• DOI:10.1086/465826
• Corpus ID: 144185501

Proto-Algic I: Phonological Sketch

@article{Proulx1984ProtoAlgicIP,
  title={Proto-Algic I: Phonological Sketch},
  author={Paul Proulx},
  journal={International Journal of American Linguistics},
  year={1984},
  volume={50},
  pages={165 - 207}
}

• P. Proulx
• Published 1 April 1984
• Linguistics
• International Journal of American Linguistics

0. Introduction. Algonqian-Wiyot-Yurok is one of the more distant genetic relationships regarded as definitely established by adherents of the "splitting" tradition of historical Amerindian linguistics (Campbell and Mithun 1979:37-40).1 Indeed, from 1913-when Sapir first proposed it-until 1958 there was academic controversy about whether the three branches were genetically related at all (Haas 1958:sec. 1). Since 1958, it has been a favorite example (in methodological discussions) of a proven language family whose protolanguage has largely defied reconstruction (Haas 1967, Hamp 1970, Teeter 1974, Goddard 1975). Besides the distance of the relationship among its branches, this language family-variously called Algic, Algonquian-Ritwan, Algo-Ritwan, and even Algonquian-is of interest as an example of what Haas (1967:138-45) calls a protolanguage of the second order (PL2): one whose reconstruction depends on data from one or more ordinary protolanguages, in this case Proto-Algonquian (PA). Detailed reconstructions of PL2s are very rare, if indeed any can be cited: Proto-Indo-European, for example, was reconstructed from records of attested languages such as Latin and Gothic, not Proto-Romance and Proto-Germanic. In North America, PL2s which await reconstruction include Proto-AthapaskanEyak, Proto-Iroquoian, Proto-Siouan, and perhaps Proto-Caddoan. Due to the time depth of Proto-Algic and to its status as a PL2, its reconstruction requires especially careful attention to the principles of scientific hypothesis formation and confirmation. Here-as in any situation where one is constrained to develop a complex model using very limited data-plausibility considerations are crucial: a hypothesis/ assumption/explanation/bridging-argument is convincing if and only if its prior plausibility is established by a close analogue, chosen from an appropriate reference class. In the case of Algic, a highly appropriate reference class is Algonquian. Hence, any change from PA to one of its