270 WILLIAM H. JACOBSEN, JR.
- I rather think that the starred example (20) given by Schachter (1985:12)
*qw?asma mamwk (man-INDIC-3 work) could occur as a nuclear cosubordina-
tion with a reading parallel to this example: "There's a man working". Schachter
is, however, excluding a predicate plus argument construction with a reading
along the lines of "a working one is a man". - Cf. Sapir (1924:89, η. 55).
16. Cf. Rose (1981:210-211), who labels this morpheme temporal and illustrates its
meaning of a specific time or time duration. However, this should not have been
categorized as a tense, as it does not locate an event along the time line.
17. Cf. Rose (1981:158-159, 253-254) for similar Kyuquot examples. - This is part of a broader system of non-indicative quotative modes, to which
belongs also the quotative form ?anvc of the subordinating particle (section 10.7).
Cf. Sapir (1924:89, η. 57, 58; 90, η. 66; 101, η. 178; 102, η. 185); Swadesh
(1939:82; 1948:110, 112); Jacobsen (1986a:18).
19. Cf. Rose (1981:183-186) on avoidance of subordinate clauses in Kyuquot.
20. See Myhill (1984:199-201, 221-226) for helpful interpretation and generalization
of Jeffers's findings.
21. Cf. the summary by Nichols (1986:109, fn. 37) of difficulties and disagreements
experienced by Soviet linguists in trying to fit chained clauses of Uralic, Altaic,
and Northeast Caucasian into the coordination/subordination dichotomy as devel
oped for Russian.
22. Van Valin's revised treatment (this volume) seems essentially to agree with this
viewpoint. Cf. Longacre (1983:186): "Often the non-finals are clearly coordinate
(in the broad sense of theword) with the final clause ...". Elsewhere (1985:238),
however, he contrasts "co-ranking" structures and "chaining" structures, and
states that for the latter "... it is simply not possible to combine two verbs of the
same rank in the same sentence."
23. This interpretation would make the term cosubordination, with its successive pre
fixes, less appropriate (and in terms of neo-Latinity one would want consul·- as in
consul·stantiation). One might have introduced a term with an entirely different
prefix, such as adordination.
24. The short Nitinat text analyzed in Swadesh & Swadesh (1933) shows a high ratio
of about 3 to 2 of quotatives to absolutives in narrative passages. My information
about Makah comes from my own collection of texts, from which indeed I first
became aware of this use of the absolutive.
25. In the short Central Yana text analyzed by Sapir (1923:265-277), there is a ratio of
1 to 3 of past quotatives to narrative infinitives. I am indebted to Herbert W.
Luthin for drawing my attention to this Central Yana pattern.