1 An earlier version of this paper was published under the title of L'obviation en Kutenai—échantillon d'une catégorie grammaticale amérindienne, BSL 47.166–212 (1951).

2 For these criteria, see Louis Hjelmslev (transl. F. J. Whitfield), Prolegomena to a theory of language 6 (IUPAL Memoir 7; Baltimore, 1953), and my review, Lg. 30.70 (1954).

3 Obtained on a field trip in the summer of 1950, under a grant from the Phillips Fund of the American Philosophical Society, supplemented by a subsidy from the Faculty Research Fund of the University of Oklahoma.

4 Franz Boas' work includes two published items: Kutenai tales (BAE.B 59; Washington, 1917); Additional notes on the Kutenai language, IJAL 4.85–105 (1926). There is also an unpublished ms.: No. 48 in the Franz Boas Collection of Materials for American Indian Linguistics; see C. F. Voegelin and Z. S. Harris, Index to the Franz Boas Collection (Lg. Monograph 22; Baltimore, 1945).

5 For a detailed treatment of Kutenai structure, see my Kutenai I: Phonemics, IJAL 14.37–42 (1948); Kutenai II: Morpheme variations, IJAL 14.87–90 (1948); Kutenai III: Morpheme distributions (prefix, theme, suffix), IJAL 14.171–87 (1948); Kutenai IV: Word classes, IJAL 17.84–97 (1951).

6 It is obvious that the long vowels e·, o·, a· could be treated as short vowels + a feature (or phoneme) of length; this solution (i·, u·, a· instead of e·, o·, a·), however, would unnecessarily increase the graphemic differences between data published by myself and earlier material published by Boas.

7 I am here repeating my description of the Kutenai morphological index, first given in Problems in American Indian lexicography and text edition, Anais do XXXI Congr. Internac. de Americanistas 1025–7 (São Paulo, 1955).

8 In the following formulae, the symbol ± indicates that the morpheme(s) immediately following the symbol may be included in the sequence in line with occurrent distributional patterns, but are not always present. Morphemes not preceded by this symbol, on the other hand, are always present in the sequences symbolized by the formulae.

9 In the Kutenai examples, the individual morphemes in sequence with the theme are separated by hyphens from each other and from the latter, to indicate theme boundaries and morpheme boundaries outside the theme. Morpheme boundaries within the theme are not indicated. Apostrophes mark morphophonemic elisions at word boundaries. Each example is followed by the morphological index formula in parentheses, in which morpheme and theme symbols are separated by commas, words by semicolons. The symbols v₁, v₂, v₃. t in the formulae stand for various classes of particles; see IJAL 17.89–90 (1951). In the literal translations, several English words standing for a single Kutenai word are joined by hyphens. The place of each example in the corpus is indicated by a capital letter with a Roman numeral after it giving the text from which it is taken, and an Arabic numeral giving the serial number of the particular sentence (impressionistically segmented) within that text. For the sake of consistency, this numbering follows my original field notes. Texts marked H are historical or pseudohistorical (legendary), A are autobiographic, C are free conversations. For a recently published text from my collection, see my Colloquial Kutenai text: Conversation II, IJAL 20.316–34 (1954).

10 A second technique for the analysis of semantic content is the rephrasing in the same language of sequences minimally different in morpheme content; again, consistent differences in rephrasing as correlated with consistent differences in morphemic composition can be considered clues for a statement of morpheme meaning. My conversational knowledge of Kutenai is much too inadequate to allow use of this technique.

11 The terms used in these definitions are taken from my article Delimitation of syntactic units, Lg. 30.345–8 (1954). They are defined as follows (346–7): one word or word sequence is bilaterally dependent on two others if both the latter are necessary conditions for the occurrence of the first; one word or word sequence is unilaterally dependent on another if the latter alone is a necessary condition for the occurrence of the former; two or more words or word sequences are mutually tolerant if they can occur together but none is a necessary condition for the occurrence of any other. A substitution is considered right if it produces no change in the previously registered dependences; sequences suitable for a right substitution are called rightly substitutable.

12 Note the interesting parallel to the use of the instrumental case in Russian, as described by Roman Jakobson, Beitrag zur allgemeinen Kasuslehre, TCLP 6.265–6 (Prague, 1936).

13 Contrast the structure of the last clause of this sentence (... c pał #-n-?in-ne· ktunaxanana ‘... and really he-was [the] little-Kutenai‘) with the examples treated in §3.2.4.V.

14 The brackets here indicate a parenthetic coordinate clause, that is, one inserted between the phrases of an adjacent clause which thus becomes discontinuous.

15 Hjelmslev 60–1.

16 Comparable to N. S. Trubetzkoy's Grenzsignale, cf. Grundzüge der Phonologie 241 ff. (TCLP 7; Prague, 1939).

17 This is an instance of what C. F. Voegelin calls inferentially marked meaning, defined as ‘meaning marked by the association of two or more morphemes across word boundaries; or, more indirectly, by associations of morphemes within word boundaries’, Linguistically marked distinctions of meaning, Selected papers of the XXIX Internat. Congr. of Americanists 3.227 (Chicago, 1952).

18 Cf. Trubetzkoy 206 ff. For a somewhat different conception of grammatical neutralization, see Martin S. Ruipérez, The neutralization of morphological oppositions, Word 9.241–52 (1953).

19 Trubetzkoy 73.

20 Cf. Roman Jakobson, Zur Struktur des russischen Verbums, Charisteria Guilelmo Mathesio quinquagenario ... oblata 74–6 (Prague, 1932).