The Arapaho Language (review)

Abstract

Reviewed by: The Arapaho Language Ives Goddard The Arapaho Language. Andrew Cowell, with Alonzo Moss, Sr.Boulder: University Press of Colorado, 2008. Pp. xx + 519. $65.00 (cloth). This book is a most welcome full description of Arapaho, packed with interesting data and analyzed examples. The language diverges greatly from other Algonquian languages, and to Algonquianists it generally appears alien and impenetrably difficult to see into. With occasional nods to Algonquian terminology, the authors justifiably take a completely fresh approach, with a functional emphasis on the interpenetration of the components of the grammar and on how to do things with the language in discourse and cultural contexts. There are major sections on phonology, inflectional morphology, derivational morphology, "usage," and "complex clauses and syntax," the last including chapters on discourse features, counting things, and variation. Morphophonemic variation is discussed throughout the book in connection with specific morphemes rather than being dealt with in a unified way. There are sequences of up to three vowel moras, and in the orthography used each mora is written separately as i, e, o ([ɔ]), or u ([Ɯ]). This convention was used by Salzmann (1956, 1963), presumably influenced by his thesis advisor, Carl Voegelin, who believed that the long vowels of Shawnee actually were sequences of two vowels and wrote them accordingly (Voegelin 1935). (Double vowels were also used in the Navajo orthography adopted in 1937, but these have always been presented as spelling long vowels.) Salzmann's convention of marking each mora for a pitch contrast between high and low (or here "normal") is also followed: he discerned six phonetic pitch levels and wrote an acute accent on moras with one of the three higher pitches, leaving the others unmarked. Salzmann placed the study of Arapaho on a firm footing and preserved an invaluable record of the language as spoken in the mid-twentieth century, but he confronted the formidable complexities of Arapaho prosody at a time when the phonological tools needed to make sense of it did not exist, as indeed they still may not, and his templates of Arapaho syllable types would not be accepted by phonologists today as even an adequate description, let alone as the basis for understanding what is really going on in Arapaho phonology. In fact, the authors concede that the section on pitch accent "is probably the most problematic in this grammar" (p. 22). Minimally, Arapaho phonology must be recognized to have the following features (I rely on sound recordings by Salzmann, Julius Moshinsky, Loretta Fowler, and myself; words in slashes are my phonemic transcriptions). There are long vowels distinct from heterosyllabic vowel sequences. Long vowels may have level pitch (/hí·siʔ/ 'day' [p. 24]) or falling pitch, e.g., /betê·č/ 'brain' (not betééc [p. 24]; audio clip in Conathan 2006); /bî·xû·t/ 'shirt' (not bíixúút [p. 24]). The level pitch may be high or low. The high and falling pitches have progressive lowering (terracing) of the pitch register in words, meaning that absolute pitch levels do not have a consistent allotonic status. There is a distinction between tautosyllabic /ei/ ( with level pitch like a long vowel, e.g., hísei 'women', hoséínoʔ 'meet') and heterosyllabic /e/ + /i/ ([ε.i] with falling or rising pitch, e.g., /béi/ 'awl, needle'; /hóteí/ 'mountain sheep (pl.)'). Open questions include whether sequences of two mores with falling pitch can be both heterosyllabic (/íi/) and tautosyllabic (/î·/), and in general what the possible pitch templates are for words with each occurring syllabic profile. (Some of these ideas were incorporated into the orthography used in Salzmann [1983], but this was later abandoned for a simpler system.) [End Page 388] The authors recognize that the postulation of underlying forms is useful in understanding and describing surface allomorphy, but their proposed abstract representations and the rules that interpret them are not integrated into a consistent system, which if anything might benefit from being even more abstract, and there are often loose ends. For example, heetneiʔtowuunoʔ 'I'm going to tell her' (from curricular materials that do not mark pitch accent) is given the underlying representation eti-eʔitowuun-oʔ (eti FUTURE, eʔitowuun 'tell', -oʔ 'first person sg. acting on third...