Much of the work on countability stems from personal observations about which nouns occur in which co-texts: e.g. "one chair" is well formed but "one plastic" is not; "much plastic" is well formed but "much chair" is not. From these sorts of observations, a great deal of work has been done in the name of explaining why such examples are well formed or not.
When it comes to diachrony, while the observations are potentially more repeatable in the sense that they come from corpus based research, I will argue that the previous studies are not exactly straightforwardly repeatable. Despite some degree of opacity, these previous studies point towards something like a consensus that countability in Old English looked very different from Present Day English. Using the OED as a corpus, we see a different picture, i.e. Old English countability appears quite similar to Present Day English. This begs questioning the extent to which bias may play a role in previous studies and/or the work of the lexicographers and the present project, to what extent we can circumvent this, and what sort of methods can be used to uncover the nature of countability in Old English.
