Having formed an opinion as to the place where Zara-- thujrtra laboured, and proceeding to the question as to when he lived and wrote the Gathas, we find ourselves under the necessity to form our estimate first as to the age of the later parts of the Avesta. While interpolated passages, or indeed whole Yajts, may be very late, I cannot place the later Avesta in its bulk later than the Cuneiform Inscriptions of Darius, for the fact that the In- scriptions preserve either a pre~Zarathujtrian Mazdaisra, or the Zarathujtflanism of the Gathas long previous as it was in its origin to that of the V endidad, has nothing whatever to do with the relative age of the Inscriptions themselves. The later Avesta, with its forbiddal of burial and crema- tion, must have existed for a long time side by side with that religion which has left sepulchral monuments, and XXXIV THE GAT II AS.
whose adherents could contemplate the burning of cap- tives; and analogous facts are universal.
But aside from the seeming difference in the. type of Mazda- worships which simply severs the religion of the Inscriptions from that of the more developed Zarathiur- trianism, and which has, as we, have seen, nothing whatever to do with the question of the relative ages of the Inscrip- tions and the later Avesta, I think that we have some signs of a later age in the language of the Inscriptions apart from their contents. As, however, Darmesteter is inclined to regard the West Iranian, or Cuneiform, as better pre- served than the Zend of the later Avesta, I make my few remarks only with great hesitation.
The termination which would otherwise be justly considered as an evidence of degeneration in the Zend, I regard as merely a wrong writing for -ahya==G&thic ahyd. The letterJO is a relic of the time when the Avesta stood in the Pahlavi character; I think that it is here merely a lengthened 10 —ya 1 * *. Terminations also seem much muti- lated in the Cuneiform, and the name Auramazda written as one word, does not seem to me so original.
We must indeed remember that a later generation, owing to an isolated position, often preserves an older dialect, as it may an older form of religion, whereas an 'earlier genera- tion, if its predecessors have lived in a compact society m smaller districts, varies the ancient forms, as the old Indian developed into Sanskrit and Prakrit. Still we have little reason to be certain that the civilisation of Media and 1 Also W 4 * is simply ayam, and should he so transliterated; so also in a throng of other words. Salemarm has noticed the origin of KJ =s 6, but gives no other indication in the present sense. I think that ft and also g, where they equal Aryan ya, should be corrected everywhere, like all other instances of miswriting. Unless indeed we can regard the m, for which » were often dearly miswritten, as itself of double significance, as in Pahlavi. * might then regularly and properly equal both e and ya; so may equal long $ or y& (aya). Other instances of miswriting in Zend would be dat. dual -by*. The Aryan -am was first written as the nasal vowel and still further carelessly reduced to -a, but never so spoken. On the contrary, in the acc. fern. &c., the nasalisation was over-written, too much expressed. The final nasal caused the scribes to write the preceding letter as if nasalised, *a/ but it was never nasalised In speech.
INTRODUCTION.
XXXV Persia was either more or less condensed and social than that of Bactria and the East. But beside a priority to the Inscriptions, we are -obliged to consider the time needed for developments. The Greeks -of the time of Herodotus probably, and those later certainly, found a form of Zara- thuJtrianlsm in full development in Media; but If the con- temporaries of Herodotus heard familiarly of a Zarathuj*- trianism there, a long period of time must be allowed for its development if It originated in Media, and a still longer period if it found its way there from the East. If, then, the bulk of the later Avesta existed at the time of Herodotus and at that of Darius, how long previously must it have been composed; for such systems do not bloom in a day?
We have the evidence- of historical tradition that the Magi 1 were influential even at the time of Cyrus, not dwelling upon the possibility of their existence at the earliest mention of Medes as the conquerors and rulers of Babylon.
Can we then, considering the recognised stagnation of ancient Eastern intelligence, ascribe to the development of the Median Zarathurtrianism a shorter period than from one to three centuries? If, then, the bulk of the later Avesta must be placed so long before the Inscriptions of Darius, where shall we place the earlier Avesta with its most important remaining fragments, the Gathas 2?
After studying the Gathas carefully in detail, and be- coming also familiar with them as a whole by frequent perusal, we must measure the time needed for the change from their tone to that of the later Avesta. Could it have been less than a century, or centuries? Was not as much time needed for the Zarathurtra of the Gathas to become the Zarathurtra of the later Avesta, as was afterwards con- sumed" by the migration of the creed from the North-east, If it really originated there? As there is undoubtedly a 1 I regard the Magi as representing the Zarathurtrianism' of the Vendidad.
This the false Bardiya endeavoured to introduce, demolishing the temples which the old Mazda- worship permitted in Persia. See the Cuneiform Inscription of Behistun Up Darius 61. ♦ 2.All 'in the Gathic dialect is old. /?
XXXVI THE GATH AS.
difference of several centuries between the dates of the newest and oldest parts of the later Avesta. so we must think of a considerable interval between the oldest parts of the later Avesta and the latest parts of the older Avesta, for there is the other consideration which imperatively con- strains us to avoid concluding for short periods in the stages of development. The Vedic Hymns, sung in metres closely similar to those in both the Gathas and the later Avesta, and naming gods, demons, and heroes so closely related, not to speak of myths, challenge us to say whether they are, the oldest of them, older or later than the oldest parts of the Avesta, and, if there exists any difference as to the ages of these ancient productions, how great that difference is. The oldest i?fks have now an established antiquity of about 4000; were the hymns sung on the other side of the mountains as old? The metres of these latter are as old as those of the i?2g-veda, if not older, and their grammatical forms and word structure are often posi- tively nearer the original Aryan from which both proceeded. If it were not for two circumstances, we should be forced to ask very seriously which were the older, and to abandon altogether our mention of later dates. Those circum- stances are the absence of the Aryan gods from the Gdthas; and, secondly, their abstract conceptions. These latter are so little offset with expected puerilities that it is often hard to believe that the Gathas are old at all. Their antiquity is placed beyond dispute by the historic mention of Zara- thastra. But, if Zarathurtra were not indisputably a living man in the Gathas, their depth and refinement, together with the absence of Mithra, Haoma. &c,, would, in them- selves considered, force us to place them rather late. As it is, the absence of Mithra and his colleagues, who reappear in the later Avesta, permits us to place the Gathas con- siderably later than the oldest Rik s. For no sudden and intentional dismissal of the ancient gods is to be accepted with Haug, nor any religious schism as the cause (!) of the migration of the Indians toward the south. The process was of course the reverse.
The migrating tribes, in consequence of their separation INTRODUCTION* xxxvu Irom their brethren in Iran, soon became estranged from them, and their most favoured Gods fell slowly into neglect, if not disfavour.
We need time to account for this change, and no short interval of time. We can therefore place the Gathas long after the oldest Rzks. While, therefore, in view of the established age of the i?2g-veda, the Gathas may possibly have been composed as- early as about 1500 B.C., it is also possible to place them as late as (say) 900-1300 B.C., while the fragments in the Gathic dialect must be considered somewhat later. The dates of the composition of the several parts of the later Avesta, on the other hand, must be sup- posed to extend over many centuries, as the various sections in the Zend dialect are so much more numerous than those in the Gathic, the G&thas themselves representing practi- cally but one date. Placing then the oldest portions of the later Avesta somewhat earlier than Darius, we are obliged to extend the period during which its several parts were com- posed so far as perhaps to the third or fourth century before Christ, the half-spurious matter contained in them being regarded as indefinitely later.
It seems necessary to state here for the information of non-specialists, and as bearing very seriously upon all the questions involved, that a very unusually severe controversy prevails upon the exegesis of.the Avesta, and that it centres in the question as to the value of the Asiatic trans- lations of it. A similar debate was once held on the Rfg- veda, but that is now silenced, all agreeing that the traditional renderings are neither 'to be slavishly followed, nor blindly ignored. Very different has been the fate of Zend philology, and in one important particular the studies are poles apart; for whereas the commentaries on the Riks are written in Sanskrit,- which is clear to experts, those on' the Zend-Avesta are written in a language upon which the lexicography is most incomplete, and the elucidation of these explanations themselves remains by far the most XXXVlll THE GATHAS.
difficult task now before us. Professor von Spiegel has accomplished much toward breaking the rough road of science in this direction, and scholars of the first order have followed his leading, while all with one accord express to him their acknowledgments. But Professor von Spiegel has not intended Ms editions and citations to represent full translations. He has, as a matter of course, taken it for granted that those who oppose him, as well as those who follow him, have studied his Pahlavi editions, not paying Mm the undesired compliment of making his commentaries the sole source of their knowledge of tradition. Moreover in no branch of science does scholarship make more rapid strides than in Pahlavi, several Important works having appeared since Spiegel’s commentaries.
In the attempt to master the Pahlavi translations of the Avesta we must consider many and difficult problems.
In the first place, and as a matter of course, they cannot be at all reasonably attempted without a full knowledge of the Gathic and Avesta texts so far as they have been as yet otherwise and approximately elucidated. The two problems hang together like the arches of a circular building, and they should be studied together word for word; for the Pahlavi used is not fully that of the books. It is often turned quite out of its course, as Pahlavi, by an effort to follow the more highly inflected Zend literally. Then, again, a question of the utmost importance meets us in estimating the glosses, which are often, but not always,- from a later hand. A translation of the Pahlavi must of course first be considered as in the light of the glosses, for the language is so indefinite as to many of Its grammatical forms, that such an indication as a gloss, if it be proved to have been written by the same person who composed the text, would be decisive in determining the rendering; but & final translation should be made more strictly In the light of the Gathic, so far as it affords on Its side positive indications, and the glosses, where they do not correspond, should be set apart as from a later hand.' Then, once more, and on the contrary, where the gloss is obviously right, and the text ^roneous, the former should be appro- INTRODUCTION* XXXIX priated unencumbered by the latter We must recognise the traces of former accurate scholarship whether we see them In text or gloss, and, from the accumulation of the correct surmises, we should construct an argument for the probability of the correctness of the hints of the Pahlavi in cases of great difficulty. In rendering the Pahlavi as a necessary prelude to rendering the Avesta, all possible help should of course be sought from the Asiatic translations of the Pahlavi, from those of Neryosangh in Sanskrit, and from the still later ones in Parsi and Persian. Here, again, those who read the P ahlavi only as rendered by N eryosangh neefl great caution. If Neryosangh is simply read like the classical Sanskrit, great errors will be committed. He needs a glossary of his own, and should be read solely in the light of the Pahlavi which was chiefly his original. So of the Pars! Persian translations, they 'must be read with especial attention to their originals. After these original translations have been fully mastered, and compared with an improved rendering of the Gathic, likewise also studied in the full light of the Veda, the patient scholar will be sur- prised at the result. He will find that to a certain exten- sive degree, the two sources of information coincide when reasonably estimated, and, moreover, that where the Pah- lavi gives us an indication differing from that derived from the Vedic, the surmise of the Pahlavi is the more often correct. I say reasonably estimated, 1 * 3 for not only Is the Pahlavi, as a less highly inflected language, incapable of rendering the Avesta literally, but its authors do not. uni- formly make the attempt to do so; nor do they always follow the order of the Gathic or Zend. Their translations generally ran w r ord for word as to their outward forms, for the ancient interpreters probably regarded such a following as essential to a complete rendering, but they found them- 1 I would here state- to the distinguished scholars who have clone me the honour to study my work on the Gathas, that the Pahlavi translations contained in it are those made in the light of the glosses. Here and there final ones will be added in a later volume, as from the Pahlavi texts sometimes considered apart from the Pahlavi glosses, and in consequence often much nearer the Gathic than those from both text and gloss.
xl THE X3ATHAS* selves compelled to resort to the most Important excep- tions. And, lastly 5 the rejection, or total neglect of the Fahlavi translations and their successors, on the ground that they contain errors, Is a policy which seems, to me defective, and to the last degree. What absurdities can Siya;za be capable of, and yet who would utter final opinions upon the i?/g-veda without either the ability, or the attempt, to read Saya^a 1?
It is hardly necessary to mention that the restoration of texts goes hand in hand with translation. For how are we to interpret a passage before we know that it exists? And of what inestimable worth are the Pahlavi translations as evidence to texts! Who does not see that where the ancient scribe is most free or erroneous as to form, or root, his rendering often shows plainly which of two words stood before him in his manuscripts. Our oldest MS. (that of Copenhagen, numbered 5) dates from the year 1323 A.D.; and what were the dates of the ancient documents before the eyes of the Pahlavi translator who writes in it?
We must now ask whether our present Pahlavi transla- tions are improvements upon their predecessors, or the reverse. That they are improvements in some few in- stances is undeniable, for, as we have seen, some of the glosses to them from later hands give the truth where the text is wide. But the glosses which show a later origin are, for the most part, inferior in richness to the texts. Here and there a talented, or fortunate, Parsi threw new light on the subject, but the general tendency was one of deteriora- tion; that is, before the revival of Parsi- learning under Neryosangh (400-500 years ago). This deterioration would naturally decrease as we approach successive periods in going back to the time when MSS. of the Gathas existed according to positive evidence, that is, to the time when, according to the An/a Viraf, Alexander’s servants found skins at Persepolis on which the Avesta had been traced in Well has Geldner mentioned the * epoch-making 5 Etudes Iraniennes of Darmesteter (KZ. vol. xxviii, p. 186). It is to-be hoped that these brilliant pieces will stimulate the study.of the relation between the Zend and the New Persian through the Ancient fersian and the Pahlavi.
' INTRODUCTION* xli gilded letters (for it Is not positively proved that the in- formants of Herodotus heard the Magian priests singing their s theogonies ’ from written books). At each of these periods scholarship is proved to have been competent by the results which it accomplished. The first of them we must place in the sixth century when, on Spiegel's estimate 1, the Zend characters were modified into their present lucid form from the Pahlavi, and distinct short vowels took the place of the unknown signs which existed previously. Then all MSS. which were to be found must have been collected and copied, and, so to speak, re-edited; and here we must accordingly place a period when the' Pahlavi translations were more valuable than those of any later date. As we go further back we come upon another period, when, under Shapur II, Adarbad Mahraspend brought the surviving portions of the Zend-Avesta together (about A.D. 33b). Still earlier the servants of Artaxerxes, the Sasanian, col- lected yet more abundant writings, when Zarathiutrianism was instituted as the state religion. Then, under the Arsa- cids (possibly under Vologeses the first), those most competent in the realm were directed' to gather the then extant documents.
While, if we hold that the entire Avesta was written originally in some character different ' from the Pahlavi, we must finally infer the existence of an early epoch, when the entire Avesta was brought over in its bulk from the earlier East (or West?) Iranian character in which it was first inscribed. If this character differed radically from the Pahlavi, this transliteration must be regarded as one of the most remarkable of literary events. Notwithstanding all the now rapidly corrected errors, the texts have been handed down with the minutest distinctions of dialect preserved 2, and this proves the existence of competent interpreters at a period practically contemporaneous with the composition of the later portions of the later Avesta. What commentaries must then have existed, not free from 1 Eranisches Alterthumslamde III, s. 767.
3 See Hiibschmann, KZ. bd. 24, s. 326.
xlii THE GATHAS.
error, as we see from the Zand of the Avesta, but, as to language and general sense, how close l Even if the degree of linguistic knowledge increases only gradually or steadily in " going back, without any epochs from tjic time of N eryosangh to the inferable date of the latest Zend writings, and if the character in which the Avesta was first recorded (after a lengthy life as an orally extended lore) differed only as to mode and fashion, and not radically, from the Pahlavi (which, so far as the later Avesta is concerned, is most probable), we have yet the transliteration of the Gathas to account for, ‘which perhaps were 'brought over (after long oral life) from the so-called Aryan character, while the existence of a gradual tradition of a scholarship does not refute the fact that this scholarship must have been at times of the highest cha- racter; it makes high scholarship more probable.
What translations, we again remark, may have ex- isted among these early sages I And, if they could once make translations fresh from the exegesis of the latest Zend writers themselves, is it not practically certain, considering the tenacity of life manifested by Zoroastrianism, that their explanations still lurk in the commentaries which have come down to us. And if these inferences be at all correct, how should we labour to discover from our present transla- tions what these predecessors were; and what scholar cannot perceive that gems of evidence as to texts and sense may yet linger in those of our present Pahlavi translations which may yet be otherwise most filled with phantastic error? And shall we not therefore conclude that their ex- pected inaccuracies, whether small or great, cannot destroy their inherent value? What, then, are we to think of it, when the New Persian, a quasi-daughter of the Pahlavi, is superficially referred to for linguistic analogies, when even the Armenian is also scanned, while the Pahlavi is left un- mastered? Is a quasi-mother language of the New Persian any, the less likely to afford linguistic analogies because an actual translation of the Avesta -has been attempted in it, and because the Avesta once stood in its characters, while it may also present claims to be considered to a certain • limit a daughter language to both the Gothic and Zend?
INTRODUCTION. xliii INTRODUCTION. xliii And should the acknowledged difficulty of the character continue to be a 'reason for avoiding all efforts to make it out 1?
In the endeavour to divide, our Avesta texts into originals and gloss, we are greatly aided by the metre. Interpolated words and phrases are often obvious at a glance, and we should never suspend our efforts to discover all the traces of metre which exist In the Avesta, as a necessary step to the restoration of the documents to their first form; but we should avoid exaggeration, and a carelessly dogmatic pro- cedure in insisting upon reducing lines to an exact, or to a supposed exact, number of syllables 2. I regard it as un- wise to suppose that the metrical lines of the Avesta, or indeed of any very ancient poetical matter, have been com- posed with every line filed Into exact proportions. The ancient poets would have brought out the measures in many a place by accent and a sandhi which are no longer known to us. The Vedic Hymns may, to a great extent, form an exception, but who would not say that where uniform even- ness is at hand, an effort to improve the metre has often corrupted the text. Priests or reciters of intelligence would here and there round off an awkward strophe, as year after year.they felt the unevenness of numbers. Metre must inevitably bring a perfecting corruption at times, as a de- ficiency in the metre must also prove a marring corruption. Cases should be carefully discriminated. The expression of passionate feeling, for instance, would be likely to cause 1 One of the most powerful tributes ever paid to the Pahlavi translators was Haug’s conversion to them. Before studying them he lost no opportunity to stigmatise their deficiencies; later, however, he followed them in many an im- portant place, and sometimes with little reserve.
As writers of the opposed extremes seem honestly convinced of the radical error of each other’s views, it is obvious that association and interest have much to do with decisions. A scholar should put himself fully under the influence first of one school and then of the other. The necessity for well-balanced studies is extremely great.
2 It is only lately that the variation from eleven to twelve syllables in the lines of Trish/up has been applied to the Gatfaic metres, nor has the possibility of a shifting caesura been acceded to till lately.
xliv THE GATHAS.
unevenness in lines. The language would be vigorous and idiomatic, and of unusual value as a fragment of ancient phrase, but the metre would have suffered.
Then as to conjectured texts; after texts have been •improved from all available relics of ancient tradition, or scholarship, as afforded by the Pahlavi translations, and from the evidence of metre, we are at times still left with readings before us which could not have been original. The composers have indeed here and there constructed sentences which they either could not, or would not, make easy, but as a general thing we may say, that where the text, as it stands, gives no satisfactory sense to us, after we have ex- hausted the resources of previous Asiatic scholarship, or direct analogy, in our efforts to explain it, it is in that cas e not the text as the composer delivered it. We are then reduced to conjecture, for how are we to translate a text before we are certain that it is integral? Our first efforts should be directed to the detection of losses; for a text may still be of great value when considered- as a mass of broken sentences, for, if we are certain that such is its character, we can often fill out the missing members with much proba- bdity, But whether we insert supplementary conjectures, or merely bracket later interpolations, we must by all means in cases of real necessity make the effort to amend the text (as also in the Veda).
Even if we fail in our attempted improvements, we are often little worse off than before, for whereas it is possible, or even probable, that the composers wrote what we sug- gest, it is sometimes not possible that they wrote exactly what stands in our texts. We should even suggest alterna- tive readings where our present ones are only less probable (tor the suggestion of an alternative is not the wholesale estimation of a sentence), while even when we declare their outcomrag meaning totally unsatisfactory, the MSS. still re- mam to other writers to begin on afresh. And in estimating what would be reasonable meanings, we should guard care- fully against both extremes, and we should especially exer- cise a strong negative criticism against the recognition of INTRODUCTION.
xlv INTRODUCTION.
xlv too much meaning, or too subtle a meaning. Profound and subtle conceptions placed where we are obliged to place the 6 G&thas, and other ancient portions of the Avesta, are indeed precious relics, as such conceptions at any age show a higher mental power, but we must doubt them only so much the more, and doubt, if we would be scientific and conscientious, till doubt becomes no longer possible. Beyond that we should turn our suspicions against our doubts themselves, which is the proper course if we would exhaust the meanings of the Gathas. Unless these are a fortuitous concourse of syllables, religiously profound modes of thought are manifest throughout. It is therefore strictly unscientific to force parts of them to express shallow details, and it is above all deplorable to change the text itself in order to produce out of it less enlarged meanings \ I say to force parts of them, for the great mass of them confessedly defies all attempts to reduce them to the statements of simple commonplace.
They can never possess the rich colour of the Riks; it is therefore the more to be deplored if we fail to see their deep, but awkwardly expressed, and oft-repeated thought. I must express my regret that until lately, when the enclitics have been more carefully considered, the form of sentences in the Gathas does not seem to have been noticed, writers conjecturing infinitives and simple accusatives at the ends of sentences. Both may, of course, fall there, but when we wish to reconstruct a word, we should not change it to a form which is not placed according to prevailing analogies. Infinitives and accusatives generally, both in the Gathas and the Rzg-ved a, avoid the end of the sentence. The accusative; when it falls there, is generally preceded by qualifying words often in apposition or agreement with it. Also in the conception of translations, authors seem to sup- 2 Non-specialists must not suppose that our texts are more apparently uncer- tain than (say) many portions of the Old Testament. Large portions of them are also as clear, at least, as the itVg-veda; and the emendations referred to need very seldom affect the doctrines. Let the learned public, however, insist on scholars making honest attempts to render the texts as they stand before their emendations, and greater harmony would Result.
x lvi THE GiTHAS.
x lvi THE GiTHAS.
pose it impossible that the lines can contain anything but lengthened prosaic sentences (too often with an accusative, or infinitive, pushed awkwardly out to the end). To me the Gdthic sentence is often very short, and so better adapted to poetic expression.
It has been already implied, and it has been taken for granted throughout 1, that the Avesta should be closely compared with the Veda, but let it never be forgotten, in the name of science, that the force and meaning of analo- gous words in the Gathic and the Vedic cannot be expected to be uniformly identical, considering the extent of territory, and the length of time, by which those who spoke the two languages were separated. The meanings of the Vedic words could not hold their own even in India, developing into the Sanskrit and Prakrit which differ widely, how truly mis- guided is it therefore to attribute necessarily the same shades of meaning to the terms of the two sister tongues. If even the Gathic hymns stood in the Indian forms, and had been discovered in India, having also reference to Indian history, no thoughtful writer would have rendered them in complete analogy with the Rzg-ve da. The Gathic usages would have been added in our dictionaries to those of the Vedic, just as the Sanskrit definitions are added.
An additional word seems called for as to the results of Zarathuitrian theology. Besides its connection with the modern philosophy through Gnosticism which has been already noticed 2, a relation between it and the Jewish theology since the Captivity has long been mentioned. The hagiology, the demonology, the temptation, the para- bles, the eschatology, have all been supposed to show traces of the time when Persian power was dominant in J erusalem, and with it, Persian literature; but the discussion of such questions requires separate treatises.
As to the general benefit which has resulted frpm Zara- thurtrlanism in the past, few reflections need to be added. If the mental illumination and spiritual elevation of many millions of mankind, throughout long periods of time, are of 1 See remarks in the Preface, p. xv.
3 See note on p. xix.
INTRODUCTION.
xlvii any importance, it would require strong- proof to deny that Zarathostrianism has had an influence of very positive power In determining the gravest results. That men should be taught to look within rather than without, to believe that suffering and sin do not originate from the capricious power of a Deity still called c good/ that the 4 good thought, word, and deed 5 should be recognised as essential to all sanctity, even in the presence of a superstitious ceremonial, that a judgment should have been expected according to the deeds done in the body, and the soul consigned to a Heaven of virtue or to a Hell of vice, its recompense being pro- nounced by the happy or stricken conscience, these can never be regarded by serious historians as matters of little moment, and if, on the contrary, they are allowed to be matters of great moment, the Zend-Avesta should be revered and studied by all who value the records of the human race.
ABBREVIATIONS.
Barth. — Barth olomae.
B.V. S. — Yendidad Sade, von Dr. Hermann Brockhaus. Leipzig, 1850.
D. « dastnr.
De inf. = De infinitivi linguarum sanskritae bactricae persicae graecae oscae timbricae latinae gotticae forma et usu, scripsit Eugenius Wilhelmus, ph.il. doctor. 1872.
G. — G&mSsp^i.
H. = Hiibschmann.
Inf. ~ Geschichte des Infinitivs im Indogermanischen, von Dr. Julius Jolly. 1873. K. — Kopenhagen MSS.