Roscher's Greek Mythology
No TL;DR found
Abstract
association, and so levelled into more or less of likeness: and at last we get something as near regularity as, e.g, the Sanskrit verb : Greek and Latin retain to late days something more of the old irregularity : absolute regularity and system is achieved by no language. Again, in II. Mr. Walker does not say why he thinks the common view unsatisfactory. According to it the perfect in its oldest ascertainable — not necessarily its oldest—form had one set of flexions : 1, sing. a (original or for orig. m); 2, tha; 3, e: the s-aorist had the common ones m, s, t: these were levelled completely in Greek, and also in Latin (into the perfects in i, and those in si) : this levelling is just what always occurs when some one or two forms, out of two sets of forms originally distinct, coalesce phonetically, and when also the functions of the two sets become more or less identical— so here, when the perfect became (as in Latin) in the main a mere preterite like the aorist. Vedic preserved the old state of things : and the difficulty arising from the Vedic is, on Mr. Walker's view, as pointed out by Mr. Snow in the April number of the Review, almost insuperable. The 'reigning theory' which Mr. Walker finds unsatisfactory is, to my mind, as clear as any scientific principle can be, though some of the details may be wrong, and difficulties may remain which cannot be solved with any certainty. But here again what is the actual objection,to the theory? Mr. Walker only mentions ' the identification of scripsem with scripsi' as ' an effort of philological despair.' No one has 'identified' them, so far as I am aware : though it may be held that scripsem is ananalogicallyaltered form of * scripsem, representing an original * scripsm, which would be a proper aorist: while scripsi is a form which, by a wholly different process, and at a different period of the language, has been substituted for the original aorist: this is an explanation which is quite tenable : it may be right or it may be wrong. But it is a mere detail: and it has nothing to do with the truth of the general theory that the identity of the personal suffixes of the Greek perfect and aorist is merely a Greek development, and is due to the common principle of form-association. Lastly, in III. I see no reason to reject the accepted Greek phonetic rule that orig. nti when n was sonant, if accented, became avri (ao-i), and if unaccented an (atri): these last were levelled out of the ordinary language. (It is true that the Homeric ir€<j>vKaa-i, \e\oyxam, were conclusively shown by Mr. Leaf in the Journal of Philology to be late; but they must have been formed on true types, though these have perished.) H e n c e forms l ike TreTrotOaa-i (for vtTroiO-VTi): then by analogy dai passes into vowelstems, like ytyaao-t, and even into reduplicated presents like nOiaa-i, for the original TiOa'Ti (riOiiai). This explanation accounts for all forms, including iOwKan. The difficulty of lo-am, which seems to have started Mr. Walker's hypothesis, is surely quite a different one: it is the presence of the first cr (instead of 8), not the second one: this can only be guessed at : the form may be worked backwards from the preterite "a-av (for *ta-uav for * fiS-aav), as Brugmann holds, or it may be due to the analogy of la-ft.a', io-re, which changed * ISacri into lucun (but this would not explain the parallel clgacri): we cannot know for certain. But assuredly this form gives us no ground to assume an original suffix si for the third pers. plur. And Mr. Walker's lengthening of an imagined * iScri into io-ao-1 ' on the analogy of tauri and tacrt' is barely conceivable. As to the Latin forms uidSront, etc., Brugmann has long ago shown (to my mind conclusively) that they are s-aorists and not perfects. Here however we shall decide accordingly as we decide on II. JOHN PEILE.