Stress Change and Initial Consonant Loss in Umpithamu: Cause or Consequence?*

Abstract

This study aims to reconstruct a change in the system of word stress in Umpithamu, a language of Cape York Peninsula, in northeastern Australia. The change involves a shift in parsing direction, from a left-aligned system as found in most Pama-Nyungan languages, to a system anchored at the right edge. The process of change is reconstructed using evidence from changes in root structure and root size, in combination with specifics of the synchronic system of word stress. The broader context for the analysis is a well-known hypothesis about stress shift as a cause of initial consonant loss, a typologically rare phenomenon attested in areal patterns in Australia, including Umpithamu and its neighbours. The reconstruction of stress change in Umpithamu shows that the picture is more complex, with the major stress shift as a consequence rather than a cause of initial loss. Umpithamu shows partial rather than complete initial loss, and thus reveals the complexities of the actual processes rather than their end point. Comparison with languages at the end point of the processes shows how their stress systems can be derived from an Umpithamu-type system simply by assuming further changes in root structure attested as following initial consonant loss. INTRODUCTIONThis study aims to reconstruct a change in the system of word stress in Umpithamu, a Middle Paman (< Paman < Pama-Nyungan) language of Cape York Peninsula, in northeastern Australia. The change involves a shift in parsing direction, from a left-aligned system as found *I am grateful to the late Florrie Bassani for teaching me about Umpithamu, and to the younger Lamalama people for continuing support and encouragement. Earlier versions of this study were presented at the Philological Society (London) and Vielfaltslinguistik III (Cologne), and in seminars in Canberra and Leuven. I thank

The mystery of word-initial consonant loss in Kera'a

“…Assuming two shifts helps explain the pattern, but the analysis is not supported -while also not contradicted -by empirical evidence. For Umpithamu, Verstraete (2021) argues that stress shift is a consequence of the initial consonant loss rather than its trigger, in the context of a detailed parsing direction change scenario. It is worth noting that in cases where stress shift is argued to be an explanation for consonant loss, the evidence for this is often impressionistic.…”

Where have all the sound changes gone? Examining the scarcity of evidence for regular sound change in Australian languages

“…To appear in: Linguistics Vanguard, Special collection on Sound change in endangered and small speech communities Word-initial deletion, or 'initial-dropping' of word initial consonants or CV sequences is not operative synchronically in any Australian language, but is widely attested as an irregular change, and in cases has been regular also (Blevins 2001b), most notably in proto-Arandic (Koch 1997;Koch 2001) and some languages of Cape York (Hale 1976b;Hale 1976c;Alpher 1976;Verstraete 2019). Like lenition discussed above, initial dropping would be also susceptible to lexical diffusion, at least in some cases.…”