Violence and Punishment

Civilizing the Body Through Time

Pieter Spierenburg

• Publisher: Polity
• Year: 2013
• ISBN:
  9780745663982
• Edition: 1
• Topic: Ciencias sociales
• Subtopic: Criminología

This innovative book tells the fascinating tale of the long histories of violence, punishment, and the human body, and how they are all connected. Taking the decline of violence and the transformation of punishment as its guiding themes, the book highlights key dynamics of historical and social change, and charts how a refinement and civilizing of manners, and new forms of celebration and festival, accompanied the decline of violence.

Pieter Spierenburg, a leading figure in historical criminology, skillfully extends his view over three continents, back to the middle ages and even beyond to the Stone Age. Ranging along the way from murder to etiquette, from social control to popular culture, from religion to death, and from honor to prisons, every chapter creatively uses the theories of Norbert Elias, while also engaging with the work of Foucault and Durkheim.

The scope and rigor of the analysis will strongly interest scholars of criminology, history, and sociology, while the accessible style and the intriguing stories on which the book builds will appeal to anyone interested in the history of violence and punishment in civilization.

PART ONE

VIOLENCE

I

Long-Term Trends in Homicide

Amsterdam, Fifteenth-Twentieth Centuries

Homicide is haunting historians. Some welcome its possibil ities for a study of the long term, calling it “the one crime for which the evidence is most reliable.”¹ Others, more skeptical, are trying to prove that any visible trend is only a chimera. This clash of arguments and approaches particularly concerns the period before the availability of national statistics. Evidence from that period is indispensable, if we want to make meaningful statements about the long term. Throughout this chapter, I will usually speak of the homicide rate, meaning the annual average, over a specified period, per 100,000 population, in a specified area.

The study of homicide rates has progressed much since the first ten years (1981–91) when England was the main focus of research and the long-term decline of violence since the middle ages was first established. Yet some of the problems emerging from the debates of that period have not been satisfactorily solved. In the most recent synthesis of the quantitative evidence about murder by Manuel Eisner (2001, 2003), for example, it is not specified which of the data are based on coroners’ records and which on prosecutions alone. He also conflates the categories of family homicide and homicide on intimates. On the other hand, recent syntheses pay much more attention to the trend reversal since about 1970, which was still hardly visible when the pioneers wrote their studies.

The Inadequacy of Trial Figures

Examining the Dutch evidence available when I started my research, we can definitely conclude that frequency rates based on prosecuted homicide alone are unreliable. Combining the work of several historians, we might construct a graph of murder and manslaughter cases tried by the city’s court from the beginning of the sixteenth century to the beginning of the nineteenth.² This enterprise leads to a surprising conclusion: the average annual level of prosecuted homicide in Amsterdam tended to remain constant, at about 1, over the entire period from 1524 to 1811 (although in the short run there were peaks and quiet decades). The city’s population, however, rose dramatically from 15,000 in 1500 to 100,000 in the early decades of the seventeenth century, and to well over 200,000 at the end of that century. Thereafter there were minor fluctuations, but the population always remained over 200,000. Thus, a graph of homicide trials per 100,000 inhabitants would simply be the inverse of the population curve. The figure would be less than 0.5 per 100,000 throughout the eighteenth century. It is highly unlikely, although not impossible, that this graph would bear any resemblance to the real incidence of killing in Amsterdam.

The evidence on default procedures, discussed by Sjoerd Faber and Marijke Gijswijt-Hofstra, is one indication that the real incidence of killing in Amsterdam was higher than the prosecution records suggest.³ Their data from two archival series allow me to calculate that the rate of homicides with identified killers (i.e. default procedures and prosecuted cases combined) lay between 1.5 and 2.1 during the period 1680–1790 and that it was 0.7 during the period 1751–1801. These figures indicate a declining trend, but the total number of homicides committed was higher than the rate of homicides with identified killers.

Outside Amsterdam, too, the available data indicate that homicide rates based on records of criminal prosecution are inadequate. A project on criminal justice in Leiden, 1533–1811, covers the next-longest period. The sentence registers of that town contain cases involving arrested killers as well as default trials. Peaks in the absolute numbers, however, are always due to increased efforts in sentencing fugitive suspects. When the number of homicide cases is low, there are almost no default trials. In another proj ect the court records of seven sample jurisdictions, mostly small towns and rural regions, are analyzed from 1700 to 1811. In those seven jurisdictions there were only 98 homicide trials altogether, including default cases, during that period. The frequency declined after 1750.⁴ In Haarlem, 1740–95, the rate of prosecuted homicide amounted to 0.7 per 100,000 per year.⁵ Two widely diverging figures for the city of Utrecht also illustrate the problem. Based on the work of Dirk Berents, we can calculate that the homicide rate in this town in the first half of the fifteenth century was 53.3 (perpetrator rate) or about 50 (victimization rate).⁶ Examining the town’s register of criminal sentences for 1550–70, however, we arrive at a rate of prosecuted homicide of only 1 per 100,000.⁷ It cannot be doubted that this immense drop within 100 years is largely due to changes in the registration of killings.

While these data from prosecution and default procedure records are suggestive, the really conclusive evidence is provided by the body inspection records discussed below. Jüngen was the first to recognize their importance and, to anticipate the discussion, let me cite his conclusion: in sixteenth-century Amsterdam, the ratio of detected to prosecuted killings was 9:1. The obvious superiority of lists of body inspections as a source for establishing the incidence of homicide lies in the fact that these records include cases in which the killer remained unidentified. To conclude, there are three types of sources, each successively yielding a higher number of homicides: court cases involving arrested killers; records listing all cases with identified killers; and lists of body inspections. Figures derived from the first two types of sources should be taken into consideration only when they are relatively high, because in that case the actual number of killings must necessarily be at least as high.

Methodological Considerations

Before I present my evidence, some methodological questions need to be cleared up. For one thing, can we infer from the homicide rate that in the early modern period, when a measure of internal pacification had been reached, violence still was socially accepted to some extent? The homicide rate refers to activities which were not tolerated per definition, since they were liable to criminal prosecution. Consequently, the cases we are collecting belong to unaccepted violence. The solution is to consider the homicide cases as representing the extreme pole of a continuum that begins with accepted violence: quarrels that are laughed about; fist fights in which bystanders encourage the participants more or less indiscriminately; a teasing with knives that everyone takes lightly; a beating within the family that the community does not consider serious. In the overwhelming majority of instances, such acts of violence do not result in the death of a participant.

Viewed from this angle, homicide only occurs when things get out of hand. Homicides are “casualties.” The frequency of these casualties is an indication of the frequency of relatively minor and common forms of violence. In the sixteenth through eighteenth centuries, lethal fights were not accepted, but moderate forms of violence were considered relatively normal. If we observe a decline in the homicide rate, and thus in the number of casualties, in the course of this period, it means that the level of routine violence has declined. When this process continues, there must be a qualitative shift at some point: acceptance of moderate violence turns into a rejection among most members of society; as routine aggression becomes rarer, the number of casualties becomes insignificant; the residual homicide rate largely consists of extreme or marginal cases, in either the sphere of erupting tensions or a criminal underworld. That is in line with the evidence on the contemporary proportion of homicides on intimates and killings related to other illegal activities.

Also, with respect to the more technical side of recons tructing the long-term trend, problems lessen if we take theory into consideration. It might be argued, for example, that in our time the cowardice of people who are mugged lowers the homicide rate artificially. Unlike 200 or 300 years ago, nowadays most victims choose not to resist, thereby allowing the instrumental violence accompanying the robbery to be limited to a light form or even just the threat of using it.⁸ However, it is doubtful whether in earlier centuries many cases of resistance to an assault on one’s property resulted in the death of the victim or the robber. More important is that, if most victims today prefer to hand over their wallet, this is, besides their being insured, precisely because they are unaccustomed to violence. So this is in line with the general trend. A similar argument applies to killing in self-defense. When a case is acknowledged as such, it is a lawful act, so we might argue that it should be excluded from the homicide statistics. Such an argument is unconvincing. Had the outcome of the event been the reverse, it would have been a regular homicide; either way there has been a fight. Manslaughter and killing in self-defense are two possible casualties arising out of the same situation.

Still another disturbing factor consists of the cases which remain entirely unknown. Historians generally assume that homicide is the crime most easily detectable, since it is so difficult to hide the body. Nevertheless, there are people who disappear permanently and some of them may have been murdered. Does this happen more frequently today than in earlier centuries? Reliable statistics are lacking. Whatever is the case, hidden bodies point at planned, rational violence rather than killing in affect. A somewhat similar problem applies to a specific type of murder: poisoning. Some historians assume that the chances that this crime remained undetected increase the farther we go back in time. This assumption, however, is highly questionable.⁹

Two of the possibly disturbing factors are peculiar to modern times. In some countries reckless driving with lethal consequences is prosecuted as a homicidal act, but this is not customary in Dutch judicial practice. In medical statistics, death in a car accident is normally listed as a traffic casualty, no matter how guilty the driver was. Although some traffic casualties may be deliberate murders, I cannot imagine these cases to be very frequent. Modern means of transport also affect the contemporary figures in the form of fast ambulances taking seriously injured victims to an operation room. Most scholars agree that increased medical skill and organization only start to play a significant role toward 1900. On the whole, it is probably safe to say that all factors affecting the homicide rate were fairly constant well into the nineteenth century.

Two classification rules, finally, follow from the demand for comparability of figures. The first is that the victims rather than the killers should be the basic unit for constructing the homicide rate. It can make a difference, even if we have the complete evidence on both. There were instances in which one assailant killed several victims simultaneously, as well as those in which one victim was killed by several persons simultaneously. Especially in the latter case, the urge to assign individual responsibility to the actual killer and his accomplices may vary over time. Moreover, in coroners’ inquests and medical statistics, we count bodies –that is, victims. The second rule is always to exclude attempted murder or manslaughter from the homicide rate. The definition of what constitutes an attempt at a person’s life varied greatly over time.¹⁰ Naturally, mere attempts are absent from medical statistics on the causes of death.

Homicide in Amsterdam, 1431-1816

Amsterdam meets another methodological requirement, that of studying a relatively populous area over time. Jan Boomgaard presents the earliest reliable and relevant data.¹¹ Information is available for eleven fiscal years during the period 1431–62. The names of 54 killers were known. Exactly a third of them were dealt with judicially, which resulted in 17 financial settlements and only 1 death sentence. Another 34 killers had fled Amsterdam; 1 had died in the brawl and the fate of 1 is unknown. In six instances the victim was killed by 2 men, and in two instances by 3 men, which means that the total number of victims was 44. These figures amount to an average of 5 killers per year and 4 victims per year. Amsterdam’s population by the middle of the fifteenth century may be estimated at 8–9,000.¹² Consequently, the homicide victimization rate was approximately 47.

The Amsterdam court records of the sixteenth century include the inspections, carried out at the orders of the judiciary, of the bodies of persons who had died under suspicious circumstances. During the years 1524–65, they numbered 646 altogether, around a third of which involved cases of drowning. There were 344 homicides; as noted above only 11 percent of these led to prosecution.¹³ The average absolute number of detected killings was 8.2 per year. With a population of almost 30,000, this amounts to a homicide rate of around 28. In an unpublished thesis Jüngen supplements these figures with data from four sample years: 1560, 1570, 1580 and 1590. In the inspection reports of those years he found an annual average of 10 homicides and 1.5 uncertain cases. Around 1575 the city’s population was probably about 47,500, which results in a homicide rate for the period 1560–90 that lies between 21.1 and 24.2.¹⁴ This suggests that the rate was declining during the sixteenth century.

Around 1600 the Amsterdam court stopped the practice of inserting body inspection reports in the registers of criminal sentences. Possibly, the court began right away to list these reports separately, but in that case the earliest records have not been preserved. There is a gap in our information until the 1660s. Five separate registers of body inspections carried out at the orders of the Amsterdam judiciary are extant.¹⁵ Faber has already consult ed this source in order to count infanticide cases.¹⁶ Because most of the inspections involved an autopsy, I will refer to them under that name. Together, the registers cover the period 1666–1817, with gaps for the years 1680–92 and 1727–51. When a register or set of consecutive registers did not begin in January or end in December, I discarded these incomplete years. Because of this, evidence is available for three periods: 1667–79, 1693–1726 and 1752–1816. In my tables the second period is split into two sub-periods and the third into four. Because the body inspection reports were copied in a bound register in chronological order, it is unlikely that cases have been lost. The problem is rather that there are too many.

Not every inspected body was actually dead. Around 1700 the registers include visitations of wounded persons, while throughout the period there are a few reports in which a person’s state of mind or the condition of a girl’s genitals were the issue. These cases are easily identified. It is different with the corpses. Information on the cause of death was based on the registered conclusions of a court committee, who did the inspections. The regist ers I studied also included cases of non-viol ent death. The problem is that the judicial conclus ion is missing. The autops ies were performed by the city’s anatomy professor assisted by two or three surgeons.¹⁷ They never stepped outside the bounds of their profession, always listing the medical cause of death only. Supposedly, the prosecutor and the judges drew their conclusions, based on information about what had happened. It is only in some cases that this information was inserted in the registers. Thus, we know, for example, that a man found drowned was murdered because the surgeons noted that his hands and feet were tied, or that a woman whose skull was smashed had committed suicide because a clerk wrote in the margin that she had jumped out of a window.

The only way to proceed was to col...