M HN AK Y 
 
 ! HI 
 
 UNIVERSITY OF CALIFORNIA. 
 
 < . I 1- I < ) 
 
 \l 
 
 99685 Cla * s 
 
 sPH 
 

A STUDY OF THE ROMANCE OF THE SEVEN SAGES 
 
 WITH SPECIAL REFERENCE TO THE MIDDLE 
 
 ENGLISH VERSIONS 
 
 A DISSERTATION 
 
 PRESENTED TO THE BOARD OF UNIVERSITY STUDIES 
 
 OF THE JOHNS HOPKINS UNIVERSITY FOR THE 
 
 DEGREE OF DOCTOR OF PHILOSOPHY 
 
 BY 
 
 KILLIS CAMPBELL 
 
 FORMERLY FELLOW IN ENGLISH AT THE JOHNS HOPKINS UNIVERSITY 
 
 IVE 
 
 BALTIMORE 
 
 THE MODERN LANGUAGE ASSOCIATION OF AMERICA 
 1898 
 
A STUDY OF THE ROMANCE OF THE SEVEN SAGES 
 
 WITH SPECIAL REFERENCE TO THE MIDDLE 
 
 ENGLISH VERSIONS 
 
 A DISSERTATION 
 
 PRESENTED TO THE BOARD OF UNIVERSITY STUDIES 
 
 OF THE JOHNS HOPKINS UNIVERSITY FOR THE 
 
 DEGREE OF DOCTOR OF PHILOSOPHY 
 
 BY 
 
 KILLIS CAMPBELL 
 
 FORMERLY FELLOW IN ENGLISH AT THE JOHNS HOPKINS UNIVERSITY 
 
 BALTIMORE 
 
 THE MODERN LANGUAGE ASSOCIATION OF AMERICA 
 1898 
 
JOHN MURPHY A CO., PRINTERS, 
 BALTIMORE. 
 
 f 
 
 
 [Reprinted from the Publication* of the Modern Language Association 
 of America, Vol. XIV, No. 1.] 
 
U* 
 
 TABLE OF CONTENTS. 
 
 PAGE. 
 
 A Word of Introduction, 1 
 
 I. The Earlier History of the Romance, 3 
 
 I (a). The Romance in the Orient, 3 
 
 I (6). Transmission of the Romance to the Occident, - - 12 
 
 I (c). The Romance in France and Italy, 20 
 
 1. The Dolopathos, 21 
 
 2. The Sept Sages de Rome, 24 
 
 n. The Romance in England, 35 
 
 II (a). The Middle English Versions, 37 
 
 1. Description of the Manuscripts, 37 
 
 2. Interrelation of the Middle English Versions, 43 
 
 3. Authorship of the Middle English Versions, - 84 
 
 4. Source of the Middle English Versions, 87 
 II (6). Sixteenth Century and Chap-book Versions, - - 91 
 
 Appendix, 94 
 
 99685 
 
A STUDY OF THE ROMANCE OF THE SEVEN 
 
 SAGES WITH SPECIAL REFERENCE TO 
 
 THE MIDDLE ENGLISH VERSIONS. 
 
 A WORD OF INTRODUCTION. 
 
 The main object of this study has been to investigate 
 thoroughly the relations of the Middle English versions of 
 the Seven Sages of Rome. 
 
 As preliminary to this investigation, a review of the history 
 of the romance in the several stages through which it has 
 passed before reaching English has been made. This survey, 
 a recapitulation of the results which modern scholarship has 
 attained in the study of the romance, has been made im- 
 partially, and with a view to set forth the most approved 
 views that have been held rather than to advance any new 
 theories of my own. Where these views are conflicting, as is 
 particularly the case with respect to the eastern versions, I 
 have endeavored to sift truth from error, though here 
 naturally some difficulty has been encountered. It is only 
 on the question of transmission of the romance that a view 
 differing from that of the best authorities has been taken. 
 
 The chapter on the French and the Italian versions has been 
 based in large part on the work of Gaston Paris, whose Deux 
 
 I 
 
2 KILLIS CAMPBELL. 
 
 Redactions has superseded all previous contributions, repre- 
 senting as it does the most recent and the best results that have 
 been attained in this branch of the study of the romance. 
 Additions which have been made consist largely in informa- 
 tion as to a number of manuscripts which were unknown to 
 Paris, or which have since been found. 
 
 The second and major part of the study has been devoted 
 to the Seven Sages in English. Here I have been preceded 
 by Petras and Buchner, the one dealing mainly with the 
 Middle English group, the other especially with the relations 
 of the Wynkyn de Worde and Holland versions. The 
 dissertations of these two scholars are the only real contri- 
 butions which have been made to the study of the English 
 versions. It is therefore not surprising that many of the 
 current theories with regard to these versions are shown on 
 closer examination to be erroneous. The most far-reaching 
 of these misconceptions is, I believe, that which regards the 
 Wright version as independent of all other English versions. 
 My investigations lead me to the conviction that at least seven 
 of the eight Middle English manuscripts are related to each 
 other through a common Middle English original. 
 
 I regret that I have been forced to forego consideration of 
 one of the Middle English versions, the Asloan. I was 
 denied access to this manuscript by its owner, Lord Talbot 
 de Malahide, and learned of the existence of a transcript of it 
 in the University Library at Edinburgh when it was too late 
 to avail myself of it. Prof. Varnhagen believes it to have 
 had an immediate basis on some Old French manuscript; 
 there are reasonable grounds for doubting this belief, however, 
 and I am unwilling to subscribe to it until a further comparison 
 with the remaining Middle English versions has been made. 
 
 This study leaves undone the most interesting, if not the 
 most valuable part of the work I had planned, a comparative 
 study of the stories themselves ; for not even the stories of 
 the Bidpai collection have enjoyed a wider vogue than those 
 of the Seven Sages. The task of tracing these in their travels 
 
THE SEVEN SAGES. 3 
 
 and of collecting their analogues will be attempted in a future 
 publication, when it is hoped that an edition of one or more 
 of the unpublished Middle English manuscripts may also be 
 attempted. 
 
 I. THE EARLIER HISTORY OF THE ROMANCE. 
 I (a). The Romance in the Orient. 
 
 It is universally held to-day that the great collection of 
 popular stories known in the West as the Seven Sages of Rome, 
 in the East as the Book of Sindibdd, is of Indian origin. 
 This was well established by Deslongchamps already in 1 838, 
 in his Essai sur les Fables Indiennes, 1 and has never since been 
 seriously brought in question. The Indian original, however, 
 has not yet been discovered, nor is it probable that it ever will 
 be; and it even admits of very considerable doubt whether 
 the romance ever existed in India in a form very near to that 
 in which it is first found. 
 
 All attempts, too, to show a kinship between the romance 
 and some surviving Sanskrit story have proved in large part 
 futile. Benfey first pointed out the analogy between the 
 introduction to the Pantchatantra and the framework of 
 the Sindibdd, 2 but he very justly concluded that the Pantcha- 
 tantra was indebted to the Sindibdd rather than the Sindibdd 
 to the Pantchatantra. In a later publication, 3 he called atten- 
 tion to the similarity between the Sindibdd and the legend 
 of Kunala and Asoka, and Cassel has boldly assumed this 
 legend to be the ultimate basis of the romance. 4 
 
 The story of Kunala is widely known in Sanskrit litera- 
 ture. Asoka, a famous Indian king, had, after the death of 
 his first wife, married one of the latter's attendants. The 
 
 1 Published at Paris, 1838, in conjunction with Leroux de Lincy's edi- 
 tion of the Sept Sages de Rome. 
 
 2 Pantchatantra, Leipzig, 1859, I, \ 8 ; also Melanges Asiat., Ill, p. 188 f. 
 
 3 Orient and Occident, in, p. 177 f. 
 'Mischle Sindbad, Berlin, 1888, pp. 10f., 62. 
 
4 KILLIS CAMPBELL. 
 
 new queen had been rejected previous to this by Kunala, 
 the son of Asoka by another wife, and bore in consequence the 
 greatest hatred toward him. The prince is sent by Asoka to 
 one of the provinces to put down a rebellion, where he wins 
 great distinction for himself. In the meantime the king is 
 stricken with a fatal disease, and determines to recall the 
 young prince and place him on the throne. The queen, 
 realizing what this would mean to her, offers to cure the king 
 provided he grant her one favor. Having been restored to 
 health through her agency, the king agrees to grant her what- 
 ever she may desire. She asks to be permitted to exercise 
 supreme authority for seven days, during which time, at her 
 instigation, the prince's beautiful eyes l are put out. Kunala 
 subsequently presents himself before his father in the guise 
 of a lute-player, and is recognized. The queen is burned in 
 expiation of her crime. 2 
 
 Such in brief outline is the legend, which, if it is indeed 
 the ultimate origin of the Sindibdd, at least does not suggest 
 an obvious relation to it. 
 
 Abundant proof of a Sanskrit origin of the Sindibdd, how- 
 ever, is had in the nature or content of its stories and, in 
 particular, of its framework, which is distinctly Buddhistic. 
 Cassel has treated this aspect of the problem at great length. 3 
 He would concede as the result of his investigations that some 
 of the many varying stories were not found in the hypotheti- 
 cal original, and that no one of the extant versions faithfully 
 represents this original. Nor is it strange that this should be 
 the case, for it would be a very miracle had the collection 
 remained intact throughout a possible half-dozen redactions. 
 It is, accordingly, impossible to determine which of the stories 
 were in the original, or which not; this, for the present at 
 least, must remain largely a matter of conjecture. Still, this 
 
 l Cf.Mi*chleSindbad,p. 10. 
 
 2 For further details of this legend, see Burnouf, Introduction d Fhistoire 
 du Buddhisme indien, Paris, 1844, pp. 144 f., 406. 
 ' Mischle Sindbad, p. 82 f. 
 
THE SEVEN SAGES. 5 
 
 much may be accepted as established, that some of the original 
 stories, the ethical purpose, and many of the general charac- 
 teristics of the Indian prototype have been preserved. 
 
 The Eastern group comprises a Hebrew, a Syriac, a Greek, 
 an Old Spanish, two closely related and a third somewhat 
 anomalous Persian, and three cognate Arabic versions. All 
 these differ more or less from each other, but, as compared 
 with the Western group, with which they have in common 
 only four stories and the framework, they distinctly stand 
 apart and make up a separate group. There are many 
 important details in which the two groups differ, but the 
 most marked features which characterize the Eastern group 
 are, first, that each sage tells two tales as against one each in 
 the western versions 1 a feature which was probably not 
 in the Sanskrit original; and, secondly, in contradistinction to 
 the entire western group with the exception of the Dolopathos, 
 that the prince has only one instructor, the philosopher Siudi- 
 bad. This illustrious teacher is the central figure of all 
 versions in the East, where by general consent the romance 
 is called after him the Book of Sindibdd. 2 
 
 The origin of the name Sindibdd is in dispute. Benfey 
 traces it back to *Siddhapatl, 3 Teza to *Siddhapala ; 4 Cassel, 
 on the contrary, holds that the word was coined first after 
 leaving India, and is neither Siddhapatl nor Siddhapala, but 
 *Sindubadhjdja = Indian teacher. 5 
 
 The name of the prince has not been preserved, but the 
 king is named in each one of the representative eastern texts. 
 In the Syriac and the Greek he is called Kurus; in the Old 
 
 1 This is the case in all eastern versions save the Seven Vezirs and the 
 version of Nachshebi : in the former some sages tell one, some two stories ; 
 in the latter each sage tells only one. 
 
 2 Prof. Rhys Davids in his work on the Jdtakas ( Buddhist Birth Stories, 
 Boston, 1880, vol. i, pp. XLI, xciv) seems to have confounded this romance 
 with the story of Sinbad the Sailor of the Arabian Nights. The two are in 
 no way related. 
 
 3 Pantchatantra, i, 5 (p. 23). 
 
 4 11 Libra dei Selte Savj, ed. D'Ancona, Pisa, 1864, p. XLVH. 
 
 6 Mischle Sindbad, p. 66. 
 
6 KILLIS CAMPBELL. 
 
 Spanish, Alcos, which may be considered a variant of Kurus 
 (Al-Curus), since the Spanish holds very closely with the 
 Greek and Syriac, and goes back to the same original. The 
 Hebrew version, on the other hand, calls the king Pai Pur, 
 or, as Ben fey has suggested, Kai (king) Pur, and Cassel 
 would identify this Pur with the Indian king Porus, ruler 
 of India at the time of the Alexandrian invasion, and third 
 before King Asoka of the Kunala story. Porus, Cassel 
 maintains, is a substitution for .the less famous Asoka of the 
 original a transference of the Asoka tradition to Porus. 1 
 The Kurus of the Greek and Syriac he would explain in like 
 manner as a similar transference, after leaving India, from 
 Porus, or Asoka, to the far-famed Cyrus of the Persians. 2 
 
 The route of transmission from India westward is very 
 generally assumed to have been through Pahlavi into Arabic. 3 
 There seems to be little evidence, however, of the existence of 
 a Pahlavi version, unless the current tradition to that effect, 
 or the fact that the Kalila wa Dimna had such an inter- 
 mediate stage, be regarded as such. Hence Cassel takes a 
 radically different view from that generally held, maintaining 
 that the lost Arabic text goes back not to a Pahlavi but to 
 a Syriac version, which, in its turn, goes back to the San- 
 skrit, the collection, then, having been transmitted westward 
 through the agency of the Manicheans in the third or fourth 
 century of our era. 4 The Hebrew and the lost Arabic versions 
 he conceives to be coordinate redactions of this early Syriac 
 version, finding support of this theory, so far as it concerns 
 the Hebrew text, in the Syriac influence which the language 
 of the latter exhibits. At the same time, although he thus 
 claims for the Hebrew version the greatest antiquity of any 
 text which lias been preserved, Cassel admits that, in addition 
 to the Syriac influence, the Hebrew text also contains traces 
 of a Greek influence (as, for instance, in the names of the 
 
 1 Ibid., pp. 63, 2 1 'J. Ibid., p. 6 1. 
 
 * So Comparetti, Noldeke, Clouston, and others. 
 4 Mutchle Sindbad, pp. 61, 310. 
 
THE SEVEN SAGES. 7 
 
 sages), 1 which is of itself sufficiently indicative of the lack of 
 conclusive proof of his thesis. 2 
 
 The Arabic text, unlike the early Syriac, is in no way 
 hypothetical, but the evidence that it once existed, even as 
 late as the thirteenth century, 3 is conclusive. Its influence 
 has been very wide, and, until Cassel, it has been generally 
 assumed to be the source, either mediate or immediate, of the 
 entire Eastern group. The Syriac Sindban and the Old 
 Spanish version are believed to be its closest representatives. 
 Its author, according to the testimony of the introduction to 
 the Syntipas, was a certain Musa, and its date has been con- 
 jectu rally placed by Noldeke 4 and others in the eighth century. 
 
 Only ten versions belonging to the Eastern type have sur- 
 vived. These are the Hebrew Mischle Sindbad, the Syriac 
 Sindban, the Greek Syntipas, the Persian Sindibdd-ndmeh and 
 its source, the text of As-Samarquandi, the Old Spanish Libro 
 de los Engannos, the three Arabic versions of the Seven Vezirs, 
 and the eighth night of the Tuti-ndmeh of Nachshebi. 5 
 
 The relative age of these is not definitely known. Early 
 scholars as a rule held that the Hebrew version antedated all 
 others ; but this view was summarily rejected by Comparetti 6 
 and his followers, who claimed greatest antiquity for the 
 Syntipas, a distinction of which it was robbed by Rodiger's 
 discovery of the Syriac version. The Nachshebi version has 
 also been held to be the oldest, 7 and Clouston in recent years 
 
 1 These are, according to Cassel (p. 219 f.), Sindibad, Hippocrates, Apu- 
 leius, Lucian, Aristotle, Pindar, and Homer. 
 
 "Mischle Sindbad, pp. 222, 310. 
 
 3 The Old Spanish version was made from it in 1253. 
 
 * In his review of Baethgen's edition of the Sindban in Zeitschrift d. d. 
 Mory. Gesellschaft, xxxui, p. 518. 
 
 5 A11 these, with the exception of the text of As-Samarquandi, have been 
 rendered accessible either in the original or in translations, and in most 
 cases in both. 
 
 6 Comparetti, Book of Sindibad, p. 53 f. Citation is made from the English 
 translation by Coote, for the Folk Lore Socy., London, 1882. The original 
 Ricerche appeared at Milan in 1869. 
 
 7 Brockhaus for example. 
 
8 KILLIS CAMPBELL. 
 
 has contended for the Sindibad-ndmeh as representing most 
 closely the hypothetical original. 1 The result of the latest 
 investigation, as has been seen, is to return to the view of 
 early scholars, which gives to the Hebrew text first place both 
 as regards date and fidelity to the lost original. Such is 
 CassePs conclusion, which, although somewhat revolutionary, 
 is arrived at by argument which at least serves to invalidate 
 Comparetti's assumption that the Hebrew text stands for a 
 late and very free version of the romance. It is hardly legiti- 
 mate to conclude, from the circumstance that the Misclde 
 Sindbad stands apart from the remaining members of the 
 Eastern group, that it is, on that account, less faithful to 
 the original tradition. Nor is Comparetti's argument for the 
 identification of the Joel to whom the work is attributed by 
 Rossi and the British Museum manuscript, with the Joel 
 who is reported to have translated the Kalila wa Dirnna into 
 Hebrew, and the consequent establishment of a thirteenth 
 century date for this version, any more valid. 2 At the same 
 time, it is to be regretted that Cassel has attained no definite 
 results as to chronology. 3 
 
 The Mischle Sindbad* contains twenty stories, three of 
 which, Absalom, The Disguised Youth, and The Humpbacks 
 (amatores), appear in no other version of the Eastern group. 
 Its first three stories come in the same order as in the Syriac, 
 Greek, and Old Spanish versions. Other agreements which 
 are evident on reference to a comparative table serve appar- 
 ently to hold these four texts together; 5 this, however, is 
 probably rather due to a more faithful preservation of the 
 
 1 ( louston, Book of Sindibdd [Glasgow], 1884, p. L f. 
 
 * Comparetti, Book of XiiM<l<l t p. 53 f. * Mischle Sindbad, p. 310. 
 
 4 The Hebrew text has undergone the following editions: Sengelman 
 (with (ierman translation), Halle, 1842; Carnioly (with French transla- 
 tion), Paris, 1849; and Cassel (German translation and copious notes), 
 Berlin, 1888. 
 
 6 For the most complete comparative table, see Landau, Quellen des Deka- 
 meron, 2d ed., Stuttgart, 1884; see also Cassel, p. 362 f., and Comparetti, 
 p. 25. 
 
THE SEVEN SAGES. 
 
 ultimate original on the part of these than to any very close 
 relationship with the Hebrew, and comparison will show not 
 only that these three have much in common which does 
 not appear in the Hebrew, but also that the latter has many 
 features (the naming of the sages, for example) which are 
 peculiarly its own. Additional importance attaches to the 
 Hebrew text from the fact that it probably bears a closer 
 relation to the Western group than any other known eastern 
 version. 1 
 
 The Syriac Sindban was discovered by Rodiger in 1866, 
 and was published with a German translation by Baethgen in 
 1879. 2 The text is unfortunately fragmentary, especially at 
 the end. Although at first doubted by Comparetti, it has 
 been satisfactorily shown by Noldeke to be the Syriac basis 
 of the Syntipas, alluded to in the prologue of the latter. 3 The 
 immediate original of the Sindban must then be the last 
 Arabic text of Musa. Noldeke believes it to belong to the 
 tenth century. 
 
 The Greek Syntipas is, in interest and importance, second 
 only to the Hebrew text. As compared with its Syriac origi- 
 nal, it is much more full and ornate, an almost unfailing 
 characteristic of a later text. Its author was, as the prologue 
 establishes, a certain Michael Andreopulos and the translation 
 was made at the command of one Gabriel //.eXowiy-to?. Com- 
 paretti would identify this Gabriel with Duke Gabriel of 
 Melitene, and thus establish the date of the work as the 
 second half of the eleventh century ; 4 but this, while a gain 
 in a measure, is little more than a happy suggestion. Far 
 less probability has Cassel's proposition that the reference is 
 to the angel Gabriel. 6 The text was first published by 
 
 1 See the next chapter on " The Transmission of the .Romance to the 
 Occident." 
 
 8 Baethgen, Sindban, oder die Sieben Weisen Meister, Leipzig, 1879. An 
 English translation by H. Gollancz appeared in Folk Lore, vin, p. 99 f., 
 June, 1897. 
 
 3 Zeitschr. d. d. Morg. Gesellschaft, xxxin, p. 513 f. 
 
 4 Book of Sindibdd, p. 57. Mischle Sindbad, p. 368. 
 
10 KILLIS CAMPBELL. 
 
 Boissonade, and has been lately critically edited by Eberhard. 1 
 A modern Greek adaptation of the older text is of little value 
 in a comparative study of the romance. 2 
 
 The Libro de los Ifagannos, like the Syriac text, was not 
 known until late in the century. It is, according to its pro- 
 logue, a translation from the Arabic, made in the year 1253. 
 The text is complete, but very corrupt. Its closest affinities 
 are with the Greek and Syriac versions, with both of which it 
 exhibits intimate agreement in content and order of stories. It 
 seems to have had no influence at all on modern Spanish litera- 
 ture. The first edition of the text appeared in Comparetti's 
 Ricerchejin 1869; a second edition, with an admirable Eng- 
 lish translation appended, appeared in the English edition of 
 this book in 1882. 3 
 
 The Persian Sindibad-ndmeh 4 ' dates from the year 1375. 
 It purports to be based on a Persian prose text which goes 
 back to the Arabic. Clouston first suggested that this origi- 
 nal was the text of As-Samarquandl, which was known in the 
 early part of the century, but which had subsequently been 
 lost sight of. By the rediscovery of a manuscript of this 
 version in 1891, he has been enabled to establish this conjec- 
 ture as a fact. 5 The As-Samarquandi text agrees closely with 
 the Sindibdd-ndme/i in content, the only important difference 
 being the substitution on the part of the latter of one or two 
 extraneous stories for those it found in its original. The 
 agreement in order of stories is close throughout. The date 
 of the prose text falls late in the twelfth century. It differs 
 considerably from the rest of the Eastern group, but is nearer 
 
 1 Eberhard, Fabulae Romanenses Greece, etc., I (Teubner), Leipzig, 1872. 
 
 2 For the Syntipas in later literature, see Murko, " Die Geschichte v. d. 
 Sieben Weisen b. d. Slaven," Wiener Akad. Sitzunysb., Ph. Hist. Cl., cxxil, 
 No. x, p. 4 f. 
 
 8 Book ofSindibad, pp. 73-164. 
 
 4 This text has not yet been edited. An abstract of it was given by 
 Falconer in the Asiatic Journal, xxxv, p. 169 f. and xxxvi, pp. 4 f., 99 f. ; 
 a complete translation into English appears in Clouston's Book ofSindibad. 
 
 5 Athenaeum for Si-pt. Ill, 1891, p. 355. 
 
THE SEVEN SAGES. 11 
 
 to the Syriac, Greek and Spanish versions than to the Hebrew. 
 There appears to be no evidence to support Clouston's sugges- 
 tion that it represents the Sanskrit prototype more faithfully 
 than any other known version; neither is Modi's contention 
 for a close relation with the story of Kaus, Sonddbeh, and 
 Sidvash 1 by any means convincing; but the tradition which 
 makes its origin in the Arabic text is doubtless well founded. 
 
 Under the head of the Seven Vezirs fall three versions which 
 have been introduced into the frame of the Arabian Nights. 
 These are the texts of Habicht and Scott, and the Boulaq 
 edition. 2 They are of late composition, and of comparatively 
 slight value for the present purpose. 
 
 The text contained in the eighth night of Nachshebi 3 is 
 one of the most interesting of the Eastern group, and has 
 given rise to much speculation. It differs considerably from 
 all other related versions, having but six stories, only five 
 of which appear elsewhere in the Eastern group. All five of 
 these in the fuller versions are second vezir's tales, and as 
 they were also found originally in the Sukasaptatl (though 
 not connected as with Nachshebi), it has been conjectured by 
 Comparetti that they were first introduced into the Sindibdd 
 after leaving India, and that Nachshebi, observing this, again 
 inserted them in his free translation of the Tutl-ndmeh, and 
 practically in the same form in which he found them in the 
 Sindibdd.* Comparetti would further identify the collection 
 before and after this addition with the ' Greater 7 and ' Lesser' 
 Sindibdd referred to by the tenth century Mohammed Ibn el 
 "VVarrak. A radically different theory has been advanced by 
 Noldeke, who maintains that the ' Greater ' Sindibdd has been 
 lost. 5 As for the version of the Sindibdd whence Nachshebi 
 
 1 Modi, Dante and Viraf and Gardis and Kaus, Bombay, 1892. 
 
 2 1001 Nights, Breslau, 1840, xv, pp. 102-172; Scott, Tales, Anecdotes and 
 Letters, Shrewsbury, 1800, p. 38 f. ; 1001 Nights, Boulaq, 1863, in, pp. 75-124. 
 
 'Brockhaus, Nachshebl's S. W. M., Leipzig, 1845; translated by Teza, 
 D'Ancona ed. of Sette Savj., p. xxxvn f. 
 
 'Book ofSindibad, p. 37 f. 
 
 8 Zeitschr. d. d. Morg. Gesellschaft, xxxin, p. 521 f. 
 
12 KILLIS CAMPBELL. 
 
 drew, both Comparetti and Noldeke concur in the belief that 
 it was the text on which the Sindibdd-ndmeh was based, or 
 that of As-Samarquandi. The date of the Nachshebi version 
 is late, as its author died in 1329. 
 
 Besides the ten versions catalogued above, the existence of 
 certain others which have been lost is proved by sundry refer- 
 ences from oriental writers. A Persian text is attributed to 
 Azraki by Daulat Shah, and there are several references from 
 the ninth and tenth centuries to works which do not seem to 
 be identical with anything which has been preserved. The 
 best-known of these, probably, is Masudi's (943) statement 
 that in the reign of Kurush " lived es-Sondbad, who is the 
 author of the book of the seven vezirs, the teacher and boy, 
 and the wife of the king. This is the book which bears the 
 name Kitab-es-Sindbad." 1 A still earlier reference is that of 
 Al-Yaqubi (880). Both of these may refer to the Arabic 
 text of Musa, though this is by no means certain. Most 
 perplexing of all is the reference, already mentioned, to a 
 ' Greater ' and a ' Lesser ' Book of Sindibdd. 
 
 Doubtless many more versions have been lost than this 
 would indicate ; but since nearly a third of the known texts 
 have been revealed only within the last generation, it may be 
 hoped that the near future has in store many revelations 
 which will materially serve to dispel the mist which now 
 surrounds almost the entire question of relations in the East. 
 
 I (b). Transmission of the Romance to the Occident. 
 
 The Greek Syntipas and the Old Spanish Libro de los 
 F.iKjtinno* are the only representatives of the Eastern group 
 which have arisen on European territory. Neither one of 
 then-, however, can he considered a connecting link in the 
 chain of transmission; nor can, in fact, with nil certainty, 
 any one member of the Eastern group claim this distinction. 
 
 1 Masfull, Meadows of Gold, translated by Sprenger, London, 1841, p. 175. 
 Alasudi was not wi-11 acquainted with the romance, as follows from the fact 
 that he attributes its authorship to Sindibiid. 
 
THE SEVEN SAGES. 13 
 
 The question of transmission is, and must doubtless always 
 remain, very much shrouded in darkness. The two groups, 
 having in common only four stories and the framework, and 
 having in these, also, many radical differences, cannot be 
 thought of as connected through free or literal translation, 
 nor by intermediate redactions; the only valid explanation 
 of the enormous gap existing between them must repose in 
 the assumption of a basis for the western original in popular 
 tradition. This alone can explain the difference between the 
 two groups. 
 
 But this assumption should not carry with it (as with 
 Comparetti apparently; 1. c., p. 2) the further assumption 
 that, since the medium of transmission was oral, all possi- 
 bility of ever determining the specific original of the Western 
 group is thereby done away with. This need not follow at 
 all. The oral tradition on which the western parent version 
 had its basis, must itself have had some basis, and this cannot 
 have been the entire Eastern group, nor with any degree of 
 probability any two of its members ; it was some one member 
 of the Eastern group. Accordingly it is legitimate to endeavor 
 to determine which one of the Eastern versions is the origi- 
 nal, or the closest representative of the original, of the Western 
 group. 
 
 Modern scholars in general have refrained from any investi- 
 gation of this stage of the history of the romance. With a 
 single exception, the only judgments upon the problem date 
 from the earlier part of the century. Dacier, Keller, Deslong- 
 champs, Wright, D'Ancona, and others put forth claims for one 
 or another of the Eastern group (some for the Greek, others 
 for the Hebrew), as the original of the western type. But 
 all these claims were unsustained by any evidence adduced, 
 and were in every case scarcely more than conjectures. The 
 modern scholar who alone has put himself on record here is 
 Landau ; l and he is, at the same time the only one of the 
 
 1 Marcus Landau, Quellen des Dekameron, 2d ed., Stuttgart, 1884. 
 
14 KILLIS CAMPBELL. 
 
 whole number who has made a serious effort to sustain his 
 position. At the basis of Landau's work, however, lies the 
 assumption that the Latin prose Historia Septem Sapientum 
 (H) is the parent version of the Western group, an assump- 
 tion which is entirely gratuitous, for surely Gaston Paris has 
 succeeded in demonstrating that H is not the original western 
 text; while the majority of Landau's arguments therefore hold 
 also in a comparison of the oldest texts with the Eastern 
 group, it is in view of this fundamental misconception on his 
 part that he has in reality proved nothing more than that the 
 fourteenth century Historia is nearer the Hebrew than to any 
 other eastern version. 
 
 With the proof of the unoriginality of H, the question 
 as to the nearness of the various sub-types of the western 
 group to the parent version has been left open. The oldest 
 text preserved is the Dolopathos ; but this is a unique version, 
 and, as will be shown in the next chapter, cannot with the 
 slightest probability be looked upon as the western original, 
 though it is assuredly connected in some way with the pre- 
 vailing type of the Western group, the Seven Sages of Rome. 
 Next to the Dolopathos the Scala Coeli (S) and Keller (K) 
 texts have been treated as the oldest by the latest and best 
 authorities ; to these, in view of its prime importance and the 
 uncertainty as to its relations, we should like to add the type 
 A*. 1 No proof of the priority of any one of these has yet been 
 brought forward ; moreover, the earliest dating proposed for 
 any of them is the first half of the thirteenth century. We 
 may begin, then, with the assumption that the immediate 
 parent version of the Western group has been lost. At the 
 same time, since the Dolopathos, which dates from the last 
 quarter of the twelfth century, is evidently based on some 
 version of the prevailing western type, we may assume for 
 
 Old French versions A, C, D of Paris (Deux Redactions) have been 
 "starred" throughout in order to avoid confusion with the Middle English 
 (M. E.) versions A, C, D. 
 
THE SEVEN SAGES. 15 
 
 this lost original a date not later than the middle of the 
 twelfth century. 
 
 A twelfth century original having been assumed for the 
 Western irrotip, the Libra de los Engannos (xin cent.), the 
 Sindibdd-nfijmeh (xiv cent.), and the Seven Vezirs (very late) 
 may be eliminated from the investigation ; likewise the unique 
 text of Nachshebi for reasons that are obvious. There remain 
 the Mischle Sindbad, the Sindban, and the Syntipas, no one 
 of which can be dated later than the eleventh century, if 
 we accept CassePs view as to the comparative antiquity of 
 the Hebrew text. Further, since the western original of the 
 Western group has been lost, comparison can be made with 
 the latter only on the basis of the constant elements appearing 
 in its most ancient versions, S, K, A.* Accordingly, the 
 comparison must be instituted between the Hebrew, Syriac, 
 and Greek versions, on the one hand, and S, K, A* on the 
 other. 
 
 The framework of the romance has undergone a radical 
 change in the course of its transmission westward. There is 
 no longer mention of a philosopher Sindibad, but the seven 
 sages of Rome become the central figures, and play the double 
 r6le of instructors and defenders of the prince. Sundry other 
 characteristic features of the Eastern group, such as the prince's 
 early stupidity, the multiplicity of the king's wives, etc., have 
 been lost; but the most far-reaching change consists in the 
 curtailment of stories, each sage telling only one story in 
 the Western group as against the prevailing number of two 
 in the Eastern. 
 
 In these variations the Greek, Hebrew, and Syriac versions 
 present essential agreement ; but there are several features in 
 which these three texts do not agree, and it is significant here 
 that where the Western group preserves any of these features, 
 it is always in agreement with the Hebrew, and in no single 
 instance with the Greek or the Syriac. 
 
16 KILLIS CAMPBELL. 
 
 The following features peculiar to the Hebrew text as 
 compared with the rest of the Eastern group reappear in the 
 oldest western versions : l 
 
 (1). The seven sages are not referred to simply as such, 
 but are mentioned by name 2 (Landau, p. 48). 
 
 (2). They vie in their efforts to secure the office of instructor 
 of the prince 3 (Landau, p. 48). 
 
 (3). These sages, and not the vezirs or counsellors of the 
 king as with the rest of the Eastern group, relate the stories 
 which preserve the prince's life 4 (Landau, p. 48). 
 
 The mode of punishment of the guilty queen offers nothing 
 determining. The eastern texts have little in common here 
 
 } A\\ these several bits of argument adduced here and on the following 
 pages, with the exception of those under the story avis, have been advanced 
 by Landau (pp. 47- 50) ; in addition to these, owing to his false hypothesis 
 of the originality of H, Landau has made use of two other features in which 
 H agrees with the Hebrew text versus the remainder of the Eastern group, 
 but which must be cancelled, since they are also peculiar to H. These are 
 ( 1 ) the disguised-youth incident of H, which Landau (p. 48 f.) inclines to trace 
 back to the seventeenth story of the Mischle Sindbad, and (2) amatores, the 
 twelfth story of the Historia, which is ultimately the same as the Hebrew 
 story of the Hunchbacks (M. S. 18; see B&lier, Les Fabliaux, Paris, 1893, p. 
 201 f.). Neither of these appears in any other western version, whence 
 the only legitimate inference that they were not in the lost western original, 
 but are late incorporations on the part of jffinto the frame of the collection. 
 
 <2 This, a characteristic feature of the Western group, appears in all 
 western texts save those (as S) which have been abridged. The names 
 of the sages in the Mischle Sindbad are Sindibad, Hippocrates, Apuleius, 
 Lucian, Aristotle, Pindar, and Homer (Cassel, p. 253) ; in the Western 
 group, Bancillas, Ancilles, Malquidras, Lentulus, Caton, Jesse, and Meros. 
 For variants of these, see Landau, Quellen den Dekamerqn, p. 60 n. 
 
 s ln the Hebrew (see Cassel, p. 255 f.) one proposes to instruct him in 
 five years, another in two years, a third in one year, and finally Sindibfid 
 offers to make him wisest of all men in six months. The term of years 
 proposed by the sages in the western versions varies from seven to one. 
 
 *Carmoly (p. Go) states expressly that these were the king's counsellors, 
 and not the sages, who, he says, were now in hiding to avoid the king's 
 anger; but, as Landau (p. 48) points out, the sage Aristotle is referred to 
 by name at the end of the third story as having saved the prince's life by 
 his stories on the preceding day (Cassel, p. 267); accordingly, although 
 there is a slight confusion, it is evident that Carmoly is in error. 
 
THE SEVEN SAGES. 17 
 
 beyond the bare outline. In the Greek and As-Samarquandi 
 texts, the woman is condemned to wander through the streets 
 on an ass, with her head shaved and her face soiled, and with 
 two criers proclaiming her shame. In the Hebrew text, she 
 is, at the prince's request, pardoned unconditionally. The 
 Syriac text is fragmentary here. Of the western feature of 
 condemning the queen to die the death prepared for the 
 prince, there seems to be no hint in the eastern versions. 
 
 A comparison of the four stories (cams, aper, avis, and 
 senescalcus) common to the two main groups also shows many 
 variations, but here, too, where the Mischle Sindbad differs 
 from the Syntipas and other versions of the Eastern group, it 
 will be seen to accord in several particulars with the Western 
 group. 
 
 (1). Canis. The story cams, the only one found in all 
 versions of the Seven Sages, both eastern and western, exhibits 
 in the earliest western versions no noteworthy variations from 
 the prevailing type of the story in the East. In the Sindibdd- 
 ndmeh it is a weasel or ichneumon which attacks the sleeping 
 child ; in all other versions it is a snake. The child is left in 
 charge of nurses in the western versions, a feature entirely 
 foreign to the Eastern group. The derivative types, Dolo- 
 pathos and Historia, introduce a bird (Dolop., a goshawk; 
 Hj a falcon) which wakes the child on the snake's approach. 
 This and several other additions, especially to the Dolopathos, 
 are not found in the types S, K, and A*, a circumstance which 
 well warrants the inference that they were not in the western 
 parent version. 
 
 (2). Aper. This story, like cam's, has been subjected to 
 considerable alteration in the course of transmission, e. g., in 
 the East, the boar comes to his death as the result of holding 
 up his head in the expectation of more fruit (the sinews drying 
 up) ; in the West, he is slain by the shepherd, who, descending 
 the tree until in reach of him, " claws " him on the back until 
 he falls asleep, and then dispatches him with his knife. But 
 2 
 
18 KILLIS CAMPBELL. 
 
 the special value in the collation of this story lies in the fact 
 that the Hebrew text coincides with the Western group in 
 having a man chased up the tree, while in the remaining eastern 
 versions it is a monkey who thus flees from the boar. This 
 coincidence, first noted by Deslongchamps (/. c., p. 110 n.), 
 is one of the most striking agreements of the Hebrew text with 
 the Western group. 
 
 (3). Senescalcus. A comparison of the various versions of 
 senescalcus reveals no eastern motive reproduced in the West 
 which is not common to the entire Eastern group. The 
 western version of the story agrees in general outline with 
 the eastern, but is distinguished from it by the introduction 
 of even more objectionable details than those which characterize 
 its oriental original. The western texts vary in the method 
 of punishing the seneschal : in S he is hanged ; in K, A*, 
 and the prevailing sub-groups, he is banished by the king on 
 pain of death in case he return. In the East the bathman 
 (= seneschal) dies by his own hand. 
 
 (4). Avis. The essential features of this famous story have 
 been preserved remarkably intact thoughout all versions. 
 There are, however, two features which occur in the East 
 only in the Mischle Sindbad which have been preserved in 
 the western texts. These are (1) that the wife goes on the 
 house-top in order to sprinkle water over the bird's cage, and 
 (2) that she is aided and abetted in her efforts to deceive the 
 bird by her maid. Of the first of these we have in no other 
 eastern version any hint ; likewise, for the second, there is no 
 real suggestion in any of the Eastern group besides the Mischle 
 Sindbadj for, although there is mention elsewhere of the maid, 
 it is only as having been suspected of informing on her mis- 
 tress, and never in the r6le assigned her in the Hebrew and 
 the western versions. 1 
 
 1 The arguments made by Landau under avis are not valid. That the 
 bird speaks Hebrew as well as Latin, is not true of any of the oldest 
 western versions, but appears to be peculiar to H; while the argument 
 from the killing of the bird in H and the Hebrew text is altogether in- 
 
THE SEVEN SAGES. 19 
 
 To recapitulate then, the features peculiar to the Hebrew 
 and the oldest western texts are as follows : 
 
 (1). The seven sages are mentioned by name. 
 
 (2). There is a rivalry between the sages in their efforts to 
 secure the tutelage of the prince. 
 
 (3). The sages, not the king's counsellors, defend the prince. 
 
 (4). In oper, the adventure happens not to an ape, but to a 
 man. 
 
 (5). In avis, (a) the deception is practised on the bird 
 through an opening in the house-top, and (b) the maid appears 
 as an assistant of the faithless wife. 
 
 A comparison with the Syntipas fails to bring out any 
 feature exclusively common to it and the Western group. 
 The same holds for the Syriac and later versions. The 
 question is then narrowed down to the significance of the 
 agreements between the Hebrew and the western texts. Are 
 they only accidental, or have they a real significance? Cer- 
 tainly they do not prove a direct relationship between the 
 Hebrew and any western version, as Deslongchamps and 
 Landau have maintained ; nor are they sufficient to justify 
 the thought of a connection of the Eastern and Western groups 
 through intermediate literary stages; indeed, they yield no 
 conclusive proof of anything with regard to the problem of 
 relationship. Nevertheless, they are in a measure significant ; 
 though some of them are in all probability accidental, yet it 
 does not seem possible that all of them can be mere coinci- 
 dences. They justify, at least, the negative conclusion that 
 neither the Syntipas (nor the Sindban) was the eastern original 
 whence sprang the tradition which culminated in the parent 
 version of the Western group. And while they do not prove 
 the Hebrew text to represent this eastern original, they 
 do, nevertheless, establish this as a probability, with the 
 only other alternative in the supposition that the eastern 
 original of the Western group has been lost. 
 
 valid, since the same feature is found in all eastern versions save the 
 Syntipas, and would be in any case of little value for the purpose to which 
 Landau would put it, since it is a simple and natural variation. 
 
20 KILLIS CAMPBELL. 
 
 I (c). The Romance in France and Italy. 
 
 Between the eastern and western types of the Seven Sages, 
 as has been seen, there is a very wide difference. Four of the 
 original stories and the main outline of the eastern framework 
 have been preserved in the western versions, but, as Comparetti 
 has aptly said, " there is no eastern version which differs so 
 much from the others as the whole Western group differs from 
 the Eastern, whether it be in the form of the fundamental 
 story or in the tales which are inserted in it. 7 ' In explanation 
 of this wide difference a basis has been assumed for the Western 
 group in oral accounts. 
 
 Where these oral accounts first took literary form has not 
 been, and probably never will be, satisfactorily determined. 
 Some have maintained an origin on Latin territory ; but the 
 probabilities favor a French origin, though it is more than 
 possible that the parent version was written in the Latin 
 language. 
 
 The oldest form, apparently, under which the western type 
 has come down to us is the Dolopathos. There can be little 
 doubt, however, that the more widely known Sept Sages de 
 Rome, of which there survive many manuscripts dating from 
 a period but a little later than that of the earliest version of 
 the Dolopathos, preserves more nearly the form and contents 
 of the western parent version. And it is under this form that 
 the romance has acquired its marvellous popularity in France, 
 whence it has penetrated into nearly every other country of 
 Europe. 
 
 With regard to the relationship of these two forms or groups 
 under which the romance appears in the West, early scholars 
 were very much in error. For a long time it was believed 
 that the poetical version of the Dolopathos found its source in 
 the Latin prose Historia Septem Sapientum-, 1 again, it was 
 always assumed as fundamental that the Hisloria antedated 
 
 1 The most widely known of all versions of our romance; see below. 
 
THE SEVEN SAGES. 21 
 
 and was the ultimate western original of the entire Western 
 group, these two misconceptions pervaded the entire litera- 
 ture on the romance during the first half of this century. The 
 error of the first was first shown by Montaiglon in 1856, 1 and 
 its utter absurdity was conclusively proved a few years later 
 by Oesterley's discovery of the Dolopathos of Johannes, from 
 which Herbert had made his poem. 2 The second was current 
 even until the appearance of Gaston Paris's Deux Redactions* 
 in 1876, in which the comparatively recent date of the His- 
 toria, and its immediate dependence on .A*, has been placed 
 beyond question. 
 
 1 . The DolopathoK. The Dolopathos exists in two versions, 
 the Latin prose of Johannes de Alta Silva and the Old French 
 poem of Herbert. The latter is preserved, so far as is known, 
 in but three manuscripts; 4 of, the former, there are known, 
 besides the original manuscript discovered by Oesterley, three 
 late copies pointed out by Mussafia, 5 an Innsbruck, 6 and 
 a British Museum MS. 7 
 
 1 In the preface to his edition of the Herbert version : Li Romans de 
 Dolopathos, ed. Brunet and Montaiglon, Paris, 1856. 
 
 3 This manuscript was discovered by Oesterley in 1873, and was published 
 by him in the same year: Johannis de Alta Silva Dolopathos . . . ., Strasburg. 
 See reviews by Paris, Romania, IT, p. 481 f. ; by Studemund, Z.f. d. A., xvn, 
 p. 415 f. and xvin, p. 221 f . ; and by Kohler, Jahrb. f. ran. u. engl. Lit., 
 xiir, p. 328 f. Several manuscripts discovered by Mussafia ( Wiener Akad. 
 Sitzungsb., Ph. Hist. CL, XLVIII, p. 246 f., 1864) prior to this, and at first 
 supposed to be original, were soon shown to be fifteenth century copies 
 of the older manuscript. 
 
 'Published in the Soc. d. Anc. Textesfr. for 1876. For the Historia, see 
 
 pp. XXVJII-XLIII. 
 
 4 See Paris in Romania, n, p. 503. A leaf of a fourteenth century MS. of 
 the Herbert version has been lately acquired by the Bibliothque Nationale 
 Nouv. Acq.fr. 934, No. 6 (Bulletin de la Soc. d. Anc. Textesfr., for 1896, p. 
 71 f.). See also Haupt's Altd. Blatter, I, p. 119 f., for a German version of 
 six stories of the Dolopathos. 
 
 5 See Wiener Akad. Sitzungsb., Ph. Hist. CL, XLVIII, p. 246 f. 
 6 Also brought to light by Oesterley. 
 
 7 Usually overlooked ; see Ward, Catalogue of Romances, London, 1893, II, 
 p. 228 f. 
 
22 KILLIS CAMPBELL. 
 
 Johannes de Alta Silva, the author of the Latin original, 
 was a Cistercian monk of the monastery of Haute Seille. His 
 work bears the title Dolopathos, sive Opusculum de rege et 
 seplem Sapientibus. It was dedicated to Bishop Bertrand of 
 Metz, who had jurisdiction over the monastery of Haute Seille 
 from 1184 (when it was transferred from the see of Toul to 
 the see of Metz) to 1212, during which period, since Johannes 
 would naturally dedicate to his own bishop, we may safely 
 place the composition of his work. Paris favors a dating 
 between 1207 and 1212 (Romania, n, p. 501). 
 
 The Old French poem of Herbert was made from the Latin 
 prose text of Johannes toward the end of the first quarter of 
 the thirteenth century (Montaiglon, 1223-1226 ; Paris, before 
 1223). 
 
 This type of the romance differs from all other western 
 types in having only one instructor for the prince. For this 
 reason it has been conjectured that it was founded on some 
 oriental original, but there is no real evidence in support of 
 this. In the suppression of the queen's stories, a feature in 
 which it agrees with the Nachshebi version, equally as little 
 indication of an immediate eastern original is to be found. 
 
 The Dolopathos has only one story (canis) in common with 
 the Eastern group, and inasmuch as this, together with three 
 other of its stories (gaza, puteus, and inclusa), is also found in 
 the Sept Sages de Rome, it is reasonably certain that the monk 
 Johannes was acquainted with some version of the latter type. 1 
 There is only one alternative supposition, viz. that both types 
 grew up independently of each other and almost contempo- 
 raneously, the one drawing only one story from the traditions 
 brought from the East, while the other drew this and three 
 others in addition, with the further coincidence that both 
 receive, as the result of like influence and environment, three 
 stories (gaza, puteus, and indusa) in common which were not 
 
 1 See Comparetti to the contrary ; Vergil in the Middle Ages, translated by 
 Benecke, London, 1895, p. 234 f. 
 
THE SEVEN SAGES. 23 
 
 in the eastern framework. That such was the case is, to say 
 the least, very improbable. 
 
 But, in any case, the prose Dolopathos was made not from 
 written, but from oral sources. This is expressly stated by 
 its author who says he wrote non ut visa, sed ut audita and 
 is borne out by the introduction of the Lohengrin story, which 
 appears here for the first time, 1 as well as by the variations to 
 which both framework and stories have been subjected. 
 
 The poetical version of Herbert is based directly on the 
 Latin prose version of Johannes. It contains many details 
 and several important episodes which do not appear in the 
 text discovered by Oesterley, chief among which additions are 
 (1) the story inclusa, which has been fused with puteus in 
 the poem, and (2) a very interesting episode with which gaza 
 has been supplemented. Gaston Paris 2 thinks that these were 
 contained in Herbert's original, which he believes to have 
 been an enlarged copy of the first draft of the work as seen 
 in the Oesterley manuscript ; but whether they are to be thus 
 explained, or are to be attributed to the independence of the 
 poet, has not yet been definitely settled. 
 
 The Herbert version is very long, containing nearly 13,000 
 lines. In both length and style it stands in striking contrast 
 to the Keller metrical version of the Sept Sages de Rome (K)? 
 which, although it has nearly twice as many stories, has only 
 5,060 lines. The Dolopathos has an introduction of about 
 4,800 lines where K has but 68. 
 
 The king in this branch of the Western group bears the 
 name Dolopathos, and rules over the island of Sicily. The 
 prince is called Lucinius. Before his birth it is predicted that 
 he will become very wise, but will undergo many hardships, 
 and will ultimately become a worshipper of the true God. 
 
 1 See Todd, La Naissance du Chevalier au Oygne, Introduction, p. in f., in 
 Publications of the Mod. Lang. Assn. of America, vol. iv, 1889. See also Paris's 
 review in Romania, xix, p. 314 f. 
 
 2 Romania, n, p. 500. 
 
 3 See the dissertation of Ehret, Der Verfasser des Roman des Sept Sages und 
 Herberz, Heidelberg, 1886. 
 
24 KILLIS CAMPBELL. 
 
 The prince's instruction begins when he has reached the age 
 of seven. He is sent to Rome, and put under the care of the 
 poet Vergil, whose figure is supreme throughout the romance, 
 and gives to it one of its strongest claims upon our interest. 1 
 The sages, who are, owing to Vergil's prominence, placed 
 somewhat in the background, come up as in the other western 
 versions, one each day and in a most mysterious fashion, 
 always just in time to save the prince's life. The prince 
 relates no story at all, but Vergil tells the eighth and last. 
 The order of stories is as follows : (1) canis (Dog and Snake), 
 (2) gaza (King's Treasury), (3) senes (Best Friend), (4) creditor 
 (the Pound of Flesh episode of the Merchant of Venice), 2 (5) 
 viduae filius (Widow's Son), (6) latronis filius (Master- Thief), 
 (7) cygni eques (the fabled origin of Godfrey de Bouillon), (8) 
 inclusa-puteus (Two Dreams and Husband Shut Out)* 
 
 2. The Sept Sages de Rome. The Sept Sages de Rome, in 
 contradistinction to the Dolopathos, comprises a very large 
 number of more or less closely related versions. Probably 
 one hundred manuscripts of its type are already known, and 
 many others, we may be sure, remain to be revealed by further 
 research. The immediate source whence these have sprung 
 has not come down to us. The date, too, of the parent ver- 
 sion is uncertain, but, in view of its influence on the Dolopathos 
 and the comparatively large number of thirteenth century ver- 
 sions, it must be placed as early as 1 1 50, and it may fall in a 
 time considerably anterior to this. 
 
 The normal number of stories in this branch is fifteen ; of 
 these the queen relates seven, the seven sages one each, and 
 
 1 See Comparetti, Vergil in the Middle Ages, p. 232 f. 
 
 Ward, Gitatogue of Romances, n, p. 122, makes the slight oversight of 
 asserting that the casket-episode of the Merchant of Venice is also intro- 
 duced into the Dolopathos. 
 
 'These stories have had a wide currency, and, in several instances, a 
 most interesting history. For the fullest collections of analogues to them, 
 see the editions of Montaiglon-Brunet and Oesterley, and the appendix to 
 the latter's edition of the Gesta Romanorum. 
 
THE SEVEN SAGES. 25 
 
 the prince the fifteenth. The scene of action is prevailingly 
 Korne, though in two instances K and D it is Constanti- 
 nople. 1 The emperor's name is Diocletian. 2 
 
 The interrelation of the various sub-types into which the 
 Sept Sages falls has been the subject of almost continuous 
 investigation for more than half a century. The first serious 
 attempt at an orderly classification was made by Goedeke in 
 1866 (Orient und Occident, m, p. 402 f). He was followed 
 two years later by Mussafia, 3 in a study which possesses great 
 merit, and which served very much to clear the way for sub- 
 sequent investigation. But it is to Gaston Paris above all 
 that credit is due here for bringing order out of chaos. The 
 Preface to his Deux Redactions is by far the most significant 
 contribution to the study of the Seven Sages which has yet 
 been made, and leaves but the one regret that he has not 
 extended his investigations so as to include the problems of 
 the origin and propagation of the romance. It goes without 
 saying that the excellence of Paris's work has been recognized 
 on all sides, and that his conclusions have been almost uni- 
 versally adopted. 
 
 Paris classifies in five sub-groups, as follows : 
 
 1. S. The Scala Coeli abridgment published by Goedeke. 
 
 2. K. The well known metrical version of Keller. 
 
 3. H. The very large group, of which the Historia is the 
 type. 
 
 4. /. The Versio Italica. 
 
 5. French prose versions (other than JET), including ^4*, 
 , Z>* ( F), and M. 
 
 1. 8. The first of these, the text contained in the Scala 
 Coeli, a compilation of the early fourteenth century by the 
 Dominican Johannes Junior, is a Latin prose abridgment of a 
 lost Liber de Septem Sapientibus. For the latter, Goedeke 
 
 1 This is only partly true of D; see Paris, Deux Redactions, p. 1. 
 
 8 There are several exceptions to this : in K he is called Vespasian ; in 
 D*, Marcomeris, son of Priam (!); in H, Pontianus, the name Diocletian 
 being transferred to the prince. 
 
 8 Wiener Akad. Sitzungsb., Ph. Hist. Cl., LVII, p. 37 f. 
 
26 KILLIS CAMPBELL. 
 
 (who has published the text according to the Scala Coeli in 
 Orient u. Occident, m, p. 402 f.) conjectures a date in the first 
 half of the thirteenth century. An extract in the Summa 
 Recreatorum (xv cent.), which agrees very closely with S, has 
 been pointed out by Mussafia (Wiener Akad. Sitzungsb., Ph. 
 Hist. Cl., LVII, p. 83 f.). 
 
 8 differs materially from H, and is almost as far from K 
 and jD*. It stands nearest to L, having in common with it 
 the two stories filia and noverca in the place of Roma and 
 inclusa of the remaining types. The agreement with Z)*, in 
 that the queen is defended on the last day by a champion, is 
 doubtless a mere coincidence (Paris, /. c., p. vin). Its only 
 influence seems to have been that exercised on L. For 
 Goedeke's claim that it is the closest extant representative 
 of the western original no sustaining argument has yet been 
 brought forward. 1 
 
 2. H. The type of the second group is the well-known 
 Historia Septem Sapientum Romae. Buchner 2 enumerates six- 
 teen manuscripts in which the Historia has been preserved. 
 Its first edition appeared at Cologne in 1472, and the bibli- 
 ographers report many of subsequent date. The latest edi- 
 tion, and only nineteenth century reprint, is that of Buchner. 3 
 An Old French translation, printed at Geneva in 1492, has 
 recently been republished by Paris as the second text of 
 his Deux Redactions (pp. 55-205). The Historia Calumnia 
 Novercali (Antwerp, 1496) differs from it mainly in the 
 omission of all Christian features. 
 
 The Historia is by far the most widely known of all 
 western versions, having had equally as great a vogue in 
 some other European countries Germany for instance as in 
 France. In English the Wynkyn de Worde text (to which 
 
 1 Ward, CaUdoyue of Romances, u, p. 200, erroneously states that Paris 
 upholds Goedeke here. 
 
 * Erlanger Beiiraye zur englischen Pkiloloyie, v, p. 1. Of these six were 
 first pointed out by Paris, /. c., p. \\.\i\, eight by Varnhagen, Eine Ilal. 
 Prosaversion d. Sieben Weisen, p. .\ v. 
 
 * Erlang. Beitr., v, pp. 7-90. An Innsbruck MS. which dates from 1342. 
 
THE SEVEN SAGES. 27 
 
 the many English chap-book versions owe their origin), the 
 Copland, and the Holland versions found in it their ultimate 
 original. With the Germans the Historia type is practically 
 the only one which has found acceptance, and the number of 
 versions, either in Latin or German, which are contained in 
 their libraries is very large. 1 It is under this form, also, that 
 the romance has acquired its popularity in other Germanic 
 and in the Slavonic languages. 2 
 
 The history of opinion with regard to this type of the 
 romance possesses much interest. Until quite recently, as has 
 been seen, H was supposed to be the oldest member of the 
 Western group. Goedeke, in 1866, was the first to break 
 with this tradition, but without showing why. Paulin Paris 
 followed in 1869, throwing the question open. 3 Comparetti, 
 also, in the same year, expressed the opinion that H was far 
 from representing the western original. 4 The matter was not 
 satisfactorily cleared up until the appearance of Gaston Paris's 
 book in 1876. The results of Paris' s investigation (/. c., p. 
 xxviii f.) are to entirely dethrone H from the position which 
 had been traditionally accorded it, and to establish for it a 
 date in the first half of the fourteenth century, and an im- 
 mediate basis on type J.*. 6 
 
 The distinguishing features of H, aside from its slight 
 difference from A* in the order of stories, are the introduction 
 
 1 For the first general discussion of the romance in Germany, see the 
 preface to Keller's Li Romans des Sept Sages, Tubingen, 1837. A more 
 comprehensive discussion of the German versions accompanies his edition 
 of the Hans von Biihel metrical version, Diodelianus LeJ)en (Quedlinburg, 
 1841). 
 
 8 Keller enumerates versions, either in manuscript or in print, in Dutch, 
 Welsh, Icelandic, Swedish, Danish, Polish, Hungarian, Russian, and Arme- 
 nian ; see the prefaces to his two editions cited above. See, also, Murko, 
 "Die Geschichte v. d. Sieben Weisen b. d. Slaven" in Wiener Akad. Sit- 
 zungsb., Ph. Hist. Cl., cxxn, 1890, and "Beitr. zur Textgesch. d. H. S. S." 
 in Zeitschr.f. vergl. Lit.-gesch., pp. 1-34, 1892. 
 
 3 Biblioph. Fran$ais, iv, p. 69 f. 4 Book of Sindibdd, p. 47. 
 
 5 It is hard to see how Landau, Quellen des Dekameron, 2d ed., p. 51 f., 
 and a few others, can still persist in their adherence to the old view. 
 
28 KILLIS CAMPBELL. 
 
 of the stories amatores and amid (the latter appended to 
 vaticinium), the fusion of senescalcus and Roma, and its 
 unusual mass of details. 
 
 3. K. The Old French metrical version, Li Romans des 
 Sept Sages, was published by Keller, at Tubingen, in 1836. 
 Of this version there exists only one complete manuscript, to 
 which its editor gives a date in the late thirteenth century. 
 A fragment of a metrical text agreeing closely with it in 
 content, but differing slightly in order of stories, is preserved 
 in MS. 620 of the Library of Chartres. 1 An edition of this 
 has been promised by Paris. 
 
 K has the same stories as Z>* and A*, but in a different 
 order. The agreement in order, as also in incident, is, as a 
 rule, closest with D* ; in the stories vidua, Roma, inclusa, and 
 vaticinium, however, K exhibits a very close, at times even 
 verbal, agreement with A*. In explanation of this, the possi- 
 bility of an influence of jfiTon^.* is precluded by the fact that 
 the former is of earlier date ; hence it is necessary to posit for 
 A* and K a common source, designated by Paris as V. 
 
 4. I. The Versio Italica was first so styled by Mussafia in 
 his study of the Italian versions, in Jahrb.f. rom. u. englische 
 Lit., iv, p. 166 f., 1862. This group consists of six versions, 
 three of which are in Latin. One of the latter has been 
 brought to light only within the last few years; 2 one was 
 published by Mussafia ( Wiener Akad. Sitzungsb., Ph. Hist. 
 CL, LVII, p. 94 f.) in 1868, and is well known; and the third 
 is the British Museum MS. Addl. 15685. 3 Of the Italian 
 versions one is in verse, 4 but of late date, Rajna in his 
 description (Romania, vn, pp. 22 f., 369 f. ; x, p. 1 f.) plac- 
 
 1 See Paris, I. c., p. in n., and Paul Meyer in the Bulletin d. I. Soc. des 
 Anc. Textes francais, 1894, p. 40 f. The order of stories here is tentamina, 
 Roma, avis, sapientes, vid.ua, Virgilius, inclusa, vaticinium. For the order in 
 K and other versions, see the comparative table, p. 35. 
 
 * By Murko ; see Romania, xx, p. 373. 
 
 1 Ward, Catalogue of Romances, n, p. 207 f. Hitherto unnoticed in this 
 connection. 
 
 4 Edited by Rajna, Storia di Slefano, Bologna, 1881. 
 
THE SEVEN SAGES. 29 
 
 ing it between 1440 and 1480. The two remaining Italian 
 versions early underwent publication, one in 1832 by Delia 
 Lucia, 1 the other by Cappelli in 1865. 2 
 
 The order of stories in J is materially different from that 
 in any other group or version. The queen in this group, 
 instead of relating the first story, follows in each instance the 
 sage, thus reversing the order, 2 becoming 1, 4-3, and so 
 on. In consequence of this innovation, the number of stories 
 is reduced to fourteen, the seventh being crowded out. 3 
 
 In the absence of the filia-noverca and amatores-amici 
 features, I groups itself with K, D*, and A*. Its closest 
 agreement in incident is with A* 9 in which recent scholars 
 believe it to have had its source. 4 
 
 The modern Italian Erasto, which at one time was placed 
 by itself as representing a free adaptation of the romance, 
 and as bearing a somewhat similar relation to the remaining 
 Italian versions as the Dolopathos to the prevailing French 
 type, is now universally acknowledged to be an offspring of 
 the Versio Italica. The Erasto has been very popular in its 
 own country, and has been translated into other languages. 
 The first edition of it appeared at Venice in 1542, the last in 
 1841. An English translation was made by Frances Kirkman 
 in 1674. 
 
 5. French Prose Redactions. The number of French prose 
 redactions is very large. Paris already in 1876 knew of nine- 
 teen manuscripts in Paris, besides the four in Brussels, and 
 one in the Cambridge University Library. A number of 
 others have been since pointed out. 5 
 
 1 Delia Lucia, Novella antica scritia nel buon sec. d. lingua, Venice, 1832. 
 
 2 Cappelli, // libro dei sette savi di Roma, Bologna, 1865. 
 
 3 Jt is interesting to note here that the story thus discarded is zenescalcus, 
 a feature in which the Versio Italica has anticipated one of the English 
 versions Cambridge Ff, n, 38 (F). 
 
 4 See, for the most recent opinion, Rajna in Romania, vii, p. 369 f. 
 
 5 These are mentioned under the discussion of the various groups into 
 which they fall. 
 
30 KILLIS CAMPBELL. 
 
 (1). Paris classifies under the sub-groups D* (T 7 ), L, A*, 
 and M. Of these M the Male Marastre is of little interest 
 other than as showing the immense popularity of the romance 
 in the thirteenth century. Only three manuscripts of it have 
 so far been brought to light. In all these the emperor is 
 Diocletian and the prince, Fiseus; Marcus, son of Cato, 
 is given prominence; and, a feature which distinguishes this 
 sharply from all other groups, six new stories are substituted 
 for a corresponding number of those in the prevailing types. 
 The original of M is believed to have been made on a very 
 mutilated manuscript of the Jl*-type. The new stories, which 
 are of a much lower order than those they displace, are proba- 
 bly the invention of the author. 1 
 
 (2). With M may be associated the numerous ' continuations ' 2 
 of the Sept Sages in French, of which the most important is 
 the Marques de Rome. This type originated in Picardy in the 
 thirteenth century. A version of it has been recently pub- 
 lished by Alton (Li Romans de Marques de Rome, Tubingen, 
 1 889). In the introduction to this edition, the editor states 
 that the romance was certainly not written later than 1277, 
 and probably even forty years earlier (Alton, p. xiv). It 
 seems to have met with considerable popularity, as Alton 
 describes ten manuscripts which still survive. It doubtless 
 had its ultimate basis in A* Alton thinks with M as an 
 intervening stage, but Paris (Romania, xix, p. 493) denies 
 this, maintaining that M is posterior to the Marques. 
 
 (3). D*. The Version Derimee, a unique prose manuscript 
 published by Paris as the first text of his Deux Redactions 
 (pp. 1-55), is thus called on account of the numerous instances 
 of rime still discernible in the text, and which prove beyond 
 doubt a metrical original. 3 
 
 1 See Paris, /. c., p. xxnr f. 
 
 * For these compare P. Paris, Les MSS.franfais de la Bibl. du Roi, Paris, 
 1836, i, p. 109 f. More accessible in Leroux de Lincy, /. c., p. x f. 
 8 This was first shown by Paris, Deux Redactions, p. v f. 
 
THE SEVEN SAGES. 31 
 
 _D* agrees more closely with K than with any other known 
 version. It cannot have been based on K, however, as Paris 
 has shown, but the two doubtless flow from a common source, 
 which Paris designates as V. From this V, also, the Chartres 
 manuscript was in all probability made (Paris, 1. c., p. x.) 
 
 (4). There remain the two families L and A*. The first 
 of these comprises all versions of the type of the first Leroux 
 de Lincy print, 1 in which the order of stories is arbor, canis, 
 , medicus, gaza, puteus, senescalcus, tentamina, Virgilius, 
 rix, sapientes, noverca, filia. Only six manuscripts (four 
 strictly according to L, and two slightly influenced by A*) 
 were known to Paris (I. c., p. 10 f.). To these must be added 
 the Catalan version in ottava rima, edited by Mussafia( Wiener 
 Akad. Denkschr., xxv, p. 185 f., 1376), and five Old French 
 prose manuscripts, partly fragmentary, enumerated by Paul 
 Meyer in Bulletin de la Soc. des Anc. Textes fr. for 1894, 
 p. 38 f. 2 
 
 In its employment of the stories filia and noverca, L at once 
 groups itself with 8. This, however, is not the only feature 
 which the two types have in common. A general comparison 
 with the rest of the Western group serves to show that (if we 
 may except A* for the time being) 8 is also nearest to L in 
 motive (Paris, 1. c., p. xn). In order of stories, too, S and 
 L fall together, the only differences being the reversal on the 
 part of L of tentamina and puteus, and the suppression of 
 vidua and vaticinium. Paris has therefore concluded that L 
 was made on a manuscript of 8 which was mutilated toward 
 the end, and that the scribe has in consequence had to trust to 
 his memory for his last stories (/. c., p. xni). 
 
 1 Leroux de Lincy, Romans des Sept Sages, Paris, 1838, pp. 1-76. 
 
 2 Meyer does not express himself definitely as to the class of but one of 
 these the Chartres MS., which he groups with L. He implies, however, 
 in his statement that the Bib. Nat. fragment (p. 39, n. 2) belongs to A*, 
 that all the rest belong to L. Nevertheless, his notices leave the impres- 
 sion that some of these manuscripts (possibly all except the two just 
 mentioned) have not been handled, and that a part of them may yet be 
 found to belong to the larger group .4*. 
 
32 KILLIS CAMPBELL. 
 
 (5). A*, the largest and most important of all French 
 groups, has been reserved for the last place. To this family 
 pertain, besides its immediate members, the groups Marques, 
 M, I, and H; it is, then, the original, either directly or indi- 
 rectly, of four-fifths of the manuscripts and prints of the 
 romance which survive. It is not only the ultimate source 
 of all Italian versions, whether direct, as with the D'Ancona 
 edition, or indirect through J, but it is also, through H, the 
 parent of almost all the manifold versions of the Sept Sages 
 outside of Romance. And, what is of prime interest and 
 importance to the English student, it was some manuscript 
 of this group which furnished the immediate original of the 
 Middle English versions. 
 
 Under group A* Paris includes all manuscripts of the type 
 of the Italian version published by D'Ancoua. 1 He enumer- 
 ates in his preface (p. xvi f.), in addition to the Italian 
 version whence the group is named, fourteen manuscripts 
 in Old French, 2 several of which date from the thirteenth 
 century. Four other manuscripts, pointed out since the 
 appearance of Paris's work (Brit. Mus. Harl. 3860 [xiv cent.], 
 St. Jno. Bapt. Coll., Oxf., 102 [xiv cent.], 3 Cambr. Univy. 
 Liby. Gg. 6, 28, 4 and a fragment in the Bib. Nat.-Nouv. 
 Acq. fr. 1263 [xm cent.]), 5 increase the number of French 
 versions to eighteen. To this family, also, belongs the British 
 Museum Italian prose version published by Varuhagen. 6 
 
 The text of A* 7 falls into two parts, the first eleven 
 stories (Ai*) being textually very close to L, while the last 
 four (-4 2 *)> as Paris has shown, agree very closely with K. 
 
 1 11 Libro del Sette Savj di Roma, Pisa, 1864. 
 
 2 One of these is the manuscript 2137 of the Bib. Nat., published in part 
 by Leroux de Lincy, pp. 79-110. 
 
 3 For these two, cf. Varnhagen, Z.f. rom. Ph., I, p. 555 f. See also for the 
 first, Ward, /. c., n, p. 199 f. 
 
 4 Romania, xv, p. 348. 
 
 6 Delisle, MSS. lat. et fr. ajoulees aux Fondes, etc., Paris, 1891, I, p. 259. 
 *Eine Ital. Prosaversion der Sieben Weisen, Berlin, 1881. 
 
 7 By this is meant the second Leroux de Lincy redaction. Other versions 
 of this type, as, e. g., MS. 6849 (new No. 189), are not so close to L. 
 
 
THE SEVEN SAGES. 
 
 The composite nature of the text Paris explains as due to the 
 fact that the scribe primarily employed a fragment of L con- 
 taining only eleven tales, and that A", or its source, F, has 
 been used for the remaining four tales. 1 And this seems to be 
 borne out by internal evidence ; for A 2 * not only falls in with 
 jSTas regards incident, but, as in the case of Z)*, there is often 
 even a textual agreement in which entire lines that appear 
 in Jv^are reproduced. 2 Yet, as already observed, this metrical 
 original of J. a * cannot have been K, since there are a number 
 of ^i*-manuscripts which antedate the latter, especially if we 
 may accept Keller, who despite his maintenance of the priority 
 of Kj ventured a date no earlier than 1 284, or later in all 
 probability than the composition of the English parent text. 
 Moreover, a comparison of Af with K and D* will show 
 that each of the latter possesses features in common with .4* 
 which are not found in the other. The original of ^4 2 * must 
 therefore be sought in some other version than K } probably, 
 as Paris assumes, in V. 3 
 
 1 Deux Redactions, p. xvm. 
 
 8 Ibid., p. xix, for a citation of parallel passages from A 2 * and K. Almost 
 as noteworthy agreement will be found in some of the remaining stories. 
 
 3 But can this be final ? Is it not possible, however improbable it may 
 seem, that the manuscripts of A* which have survived were ultimately 
 based on a metrical text which preserved the -A*-order of stories (or, at 
 least, was nearer the ^4*-order than the K-, C*- or _D*-order), and which was 
 closely related with V? In this case, of course, L (the first eleven stories), 
 would have to be explained as based on A* (rather than the reverse, as 
 with Paris), and A 2 * as representing a prosing of a portion of the metrical 
 A*, to which K has very nearly approached. Against this view would 
 be the strong evidence submitted by Paris. In favor of it, however, are 
 the considerations (1) that this would better account for the popularity 
 of the ^4*-type during the first half of the thirteenth century ; (2) that the 
 Middle English versions both favor a metrical original and were based on 
 a text nearer to K in many details than is the De Lincy print of A* ; (3) 
 that to base A* on L, and consequently, as Paris maintains, ultimately on 
 S, is to connect it with a different line of tradition from that which it 
 seems to follow (cf. certain textual agreements with K which A*, L exhibit : 
 p. 16 : " comme il fist au cheualier de son leureier " =: JT 1141-2 : " Comme 
 il fist au cheualier, Ki atort occist son leurier ; " p. 39 : " II apela son senes- 
 chal "= K 1509 : " Lors apiela son seneschal ; " p. 40 : " Vos gerrez auec le 
 
 3 
 
34 KILLIS CAMPBELL. 
 
 Resume*. Looked at externally the Western group falls 
 into two main sub-groups, the Dolopathos and the Sept Sages 
 de Rome. The Dolopathos, however, did not develop from 
 the Eastern group independently, but must have had an 
 ultimate basis (doubtless through an oral medium) on some 
 version of the larger group. 
 
 The Sept Sages de Rome, as regards order and content of 
 stories, also falls into two groups, one represented by S and 
 L, the other by K, Z)*, C*, ( F), and A* and its variants, /, 
 Hy M, and Marques. Peculiar to the former group (8, L) are 
 the stories filia and noverca, to the latter the stories Roma 
 and inclusa. 
 
 Which of these groups represents most faithfully the lost 
 western original is, at the present stage of our knowledge, 
 impossible to determine, but the fact that the Dolopathos of 
 Herbert contains the story inclusa seems to point to the 
 priority of the K-, D*-, J.*-group. 1 
 
 With respect to the separate sub-groups, L may have been 
 based on A* and 8, though the view of Paris, that it had its 
 basis in 8 alone, carries with it greater probability. Either 
 explanation leaves the origin of 8 unexplained. K, D*, (7* go 
 back to the same lost metrical original, V. A* is probably to 
 be explained with Paris as having its source in L and V 9 though 
 this, as yet, has been by no means established. It is not 
 improbable that a metrical version of A* existed at some time. 
 
 roi"=K\53l: "Auoeques le roi vous girois;" p. 50: "Qui me ferra, je 
 trerai fo"= K 3938: "Ki me ferra, je trairai ia"); (4) that we may still 
 find in A*, what appear to be reflections of a versified original ; thus, p. 15 : 
 " Celz que je mout amoie et en qui je me fioie ; " p. 23 : " Li sangliers vint 
 vers 1'alier, si commenya fl mengier," and " quant il vit le sanglier, si s'en 
 volt aler;" p. 33: "Quant eles virent lor pi-re trainer, si commenci6rent 
 (a. brre et) it crier;" p. 50: "Sire, il ot en ceste vile un clerc qui ot non 
 Vergile." When all this is said, however, the case is by no means strong, 
 and we would not presume to insist on this theory as presenting the proba- 
 bility, by any means, which attaches to the view set forth by Paris; it is 
 merely suggested as an alternate possibility, which has not yet been dis- 
 posed of. 
 
 I 8ee also, Paris, Romania, iv, p. 128, for the additional evidence in 
 support of this view drawn from the story Roma. 
 
THE SEVEN SAGES. 
 
 35 
 
 Table of Stories in the Western Versions. 1 
 
 A* 
 
 L 
 
 S 
 
 K 
 
 D* 
 
 H 
 
 / 
 
 M 
 
 Dolopathoa. 
 
 arbor 
 
 arbor 
 
 arbor 
 
 
 
 srDor 
 
 
 \r\\i\r 
 
 
 canis 
 
 canis 
 
 canis 
 
 flrDor 
 canis 
 
 m*i )i > r 
 canis 
 
 canis 
 
 canis 
 
 ctruur 
 
 canis 
 
 canis 
 
 aper 
 
 aper 
 
 aper 
 
 senesc. 
 
 senesc. 
 
 aper 
 
 arbor 
 
 aper 
 
 
 
 medicus 
 
 medicus 
 
 medicus 
 
 medicus 
 
 medicus 
 
 puteus 
 
 medicus 
 
 medicus 
 
 gaza 
 
 gaza 
 
 gaza 
 
 gaa 
 
 aper 
 
 aper 
 
 gaza 
 
 aper 
 
 gaza 
 
 
 
 puteus 
 
 puteus 
 
 tentam. 
 
 puteus 
 
 puteus 
 
 avis 
 
 tentam. 
 
 avis 
 
 senes 
 
 senescalcus 
 
 senesc. 
 
 senesc. 
 
 sapient. 
 
 sapient. 
 
 sapient. 
 
 sapient. 
 
 filius 
 
 
 
 tentamina 
 
 ten tarn. 
 
 puteus 
 
 tentam. 
 
 tentam. 
 
 tentam. 
 
 avis 
 
 vidua 
 
 creditor 
 
 Virgilius 
 
 Virgil. 
 
 Virgil. 
 
 Roma. 
 
 Roma. 
 
 Virgil. 
 
 gaza 
 
 nutrix 
 
 
 
 avis 
 
 avis. 
 
 avis. 
 
 avis 
 
 avis 
 
 medicus 
 
 inclusa 
 
 Antenor 
 
 vid. fil. 
 
 sapientes 
 vidua 
 
 sapient, 
 noverca 
 
 sapient, 
 vidua 
 
 gaza 
 vidua 
 
 gaza 
 vidua 
 
 sen. Rom. 
 amatores 
 
 Roma 
 vidua 
 
 spurius 
 cardamum 
 
 latro. fil. 
 
 Roma 
 
 filia 
 
 filia 
 
 Virgil. 
 
 Virgil. 
 
 inclusa 
 
 Virgil. 
 
 assass. 
 
 
 
 inclusa 
 
 
 
 noverca 
 
 inclusa 
 
 inclusa 
 
 vidua 
 
 puteus 
 
 inclusa 
 
 cjg- eq. 
 
 vaticinium 
 
 
 
 vaticin. 
 
 vaticin. 
 
 vatic, -f- 
 
 vat. amici. 
 
 vaticin. 
 
 vaticin. 
 
 incl. put. 
 
 II. THE ROMANCE IN ENGLAND. 
 
 The enormous popularity of the Seven Sages in French 
 found but a faint reflection in early English. So far, only 
 eight Middle English versions have been brought to light, 
 and as at least seven of these go back to the same lost origi- 
 nal, it appears that the romance did not at first take a very 
 firm root in English soil. Nor has it in more recent times 
 acquired the popularity in England that it enjoyed in other 
 countries of Europe ; for, besides the numerous chap-book 
 versions, all which are of a low order of excellence, there 
 have survived only two versions belonging to the Modern 
 English period. 
 
 Yet, despite this comparatively small popularity of the 
 romance in England, it is very evident that the English 
 
 1 The order of the fragmentary Old French metrical version C* is as 
 follows : tentamina, Roma, avis, sapientes, vidua, Virgilius, inciusa, vaticinium. 
 In the Varnhagen Italian prose version, puteus has been supplanted by a 
 new story, which V. calls mercator. All the Middle English versions save 
 F (for which see p. 62 of this study) follow the 4*-order. The later Eng- 
 lish versions belong to group H. 
 
36 KILLIS CAMPBELL. 
 
 versions have not received attention commensurate with their 
 importance. Indeed, there is no department of the study of 
 the Seven Sages, much neglected though all have unfortu- 
 nately been, which has been more neglected than the English. 
 Weber, the first in the field, offered with his edition of the 
 Auchinleck text practically no introduction at all. 1 Likewise 
 Wright, in the essay which accompanied the Cambridge text 
 (Dd, I, 17), while he presented an abstract of the Historia, 
 confined the discussion of his own text, singularly enough, to 
 less than two pages. 2 Besides these, Ellis in his Specimens? 
 Clouston in his Book of Sindibadf and Gomme in the preface 
 to his reprint of the Wynkyn de Worde edition 5 have sub- 
 mitted analyses of the Weber, Wright, and Wynkyn de Worde 
 editions respectively, and sundry others have made incidental 
 references ; but there has so far appeared only one detailed 
 and serious investigation of the problems which the English 
 versions present the dissertation Ueber die mittelenglischen 
 Fassungen der Sage von den sieben weisen Meistern, Breslau, 
 1885, by Paul Petras. This scholar, in dealing with the 
 source and inter-connection of the English versions, has 
 arrived at some very gratifying results, but his work leaves 
 much to be desired. Three of the eight Middle English 
 versions have escaped notice at his hands, as also, for some 
 unaccountable reason, the well-known edition of Wynkyn de 
 Worde, and a good half of his conclusions may be overthrown 
 by a more thorough investigation. In view, then, of this 
 manifest neglect of the English versions another detailed 
 study of them especially of the relations of the Middle 
 
 1 Metrical Romances, Edinburgh, 1810, I, p. LV and in, pp. 1-153. 
 
 * The Seven Sages, Percy Society Publications, vol. xvi, p. LXVIII, London, 
 1845; also in Warton's History of English Poetry, ed. Hazlitt, London, 1871, 
 I, p. 305 f. 
 
 3 Specimens of Early English Metrical Romances, London, 1811, in, pp. 
 1-101. 
 
 4 Book ofSindibdd [Glasgow], 1884, p. 327 f. 
 
 6 The History of the Seven Wise Masters of Rome, published for the Villon 
 Society, London, 1885. 
 
THE SEVEN SAGES. 37 
 
 English manuscripts will not, it is believed, be deemed 
 untimely. 
 
 II (a). The Middle English Versions. 
 
 The Middle English group comprises eight known versions, 
 in as many different manuscripts. All these are in verse, 
 and in the octosyllabic or four-stressed couplet. 
 
 They are as follows : Auchinleck (^4.), Arundel 140 (Ar), 
 Egerton 1995 (E), Balliol College 354 (5), Cambridge Ff, 
 n, 38 (F), Cotton Galba E, ix (C), Cambridge Dd, i, 17 (D), 
 and Asloan (As). 1 
 
 1. Description of the Manuscripts. 
 
 A. The Auchinleck MS. of the Advocate's Library, Edin- 
 burgh, denoted throughout as A. For a general description 
 of this manuscript, see Kolbing, Englische Studien, vu, p. 
 185 f. The text of the Seven Sages occupies ff. 85a-99d, 
 and is fragmentary at both beginning and end, only 2645 
 lines remaining. It has been published by Weber, Metrical 
 Romances j Edinburgh, 1810, in, pp. 1-153, where it com- 
 prises lines 135-2779, the Cotton MS. (C) having been used for 
 the remainder. For a collation of this edition with the manu- 
 script, see Kolbing, Englische Studien, vi, p. 443 f. Copious 
 extracts with an analysis may be found in Ellis's Specimens, 
 London, 1811, in, pp. 1-101. With regard to date of com- 
 position there is no internal evidence other than linguistic; 
 since, however, the Auchinleck MS. dates from about 1330, 
 the composition of A must fall before that time. 2 The form 
 
 1 1 have handled and made transcripts of all these manuscripts save those 
 which have been printed and the Asloan. Five of them (A, E, (7, F, and 
 D) have been studied either in whole or in part by Petras, and the Asloan 
 MS. was also known to him through Laing's very incomplete description of 
 it in the preface to his edition of the Holland text, p. xn. Of the Arundel 
 and Balliol manuscripts Petras was apparently unaware. 
 
 2 Cf. Morsbach, M. E. Grammatik, Halle, 1896, p. xi, and Brandl in 
 Grundriss, n, 1, p. 635. 
 
38 KILLIS CAMPBELL. 
 
 hardly justifies a dating earlier than 1300. In text and 
 metre A is, as a rule, very good, though in both there are 
 occasional imperfections and corruptions. 1 The dialect is 
 Kentish, though not of the strict type. 2 
 
 Ar. MS. Arundel 140 of the British Museum, cited as 
 Ar. Paper, dating from the first half of the fifteenth century. 
 For general description, see Ward, Catalogue of Romances, n, 
 p. 224. This text occupies ff. 152-1655, and is fragmentary, 
 beginning with the conclusion of aper (3) and ending with 
 the 21st line of vaticinium (15); 2565 lines remain. It is 
 very much faded, and in many cases illegible, especially at 
 the end of the b- and at the beginning of the c-columns. With 
 regard to initial capitalization, it is very irregular. A line 
 has been lost after 1. 618 ; after 1. 919 an extra line has been 
 introduced with no corresponding rime. The text is metri- 
 cally very poor, and many final e's have to be inserted in 
 order to secure the required four stresses ; there are also a 
 number of imperfect rimes (such as yspede : saue, 243-4) 
 and other textual irregularities ; nevertheless, Ar, as is shown 
 below, is the closest representative of the lost M. E. original. 
 The dialect is Kentish. 3 The text has not been published. 
 
 1 There are many emendations which lie on the surface and which are 
 sustained by the closely related versions Ar, E, etc. Some of these are : 
 (1) for schild 1016 read sckuld(e) cf. F 1487, Ar, B, E] (2) for swich 1031 
 read syke or seke cf. Ar 91, etc.; (3) for tol of 2050 read to lofcf. E 2082, 
 etc. ; (4) for to-ddue 2417 read go deluecf. B 2509, etc. ; (5) after He 2657, 
 insert \>ou$t cf. Ar 1782, etc. 
 
 8 A. S. y is regularly represented by the e-sound, though this may not 
 always be graphic. Of the 27 determining rimes, 22, or 81 per cent., 
 have the e-coloring. There is nothing in other developments to contradict 
 this result. The only Northern forms in the rime are a pres. part, in 
 -and, 1977-8, and two instances of the third pers. sing, of the present tense 
 in , 615-6 and 937-8. 
 
 3 To the development of A. S. y (stable or unstable, long or short) into e, 
 there is only one certain exception : vn/ne ; syne, 691-2. Elsewhere we find 
 only the e-quality; cf. nede: hyde, 383-4; ifet: iknet, 601-2; gardyner : fyr, 
 863-4, 872-3; also 892-3, 939-40, 979-80, 1433-4, 1515-6, 1535-6, 1541-2, 
 1583-4, 1761-2, 1847-8, 2059-60. The additional rime-evidence is alto- 
 gether confirmatory of a Southern scribe : A. S. a > o unexceptionally, the 
 
THE SEVEN SAGES. 39 
 
 E. MS. Egerton 1995 of the British Museum, 1 cited 
 throughout as E. Ff. 3-54b. Paper, dating from the fif- 
 teenth century, probably the second half. 2 Written in single 
 columns, with initials in red. Very regular as regards capital- 
 ization. Complete, containing 3588 lines, and bearing the 
 title Seven Sages of Rome, with the colophon Expliciunt Septem 
 Sapientes. Before the first story, arbor, stands the simple 
 rubric, " He[re] begynnythe the fyrste tale of the Emperasse ; " 
 before nine others, there is substituted for this a couplet indi- 
 cating the contents of the story which follows, as e. g., canis 
 
 (695-6) : 
 
 * Here begynnythe the tale of a knyght 
 That cylde hys grehounde with unryght.' 
 
 The stories avis, vidua, Roma, inclusa, and vaticinium have 
 nothing corresponding to this. The dialect is Kentish, though 
 less strongly marked than in Ar. 3 No edition of E has yet 
 appeared. An extract, including 11. 2251-2358, accompanies 
 the monograph of Petras, "Anhang," p. 54 f. 
 
 B. MS. No. 354 of Balliol College Library, Oxford, 
 denoted as -B. 4 Ff. 18a-54b. Paper, belonging to the early 
 
 pres. part, (except buland : blynd, 1589-90) ends in -ng, the verb is Southern 
 (save cryen: mene, 2556-7, where we have a Midland form), the past part, 
 preserves, as a rule, the prefix, and rejects (in the case of the strong verb) 
 the ending, etc. Within the line, however, there are occasional Northern 
 forms, particularly of the pres. part., as buland) 1588, 1591, 1599, brynand, 
 1922 ; but these are by no means the rule, the Southern form being in 
 general preserved as well within the line as in the rime. 
 
 1 For a general description of this manuscript, see Ward's Catalogue, u, 
 p. 218 f. 
 
 8 See the sixth article: "Gregory Skinner's Chronicle of the Mayors of 
 London, ending in 1469," ff. 113-122b. 
 
 'The usual development of A. S. y is e, or the e-quality, see the rimes 
 of 11. 245-6, 577-8, 783-4, 845-6, 1323-4, 1545-6, 1799-1800, 1821-1822; 
 but occasionally y, cf. kynne: lynne (O.N. linna), 1317-8 and wynne: syne, 
 1635-6. The evidence is otherwise strongly indicative of a Southern scribe, 
 though a few Northern forms are borne out by the rime ; cf. hondys : stondys 
 (3d sing.), 439-40, also kynge: yonge, 93-4, and yonge: connynge, 3581-2. 
 
 4 The existence of this version of the Seven Sages was first pointed out by 
 Varnhagen, in his Eine Ital. Prosav. d. Sieben Weisen, Berlin, 1881, p. xi ; 
 see in the same connection his review of Petras, Eng. Stud., x, p. 279 f. 
 
40 KILLIS CAMPBELL. 
 
 sixteenth century. 1 In single columns ; irregular in capitali- 
 zation. Described in Coxe's Catalogus, i, p. 110, as in the 
 hand of John Hyde. The text is complete, containing 3708 
 lines. The first rubric, which contains the title, reads as 
 follows : " Here begynneth ]?e prologes of the VII. sagis or 
 vii. wise masters which were named as here-after ffollowing." 
 Each story has a heading or title, as e. g. y arbor : " The 
 empresse tale off the pynote tree." At the end of the text 
 stands the colophon : " Thus endith of the vii. sages of Rome, 
 which was drawen owt of crownycles and owt of wrytyng of 
 old men, and many a notable tale is ther-in, as ys beffore 
 sayde. Quod Kichard Hill." This manuscript contains very 
 few abbreviations, and the language is much modernized. In 
 line 1761 : "On the ffall suche as fell to a old man by his 
 wif," we have two lines in one. The rime is, if anything, 
 slightly better than in A, Ar, and E, but is, nevertheless, 
 occasionally imperfect, cf. visage : noyse, 459-60; assonance, 
 as in all other related M. E. texts, abounds ; often four lines 
 rime together, and occasionally six, cf. 2583-8. The dialect 
 is Southern. 2 No edition of the text has yet appeared, but 
 the E. E. T. S. has for some time been advertising the entire 
 manuscript as needing editing. 
 
 P. MS. Ff, n, 38 (formerly marked More 690) of the 
 Cambridge University Library, denoted as F. s Ff. 134a- 
 156d. Paper, dating from about the middle of the fifteenth 
 century. Written in double columns of about 40 11. to the 
 column. Handwriting uniform ; irregular as to capitaliza- 
 tion, though most lines begin with a capital. The beginnings 
 of stories indicated merely by large initial capitals in red. 
 
 1 Cf. Art. 31, "Memoranda of Richard Hill," and Art. 98, "Names of 
 Mayors (of London)." 
 
 'Southern forms are sustained by the rime almost without exception. 
 A. S. y is represented by both y and e, in about equal proportion ; the rimes 
 in e are probably to be explained, however, as reminiscences of a Kentish 
 original. 
 
 3 Cf. Halliwell, Thornton Romances, Camden Society, vol. xxx, p. xxxvi f., 
 and the Cambridge Univ. Lib. Catalogue of xss., II, p. 408. 
 
THE SEVEN SAGES. 41 
 
 The text is fragmentary ; ff. 141 and 144 (or less than 400 
 11.) have been lost, and fol. 135 is in a mutilated condition; 1 
 2555 11. remain. Criteria for determining the dialect are not 
 abundant, as the manuscript is late and the forms are some- 
 what mixed ; but the bulk of the evidence favors a Southern 
 dialect. 2 The text has not been edited, although, in view 
 of its uniqueness, it is not uninteresting, and in its last four 
 stories is of considerable value. Extracts are given by Halli- 
 well, Thornton Romances, p. XLIII f., Wright, The Seven Sages, 
 p. LXX f., and Petras, /. c., p. 60 f. 
 
 C. MS. Cotton Galba E, ix, of the British Museum, 
 denoted as C. 3 Ff. 25b-48b. Vellum ; in double columns, with 
 initials in blue and red, and in a very plain hand of the first 
 third of the fifteenth century. Complete, in 4328 11. Bearing 
 the title pe Proces of ])e Seuyn Sages. Each prolog and each 
 story marked off by rubrics : in the case of the former, such 
 as " Here bigins ]?e fyrst proces " (called " prolong " after the 
 fourth story), with the latter, " Here bygins j?e first tale of 
 J>e whyfe," etc., the number being given in each instance, 
 and, in the case of the masters' stories, their names also. 
 The dialect is Northern. Both text and "metre are very 
 pure ; 4 the rime, especially, stands in marked contrast to the 
 Southern versions, being almost free of assonance and the im- 
 
 1 The Cambridge Catalogue fails to specify the leaves which have been lost. 
 Petras (p. 8) and others go to the other extreme in asserting that the text 
 is very incomplete. 
 
 * A. S. a > o, and the forms of the verb, with the exception of the strong 
 past part., where -en is the usual ending, are Southern. The scribe, how- 
 ever, probably belonged rather to the middle or western South than to 
 Kent, or its neighborhood ; cf. the rimes in y where the w-quality prevails : 
 tyme : kynne, 813-4 ; wytte : pytte, 845-6 ; hym : kynne, 871-2 ; 1348-9, 1636-7, 
 etc. The rimes bedd: hydd, 200-1, and kende: sende, 1890-1, are probably 
 to be traced to the Kentish original. 
 
 3 Cf. Ward's Catalogue, n, p. 213 f., for a general description of this manu- 
 script. 
 
 4 There are very few verses that are too short (among these are 84, 443, 
 911, 1868, 1901, 1918, 2973), and almost none that are too full (cf. 843). 
 Among the few inexact rimes are sages: message, 355-6 ; brend: assent, 2321- 
 2; hew: mowe, 2842-3. 
 
42 KILLIS CAMPBELL. 
 
 perfections in which the latter abound. No complete edition 
 of C has so far appeared; but lines 1-134 and 3108-4328 
 are printed in Weber, Metr. Rom., m, pp. 1 f. and 108 f., 
 where this text has been employed to supplement A. The 
 story avis, comprising lines 2411-2548, appears in the "An- 
 hang " to Petras's monograph, p. 56 f. 1 
 
 D. MS. Dd, I, 17 of the Cambridge University Library, 
 cited as D. 2 Ff. 54a, col. 1 63a, col. 3. Parchment; in 
 treble columns; appears to belong to the end of the fourteenth 
 century. 3 Textually very imperfect, and plainly the work of 
 a careless scribe. Thirteen lines have apparently been lost, 
 after 1312, 1417, 1696, 1719, 2094, 2293, 2695, 2840, 2960, 
 3057, 3134, 3365, 3395. Irregularities in rime are numerous, 
 but in most cases easily emended. 4 The dialect is southeast 
 Midland, with an intermixture of Northern forms. 5 The 
 text has been edited by Wright (Percy Society for 1845, vol. 
 xvi, pp. 1-118). For a collation of this edition with the 
 manuscript, see Kolbing in Englische Studien, vi, p. 448 f. 
 An analysis of the romance on the basis of this text appears 
 in Clouston's Book of Sindibdd, p. 327 f. 
 
 As. MS. Asloan, in the possession of Lord Talbot de 
 Malahide, Malahide Castle, Ireland, denoted by As. For a 
 general description of the manuscript (quoted from Chalmers), 
 
 *An edition of this manuscript by the lamented Dr. Robert Morris was 
 announced by the E. E. T. S. many years ago ; and an editor was advertised 
 for for some time after Dr. Morris's death, but in the recent issues of the 
 publications this advertisement no longer appears. It is the purpose of 
 the present writer to prepare a critical edition of this text within the near 
 future. 
 
 * For a general description of this manuscript, see the Cambridge Cata- 
 logue, I, p. 15 f.; Skeat, Publications of E. E. T. S., vol. xxxviu, p. xxm f.; 
 and Halliwell, Manuscript Rarities of Cambridge, p. 3. 
 
 "Morsbach, for some unknown reason, would place it earlier, "1300?"; 
 see his M . E. Qrammatik, p. 9. 
 
 4 Lines 337-9 may be explained as a triplet, but it is better to suppose 
 that a verse has been lost. A more probable example of the triplet in 
 M. E. is found in A, 915-7. 
 
 6 See Skeat, E. E. T. S., vol. xxxvni, p. xxv, and Brandl, in PauPs 
 Grundriss, n, 1, p. 635. 
 
THE SEVEN SAGES. 43 
 
 see Schipper's Poems of Dunbar, Vienna, 1891, Pt. 1, p. 5 f. 1 
 The text of the Seven Sages occupies ff. 167-209, and bears 
 the title The Buke of the sevyne Sagis. According to Laing 2 the 
 text is incomplete, extending to only about 2800 lines, and the 
 twelfth and thirteenth stories are wanting entirely. It begins, 
 
 'Ane Empriour in tymes bygone 
 In Rome callit Dioclesiane ' 
 
 and ends, 
 
 ' Syne geld till heuyn and sa do we 
 Sayis all Amen for cherite.' 
 
 Its dialect is Scottish. 3 A complete transcript, made by D. 
 Laing in 1826, exists in the University Library, Edinburgh. 
 An edition, long ago promised by Varnhagen, is expected to 
 appear shortly in the Scottish Text Society Publications. 
 
 2. Interrelation of the Middle English Versions. 
 
 With regard to the relationship of the Middle English 
 versions there has been a variety of opinions, and, as in the 
 case of the French versions, there has existed no little ignor- 
 ance and error. The general tendency has been to consider 
 any and all versions of the M. E. period independent trans- 
 lations from the French. This has been nowhere better 
 demonstrated than in Petras's dissertation, where it has been 
 boldly maintained that at least four of the M. E. versions 
 (A, Cj Fj D) are unrelated save through a common foreign 
 original. And while others have been more conservative 
 than Petras, the prevailing opinion seems to have been that a 
 majority at least of the M. E. group are independent of each 
 other. It will be one of the results of this study, however, it 
 is believed, to show that seven of the eight M. E. versions 
 
 *A further description, together with an extract containing the story avis, 
 has recently appeared in Englische Studitn (xxv, p. 321 f.), through the 
 kindness of Prof. Varnhagen. 
 
 *The Seven Sages in Scottish Metre (Holland), Edinburgh, 1837, p. xir. 
 
 3 Chalmers says of it: "Evidently written by a Scotish versifier in the 
 reign of James IV, as a number of Scotish terms occur, which would not 
 have been introduced by a Scotish transcriber of an English work." 
 
44 KILLIS CAMPBELL. 
 
 are ultimately related through a common M. E. parent ver- 
 sion (x), and it is held not improbable that the eighth (As) 
 is also thus related to x. 
 
 All the M. E. versions, however, do not represent the same 
 line of tradition. One of the texts, Z), as later shown, is a 
 development from #, independent of the rest of the M. E. 
 group, and Varnhagen holds that As was made directly from 
 the Old French. The remaining versions fall together into 
 one connected group, all related through a common original 
 (y), which goes back to x, but which was not identical with 
 it. This group will be designated as Y. 
 
 The close relationship of the texts which constitute this 
 group Y is confirmed by evidence from all sides, but it can be 
 no more effectively illustrated than by a comparative table of 
 lines. For this purpose a line-for-line comparison of the 
 section which the five most important texts of this group (A, 
 Ar, E, B, C) have in common has been made, the comparison 
 being restricted to identical lines and similar rimes, with the 
 following results : l 
 
 (1) A = 1816 11. (4) B = 1931 11. 
 
 Total II. Ident. II. Sim. rimes. Total II. Idenl. II. Sim. rimes. 
 
 Ar 1916 234 722 A 1816 154 537 
 
 E. 1843 125 636 4r...l916 137 646 
 
 B 1934 154 537 E.... .1843 83 558 
 
 C 2067 26 2 C 2067 13 281 
 
 (2) Ar = 1916 11. (5) C= 2067 11. 
 
 A 1816 234 722 A 1816 26 
 
 E. 1843 169 746 ^r...l916 19 413 
 
 B 1931 137 646 E 1843 11 352 
 
 C 2067 19 413 B 1931 13 281 
 
 (3) E= 1843 11. 
 
 A 1816 125 636 
 
 Ar 1916 169 746 
 
 B 1931 83 558 
 
 C 2067 11 352 
 
 1 An illustration of the method by which these figures have been arrived 
 at may be found in the appendix to this study. F, owing to special features 
 which are discussed below, is excluded from this comparison. 
 
 'Petras, p. 11, finds A and (7, the entire texts being compared, to have 
 1096 similar rimes. 
 
THE SEVEN SAGES. 
 
 45 
 
 But this comparison, while valuable as far as it goes, serves 
 only to show a connection between the texts compared ; it 
 does not suffice to show the nature of this connection. 
 Accordingly, in addition to this, a comparison of motive or 
 incident as a safer basis for classification has been made 
 for the entire Middle English group ; and it is by means of 
 this, in the main, that our results as to the interrelation of the 
 M. E. versions have been reached. The limits of this publi- 
 cation, however, preclude the submitting this except in part, 
 so that only the tabulation for the story vidua (Matron of 
 JEphesus) appears here. 
 
 (1) A certain knight had a 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 A* t 80, " un 
 
 wife. (A, Ar, B, D state 
 
 
 
 
 
 
 
 
 vicomte en 
 
 that he was a sheriff.) 
 
 
 
 
 
 
 
 
 Loherainne." 
 
 (2) They loved each other 
 
 A 
 
 (Ar) 
 
 E 
 
 B 
 
 C 
 
 F 
 
 
 A* 
 
 exceedingly. (Ar only 
 relates that he loved her. 
 
 
 
 
 
 
 
 
 
 In F, he will not permit 
 
 
 
 
 
 
 
 
 
 her to go half a mile 
 
 
 
 
 
 
 
 
 
 from him, " neither to 
 
 
 
 
 
 
 
 
 
 church nor to cheping.") 
 
 
 
 
 
 
 
 
 
 (3) A new sharp knife is 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 A* 
 
 given them. 
 
 
 
 
 
 
 
 
 
 (4) While playing with this, 
 
 
 
 
 
 
 
 
 
 {in the 
 
 
 
 
 
 
 
 
 
 til 11 111 L) 
 
 in the 
 
 
 
 E 
 
 B 
 
 
 
 
 
 womb. 
 
 A 
 
 Ar 
 
 
 
 
 
 
 A*, "el 
 
 (C, in the finger; F t in 
 
 
 
 
 
 
 
 
 pouce." 
 
 the hand ; D is silent 
 
 
 
 
 
 
 
 
 
 as to part. F adds that 
 
 
 
 
 
 
 
 
 
 the wife was paring a 
 
 
 
 
 
 
 
 
 
 pear.) 
 
 
 
 
 
 
 
 
 
 (5) For dole he dies on the 
 
 A 
 
 Ar 
 
 E 
 
 
 C 
 
 
 D 
 
 A* 
 
 morrow. (F adds that 
 
 
 
 
 
 
 
 
 
 he asks for a priest be- 
 
 
 
 
 
 
 
 
 
 fore he dies.) 
 
 
 
 
 
 
 
 
 
 (6) This was great folly. 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 
 
 
 A* 
 
46 
 
 KILLIS CAMPBELL. 
 
 (7) He was richly buried 
 
 A 
 
 Ar 
 
 E 
 
 (*) 
 
 C 
 
 
 W 
 
 A* 
 
 on the morrow. (JB does 
 
 
 
 
 
 
 
 
 
 not specify that it was 
 
 
 
 
 
 
 
 
 
 on the morrow. E, B, C 
 
 
 
 
 
 
 
 
 
 state that this occurs 
 
 
 
 
 
 
 
 
 
 after a mass. D adds 
 
 
 
 
 
 
 
 
 
 that the place of burial 
 
 
 
 
 
 
 
 
 
 was outside the city, 
 
 
 
 
 
 
 
 
 
 since there were objec- 
 
 
 
 
 
 
 
 
 
 tions to his being buried 
 
 
 
 
 
 
 
 
 
 within the city.) 
 
 
 
 
 
 
 
 
 
 (8) The wife refuses to leave 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 A* 
 
 the grave. 
 
 
 
 
 
 
 
 
 
 (9) Her friends try to com- 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 D 
 
 A*, "ses lig. 
 
 fort her. 
 
 
 
 
 
 
 
 
 nages." 
 
 (10) They suggest that she is 
 
 A 
 
 Ar 
 
 
 B 
 
 C 
 
 
 
 (4*),"jueneet 
 
 young, and may marry 
 again, and beget chil- 
 
 
 
 
 
 
 
 
 bele." (No 
 mention of 
 
 dren. 
 
 
 
 
 
 
 
 
 marry ing in 
 
 
 
 
 
 
 
 
 
 A* t but see 
 
 
 
 
 
 
 
 
 
 K and the 
 
 
 
 
 
 
 
 
 
 D'Ancona 
 
 
 
 
 
 
 
 
 
 text.) 
 
 (11) She rejects their sugges- 
 
 A 
 
 Ar 
 
 
 B 
 
 C 
 
 
 
 A* SI. 
 
 tions, assuring them that 
 
 
 
 
 
 
 
 
 
 she will die on his grave. 
 
 
 
 
 
 
 
 
 
 
 They are sorry. 
 
 
 
 
 
 
 
 
 
 (12) They make for her a 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 A*, "une 
 
 "logge" on the grave. 
 
 
 
 
 
 
 
 
 loge." 
 
 (13) Also, a fire. (D, she 
 
 A 
 
 Ar 
 
 E 
 
 
 C 
 
 
 (D} 
 
 A* 
 
 makes the fire herself. 
 
 
 
 
 
 
 
 
 
 An addition of D is 
 
 
 
 
 
 
 
 
 
 that she sends for her 
 
 
 
 
 
 
 
 
 \ 
 
 clothes.) 
 
 
 
 
 
 
 
 
 
 (14) Her friends leave her; 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 D 
 
 \A* 
 
 she moans. 
 
 
 
 
 
 
 
 
 
 (15) On the same day three 
 
 A 
 
 Ar 
 
 (E) 
 
 (B) 
 
 C 
 
 (F) 
 
 D 
 
 A*, "& celui 
 
 thieves have been taken. 
 
 
 
 
 
 
 
 
 jour." 
 
 (E, on a day before ; , 
 
 
 
 
 
 
 
 
 
 silent; F, one thief.) 
 
 
 
 
 
 
 
 
 
 (16) They were knights who 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 A* 
 
 had wasted the country, 
 
 
 
 
 
 
 
 
 
 and had been hanged as 
 
 
 
 
 
 
 
 
 
 soon as captured. 
 
 
 
 
 
 
 
 
 
THE SEVEN SAGES. 
 
 47 
 
 (17) A certain knight was to 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 -<4*,"un chev- 
 
 guard the bodies for the 
 
 
 
 
 
 
 
 
 alier la 
 
 first night. (A adds that 
 
 
 
 
 
 
 
 
 I>ivmiere 
 
 he was to watch for three 
 
 
 
 
 
 
 
 
 nuit." 
 
 nights.) 
 
 
 
 
 
 
 
 
 
 (18) Becoming cold, he spies 
 the fire in the " church- 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 A*, " cime- 
 
 haw," goes thither, and 
 
 
 
 
 
 
 
 
 
 finds the lady. 
 
 
 
 
 
 
 
 
 
 (19) He asks to be let in. 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 
 F 
 
 
 ,4*82. 
 
 (20) She refuses his request. 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 A* 
 
 In A she swears by St. 
 
 
 
 
 
 
 
 
 
 (John, in Ar, E, B, by 
 
 
 
 
 
 
 
 
 
 " St. Austyn.") 
 
 
 
 
 
 
 
 
 
 (21) He assures her that he 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 A* (K 3768, 
 
 will do her no harm, 
 
 
 
 
 
 
 
 
 "Je suiGe- 
 
 and that he is a knight. 
 
 
 
 
 
 
 
 
 rart le fil 
 
 
 
 
 
 
 
 
 
 Guion;" 
 
 
 
 
 
 
 
 
 
 also D* 37.) 
 
 (22) She lets him in ; he 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 A* 
 
 warms by the fire. (In 
 
 
 
 
 
 
 
 
 
 D there is no mention 
 
 
 
 
 
 
 
 
 
 of the wife's refusing 
 
 
 
 
 
 
 
 
 
 to permit the knight to 
 
 
 
 
 
 
 
 
 
 enter.) 
 
 
 
 
 
 
 
 
 
 (23) He sees her making 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 A* 
 
 dole, and tells her she 
 
 
 
 
 
 
 
 
 
 is foolish to do so, that 
 
 
 
 
 
 
 
 
 
 she may yet marry some 
 
 
 
 
 
 
 
 
 
 knight. She replies that 
 he was so kind that she 
 
 
 
 
 
 
 
 
 
 may not love any other. 
 
 
 
 
 
 
 
 
 
 (D adds that she begins 
 
 
 
 
 
 
 
 
 
 to love him when she 
 
 
 
 
 
 
 
 
 
 finds him to be a knight ; 
 
 
 
 
 
 
 
 
 
 and that he lies with 
 
 
 
 
 
 
 
 
 
 her.) 
 
 
 
 
 
 
 
 
 
 (24) By and by he thinks of 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 (A*} 
 
 his charge. 
 
 
 
 
 
 
 
 
 
 (25) And fearing guile, he 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 ^4*83. 
 
 rides fast to the gallows, 
 
 
 
 
 
 
 
 
 
 only to find one of the 
 bodies stolen. (^4, Ar, E, 
 
 
 
 
 
 
 
 
 
 B, he rides on a foal. ) 
 
 
 
 
 
 
 
 
 
48 
 
 KILLIS CAMPBELL. 
 
 (26) He fears he will lose his 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 D 
 
 A* 
 
 advancement if unable 
 
 
 
 
 
 
 
 
 
 to recover the body. 
 
 
 
 
 
 
 
 
 
 (27) Bethinks himself that 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 
 A* (the order 
 
 "wirnmen cou>e red." 
 
 
 
 
 
 
 
 
 of 26-7 re- 
 
 
 
 
 
 
 
 
 
 versed in the 
 
 
 
 
 
 
 
 
 
 French.) 
 
 (28) So going to the widow, 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 A* (cf. K, 
 
 he asks counsel of her. 
 
 
 
 
 
 
 
 
 3817). 
 
 (29) She agrees to help him if 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 
 A* 
 
 he will marry her. (B, 
 E, she proposes only that 
 
 
 
 
 
 
 
 
 
 he be her "leman," 
 
 
 
 
 
 
 
 
 
 he suggests matrimony. 
 
 
 
 
 
 
 
 
 
 In C, she asks if he has 
 
 
 
 
 
 
 
 
 
 a wife. ) 
 
 
 
 
 
 
 
 
 
 (30) This being agreed to, she 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 A* 
 
 advises that they dig up 
 the body of her husband, 
 
 
 
 
 
 
 
 
 
 which is done. 
 
 
 
 
 
 
 
 
 
 (31) But the knight objects 
 
 A 
 
 Ar 
 
 E 
 
 
 C 
 
 F 
 
 
 A* 
 
 to hanging up the body. 
 
 
 
 
 
 
 
 
 
 (32) The lady puts a rope 
 round the neck of the 
 
 A 
 
 Ar 
 
 (E) 
 
 B 
 
 C 
 
 F 
 
 
 A* 
 
 corpse. (E, the knight 
 
 
 
 
 
 
 
 
 
 does it.) 
 
 
 
 
 
 
 
 
 
 (33) She draws the body up, 
 
 A 
 
 Ar 
 
 
 
 C 
 
 F 
 
 
 A* 
 
 and hangs it fast. 
 
 
 
 
 
 
 
 
 
 (34) The knight is aghast at 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 
 
 
 
 this. 
 
 
 
 
 
 
 
 
 
 (35) The knight recalls that 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 F 
 
 D 
 
 A* 84, " une 
 
 the thief had a wound in 
 
 
 
 
 
 
 
 
 plaie en la 
 
 his head, and fears that 
 
 
 
 
 
 
 
 
 teste." 
 
 the " guile may be per- 
 ceived" unleas the nus- 
 
 
 
 
 
 
 
 
 
 band have a similar one; 
 
 
 
 
 
 
 
 
 
 this the wife advises him 
 
 
 
 
 
 
 
 
 
 to make with his sword. 
 
 
 
 
 
 
 
 
 
 (36) He declines to do it. 
 
 A 
 
 Ar 
 
 E 
 
 W 
 
 (C) 
 
 
 D 
 
 
 (37) She asks for his sword, 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 
 
 F 
 
 
 A* 
 
 proposing to do it herself. 
 
 
 
 
 
 
 
 
 
THE SEVEN SAGES. 
 
 49 
 
 (38) She smites with all her 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 (C) 
 
 F 
 
 D 
 
 A* 
 
 strength " amid the 
 
 
 
 
 
 
 
 
 
 brayn." (In Z>, she 
 
 
 
 
 
 
 
 
 
 wounds him with a 
 
 
 
 
 
 
 
 
 
 knife.) 
 
 
 
 
 
 
 
 
 
 (39) The knight now knows 
 
 A 
 
 AT 
 
 E 
 
 B 
 
 C 
 
 
 
 
 her to be false. 
 
 
 
 
 
 
 
 
 
 (40) He remembers that the 
 
 A 
 
 AT 
 
 E 
 
 B 
 
 c 
 
 (F) 
 
 (D) 
 
 A* 
 
 thiefs fore-teeth had 
 
 
 
 
 
 
 
 
 
 been broken out. (Z>, 
 
 
 
 
 
 
 
 
 
 F, in agreement with 
 
 
 
 
 
 
 
 
 
 A* t K, have two teeth; 
 
 
 
 
 
 
 
 
 
 but see D* 39, toutes les 
 
 
 
 
 
 
 
 
 
 dens.) 
 
 
 
 
 
 
 
 
 
 (41) She proposes that he dis- 
 figure her husband in like 
 
 A 
 
 AT 
 
 
 
 C 
 
 
 D 
 
 (A*) 
 
 manner, but he refuses. 
 
 
 
 
 
 
 
 
 
 (42) She does it herself with 
 
 A 
 
 AT 
 
 E 
 
 B 
 
 c 
 
 F 
 
 D 
 
 A* 
 
 a stone. (In A, Ar, E, 
 
 
 
 
 
 
 
 
 
 B, F, she knocks out all 
 
 
 
 
 
 
 
 
 
 his teeth ; in Z), only two. 
 
 
 
 
 
 
 
 
 
 F inserts here another 
 
 
 
 
 
 
 
 
 
 disfiguration the loss of 
 
 
 
 
 
 
 
 
 
 two fingers. In D, the 
 
 
 
 
 
 
 
 
 
 body is not hung up till 
 
 
 
 
 
 
 
 
 
 after the mutilation. ) 
 
 
 
 
 
 
 
 
 
 (43) The wife states that she 
 
 A 
 
 AT 
 
 E 
 
 B 
 
 c 
 
 F 
 
 
 (A*) 
 
 has now won his love, 
 
 
 
 
 
 
 
 
 
 which he denies, adding 
 
 
 
 
 
 
 
 
 
 that he would marry her 
 
 
 
 
 
 
 
 
 
 for no treasure, lest she 
 
 
 
 
 
 
 
 
 
 serve him as she has 
 
 
 
 
 
 
 
 
 
 served her lord. 
 
 
 
 
 
 
 
 
 
 (44) The sage wishes Diocle- 
 
 A 
 
 AT 
 
 E 
 
 B 
 
 c 
 
 F 
 
 D 
 
 A* 85. 
 
 tian such fortune if he do 
 
 
 
 
 
 
 
 
 
 not respite the prince. 
 
 
 
 
 
 
 
 
 
 (45) He asks that judgment 
 
 A 
 
 AT 
 
 (E) 
 
 B 
 
 c 
 
 
 
 A* 
 
 be suspended till the 
 
 
 
 
 
 
 
 
 
 morrow, when the prince 
 
 
 
 
 
 
 
 
 
 will speak for himself. 
 
 
 
 
 
 
 
 
 
 (46) The emperor agrees to 
 
 A 
 
 AT 
 
 
 B 
 
 c 
 
 
 D 
 
 A* 
 
 this, and the crowds dis- 
 
 
 
 
 
 
 
 
 
 perse. 
 
 
 
 
 
 
 
 
 
 4 
 
50 
 
 KILLIS CAMPBELL. 
 
 (47) The emperor goes to 
 
 A 
 
 Ar 
 
 E 
 
 E 
 
 C 
 
 F 
 
 D 
 
 A* 
 
 his bower; the empress 
 
 
 
 
 
 
 
 
 
 " lours " on him. (A, Ar 
 
 
 
 
 
 
 
 
 
 add that his "sergeants 
 
 
 
 
 
 
 
 
 
 make solace " with him.) 
 
 
 
 
 
 
 
 
 
 (48) The emperor is brought 
 
 A 
 
 AT 
 
 
 
 
 
 
 
 abed with riche baudekines. 
 
 
 
 
 
 
 
 
 
 (49) The empress is silent till 
 
 A 
 
 Ar 
 
 E 
 
 B 
 
 C 
 
 
 D 
 
 
 the morrow. 
 
 
 
 
 
 , 
 
 
 
 
 (50) When she asks if he has 
 
 A 
 
 (Ar} 
 
 E 
 
 B 
 
 C 
 
 
 (D) 
 
 A*, K 2347, 
 
 heard the "geste," etc., 
 
 
 
 
 
 
 
 
 " feste aus 
 
 why men made a feast of 
 
 
 
 
 
 
 
 
 fox." 
 
 fool*. 1 (Ar, "HowKome 
 
 
 
 
 
 
 
 
 
 was in great dread." D 
 
 
 
 
 
 
 
 
 
 likewise makes no men- 
 
 
 
 
 
 
 
 
 
 tion of the feast of fools.) 
 
 
 
 
 
 
 
 
 
 A. A is naturally the most valuable of all Middle Eng- 
 lish versions, since it is found in the oldest manuscript which 
 has come down to us, and doubtless in many respects best 
 preserves the original. In view of its age one would at least 
 hope to find in it either the parent English text or the closest 
 representative of it, but a close collation with the remaining 
 manuscripts shows that it is neither the one nor the other. 
 It is not even a link in any one of the chains of development. 
 This is established by the fact that A often abridges where all 
 the other texts of Fare true to the French. 2 
 
 There are, however, some features in which A appears to 
 reflect the original more faithfully than any other member 
 of its group. Thus, we find in A 666, "Deu vous doint 
 bonjour" = i 15, " Diex yos doint bon jor," where none 
 approximate A save B 652, "And sayde, deux vous garde 
 bonjour;" or, in A 743, "The levedi stod in pount tournis"= 
 
 1 For the origin of this feature, see Paris, Romania, iv, 128. 
 
 *This phenomenon does not seem to be confined to our text, but appears 
 also in other poems of the Auchinleck MS., as has been already observed 
 by Kolbing ; cf. his Arthour and Merlin, iv, p. CLIII, and his Bevis of Ham- 
 loun, E. E. T. S., Ex. Ser., LXV, p. XLI. 
 
THE SEVEN SAGES. 51 
 
 L 17, "sur le pont torneiz," where C reads "on a vice," and E y 
 B, " in the castle on high." And there are sundry details of 
 the original which A reproduces in common with only one 
 other text; but these are easily explained by the circumstance 
 of A's closer proximity in time to the parent text, in conse- 
 quence of which it has suffered less from the ravages of time, or 
 at the hand of the modernizer, than have some of the later texts. 
 
 The abridgments of the original which characterize A fall 
 chiefly in the conclusions of certain stories. In fact it is a 
 noticeable feature due probably to the desire to avoid repeti- 
 tion that it is almost entirely in the 'epilogaciouns' (as some 
 of the H- texts name them) that A has made any serious altera- 
 tions, while there is a very marked agreement, and only 
 occasional freedom, exhibited in the body of its stories. 
 
 This tendency to abridge is manifest throughout the ^4 -text. 
 It is most violent, however, in the stories aper, gaza, Virgilius, 
 and avis. Chief among the passages in other versions which 
 find nothing corresponding in A, are the following : (1) aper, 
 Ar, 1-20 = E 949-968 = B 933-948 = C 1041-1058 = 
 L, p. 25 ; (2) Virgilius, Ar 1280-1288 = E 2204-2212 = B 
 2244-2252 = C 2370-2376 = L, p. 55 ; (3) avis, Ar 1433- 
 1446 = E 2367-2372 = B 2401-2414 = L, p. 59. 
 
 There is, in addition to these, in the conclusion of gaza, a 
 fourth passage which A abridges radically, and which, since it 
 is a comparatively close paraphrase of the Old French, may 
 be cited here as giving a graphic illustration of this pecu- 
 liarity of A 9 and, at the same time, as showing once for all 
 its unoriginality, and its subordinate importance in settling 
 the question of the interrelation of the English versions. 
 This passage is, in Ar, 11. 456-479 ; the corresponding lines 
 are, in E 1401-1426, B 1393-1420, and C 1472-1490. Cita- 
 tion is made from Ar as best representing the lost text Y. 
 
 Ar 456 ' Loude >ei gonne on hym to crye, L 34. ' Chascun li escria : 
 
 And saide, lentylyon kybe J>y mastry, Ha! mestre, or pansez de 
 Helpe by disciple at Ms nede. vostre deciple.' 
 
 pe master a-lygt \>o of his stede, . . . ' et descent de son 
 
 cheval.' 
 
52 
 
 KILLIS CAMPBELL. 
 
 460 And grete be Emperowr on his kne. 
 Unnebe wold he hym see. 
 pe Emperour saide, bou fals man, 
 Be hym bat al men-kynde wan, 
 pou art fekell and fatour, 
 
 465 Losenger and eke traytowr. 
 A, why syr leue lord ? 
 So nas I neuer, saue by word. 
 Syr, by gentyll wyue late us her, 
 And vfith goddes helpe we schull us 
 skor. 
 
 470 I sow toke my son to loke 
 
 And for to tech hym on boke, 
 And bou first bygan to tech, 
 By-nome his tong and his spech, 
 And taugt hym sith with mor stryf, 
 
 475 Ffor to nyme forth my wyf. 
 ge schull wite beir-of nougt ; 
 Bot when he is to debe brougt, 
 I schull dampne be and by feren 
 
 479 To drawe and honge by be swyren.' 
 
 '. . . et s'en vient devant 
 1'emperdeur, si le salue : . . 
 Li empereres respont au 
 salu qui li a dit: Ja dex 
 ne vos beneie.' 
 
 'Avoi ! fet messires Lan- 
 tules, pourcoi dites vos ce? 
 
 ' Ge le vos dirai, fait li 
 empereres, je vos avoie 
 baillie mon fil & aprendre 
 et a endoctriner, et la pre- 
 miere doctrine que li avez 
 faite, si est que vos li avez 
 la parole tolue ; 1'autre qui 
 veult prendre ma fame a 
 force. Mes ja Dex ne vos 
 en doint joir ; et bien sa- 
 chiez que tantost comme 
 il sera morz, vos morroiz 
 apres, et seroiz destruit 
 ensement.' 
 
 As against this A has only the following lines (1387-92) : 
 
 'And th' emperour wel sone he fond : 
 He gret him faire, ich understond. (= Ar 460) 
 Th' emperour saide, so God me spede, (= AT 462) 
 Traitour, the schal be quit thi mede ! 
 For mi sones mislerning, 
 Ye schulle habbe evil ending ! ' 
 
 Other less important omissions occur in the conclusions to 
 aper and puteus : aper the people invoke the master to help 
 his disciple (L 25, C 1064, E, J5); puteus the empress 
 threatens, on learning of the respite of the prince, to leave on 
 the morrow. Ar 624-5, "And saide scho wold away at 
 morowe. Nai dame, he saide, jef God it wyll. . . ." = L 38, 
 "je m'en irai le matin. Non ferois, dame . . . . se dieux 
 plest." The same incident is omitted in the J.-text of avis ; 
 cf. L 59, Ar 1440-1. 
 
 In the body of the stories, as already observed, this tendency 
 is not nearly so marked. There is in fact no significant 
 
THE SEVEN SAGES. 53 
 
 feature of the stories of the original which has been preserved ' 
 in any other English version that does not appear also in A. 
 The nearest approaches to such are the following, both from 
 the story Roma: (1) An old wise man (== A* 86, " un home 
 viel et ancien. . f .") makes the proposition that the city be put 
 in charge of seven sages, a bit of detail which is omitted by no 
 other English version ; (2) after these sages have kept the 
 city for a month, the food supply is exhausted ; cf. Ar, E, B, 
 C, F, and A* 86, " vitaille failli a ceuls." In addition to 
 these there are certain other minor details in which one or 
 more of the related English versions preserve the French 
 more closely. For example, in medicus (A 1149), Ypocras 
 pierces the ton in 1000 places, as against Ar (208), E, B, F, 
 which agree with L 28, -c- broches. Likewise in Virgilius, 
 A (1977-8) translates the O. F. "arc de coivre et une sajete, 
 bien entesse " (L 50) as " arblast .... and quarel taisand," 
 while the remaining members of group ^render more literally 
 bow and arrow; in sapientes, C, Ar, E, B have the masters 
 ask Merlin his name, in agreement with L 60, " et li demand- 
 fcrent commant il avoit a non," where A abridges ; to which 
 add that A makes no mention of the divine service at the 
 burial of the husband in vidua, where E, B, C, fall in with 
 A* 80, and that in the same story, A (2618) has the knight 
 come to the gallows to watch three nights, while Ar, E, B, C 
 fall together in their adherence to the French A* 81, " la 
 premiere nuit," and we have the sum of A's noteworthy 
 variations within the body of its stories. 
 
 Additions in A are even less numerous. An occasional 
 extra couplet (so far as the evidence of the remaining English 
 versions goes) now and then crops out, as e. g., 645-8, and we 
 also find here and there additional details, such as (1) in Vlr- 
 gilius, where the poor, in addition to warming themselves at 
 the magician's wonderful fire, are represented as also prepar- 
 ing their food by it (A 1973); and as (2) in sapientes, Herod 
 is described as the richest man in Christendom (A 2340), 
 neither of which appears in any other text, whether English 
 
54 KILLIS CAMPBELL. 
 
 or Romance. But such additions are very few in number, 
 and, in any case, too insignificant to play a prominent part in 
 solving the problem in hand. They are, nevertheless, con- 
 firmatory of the evidence already adduced, with which they 
 unite in demonstrating conclusively the unorjginality of A. 
 
 We have, then, in A a secondary development from the 
 lost y. It cannot have been based on any manuscript of which 
 any other text of Y is a close transcript, since it preserves 
 the original in some places more faithfully than any other 
 M. E. text. On the other hand, it cannot have been the 
 source of any of the known M. E. manuscripts, since all these 
 preserve features of the French which A omits. 
 
 Ar. Nearest to A stands the fragmentary text from MS. 
 Arundel 140. This version, while most important as repre- 
 senting in all probability the lost y more closely than any 
 other known text, has been singularly neglected by former 
 investigators. Petras makes no mention of it, whence we 
 draw the inference that he was unacquainted with it. And 
 apparently the only notice which has been accorded it, beyond 
 Varnhagen's several references to it, 1 is that of Ward in his 
 Catalogue of Romances (u, p. 224 f.). From a comparison 
 of the introductory lines of Ar with the corresponding passages 
 in A, E, C, Ward observed that its affinities seemed closest 
 with E; and this indeed holds for the conclusions of several 
 of the stories (Ward deals with a conclusion ; cf. our parallel- 
 ling of lines for medicus, in Appendix), where A has been seen 
 to be often free, and where Ar, in consequence, frequently 
 agrees more closely with any other text than with A. It does 
 not hold, however, as regards the stories themselves, where 
 E yields the first place to A. 
 
 Except in these conclusions, Ar agrees with A very closely. 
 Their intimate relation is evident at once from our line-for- 
 line statistics on p. 44. Of the 1916 lines of the ^4r-section 
 (A 1816), 234 are identical with lines in A, and there are 
 
 1 First referred to in his Eine Ital. Prosaversion d. Sieben Weisen, p. xi, 
 and later in his review of Petras, Eng. Stud., x, p. 279. 
 
THE SEVEN SAGES. 55 
 
 722 similar rimes. Next comes E (1843 11.) with 169 iden- 
 tical lines and 746 similar rimes, a slightly larger percentage 
 of rimes than for A, and an apparent discrepancy, which is, 
 however, easily reconciled by the fact of A's characteristic 
 curtailments; B (1931 11.) has 137 lines identical with Ar 
 and 646 like rimes, and C, which comes last, has only 19 lines 
 identical and 413 similar rimes. 
 
 But the closer relationship of Ar to A develops conclusively 
 only from a comparison of details. Here, while a careful colla- 
 tion of Ar with all other members of Y reveals no noteworthy 
 bit of detail in common with any other single text when con- 
 trasted with A, there are several interesting and significant 
 agreements of Ar with A against the rest of F. Among these 
 are the following : (1) A 1462, " Ich wille bicome wod and 
 wilde," which is identical with Ar 552; in E 1498, the 
 empress (who is speaking here) seeks to slay herself (cf. L 36, 
 "seroie-je rnorte"). (2) A 1580, "And he com als a 
 leopard " = Ar 668, "pane cam he rynnyng as a lyvarde." 
 (3) A 1588, "Bihote hem pans an handfolle " = Ar 676, 
 " Behote heme pens a pours full." (4) A 2396, "Al to loude 
 thou spak thi latin " = Ar 1518, "To loude ]xm spake j?y 
 latyn." (5) A 2744, "Withe riche baudekines i-spredde" = 
 Ar 1868, "With rich clones all byspred." None of these 
 verses have anything corresponding in any other English text. 
 Doubtless some of them are only accidental, but such cannot 
 be the case with all. Their evidence is well supported by 
 such further agreements as in senescalcus, where A and Ar 
 unite in retaining the twenty marks of the original, other 
 M. E. texts varying, or as in vidua where these two agree in 
 that the wife is cut in the womb, while E, B preserve the 
 French in the thumb (A* 80, el pouce), C states that the 
 wounded part is a finger, E the hand, and D is indefinite. 
 Of these agreements there can be only one explanation, namely 
 in the assumption of a connection between the two texts. 
 What the nature of this relation is, however, can be best 
 
56 KILLIS CAMPBELL. 
 
 determined after a collection of corresponding data for the 
 other manuscripts. 
 
 In comparing the remaining texts with Ar, one is at once 
 struck with the remarkable agreement of B, E with A, Ar. 
 These four versions have a number of features in common 
 which do not survive in C, F, or D. Thus (1) in gaza, the 
 son stabs himself in the thigh (= L 33, en la cuisse), where 
 C, F are free, the one reading cheke, the other honde. (2) In 
 senescalcus, the king falls sick " by God's vengeance " (not in 
 L ; also omitted by C, D, F omitting the entire story). (3) 
 Again in the same story, the king offers twenty marks or 
 pounds for a lady to lie with (= L 40, xx mars), where C 
 reads ten pounds, and D simply " gold and silver." And this 
 is still more apparent in a line-for-liue collation, as is suffi- 
 ciently demonstrated in the Appendix. 
 
 At the same time, also, one cannot but remark certain 
 occasional agreements of Ar with E, B in opposition to A. For 
 instance, (1) the king in senescalcus, with the former, has 
 great delight in women, where A on the contrary, in agree- 
 ment with the O. F., as also with C, D, describes him as 
 disdaining women above all things (L 39, " II desdaingnoit 
 fame seur toutes riens "). And (2) in sapientes, the sages in 
 Ar, E, B ask respite for seven days, where A, C give four- 
 teen days, F 12, L 4-8, and K 15. Likewise (3) the servants 
 of the king in sapientes dig under his bed " four feet or five " 
 in Ar, E, B, while A makes no mention of the distance, but 
 says ten or twelve men dig ; so L 62, xx homes. To which 
 is to be added (4) the agreement of Ar, E, B in having the 
 husband in vidua (Ar 1756) swear by St. Austyne ; by 
 St. Johain in A (2630). Nevertheless, these are not of such a 
 nature as to contradict the classification of Ar with A, but 
 merely indicate that in such cases, Ar best preserving the 
 original, independence has been asserted by the poet of A. 
 
 But in view of these and of A'a frequent abridgments, we 
 cannot look for the basis of Ar in A, nor as it is hardly 
 necessary to add, after the citation of textual agreements with 
 
THE SEVEN SAGES. 57 
 
 A in E or jB, and still less, for even more obvious reasons, 
 in Cor F. The marked agreement of Ar with A, however, 
 begets the assumption of a development of the former, parallel 
 with the latter, from a common source r, through which 
 they both go back to y. 
 
 Certain agreements of Ar with E against all other versions 
 including A (treated more at length under E) are not alto- 
 gether easy to reconcile, but owing to Ar's nearness to other 
 texts A in particular as against E, it is impossible to con- 
 sider Ar as derived from it ; we are led rather to the converse 
 assumption, of a partial connection, or contamination, of E 
 with Ar, or, in more likelihood, with the latter's immediate 
 source r. 
 
 That Ar so far as it goes, best preserves the lost M. E. 
 original is borne out on all sides: (1) by its close agreement 
 with the texts A and E, which otherwise best reproduce this 
 source ; (2) by the fact that F in the last four stories (in 
 which we should expect a close adherence to its original) is 
 closer to it than to any other text; and (3) that while A, 
 especially, and E, B, in a less degree, often add or omit lines, 
 Ar almost never adds, and in only rare cases abridges. 1 
 
 However, that no manuscript which has survived was based 
 on Ar follows from its occasional freedom, as e. g., (1) its 
 rimes to 171-2, 227-8, 463-4, etc., which are parallelled by 
 no other text, and (2) in Roma the names of Julius and July, 
 where all other texts better preserve the Genus (Janus) and 
 January of the French. 
 
 E. With the exception of Ar, the Egerton MS. would be 
 of most value in preparing a normalized text, since it next 
 best preserves the original, and especially since it is complete. 
 
 The value of E is considerably impaired, however, by the 
 fact that its author or more probably its scribe has made 
 an unusual number of textual abridgments, as a rule for 
 
 1 The only addition in the first 1900 11. is 1871-2 : 
 
 ' When day bygane to sprynge, 
 And )>e foules mery to synge.' 
 
58 KILLIS CAMPBELL. 
 
 single couplets only, yet in a few cases for a half-dozen or 
 more lines. Some of these are the following : (1) after 996 
 = A 991-2, (2) 1024= J. 1019-20, (3) 1216 = A 1211-2, 
 (4) 1400 =A 1385-6, (5) 1500= A 1465-6, (6) 1530 =A 
 1500-1, (7) 1558 =A 1529-30, (8) 1578 =A 1549-50, (9) 
 1646 =J. 1615-6, (10) 1652 =J. 1623-4, (11) 1662 =.4 
 1633-4, (12) 1784 =.4 1749-50, etc., and, most radical of 
 all, (13) after 2472 =,4 2424 f., where ten lines have been 
 lost. 1 In consequence of this, E is somewhat shorter than 
 either of the other complete texts, B and C. For the 2564 
 lines of the Arundel fragment, it has only 2365 ; and this 
 number in reality should be reduced 18 lines, since the couplets 
 with which E heads nine of its stories, and which have been 
 included in this numbering, did not belong to the original, it 
 is safe to assume, and should not, for purposes of comparison, 
 be regarded as part of the text. 
 
 But beyond these slight abridgments, the author of E has, 
 in the handling of his original, exhibited almost no independ- 
 ence. One looks in vain for such abridgments as characterize 
 A, as also for significant additions such as are found in F and 
 C. Excepting such occasional freedom as the assigning to 
 the incident in Roma the date of the first of January, and the 
 changing of the barber in tentamina into a borowe a scribal 
 error, doubtless we shall find scarcely one other feature ex- 
 clusively peculiar to E, until we have reached almost the end 
 of the poem, when the poet for once appears to assert his inde- 
 pendence, and we have in consequence the very interesting 
 addition that 
 
 ' whenne that his fadyr dede was, 
 
 He lete make a nobylle plas, 
 
 1 The additions are less numerous. Among those which are parallelled 
 by no more than one other text, or are peculiar to E, are (1) 986-7 (after 
 A 974), (2) 1015-6 (a. A 1012), (3) 1245-6 (a. A 1238), (4) 1621-2 = 4 
 1591-2, (5) 1693-6 (a. A 1664), (6) 1761-2 (a. A 1726), (7) 1809-10 (a. A 
 1780), (8) 2097-2103 (a. A 2068), (9) 2291-4 (a. A 2246), and (10) 2349- 
 51 (a. A 2298). 
 
THE SEVEN SAGES. 59 
 
 And a fayre abbeye he lete begynne, 
 And vu. schore raonkys brought thereyn, 
 And euyr more to rede and synge 
 For hys fadyr witA-owte lesynge.' (3561-6) 
 
 All other important variations in E are repeated in some 
 one or more of the related M. E. versions. The agreement 
 here is closest with B and Ar. Its near relation to the latter 
 has already been shown, and it has been pointed out that 
 there are features in which the two are alone ; and there are 
 also cases in which the two are alone in textual abridg- 
 ments : e.g. Ar 227-8 =E 11 71-2. It has also been seen 
 under Ar, that B in several instances falls in with E, Ar, as 
 against A, C, F. 
 
 It remains to point out some of the motives common to 
 E, B versus the remaining texts of Y. The most important 
 of these are the following: (1) arbor lords and ladies begin 
 to weep when they see the prince led forth to be hanged ; 
 (2) arbor Bancyllas assures the emperor that the prince 
 will recover his speech; (3) sapientes both omit the detail 
 of A, Ar, C that Merlin declines the offer of money made 
 by the man whose dream he has interpreted ; (4) vidua 
 the wife is cut in the thumb, where other texts have vari- 
 ously womb, finger, and hand ; as also (5) vidua the knight's 
 disregarding the widow's suggestion that he knock out her 
 husband's teeth ; (6) Roma the sage who makes the propo- 
 sition for saving Rome is called Junyus (A, C, F, Gemes ; Ar, 
 Julius; D, Gynevcr). In several of these, to wit 3, 4, 5, it will 
 be observed, E, B are truest to the French. 
 
 Such evidence as this precludes the thought of a basis of E 
 in Ar, but in view of the agreements between the two already 
 noted, and, especially, of the fact that there is a greater num- 
 ber of Ar-\ines than of 5-lines identical with E's (cf. p. 44), 
 it does not seem improbable though I am unable to prove it 
 that the author of E has known and been partly influenced by Ar. 
 
 On the other hand there is abundant evidence of an all but 
 immediate connection between B and E: (1) in the agree- 
 ments in details just cited, and (2) in the textual omissions 
 
60 KILLIS CAMPBELL. 
 
 and additions which the two have exclusively in common. 
 Thus, of the thirteen Remissions collected above, six (1, 7, 
 9, 10, 11, 12) are also in B; and of the ten additions cited in 
 the foot-note (p. 58), three (1, 8, 9) are common to B, or a 
 total of 9 out of 23 a remarkable showing when it is borne 
 in mind that in ten of these cases E is alone, agreeing in only 
 one case (abridgments 9) with any other text than B. 
 
 Despite these, however, E cannot have been based on B, 
 since it preserves in agreement with other texts notably 
 AT features of the original which B omits. 
 
 In the next section it will be shown, also, that B was not 
 based on E, and it will be further demonstrated that the two 
 are related through a common source. 
 
 B. The Balliol text, like E, is complete and of late com- 
 position. The analogy between the two does not stop here, 
 however ; there are many things which bind them together, 
 not only when looked at externally, but also from an interior 
 point of view. One of the most striking phenomena which 
 they have in common, and which one cannot but remark in 
 comparing them with Ar and the remaining F-texts, is the 
 tendency to reverse the order of words, or to substitute 
 synonymous or analogous expressions, in consequence of 
 which the identity of the line and often the rime is destroyed. 
 This is equally as prominent in B as in E t if not more so. 
 In B especially, the change of epithet often flows, one feels, 
 from a desire to modernize, rather than from a conscious 
 effort, as might be supposed, to conceal the source. 
 
 In some other respects, however, B and E are very unlike. 
 For instance, while it is characteristic of E to drop out one 
 or more couplets for every column, B is exceptionally free 
 from such slight curtailments, while its additional couplets 
 are comparatively numerous. 1 Moreover, while E is at first 
 
 1 In the first 1000 lines of the part selected for a line-for-line comparison 
 (= B 933-1951), B has 16 couplets which do not appear in any other 
 manuscript, and which were accordingly, in large part in all probability, 
 its own additions. E, on the contrary, has only 4, or one-fourth as nuiny 
 (1015-6, 1245-6 and 1693-6). 
 
THE SEVEN SAGES. 61 
 
 close to the original more so by far in the first thousand 
 lines than anywhere else and becomes more and more free, B 
 exhibits just the reverse tendency, and we find it in the last 
 third of the poem textually almost as close to the original 
 as is E. 
 
 As regards incident, B is usually more free than any 
 one of the texts so far treated. Its chief variations in the 
 nature of additions largely are the following : (1) aper 
 the herd fills both arms and sleeves (later laps) with the haws; 
 
 A, E, laps = L 23, girons; C, D, hood. (2) medicus the ille- 
 gitimate father of the sick prince, called in the remaining 
 members of Y either the earl or the king of Naverne (== L 27, 
 li quens de Namur) is not named. (3) puteus besides the 
 feature peculiar to Y y viz. that the burgess would only marry 
 some one from a distance, B adds that he also would marry 
 no poor woman, with the additional information that he 
 already had had two wives. The feature of A, E, Ar, that 
 he made a covenant with the bride's father, does not appear in 
 
 B. (4) senescalcus while in the remaining texts the steward 
 is banished, in B he is put to death and by pouring molten 
 silver and lead down his throat. This incident, which consti- 
 tutes the most violent freedom of -B, is apparently borrowed 
 from Virgilius, where Crassus dies a similar death. The 
 punishment in either case is fitted to the crime. (5) tenta- 
 mina the wife wishes to love the parish priest, where A, A r, 
 E, F y C have simply priest = L, provoire (but see jD* 27, 
 Messire Guillaume le chappelain de la parroise). (6) sapientes 
 they meet with the old man after two days; other texts not 
 definite as to time. (7) Roma the town is put in charge of 
 two wise men ; in other texts it is seven. (8) inclusa the 
 knight has travelled only one month before he comes into 
 the land of his lady ; according to other M. E. versions it is 
 three months (K, Z)*, A* 89, trois semaines; but cf. Varn- 
 hageu's Ital. Prosaversion, p. 36, ire mesi. (9) inclusa the 
 wife's ring had been given her as a New Year's gift, an 
 invention of B. 
 
62 KILLIS CAMPBELL. 
 
 But while B has thus many features peculiar to itself, it 
 possesses very few exclusively peculiar to itself and any one 
 other text, a circumstance which renders the problem of its 
 relations somewhat difficult of solution. We may resort, 
 however, to the verse-omissions or additions, and it is signifi- 
 cant here that the evidence from motive-comparison (submitted 
 already under E) which pointed to a relation with E, receives 
 very strong confirmation. In almost every instance in which 
 B agrees in an addition or omission with only one other text, 
 this text is E. Thus, in the first thousand lines of the con- 
 stant element in Y (= B 934 f.), there is a total of ten such 
 variations, of which nine are in agreement with E the tenth 
 being with (7, an agreement which can only be explained as a 
 coincidence or, at least, as signifying nothing. The agreements 
 with E, however, cannot well be accidental. They offer strong 
 confutation of the evidence of the line-collation (p. 44), which 
 seems to indicate a closer relationship with A or Ar. 
 
 That B was not based on either of the latter A, Ar 
 follows from the fact that it preserves certain features of the 
 original (cf. 3, 4, 5 of motive-agreements of E, J5, p. 59) which 
 they have either lost or altered. 
 
 And that both B and E go back to y independently of each 
 other is rendered improbable in the highest degree by their 
 agreements in omissions and additions. We are forced then to 
 the assumption of the existence at some time of a manuscript 
 denoted by s which served as the common source of B 
 and E. 
 
 F. There is no one of the M. E. texts of the Seven Sages 
 which has been more imperfectly reported than that contained 
 in the Cambridge University MS. Ff, II, 38. Wright as early 
 as 1845 was acquainted with this version, and printed in the 
 introduction (p. LXX) to his edition of D the opening lines, 
 but vouchsafed no further description of the text than that it 
 presented many different readings from A and was much 
 mutilated. And Petras, on the basis of this description, and 
 with the aid of about 190 lines of the text, has inclined to the 
 
THE SEVEN SAGES. 63 
 
 view that F is nearer to C than to any other M. E. version. 1 
 Neither Wright nor Petras, however, has made reference to 
 the description of Halliwell in his Thornton Romances (Cam- 
 den Society Publications, xxx, p. XLII f.), and both were 
 evidently ignorant of it. 
 
 The description of Halliwell is the most reliable which has 
 up to this time appeared ; yet in one or two instances it, too, 
 is inaccurate. For example, the thirteenth story of F has 
 been overlooked entirely; again it implies that there is only 
 one new story introduced into this version, the one which 
 he prints on p. XLIII f. In reality there is a second story in 
 F which is peculiar to it, the ninth story, to which Halliwell 
 gives the name The Squyer and his Borowe. This tale is 
 complete and runs as follows : 
 
 ' Hy t was a squyer of thys centre, 
 1115 And full welbelouyd was he. 
 
 Yn dedys of arrays and yn justyng [145 b.] 
 
 He bare hym beste yn hys begynnyng. 
 
 So hyt befelle he had a systur sone, 
 
 That for syluyr he had nome, 
 1 120 He was put yn preson strong, 
 
 And schulde be dampned, and be hong. 
 
 The squyer faste thedur can gon, 
 
 And askyd them swythe anon 
 
 What >yng he had borne a-way ; 
 1125 And they answeryd, and can say, 
 
 He had stolen syluyr grete plente ; 
 
 Therfore hangyd schulde he bee. 
 
 The squyer hym profurd, permafay, 
 
 To be hys borowe tyll a certen day, 
 1130 For to amende that he mysdede, 
 
 Anon they toke hym yn that stede, 
 
 And bounde hym faste fote and honde 
 
 And caste hym yn-to preson stronge. 
 
 They let hys cosyn go a-way 
 1135 To quyte hym be a certen day. 
 
 Grete pathes then used he, 
 
 And men he slewe grete plente. 
 
 Moche he stale and bare a-way, 
 
 And stroyed the centre nyght and day. 
 
 a See his dissertation, p. 31. Cf. also Varnhagen, in his review of Petras, 
 Englische Studien, X, p. 281 f. 
 
64 KILLIS CAMPBELL. 
 
 1140 Bot upon )>e squyer )>oght he nothyng 
 
 That he yn preson lefte lyeng, 
 
 So that tyme came as y yow say, 
 
 But for the squyer came no paye. 
 
 He was hanged on a galowe tree. 
 1145 For hym was dole and grete pyte, 
 
 When the noble squyer was slon, [145 c.] 
 
 For hym morned many oon. 
 
 That odur robbyd and stale moche }>yng, 
 
 And sethyn was hangyd at hys endyng. 
 1150 Thus schall be-tyde of J>e, syr Emperour, 
 
 And of thy sone, so gret of honour.' 
 
 Otherwise Halliwell's description is characterized by the 
 strictest accuracy, and leaves no room for the assumption, 
 apparently made by Petras, of an identity in the order of 
 stories between ^and the remaining M. E. versions. 
 
 The correct order of stories in F is as follows : (1) arbor, 
 (2) puteus, (3) aper, (4) tentamina, (5) gaza (end of), (6) vidua, 
 (7) Riotous Son (beginning of), (8) canis (end of), (9) Squyer 
 and Borowe, (10) avis, (11) sapientes, (1 2) medicus, (13) Roma, 
 (14) inclusa, and (15) vatidnium. Eight stories then (1, 3, 5, 
 10, 11, 13, 14, 15) retain their usual order. The two new 
 stories, 7 and 9, supplant senescalcus and Virgilius, taking their 
 respective order. For the remaining five stories, 2 changes 
 place with 8, 4 with 12, 6 with 2, 8 with 4, and 12 with 6. 
 For this order there is no parallel either in other English or 
 in foreign versions, and there can be little doubt that it was 
 original with the .F-redactor. 
 
 In content, also, F is very unique. In some cases the orig- 
 inal story has been altered almost beyond recognition. This 
 alteration consists largely in textual abridgments, but it is also 
 very evident in the many new incidents that have been intro- 
 duced. 
 
 The introduction, in contradistinction to the stories of the 
 first part, is but slightly abridged. It exhibits several more 
 or less interesting variations, but the only one of any signifi- 
 cance is the assigning to the king's steward the distinction 
 
THE SEVEN SAGES. 65 
 
 (accorded the king's retinue in the other texts) of making the 
 petition which saves the prince's life the first day. 
 
 1 Then come forthe the steward, 
 And seyde, st/r, thys was not forward, 
 When that y helde the thy londe, 
 When ii. kynges bade >e batell with wrong, 
 And then hou swere be heuen kyng 
 Thou schuldest neuer warne me myn askyng. 
 Geue me thy sones lyfe to-day, 
 Yentyll Emperour, y the pray, 
 And let hym to-morowe be at J?y wylle, 
 Whethur K>u wylt hym saue or spylle. 
 I graunt the, seyde the Emperour, 
 To geue hym lyfe be seynt sauyour.' (380-391) 
 
 Arbor is very much abridged, the story proper comprising 
 only twenty lines. There is no mention of the burgess's going 
 away from home, nor of the trimming away of the branches 
 of the old tree. 
 
 Of cam's only a short fragment is left, for which compare 
 Halliwell, Thornton Romances, p. XLIV. 
 
 Aper has to do with a "swynherde" who has lost a "boor," 
 
 and who 
 
 1 durste not go home to hys mete 
 
 For drede hys maystyrs wolde hym bete,' 
 
 but climbs a tree, and is making a repast of acorns when the 
 wild-boar of the forest comes up. 
 
 Medicus is one of the last four stories, hence agrees faith- 
 fully with its original. 
 
 Only the conclusion of gaza has been preserved. 
 
 Puteus has undergone radical alteration: (L) The curfew 
 of the original is omitted. Instead of it there is a law in 
 Rome that whosoever shall be found away from home at 
 night with any woman other than his wife shall be stoned 
 to death on the morrow. (2) The lover here is a " squire of 
 great renown." (3) The burgess uses a rope in trying to get 
 his wife from the well. (4) He has already had two wives 
 before his marriage with the one who figures here. This 
 5 
 
66 KILL1S CAMPBELL. 
 
 feature has been transplanted from the introduction to tenta- 
 mina, where it properly belongs. 
 
 Senescalcus and Virgilius do not appear in F. 
 
 Tentamina is characterized by the addition of a fourth trial, 
 the killing of the knight's hawk. Other features are (1) the 
 assigning to the wife the office of the gardener in the first 
 trial (she fells the tree, and sets " dokys and nettuls " in its 
 stead), (2) the omission of mention of the church as the meet- 
 ing-place of mother and daughter, and (3) the transference to 
 puteus of the ' two-wives '-feature. 
 
 Avis, though textually free, contains no unusual details 
 other than (1) that the lover is a priest, and (2) that the wife 
 is killed by the enraged husband. 
 
 In sapientes, however, there are several striking variations : 
 (1) The sages build a " horde-house" just above the city gate, 
 which renders the emperor blind whenever he tries to pass it 
 in going out of the city. (2) There is no mention of Merlin's 
 first dream-interpretation, a feature in which F agrees with 
 D, an agreement, however, which can only be accidental 
 since F contains the search for and meeting of the sages with 
 Merlin, which we find no hint of in D. 
 
 Vidua has the following peculiar features: (1) The husband 
 will never let his wife go a half-mile from him, " neither to church 
 nor to cheping." (2) The wife is paring a pear when she cuts 
 herself. (3) There is mention of only one thief, and he is not 
 alluded to as a knight. (4) A "pyke and spade" are used in 
 digging up the corpse. (5) In addition to the mutilations 
 usually recorded, F adds a fourth, the cutting off of two 
 fingers which the knight claimed that the thief had lost. 
 
 The last three stories, Roma, inclusa, and vaticinium, offer 
 essential agreement in detail with the other texts of Y. 
 
 The variations of -F are thus seen to be very numerous. 
 Yet, significant though many of them are, they tell only half 
 the story. The whole truth is revealed only when it is con- 
 sidered that along with these, and partly consequent upon 
 
THE SEVEN SAGES. 67 
 
 them, the length of the poem has been reduced by about 
 one-third, or to little more than 2500 lines. 
 
 And what is most noteworthy about this abridgment is 
 that it is not carried through the entire text, but extends only 
 through the eleventh story. Up to the conclusion of this 
 story the greatest freedom prevails, old incidents are rejected 
 and new ones introduced at will, and, again resorting to 
 figures for forcible illustration, the text is reduced from a 
 normal 2500 lines to scarcely more than 1000. 1 In the 
 remaining four stories, however, there is, as has been seen, 
 close agreement with the remaining texts of Y. 
 
 How to account for this wholesale mutilation to which F 
 has subjected its original is not an easy problem. One would 
 think of a basis for the first part in oral accounts, but this is 
 rendered extremely improbable by the fact that throughout 
 this part there is frequent agreement of rimes, and not unusual 
 identity of lines, with other M. E. versions. Or again, there 
 is a possibility that J^ was made from some very fragmentary 
 manuscript, but there is no substantial basis for this supposi- 
 tion, and the changed order of stories is distinctly against it. 
 The most probable view, by far, seems to be that the poet had 
 before him a complete manuscript, which, for some reason, 
 possibly to conceal his source, he has for the first eleven stories 
 arbitrarily altered ; and that beginning with the twelfth story, 
 having grown tired of his task, he has for the remaining stories 
 reproduced his original with fidelity. 
 
 With the acceptance of this explanation, the problem of F'a 
 relationship is rendered comparatively simple ; for, if the 
 variations of the first part are attributable to the poet, this 
 part is of little value for purposes of comparison, and we are 
 accordingly restricted to the last part as the basis for any 
 investigation. 
 
 For this part there is comparatively close textual agreement 
 with E, B, C, Ar, and A (the last two unfortunately frag- 
 mentary here in part). No single important detail and a very 
 
 'For the corresponding part, E has 2593 lines, and B, 2658. 
 
68 KILLIS CAMPBELL. 
 
 small percentage of the rimes have been changed, while lines 
 identical with one or more of the other texts are numerous. 
 The agreement is closest with Ar as a rule, with E next in 
 order; thus, for the 845 lines (F 1440-2285) which the three 
 texts have in common, only 53 lines of F are identical with 
 lines in E, while the corresponding figure for Ar is 116. 
 Again, for this section Ar has agreement with F in 26 couplets 
 which do not appear in ^(^1476-7, 1490-1, 1694-5 [, A], 
 1714-5, 1726-31, 1738-9, 1754-5, 1774-7, 1790-1, etc). 
 But despite this affinity with Ar, .F cannot have been based 
 on it, for in one case (.F 2280-1) Ar lacks a couplet which both 
 E and .Fhave preserved, and in other cases, it has made inde- 
 pendent additions (cf. Ar 1896-7, 2374-7, 2384-5). This 
 slight evidence is everywhere well supported : on the one 
 hand we find B, though much farther removed than E or Ar, 
 nearest F (cf. B 1095 = F 1578) ; again A will be found to 
 be, nearest (cf. A 997 = F 1464, A 1016 = F 1487, A 1048 
 = F 1518, A 1088-9 = F 1553-4) ; while in other instances 
 several will agree as against Ar (cf. A 2762 = B 2848 = F 
 1679, and A 2751 = E 2762 = B 2833 = F 1662). 
 
 In the face of this otherwise contradictory evidence, it is 
 impossible to find the source of ^ 7 in any one known manuscript. 
 At the same time there is nothing to indicate a partial basis 
 on any two of them, since some exclusive agreements with 
 each of the other closely related texts are found. On the 
 contrary, the evidence from all sides combines to show that F 
 goes back to y independently of any other known manuscript. 
 
 C. Petras, although he showed a close agreement of C 
 with A 52 lines identical and 1296 with similar rimes 
 classed it apart from A, and as only related with it through 
 a common O. F. source. 1 His own figures, however, as 
 Varnhagen has already pointed out, justify quite another 
 conclusion ; for it is inconceivable that two independent trans- 
 lations from a foreign source should have 52 out of about 
 2500 lines identical, or 1300 with like rimes. The rather are 
 
 'See his dissertation, p. 21. 
 
THE SEVEN SAGES. 69 
 
 we to conclude that C is ultimately based on the ultimate 
 common original of A, Ar, E, B y F, and belongs with them 
 to group Y. 
 
 Of all M. E. texts C is the fullest and, from a literary 
 point of view, the most perfect. At the same time it is, with 
 the exception of F, the freest of the texts which comprise Y. 
 This freedom, however, does not consist in the changed order 
 of stories nor the wholesale mutilation of text which charac- 
 terize F] nor is it violent or spasmodic. It flows from 
 an independence or individuality of a much higher type, which 
 neither eliminates old motives nor introduces new ones of a 
 startling nature, but which contents itself, on .the one hand, 
 with a slight variation of the episode (generally in the nature 
 of additions), on the other, with the enlargement and embellish- 
 ment of the often more or less lifeless language of its original, 
 in both cases with the purpose of heightening the poetic effect. 
 So that, while we see in A the most important of the M. E. 
 texts from an historical viewpoint, in Ar the most faithful 
 representative of the lost y, we have in C preeminently the 
 most perfect poem, holding, as it does, in language, style, and 
 metre, the first place in the early English group. 
 
 As regards fidelity to the original, as already suggested, C 
 does not occupy a very high rank. Its variations, however, 
 consist rather in amplification than in invention, as is well 
 illustrated by the fact that, while 600 additional lines have 
 been interwoven into the text, there are only the following 
 noteworthy variations of incident : (1) The step-mother in 
 bringing about the prince's downfall seeks counsel and assist- 
 ance from a witch (297). (2) In arbor, the tree with which 
 the story deals is a pineapple-tree ; A, E, B, F read pynnote- 
 tree, and D, apple-tree. (3) The queen in medicus states that 
 it has been twelve years since the Earl of Naverne had visited 
 her (1167) ; other texts indefinite. (4) The patient in the same 
 story is advised to " Ete beres fless and drink J;e bro " (1184). 
 A, Ar, E, B, " beef's flesh with the broth " (E, " with the 
 blood ") ; L 27, char de buef. (5) There is mention of only 
 
70 KILLIS CAMPBELL. 
 
 two clerks in gaza, where the remaining English and the French 
 texts have seven, five of whom are stationed away from the city 
 (1319). (6) In the same story the father alone goes into the 
 tower Cressent, while in the other texts both father and sou 
 go (1340). (7) In tentamina, the history of each of the two 
 deceased wives is related separately; in other texts it is simply 
 stated that the husband had survived two wives (1879). (8) 
 In the same story, also, it will be noted that only the right 
 arm of the wife is bled. (9) In Virgilius, the two brothers them- 
 selves fill the two "forcers" ; elsewhere the King has them filled. 
 Other variations here are the changed order of incident in 
 burying the treasure, and the omission of the name of the 
 Emperor (Crassus). (10) There is, in avis, no mention of a 
 maid as assisting the faithless wife. (11) The lord of the 
 castle in inclusa is playing chess when the knight rides up 
 (3294). (12) The son in vaticinium learns of the whereabouts 
 of his father through a vision (4135). 
 
 We may judge from this enumeration how faithfully C has 
 reproduced the subject-matter of the original. It has altered 
 very few details, and none radically, while no single significant 
 feature, either from the body or from the end of its stories, 
 has been omitted ; at the same time, only an occasional bit of 
 detail has been added, a remarkable showing, indeed, when 
 the large increase in the number of lines is considered. 
 
 But there is more specific evidence of C's fidelity to its 
 original. There are certain details in which it appears to give 
 a more faithful reflex of the Old French than any other M. E. 
 text. Thus, in aper, the boar on reaching the tree finds 
 " hawes ferly fone " (987) ; cf. L 23, " s'il se merveille mult 
 durement de ce qu'il ne pot autretant trover des alies comme 
 il soloit faire devant." According to other M. E. versions the 
 boar finds no haws at all. Another illustration may be had 
 from inclusa, where C (3264) preserves the Hongrie of the 
 French (A* 89) as the land into which the knight finally 
 comes in search of his lady ; M. E. variants are Pletys in Ar, 
 and Poyle in E, F, and D. 
 
THE SEVEN SAGES. 71 
 
 And there are also instances in which C is in agreement with 
 only one other text in its preservation of the French : (1) With 
 A in its rendering blanche leuriere (K 2604 ; L 45, only leu- 
 riere) by gray bitch, where Ar, E, B render greyhound, F 
 simply hound. (2) With F in giving, in Roma, the informa- 
 tion as to the origin of the word January at the beginning of 
 the Janus-episode ; other M. E. versions, where they preserve 
 this detail, depart from the O. F. order in placing it at the 
 conclusion of the story. 
 
 It is to these facts in the main that we have to resort to 
 determine C's immediate relations ; for the theory of a direct 
 translation from the O. F. can no longer be defended in the 
 face of the evidence from a comparison of rimes, etc. From 
 this comparison it is evident that C is nearly related to the 
 other versions of group Y. That it cannot have been based on 
 any one of them, however, follows from its agreements (just 
 cited) with the French where the remaining M. E. texts are 
 free. And this also derives confirmation from the features 
 which it has exclusively in common with only one M. E. ver- 
 sion and the O. F., for neither of the two M. E. versions in 
 point here (A and F) can possibly have been its original. 
 
 We have, accordingly, to assume for Can independent basis 
 in the lost text y. Whether one or more manuscripts inter- 
 vene between C and y cannot be determined so long as they 
 are not forthcoming ; in any case there seems nothing to sup- 
 port Varnhagen's proposition (Eng. Stud., x, p. 280) of a 
 " mundliche Ueberlieferungsstufe " between the two. 
 
 D. Version D, as compared with the texts so far con- 
 sidered, is unique, and cannot be classed with them in group 
 Y. Though it is written in the same metre as the remaining 
 M. E. versions, and while it preserves, also, the ^.-order of 
 stories, it differs from each and every text of Y much more 
 radically than any one of these differs from any other. And 
 so great has this difference seemed that scholars have been 
 unanimous in assuming for D an immediate basis in the Old 
 French. The thought of a near kinship with any other M. E. 
 
72 KILLIS CAMPBELL. 
 
 version appears never to have been entertained. Wright's 
 testimony is to the effect that "The two English metrical 
 versions (by which he meant A and D) are altogether different 
 compositions ; but .... were evidently translated from the 
 same original. . . ." l And the views of Petras (p. 44 f.) and 
 others are of like import. Scholars without exception seem 
 to have blindly accepted Wright's view, with no effort what- 
 ever to test its validity. 
 
 That Wright's assumption is unwarranted, however, may 
 be demonstrated, it is believed, beyond question. And it 
 will be the purpose of the following pages to make good this 
 assertion. With this end in view, we may first bring together 
 the chief variations in incident which D exhibits. 
 
 The introduction of D contains no significant alteration 
 of the original. A unique feature is the naming of the queen 
 Helie (variant Elye, 223) where the French is silent, but 
 where Fhas the name Milicent (or Ilacent). In not giving a 
 name to the prince it falls in with the French ; other M. E. 
 texts call him Florentine. There is a slight enlargement in 
 the account of the meeting of the father and son, in which 
 we have possibly a more faithful preservation of the French 
 than in Y. Other slight variations are the additional nature- 
 touch in having the queen ask to see the prince " In a myry 
 mornyng of May" (261), and the requiring the sages to 
 come to court within three days after the receipt of the royal 
 message (312). 
 
 Arbor preserves all the essential motives of the French. 
 A slight abridgment is the omission of mention of the knight's 
 going away for the sake of " chaffare " (A, E, B, C y L). 
 
 Canis, on the other hand, contains a number of interesting 
 variations : (1) The infant has only two nurses ; in A, E, B, C, 
 K y L, there are three, cf. L17, "Li enfes avoit -m. norrices." 
 (2) D also fails to catalogue the duties of the nurses, which is 
 otherwise a constant feature in both English and French (cf. 
 y, K, L 17). (3) A third curtailment is the complaint of the 
 
 1 See the preface to his edition of the D-text, Percy Soc., xvi, p. LXVIII. 
 
THE SEVEN SAGES. 73 
 
 knight against women when he finds his child alive. (4) A 
 very original addition is that the knight drowns himself for 
 sorrow in a fache-pole in his garden (883) ; L 21 and Y have 
 him go on a pilgrimage by way of atonement. 
 
 Aper exhibits comparative agreement with F, except in the 
 conclusion which has been much abridged. 
 
 The tale medicus is very much condensed. The ton-motif 
 is cancelled altogether (L 28 f., A 1142 f.), and there are 
 numerous less important omissions: e.g. (1) mention byname 
 of the Earl of Navern ( F, L 27, " li quens de Namur ") ; (2) 
 the cure of the invalid (F, "beef's flesh/ 7 etc. ; L 27, "char 
 de buef"); (3) specific allusion to the prince as an avetrol 
 (L 27, avoltres, so F, except F, C read horcopp). A single 
 addition is that the queen of Hungary is accompanied by ten 
 or twelve maids (1082). 
 
 Gaza. Omissions are (1) the names of both emperor and 
 tower (Octavian and Oressent, respectively, in J, Ar, E, j, C 9 
 L 30), and (2) the warden's finding the headless body, and his 
 endeavor to identify the same, a feature which is preserved 
 and worked out in detail in all other related versions (cf. L 
 32 f., A 1319-48). 
 
 Puteus. (1) No mention of the Roman law until late in 
 the narrative (1413 f.) ; in other versions it appears at the 
 beginning of the story ( F, L 36). (2) This law is not alluded 
 to at all as curfew (cf. L 36, coevrefeu). (3) The wife makes 
 no threat of drowning herself in the well ( F, L 37). (4) The 
 husband's excuse for being out thus late is that he thought he 
 heard a spangel, which he had " mysde al thys seven-nyght " 
 (1448-9). 
 
 Senescalcus. (1) Much abridgment of the scene between 
 the seneschal and his wife on the former's announcing his 
 infamous purpose. (2) Abridgment also of the early morning 
 scene, notably the dialogue between the king and his seneschal. 
 (3) An omitted detail is the bestowing the wife on a rich earl, 
 which is found in F, but which seems not to have been in 
 the Old French. 
 
74 KILLIS CAMPBELL. 
 
 Tentamina variations are (1) the wife herself contrives the 
 " tentamina." In all the related versions, they are proposed 
 by the mother. (2) A brother of the sage assists in the blood- 
 letting. Omissions are (1) mention of the sage's having sur- 
 vived two wives (cf. L 43 and all M. E. versions except F), 
 and (2) the wife's third visit to her mother, and the implied 
 r6le of the parish-priest of the original and the remaining 
 M. E. versions. 
 
 Virgilius. (1) A striking and altogether unwarranted alter- 
 ation is the substitution of Merlin for Vergil (1880). (2) 
 Allied with this is the very radical variation probably the 
 most radical of all in D in the omission of the entire first 
 episode, the incident of the mirror-pillars alone being preserved. 
 Other less striking variations are (3) the two coffers of gold 
 are buried, not as in the remaining M. E. versions, at the gates 
 of the city, but in " lyttyl pyttys twaye" (1926); (4) the 
 emperor is not asked to divide half with the brothers, nor does 
 he accompany the latter to their place of digging, but sends 
 one of his men with them (1932 f., 1950) ; (5) the brothers set 
 fire to the foundation of the pillar before going to their inn, 
 and even visit the emperor to bid farewell before taking final 
 leave of the city; (6) instead of pouring molten gold down the 
 emperor's throat, a ball of gold is ground to powder and his 
 eyes, nose, and throat are filled with it (2067-71). 
 
 Avis. Instead of the pie of other texts we have a popynjay 
 (2145), and (2) instead of the maid, a boy as the wife's assist- 
 ant. (3) Only the boy goes on the house-top. (4) He breaks 
 great blown bladders in imitation of thunder. (5) There is 
 no mention of the husband's discovery of the wife's deception. 
 
 Sapientes. Important omissions are the search for, and find- 
 ing of, the child Merlin and the incident, dependent thereon, 
 of the interpretation of the dream. 
 
 Vidua. (1) An interesting invention is the husband's burial 
 " withouten the toun at a chapel " (2484), since, in view of 
 the manner in which he met his death, "In kyrkejarde men 
 wolde hym nout delve " (2482) ; A* 80, simply au moustier. (2) 
 
THE SEVEN SAGES. 75 
 
 The wife herself kindles the fire and makes her bed beside the 
 grave (2502 f.), having first sent after her clothes (2500). (3) 
 The knight is permitted to enter immediately on knocking ; 
 in other texts, he has to repeat his knocking and petitions. 
 (4) The wife does not, as in other texts, propose matrimony 
 to the knight. 
 
 Roma. (1) There are three heathen kings instead of seven 
 as in the original (2649). (2) The page is not named till 
 towards the end of the story, when he is called Gynever (2730); 
 cf. A* 86, Genus; A, B, C, F, Gemes; E, B, Junyus; Ar. 
 Julius. 
 
 Inclusa. This story presents remarkable agreement with 
 F, the chief and only important variation being the temporary 
 omission of the knight's explanation of the reason for his flight 
 from his native laud in that he had slain there another knight. 
 This excuse is employed later in the story, but originates with 
 the lady (2961). 
 
 Vaticinium. (1) The father also has the power of inter- 
 preting the language of birds (3138). (2) The name of the 
 father is omitted (A* 101, 7T4919, Girart le fits Thierri; B, 
 C, Fj Jerrard Noryes sone; E, Barnarde Norysshe), and there 
 is otherwise much condensation of the narrative. 
 
 Such are some of the variations of D. And these are doubt- 
 less what led Wright to his classification of this version. But 
 since all these variations are peculiar to D they can in no way 
 be held to confirm Wright's view. They are in fact of no 
 value whatever in determining D's relations, except in so far 
 as they put one on guard against laying too much stress on 
 any agreements which D may be found to have exclusively in 
 common with any particular group or version. 
 
 Wright's theory, however, does seem to derive some sup- 
 port from another quarter, namely that Z), in a number of 
 instances, preserves the Old French more faithfully than any 
 other M. E. version. 1 These are as follows : (1) In senescalcus, 
 the king rules in Apulia (so L 39); in F, he rules over both 
 
 1 Wright, however, has not adduced any of this evidence. 
 
76 KILLIS CAMPBELL. 
 
 Apulia and Calabria. (2) In sapientes, after all the sages 
 have been slain and the cauldron has become clear, Merlin and 
 Herod ride out of the city by way of testing results ; the king, 
 on reaching the gate, regains his sight (D 2409 f., L 63). 
 Other M. E. texts omit this feature. A less significant agree- 
 ment of D with the Old French in the same story is that the 
 king remains blind from the time he goes outside the city 
 gates, where F represents him as being blind only when with- 
 out the city, and as always recovering his sight on his return. 
 (3) D 2803, A* 89 have the knight in indusa travel three 
 weeks in a fruitless search for the lady of his dream. Ar, E, 
 C, F have him travel three months, B, one month}- (4) In 
 vatidnium, the father and the son, at the beginning of the 
 story, are on their way to visit a hermit on an island in the 
 sea (3141 f.). This feature is suppressed in the remaining M. 
 E. versions, but appears in all the important O. F. versions ; 
 .A* 98, " por aler a -i- reclus qui estoit seur -i- rochier," and 
 K 4693-4, " Naiant en vont a un renclus, ki en un rochier ses- 
 toit mis." (5) In the same story (3327), the city to which 
 the father comes in his poverty, is, in agreement with A* 101, 
 Plede (cf. also JT4918, "Ales moi tost au plaseis," which 
 Godefroy identifies with plaisseis = cldture). The city is not 
 named in Y. 
 
 Of these agreements two (the 2d and 4th) are very signifi- 
 cant, and serve at least to show that D was not based on the 
 common original (y) of the six versions so far treated. They 
 do not prove, however, that D goes back to the French unre- 
 lated with these, for there still remains the possibility of a 
 connection of D with y through a common M. E. original (x), 
 which y does not for these features faithfully reproduce. Yet 
 it must be granted that this explanation would seem to 
 have little in its favor could not some agreements of D 
 with certain members of Fas against the French be shown. 
 
 1 The Italian prose text published by Varnhagen agrees here with the 
 M. E. versions ; see p. 36, tre mesi. 
 
THE SEVEN SAGES. 77 
 
 Among these agreements are : (1) with A and (7, in canis, 
 in that the knight cuts out the dog's rygge-boon (D 859) ; in 
 the French, he cuts off his head (L 20, " si li cope la teste ") ; 
 (2) in ape?-, with (7, in that the herd fills his hood with haws 
 (jD 945), A, E, B, L, his laps; (3) in Virgilius, with the entire 
 group Y, in that there are only two brothers who bring about 
 the overthrow -of the image (D 1899) ; L 51, on the contrary, 
 "in- bachelors"; (4) in vidua, with F, A* 84, in that the 
 wife is called on to knock out only two of her husband's teeth 
 (D 2592); according to A, Ar, E, B } C, all are knocked out; 
 see also D* 39, toutes les dens; (5) in indusa, (a) with the 
 entire group Y, in the substitution of Hungary for the Mon- 
 bergier of A* 89, K, as the land whence the knight comes (D 
 2787), (b) with E, F in the substitution of Poyle for the 
 illogical Hungary of the French (A* 89, K) as the land into 
 which the knight finally comes (D 2805), and (c) with Fm 
 the additional detail, that the earl had been warred against 
 for two years (D 2849). 
 
 But here it is possible that these agreements were accidental. 
 Furthermore, inasmuch as the ultimate O. F. original of the 
 M. E. versions has in all probability been lost, 1 it may be 
 argued that those features in which D and other M. E. 
 versions are in accord as contrasted with the Old French may 
 have been just those in which their common original varied 
 from the known O. F. manuscripts. Hence no final conclu- 
 sion may be had from this quarter. 
 
 There remains the evidence of phraseology and of rime, and 
 it is in this that we have a final proof of the error of Wright's 
 assumption. 
 
 The following are some of the parallel passages revealed by 
 a comparison of A and E with D. 2 Others might be cited, 
 but these will suffice for the purpose. 
 
 1 See the section devoted to a study of the source of the M. E. versions. 
 * Where A is fragmentary, E has been selected in preference to Ar, since 
 the latter is also largely fragmentary. 
 
78 KILLIS CAMPBELL. 
 
 D. E. 
 
 In Rome was an emperour, Sum tyme J>ere was an Emperoure, 
 
 A man of swyth mikil honur. That ladde hys lyfe with moche 
 
 Is name was Deocclicius. honowre. 
 
 Hys name was Dioclidan. 
 
 (1-2,4) (3-5) 
 
 Uppon his sone that was so bolde, The chylde wax to -vii- yere olde. 
 
 And was bot sevene wyntur olde. Wyse of speche ande dedys bolde. 
 
 (13-14) (15-16) 
 
 The emperour for-thoght sore Hys ffadyr was olde and ganne to 
 
 Tha the child ware sette to lore. hoore, 
 
 His sone thoo he sette to lore. 
 
 (15-16) (19-20) 
 
 Whilk of thaym he myght take To hem he thought his sone take 
 
 Hys sone a wyes man to make. Forto knowe the letters blacke. 
 
 (23-24) (23-24) 
 
 The thirde a lene man was. The -lii' mayster was a lyght man. 
 
 (49) (51) 
 
 And was callid Lentulus. His name was callyd lentyllous. 
 
 Hee sayed to the emperour thus. He sayde a-non to the kyng. 
 
 (51-2) (54-5) 
 
 And er ther passe thre and fyve, Uppon payne of lemys and lyfe, 
 
 Yf he have wyt and his on lyve, I shalle teche hym in yerys -v. 
 
 (55-6) (59-60) 
 
 And inred man he was, The -mi- mayster a redman was. 
 
 And was callid Maladas. Men hym callyd Malquydras. 
 
 (61-2) (61-62) 
 
 The sevent mayister answerd thus, The -vii- mayster hette Maxious, 
 
 And was hoten Marcius. A ryght wyse man and a vertuous. 
 
 (91-2) (99-100) 
 
 D. A. 
 
 Evermore wil he wooke, Whan o maister him let, another him 
 When on levede, anothir tooke. tok ; 
 
 He was ever upon his bok. 
 (159-60) (189-90) 
 
 By God, maister, I am noght dronken, Other ich am of wine dronke, 
 
 Yf the rofe his nougt sonken. Other the firmament is i-sonke. 
 
 (209-10) (211-2) 
 
 Hym byfel a harde caes. Ac sone hem fil a ferli cas. 
 
 (222) (222) 
 
 And to have anothir wyf, Ye libbeth an a lenge lif : 
 
 For to ledde with thy lif. Ye sholde take a gentil wif. 
 
 (231-2) (227-8) 
 
THE SEVEN SAGES. 
 
 79 
 
 A good childe and a faire, 
 That sal be oure bothe ayere. 
 For sothe, sire, I hold hym myn, 
 .4/30 wel as thou dost thyn. 
 
 (267-70) 
 
 Than sayd mayster Baucillas, 
 "For soth this his wondir cas: 
 Tharefore take counsel sone 
 What his best to don, 
 The childe answerd ther he stood, 
 " I wyle gyf sou counsel good ; 
 Seven dayes I mot forbere 
 That I ne gyf no answere ; 
 
 (360-3, 368-71) 
 I schal saue thy lyf a daye. 
 
 (381) 
 
 Thus they were at on alle, 
 And wenten agayen into the halle. 
 
 (388-9) 
 
 By hym that made sone and mone, 
 He ne hade nevere with me done. 
 
 (464-5) 
 
 " Kys me, yf thy wylle bee, 
 Alle my lyfe hys longe on the." 
 
 (474-5) 
 Callid to him a tormentour. 
 
 (509) 
 
 Also mote bytide the 
 As dyde the fyne appul-tre. 
 
 (582-3) 
 
 Than sayde Baucillas, 
 "A ! sire emperour, alas ! " 
 
 (688-9) 
 
 And hir clothes al to-rent, 
 Afte the thef wold hir have shent. 
 
 (700-1) 
 
 That knave kest hym frnyt y-nowe, 
 And clam a-doune fra bough to boghe. 
 
 (972-3) 
 
 And rent hys wombe with the knyf, 
 And bynam the bore hys lyf. 
 
 (982-3) 
 
 "A ! sire," quod mayster Ancilles, 
 " God almighty send us pees ! " 
 
 (1018-9) 
 
 Hit is thi sone, and thin air ; 
 A wis child, and a fair. 
 For thi sone I tel mine, 
 Alse wel als ton dost thine. 
 
 (283-4, 289-90) 
 Than seide master Bancillas 
 Here is now a ferli cas I 
 Counseil we al herupon; 
 How that we mai best don. 
 Than seide the schild, Saunz fail, 
 Ich you right wil counseil, 
 This seven daies I n'el nowt speke; 
 Nowt a word of mi mowht breke; 
 
 (371-8) 
 I schal the waranti o dai. 
 
 (389) 
 
 With this word, thai ben alle 
 Departed, and comen to halle. 
 
 (401-2) 
 
 I swere bi sonne and bi mone 
 With me ne hadde he never to done. 
 
 (451-2) 
 
 Kes me, leman, and loue me, 
 And I thi soget wil i-be. 
 
 (457-8) 
 And cleped forht a turmentour. 
 
 (498) 
 
 Ase wel mot hit like the 
 Als dede the pinnote tre. 
 
 (543-4) 
 
 Than seide maister Bancillas, 
 Sire, that were now a sori cas. 
 
 (683-4) 
 
 Th' emperour saide, I fond hire to- 
 rent: 
 Hire her, and hire face i-schent ; 
 
 (689-90) 
 
 He kest, the bor doun hawes anowe 
 And com himself doun bi a bowe. 
 
 (921-2) 
 
 The herd thous with his long knif 
 Biraft the bor of his lif. 
 
 (933-4) 
 
 Than saide maister Ancilles, 
 For Godes love, sire, hold thi pes. 
 
 (977-8) 
 
80 
 
 KILLIS CAMPBELL. 
 
 That 56 bytyde swilk a cas 
 
 As bytyde Ypocraa, 
 
 That slow hys cosyn withouten gylt. 
 
 (1026-8) 
 
 With my lordefor to play, 
 And love wax bytwen us twey. 
 
 (1100-1) 
 
 Oppon a day thay went to pleye, 
 He and hys cosyn thay twey. 
 
 (1118-9) 
 And mad hym myry, and spendid 
 
 faste, 
 
 Al the wylle that hit wolde laste. 
 He that lokyd the tresour, 
 Come a day into the tour. 
 
 (1220-3) 
 Bot hastilich smy t of my hede. 
 
 (1255) 
 
 Byfore the dore, as I sow telle, 
 Thare was a mykyl deppe welle. 
 
 (1381-2) 
 
 To do thy wyl by a-night, 
 Yf 1 schal helle the aryght. 
 
 (1546-7) 
 
 Now he slakys to lygge above ; 
 I wyl have another love. 
 
 (1686-7) 
 
 Er the myrrour be broght a-doune, 
 And than gyf us oure warrysoun. 
 
 (1906-7) 
 
 And sayed, we wyte, sire emperour, 
 About this cite gret tresour. 
 
 (1932-3) 
 And dolvyn a lytyl withinne the 
 
 grounde, 
 And the tresour was sone founde. 
 
 (1952-3) 
 
 The ton sayed, sire emperour, 
 Undir the pyler that berys merour. 
 
 (2002-3) 
 
 Gladlich, sayed scho, 
 The bettyr yf hyt wylle bee. 
 
 (2287-8) 
 And hadde seven clerkys wyse, 
 
 (2293) 
 
 On the falle swich a cas 
 Als fil on Ypocras the gode clerk, 
 That slow his neveu with fals werk. 
 (994-6) 
 
 With mi louerdfor to plai; 
 And so he dede, mani a dai. 
 
 (1083-4) 
 
 So bifel upon a dai 
 He and his neveu yede to plai. 
 
 (1113-4) 
 
 And beren hit horn wel on hast, 
 And maden hem large whiles hit 
 
 last. 
 
 Amorewe aros that sinatour, 
 And sichen to-bregen his louerdes 
 tour. 
 
 (1265-8) 
 And hastiliche gird of min heved. 
 
 (1299) 
 
 But thou me in lete, ich wille telle, 
 Ich wille me drenchen in the welle. 
 
 (1463-4) 
 
 Have womman to pleie aright, 
 Yif ye wil be hoi aplight. 
 
 (1577-8) 
 
 Ich moste have som other love ! 
 Nai, dowter, for God above ! 
 
 (1753-4) 
 
 Who might that ymage fel adoun, 
 He wolde him yif his warisoun. 
 
 (2029-30) 
 
 And said, al hail, sir emperour ! 
 It falleth to the to lof tresour. 
 
 (2049-50) 
 
 And ther thai doluen in the gronde; 
 A riche forcer ther thai founde. 
 
 (2079-80) 
 
 Than saide the elder to the emperour, 
 
 Under the ymage that halt the mirour. 
 
 (2091-2) 
 
 Bletheliche, sire, so mot ich the, 
 So that ye wolde the better be. 
 
 (2337-8) 
 He hadde with him seven wise. 
 
 (2343) 
 
THK SEVEN SAGES. 
 
 81 
 
 Who so anny swevene by nyght, 
 O morne when the day was bryght. 
 
 (2296-7) 
 
 The einperour and Merlyn anoon 
 Into the charabyr thay gonne gone; 
 (2339-40) 
 
 Hyt was a knyght, a riche schyreve, 
 That was lot hys wyf to greve. 
 He sate a daye by hys wyf, 
 And in hys honde helde a knyf. 
 
 (2471-4) 
 
 Bot saved for non worldlys wyne 
 Schulde no man parte horn a-twyne. 
 
 (2487-8) 
 
 In hyr hoond scho took a stoon, 
 And knockyd out twa teth anoon ! 
 
 (2601-2) 
 
 D. 
 
 Made to fle with hys boste 
 Thre kyngys and hare hoste. 
 
 (2732-3) 
 The knyght that met that sweven at 
 
 nyght 
 
 Of that lady was so bright, . . . 
 Kyght a lytyl fram the toure 
 Thare was the lady of honour, 
 And ate the wyndow the lady he see. 
 (2822-3, 2826-7, 2831) 
 He bytoke undyr hys hond, 
 And made hym stywarde of al hys londe. 
 Oppon a day he went to playe, 
 TJndir the tour he made hys waye. 
 
 (2869-72) 
 
 Lenand to the mykyl toure, 
 To do in hys tresour. 
 Thorow a q weyntyse he thout to wyne 
 The lady that was loke there-inne. 
 
 (2895-8) 
 
 That who that mette a sweven anight, 
 
 He scholde come amorewe, aplight. 
 
 (2349-50) 
 
 The emperour him ladde anon, 
 Into his chaumbre of lira and ston; 
 
 (2453-4) 
 
 Sire, he saide, thou might me leue, 
 Hit was a knight, a riche scherreue, 
 So, on a dai, him and his wif 
 Was i-youen a newe knif ; 
 
 (2563-4, 2569-70) 
 The leuedi saide, for no wenne, 
 Sche ne wolde neuer wende thenne. 
 
 (2581-2) 
 Than wil ich, she saide, and tok a 
 
 ston, 
 And smot hem out euerichon. 
 
 (2713-4) 
 
 E. 
 
 And made more noyse and boste 
 Thenne wolde a kyng and hys hoste. 
 (2812-3) 
 And soo there come rydyng thys 
 
 knyght 
 
 That had sought the lady bryghte. 
 He lokyd uppe into the toure, 
 And say that lady as white as flowre; 
 And anon, as he hyr say, 
 
 (2914-8) 
 And toke hym hys goodys in-to hys 
 
 hande, 
 And made hym sty war de ouyr alle hys 
 
 lande. 
 
 So oppon a day, with moche honoure, 
 The knyght come playnge by the 
 
 toure. 
 
 (2944-7) 
 
 To make a chambyr byfore the toure 
 That may ben for my honoure* 
 Thenne thought he uppon sum quent 
 
 gynne 
 
 Howe he myght to that lady wynne. 
 (2962-3, 2968-9) 
 
82 KILLIS CAMPBELL. 
 
 Oppon a day stylle as stoon The knyght toke workemen a-non, 
 
 He sent eftyr masons anoon. And made a chambyr of lyme and 
 
 8 ton. 
 
 (2901-2) (2966-7) 
 
 And sate stille and made hym glade, And bade hym ete and be glad, 
 
 And thus hys wyf made hym made. And euyr he sat as he were mad. 
 
 (3021-2) (3110-1) 
 
 Into Plecie when he was comen, Amorowe the kyng thedyr came, 
 
 Ner hysfadir hys in was nome. And with hysfadyr hys in he name. 
 
 To mete when he was redy to gon, He and hys baronys euerychone 
 
 After hys fadir he sent anoon. Wente to mete with hym a-non. 
 
 (3336-9) (3473-6) 
 
 It is impossible to account for these agreements as mere 
 coincidences, or as flowing from a translation from the same 
 O. F. source. Some of them may indeed be, and doubtless 
 are, due to the often stereotyped style, or the fondness for like 
 epithets or collocations which characterize the M. E. romance; 
 but all of them cannot be so explained. They warrant this 
 assumption alone, that D and y are related either through 
 the derivation of one from the other, or through a common 
 M. E. original. 
 
 And inasmuch as D cannot have been based on y or on any of 
 the texts which have developed from it, since in all the latter 
 some of the O. F. features are lacking which are preserved in 
 D, or, conversely, y on Z), in view of the very many inde- 
 pendent variations of the latter where y is faithful to the 
 French, we can only conclude that both y and D go back to 
 the same lost M. E. version x. 
 
 \Ve may accordingly sum up our results as to D as follows : 
 
 (1) it is remarkably free, and exhibits many unique variations; 
 
 (2) it does not represent an independent translation from the 
 French, but is connected with at least six other M. E. versions 
 through a common M. E. source; (3) this source was not the 
 same as the more immediate common original of these six 
 versions (y), but was a version one or more stages nearer the 
 Old French. 
 
THE SEVEN SAGES. 83 
 
 As. The Asloan version is at present inaccessible in the 
 original manuscript, 1 and, as only about 200 lines of it have 
 been printed, 2 any discussion of its relations must be very 
 unsatisfactory. We may be permitted, however, to bring to- 
 gether the few facts which are known about it, and to draw 
 from these such conclusions as their evidence may justify. 
 
 From the descriptions which have appeared, it is established 
 that As, so far as it is not fragmentary, preserves the usual 
 M. E. order of stories, but that beyond this it is, in many respects, 
 extremely free. The names of the sages are much garbled, 
 and they vary in the introductory enumeration from their 
 form in the stories themselves. They are, moreover, in no 
 case close to those of any version now in print, or to those of 
 the remaining M. E. manuscripts. 
 
 Avis, too, the story which has been printed, exhibits very 
 radical variation from other versions, both textually and as 
 regards incident. There are apparent no significant agree- 
 ments in rime or phraseology with any other M. E. version, 
 while two new episodes, 3 well-known in other collections, but 
 otherwise foreign to the Seven Sages, are woven into the narra- 
 tive. And there are other variations, besides, such as the intro- 
 duction of the wife's mother as a go-between, and mention of 
 the burgess's name first Annabill, later Balan. 
 
 But none of these serves to shed any light on the question 
 of relationship. All the new features of As, as compared with 
 the remaining M. E. versions and the accessible Romance ver- 
 sions, are peculiar to it, and hence afford no grounds for deter- 
 mining its connections. 
 
 1 As already stated in my " Word of Introduction " (p. 2), Lord Talbotde 
 Malahide declined to permit my consulting this manuscript. His reasons 
 for doing so are, I understand, the same as those given by certain other 
 possessors of valuable M. E. manuscripts, for which I beg to refer to Dr. 
 Furnivall, Temporary Pref. to the Six-Text Ed., Chaucer Soc., 1868, Pt. I, p. 6. 
 
 * In a contribution by Prof. Varnhagen (Englische Studien, xxv, p. 
 321 f.), who will edit the text for the Scottish Text Society. 
 
 3 See Englische Studien, xxv, p. 322. 
 
84 KILLIS CAMPBELL. 
 
 Prof. Varnhagen claims that As was made directly from 
 some O. F. version, 1 and the lack of textual agreement between 
 it and other M. E. versions in the story avis may seem to offer 
 some support to this view, but by no means necessarily, since 
 it is evident that the author of As worked very independently. 2 
 And that the evidence offered by Varnhagen in support of his 
 claim, viz., the agreement in order of stories with the O. F. 
 _4*-type, is not adequate, he himself, I believe, will concede 
 on reconsideration. 
 
 3. Authorship of the Middle English Versions. 
 
 It has been assumed in the preceding chapter that the Eng- 
 lish original (x) of the seven M. E. manuscripts A, Ar, E, B, F, 
 (7, and D, has been lost. It remains to inquire when, where, 
 and by whom this original was made. For this purpose we 
 unfortunately have almost no data at all, and can only resort 
 to indirections to find directions out. 
 
 (1) For the determining the date of x the Auchinleck MS. 
 (A) is of first importance. This manuscript dates from around 
 the year 1330; this, then, must be the superior limit for the 
 dating of y. And since, as has been shown, A was not derived 
 directly from y, but rests in all probability on a lost manuscript 
 r, which may have been based on y directly or through an inter- 
 vening manuscript, and since, moreover, it is highly credible 
 that A had already been composed some time before the Auch- 
 inleck copy was made, it is not probable that the date of y 
 would fall later than the beginning of the fourteenth century. 
 
 And inasmuch, now, as y cannot have been this parent version, 
 since D, though closely akin to it, was neither based immedi- 
 ately on it nor on any of its derivatives, but was connected with 
 it through a common source, which source we may assume to 
 be either identical with, or based directly on, the translation 
 
 * Ibid., xxv, p. 322. 
 
 * F offers even more radical variation from other M. E. versions in some 
 of its stories than does As in avis. 
 
THE SEVEN SAGES. 85 
 
 from the French, it is necessary to assign to this parent ver- 
 sion a date before the year 1300. The year 1275 would, it 
 is believed, represent a conservative conjecture. 
 
 (2) Available material for determining the place of transla- 
 tion of this parent text is somewhat more satisfactory. Of the 
 entire group of seven versions which have been shown to be 
 based on x, only one is in the Northern dialect, and this (C) 
 is of comparatively late date. One other (D) belongs to the 
 south-east Midland, while the rest (A, Ar, E, B, F) belong to 
 the South, a fact which well justifies the assumption that x 
 was also Southern. Furthermore, inasmuch as three of these 
 versions (A, Ar, E) possess marked Kentish features, and two 
 others ($, F) show a Kentish influence, but less marked, we 
 seem justified in a further restriction to the eastern South 
 Kent or its neighborhood as the home of the parent text. It 
 is further confirmatory of this view that just those versions 
 (Ar, E) which are most faithful to x are most distinctly 
 Kentish. 1 
 
 (3) But while we are thus justified in indulging in conject- 
 ure as to the time and place of composition of x, in the mat- 
 ter of its authorship we have no grounds for such an indulgence. 
 The nature of the subject might establish a slight probability 
 in favor of lay authorship, but not at all necessarily ; and the 
 same is true of the references to priests, in tentamina and avis, 
 as adulterate lovers, especially since in the only story in 
 which it is a constant feature (tentamina), it was also in the 
 Old French ; so that, in respect to this side of the problem 
 in hand, we have, for the present at least, and probably for all 
 time, to content us with absolute ignorance. 
 
 With regard to the authorship of the texts which have been 
 preserved, we are equally at a loss for definite information. 
 
 An ingenious and praiseworthy effort has been made by Dr. 
 Kolbing to demonstrate a community of authorship for the 
 ^L-text and the Auchinleck texts of the Arthur and Merlin, 
 
 1( rhe dialect of D southeast Midland also offers support to this view. 
 
86 KILLIS CAMPBELL. 
 
 Kyng Alisaunder, and Richard Coer de Lion ; l but without 
 meaning to discredit his conclusions in general, it is necessary, 
 we regret to say, to reject them in so far as they concern the 
 Seven Sages. Kolbing's argument is made on the basis of 
 features (rime, language, etc.) exclusively, or almost exclu- 
 sively, peculiar to these poems. The only part of his argu- 
 ment which holds is that which concerns the expletives cert 
 and vair. These appear only in the J.-text, being either orig- 
 inal with it, or, if in y, having been displaced in the remaining 
 texts by other rimes. On the other hand, of the 18 rimes 
 which Kolbing cites 2 (one of which, 2803-4, bataille: mer- 
 vaile, should be cancelled, since it is taken from C), a com- 
 parison with the remaining members of Fshows 12 to reappear 
 in the corresponding lines in Ar, 9 in E, etc. The evidence 
 to which Kolbing attaches most importance, that of certain 
 textual agreements between Arthur and Merlin (1201 f.) and 
 A (2389 f.), 3 is likewise not valid, as is manifest from the 
 following parallel comparison of these passages with Ar and 
 E. Compare 
 
 1 Merlin in J>e strete J>o pleyd, ' On a dai J>ai com ber Merlin pleid, 
 
 And on of his felawes him trayd.' And on of his felawes him traid.' 
 
 (A. M. 1201-2). (A 2389-90). 
 
 with 
 
 ' So >ei come |>eir J>e child played, 'Thenne come they thorowe happe 
 And on of his felawes hym by trayed.' there he playde, 
 
 One of his felowys hym myssayde.' 
 (Ar 1511-2). (E 2437-8). 
 
 Compare further, as against his citation of 
 
 'Foule schrewe fram ous go ! ' 'And cleped him schrewe faderles.' 
 
 ' pou hast yseyd to loude J>i roun.' 'Al to loude f>ou spok bi latin.' 
 
 pat haj> me sougt al Hs jer.' ' pat han me sought al fram Rcme.' 
 (A. M. 1204, 18, 20). (A 2392, 6, 8). 
 
 1 Arthur and Merlin, Leipzig, 1890, p. LX f. 
 
 f Ibid., p. LXXXII. 3 Ibid., p. civ. 
 
THE SEVEN SAGES. 87 
 
 the following from Ar and E: 
 
 'And clepyd hym schrewe faderlese.' 'And calde the chylde fadyrles.' 
 
 ' To loude Jx>u spake by latyn.' 
 
 1 pat haue me soujjt fro gret Rome.' ' That have Bought me fro Rome.' 
 
 (Ar 1514, 18, 20). (#2440,6). 
 
 From these it is evident that any inference as to A'a author- 
 ship made on this basis will apply equally as well to Ar and 
 E. Accordingly the parallels pointed out by Kolbing must 
 either be explained as accidental, or as traceable either to an 
 influence of Arthur and Merlin on the source of A, Ar, and E, 
 or, conversely, of some one of these on the Arthur and Merlin. 
 
 4. Source of the Middle English Versions. 
 
 The question of the ultimate source of the M. E. versions 
 has, to all intents and purposes, been settled by Petras. 1 We 
 need only present here his general argument and his conclu- 
 sion, inserting where deemed expedient additional proofs, and 
 adding here and there details which he has omitted. 
 
 But first of all it is necessary to state that such expressions 
 (which Petras [p. 32] inclines to accept as evidence) as A 2771, 
 ' So seigh ]?e rime' 2 (to which add .F1690, 'as seyj> J>e ryme') 
 proves nothing, for by a like reasoning we might, on the basis 
 of Ar 1906, ' as it saty in latyn/ prove a Latin source for the 
 M. E. versions. It is not on such formulae that the pre- 
 sumption in favor of a metrical original of the lost M. E. 
 original must repose; this must rather rest on the fact that 
 
 1 See his dissertation, p. 31 f. Our investigation must differ from his, 
 however, in that we are concerned only with the source of the parent ver- 
 sion, x (As being disregarded), while Petras has assumed each of four ver- 
 sions (A, C, jP, _D) to be independent translations from the French. Since, 
 however, he begins with the assumption that the same O. F. version was 
 the source of all these, his argument is essentially the same as ours. 
 
 2 References to source in the M. E. versions are numerous: A 317, 1245, 
 2766, 2770; Ar 1900, 1906, 2206, 2261, 2442; #1253, 2779, 2784, 3445; 
 295, 1235; I? 7 928, 1683, 1690, 1973; C 622, 1324; D 1385, 1520, 2690, 
 2922. 
 
88 KILLIS CAMPBELL. 
 
 this original (x) was itself in verse, and, hence probably made 
 from a metrical text, and that this does not permit of any 
 definite conclusion it is hardly necessary to add. 
 
 It is not improbable, however, that this original of x was, 
 like itself, composed of octosyllabic couplets, and it is needless 
 to state that it was in the French language. 
 
 There exist three O. F. metrical versions, the Dolopathos, 
 the Keller text (K), and the fragmentary version C*. The 
 first of these, the Dolopaihos, must, for obvious reasons, play 
 no part in this investigation. The unique version D* should, 
 however, since it represents a prosing of a lost metrical ver- 
 sion, receive equal attention with K and (7*. 1 
 
 The only one of this group which has ever been proposed 
 as a possible source of the M. E. versions is K; but a com- 
 parison of the two types as regards order of stories 2 reveals a 
 considerable difference between them, only ten stories (1, 2, 4, 
 6, 8, 10, 11, 12, 14, 15) having the same position in each. 
 Such a comparison, however, while bearing with it much 
 weight, can in no wise be accepted as determining, as it would 
 be quite natural for the redactor, or even the translator, to 
 change about the stories at will, either with artistic purpose or 
 with a view to making his source less apparent. Hence the 
 safest test of relationship should be from the consideration of 
 content, rather than of order of stories. And it is on this basis 
 that Petras's comparison has been made. The Cotton- Auchin- 
 leck (C-A), or Weber, text he finds to contain only 460 lines 
 which could be possible translations from the Keller text. 8 
 And since the latter contains over 5000 lines, it is not probable 
 that even numerous intermediate redactions could have made 
 such a difference. Besides this, there are many variations in 
 incident, all which unite in making it extremely improbable 
 that K was used by the English translator. 
 
 1 For the Dolopathos, K, C*, and D*, see the chapter on " The Romance 
 in France and Italy." 
 
 1 For the order of stories in the various sub-types of the Western group, 
 see our comparative table on page 35. 
 
 3 See p. 33 of his dissertation. 
 
THE SEVEN SAGES. 89 
 
 The fragmentary text (7*, though differing somewhat from 
 JTin order of stories, seems, nevertheless, to be much nearer 
 to it than it is to the English. 
 
 The prose version Z>*, representing a lost metrical version 
 F, exhibits still less agreement with the M. E. type, and 
 possesses many unique features. In the content of its stories, 
 however, it is comparatively close to K, so that in denying 
 the claims for it, the legitimacy of any claim for .D* is also 
 denied. 
 
 JT, C*, and D* having been eliminated from the problem, it 
 is necessary to conclude that the O. F. original, if metrical, 
 has been lost. It remains to show whether or not the M. E. 
 parent text was based on any of the prose texts which have 
 come down to us, or, at least, which one of them nearest 
 approximates the lost original. 
 
 The most widely known of the prose versions, the Historic*, 
 must be ruled out at once, since Paris has shown that the 
 earliest date which can be given it is around the year 1330, 
 or some time after the composition of the derivative M. E. 
 version A. Other circumstances, such as the order of stories, 
 the introduction of amatores, and the amicws-legeud, as well 
 as the fusion of Roma and senescalcus, together with its many 
 modern touches, all unite in invalidating any claim for H. 
 
 The Scala Coeli (S) also exhibits many features at variance 
 with the M. E. type, and its two new stories, filia and noverca, 
 are sufficient to exclude it from the list of possibilities. 
 
 Likewise the first Leroux de Lincy (L) version, although 
 it agrees very closely with the Middle English versions for 
 the first eleven stories, cannot be considered their source, 
 since it also contains the stories filia and noverca. 
 
 Nor to the Versio Italica does there attach any more proba- 
 bility, its distinguishing feature the reversal of the order 
 of stories finding no parallel even in French. 
 
 There remains group A*, or the family represented by the 
 second text of the Leroux de Lincy edition. A presumption 
 in favor of some member of this family is at once established 
 
90 KILLIS CAMPBELL. 
 
 in the fact that it has the same order of stories as the M. E. 
 group. This circumstance has led Paris and others to see in 
 this group the source of the M. E. texts, but no explicit claim 
 has been made as to which one of the J.*-manuscripts served 
 as this original, though Petras has made a detailed investiga- 
 tion with a view to arriving at some definite conclusion. 1 
 
 The results which Petras reaches, 2 however, are wholly 
 negative. He shows in the first place that MS. 6849 [new 
 No. 189] of the Bibliotheque Nationale, which Ellis had 
 suggested as the probable source of the M. E. versions, is not 
 even a possible source, but belongs to group L. He next 
 endeavors to show that the Leroux de Lincy text of A* (the 
 only one of the O. F. manuscripts of this type yet published) 
 is not as close to the M. E. versions as are some of the 
 unpublished manuscripts belonging to this family. Among 
 the latter, he finds the MS. 4096, Laval. 13, to be nearest 
 to the M. E. versions ; thus, by way of illustration, where L, 
 A*- call the seventh sage Herons, this manuscript names him 
 Meceneus, which approximates the M. E. Maxendus much 
 more closely. Despite this fact, however, he is not willing to 
 concede that this text was the source of the M. E. group, but 
 maintains that the latter had its basis in a lost manuscript 
 which is connected with the former through a common lost 
 source. 
 
 And in this conclusion Petras is probably correct, and 
 assuredly so as regards the Leroux de Lincy text, as is estab- 
 lished by certain features, which are not in J.*, but which the 
 M. E. texts have in common with JTand other O. F. versions. 
 A few of these are the following : (1) in tentamina, A, (7 read 
 gray bitch = K 2604, blanche leuriere; L (A* 45), only une 
 leuri&re ; (2) in Virgilius, L (A* 51) has lost the feature of 
 Vergil's casting images also for the east and west gates of 
 Rome, which has been preserved in K 3960 f. and the M. E. 
 group ; (3) in vaticinium, the child, when discovered alone on 
 the island, has had nothing to eat for four days in E, B, C y 
 
 Petras, p. 37 f. * Ibid., p. 44. 
 
THE SEVEN SAGES. 91 
 
 and JT4725; A* 99 and D*, only three days. These suffice 
 to indicate the result which would follow from a detailed 
 comparison. 
 
 In view of this conclusion, the problem of the source of the 
 M. E. parent text must, so far as a specific source is con- 
 cerned, remain for the present unsolved. Examination of all 
 ^4*-nianuscripts will doubtless bring us nearer to the truth, 
 and, it is hoped, settle the question. 
 
 II (6.) Sixteenth Century and Chap-book Versions. 
 
 Under this head fall the Wynkyn de Worde version and 
 the many chap-books founded on it, the lost Copland text, and 
 the Holland metrical version, all which fall together into one 
 distinct group apart from the M. E. group. 
 
 1. The Wynkyn de Worde text is in prose. Its date is not 
 definitely known ; in the British Museum catalogue it is 
 entered as 1520, though Hazlitt (Handbook, p. 660) gives it 
 a dating fifteen years earlier. Only one copy of the original 
 text has been preserved, and that is imperfect. A reprint 
 made by Gomme for the Villon Society (1885) makes the text 
 accessible. 1 
 
 This version seems to have been the first prose version made 
 in English, and, as already noted, it can in no way be related 
 with the M. E. metrical versions which antedate it. In length 
 alone the contrast is sufficiently striking to justify a serious 
 doubt as to any immediate relationship between them, the 
 prose version comprising 180 pages in Gomrne's edition. It 
 is based on some member of the Historia family probably a 
 Latin 2 rather than an O. F. text. As a translation of H it 
 
 l The History of the S. W. M. of Rome, London, 1885. A few pages missing 
 from the Wynkyn de Worde text are supplied from a chap-book version 
 printed in 1671. 
 
 ' 2 Graesse enumerates a half-dozen or more prints between 1483 and 1495, 
 any one of which may have served as the basis of this version. 
 
92 KILLIS CAMPBELL. 
 
 is comparatively close, though it abridges at times, and also 
 makes occasional independent additions. 1 
 
 2. The Wynkyn de Worde edition served as the basis of a 
 second prose edition, attributed to the printer Copland, which 
 has been lost. The superscription to this edition, which alone 
 has been preserved, agrees almost word for word with that 
 of the Wynkyn de Worde edition, and it is more than 
 probable, as Buchner suggests, 2 that it is only a reprint of it. 
 The date of the Copland text is variously placed between 1548 
 and 1561. 
 
 3. The Holland version is a very long poem written in 
 heroic couplets, and in the Scottish dialect. The original edi- 
 tion bears the date 1'578, but Laing has shown it to be probable 
 that its composition dates from the year 1560. It seems to 
 have been very popular in its day, undergoing at least five 
 editions (1590, 1592, 1599, 1606, 1620) in little more than 
 half a century after its first publication. A modern reprint 
 was edited by Laing for the Bannatyne Club in 1837. 
 
 Sundry conjectures as to the source which Holland employed 
 have been made. Laing maintained that he used either the 
 Copland print, or some O. F. or Latin text of H. Petras, 
 who did not know of the Wynkyn de Worde version, and who 
 makes the Holland version his " Redaction C," investigated 
 the question at some length, 3 and concluded in favor of the 
 O. F. translation of H as Holland's original. 4 But that 
 neither of these views is correct, and that the Holland text 
 was the rather based on the Wynkyn de Worde version, has 
 been conclusively proved by Buchner in his dissertation in the 
 Erlanger Beiirage, v, p. 93 f. This he established by show- 
 ing that where there are differences between the three versions 
 H (either Latin or French), the Wynkyn de Worde, and the 
 Holland the last two are in almost every instance in accord 
 
 1 See Buchner, Erlang. JBeitr., v, p. 95. 
 
 *Erlang. Beitr., v, p. 96. ' See his dissertation, p. 47 f. 
 
 4 The second text of Paris' s Deux Redactions. Its date is 1492. 
 
THE SEVEN SAGES. 93 
 
 with each other. A large number of textual parallels be- 
 tween the two English versions are cited in further support 
 of this. 
 
 (4) The English chap-book versions merit but little atten- 
 tion. They have been numerous, but of poor quality, the 
 later versions especially having deteriorated from the original. 
 In some of these, new stories have been introduced, and in 
 almost all of them the old stories have been abridged in 
 some of them, so as to be scarcely more than epitomes of their 
 prototypes. That they were very popular for a long time, 
 however, is indicated by the fact that the British Museum 
 alone contains at least twelve various prints, one of which 
 purports to have reached its twenty-fifth edition. Another 
 was published at Boston in 1794, the most recent at War- 
 rington in 1815. 
 
 All versions of the chap-book group contain the distinctive 
 features of H. They doubtless go back to the Wynkyn de 
 Worde, or to the Copland, text. 
 
 In addition to the four versions or groups already described, 
 there is evidence that there once existed another sixteenth cen- 
 tury version, which, like the Copland text, has not survived. 
 This is a dramatic version, bearing the title The Seven Wise 
 Masters of Rome, which is mentioned in Henslowe's Diary 1 as 
 having been made by Dekker, Chettle, Haughton, and Day, 
 and as having been acted at London in March, 1599-1600. 
 No later notice of its presentation has been pointed out, how- 
 ever, and it is altogether probable that the work was lost 
 without undergoing publication. 2 
 
 1 Ed. Collier, London, 1845, pp. 165, 167. See also the Dramatic Works 
 of Dekker, ed. Shepherd, London, 1873, I, p. xn. 
 
 1 The enumeration of the late English versions should also include refer- 
 ence at least to the Seven Wise Mistresses of Rome, a chap-book modelled 
 after the chap-book version of the Seven Wise Masters of Rome, and a sort of 
 counterpart to it. The English libraries contain several versions of this 
 type, but, though very interesting, they possess little value, 
 
94 KILLIS CAMPBELL. 
 
 APPENDIX. 
 
 [Containing the story medicus according to Ar (1-228), with a tabulation 
 of the corresponding lines in A t E, B y C y F.~\ 
 
 Hys comaundement )>ei dide be-lyve. 152a. 
 
 jjaue wex ]?ei? mochel stry ve 
 
 Be-tuen kynge and baron, 
 
 ffor J?e Emperowr wold scle his son, 
 5 J?e Empero^r hym nold save. 
 
 He lete a-none to spoile ]>at knaue, 
 
 And witA scourges hys body swynge ; 
 
 To foul dethe thei wold hym brynge. 
 
 A-none after that, god it wote, 1 
 10 He bade hem to hange hym fote hote. 
 
 With scourges J>ei dide hym swynge, 
 
 To foull de]?e J>ei wold hym brynge. 
 
 He was lade forj?e with-oute pite 
 
 ]>orouj-oute all ]?at fai? cite ; 
 15 J>ei? be-gan a rewfull cry 
 
 Of many gentyll lady. 
 
 All J?e folke oute of Rome 
 
 A-jeyne J>at gentyll child come. 
 
 Waleway, ]?ei saide, with wronge 
 20 Schall ]>is child nowe be honge. 
 
 Ryjt a-mydward J^at ilke pres 
 
 Come rydynge Maxilles, 
 
 And he sawe J>at rewfull cas ; 
 
 Hys second master forsoj>e he was 
 25 Hys scole? to helpe and to rede 
 
 All )>e folke to hym pei bede ; 
 
 A-none to court he gan ryde, 
 
 And wit/i ]>e Empe?'owr in reson chide 
 
 ffonde to let J;e Emperoitr wronge 
 30 );at his son be noujt an-hange. 
 
 1 This line is repeated after 1. 12, but is erased. 
 
THE SEVEN SAGES. 95 
 
 SwyJ>e fast fro }>e folke he rode, 
 
 His palfray a-none to pe paleys glode : 
 
 J>o come he by-fo? ]>e Empmwr, 
 
 And grete hym fai? with honour. 
 35 j>e emperour by hym styll stode, 
 
 And by-helde hym with steren mode 
 
 he saide to hym, " master, }>ou haue 
 
 J?e cors of god for techyng of J>is knaue. 
 
 je haue by-nome my sone his spech ; 
 40 ]>e devyll of hell I J?e be-tech, 
 
 Thyn felows and ]?ou be my swye? ! 152b. 
 
 je schull haue lytyll hye?." 
 
 " O Syr Emperow, knyjt of prys, 
 
 In dedes ]>ou schold be wa? and wyse. 
 45 It is no wysdome no lyuys hale 
 
 To by-leue no womans tale. 
 
 Mo? to harme ]?ane to note 
 
 A womans bolt is son schote. 
 
 ffor jef ]>ou sclest hym, I be-sech 
 50 On ]?i heued fall J;at ilke wrech 
 
 )?at fell on Ypocras, ]>e good clerk, 
 
 )?at sclewe his scole? J^orouj fals werk." 
 
 " Master, I pray f>e, tell j?at cas 
 
 Of J?at clerke Ypocras." 
 55 " Sy?, ]>is tale is noujt lyte ; 
 
 ffor jef );ou wyllt jef J?y son respyt, 
 
 A-for to-morowe day lyjt, 
 
 I wyll J;e tell a-none ryjt, 
 
 A-jenst J?e lawe, with grete wowe, 
 60 How Ypocras his nefew sclowe." 
 
 " I jeue hym respyt," said J>e Emp^-our, 
 
 And saide anone with-oute soiou?, 
 
 Mon schold a-jeyne feeche his son, 
 
 And put hym in-to preson. 
 65 J?e chyld was broujt oute of ]>e ton 
 
 With well grete procession. 
 
96 KILLIS CAMPBELL. 
 
 po he cam to pat hall, 
 
 He a-loutede pe barons all ; 
 
 And in to prison y-put he was. 
 70 Now tell we forpe of Ypocras. 
 
 y?," saide Maxillas, "paramour, 
 
 Ypocras was a clerke of grete honnour ; 
 
 Of lechcraft was none his pe? 
 
 Neuer jit in pis londe he?. 
 75 He hade wit A hym his nefewe 
 
 pat he schold leren of his vertue. 
 
 He savv pat child comyng of lo?, 
 
 pat he nold tech hym no mo? ; 
 
 ffor he poujt, and saide also, 
 80 pat he in lo? wold to-fo? hym go. 
 
 pe childe perseuyd full well, I-wis, 
 
 And hid it full wele in hert his. 
 
 His nefys herte he gan a-spye, 152c. 
 
 When he coupe all pe mastrye. 
 85 Ypocras gins understonde, 
 
 porouj werkes of pe childes honde, 
 
 pat he coupe all his mastrye. 
 
 He ba? to hym grete envye. 
 
 Sy by- fell apon a pynge, 
 90 Of hongre pat ilke kynge, 
 
 Hade seke a son gente ; 
 
 To Ypocras a rnessenge? sente, 
 
 pat he schold come his son to hele, 
 
 And haue he schold of gold full a male, 
 95 Ipocras wend ne myjt ; 
 
 He clepyd his nefewe anone ryjt, 
 
 And bade hym wende to pat londe, 
 
 To nyme pat chylde under honde ; 
 
 And whane he hade so lo, 
 100 He schold come ajeyne hym to. 
 
 pe child was set on a palfray, 
 
 And rode hym forpe on his way. 
 
THE SEVEN SAGES. 97 
 
 po he to pe kynge came 
 
 pe kynge hym by pe honde name, 
 105 And lade hym to pe seke childe. 
 
 Ihesus cryst to us be mylde ! 
 
 pat jonge man sawe pe childes payne, 
 
 He tastes his arrnes and his veyne; 
 
 He asked an urynall, as I wene, 
 110 And schewed pat uryn kenge and qwen, 
 
 Of pe childe all god it wyt, 
 
 And saide it was mys-by-get. 
 
 He gan pe qvvene on side drawe, 
 
 And saide, " dame, a-knawe, 
 115 What man hape by-gete pis childe?" 
 
 " Bel amy," scho sayde, " art pou wylde? 
 
 Who schold bot pe kynge?" 
 
 " Dame, say pou for no pynge, 
 
 He was neue? of kyngges streen." 
 120 " Lat," scho saide, " soch wordes ben ; 
 
 Or I schall do pe bete so, 
 
 pat pou schalt neuer ryde no? go." 
 
 " Dame," he saide, " witA soch tale, 
 
 py childe schall neue? be hale. 
 125 Tell me, dame, all pat cas, 
 
 How pe childe by-gete was." 
 
 " Bel amy, saist pou so ? " 
 
 "Sertes, dame," he saide, "no." 152d. 
 
 He schoke his hede upon pe qwene, 
 130 And saide, u pouj pou do me to-scleyne, 
 
 May I noujt do py childe bote, 
 
 Bot je me tell hede and rote, 
 
 Of what man he was be-geten." 
 
 " No man," scho saide, " may it weten ; 
 135 ifor jef n counseill we? un-hele, 
 
 I schold be sclowe with ryjt skyll." 
 
 " Dame," he saide, " so mot I the, 
 
 No man schall it wyt for me." 
 
98 KILLIS CAMPBELL. 
 
 " Syr/ 7 scho saide, " it so by-fell, 
 140 ]>ls oj>er day in Auerell, 
 
 J>e kynge of nauerne come to ]>is }>ede, 
 
 On fai? hors and in rich wede, 
 
 With my lord for to play, 
 
 And so he dide many a day. 
 145 I gan hym son in herte to lone, 
 
 Ouer all J?ynge so god aboue ; 
 
 So J>at for grete drewrye, 
 
 I late J;e kynge be me lye ; 
 
 So it was on me by-gete : 
 150 Sy?, late no man ]>at i-wete." 
 
 " Nay, madame, for so}>e, i-wys, 
 
 Bot for )?at childe was gete a-mys, 
 
 He mot both drynke and ete 
 
 Contrarious drynke and contrarious mete, 
 155 ffresch beef and drynke ]>e bro)?e." 
 
 He jaf a-none ]>Q child forso^e. 
 
 J?e childe was heled fai? and' \\e\e. 
 
 )?e kynge hym jaf many Jewell, 
 
 A wer hors i-charged \\iih s\\uer and gold, 
 160 Als moch as he nyme wold. 
 
 He dide hym forpe a-none ry^t, 
 
 And come home in ]n\t nyjt. 
 
 |;e master hym asked jef he we? sond 
 
 " ja si?/' he saide, " be seynt Symond ! " 
 165 ]>o asked he, " what was his rnedecyne?" 
 
 He saide, u fresch beef good and fyne " 
 
 "]?an was he a nauetroll." 
 
 "]?ou saist sol?e, be my poll !" 
 
 " O," (\uod Ypocras, " be goddes dome ! 
 170 )?ou art by -come a good grome." 
 
 )?o by-gan Ypocras to pencil 
 
 To scle his nd'cwc with some wrench. 
 
 ];ei?-aftc?, |>t- ]'ridc day. 153a. 
 
 W/tA his nt- few he went to play, 
 
 
THE SEVEN SAGES. 99 
 
 175 Yn-to a fai? grene gardvn ; 
 
 fei? wcx many an erbe 1'yii. 
 
 fe childe sawc an erl>(> on fo gffOUDde, 
 
 fat was iny^ty of mochrl! nionde ; 
 
 He toke it and srhcwcd to Ypocras, 
 180 Bot he saidc a better fei? was; 
 
 For he wold fat child be-cach. 
 
 He stoupyd soch on to rech. 
 
 fo fyle Ypocras with a knyf, 
 
 He nomc his ncfewe of his lyf. 
 185 He dide hym bury unkonnynglych, 
 
 As he had dyed sodeynlych, 
 
 And afte?-warde, swyf e jerne, 
 
 He dide his bokes all to-bryne. 
 
 God of heuen, ]>e hyje kynge, 
 190 fat is oue?-sea? of all fynge, 
 
 Sende Ypocras for his treson, 
 
 fe foul rankkeland menyson. 
 
 Ypocras wyst wele, for his quede, 
 
 fat he schold son be dede ; 
 195 Bot for no Jwige fat he coufe fynch 
 
 fe meuyson he no mygt quench. 
 
 A nerapty ton he dide forfe fett, 
 
 And full of olene water he it pyt, 
 
 Also full to f e mouf e ; 
 200 ffor he wold it we? couf e, 
 
 And dide after sende rnochell and lyte, 
 
 Nejbowrs hyni to bysyte. 
 
 He saide to-fore hem euerchon 
 
 fat f e def was hym apon, 
 205 All with ryjt and noujt with wouje, 
 
 ffor his nefewe fat he sclowge. 
 
 fat treson he gan hym reherce. 
 
 On f e tone a C. holes he gan perce. 
 
 When f e holes we? mad so fell, 
 210 He dide hem stope with dosell, 
 
100 KILLIS CAMPBELL. 
 
 And saide to hem once or tweye, 
 
 " je schall see of my mastrye." 
 
 He smered J?e dosells all a-boute, 
 
 And made he me after- ward drawen oute. 
 215 A dro];e J?ei?-of oute ne came; 
 
 )>a?-of merveiled many man. 
 
 Ypocras saide, " water y can stope, 
 
 J>at it ne may une];es drope; 153b. 
 
 But y ne may stope my menyson. 
 220 All it is for J?at foul treson, 
 
 }>at y my nefewe sclewe vylengly, 
 
 ifor he was wyse? man ];ane y. 
 
 I no? no man unde? eon 
 
 geue me helpe ne can, 
 225 Bot my nefewe o-lyue we?. 
 
 Ryjt it is J>at y mys-fai?. 
 
 To soifre wo it is skyll 
 
 ffor y sclouj my lyuys hele." 
 
 TABLE OF CORRESPONDING LINES/ 
 
 AT A E B C F 
 
 949 933 1041 
 
 950 934 1042 
 
 951 1043 
 
 952 1044 
 
 5 953 (1045) 
 
 954 (1046) 
 
 955 (1047) 
 
 956 (1048) 
 
 (957) (1436) 
 
 10 (958) (1438) 
 
 J An identical line is indicated by an asterisk (*), an omission by a dash 
 
 ( ), an addition by brackets ([]), a corresponding but not similar line 
 
 by leaders ( ), and altered rimes by parentheses (). 
 
THE SEVEN SAGES. 101 
 
 AT A E B G F 
 960 (1052) 
 
 942 1054 
 
 1057 
 
 964 944 1058 
 
 965 945 
 
 966 946 
 
 947 
 
 20 968 948 
 
 963 969 949 1059 
 
 964 970 950 1060 
 
 966 972 952 1062 1440 
 
 965 971 951 1061 1441 
 25 974 
 
 973 
 
 (967) 976 957* (1065) 
 
 (968) 975 958 
 
 977 959 
 
 30 978 960 
 
 979 (961) 
 
 969 981 963 1067 (1442) 
 
 970 982 964 1068 (1443) 
 35 (971) 983 965 (1069) (1444) 
 
 (972) 984 966 (1070) (1445) 
 
 (973) 985 967 1071 (1446) 
 
 (974) 986 968 1072 (1447) 
 [87-88] [69-70] [48-49] 
 
 (975) 989* 971 " 1450 
 40 990 972 1451* 
 
 991 (973) (1073) (1452) 
 
 (976) 992 (974) (1074) (1453) 
 
 (977) 993 975 1454* 
 
 (978) 994 976 1455* 
 [79-88] 
 
102 
 
 KILLIS CAMPBELL. 
 
 AT 
 
 A 
 
 E 
 
 B 
 
 C 
 
 F 
 
 45 
 
 989 
 
 995 
 
 (977) 
 
 
 
 1456 
 
 
 990 
 
 996 
 
 (Ok f 7Q\ 
 
 
 14^7 
 
 
 v 7 V/ 
 
 QQ9 
 
 C7 U \J 
 
 * ' 
 
 
 J.TCU ( 
 
 
 t/t/ft 
 
 QQ1 
 
 
 
 
 
 
 J77 
 
 /OQO^ 
 
 
 Q7Q 
 
 
 
 50 
 
 (994) 
 
 
 
 V I *J 
 
 980 
 
 (1085) 
 
 1461 
 
 
 995 
 
 
 
 981 
 
 (1086) 
 
 1462 
 
 
 996 
 
 
 QS9 
 
 ( \ o&7^ 
 
 1 4fi*-i 
 
 
 *7 t7 \J 
 
 997 
 
 997 
 
 f/O -> 
 
 983 
 
 (1088) 
 
 1464 
 
 
 998 
 
 998 
 
 984 
 
 
 1465 
 
 55 
 
 999 
 
 1001 
 
 985 
 
 
 (1466) 
 
 
 
 
 
 
 v / 
 
 
 1000 
 
 1002 
 
 986 
 
 (1091) 
 
 (1467) 
 
 
 1001 
 
 1003 
 
 (987) 
 
 
 1468 
 
 
 
 
 V / 
 
 
 
 
 1002 
 
 1004 
 
 /Q00\ 
 
 
 1 4fiQ 
 
 
 J. \J \J+J 
 
 lOO'l 
 
 J. \J \J A 
 
 100^ 
 
 ^ ' 
 
 
 1470 
 
 60 
 
 -I V/V/O 
 
 1004* 
 
 1 \J\JO 
 
 1006 
 
 
 
 (1093) 
 
 J.TC I \J 
 
 1471 
 
 
 1005 
 
 1007 
 
 989 
 
 (1094) 
 
 1472* 
 
 
 1006 
 
 1008 
 
 990 
 
 
 
 1473 
 
 
 1007* 
 
 1009 
 
 (991) 
 
 1095 
 
 1474 
 
 
 1008 
 
 1010 
 
 (992) 
 
 1096 
 
 1475 
 
 65 
 
 1009 
 
 _____ 
 
 993 
 
 
 1476 
 
 
 1010 
 
 
 994 
 
 
 1477 
 
 
 -i \y JL vr 
 
 1011 
 
 c/ u rt 
 
 995 
 
 
 1478 
 
 
 
 1012 
 
 996* 
 
 
 1479 
 
 
 1011 
 
 1013 
 
 997 
 
 
 
 1480* 
 
 70 
 
 1019 
 
 1014 
 
 QQK 
 
 
 1481 
 
 i \j 
 
 j. \j i & 
 
 -I \J 1 Tt 
 
 1017 
 
 7 t/O 
 
 999 
 
 1101 
 
 J- TtO i 
 
 1482 
 
 
 
 
 1018 
 
 1000* 
 
 1102 
 
 1483 
 
 
 1014* 
 
 1019* 
 
 1001* 
 
 (1103) 
 
 1484* 
 
 
 1013 
 
 1020 
 
 1002 
 
 (1104) 
 
 1485 
 
 75 
 
 1015* 
 
 (1021) 
 
 1003* 
 
 1105 
 
 1486 
 
 
 1016 
 
 (1022) 
 
 1004 
 
 1106 
 
 1487 
 
 
 
 
 
 [1107-8] 
 
 
 
 1017 
 
 1023 
 
 1005 
 
 1109 
 
 1488 
 
 
 1018 
 
 1024 
 
 1006* 
 
 1110 
 
 1489* 
 
TIIK SEVEN SAGES. 103 
 
 Ar A E B C F 
 
 (1019) (1007) (1111) (1490) 
 
 80 (1020) (1008) (1112) (1491) 
 
 1021 1025 1009 (1113) 1492 
 
 102-2 1026 1010 (1114) 1493 
 
 1023 1027 1011 1116 1494 
 
 1024 1028 1012 1115 1495 
 85 1025 1029 1013 1496 
 
 1026 1030 1014* 1497 
 
 1027 1031 1015* 1498* 
 
 1028 1032 1016 1499 
 
 1029 1033 1017 1117 1500 
 90 1030 1034 1018 1118 1501 
 
 1031 1035 1019 (1H9) 1502* 
 
 1032 1036 1020 (1120) 1503 
 
 1033 (1037) 1021* 1121 1504* 
 
 1034 (1038) 1022 1122 1505 
 95 1035 1039 (1023) (1123) 1506 
 
 1036 1040 (1024) (1124) . 1507 
 
 1037* 1041 1025 1125 1508* 
 
 1038 1042 1026 1126 1509 
 
 1039* 1043* 1027* 1510* 
 
 100 1040 1044 1028 1511 
 
 1041* 1045 1029* 1129 1512* 
 
 (1042) 1046 1030 1130 1513 
 
 (1043) 1047 1031 (1131) 1514 
 1044 1048* 1032 (1132) 1515* 
 
 105 1045 1049* 1033* 1133 1516* 
 
 1046 1050 1034 1134 1517 
 
 1047 1051 (1035) 1135 1519 
 
 1048 1052 (1036) 1136 1518 
 
 1049 1053 1037 (1137) 1520 
 110 1050 1054 1038 (1138) 1521 
 
 1051 1055 1039 (1522) 
 
 1052 1056 1040 (1523) 
 
 [41-42] 
 
104 KILLIS CAMPBELL. 
 
 Ar A E B OF 
 
 1053 1057 (1041) (1143) (1524) 
 
 1054 1058 (1042) (1144) (1525) 
 115 1055* 1059 1043 (1145) (1526) 
 
 1056 1060 1044 (1146) (1527) 
 
 1057 1061 1045 1147 1528 
 
 1058 1062 1046 1148 1529 
 1059* 1063 1047 1149 1530 
 
 120 1060* 1064 1048 1150 1531 
 
 1061* 1065 1049 1151 1532 
 
 1062* 1066 1050 1152* 1533* 
 
 1063 (1067) 1051 (1153) 
 
 1064 (1068) 1052 (1154) 
 125 1065 1069 1053* 1155 
 
 1066* 1070* 1054* 1156 
 
 1067 . 1071 1157 
 
 1068 1072* 1158 
 
 1069 1073* 1534 
 130 .1070 1074 1535 
 
 1071 1075 1055 1159 1536 
 
 1072 1076 1056 1160 1537 
 1073* 1077 1057* (1161) 1538* 
 1074 1078 1058 (1162) 1539 
 
 [59-60] 
 
 135 1075 (1079) (1061) (1163) (1540) 
 
 1076 (1080) (1062) (1164) (1541) 
 
 1077* 1081 1063 1165 (1542) 
 
 1078 1082 1064 1166 (1543) 
 
 1079 1083 1065 1167 1544 
 140 1080 1084 1066 1168 1545 
 
 1081 1085 1067 1169 (1546) 
 
 1082 1086 1068 1170 (1547) 
 1083* 1087 1069* 1171 1548* 
 1084* 1088 1070 1172 1549 
 
 145 1085 1089 1071 1173 1550 
 
 1086 1090 1072 1174 1551 
 
THE SEVEN SAGES. 105 
 
 Ar A E B C F 
 
 1087 1091 1073 1175 1552 
 
 1088 1092 1074 1176 1553 
 
 1089 1093 1075 1177 1554 
 150 1090 1094* 1076 1178 1555 
 
 [56-59] 
 
 1091 1095 1077 (1179) " 1560 
 
 1092 1096 1078 (1180) 1561 
 1093* 1097 1079* 1181 1562* 
 1094 1098* 1080 1182 1563* 
 
 155 1095 (1099) 1081 1183 1564 
 
 1096 (1100) 1082 1184 1565 
 
 [85-90] 
 
 1097 1101 1083 1191 1566 
 
 1098 1102 1084 1192 1567 
 
 1099 1103 1085 1193 1568 
 160 1100* 1104 1086 1194 1569 
 
 1101 (1105) (1087) (1195) 1570 
 
 1102 (1106) (1088) (1196) 1571 
 
 1103 1107 1089 (1197) (1572) 
 1104* 1108* 1090 (1198) (1573) 
 
 165 1105 1109 1091 1199 1574 
 
 1106 1110 1092 1200 1575 
 
 1107 1111* 1093* (1201) (1576) 
 
 1108 1112 1094 (1202) (1577) 
 
 1109 1113* 1095 1203 1578 
 170 1110 1114 1096 1204 1579 
 
 (1111) (1115) (1097) (1205) (1580) 
 
 (1112) (1116) (1098) (1206) (1581) 
 
 1113 1117 1099 1'207 1582 
 
 1114 1118 1100* 1208 1583 
 175 1115 1119* 1101 (1209) 1584* 
 
 1116 1120 1102 (1210) 1585 
 
 1117 1121 1103 (1211) (1586) 
 
 1118 1122 1104 (1212) (1587) 
 
 1119 1123 1105 1213 1588 
 
106 KILLIS CAMPBELL. 
 
 Ar A E B C F 
 
 180 1120 1124 1106 1214 1589 
 
 1121 1125 1107 (1215) (1590) 
 
 1122 1126 1108 (1216) (1591) 
 
 [17-20] [92-93] 
 
 1123 1127 1109 (1221) " 1594 
 
 1124 1128 1110 (1222) 1595 
 185 1125 1129 1111 1223 1596 
 
 1126 1130 1112 1224 1597 
 
 1127 (1131) (1598) 
 
 1128 (1132) (1599) 
 
 1129 1133 1113 1225 1600 
 190 1130 1134 1114 1226 1601 
 
 1131 1135* 1115* 1227 1602* 
 
 1132 1136 1116 1228 1603 
 
 1133 1137 1117 (1229) 
 1134* 1138* 1118 (1230) 
 
 195 1135 1139 1119 1231 
 
 1136 1140* 1120 1232 
 [cf. 1142] [33-34] 
 
 1143 (1141) 1121 " 1235 1604 
 
 1144 (1142) 1122 1236 1605 
 
 1145 1143 1123 1606 
 200 1146 1144 1124 1607 
 
 1137 1145 (1125) (1237) 1608 
 
 1138 1146 (1126) (1238) 1609 
 
 1139 1147* 1127 1239 1610 
 
 1140 1148 1128 1240 1611 
 205 1141 1149 (1129) (1241) (1612) 
 
 1142* 1150 (1130) (1242) (1613) 
 [cf. 1136] 
 
 1147 1151 1131 1614 
 
 1148 1152 1132 1615 
 
 1149 1153 1133 1243 1616 
 210 1150 1154 1134 1244 1617 
 
 1155* 1135* 
 
THE SEVEN SAGES. 
 
 107 
 
 AT 
 
 215 
 
 220 
 
 225 
 
 A 
 
 E 
 
 1156 
 1157 
 
 B 
 1136 
 1137* 
 
 C 
 1245 
 
 F 
 
 1618* 
 
 1151 
 
 1152 
 
 1158 
 
 1138 
 
 1246 
 
 1619 
 
 1153 
 
 1159* 
 
 1139 
 
 1247 
 
 1620 
 
 1154 
 
 1160 
 
 1140* 
 
 1248 
 
 1621 
 
 1155 
 1156 
 1157 
 
 1161 
 1162 
 1163 
 
 1141 
 1142 
 1143 
 
 (1249) 
 (1250) 
 1251 
 
 1622 
 1623 
 1624* 
 
 1158 
 
 1164 
 
 1144 
 
 1252 
 
 1625 
 
 [59-60] 
 1161 
 
 1165 
 
 1145 
 
 [53-54] 
 1255 
 
 1626 
 
 1162* 
 
 1166 
 
 1146 
 
 1256 
 
 1627 
 
 1163* 
 1164 
 1165* 
 
 1167 
 1168 
 1169 
 
 1147 
 1148 
 1149* 
 
 
 
 1628 
 1629 
 1630 
 
 1257 
 
 1166 
 
 1170 
 ( \ 1 79^ 
 
 1150 
 
 1258 
 
 1631 
 
 This partial table will serve to illustrate the correspondences 
 between the various members of group Y. The array of 
 figures may look repellent, but I have preferred to submit 
 the tabulation for an entire story rather than to give only a 
 part of it, or to resort to any printer's devices to compress it, 
 and thereby incur the risk of impairing its value. 
 
 KILLIS CAMPBELL. 
 
LIKE. 
 
 I was born June 11, 1872, at Enfield, King William Co., Va. 
 My early training was acquired in the public schools of my native 
 township. In the fall of 1888 I entered William and Mary Col- 
 lege, where I remained for two sessions. Securing a Peabody 
 scholarship at the University of Nashville in 1890, 1 entered that 
 institution, where I remained for two years, and was graduated in 
 1892 with the degree B. L. In the fall of 1893 I returned to 
 William and Mary College, and received there in June, 1894, the 
 degree B. A. In October of the same year I entered Johns Hop- 
 kins University, where I have studied for four academic years, 
 pursuing courses under Professors Bright and Browne in English, 
 Professor Wood and Dr. Learned in German, and Drs. Menger, 
 Marden, and Rambeau in French, and holding during the session 
 1897-98 a fellowship in English. 
 
 I take this occasion to acknowledge my obligations to all my 
 instructors, and in particular, to Professor Bright, to whom I am 
 deeply indebted both for guidance and encouragement in my 
 academic work and for much valuable assistance in the prepara- 
 tion of this study. I wish also to thank the authorities in the 
 British Museum, Balliol College, and Cambridge University 
 Libraries (especially Mr. Bickley of the British Museum) for 
 courtesies extended, and for valuable suggestions and informa- 
 tion concerning the Middle English manuscripts of the Seven Sages. 
 
 KILLIS CAMPBELL. 
 
 JOHNS HOPKINS UNIVERSITY, 
 May 1, 1898. 
 
8 
 
 RETURN CIRCULATION DEPARTMENT 
 
 202 Main Library 
 
 LOAN PERIOD 1 
 HOME USE 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 ALL BOOKS MAY BE RECALLED AFTER 7 DAYS 
 
 RENEWALS AND RECHARGES MAY BE MADE 4 DAYS PRIOR TO DUE DAI i 
 LOAN PERIODS ARE 1-MONTH. 3-MONTHS. AND 1-YEAR. 
 RENEWALS: CALL (415) 642-3405 
 
 DUE AS STAMPED BELOW 
 
 mm JUN 1 3 
 
 UNIVERSITY OF CALIFORNIA, BERKELEY 
 FORM NO. DD6, 60m, 1/83 BERKELEY, CA 94720 
 
I CN f\ 
 
 U.C. BERKELEY LIBRARIES 
 
 90