EDITORS' NOTE

The Revised Edition of Sri Aurobindo's Savitri is the result of a systematic comparison of the previously printed text with the manuscripts. The checking has included a detailed study of the various stages of copying, typing and printing — processes involving persons other than the author—through which the poem reached its published form. A substantial number of discrepancies due to accidents in the process of transmission have been discovered. The editors have critically examined these variations and have restored the original readings in most cases. The authenticity of the text, rather than subjective preference, has been the guiding criterion. This work, begun in 1979, has been carried out by members of the Sri Aurobindo Ashram Archives under the supervision of Nirodbaran and K. D. Sethna (Amal Kiran), who have made the final editorial decisions.

THE NEED FOR A REVISED EDITION

The first known draft of Savitri is dated 1916. Originally conceived as a medium-length narrative poem, by the early 1930s Savitri was assuming epic proportions and the status of a magnum opus. From August 1946 it started to appear in print, canto by canto, in journals of the Sri Aurobindo Ashram as well as in separate fascicles. The first edition of Part One was published towards the end of 1950, shortly before Sri Aurobindo's passing. The remainder of the poem, not all of which received final revision, appeared as a second volume the following year.¹

It might be thought that an ideal edition of Savitri should strictly reproduce the first edition. This seems plausible at first sight, at least with regard to Part One, since the proofs of this part were revised by Sri Aurobindo himself. Yet in fact the first edition contained serious errors. Some of these were removed in subsequent editions. The presence of these errors requires some explanation.

Until the mid-1940s, Sri Aurobindo continued to write out version after version of Savitri in his own hand, tirelessly expanding and

1 The first edition was described in the Publishers' Note as "extensively revised and enlarged" (with reference to the fascicles previously published). "Revised" here referred to the author's own alterations. In the present edition, revision in a different sense has been undertaken. The editors' intention has been not to improve upon what Sri Aurobindo wrote, but to remove accidental distortions for which he was not responsible.

Page-1

perfecting it. But when he began to prepare the poem for publication, he could no longer do all the work unaided. He took the assistance of two disciples, one of whom, Nirodbaran, made the final handwritten copies and the other, Nolini Kanta Gupta, the typescripts.

The deterioration of Sri Aurobindo's eyesight in these last years had two consequences affecting the text of Savitri. First, his later handwriting became increasingly difficult to read. This resulted in almost inevitable mistakes by the scribe who was asked to copy the hundreds of pages of manuscript. In the end, Sri Aurobindo came to rely entirely on dictation (to the same disciple) for the composition and revision of the poem. This opened the door to occasional inaccuracies of another kind.

The present edition is not the first to contain corrections. Each previous edition of Savitri has emended a number of errors noticed by the editors or brought to their attention by readers. Once a likely mistake had been observed, the manuscript was sometimes consulted for confirmation. But a systematic search for errors was not conducted until work began on the present edition.

EDITORIAL METHOD

The method of checking the text has been to trace the source of each difference between Sri Aurobindo's last manuscript and the printed version. Difficulties in the final manuscript were solved by reference to earlier manuscripts. This meticulous procedure confirmed the accuracy of much of the existing text. Most of the differences between the manuscript and the first edition were found to be changes dictated by Sri Aurobindo at some point in the process of continual revision through which Savitri assumed its final form. But a significant number of divergences due to slips or misreadings on the part of Sri Aurobindo's assistants were also identified.

The number of intermediate stages between the manuscript and the printed work varies in different parts of the poem. The manuscript itself had usually been revised by dictation before the scribe was asked to copy it. This included sometimes the addition or substitution of passages written on small note-pad sheets pinned to the MS. In some cantos the manuscript version was entirely replaced by a scribal version, partly copied from the manuscript and partly dictated, which was written on the backs of the MS pages or in a separate notebook. After this, there were generally several stages. Commonly we find a scribal copy, one or sometimes two or three typed copies, then the first printed version published as a journal instalment or

Page-2

fascicle, and finally the text of the first edition.

Sri Aurobindo normally revised each stage by dictation before the next transcript was made. The revision was often extensive and sometimes almost bafflingly complex. Partly because of this complexity, transmission of the text from one stage to another was liable to be less than perfectly exact. Some errors were later caught by Sri Aurobindo and corrected, not always in accordance with the original reading. Others passed unnoticed and remained in the published version.

Fortunately, the documents for almost all of the stages have been preserved. Thus it is usually possible to make a clear distinction between inaccuracies in transmission and deliberate changes made by Sri Aurobindo. The rule followed by the present editors in all but exceptional cases has been to accept the author's intentional revision, but reject variants due to accidents of transmission.

TREATMENT OF UNCERTAIN AND COMPLEX SITUATIONS

Most readers of Savitri would presumably wish to have a text in which each word is Sri Aurobindo's own. If so, it is clearly desirable that corruptions due to the vagaries of transmission be removed as far as possible. There is ordinarily no difficulty in identifying errors by the method described above. But uncertainties may arise where a stage is missing. The proofs of most of the journal instalments and fascicles and of the first edition of Part One have not survived. A comparison of the existing documents shows that substantial changes were made by Sri Aurobindo in revising the proofs. Some minor differences, however, could be due to compositors' errors. In the absence of direct evidence, such changes have as a rule been given the benefit of the doubt. They have been rejected as typographical errors only in a relatively few instances where the apparent inferiority or inconsistency of the printed reading makes it seem unlikely that the author's proof-revision was involved.

Even when there are no gaps in the evidence, the editors are sometimes faced with a complication. After a transmission error occurred, Sri Aurobindo may have revised the passage, yet left the variant intact. The reading may then be assimilated into the altered context in such a way that editorial intervention becomes problematic. The variant may have a claim to be retained in the text despite its illegitimate origin. Such cases and a few other exceptional situations are discussed in the Introduction to the Supplement in connection with the Table of Alternative Readings.

Page-3

THE PRINCIPLE OF THIS EDITION

Leaving aside special cases, it is the editors' view that textual corruptions that happened to escape Sri Aurobindo's notice do not thereby acquire a value equal to or greater than that of his own conscious choice of words. Copies containing transmission variants were orally revised by him without referring again to the documents from which they were copied. When he let plausible variant readings pass, there is no reason to think he was aware that his text had been altered.

This view could be challenged if it were believed that Sri Aurobindo remembered virtually every word of what he had written over the many years in which he worked on Savitri. In that case we could assume, when he did not correct inadvertent substitutions by his scribe and typist, that he actually preferred their readings to the words of his own inspiration. According to this hypothesis, he would have consciously accepted the unintentional creative collaboration of his assistants in the composition of the poem. But this seems improbable. Moreover, the scribe and typist themselves have been consulted on this point. They have made it clear that they do not wish their unwitting contributions to remain as a permanent feature of Savitri.

As has been mentioned, some errors in the first edition of Savitri were emended in subsequent editions. The most notable instances relate to dictated lines in which similar-sounding words had evidently been confused by the scribe. A few further emendations of the same type have been made in this edition. In locating cases of scribal mishearing, the present editors have had the advantage of knowing which passages were dictated and they were alert to the possibility of such slips. But this type of emendation, lacking concrete evidence, has been resorted to with the utmost caution and only where there is no reasonable doubt about Sri Aurobindo's intention.

PUNCTUATION AND CAPITALISATION

Details such as punctuation and capitalisation have been handled in this edition in essentially the same manner as the words of the text. The manuscripts for Savitri were in general carefully written with regard to all details, apart from some late fragmentary drafts which are nearly illegible because of Sri Aurobindo's failing eyesight. But non-verbal details suffered even more frequently than words from accidents of transmission. As a rule, the editors have restored the punctuation, hyphenation and capitalisation of the manuscript wherever

Page-4

these were accidentally altered in transmission, unless later revision of the context by Sri Aurobindo has made the original details inappropriate or irrelevant.

With regard to dictated matter a slight modification in the approach was required. Even when he dictated, Sri Aurobindo very often attended to the details as well as to the words. But this was not invariably the case. In dictating a new passage, his primary concern was with getting the words on paper. The details were sometimes left for later. The transmission of dictated matter has to be viewed with this in mind. On occasion the scribe and typist filled on their own the gaps in punctuation left in the process of dictation, for it was not always possible to consult Sri Aurobindo directly. Sri Aurobindo revised some of this punctuation, but much of it he tacitly accepted when the copies were read to him at a later time.

The editors have accordingly refrained from undue tampering with appropriate punctuation or capitalisation supplied by Sri Aurobindo's assistants in transcribing dictated passages. Further details introduced in previous editions have been accepted where they seemed justified. Some new editorial punctuation has been added in a few places where it appeared necessary. Where capitalisation, hyphenation and spelling in dictated matter were found to be at variance with Sri Aurobindo's practice in passages in his own hand, the editors have considered it legitimate to normalise these details.

SPELLING OF SANSKRIT NAMES

Sri Aurobindo's spelling of the names of the characters of his epic changed over the years without arriving at complete consistency. However, the change was in a definite and easily discernible direction which agrees with general modern practice in the transliteration of Sanskrit. In the present edition, consistent spellings have been adopted which are supported to some extent by the manuscripts.

In early manuscripts, the names of the three principal human characters in the story were written "Uswapathy", "Savithri" and "Suthyavan". Later, Sri Aurobindo settled on the spellings "Savitri" and "Satyavan", but in the relatively few places where Savitri's father is mentioned by name, he wrote variously "Aswapathy", "Aswapaty" and "Aswapati". Satyavan's father appears even in late manuscripts as "Dyumathsena" as well as "Dyumatsena". Since the "th" in "Aswapathy" and "Dyumathsena" belongs to the old style of transliteration which Sri Aurobindo was gradually abandoning, spellings without the "h" have been adopted in this edition. "Aswapati" has

Page-5

been chosen over "Aswapaty", though the number of manuscript occurrences of the two is about the same. The spelling "Aswapati" occurs as early as 1936-37 in some of Sri Aurobindo's letters on Savitri. It is also found in his handwritten note (c. 1946-47) printed as the "Author's Note" at the beginning of this edition.

SRI AUROBINDO'S LETTERS ON Savitri

In a letter dictated in 1946, Sri Aurobindo mentioned that he wanted to write "an introduction to Savitri when it is published".² This introduction never materialised. However, in the remainder of his letter Sri Aurobindo dwelt at some length on "questions of the technique of mystic poetry" which were to have been discussed in the intended introduction.

Other letters of the same period dealt with these questions in further detail. About ten years earlier, Sri Aurobindo had written his comments on points raised by a disciple to whom passages in the first two books of Savitri had been privately sent. Taken together, these letters give the author's own insights into the poem and provide the best available substitute for the unwritten introduction.

A selection of Sri Aurobindo's letters on Savitri was first published in 1951, separately from the first edition of the poem. In the editions of 1954 and 1970 a somewhat different arrangement of the letters was included at the end of the book. This is retained in the present edition, with a few additions in the first section. There has been some revision of the footnotes and a few textual corrections have been made after consulting the manuscript.

2 Savitri (1993), p. 735.

Page-6

INTRODUCTION TO THE SUPPLEMENT

The Revised Edition of Savitri has been published without the encumbrance of footnotes and appendixes, keeping the normal reader in mind. This Supplement provides information for those with a special interest. The principles and methods on which the edition is based have been explained in the Editors' Note. The remainder of the Supplement consists of a Table of Alternative Readings, a Table of Emendations, a Table of Line Numbers by Canto, and a selection of unused versions and omitted passages. Each of these sections is introduced below.

TABLE OF ALTERNATIVE READINGS

The Table of Alternative Readings is a list of variants deserving the special attention of the careful reader. It does not include the numerous simple transmission errors corrected in this edition; for these, see the Table of Emendations. Many variants of punctuation and capitalisation might have been treated as alternatives, but these have not been listed unless the meaning is significantly affected.

Various textual situations have given rise to the pairs of readings listed in this table. These are described below in terms of eight categories. An example of each type is discussed in the next section of this Introduction.

(1) In the most frequent cases (categories 1 and 2), two authentic readings exist, with no evidence that Sri Aurobindo made an explicit choice between them. This happens most often when he corrected a transmission error by changing it to something different from his original wording. It cannot ordinarily be assumed that he remembered the earlier version and consciously altered it.¹ Traces of the influence of the scribal error may even survive in the final revision. Thus the restoration of the original reading may sometimes be justified. Whichever reading is selected for the text, the other version is listed as an alternative. (Cases in which the earlier reading has been

1 In contexts where only two or three obvious words would be appropriate, the original word might easily have suggested itself again to Sri Aurobindo, whether he remembered the line or not. His last version has normally been accepted in such instances. The assumption that Sri Aurobindo did not always remember the exact wording of his own lines is based primarily on the fact that plausible yet clearly inferior readings due to miscopying or mistyping were sometimes not corrected by him when he revised the transcripts.

Page-7

restored are items 8, 12, 22, 29, 32, 39, 41, 46, 52, 62, 68, 71, 74, 81, 103, 117, 118, 124, 126, 132, 137, 139, 142, 158, 162, 164, 166, 176. The later reading has been retained in the text in nos. 2, 13, 34, 40, 42,.44, 45, 48, 66, 73, 75, 92, 102, 128, 131, 133, 138, 140, 145, 147, 153, 154, 156, 159, 163.)² A special case is where the revision was prompted by an apparent slip of the pen, inadvertent substitution or omission by Sri Aurobindo himself (27, 58, 79, 80).

(2) Occasionally a word, phrase or line was written between the lines or in the margin as a possible replacement for the original reading, but Sri Aurobindo's final choice was not indicated by the cancellation of either reading. If so, one version has been selected for the text of this edition, listing the other as an alternative. (Instances of this type are items 38, 51, 63, 89, 115, 135, 146, 171, 174, 186.)

(3) Sometimes a reading which had originated due to a transmission error was left intact by Sri Aurobindo in the course of significant revision of its surroundings. It may then merit special treatment despite its illegitimate origin. Depending on the nature of the later revision, it may be listed as an alternative (3, 20, 30, 43, 70, 77) or, if it has been virtually assimilated into a new context, it may even be retained in the text (10, 35, 49-50, 53, 59, 82, 175). The original reading is then recorded in the list of alternatives.

(4) Very rarely, a transcription error happened to reproduce a reading found in one or more earlier versions. It then has the effect of accidentally undoing Sri Aurobindo's latest revision. The editors have assumed that this revision represents his conscious choice, and the text is emended accordingly. Yet the reading which reappeared by chance and passed into early editions of the poem is not an altogether inauthentic one and may deserve mention as an alternative (5, 23, 78, 85, 119).

(5) The same reasoning applies to the more common situation where an alteration made by Sri Aurobindo was not carried out in the next stage of the transmission of the text. This happened most frequently when dictated revision was transferred from one copy to another, such as a carbon copy of the same typescript. In the process of transferring the changes, something was occasionally missed. These previously overlooked revisions have normally been incorporated in the text of this edition; the original readings, however, are mentioned as alternatives (4, 6, 7, 9,15-17, 21, 26, 28, 31, 36, 54-57,

2 In one instance (178), Sri Aurobindo's revision is believed to have been influenced by the scribe's mishearing of a dictated word rather than by a transmission error in the usual sense. Here the emended original line is given as an alternative to the final version.

Page-8

61, 69, 76, 83, 90, 93-95, 97-99, 101, 104-8, 111-14, 116, 117, 121-23,125,129,144,155,157).³ In a few cases, this order is reversed or a differently revised version is accepted as the final text (24, 96, 120, 127).

Other alternatives arise when there is insufficient evidence to be sure of the intended reading (categories 6 and 7):

(6) One type of uncertainty arises when the evidence for a stage of revision is lost. Some changes appearing in the fascicles and journal publications or in the first edition, such as variations involving a single letter, might be due either to compositors' errors or to intentional revision of the proofs. Most of these proofs have unfortunately not survived. In cases of genuine doubt, the editors have selected the reading which seems most likely to represent Sri Aurobindo's intention, listing the other reading as an alternative (1, 18-19, 25, 33, 64, 67, 86, 109, 110).

(7) In taking dictation the scribe sometimes misheard, confusing two words that are similar in sound, or otherwise misunderstood Sri Aurobindo's intention. Often such errors are obvious and can be corrected without hesitation; most emendations of this kind were made before the present edition. But in a few cases there is some doubt. If the suggested reading is only a plausible conjecture, it may be listed as an alternative rather than being adopted in the text (72, 136, 160, 168, 177, 179, 181, 183-85).⁴ An emendation for which there is strong but perhaps not unquestionable support may be printed in the text, what was written by the scribe being mentioned in the Table of Alternative Readings so that the readers can judge for themselves (11, 37, 60, 65, 148, 150, 167, 169, 170, 173, 180).⁵

(8) In rare instances, a final change made by Sri Aurobindo (or by the scribe acting on his instructions) appears authentic, yet its sense is problematic. An earlier straightforward reading may then be mentioned as an alternative (149, 151, 152, 165) or even accepted in the text if there is very strong justification for it (87).

Still other special cases occur, which have to be dealt with on an individual basis (47, 84, 88, 91, 100, 130, 134, 141, 143, 161, 172,

3 Some of the revised readings (76, 93-95, 97, 99) were introduced in the second edition.

4 In 72, 177 and 181, readings found in one or more past editions have been rejected from the text but are noted as possible conjectural emendations.

5 Item 14 is similar but does not involve an emendation of the text of previous editions. Here the reading introduced by the typist has not been dismissed as a transmission error but accepted as a plausible correction, the word written by the scribe being noted as an alternative.

Page-9

182).⁶ Often the editorial decision between alternative readings cannot be made by any simple rule. In view of the complexity, diversity and importance of these items, a brief description of the circumstances of each case is given in the notes at the end of the table. A full justification of the choice of textual readings would require a more detailed discussion of many of the items and sometimes the reproduction of facsimiles of the documents in question. This is not attempted here.

EXAMPLES OF ALTERNATIVE READINGS

Textual situations giving rise to alternative readings are illustrated below with one example of each type defined above. Some of the most complex and important cases have been selected, especially those which could not be described adequately with the concise form of presentation used in the notes following the table.

(1) Near the end of Book Six, Sri Aurobindo wrote the line:

Holding the ideal's ringed and battered fort

(459.7)⁷

This line is found in one of the small note-pads Sri Aurobindo used in the last years before he stopped writing in his own hand altogether. When the scribe transferred the line from the note-pad to his copy of an earlier version, the just legible word "ringed" was misread as "seized". After this copy was typed, Sri Aurobindo revised the typescript by dictation. He deleted the words "seized and". He preserved the metre by inserting "Or" at the beginning of the line, so that it became:

Or holding the ideal's battered fort

This version was the result of Sri Aurobindo's final revision. Yet it is doubtful whether he would have dropped the word "ringed" if it had been correctly deciphered when the line was transferred from the note-pad. He does not seem to have deliberately altered the original line to the later one. There are, therefore, two authentic versions and it does not appear that Sri Aurobindo consciously chose one over the

6 A common feature of several of these items is some unclearness or ambiguity in Sri Aurobindo's or the scribe's markings. For example, in three cases a word was altered to another word in such a way that it is not certain which is the final reading (84, 161, 172). Other items are closely related to one or more of the situations described above, and have been handled according to the same principles.

7 References to the Revised Edition of Savitri will be indicated by the page and the line number counting from the top of the page (excluding headings and blank lines).

Page-10

other. In this edition the original, more forceful line has been printed in the text, while the later one is listed in the Table of Alternative Readings.

A similar choice has been made in a number of other cases of the same type. However, the editors have not made it a fixed rule to select the earlier reading in such situations. The final reading has often been retained in the text if there is no good reason for changing it. If so, the reading that preceded the transmission error is mentioned as an alternative.

(2) The following line occurs in the final manuscript of Book Two, Canto Six:

This to Life's music gives its anthem swell.

(194.34)

In the margin next to it, Sri Aurobindo wrote another version of the same line:

A million motives in Life's music swell.

The original line was not cancelled. When the scribe made his fair copy of this canto, he copied the first version of the line and disregarded the one written in the margin. The line copied by the scribe was later left unaltered when Sri Aurobindo, in revising the fascicle, dictated new lines to be inserted just before it. This version has been kept in the text of the present edition. The other line is noted in the Table of Alternative Readings.

(3) In his last handwritten version of Book Three, Canto Three, Sri Aurobindo wrote these lines:

A plenitude of illimitable Light

Inspired the passing act, the moment's thought.

A wisdom worked in all self-moved, self-sure,

An authenticity of intuitive Truth

In a glory and passion of creative Force.

When the manuscript was revised by dictation, this passage was rearranged using arrows and transposition signs. Punctuation was partially overlooked in the process, but if we supply commas after "Light", "Truth" and "Force", the revised passage reads:

A wisdom worked in all self-moved, self-sure;

A plenitude of illimitable Light,

An authenticity of intuitive Truth,

A glory and passion of creative Force,

Page-11

Inspired the moment's thought, the passing act.

The most important changes were in the order of the lines. But in the line which now comes at the end, the order of the words was also altered: "the passing act" and "the moment's thought" were interchanged with a transposition sign. However, in copying this line the scribe misinterpreted the mark and wrote:

The moment's thought inspired the passing act.

This introduced an idea quite different from what Sri Aurobindo had intended. Yet when he revised the copy, he apparently did not notice the mistake. Instead of correcting it, he worked the line into the passage by adding a new line before it:

Infallible, leaping from eternity,

The preceding lines were punctuated to make a separate sentence, so that the passage finally became:

A wisdom worked in all, self-moved, self-sure,

A plenitude of illimitable Light,

An authenticity of intuitive Truth,

A glory and passion of creative Force.

Infallible, leaping from eternity,

The moment's thought inspired the passing act.

(324.32-325.1)

The idea in the last line was not originally Sri Aurobindo's. An unusual scribal error made "the moment's thought" the subject of the verb "inspired" instead of its object. Such mistakes are normally corrected in this edition. But here, Sri Aurobindo's creative solution to the problem justifies another editorial approach.

The new line, "Infallible, leaping from eternity," was added to explain the miscopied line. It is the "moment's thought" which Sri Aurobindo described as "leaping from eternity". If "inspired" was restored to the beginning of the last line and a comma was put after "Force", the description "Infallible, leaping from eternity," would come to apply to the "plenitude", etc., in the preceding lines. This might be possible, but it was not what Sri Aurobindo meant to say when he inserted the new line. The added line cannot very well be omitted. Therefore, in this edition the passage is kept intact as it stood after Sri Aurobindo's final revision. The earlier sentence, as it was intended to read before the copying error occurred, is given in the Table of Alternative Readings.

Page-12

(4) In the final manuscript of the last canto of Book Three, Sri Aurobindo wrote the line:

No more can earthly limits pen thy force;

(340.34)

The scribe copied the third word as "in" instead of "can". When the copy was read to Sri Aurobindo, he did not notice anything wrong with it. Indeed, in the penultimate manuscript he himself had written "in". There the line had been worded exactly as it again became in the scribe's copy of the final MS:

No more in earthly limits pen thy force;

This is how the line was printed in the first edition. The reversion to this earlier version seems to have been purely accidental. Since in his final manuscript Sri Aurobindo deliberately wrote "can" in place of "in", the editors have adopted "can" as the textual reading in the present edition. This agrees with the general policy of rectifying transmission errors. In this case, no significant changes in the context intervened later to complicate the picture. Nevertheless, it is a fact that Sri Aurobindo himself had written "in" at an earlier stage, as well as letting it pass when it happened to reappear in the final version. It is not a wholly inauthentic reading like most transmission variants. In view of this, it is listed as an alternative reading.

(5) The following sentence for Book Two, Canto Four was written by Sri Aurobindo in a small note-pad:

Life had for them no aim but natural joy;

They worked for the body's wants and craved no more,

Satisfied to breathe, to feel, to sense, to act,

Identified with the being's outer shell.

The three-page passage in the note-pad where these lines occur was read to Sri Aurobindo by the scribe, who clarified some of the less legible words in his own hand and marked a few dictated alterations. The scribe then copied the passage neatly on small sheets similar to those used by Sri Aurobindo. The copy of the above lines, in which there were no dictated changes, was exact except that "breathe" was misread as "battle".

The fair copy was revised by dictation and attached to the manuscript containing Sri Aurobindo's last handwritten version of most of this canto. In the first of the above lines, "but" was altered to "save". The miscopying of "breathe" as "battle" was corrected. In addition, arrows were marked shifting the second line to the end of the sentence

Page-13

and the third line to the beginning, punctuation being adjusted accordingly:

Satisfied to breathe, to feel, to sense, to act,

Life had for them no aim save natural joy;

Identified with the being's outer shell,

They worked for the body's wants and craved no more.

Normally, the next step would have been for the scribe to copy the passage along with the rest of the canto into the large ledger used for his copy of the whole of Part One. This copy was ordinarily made after all revision of the manuscript and attached passages was finished. But in this case, the sheets pinned to the manuscript were for some reason revised after the passage had already been copied into the ledger. This meant that the changes marked on them had to be transferred afterwards. The verbal changes were correctly transferred to the final copy, but the arrows in the margin were overlooked in the process.

The final scribal copy of the later portion of this canto was not revised, but the typescript made from this copy received significant revision. In the present lines, "natural" was changed at this stage to "Nature's", "Satisfied" to "Content" and "being's outer" to "spirit's outward". The order of the lines and the punctuation remained as in the first untransposed version:

Life had for them no aim save Nature's joy;

They worked for the body's wants and craved no more,

Content to breathe, to feel, to sense, to act,

Identified with the spirit's outward shell.

When Sri Aurobindo revised the proofs of the 1950 edition, he amplified the idea in the first line by adding a new line and altered "wants and" to "wants, they" in the line beginning "They worked":

Life had for them no aim save Nature's joy

And the stimulus and delight of outer things;

They worked for the body's wants, they craved no more,

Content to breathe, to feel, to sense, to act,

Identified with the spirit's outward shell.

This is the text printed in all previous editions. It contains Sri Aurobindo's last revisions and is not marred by any positive transcription error introducing an unauthentic element. The order of the lines is based on the original manuscript and Sri Aurobindo himself did not find anything wrong with it when he revised the typescript and

Page-14

final proofs. Yet at a previous stage he had seen that the sentence could be improved by rearranging the lines. It may be noted that the changes he dictated later are such as could equally well have been made if the intended transposition had been transferred to the scribal copy and carried out in the typescript. The sentence would then have read:

Content to breathe, to feel, to sense, to act,

Life had for them no aim save Nature's joy

And the stimulus and delight of outer things;

Identified with the spirit's outward shell,

They worked for the body's wants, they craved no more.

(143.12-16)

This incorporates the results of all of Sri Aurobindo's work on these lines and is the text printed in the present edition. The transposition he intended to make has been combined with his later revision. It is an "eclectic" version in the sense that the entire sentence does not occur in this exact form at any single stage of its transmission and revision. The advantages of such a text seem in this case to outweigh the possible objections. The reasons for preferring the transposed order of the lines—such as the stronger ending it gives to the sentence — are as applicable to the lines in their final form as they were earlier. Nevertheless, the version found in previous editions has a good claim to be listed as an alternative.

(6) In Canto Three of the 1947 fascicle version of the first six cantos of Book Two, the following lines appeared:

Our human ignorance moves towards the Truth

That Nescience may become omniscient:

Transmuted her instincts change to divine thoughts,

Thoughts into infallible immortal sight

And Nature climbs towards God's identity.

Before the first edition of Part One was printed, Sri Aurobindo revised a copy of this fascicle by dictation. He instructed the scribe to make a few changes in this passage. Among these was the cancellation of the "s" of "climbs" in the last line. A comma at the end of the second line was implied by this change but was not marked. With the other revisions, the lines were thus intended to read:

Our human ignorance moves towards the Truth

That Nescience may become omniscient,

Transmuted instincts shape to divine thoughts,

Page-15

Thoughts house infallible immortal sight

And Nature climb towards God's identity.

(121.23-27)

The fact that the original colon remained after "omniscient" in the revised copy of the fascicle was clearly an oversight, since "climb" in the last line has to be connected with "may" in the second. Yet when the first edition was printed, the colon was still there. So was the "s" of "climbs", which Sri Aurobindo had instructed the scribe to cancel. In the first and later editions, the passage reads:

Our human ignorance moves towards the Truth

That Nescience may become omniscient:

Transmuted instincts shape to divine thoughts,

Thoughts house infallible immortal sight

And Nature climbs towards God's identity.

It is conceivable that Sri Aurobindo changed his mind about "climbs" when he revised the proofs of the first edition. Since these proofs were not preserved, there is no objective way to distinguish a typographical error or unauthorised change from Sri Aurobindo's proof revision. But a careful analysis of his revision of the fascicle version shows that the deletion of the "s" of "climbs" was closely connected with the significance of the other verbal changes he made at the same time. A suspicion arises that the preservation of the "s" in the first edition was a mistake due to the presence of a colon after "omniscient".

The changes marked in the fascicle are the last indisputable indication we have of Sri Aurobindo's intention with regard to this passage. A comma after "omniscient", though not marked, was certainly implied. This, as in the second quotation above, is the form in which these lines are printed in the Revised Edition. The editors assume that the overlooked colon after "omniscient" in the fascicle may have contributed to the retention of the "s" in the first edition. However, since Sri Aurobindo's involvement at this stage cannot be ruled out, the version with "climbs" and a colon after "omniscient" is given as an alternative.

(7) In a passage dictated by Sri Aurobindo in the Epilogue as part of Savitri's final speech to Satyavan, these lines occur in the scribe's hand:

Heaven's touch fulfils but cancels not our earth:

Page-16

Our bodies need each other in the same last;

(719.24-25)

The expression "same last" has baffled attentive readers ever since it first appeared in print. Taking into account the implications of the first part of the line, the suggestion has been made that Sri Aurobindo might have said "lust", using this word as he sometimes did to mean simply desire — particularly the normal physical desire between man and woman — with no pejorative connotation intended. The scribe, it is supposed, could have heard it as "last".

Analogous confusions of similar-sounding words did occur in the dictated portions of Savitri. A few examples corrected prior to this edition are: "feeling" written by the scribe for "filling" (44.35), "lens" for "lance" (252.12), "wants" for "once" (256.3), "keen" for "kin" (264.35), "wicks" for "weeks" (350.27), and "melodies" for "maladies" (439.5).⁸ The relation between the vowels of "wants" and "once", one of the pairs of words confused, is somewhat similar to the difference, between "last" and "lust".

In a few places in his writings, Sri Aurobindo used the word "lust" in ways that might seem consistent with the present context. For example, in two of his letters to disciples he mentioned it in connection with Avatarhood:

Why should not Rama have kama (lust) as well as prema (love)? They were supposed to go together as between husband and wife in ancient India.⁹

What do you mean by lust? Avatars can be married and have children and that is not possible without sex....¹⁰

Of course, "lust" can also mean a strong desire which is not at all sexual. In a philosophical context, one even finds Sri Aurobindo giving it a higher sense:

The impulse to realise that secret consciousness is the spur of the cosmic Divine, the lust of the embodied Self within every individual creature ....¹¹

⁸ Some of these were discovered before the first edition came out and so do not appear in the Table of Emendations. The correction of "wants" to "once" (256.3) was made in the 1976 impression of the third edition.

9 Letters on Yoga (1970), p. 418.

10 Ibid., p. 422.

11 The Life Divine (1970), p. 194.

Page-17

In Sri Aurobindo's poetry, the phrase "lust of the infinite skies"¹² may be noted, as well as a prominent occurrence of "lust" at the end of a line in his short poem, "The Mother of God":

She forces on the cold unwilling Void

Her adventure of life, the passionate dreams of her lust.¹³

It may be possible, then, to imagine Sri Aurobindo putting this word in the mouth of the heroine of his epic. Although his Yogic teaching calls ultimately for the elimination of desire, he did not share the Christian contempt for the flesh which has given "lust" its usual associations in English. Savitri's speech to Satyavan in the Epilogue does not evoke an impression of passionless sainthood. Two occurrences of the word "desire" are worth noting:

I am thy kingdom even as thou art mine,

The sovereign and the slave of thy desire,

(719.35-36)

Thy body is my body's counterpart

Whose every limb my answering limb desires,

(720.5-6)

Attempts have been made to explain "last" as a shoemaker's model or to take "in the same last" as an unusual way of saying "in the same way as before". These and other interpretations seem forced and unconvincing. Nevertheless, objections to the proposed emendation can be made on grounds other than mere prudery. The emended line would read:

Our bodies need each other in the same lust;

It has been observed that "with" rather than "in" might have been expected, as in these lines from Sri Aurobindo's Ilion.¹⁴

So might a poet inland who imagines the rumour of Ocean,

Yearn with his lust for the giant upheaval, the dance as of hill-tops.....

This is not a conclusive argument, however, for "in" is surely possible in the present line and would give a different shade of meaning which may be appropriate here.

We have seen that Sri Aurobindo could speak of lust, in the sense

12 Ilion (1989), p. 91.

13 Collected Poems (1971), p. 105.

14 Pp. 15-16.

Page-18

of physical desire,¹⁵ as an acceptable element in the relations between husband and wife, even in referring to an Avatar like Rama. His neutral use of this word in such a context is noteworthy in connection with Savitri's speech. But it must be acknowledged that he more frequently used the word in the usual manner to denote a form of desire, usually sexual, regarded as impure and degraded. Apart from the present debatable instance, there are fourteen occurrences of "lust" and five of "lusts" in Savitri. Everywhere its negative associations are felt to some degree. This is true even where the body's lust is recognised as serving a necessary evolutionary purpose, as in these lines:

The secret crawl of consciousness to light

Through a fertile slime of lust and battening sense,

(138.10-11)

After all has been said in its favour, Savitri's use of the word "lust" in her speech to Satyavan in the Epilogue would be a little unusual.

The rule followed in this edition with regard to such cases is that the text is not emended unless the editors feel quite certain of Sri Aurobindo's intention. Since some misgivings linger here, the cautious approach has been taken of suggesting "lust" as a conjectural emendation of "last" in the Table of Alternative Readings.

(8) The concluding passage of Book Three is found in more than two dozen versions in Sri Aurobindo's hand. In one of the later manuscripts, a sentence which had gradually taken shape through many previous versions was written in the following form (cf. 347.29-33):

Once more he moved amid material scenes

Lifted by intimations from the heights

And in the pauses of the building brain

Touched by the thoughts that skim the fathomless surge

Of Nature and wing back to hidden shores.

In subsequent copies of this passage, Sri Aurobindo changed the wording of the first line slightly, substituting at various times "lived" for "moved", "among" for "amid", and "things" for "scenes". In the last line, one manuscript has "swim" instead of "wing". Most versions had a comma at the end of the first line and some had commas in the third line after "And" and "brain". Otherwise, the lines remained the same until the final MS was reached. Here they returned to the form

15 Sri Aurobindo himself gives this definition: "This physical element [of the heart's love] may be purified of the subjection to physical desire which is called lust...." (The Synthesis of Yoga [1970], p. 623).

Page-19

quoted above in all but two details: a comma after "scenes" and the word "twixt" instead of "in" in the third line.

The latter change is puzzling. This line had first been inserted in a manuscript which represents roughly the mid-point in the evolution of the passage. After the original "his" before "building brain" was changed to "the", its wording had remained the same in a dozen manuscripts. But the last version reads:

And twixt the pauses of the building brain

Logically, the phrase "twixt the pauses" should mean the opposite of the original "in the pauses". For "twixt" means "between". The times between the pauses of the brain would be the periods when it is active. But this is probably not what Sri Aurobindo meant. It seems unlikely that he intended to give a contradictory sense to a line which he had written out consistently so many times. Moreover, in all of his writings on Yoga it is the quieting of the brain-mind, rather than the continuation of its normal activity, which is considered most conducive to the reception of higher influences like the thoughts from "hidden shores" in this- passage.

The replacement of "in" by "twixt" cannot quite be dismissed as a mechanical slip of the pen. However, it may be supposed that Sri Aurobindo made the substitution without noticing its misleading effect. Though "twixt" occurs in the last manuscript, it can be plausibly maintained that it does not convey the intended meaning as aptly as the earlier reading did. If so, there would seem to be good reason in this instance for making an exception to the rule that the text should follow the author's latest version. Because of the problems of interpretation raised by "twixt the pauses", the long series of manuscripts with the more straightforward phrase, "in the pauses", deserves special consideration. In the present edition, the text is printed with "in", while "twixt" is given as an alternative reading.

NOTES ON THE ALTERNATIVE READINGS

The notes at the end of the Table of Alternative Readings give details about the textual history of each item. The relevant facts of each case are presented in a concise form, employing for the sake of brevity and precision certain conventions which are illustrated by the following example:

13. (98.22) MS "Made for"; copied "Made from"; scribal copy altered to "Made by"

Page-20

This is the note on item 13 in the table, which concerns page 98, line 22 of the present edition:

Made by an interpreting creative joy

The absence of an asterisk in the table after the reading in the Text column, "Made by", means that the Revised Edition has the same reading as previous editions. "Made for" is listed as an alternative. The note describes the three stages in the history of this line that are essential for understanding the reasons for the alternative. Some formal features of the note may be observed: (1) distinct stages in the transmission and revision of the text are separated by semicolons; (2) the nature of each stage or operation is indicated by a word or phrase in italics; (3) words of the text at each stage (including any punctuation) are enclosed in quotation marks.

The above note indicates that the reading "Made for" occurs in Sri Aurobindo's last handwritten manuscript.¹⁶ When the scribe copied the manuscript, he misread "for" as "from". His copy was later read out to Sri Aurobindo, who dictated "by" in place of the miscopying, "from". This remained unchanged in all subsequent stages of the transmission and revision of this canto; these later stages are therefore not mentioned in the note on this item. The editors have assumed it to be possible that Sri Aurobindo did not remember the manuscript version when the scribe's copy was read to him. If so, he did not deliberately reject "for" when he changed "from" to "by". This is why "for" is listed as an alternative reading. Since this common type of alternative has been discussed in the first section of the Introduction (category 1), the facts of the particular case are presented in the note without further comment. Wherever necessary, however, explanations and clarifications are given between square brackets.

In order to understand these notes correctly, the reader should know the exact use of the following terms:

16 Apart from the words quoted, the rest of the line and its immediate context are in this case the same in the manuscript as in the printed version. Significant differences relevant to the reading in question would have been mentioned in the note.

Page-21

17 In Part One, which appeared in book-form in 1950, dates from 1946 to 1948 refer to the fascicles in which the cantos were first brought out. (Some cantos appeared at the same time in Sri Aurobindo Mandir Annual, The Advent, or Sri Aurobindo Circle.) Several cantos of Parts Two and Three were published between 1948 and 1951 in fascicles and journal instalments. The second volume of the first edition came out in 1951; the date 1951 refers to this edition unless a journal is specified.

Page-22

TABLE OF EMENDATIONS

The differences between the Revised Edition of Savitri and previous editions are listed in the Table of Emendations. The numerous corrections made in this edition are the result of a systematic checking of the text against the manuscripts, as explained in the Editors' Note. The more sporadic differences between the three previous complete editions (1950-51, 1954 and 1970) are also listed in the table. The 1954 and 1970 (Centenary) editions each contained some legitimate emendations of the first edition which have been accepted in the present edition, as well as some typographical errors and other changes which have been rejected. It should be noted that some readings identified in the table as appearing in 1970 are found in the 1968 reprint of Part One of the first edition. A few of the corrections made in the Revised Edition were introduced in later impressions of the Centenary Edition.¹⁹

The following typical entries illustrate how differences between editions are shown:

The first example represents the most frequent situation. This is an emendation made for the first time in the present edition. The absence of a date by either reading means that "peered" is the reading in the Revised Edition, while "pierced" is found in all earlier editions. The

18 Usually typed from the scribal copy, except in most of Book Four and part of Book Five, where there was no copy by the scribe and the typescript was made directly from the MS.

19 There are small differences even between the texts published in the Popular Edition and the De Luxe Edition of the Sri Aurobindo Birth Centenary Library. The date "1970" in the Table of Emendations refers to the De Luxe Edition. This text was reprinted in reduced format as a single-volume in the same year.

Page-23

correction, as in most such cases, is based on a comparison of the manuscript with the stages through which the text was transmitted. Here a copying error was found. The word "peered" was written by Sri Aurobindo in several successive versions. It was misread by the scribe, in copying from the last and most difficult manuscript, as "pierced".

In the other examples, dates in one or both columns show variations between previous editions. "Nature's soul" of the first edition was corrected to "Nature sole" in the second edition, as indicated by the date next to the present reading. The 1954 emendation has been accepted since it has the support of several manuscripts in Sri Aurobindo's hand. "Nature's soul" was evidently the scribe's mishearing of "Nature sole" when the final version of the passage was dictated to him by Sri Aurobindo.

The apostrophe in the phrase "the ages' weltering flood" (69.16) was shifted due to a typographical error in the 1954 edition so that the second word became "age's". This mistake was repeated in the Centenary Edition. The date "(1954)" next to "age's" in the last column of the table indicates that this reading did not occur in the 1950 edition. Since no other previous reading is mentioned, it is implied that the first edition had the present reading.

Sometimes two previous readings are listed, as in the fourth example. In this case, the first edition had the reading "black board", which was emended in 1954 to "black-board". In the Centenary Edition it became "blackboard"; this is shown by the date, "(1970)", next to the present reading. This reading has been accepted in the Revised Edition since it is also found in the manuscript, the scribal and typed copies and the fascicle. Though "black board" in the first edition could theoretically have been a change made by Sri Aurobindo in revising the proofs, it was more likely a typographical error.

Items marked with an asterisk (*) after the present reading are also listed in the Table of Alternative Readings. The alternative reading given in that table is usually the same as the previous reading noted in the Table of Emendations. Most of these items are more complex than the simple transmission errors primarily corrected in this edition. Detailed information about them is provided in the notes at the end of the Table of Alternative Readings.

TABLE OF LINE NUMBERS BY CANTO

It may be desirable for some purposes to make references to Savitri by book, canto and line number, independent of the pagination of

Page-24

the particular edition, which has varied in the past and may change in the future.²⁰ However, line numbers have not been put in the margin in this edition as these would be an unwanted distraction for many readers. To facilitate references, a Table of Line Numbers by Canto is provided in the Supplement. Each page number is listed along with the line number of the first line on that page in this edition, counting from the beginning of the canto. The number of lines in each canto is also mentioned, and the total number of lines in the poem is given at the end.

UNUSED VERSIONS AND OMITTED PASSAGES

Each book and canto of Savitri normally passed through a series of versions, often a long series. After each version was revised, a fair copy of it was made, by Sri Aurobindo himself or later by an assistant, which was revised in its turn. Every stage of this process usually contributed in some way to the final text. But occasionally, especially in the period when Sri Aurobindo's eyes had begun to fail and he was working with the help of a scribe, a version not containing his most recent changes and additions was taken as the starting-point for subsequent revision. Thus, the results of one or more stages of revision were not incorporated in the published version.

The unused revision of certain passages includes substantial alterations and new lines which Sri Aurobindo might have wished to utilise in the poem. Yet it cannot be proved with certainty that these versions were merely overlooked; in some cases they may have been deliberately rejected. Moreover, an unused version, however attractive its readings, cannot always be combined with the differently revised final version. Therefore, in this edition, lines and readings from versions not used by Sri Aurobindo have generally not been introduced into the body of the text. The most important passages are printed in this Supplement for their intrinsic value and as illustrations

20 The first three editions of Savitri differed from each other in pagination. For the readers' convenience, the pagination of the Revised Edition has been kept as close as possible to that of the Centenary Edition (1970), which has been widely referred to for more than twenty years. (This has meant perpetuating an anomaly in pagination, due to the fact that the title pages at the beginnings of Part One and Part Two of the two-volume Centenary Edition were not counted, but the corresponding pages in Part Three — which came in the middle of the second volume — were counted.) Each canto begins on the same page as in the Centenary Edition. Most lines will be found on the same page as previously, though the insertion and deletion of lines and blank spaces, the removal of footnotes, and changes in the typography of canto titles have caused some slight shifting of lines within cantos.

Page-25

of the complex process by which Savitri took shape. They include passages intended for Book Two, Cantos Six and Seven (sections A and B), Book Four, Canto Two (C), and Book Six, Canto Two (E).

A somewhat different case (D) is that of a passage existing in only one version which was omitted for unknown reasons. A dictated passage is found at the end of Book Five, Canto Three, in an otherwise unrevised duplicate copy of the typescript. Direct evidence is lacking that would show whether Sri Aurobindo finally rejected these lines or whether they were left out due to an oversight. The editors have thought it safer not to incorporate them in the body of the poem, but to print them separately in the Supplement.

In Book Seven, Canto Three, a footnote in the first edition gave a passage of nine lines (F) as an alternative version to the first twelve lines of the third section of the canto. The twelve-line passage, which was probably dictated, first appears in a typed copy. No record of the shorter version survives prior to its appearance as "another version" in the footnote in the first edition. In the present edition, the version previously printed as a footnote has been shifted to the Supplement. A verbal emendation introduced in previous editions has been accepted. The third line was printed in 1951 in a metrically defective form as "Where soul was not and thinking mind". It was later emended to "Where soul was not nor spirit, and thinking mind". The words "nor spirit" were taken from the fourth line of the other version, whose similarity to this line is the only close verbal resemblance between the two passages.

The misplacement of the original dictated ending of Book Seven, Canto Seven, led Sri Aurobindo to dictate a new and quite different ending. Before the first edition was printed, the original ending (555.11-556.13 in this edition) was found again and inserted before the new ending. A footnote, "Alternative version", was put in the first edition at the beginning of the last section of the canto, without indicating what was an alternative to what. The footnote was dropped in the second edition. Since there is no verbal repetition between the two passages and they read well one after the other, both have been retained in the present text as in previous editions. This explanation of their history is offered in place of the earlier footnote.

Page-26

TABLE OF ALTERNATIVE READINGS

TABLE OF EMENDATIONS

TABLE OF LINE NUMBERS BY CANTO

UNUSED VERSIONS AND OMITTED PASSAGES