12 APPENDIX

Agrayana and Orion.

I have already stated in brief my reasons for provisiopally identifying Sk. Agrayana with Gk. Orion, and here I wish to examine the point more fully, not because my case rests upon it, bnt simply with a view to indicate the real nature of the objections that may be urged against the proposed identification. If philologists are still inclined to hold that the identification is not even probable, we shall have to look for some other Aryan derivation, as the similarity of the Eastern and Western traditions of Orion is, in my opioion, too strong to be accidental. Agrayana is evidently derived from agra and ayana, Of these ayana, which is derived from i, to go, may be represent ed by ion in Greek ; cf. Sk. @yus, Gk, aiön Sk. comparative termination (nom. sin.) iyān Gk. ion; Sk. termination dyana, as in Gargyayana, Gk iön, as in Kronion, ’the descen dant of Kronos’. The initial ā in Sk. Agrayana may also become in Greek; as in Sk. ashayana, Gk õkeanos; Sk. dshu Gk, ökus, Sanskrit Ägrayana inay therefore be represent. ed by Ogriön in Greck, and we have now to see if g may be dropped before and Ogrion can be changed into Orion. It is a general phonetic rule in Teutonic languages that a gutteral may disappear before a liquid, whether initially or medially ; cf. Ger, nagel, Eng. nail ; hagel and hail; regen and rain; Sk, kravis, O. H. G. rõ. Prof. Max Mūller has extended the application of this rule 10 Latin and Greek, and Latin and French in his Lectures on the Science of Language, Vol. II., p. 309. He compares Lalin paganus with French paien, Gk, lāchne’ With Lat, lāna; and points 222 APPENDIX. out that on the same principle luman stands for lucmen, examen for exagmen, Hamma for fagma. K. Brugmann (Oom. Gr. I., 9023.) would derive 0, Ir. dr, Cymr, aer from “agro on the same principle. This shows that Sk. agra may be easily represented by ār in Teutonic languages. We may account for the change in two ways. We may cither suppose that the final gutteral of a root is sometimes dropped before terminations begioning with a liquid and thus put luc-men = lu-men, fulg-men-ful-men, Ang-mens Aa-men, ag-men-d-men, ag-11 - dura ( with compensation vowel lengthening; Bopp derives Sk, roman, a hair from ruh-man growing, on the same principle); or we may suppose that the change is in accordance with general phonetic rule which sanctions the omission of a gutteral before a liquid in such cases. But whichsoever explanation we adopt, there no question as to the change itself, It must pot, however, be supposed the rule is an uninflexible one, and that a gutteral must always be dropped before a liquid; for we find that gutteral in such cases is often cither retained or labialised. cf. Sk. gravan, O. Ir. broo, bro, (gen. broon), Cymr, breuan; Sk. spindmi, Q. Ir. gair. The proper rule to deduce from these instances would therefore be, thal gr in Sanskrit may be represented by gr, br or r in Teutonic languages, and that all the three changes are possible. Can we pot extend the rule to Greek and Sanskrit? is the next question we have to coosider. I do not mean to deny that there are phunetic rules which are not univer. sally applicable to all languages. But the present rule can be easily shewn not to belong to this clāss. Prof. Max Māller has already extended it to Greek and Latin and APPENDIX. Vararuchi, in his Prakrita Prakasha II. 2, lays down that & in ga may be medially dropped as between Sanskrit and Prakrit, 6. 8. Sk. sagara, Pk. sa-ara; Sk. nagara, Pk. na-ar, eventually corrupted into nara, as in Jun-nara and other names of cities. This is, in fact, the same rule which, when applied to Teutonic languages, accounts for the change of segel into sail, nagel into nail and so on. Comparison of Avestic tigra with Mod. Per. tir shews that a similar change may also take place between those languages. We may, therefore, fairly say that the rule about the omission of gutteral before a liquid obtains not only in Teutonic langa. ages, but also between Greek and Latio, Latin and French Sanskrit and Old Irish, Sanskrit and Prakrit, and Avestir. and Modera Persian. In the face of these facts it would, I think, be uaduly restricting the applicability of the phonetic rule if we refuse to apply it to Sanskrit and Greek. There is at any rate no a priori improbability in expecting that a similar change may take place as between Greek and Sanskrit. Let Us Dow see if there are any instances as between Greek and Sanskrit to support such a conclusion. Prof. Benfey compares Sk. grāvan with Gk. laas (Lat. lapis); and Sk. ghrāno with Gk, nis, rinos. If this comparison is correct, here at least we have two instancos where a gutteral before r in Sanskrit is lost in Greek. It is some. times labialised, as in Sk. krinami, Gk, priami ; Sk guru, Gk, barus ; and sometimes retained as it is, as in Sk, kralu, Gk, kratus ; Sk, gras, Gk, grao, to swallow. From these instances we may therefore infer that as belween Greek and Sanskrit, the initial cutleral in kr or gr in Sanskrit may be either retained as it is, or labialised or dropped in rannan nama polim mhich holds good, as shown above, 224 APPENDIX in Teutonic languages. It may be noticed here ihal while grdvan becomes laos in Greek, it is broon in Old Irish, that is, while the initial g of a Sanskrit word is labialised in Old Irish it is.dropped in Greek. This shews that the initial kr or gr in Sanskrit may be differently represented in different languages. Sanskrit kṣimis, Lat, vermis, Gk, elmis, and Sk. Rlipta, Avestic kerepta, Gk, raptos, may, I think, also be regard. ed as further illustrations of the same rule. I know that the connection between the words last quoted is still considered doubtful, but that is because the rule about the omission of a guiteral before a liquid, as between Greek and Sanskrit, is not yet recognised by scholars. If the examples I have given at the beginning of this paragraph are, however, sufficient to justify us in applying the rule to Greek and Sanskrit, the instances last cited may be taken as further supporting the same view. With these instances before us, it would be unreasonable to deny that the three possible changes of kr and gr, which obtain in Teutonic languages, do not take place as between Greek and Sanskrit, at least initially; and if these changes take place initially, a nalogy at once suggests that they would also take place medially. At any rate there is no reason why they should not. It may be urged that a comparison of Sk, chakra with Gk, kuklos shews that a medial kr is retained as it is. But as pointed out above the argument is not conclusive. There may be cases where kr is retained as it is. But we have seen that by the side of such cases, instances can be quoted where it’ is changed to pr or r initially; and we may expect the same Lhreefold possible change medially. It is admitted that labiali. sation takes place medially; and we have therefore to see if there are any instances where a gutteral is dropped before a liamid io tome body of word. K. Brmamann tells us that APPENDIX. 225 at one period gn and gm came to be represented by n and m in Greek; cf. gignomai and ginomai, stugnos and stunos. Now this change in the budy of a word is exactly similar to that of agnien inlo amen, and is evidently due to the same rule, which accounts for the latter change. Similarly Gk. anoos may be compared with Sk. ajna, aad Gk. arinos 10 Sk. aghrana. But I do not lay much stress on these inasmuch as these words may be supposed to have been derived by the addition of the prefix alpha to the already existing Greek forms, and not directly obtained from Sanskrit ajna and aghrāna. The chaoge of gignonai into ginomai or of gignosko into ginosko cannot, however, be so accounted for, and if so before n is dropped in the body of a word, there is no reason why it should not be dropped before r on the analogy of the phonetic rule given above. Works oa philology do not give any more instances of such changes, but as observed above, the attention of scholars does not appear to have been direct. ed to this point. Otherwise I do not think it was difficult to discover the similarity between Gk. turos and Sk. takra. Takra is derived from tanch (*teng) lo coatract, to coagulate or curdle, and according to Fick the root is Iado Germanic. It is an old Vedic root, and we have such expressions as dadhna dlanakti coagulates (milk) with curds’ in the Taittirīya saṁhitā II. 5. 3.8. Takra therefore literally means ‘curdled milk’ and not curds mixed with water ‘as the word is understood in modera Sanskrit. Now, if we suppose that the rule, which saactions the omission of g or k before y or m in other languages, also holds good as between Greek and Sanskrit, not only initially (as in grivan aad laos) but also medially, as in ginomai and ginomai, Sk takra may be casily identified with Gk, duros meaning “cheese’. Takra may hanne lova, amell ob man maono mariniramo in na rin anim a le 226 APPENDIX. Turas is an old Greek word used in the Odyssey, and it ha not yet been explained by anything in Iado-Germanic. D Schrader therefore records a suggestion that it should b derived from Turko Tataric turak. But if Sanskrit sāra an sarpis are found in Greek oros (whey) and elphos (bulter) is not reasonable to suppose that turos alone was borrowe from a pop-Aryan source. Takra in modern Sanskrit meat curds mixed with water and churned’ and perhaps it may b contended that we cannot identify it with furos, which mean *cheese, I have, however, shewn that takra etymologicall means ‘curdled milk’ and aot ‘curds dissolved in water’ whic is evideatly its secondary meaning. Besides when we see the sara which in Sanskrit denotes ‘curdled milk,’ has becom oros - whey in Greek, and serum in Latin, there is nothin unusual if we find takra and turos used in slightly differen seases in the two languages. I have already suggested in th body of the essay that we may ideatify Sk. Shukra with Gk Kupris. Chakra kuklos, Shukra. Kupris, and takra=turos may thus be taken to illustrate the application of the rul above discussed, regarding the three-fold change of kr or g to Greek and Sanskrit medially; and instances have bee already quoted to show that the rule holds good initially a between Greek and Sanskrit. We may therefore conclude tha the change of signomai into ginomai is not a solitary instance and that as a general rule & may be dropped, labialised o retained before a liquid as between Greek and Sanskrit whethe at the beginning or in the body of a word. We might eve discover further instances of the applicability of this rule for, if takra is thus correctly identified with turos, we may, oi the same principle, identify 8k, agra with Gk. oros, meanin, top summit. It was impossible to represcot Sanskrit agri frumemmomnandha mamlama mo laman 1 million con amor APPENDIX. 227 sented by agros in Greek as the latter word corresponded to Sk, ajra, a field ; nor can agra be changed to akris which represented Sk. ashri. Sanskrit agra, therefore, naturally came to be represented by oros. Oros, meaning top or sum. mit, has not yet been satisfactorily derived in any other way. It will be seen from the above that we have sufficient grounds to hold that the rule about the omission of a gutteral before a liquid, whether initially or medially, applies to Greek and Sanskrit in the same way as it does to other languages; and if so, Sk. Agrayana can be represented by Orion in Greek. I have already quoled Brugmann to show that *agra becomes ār or aer in Teutonic languages. Now further com paring Lat. integri, integer with Fr. entier ; Gk. dakru, Goth. tagr with Eng. tear; pagan with paien and regen with rain, we are led to infer that where k org is dropped before y or a liquid we may expect two contiguous vowels, probably because this gr is at first optionally altered into ger or gar. We can now understand why Orion was sometimes spelt as Oarion; and the existence of this double form cooörms, in my opinion, the derivation above suggested. As for union alone we might derive it from oros, limit, or āra, spring, and ion, going thus, giving the same meaniog, viz., the lirit of the beginning of the year or spring, as Agrayana in Sanskrit. But this does not account for the double form Orion and Qarion–unless the latter be taken for a poetic or a dialectic variation of Orion. I therefore prefer to derive the word from Sanskrit Agrayana.