^{Note 1 in page 1141} A useful introduction to the whole problem is W. W. Lawrence's chapter, “The Sequence of the Tales,” in his Chaucer and the Canterbury Tales (New York, 19S0); to this I owe the initial impulse for the present study, and invaluable guidance along the way. My debt to others is also great, for my paper is in part a selective synthesis of various conclusions of recent scholarship.

^{Note 2 in page 1141} See F. J. Furnivall, Temporary Preface, Chaucer Soc. Publ., 2nd Ser., No. 3 (London, 1868), esp. pp. 9, 20–22.

^{Note 3 in page 1141} Fragment rv-v is clearly a unit; see J. S. P. Tatlock, PMLA, L (1935), 122 and nn. 58, 59; J. M. Manly and E. Rickert, The Text of the Canterbury Tales (Chicago, 1940; hereafter designated as “Manly”), ii, 489. (All quotations from the Canterbury Tales are based on Manly's Text.) The “Fragments,” indicated by Roman numerals, were numbered according to the order in the Ellesmere MS. by Manly, Canterbury Tales by Geoffrey Chaucer (New York, [1928]), and by F. N. Robinson, The Complete Works of Geoffrey Chaucer (Boston, [1933]). The capital letters designate the “Groups” in the order devised by Furnivall for the “Chaucer Society.”

^{Note 4 in page 1142} “B,” of course, stands for “B¹” plus “ ⁸,” i.e., Fragments ii and vii.

^{Note 5 in page 1142} I quote Tatlock, p. 131; see Manly, ii, 475–489; Germaine Dempster, PMLA, LXTV (1949), 1123–42.

^{Note 6 in page 1143} George Shipley, MLN, ii (1895), cols. 259–279; R. K. Root, The Poetry of Chaucer (Boston, 1906), pp. 153–158; W. W. Lawrence, MP, xi (1913), 256; F. Tupper, Nation, 16Oct. 1913, p. 355; J. L. Lowes, PMLA, xxx (1915), 365; S. Moore, PMLA, xxx (1915), 122; Manly, ii, 490-493; W. W. Skeat, Academy, XL, NO. 1004 (1 Aug. 1891), 96, and The Evolution of the Canterbury Tales, Chaucer Soc. Publ., 2nd Ser., No. 38 (London, 1907), pp. 17–19.

^{Note 7 in page 1143} Page 126; Tatlock wrote as follows: “[Fragment vi] might possibly, but in each case inappropriately and with difficulty, come after [ii,iii, rv–v, or viii], and for none of these positions is there the slightest internal evidence anywhere. With perfect smoothness it may be put between [i and ii] or between [vii and viii]; the latter, Furnivall's position, cannot be improved on.” This is all Tatlock had to say concerning vi. Skeat (iii, 379) stated that vi could go anywhere between i and viii.

^{Note 8 in page 1144} J. M. Manly, Canterbury Tales by Geoffrey Chaucer (New York, [1928]), p. 655; Root, MLN, XLTV (1929), 493–496; Manly, SP, xxviii (1931), 613–617; J. A. Work, JEGP, XXX (1932), 62–65; Tatlock, pp. 122–125,138–139; Manly, ii, 493, and 1,277.

^{Note 9 in page 1144} Manly, iv, 361, 527. In the early, conservative Hg, “maunciple” is written over erasure in ii 1, the first line of a quire. The Manciple's Tale, which ends at the bottom of the last page of the preceding quire (i.e., before certain quires were misplaced; see Manly, 1,270), is, like NPT and the iv–v links, in an ink now turned yellow, and was probably added late to the MS. (Manly, 1,271–273). The word “maunciple” in ii 1, however, is not in this yellow ink—perhaps an argument against the belief that the reading was originated by the scribe of Hg after McT had been inserted in the MS. See Dempster, PMLA, LXTV (1949), 1129; and ii. 22, pp. 1129-30; LXI(1946), ii. 77, p. 394. The reading “Ten” for “Foure” in ii S in ¹ cannot with any security be attributed to Chaucer since Arabic “4” and Roman “x” were so frequently confused by scribes.

^{Note 10 in page 1145} F. G. Fleay, Folk-Lore Record, ii (1879), 162; J. Koch, The Chronology of Chaucer's Writings, Chaucer Soc. Pub., 2nd Ser., No. 27 (London), pp. 56–59 (Koch was willing to put vi between MLT and ML Endlink); Shipley, MLN, ii (1895), cols. 267–272, and xi (1896), cols. 290–293.1 have found no support for J. E. Wells's repeated statement in A Manual of the Writings in Middle English: 1050–1400 (New Haven, 1916), p. 679, that the “present preference” is for Shipley's order.

^{Note 11 in page 1146} Skeat, v, 141; Furnivall, Academy, XLVHI, NO. 1223 (12 Oct. 1895), 296–297; Tupper, JEGP, xxxni (1934; hereafter designated as “Tupper”), 355; C. Brown, SP, xxxrii (1937) 18.

^{Note 12 in page 1147} Koch, pp. 56–57; see also Skeat, v, 132; E. P. Hammond, Chaucer: A Bibliographical Manual (New York, 1908), pp. 258,281; Manly, ii, 491.

^{Note 13 in page 1147} It may be, as Carleton Brown suggested (SP, xxxrii [1937], 23–33), that Chaucer at first intended the General Prologue (11.1-821) to be immediately followed by the ML Head-link and the Tale of Melibee, thus beginning the story-telling of the first day at 10 A.M. If so, and if Chaucer later decided to begin with the Knight's Tale and the rest of Fragment i, we can understand how in the new opening (1. 822) he aroused the pilgrims at dawn in order to get these new stories (the Knight's Tale and the fabliaux) out of the way before the ML Headlink. (Brown assumes that the arising at dayspring was in the first version; but this simply leads to difficulties.)

^{Note 14 in page 1147} Moore, pp. 116–123; Manly ii, 490–494.

^{Note 15 in page 1148} Manly, il, 492.

^{Note 16 in page 1148} Manly, iii, 230; v, 552–553; for the tale orders of the MSS., see the Charts after p. 494 of Vol. n. Manly (ii, 189) offers an explanation for the reading “shipman” in Se.

^{Note 17 in page 1148} Manly, v, 552–553.

^{Note 18 in page 1148} Manly, I, 271–275; ii, 477–478; Dempster, PMLA, Lxni (1948), 470–473 (in line 8, p. 473, read “Me-Sq” for “ME-Sq”; in line 10, read “ML-Sq sequence” for “Me-Sq sequence”); Lxrv (1949), 1123–26, 1131–35, 1140–42. Hg, like El and the MSS. of group a, lacks the ML Endlink.

^{Note 19 in page 1149} Ha⁴, Ln, Mc, Py, Ra', Ry¹. In a seventh MS., Ra², the reading “somnour” was changed to “Squyer.”

^{Note 20 in page 1149} MP, III, 163–164.

^{Note 21 in page 1149} Tatlock's figuring went somewhat astray, but his conclusion is none the less sound. In the first place, certain pilgrims in addition to the Sergeant of Law must be definitely eliminated: the Parson, who plays another role in this Link; his brother the Plowman, who would not thus have injected himself into the altercation; Chaucer, who is the narrator; and the Prioress, whose personality is poles apart from the interrupter's. There may be others; in any event, without these, and without the other “Nonne” and the “Preestes thre,” we are left with a total of 22 pilgrims as possible candidates for the role of interrupter. Second, our question is not the chance that getting precisely the 3 “s—” pilgrims out of 22 is significant, but the chance that, given 1 “s—” pilgrim (to produce the archetypal “s . .,” reading), our getting precisely the other 2 out of the remaining 21 is significant. I therefore rephrase Tatlock's statement in the following manner. Given 1 “s—”pilgrim, let us suppose the appearance of the other 2 pilgrims to be pure accident. Then the chance that they both be “s—” pilgrims is 1 to 210 (2 · 1/[21 · 20]). Finally, if the MLT—SqT sequence originated in Hg (see my notes 18 and 45), then the reading “squyer” must probably be omitted from our calculations as due to the sequence of tales; hence, given 2 “s— ” pilgrims (“shipman” and “squyer”), let us suppose the appearance of the 3^rd to be pure accident. Then the chance that he be an “s—” pilgrim (“somnour”) is 1 to 20.

^{Note 22 in page 1150} I give the Manly text, but instead of “somnour” offer the reading “s ...”

^{Note 28 in page 1151} Manly, ii, 190; R. F. Jones, JEGP, xxrv (1925), 546; R. K. Root, MLN, XLViii (1933), 470; SP, xxxvni (1941), 12; Manly, Canterbury Tales by Geoffrey Chaucer (New York, [1928]), p. 573; see also Robinson, p. 6.

^{Note 24 in page 1151} PMLA, Lvn (1942), 36–37. The arguments offered in favor of the Squire by C. R. Kase (Ch. m) in Three Chaucer Studies (New York, 1932), have been answered by Tupper, pp. 352-372, and Tatlock, ii. 44, pp. 115–116. The plea for the Yeoman by A. Brusendorff in TheChaucer Tradition (London and Copenhagen, [1925]), pp. 72-73, came stillborn into the world.

^{Note 25 in page 1151} To ask whether he would call his own body “joly” is to open a Pandora's box of psychological speculation. Manly (ii, 195) favored this “jocose rascal who rode garlanded with flowers, armed with a buckler of a cake and accompanying his friend and compeer in lusty love songs.”

^{Note 26 in page 1152} MLN, xxviii (1933), 470; on the Summoner's Latin, see also Jones, p. 531, ii. 37; opposed to Root's interpretation were Tupper, p. 364, and Tatlock, ii. 44 on p. 116.

^{Note 27 in page 1152} See, e.g., Tupper, pp. 361–362. By 1907, after studying the MSS., Skeat had given up the “Bradshaw shift” and felt that Chaucer had never intended the ML Endlink to introduce the Shipman; he nevertheless wrote, “I do not object to its being made into a Ship-man's Prologue, because, as he wants a Prologue and the Prologue suits him, it is the best thing to do with it” (Evolution, p. 13). In 1910 Skeat held that ML Endlink had been originally “composed with the Shipman in view” (MLR, v, 433).

^{Note 28 in page 1153} G. L. Kittredge, The Poetry of Chaucer, pp. 168–170.

^{Note 29 in page 1153} R. D. French, A Chaucer Handbook, p. 232.

^{Note 30 in page 1154} 5, xxxviii (1941), 12.

^{Note 31 in page 1155} Page 522 and n. 21. I am not here concerned with Jones's conjectured history for the Wife of Bath's Prologue.

^{Note 32 in page 1156} See Jones, n. 29, whose evidence I bring up to date from Manly in, 230–231 (or v, 553).

^{Note 33 in page 1158} iii (D) 166; nor do the Pardoner's later words to the formidable Alice (11. 186-187), “spareth for no man,/And teche us yonge men of youre praktyke,” belong 200 lines after his humiliation by Harry Bailey.

^{Note 34 in page 1159} Kittredge, MP, ix (1912), 435-467; Lawrence, MP, xi (1913), 247–258; Tupper, Nation, 16 Oct. 1913, p. 355; J. S. Kenyon, JEGP, xv (1916), 282-288; S. B. Hemingway, MLN, xxxi (1916), 479–483. See also J. B. Severs, SP, XLHI (1946), 28–29, 36–37; Lawrence, Chaucer and the Canterbury Tales, Ch. v.

^{Note 35 in page 1159} Tupper, however, wrote (PMLA, xxix [1914], ii. 6, p. 97), “It seems to me a potent additional argument for the order here adopted, that the Physician's story of oppressed virginity courting death rather than disgrace follows so naturally upon the Franklin's many illustrations of this pathetic theme (F. 1364 f.).”

^{Note 35a in page 1159} Speculation concerning where Chaucer—had he lived longer—might have had the pilgrims eat and sleep, is beyond the scope of this study, which is concerned with Chaucer's final intention insofar as it is revealed by the internal evidence and the evidence of the MSS. Although the speculations on the overnight stops of the pilgrims offered by Charles A. Owen, Jr.—PMLA, LXVI (Sept. 1951), 820–826—are interesting and stimulating, I feel that the evidence for the VH-III sequence is sufficiently cogent to discourage our supposing that Chaucer could have intended putting in—iv-v on a conjectural return journey. Indeed, we cannot be sure that Chaucer might not have made Rochester an overnight stopping place on the trip from Southwark to Canterbury (i.e., between vii and in) since the Host's words, “Lo, Rouchestre stant heer fast by,” might, for example, be directing the Monk's attention to the view of the Castle and the Cathedral as the company approaches Rochester; they need not necessarily mean that the pilgrims are departing, or about to depart, from Rochester (vii 1926 [B² 3116]). Actually, however, Chaucer left the Canterbury Tales in a form which reveals his continued concern with the route and with the sequence of the tales, but with an almost total neglect of overnight stops along the way.

^{Note 36 in page 1160} Skeat, iii, 434; A. W. Pollard, Chaucer (London, 1893), pp. 108–109; Lawrence, MP, xi (1913), 247–258, and Chaucer and the Canterbury Tales, pp. 90–118, 121–127; Tupper, Nation, p. 355; Lowes, PMLA, xxx (1915), 365. Professor W. W. Lawrence has written me as follows, and at my request has given me leave to quote him. “Although believing that the Wife of Bath (D) should directly follow the Nun's Priest (B²), except in a rigid transcript from the MSS., I advocated in my Chaucer and the Canterbury Tales, as a practical measure, the Chaucer Society's order, in which the Physician and Pardoner (C) intervene, because that order has been adopted in many of the current editions. The careful study of references and echoes which you have made convinces me, however, that there is a better place for C. To Root's objection in his review (Speculum, Jan. 1951, p. 171) that we should expect ‘some sort of cross-reference’ linking the Wife of Bath to Melibeus and the Nun's Priest, I would point out that we do have just that. 1) The Wife's 'it is an impossible That any clerk wol speke good of wyves,' etc. (W. B.'s Prologue 688 ff.) must, it seems to me, be a hit at the Nun's Priest, not at the Clerk of Oxford, who has not yet spoken, and whom the Wife has no reason to attack (see my book, p. 138). 2) The Wife's opening words, in which she says that experience rather than authority has convinced her of the woe in marriage, follow naturally after the Nun's Priest's 'If I conseil of wommen wolde blame, . . . Rede auctors, wher they trete of sivich mateere, And what they seyn of wommen ye may heere' (B²4451ff.).”

^{Note 37 in page 1161} PMLA, ixiv (1949), n. 32 on p. 1134. The continuity of these genetic groups may be seen by comparing the “Classifications” on pp. 65, 83,108,125, and 138 of Manly, rv, with those of the other tales (in Vols, in and iv). See also Dempster, p. 1127 and ii. 17. Manly, of course, believed that most of our CT MSS. derived chiefly from pre-1400 gift copies; see Dempster, PMLA, LXI (1946), 384–392. Evidence against this theory has been offered by Dempster, p. 385, n. 32; PMLA, ixiii (1948), 464, ii. 29; p. 478, and n. 12;MLN, lxin (1948), 325–330; PMLA, LXIV (1949), 1127, and n. 18 on pp. 1127–28. See also Severs, Speculum, xxi (1946), 295–302. In these articles are described conditions which are incompatible with Manly's theory of pre-1400 versions. The present study is postulated on the rejection of this theory, both in the evidence for the “1400” vi–vn sequence, and in the conclusions which follow.

^{Note 38 in page 1162} There is no use in seriously considering the alternative supposition, namely, that vi was accidentally shifted to the position between ii and vii; for the internal evidence has shown that vii (not vi) was wrenched loose from its moorings.

^{Note 39 in page 1162} On the fascicle theory, see Hammond, Chaucer, pp. 159, 243; Brusendorff, p. 126; Dempster, PMLA, LXI (1946), 383 and ii. 23; LXIV (1949), 1127 and ii. 18 on pp. 1127–28.

^{Note 40 in page 1162} Such a misplacement was clearly the basis (and obverse) of Henry Bradshaw's shift of vn; the proposition was more recently brought forward and defended by Tupper, JEGP, xxxiii (1934), 366: “The modern English editors . .. restored the original sequence (still extant in Selden), which had been destroyed by the accidental shifting, in an archetypal manuscript. . ., of the fascicule which contained [vn] to a much later place (behind [vi]) in the collection.”

^{Note 41 in page 1162} Manly, Chart after p. 494, Vol. ii. It has been suggested, but without conclusive evidence, that Chaucer cancelled the ML Endlink and the NP Endlink. For arguments pro and con, see, e.g., Hammond, Chaucer, p. 277; Robinson, pp. 800,1013; Tupper, n. 38 on pp. 371-372; Tatlock, pp. 112–119; Manly, ii, 188–190, 422–423, 480–481; iv, 517; Brown, MLN, LV (1940), 616; Dempster, PMLA, LXI (1946), ii. 18 on pp. 382–383; LXTV (1949), ii. 31 on p. 1134; ii. 48 on p. 1139. Further support for the shift of vii is perhaps offered by the MSS. of group a. Their common ancestor possessed certain lines which suggest that for the Nun's Priest's Tale and the Wife of Bath's Prologue “the scribe or some predecessor had access to special copies.” Some of these lines appear to offer late additions by Chaucer; others have “correct readings as against the other MSS.” (Dempster, LXI [1946], 398-399; Manly, I, 361; ii, 191–194, 421-423; iv, 257; vi, 558.) Now group a, which thus brings us close to Chaucer's desk, preserves the genuine but fragmentary Nun's Priest's Endlink, even though this link does not introduce a specific tale. In the last line (vii 3462 : “Seide un to another, as ye shullen heere”) the word “another” is suspect (see, e.g., Tatlock, p. 114; Robinson, p. 862) and sounds like the a scribe's attempt to patch things up for the best. The unsatisfactory state of the NP Endlink, and its absence from most non MSS., I attribute to the shift of Fragment vn from its proper position just before Fragment m. The NP Endlink as preserved in group a, may have been the opening of a completed “Nun's Priest—Wife of Bath” Link, fractured and half lost because of the accidental dislocation of vn. Line 3462 may originally have read, “Seide un to the Wyf, as ye shullen heere,” or the like, and have been followed by lines intended to introduce Alice but not suitable for the Second Nun, whose tale follows Fragment vn in the MSS. of group a. The o-scribe's rejection of the second half of such a link, and the retouching of 3462, are comprehensible.

^{Note 42 in page 1163} Page 119, ii. 49; p. 112; and, in general, pp. 112-119; and Dempster, PMLA, LXTV (1949), ii. 48 on p. 1139; Tupper (ii. 38 on p. 372) wrote: “If, as I have sought to show, the [Man of Law's Endlink] was separated from its sequel, the Shipman's tale, by the accident of a misplaced booklet, the probabilities are that it was cancelled by some copyist. .. when it had ceased to perform any cohesive function. The suppression of this significant Link seems to me post-Chaucerian.”

^{Note 43 in page 1164} ', lxiii, 456–484.

^{Note 44 in page 1164} PMLA, lxiv, 1123–42 (quotationsfrom p. 1124, and n. 1 on p. 1123).

^{Note 45 in page 1164} Manly, i, 266–277; ii, 477–478; Dempster, LXTV, 1129–40. A complete analysis of the compilation of Hg would have to be lengthy indeed. Hg is clearly the work of the scribe who later wrote El. Hg begins with Fragment and ends with x. Overlapping of quires and tales reveals that the positions of the tales of Fragments ii, iv-v, vi, vii, and ix, must have been determined by the order in which the Hg scribe obtained copies for transcription. We must note that although he obtained the iv-v tales together, their order in Hg is incorrect: Sq—Me—Fk—CI. Signatures, inks, and the make-up of quires show that only after he had prepared the bulk of the MS. did the Hg scribe receive vi (NPPT), ix (McPT), and two of the iv-v links (all in an ink now turned yellow); vrrr (SNPT: extra folios in the center of the quire; Hg never had vm^b—CYPT); and Fragment in (in an ink now turned brown). The simplest explanation of these late additions to the MS. is that the “yellow ink” materials (Manly, 1,268,271-273; ii, 477–478) were added first, and that along with the Sq—Fk and Me—Sq Links (which he adapted as “Sq—Me” and “Me—Fk” Links; Manly, i, 272) the Hg scribe also received the other rv-v link, the CI—Me Link. From these three links the Hg scribe must have learned one simple fact: that his tale order for rv-v (Sq—Me—Fk—CI) was incorrect, the correct order being CI—Me—Sq—Fk; hence it must have been amply clear to him that he would never find a Fk—CI link for the blank he had left between FkT and CIT. At this point, accordingly, he later inserted the extra folios needed for the short Second Nun's Prologue and Tale (Manly, i, 274–275; Dempster, LXUI, 474, ii. 59). This was the only place in the entire MS. where he could put SNPT without blocking an already existing or a possible future continuity. Finally, when Fragment m reached him, he gave up hope of finding the rest of the Cook's Tale, and inserted m in the mechanically easy position between Fragments i and ii (Manly, i, 273–274). Hence, except for the final additions (viii and iii), which were obviously misplaced, the Hg scribe placed the Fragments which he received, in the “1400” order: I iv-v vi vii ix x. I therefore suggest that the Hg order somehow reflects the “1400” order before it became completely disintegrated; for the person who provided the Hg scribe with the copies of the CT pieces seems to have supplied him with the central materials (ii, iv-v, vi, vii, ix) in precisely the one order which can have any claim to “authority” from any MSS. : the “1400” order, found also in El and group .

^{Note 46 in page 1165} Another theory (briefly sketched as a possibility by Mrs. Dempster, Lxrv, 1138–39) is that the order of El represents a revision of the order of Hg (written earlier by the same scribe) based in part on the additional materials that became available to him in his later role as El-scribe. The El-scribe's new position for Fragment in, his correct ordering of the iv-v tales, and his position for vin, are innovations which (as Mrs. Dempster notes, p. 1138) could have been “dictated by internal evidence” (allusions, links, notes of place and time). However, the elements which the Hg-scribe placed incorrectly are precisely those which he received last (see the footnote preceding). On the other hand, except for the incorrect arrangement of the rv-v tales, the materials which the Hg-scribe received first are in the El order. The order of these materials may well be due to pre-Hg factors: we must note the El-scribe's restoration of the original Hg i— sequence, the presence of the vi-vii sequence in each MS., and of the ix— sequence (see Dempster, LXIV, 1138). Indeed, the only features of Hg which could possibly represent an earlier tradition coincide exactly with the “1400” order. Conversely, the three innovations supposedly made in El (if we imagine it to offer a revision of the Hg order) correspond precisely to the three late incorrect insertions made in Hg. It is presumably easier to believe that the similarities of Hg and El stem independently from an earlier tradition (the “1400” order) than that Hg just happens to be right in its original sequence. The early short form of Hg may reflect the “1400” order because the scribe was at first provided with materials in that order insofar as they were available; or it may reflect what was left of the “1400” order after copyists had started to demolish the ordered pile of fascicles and sheets.

Although the changes necessary to convert the Hg order into the El order could have been dictated by internal evidence, the complicated history of the assembling of Hg does not suggest that the Hg-El scribe (in his earlier Hg-role) was interested in following such evidence. For example, he accepted the MLPT-SqT sequence even though the two tales are clocked at “ten” and “prime” respectively (ii, 14; v, 73), and he went so far as to leave a space in anticipation of a link to tie the two tales together (Manly, i, 271; ii, 478; Tatlock, ii. 6 on p. 135). Again, when the iv-v links came to hand, he adapted two of them to the tale order already existing in his MS. by the expedient of changing the name of the second pilgrim mentioned in each (Manly, i, 268, 271–272; Dempster, LXHI, 466 and ii. 36; Tatlock, ii. 6 on pp. 134-i135). Finally, in his later role as El-scribe, he did not rectify the geographical inconsistency by putting Rochester (vii) before Sittingbourne (viii), even though such a move would have disrupted no other piece of “internal evidence.” Like his 15th-century successors (as I understand them after reading Dempster, Lxra and LXTV), the Hg-El man seems to have chiefly aimed to welcome preeedent and to achieve and preserve continuity. On the whole it seems to me more reasonable to suppose that the order found in a and El is a direct descendant of the “1400” order, than to attribute to the Hg-El scribe the interests of latter-day Chaucerians in the interpretation of internal evidence. Examination of Manly's dating of the various MSS. involved, reveals no evidence whatever against this supposition.

^{Note 47 in page 1167} For example, in defending his use of the El order, Robinson (p. 1005) states flatly that Chaucer “never made a final arrangement of what he had written”; elsewhere (p. 2) he writes: “Nor, apparently, did he get to the point of arranging the tales he had written.” These statements and the remarks on p. 800 do not seem wholly compatible with Robinson's selection of the reading “Shipman” for 1179 (ML Endlink).