Volume 76, Issue 2 - Special Issue: The emergence of middle voice systems: A typological diachronic perspective; Issue Editors: Guglielmo Inglese and Andrea Sansò

Much of the typological research on the middle voice has so far largely focused on the cross-linguistic definition and status of the middle as a voice category and its relationship to other voice operations such as reflexives and passives. Diachronic research on the middle, especially in a cross-linguistic perspective, remains to these days comparatively marginal. Overall, the existing studies have argued that middle markers generally originate from reflexive markers, and that the reflexive > middle diachronic path is unidirectional. This special issue collects a number of papers that address the diachrony of middle markers cross-linguistically, with the goal of either contributing to refine our understanding of already known diachronic pathways leading to middle markers or challenging assumptions on their possible historical sources (which go well beyond reflexives only) and the purported unidirectionality in their development.

This paper seeks to explore the possibility that Spanish intransitive verbs showing middle marking with no associated change of valency are a source of non-oppositional middle verbs. In order to do so, a study is conducted of the historical evolution of the middle-marked Spanish verb reír(se) ‘to laugh’ as well as other semantically related verbs (such as alegrar(se) ‘to be happy, pleased’, regocijarse ‘to rejoice’ and burlar(se) , mofar(se) , escarnecer(se) ‘to mock, to ridicule, to make fun of’), where the presence of the middle marker is associated with a valency-reducing diathesis.

The goal of this paper is to recharacterize the distribution of two suffixes in Common Turkic, the reflexive -(I)n and the reciprocal -(I)š . Despite having quite distinct sources (the former as a suffix employed in the derivation of intransitive verbs, the latter a plural marker), these two suffixes have fallen into a somewhat overlapping distribution in the modern Turkic languages. They share a range of middle-like functions, and so they are characterized here as a bipartite middle voice system. Many of these functions have become peripheral, however, and the system increasingly consists of a class of verbs with deponent paradigms.

This paper explores the middle voice in Kagulu, a Bantu language of Tanzania. Although not traditionally recognized in Bantu languages, recent research has asserted that middle voice is attested in some Bantu languages. We propose that of eight affixes that might be considered middle markers, Kagulu has two affixes that are part of the middle voice system, each one coding two different detransitivizing voices. We argue that, from a diachronic viewpoint, the underdeveloped voice syncretism of Kagulu’s middle markers is the result of competing morphology and minimal functional innovations towards voice syncretism in the derivational system of Kagulu.

In Malayic languages, prefixes derived from Proto-Malayic *(mb)AR- are characterized by an unusual multifunctionality pattern. On the one hand, continuators of *(mb)AR- may encode functions that are typical of middle markers across languages. On the other hand, they are used as denominal verb formatives. The aim of this article is to provide a diachronic explanation for such an unusual combination of functions. The denominal verbalizing function of *(mb)AR- appears to be primary, and a diachronic scenario can be sketched through which a denominal verb formative has extended to cover a portion of the functional space cross-linguistically encoded by middle markers.

Middle voice systems are characterized by the distribution of a middle marker over two macro-classes of verbs: oppositional and non-oppositional verbs. In diachronic studies, it has been proposed that the historical link between the two groups is unidirectional, with middle marking spreading from oppositional to non-oppositional verbs. In this paper, we challenge this assumption and discuss two case studies, one from Anatolian and one from Paman languages that show the opposite development, that is, from non-oppositional to oppositional. In both cases, we argue, constructions that originally had a lexically determined distribution develop grammatical functions connected with valency reduction.