On reading late modern intentions: a corpus-based analysis of the late modern English subjectification of be going to

In this talk we provide a detailed account of the formal, functional and semantic changes that be going to underwent in Late Modern English. We know the history of be going to, a classic of grammaticalization studies (see Traugott 2012 for a recent overview), in quite some detail. Specifically for the late modern period, the period of be going to’s maturation, Disney (2009) hypothesises that be going to extended from encoding intention to encoding prediction through increased guessing of other people’s intentions, resulting in its evidential semantics. While invaluable, Disney’s study is based on a limited corpus, and mostly qualitative in nature. We test and refine Disney’s hypothesis, and integrate it into the theory of subjectification (in the sense of Traugott 1989), on the basis of a large-scale corpus study. Despite be going to’s popularity, such studies remain rare. A recent large-scale study has focused on the initial stages of this process (Petré & Van de Velde 2014), but similar studies are still lacking for the late modern period. Hilpert (2008) provides some insight in the extension to new types of infinitive, but does not really go into details regarding the underlying mechanisms of change. Data are based on an extensive sample from CLMETEV (De Smet, Diller & Tyrkkö 2011) complemented by exhaustive data from PPCMBE (Kroch, Santorini & Diertani 2010). In total, 1257 attestations were analysed on quantifiable, mostly formal features at various levels of the construction, such as contraction of be (substantive level), extension to new subject types (e.g. raised, empty subjects) and types of infinitives (e.g., states rather than actions) (schematic level), and the sentence type and presence of epistemic and evidential marking (level of the host clause). The corpus data reveal that the shift from intention to prediction started to take place in the first half of the eighteenth century, and originated in contexts with third person subjects, often in past tense narratives. Naturally, reporting the intention of others generally involves a certain amount of guesswork. This amount increased by the middle of the eighteenth century, when be going to started to occur with non-imminent infinitival complements. This naturally resulted in an additional, epistemic layer of prediction that gradually gained strength during the second half of the eighteenth and the beginning of the nineteenth century, allowed the underlying meaning of intention to wither, and the construction to become more and more frequent in the present tense. The shift from intention to prediction is arguably an increase in subjectivity, as the emphasis gradually moved away from the grammatical subject to the speaker: what mattered was no longer the intentions of the subject, but the knowledge of the speaker about them. Traugott (1989) shows that other English epistemic auxiliaries such as will and shall go through a similar stage of subjectification. Interestingly, while there are significant differences between will, shall, and be going to, it appears that each goes through an intermediary stage that involved past tense uses with reference to a future in the past, which was already known to the speaker. From a theoretical point of view, our analysis provides further evidence for recent claims (e.g. De Smet 2012) that grammaticalization follows minimally disruptive pathways, taking the smallest steps possible in the development, and at the same time shows that recurring patterns may be found at this smallest level as well.status: publishe