It has long been recognized that the Semitic suffix conjugation and the Berber adjectival perfective suffix conjugation have striking similarities in their morphology, which has been correctly attributed to be the result of a shared inheritance from Proto-Afro-Asiatic. Nevertheless, the function of these conjugations in the respective language families is quite distinct. This article argues that ultimately this suffix conjugation is a predicative suffix in the common ancestor of Berber and Semitic, and moreover shows that Semitic and Berber have significant overlap in the stem formations of adjectives. It is argued that these formations must likewise be reconstructed for their common ancestor.

The article investigates an overlooked development in the history of the English modals, namely the regularization of their plural inflection in Middle English (e.g. prs.ind.pl shulleþ for expected shullen). Using the LAEME and eLALME atlases and a number of electronic corpora, I document the frequency and dialectal distribution of such regularized modal verbs. It is shown that regularized shall was fairly common in Late Middle English, regularized can less so, and regularized may only very sporadically attested. The distribution of these forms shows a clear areal pattern, being most numerous in manuscripts from the southwest Midlands. I suggest that the most likely explanation for the observed patterns is interparadigmatic analogy with the ‘anomalous’ verb will, which in some dialects had developed the same stem vowel as plural shall.