12

Using a Mixed Approach to Content Analysis

The Case of Apologetic Rhetoric in the Modern Presidency

Rico Neumann and Kevin Coe

Within a few months of taking office in 2009, President Barack Obama traveled abroad to announce his presidency to the world and to begin rebuilding some of the relationships that his predecessor's administration had strained. On April 3, for example, he was in Strasbourg, France, where he critically referred to earlier times when “America has shown arrogance and been dismissive, even derisive.”¹ The following day, Obama was asked about America's place in the world, to which he responded:

Two days later Obama addressed the Grand National Assembly of Turkey in Ankara, where he characterized relations between the United States and Turkey as follows:

Obama's self-critical rhetoric did not go unnoticed. US journalist James Kirchick, for instance, wrote in the Los Angeles Times that the president had undertaken a “grand global apology tour” during which he was guilty of “flagellating the nation he leads” (Kirchick, 2009).

President Obama's words, whether a true apology or not, are but one example of the brand of apologetic rhetoric that is common in the political arena. Indeed, political leaders' statements of remorse for historical transgressions, injustices, and wrongdoings, have been an essential part of many societies and cultures (Barkan, 2000; Digeser, 2001; Negash, 2006; Weiner, 2005; Yamazaki, 2005). In the context of the US presidency alone there are numerous examples. Among the most widely known are George W. Bush's attempt to restore national reputation after tragic incidents at the Iraqi prison Abu Ghraib and the detention camp in Guantanamo Bay; Bill Clinton's attempts to repair his personal image after the Monica Lewinsky scandal; Ronald Reagan's public statement on his role in the Iran–Contra affair; and Richard Nixon's response to Watergate. Given these and other high-profile instances, it is not surprising that there has been much recent discussion about the “age of apology” (Brooks, 1999; Cunningham, 1999; Gibney, Howard-Hassman, Coicaud, & Steiner, 2008; Govier, 2002; Harris, Grainger, & Mullany, 2006), accompanied by labels such as a “sorry world” (Nobles, 2008), and “America the apologetic” (Brookhiser, 1998). The body of research on apology is growing rapidly (Gibney et al., 2008; Negash, 2006; Nobles, 2008).

Notably, the literature on apologies – and public apologies in particular – is almost exclusively qualitative in nature, drawing on detailed analyses of individual cases or single presidencies (e.g., Edwards, 2008a; Harter, Stephens, & Japp, 2000). Very few scholars have investigated the topic using quantitative methods or mixed methods approaches (notable exceptions are Blatz, Schumann, & Ross, 2009; Kampf, 2009a). Given this imbalance in the literature, apologetic rhetoric is a particularly useful context in which to demonstrate the utility of a mixed approach to content analysis, one that precisely specifies quantitative trends in large quantities of text by blending computer-assisted content analysis (CCA) with traditional manual coding. In explaining and demonstrating this approach, this chapter also addresses several unanswered questions about apologetic rhetoric in one key context: the modern American presidency. Toward this end, we first briefly review the literature on apology and offer a conceptual definition of “apologetic rhetoric.” We then explain our method in some detail, so as to fully outline the steps that are necessary to undertake a mixed approach to content analysis. We then present the findings of our analysis, before concluding by considering the advantages and disadvantages of this particular approach to content analysis.

Apologetic Rhetoric

At the interpersonal level, many people do not even notice how often they use the word “sorry” to help undo a hurtful act. People apologize for myriad reasons. Whatever the motive, the goal is often to heal the relationship between the victim and the wrongdoer and to repair the wrongdoer's image (Benoit, 1995; Smith, 2008). Just as important as interpersonal apologies are those on a national or international level, where a genuine apology can be of profound value for the future relationship between the wrongdoer and the wronged (Lazare, 1995). The act of apologizing, however, becomes more complex (and therefore also less frequent) when apologies are offered on behalf of a group, institution, company, or nation (Tavuchis, 1993). Not surprisingly, literature on public apologies is more limited than writings that deal with the interpersonal sphere, yet the body of research has been growing rapidly throughout the past decade. According to Lind (2008), a great number of activists and academics have contributed to an increasing body of scholarship on contrition and apologies on the international level.

As a response to historical transgressions, leaders around the world have offered public apologies in an effort to heal relations and seek reconciliation; for example, Japan apologized to South Korea for military invasions before and during World War II (Yamazaki, 2005); Germany apologized for its colonial past in Namibia (Jamfa, 2008) and for the Holocaust (Negash, 2006); and Kofi Annan apologized for the United Nations' inaction during the Rwandan genocide (Edwards, 2008b). Several examples from the American context have already been noted. In all cases, such public apologies represent rhetorical attempts to (re)negotiate “membership in a group, community, nation, or humankind” (Towner, 2009, p. 431). National leaders, whose words carry particular weight, thus act as mediators by representing the interests of the national community on the international stage. As Kellerman (2006, p. 74) states, “a leader's apology is a performance in which every expression matters and every word becomes part of the public record.” Moreover, the willingness to offer a public apology is not only of high emotional value and may raise “the moral threshold of a society” (Brooks, 1999, p. 3), but it also bears a clear strategic value for the wrongdoer (Lazare, 2004).

Apologetic rhetoric can take many different forms and invites a variety of rhetorical approaches, strategies, and tactics. Specifying its usual form is made difficult by a lack of systematization and the explosive growth in literature on public apologies (Benoit, 1995). An early conceptual distinction was made between “apologia” and “apology.” The Greek term apologia originally appeared in Plato's work Apología Sokrátus and refers to the speech Socrates gave in defense against the accusations that were brought against him. Hence, it has been framed as a defensive rhetorical response to an attack on one's character and is therefore primarily focused on the self, to restore a damaged image (Ware & Linkugel, 1973). Despite its etymological similarity, an “apology” typically occurs without any defense and is “a form of self-punishment that cuts deeply because we are obliged to retell, relive, and seek forgiveness for sorrowful events that have rendered our claims to membership in a moral community suspect or defeasible” (Tavuchis, 1993, p. 8).

A notable effort to systematize the field was undertaken by Benoit (1995), who proposed a theory of image restoration consisting of five strategies which rhetors can rely on when faced with a crisis: denial, evading responsibility, reducing the offensiveness of the event, taking corrective action, and mortification. While denial – often accompanied by some form of blame shifting and self-defense – is most similar to the traditional apologia rhetoric, a mortification strategy usually strives for a true apology. Similarly, Kampf (2009a, p. 2258) focuses on the discourse of minimizing responsibility by utilizing “creative forms of apologetic speech.” Based on a detailed discourse analysis of apologies made in the Israeli public sphere, he demonstrates a variety of rhetorical tactics which can be grouped into four main categories: blurring the nature of the offense, compromising the apology's performative verb (e.g., “sorry” versus “apologize”), questioning the identity of the offender, as well as questioning the identity of the offended.

In a narrative synthesis, Towner (2009, p. 457) argues that although the extant literature provides many unique and detailed analyses, it “may have the unintended consequences of alienating scholars who might otherwise learn from and support each other's research.” In other words, the body of research has grown so rapidly that scholars following one approach tend to overlook important research from the other end of the spectrum. One way to bridge that gap is to embrace a broader conception of apology, which takes the many different creative forms of apologies into account. Various efforts have been made to describe (and prescribe) the “perfect” or “true” apology (Kellerman, 2006; Lazare, 2004; Tavuchis, 1993). However, given that conflict situations take multiple forms and offenses are perceived differently, different rhetorical strategies are required. With this in mind, we draw on previous conceptualizations of apology and apologia to offer a broad definition of apologetic rhetoric.

Our goal in doing so is to help bridge the gap between the extant literature, which is largely qualitative and specific, and the broad patterns that might exist across apologies for a range of different wrongs across an extended period of time. The large body of qualitative and normative writings is important in that it sheds light on specific cases. Still, broader understandings that might be gleaned from considering apology more broadly are needed. With this goal in mind, we define apologetic rhetoric as a strategic and value-driven communicative process in which an apologizing entity publicly and responsibly acknowledges a past misdeed, negotiates a shared understanding of the conflict situation with the harmed party, and ultimately asks for forgiveness.

Several elements of this conception are noteworthy. That apologetic rhetoric is strategic and focused on responsibility highlights that a public figure may employ a variety of rhetorical tactics in order to minimize – or “democratize” – the degree of responsibility (Kampf, 2009a, 2009b). It is value-driven because it is rooted in basic moral values that are reaffirmed, which allows an individual, a community, or a nation to grow morally and re-establish the moral balance and relationship between victim and offender (Hearit, 2005). In the political realm, it may also serve as a “platform for announcing new policy directions” (Nobles, 2008, p. 111). Apologetic rhetoric takes place in public because it is a collaborative process and shapes the discourse between national and international leaders, citizens, activists, interest groups, and the media. It seeks to negotiate a mutual (re)interpretation of the crisis situation “rather than an inherently psychological or scripted response to a conflict” (Paul, 2007, p. 4) in order to (re)interpret membership between the involved parties (Nobles, 2008). In a similar manner, Tavuchis (1993, pp. 57–58) notes that apologetic rhetoric “presupposes cognitive and evaluative congruence in the form of shared definitions of the violation, its severity, history, and implications.” In contrast to previous conceptions of apologia and apology, this broader definition of apologetic rhetoric allows for a diverse collection of apologetic expressions to be examined over time – a task that certainly invites CCA but also relies on the merits of a traditional approach via manual coding.

Research Questions

The above discussion makes clear the need for a broader analysis of apologetic rhetoric. We took a step in this direction by examining eight decades' worth of presidents' apologetic rhetoric. Our focus was the American presidency because US presidents are key figures in the nation and the world, and their apologetic rhetoric therefore has great potential for attention and impact. To keep the scope manageable, we focused particularly on situations where presidents face a foreign audience. Such situations may require a higher degree of rhetorical sensitivity, especially when relations with the nation in question have been strained. Compared to other addresses, speeches abroad are also the most likely to include some form of apologetic rhetoric. Four specific questions guided our analysis:

RQ1	To what extent have presidents used apologetic rhetoric in their public messages given to foreign audiences?
RQ2	What forms of apologetic rhetoric have presidents used in their public messages given to foreign audiences?
RQ3	How does the use of apologetic rhetoric differ in communications that are high or low in “message control”?
RQ4	To what extent have presidents used common rhetorical strategies in their public messages given to foreign audiences?

Planning a Mixed Approach to Content Analysis

Three basic steps are necessary in conducting a mixed approach to content analysis: (1) selecting an appropriate body of text; (2) administering the computer-assisted content analysis; and (3) administering the manual content analysis. This section details each step in turn.

Selection of Text

We made two choices to narrow the scope of texts for this analysis. The first, discussed above, was to focus on presidential speeches delivered abroad – a context in which apologetic rhetoric often occurs. The second was to limit our analysis to the modern presidency, which scholars have generally defined as beginning with the presidency of Franklin D. Roosevelt in 1933. This period of analysis is useful because the presidency changed significantly beginning with Roosevelt. Federal and national power grew, as did the United States' worldwide recognition as the leader of the free world as well as advocate of human rights and free markets, particularly highlighted during the Cold War and beyond (Greenstein, 1998, 2004; Medhurst & Brands, 2000). As Leuchtenburg (1988, p. 7) notes, “by almost all accounts, the presidency as we know it today begins with Franklin Delano Roosevelt.” Thus starting with Roosevelt allows for 13 presidents to be included in the content analysis but retains some consistency in the cultural context in which each president governed.

Our corpus of presidential communications therefore consisted of 2,186 messages delivered abroad, beginning with Roosevelt's visit to Haiti in July 1934 and ending with Obama's visit to Prague, Czech Republic, in April 2010. To identify foreign speeches, we first consulted the official website of the US Department of State (http://history.state.gov/departmenthistory/travels/president), which provides an overview of all visits abroad for every president included in our analysis. Using information about the specific date and country provided in these lists, we were able to locate and retrieve the transcripts from the National Archives' “Public Papers of the Presidents” (n.d.), the comprehensive record of presidential communications. For example, once we found out that Obama visited Egypt on June 4, 2009, to deliver his address “A New Beginning” at Cairo University, we could easily identify the corresponding transcript. Additional remarks delivered during his stay in the country were retrieved as well (in that case, an interview with foreign journalists and remarks prior to a meeting with then President Mubarak). The procedure was repeated for every foreign visit undertaken by each of the 13 modern presidents.

Notably, this sampling approach excluded some speech acts that are known to include formal apologies, such as Congressional statements (e.g., the apology for the overthrow of the Hawaiian kingdom) or presidential speeches given to groups of people within the nation's borders (e.g., apologies for Japanese American internment during World War II). Also excluded were informal remarks during the president's vacations, brief remarks made on arrival at or departure from foreign airports, and informal exchanges with reporters abroad. The final corpus thus consisted of the following message types:

• First, perhaps one of the most significant speech occasions, is a major address given to a foreign legislative body (assemblies or parliaments), which accounted for 5.1% of all messages considered in this analysis (n = 111). Examples included Carter's address before the Israeli Knesset in March 1979 and George W. Bush's address before the Japanese Diet in February 2002.
• Second, presidential remarks during state dinners are often given in honor of the host, usually the head of state or executive leader. These were once famous speech occasions (particularly during the Cold War), but the significance and frequency with which presidents communicate in such forms have decreased markedly. However, due to generally tight time schedules these are sometimes the only major remarks a president delivers while visiting a foreign nation. For example, while the early toasts made by Roosevelt were often only brief statements relying on rather typical rhetoric for such occasions, the toasts offered by Carter were lengthy speeches dedicated to foreign leaders, and therefore worthy of exploration. Across all presidents, such speeches accounted for 13.7% (n = 299). Examples included Nixon's remarks at a state dinner hosted by the shah of Iran, Mohammad Reza Pahlavi, in Tehran in May 1972, and Reagan's remarks at a dinner hosted by Queen Elizabeth II at Windsor Castle during his visit to the United Kingdom in June 1982.
• Third, no less significant, but more numerous are presidential speeches and remarks made to foreign citizens in public places (e.g., city halls, universities, market squares). Often, those follow meetings with the head of state and are intended to address a larger audience more directly. In our study, those accounted for 28.8% of all presidential communications abroad (n = 629). Examples included Kennedy's 1963 and Reagan's 1987 speeches in front of the Brandenburg Gate in Berlin, Germany, and Obama's address at Cairo University in Egypt in June 2009.
• Fourth, major documents, such as (joint) statements and declarations released during or after a president's foreign visit, were considered for analysis. Previous research suggested that apologies (and other expressions of regret) are sometimes offered intentionally as brief, written statements only – in order for presidents to bypass such comments in their speeches (see Nobles, 2008). Written messages account for 14.1% of all messages (n = 309). Examples included Clinton's joint communiqué on the Summit for Peace and Prosperity, which was released in March 1998 in Uganda, and G. W. Bush's statement calling for a new strategic partnership with Pakistan in March 2006.
• Fifth, foreign press conferences, town hall meetings, and interviews with foreign journalists were subject to analysis. These contexts differ from others in that speakers are forced to be less deliberate and strategic in the words they choose (by virtue of the fact that they take place in an interactive, two-way communication forum). For example, in their conception of major presidential addresses, Coe and Neumann (2011, p. 733) delineate “message control” as a major criterion by referring to situations where the president has full control over the message; that is, “the words that a president speaks are a matter of choice, usually carefully planned and rehearsed well in advance of the speaking event.” This loss of “message control” in news conferences may lead presidents to employ different rhetorical tactics from what they would use in the more deliberate speech acts listed above. Press conferences and interviews with the president accounted for 26.4% of all presidential communications abroad (n = 577). Examples ranged from Roosevelt's famous World War II press conferences in Casablanca (January 1943) and Yalta (February 1945) to the above-mentioned news conference of Obama in Strasbourg, France, in April 2009.
• Sixth, although not given on foreign soil, speeches to the United Nations General Assembly or the Security Council (in New York City) have become a regular (i.e., annual) speech tradition for the US president since Reagan. Given that in those speech situations the president also responds to concerns of the international community in general and of foreign leaders representing their nations in particular, these speeches were considered worthy of analysis. UN speeches accounted for 2.9% of all presidential messages considered in this study (n = 64). Examples included Truman's first address to the UN General Assembly during its opening session in October 1946, and G. W. Bush's “An Appeal to Muslims” speech in September 2006.
• Lastly, speeches given on foreign soil also encompass remarks directed to Americans living abroad (e.g., embassy staff, troops, US communities) and to the American audience at home (e.g., radio addresses to the nation). Those messages accounted for 9% of all presidential communications abroad (n = 197). Examples included Ford's remarks to American troops stationed in South Korea in November 1974, Clinton's remarks to military personnel at Ramstein Air Base, Germany, in May 1999, and G. W. Bush's speech to US troops in Baghdad, Iraq, at a Thanks-giving dinner in 2003.

Administering CCA

At the outset of the study, all the necessary transcripts were compiled in a raw text file (e.g., a Word document) and organized around simple descriptive variables that depended on the goal of the analysis (in our case, this included date, president, location/country, and message type). This information – called the “control sequence” – was entered manually for every transcript as a string of external variables (separated by dashes) beginning with a four-digit, ascending sequence number that each text unit received at the outset (the length of the sequence number depends, of course, on the number of texts the researcher considers for analysis). The control sequence was placed before each transcript so that the raw text in the entire file could be easily separated into text units. This process is time consuming, but manually entering these external variables allows researchers the flexibility of more varied analysis (e.g., by comparing Democrats and Republicans, distinguishing between written and spoken communication, grouping countries into regions). Additionally, certain transcripts required some pre-editing to remove all substantive nonpresidential content (such as lengthy responses of another country's leader in a joint press conference).

Once the raw text file was compiled, the content was subject to computer-assisted content analysis using the software TextQuest,² which allows processing of large quantities of text. This software requires that the researcher develop another file in the form of a dictionary of search patterns that the computer will retrieve from the entire text file once identified. For the present study, relevant semantic units included in our dictionary file (in the following referred to as lexemes) were as follows: amends, apology, atonement, blame, blunder, compunction, confession, contrition, deficiency, deplore, error, excuse, expiation, fail, fault, flaw, forgiveness, guilt, inadvertence, injustice, lapse, misconduct, misdeed, misdemeanor, mistake, pardon, penance, penitence, plea, reconciliation, regret, remorse, repentance, ruefulness, shame, shortcoming, sin, sorrow, transgression, and wrongdoing.

Words appear in various inflected forms (as noun, verb, adverb, adjective, etc.). We therefore focused on a word's morphological stem (or lemma) to allow for its linguistic derivatives. For instance, instead of entering the term “apology” into our dictionary file, we used the root morpheme “apolog-,” which captures the noun “apolog g,” its plural form “apologies,” its adjectival form “apologetic,” the verb “apologize” (and given its regular verb form, its past and future tense too), its adverbial form “apologetically,” and related nouns such as “apologist(s)” and “apologia.” In doing so, the analysis can be refined and focused on the unit of meaning (semantic content) rather than a unit of speech (predefined words).

Additionally, we consulted the online Merriam-Webster dictionary to retrieve a term's definition and perform a search for synonyms and related words. For example, Merriam-Webster provides two definitions of an apology: “an explanation that frees one from fault or blame” and “an admission of error or discourtesy accompanied by an expression of regret.” Hence, the terms “fault,” “blame,” “error,” and “regret” subsequently became subject to the same thesaurus search and their stems were entered into the dictionary file. Other suggested surrogate and related terms for “apology” were “plea (of guilty),” “atonement,” and “confession,” among others. Admittedly, the thesaurus search provides many more synonyms and related words, but researchers who aim to conduct a computer-assisted content analysis of a significant corpus of speeches are forced to make certain practical decisions. For example, the words “reason” and “justification” were suggested surrogate terms for “apology,” while “responsibility” and “accountability” were offered as a substitute for “blame” by Merriam-Webster. Entering the stems of those terms would make the analysis less precise because they capture a large number of rhetorical moments that are not related to apologetic rhetoric. To be clear, then, these terms are not exhaustive, but rather emerged as the most common ways public officials acknowledge their mistakes and verbalize their apologies, based on their presence in the extant literature, a thesaurus search for relevant synonyms, and several readings of randomly selected presidential speeches prior to conducting the CCA.

The main purpose of developing such dictionaries is twofold. First, in analyzing a set of transcripts it provides some initial trend data based on the usage of specific terms (e.g., to assess the overall use of the lexeme “apology” in presidential rhetoric) or frequency of the variables entered into the string (e.g., in our case the frequency with which presidents have traveled to certain countries). Second, the method provides the opportunity to narrow the search down to relevant content – those messages that include some form of apologetic rhetoric. In other words, a certain number of speeches did not qualify based on these exclusion criteria, and thus did not require further analysis. TextQuest makes it possible to locate such key terms within a given text or a set of texts. In addition, it provides contextual information (all surrounding words and sentences called “concordances”) that is clustered around the lexeme. By using such a filter function, a printable output file is created and the manual coding procedure is made manageable.

Administering the Manual Coding Procedure

All remaining speeches were subject to manual content analysis, with the unit of analysis being a single mention of any of the identified lexemes. The coding scheme for this analysis was in part based on previous research and conceptualizations of public apologies (Abadi, 1990; Blatz et al., 2009; Kampf, 2009a) and newly added measures. As mentioned in the preceding section, each text received a number at the outset. Units that qualified based on the filter search via keywords received an additional number (in ascending order). For example, a text may include no such term at all, while another text includes several of the lexemes representative of apologetic rhetoric. Once those units were identified, one essential criterion for inclusion was to code for self-reference, that is, while using the lexeme, presidents refer either to themselves or the United States (e.g., “my fault” or “mistakes we've made in the past as a country”) or to another entity (e.g., “guilty nations”). This aspect of the process requires human, rather than computer, judgment. If the latter was the case, the instance did not qualify as self-referential apologetic rhetoric. However, if the former was the case, the following additional items guided the rest of the content analysis.

• Creative use of tense and conditional form. Sometimes speakers make use of so-called “non-performative” apologies (Kampf, 2009a). The purpose of this category was to differentiate between actual statements (the act of confession occurs as part of the speech act, or a reference to a past confession is made), and hypothetical situations (including conditional and future tense). An example of the former is “The uncertainty might come from our failure to control and to regulate the arms competition” (Carter, 1979) whereas an example of the latter is “If we, the U.S., have been at fault we stand ready to be corrected” (Eisenhower, 1958).
• Willingness to engage in self-criticism. The purpose of this code was to distinguish between instances in which the speaker makes some criticism of the nation (by publicly admitting mistakes or offering an explicit apology) and instances in which the speaker refuses to do so (an explicit refusal to apologize or to acknowledge mistakes). This does not refer to the responsibility taken for the act. Examples are “We shall not be so pious or so hypocritical as to pretend that we have not made mistakes” (Nixon, 1970) with regard to Vietnam, or “Today we rejoice also in another event for which we need not apologize” (Roosevelt, 1943).
• Use of the lexeme “apology.” The performative verb “to apologize” is considered the strongest of all forms of apologetic rhetoric and therefore deserves emphasis. This code was used to distinguish between uses of “apologize” and any other form, for example, “I am sorry” or “I regret.” Examples are “I am sorry, I omitted that from my comments” (G. H. W. Bush, 1991) and “To the extent I've caused this, I apologize” (G. W. Bush, 2007).
• Specifying the nature of the act. The purpose of this code was to distinguish between generic, superficial remarks and specific, clear-cut remarks. Kampf (2009a) proposed this rhetorical strategy to blur the nature of the offense and minimize responsibility. An example of the former is “we make mistakes and we're not perfect” (Clinton, 1994), whereas an illustration of the latter form is “It's a shame in America that children getting out of high school cannot read sufficiently” (G. W. Bush, 2002).
• Specifying the identity of the wrongdoer. The purpose of this category was to assess whether the offender is mentioned in the discourse or not. The rhetor may use the first-person plural form (“we”) or, to concretize responsibility, may mention the wrongdoer (e.g., “we Americans”), or refer to himself as the principal offender (“I”). Examples from the discourse include “We failed to bring Saddam Hussein back into the family of nations” (G. H. W. Bush, 1992). Examples emphasizing the nation include “We in the United States have made many mistakes in our relations with Latin America” (Kennedy, 1961). Examples highlighting the speaker include “my failure to avoid mentioning several immediate problems that are troubling the U.S.” (Eisenhower, 1960).

A total of 2,362 potential instances of apologetic rhetoric were identified by Text-Quest. However, this number also included remarks uttered by reporters or hosts (mostly during interviews or press conferences). Recall that substantive remarks by other foreign leaders (which are typically part of the speech files released by the American Presidency Project) were already excluded before conducting the CCA. However, the typical question and answer form of interviews and news conferences made it somewhat difficult to exclude all nonpresidential content at the outset of the study. Moreover, important contextual information may have been lost in so doing. To clean the data from those instances, an analysis of the TextQuest concordances – each consisting of 300 characters in our case, that is, 150 characters occurring before and after the particular lexeme – made it possible to identify those apologetic lexemes that were not spoken by a president (we considered 300 characters to be sufficient to determine the speaker and major context; others might choose a smaller or larger range depending on their study goals and design). This procedure allowed 237 such instances (reporter remarks, interviewer questions) to be identified and excluded from further analysis. Thus, a total of 2,125 apologetic lexemes ultimately served as the basis for this analysis.

One person completed the content analysis. As a check of reliability, a second person coded a randomly selected subsample consisting of approximately 8% of the lexemes. Overall, chance-corrected rates of agreement (that is, taking into account the amount of agreement between multiple coders that could occur based on chance as opposed to deliberate agreement) ranged from 0.72 to 1.00 (see Cohen, 1960). Specifically, chance-corrected agreements were as follows: self-reference 0.82, tense 0.83, willingness 0.92, specifying the act 0.73, specifying the wrongdoer 0.72, and use of the apology lexeme 1.00.

Analysis

Research Question 1

RQ1 explores the extent to which modern presidents have used apologetic rhetoric in their public messages to foreign audiences. Apologetic rhetoric is indicated by the use of respective lexemes that reveal the act not only of apologizing, but also of regretting own behaviors and other events, admitting past mistakes, or asking for reconciliation. Figure 12.1 (based on n = 2,125) presents presidents' overall use of these lexemes, controlled for total word count.

Figure 12.1 Apologetic rhetoric in presidential messages abroad, per 10,000 words (n = 2,125)

The findings suggest that modern presidents have remained fairly consistent in their overall use of apologetic rhetoric. They averaged 5.6 apologetic lexemes per 10,000 words, with Johnson (8.9), Eisenhower (7.6), and Roosevelt (6.8) scoring above that average, and Ford (3.4), Nixon (3.6), and Carter (5.1) scoring below average. While the biggest changes occurred in the 1960s and 1970s, the use of apologetic rhetoric has remained fairly constant since Reagan.

Although this initial analysis provides some preliminary insight into presidents' use of apologetic rhetoric in general, narrowing the parameters of this rhetoric provides a more precise look at presidential behavior. A major distinction can be made by separating self-referential remarks (e.g., “our mistakes”) from other-referential utterances (e.g., “their fault”). In roughly 40% of all cases (n = 843) presidents referred to either themselves or the nation, as opposed to 60% where presidents applied any of the given lexemes to either refer to the injustices or mistakes of others, or to state general phrases without further specification of the source (n = 1,282). Specific examples of other-referential units are “The Taliban and Saddam holdouts are desperately trying to stop our progress; they will fail” (G. H. W. Bush, 1990) and “It's easy to point fingers and to pin the blame of these problems on others” (Obama, 2009).

Note that those instances were then separated from “minor” apologies and hypothetical statements. For example, presidents often apologize for arriving late at a scheduled speech occasion or departing early from the visited country, or presidents say “sorry” or “excuse me” in press conferences when they do not understand the question. While this may be indicative of a speaker's politeness, these instances are not meaningful in the context of apologetic rhetoric; 309 such minor remarks (36.7% of all self-referential instances) were identified and excluded from further analysis. Expressions of regret and sorrow in case of natural disasters (e.g., a country's victims struck by a hurricane or an earthquake) or the death of a foreign leader (n = 84, 10%) were also excluded. Likewise, presidents apply another rhetorical strategy: the use of the conditional tense to describe potential, not actual, failures. Examples include “if we fail to act in the face of danger, the people of Iraq will continue to live in brutal submission” (G. W. Bush, 2002) or “our children and grandchildren will never forgive us if we allow new and unstable nuclear standoffs to develop around the world” (G. H. W. Bush, 1992). While these instances describe possible failures the United States could commit, they are nonetheless hypothetical situations and therefore not truly representative of apologetic rhetoric; 115 such phrases were excluded from the final analysis (13.6% of all self-referential remarks). Taken together, the result of these necessary exclusions is that a total of 333 units (only 15.7% of all mentions originally identified via CCA) exist that are truly representative of modern presidential apologetic rhetoric. These units serve as the basis for further analysis.³

The trend (see Figure 12.2) – looked at in this more nuanced fashion – reveals a decrease in apologetic rhetoric throughout the modern presidency. Two of the three early modern presidents have been most likely to apply apologetic rhetoric. Roosevelt averages 2.6 mentions per 10,000 words, followed by Eisenhower (2.0) and Reagan (1.4). A first change occurred in the mid-1960s, determined by four consecutive presidents who were among the least likely to use such rhetoric: Johnson averages 0.6, Nixon 0.4, and Ford 0.2 mentions per 10,000 words. A second change took place with Reagan entering office. Apologetic rhetoric was somewhat revived and has since remained fairly consistent. Although both Bush presidencies account for slight downward trends, Obama appears to be more similar to Reagan and Clinton when it comes to the extent of using apologetic rhetoric abroad, averaging 1.1 mentions per 10,000 words. However, the more recent presidents clearly do not reach the levels of early modern presidents. Additional differences exist with regard to party affiliation. While Democrats average one apologetic instance per 10,000 words across all messages (SD = 0.66⁴), Republicans average 0.76 instances per 10,000 words (SD= 0.69), thus suggesting that Democrats have been more likely to apply apologetic rhetoric. It is impossible to know for certain what caused the downward trend in the 1960s. One possibility, though, is that the general lack of media coverage and the absence of journalists abroad might have caused early modern presidents to employ such language more easily than succeeding presidents. With a general increase in media coverage, presidents may have become more cautious about their message content. This is consistent with the Cold War context: any self-critical statement made by the president could potentially be interpreted as a sign of weakness by the ideological opponent.

Figure 12.2 Self-referential apologetic rhetoric in presidential messages abroad, per 10,000 words (n = 335)

To answer RQ1, then, it can be said that apologetic rhetoric in presidential messages abroad did not increase when controlled for president's total word count. In fact, it has remained fairly stable since Reagan. However, as demonstrated, much of this apologetic rhetoric is barely apologetic by the standards set forth above. Thus an analysis using a more rigorous focus on apologetic rhetoric revealed that modern presidents have decreasingly used such language when referring to the nation. Democratic presidents have been more likely than Republican presidents to admit the nation's past mistakes and to express sorrow about them. Individual presidential differences become manifest when contrasting Roosevelt and Eisenhower, the presidents most likely to use apologetic rhetoric, with Nixon and Ford, who were least likely to do so.

Research Question 2

RQ2 focuses on specific terms in apologetic rhetoric that presidents have used in their public messages delivered to foreign audiences. Based on the mixed approach to content analysis utilized in this study, we found that the most frequently occurring lexeme turned out to be “reconciliation” (n = 107, or 32% of all units considered herein), followed by “mistake” (n = 66, 19.8%), “fail” (n = 43, 12.9%), “regret” (n = 19, 5.7%), and “error” (n = 12, 3.6%). The strongest indicator of apologetic rhetoric, the term “apology” including its derivatives, was mentioned only 13 times in all 2,186 public messages. Religiously connotated terms such as “forgiveness” (n= 3), “confession” (n = 3), “sin(s)” (n = 7) were barely used by presidents while others (“atonement,” “remorse”) were not used at all when referring to the United States. Table 12.1 summarizes these frequencies.

Table 12.1 Frequencies of self-referential apologetic lexemes (n = 335)

To answer RQ2, then, it can be noted that presidents have mostly framed their apologetic statements by referring to mistakes and reconciliation. Both account for more than half of all lexemes included in this analysis. One might argue that by using the word “mistake” (instead of a stronger term such as “injustice” or a clear acknowledgment of guilt or shame) presidents attempt to downplay a wrongful act's significance. To further illustrate, consider two examples from the discourse: “The errors we made, or at least the political decisions we made, caused us problems” (Clinton, 1997) and “We are not saints, we know we make mistakes” (Eisenhower, 1960). With regard to “reconciliation,” one might argue that instead of addressing the nation's past, presidents prefer to draw attention to the present and the future. Instead of focusing on the roots of a conflict, presidents have chosen to highlight reconciliation as the healing process or the outcome of that process. An additional analysis shows that it is more common for presidents to position the United States as a mediator or an active promoter of reconciliation between others (“America is using its influence to foster peace and reconciliation in the Holy Land,” G. W. Bush, 2008) and less common to present the United States as a nation that needs to be reconciled with other nations (“Americans and Germans came together, reconciled, and united for freedom,” Reagan, 1985). The almost complete absence of the term “apology,” however, casts some doubt on the idea that the US presidency has entered an “age of apology.”

Research Question 3

RQ3 asked about the differences of apologetic rhetoric regarding the message format. Messengers can have varying levels of control over the message. While news conferences are considered low in message control, major addresses before a foreign parliament, a foreign public, or the UN General Assembly are typically crafted messages, scheduled ahead of time and therefore prepared with diligence, often with the support of professional speechwriters. The overall trend suggests that, for the most part, presidents are more likely to apply some form of apologetic rhetoric when they have higher message control (i.e., speeches to political entities or citizens). Figure 12.3 illustrates that trend.

Figure 12.3 Self-referential apologetic rhetoric and message type (low versus high message control) (n = 333)

In particular, in their news conferences abroad presidents averaged 0.7 apologetic mentions (per 10,000 words), as opposed to 1.0 such mentions in all other spoken and written remarks. Presidents highest in apologetic rhetoric in news conferences were Roosevelt (2.6), Eisenhower (1.4), Clinton (1.0), and Obama (0.9), whereas Truman, Kennedy, Johnson, Nixon, and Ford were lowest, with apologetic rhetoric being virtually absent from the discourse.⁵ Beginning with Carter, news conferences abroad became an essential part of modern presidential rhetoric. When presidents had more control over the message they were also more likely to use such rhetoric. However, the overall trend for those messages mainly follows the trend for messages low in control. Once again, two early modern presidents, Roosevelt (2.5 times per 10,000 words) and Eisenhower (2.0), and Reagan (1.6) were most likely to apply apologetic rhetoric in crafted speeches, as opposed to Ford (0.2), Nixon (0.4), and Johnson (0.6) who were the presidents least likely to do so. One might argue that this increase is due to reporters' behavior becoming increasingly “aggressive” in broaching the issue of past wrongdoings (e.g., a standard question asked by reporters is “What is your greatest regret of your presidency?”), thus urging the president to respond carefully. In an effort to avoid such questions, presidents might have adopted a rhetorical style and become more apologetic in their crafted speeches to pre-empt those attacks.

Research Question 4

The final research question focuses on the units that were found to be self-referential based on the discussion provided under RQ1 (n = 333). A first rhetorical strategy that has been discussed in the extant literature is the willingness to admit mistakes, to express sorrow and regret at past injustices, to apologize, or – more generally – to engage in self-criticism. In the following, we will refer to this as self-critical instances of apologetic rhetoric, thus excluding all mentions of reconciliation, refusals to apologize, and other defensive acts. For all further analyses we will also draw several comparisons by dividing the modern presidency along party lines (Democrats and Republicans) as well as in three broad phases (pre-Cold War, Cold War, and post-Cold War). ⁶

Generally, it can be assumed that presidents who apply a more reconciliatory rhetoric demonstrate more willingness to admit past wrongs, whereas presidents who apply a more defensive rhetoric are less likely to do so. Presidents have admitted mistakes and acknowledged flaws to varying degrees: seven times in the early stages of the modern presidency (the pre-Cold War era), 68 times during Cold War, and 115 times in the post-Cold War phase. From a relative perspective, however, postCold War presidents have been slightly less willing than Cold War presidents to confess: during the Cold War presidents averaged 0.51 signals of willingness (per 10,000 words) while presidents governing in the 1990s and 2000s averaged 0.48 such signals.⁷

Figure 12.4 Self-referential (n = 335) and self-critical instances (n = 191) of apologetic rhetoric in presidential messages abroad

The distinction between self-referential units (those rhetorical moments where presidents broadly refer to themselves and the nation as opposed to others) and – being a subsection of it – self-critical units (those where presidents publicly made some form of criticism of the nation, for example, by admitting mistakes) is illustrated in Figure 12.4. The dashed gray line represents all self-referential units (n = 333, congruent with the line in Figure 12.2; note that this also includes few cases where presidents refuse to apologize or deny that mistakes happened), while the solid line marks all self-critical units (n = 190; both per 10,000 words). Additionally, given the prevalence of the reconciliation lexeme among all self-referential units, a third (dashed black) line is shown that refers to all self-referential units excluding all mentions of reconciliation (n = 226). The previously observed party differences remain when focusing on just the self-critical instances of apologetic rhetoric (n= 190). In particular, Democrats averaged 0.6 public confessions per 10,000 words (SD = 0.57) whereas Republicans averaged 0.4 public confessions per 10,000 words (SD = 0.53).

Recall that one of the major conclusions of RQ1 was a downward trend that began in the 1950s and reached its low point in the mid-1970s, followed by an upward trend in the 1980s that has since remained fairly consistent. This trend is represented by the gray dashed line in Figure 12.4. It is apparent that the increased usage of the term “reconciliation” (especially by Reagan, Clinton, Johnson, and Eisenhower) is responsible for some of the variation, as indicated by the gaps between the two dashed lines in Figure 12.4. Excluding reconciliation from this trend thus results in a more flattened curve in the recent past. This in turn strongly follows the trend line resembling only the self-critical instances. When focusing on those specific instances the tendency for most modern presidents to admit mistakes is not entirely different from the observation made under RQ1. However, the differences between post-Cold War presidents are now more pronounced, particularly when comparing Obama to his predecessors. While G. H. W. Bush, Clinton, and G. W. Bush rather resemble Reagan, there was an upsurge in 2009 when it came to presidential criticism of the United States (its previous administration, in particular).⁸This in turn is more reminiscent of the early days of the modern presidency.

A second rhetorical strategy is to focus on the type of wrongdoing, particularly the degree of specificity with which modern presidents have admitted past mistakes and injustices. In the past, presidents have acknowledged specific wrongdoings and historical injustices, ranging from slavery and segregation to support of corrupt regimes and military dictatorships, and policy failures regarding energy, immigration, or foreign affairs. In so doing, presidents may be more specific (e.g., “it's one of my great regrets that it did take so long in Bosnia” (Clinton, 1999)) or less specific (“whatever mistakes we have made” (Obama, 2009)) in acknowledging wrongdoings. In general, modern presidents have been willing to admit specific mistakes in nearly two-thirds (n = 120) of the cases in which mistakes were admitted; one-third relates to rather general phrases and superficial remarks (n = 70). In particular, presidents more likely to specify their own or the nation's mistakes were Kennedy and Johnson (both 100%), G. W. Bush (80.8%), and G. H. W. Bush (73.3% of all mistakes acknowledged were specific). Presidents less likely to specify their own or the nation's mistakes were Eisenhower (68.8% of all mistakes acknowledged were rather general confessions), Reagan (50.0%), and Obama (50.0%). On average, presidents governing in the post-Cold War era were slightly more likely to specify their mistakes (67.8%, n = 78) than their predecessors (55.9%, n = 38) in cases where those were admitted. However, when controlled for number of words, those differences disappear: Whereas Cold War presidents admitted 0.29 specific mistakes on average (per 10,000 words), post-Cold War presidents admitted 0.32 specific mistakes.⁹ Interestingly, although post-Cold War presidents have been slightly less willing than Cold War presidents to confess, they have been slightly more specific than Cold War presidents in acknowledging mistakes, thus indicating a more reconciliatory, less defensive rhetoric prevalent in the 1990s and 2000s.

A third rhetorical strategy is to focus on the identity of the wrongdoer, particularly the degree of specificity with which modern presidents reveal their own or their nation's role in committing mistakes. Previous research suggested that omitting, concealing, or de-emphasizing certain information (such as the wrongdoer's identity) may be one form of reducing responsibility (Fairclough, 2000; Kampf, 2009a). To do this, presidents might use the personal pronoun “we” without explicitly linking the transgression to the United States. For example, presidents may refer to a failure committed by a broader alliance of nations including the United States (e.g., G. W. Bush's “Coalition of the Willing”; the UN in its entirety; or simply the West), as in “the failure of the Allied nations to make more progress in their common search for a lasting peace” (Truman, 1946) or “inescapable human errors in our long record” (Eisenhower, 1956). Post-Cold War presidents were less likely (56.5%, n = 65) to use such general forms than their predecessors (72.1%, n = 49) in cases where mistakes were acknowledged during that period. In fact, post-Cold War presidents were more likely to engage in self-criticism emphasizing the United States (26.1%, n = 30) than their predecessors (17.6%, n = 12), as illustrated, for example, in “There's plenty of blame to go around for what has happened, and the United States certainly shares its blame for what has happened” (Obama, 2009) or “perhaps the worst sin America ever committed about Africa was the sin of neglect and ignorance” (Clinton, 1998). Similarly, Clinton, G. W. Bush and Obama were also more likely than their predecessors to admit personal mistakes (16.5%, n = 19 versus 10.3%, n = 7), such as “I've already said I made a bad mistake, it was indefensible” (Clinton, 1998). Compared with Cold War presidents, the four presidents governing in the 1990s and 2000s have been slightly less likely to signal willingness to admit past mistakes and express regret, but slightly more likely to engage in self-criticism highlighting themselves or the United States as the principal wrongdoer and clarifying the nature of the transgression.

Nonetheless, modern presidents have primarily avoided highlighting the United States as a principal wrongdoer, and instead made generalizations about the offending parties by not outlining and further specifying the role of the United States in previous conflicts and wrongful acts. Presidents seem to reduce responsibility and “democratize” blame by applying certain rhetorical strategies that are used to blur the identity of the offender, or distort the nature and dimension of the act itself.

Conclusion: Advantages and Disadvantages of a Mixed Approach

This chapter aimed to illustrate the process of a mixed approach to content analysis through a study of modern American presidents' apologetic rhetoric. The study presented here has, we hope, demonstrated the many advantages of this approach. We briefly reflect on these advantages below, and also highlight some of the disadvantages of adopting this method.

The mixed approach shares one of CCA's greatest strengths: the ability to process large quantities of data. Rather than having to rely on a sample of relevant moments, the mixed approach allows scholars to speak with authority about over-time trends in the census of relevant text. In the case of presidential apologies abroad, for instance, we can be more certain of our generalizations than we would be if we were manually coding a smaller sample of texts, or if we had limited ourselves to fewer lexemes. The use of the automated procedures discussed above made such certainty possible. Just as clear, however, is that had we only relied on these automated processes, our ability to fully understand presidents' apologetic rhetoric would have been severely constrained. Additional manual coding provided a much more nuanced view of the data. In this case, at least, neither of the two methods would have been sufficient by itself: while manual coding of more than 2,000 documents would not have been manageable given pragmatic considerations, relying on CCA alone would have generated very different findings – and provided even misleading trends. For example, recall that our data showed a considerable number of “minor apologies” and hypothetical statements. These common speech elements are not truly representative of apologetic rhetoric, but would nonetheless have been captured by the computer. Without manual coding to eliminate such instances, the data would have been far less accurate and useful.

The possibility of narrowing the search results down to a limited number of relevant texts (while excluding irrelevant entities) also facilitates the search for qualitative evidence that can help provide greater meaning to quantitative trends. This can be useful to substantiate findings when interpreting or explaining the results of a study, for example, by facilitating the process of finding illustrative quotes in the texts. Simply put, instead of wading through a sea of text, the content analysis software generates a compressed file of relevant utterances (using concordances) from which the researcher can easily choose. This feature also allows the researcher to double check data for negated statements and ambiguous remarks – elements that can mask crucial bits of data if not identified. The mixed approach thus allows rhetorical trends to be interpreted more accurately than they might otherwise be, and is therefore especially useful for many diachronic research endeavors.

Researchers adopting this approach should also be aware of a few disadvantages. Central among these is that a mixed approach is more time-consuming than traditional CCA, given that it adds the manual component. But it is also worth knowing that even the automated components are not possible without considerable time spent preparing the files for the computer. For instance, creating a text file with all sample documents and “data cleaning” may sometimes be necessary to make the actual analysis more manageable and meaningful. Setting up a dictionary of key terms can be a time-consuming endeavor too, but it is essential to make the initial term search as exhaustive as possible. Some studies (e.g., Coe, 2011) generate the dictionary of search terms directly from a complete list of every distinct word used in the corpus of texts. This approach is the surest means of making the dictionaries valid but is also extremely time-consuming.

Additionally, there will always be in the mixed approach some degree to which the manual coding is limited by lack of context. For example, a traditional content analysis of presidential speeches would have human coders reading entire speeches and coding for certain categories as they appear in the texts. In contrast, the mixed approach isolates key terms and has human coders code those terms within some limited amount of contextual material (e.g., 150 characters on either side of the term, our chosen amount of context in the present study). Coding decisions made by humans who have read an entire speech have the potential to be more sophisticated than those made by humans who have read only a small passage. In nearly all cases, however, this limitation of the mixed approach can be compensated for by carefully selecting the categories that will be central to the analysis. It is simple, for example, to understand from just a small amount of context whether a president is apologizing for his own actions or for those of others. Thus the mixed approach is effective in this case. If the coding categories require judgments that are quite complicated, however, the mixed approach is probably less useful than traditional manual coding.

Obviously, our focus has been on just one narrow context: modern presidents' apologetic rhetoric. But the mixed approach has broad applicability to various other contexts. For example, future studies might focus on the corporate level where apologetic rhetoric is used as a means of image repair, especially when relations between a company and its customers have been strained due to an environmental, safety, or health crisis. Numerous case studies have been conducted (e.g., Benoit & Lindsey, 1987; Hearit, 1994) but a quantitative assessment has been largely absent from the literature. A mixed approach could thus be used to examine companies' crisis response messages over time or by sector. Similarly, given that political discourse happens in interaction between political elites, the public, and the media, the mixed approach can also be used to incorporate citizen blogs or news media content. In so doing, one can analyze how particular moments of presidential apologetic rhetoric are perceived and interpreted by citizens or reported and commented on by the (international) news media. Such studies, for example, may focus on a single presidency but allow for a number of media responses (to presidents' apologetic rhetoric) to be included. Scholars can thus gain additional insight into which media outlets turn a particular moment of apologetic rhetoric (as the forms identified herein) into an “apology” per se, which would perhaps explain much of the recent discussion about the “age of apology” (Gibney et al., 2008; Kampf, 2009b).

More broadly, the methodology outlined in this chapter can be applied to analyze discourse beyond topics related to political communication and can be conducted in every language (assuming the content analysis software is able to process non-English language or non-roman letters). The mixed approach to content analysis employed and explained herein is suitable for scholars interested in processing large quantities of text (to examine broader rhetorical trends) yet do not want to refrain from taking a closer look at rhetorical nuances. Research might focus on such areas as environmental communication (e.g., framing of climate change in news coverage), health communication (e.g., examining arguments used in public health messages), or mass communication (e.g., portrayal of minorities and use of stereotypes in mass media). Whatever the context, the mixed approach should give scholars a useful tool for better understanding various forms of spoken and written communication.

NOTES

1 Texts of presidential speeches, from which all quotations and analyses are drawn, were retrieved from the American Presidency Project (University of California, Santa Barbara, n.d.).

2 For further information see www.textquest.de.

3 Although those other-referential units constitute apologetic rhetoric to a great extent, they were not part of further analysis. That presidents have been more likely to refer to the mistakes and faults of others is nonetheless interesting, but in order to give an accurate account of presidential apologetic rhetoric, only self-referential remarks are included in further analyses.

4 SD = standard deviation.

5 It should also be noted that those five presidents held very few press conferences abroad compared to all other modern presidents.

6 For this study, the pre-Cold War era was defined as Franklin Roosevelt's inauguration (March 4, 1933) through the end of World War II (specifically the Pacific War on September 2, 1945). That was followed by the Cold War era, which lasted until the dissolution of the Soviet Union on December 25, 1991. The post-Cold War era then continued to the present.

7 World War II presidents Roosevelt and Truman averaged 1.55 apologetic lexemes per 10,000 words, thus confirming previous trends suggested earlier.

8 However, caution is recommended in judging the rhetorical trend President Obama has set in his first 15 months in office. Presidents have generally traveled abroad more during their first year in office than in the years afterwards. In the case of Obama, for example, roughly half of the country visits made during his first term were in his first year in office (25 out of 51). Hence, it remains to be seen if the rhetorical signals Obama sends in his second term are compatible with his first year's rhetoric.

9 World War II presidents Roosevelt and Truman fall somewhere in between (57.1% of all mistakes acknowledged were specific, n = 4) and averaged 0.89 specific mistakes per 10,000 words.

REFERENCES

Abadi, A. (1990). The speech act of apology in political life. Journal of Pragmatics, 14, 467–471.

Barkan, E. (2000). The guilt of nations: Restitution and negotiating historical injustices. Stanford, CA: Stanford University Press.

Benoit, W. (1995). Accounts, excuses, and apologies: A theory of image restoration strategies. Albany, NY: State University of New York Press.

Benoit, W., & Lindsey, J. (1987). Argument strategies: Antidote to Tylenol's poisoned image. Journal of the American Forensic Association, 23, 136–146.

Blatz, C. W., Schumann, K., & Ross, M. (2009). Government apologies for historical injustices. Political Psychology, 30(2), 219–241.

Brookhiser, R. (1998). America the apologetic. American Heritage, 49(8), 42.

Brooks, R. L. (1999). The age of apology. In R. L. Brooks (Ed.), When sorry isn't enough: The controversy over apologies and reparations for human injustice (pp. 3–11). New York, NY: New York University Press.

Coe, K. (2011). George W. Bush, television news, and rationales for the Iraq War. Journal of Broadcasting & Electronic Media, 55, 307–324.

Coe, K., & Neumann, R. (2011). The major addresses of modern presidents: Parameters of a data set. Presidential Studies Quarterly, 41, 727–751.

Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37–46.

Cunningham, M. (1999). Saying sorry: The politics of apology. Political Quarterly, 70(3), 285–293.

Digeser, P. E. (2001). Political forgiveness. Ithaca, NY: Cornell University Press.

Edwards, J. A. (2008a). Navigating the post-Cold War world: President Clinton's foreign policy rhetoric. Lanham, MD: Rowman & Littlefield.

Edwards, J. A. (2008b). The mission of healing: Kofi Annan's failed apology. Atlantic Journal of Communication, 16, 88–104.

Fairclough, N. (2000). New Labour, new language? London, UK: Routledge.

Gibney, M., Howard-Hassman, R. E., Coicaud, J. M., & Steiner, N. (2008). The age of apology. Facing up to the past. Philadelphia, PA: University of Pennsylvania Press.

Govier, T. (2002). Forgiveness and revenge. London, UK: Routledge.

Greenstein, F. I. (1998). Leadership in the modern presidency. Cambridge, MA: Harvard University Press.

Greenstein, F. I. (2004). The presidential difference: Leadership style from FDR to George W. Bush (2nd ed.). Princeton, NJ: Princeton University Press.

Harris, S., Grainger, K., & Mullany, L. (2006). The pragmatics of political apologies. Discourse & Society, 17(6), 715–737.

Harter, L. M., Stephens, R. J., & Japp, P. M. (2000). President Clinton's apology for the Tuskegee syphilis experiment: A narrative of remembrance, redefinition, and reconciliation. Howard Journal of Communications, 11, 19–34.

Hearit, K. M. (1994). Apologies and public relations crises at Chrysler, Toshiba, and Volvo. Public Relations Review, 20(2), 113–125.

Hearit, K. M. (2005). Crisis management by apology. Mahwah, NJ: Lawrence Erlbaum.

Jamfa, L. (2008). Germany faces colonial history in Namibia: A very ambiguous “I am sorry.” In M. Gibney, R. E. Howard-Hassman, J. M. Coicaud, & N. Steiner (Eds.), The age of apology. Facing up to the past (pp. 202–215). Philadelphia, PA: University of Pennsylvania Press.

Kampf, Z. (2009a). Public (non-)apologies: The discourse of minimizing responsibility. Journal of Pragmatics, 41, 2257–2270.

Kampf, Z. (2009b). The age of apology: Evidence from the Israeli public discourse. Social Semiotics, 19(3), 257–273.

Kellerman, B. (2006). When should a leader apologize and when not? Harvard Business Review, 84(4), 72–81.

Kirchick, J. (2009, April 28). Squanderer in chief. Los Angeles Times, p. 23A.

Lazare, A. (1995). Go ahead, say you're sorry: Psychological aspects of apologies. Psychology Today, 28(1), 40–46.

Lazare, A. (2004). On apology. New York, NY: Oxford University Press.

Leuchtenburg, W. E. (1988). Franklin D. Roosevelt: The first modern president. In F. I. Greenstein (Ed.), Leadership in the modern presidency (pp. 7–40). Cambridge, MA: Harvard University Press.

Lind, J. (2008). Sorry states: Apologies in international politics. Ithaca, NY: Cornell University Press.

Medhurst, M. J., & Brands, H. W. (2000). Critical reflections on the Cold War: Linking rhetoric and history. College Station, TX: Texas A&M University Press.

National Archives. (n.d.). Public papers of the Presidents. Retrieved August 6, 2013, from http://www.archives.gov/federal-register/publications/presidential-papers.html

Negash, G. (2006). Apologia politica: States and their apologies by proxy. Lanham, MD: Rowman & Littlefield.

Nobles, M. (2008). The politics of official apologies. Cambridge, MA: Cambridge University Press.

Paul, G. (2007, November 6). Framing accounts: Understanding the account episode as negotiated interpretation. Paper presented at the annual meeting of the National Communication Association in Chicago, IL.

Smith, N. (2008). I was wrong: The meanings of apologies. Cambridge, UK: Cambridge University Press.

Tavuchis, N. (1993). Mea culpa: A sociology of apology and reconciliation (1st rev. ed.). Stanford, CA: Stanford University Press.

Towner, E. B. (2009). Apologia, image repair, and reconciliation: The application, limitations, and future directions of apologetic rhetoric. Communication Yearbook, 33, 430–468.

Ware, B. L., & Linkugel, W. A. (1973). They spoke in defense of themselves: On the generic criticism of apologia. Quarterly Journal of Speech, 59, 273–283.

Weiner, B. A. (2005). Sins of the parents: The politics of apologies in the United States. Philadelphia, PA: Temple University Press.

Yamazaki, J. W. (2005). Japanese apologies for World War II: A rhetorical study. New York, NY: Routledge.