English
The services of Professor Haxxs Oertel of Yale University, and of Professor Friedrich Knauer of the University of Kiew, and in particular of Pro¬ fessor Leopold vox Schroeder of the University of Vienna I have already described above, at page ix b , and acknowledge them gratefully. The editor of this Series, Professor Charles R. Laxmax of Harvard University, teacher, colleague, and friend, almost from the day when I began my Sanskrit studies in 1877, has brought to bear his great editorial talents upon this Concordance in every possible way. He has also found time to read one proof of a number of sheets (1-10 and 49-50). I can only regret that I have not had the benefit of his trusty eye and his sound scholarship for the bulk of the work, to the exclusion of many a blemish. It is a pleasure, finally, to remember my former pupil and friend, Mr. Henry C. Warren, of Cam¬ bridge, Massachusetts. Mr. Warren gained the respect of students of India by his scholarly volume, entitled Buddhism , a series of illustrative translations from the Sacred Books of the Buddhists (volume iii. of the Harvard Oriental Series). To his love of Oriental learning it is due that this Concordance and the other volumes of this Series are issued in a suitable and dignified form. At the close of this long labor, I record the kindly and substantial aid which he gave, while alive, to this undertaking; and I am glad to say that the enlightened provisions of his last will now further the studies that he loved in life. Baltimore, October, 1906. MAURICE BLOOMFIELD. EXPLANATIONS INTRODUCTORY TO THE USE OF THE CONCORDANCE General scope of these explanations,— Since this Concordance has to do with no less than one hundred and nineteen different works, it is, in the first place, obvious that a complete list of those works should be given, in a systematic order and with proper biblio¬ graphical descriptions of the editions used, and that the bibliographic list should be followed by an alphabetic list of the abbreviations by which the several works are designated. This list of abbrevia¬ tions is printed just before the beginning of the main body of the work. In the second place, while it is probable that the several more or less arbitrary devices employed in reporting the variants will be readily understood by every Vedic scholar, it is never¬ theless possible that uncertainty of one kind or another should arise in some mind; and I have accordingly thought it worth while to explain these devices systematically and in connection with a few illustrative examples. GENERAL REMARKS ON THE CHARACTER OP THE VARIANTS No hard and fast line between truly variant and vaguely similar passages.— The similarities be¬ tween related Vedic passages are of very varying decrees, and range from a likeness so close that one may be called a true variant of the other, to a like¬ ness so slight or vague as hardly to be worth report¬ ing. Nevertheless, if we take a few typical examples, the differences of character as between several of the many possible classes are palpable enough. True variants.—Uha-padas. —Thus, if we take the two padas occurring in the Rig-Veda and the Atharva- Veda respectively, acittl yat tava dharma yuyopima, and acittya cet tava dharma yuyopima, it is plain that the latter is a true variant of the former. In this instance even the motive of the changes is plain: the Atharva-Veda transfuses the hieratic language of the Rig-Veda into its own more popular wording. Another well-defined, though not common, class of true variant passages includes the uha-padas, which consist of an otherwise identical phrase varied, for instance, by the substitution of the name of one deity for another. Such are mam agne bhaginaui kuru, and mam indra bhaginarii kuru. Variants in padas other than the first. —It is im¬ portant, next, to draw attention to the cases where the differences between two or more stanzas crop out in some other pada than the first. Since, for the Concordance, the unit of comparison is the pada and not the stanza, the Concordance reveals very many such cases which would of course never come to light in a mere index of pratikas. For example, stanza 3 of the RVKh.10.127 is in part parallel to ArS.3.7; but the first three padas of each stanza are so unlike those of the other that no index hitherto available could have brought to light the parallelism. Not until we come almost to the end do we find the correspondence. The fourth padas are vigvasya jagato nigam, RVKh., and vigvasya jagato ratri, ArS., a parallelism which the alphabetic arrangement of the single padas in the Concordance makes apparent automatically. Similarly the stanza RV.1.141.13 shares its second and third padas with several Yajus texts (KS.7.12 ; ApQ.5.9. 10 ; MQ.1.5.2.11), while in padas one and four it is wholly different in subject-matter and diction. The RV. stanza reads: astavy aguih gimivadbhir arkaih samrajyaya prataram dadhanah, ami ea ye maghavano vayam ca miham 11a suro ati nis tatanyuh; while the Yajus stanza reads : ut samudran madhuman urmir agat samrajyaya prataram dadhanah, ami ca } T e maghavano vayaiii ea isam urjam madhumat sambharema. Accordingly it is clearly worth while to consult the Concordance for every pada of a given stanza, if such parallelisms are not to escape notice. Essential similarity with wide divergence of word¬ ing. —Once more, it sometimes happens that the METHODS USED IN REPORTING THE VARIANTS [Explanations xiv wording of one and the same stanza is so vaiiod in the different texts that not one of the lines of the one text is quite like the corresponding line in the other. A striking illustration is afforded by AV.5.25.8 and QG.T.19. The text of AV. reads : adhi skanda virayasva garbkam ii dhehi yonyfim, vrsfisi vrsnyavan prajayai tva nayamasi; while the text of QG. reads: abhi kranda vilayasva garbham a dhehi sad hay a, vrsanam vrsann a dhehi prajayai tva kavamahe. Of the fourteen words of the latter stanza, only five correspond in form and place with words of the prior stanza; and yet the two stanzas are beyond a doubt alike in their purpose and meaning, and they are in all probability two versions of an originally identical stanza. The like is true, for example, of SV.2.1059 and AV.6.36.2 (AQ.8.9.7 ; QQ.10.11.9). The SV. text reads: ya idarn pratipaprathe yajhasya svar uttiran rtun utsrjate vaul; while the AV. text reads : sa vi^vii prati cfiklpe rtimr utsrjate va^i yajnasya vaya uttiran. Cases of partial correspondence: At the begin¬ ning ; at the end ; in the middle. —But so numerous are the eases of partial correspondence or minor parallelism between padas or passages, that no amount of mere attentive observation can be expected to dis¬ close them all. Set phrases, groups of two or three words what Bcrgaigne used to call formulas—are, as every Vedist knows, the commonplace of Vedic technique. These set phrases may of course be located at the beginning or in the middle or at the end of a pada or passage. Those at the beginning are re¬ vealed by this Concordance. Those at the end would be revealed by a reverse index (discussed above, in the Preface, at page x b )d For those located neither at the beginning nor at the end, nothing short of a complete verbal index of the Vedic mantras, in which every pair or longer group of words should he placed side by side, could exhaust this important and interesting information about the texture of the mantras. 1 As an interesting case of parallelism under this rubric, we may cite 9 atam te santv avrtah (AV.) and agne angirah ^atam te santv avrtah (VS.). No less a personage than Panini has made the acute observation (iii. 2 . 66) that the stem havyavahana occurs only at the end of a pada. METHODS USED IN REPORTING THE VARIANTS 1 . Variants for which the alphabetic order does not require more than one entry. —These cases are so simple that little need be said about them. When, for example, a single word of a passage differs in two or more texts, the reading of that text is put first which stands first in the usual order of the texts,— RV., AV., SV., Yajus texts, Brakmanas, and so on, -and the variant reading or readings follow in parenthesis. Thus: yah praniti (AV. pranati) ya liii 9rnoty uktarn RV.10. 125-4 b ; AV.4.30.4 b . Similarly, when variants occur at more than one place, they follow in parenthesis, each in its proper place, and with an indication of its source. Thus : vasuni carur (SMB. carye; ApMB. caryo; HG. cayyo) vi bhajasi (SMB. bkrjasi; HG. bkaja sa) jivan AV.19.24.6 11 ; SMB.i.i.6 d ; HG.i. 4 .3 d ; ApMB. 2 . 2 . 8 d . This means that the four texts concerned, if written out in full, would read as follows : AV.: vasuni carur vi bhajasi jivan; SMR.: vasuni carye vi bhrjasi jivan ; HG.: vasuni cayyo vi bkaja sa jivan ; ApMB.: vasuni caryo vi bhajasi jivan. So much for variants that concern single words. The case of a variant which concerns two or more successive words is as a rule no less simple: the device employed ought to leave no doubt as to the precise wording of the several texts. Here, again, the variant is placed within the parenthesis as before. Thus : sa no mayoblnih pito av^asva (QG. PG. pitav av^asva ; AQ. pitav av^cha; MQ. pitur avive^ ; SMB. pitevavijasva) TS.5.;.2.4 C ; TB.2.4.8.7 C ; AQ. 2.9.10°; MQ. 1.6.4.25°; QG.3.8. 3 c ; SMB.2.i.i3 c ; PG.3.i. 4 c . See sa nah pito. This item, if wc include the cross-reference at the end, gives us the reading of a pada in nine different texts. The device requires no explanation.