Diachronica - Volume 43, Issue 1, 2026

This paper describes and analyses a historical alternation in what contemporary Mapudungun has developed as a set of transitive roots ending in [Cɨm], where C can represent a labial stop, a retroflex affricate or a velar stop. These forms are historically related to a causativisation process formed with a root ending in what is now a fricative plus the now-unproductive suffix [-ɨm]. An example of this pattern today is the adjective [laf] ‘(adj) flat’ vis-à-vis the transitive verb [lapɨm-] ‘to stretch (out)’. This paper posits that the word-internal stop consonant preceding [ɨm] is the unaffected remnant of an ancestral fricativisation process that targeted final stops in older Mapudungun. By presenting evidence from the historical record, the trajectory of these linguistic patterns is analysed in relation to the theoretical assumptions of the Life Cycle of Phonological Processes. This work contributes to theoretical questions about what is possible in language change as well as to the internal reconstruction of this language isolate.

The development of the English auxiliary do is a well-studied phenomenon, but there is no agreement on its source (see Denison 1993: Chapter 10). It has been suggested in the literature that a mono-causal account cannot satisfactorily capture the development (e.g., Fischer et al. 2017; Van der Auwera & Genee 2002), but no worked out alternative has been provided. In this article, we focus on the pre-innovation Old English stage and explore what properties of the verb itself and the verbal system of which it formed a part conspired to lead to the innovation. On the basis of this we propose a multiple source account in which both primary sources — uses of Old English don that form the building blocks of the innovative use — and secondary sources — properties of the language system that facilitated the development — play a crucial role. We also argue that our account expands the inventory of types of multiple source explanations (cf. De Smet et al. 2015).

Many word forms from different classes in Central Pame (cent2145, Otomanguean) allow two synonymous forms, one containing a prefix with the vowel /a/ and another one with /u/ (e.g. wa-ttsáuʔ~wu-ttsáuʔ ‘(s)he feels’). I conduct historical corpus research and elicitation to throw light on the diachronic origin and contemporary profile of this unusual phenomenon. Evidence suggests that it started as a sound change /a/ > /u/ between bilabial consonants. However, paradigmatic pressures have largely dismantled the original distribution of allomorphy synchronically, generalizing free variation (i.e. overabundance). In addition, other phonological and morphosyntactic cues have emerged for the probabilistic prediction of these allomorphies. The case provides an extraordinary window into the cognitive underpinnings of sound change, allomorphy, and the paradigm in a highly-inflecting and understudied language.

This paper deals with the historical development of the syllable-onset clusters /kl gl pl bl fl/ in Romance languages and dialects and with their articulatory and/or perceptual motivations. Several diachronic pathways are identified which depart from articulatory unstable [Cʎ] sequences, most distinctively lateral vocalization (e.g., [kʎ] > [kj]) and obstruent lenition (e.g., [kʎ] > [çʎ]). Most sound changes are attributed to articulatory variation insofar as they require adjustments in constriction degree and location. In a few cases the replacement of one consonantal sound by another appears to have been induced by acoustic-perceptual equivalence, as for example the substitution of [θ] by [f] in Franco-Provençal and of palatalized labial stops by palatal stops in southern Italy. Of special interest is the one-to-many derivation problem by which a given phonetic outcome may be achieved through more than one pathway as exemplified by the two phonetic developments /kl/ > [kʎ] > [kj] > [c] > [tʃ] > [ʃ] and /kl/ > [kʎ] > [çʎ] > [çj] > [ʃj] > [ʃ].