On the history of Portuguese /u, ēi/: The case for spontaneous diphthongization

Abstract

During the 14th and 15th centuries, the Portuguese diphthong /āu/ developed from Old Portuguese /ã/ and /õ/, and the diphthong /ēi/ developed from Old Portuguese /ē/, and the nature of these sound changes has been the subject of a long-standing debate in Portuguese linguistics. Two types of theory have been presented, one phonological and the other analogical. The analogical explanations fail to be convincing for purely technical reasons related to the nature of morphological analogy itself. The phonological accounts employ a rule of spontaneous diphthongization which entails the prior lengthening of the three vowels. Carvalho (1989) rejects this type of phonetic process and proposes, instead, that the Old Portuguese glides arose through a rule of vocalization, i.e., *[-ãη] > [-ãw], *[-õη] > [-õw] > [-ãw], and *[-ēη)] > t-ēy]. In this paper, I weigh the evidence for each of the phonological theories and present some new cases of V diphthongization via spontaneous diphthongization; I conclude that the traditional theory is better supported by the evidence than Carvalho’s theory of vocalization. I also discuss the problem of diachronic rule exceptions within the framework of lexical diffusion.