Enhancement of quantitative contrast by place as fortition

(Poster presentation)

In this paper I consider multiple cases where a consonant quantity contrast is enhanced by place specifications. Focusing on coronal non-glide sonorants, I argue that this process should be analysed as the counterpart of manner changes that are involved in the fortition of glides, as described by Bybee & Easterday (2019).

I consider several cases where a historical distinction between short and long (singleton and geminate) sonorants, especially laterals and coronal nasals, and in some cases also rhotics, is reflected as a qualitative distinction, usually involving place. Specifically, ‘long’ coronal sonorants are often reflected as laminals: examples include varieties of Ibero-Romance (Haudricourt & Juilland 1949, Martinet 1955) with palatal [ɲ ʎ], Old West Norse (Noreen 1923) with laminal n̪ l̪ , and Goidelic Celtic (Nance & Kirkham 2020) with [n̪ˠ ɲ l̪ ˠ ʎ] all reflecting original geminates. By contrast, singleton sonorants can be reflected as distinctively apical, notably Norwegian and Swedish varieties with [ɭ] for short l

In many languages we find the ‘long’ reflexes of sonorants in word-initial position, where quantity was not originally distinctive. As pointed out already by Martinet (1955), the presence of ‘long’ reflexes follows an implicational hierarchy: r (Spanish) > l (Catalan) > n (Asturian). The same sonority-based hierarchy regulates many other processes (Brandão de Carvalho 1989), but it is not clear whether the word-initial ‘long’ outcomes reflect the historical presence of phonological length (presupposing a process of initial gemination) or arise for other reasons. Very similar generalizations are in force in both Brythonic and Goidelic Celtic, where the changes are also involved in initial mutation.

I argue that the understanding of laminalization (enhancement of quantity by place) as fortition allows us to unify multiple processes whose parallel nature has not always been recognized before. This includes initial strengthening in Ancient Greek, which affects primarily the rhotic (Jatteau 2016), but also the palatal glide as well as, in some sources, nasals and laterals; the developments of palatal glides to obstruents in Romance, Germanic (Holtzmann’s Law, Faroese skerping), and Celtic; strenghtening of initial w- to gw- in Brythonic Celtic and Gallo-Romance.

Furthermore, I argue that this unified approach to fortition as enhancement accounts for how initial strengthening comes about without actual lengthening/gemination in strong (especially word-initial) positions, where it is not always phonologically motivated or supported by evidence, contra, for instance, Cravens (2002); however, I also reject the necessity of word-medial lenition as a prerequisite for initial strengthening, contra Hock (1992). I argue that these insights fit most naturally in a view of phonologization along the lines of Kiparsky (2015), where predictable information is integrated into the phonological grammar as part of the life cycle of phonological processes.

← In search of (lost?) contact-induced change in phonology

|  All presentations  |

Phonological grammar and uncovering past language contacts →