lived sequences of ac-
tion. For the manipulation of meaning in order to sustain
agency and hence the accomplishment of programmes of action
can occur at any level, from the briefest of actions up to a com-
plete programme for life. Weber himself supplies examples of
the latter in the course of his discussion of the problems that
Calvinists faced in their determination to stick to their excep-
tionally demanding regime of ascetic self-control. These efforts
were, as he suggests, consistently undermined by doubts about
10To contrast action with behaviour is not to imply the necessity of employ-
ing contrasting descriptive schemes or paradigms (let alone incompati
b
le
epistemologies and methodologies); it is merely to refer to empirically
differentiated forms of c onduct.
11This is not to say that such individualistic and covert or co-overt actions
are the on l y mean s of co mb at i n g threats to agency. In d eed the main t h rust o
f
Roy’s article is his account of how his fellow-workers combated their ex-
p
eriences of boredom and monotony by developing their own interactive
“games”. Th es e were no t h o wever, as he notes, in p l ace of co v er t d ef i ni t io ns
like his game of work but rather in addition to them.

C. CAMPBELL
their status as members of the elect. In other words their ability
to act was threatened, not by boredom and fatigue as in Roy’s
case, but by paralyzing panic, despair and “the anxious fear of
death” (Weber, 1965: p. 107). Consequently they were advised
to view any doubts they experienced as the work of the devil
and exhorted to consider themselves saved in order to generate
the self-confidence necessary to cope with the “daily struggle
of life” (ibid, 111). In other words, just like Roy, they worked
to impose a definition upon the situation in which they found
themselves that served the function of maintaining their power
of agency. All of which suggests that in arguing for the need to
study essentially intra-subjective or covert processes, the most
that could be claimed is that one is recognizing sociology’s
necessary dependence on psychology rather than actually re-
ducing the former to the latter.
Methodological Significance
One final observation worth noting is that a dynamic inter-
pretive approach to the study of action carries with it a clear
implication concerning the appropriate method for studying
action, one that counters the skepticism that has so often been
expressed concerning the accuracy of actors’ reports of their
conduct, a skepticism that has led some sociologists to claim
that self-reports can tell us little or nothing about the events
they purport to describe (Mills, 1940; Semin & Manstead, 1983;
and Gilbert & Abell, 1983). But then, in line with the argument
advanced above, there is bound to be a disjunction between
people’s actions and their reports of them if anything other than
the actors’ own descriptions are employed for the purposes of
identifying what counts as their actions. The particular advan-
tage of a dynamic interpretive approach that employs an atten-
tional criterion for identifying action is that this means there are
excellent reasons for believing that actors are able to give full
and accurate descriptions of what they are doing. The reason
for this confidence is that people generally know what it is that
they have been attending to or consciously doing. Whilst they
may well find difficulty in reporting some aspects of their be-
haviour and perhaps in some instances also find it difficult to
specify exactly what their intentions, motives or purposes might
have been, they will always know what has been the focus of
their consciousness. For as White says, “to ask a person what
he (sic) is doing now is usually to ask what is at present occu-
pying his (sic) whole energies, that is what he (sic) is con-
sciously doing. It is, therefore, a question which he (sic) can
immediately answer” (1964: p. 70)12. Individuals will, for the
same reason, find it relatively easy to report on past actions,
provided that is they are not too far in the past. This is because
there is ample evidence to suggest that people remember what
they attend to whilst correspondingly forgetting that which they
never attended to properly in the first place. This is a maxim
that applies to one’s own conduct just as much as to events in
the external environment. Thus it is, once again, that by defini-
tion, actors should have little difficulty recovering details about
their former actions whilst being unable to report very much
about their previous behaviour13. It follows from this that
studying action requires the investigator to always ask respon-
dents what it is that they are or have been attending to, or con-
centrating on, and never prejudge the nature of their actions by
naming them in the questioning process itself. They can then be
confident not only of the respondent’s ability to answer but also
that the answer will itself be a true description of the action in
question.
Conclusion
One sociologist’s personal account of the way in which he
resolved a difficulty encountered in the process of undertaking
research has been taken as the material for a reexamination of
the way in which both action and agency are conventionally
conceptualized within sociology. Although not normally in-
cluded in the subject-matter of the discipline it is argued that
personal covert, semi-covert (or co-overt) activities, such as
Roy’s game of work, represent a vital dimension of all actions
or programmes of action; indeed that they frequently constitute
the essence of subjective meaning and hence not only that their
study should be included in any genuine theory of action but
that their significance for the successful accomplishment of
what are conventionally identified as social actions needs to be
acknowledged. However it is argued that it is necessary to rec-
ognize, if this crucial phenomenon is to be the subject of socio-
logical analysis, that action can only be adequately identified
and defined by the actor and consequently should be equated
with that portion of an actor’s conduct to which he or she is
paying attention, the employment of an attentional criterion
helping to resolve many of the ambiguities that are inherent in
the use of intentionality alone. Approaching the study of action
in this way also helps to shed light on the manner in which
agency and action are interrelated, focusing attention on the
way in which actions help to maintain agency as well as agency
actions, the distinctive category of agency control actions being
the fact that goals are explicable in terms of the actions they
serve to maintain rather than vice versa. This then in turn points
to the necessity, if the manner in which action is accomplished
is to be fully understood, of applying a dynamic behavioural
analysis, one that focuses on stimuli and their responses—to-
gether with motivational structures in general—rather than
simply the eliciting of meaning. However this does not imply
any rejection of the actor’s perspective, let alone any abandon-
ment of the assumption of voluntarism, given that the nature of
this dynamic understanding is derived directly from the actor’s
own account.
Weber described sociology as a science that “attempts the
interpretive understanding of social action in order thereby to
arrive at a causal explanation of its course and effects.” (1964:
p. 88), yet sadly the history of the discipline suggests that this
prescription has not been followed with “interpretive under-
standing” and “causal explanation” effectively going their
separate ways. The consequence of this division has been that
the ability to act voluntarily has been equated with the freedom
to make choices or construct meaning and hence with thought
and reasoning, whilst emotion, effort and imagination have
been allotted little or no role in this process, being seen as pri-
marily associated with deterministic assumptions about the
nature of human conduct. Yet as the example of Roy’s game of
work clearly shows the ability of individuals to act is just as
dependent on the capacity to emote and to imagine as it is upon
12This is still true even if an individual has not been acting at all, as they
should have little difficulty in reporting this fact, via some formula such as,
“Oh, I wasn’t doing anything, I was just day-dreaming.”
13As Stendhal observed, “
I
t is very difficult to describe from memory wha
t
was natural in your behaviour; it is easier to e voke what was art ificial o
r
affected s ince the effo rt needed to put on an act also engraves it in mem-
ory.” (Elster 1983: p. 52. Italics in original).
360

C. CAMPBELL
the ability to reason; a mistake that sociologists would not have
made had they concentrated on studying action-making instead
of decision-making or meaning-making. That is to say if they
had taken the actor’s definition of the situation rather than the
conventional social definition as their starting point and con-
centrated on studying the “meaningfulness” rather than simply
“the meaning” of action. A dynamic interpretive approach to
the study of action aims to redress these deficiencies by seeking
to understand action not by placing it in a larger framework of
meanings, but rather through an examination of its constituent
elements, and hence to shed light on how precisely individuals
really do ma na ge to act.
REFERENCES
Abercrombie, N., Hill, S., & Turner, B. (1984). The penguin dictionary
of sociology. London: Penguin Books.
Ahearn, L. M. (2001). Language and agency. Annual Review of An-
thropology, 30, 109-137. doi:10.1146/annurev.anthro.30.1.109
Alexander, J. C. (1992). Recent sociological theory between agency
and social structure. Schweiz. Z. Soziol./R ev. Sui sse Soc ial., 1, 7-11.
Berger, P. L., & Luckman, T. (1966). The social construction of reality.
Harmondsworth: Penguin.
Campbell, C. (2009). Distinguishing the power of agency from agentic
power: A note on weber and the “black box” of personal agency. So-
ciological Theory, 27, 407-418.
doi:10.1111/j.1467-9558.2009.01355.x
Campbell, C. (2011). Limits to agency: Exploring the unintended (and
unattended) consequences of action. In A. Mica, A. Peisert, & J.
Winczorek (Eds.), Sociology and the unintended: Robert merton re-
visited (pp. 45-62). Frankfurt am Main: Peter Lang.
Elster, J. (1983). Sour grapes: Studies in the subversion of rationality.
Cambridge: CUP.
Fuchs, S. (2001). Beyond agency. Sociological Theory, 19, 24-40.
doi:10.1111/0735-2751.00126
Gilbert, G. N. & Peter, A. (1983). Accounts and actions: Surrey con-
ferences on sociological theory and met hod s. Aldershot: Gower.
Jary, D., & Jary, J. (1991). Collins dictionary of sociology. London:
HarperCollins.
Marshall, G. (1994). The concise Oxford Dictionary of sociology. Ox-
ford: OUP.
Mills, C. W. (1940). Situated actions and vocabulary of motives.
American Sociological Review, 5, 904-913. doi:10.2307/2084524
Roy, D. F. (1959). Banana time: Job satisfaction and informal interac-
tion. Human Organization, 18, 158-168.
Schutz, A. (1967). The phenomenology of the social world. Evanston,
IL: Northwestern University Press.
Schutz, A. (1973). The problem of social reality. The Hague: Martinus
Nijhoff.
Silverman, D. (1970). The theory of organisations. London: Heinemann
Educational Books.
Weber, M. (1964). The theory of social and economic organization.
New York: The Free Press.
Weber, M. (1965). The protestant ethic and the spirit of capitalism.
London: Unwin University Books.
White, A. R. (1964). Attention. Oxford: Blackwell.