Intertextuality and the 24-Hour News Cycle: A Day in the Rhetorical Life of Colin Powell's U.N. Address

Appendix D. Attitudinal Discourse in Linguistic and Multimodal Texts

239 Appendix D Attitudinal Discourse in Linguistic and Multimodal Texts x T he three main regions of attitude are affect, judgment, and appreciation . Table D.1 shows how these regions are broken down into subcategories. In this research, I distinguished between inscribed and invoked attitudes . In the first case, explicitly attitudinal lexis is employed. In the latter case, “The selection of ideational meanings is enough to invoke evaluation, even in the absence of attitudinal lexis that tells us directly how to feel” (62). As Martin and White (2005) explain, explicit inscriptions of attitude “act as sign-posts . . . telling us how to read the ideational selections that surround them” (64, my emphasis). In other words, an inscription of attitude can tell us how to discern invoked attitudes elsewhere in a text even when, strictly speaking, there is no evaluative lexis to speak of. Take, for instance, the following lines from Powell’s address: 240 Appendix D Table D.1. Subcategories of Affect, Judgment, and Appreciation Attitude Categories Domain Positive Examples Negative Examples Affect Un/happiness Affairs of the heart Happy Sad In/security Ecosocial well-being Anxious Confident Dis/ satisfaction Pursuit of goals Fascinated Bored Judgment Normality How special? Normal Odd Capacity How capable? Competent Incompetent Tenacity How dependable? Careful Reckless Veracity How honest? Honest Deceptive Propriety How far beyond reproach? Law-abiding Corrupt Appreciation Reaction: impact Did it grab me? Engaging Tedious Reaction: quality Did I like it? Beautiful Ugly Composition: balance Did it hang together? Logical Contradictory Composition: complexity Was it hard to follow? Detailed Simplistic Valuation Was it worthwhile? Convincing Unconvincing Source: Adapted from Martin and White (2005). Attitudinal Discourse in Texts 241 1. Saddam Hussein’s use of mustard and nerve gas against the Kurds in 1988 was one of the 20th century’s most horrible atrocities. 2. 5,000 men, women and children died. (UN/2.5/CP) Out of context, the second clause might be taken as non-attitudinal. However, the bolded inscription “horrible atrocities” in the first clause works as a signpost that directs the reader to interpret the second clause as evaluative (and not merely factual). We immediately understand that the 5,000 people described in clause 2 did not merely die; they were killed horribly and atrociously. Thus, even without any explicit evaluation, clause 2 undoubtedly reinforces a negative judgment of Saddam Hussein. Coding attitudinal lexis is not always straightforward. As Martin and White (2005) explain, “The realisation of an attitude tends to splash across a phase of discourse, irrespective of grammatical boundaries” (10). Thus, while a single word is prototypically the coding unit for inscriptions of attitude, they may also be realized in phrases. This can lead to coding ambiguities. Take, for example, the following sentence: “The design was deceptively simple.” As Read, Hope, and Carroll (2007) suggest, one might be tempted to code the bolded phrase as two separate attitudes: an instance of negative judgment (deceptively) followed by an instance of positive appreciation (simple). However, a better alternative would be to code the entire phrase as one instance of positive appreciation . Because of the potential for uncertainty, I enlisted the help of a second, independent coder familiar with Martin and White’s scheme. Based on a 15 percent sample of the linguistic data, we achieved an inter-rater agreement of .885. Coding differences were discussed and reconciled. I also consulted the other rater to discuss data that did not appear in this 15 percent sample, if I felt the coding was particularly ambiguous. Coding invoked attitudes was trickier. These attitudes can span a word, a phrase, or an entire clause—and they are not explicitly suggested by the presence of core evaluative lexis (e.g., good/bad).As Martin and White (2005) suggest, such implicit evaluations can introduce an “undesirable element of subjectivity into the analysis,” but ignoring them is even more problematic since it “amounts to a suggestion that ideational meaning is selected without regard to the attitudes it engenders ” (62). I once again enlisted the help of a second coder. This time my colleague and I took a collaborative coding approach. The idea here 242 Appendix D is to “reach agreement on each code through collaborative discussion rather than independent corroboration” (Smagorinsky 2008, 401). This is a more painstaking approach to inter-rater agreement, but a thoughtful exchange about each segment of data is particularly constructive when trying to tease out linguistically variable implications of...