Diachronica - Volume 35, Issue 2, 2018

It is commonly held that Present-Day English they, their, them are not descended from Old English but derive from the Old Norse third-person plural pronouns þeir, þeira, þeim. This paper argues that the early northern English orthographic and distributional textual evidence agrees with an internal trajectory for the ‘þ-’ type personal pronouns in the North and indicates an origin in the Old English demonstratives þā, þāra, þām. The Northern Middle English third-person plural pronominal system was the result of the reanalysis from demonstrative to personal pronoun that is common cross-linguistically in Germanic and non-Germanic languages alike.

Many studies on Spanish verbal periphrases either discuss the general features of the syntactic category or analyse the semantic and functional differences between specific members of that category. The present paper focusses on one particular periphrasis, parecer “to seem” + infinitive but, at the same time, takes into account a similar, impersonal construction with parecer. Adopting a constructionist, usage-based approach, data are drawn from a large diachronic corpus, which makes it possible to describe the interplay between the two competing constructions and identify different semantic and syntactic contexts that favoured the increase of the periphrasis in the 19th century at the cost of the impersonal construction. More generally, the paper addresses the relevance of studying individual constructions in the context of other semantically and/or formally related constructions, since a change in one particular area of the constructional landscape can have repercussions on other areas as well.

Recent decades have seen a surge of interest in grammaticalization. In this paper, however, we are not concerned with reaching a better understanding of the nature of grammaticalization phenomena or their triggering factors, but we ask under what circumstances grammaticalization does not take place, even if it would have seemed likely to – a topic that has scarcely been addressed in the literature. Based on a comparative investigation of the historical development of a class of Indo-European spatial adverbs, we argue that mismatches between layers of linguistic structure present one type of situation in which grammaticalization may be blocked. For grammaticalization to occur, the outer semantic-syntactic boundaries of the potentially grammaticalizing construction must be matched by prosodic boundaries. If prosodic chunking is shifted in relation to semantic-syntactic chunking, grammaticalization may be prevented.