Some of youu may be familiar with Pitrim Sorokin's work on historical cycles (Sorokin, Pitrim. 1937. Social and Cultural Dynamics. American Book Company, New York). I used his methods to measure the severity of internal instability (social unrest) in America. Up until a few months ago I was inclined not to count mass shootings at all. But in recent times it has become a big issue on both sides of the political aisle, with the Left pushing the gun control angle and the Right talking about terrorism after San Bernardino. In the spirit of actual theory posts, would anyone be willing to check out Sorokin’s book from their library and propose a a modified rule to account for changes in the world since 1937 when Sorokin was working. After all, the whole phenomenon of rampage/spree shootings was after 1937 and so Sorokin could not have considered this. My original analysis is outlined below:

I constructed a database of violent internal instability events for American since 1700. I divided the violent events into one of four kinds: political, labor, racial and rampage. Examination of each kind of violence revealed clusters of these events. These clusters were evaluated as a single episode of internal unrest using the methods of Sorokin (1937). Two discrete events (Revolutionary & Civil War) were also scared using Sorokin’s methods . The table below presents American history in a way that looks very much like the tables for English history presented by Sorokin .

Date	Outbreak	Events	Death	Sorokin
1770-74	Colonial sedition	7	28	11.4
1775-81	American Revolution	1	33000	67.2
1786-87	Shays Rebellion	5	20	9.1
1794	Whiskey Rebellion	3	5	9.1
1800	Prosser Revolt	1	26	3.7
1811	German Coast Uprising	1	65	7.7
1834-35	The Great Riot Year	5	8	10.0
1830-1841	Nat Turner Revolt	18	291	20.1
1834-38	Wave of Economic Unrest	4	20	6.7
1850-53	Wave of Economic Unrest	5	13	6.7
1855-58	Bloody KS/Knowthings	22	95	19.6
1861-65	Civil War	1	750000	66.9
1866-1876	Klan Insurgency	177	1933	46.4
1877	Great Railroad Strike	7	82	17.1
1866-74	Reconstruction	78	374	37.6
1886-1898	Jim Crow Terror	78	451	16.3
1894-5	Miner & Railroad Strikes	12	39	21.5
1912-14	First Miner's War	25	309	19.6
1915-19	Radical Labor unrest	36	81	27.1
1917-1921	Red Summer	99	924	33.5
1920-22	Second Miner's War	17	101	19.6
1963-1970	Post Civil Rights Unrest	62	365	42.4
1967-71	Anti-War Violence	12	23	12.6
1982-1989	Rampage Wave	14	154	6.3
1994-97	Terror Wave	8	177	7.9
1991-2000	Rampage Wave	28	178	10.0
2002-2012	Rampage Wave	46	351	12.6
2013-2015	Rampage Wave	123	583	10.0

The Sorokin measures were used to produce an annual index of Sorokin-scaled violence over time and an exponential average run through them. Very distinct peaks in instability were obtained that revealed a regular pattern:

Date	Instability	Turning
1741	1.2	2T
1781	5.6	4T
1838	2.4	2T
1865/77	6.6/6.2	4T
1895	3.8	2T
1920	6.6	3T
1970	3.9	2T
1997	1.6	3T
2015	1.8	4T

All but one of these peaks occur in a social moment turning (2T or 4T). The exception was the 1920 peak which occurred in a 3T. Also of interest is the absence of a peak for the last 4T. It looks like the violence peak for that happened early, but did not coalesce into a 4T because the generations were wrong. Hence the necessary changes did not happen and we had to wait for the 1929 economic collapse for a 4T.

The recent period is confusing. There was a definite peak in instability around 1970, as well as a host of other events that indicate s social moment was happening. Since then the instability measure has been low. The reason for this primarily has been the disappearance of riots as an expression of unrest. This may reflect the very aggressive police tactics that have come into use since the 1970’s. This matters because the Sorokin measure considers the number of people involved and considered property damage as a factor. Since riots involve groups and property values, for a given number of fatalities riots will score higher on the Sorokin scale that would other kinds of events such as terrorist attacks involving a small number of perpetrators. A common type of such other events today are mass shootings that feature small numbers of people involve and so get low Sorokin scores.

Thus, the current value of the exponential average of Sorokin violence at 1.8 is the highest reading since 1978, but it is barely higher than the 1997 peak. This peak is partially due to mass shooting violence, but also to a wave of domestic terrorism that Bob Butler was tracking as one of his “spirals of violence” a concept I have found very useful in my thinking about the cycle.

The reason for the slightly higher peak today is that there have been a LOT of these mass shootings. However the current “peaks” are quite small compared to previous peaks. The casualties from the modern events are large, for example the toll from the rampage wave over the past three years exceed the toll for the 1970’s violence that rates a much higher Sorokin score. On the other hand the deaths associated with most of these events (except for the two wars) are small compared to the homicide rate. For example, 583 dead from mass shooting over the past 3 years sounds really bad, but over this time there were about 34000 gun homicides, so we are talking about 1-2% of all the killings. When we consider that these 583 were killed in 123 separate events, the question arises, in this internal instability or crime? A riot is an explicitly political event involving many people, most of whom are presumably not criminals. But a spree shooting? Hard to tell.

Without being familiar with Sorokin's work, I have been thinking thoughts along these lines recently, as well. I am inclined to think these mass shootings are an example of despair in an age of social alienation and much more effective law enforcement.

That being said, riots have disappeared. Baltimore and Ferguson both saw riots recently, and Trump's rallies have started to see some brawls as well. It would be interesting to see how these things evolve, if Turchin's predictions are right.

It's also worthwhile to point out that during the 3rd Party system riots and public brawls potentially had political benefits for the organizers, in a ways that doesn't hold true now.

Originally Posted by JordanGoodspeed

Without being familiar with Sorokin's work, I have been thinking thoughts along these lines recently, as well. I am inclined to think these mass shootings are an example of despair in an age of social alienation and much more effective law enforcement.

I think the same thing. The question is how important are these? As I showed above, instability in the form of riots, assassinations, lynching, labor violence and internal warfare when viewed through a concept of clusters of such events forming discrete events analogous to baronial revolts and such in earlier times that can be analyzed by a common method over many centuries of time have shown generational cycle patterns that match the turnings in America (and have been called "father and sons cycles" by Turchin for earlier eras in Europe). The question is whether mass shootings are the same kind of events that define turning/fathers and sons cycles.

Originally Posted by Mikebert

I think the same thing. The question is how important are these? As I showed above, instability in the form of riots, assassinations, lynching, labor violence and internal warfare when viewed through a concept of clusters of such events forming discrete events analogous to baronial revolts and such in earlier times that can be analyzed by a common method over many centuries of time have shown generational cycle patterns that match the turnings in America (and have been called "father and sons cycles" by Turchin for earlier eras in Europe). The question is whether mass shootings are the same kind of events that define turning/fathers and sons cycles.

You could certainly make a fair argument that they are. Just as industrial age riots and labor unrest were different from baronial revolts and peasant uprisings, mass shootings (targeted primarily for media consumption) and terrorism could be their information age equivalents.

Originally Posted by JordanGoodspeed

You could certainly make a fair argument that they are. Just as industrial age riots and labor unrest were different from baronial revolts and peasant uprisings, mass shootings (targeted primarily for media consumption) and terrorism could be their information age equivalents.

Again, that is how I am thinking. The question then becomes how do you measure it? Here is summary of the measure Sorokin used:

The scale of instability events varies enormously from one-day local riots to multiyear civil wars. Sorokin provides a measure of severity. This measure was a geometric average of three factors representing extent, duration and intensity. Extent was 1 for a single rural location; 3 for a small town or several rural locations; 5 for a larger town, 10 for several such towns, one city, or a small rural county; 20 if it is in the capital or a large rural county; 40 for the whole capital or several large counties, 60 if both in capital and several counties; 80 if it involved the majority of the country and 100 if the whole country was involved (Sorokin 1937:394). Duration was 1 for a few days; 3 for a few weeks, 5 for several months, 10 for a year; an additional 5 per year up to 30 for five years, an additional 4 per year for 6-15 years (yielding a value of 70 at 15 years) and 3 more for each year after 15 (Sorokin 1937:395).

Table 1. Intensity values given to internal disturbances (Sorokin, 1937:396)

Number involved index value	Amount of violence index value
I	II	III	IV	V
I	1	3	5	7	10
II	3	10	15	20	30
III	5	15	25	35	50
IV	7	20	35	50	70
V	10	30	50	70	100

Intensity was based on the number of people involved and the amount of violence (Table 1). For the number of people involved the indices were: I for one or a few persons; II for a group; III if it involved a large social class, extensive occupational class, or economic, racial/ethnic, political, or religious group; IV if it involved several extensive classes; V if it involved practically the whole populace. For amount of violence they were: I for no violence; II for some violence; III for significant destruction of life and property; IV for an even larger amount of violence involving the overthrow of the government in various centers, but without lasting sociopolitical effects; V full-scale civil war leading to overthrow of the central government with lasting sociopolitical effects (Sorokin, 1937:395).

For the modern era when numerous events like riots or terrorist attacks occurred which were of a much smaller impact as medieval revolts, the nature of the “events” being analyzed had to change. Consider 30 years in a row in which a riot occurred. Each riot takes place in a single location during a single day and so rates a 1 for both both extent and duration. For intensity is a II for numbers involved and a II for amount of violence (property damage and injuries but no deaths) giving it an intensity of 10. The geometric average of these three is 2.15. Thirty of them in a row sum up to 65. In comparison the Civil War rates a 67. My database features riots very year, this analysis implies that the Civil War is like 31 years of ordinary run of the mill-low level unrest. This is nonsense.

If we break up the Civil War into individual conflicts, rated each one individually and then summed them up, the Civil War would have a huge score, but this isn’t how the Sorokin measure works. Since such details are not known for most of the events Sorokin considers, he looks at wars holistically and applies the rules to the period as a whole. To really use this method I should first assemble my instability database into war-like conflict periods. If you plot the frequency of certain kinds of violence such as racial violence, or labor violence you can see they come in waves. Each wave can be analyzed like a war or revolt might be. So the Great Railroad Strike of 1877 comes across as a very big deal because it paralyzed the nation’s transportation system, affecting a large portion of the country giving it a 40 for extent. And since it featured an occupational class is gets a III for number of people involved. Since 82 died it rates a III for violence, which gives it a 25. Duration (45 days) was more than a few weeks but less than a year so that is a 5. All together I get an intensity of 17, same as Jack Cade’s rebellion in 1450, a major event of its time, but dwarfed by the nearby civil wars (US Civil War at 67 or the four wars of the Roses, about 40 each).

This strategy seems to work to put all episodes of instability on the same (logarithmic) scale. But it breaks down with these rampages. First of all they don’t cluster neatly in waves. They are far more bloody than riots, they are mass murders after all. On the other hand, they are the perfect capitalist form of violence, they destroy lives, but leave property untouched, so in that sense they aren’t really very severe at all. This idea is supported by the fact that there is no significant response by elites to these events. Compare the current rampage wave since 2002, which I have with a total intensity of 22.6. This is higher than the 2^nd miner’s war, yet the elite response in the Battle of Blair Mountain the largest skirmish in the war was way beyond anything to date for the ongoing rampage violence.

At present the way I have rampages measured they show up as low-level “background” violence, of no real account.