publisher colophon

Comparative evidence for the Penutian hypothesis is very thin, but more evidence has been presented in the literature for the validity of two smaller units: Plateau Penutian, consisting of Sahaptian, Cayuse, Klamath-Modoc, and probably Maiduan, and Yok-Utian, consisting of Yokuts and Miwok-Costanoan. However, some evidence for a genetic relationship between these two units are provided here. Several lexical comparisons are presented that involve correspondences in internal word structure, larger word families, or both. Two morphological comparisons, of a locative construction and a numeral suffix, help to explain irregularities in the daughter languages.

1. Plateau Penutian, Yok-Utian, and Inland Penutian

Sapir’s (1990a, 1990b) proposed Plateau Penutian family, including Sahaptian (Sahaptin and Nez Perce), Molala, and Klamath-Modoc (and Cayuse, which is not sufficiently well documented for this hypothesis to be evaluated), is well supported by lexical (Aoki 1963; DeLancey, Genetti, and Rude 1988) and morphological evidence (Rude 1987; DeLancey 1992; Berman 1996). DeLancey and Golla (1997) present some evidence to link Maiduan to this group. Yok-Utian, a genealogical unit consisting of Yokuts and Miwok-Costanoan (Utian), is a more recent proposal (Gamble 1991), for which substantial evidence has been presented (Callaghan 1997, 2001, 2014). In this article, I present various pieces of evidence linking Plateau Penutian (hereafter, “Plateau”) and Yok-Utian.

DeLancey and Golla tentatively suggest that “all of Wintuan, Plateau Penutian, Maiduan, and Yok-Utian have a common ancestor more recent than Proto-Penutian” (1997:185). The inclusion of Wintuan here is dubious (Whistler 1977; Golla 1997), and is not considered in this article. I present evidence linking Plateau languages with Yokuts, Miwok-Costanoan, or both. My concern here is purely with this postulated relationship between Plateau and Yok-Utian; I do not venture into the larger question of the relationship of these families to other proposed members of a Penutian family. Nonetheless, I label the grouping of Plateau and Yok-Utian as “Inland Penutian,” assuming the validity of the theory that these belong in turn to larger Penutian family. Most of the comparisons presented in this article involve only two or three languages, usually Klamath, representing Plateau, and one or the other of Yokuts or Miwok. Obviously, inferring a Plateau—Yok-Utian relationship on the basis of correspondences between, say, Klamath and Miwok can only be as valid as the assumed higher units. For purposes of this article, I assume the validity of Plateau and Yok-Utian. [End Page 95]

Any genealogical connection between these groups must be at fairly considerable time depths. Moreover, the Plateau group appears to have undergone drastic grammatical restructuring (see section 4.1). Thus, we cannot expect to find extensive correspondences that will make the relationship obvious. In this article, I present a few lexical and grammatical correspondences that, I submit, can only be explained by a hypothesis of genealogical relationship. In section 2, I show evidence for a regular phonological correspondence between Klamath and Yok-Utian. In section 3, we see two lexical comparisons for which I propose deeper etymologies. Two morphological comparisons are presented in section 4. Finally, section 5 summarizes the evidence and the argument for recognizing these comparisons as demonstrating genealogical relationship.

2. Plateau and Yok-Utian lexical comparisons: phonology

The literature on Penutian includes a great many proposed lexical comparisons, including quite a few with both Plateau and Yokuts or Miwok-Costanoan comparanda. There are certainly many valid comparisons among them; for example, the well-known ‘lightning’ set—Klamath wlep’al–s, Proto-Utian *wilep, etc.1 (Shipley 1966)—sharing three successive consonants, must represent either genealogical relationship or some kind of horizontal transmission (i.e., borrowing or diffusion). The Proto-Yokuts and Proto-Utian reconstructions proposed by Golla (1964) and Callaghan (2014) include many forms for which there are Klamath or Sahaptian resemblants. But rather than simply add to the list of resemblant sets in the literature, in this section I present only a few comparisons whose phonological validity is clear. We see a regular nonidentical correspondence in section 2.1; section 2.2 presents a comparison that illustrates why not-quite-regular correspondences are common in comparative Penutian.

2.1. A regular correspondence

In table 1, I give ten sets representing a nonidentical phonological correspondence: Klamath alveopalatal c, c’ regularly corresponds to retroflex, *ţ’ in Proto-Yokuts, Proto-Miwok-Costanoan, or both, in both initial and medial position (this is one of the earliest “Penutian” comparisons; see Hymes 1964).2

Table 1. Klamath c, c’ Corresponding to Yokuts and Miwok-Costanoan *ţ, *ţ’
Click for larger view
View full resolution
Table 1.

Klamath c, c’ Corresponding to Yokuts and Miwok-Costanoan , *ţ’

[End Page 96]

Obviously these are not all of equal value. ‘Yellowhammer’ is obviously onomatopoeic, and Callaghan notes that it is a likely Wanderwort. But it fits on the list because it participates in a demonstrably regular correspondence. (The equation Proto—Miwok-Costanoan *w = Klamath w is supported by the ‘lightning’ set mentioned above.) On the other hand, ‘peel’ and ‘skin, bark’ show deeper systematic correspondences that are explored in section 3.

Several of these comparisons (‘bark’, ‘twist’, ‘shine’, and ‘star’) have an extra initial consonant in the Klamath form, which requires explanation. In comparisons with other Penutian languages, Klamath forms consistently show syncope of the vowel of the initial syllable (Shipley 1966), and the forms in question here presumably are instances of this. The common claim that Penutian languages have few prefixes (Callaghan 2001:314), is true for Yok-Utian, but not for the Plateau languages; these, including Klamath, have elaborate systems of lexical prefixes (DeLancey 1999), which, notoriously, are phonologically reduced. So while I do not have explanations available for any of these extra consonants, it is at least plausible that explanations might be found; noncorresponding initial segments in Yokuts or Miwok-Costanoan would be problematic, but in Klamath they are at least potentially explainable.

2.2. Where irregular correspondences come from

In comparing any set of Penutian languages, one encounters plausible resemblant sets in which forms do not correspond in the glottalization of a consonant. In the etymology for ‘cloud, cloudy’ in table 2, we see one way in which such miscorrespondences originate.

Table 2. Words for ‘cloud, cloudy’
Click for larger view
View full resolution
Table 2.

Words for ‘cloud, cloudy’

The comparable element is Miwok pa– : Nez Perce pe– : Klamath p’a–; moreover, the Miwokan and Nez Perce show a CVCV form in which both consonants (ʔ in the initial syllable, and p in the second syllable) correspond. To include Klamath in this comparison, we need to account for the missing initial syllable, the glottalization of the root initial, and the final two segments. The first two differences are in fact the same difference: given a root of the form *ʔVpV, syncope of the first vowel, a regular change in Klamath (section 2.1) would bring the initial in contact with the stop, a change that is a regular source of glottalized consonants in Klamath; compare Maidu ʔo–no, Klamath n’o– ‘head’. [End Page 97]

There is a plausible explanation for the remainder of the Klamath word, although it is not as clean as the story of the initial. Klamath has a large number of nouns in –i:s, morphologically {–y–s} (e.g., waqi:s ‘ladder’). As well as occurring frozen in a large number of noun stems, this form functions semiproductively as a nominalizer of certain verbs, mostly sonorant-final stems. The stem {p’ays} cannot be synchronically analyzed as containing this nominalizer, since {p’ays} not only takes the normal verb suffix –a in the verbal form p’aysa, but requires a further nominalizer to function as a noun in p’aysas ‘cloud’. Still, the comparison with Nez Perce, and more distantly with Miwokan, supports segmenting the –ys; ays– could then plausibly be a frozen nominalization that has been reinterpreted as an unanalyzable stem.

3. Plateau and Yok-Utian etymologies

One way in which comparison of lexical resemblants can be made more convincing is by finding regular phonological correspondences, or ways to explain irregular correspondences, as in section 2 above. Another is when we can internally reconstruct a history that explains apparent problems in comparing synchronic forms. In this section, I present two etymologies in which we can identify cognate material undergoing divergent derivational histories in the two groups. The first, ‘peel/skin/bark’, is a mainstay of the Penutian literature and one of the strongest pieces of evidence for the relationship of Plateau and Yok-Utian. The second, ‘medicine/shaman’, involves three etyma, of varying certainty, that show deep semantic correspondences.

3.1. ‘Peel’, ‘skin’, ‘bark’

The word family ‘to peel’, ‘skin’, ‘bark’ is attested in other Penutian branches as well (Hymes 1964:236). For Penutian as a whole there may be enough data for a monograph-length treatment; here I present only enough Klamath and Yok-Utian data to establish the argument for cognacy. Let us first look at forms meaning ‘to peel’, in table 3.

Table 3. Words for ‘to peel’
Click for larger view
View full resolution
Table 3.

Words for ‘to peel’

The Proto—Miwok-Costanoan form is “the Simplex Grade of a longer stem, probably *ţol:e– or *ţol:a–”; the Proto—Sierra Miwok form [is] “a Lengthened Simplex Grade plus –u– < *–y– [i.e., – i–] ‘verbalizer’” (Callaghan 2014:417).

The apparent variation in vowel length may be illusory; quantitative stem alternations are common in all of the languages under consideration. Note that while Barker (1963) gives the Klamath root with a short vowel, two of his three [End Page 98] entries under this lemma have a long vowel: doc’o:l’a ‘peel off an outer layer’, sqac’o:l’a ‘pull off easily’. Newman (1944:57) calls vowel length in Yokuts a “retardative” morpheme. Barker glosses morphological lengthening of a stem vowel in Klamath as “action upon plural, repeated and intensive action (but not distributive action)” (1964:120), but this sense is not always apparent in his examples. In particular, the pairing of hoḏ– ‘run, jump’ with ho:ḏ– ‘run, jump slowly’ is explained much better by Newman’s Yokuts gloss. (For this reason, Barker [1964] does not analyze this pair as exemplifying his length morpheme. But the length alternation, and attendant semantics, are particularly interesting since, as is seen in section 4.1 below, these Klamath forms have good Yok-Utian cognates.)

Matching not only in consonants, but in stem vowel as well, and also in quantitative alternation, this is quite a nice set. It is improved further by being part of an extensive word family in both branches. Obviously related, through ablaut and a stem-extending *–k, are Proto–Eastern Miwok *ţalka–, Klamath c’elk–s ‘hide, skin’. Proto-Yokuts *č’uluy ‘hide, skin’ must fit here somehow as well, but perhaps not directly; the Miwok—Klamath comparison, however, is nearly perfect—the difference in manner of the initial is not a problem, since Miwok lacks a glottalized series, so only the quality of the stem vowel remains to be accounted for.

A third set of forms is seen in Proto-Yokuts *ţ’o:tot ‘skin, bark’, Klamath nc’et–s ‘inner bark’. Callaghan (1997 fn. 22) notes the probable relevance to this set of Proto—Sierra Miwok *ţol:e–ča– ‘peeling, skin’, but, like Proto-Yokuts *č’uluy ‘hide, skin’, this may be better compared with the ‘peel’ forms discussed above.

The word family so far is summarized in table 4.

Table 4. Forms Meaning ‘peel/skin/bark’
Click for larger view
View full resolution
Table 4.

Forms Meaning ‘peel/skin/bark’

In this group of etymologies, we have correspondence not simply of segments and CV structure, but of word structure and fossilized derivational morphology.

Other words in various languages appear to belong to the same word family, but their place in the scheme is not yet evident, e.g., Klamath c’eL–as ‘peeling, skin of fruit’, c’iLaq ‘shell, pod hull, scale (of a fish)’, Lake Miwok luk ‘skin, peel, shell, outer covering’, ṣúlik ‘to skin’. The equation Klamath c : Lake Miwok is not regular (or, at least, I do not at present have other sets to support it), but inexplicable alternations among the palatal initials are an unhappy fact of life in Penutian (see, e.g., Callaghan 1997:31). If we allow Klamath c to correspond to Lake Miwok , then Lake Miwok ṣulih ‘to shed, peel off’ corresponds to Klamath [End Page 99] –c’oli ‘peel’. Proto-Yokuts *č’uluy ‘hide, skin’ most closely resembles this set in form, although its gloss better matches the **ţV–l–k forms.

3.2. ‘Medicine’, ‘shaman’

In the lexicon relating to medicine and shamanism, we find three different roots, hypothetically *yoq, *tui–, and *ni:q, all of which are present in both families, and combine in different ways in the daughter languages. The first root appears in Plateau, in Klamath yawG–a ‘to doctor’ and qyoq–s ‘shaman’ (the source for the initial q in qyoq–s is not clear to me). Nez Perce wé:yek ‘go to receive a guardian spirit’ might be related, but since the coda consonant does not correspond in place of articulation (it is velar rather than uvular), we cannot consider it for now. (Nez Perce contrasts velar and uvular stops.)

The hypothetical root *tui offers better correspondences, at least within Yok-Utian: Proto-Yokuts *t’uyox ‘medicine’, Proto—Sierra Miwok *tuy:uk ‘shaman, poison doctor’. (Miwok does not have an uvular series.) Callaghan (2014:408) suspects the Sierra Miwok form of being a borrowing from Yokuts, so for our purposes the Miwok form does not count as an independent witness. Moreover, the Yokuts and Miwok form is a much stronger witness if we can account for the first syllable.

A potential comparison for these Yok-Utian forms in Klamath is the first syllable of t’wini:q–s ‘berdache’ (transvestites who “took the role of shaman and were credited with great spirit power” [Barker 1963:424]). The separability of the first syllable is supported by comparison with Proto-Sahaptian *tiwé:t ‘shaman’ (Aoki 1962:180). Although the root is not analyzable within Sahaptian, nominalizing *–t is widely and deeply attested in Penutian; in both Klamath and Nez Perce its reflex –t derives participles and other grammatically nominalized verbs, so identifying the final *t in *tiwé:t as this suffix is not entirely ad hoc, although it is also not provable. If this is accepted, it gives us a Proto-Plateau stem *tiwé—a form that we would expect to undergo syncope in Klamath, as we have seen already (something similar has happened in Yakima Sahaptin twáti ‘Indian doctor’). The same root may, in turn, be comparable to Proto—Valley Yokuts *tʰe:yiṣ < **tʰi:yiṣ ‘curing doctor, sucking doctor, shaman’ (Callaghan 1997:58). This second Yokuts form and the Proto-Sahaptian form differ from the Klamath form and the first Yokuts form in not showing glottalization of the initial segment, so these comparisons are problematic.

Nonetheless, the Klamath-Yokuts comparison is promising if we can also explain the second syllable of the Klamath form. In fact, Hymes (1983:130–31) long ago suggested such an analysis of t’wini:q–s, comparing it to forms across Penutian that refer to mythic or supernatural female beings (ogres or bears), including Wasco (Chinookan) a–t’únaqa ‘female ogre’ and Southern Sierra Miwok tunak:a– ‘bear moiety’. Hymes sees the putative Penutian form as (diachronically) segmentable, identifying the part of it after the first syllable with “Tsimshian . . . mythical bear power names that appear to contain the [End Page 100] –naqa element” (1983:131); this segmentation is supported by Klamath n’a:q’a ‘cinnamon bear’.

Some of these connections are very strong, others, particularly Hymes’s account of “the –naqa element,” are more tenuous. (Another possible comparison here is the last syllable of Sahaptin tawt–núk ‘medicine’, formed on the stem *tiwé:t.) However, putting all these proposals together, we have an account that provides a plausible place for every component of every Plateau and Yok-Utian word relating to medicine and shamanism mentioned in the present section—even if this set of words ends up illustrating many of the chronic problems of comparative Penutian: sporadic glottalization, vocalic ablaut grades or possibly relict vowel harmony, and quantitative stem alternations.

4. Klamath—Miwok morphological comparisons

The gold standard for demonstrating genealogical relationship is paradigmatic and syntagmatic morphological correspondence, ideally including corresponding irregularities. As noted in section 1.2, the grammatical structures of the Plateau and Yok-Utian groups are very different, probably as a result of dramatic restructuring of the grammar in Proto-Plateau; the basic differences are briefly outlined in section 4.1. As a result, we do not find full paradigmatic correspondences. What we can find, in principle, are relict irregularities reflecting older syntagms that may be comparable. In this section, I propose two morphological comparisons between Klamath and Miwokan: a shared locative formation in section 4.2, and a cognate plural formative in section 4.3.

4.1. Yok-Utian and Plateau grammatical structure

The grammatical structures of the Plateau and Yok-Utian groups are very different. Yokuts and Miwok-Costanoan both have a verbal system based on templatic stem alternations of triconsonantal roots. Kramer (1995) shows that the templatic systems of Miwokan and Yokuts correspond in close detail. Compare the long and weak stems of cognate verbs for ‘run’ in Southern Sierra Miwok and Yawelumni Yokuts ‘run’ (Kramer 1995:440), shown in table 5; in both languages, the second vowel of the root is long in the long stem, while the first vowel of the root is long in the weak stem.

Table 5. Corresponding Alternations for Verb Stem ‘run’ in Miwok and Yokuts
Click for larger view
View full resolution
Table 5.

Corresponding Alternations for Verb Stem ‘run’ in Miwok and Yokuts

This is the general structure of Takelma as well, and can be attributed to Proto-Penutian.3 (Indeed, the length alternation in the evidently cognate Klamath hoḏ– ‘run, jump’, ho:ḏ– ‘run slowly’ [section 3.1], might be a distant [End Page 101] reflex of this stem alternation.) Consistent with this inference, much of the morphology of the Utian languages seems to be old (Callaghan 2014).

Plateau languages, in contrast, have undergone a major grammatical reorganization from the archaic pattern. They have bipartite or even tripartite verbs that reflect old compounding constructions. This is an areal feature (Dixon and Kroeber 1919; DeLancey 1996) that can probably be reconstructed to Proto-Plateau. We find a few members of both initial and final stem classes corresponding between Klamath and Nez Perce, and even a few correspondences between complete bipartite stems. These are not numerous, but there are a few examples, such as those in table 6, which involve Yakima wá– ‘with instrument’, Nez Perce we– ‘with an implement, with a stick, axe, beak’, and Klamath w– ‘with a long instrument’.

Table 6. Corresponding Bipartite Stems in Klamath and Sahaptian
Click for larger view
View full resolution
Table 6.

Corresponding Bipartite Stems in Klamath and Sahaptian

Still, even if this structure, the relevant categories, and some of their members are cognate across the Plateau languages, the majority of bound stems are nonetheless not cognate; that implies that this type of stem form represents an innovation at the Proto-Plateau level that was then independently elaborated in the daughter languages. Consistent with this picture, much of the morphology of Klamath (DeLancey 1991), Sahaptian (Rude 1996, 1997), and Maiduan appears to be relatively new.

4.2. Irregular locatives in Klamath and Miwok

The productive Klamath locative is a suffix {ḏat}, ({ḏ} represents lengthening of a stem-final consonant, but td—phonetically, [tt]—after a vowel: som ‘mouth’, sommat ‘in/on the mouth’; ʔambo ‘water’, ʔambotdat ‘in the water’). This is a morphophonemic anomaly; no other suffix in the language shows similar behavior. There is also a relict locative –t, which occurs in a few adverbials and demonstratives such as ge:–t ‘over there’. A third locative form, –t’a, occurs in a small set of lexicalized adverbs, including gina:t’a ‘on this side’, goni:t’a ‘on the other side’, wik’a:t’a ‘near’, and tabi:t’a ‘behind’. These appear to be historically related, but as yet, no Klamath- or Plateau-internal clue has been discovered as to how. However, such a clue can be found in Miwok. In Sierra Miwok, the locative is –t, with a partially predictable alternant –t:oʔ. This is the reflex of Callaghan’s (2014:180) Proto-Utian *–to ‘allative’; for Proto-Miwok, she reconstructs *–to after consonants, *–t:o ~ –t after vowels—that is, the length alternation in the consonant [End Page 102] has the same conditioning as in Klamath. Sierra Miwok also has a “directional suffix” –t: it, formally quite similar to Klamath {ḏat}, including the gemination associated with the first segment. Freeland explains this form as follows: nominals in locative case form nominal derivatives with the suffix i, e.g., né:–t ‘at this place, in this locality’, né–t:– i ‘this place, this locality’; the “directional suffix” –t: it “apparently arises through the re-adding of the locative suffix –t to nominal endings of this type” (1951:150 n. 148).

This, then, provides an etymology for Klamath {ḏat} that explains both its phonological bulk and its idiosyncratic morphophonemics. The difference between Klamath –dtat and Sierra Miwok –t:ït is the reverse of the difference between Klamath and Sierra Miwok –t; {ḏat} is the general, productive Klamath locative, while Sierra Miwok –t:ït is restricted to an apparently small set of adverbials (Freeland 1951:172–73). Interestingly, Klamath also has a secondary locative formation of exactly the same structure. In forms like wik’a:tant ‘close to’, the final syllable consists, on Barker’s analysis, of locative –t followed by a “nonnominative theme formant” (Barker 1964:287) –an, which forms oblique nominal themes from forms that are not nominal and thus not eligible for case inflection; this is in turn followed by a second locative –t. This is exactly Freeland’s analysis of –t: it, wherein Sierra Miwok i, which forms noun stems from locative and genitive forms, functionally corresponds to Klamath –an. Thus, this construction is repeated historically. Sierra Miwok is more conservative than Klamath; the latter extended the first version of this formation (i.e., locative plus nominalizing suffix plus locative) to become the ordinary locative, and then innovated a new version, which differs from the old only in using a new oblique theme formative.

The most striking aspect of this comparison is the correspondence of the morphophonemically lengthened suffix-initial t in both languages, explicable in Miwok (when the nominalizer i is added to a case form, “the consonant of the case suffix . . . is lengthened” [Freeland 1951:150 n. 48]) but completely mysterious in Klamath. This is especially striking, given that the only other remotely similar phenomenon in Klamath is the behavior of the stems {hoḏ} ‘run, jump’ and {ho:ḏ} ‘run, jump slowly’. Both of these are bound motion stems that require a locative-directive stem (DeLancey 1999), and induce gemination in the initial consonant of the locative-directive stem: hokkyamna ‘run around something’, hollac’wi ‘run right up to’, ho:cc’na ‘run slowly’, howwa ‘run into water’, etc. These have obvious comparanda in Sierra Miwok h iwá:t–, h iwát:– ‘run’ (section 4.1); additionally, the gemination induced by the Klamath stem corresponds to the lengthened final t: in the second stem in Miwok in the same way that the gemination induced in a preceding consonant by Klamath {ḏat} corresponds to the lengthened initial t in –t: it.

4.3. An ancient plural

Klamath and Miwok share a plural formative in –k, as seen in table 7 below. In both languages it forms plural demonstrative stems; on Barker’s analysis (1964:250), this is its only function in Klamath. In Southern [End Page 103] Sierra Miwok, –ko–, –k:o– ‘third person plural’ (Freeland 1951:27, 31–32) has the same function. (The Molala demonstrative roots, ni:– ‘this’, nu:– ‘that’, correspond paradigmatically to Miwokan and Costanoan [Berman 1996:25]. Sahaptian languages have forms with an initial velar consonant for both, e.g., Nez Perce kí:– ‘this’, ku– ‘that’; Klamath appears to have one from each set.)

Table 7. Klamath and Southern Sierra Miwok Demonstratives
Click for larger view
View full resolution
Table 7.

Klamath and Southern Sierra Miwok Demonstratives

Another form in which Freeland identifies this suffix is the numeral ʔotí:ko– ‘two’; the same suffix is also clearly present in Sierra Miwok *tolo:–ko... ‘three’ (Broadbent and Callaghan 1960). Callaghan (2014) reconstructs the suffix for Proto-Miwok. It does not occur with other numerals in Sierra Miwok, but in Plains Miwok the same element occurs in the numerals ‘two’ through ‘five’, but in no others: ʔo:yok:o– ‘two’, tel:ok:o– ‘three’, ʔoysek:o– ‘four’, kas:ok:o– five’ (Callaghan 1984). On purely Miwokan evidence, we cannot go farther than this. Plains Miwok could represent the original Proto-Miwok situation, with the suffix lost irregularly in the other languages. Alternatively, the Plains situation could represent analogical spreading of the suffix through the paradigm based on its irregular occurrence in one numeral.

The Plateau languages show a form that corresponds to this, as well as to the Klamath demonstrative pluralizer (see table 8), although Barker does not connect the two Klamath forms. In Klamath and Sahaptin, the lower numerals occur with an apparently cognate inclusive suffix (DeLancey 1992). In Klamath, this is (’)ok,4 used with the numerals ‘two’ through ‘five’, and probably also identifiable in the synchronically unanalyzable na:nok ‘all’ (Barker 1964:271). The corresponding Sahaptin form is the suffix –(w)ə́k ‘all’, which occurs only with numeral stems (Jacobs 1931:241). Jacobs does not specify which numerals this occurs with, but gives examples only with ‘two’ through ‘four’.

Table 8. The “Inclusive” Suffix with Lower Numerals in Klamath and Sahaptin
Click for larger view
View full resolution
Table 8.

The “Inclusive” Suffix with Lower Numerals in Klamath and Sahaptin

This looks something like the suffixed numerals in Miwok, although the forms are analyzed quite differently by Freeland and Barker. But a simple comparison [End Page 104] of the relevant numeral forms in the two languages in table 9 suggests that the final syllables might be cognate. And if they are, then Klamath –’ok ‘inclusive’ and –k ‘demonstrative pluralizer’ are related, since they are both cognate to the same morpheme in Miwok.

Table 9. *–ko ‘plural’ with Lower Numerals in Miwok and Klamath
Click for larger view
View full resolution
Table 9.

*–ko ‘plural’ with Lower Numerals in Miwok and Klamath

If this analysis is to work, the first requirement is an account of the difference in the Klamath forms; in fact, we can find such an account within Sahaptian. The w that occurs after vowel-final stems in Sahaptin is the personal classifier used in Sahaptian languages with numerals when counting people (Jansen 2010:147), still attested in Nez Perce as well, although Nez Perce lacks the inclusive suffix: ʔuy–né:p ‘seven’, ʔuy–né:p–we ‘seven people’ (Aoki 1970:59). While there is no such form in Klamath, we can identify the vowel o in the Klamath suffix as corresponding to Sahaptin w, and thus recognize it as a Plateau innovation. We are then left with something like *–ak in pre-Proto-Sahaptian, which is close enough to the demonstrative pluralizing –k that we can entertain the idea that they have the same origin. The second requirement is an explanation of the differences between the Plateau and Miwok forms and constructions. I have no explanation for the difference in form, save to guess that the absence of a final vowel in Plateau could be a case of apocope.

An obvious problem here is the fact that the numeral stems themselves do not correspond between Plateau and Miwokan; a solution to that may be found as we reconstruct the function of the suffix. The demonstrative suffix is unambiguously plural in both Klamath and Miwok. It is not obvious why a language would mark numerals for plurality, but one can imagine such a construction using something like English all, emphasizing that the speaker means exactly that number. In turn, this function has been preserved in Sahaptin and Klamath, so it must have still been productive in Proto-Sahaptian and could be applied to new numeral stems. The Miwok data suggest that the functionality of the construction has decayed—presumably reflecting the common process whereby originally intensifying expressions diminish in force—with the result that the suffix has been preserved in the daughter languages in this or that numeral. That is to say, if we can use the Plateau facts in our reconstruction of Miwokan, we can postulate that in Proto-Miwok the *–k(:)o suffix had the same distribution and behavior as Proto-Plateau *–ok—that is, this suffix occurred optionally, but only with the numerals ‘two’ through ‘five’. Subsequently, the daughter languages merged the plain and suffixed numerals into a single [End Page 105] paradigm, retaining some members of each. In this scenario, it seems intuitively plausible that the suffixed form of ‘two’ should be the most widely attested suffixed form across Miwok; ‘two’ seems to have a strong affinity for special emphatic or inclusive forms (e.g., English both).

5. Conclusion

As the case for a genealogical relationship, the evidence presented in this article is sparse, but strong. A dozen or so lexical comparisons and two corresponding grammatical morphemes are slender evidence for a genealogical relationship. However, it is difficult to explain the data presented here any other way. We have evidence not simply for a set of shared lexical items, but phonological correspondences, deep etymologies, and shared word families.

While I consider that most of the other correspondences in this article necessarily represent shared inheritance rather than horizontal transmission, the lexical comparisons in section 3.2, even if all valid, could be explained as borrowing. But they would have to reflect sharing, at a considerable time depth (since the roots have had time to combine into different stems in different languages), of a detailed (because several roots are involved) semantic mapping of a particular cultural domain. So they are at least evidence that at some point in the distant past Proto-Plateau and Proto—Yok-Utian were in close contact, geographically and culturally. But since we have both lexical and morphological evidence which cannot easily be interpreted as horizontal transmission, shared inheritance becomes the most economical explanation here as well.

The validity of a genealogical relationship between Plateau and Yok-Utian is independent of the question of a higher Penutian grouping including any or all of Takelma, Kalapuyan, the Oregon Coast languages, Chinookan, and Tsimshianic. Assuming that such a higher grouping is valid, the evidence that I have presented here suggests an Inland branch as a primary unit.

Scott DeLancey
University of Oregon


Abbreviations. C = consonant; PMC = Proto—Miwok-Costanoan (Callaghan’s Proto-Utian); PMi = Proto-Miwok; PMie = Proto—Eastern Miwok; V = vowel.

Sources of data. Data in the present article are taken from the following sources, unless otherwise specified: for Klamath, Barker (1963, 1964); for Nez Perce, Aoki (1970, 1994); for Plains Miwok, Callaghan (1984); for Proto-Utian, Callaghan (2014); for Proto—Yok-Utian, Callaghan (1997, 2001); for Proto-Yokuts, Golla (1964), Whistler and Golla (1986); for Sierra Miwok, Freeland (1951); for Yakima Sahaptin, Beavert and Hargus (2009), Jansen (2010).

Transcription. Forms presented in this article have been retranscribed into a standard Americanist orthography; length is indicated by a colon (:). In Yok-Utian languages, t (and its aspirated and glottalized counterparts) represents a dental stop, while ţ represents a postalveolar or retroflex stop. In Klamath data, following the practice of Barker (1963, 1964), the major source of data on the language, unaspirated stops are represented by voiced-stop symbols b d g G, and aspirated stops are represented by the voiceless-stop symbols p t k q (and c for the aspirated alveopalatal affricate).

1. This certainly also includes Nez Perce ʔipelí:kt ‘thundercloud’ (Aoki 1994:1057), which appears again in section 2.2, but since only two consonants correspond, this form does not belong in this list.

2. Morphophonemic symbols occurring in Klamath forms include underlined i and , which represent alternations between i and zero and between n and zero, respectively; underlined , representing alternations between intervocalic td and lengthening of a preceding consonant; and , representing an alternation between intervocalic glottal stop and glottalization of a preceding consonant (Barker 1964: 54, 55, 77, 93–95). (These alternations are more complex than indicated here, but the full details are not relevant for present purposes.) Also, in Barker’s system {} represents the morphemic level and || the morphophonemic.

3. Kramer assumes a priori that the languages are not genealogically related, so that these correspondences must represent horizontal transmission. However, since Kramer’s work, Callaghan (1997, 2001, 2014) has supplied many Yok-Utian cognate sets.

4. Barker’s underlying form for this morpheme is {’ok’}, with glottalized final k’. As far as I can see, there is no reason to think that the final velar stop of this morpheme is actually glottalized. Word-finally, Klamath allows only the aspirate series of stops, so Barker’s glottalization can only be based on forms in which k’ occurs before some other suffix. But only two suffixes occur following ’ok, and Barker analyzes both as having glottalization-inducing allomorphs—so including glottalization in an inferred underlying {’ok’} is unnecessary.


Aoki, Haruo
1962 Nez Perce and Northern Sahaptin: A Binary Comparison. International Journal of American Linguistics 28(3):172–82.
1963 On Sahaptian-Klamath Linguistic Affiliations. International Journal of American Linguistics 29(2):107–12.
1970 Nez Perce Grammar. University of California Publications in Linguistics 62. Berkeley: University of California Press.
1994 Nez Perce Dictionary. University of California Publications in Linguistics 63. Berkeley: University of California Press
Barker, M. A. R.
1963 Klamath Dictionary. University of California Publications in Linguistics 31. Berkeley: University of California Press.
1964 Klamath Grammar. University of California Publications in Linguistics 32. Berkeley: University of California Press.
Beavert, Virginia, and Sharon Hargus
2009 Ichishkíin sínwit: Yakama/Yakima Sahaptin Dictionary. Seattle: University of Washington Press.
Berman, Howard
1996 The Position of Molala in Plateau Penutian. International Journal of American Linguistics 62(1):1–30.
Broadbent, Sylvia, and Catherine Callaghan
1960 Comparative Miwok: A Preliminary Survey. International Journal of American Linguistics 26(4):301–16.
Callaghan, Catherine
1984 Plains Miwok Dictionary. University of California Publications of International Linguistics 105. Berkeley: University of California Press.
1997 Evidence for Yok-Utian. International Journal of American Linguistics 63(1):18–64.
2001 More Evidence for Yok-Utian: A Reanalysis of the Dixon and Kroeber Sets. International Journal of American Linguistics 67(3):313–45.
2014 Proto Utian Grammar and Dictionary: With Notes on Yokuts. Berlin: De Gruyter Mouton.
DeLancey, Scott
1991 Chronological Strata of Suffix Classes in the Klamath Verb. International Journal of American Linguistics 57(4):426–45.
1992 Klamath and Sahaptian Numerals. International Journal of American Linguistics 58(2):235–39.
1996 Penutian in the Bipartite Stem Belt: Disentangling Areal and Genetic Correspondences. Proceedings of the Twenty-Second Annual Meeting of the Berkeley Linguistics Society: Special Session on Historical Topics in Native American Languages. Berkeley: Berkeley Linguistics Society.
1999 Lexical Prefixes and the Bipartite Stem Construction in Klamath. International Journal of American Linguistics 65(1):56–83.
DeLancey, Scott, Carol Genetti, and Noel Rude
1988 Some Sahaptian—Klamath—Tsimshianic Lexical Sets. In In Honor of Mary Haas: From the Haas Festival Conference on Native American Linguistics, edited by W. Shipley, 195–224. Berlin: Mouton de Gruyter.
DeLancey, Scott, and Victor Golla
1997 The Penutian Hypothesis: Retrospect and Prospect. International Journal of American Linguistics 63(1):171–202.
Dixon, Roland B., and A. L. Kroeber
1919 Linguistic Families of California. University of California Publications in American Archaeology and Ethnology 16(3):47–118. Berkeley: University of California.
Freeland, Lucy S.
1951 Language of the Sierra Miwok. Supplement to International Journal of American Linguistics 17(1). Indiana University Publications in Anthropology and Linguistics, International Journal of American Linguistics Memoir 6. Baltimore: Waverly Press.
Gamble, Geoffrey
1991 Yokutian: Old California Penutian? Paper presented at the 1991 Hokan-Penutian Languages Workshop, University of California, Santa Cruz.
Golla, Victor
1964 Comparative Yokuts Phonology. In Studies in Californian Linguistics, edited by William Bright, 54–66. Berkeley: University of California Press.
1997 The Alsea—Wintuan Connection. International Journal of American Linguistics 63(1):157–70.
Hymes, Dell
1964 Evidence for Penutian in Lexical Sets with initial *c– and *s–. International Journal of American Linguistics 30(3):213–42.
1983 Victoria Howard’s ‘Gitskux and His Older Brother’: A Clackamas Chinook Myth. In Smoothing the Ground: Essays on Native American Oral Literature, edited by Brian Swann, 129–70. Berkeley: University of California Press.
Jacobs, Melville
1931 A Sketch of Northern Sahaptin Grammar. University of Washington Publications in Anthropology 4(2). Seattle: University of Washington Press.
Jansen, Joanna
2010 A Grammar of Yakima Ichishkíin / Sahaptin. Ph.D. diss., University of Oregon.
Kramer, Marvin
1995 Yokuts as a Target Language in a Shift from Miwok. In Proceedings of the Twenty-First Annual Meeting of the Berkeley Linguistics Society, 438–49. Berkeley: University of California.
Newman, Stanley S.
1944 Yokuts Language of California. Viking Fund Publications in Anthropology 2. New York: [n.p.].
Rude, Noel
1987 Some Klamath-Sahaptian Grammatical Correspondences. Kansas Working Papers in Linguistics 12:67–83. Lawrence: University of Kansas, Department of Linguistics.
1996 The Sahaptian Inflectional Suffix Complex. In Proceedings of the Hokan—Penutian Workshop: July 8–9, 1994, University of Oregon, Eugene, and July 5–6, 1995, University of New Mexico, Albuquerque, edited by Victor Golla, 51–89. Survey of California and Other Indian Languages Report No. 9. [Berkeley, Calif.] : Survey of California and Other Indian Languages.
1997 On the History of Nominal Case in Sahaptian. International Journal of American Linguistics 63(1):113–43.
Sapir, Edward
1990a The Problems of Linguistic Relationship in America. In The Collected Works of Edward Sapir. Vol. 5: American Indian Languages 1, edited by William Bright, 83–91. Berlin: Mouton de Gruyter. (Originally published in 1920.)
1990b Central and North American Languages. In The Collected Works of Edward Sapir. Vol. 5: American Indian Languages 1, edited by William Bright, 95–104. Berlin: Mouton de Gruyter. (Originally published in 1929.)
Shipley, William
1966 The Relation of Klamath to California Penutian. Language 42(2):489–98.
Whistler, Kenneth
1977 Wintun Prehistory: An Interpretation Based on Linguistic Reconstruction of Plant and Animal Nomenclature. Proceedings of the Third Annual Meeting of the Berkeley Linguistics Society, 157–74. Berkeley: University of California.
Whistler, Kenneth, and Victor Golla
1986 Proto Yokuts Reconsidered. International Journal of American Linguistics 52:(4)317–58.

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.