Recurring Cycles

Early modern patterns of lethal violent crime reflected the rou-tine activities of the age. This can be seen in the daily, weekly and annual fluctuations of homicide.

Day

In terms of daily cycles, the most prevalent time of homicide was between 18.00 and 24.00 hours. Almost half of the homicides (47 per cent) happened in that time bracket. The evening concen-tration was more pronounced in Denmark and Sweden than in Finland, where a higher proportion of incidents took place during the day. Fewer crimes took place during the night (18 per cent) or in the morning hours (12 per cent).

The relative and comparative scarcity of night-time incidents is of interest as it contrasts with modern patterns (see Chapter 11).

It can relate to lack of lighting, fear of darkness, and deterrence, since committing the homicide in darkness could be an aggravat-ing circumstance. Even fear of darkness could have been involved (Duedahl & Langen, 2015). While the evening hours were in all three regions the most likely time of homicide, in other respects there was some variation (Kivivuori et al., 2020b). The partially dif-ferent daily profile of Finnish homicide may reflect geographical

difference, as the Finnish study region was nearly 1,000 kilome-tres to the north from the two other study regions, a fact that also affects natural lighting conditions.

Week

If homicides were distributed equally over the weekdays, each weekday would have 14.3 per cent of the crimes. The early mod-ern distribution of homicides was not flat in this manner. There was a clear cycle connected to the alternation of workdays and the day of rest (Sunday). In the overall distribution, Sunday was clearly the weekly homicide risk peak. In the total data, 23 per cent of incidents took place on Sundays. This was nearly ten percentage points more than would be expected if each day had the same number of cases. In Denmark and Sweden, Sunday was the peak homicide day of the week. Finland, in contrast, mani fested a slightly deviant pattern, as Saturday had the highest share of homicides in that country.

The basic weekly pattern was, in all three regions, linked to the alternation of working days and the day of rest (Sunday). In all three areas, the post-peak day was the second most risky day:

Monday in Denmark and Sweden; Sunday in Finland. Thus, in Denmark and Sweden, the Sunday homicide peak extended to Monday, while the Finns had already jump-started to the ‘week-end’ on Saturday. However, the link between the main partying night and the day after was not straightforward. The homicides of the ‘day after’ were often committed later in the day, not during the early hours. Some of them may have reflected disagreements triggered in the previous day or evening. In the overall data, the percentage of offences committed under the influence of alcohol was highest in crimes committed on Sundays and Mondays (Kivivuori et al., 2020b).

In terms of routine activity theory, the Sunday social activi-ties seemed to have functioned as a crime attractor in the early modern age. It is not easy to tease out the role of church services in this, as not all Sunday incidents were connected to attending

church services. The aberrant Finnish week cycle pattern was probably linked to the specific conditions in the northern Finnish research region. Owing to long distances, going to Sunday church could require an overnight stay. The tradition of Satur-day bathing in the sauna could also impact the weekly pattern (Kivivuori et al., 2020b). The social nature of sauna-related activ-ities (Vuorenjuuri, 1967) was such that they could be linked to risk of conflict.

Year

If homicides were equally distributed over the calendar year, each month would have approximately 8.3 per cent of the crimes (Figure 8.2).¹⁰⁹ In reality, the months were not identical in terms of how many homicides occurred during them. In the monthly dis-tribution of homicide incidents, April was the peak month in our combined data (12 per cent of incidents). There was some varia-tion between the three study regions, but the overall pattern indi-cates a distinct period of heightened homicide occurrence. This period lasted from March to July, even though the difference from expected frequency is not very extensive.

Within the spring/summer period, the peak homicide month differed in the three study regions. The peak was March in Finland, June in Sweden, and July in Denmark. The earlier Finnish peak may have been linked to multiple factors, such as slash-and-burn agriculture and naturally occurring lighting and weather differen-tials in the north. The possibly relevant natural conditions include darkness during winter in areas close to the Arctic Circle, long distances, and post-winter weather conditions making transpor-tation and control functions more difficult. If these are relevant, they likely interacted with social-economic factors. However, it needs to be added that the Finnish outlier position is based on the single months of March and September. Overall, there seems to have been a risk period in the spring and early summer, and a lull in the autumn.

109 Figure 8.2 adjusts for month duration in days.

–15

Difference to equal distribu�on, percentage points

Monda

Difference to expected distribu�on, percentage points

Januar

Figure 8.1: Homicide by weekday in the early modern period: difference to equal distribution, percentage points, Nordic countries, 1608–1699.

Figure 8.2: Homicide by month in the early modern period: difference to expected distribution, percentage points, Nordic countries, 1608–1699.