Scholars broadly agree that there was an important evolution in popular presidential communication during the eighteen and early nineteenth centuries. However, the causes behind the increased frequency of spoken presidential rhetoric are the subject of vigorous scholarly debate. This disagreement can be distilled to two essential arguments: changes in presidential rhetorical behavior either were the product of a changing set of cultural norms and ideas or were the result of the presidents taking advantage of the emergence of new strategic opportunities provided by technology and changes in the structure of political parties. I propose to clarify the issues in this debate by presenting an analysis based on new data and a multimethod research strategy.
In this article, I challenge many of the common assumptions about the rhetorical behavior of early presidents and look closely at a number of competing claims surrounding the mode (spoken vs. written), content (intellectual quality), audience (who the president is addressing), and frequency (number of times the president speaks) of spoken popular presidential communication (SPPC). (1) I show through a number of empirical tests that the political opportunities and incentives that presented themselves in the nineteenth and early twentieth century shaped presidential rhetorical behavior rather than the president's personal or partisan ideology. I present a coherent, alternative view to the claim that ideas were responsible for changes in the communication strategies of American presidents. In sum, the rhetorical practices of American presidents evolved over time in response to a changing political landscape.
Presidential Popular Communication
Jeffrey Tulis (1987) hypothesized that Woodrow Wilson and his innovative conception of the constitutional role of the executive was fundamental in the development of the twentieth-century rhetorical presidency. Frequently cited with Tulis, is Samuel Kernell (1993) whose causal mechanism was different: a decline in party strength and a changing media environment led presidents to bypass the bargaining processing in DC and "go public" with their policies instead. Mel Laracey (2002, 2008, 2009) argued that the origins of the rhetorical presidency were much older and found in the use of newspapers by nineteenth-century presidents. Laracey proposed that political party philosophy led nineteenth-century Republicans to be more restrained than their Democratic counterparts in their public communications (Laracey 2002, 2008, 2009). Richard J. Ellis found that early presidents were not necessarily constrained from speaking to the public but that norms discouraged them from appearing as partisans (Ellis 2008). Scholars such as Kernell (1993), Gamm and Smith (1998), Korzi (2004), and Klinghard (2005, 2010) emphasized opportunities and incentives generated by changes in the functioning of political parties, technological innovation, and the commercialization and professionalization of journalism and the press. These changes encouraged presidents to speak to the public more frequently for political gain. (2)
While many authors have challenged Tulis, the debate continues, in part because of the lack of systematic analysis over time (but see Bimes 2009). Many challenges have focused on individual presidents and their rhetorical behavior (Arnold 2009; Ellis 1998; Ellis and Walker 2007; Medhurst 2008; Saldin 2011). Elvin T. Lim's The Anti- Intellectual Presidency (2008) stands out for providing systematic, quantitative support for Tubs' conclusion that Wilson was an inflection point in the development of SPPC. Similarly, to Lim's argument, many other scholars lament the decline of presidential rhetoric (including Hart 1989; Lowi 1985; Stuckey 1991). With this in mind, Lim's work describing the systematic decline in intellectual quality of SPPC is an important part of the larger debate surrounding presidential rhetoric. (3)
Finding Popular Presidential Communication
The opportunity to advance debate comes in part because of technological advances. As more historical newspapers have been digitized, it has become markedly easier to search them. In this case, it has made it possible to find many additional instances of presidential rhetoric that earlier scholarship may have missed. Similarly, it is simpler to find contemporary reaction to these speeches.
In order to find instances of SPPC, I searched digitized historical newspapers using key word searches. (4) These searches returned headlines with related articles that included these key words. For the early presidents, I used the Early American Newspaper collection, the Nineteenth Century Newspaper collection and the American Periodicals series. I searched the Chicago Tribune beginning with 1851, the New York Times beginning with 1855, the Washington Post and Los Angeles Times beginning with 1877, and the San Francisco Chronicle beginning with 1881. Once all of these newspapers were in existence (1881), I stopped searching the other collections. (5)
I coded information about each speech, including the date of the speech, the greeting, the newspaper in which it was published, the date it was published, the location in which the speech was given and the topic that the speech was addressing. I also coded the speech as having substantive or ceremonial content. In the end, I collected over 3,000 instances of SPPC. While most of these speeches have been discovered and reported before, they have never been compiled and systematically coded.
Challenging the Assumptions of the Rhetorical Presidency
In order for Tubs and those that subscribe to his version of ideational change in rhetorical behavior of the president to be right, a number of things must be true. Firstly, there should be a marked increase in the frequency of SPPC following Wilson's articulation of his ideas about presidential leadership. This assumption is presented as Hypothesis 1:
H1, (The Frequency Hypothesis); The annual frequency of popular presidential communication is greater beginning with Woodrow Wilson. Alternatively, more frequent SPPC appeared gradually in a way consistent with the gradual diffusion of new communication and transportation technology.
Tulis' main source of evidence was the official record that he obtained from the Richardson papers (Tulis 1987, 137). (6) By taking a sample of these official documents and combining it with all Annual Messages and Inaugural Addresses, Tulis concluded that a very small amount of presidential rhetoric before 1913 was spoken. The amount of spoken rhetoric in the twentieth century swelled dramatically, and Tulis attributed this to Wilson's philosophy. However, there was a significant amount of pre-Hoover SPPC excluded from the Richardson papers and therefore from Tulis' estimates (Pluta 2013).
Contrary to Tulis' argument, there was no significant increase in the frequency of SPPC coinciding with Wilson's presidency. Instead, the upward trend began in the post--Civil War period. Importantly, a number of Wilson's predecessors averaged more speeches per year than Wilson (see Figure 1).
The trend was not entirely consistent, as Chester A. Arthur and Grover Cleveland were relatively quiet. There are a number of competing explanations as to why Arthur and Cleveland spoke less than their contemporaries did. Importantly, both had chronic illnesses that limited their ability to tour extensively. Cleveland was exceptionally quiet in his second term. This was most certainly because of mouth cancer surgery that required Cleveland to have an artificial jaw. Cleveland was concerned with keeping his illness from the public (Welch 1988, 120-21). Some scholars argued that Cleveland used written communication to express his policy preferences. (7) According to his primary biographer (Howe 1934), Arthur was uninterested in being president and suffered from Bright's disease, which led to constant fatigue. With these exceptions noted, there was a significant increase in SPPC before Wilson. A number of Wilson's contemporaries including Theodore Roosevelt, William McKinley, and William Howard Taft spoke more frequently than Wilson did.
Was There a Constitutional Norm Regarding Presidential Rhetorical Behavior?
The mere fact that there was more SPPC than estimated does not itself preclude the existence a constitutional norm. A more likely indicator of a constitutional norm would be an acknowledgment of such an expectation by contemporary commentators. This leads me to my second hypothesis:
H2 (Constitutional Norm Hypothesis): A challenge to the dominant norm regarding presidential rhetoric should result in widespread critical commentary from newspaper reporters and editors. Alternatively, if no norm exists newspapers should focus on other reasons for changes in presidential rhetorical behavior.
In an effort to find evidence of such a norm, I read newspaper reports surrounding two critical inflection points for Tubs' theory: The first instance was Thomas Jefferson's decision in 1801 to stop appearing before Congress (both George Washington and John Adams had appeared before Congress to deliver the Annual Message). The second involved Wilson delivering the State of the Union for the first time in 1913- While newspapers widely reported and reprinted Jefferson's message, I was unable to find commentary directly related to the mode of the speech. Of course, I may have missed something, but if Jefferson's innovation was so profound as to be constitutional, such commentary should have been widespread. Biographers suggest that Jefferson instituted the change because he felt a written address was more efficient and that it eliminated the spectacle of the presidential entourage arriving at the Capital (Cunningham 1988, 246-7; Ellis 1998, 227; Malone 2006, 94-95; Randall 1993, 559). (8) The Federalists were unhappy with the tone of the message and skeptical of Jefferson's vision of government, but no...