Werner 2002: II, 380-381; Werner 1993: 113. Quoted as uˑsʸ1 ~ uːsʸ3, pl. uˑsʸ-eŋ1 in [Werner 1977: 188]; as uːsʸ in [Castrén 1858: 166]. Lexically distinct from aˑ 'hot / heat' [Werner 2002: I, 92].
Werner 2011: 330. Plural form: ˈus-eŋ. Quoted as uˑs1 ~ uːs3, pl. us-eŋ1 in [Werner 1977: 188]. Lexically distinct from ˈafɨŋ 'hot' [Werner 2011: 170] (although the latter word is also sometimes translated as 'warm').
KYU_NOTES:
Proto-KY *ʔuˑs ~ *ʔuːs 'warm' (free variation between long and semi-long vowels is not easily explicable), opposed to *ʔap- 'hot'.
Castrén 1858: 224. Meaning glossed as 'hot, warm'; it remains unclear whether the two meanings were or were not distinguished lexically in Kott. Cf. in older sources: pal-tu '(it is) warm') (Kh.) [Verner 1990: 378] (-tu is the 3rd p. predicative suffix).
Dulzon 1961: 167 (Pal.). Also attested in (Dict.) in the extended variant ˈuttɨči-dˈin. Meaning glossed as Latin 'calidus'. The quasi-synonymous form afam (Pal.) is really Yugh 'hot'.
YEN:*xus-
S. Starostin 1995: 299. Alternately reconstructed as *usǝ ~ *utʸǝ ~ *kusǝ ~ *kutʸǝ in [Werner 2002: II, 380-381]. Distribution: Preserved in all daughter languages, with the probable exception of Kott. Replacements: In Kott, there may have been a merger of the lexically distinct Proto-Yeniseian meanings 'warm' and 'hot': Kott fal ~ pʰal is compared both by S. Starostin and H. Werner to such forms as Yugh aːp, Ket aː 'heat', Yugh af-ɨŋ, Ket aː-ŋ 'hot', etc. < Proto-Yeniseian *ʔap- 'hot' [S. Starostin 1995: 182; Werner 2002: I, 91-92], with the somewhat weakly founded assumption of development from *ʔap-al (neither the suffix nor the word-initial vowel reduction are well explained, but the word has no alternate etymology). Reconstruction shape: Correspondences are regular; initial *x- is reconstructed on the basis of k- in Arin. Final -či in Pumpokol is equivalent to the predicative suffix -sʸ in Ket-Yugh. Semantics and structure: The semantic opposition *xus- 'warm' : *ʔap- 'hot', best attested in Ket-Yugh, is quite probably of Proto-Yeniseian origin.
Dulzon 1961: 160 (Dict., Pal., Kl.). There is also a strange quasi-synonymous form dok 'water' (Pal., Kl.) with no supporting parallels whatsoever [ibid.].
YEN:*xur1
S. Starostin 1995: 298. Alternately reconstructed as *(k)uʎ ~ *(k)uʎǝ in [Werner 2002: II, 378]. Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular. Initial *x- is reconstructed based on the velar refelxation k- in Arin. For the difference between 'water' and 'rain', see notes on 'rain'.
Werner 2002: II, 419; Werner 1993: 129. This form is clearly connected with at 'I', but cannot be derived from it synchronically. Cf. also the possessive pronoun ˈǝtn-a 'our' [ibid.]. Quoted as ǝˑt1 / ǝˑt1 ~ ǝˑtn1 (S.-Imb.) in [Werner 1977: 173]; as ɜtn in [Castrén 1858: 48].
Werner 2011: 342. This form is clearly connected with at 'I', but cannot be derived from it synchronically. Cf. also the possessive pronoun ǝn-na (< *ǝt-n-da) 'our' [Werner 2011: 313]. Quoted as ǝtn1 in [Werner 1977: 173].
Dulzon 1961: 173 (Dict.). The quasi-synonymous form etn-ɨn 'we' (Pal., Kl.) is really Yugh, not proper Pumpokol.
YEN:*ʔaʒ-ǝŋ
S. Starostin 1995: 185 (forms quoted in association with *ʔaʒ 'I' q.v.). Alternately reconstructed as *ad-ǝŋ in [Werner 2002: II, 419]. Distribution: Preserved in all daughter languages. Reconstruction shape: The Proto-Yeniseian equivalent for 'we' was clearly the regular plural form of 'I'; hence, see notes on 'I' for reconstruction peculiarities. The plural marker was probably *-ǝŋ (*ʔaʒ-ǝŋ), with contraction and assimilation in Ket-Yugh (> *ʔǝʒ-ŋ > *ʔǝʒn > *ʔǝtn).
Werner 2002: I, 66; Werner 1993: 13. This default inanimate object pronoun may be used both independently (e. g. aksʸkˈu=b=bet "what are you doing?") and as a "modifier" in the incorporating verb 'to do what?', e. g. d=ˈakusʸ-iˑ-vet "what am I doing?". The earlier source of [Castrén 1858: 51] lists the following forms: assa ~ ai 'what?' (at least the former is probably Yugh), aːkusʸ ~ aːku 'what then'?
Werner 2011: 332. Cf.: u ˈassa ku=b=betʸ "what are you doing?" Quoted as assa in [Castrén 1858: 51].
KYU_NOTES:
Ket ak/u/sʸ is most likely the result of a metathesis from *ʔas-k; together with Yugh assa, where, on the contrary, an assimilation may be supposed, both are traceable back to Proto-KY *ʔas-ka 'what?', itself an extended variant of simple *ʔas 'what? (adj.)', as in Ket asʸ kɛʔt 'what (kind of) man?' [Werner 2002: I, 66].
Castrén 1858: 55. The sequence -na is tentatively segmented out as a suffix due to (most likely) the same root morpheme *ši in a=ši-x 'who?' q.v.
Arin:
Not attested.
Pumpokol:
Not attested.
YEN:*si ~ *ʔa=si
S. Starostin 1995: 183 (*ʔas- / *sV-). Distribution: Preserved in all daughter languages where attested, but not found in Arin or Pumpokol. Reconstruction shape: The Proto-Yeniseian morpheme clearly had the fricative *-s- as its main distinctive component, but the vocalic "framing" differs in between Ket-Yugh and Kott and is hard to reconstruct convincingly. The hypothetical variant *ʔasi surmises reduction of the second syllable in Ket-Yugh (> *ʔasʸ) and of the first syllable in Kott (*ʔasi-na > *si-na > ši-na). However, it is more probable that *ʔa= was a separate prefixal morpheme, considering that it does not elide in Kott ašix 'who?' (if this were a strictly phonetic process, one would expect reduction in both of the interrogative pronouns).
Werner 2002: II, 249; Werner 1993: 94. The suffix -ɨm ~ -am is an adjectival formant, but the word is not properly segmentable on the synchronic level (there is no separate noun from which it could be derived; see, however, notes on Common Ket-Yugh for additional discussion). Quoted as taɣam5 in [Werner 1977: 179]; as taum in [Castrén 1858: 175].
Werner 2011: 335. Transparently derived from tik 'snow' with the prosecutive suffix -bɛʰːs (cf. the same model in the formation of 'red' q.v.). Cf. also the verb: tig-ey 'to become white' [ibid.]. Quoted as tig4-bɛʰːs in [Werner 1977: 179]; as tig-bes in [Castrén 1858: 177].
KYU_NOTES:
This case is extremely similar to that of 'red' q.v.: just as 'red' is formed in both languages from the noun 'blood' with, respectively, the adjectival formant -ɨm ~ -am in Ket and the prosecutive suffix -bɛʰːs in Yugh, so the equivalent for 'white' can be formed in the exact same way from the word for 'snow': Ket tiˑk, Yugh tik [Werner 2002: II, 269]. There is, however, a significant problem here, namely, the vocalism in Ket, where *tiɣam or *tiɣɨm would be expected instead of the actual taɣ-am ~ taɣ-ɨm. There is no way to account for this discrepancy in vocalism, and since it is the Ket rather than the Yugh form that is further supported by external data, at this point it may only be suggested that Yugh tig-ey, tig-bɛʰːs is an innovation.
Castrén 1858: 218. Formally, could be analyzed as an adjectival derivate from tʰiːk 'snow' [Castrén 1858: 219]. Cf. in older sources: tegama 'white' (M., Dict., Pal., Kl.), inka=tekam-a 'it is white' (Kh.) [Verner 1990: 286].
Dulzon 1961: 158 (M., Kl.). Quoted as tʸaːma in (Dict., Pal.); as tamo ~ tama 'is white', berik=tam-tu 'is (very) white' (Kh.) in [Werner 2002: II, 249].
S. Starostin 1995: 282 (*täk-). Alternately reconstructed as *tʰeg/am/ in [Werner 2002: II, 249]. Distribution: Preserved in all daughter languages, but morphologically restructured in Yugh. Reconstruction shape: The situation here is complicated. Most of the attested forms can rather easily be analyzed as derived from Proto-Yeniseian *tik (= S. Starostin's *tiχ) 'snow' > Ket tiˑk, Yugh tik, Kott tʰiːk, Arin teː, Pumpokol tɨg [S. Starostin 1995: 285] with the aid of the productive derivational suffix *-Vm. This scenario is further corroborated by the absolutely identical situation with the adjective 'red' q.v., derived in the exact same way from Proto-Yeniseian 'blood'. The only fact that goes directly against such a scenario is the root vocalism in Ket: tˈaɣ-ɨm instead of the expected *tiɣ-am. This is such a serious argument that it caused S. Starostin to separate Proto-Yeniseian *tiχ 'snow' from *täk- 'white', and then look for separate external comparanda for each of these roots. However, there is an alternate explanation: namely, that Ket taɣ-ɨm actually reflects a different "Ablaut" grade of *tik 'snow' (parallel to such Proto-Yeniseian variants as *čip ~ *čaʔp 'dog', *ʔes ~ *ʔas 'sky', etc.). In fact, the same grade of vocalism may be implicitly present in the contracted Arin and Pumpokol forms as well (i. e. it is possible to regard Arin taːma, Pumpokol tamxo as contracted from *tak-am rather than *tik-am), while Kott may have restructured the vocalism by analogy with the noun 'snow' - and Yugh has carried the restructuring even further, by replacing the old suffix with a new one. Whatever be the case, it seems that Ket vocalism simply does not constitute sufficient evidence for postulating an extra Proto-Yeniseian root; the more economic solution is to propose the derivation {'snow' > 'white'} already at an early stage of Proto-Yeniseian.
Werner 2002: I, 134; Werner 1993: 25, 28. Masculine and feminine variants respectively. According to E. Vajda, these forms, "...being gender-specific, can be used in rhetorical questions or when the speaker knows the gender of the person asked about" [Vajda 2004: 31]. Quoted as bisʸsʸe5 / bitsʸǝ5 ~ bɛtsʸǝ5 (masc.), bɛsʸa6 / bɛsʸɛ6 (Kur.) (fem.) in [Werner 1977: 139].
Werner 2011: 337, 338. This stem forms the feminine gender interrogative pronoun (ˈasɛ-ra) and the plural form of the interrogative pronoun, indifferent to gender (asˈe-in). Quoted as asɛra5 'who (of woman)' in [Werner 1977: 135].
KYU_NOTES:
Proto-KY *ʔan- 'who?' (the coda is difficult to reconstruct because of the multiple variants). The other forms look like local innovations: Ket bitsʸe, etc. is formed from the general interrogative stem bi-, which is most frequently used in the formation of adverbial interrogatives in Yeniseian languages [S. Starostin 1995: 294], and Yugh asɛ-ra, ase-in are formed from the second general interrogative stem *ʔas-, discussed in more details under 'what?' q.v.
Castrén 1858: 55. Plural form: a=ši-g-an. Possible internal analysis of the form: a= is a common Yeniseian pronominal prefix (same as a= in Ket *ʔa=s- 'what?', *ʔa=n- 'who?', etc.); =ši- is the main interrogative morpheme (same as in ši-na 'what?' q.v.); -x is the final suffix that conveys the meaning of animacy, possibly < ig 'male' [Castrén 1858: 200] (this would mean that the form originally referred only to the masc. gender).
Arin:
Not attested.
Pumpokol:
Not attested.
YEN:*ʔan- #
S. Starostin 1995: 181. Distribution: Preserved only in Ket-Yugh. Replacements: Apart from Ket-Yugh, the animate interrogative pronoun is attested only for Kott (ašix), where it shares the same root with the inanimate pronoun (ši-na) and has a probable internal etymology: < *a=ši-ig, lit. 'what male?' (not secure, since the final velar consonant may have other origins as well). In Ket-Yugh itself, both Ket and Yugh show two sets of equivalents for 'who?', one of them common and probably inherited from the common ancestor (*ʔan-), the other one different in the two languages and probably innovated (see notes on Ket-Yugh). Tentatively, we choose *ʔan- as the original 'who?', since there are no reasonable scenarios of its secondary origin. The overall confusion may have been caused by a tendency to reform the original system of interrogative pronouns by introducing new "gender-sensitive" stems, formed from alternate interrogative morphemes (*si, *bi, etc.).
Werner 2002: I, 34; Werner 1993: 16. Free variants. This is the gender-less interrogative pronoun, used in situations where there is no need to indicate masculine or feminine sex of the referent. Accordingly, we treat bˈitsʸe and ˈana as synonyms. Quoted as aney6 ~ ana6 (S.-Imb.) in [Werner 1977: 134]; as anet ~ ana ~ anasʸ in [Castrén 1858: 51].
Werner 2002: II, 90; Werner 1993: 65. Feminine gender. Plural form: qiˑm-n {қимн}. Quoted as qiˑm1, pl. qiˑm-n1 in [Werner 1977: 161]; as qiːm ~ qim, pl. qiːm-eŋ in [Castrén 1858: 170].
Werner 2011: 141. Feminine gender. Plural form: χem-n ~ χim-n. Quoted as χem1 ~ χim1, pl. χem-n1 ~ χim-n1 in [Werner 1977: 161]; as xim, pl. xiːm-en in [Castrén 1858: 172].
Castrén 1858: 196. Plural form: alit-n ~ aliːt-n. It should be noted that al= could quite possibly be the same fossilized prefix here as in 'dog', 'bird', 'star' q.v. However, since the word has no parallels in Ket-Yugh, this time there is no external evidence to justify this segmentation. Cf. in older sources: alat 'woman', alit 'wife' (Kh.); alˈit 'wife' (M., Dict., Pal.) [Verner 1990: 310].
Werner 2002: I, 25. Most of the sources only record the word 'wife' for Arin, which seems to be a compound form: bɨ=qam ˈalte (M.), bi=qam-ˈalte (Dict.), bi=qam-al (Pal.) [Dulzon 1961: 167], where bɨ= ~ bi= is the possessive prefix 'my', =qam- corresponds to Ket qiˑm, etc. 'woman', and -al ~ -alte is to be equated with Kott alit 'woman' q.v. Only in (Kh.), next to the form kek melte 'wife' (= ke=km-elte 'your wife'?), we also find kem-elʸa 'woman'. Since Proto-Yeniseian seems to have had a firm lexical distinction between 'husband' / 'man' and 'wife' / 'woman', the Arin forms, for the most part, seem like collocations of both terms (i. e. 'wife' = 'wife-woman').
Pumpokol:
Not attested. Cf. ils=ˈem (Dict.), ilz=em (Pal.) 'wife' (where ils= is a semantically obscure component, also attested in ils=ˈet (Dict., Kl.), ilz=et (Pal.) 'husband') [Dulzon 1961: 167, 173].
YEN:*qem
S. Starostin 1995: 266 (*qVm). Distribution: Preserved in Ket-Yugh and, most likely, in Arin; possibly also in Pumpokol, if the word for 'wife' in that language had the same root as 'woman'. Replacements: In Kott-Arin, there is another stem for the meanings 'woman' and 'wife', functioning on its own in Kott (alit) and as part of a compound with the older word for 'woman' in Arin (*qem-alit, with various assimilations and reductions in the actual attested dialectal forms). There are no parallels for this *ʔalit in Ket-Yugh, and it is not clear why Arin turned the old word into a compound, and Kott retained only the newer part of this compound, but from the point of view of cognate distribution, this is the most economic scenario. Reconstruction shape: Consonantal correspondences between all the languages are fully regular (in Pumpokol, the root-initial uvular is lost inside a compound formation: *ils-qem > ilsem). Reconstruction of the vocalism is less secure: *qem seems like the optimal variant, as it is attested in both Yugh and Pumpokol, but the discrepancy between *qem and *qim in Ket-Yugh lacks an explanation so far.
Dulzon 1961: 167 (Dict.). Somewhat dubious, since this is the same word as 'red' q.v.
YEN:
Not reconstructible: the Ket word is transparently derived from 'gall', the Kott word is the same as 'moon', the Pumpokol word is the same as 'red', and the Arin word has no etymological parallels whatsoever.
Werner 2002: I, 125; Werner 1993: 26. Adverbial form; the predicative variant is bilʸ-am. Quoted as bilʸ4 (S.-Imb.) / biːlʸ4 ~ biːlʸ3 (N.-Imb.) in [Werner 1977: 139]; as biel ~ biel-ɛs ~ bielʸ in [Castrén 1858: 189].
Werner 2002: I, 125. Attested only in (Kh.), as an adverbial form; -ta seems to be a segmentable suffix.
Pumpokol:
Not attested.
YEN:*biːr1
S. Starostin 1995: 211. Alternately reconstructed as *biʔǝʎǝ in [Werner 2002: I, 125]. Distribution: Preserved in Ket-Yugh and Kott; preservation in Arin is under serious doubt. Replacements: Arin pa-ta 'far' is regarded by both S. Starostin and H. Werner as cognate with Ket-Yugh *biːr and Kott piːl, but this is a problematic judgement. Even if the segmentation into pa-ta is correct (cf. u-ta 'long' which seems to corroborate this assumption), vocalic and consonantal correspondences are not easily confirmed: in particular, elision of the root resonant seems quite suspicious, since Arin usually preserves resonants in clusters. We prefer to count the Arin form as a lexical replacement for the moment, albeit without any reasonable etymology. Reconstruction shape: Correspondences between Ket-Yugh and Kott are quite regular.
Werner 2002: II, 221; Werner 1993: 92. The non-predicative form is sʸɜː. Quoted as sʸɜː-sʸ3 (predicative), sʸɜː3 (attributive) in [Werner 1977: 177]; as sɜː ~ sʸoaga in [Castrén 1858: 186].
Werner 2002: II, 221. Attested only in (Kh.) as an adverbial (or predicative) form ('it is heavy').
Pumpokol:
Not attested.
YEN:*sǝɢ-
S. Starostin 1995: 273. Alternately reconstructed as *sǝkǝ ~ *sǝgǝ in [Werner 2002: II, 221]. Distribution: Preserved in all daughter languages, but not attested in Pumpokol. Reconstruction shape: Consonantal correspondences are completely regular; vocalic reconstruction is approximate, but generally indicative of Proto-Yeniseian *ǝ.
Werner 2002: II, 371; Werner 1993: 115. Final -sʸ here is not the usual predicative suffix, since it never disappears in attributive forms, cf. ˈutisʸ baʔŋ 'nearby land', 'vicinity'. Quoted as uˑtisʸ1 in [Werner 1977: 189]; as utɛs in [Castrén 1858: 165]. The word should not be confused with ɨˑlʸ ~ ɨlʸga 'near' = 'close to', 'beside' (smth. or smbd.) [Werner 2002: II, 434-435]).
S. Starostin 1995: 201 (*ʔut-). Distribution: Preserved only in Ket-Yugh. Replacements: In Kott, PY *ʔuti 'near' is preserved in the adverbial form uti-ga 'here' (directional) [Castrén 1858: 204], i. e. {'near' > 'here'}. Kott iːma-ŋ 'near' is compared by S. Starostin with Ket im-da 'small; (to grow) thickly, densely' [YED # 100], implying the semantic shift 'small (gen.)' > 'small (of distance)' > 'close, near'. The plausibility of such a shift may be put under doubt, but this is currently the best etymology for the Kott word anyway. Reconstruction shape: Lack of parallels in Arin means that the Proto-Yeniseian equivalent of the Ket-Yugh forms could have been *ʔuti or *xuti. Semantics and structure: Although final -sʸ is not segmentable on the Ket-Yugh level, it is not likely that it could have constituted part of the root on the Proto-Yeniseian level, for general structural reasons.
Werner 2002: II, 301; Werner 1993: 104. Neuter gender. Quoted as tɜʔ2 in [Werner 1977: 183]; as tʸɜʔɛ ~ tʸaʔa ~ tɜʔa, pl. tʸɜʔɛ-n in [Castrén 1858: 178].
Castrén 1858: 214. Plural form: šinčeːt-aŋ. Phonotactic considerations and external comparison suggest the analysis of this form as a compound. Cf. in older sources: šinčet (M., Pal., Dict., Kl.), šinšet (Kh.) [Verner 1990: 370].
Arin:tus-1
Dulzon 1961: 183 (M., Dict., Pal., Kl.). A transparent Turkic borrowing (< Common Turkic *tuz 'salt).
Pumpokol:tus-1
Dulzon 1961: 183 (Dict.). A transparent Turkic borrowing (< Common Turkic *tuz 'salt). The other quasi-synonymous form, če (Pal., Kl.), is most likely Yugh rather than proper Pumpokol.
YEN:*čǝʔ
S. Starostin 1995: 216. Alternately reconstructed as *tʸǝʔ < *tʸǝgǝ in [Werner 2002: II, 301]. Distribution: Preserved in Ket-Yugh and in Kott (as part of a compound). Replacements: Arin and Pumpokol tus 'salt' are borrowed from Turkic; H. Werner's attempts to relate them to Ket-Yugh and Kott forms are unnecessary. Reconstruction shape: Correspondences between Ket-Yugh *čǝʔ and Kott ši- are quite regular. Semantics and structure: Kott šinčeːt is not a particularly transparent form. Its first part may be the original plural form of 'salt' (i. e. ši-n = Ket tʸɜʔɛ-n in Castrén's notation), but the component -čeːt has no reasonable etymology. There is also an alternate possible analysis, completely different: šinčeːt = *ši-ŋ ~ še-ŋ 'stones, rocks' q.v. + *čǝʔ or maybe *čǝʔč 'salt'; the main obstacle here is the final consonant, which then finds no equivalent in Ket-Yugh and whose disappearance there has to be ascribed to an irregular / sporadic process (e. g. an original *čǝʔč with dissimilation?). In any case, there are multiple possible scenarios that allow to etymologize the Kott word on the same basis as the Ket-Yugh item, but no clear preference for an optimal one.
Werner 2002: I, 324; Werner 1993: 120. The non-predicative form is hɔʔlʸ. Plural form: hˈɔlʸ-aŋ {холяӈ}. Quoted as hɔʔlʸ2, pl. hɔlʸ-ǝŋ5 in [Werner 1977: 191]; as hoːli in [Castrén 1858: 174].
Not reconstructible: Ket-Yugh *pɔʔl 'short' and Kott tʰuːki (< *tuk- ?) have more or less equal chances at representing the Proto-Yeniseian item. Arin kamara is slightly more suspicious: in [YED # 100], it is tentatively compared by S. Starostin with Ket im-da 'small', Kott imɨraŋˈaga ~ imgara id. Vocalic correspondences are suspicious, but if kamara is indeed a "corrupt" variant of *kimara, it is then easily comparable with imgara < Kott-Arin *xim-gara, and in this case, the semantics of 'short' for Arin is most likely secondary.
Werner 2002: II, 267; Werner 1993: 98. Masculine gender. Plural form: tˈiɣ-inʸ. Quoted as tiɣ4 (S.-Imb.) / tiːɣǝ4 (N.-Imb.), pl. tiɣ-inʸ5 ~ tiˑɣ-inʸ5 in [Werner 1977: 181]; as tieɢ, pl. tieɢ-en in [Castrén 1858: 176].
Castrén 1858: 202, 252. Plural form: oŋxo-n ~ oŋxot-n. The second part in this compound formation is clearly hoi 'worm' q.v.; the first part remains unclear. Cf. in older sources: onxoy (Kh.) [Verner 1990: 315].
Werner 2002: II, 47. Attested only in (Kh.). The structure of this compound is exactly the same as in Kott q.v.
Pumpokol:
Not attested.
YEN:
Not reconstructible. The original meaning of Proto-Ket-Yugh *čiːk, considering the external, evidence and distribution of cognates, must have been 'fish' q.v. As for Kott-Arin *ʔaŋ-koy, it is clearly a composite formation where the second component is *koy 'worm' q.v.; the first component is tentatively equated by H. Werner with *ʔaŋ 'rope' [Werner 2002: II, 47], thus, 'rope-worm'? (this is by no means a finalized etymology). In both cases, it seems as if the original Proto-Yeniseian form for 'snake', whatever it might have been, has undergone different paths of "tabooization" in Ket-Yugh and Kott-Arin.
Werner 2002: I, 293; Werner 1993: 116. Said of flat objects (paper, leaves, bread, etc.). Quoted as haksʸ-em5 in [Werner 1977: 189]; as haːqs-em in [Castrén 1858: 173].
S. Starostin 1995: 245 (*pak-si-m). Alternately reconstructed as *phaksǝm in [Werner 2002: I, 293]. Distribution: Preserved in all daughter languages where attested, but not found in Arin or Pumpokol. Reconstruction shape: Correspondences are regular, although vocalism of the second syllable is hard to determine. Semantics and structure: In Proto-Yeniseian, as in attested languages, the word must have been applied to flat objects. S. Starostin's morphological segmentation of the stem into *pak-si-m is conditioned by external comparison; Yeniseian-internally, *-m is indeed a derivational suffix, but *pakse- (or *paksi-) functioned as a monolithic stem already in Proto-Yeniseian.
Werner 2002: II, 276; Werner 1993: 99. Meaning glossed as 'schmal, eng'; applied to 'snowshoes', 'ropes', i. e. 3D-objects. Quoted as tɔʁǝ5 in [Werner 1977: 182]; as toɢalʸa ~ toːɢalʸa ~ toɢo in [Castrén 1858: 177] (the first two forms contain an extra suffix and are probably Yugh).
Castrén 1858: 217. The difference between fačam and tʰaːge 'thin' is not made clear in Castrén's description, but it is quite likely that it was the same as in the etymologically related Ket-Yugh pair.
Arin:
Pumpokol:
YEN:*tɔq-
S. Starostin 1995: 287 (*tɔqV-). Distribution: Preserved in all daughter languages where attested, but not found in Arin or Pumpokol. Reconstruction shape: Consonantal correspondences are regular; vocalism of the second syllable is not reconstructed with any degree of certainty. Semantics and structure: In Proto-Yeniseian, as in attested languages, the word must have been applied to 3D-objects.
Werner 2002: I, 122-123; Werner 1993: 24. Neuter gender. Plural form: bey-eŋ. Quoted as beˑy1 in [Werner 1977: 138]; as bei, pl. bey-eŋ in [Castrén 1858: 189].
Dulzon 1961: 160 (Dict., Pal.). Quoted as boi in (Kl.).
YEN:*bey
S. Starostin 1995: 208. Alternately reconstructed as *bay in [Werner 2002: I, 122-123]. Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are fully regular.
Werner 2002: II, 371; Werner 1993: 115. Masculine gender. Plural form: ˈuti-n {утин}. Quoted as utiɣ6 ~ utiɣ5, pl. utin5 (S.-Imb.) in [Werner 1977: 189]. A highly localized word, not encountered beyond the Southern area. Less viable candidates include (a) kɨnʸsʸ 'caterpillar, earthworm', a rare, not too well confirmed word that is completely homonymous with kɨnʸsʸ 'Russian' (!) [Werner 2002: I, 477]; (b) kiˑnʸ, pl. kînʸ ='worm' [Werner 2002: II, 436], which is actually 'maggot' rather than the required 'earthworm'.
The word utˈiɣ is clearly a composite formation: the second part is easily identifiable as tîɣ 'snake' q.v. The first part is more problematic; possibly = uː 'meadow' [Werner 2002: II, 376] (i. e. 'worm' as 'meadow-snake').
Werner 2011: 344. Feminine gender. Polysemy: 'worm / small insect'. Plural form: ˈɔllɨ-n ~ ˈɔlɨ-n ~ ˈɔlɨ-ŋ. Quoted as ol, pl. ol-en ~ ol-an in [Castrén 1858: 163].
KYU_NOTES:
Not easily reconstructible. All of the Ket forms seem like innovations, whereas Yugh ɔllɨ ~ ɔʔl is polysemous, and only corresponds to the first component in Ket ɔlǝŋgǝs 'spider' [Werner 2002: II, 48]. Still, based on both internal and external evidence, *ɔʔl is currently the best candidate for Proto-KY 'worm'.
S. Starostin 1995: 242 (*kVy). Distribution: Preserved only in Kott (although the Arin and Pumpokol equivalents are simply not attested). Replacements: (a) In Ket, replaced in the meaning 'worm' by utˈiɣ, a compound of 'snake' with an unclear first component (see notes on the Ket form); (b) in Yugh, replaced by ˈɔllɨ 'worm / small insect', cognate with Ket ɔlǝŋgǝs 'spider' [Werner 2002: II, 48], indicating a more generic term than simply 'worm'. Reconstruction shape: The reconstruction depends almost exclusively on the Kott form, meaning that reconstruction of the vocalism is quite approximate. Semantics and structure: Kott hoy, pl. ho-n is cognate with Ket kiˑnʸ 'maggot, larva', reflecting the Proto-Yeniseian paradigm *koy, pl. *koy-n. In Ket, the latter form was generalized as a (collective) singular and contracted.
Werner 2002: II, 223; Werner 1993: 92. Neuter gender. Plural form: sʸɨk-ŋ {сыкӈ}. Quoted as sʸɨː3, pl. sʸɨk-ŋ5 in [Werner 1977: 178]; as sɨ, pl. sɨk-ŋ in [Castrén 1858: 187].
S. Starostin 1995: 275. Alternately reconstructed as *sǝ́gǝ in [Werner 2002: II, 223]. Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are generally regular, although vocalic reconstruction in both syllables is somewhat questionable. Semantics and structure: S. Starostin has proposed that *-ɢa is an old suffixal element, present also in such words denoting time as *si-ɢ 'night' q.v., *xiʔ-ɢ 'day' (see under 'sun'), but this is questionable.