In this section I will compare the saeculum to rising and falling social trends in an effort to establish correlations. Strauss and Howe mention substance abuse as an indicator:¹⁰

Figure 1 presents trends in per-capita American alcohol consumption^11,12 and British spirit production¹³ (assumed to reflect consumption) as examples of this indicator. Strauss and Howe's observations about the US patterns can be easily seen in the figure. Nineteenth century trends in alcohol use in both counties are broadly similar. The highest consumption levels of all time were seen during the 1830's in both countries. Consumption fell to a trough mid century and rose to a lower peak around the beginning of the twentieth century. This broad correspondence suggests 18^th century patterns in Britain may be indicative of the situation in America. If this is true I note that alcohol use peaked in 1743 (the end of the Great Awakening) in accord with Strauss and Howe's correlation of alcohol use with spiritual awakenings. On the other hand, there is some evidence of another period of heightened alcohol consumption around 1800 that does not correspond to an Awakening.

Figure 1. Alcohol consumption trends in the United States and Britain 1700-2000

Strauss and Howe describe the spiritual awakening period as one in which "sex role distinctions narrow, public order deteriorates, and crime and substance abuse rise."¹⁴ This suggests another indicator might be crime rates. Figure 2 presents crime rates for the United Kingdom¹⁵ and the United States.^16-18 Crime rate did rise during the Consciousness Revolution (1964-1984) in both countries, but not during the Third Great Awakening (1886-1908). One the other hand, a rising trend in crime was certainly present in Britain during the Transcendental Awakening (1822-1844).

Figure 2. Trends in crime in the US and UK, 1800-2000

Figure 3 shows four measures of 18th century crime incidence from four British localities: (1) the incidence of property crimes in the urban parishes of Surrey over the 1736-53 period,¹⁹ (2) the annual committals to Newgate (the City of London jail) in the period 1748-62,²⁰ (3) rates of theft in Staffordshire over the 1743-1805 period,²¹ (4) the incidence of game offenses in Wiltshire in the 1756-1800 period.²²

Figure 3 Crime rates in Great Britain 1735-1855

The United Kingdom has collected national crime statistics of varying degrees of detail since 1805.¹⁵ This crime index is also shown in Figure 3. The five data sets in Figure 3 are presented as indices with the year of maximum incidence set equal to 1.0. A composite index was constructed out of the four sets of local data to obtain a consensus crime index over the 1736-1800 period. The data in Figures 2 and 3 suggest falling crime beginning around the end of the Great Awakening (1727-46), the Transcendental Awakening (1822-44), and the New Consciousness Awakening (1964-84) in accordance with Strauss and Howe's observations about crime. There are also apparent periods of rising crime in the late eighteenth and early twentieth centuries that do not correspond to Awakenings.

Strauss and Howe also mention breakdown of public order. Figure 4 plots annual days lost from strikes relative to the maximum year (1946) and a 15-year moving average of the frequency of race riots, also relative to a maximum year (1965). These two indicators tend to move together and indicate that the 1960's, the 1940's and the 1910's were periods of social turmoil. Two of these three occur in social moments.

Figure 4 Labor Unrest and Racial Violence 1860-2005

Table 4 summarizes the trends in alcohol use, crime, strikes and race riots shown in Figures 1-4. Rising trends in unrest and antisocial behavior appear to correlate with social moments or at least with Awakenings. The correspondence is far from perfect and the alignment between trends and turnings abruptly start and stop. For example, rising trends in crime and alcohol consumption in Britain appear to correlate with both crises and awakenings in the eighteenth and early nineteenth century, after which the correlation breaks down. The same trends in America during the twentieth century roughly correlate with Awakenings, as Strauss and Howe relate.

Table 4. Trends* on alcohol use, crime, labor and racial unrest compared to the saeculum

Table 4 presents some intriguing relations, but the significance of the correlations is open to question because of the "patchy" nature of the regions where the correlations exist and my ad hoc combinations of trends to obtain matches with the Strauss and Howe turnings. The early period of correlation between crime and alcohol use lasts only a little more than one saeculum, and then breaks down. The later correlation between labor/racial strife, alcohol use and crime with the saeculum is not perfect and lasts just one saeculum. In the mid-to-late 19^th century all the correlations break down and there is little relation between Strauss and Howe's turnings and any trend so far discussed.

On the other hand, historian David Krein has demonstrated statistically significant effects of generational membership in voting patterns in the British House of Commons over the 1842-1858 period.²⁴ Krein identified four generations who were active in Parliament over this period (Table 5). Because politics normally cuts across generational lines, one would only expect generationally significant voting when parties are in disarray. When Whips have trouble keeping party members in line and members value and are able to exercise independence, voting along generational lines should be detectable.

Krein identifies the 1842-57 British Unraveling turning as just such a period:

He selected 270 divisions taken in the House of Commons from September of 1841 to May of 1859. They were not randomly selected but were chosen by the issues they raised: all divisions that were regarded as votes of confidence; all divisions taken in a full House and which involved major issues that attracted the attention by contemporary commentators or of historians; plus some highly emotional issues. Fifty-three of the 270 divisions, taken over the 1842-1858 period, showed statistically significant generational voting behavior.

Table 5. Four British generations used by Krein in his Parliament voting study

This demonstration establishes that generational membership is a real thing and relevant to political behavior, that is, generations create history. Not only that, but it establishes generational behavior in Britain (which supports my use of British trends to supplement American ones) and by men who would be active during much of the nineteenth century (when the trend data in Table 4 is most sparse). Taking both Table 4 and Krein's study together, there is a fair amount of evidence that cyclical history along the lines of what Strauss and Howe suggest has occurred since the eighteenth century in Britain and America.

A more compelling demonstration would be correspondence of the saeculum with historical trends over a great length of time without the correspondence breaking down. To look for such a correspondence, it is useful to work with as long a time span as possible. Strauss and Howe supply dates for more than five centuries, from 1435 to the present. They claim the saeculum began in the fifteenth century, but this belief may simply reflect the inability of their biographical methodology to detect the saeculum before the fifteenth century. Historian David McGuinness has developed a saeculum spanning five millennia.²⁵ A portion of his saeculum is reproduced in Table 6, where it is compared to that of Strauss and Howe.

Table 6. Construction of a "working saeculum"

The two versions of the saeculum are quite close. Thus, I have created a "working saeculum" for the purposes of this study by concatenating Strauss and Howe's saeculum with the pre-1435 McGuinness saeculum. McGuinness's Bohemian Unraveling runs from 1445-1471, whereas the corresponding Strauss and Howe Unraveling runs from 1435-1459. The Strauss and Howe dates are based on an identified procedure (cohort biographies). In contrast, the McGuinness dates are simply given with no indication of how they were determined. I favor the Strauss and Howe dates because of this difference in methodology and so will use the 1435-59 dates for the Unraveling. Doing so shortens the Hussite Awakening to just 20 years (1415-1435), compared to the 37 year length of the prior Florentine High (the longest of all the turnings in Table 6). I chose to change the divide between the Florentine High and the Hussite Awakening to 1406, which gives more uniform 28 and 29-year lengths for these turnings, consistent with typical lengths of surrounding turnings. Other than this, I use the McGuinness dates for the pre-fifteenth century period in my working saeculum.

Having now obtained dates for a full nine saeculae since the 12^th century, I can now look for long-term correspondences with social trends spanning centuries. There are no official statistics covering crime, recreational drug use, labor actions, etc. that span the time period of interest, so I have to take another approach than the prior analysis. Indeed, there are very few time series of any kind that stretch back to the Middle Ages. It will be necessary to create one.

The strategy I chose was something I call generational density analysis, which is really just various moving averages of the annual frequency of certain kinds of historical events. The kind of events chosen is based on the definitions of turnings. For example, it should be possible to track Awakenings by looking for events of the type McLoughlin (and Strauss and Howe) consider the very stuff of Awakenings. McLoughlin defines Awakenings as periods of cultural revitalization caused by a crisis in beliefs and values that produces a reorientation in those values and beliefs. This reorientation produces changes in institutions, worldviews, and cultural mores that reflect a new values/beliefs regime.

As examples of changes in belief regime one can look for the start of new off-shoots from mainline Protestant denominations, heretical movements springing from the Roman Catholic church, and the writing of treatises expounding new ideas. Institutional changes would include the founding of new religious orders, organizations, and churches. As evidence of spiritual crisis one might see an increase in visionary, apocalyptic or even hysterical behavior. Events like visions of the supernatural (e.g. apparitions of Mary in the Roman Catholic tradition or prophetic visions like Joseph Smith's encounter with the Angel Moroni) and mass-movements like the Crusades should be more frequent during Awakening turnings.

Figure 5. Spiritual event frequency 1200-2000

I call events like these "spiritual events". By constructing a timeline of such events and plotting the annual frequency of such events over time, it might be possible to identify periods with an unusually high incidence of spiritual events. Presumably, such periods would be spiritual awaekenings. Appendix A contains a spiritual event timeline from which annual frequencies of spiritual events were obtained. These were analyzed using moving averages to identify eras showing unusually high frequencies, which should constitute awakening eras. A 25-year moving average before 1820 and a 15-year average was used after 1820. As the numbers of events recorded rises, a shorter averaging period provides improved resolution while still having sufficient numbers of events to analyze. The results appear in Figure 5. Nine periods of high frequency of spiritual events can be seen. With one exception, these periods correspond extremely well with awakening turnings in Table 6.

To look for social moment tunings in general I considered the sort of events that can radically change the social environment, which is what a social moment is about. Events that can produce radical change are necessarily challenges to the status quo; events that upset social stability. Another term for such events would be "unrest". I would expect social moments to be periods unusually rich in "unrest events", events like peasant uprisings, urban riots, labor strife, slave revolts/race riots, financial collapse, revolutionary/civil wars and the like.

Figure 6. Unrest event frequency (1650-1700)

Appendix B contains an unrest event timeline and Figure 6 shows a normalized column plot of the annual frequency of such events over the 1650-2000 period. The frequencies in Figure 6 were scaled by dividing the individual values by its average value over the preceding 25 years. Scaling accomplishes two purposes. Firstly, it makes the range of frequencies roughly the same over time which aids is visualizing the cycles. I found many more recent events than older ones and a "raw plot" of the data would be dominated by the 20^th century peaks. Secondly, it weights events occurring in the beginning of period of high unrest more heavily than events occurring at the ends of such periods. It seems to me that unrest following periods of relative calm are more significant that unrest occurring in the midst of lots of other unrest. Some clusters of high frequency events in Figure 6 can be discerned by inspection, although they are not very clear. A centered moving average is plotted to help the detection of clusters. This moving average is 25 years long before 1810 and 15 years long after 1810. Eight clusters of high unrest can be seen.

Figure 7 presents scaled unrest event frequencies for the 1200-1700 period. Events were scaled to the average frequency in the previous 50 years before 1500 and previous 25 years after 1500. The longer period before 1500 was used to prevent division by zero because there are 25-year periods devoid of events prior to 1500. The scaled annual frequency of unrest events from Appendix D was subjected to a 25-year centered moving average, which is plotted as the black line in the figure. Periods of relatively high and low levels of unrest are labeled in the figure.

Figure 7. Unrest event frequency (25-year moving sum) 1200-1700

The composite dates in Table 4 and trend dates from Figures 5-7 are reproduced in Table 7. I averaged the data together to obtain a composite cycle which I compare to the working saeculum. The composite cycle identifies empirical social moments that closely align with the corresponding turnings from the working saeculum for 13 successive social moments. This alignment is statistically highly significant (>99%). The alignment breaks down for the next two social moments and then begins to align again after that.

Table 7. Construction of the empirical saeculum

This lack of correspondence between the empirical data and the saeculum is real.. There is no shortage of unrest or spiritual events, the expected peaks simply do not occur at the right places. Not only that, but the early twentieth century unrest peak is confirmed by the official strike statistics (Figure 4). Furthermore, the rising spiritual frequency from ca. 1895 to 1933 (Figure 5) is corroborated by Robert Fogel who dates the Third Great Awakening from 1890 to 1930.²⁸ There doesn't seem to be an error in my event analysis, the historical trends simply do not match with the generational biographies over the 1850-1930 period.

After 1930, the old pattern of alignment appears to re-establish itself. If the 1850-1930 period is left out, the characterization of the saeculum in terms of empirically-determined trends has been highly successful for eight centuries of history. This observation is sufficient to establish that the saeculum is a real historical cycle despite the problems over the 1850-1930 period. Generations do create history as Strauss and Howe claim.

The problems over the 1850-1930 period reflect what Strauss and Howe call the Civil War anomaly. Examination of Table 3 shows that the Civic generation associated with the 1844-1860 unraveling is missing. The regular cycle of generations experienced a "hiccup" at the exact same time as did their correspondence with the empirical saeculum in Table 7. Clearly, something happened to the saeculum around the time of the Civil War. As we shall see later, this "hiccup" is a consequence of a fundamental shift in the basic mechanism of the saeculum.

Return to index