Absence. You don't know what you're missing. Or do you?

Since the inception of corpus linguistics (CL) the issue of absence has preoccupied both its practitioners and its detractors. To the latter it is self-evident, a truism, that a corpus can yield no information about phenomena it does not contain, a criticism which we hope to demonstrate is based on a failure to grasp the complexity of the notion of absences and an underestimation of the flexibility of corpus techniques. However the former, the exponents of CL, have also worried greatly about the significance of not finding something, say, a particular set of lexical items or a certain syntactic structure in their corpus. Is this (non) discovery telling me something about the discourse type(s) under study or about what is usually termed the ‘representativity’ of the corpus (i.e. how typical of the discourse type is the subset of it contained in the corpus)? And the CL literature is replete with warnings ‘not confuse corpus data with language itself’ (McEnery & +-ètHardie 2012: 26), to which we would add that observations arising from corpus data can only be generalised with the utmost care. Following Plato and Kant , we must not confuse the tangible, the phenomenal (corpus) with the intangible noumenal (language). In this chapter we will discuss, on the basis of a number of case studies, what can reasonably be inferred about discourses from corpus analysis, with regards to absences. Along with Scott, we maintain that ‘much can be inferred from what is absent’ (2004) and, following Taylor (2012), we will argue that corpus tools provide an ‘armoury’ for locating and verifying absence. In particular, the comparison and contrast across different corpora can firstly reveal absences, both those being searched for and others accidentally stumbled upon, and then allow the analyst to track the appearance and disappearance of linguistic elements or discoursal notions once they have come in some way to the analyst’s attention.