Journal of Historical Linguistics - Volume 6, Issue 1, 2016

This article deals with the patterns of variation and change undergone by a syntactic variable in Early Modern Spanish grammar, namely the alternation between deber ‘have to, must, should’ and deber de + infinitive ‘have to, must, should’ as a modal periphrasis. Based on a 1,500,000-word corpus of immediacy texts (private letters, memories) the results of this variationist study suggest that throughout the 16th century, but more especially during its second half, the prepositional periphrasis gradually became more common, above all in stylistic contexts predominated by the spontaneity and proximity of the relationships between the interlocutors. It was also more frequently found in contexts involving members of northern speech communities, particularly males, the young and the middle-low social strata, the incidence being especially high at the points where some of these groups intersect. All this suggests a change from below in the Golden Age period, which reached considerable dimensions in a relatively short time but was destined to stagnate and later decline just as quickly in the centuries that followed as a result of some structural features, such as the special “visibility” of the preposition, which could have led to the stigmatization of the periphrasis in a similar way to what happened in other Spanish syntactic phenomena.

Old Finnish (1540–1809) shows variation between non-promotional and promotional passives. Historically, there is reason to suppose that the Finnic passive was originally promotional and that a shift to a non-promotional passive already took place during Proto-Finnic times. At first sight, the Old Finnish promotional passive could be based on contact with German, Swedish, and the classical languages — but it could also be a Proto-Finnic remnant conserved partially through these language contacts. To ascertain which is the case, I apply the notion advanced by Timberlake (1977) that a syntactic process of reanalysis and extension will first proceed in contexts which are least marked from the viewpoint of the new analysis, to a corpus of Old Finnish texts. The result is that some texts do indeed exhibit a stepwise generalization of non-promotional passives with, for example, indefinite and divisible arguments. This generalization is restricted to non-periphrastic passives: periphrastic passives are generally promotional. I argue that this state of affairs is best explained by assuming that a Proto-Finnic promotional passive remained in use in the language of the Finnish educated elite until early Old Finnish times, and that the non-promotional passive of Old Finnish is innovative.

The Eskimo-Uralic hypothesis of a genetic link between Eskimo-Aleut and the Uralic languages is now reaching its second centenary. Two major problems with its advancement since Bergsland’s (1959) summary of its status are addressed in this article. The first of these is the lack of an obvious correlate of the ubiquitous Eskimo-Aleut (EA) relative case marker -m in Uralic; the other is the lack of an m-initial first person singular morpheme in EA to correlate with that of the Uralic languages. That the EA singular genitive/relative marker -m — as well as the instrumental/accusative singular -mək based on it — might be cognate with Uralic singular accusative -m was suggested already by Sauvageot (1953) , but no firm conclusion on the matter has since been reached. This has remained a tantalizing possibility, despite the conflicting semantics. However, the remarkable morphosyntactic parallels between Eskimo-Aleut and Samoyedic in particular have grown more apparent with recent publications. A solution is proposed, linking the emergence of ergativity in the Eskimo-Aleut family with a reanalysis of the original nominative-accusative case marking system.

The Gothic Bible offers valuable secondary evidence for the pronunciation of Greek in the fourth century AD. However, inferences based on such data may result in a vicious circle, as the interpretation of Gothic is, to a great extent, dependent on the historical details of contemporary Greek. I show that a circular argument can be avoided by using a novel method, which is based on the comparison of transcription correspondences of Greek loan words and biblical names occurring in the Greek original and the Gothic version. I test the method by applying it to three example cases. The first concerns the aspirated stops φ, θ, χ: Gothic evidence confirms the fricativization of these stops. The second case concerns the potential fricativization of voiced stops β, δ, γ: the results are inconclusive, which is an important finding, since this shows that Gothic cannot be used as evidence for the fricativization of these stops. The third case concerns front vowels: Gothic evidence confirms the coalescence of αι and ε on the one hand, and ει and ῑ on the other, while it also indicates that η was not (yet) pronounced as [iː] in the fourth century AD.