Understanding Agency: Social Theory and Responsible Action

by S Barry Barnes

In this penetrating and assured book, one of the leading commentators in the field argues that social theory is moving in the wrong direction in its reflections on human freedom and autonomy. It has borrowed notions of ′agency′ and ′choice′ from everyday discourse, but increasingly it puts a misconceived individualistic gloss upon them. Against this, Barnes unequivocally identifies human beings as social agents in a profound sense, and emphasises the vital importance of their sociability. Notions of ′agency′, ′freedom′ and ′choice′ have to be understood by reference to their role in communicative interaction; they are key components of the discourse through which human beings identify each other, and have effects upon each other, as soci

176 pages, Hardcover

First published January 1, 1999

Reviews

[Sharad Pandian · Rating 5 out of 5]

Barry Barnes has quickly become one of the most interesting and important thinkers for me - although he's a theoretical sociologist mainly associated with the Edinburgh school of the Sociology of Scientific Knowledge, he engages with a broad range of scholarship, including philosophers and historians. This allows him to draw from and make important contributions to multiple areas, overturning presumptions of individual fields. This is a short book (~150 pages), but it's excellent and deals with so much that I'm simply going to quote as much as possible here (for future reference) until my character limit runs out:

Part I

Arguing that the model of individuals who reason solely through their own solitary powers of reason is fundamentally flawed, since it does not account for how socially-coordinated action would be possible, he argues instead for a model of human beings where we are accountable and mutually susceptible.

Where there is social order on a large scale there is agreement in practice amongst participants in the order. But that agreement is not the happy coincidence of the individual practices of appropriately socialised or informed agents, nor the fortunate consequence of their separate access to identical rules. It is rather the creation of agreement out of difference as a continuing ubiquitous project. The agents who successfully engage in this project must be actively oriented to any (instances of) rules they relate to. But they may not be independent individuals with no particular inclination to co-ordinate their actions with others, and no incentive to engage in the communicative interaction necessary to keep shared rules shared. Social agents are necessary here, agents with a prior non-rational inclination toward agreement and coordination, agents who by virtue of this inclination possess collective agency. There is no place here for agents who will interact only when they have calculated that it is profitable, and make that calculation only if that also is to their benefit, and consider the extent of that benefit only. (56)

In brief, the basic competence members presume in each other is that of accountability, of being able to give an intelligible account of what has been done. This basic mutual expectation informs communicative interaction, and propels it along as a series of matter-of-course exchanges, not as the calculated means to any particular end. (65)

And as the co-ordination being spoken of is the condition of continuing mutual intelligibility, the interaction that secures it cannot be understood as a rational dialogue wherein mutual intelligibility had already been achieved and is no longer a problem. It is necessary here to understand the relevant interaction causally, and to recognise that members affect each other therein in a causal sense. Agents who are disposed to co-ordinate their understandings with others have to be agents who are affected by others. Mutual accountability implies co-ordinated understanding, which implies agents who affect each other. It implies, we might say, mutual susceptibility.(67)

But if voluntary actions have causal antecedents sufficient to account for them, what is it about them that makes them voluntary?...kinds of actions designated as voluntary are generally those we believe to be modifiable by symbolic intervention, perhaps through the operation of the deference-emotion system. If this is correct, then in the last analysis violent assaults are commonly regarded as voluntary actions because verbal intervention, persuasion, calming the person down, and so forth, offer hope of forestalling them, even where men are involved.

With this background, he goes on to think about agency. He points to how for any action, there seem to be both causal stories as well as those about agents' choices. He points to how some people try to model the statuses of agency and responsibility (in judicial and as well as philosophical contexts) as existing if appropriate brain/psychological states are present, but argues that the positing of invisible internal states is subject to "back-pressure" from whatever status we want to endorse ("The move that is supposed to be made is from state to status, but it is clear that strong back-pressure, to say the least, exists from status to state."). If status' cannot be grounded in the presence of states, what does agency mean? Focusing on how it's connected to the sense that an actor could have acted otherwise, choice is attributed when it is believed that "the action was performed by a normal, susceptible agent, and that it was accordingly an action possibly modifiable by symbolic intervention."(70)

Part II

With this background, Barnes extends his framework to a number of fields/debates:

-On Weber's treatment of class-based societies as distinct from status-based societies:

Theorists have been slow, and often deeply reluctant, to take full account of how most instances of supposed 'class action' are actually products of the collective agency of status groups or of localised interacting communities, even though this has long been evident from the work of historical sociologists. (83)

-On bureaucracy [a constant emphasis of his is against talking about "the system" as though it were present independent of the humans who sustain it]

It is interesting to notice how, over time, sociological research has identified more and more difficulties for the individualistic elements of Weber's account. It has demonstrated the need for power and discretion to reside at every level of the hierarchy, to overcome information bottlenecks, to deal with the 'unforeseen circumstances' that are actually normal circumstances in all bureaucracies, and to carry out the 'staff' functions that Weber neglects as well as the 'line' functions he highlights. The need for discretion in the interpretation and application of supposedly 'impersonal' rules is especially obvious in work of this kind. It makes it clear as well that bureaucrats have the capacity to act instrumentally for their own collective good as bureaucrats, a capacity strikingly documented in study after study of 'bureaucratic politics'. And it is evident from the same studies that bureaucratic and administrative organisations are not so much mechanisms as cultures, which opens the way to seeing them not only as instruments of rationalisation but as targets of it. In general, empirical accounts of bureaucratic administration are consistent with its being the accomplishment of responsible agents, not isolated individuals. (86)

-On Habermas

Habermas' famous contrast of 'system' and 'lifeworld' is, of course, the basis of a powerful political vision. 'System' menaces 'lifeworld'. In the form of ever-increasing instrumental control and regulation by administrators and technical professionals, it threatens to encroach on a 'lifeworld sphere' currently the realm of communicative action and human agency, and to 'colonise' it. This theory has obvious resonances with the everyday stereotype of 'the system' that presses upon everyday life from without. And, like much sociological theory, its mythical form and implicit moral message are congruent with those of the everyday stereotype: 'system' / 'power ' opposes 'lifeworld' /'agency'; so that 'system' is easily cast as villain, and 'lifeworld' as hero, in a drama of conflict. It is indeed an unwritten rule of sociological theory that 'system' is evil and 'agency' good, just as it is a characteristic 'attribution bias' of ordinary members of society that externalities are evil and inner states are good. (87)

By switching whenever expediency requires between intentional/voluntarist and systemic/ functionalist accounts of action, in effect between accounts of human beings as social agents and accounts of them as impoverished asocial individuals, an illicit contrast of 'system' and 'lifeworld' is conjured into existence, albeit one that can claim spurious justification from its correspondence with much informal thinking. (88)

-On the "de-traditionalisation thesis"

What will be suggested here is in some ways the opposite of the current view. The proposal is that differentiated societies are remarkable for how little they rely upon individual responsibility in the usual sense (although the attribution of individual responsibility remains as important as ever), and how much and how successfully they have come to rely, not upon collective or group responsibility as usually understood, to which they are indeed averse, but upon institutional responsibility, and in particular institutional response. (93-94)

References to a development of this kind are implicit in the familiar story of the emergence of specialised occupational roles in modem societies, and the proliferation of administrators, skilled professionals and technical experts. Many responsibilities hitherto located in the family have been passed to them: most of the responsibilities attendant upon birth and death are in their hands, as are those relating to the health, education and basic well-being of children; and they share responsibility, so to speak, for the basic economic provision in family units. (94)

In the docile, normalised collectives of modem societies, life is more peaceful, predictable and riskfree than almost ever before. Because of the shift of liabilities to institutions, the individual costs of acting in one way rather than another have been radically reduced. In comparison with the life-or-death consequences of many decisions that have to be made in simple societies, about when precisely to sow seed or where to trek to find good grazing, for example, the penalties exacted for unfortunate individual decisions in modem societies are generally trivial. And it is just when alternatives differ little and the risk inherent in choice is consequently slight that individuals will be inclined to say that they 'have a choice' and to regard what they do as more than an obligatory response to circumstances. Then it is notable as no more than a feature of language that a society replete with choice and diversity may equally well be rendered as one of insipid uniformity, and vice versa. (95-96)

-On why causal stories are not allowed to replace choice-stories, even when we have statistics for (say) how men have a greater tendency for violence:

There are two fundamental options open to those who face difficulties coping with the expectations associated with a given status. One is to perform in status nonetheless, and to accept whatever in the way of honour and contempt, praise or blame, one's performance attracts, along, of course, with whatever more substantial sanctions might exist. The other is to take up whatever available causal stories will serve as the basis for a plea of impaired agency. A plea of this kind may be successful in modifying the expectations of others, and hence in bringing the individual more praise and less blame for a given performance. But to offset this gain there will be a corresponding loss. The individual who successfully claims impaired agency is liable to be moved to a lower, less honourable status - in the extreme case to lose the basic status of responsible agent altogether and be assigned some special lesser alternative...The dangers of blame diminish but those of stigmatisation increase. (115-6)

-On whether genetics destroys free will

What the new biotechnology is in the course of doing is identifying greatly increased numbers of such causal antecedents and implying the existence of many more, yet to be identified. In this way it is increasing our understanding of how individual human beings are prone to act, and our awareness of the extent of individual difference and diversity amongst human beings. But none of these revelations of difference need be taken as grounds for denying specific persons the licence permitting their participation in the institution of responsible action. And such participation precisely involves their social interaction through the medium of a voluntaristic discourse, discourse in which others orient to their actions as chosen ones for which they are accountable. Indeed, it is precisely interaction in this voluntaristic idiom that allows the orderly social life in the context of which we exist to be enacted by the remarkably diverse organisms that we are. (117-8)

-Extending "status" to all categorization

Statuses may be attributed to all manner of things and objects. Such attributions classify things not by anything internal to them but by what lies outside them - what is directed toward them, or attached to them, or associated with them. Indeed, what constitutes a status may not merely lie beyond the boundary of the object classified; it may be smeared across the whole of the collective which recognises it. That something counts as having a status is what constitutes its having the status. Status classifications have strange self-referring and self-validating properties that make them hard to understand, a problem which is intensified by their representing a radical departure from the accepted paradigm of classification in general use amongst us. Paradigmatically, we think of the classification of things as being based on their own internal, intrinsic properties, on aspects of their own nature as it were, not on aspects of the context in which they reside. But in truth this paradigm fails to represent not just some but most actual acts of naming and classification. It is commonplace to characterise even physical objects by reference to what surrounds them rather than what constitutes them, as with holes and ridges and cavities, valleys, lakes and islands. And it is even more commonplace to classify things by reference to function or utility, and hence by the orientation toward the things of the people who surround them, as with tools and furniture, doors and fences, crops and weeds, and of course social statuses of all kinds. Even in the natural sciences it is commonplace to classify by reference to the externalities of context (planets, parasites) and function (reagents, enzymes, counters, spectrometers). Just like ordinary members, natural scientists classify the things around them, and particularly the things they work with, in a profoundly teleological way, but this kind of classification is even less remarked upon in the context of science than it is in the context of everyday life. (148)

[Paul Ernest · Rating 5 out of 5]

Deep and important book.See longer review on this site.