Patterns of Affect Across Corpora

Abstract

In the emotion profiles described in Chapter 2, the focus was solely on individual emotion terms. This means that these profiles include occurrences for all sorts of different types of affect (emoted, unemoted, directed, undirected and so on), as there is no way of automatically knowing how an emotion term is used in discourse by looking at its frequency. As suggested in Section 3.3.5, it is only by looking at context and patterns that we can make some hypotheses about the meaning and functions of emotion terms. This is the aim of this chapter, which reports some of the findings of an analysis of the lexico-grammatical patterns of 15 emotion terms (see Table 3.1) in the BRC. The description focuses on L1 and R1 patterning, and is limited to detailing those patterns that are most frequent in each sub-corpus.¹ for example, ‘m is the most frequent L1 collocate of surprised in conversation, but not, be and was are also common. Nevertheless, the main focus is on ‘m surprised. Furthermore, I only mention functions of emotion terms that seem to cut across the usage of several terms — meaning that for example the noun affection is not discussed in detail in news reportage and academic discourse because it seems to have a distinct patterning.