Thornton, R. (1995). The Peculiar Temporality of Violence. Paper presented at the Centre for the Study of Violence and Reconciliation, Seminar No. 1, 29 March.
Robert Thornton is head of the Anthropology Department at the University of the Witwatersrand
Date: 29 March 1995
Venue: Centre for the Study of Violence and Reconciliation, Johannesburg, South Africa
Violence happens. This paper offers a radical exposition of this view. By "violence", I mean literal empirical behaviour that causes damage or death. By "happens" I mean it is an "event"; it takes place in time, and is assigned significance or meaning by observers, but I do not mean to imply that its eventuality is a consequence of other events or "causes". Violence "happens" means that when it occurs it is a significant event of empirical behaviour of groups or individuals that causes measurable damage or death to others. This may seem unexceptional, but it is not. The accounts we most often give of violence either imply or state explicitly that violence is the consequence of previous events or that it is the cause of other events. The more significant the "events" with which violence is associated, the more significant, or meaningful the happening of violence is held to be. Violence that "founds a nation" or that occurs between salient "races", "classes", "religions" or "groups" is assigned special significance because it is held to be either the cause of these particular social forms (nations, races, classes, religions, or groups) or that it is the consequence of conflict between them. It is this seemingly obvious truth that I wish to demonstrate is false. If it is true, then we must accept that violence has a legitimate, even necessary, role in human affairs; but if it is false, then we must seek to understand, first, why we have believed it to be true, and second, we must explain violence in some way other than as the (necessary) consequence of conflict between groups or individuals.
My argument hinges on the fact that all views of violence are retrospective. After they occur, people tend to see them as having been "necessary", or at least as playing a role in a political "process". This view looks backwards in time, and attempts to say why violence happened in terms of what was there before social structures, "structural violence", personal anger, vendettas, and so on. Before violence happens however, we are never able to say precisely
except under special or very unusual circumstances such as planned military battles or executions when, where or how violence will occur, or who will be involved. For instance, we can say with some certainty that a violent death will occur in Johannesburg in the next 24-hour period. We can, however, say nothing about when, where, or how it will occur, or who it will involve. All we state is certain "probabilities" within certain artificially constructed statistical "populations".
Thus, if retrospective accounts of violence appear clear and decisive, it is only after the fact, and never before that we can achieve such clarity. Before the fact, violence is always a "risk" or a "probability", but never empirically predictable or confinable to one moment, place, person or mode. This paradox is what I call the peculiar "temporality of violence", that is, that it is only clear that it happened, and only clear that it was really violence in retrospect. Violence, then, is statistically probable, but fundamentally unpredictable and is therefore "chaotic".
Nevertheless, the threat of violence and stories of past violence are always part of political discourse. Violence has a very powerful role in human representations of power, and in attempts to wield power over others. Indeed, we tend to see violence as a political instrument, that is, as an act which causes other people to do certain things, or to refrain from doing others. If violence is fundamentally unpredictable, uncontainable and chaotic, how can it function as an instrument? How is it possible that people generally believe that violence is not only causally efficient (that is, that it "works"), but also that it is the "original" or the final cause of social forms and actions? This belief in the efficacy of violence, in the light of the argument presented here, appears rather mysterious!
I argue then that it is not violence itself the act and its destructive consequences that is causally effective or the "instrument of power", but rather our narratives about violence that we construct after the event of violence. Unlike other planned or and emotionally charged social interactions (such as eating, sexuality, ritual) acts of violence interrupt (disrupt, breach, rupture, break, etc.) and terminate parts or all of previous social relations. The act of violence therefore requires that a new story be told to explain the loss, to account for the disruption, and to rebuild social relations after its occurrence. This makes violence appear to be located at the "beginning" of new social forms, new behaviours, and new accounts, and thus to appear as their "cause", but this is a false perception based on the peculiar temporality of violence itself, and its chaotic nature.
For instance, stories of identities (tribe, nation, volk, ethnic group, class, religious sect, etc.) often begin with acts of violence (the mfecane, Blood River, the fall of the Bastille, for instance) as if it were the violence that caused the coming into being of a new order, a new identity, nation, group or class. In fact, it is merely the fact that a new story must begin to be told after violence has happened that makes it appear that violence is the "origin" of the social group. Instead, violence should be seen as merely the condition for a new narrative of identity, but not its cause. Violence is not the social cause, but rather the random condition under which identities emerge.
Violence as social fact
Violence is a social fact of a peculiar kind. Most social scientists would agree that violence has a high emotional salience,1 and that episodes of violence, like wars or assassinations, provide the narrative pivots of histories of all kinds,2 that social power is exercised by means of it3 and that it remains deeply perplexing nonetheless.4 Most would assert, however, that violence is ultimately explicable.5
The aim of this paper is try to show, to the contrary, that events or episodes of violence have a special temporal character that make them necessarily perplexing, and that the perplexity that violence occasions is not due to "inadequate" social theory, but rather derives from the fact that violence is genuinely "original" or "emergent" in a fundamental sense.6
The peculiarity of violence as a social fact derives from its special temporality. Violence necessarily disturbs all structural, causal or narrative sequences and continuities. Since the kinds of explanatory or narrative accounts that virtually all of the social sciences and history often depend on some notion of spatial or temporal continuity, and since violence disrupts these, we can find in this fact some reasons to explain our perplexity about violence. And since the notion of social power depends crucially on the question of violence, these arguments will have a bearing on the concept of social power. In this paper, I am challenging this assumption of continuity of "structural" or "historical" causation with the proposition that violence, considered in relation to the continuities of time, place and social structure, violates precisely those continuities on which most accounts depend. Therefore, the same kinds of accounts that we use to explain, for instance, the flow of goods by gift and exchange or the structure of kinship or the constitution of political parties or mining companies must fail to explain violence.
The arguments I offer here are also, perhaps, more applicable cross-culturally and across historical epochs since the understanding of violence that I offer depends on what seem to me to absolute characteristics of time and social process rather than on more culturally-relative concepts such as the value of human life, the notion of personal autonomy or identity, the idea of the self, or of violation of personal space or boundary. While all of these issues are what make violence a pressing political issue today, they cannot be assumed to be human universals. This analysis, then, deals with the question of why violence is so often the "zero-point" of history.
The event of violence and any outcomes of violence are inherently uncertain from the point of view of any observer, and are therefore unpredictable. This "uncertainty principle" means that violence is only apparently explicable in retrospect. Invariably, however, explanations of violence are construed as if previous states of affairs caused the violent event they seek to explain. Instead, each episode requires interpretation each time. This accounts for the perennial perplexity of the social sciences in the face of violence. It does not, however, account for violence.
Violence and the attribution of causation: an unexamined problem
We usually see violence as the consequence of conflict or the as the instrument of power, but there are problems with both views. First, violence is unpredictable: social scientists can not predict when or where violence will actually occur. Second it is likely to have unintended consequences. Third, violence invariably interrupts social processes and regularities, on the existence of which all causal and structural accounts depend. Finally, most accounts do not distinguish between violence which is the result of causes — what we shall call "consequential violence" and violence which is the instrument of power or will that is "instrumental violence". It is possible that two ontologically different types of violence do exist, but since perpetrators or victims seem to experience violence as only one kind of behaviour — that is, as acts of violation, hurt or damage — such philosophical distinctions seem unnecessary. Instead, it appears that violence may in fact be essentially and necessarily unpredictable for a number of reasons.
If this is so, then we must abandon theories about social structures or kinds of social relationships which "cause" violence on the grounds that such causality can not be demonstrated, but only asserted. In other words, the accounts of violence which purport to show structural or social causes for violence are simply fictions that have more in common with other types of fictional narratives than they do with true empirical accounts. While fictional narratives have a crucial and central role in human culture and behaviour, they are, nevertheless, not the adequate causal, empirically demonstrable, accounts that they purport to be. Rather, they are like other naïve accounts of violence, post facto constructions based on culturally constituted and social transmitted "master" narratives.
This argument, if true, has important consequences for our understanding of the role of violence in the rather mysterious entity we call social or political "power". It also has consequences for the way we do social science and the way we may write empirically valid accounts of violence.
With respect to the ways in which we may write valid accounts of violence, we must recognise that the four features of violence, outlined above — its unpredictability, its productivity of unintended consequences, its emergence into and disruption of regular social processes, and the apparent divergence between two modes of explanation (as consequential or instrumental) notwithstanding the unity of experience — all suggest that the only kinds of accounts of violence that we may in fact have are accounts after the fact of violence. Violence then must be seen as a genuinely new or emergent phenomenon, and invariably disruptive of regular, institutionalised, customary or habitual behaviour. Whatever its cause, however, violence must receive an account after — but only if — violence actually occurs. In empirical terms, then, what we can know about violence is its cultural and social consequences, such as the way it is appropriated symbolically by the political process or dealt with in narrative and memorials, for instance. But we can only speculate about causes which are, in any case, rarely efficient, never unitary, and very often not recoverable from evidence or memory.
Both naïve accounts of violence, and those of the philosophers and social scientists, however, agree that violence may be both instrumental and consequential. In view of the observations that have already been made, this divergence is likely to be a property of the narrative genres according to which violence is given account, or to the modes of its symbolic appropriation into other kinds of cultural formulae or frames. Indeed, the maintenance of such distinctions in our "ordinary", culturally approved, accounts is necessary for the way violence may be construed as the instrument of power. Again, we find that there is a temporal anomaly in the nature of power that helps to explain the role of violence in the construction of political power. Briefly, those who use violence instrumentally can only say that the exercise of power has been successful after the act of its exercise — that is, that power existed. Empirically, however, we can not say, that power existed before its exercise since its prior existence depends uniquely on its effects. If there are no effects of power, then there can be no power. But effects must come after causes. If this is true, then power must exist prior to any evidence of its existence. But this is no more than metaphysics. Therefore, power is not an empirical phenomenon. Moreover, if the existence of power depends on the existence of its effects, it cannot be true to say that "power" causes violence, or that violence is the instrument of power. Instead, these relationships are always culturally, not materially or causally, constructed.
To "use" violence, then, is to control the modes of narrative by which it is explained, and the symbolic forms in which it is incorporated into the memories and behaviours of individuals and into the social environment and political process.
One final point emerges from this argument. Since violent events frequently figure as the points of origin in the narratives of social power (such as histories of the foundation of dynasties, nations or People) or at the origin of social statuses (purity achieved through sacrifice, or masculinity achieved through combat or circumcision, for example), it appears that violence must somehow be the cause of these social formations and statuses. It seems to stand at their origins. This view seems characteristic of naïve, formal historical or social scientific accounts of States, ethnicities, dynasties or of statuses such as warrior-hood or masculinity. Since causes, as we commonly understand them, must precede and be proximate to the effects they are said to cause, we seem to have another reason to suppose that it is violence itself which is instrumental. As instrument, violence seems to originate social forms or cognitive states. According to the account given here, however, since violence is understood to be disruptive of customary or institutional practices, violent events constitute a sort of natural "zero point" for social narratives of many types. Violent events are moments against which other social processes, and other moments, may be calibrated. Thus violence may commonly represent the causes of social forms because it is the emergent moment of ruptured time according to which social process are calibrated and fixed in temporal relation with other social processes. Although violence is present at the "beginning", it is not valid to attribute to it instrumentality of causal efficiency.
This recognition, finally, allows us to re-imagine a political philosophy in which violence is not taken to be either instrumental or consequential but rather to be seen as "causally isolated" from the cultural processes through which its political or cognitive significance is constituted.
Violence as rupture of social process
My fundamental claim is that because violence erupts into the flow of the ordinary and disrupts it, violence inevitably disrupts the continuity of intentions and actions, or the regularity of behaviours on which causal, historical or structural accounts are based. Violence occurs, therefore, but any account of it is necessarily retrospective. Therefore, the sociologically significant aspects of violence and, more importantly, the only aspects that are genuinely accessible to observation are the consequences — never the causes — of violence. But if this does not offer an account of violence in the causal, sequential, or "historical" terms that we expect, it is because no such account of violence is possible.
The methods used in this paper are chiefly philosophical and not empirical though they are motivated by empirical instances. The empirical details of particular cases are not germane here since I am advancing some general ideas about the temporal nature of social violence and what I believe to be universal characteristics of all forms of violence. I hope the reader will grant me his philosophical space.7
The arguments that I am advancing here, however, pertain especially to a type of violence that we see more an more of today, namely the "senseless" or "meaningless" violence of Kampuchea, Sri Lanka, Central America, Lebanon, Mozambique and South Africa. When a bus load of peasants is shot to pieces by unknown assailants, when victims and killers are most likely complete strangers to one another, when bodies are found mangled in mass graves in the midst of inchoate violence, when "sides" and "factions" change membership and identity frequently but without diminution of their lethal potential — then a kind of violence exists in which not just standards of humanity break down. Social scientists and historians are frequently at a loss to explain them; politicians do not know how to control them; the innocent know no escape, and the guilty achieve neither resolution nor absolution. It seems to me that there must be room for a new conceptual effort.
My effort, then, will focus on the forms of inchoate violence of which "Lebanon", "Cambodia/Khmer Rouge", "Uganda/Idi Amin" and, more recently, "South Africa" have come to be the outstanding representatives in the public mind. In these cases there has been no formal declaration of war, no judicial recourse, no limits to the involvement of any categories of population, no battle lines nor formal chains of command. Rather there have been "killing fields"; the key notions of liberal social science and history like community, intention, polity, humanity and responsibility seem no longer to apply. These are situations in which violence has indeed become "ordinary", but the mark of its profound disruption of ordinary explanation can be seen in the difficulty that all reasonable people have in offering coherent accounts of other aspects of social life under these conditions.
Violence and the ordinary
In a sense, violence is always "ordinary". The concepts we tend to use in order to talk about violence treat it as a static social fact — something that can be counted, measured, typed, structured — not as a many-stranded complex of processes involving actions and judgements in time and space.8 If we treat violence as a kind of social pattern that is genuinely temporal in character, and contingent though unpredictable, then several important consequences emerge.9
Much literature and the contemporary mass media treat it as drama, or worse, as entertainment.10 The political scientist or sociologist emphasise "structures", the regular institutionalised patterns of government and administration that use violence to secure their political goals. The criminologist treats it as "abnormal", or deviant. The masters of the horror film and the mystery novel, however, know that violence evokes terror when it happens in the context of daily life: the knifing in the shower, the rapist in the kitchen, the death at the hands of a lover, the devil's child in the cradle. Unlike most academic approaches to violence, those who suffer it know that what really counts as violence is that which takes place in the context of daily life. Violence which suddenly obtrudes into the flow of existence is truly fearful precisely because it occurs in the midst of the normal. In order to understand violence, the terror it evokes, the fear it engenders, we must seek to understand it in the context of ordinary life, in the flow of the day-to day, and in the minds of ordinary people … because this is where it really happens.
What turns the ordinary into violence? This is the genuinely anthropological question about violence since it asks, What is the nature and function of violence in the activity of being human. For the most part, anthropologists and sociologists have reported the facts about the relationship between power and violence in the context of other kinds of relationships. Indeed, violence manifests itself in so many ways and in so many social contexts, that virtually all holistic monographs contain some reference to violent interactions. These are frequently related to "political" activity: defence or acquisitions of resources or women, or enforcement of legal or quasi-legal decisions, the executions of sentences or punishments. Sometimes, however, we find that violence in practice has a close relationship to the aesthetic or the religious, as it does for instance in sports such as boxing or cock-fighting, or in religious icons and sacrifices. Sometimes violence is also closely related to rites of passage, as for example when young males achieve the status of manhood through performances that are often violent and bloody.
Violence, in other words, takes place in complex cultural contexts which frequently have nothing (apparently) to do with conflict, institutional power, or with damage and violation. Violence, seen in its anthropological perspective, goes well beyond politics and power, and well beyond the concepts of forcible damage. Anthropologically, then, violence must be understood as part of being human in the contexts in which we all must live. in other words, the investigation of violence must consider the meaning of violence as part of the process of social life. Unlike many other social acts which come about as the realisation of intentions, or the attempt to make certain concepts concrete, violent acts are often meaningfully constituted only at the moment of their commission. This fact makes violence, especially lethal violence, difficult to accommodate in the ordinary narrative history or the sociological account of social "causes" or "forces". Indeed violence is what it is, socially and historically, precisely because of its temporal peculiarity.
The "uncertainty principle" of social violence
In practice, violence is notoriously unpredictable. First of all, because the proximate causes of violence are often so complex that is rarely possible to know exactly or precisely what triggered a violent event. What is knowable, on the other hand, is that violence has happened. Our knowledge of violence, therefore, is necessarily past tense. Second, the means and instruments of violence are not easily controlled. Whether pens are sharper than swords, it is just as possible to misconstrue a word as it is to mis-aim a gun or to discover that a knife cuts both ways. In simple physical (or "material") terms, instruments of violence are designed so as to concentrate a maximum of force in the minimum of space in order, precisely, to disrupt, sever, damage or destroy. Any such instrument is necessarily — in thermodynamic terms — unstable irrespective of the stability of intent or the balance of reason among those who resort to their use.
All societies possess judicial or adjudicative institutions designed to deal with the first of these features of violence, namely, that the proximate causes or triggers of violence are complex and frequently not knowable in any absolute sense. Formal and informal limits are placed on the pursuit of "causes" and every effort is made to generate an authoritative account of any death, but especially of violent death. The social instability of violence is controlled by judicial proceedings, religious sanctions, institutions of "honour", feuding or shame, or resort to ordeals, oracles and so on. The practical or thermodynamic instability of the means of violence is recognised implicitly by everyone: these things are dangerous. These instabilities of both the means and the meaning of violence, however, have social consequences which are not widely recognised.
Consider: A bullet that misses is not the cause of murder; it merely signals an intention — or perhaps an accident. A blow that lands wide of the mark is not assault; if not meant in jest — this is a possibility, after all, that only context and post-facto verbal accounts ("just joking!") can tell — there is still the possibility of flight or negotiation. In other words, what seems to be a perception of violence is after all a judgement that we make after the fact. It's very often not possible to decide before the fact whether violence is going to occur or not. People back down, have second thoughts, apologise, and go their separate ways … sometimes!
Of course, when violence does occur subsequent to a build up of tension, it is always understood retrospectively, and by those who were involved in it, to have been caused by those tensions. There are, however, many situations where violence might have erupted, but didn't. There are many situations where it seems that violence should not have happened but did. It is plain, in other words, that we can not predict the outbreak of violence. What we do have, empirically, is judgements about events that have already happened. "Now, that's violence", someone might say in response to a newspaper story, or to having seen a violent robbery in the street. What we do not have, empirically speaking, is clear concepts of what causes these events to happened. The causal accounts and the typological account are not very useful. Neither predicts, nor accounts for the process by which violence unfolds.11
Violence does happen, however, and if it is caused the causes are extremely complex — in fact, they are probably too complex to comprehend in terms of a theory which will predict violence in real situations.12 Any useful account of violence that we offer, then, must be an account of the processes that lead to a situations in which the possibility of violence does or does not exist, and, to try to explain, once violence has happened, the manner in which it is integrated into the politics, the meanings, the concepts, and the culture of the people who have experienced that violence. This would not be either a causal or a typological account, but rather an account based on probabilities and interpretations, and thus either inherently uncertain or contestable, or both.
Boxing, for example, is a violent sport in which people are usually injured and, from time to time, killed. Violence, in this case, is a sport, not assault. Here the nature of violence, and the judgements that we make about it, depend on the context and ceremony of the event. If a boxer is killed in the ring, the kinds of judgements that are made depend on the fact that he was killed. In other words, they are post-facto judgements. It is only after the fact that the episode of violence becomes integrated into a narrative of rationales, a narrative that seeks to explain after the fact everything that came before the fact as part of the cause of violence. Now until that unfortunate and probably unexplainable blow that actually killed the boxer happens, its just a boxing match and all of the previous events are sport. Once the violent event has happened, that is, once the death has occurred, then everything is reinterpreted. We go back in time — we quite literally seek to reverse time by means of the cultural tools of narrative and rhetoric — in order to construct a meaning for the violent event itself.
Violence in this case is more that the act of violence itself — the blow that laid the opponent out for the long count and beyond; it is rather a process of setting up contexts and expectations and reinterpreting them once — and only if — an event we consider violent actually occurs. Frequently enough, the actual proximate cause of death — an unfortunate or unexpected blow — is itself insignificant in the subsequent narrative that "explains" the death.13 Violence is only violence — that is, it is only judged to be violence and thus called "violence" — within an ongoing series of evaluations in which meanings and acts are brought into relationship with one another.
Violence as process in time and narrative
These considerations suggest that more attention must be paid to the temporal dimensions of the context of violence. Let us consider for a moment then not on the damage that violence does but rather the framing or staging of violence.14 Because the timing or outcomes of acts of violence are unpredictable, the "framing" of the violent event must be in place before the eventuality it may eventually "frames" actually occurs. Armies hurry up and "wait" precisely for this reason. The violence of warfare and planned "political violence" usually takes place in "fancy-dress". The dispositions and appearances of social actors who engage in potentially violent activity array themselves in uniforms, carnival masks, guerrillas camouflage, war-paint or other distinctive dress precisely in order that the violence they expect will be immediately imbued with meaning when it happens. Otherwise it is accidental, beside the point, not "political", that is, it is not construable as socially causal. In fact, most participants in violence dress for the part. This dressing for the part may include their clothes (boxing shorts, military tunic, karate gown), but also their positions (in a ring, in a fox-hole or trench, or in bed), and their postures or gestures ("macho", threatening, poised, bullying, menacing). The uniforms, fancy-dress and war-paint do not, of course, cause violence: they merely wait, as potential "frames" to make its eventuality explicable when it does occur.
The rules of specific culturally-sanctioned or traditional narratives intensify the climate of meaning in which the expectation of violence informs the event of violence. The ideas that "a man should defend his honour" or that "cuckolds become violent" are characteristic of narrative genres in which the plot is moved by the generic expectation that men are moved to violence by dishonour. These narrative plots move from resentment to planned revenge to satisfaction and closure in an act of violence. The violence of head-hunting, for instance, is motivated — in both the narrative and emotional senses — by stories of previous head-hunting expeditions. The narrative motif can not be separated from the emotional and social motives. They provide templates that guide our thinking about other events of violence. The statement of intentions to do violence, then, may be seen as attempts to set the stage so that when violence does occur it will be explicable in terms that have already been written into the cultural "script". The paroxysm of jealously, the achievement of revenge, the exercise of righteousness, the execution of just punishment, or the act of war all have their templates in genres of narrative which construct causal sequences and then move to satisfy the expectations by demonstrating that the proposed causes have just these expected effects. Of course, while "real life" is not fiction, "real life" resembles these fictional sequences often enough to make them both — life and fiction — convincing. The problem is not that there are no causal sequences of these kinds in sociological reality, but rather that we do not, and probably can not, know when or where one of these sequences will actually happen.
Violence by its very nature happens suddenly, but it comes into being and acquires its meaning as a result of a narrative process. It is not there as a part of the structure or as an innate biological endowment. Violence comes into being over time, and through time. Intentions which were there beforehand, suddenly become rationales after the bullet finds its mark and the opponent dies. The intentions of the shooter who misses remain wishful, though malevolent, thoughts. The status of intentions, whether they are wishful thoughts (the bullet missed) or rationales (the bullet hit), depends on the occurrence of an event. In many cases it's a chance occurrence. Suddenly intentions become rationales depending on which temporal "side" of the event they fall on.
Violence, then, is like a full stop in time. The "." at the end of sentence carries no meaning itself, but rather forces a conceptual halt and an evaluation: What did the sentence mean? … now let's being another. And so on. Violence is a sudden eruption into the flow of social time which has to be accounted for. It's not there beforehand. Violence, then, can not be part of the general structure of culture or society. It is something which is genuinely time bound, becoming genuinely part of social and cultural time as it is interpreted and incorporated into narratives, memorials, tombstones, poems, photographs. It must therefore be conceptualised as part of the temporal process of society, not its structure.
The crucial result of violence then is that it must be inserted into the narratives and subsequent frameworks of meaning that we cast for it. This is part of the process of violence and extends our definition well beyond the simply behavioural or physical manifestations, well beyond the philosopher's or the sociologists definitions into the realm of social processes involving meanings and narratives, intentions and rationales of real responsible individuals.
Violence, then, is temporal, not structural or categorical. Violence is not a priori, but depends on whether an act has been committed in which motivations may be later construed as rationales, and thus constructed as explanations. Thus, a statement, "I want to kill/injure him because … " must not be taken, un-problematically, as merely the previous tense of "I killed/injured him because … . " Violence is even more complexly temporal than this: it is a process the relies on the multiple and on-going evaluations of meanings, complex sequences of processes partly entrained with one another, and partly independent; it is not analytically appropriate to reduce it to sequences of causes.
Violence, inevitably, is an orphaned moment that must be claimed by meaningfulness if it is to grow to social significance.
Violence as social instrument
If the actual event of violence is uncertain and if, as I have claimed, it possesses a uniquely temporal character, then these claims raise serious problems for the notion that violence can be (or is) a political instrument. This is because the presumed social instrumentality of violence depends on the notion that violence itself is and efficient social cause, something which can be "used" to achieve political goals. If violence is not causal in this way, then the account of social power that rely on this account of the social instrumentality of violence may fail.
In short, violence itself is not a political instrument primarily because its occurrence is unpredictable, and therefore not properly "instrumental" in any common-sense way. Moreover, violence does not participate directly in the achievement or exercise of what we commonly call "(social) power". This latter claim, I think, will be most surprising, since most people believe that violence is the essence of social power. The apparent instrumentality of violence depends rather on the relationship that any event of violence comes to have with the cultural frames and narrative expectations in which violence must be clothed before it can play this role. The notion of violence as a political act is logically flawed despite its powerful and pervasive influence in both scholarly and ordinary accounts of violence. Violence is an act which is often most directly caused at a pre-verbal, non-social level, even as it is motivated and rationalised by elaborate verbalisation, in terms of complex ideologies maintained by means of massive social organisation and investment of time and resources. In this sense the word "motivated" fails to indicate the appropriate range processes which we believe result in a violent outcome. If motivation to commit violence were able to fully account for violence, then surely we could account for it more precisely, more satisfactorily, with a surer sense of closure, and of course, with less massive investment of time and energy (such as armies, armaments, police forces all of which represent a unimaginably vast store of human effort and material). But we cannot do so. The processes which actually cause violence are apparently more vague, and less accessible to consciousness than the desires and hungers and even hates or feelings of revenge which motivate us to declare that violence is the consequence of conscious intentions, and that political results (submission, domination, exploitation) are the caused consequences of violence.15
Cause and violence
These considerations must force us to conclude that violence can not, at least logically, be given a genuinely causal account. This means, in effect, that there is no history or violence, only accounts of violence in history. Our belief to the contrary depends on metaphors and images of causes and consequences that satisfy us rhetorically and narratively – that is politically and aesthetically – but can never be demonstrated or recreated, and are thus not empirically grounded.
Rather, it is what we might call the imagination that makes violence an instrument of political will or a cause of submission and subordination. It is imagination which motivates the rhetoric which accompanies violence and which gives it its meanings. Here, the word "motivation" makes sense in describing the generation of rhetoric in which violence is claimed and stories of violence are sold and consumed, since it is usually a conscious decision to buy a newspaper or a book in which violence occurs, or to plan a strategy of violence which is intended to cause the fall of a regime, to use violence in order to encourage intervention by means of more or different violence, or to employ violent force in the commission of theft. It is a conscious decision to accept the blame or to claim the glory of violent acts. These fit quite easily into the narratives of "political development", the declarations of the "struggle against oppression" and other kinds of stories by which violence is explained.
These are all verbal accounts and rationales of violence. They are not violence itself. Violence, in principle, is the end of a process, but it is also a pattern, it has a "figure" or "signature" that can be described and which differs from episode to episode, place to place, war to war and criminal to criminal. This patterned nature of violence is crucial because it is by means of the pattern that meaning can be attached to it. Violence which is seen to be without pattern is most difficult of all to "explain", that is, it does not fit the pre-figured narratives by which violence is appropriated to the domain of language and into the narratives of intention.
It is because of this that violence itself cannot properly be said to "cause" submission to authority, nor can it be understood as an "instrument" of power. The notions that violence does "cause" submission, or that it is an "instrument", either of power or the powerless, belongs to the domain of rhetoric, not mechanics or physics. Violence itself only causes pain, injury or death. The extent to which behaviour is affected by it, or that changes come about subsequent to it is a consequence of the rhetorical forms in which violence and its patterns or "signatures" are integrated into a narrative of antecedents and consequences, and the extent to which these patterns of succeeding events are assimilated to a metaphor of causation. The nature of this kind of causation does not suggest that it may be "used" as a screwdriver or surgeon's scalpel is used. Violence is not an instrument. These problems are cultural problems, and the analysis of violence must depend on a nuanced understanding of metaphor, audience response, rhetorical, poetic and prosodic strategies, relative status positions of hearer and speakers. It cannot be taken to depend on an "architecture" of social relationships, a social structure, or a concept of social causation which is nothing more than a metaphorical use of physical systems (in which such mechanical causes and effects can be reasonably said to apply).
Initially, there are two problems which we must deal with. The first is that violence is (or can be) the instrument of power (or the "powerless"). The second is that violence compels (or coerces) submission to authority (or power).
The first problem here must be divided into its sub-problems. First of all, the statement is profoundly metaphorical. The notion of an instrument refers, in its primary sense to a physical object which is used to cause some physical effect. A violin is an instrument that causes a physical effect we perceive as sound, and it can be used to create music if the patterns of causes (the movements of the musician's bow against the strings, and the resonance of the body of the instrument itself) create the expected effects of patterned sound which the musician intends to produce. If the instrument is tuned and in good shape, and the musician is skilful, music will result. Of this we can all be reasonably certain. Similarly, a screwdriver is an instrument for turning a special kind of fastener into wood or other material. Other things being equal, a properly applied motion will cause the expected physical results, and the intentions of the carpenter will be realised. There is little to take issue with here. This direct cause and effect relationship between instruments and their effects, however, does not apply in the case of violence. "Power" in the case of the violin player or the carpenter can be simply and un-problematically understood as the sets of forces, the patterns of their application and durations of force that are applied to the instruments. Applying power, musicians and carpenters achieve their intentions. "Power" in the domain of social relations, however, is judged not by the patterns and durations of its exercise but by its effects. There is no power without effects of power. With violins and screwdrivers, however, their can be an exercise of power that produces cacophony and skew or broken screws. Power is not judged in physical systems by its effects (although these may be indexes or signs of quantities of power that was applied previously to achieve the observed state), but by the measurements of the scalar quantities of its process: time, distance, amplitude, mass, frequency. The judgement that social power exists can only be made in the presence of its effects. There is no such thing as "power that didn't work" but which is nevertheless still (social) power. A miscarriage of power may reveal the intention to exercise power, or the delusion that power exists on the part of one who attempted to exercise power and failed, but this is not power. It is the failure to achieve the goal of one's intentions. Since this is always possible in social power, and never possible in the domain of physics, this is precisely the point where the presumed analogy (on which the metaphor of social "power" is based) breaks down.
In other words, social power can never exist at the moment of its own "exercise," but exists only as a post hoc judgement which attributes its existence to a time before the existence of its supposed effects. If the order is not obeyed, if the gesture of obeisance is not delivered, then power did not exist. But we only know this after the coming into being of the effect of which power is supposed to have been the cause. Without the effect, no power. Precisely the same gestures of the exercise of power, precisely the same intentions, and precisely the same acts of violence or threat may be carried out, but unless they elicit a response of submission, compliance or deference, then power did not exist. This however is a profoundly troubling paradox since power, as we commonly use the term, can only be said to have existed in the past if it is to be cited as a cause, that is, it must exist previous to the events it is said to have caused. But according to the analysis presented here, it does not "exist" in the present since it may always be contingently disobeyed, and its existence depends on the existence of its future "effects" which may or may not exist! How can something exist only in the past without the possibility of it existing in some previous present? How can the present existence of something depend uniquely and exclusively on its future?
Power in social systems is a word that rests on a metaphorical extension of what we mean by "power" in the physical systems we may all observe by sight, hearing, touch or smell. In social systems, we "see" power by the size of the monuments that the powerful build to themselves or to their exploits, or by means of the buildings and clothing they construct and own. We hear it on the radio and television, in discussions, or in stories. What we experience are the effects of "power". The process of getting these effects is often too messy, too protracted, and too secret to observe. What we see, however, is what we expect to see. We see and hear patterns that we interpret as "power". Seeing the effects, we assume that a process, akin to the process of making music on the violin or furniture with a screwdriver has existed in the immediate past and which is responsible for the effects of social power. But that, of course, is what we call a metaphor. We apprehend the social in terms of the physical. That does not make the social patterns identical to physical patterns. All it does is alert us to the existence of patterns. The attribution of causality is another matter entirely.
Thus the "instrumentality" of violence depends on the metaphor of the physical instrument which directly and un-problematically causes its physical effects. Those who employ these physical instruments skilfully, are capable of realising their intentions — the music "in their heads" or the house or chair in the drawing — because they control the patterns of causes (pushes, turns, pulls, blows) that result in patterns of effect (music, chairs, houses) which they envision. Violence is not like this at all.
In fact, far from being an instrument of power, violence is perhaps the most troubling and disruptive problem for any social holder of "power". Power does not control violence in any causal way but rather generates the patterns of discourse and interpretation in which episodes of violence will be understood by its observers, executors and victims as unique and necessary components of the political state-of-affairs. The states-of-affairs that violence is supposed to cause — submission to authority, compliance with edicts directives and wills, or other actions and thoughts — are no more consequences of violence than is violence itself. The narratives of violence, ideologies of power, and representations by which we picture and observe violence second-hand, are the means by which we come to believe that violence is an effective and immediate cause of certain consequences and social states of affairs. Violence, however, does not produce either power or social states of affairs, nor does power cause violence. Rather, violence together with those other states of affairs and patterns of behaviour that we call trade (or theft), trust (or deceit), reciprocity (or chicane) love (or war) are all consequences of a deliberate and intentional creation of patterns. They are all the ends of social process and as such are capable of being represented, but not of mutual causation. In other words, violence is itself a social state-of-affairs and does not cause other states of affairs any more than a song causes other songs, a picture causes other pictures, or a chair (or a drawing of a chair) causes other chairs. What causes other songs, other pictures, other chairs, is a process of intentions and manipulations of patterns of physical cause and effects. What makes the results of these processes "songs", "pictures" or "chairs" is the complex of values, words, objects and intentions that we call "meaning", or sometimes "culture". Violence itself, as a physical process which causes hurt or damage to those people or things whom we or others value, is a process of physical cause and effects like any other. It is a pattern which has a signature and a recognisable character, and it involves skills which those who would have "social power" frequently value very highly. But what gives violence character as the "instrument" of "power" is the images and narratives within which its episodes are placed. Those who experience violence may change their behaviour or their thoughts or both, but they do so as a consequence of persuasion, not of violence itself. Persuasion is the real instrument of power, and violence merely the condition under which it may operate.
Violence as cause of submission
This brings us to the second part of the our decomposition of the problem of violence as a political act, namely the notion that violence causes the submission to authority or will of those who execute it. This notion is similarly confused by the pervasive metaphors of physical causes and their (physical) consequences.
Lethal violence causes death. Once dead, of course, the victim of violence no longer participates in politics as a person. The name, of course, as a symbol of a position or a principle or an act of defiance or bravery may certainly participate in politics. This however, is a "participation" in a different kind of metaphorical (or tropological) sense. It is not the person who participates but the representation of the name or the "memory" of the person's deeds. It is this participation of memory that Milan Kundera has in mind when, in The Book of Laughter and Forgetting he says that the struggle of memory against forgetting is the struggle of mankind against power. The victim of lethal violence may indeed be remembered, and the memory may be more potent that any personal political activity he/she actually accomplished while living. But, to clear the decks for argument's sake, it seems clear that the dead do not participate as conscious physical persons in politics. Lethal violence, therefore, does not cause the submission to authority or compliance with the will of its perpetrators for its direct victims. Those whose compliance it is said to cause are not those whom it acts on directly. Already, it is clear that it is not a "causal" instrument in the way that a violin or a screwdriver is. Imagine playing a violin which produced no sound but caused other violins to play. Again, the analogy breaks down here, revealing rather precisely the nature of the metaphor by which we apprehend it. Violence does not cause compliance in any direct analogy with the way a violin causes sound. But let's not beat a dead trope any longer than necessary. Lethal violence only causes compliance, and to the degree that it does so, only as far as it is represented to an audience. The audiences responds to it only in so far, and to the degree, that they are able to see themselves as potential victims of violence, that is, to the extent and degree that they foresee it causing their deaths or the deaths of those with whom they have intense social relationships (or imagine that they do). The efficacy of lethal violence, then is not the consequences that it causes directly but the representations that people produce in order to bring it into account. Bringing violence into account implies that it is represented in the form of stories or pictures which necessarily imply a causal linkage between violence and political states of affairs. It is the representation of causality, then, not the proximate physical cause nor the efficient social/psychological cause, that is in question here. The necessary pictorial or narrative relation between "cause" and "effect" is provided by news stories, gossip, traditions, histories and other kinds of narratives. Quite routinely and by means of the expectations of plot structures and sequences that we all share, these narratives and representations present a case for causality, but they do not demonstrate it. They deploy the tropes of comedy and tragedy, and the metaphors of physical or organic systems that we have mentioned, but they do not constitute any empirical evidence of a material causality that is in any way different from any other concepts or symbols that we hold or believe. Pictures of violence focus necessarily on the physical acts. A picture can not show a political state of affairs, but only physical visual states of the world in which political states are represented by means of conventional symbols. In either case, the "causal" link that we think we see is clearly rhetorical or mimetic or otherwise representational.
It is precisely because people routinely represent violence to themselves in terms of causal schemas that politics can appear to "use" violence instrumentally. What it uses, however, are the rhetorical schemas which people construct in order to understand violence and to give it meaning, not the raw stuff of violence itself. The same may be said of non-lethal violence which causes pain or physical damage to an individual. Such an individual may continue to live and may continue to participate politically in the same arena as those who have inflicted violence on him or her. Surely we must assume that the violence that has been done to them will have changed their behaviour in a way which we may call properly "causal", that is, the violence has had an instrumental effect on the individual who has suffered it. This may well be so, but only if the victim equates the ultimate source of "power" (essentially a metaphysical quantity) with the immediate cause of his suffering. This can be seen by comparing political violence to the violence of storms. The comparison is suggested, of course, by the identical vocabulary which we use to talk about both, but, more than this, they share a common narrative of cause and power. The difference lies in whether the cause of the violence (political or natural) is equated with the locus of will and desire.
Hurricanes and earthquakes are described as violent, but the violence which they do is not instrumental to the achievement of their will. In this sense, it is not " violent" in the way we have been using the term here, not because it does not kill, damage, or hurt people and property, but because it is not ascribed any property of will. Hurricanes, we know are caused by "weather", and earthquakes are caused by the cumulative pressures that build up in the earth's crust due to plate tectonic. Those who think of these things as in some way willed, ascribe the will to God or Nature, not to the event of the hurricane or the earthquake. The causes of natural violence, then are not conjoined with the locus of power in the contemporary, "modern" imagination. Hurricanes are only violent with respect to the damage they do, not the violence they intend. Political violence is violence by virtue of the damage it does (and only in the actual event of that violence), as well as by virtue of the violence it intends. But, since "politics" or "structures" are not people but abstractions or descriptions of patterns, and thus do not have intentions any more that storms or plate tectonic do, a narrative must be constructed which will provide the satisfying explanatory linkage to "power".
If this linkage to a culturally constructed locus of power is not constructed, or can not be constructed, then the violence is not "political".16 The inflictor of violence has, quite literally, no power over his victim, and violence is not construable as instrumental. Violence may be employed instrumentally to the extent that people can be persuaded that violence is an instrument, that is, only in terms of a particular image or "imagining" of the nature of the power and the way it works. Its "instrumentality" is not immanent in violence itself, but must be constructed in the narratives of violence which explain it.
Violence, then, far from being an instrument by which experience is shaped, or wills are coerced, is merely the raw "stuff" of experience, but is only contingently the consequence of will, intention, or deliberate social process. Violence does indeed play a central role in the way we construct our social and psychological worlds because it creates a unique moment when the time of its occurrence — embedded as it is in the time of the ordinary social process — stands out in our consciousness and thus seems to stand still. But the moment of violence is not part of the time of the ordinary routine. It emerges as something new: it is a genuine and incontrovertible becoming. In this sense the event of violence creates time and space by creating events and places that stand out because these seem to stand outside of ordinary time. The violent event creates a moment that must be interpreted, that must be given new meaning, that must be accommodated to the scripts that culture has prepared, and which people are prepared to accept. It must fit the frames that are intended to encompass it, but it is not itself directly explicable in terms of intentions.
These considerations should allow us to see the error of reasoning that justifies the terrorist's act of violence against "the system", the racist's violence against a "race", the nationalist's attack against the republican, or the assassin's assault on the representative of a polity. Violence is not instrumental in the way they suppose it to be. The error that the revolutionary makes in supposing that violence itself will cause change is flawed for the same reason. While it is true that violent events do indeed stand at the beginning of most major social or conceptual changes, it is an error of reasoning to attribute the violence a causal or creative role.
The creative or constructive aspect of the violent event is not the occurrence of violence itself, but rather the interpretations, memories and memorials that violence evokes after the fact. This is a peculiar kind of creativity, then, for the act of violence itself almost always precludes the "normal" sequences of expectations, narratives or social causality. If violence is the moment when the normal routines of life break down, it can not be explained simply in terms of those routines. Only post-facto narratives constructed after the fact in terms of narrative schemas which exist in different kinds of "memory" can construct the meaning of violence. Violence itself precludes social causality precisely because it disrupts the normal sequences of social activity. It activates new social patterns and processes precisely because it does disrupts social and cognitive patterns, but it cannot do more than this: it cannot yield direction or form. Violence, therefore, stands at the origin of new social forms, but it does not cause them.
2 On assassination as political pivot, see Franklin L. Ford, Political Murder: From tyrannicide to Terrorism (Harvard U. Press, 1985) for a general survey from early Antiquity to the present. On wars as political pivots, nearly any political history of Europe will suffice.
3 In our times the slogan, from Mao Ze Dung, that "political power grows out of the barrel of a gun" is taken as a truism. This rhetorical formulation comes through Marx from Hegel who understood political conflict, and thus history (the synthetic outcome of series of conflicts), as necessarily violent. In short, the fear of violent death that any participant to conflict must feel (given that all conflicts in Hegel's terms are life-or-death struggles), will lead one party to prefer submission to death, and thus to found the facts of inequality in this elemental "rational" choice. See Leo Strauss and J. Cropsey, History of Political Philosophy, 733; G.F.W. Hegel, Encyclopaedia of the Philosophic Sciences, (1959) paragraph. 433, p.352. Other "republican" or "nationalist" formulations of this same fundamental political tenet — I would call a belief — could be easily multiplied: Clausewitz On War or Chesterton's famous remark, and Hitler's Mein Kampf might serve as the extreme data points in the topography of this "discursive space".
4 Garver remarks, for instance, that "it is never entirely clear what violence is". (1968:223). Glover's (1977) consideration of the depths of questions about "violent" or personally-violating medical and legal interventions reveals more complexity. Marsland (1985) points to the "neglect" of violence by sociologists and argues that ideological blinkers (utopianism, Marxism) have rendered us perplexed (cf. Ashworth 1986). Kekes (1990) argues that evils, including violence, are traditionally understood in a tragic way (beyond control) or in a rationalistic/moralistic way (insufficient control), and that either mode of understanding is insufficient.
5 The introduction of a 1981 UNESCO volume entitled The Causes of Violence asserts precisely this: "At all levels of research into the causes of violence, scientific discussion generally holds that violence is explicable and that, because it is explicable, it is avoidable" (Joxe 1981: 15). In fact, this statement could stand as an apt and concise statement of the "normal science" consensus among social scientists. Other, broadly similar statements abound: American Anthropological Association 1987; Apter 1987; Brown and Schuster 1987; Coady 1986; Riches 1986; Townshend 1987; Scott 1976; Smith 1975; Wolff 1969).
7 Personally, my empirical and ethnographic experience of lethal violence includes local cattle-raid warfare in Karimoja and northern Uganda (1969-71), the coup d'etat of Idi Amin Dada in Kampala Uganda and its immediate aftermath (1971), local fighting in central Tanzania (mid-1970s; see Thornton 1980), and violence in South Africa, both "political" and "criminal" since 1979; see, for instance, Thornton 1989.
8 The fact that we have a word, "violence", which is nevertheless extremely difficult to define precisely is evidence for the "facticity" of violence — that is, the existence of the word proves that there is something there conceptually and phenomenologically, even if we can not define it precisely. The fact that judicial proceedings in all societies consume much time and energy attests to the perplexity that violence must inevitably occasion. Moreover, what violence is not adjudicated becomes the focus of gossip and "public opinion".
11 It might be remarked that our "causal" accounts are probabilistic, that is that they predict likelihood, not events. This, of course, only underscores my claim that such accounts are useless. While a physicists account of atomic behaviour makes appropriate use of the concept of probability because his sample is very large, and because no individual is of any interest to him at all, the opposite is the case with the social scientist. It does not help a police officer to be told that a murder is almost certain to occur within a mile radius of his position, or that a shooting death is certain to occur at the corner cafe but without telling him when.
13 In episodes of witchcraft accusations, the proximate cause is deemed irrelevant; in more sophisticated judicial proceedings, the proximate physical cause is merely "evidence", that is, an index that points to a larger-scale narrative of intentions, trajectories, motives, situations, opportunities and plans.
15 The problem of slippage between temporal scales, understood fairly precisely as the actual metric relationship between the frequencies of events relative to the relatively fixed human frequencies of chemical, neural, diurnal and life-time events, is given more extensive treatment in my "Time scales and social thought" (1989).
16 This notion of the "political" follows that of Max Weber in, for instance, "Bureaucracy and political leadership" where he says very simply "politics means conflict" (Weber 1968:1399); or G. Simmel in "Conflict" (1971).
American Anthropological Association, 1987, "Endorsement of the "Seville statement on violence"", Anthropology Newsletter 28(1)24, Jan., 1987 and 28(9):1, Dec., 1987.
Apter, David, 1987, "Violence as a post-modern condition". In Rethinking Development: Modernisation, Dependency and Post-modern Politics. Newbury Park, California: Sage Publications.
Ashworth, T., 1986, "Neglect and Betrayal — War and Violence in Modern Sociology" [A review of War and Violence in Modern Sociology, by David Marsland.]. Sociology 20(2):308-309.
Bateson, Gregory, 1979, "Mind and Nature, a necessary unity". New York: Dutton.
Brown, Paula and I. Schuster, 1987, "Introduction: Culture and Aggression" [from a Symposium of "human Aggression", American Anthropological Association Meetings of 1985; see also the papers it introduces]. Anthropological Quarterly, 59(4):155-159.
Coady, C. A. J., 1986, "The Idea of violence". Journal of Applied Philosophy, 3(1):3-19.
Ford, Franklin, 1985, "Political Murder: from tyrannicide to terrorism". Cambridge, Mass. Harvard University Press.
Galtung, Johan, 1981, "The specific contribution of peace research to the study of violence: typologies". In Violence and its Causes, pp 83-96. United Nations Education, Scientific and Cultural Organisation. (UNESCO)
Garver, Newton, 1968, "What violence is". The Nation, June 24, 1968, pp. 817-822. Reprinted in Philosophical Issues, edited by James Rachels and Frank Tillman, pp. 223-228. New York: Harper and Row, 1972.
Glover, Jonathan, 1977, "Causing Deaths and Saving Lives". London: Harmondsworth.
Hegel, G. F. W., 1959, Encyclopaedia of Philosophic Sciences [Philosophisches Bibliotek]. Translated by Pierre Hassner and Alan Bloom. Hamburg: F. Meiner.
Hume, David, 1969,  "A Treatise of Human Understanding". Edited with an introduction by Ernest C. Mossner. London: Penguin.
Joxe, Alain, 1981a, "General Introduction", In Violence and its causes, pp 9-22. UNESCO
Joxe, Alain, 1981b, "A critical examination of quantitative studies applied to research in the causes of violence". In Violence and its causes, pp. 63-82 . UNESCO.
Kekes, John, 1990, "Facing Evil". Princeton University Press.
Kundera, Milan, 1978, "The Book of Laughter and Forgetting".
Marsland, David, 1985, "War and violence in modern sociology". London: Institute for European Defence and Strategic Studies.
Nietzsche, Friedrich, 1967, "The Genealogy of Morals". In The Essential Nietzsche. Translated by Walter Kaufmann. New York: The Modern Library.
Riches, David, 1986, "The Phenomenon of Violence". In The Anthropology of Violence, Chapter 1, D. Riches, ed. Oxford: Basil Blackwell.
Scott, James C., 1987, "Resistance without protest and without organisation: Peasant opposition to the Islamic Zakat and the Christian tithe". Comparative Studies in Society and History, 29(4):417-452.
Scott, James C., 1976, "The Moral Economy of the Peasant: Rebellion and Subsistence in Southeast Asia". New Haven: Yale University Press.
Simmel, Georg, 1971, "Conflict". In Georg Simmel on Individuality and Social Forms, D. H. Levine, editor, pp. 70-94. Chicago: The University of Chicago Press.
Smith, Estellie, 1975, "Cultural variability in the structuring of violence". In Psychological Anthropology, T.R. Williams, ed. The Hague: Mouton.
Strauss, Leo and Joseph Cropsey, 1987, "History of Political Philosophy". Third edition. Chicago: University of Chicago Press.
Taussig, Michael, 1984, "Culture of terror — Space of death: Roger Casement's Putumayo Report and the explanation of Torture". Comparative Studies in Society and History, 26(3):467-497.
Taussig, Michael, 1986, "Colonialism, Terror and The Wild Man". Chicago: University of Chicago Press.
Thornton, Robert, 1980, "Space, Time and Culture among the Iraqw of Tanzania". New York: Academic Press.
Thornton, Robert, 1990, "The Shooting at Uitenhage, South Africa, 1985: the context and interpretation of violence". American Ethnologist 17(2):217-236.
Townshend, Charles, 1987, "The necessity of political violence". Comparative Study of Society and History.
UNESCO, 1981, "Violence and its causes". United Nations Education, Scientific and Cultural Organisation.
Weber, Max, 1969, "Economy and Society". Edited by Guenther Roth and Claus Wittich. 2 volumes. Chicago: The University of Chicago Press.
Wolff, Robert Paul, 1969, "On violence". Journal of Philosophy 66:601-616.