In this group of cases the corresponding words are found in Tungus, but for various reasons they cannot be maintained for the parallels.
Case 1. G. Ramstedt's parallels and hypothetic stem *pY or *pY in Finno-Ugrian means «tree,» in Tungus «birch-tree,» etc The equation seems to confirm the existence of the same stem in Goldi and Manchu ba, ba (Goldi) which is fa (Manchu). Theoretically, if we suppose this stem to be a Tungus one, it may be reduced to ja, and even a, ae, unless it is borrowed by the Manchu in the form pa from some other language; e. g., Paleasiatic [102]. However, such a stem with the meaning of «birch» and «tree» is unknown in Northern Tungus. The «birch» is chalban (Bir., Kum., Khin., RTM, Ner., Barg., Turn., Lam.), etc.; cf. also chalfa (Manchu Writ.) with a special meaning—«the raw birch-bark.» The other parallels are inconsistent; namely, hijika (W)—«wald,» and hijela (W)—«im dickicht.» [103] Both of them are well-known words in this particular dialect being modified, namely, siyi (RTM, Khin., Ner.), sigi (Neg., Sch.), hegi (Turn.) and even egi (Lam.), whence siji with the alteration s->h hiji (W). The suffixes are -ka and -la, the first one perhaps diminutivus [104] and the second one directivus. These words are invariably used for «dense, thick forest,» «thickness in forest,» etc., occupying a usually limited area and consisting usually of young larch and cedar («spreading») trees. This cannot refer to the birch-tree forest, for the birch-tree never forms such a thick forest, yet it cannot refer to the «forest» in general. Being different as to the meaning (thickness) and stem (sigi), it cannot be connected with pYY and pY. The only possible connexion is Manchu fa, which is not perhaps a Tungus word at all, for it is confined to the Manchu Writ, only, whence borrowed by the Goldi.
Case 4. G. Ramstedt's stem *pYiYngY *pYlYngY, which is compared with the Finno-Ugrian stem *pYngY—«the palm of the hand.» The Tungus parallels are falango (Manchu Writ.) (the spelling is rather complex), paina, xanga, hanga quoted by A. Sauvageot and compared with Mongol halaqan (P. Pelliot), alaya, etc. However, the situation is not so simple as that. Other Tungus dialects give us two series; namely, ongan, angachan (Bir.), anga (Neg., Sch), by the side of xanga (Neg., Sch.), xaqi (Turn.), which can be reduced to the stem anV or even ang — in some dialects aspirated, in some other dialects bilabialized. Another series is aliya (Mank.), aliga (Mank., Castr.), probably Manchu falango (cf. aliy of Buriat, Castr.), which ought to be compared with the Mongol stem alay, etc. So that if we leave aside the hypothesis of the loss of p, which is a mere hypothesis, then we have to deal with two stems angV and alVg. To identify these two stems as one and the same, it would mean a building up of a new hypothesis; namely, n~l, or l~0, etc. Of course, it is not impossible, but it must be supported by solid evidences. On the other hand, the Finno-Ugrian series has words with the initial p forming the first syllable with i or e, and the second syllable varying between v and ng reduced to the stem pYngY. Owing to the fact that this reconstruction and connexion is possible only on the condition of adopting several hypotheses, I think that the common stem, if it ever existed, is not definitely shown.
Case 5. A. Sauvageot produces the stem *pYrY, i.e., always the same as in Cases 3 and 6, this time with a broad meaning «toute armature de bois.» Naturally, the finding of words is not difficult at all and so they are found in all possible languages (e.g., in the sense «grenier,» «combles d'une maison,» «schwimmende insel,» «runder hasten,» «floss am netz,» «sledge,» «island,» «archipelago,» «steppeninsel,» «groove.» etc.). In this series are found fara (Manchu Writ.), para (Goldi, W. Grube) (Tungus, Gerstf.), the latter being naturally borrowed from Manchu. This case is rather analogous to Case I, for in various Tungus languages different words are found, e.g., tolgoki (Ner., Mank.) (Irk., Tit.), tolgoki (Neg., Sch.) [tuki (Or., Sch.), tuli (Olcha, Sch.), toki (Goldi, Sch.), which P. P. Schmidt compares with Gilak tu]. Yet there are several other words for designating the reindeer sledge; e.g., nolima (Lam.), sirga (Turn.) (cf. Yakut syrga, syarya, Pek. compared with sor, sorga, charga, and chirga, some of Turk groups and some of Mongol groups). If the Gilak word tu is a contracted form of tuki, etc., and the latter is a contracted form of tolgoki, etc., it is not perhaps a Tungus stem, but a Paleasiatic (Gilak) one. Some other suggestions can be made, but since they are not sure I will leave the question where it is. Let us remark that the sledge is an element confined to the geographical areas where there is good snow and draught animals. Not all Tungus use the sledge, and amongst some of them the method of harness is the same as that used for dogs. Yet the words compared, fara~para, are found only in Manchu and in dialects influenced by Manchu, so the Goldi possesses two terms. Since the Manchu word is connected through a series of extremely extensive semantic operations, and since no such word is known in other Tungus languages and the parallels may be accepted only on the agreement with G. Ramstedt's hypothesis, I consider that these parallels are not convincingly shown.
Case 10. The stem *pYr- means «black» in Samoyed. The Tungus parallels quoted are farxun (Manchu Writ.) — «dark,» and a series of words of the stem akt — «twilight,» well known in most of the Northern Tungus dialects. The comparison of farxun and akt is reached by advancing another hypothesis: namely, r before k is lost in Tungus and the original stem must be pYrk. What t is, and where it is, is not shown, but it is merely dismissed. However, farxun may be understood as far + suffix -xun (cf. I. Zaxarov, «Grammar,» op. cit., p. 83, Sec. 53), and in the original form it might be ar. In this form it may be compared with the Mongol haru (proposed by P. Pelliot, «Les Mots, etc.» op. cit., p. 216), while the Northern Tungus stem akt will remain as it is. Since this stem is found only in Samoyed, cannot be connected with the Tungus Northern Tungus stem, and the Manchu stem may perhaps be connected with the Mongol stem, I cannot accept A. Sauvageot's parallels as valid ones.
Case 12. The same stem *pYr—this time with the new meaning «tranchant,» «coupant.» Only one Tungus word is found— furu (Manchu Writ.)—«to reduce to pieces,» e.g., «meat,» «noodles,» «tissue into small pieces of thread,» etc., i.e., «to mince,» etc., and compared with Mongol hurtdsun (P. Pelliot), oertasan—«chiffons,» «morceaux d'etoffe de sole coupes en fragments,» whence urtahun (Ur., Castr.), urtasun (Mank., Castr.), evidently borrowed from Mongol in the sense of «patch,» «piece.» Since no words are found in other Tungus dialects and the Manchu word has rather a different meaning, the Ural-Altaic value of the Manchu word seems to be very little if none.
Case 13. A new stem *pYl for rendering the idea of «touffu,» «dru»; if the hypothesis of the loss of the initial p is accepted, then the Tungus and Mongol parallels are valid. However, in the Northern Tungus and Southern Tungus dialects it seems to be somewhat confined to a limited number of words and dialects, in which it looks to be of Mongol origin when met in the form utek, and of Manchu origin when it has the form of fuli, pule, etc. It is difficult to show whether all of them may be brought to one and the same source or not. I. Zaxarov, I think wrongly, compared Manchu fulu with a Chinese word, but even in European languages it is met with as containing about the same consonants, as, e.g., «full,» «plenus,» «polnyi,» etc. However, if the Manchu stem and the Northern Tungus stem are borrowed from the Mongol, then it is very likely that the actual stem is ul and thus the whole discussion is confined to the consonant l. In addition to this, it ought to be pointed out that in the Tungus language there are other stems (e.g., dzat) with the meaning «full,» etc.; while the stem ul figures in the words with the general idea of «exceeding.»
The above-demonstrated six cases may suffice to show the reasons why the class of these cases is not analysed in the coming sections. My intention does not consist in correcting, from the point of view of Tungus languages, the authors who have proposed various parallels, and therefore I do not need to go through all the cases. For the time being, I need only those cases which may be suspected of being common for the Tungus and some other languages. Yet they will suffice for penetrating the analytical methods of theoreticians of uralaltaicology.
102. With certain imagination it may perhaps be connected with Gilak «birch-bark» xiv, xip (Gilak, W, Grube), cf. kivo (Ner.) huva (Lam.), kiwa (Turn.), kiva (Ur., Hank., Castr)—«the birch-bark» (ft appears instead of Gilak x according to the phonetic character of the Northern Tungus and in Gilak the final vowel very often disappears). Besides this stem there is another stem in Tnngus; namely, lalu (Bir., Kum., RTM, Mank.) (Goldi, Orcchi, Sch.), talo (Khin.) (Neg. Sch.), cf. Manchu Writ, tolxon. However, it must be first shown that xiv, xip may be brought to ja, a, or ae. Indeed, it is absolutely uncertain and I do not naturally propose it. The fact is that we have two stems in Tungus.
103. These parallels are brought forth probably owing to the Manchu orthography, which renders f'a as fiya (or fija).
104. It is possible that in hijika -ka means something else, e.g., «shrubs,» as is met with in some other dialects. However, in the given case it is not essential for us.