266 WILLIAM H. JACOBSEN, JR.
Grammar Conference, University of California, Davis, March 15, 1986. Sections
relating to clausal cosubordination were also presented to the summer meeting of
the Society for the Study of the Indigenous Languages of the Americas, Univer
sity of Arizona, Tucson, July 1, 1989, under the title "The Nootka Absolutive and
Generalized Clause Chaining". I am indebted to Michele Emanatian, John
Haiman, Herbert W. Luthin, Sandra A. Thompson, and Robert D. Van Valin, Jr.
for their comments.
- Greenberg calculates numerical typological indices based on sample 100-word
passages for eight languages. I have done the same for Nootka, using the first 100
words in Text 2, pp. 19-21 of Sapir & Swadesh Nootka Texts (1939). These figures
should be taken only as ball-park indications, given the small sample. Each index
is a ratio of two units. That for synthesis is that of morphemes to words. For the
Nootka passage I find this to be 3.66, which is close to the 3.72 for Eskimo, often
cited as the type-case for polysynthesis. Greenberg mentions that values above
3.00 are infrequent, and that they would define a polysynthetic language (English
shows 1.68). This Nootka passage is atypical in that it contains called-out dis
course, entailing vocative forms (cf. Jacobsen 1986b), which add to the number of
morphemes. I have also counted -0 suffixes on certain absolutive forms, as
explained below (section 4). Excluding both these factors lowers the index to 3.39.
The other Nootka indices are as follows: agglutination (agglutinative junctures/
morph junctures) .70, compounding (roots/words) 1.00 (i.e., there is none), deri
vation (derivational morphemes/words) .97, gross inflection (inflectional mor
phemes/words) 1.69, prefixation (prefixes/words) .09, suffixation (suffixes/words)
2.57, isolation (uninflected words/nexus) .15, pure inflection (non-concordant
inflectional morphemes/nexus) .47, and concord (concordant morpheme features/
nexus) .38 (nexus is the sum of the last three relations). The indices of derivation,
gross inflection, and suffixation are not independent of that of synthesis, and these
are also high for Greenberg's sample, being exceeded only by Eskimo's 1.25, 1.75,
and 2.72. Imposing a derivation/inflection dichotomy on the multiple layers of suf
fixes is rather Procrustean; I have equated this with Sapir & Swadesh's (1939:236)
distinction between formative and incremental suffixes, which gives to inflection
some categories (such as causative and diminutive) that might otherwise be con
sidered derivational. Although the only prefixation in Nootka consists of redup
licative morphs, the prefixation index, while far below the 1.16 of Swahili, is yet
exceeded by those of only two languages in Greenberg's sample. The index of iso
lation (.15) is among the lower ones in the sample, the only lower being Eskimo's
.02. Those for pure inflection and concord are close to those of Eskimo and older
Indo-European (Sanskrit, Anglo-Saxon). Agglutination is difficult to compute;
my index of .70 is higher than others in Greenberg's sample, being approached
only by Swahili's .67; Eskimo has only .03. Cf. also Kroeber (1960) on the rank
orderings for these indices.
- Eskimo shows much the same pattern, which Sadock (1980, 1986) considers a kind
of noun incorporation, while Mithun (1986:32), after Sapir (1911b:254), prefers to
differentiate this as denominal verb formation, whereas canonical noun incorpora-
