NOTE ON PRIMARY SOURCES
THE HISTORICAL RECOVERY OF IDEAS—as well as practices and habits—has been a long-standing problem philosophical and methodological for both historians and social scientists. Foucault famously challenged genealogists to “read everything” in generating their archives. There are many cases of Herculean efforts undertaken by solo researchers to recover pieces of history. For example, in his historical reconstruction of French ideas circa 1899, Marc Angenot told his “story of the year” from an archive of 1, 200 books, 150 daily newspapers, and 400 periodicals.1 My genealogy was less ambitious. To tell six stories of the year from five national perspectives, I relied on fifty books, forty daily newspapers, and eighty periodicals. In the end, the archive of primary source texts subjected to analysis came to consist of nearly 3,000 discreet units, ranging from a short newspaper article to a novel: 800 for Chapter 2, 500 for Chapter 3, 900 for Chapter 4, 700 for Chapter 5. This appendix describes the sampling strategies I used in accessing data in each chapter.
As I explained in Chapter 1, I executed my case studies in three analytical steps, each corresponding to a level of the content and contestation of identity: society, debates, and decision. The selection of texts in steps one and two was based on circulation, region, ownership structure, and political or ideological orientation. (These dimensions are referenced in social history and in the standard guides on national literatures). I excluded publications the contemporaries regarded as extreme and thus unrepresentative of the discourses and debates of interest. For example, for the British archive on Venezuela II, I excluded the pro-American The Spectator and the antiAmerican Saturday Review. Because documentary sources like these favor the voices of journalists, an attempt was made to mine letters written by more ordinary citizens (“letters to the editor”) wherever possible.
This type of historical research has become easier in the digital age.2 To generate a basic archive of newspaper and newsmagazine articles, I started with the electronic, full-text,