that much light must be thrown by them upon still other forms of Avesta Grammar. As to the unsightly forms above mentioned with their irrational e in -yemi (yeimi), yehi (yeihi), etc., I should say that I once indeed surmised that the presence of these supposed e sounds in the terminations cited looked backwards towards the original Indogermanic 'e' which has thrust the Indian 'a' and 'a' for the moment aside from our consideration in our search for the original vowel sounds of the verbal terminations. But after we have seen that the once reported sounds in the very familiar and important terminations of the gen. s. masc. and neut. have no existence whatsoever, and that they never had any place among real linguistic features in use by persons speaking their vernacular, we look for another explanation of the phenomena in the verbal terminations. And we soon see that the explanation of the origin of our forms there is similar to that pursued before. The character present in each of these cases of verbal termination, as in many others (see below) positively arose from the misapprehension of a Pahlavi sign which indeed, as I hold (see below), may have expressed a+i which are the elements of an ẹ, but which also still more prominently expressed and expresses ya in the original Pahlavi forms out of which our remarkable Avesta Alphabet was developed. So that the terminations which we have universally read as yemi (-yeimi), -yehi (-yeihi), yeiti, etc., etc. (much to the bewilderment of all who have given the matter interior attention) are the result of a complete mistake; and of one quite as irrational as our supposed -ahe for a gen. sg. m. n. of the 'a' declension, and perhaps even more offensive to the philological instinct, though this last might be saying too much. And this explanation is fully corroborated by the occasional usages of our Avesta text itself, which sometimes leaves us no loophole to escape into our blunder; for the Avesta itself at times actually reads the normal forms -yāmi, -yahi, etc. fully in accordance with the Indian, see below on page 322. The sign which we have supposed to be e! in the yẹimi (!) of the -ya verbs is simply -ya, a survival of the original familiar Pahlavi form without which -ya at one time could hardly have been written, and without which it could never have been naturally written. I do not produce the original characters here, as this might cause inconvenience or expense, but the reader will find them used in the Zeitschrift of the German Oriental Society as cited in the note above.1 The first item which an objector should notice is that I am obliged to render our compositum as -yā for the first personals and as -ya for the 2d and 3d. To this we must say at once that uniformity in equivalence is precisely what we must not look for in the developments of these transitional forms, and above all not just here and in reference to this a and a, for the Avesta scholars of the early centuries were in fact just beginning to use the Pahlavi form which stands for long a as the Avesta form now used for the shortened a, as the equivalent of which it now stands, for they were just inventing all the now prevalent regular signs for the shortened vowels, they (the short vowels) having never before been expressed in writing, so far as we are now aware. All the short vowels (and not merely the short 'a', as in Indian) have been always inherent in the consonants of the Pahlavi Alphabet, save the 'a' privative which has no consonant before it. The usage, therefore fluctuated, as a matter of course, at the formative, or reformative, period of reconstruction, when the Avesta Alphabet was just emerging from the Pahlavi. And the Avesta Alphabet partook and still partakes to some degree of the notorious characteristic of its parent, the Pahlavi Alphabet, which was polyphonic in the extreme. That the Avesta Alphabet partook only to a slight degree of this characteristic of its mother was all the more natural because it (the Avesta Alphabet) was especially invented for the purpose of removing the original and long prevalent ambiguities, though it could not be expected to have retained no lingering trace of the agitation. Another objection should be at once pointed out, which is that by this rectification we are left with a redundant 'y'. We have already here in our erroneous forms, -yeimi, etc. before us the fully written Avesta 'y' of the -ya verb; when therefore we decipher our character (until lately mistaken for an e) as being simply in fact our needed y +ā, an expected relic of Pahlavi usage, we have on our hands an additional unnecessary and redundant 'y'; for beside the fully written Avesta 'y' we have our ancient compositum with the same 'y' in its meaning, which meaning is y + ā (or ‘y + a'). To this the answer is that here we have an instance of a 1As well as in my Gāthas, Vol. II, and IIIa; see also at many places in the Latin notes to the texts in Vol. I. redundancy not nearly so irrational as some others which appear in the still halting Zend alphabetical forms, with our still more imperfect transcriptions of them. For our lingering and redundant 'y' was originally by no means so useless as one might at the first glance suppose. It was obviously left in its place in the endeavor to save us from the very error into which we have fallen. The intention to abandon the use of our compositum for y+a (or 'y+a') was beginning to be formed within the scattered schools of Avesta Philology, and an Avesta y + a was purposely inserted before our character here to guide the readers against the coming use of this sign for e, which use was not applicable here in the first, second, and third personals of the -ya verbs, and therefore not to be written with that slight prolongation of one extremity which alone distinguishes the form at present known as 'e' from its fellow character of the same shape; that is to say, from the actual Pahlavi original of our y+a (or y+a) and from the same form when standing for "š', sh. The fully expressed Avesta and later characters for y + ȧ were inserted to save us from writing 'ye', just as the otherwise so irrational 'a' was left before '' in Daeva, etc. (as to which see Gathas Vol. IIIa, p. xv). This correctly written Avesta 'y' is a portion of that once wisely inserted 'ya', and it has been left in the word like other fragments as débris, and just as also a now useless and extinct 'a' has been left clinging to our sign for ñ, which originally represented an 'an'; see also the unnecessary nasalism long a before m in yām =yām, etc. This last is especially useful as an illustration; for the redundant nasalisation once pointed to the now so well expressed nasal 'm'. The a which originally stood after our redundant 'y' disappeared in consequence of its having been already and originally expressed in the sign which we have mistaken for an e (in the falsely deciphered yeimi). Or, vice versa, we might say (on the other hand) that the redundancy consists in the presence of the first half of this our sign till lately mistaken for 'e', but which cannot possibly mean anything but y + a here; so that we have in fact in either case a doubled 'y' upon our hands as in yyāmi, yyahi, etc., just as we have a superfluous 'a' in our sign for 'ñ' which was originally 'an', a superfluous anticipation of the nasalisation in our 'a' for a before 'm' in -yām, and in our 'a' before 'e' in d(a)ēva, etc. My answer is that we should simply do as our predecessors did with the superfluous a in our present sign for 'ñ' which was 'an'; we should in our transcriptions ignore and omit our first superfluous sign for 'y', or we should bracket it, as our form is -yāmi. There is most certainly no -yçimi (nor -yẹmi) whatsoever, nor ever were there such forms, nor is there any yehi (-yẹihi), nor any yeiti, nor -yẹiñti in any -ya verbs in the Avesta language. The words end in -y(y)āmi, -y(y)ahi, -y(y)ati, -y(y)anti; see Indian -yāmi, -yasi, -yati, -yanti; and also the actual Avesta forms quite rationally written as -yāmi, -yahi, -yati, -yañti; and all the forms should be so transcribed, if transcribed at all. Although wishing to avoid statistics here, I cannot refrain from citing Avesta and the other nine (odd) forms of the first person pl.; so of the conjunctive e. g. taurvayama, etc. all read -ya in the -ya verbs. jai@yanti davayanti yanti yanti syanti $yanti So of the third pl. med. all (?) the -ya verbs have -yanta, each form being perfectly rational. How could an ancient Iranian have said -yāmi at one moment like his Indian kinsman, and then this yeimi at the next at the next (no dialectical influence being here present in any shape). P. S.-Objectors should not ask for instances of absolutely pure Pahlavi of an unequivocal character in the Avesta Alphabet; yet we have these in the otherwise exclusively Pahlavi characters, in the compositum for h' (or 'hw', which in no sense requires the elevation of the v (or w)). Both the signs are absolutely pure Pahlavi, and both are otherwise totally excluded from the Avesta Alphabet, which has but one and a different sign for 'h' (though at times abbreviated) and but one and a different sign for this particular interior v (or 'w') in other occurrences. UNIVERSITY OF OXFORD, May, 1903. L. H. MILLS. V. VICA POTA. In the Ludus de morte Caesaris, that scurrilous but amusing diatribe against the deceased Claudius, Seneca briefly mentions Vica Pota, one of the obscurest of the many obscure Roman deities. The passage in question occurs in 9, 4 and reads as follows: "Proximus interrogatur sententiam Diespiter Vicae Potae filius, et ipse designatus consul, nummulariolus: hoc quaestu se sustinebat, vendere civitatulas solebat." The earliest reference to the goddess is found in Cicero, De Leg. II, 11, 28, "Virtutes enim, non vitia consecrare decet. araque vetusta in Palatio Febris et altera Esquiliis Malae Fortunae detestanda, atque omnia eius modi repudianda sunt. quod si fingenda nomina, Vicae Potae potius vincendi atque potiundi, Statae standi cognominaque Statoris et Invicti Iovis, rerum expetundarum nomina, Salutis, Honoris, Opis, Victoriae." Such at least is the reading of the codices, but the majority of editors would bracket as glosses the gerunds vincendi, potiundi and standi. Whether we grant or deny that Cicero derived the name Vica Pota from vinco and potior has little bearing on the aim of this investigation, for we have certain proof that in the early Empire the goddess Vica Pota was completely identified with Victoria. The evidence may be seen in the passages next cited. Livy II, 7, 12, after telling how, in the first year of the Republic, P. Valerius aroused suspicion by beginning to erect a house on the summit of the Velia, and how, in indignation at the fickleness of the people, he decided to build at the foot of the hill, says: "Delata confestim materia omnis infra Veliam, et, ubi nunc Vicae Potae est, domus in infimo clivo aedificata." Plutarch, Poplicola Io, gives the same tradition (ὅπου νῦν ἱερόν ἐστιν Οὐίκας Πότας); and Asconius, in his commentary on Cicero's oration In Pisonem 52, shows the popular identification of Vica Pota and Victoria by his statement: "Iulius Hyginus dicit . . . P. Valerio Volesi filio Publicolae aedium publice locum sub Veliis, ubi nunc aedis Victoriae est, populum in lege quam ipse tulerat concessisse." ... Without exception, every scholar who has treated the subject identifies Vica Pota with Victoria and follows the lead of Cicero, |