Journal of Historical Linguistics - Volume 10, Issue 2, 2020

This introduction discusses the construal of temporal categories, their connection with spatial categories and general principles of the patterning of time in language. Transfers between categories and up language hierarchies are frequent types of change. It is shown how aspect may either lose or increase some of its functional load, and how tenses may develop from aspects, or extend into modal domains. General tendencies (such as developments from spatial to temporal categories) and systemic constraints are discussed in relation to the papers in this volume with additional Indo-European examples.

This paper presents an overview of the tens-aspect system in the Sogeram languages of Papua New Guinea. Taking the Proto-Sogeram reconstruction in Daniels (2015, 2020) as a starting point, I outline the innovations that have taken place in daughter languages and discuss the patterns of change that emerge. The study confirms a variety of known cross-linguistic tendencies, such as the common occurrence of the analytic-to-synthetic and aspect-to-tense pathways of change. More notable trends include the diachronic stability of the present and most remote past tenses; the instability of the middle pasts and future; the stability of the relative semantic ordering of tenses; the absence of a pathway leading from relative-tense to absolute-tense marking; and the ability of innovative tenses to be inserted anywhere into the five-way tense system of Proto-Sogeram. The study also illustrates how featural systems can interact over time, at first by introducing a new feature value in one system which can combine with values from another (as with the Manat habitual), and then, if the featural distinction is lost, creating a pattern of distributed exponence (as in Mum).

Languages of the Arandic subgroup of Pama-Nyungan languages of Australia have developed markers of aspect from a variety of sources, including verb phrases with stance auxiliaries and reduplicated forms. Other origins involve nominalised verb forms and the refunctionalisation of tense suffixes. Some unusual diachronic developments have to do with interactions between aspectual markers and those of the highly developed verbal category of associated motion. There are shifts in both directions – from aspectual to associated motion values as well as extension of associated motion to aspectual meanings. All the posited diachronic changes are inferred by means of reconstruction, since there is virtually no corpus of documents from which changes in real time can be traced.

The North-Angolan Bantu language Kisikongo has a present tense (Ø-R-ang-a; R = root) that is morphologically more marked than the future tense (Ø-R-a). We reconstruct how this typologically uncommon tense-marking feature came about by drawing on both historical and comparative evidence. Our diachronic corpus covers four centuries that can be subdivided in three periods, viz. (1) mid-17th, (2) late-19th/early-20th, and (3) late-20th/​early-21st centuries. The comparative data stem from several present-day languages of the “Kikongo Language Cluster.” We show that mid-17th century Kisikongo had three distinct constructions: Ø-R-a (with present progressive, habitual and generic meaning), Ø-R-ang-a (with present habitual meaning), and ku-R-a (with future meaning). By the end of the 19th century the last construction is no longer attested, and both present and future time reference are expressed by a segmentally identical construction, namely Ø-R-a. We argue that two seemingly independent but possibly interacting diachronic evolutions conspired towards such present-future isomorphism: (1) the semantic extension of an original present-tense construction from present to future leading to polysemy, and (2) the loss of the future prefix ku-, as part of a broader phenomenon of prefix reduction, inducing homonymy. To resolve the ambiguity, the Ø-R-ang-a construction evolved into the main present-tense construction.

Modern Japanese has a formally very simple tense-aspect system, which at its core has only three forms that are complemented by a number of peripheral markers and constructions. The core of the tense-aspect system was much more elaborate and complex in Classical Japanese. This paper discusses the systems of Modern and Classical Japanese, and then sketches the development from the latter to the former. This development involves the grammaticalization from aspect to tense, the recruitment of lexical means and constructions to renew aspectual categories, and category climbing. Two major paths of grammaticalization can be distinguished.

This paper examines the development of the aspect systems in the Indo-European languages Vedic and Latin. Even though aspectual distinctions are central in the verbal systems of both of these languages at the beginning of their attested traditions, they undergo quite different developments in the course of their history. The Vedic verbal system instantiates a classic case of the development from aspect to tense, whereas Latin maintains an aspect-based verbal system, which survives in the Romance languages. The paper explores the semantic properties of the Vedic and Latin past tenses in some detail from two distinct perspectives: a neo-Reichenbachian model, where aspect is regarded as a type of relation between reference time and event time (cf. e.g., Dahl 2010, 2015), and a model where aspect involves different types of partitive operators (cf. Altshuler 2013, 2014). Although these two approaches may first appear to be in conflict, the paper attempts to show that they represent complementary perspectives highlighting different dimensions of aspectual meaning.