Reader “Albert Nock” was kind enough to send me notice of some new research on violence on the American frontier. This relates back to an article I wrote for Mises.org in 2004 and to extensive work done by Terry Anderson and Peter Hill (including an article in the JLS). There is also a chapter on the topic in Thomas Woods’ latest book.
The article, “Guns, Murder, and Probability” by Randolph Roth, makes the case for a very murderous West based on some reseach done on the California and Oregon frontiers in the late 19th century. Roth provides a statistical case for why the California and Oregon numbers, which show a high number of murders per capita, should provide a more accurate view of the frontier overall than the work done by Dykstra and Hollon, for example, showing that the frontier was quite mild in its violence.
Roth clearly is in search of the best data he can find, but he draws conclusions that may be far too broad.
First of all, I would have to point to the methods in the Oregon research right away. If the research is based on newspaper articles as Roth notes, then that could be thoroughly unreliable. As noted here about killings in Bleeding Kansas, contemporary accounts of killings on the frontier can be notoriously unreliable due to exaggeration. The same may be true of the California data as well, depending on its reliance on contemporary journalism.
The problem with drawing broad conclusions, as Roth does, is that the California case may be unique, and gives some indications that it is in fact unique, and not suited to be used as a generalization for the entire Western frontier. Roth notes that the areas he examines are areas of high Anglo-hispanic conflict, and that even among the areas he surveys, there are wide variances in the murder rate. In other words, we can’t even generalize a murder rate to the sample area, let alone all of California or the entire West.
Also, if we really see such a large difference in murder rates between that of Kansas (as shown by Dykstra) and the murder rates in Southern California or Oregon (as cited by Roth) doesn’t this all the more call broad conclusions into question?
But then, one could easily say that we have no reason to believe that the Kansas numbers, to name one “low-violence” case, are better than the California case. Fair enough, but which is really more typical for the West? Is a cattle town more typical or is Southern California?
Is Southern California even the “frontier” at this point? After all, it had been settled by Hispanics a century earlier, so we really have conflict betwen 2 European-derived groups, not just whites and Indians, as was generally the case elsewhere on the frontier. Also, it appears that the research cited by Roth is counting murders perpetrated on both Hispanics and on Anglos.
Thus, we’re looking at an area where the politics of the Mexican war and the resulting oppression of the native Hispanic population is a salient issue, undoubtedly leading to considerable violence between non-Indian citizens of California. Can we really extrapolate from this to the rest of the frontier?
These are just some initial observations, but It’s not clear that the theory of the “Mild West” should be overturned quite yet.