161 Chapter 6.1 Middle English phonology in the digital age: What written corpora can tell us about sound change Nikolaus Ritt 1. Introduction While digital corpora have revolutionized the study of historical morpho-syntax and lexicology, their impact on the study of phonological change has so far not been overwhelming (cf. Laks 2008). This has various reasons. First, the relationship between spelling and phonology is hardly ever straightforward. Just how complex the relation between letters and sounds can get has been demonstrated in recent studies by Laing and Lass (e.g. 2003, 2009). Second, when spelling gets conventionalized, those conventions are transmitted—at least partly—independently of speech and standardize more easily. Therefore, they tend to mask the phonetic variability that drives phonological change. Since the tendency of spelling systems to standardize makes them also more conservative than phonological systems, they evolve more slowly than the latter, so that writing often lags behind phonology. Third, corpora are often based on printed editions of texts rather on than manuscripts. Potentially revealing scribal idiosyncrasies fail to be represented in them. Therefore, the vast amounts of quantitative data that can be so easily retrieved from corpora appear to require detailed item-by-item interpretation before much can be inferred from them about historical phonologies and the changes they underwent. Clearly, this defeats the whole purpose of using corpora in the first place. Of course, this may change in the future. The Middle English Grammar Project (Horobin 2000), for example, envisages the establishment of a corpus cum database, in which manuscript spellings are orthographically pre-analyzed. That will facilitate their phonological