Towards a Multilingual Approach to the History of English

construct (Myers-Scotton 2002; Auer and Muhamedova 2005). If this discussion is extrapolated beyond present-day communities, what becomes particularly noticeable, especially after reading the contributions in the two volumes under review, is that our contemporary notions of languages as clearly delineated and distinguishable entities do not apply to the Middle Ages. The processes of standardisation were not underway, so it is not uncommon to find great variability and hybridity across languages in contact. In the case of late medieval England, the primary focus of a significant proportion of Pahta et al. and, to a lesser extent, Louviot and Delesse, the simultaneous coexistence of not two but at least three languages (i.e., Medieval Latin and the two vernaculars, Middle English and Anglo-French) with an asymmetrical sociolinguistic distribution, complicates the disentangling of the origin of the lexical material used. Myers-Scotton (2002) postulated that frequency rather than integration could help to establish distinctions between code-switching and borrowing. Consequently, one imperative guiding code-switching research could be to obtain the largest possible data sample, something which now seems more feasible than a few decades ago, as can be gleaned from the contributions in Pahta et al. Nevertheless, attestations are still challenging: some lexical items may be hapax legomena or scribal verbatim copies from exempla (thereby not representing the actual language use of the individual); others may be so widely spread (internationalisms) that they cannot be easily traced. Faced with such taxonomical difficulties, each author in the present volumes proposes different methodological frameworks. For example, following Yaron Matras (2009, 110-14), Schendl puts forward five criteria for identifying one-word switches (Pahta et al. 2017, 49-50), and Queiroz de Barros identifies nine types of foreignisms applying graphemic, typological, phonological and grammatical criteria (Pahta et al. 2017, 70-72). The particulars of these classifications vary, but most of them—if not all—acknowledge the continuum on which code-switching and borrowing should be placed. Even relatively