925 
 
 L282 
 
 K7 
 
 &NRLF 
 
 ^ 3 54-:? 1^3 
 
I 
 
Trr~r - ^^g •anivcrsit^ ot CbtcaQo 
 
 AN ESSAY TOWARD THE CRITICAL TEXT 
 
 OF THE A-VERSION OF "PIERS 
 
 THE PLOWMAN" 
 
 A DISSERTATION 
 
 SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF ARTS 
 
 AND LITERATURE IN CANDIDACY FOR THE DEGREE 
 
 OF DOCTOR OF PHILOSOPHY 
 
 (department of ENGLISH) 
 
 BY 
 
 THOMAS A. KNOTT 
 
 {? R ,\ 
 OF IK 
 
 Reprinted with additions from 
 
 Modern Philology, VoL XII, No. 7 
 
 Chicago, 191 5 
 
XTbe lllniversit^ ot CbicaQo 
 
 AN ESSAY TOWARD THE CRITICAL TEXT 
 
 OF THE A-VERSION OF "PIERS 
 
 THE PLOWMAN" 
 
 A DISSERTATION 
 
 SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF ARTS 
 
 AND LITERATURE IN CANDIDACY FOR THE DEGREE 
 
 OF DOCTOR OF PHILOSOPHY 
 
 (department of ENGLISH) 
 
 BY 
 
 THOMAS A. KNOTT 
 
 Reprinted with additions from 
 
 Modern Philology, VoL XII, No. 7 
 
 Chicago, 191 5 
 
ACKNOWLEDGMENT 
 
 The execution of such a piece of work as this would have been 
 impossible without my incurring many obligations. I wish to 
 express my warmest gratitude to the authorities and the librarians 
 of the British Museum, the Bodleian, Trinit}^ College, Dublin, 
 Trinity College, Cambridge, and Lincoln's Inn, for their courtesy 
 and cordiality, and to the Duke of Westminster, the late Sir Henry 
 Ingilby, and Sir William Ligilby, for the courteous and liberal spirit 
 with which they made their MSS accessible. Dr. Furnivall and 
 Professor Skeat gave me valuable advice and assistance (and every 
 student of "Piers the Plowman" is under a heavy debt to them for 
 their labors in locating the MSS long ago). Professor Kittredge 
 was kind enough to give many hours to the examination of some of 
 my results. My friend. Dr. J. R. Hulbert, read my MS and gave 
 me advice. Professor C. H. Beeson also read the article and criticized 
 it. It is impossible for me to acknowledge adequately what I owe 
 to my teacher and friend. Professor John M. Manly, to whose 
 inspiration this study is due, and whose training and assistance 
 have given my work whatever value it may possess. 
 
^7- 
 
 l 
 
 
 Modern Philology 
 
 Volume XII January 10 1 ^ Number 7 
 
 AN ESSAY TOWARD THE CRITICAL TEXT OF THE 
 A-VERSION OF ''PIERS THE PLOWMAN" 
 
 The interest of students of Middle English literature in the Piers 
 Plowman poems was greatly stimulated several years ago by two 
 articles by Professor John M. Manly, "The Lost Leaf of 'Piers the 
 Plowman'"^ and " 'Piers the Plowman' and Its Sequence."^ 
 I was so fortunate as to be a student under Professor Manly in 
 1905, when his belief in the diversity of authorship of the several 
 versions was daily receiving fresh confirmation from his investi- 
 gations, and we recognized the need for an adequate critical text 
 in order that the differences between the three versions might be 
 determined satisfactorily. Accordingly, in my first subsequent 
 vacation, in the summer of 1907, I began the necessary work by 
 collating the fourteen MSS of the A-version as far as 8 . 130 (Skeat's 
 numbering) ,3 with the object of studying their relationship to one 
 another, and attempting to settle the existing uncertainties of the 
 text. This work I have since been carrying on as time and oppor- 
 tunity offered, and the results I now publish in this essay. The 
 critical text, with the collations, must wait until similar work on 
 the B- and C-versions has been finished (when all will be printed 
 
 I Modern Philology, III (January, 1906), 359-66. 
 
 - The Cambridge History of English Literature, II (1908), 1-42. 
 
 ' This study of the critical text covers only the prologue and the first eight passiis 
 to 8 . 130 becaixse it is at this point that Mr. Manly (and I) believe the work of Al ceases. 
 This line marks the close of the most vigorous, the most readable, and the best organized 
 part of the A-text. 
 
 389] 129 [Modern Philology, January. 191.5 
 
130 , TpoMAS A. Knott 
 
 together), but the text I hope to pubUsh in a short time in the form 
 of a reading edition. 
 
 For the A-text there are fourteen MSS, some of which unfortu- 
 nately are not complete, and some of which are not pure A-text 
 throughout. A table of these MSS, showing what each contains, 
 and where each is defective, and where any one is B- or C-text, may 
 be helpful to the student, and therefore is appended.^ 
 
 I. Vernon Codex. Omits 1 . 176-83 (178-85) and 2. 106-21 (111-27). 
 
 II. Harleian 875. Omits 6 . 49—7 . 2 (6 . 52—7 . 2) . 
 
 III. Ingilby. IV. Lincoln's Inn 150. V. Trinity College, Cambridge 
 R 3.14. VI. Rawlinson Poet. 137. All practically complete save for a 
 few sporadic omissions of single lines. 
 
 VII. University College, Oxford, 45. Omits 1 . 33-99 (folio torn out). 
 
 VIII. Douce 323. Omits 3. 120-34 (128-42). 
 
 IX. Harleian 6041. Parts of ff. 23, 24, 26, and 27 are torn out, thus 
 causing the loss of 7.59-74 (60-79), 82-105 (87-110), 115-36 (120-41), 
 145-87 (150-94), 198-218 (205-25), 228-47 (235-54), 258-78 (265-86); 
 and the loss of parts of 7.53-58 (54-59), 77-81 (82-86), 108-14 (113-19), 
 139-44 (144-49), 188-97 (195-204), 219-27 (226-34), 248-57 (255-64), 
 279-89 (287-97). 
 
 X. Trinity College, Dublin, D 4.12. Omits 7.45-69 (46-70) and 
 7.210 (217) to the end. 7.44 (45) is actually the final line in the MS, but 
 7.69a-209 (71-216) had been transposed in an archetype to a position 
 before 1.180 (182), and therefore were preserved. 
 
 XI. Ashmole 1468. Begins at 1 . 142, because the preceding leaves have 
 been cut out; then omits 2.18-145 (18-158); 3.30-33 (32-35), 112-226 
 (120-235); 7.33-81 (34-86); 8.32-80 (32-81), all but 3.30-33 because 
 leaves have been cut out. 
 
 XII. Harleian 3954. Is B-text to (B) 5.128, then A-text from (A) 
 5. 106— 8. Ill (5. 107— 8. 113), then omits to 9.97. No extended omissions. 
 
 XIII. The Duke of Westminster's MS. Inserts a large number of 
 lines and passages from the B- and C-texts : B 1 . 32-33 after A1.31; B1.113- 
 16 after A 1.111; C 3.28-29 after A 2.20; C 1.84-87, 89, 92, 98-100, 
 102-4 after A 2.65 (68); C 3.185-88 after A 2.130 (140); C 3.243-48 
 after A 2 . 194 (208) ; C 4 . 32-33 after A 3 . 33 (35) ; B 4 . 17-18 after A 4 . 17 ; 
 B 4.62 after A 4.48; B 4.119-22 after A 4.105; then follows A 108, then 
 B 4. 123-25; B 4. 152-56 after A 4. 143; B 4. 165-70 after A 4. 145; B5.36- 
 41 after A 5.33; B 5 . 49-56 after A 5 . 39 ; B 5.60 after A 5.42; B 5.87-93 
 after A 5 . 68 (69) ; B 5 . 120-21 after A 5 . 98 (99) . 
 
 ' The line numbers in this paper refer to the Critical Text, but in order to facilitate 
 reference until that is published, I give in parentheses the line numbering of Skeat's 
 Early English Text Society edition, wherever the number in the CT differs from Skeat's. 
 
 390 
 
The A- Version of "Piers the Plowman" 131 
 
 XIV. Digby 145. No extended omissions. Has several contaminations 
 from the C-text, especially in the prologue, which is chiefly C, with some 
 readings from A. The other insertions are B 3.52-54, 56-58 after A 3.45 
 (47); C 7.423—8.55 after A 5.220 (228); then A 5.215-20 (223-28) is 
 repeated; (Digby changes 214 (222) so that it reads "this glotoun" for 
 "sleul^e"); C 8.70-154 after A 5.251 (259); C 8.189-306 substituted for 
 A 6.31-123 (34^126). 
 
 As the basis of my text I have used MS R 3 . 14 in the library 
 of Trinity College, Cambridge, represented in my paper by "T." 
 By "basis" I mean, of course, not that I shall print that MS as it 
 stands, nor with such occasional readings from other MSS as may 
 "seem better" to me. On the contrary, the readings adopted into 
 the CT must always be the critical readings, as attested in every 
 case by the weight of evidence, genealogical and other. No matter 
 how plausible the reading of T may seem, it must not be retained if 
 not supported. By "basis" I mean, therefore, little more than the 
 basis for spelling and dialect, for whenever the reading of T is replaced 
 by the critical reading, it seems better to make the latter conform 
 in spelling and dialect to T. Otherwise we should have a critical 
 text containing too may inconsistent forms and spellings. In every 
 case, of course, when the apparatus is printed, the footnotes will 
 record all variants from the CT, including those of T. 
 
 MS T was chosen as the basis of the CT because it is early 
 (shortly after 1400), because it is well spelled, and because it con- 
 tains comparatively few individual deviations and errors, and there- 
 fore probably requires less changing to make it a critical text than 
 any other MS. It should be said that the CT would have been 
 exactly what it is, save for dialect and spelling, no matter what 
 particular MS had been chosen for a basis. 
 
 The numbering of the lines differs in this paper from that of 
 Skeat in his E.E.T.S. and Oxford editions, because I have numbered 
 the lines of the CT, and of course the CT does not contain the 
 unsupported expansions and the spurious lines, contained in only 
 one MS, some of which Skeat admitted into his text. 
 
 The following lines in the E.E.T.S. edition have been rejected 
 in the CT because they occur only in MS Harleian 875:1.176-77; 
 2.31, 34, 48, 96, 118, 136-39, 141-43, 182; 3.19-20, 66, 91-94, 98, 
 
 391 
 
132 Thomas A. Knott 
 
 234; 5.182; 6.1-2, 5; 7.26; 8.46, 101, 125-26. One line is re- 
 jected because it is in V only: 7.286. In two cases one line of the 
 CT has been expanded into two by V: 5.55-56; 7.157-58; the 
 CT numbering in each case is reduced to one line. One line is in 
 H, and, with some differences, in H2: 2.79. 5.202-7 are in only 
 UT2AH3; that is, in one small sub-subgroup, often contaminated 
 from the B-text, and one other MS; the lines are a contamination 
 from the B-text, and are therefore rejected from the CT. Lines 
 7.71-74, containing the names of Piers's wife and children, are an 
 interpolation, and are therefore omitted.^ Lines 7.180-81 are an 
 expansion of one line, and, though contained in MSS V, H, and I, 
 are reduced to one line in the CT. 
 
 It is hardly necessary to recount here in great detail the processes 
 that must go toward the determination of a critical text. Adequate 
 expositions of these processes have long been accessible, especially 
 in the Introduction to Westcott and Hort's Greek New Testament, 
 and in Edward Moore's Contributions to the Textual Criticism of the 
 '^Divina Comedia," "Prolegomena"; and the principles have been 
 admirably stated recently by Dr. Eleanor Prescott Hammond in 
 her Chaucer: A Bibliographical Manual, pp. 106-13. 
 
 The older method of printing a text was to select an old, well- 
 spelled, well-written MS, the readings of which seemed to the editor 
 to give "the best sense." In case of dissatisfaction with a reading, 
 support for it was looked for in other MSS, and, if support failed, 
 a reading was adopted from some other MS or MSS which the editor 
 thought gave the ''best sense." This "eclectic" method was 
 unscientific and unreliable for two reasons: The editor left in his 
 text a large number of readings which gave "good smooth sense," 
 but some of which were sophisticated, that is, introduced by copy- 
 ists who were practicing conjectural emendation; and others of 
 which (introduced carelessly) were intelligible, but which could not 
 be supported by scientific proof. Secondly, this method laid too 
 much responsibility on the unchecked discretion of the editor, who 
 often adopted a reading merely because it was in the greater num- 
 ber of MSS, and who, on the other hand, often adopted readings 
 merely according to his whim or his personal taste. 
 
 > Cambridge History of Enylish Literature, II, 33; and my forthcoming article in 
 Modern Philology. 
 
 392 
 
The A-Version of "Piers the Plowman" 133 
 
 The dangers arising from the exercise of personal taste or whim, 
 and from reUance on mere number of MSS, are avoided by the 
 critical method. A reading must not be valued according to the 
 number of supporting MSS, for a large number of MSS may be, 
 and often are, descended from one common ancestor, from which the 
 reading has been transmitted to its descendants. 
 
 The necessity is therefore evident for classifying all extant MSS 
 according to their family relationships, and for constructing a family 
 tree, before anything is done toward determining what readings 
 ought to be adopted in the text. 
 
 Two or more MSS, or two or more groups of MSS, are assigned 
 to an identical, hypothetically reconstructed ancestor, or archetype, 
 if they possess in common a number of clear errors, omissions, and 
 additions. Common errors, deviations, and omissions in two or 
 more MSS must be due to coincidence, or to contamination, or to 
 their existence in the MS from which copies were made. If there 
 are more than a very few significant errors, the laws of probability 
 forbid attributing them to coincidence. If two MSS, copied from 
 two entirely different archetypes, were afterward compared, and 
 a number of erroneous readings were transferred from one to the 
 other by the collator, the position of descendants of the contami- 
 nated MS in the family tree would be very difficult to determine. 
 For these descendants would contain the erroneous readings and 
 deviations which were their legitimate inheritance, and also those 
 which resulted from the contamination, and the text critic would 
 find it difficult, if not impossible, to determine the real position of 
 the MSS. But one characteristic will enable him to locate such 
 MSS with some degree of certainty, and thus to determine which 
 are contaminations and which are legitimately descended errors. 
 Omissions are not the result of contamination. We have a number 
 of A-text MSS of "Piers the Plowman" which contain readings 
 inserted or substituted in different or later hands or inks than the 
 original hand and ink.^ In several of these MSS lines or words are 
 inserted which were omitted by the original scribe, or lines are 
 inserted from the B-text. But in no case is any Une or word or 
 passage expunged. Possessors of MSS who compared them with 
 
 I E.g., T, H2, D. W, Di. Ts, R, I, H. 
 
 393 
 
134 Thomas A. Knott 
 
 other MSS seem to have thought that their own MSS were defective 
 or wrong whenever they differed or omitted anything, but not 
 when they contained hnes, words, or passages which the other 
 MSS omitted. The possession of any considerable number of 
 common omissions, therefore, unless they can be accounted for on 
 some other definite grounds, makes a very strong case for common 
 descent. 
 
 Common ancestry is of course rendered more certain if all the 
 MSS of a group possess also a considerable number of other variants 
 (not necessarily errors) different from the readings common among 
 the MSS of other groups. 
 
 It needs to be especially emphasized that the common possession 
 of the correct reading by several MSS is no proof at all that these 
 MSS are members of a group. 
 
 After the genealogical tree of the extant MSS has been plotted, 
 the determination of the reading of the Original in a given passage 
 is usually comparatively simple, especially if more than two inde- 
 pendent lines of descent from the Original copy have been estab- 
 lished. In the latter case, the agreement between all lines of descent 
 but one settles the text. In case, however, each one of three lines 
 of descent has its own peculiar reading, the determination of the 
 original reading is beset with greater difficulty. The three readings 
 must then be carefully examined to see whether one of them may 
 have been based on one of the others. If so, that settles the text. 
 Sometimes, however, the three readings all look equally hke the 
 reading of the Original. In such a case, if one of the three main 
 groups has a smaller total number of errors and deviations than 
 either of the others, that group should be followed here, because, as 
 a matter of probability, it is here less likely to be in error than either 
 of the others. 
 
 A distinction should be made between the Critical Text and the 
 Genealogical Text. The Genealogical Text may contain some 
 errors, as all extant MSS may be derived eventually from a copy 
 of the Author's Original that itself contained some errors. In a 
 few cases our Genealogical Text is not the Critical Text, which 
 must conjecturally go farther back than the Original of all extant 
 MSS. 
 
 394 
 
The A- Version of "Piers the Plowman" 135 
 
 The Genealogical Tree 
 General Survey 
 
 It will probably help the reader to follow the detailed study of 
 main groups, subgroups, and sub-subgroups that must now be 
 undertaken if the principal conclusions are briefly summarized in 
 advance. 
 
 First of all, the Original of all the extant MSS of the A-text of 
 "Piers the Plowman," and naturally of all the hypothetically recon- 
 structed archetypes, was not the Author's Copy. That it was not 
 is shown by the presence in all MSS of two breaks or gaps where 
 extended passages have been omitted;^ by the insertion of a scribal 
 marginal note into the wrong place in the text;^ by the very prob- 
 able omission of one line;' and by the omission of part of one line, 
 making imperfect sense.* 
 
 The fourteen MSS of the A-text fall into two main groups. 
 Vernon (V) and Harleian 875 (H) belong to the first, x. 
 
 Trinity College, Cambridge, R 3.14 (T), Harleian 6041 (H2), 
 Douce 323 (D), University College, Oxford, 45 (U), Rawhnson Poet. 
 137 (R), Trinity College, Dublin, D 4.12 (T2), Ashmole 1468 (A), 
 Harleian 3954 (H3), Digby 145 (Di), the MS belonging to the Duke 
 of Westminster (W), Sir William Ingilby's MS (I), and MS 150 in the 
 Library of Lincoln's Inn (L) belong to the second main group, y. 
 
 y comprises four subgroups; (1) L, (2) I, (3) W and Di, and 
 (4) TH2DURT2AH3. 
 
 The subgroup TH2DURT2AH3 falls into two further subgroups, 
 one containing TH2D throughout, the other containing T2AH3 
 nearly throughout,^ while UR fall with the latter group at the begin- 
 ning of the poem, and with the former group throughout the 
 remainder of the poem. 
 
 1 With 5-105 (106) the account of Envy is left incomplete, and the account of 
 Wrath is omitted, probably at this point. Between 5.227 (235) and 228 (236) are lost 
 some lines containing the close of Sloth's vow and a transitional passage leading up 
 to the line "And jet wile I jclde a5en jif I so muchel haue." (See Modern Philology, 
 III, 359-66.) 
 
 2 The foiu-'Une passage, 7.69a, 6, c, d (7.71-74), giving the names of Piers's wife, 
 daughter, and son, inserted quite erroneously into Piers's remarks about his pilgrimage 
 and his will. 
 
 3 A line about Wrath in the feflfement, passus 2. 
 
 * The Genealogical Text of 4 . 61 is : " For of hise handy dandy payed." 
 5 From 7.69o to 7.209 (71-216), T2 goes with UR, while AH3 form a sub-subgroup 
 of equal genealogical weight with TH2DURT2. 
 
 395 
 
136 
 
 Thomas A. Knott 
 
 Within the sub-subgroup TH2D we have a still further subgroup, 
 TH2. 
 
 U and R form a separate sub-subgroup throughout, both when 
 they go with TH2D and when they belong with T2AH3. 
 
 The B-text is available to settle doubtful questions, as it is 
 derived from a MS of A not belonging to either x or y. The arche- 
 type of B we may call z. 
 
 The family tree of the A-text then is as follows: 
 
 Original (not the Author's) 
 
 V H T H2 D U R T2 A H3 I L W Di B-text 
 
 For the modifications among TH2DURT2AH3 see the subsidiary 
 tables on p. 142. 
 
 MSS V AND H— The Group x 
 
 The readings proving common ancestry for MSS V and H fall 
 into four classes: (1) clear errors; (2) peculiar deviations; (3) cases 
 where y alliterates and x does not; (4) lines omitted by x. 
 
 Belonging to the first class are : 
 
 Prol. 63 : But holy chirche and t'ei holden bet togidere] 
 
 But holy chirche bi-ginne holde bet to-gedere V; 
 
 An but hooly churche bygynne pc bettere to holde to-gedre H. 
 
 1.54: tutour] toure HV. 
 
 1.104: and such seuene ot'ere] an al Ve foure ordres VH(Di). 
 
 2 . 66 (69) : seignourie] seruyse HV. 
 
 3.166 (174): half] nekke VH. 
 
 5.99-100 (100-101): H and V transpose the two second half-lines. 
 
 Other errors occur in 2.30, 73 (76), 97 (102); 3.31 (33), 235 
 (244); 5.128 (129), 163 (164); 8.78 (79), 103 (105). 
 
 396 
 
The A- Version of "Piers the Plowman" 137 
 
 Peculiar deviations are : 
 
 Prol. 53: from of ere] for bretheren, HV. 
 
 1.21: Nam none nedful but fo] Her naames beot* neodeful, HV. 
 5.33: Let no wynnyng for-wanye hem] Let hem wonte non ei3e, VH. 
 7.91 (96): putte, TDiD; pyche, LURT2AIH3; posse, H; posschen, V. 
 Also prol. 76: 1.162; 2.182 (196); 3.15, 143 (151); 4.19, etc. 
 
 Alliteration occurs in y and not in x, presumably by error, in : 
 
 Prol. 14: I sai5 a tour on a toft, tritely I-makid] 
 
 I sauh a Tour on a Toft, wonderliche I-maket, VH. 
 
 Prol. 41: Til here belyes and here bagges were bretful ycrammid] 
 
 Til heor Bagges and heore Balies weren faste I-crommet, VH. 
 
 Also prol. 21, 76; 1.4, 34, 88, etc. 
 
 Lines omitted in x are: 
 Prol. 50-51, 99-100, 109; 2.28-29; 4.119. 
 
 The rest of the readings distinguishing the group VH are : 
 
 Prol. 32, 52, 58. 
 
 1.9, 22, 39, 68, 72, 78, 90, 98, 105, 121, 127, 134, 139, 155, 168. 
 
 2.4, 9, 23, 58 (61), 64 (67), 70 (73), 80 (84), 81 (85), 84 (88), 104 (109), 
 128 (134), 131 (144), 190 (204). 
 
 3.1, 10, 14, 25 (27), 32 (34), 33 (35), 39 (41), 69 (72), 78 (81), 84 (87), 
 114 (122), 117 (125), 144 (152), 175 (183), 191 (199), 206 (214), 212 (220), 
 214 (222), 223 (231), 242 (251), 251 (260). 
 
 4.1, 17, 24, 39, 50, 66, 69, 77, 78, 112, 128, 140, 144. 
 
 5.8, 50, 57 (58), 66 (67), 98 (99), 133 (134), 158' (159), 170 (171), 175 
 (176), 205 (213), 209 (217), 220 (228), 243 (251), 251 (259). 
 
 6.2 (4), 24 (27), 30 (33), 35 (38). 
 
 7.3, 26 (27), 32 (33), 41 (42), 62 (63), 69 (70), 73 (78), 124 (129), 127 
 (132), 148 (153), 179 (186), 206 (213), 221 (228), 243 (250), 247 (254), 252 
 (259), 253 (260), 274 (281), 278 (285), 281 (289), 284 (292), 294 (302), 
 296 (304). 
 
 8.5, 17, 44, 53 (54), 54 (55), 58 (59), 61 (62), 72 (73), 81 (82), 110 (112). 
 
 WD1ILTH2DURT2AH3— The Group y 
 
 All the remaining twelve MSS belong to one other main group, 
 y, though the number of common errors and deviations is small 
 compared to those of x. The small number of common errors in its 
 descendants, however, means only that y was a very good transcript 
 of the Original. 
 
 397 
 
138 Thomas A. Knott 
 
 The errors common to all, or practically all, of the MSS of y are 
 as follows : 
 
 An erroneous omission occurs in 5.152 (153). VH and the 
 B-text read : 
 
 Hastou ou3t I pi pors quod he, eny hote spices ? 
 
 THaDRTaHgWDi omit "ou^t I pi pors." UAIL omit "I pi pors." 
 *'I pi pors" was accidentally omitted in the source of all twelve 
 MSS, while various archetypes and individuals thereupon each 
 intentionally omitted "ou^t," feeling it to be superfluous and 
 meaningless. 
 
 The omission of one line, 5.162 (163), from all the MSS but one 
 small subordinate subgroup is further evidence of common ancestry. 
 The line reads : 
 
 Sire pers of pridye and pernel of Flaundres. 
 
 It is present in the B-text, in VH, and in T2AH3, but is omitted in 
 TH2DURWDiLI. The subordinate position of the little group 
 T2AH3 renders it impossible that the presence of the line in the 
 ancestor of that group represents the tradition from y, and the fact 
 that the archet3T)e of these three MSS was not infrequently con- 
 taminated from the B-text explains the presence of the line in the 
 descendants of that archetype. 
 
 In 5.99 (100) X has "aswagen hit vnnepe." For "vnnepe" 
 y has "an vnche." The reading in y seems to be of the sort more 
 probably derived from that of x by scribal sophistication than vice 
 versa. If this is so, then the reading of the twelve MSS is evidence 
 of the group. In 6.88 (91) VT2H3 (H and A defective) correctly 
 have "ones" at the end, while the MSS of y (except T2H3) have it 
 erroneously at the beginning of line 89 (92). The presence of the 
 correct reading in the minor subgroup T2H3 means nothing but con- 
 tamination from B, or perhaps conjectural restoration in their 
 ancestor. 
 
 In 2.87 (91) X has "hure," while y has "mede." The reading 
 of X alliterates, making the line read: 
 
 WorK is Ipe werkman his hure to haue. 
 
 This alliteration within each half-line is not unknown in the A-text. 
 Cf. 1.1; 3.199 (207). 
 
 398 
 
The A- Version of "Piers the Plowman" 139 
 
 8.62 (63) reads: 
 
 Sipen 5e sen it is so, sewip to Pe beste. 
 "So" is the reading of x and z. y has "]>us" ("sop," W; "this," 
 Di). X alliterates, but y does not. The reading of W is merely- 
 conjectural emendation, for Di, W's sister MS, has a reading obvi- 
 ously based on "thus." 
 
 In 1 . 148 the Critical Text is: 
 
 To hem fat hongide him hy^e & his herte t^irUde. 
 For "hy^e" THgDURWDil read "by." L reads ''on cros." For 
 "him hy^e" A has "on hym." For "hongide him hy^" T2 has 
 "hym hangyd." T2A are a minor subgroup, and their readings 
 are an obvious attempt to avoid the unintelligible "by," by omitting 
 or changing it. The reading of L ("on cros") is quite clearly of the 
 same sort. In view of L's well-known habit of revising lines to gain 
 superfluous alliteration,^ it cannot be asserted that L is here deliber- 
 ately substituting "on cros" for "hyge," that is, removing alliter- 
 ation. The reading might of course be a careless substitution of 
 what amounted to a synonym, but it seems far more likely to be an 
 intentional attempt to give sense to an unintelligible word.^ 
 
 W AND Di 
 On the basis of twenty-one deviations and errors, W and Di 
 must be assigned to the position of one subgroup of y: 
 Prol. 44: For "knaues" W has "hyne," Di "hewyn." 
 1.162: For "wi}:^oute" W has "sanz," Di has "sauns." 
 1.163: For "lewid as a laumpe Pat no ligt is inne," WDi read "lewed a 
 
 ):^ing as a lampe wip outen lyght." 
 2.80 (84): The CT is " sorewe on )^i bokes " (for "bokes"HV have" lockes"). 
 For "bokes" W has "chekes," Di has "bokes chekes" (sic), with both 
 words in the original hand and ink, and with "bokes" crossed out in 
 the original ink. "Chekes" must have been in the archetype of WDi, 
 but the Di scribe had read or copied the poem enough times from 
 some other archetype to have a strong recollection of "bokes," which 
 he at first wrote. Then looking at his copy, he saw that the reading 
 there was "chekes," and he changed his reading accordingly. 
 4.70: WDi both omit "king," though in each MS the word is inserted in 
 a different hand from the original. 
 
 1 See Skeat's account of this MS in the E.E.T.S. A-text, p. xxii. 
 
 2 All MSS of y except UT2A omit "lyk A gleo monnes bicche" in 5.195, and mis- 
 arrange 195-96 (197-98). y omits "hom" in 5.201 (209). 
 
 399 
 
140 Thomas A. Knott 
 
 5.165 (166): For "redyng king" WDi have "redekyng." 
 7.220 (227): Omitted in both MSS. 
 
 Other readings where both agree in a deviation are in 3.137 
 (145); 4.45; 5.9, 125, 145 (146); 6.8 (11); 7.110 (115), 164 (170), 
 192 (199), 232 (239); 8.55 (56), 63 (64). Still other evidence is in 
 5.246 (254) and8.28.i 
 
 The Group TH2DURT2AH3 
 
 Within the group y, the MSS TH2DURT2AH3 constitute a sub- 
 group. In 2.83 (87), the CT reads: "For Mede is moylere of 
 mendes engendrit." 
 
 For "of mendes engendrit" TH2DUR (H3A defective) read 
 "of frendis engendrit"; T2, obviously attempting an emendation 
 on the basis of this, reads "fendes." W omits the line. VH have 
 "a mayden of gode" — clearly from 2.96 (101). 
 
 In 5.240 (248) the CT is: 
 
 H wil worl? vpon me as I haue wel desemid. 
 
 For "worp" TH2UA have "werche," DR have "wirche," T2 has 
 "wirke"; "worth" is in LWDiHsIHV. Here H3 agrees with the 
 MSS outside the group, but is undoubtedly restoring conjecturally, 
 or perhaps has a contaminated reading. That it belongs with the 
 main group is proved by its membership in the sub-subgroup T2AH3. 
 In 3.257 (270) VHLWAI read "kuynde wit." TH2D have 
 "kynde it"; URT2 have "reson it"; Di has "kynde," omitting 
 "wit" (H3 defective). The only difficulty here is the reading of A. 
 As this MS is throughout this part of the poem (1 . 145 — 6.80) closely 
 related to T2, its reading must be due to contamination or conjectural 
 emendation (probably the former). The reading of Di consists of 
 the omission of a word, and is not at all the same as that of the group 
 under discussion. 
 
 J 5 . 246 (254) . The CT reads ' ' not f aire. " W has ' ' no f erl)er , ' ' Di has ' ' no f arder , ' ' 
 T2 has "no f error." 
 
 8.28. The CT has "mysoise." WH have "mesels," Di has "mysselles." When 
 a group of MSS appears as a fixed or constant element in combination with various scat- 
 tered MSS, if the latter are clearly constituted members of other well-estabhshod groups, 
 then the evidence, I take it, tends to argue in favor of common descent for the fixed MSS. 
 For example, if we have such agreements as AB, ABC, ABD, ABE, and if we know that 
 C, D, and E belong to other groups, the evidence confirms the group AB. This is the 
 sort of evidence we have to deal with hero. 
 
 400 
 
The A- Version of "Piers the Plowman" 141 
 
 In 6.94 (97) the CT reads: 
 
 And lere Pe forto loue hym & his lawes holden. 
 TH2DURT2H3 omit "hym," which is in VLWI (though V has the 
 first half-line somewhat changed). (H defective. Di is C-text here. 
 A omits this line.) 
 
 In 5.160 (161) the CT has "nedelere" ("neldere" HV). 
 TH2DURT2 have "myllere" ("mylner" URT2). AH3 must show 
 contamination, for they are here, as elsewhere (5.108 — 7.8), closely 
 connected with T2. (L changes the whole half-Une.) 
 
 In 4.84 lack of alliteration characterizes the subgroup, which 
 reads in the second half-line, "he shal do so nomore." WIA have 
 "wil" for "shal," and LDiHV have "wol." "Wil, wol" alliterates. 
 
 The CT for 7. 112 (117) reads: 
 
 We haue no lymes to laboure wip, lord pankid be gc. 
 
 For "lord pankid be ye," TDRH3 read "lord ygracid be je"; U has 
 "lord ygraced be pe"; T2 has "lord gyff vs grace." A has "lord 
 grace be ^e"; H2 is defective. The CT is determined by VHWLI, 
 which read: "lord ponked be pcrw," W; "lord I thanked be ^e," L; 
 "lord pankyd be pe," I; "vr lord we hit ponken," VH. (Di has 
 "lord I graced be thou," which must be a result of contamination.)^ 
 In 1.153 the CT reads: 
 
 For Pel? 56 be trewe of joure tunge, & treweliche wynne. 
 
 Instead of "For pei^ ^e" TH2DRT2 read "For pi." While U agrees 
 with the other MSS, its position in the subgroup URT2, and in the 
 sub-subgroup UR, shows that its reading here must be a result of 
 contamination or of emendation by a scribe. T2 in fact has such 
 a contamination or conjectural emendation. The original has 
 "For pi," changed in a contemporary hand to "For pof 5e." The 
 reading of A, "Thow ^e," must be due to the same sort of reason. It 
 is a member of the subgroup URT2A, and of the sub-subgroup AT2. 
 In 7.209 (216) TDURT2A (H2 defective) have the first half- 
 line wrongly arranged so as to follow the Latin of the preceding line. 
 TheCTis: 
 
 1 The exact situation here perhaps might be regarded as less certain because of the 
 compUcations furnished by the B- and C-texts. B has "lorde y graced be ,ie"; while 
 C has "lord god we l)onket)" (C 9.135). But the genealogical positions of the MSS 
 attesting "tsonked, fionken," in A render the CT of that version certain. 
 
 401 
 
142 
 
 Thomas A. Knott 
 
 Facite vohis Amicos. 
 
 I wolde not greue god quaP peris for al the gold on ground. 
 
 TDURT2A read: 
 
 Facite vobis Amicos I wolde not greue god 
 QuaP peris for al Pe gold on ground.^ 
 
 H3 by conjecture or contamination has the correct arrangement, 
 but that it belongs to the group is proved by its closeness in many 
 readings throughout here to A. AH3 form a sub-subgroup from 
 7 . 69a to the end. 
 
 The Subgkoups among TH2DURT2AH3 
 For MSS TH2DURT2AH3 three different genealogical trees are 
 
 necessary in the different parts of the poem, as follows: 
 
 I. Prol. 1 to 1 . 183 (185). [The last reading for URT2 is 1 . 167. 
 
 A defective to 1 . 142.] 
 
 H2 
 
 D 
 
 U 
 
 R 
 
 II. 2.1 [first reading for TH2DUR is 2.163 (176)] to 7.69 (70); 
 7.210 (217) to 8.126 (130). [H3 becomes A-text at 5.106 (107); 
 T2 defective from 7 . 210 to the end.] 
 
 T H2 D U R 
 
 III. 7.69a (71) to 7.209 (216). 
 
 H3 
 
 Ho 
 
 D 
 
 U 
 
 R 
 
 Ha 
 
 > URT2 omit "I." UT2 omit " Qua^ .... groimd." For "Qua]) .... ground" 
 T has: "Qua^ peris for al lie gold on Ms ground"; D has "Quod piers for al ^e gold Irnt 
 grouel) on grounde"; R has "for al I)e gold quod peris I)at growet) on lie grounde"; 
 A has " Quot peris for alle l>e gold on lie groimde." 
 
 402 
 
The A- Version of "Piers the Plowman" 143 
 
 It is most important to note the general principle of textual 
 criticism according to which we require these three different trees 
 for the subgroup comprising our eight MSS. Briefly, this principle 
 depends on the fact that different parts of the same MS were some- 
 times copied from different ancestors. MS H3, for instance, is of the 
 B-text to B 5.128, and of the A-text from A 5.106 (107) to 8.111 
 (113). In a similar manner, the common ancestor of UR was 
 copied in the first part of the poem from a transcript of the ancestor 
 of T2A. But from about the beginning of passus 2 to at least 8 . 126 
 (130) the immediate source of UR was a MS which was a sister of 
 the ancestor of TH2D. From 7.69 (70) to 7.209 (216), moreover, 
 T2 was copied from the ancestor of UR, which still belongs with the 
 group TH2D, while AH3, still belonging to one subgroup attested 
 by numerous readings, go back to an ancestor which was a sister to 
 the archetype of TH2DURT2. 
 
 TH2D 
 
 Of the sub-subgroups in this subgroup, we may first discuss 
 TH2D. Their common errors and deviations run consistently 
 throughout the poem, and are many and important. Some of the 
 most significant errors are: 
 
 3.206 (214): "mede," "nede," TH2D. 
 
 7.171 (177): "an hepe"; TD have "In helpe" (H2 defective). 
 
 8.70 (71): "defraudeth"; H2D have "Gyleth"; T has "kiUp." 
 
 In 3.82 (85) TH2D omit "meires and." 
 
 TH2D omit 3.100 (108) and 7.174 (180-81) (H2 defective). 
 
 Deviations clear and important appear in 3.169 (177); 4.24, 
 148; 5.16, 233 (241); 6.6 (9), 82 (85), 106 (109); 7.168 (174). 
 
 Other deviations, not quite so convincing individually, but in 
 their total supporting the group weightily, are: 1.104; 2.82 (86), 
 123 (129); 4.58, 106, 145, 153; 5.41, 89-91 (90-92), 182 (184), 
 237 (245); 6.29 (32); 7.160 (166), 116 (121), 218 (225), 262 (269), 
 
 112 (117), 140 (145), 192 (199), 302 (310); 8.26, 46 (47), 61 (62), 
 
 113 (115). 
 
 TH2 
 
 Of the group TH2D, T and H2 form a subgroup. Clear errors 
 occur in: 3.71 (74), "richen," "risen," TH2; 5.252 (260), "po 
 
 403 
 
144 Thomas A. Knott 
 
 prongen," "pe wrong," THa; 8.10, "rijtfulliche," "rewfuUiche,'' 
 TH2. 
 
 Common deviations supporting the grouping are in 1.159; 
 2.144 (157); 5.7, 17, 197 (199), 163 (164); 8.125 (129). 
 
 Other deviations, some of them weighty, and in the sum total 
 constituting conclusive evidence, are in 1.59, 72, 110, 135, 138, 171; 
 2.4, 7; 3.90 (97), 107 (115), 116 (124), 210 (218), 239 (248), 255 
 (264); 4.48, 73, 119, 129; 5.29, 56 (57), 57 (58), 182 (184), 215 
 (223), 251 (259), 254 (262); 6.6 (9), 53 (56), 67 (70), 104 (107); 
 7.35 (36), 80 (85); 8.7, 32, 45, 61 (62), 118 (120). 
 
 URT2 FROM Prologue 1 to about 1.183 (185) and from 7.69a (71) to 
 
 7.209 (216) 
 
 The evidence grouping URT2 from the beginning of the poem to 
 about 1.183 (185), and from 7.69a (71) to 7.209 (216) [T2 defective 
 from 7.210 on], appears quite conclusive. Some of the most impor- 
 tant readings are: Prol. 71, "bunchide," "blessid," URT2; 1.98, 
 "professioun," "prophecye," RT2 (U defective); 7.91 (96), "pote," 
 URT2 omit; 7.147 (152), "ordre," "lord," URT2; 7.173 (179), 
 "bedrede," "Were ey^ed," URT2; 7.173, "botnid," "aboute," 
 URT2. 
 
 Other readings are in Prol. 1, 13, 14, 17, 21, 29, 32, 37, 77; 1.23, 
 52; 7.76 (81), 98 (103), 99 (104), 104 (109), 116 (121), 127 (132), 
 159 (165), 164 (170), 169 (175), 172 (178), 191 (198), 197 (204), 
 206 (213), 209 (216). 
 
 After Prol. 54, R adds two lines, and at the same point T2 adds 
 four, including the same two : 
 
 . on fele halue fonden hem to done 
 lederes Pei be of louedayes and with Pe la we medle. (R) 
 
 Parsons with P&ir proourases [prouisours ?] permutyn pair chirches 
 
 With al pe besynes of Pair body Pe better to haue 
 
 Vicars on fele halue fandyn paim to Done 
 
 Leders Pai ben of lovedays & with Pe lawe mellyth. (ToY 
 
 » U not only does not contain any of these four lines, but also omits line 54. From 
 the fact that in R the first word of the first added line is omitted, and a blank space left 
 for it, it has been argued by Skeat and Chambers that the word must have been illegible 
 in the ancestor of U and R. and that U must have omitted both lines because of the 
 resultant unintelligibility. The fact, however, that U also omits hue 54 seems to me 
 to point to a piu-ely accidental omission of all three lines, rather than to an intentional 
 omission of the two spm-ious lines on account of the obscurity of one word in the first of 
 
 404 
 
The A- Version of "Piers the Plowman" 145 
 
 Further evidence for the relation of URT2 is to be found in the 
 transposition, common to all three, of 7.69a-209 (71-216) to a posi- 
 tion immediately preceding 1.180 (182).^ This dislocation is due 
 to the accidental transference, in the archetype of URT2, of the 
 inside leaf (two folios, or four pages) of the third quire of four leaves 
 into the middle of the first quire, also of four leaves.^ 
 
 The Group TH2DUR from 2.1 to 7.69 (70) 
 
 MSS TH2DUR form a sub-subgroup from about 2. 1 to 7.69 (70), 
 and MSS TH2DURT2 form a sub-subgroup from that point to the 
 end of the poem (8.126). 
 
 In 2.163 (176) LWDilATaDVH read "And gurdeth of gyles 
 hed." TH2UR quite erroneously read "gederip" for "gurdeth." 
 Though D agrees with the other MSS outside the group, its reading 
 
 them. Scribal habit was to "edit" an obscure or semi-obliterated word into an intelli- 
 gible word. We have enough independent deviations in U to prove that U had this 
 common habit of editing. It is certainly hard to imagine a scribe who would inten- 
 tionally omit two lines for such a reason, or, indeed, for any reason. 
 
 1 Although the dislocated passage in all three MSS precedes the same line (1.180), 
 however, the situation in two MSS (U and T2) is involved in some difficulty, which has 
 never been adequately noticed or discussed. In U the line preceding the shifted matter 
 is not 1 . 179 (181), as we should expect, but 2 . 23. That is, 1 . 180 to 2 . 23 is given twice, 
 once before, and again after, the transposed passage. In T2 the transposed passage 
 follows not 1 . 179, but 1 . 182. Thus in this MS three lines (1 . 180-82) are repeated. 
 In R the line preceding the dislocation is correctly 1.179. The latter MS undoubtedly 
 represents the original condition of the archetype of URT2. But how account for the 
 repetitions in the other two MSS ? Skeat, who discusses the problem in MS U, believed 
 that U had been copied from two MSS at this point (E.E.T.S. A-text, p. xx). But 
 against this it must be urged that U and the repeated fragment (called U by Skeat) 
 have several peculiar deviations and errors in common, thus postulating a single arche- 
 type. The most probable explanation seems to me to be that in an immediately pre- 
 ceding ancestor of each MS, independently, someone noticed the dislocation between 
 1.179 and the adjacent 7.69o. Then, either remembering (from some previous famil- 
 iarity with the poem), or discovering that there were some lines about four pages later 
 that fitted in after 1. 179, the possessor of the MS in each case began copying the cor- 
 rectly following lines into the proper place in the margin, or, perhaps, in the case of U 
 (as the repetition is so long) , on a bit of inserted parchment. In the case of the ancestor 
 of T2, the corrector stopped after three hues, the end — practically — of a logical speech 
 and but one line short of the end of the passus. In the case of U, the corrector kept on 
 into passus 2, for some reason, perhaps because his bit of inserted parchment was just 
 large enough, to line 23. 
 
 2 If we calculate about 36 lines to a page, the archetype contained just enough lines 
 to fill the foiu" folios preceding the point of incorrect insertion. At 36 lines per page the 
 dislocated passage would fill four pages, or two folios. At 36 lines per page, the passage 
 between the insertion and 7.69a would fill fifteen folios, that is, it would fill the second 
 four folios of the first quire, the eight of the second, and the three of the tliird preceding 
 the point of removal of the dislocated passage. The quires must have been left unsewed 
 while being copied. The inside leaf of the third must have fallen out, and then must 
 have been stuffed back into the center of the first, instead of the third, quire, and then 
 must have been sewed there. Afterward the copies were made, perpetuating the error. 
 
 405 
 
146 Thomas A. Knott 
 
 must represent contamination or conjecture, as this MS is closely 
 connected with TH2. 
 
 In 3.83 (86) the other MSS have "And told hem pis teeme" 
 (except H, which has "lo f>is was his teme"), but TH2DUR have 
 "And tok hym pis teeme." 
 
 In 4.19 the CT is "wytful gerpis," the reading of LWDiAID 
 ("full wyght girthes," T2; "swipe fele gurpis," HV). TH2UR 
 read "ri^tful gerpis." D again disagrees with the group, but 
 through contamination or conjecture. 
 
 3 . 240 (249) is omitted in TH2DU. It must have been restored 
 in R by contamination from a MS outside the group. 
 
 In 3.137-8 (145-6) the CT has "to holde" correctly at the 
 beginning of hne 138, while TH2DUR have the phrase incorrectly at 
 the end of line 137. 
 
 In 4. 113 the CT has "graue wip kynges coroun," but TH2DUR 
 have incorrectly "ygraue wip kinges coyn." 
 
 In 5.43 the words "ran" and "and" are omitted in TH2DUR. 
 
 Other agreements, most of them striking, are in 3.99 (107), 
 113 (121), 119 (127), 130 (138), 254 (263), 266 (275); 4.4, 154; 
 5.16, 17, 34, 71, (72), 90 (91), 92 (93), 94 (95), 108-9 (109-10), 113 
 (114), 130 (131), 215 (223); 6.120 (123); 7.18, 31 (32), 29 (30). 
 
 Several agreements group TH2DURT2 from 7.69a (71) to 7.209 
 (216). In 7.72 (77) TDURT2 (H2 defective) omit "pis." In 
 7.75 (80) TDURT2 (H2 defective) add "For" at the beginning of 
 the line. In 7.140 (145) TDUR (H2 defective) add "awey" (T2 
 omits the line) . In 7 . 181 (188) TDUR (H2 defective) have " asserue" 
 for " deserue " ; T2 has " serue." In 7 . 139-40 (144-45) TH2DURT2H 
 have "of pi flour" incorrectly at the end of line 139. (V has it in 
 the middle of the line, considerably changing the rest of the Une, 
 as does also H.) AHsLWDil have the phrase correctly at the begin- 
 ning of line 140. The error must have occurred in VH independently 
 
 of TH2DURT2. 
 
 In 7.161 (167) the CT is "he jede hem betwene" ("he jed 
 hem," H3I] "5ede hem," H; "he wente hem," A; "he 5ede so," W; 
 "had hyhyd," Di; "busked heom," L; "I bot hem," V); TDUR 
 read "he hadde"; T2 reads "pai abade," which seems to be a cor- 
 ruption based on the reading of TDUR. (T2 omits "betwene.") 
 
 406 
 
The A-Version of "Piers the Plowman" 147 
 
 TDURT2 omit 7.207 (214) (H2 defective). 
 
 DURT2 (H2 defective) misarrange 7.204-9 (211-16) similarly. 
 D, with which the other three substantially agree, has them as 
 follows : 
 
 And alle maner men t^at t'ou myjt aspyen Pat nedy ben or naked 
 And nou5t haue to spende with mete or with mone 
 late make Pe frendes }^er with & so Matheu vs teches 
 Facite vobis amicos I wold not god greue. 
 
 T, which had the line similarly misarranged in its "copy," has 
 arranged them more nearly correctly, but has had to supply a con- 
 jectural second half -line for line 206 (213). T reads: 
 
 And alle maner of men fat t^ou mijte aspien 
 
 Pat nedy ben or nakid & nou5t han to spende 
 
 Wip mete or mone let make hem at ese 
 
 And make Pe Frendis Per mip for so matheu vs techip. 
 
 The CT for these lines reads : 
 
 And alle maner of men Pat Pou mi^te aspien 
 Pat nedy ben or nakid & nou3t han to spende 
 Wip mete or wip mone let hem be Pe betere^ 
 Or wip werk or wiP word whiles Pou art here 
 Make Pe Frendis Per wip and so matheu vs techip. 
 
 The Minor Group UR 
 
 The citations proving close connection between U and R are 
 probably more numerous and convincing than for any other group, 
 except perhaps VH. 
 
 The clearest errors, some of them mere absurd blunders, occur 
 in Prol. 85, where for "seruide" UR have "pletiden"; in 2.42 (44) 
 for "teldit" UR have "tight"; in 2.168 (181) for "preyour" UR 
 have " tresour " ; in 3 . 74 (77) for " burgages " UR have " bargaynes " ; 
 in 3.169 (177) for "menske" UR have "mylde"; in 2.121 (127) 
 for "ioye" UR have "lawe"; in 3.183 (191) for "mournyng to 
 leue" UR have "fro morwe til eue"; in 5.60 (61) for "in pe palesie" 
 U has "palatik," R has "paltyk"; in 5.131 (132) for "aunsel dede" 
 UR have "almesdede"; in 6.29 (32) for "to so wen and to setten" 
 UR have "now and sithe"; in 7.30 (31) for "wastours" UR have 
 "watris." 
 
 1 For "let hem be I)e betere," the reading of LWDiAHal, V has "mak hem fare te 
 betere"; H has "lete hem fare pe better." 
 
 407 
 
148 Thomas A. Knott 
 
 Both MSS omit lines 2.11, 2.24-25, 5.220 (228), 6.108 (111). 
 
 Other common deviations are to be found in Prol. 86, 102, 106; 
 1.1, 130, 152; 2.45 (47), 53 (56), 54 (57), 58 (61), 77 (81), 156 (169); 
 3.1, 44 (46), 62 (64), 120 (128), 164 (172), 174 (182), 175 (183), 
 231 (240), 259 (266), 267 (276); 4.24, 30, 60, 61, 131; 5.37, 57 (58), 
 87 (88), 91 (92), 92 (93), 113 (114), 117 (118), 142 (143), 167 (168), 
 177 (178), 179 (180), 184 (186), 222 (230) 224 (232), 230 (238), 
 252 (260); 6.35 (38), 52 (55), 55 (58), 67 (70), 97 (100), 98 (101), 
 103 (106), 104 (107); 7.10, 29 (30), 39 (40), 52 (53), 66 (67), 119 
 (124), 140 (145), 213 (220), 233 (240), 238 (245), 279 (287), 283 
 (291); 8.21,84 (85), 118(120). 
 
 The Minor Group T2AH3 
 
 The evidence for grouping T2 and A from 1 . 143, where A begins, 
 to 5.105 (106), to which H3 is B-text, includes: 1.145: T2A omit 
 "pite," though in T2 a different hand, in a blacker ink, has inserted 
 it after "peple"; 2.9: for "I-purfilid" T2 has "puryd," and A has 
 " I purid." In 3 . 87 (90) for '' in ^oupe or in elde" T2 has "in thoght 
 or in dede," and A has "in pouth or indede." In 4.38 for "gade- 
 lynges" T2A have "goslynges." Other readings supporting the 
 group are in 1.151, 152, 157, 180 (182); 2.5-6, 16, 148 (161); 3.11, 
 21 (23), 259 (266), 270 (279); 4.24, 42, 50-51, 58, 47, 67, 82, 98, 
 100, 129, 130, 147; 5.23, 31, 32, 77 (78), 78 (79). 
 
 Some of the strongest evidence for grouping T2AH3 from about 
 5.106 (107) to about 7.69 (70) is: In 5.145 (146) instead of "forto 
 go to shrift," T2 reads "to gang on hy way" (sic), A has "to gon 
 his wey," and H3 has "to gon on hys weyje." After 6.81 (84) 
 T2AH3 add three lines, the first unique, the second and third from 
 the B- or C-text. Other readings are in: 5.108-9 (109-10), 114 
 (115), 115 (116), 129 (130), 136 (137), 141 (142), 146 (147), 158 (159), 
 189 (191), 206 (214), 216 (224), 242 (250); 6.1 (3), 28 (31), 58 (61). 
 
 In passus 7, after T2 has become defective at line 209 (216), 
 some of the evidence for grouping AH3 is: In 7.218 (225) WDiH 
 have "mouthed," URI have "mouthith," L has "techeth," T has 
 "nempnip," D has "nemened," V has "Mommcp," but AH3 read 
 "mevith, meuyth." Other readings occur in: 7.219 (226), 239 
 (246), 266 (273); 8.18, 21, 98 (99), 99 (100). 
 
 408 
 
The A- Version of ''Piers the Plowman" 149 
 
 The evidence for the minor subgroup ToA within the subgroup 
 T2AH3, after H3 has become A-text, includes: In 5.163 (164) for 
 "dykere" T2A read "Drinker" (VH read " disschere ") ; T2A omit 
 line 5.165 (166). In 5.189 (191) for "ygulpid" T2A read "gobbyd." 
 (H3 has "I clobbyd," H2D have "gluppid," HV have "ygloppid.") 
 In 5.248 (256) for "gilt" T2A read "coulpe." In 6.2 (4) for "ouer 
 valeis" T2A read "oure bankes." In 5.125 (126) for "a pakke 
 nedle" T2 reads "bat nedyls," A reads "abatnedil" (H3 has "a 
 betyngnedyl"). 
 
 The group AH3 subsequent to 7.69 (70) is attested by the 
 following: In 7.82 (87) for "mynde" AH3 read "messe." In 7. 138 
 (143) for "pilide" H3 reads "pynyd," A reads "foule pyne" (T2 has 
 "pelyd"). Other evidence, just as strong, is in 7.172 (178), 174 
 (180-81), 183 (190), 189 (196). 
 
 The Third Main Group — z 
 
 A careful collation of the B-text, so far as the CT of B can be 
 safely determined from the variant readings of the E.E.T.S. edition, 
 has shown that B contains none of the errors and omissions of x, and 
 none of those belonging to y or any of the subgroups of y. This 
 leads us to the obvious conclusion that z — that MS of A which 
 B used as the basis of his recension — must have been derived from 
 the Original in a line of descent independent of either x or y. Con- 
 sequently, whenever x and y differ, but when neither is clearly in 
 error, we have the independent evidence of B to help us in deter- 
 mining the reading of the Original, for when two independent lines 
 of MS descent agree against a third, the agreement of the two must 
 determine the critical reading. In spite of the large number of indi- 
 vidual changes introduced by B, and in spite of the number of cor- 
 ruptions and errors in its A-text original, 2, which can be discovered 
 because the MSS of A generally agree unanimously, or nearly so, 
 whenever B de\'iates, B is thus of the greatest value to the student 
 of the A-text. It is only when we have three different readings, one 
 in X, one in y, and another in z, and when neither x nor y is obviously 
 correct, that we are without reliable genealogical evidence of the 
 reading of the Original of A. In cases like this, we are logically 
 obliged to follow the readmg of that group which is less often in 
 
 409 
 
150 Thomas A. Knott 
 
 error whenever error can be determined. As y furnishes a much 
 better tradition than x, we must therefore rely upon y in cases of 
 this sort. The CT, of course, can never safely adopt the reading 
 of B alone, however tempting that reading may appear.^ 
 
 Chambers and Grattan on the Critical Text 
 
 Students who compare this account of the genealogical relations 
 of the MSS of the A-version of "Piers the Plowman" with that 
 given several years ago by R. W. Chambers and J. H. G. Grattan,^ 
 and who have read Mr. Chambers' later paper^ in which he discusses 
 some matters of the text, will note some wide discrepancies between 
 their results and mine. Nine MSS (LIWDiDURAHs) are either not 
 located in their genealogical tree or in whole or in part are located 
 elsewhere than in mine. Such different results cannot be due to 
 mere difference in opinion. How then are they to be accounted for ? 
 
 First, the method employed by these students has been at fault; 
 secondly, they have stated their opinions before they have had the 
 necessary material in hand to formulate sound opinions; and, thirdly, 
 they have not collected the evidence afforded by MS readings which 
 were perfectly accessible. 
 
 The most serious fault in their presentation is that they cite 
 almost no specific evidence whatever for their classification of the 
 MSS. They cite none of the errors and deviations, either by quota- 
 tion or line number, which they have made the basis of their classi- 
 fication. Consequently, other students who would like to know 
 what Chambers and Grattan regard as errors, significant or insig- 
 nificant, are left absolutely in the dark.* 
 
 In view of their subsequent erroneous location of several MSS, 
 one would say that the establishment of their group Tau (which 
 comprises part of my group y) is a matter of the greatest importance. 
 
 1 It is well to call attention to the insecurity of the text of B. Skeat has collated 
 only six of the fifteen MSS of that text, and our information may therefore sometimes be 
 inadequate to settle the critical reading. It will require some years, however, to collate 
 all the other MSS of B, and in the meantime we shall have to rely on the tentative text 
 ascertainable from our incomplete materials. It may ultimately, therefore, bo necessary 
 to revise a few of our readings of A, which are sound only in so far as we can now determine 
 the CT of B. 
 
 J '"pjjg Text of 'Piers Plowman,' " Modern Language Review, IV, 357-89. 
 
 3 "The Original Form of the A-Text of 'Piers Plowman," " ibid., VI, 302-23. 
 
 « Soo Modern Language Review, IV, 372, 380, 382. 
 
 410 
 
The A- Version of "Piers the Plowman" 151 
 
 Yet for their grounds we are merely referred to Skeat's E.E.T.S. 
 footnotes: "There is no necessity to argue, what has been recognized 
 by all students of the subject, that V and H form one group, and 
 T and U another. If anyone wishes to satisfy himself of this afresh, 
 five minutes' study of Skeat's footnotes, taken at random anywhere, 
 would prove it."^ 
 
 Five minutes of study at random has failed to reveal the whole 
 truth about T and U to Chambers and Grattan. If Skeat's cited 
 MSS were the only ones extant, then V and H would form one main 
 group, and T and U would be the main representatives of the other 
 main group. But many more MSS are extant. And attentive 
 study shows that T and U are not primary representatives of a main 
 group, but are merely members of a subgroup of that main group. 
 The latter fact is of the most vital importance when the text critic 
 attempts to locate other MSS in his tree, and when he begins to use 
 his tree to establish the CT. Chambers and Grattan derive MSS 
 L and I, for example, immediately from the Original, because L and 
 I do not seem to them to possess the most striking errors in T and U. 
 But the most frequent and most striking errors common to T and 
 U are due to several intervening layers of MSS between TU and y, 
 while, as we have seen, L and I are descended from y in lines separate 
 from the TU line. 
 
 Still another imperfection in Chambers and Grattan's method 
 is disclosed in their method of classifying MSS W and Di. They 
 assign these two MSS to a subgroup along with T and H2 because 
 all four add the C-text, from C 12.297 on, to the end of the eleventh 
 passus of the A-text. This grouping is made by Chambers and 
 Grattan in ignorance of what would have been immediately dis- 
 closed by a line-by-line collation of T, H2, W, and Di, namely, that 
 W and Di throughout their A-text parts, not only are not members 
 of the little sub-subgroup TH2, but are not even members of the 
 much larger group TH2DURT2AH3. W and Di are descended, as 
 I have shown, from y in a line independent of any other subgroup 
 of y. 
 
 There are many other instances of faulty method, but I select 
 only a few. For example, Chambers and Grattan assume that 
 
 I Ibid., IV, 373. 
 
 411 
 
152 
 
 Thomas A. Knott 
 
 MS D occupies a "middle" position between TH2 and RUTg. 
 "Here and there, though rarely, D will enable us to get a better 
 reading than either T or RU supply; but D's chief function will be 
 to decide the balance between the readings of T and of RU, where 
 these differ without a clear advantage on either side."^ Their tree 
 for the MSS would be as follows: 
 
 Ho 
 
 D 
 
 U 
 
 R 
 
 T, 
 
 But we have seen that D is grouped with TH2 by thirty-six errors 
 and deviations. The correct tree therefore must be : 
 
 T H2 D U 
 
 While, through the middle of the poem it is : 
 
 R 
 
 H2 
 
 D 
 
 U 
 
 R 
 
 T2 (AHs) 
 
 According to this correct tree, whenever D at a given point agrees 
 with URT2 (AH3) that agreement establishes the CT for the whole 
 subgroup. Before 2.1, and after 7.70 (75), the two readings, 
 TH2D and URT2, are of equal genealogical weight, and the choice 
 between them must be made on the basis of readings outside the 
 subgroup. From 2.1 to 7.69, however, the reading of TH2D is 
 subordinate genealogically to that of URT2 (AH3), and the latter 
 three (or five) establish the reading of the archetype common to all 
 six (or eight). 
 
 Again, by their own admission. Chambers and Grattan seriously 
 disturb one's confidence in their ability to distinguish between 
 
 1 Modern Language Review, IV, 379. 
 
 412 
 
The A- Version of "Piers the Plowman" 153 
 
 inferior and superior readings. They even admit that they cannot 
 always observe when a reading which they suppose to be "inferior" 
 is peculiar to a group, and when it is actually attested in the CT by 
 its presence in half of the MSS of the other main group. In dis- 
 cussing the position of L, for example, they say that they "have 
 judged TU inferior to VH" in twenty-two instances.^ They find 
 that L agrees with VH in fifteen cases, is wanting in one, and agrees 
 with TU in six. Then they scrutinize these six, and discover that 
 they are not really inferior after all. Furthermore, they point out 
 themselves that in two of these six the reading which they have 
 adjudged inferior is not pecuhar to TU alone, but is supported by 
 H of the other main group. The readings which they beheved 
 inferior were the critically attested readings! 
 
 But they practically destroy whatever confidence one has left 
 in their judgment when they come to the discussion of the position 
 of I. They find that I agrees with TU in eleven of the twenty-two 
 readings in which they beheve TU to be inferior, the eleven including 
 the above-mentioned six doubtful, which they again dismiss as 
 "inconclusive." Then they say: "The five remaining cases are 
 not very conclusive either." So that just one-half their "inferior" 
 readings seem inconclusively inferior when Chambers and Grattan 
 wish to be rid of them. Here again the reader wishes that citations 
 of these readings had been given, so that some means of testing 
 these curious results might be available. 
 
 A case of loose thinking or loose phrasing appears on p. 381: 
 "Further, there are passages where a very early corruption has crept 
 in, which is common to both the VH group and the TU group [italics 
 mine]. Here L sometimes shows a reading superior to that of either 
 group. An example is the line referred to above, p. 368 [2.83 (87)]: 
 
 For Mede is moylere of Amendes engendred. 
 Let us see the "corruption .... which is common to both the 
 VH group and the TU group." THsDURTa^ read: 
 For mede is molere^ of frendis* engendrit. 
 
 ^ Ibid.. IV, 381. 
 
 - A defective here; Hs is B-text at this point; W omits the line. 
 ' "molere] medlere," H2; "mulyer," D; "muliere," URT2; "moilloMJ*," L; "a 
 mewliere," Di; "moylere," I. 
 
 « "frendis] fendes," Ts; "frendis of frendis," U (.sic). 
 
 413 
 
154 Thomas A. Knott 
 
 VHread: 
 
 For Meede is a luwelere* A Mayden of goode. 
 
 LDil read : 
 
 For mede is moillour^ of mendes engendred. 
 
 One may legitimately ask, Where is the corruption common to 
 both the VH group and the TU group ? For there is no connection 
 whatever between the VH reading and the TU reading. The former 
 is only the second half -line of 2.96 (101), while the latter is obviously 
 descended from an entirely different error, or a gratuitous scribal 
 emendation, in the parent MS of the TU group. There is no "com- 
 mon corruption" in the two groups. If there were, and if the cor- 
 ruption were not present in LDil, the consequences upon the tree 
 would be very great. 
 
 The reason for many of the faults which we have seen is not 
 far to seek. The study was printed before all the evidence wajS 
 adequately examined, or even collected. In their first paper the 
 authors say: "Many of the above suggestions are put forth only 
 tentatively; for we have not yet had time to sift thoroughly our 
 transcripts and collations."' Over two years later Chambers says: 
 "I have not yet collated W up to this point." [The end of 
 passus ll.p 
 
 Finally we may mention some of the conclusions which Chambers 
 and Grattan would never have reached if they had collected and 
 examined all the evidence. 
 
 Because U and R are clearly to be grouped with T2 (their E) 
 in the early part of the poem, the three MSS are indiscriminately 
 grouped together throughout as one subgroup of y (their Tau). 
 But, as I have shown above, for over five passus (2. 1 — 7.69) U and 
 R belong in a subgroup with TH2D, while elsewhere UR belong 
 in a subgroup with T2AH3. 
 
 The Ashmole MS is dismissed in less than two lines: "Ashmole 
 1468 combines all possible faults. It is imperfect, corrupt, and 
 contaminated by B- or C-influence."^ 
 
 Regarding H3, Mr. Chambers writes: "Mr. Grattan and myself 
 have so far been unable to trace any special affinities of Harleian 
 
 1 "luwcler] mcdoler," H. ' Modern Language Review, IV, 383. 
 
 2 See note 3, p. 153. « Ibid.. VI. 313. ' Ihid., IV, 383. 
 
 414 
 
The A- Version of "Piers the Plowman" 155 
 
 3954 to any other MS or group of MSS. It is therefore an inde- 
 pendent witness," etc.^ 
 
 But as I have shown above, H3, A, and T2 are bound together 
 into a minor subgroup by a very large number of common errors 
 and variations; and most of the contaminations from the B-text 
 in MS A are in the source of all three MSS. 
 
 It is especially important to consider how Chambers and Grattan 
 have dealt with MSS L and I. They wish, if possible, to discover 
 a MS which does not belong to either of their two main groups, for 
 they wish to use this independent MS to determine the CT when- 
 ever their main groups differ, but when neither one is clearly in error. 
 Such a MS they believe they have found in L. In discussing the 
 genealogical position of this MS,^ they point out that it usually 
 agrees with TU when VH are in error, and with VH when TUD 
 seem to them to have readings "inferior to VH." 
 
 But L cannot be excluded from group y merely on the ground that 
 it usually seems to have the correct reading when TUD appear to 
 be wrong. For I have shown that TUD are three members of one 
 subgroup of y, and for over five passus members of a sub-subgroup, 
 and that WDi and I comprise two independent subgroups of y. 
 Hence the important question is. Does L invariably contain the 
 correct reading when WDi, I, and TH2DURT2AH3 all in common 
 contain a wrong reading ? I have shown that in a number of cases 
 all these MSS, including L, have an incorrect reading or an erroneous 
 omission. As L contains no significant deviations in common with 
 any of the other three subgroups of y, aside from those possessed by 
 all, it therefore constitutes a fourth subgroup of y. 
 
 In MS I Chambers and Grattan believe they have found another 
 independent line of transmission from the Original, with perhaps a 
 few deviations in common with x, and a few in common with y.^ 
 I have shown, however, that I is a member of y, though not of the 
 subgroup TH2DURT2AH3. 
 
 The essential unsoundness of any critical text based on the 
 assumption that L and I are descended from the Original in a line 
 of descent independent of any other group of MSS is so obvious as 
 to require no comment. 
 
 1 Ibid., VI, 312. 2 Ibid., IV, 380 fl. « Ibid., IV, 382-83. 
 
 415 
 
156 Thomas A. Knott 
 
 Determining the Critical Text 
 
 In this last part of my study I shall discuss several concrete 
 problems of the sort that are encountered in the actual construction 
 of the critical text, after all the preliminary work of determining the 
 genealogical tree has been completed, and shall try to show how the 
 tree is used in deciding the critical readings. 
 
 It would take too long for me to speculate fully on all the psy- 
 chological and mechanical reasons for the various sorts of errors, 
 but one or two remarks may be offered. The "average" mediaeval 
 scribe utterly lacked the modern typographical compositor's ideal 
 of conscious fidelity to his "copy." If he was careless or stupid, he 
 introduced several kinds of misreadings into his copy, or omitted lines 
 or words. If, on the other hand, he was a careful or critical reader 
 of his "copy," he was likely to change the sense if he thought it 
 could be improved, thus indulging in what we now call "conjectural 
 emendation," or even editorial rewriting. Sometimes the possessor 
 of a MS compared it with another copy of the same work, and, 
 noticing differences between the two copies, scratched out the words 
 of one MS and substituted those of the other, or added lines not 
 in his MS. We have a great many cases of this in MSS H and H2, 
 and sporadic cases of it in a number of other MSS of the A-text. 
 A later copy of a MS which had thus been "corrected" would nat- 
 urally reproduce only the "revised" readings, and the modern text 
 editor would perhaps encounter considerable difficulty in placing 
 such a contaminated MS in his tree. Some contaminations got in 
 unconsciously because the scribe was previously familiar with the 
 work through copying it or reading it in a MS belonging to some 
 other family branch. While carrying a line in his mind between 
 reading it in his "copy" and writing it down, he might unconsciously 
 substitute a formerly read or written term for the one in his "copy." 
 The substitution of inferior or non-alliterating readings, often 
 synonyms, must have been an unconscious process. The scribe 
 merely reproduced the meaning of the line substantially, without 
 caring for exactness. And there must have been many shades and 
 sorts of errors between the conscious emendations and contaminations 
 and the unconscious substitutions. 
 
 416 
 
The A-Version of "Piers the Plowman" 157 
 
 Then there are errors due to mis-seeing, or to mishearing, or to 
 purely mechanical miswriting. One letter may be misread for 
 another with a similar shape, "h" and "b," "t" and "c," "n" 
 and "u," "e" and "o," "b" and "1" are pairs of letters one of 
 which might be easily misread or miswritten for the other. Words 
 such as "lene," "leue," "loue" might thus be substituted for one 
 another. Or "hye," "by," or "ac," "at," or "beste," "leste" 
 might be confused. All these are variants to be found in our MSS. 
 
 The so-called "errors of mishearing" might occur in one of two 
 ways. "Copy" may have been read aloud by one scribe and written 
 by another or others, though there is little positive evidence that 
 this method was practiced in the Middle Ages. These errors seem 
 to me to have much more probably occurred in the work of scribes 
 who belonged to what psychologists call the "auditory type" — 
 individuals who remember in auditory images. Such persons most 
 naturally read aloud, or imagine vividly that they read aloud, mate- 
 rial that they wish to copy. My theory is that most errors of this 
 type in mediaeval MSS occurred in this way, and not through a 
 mishearing of what was being read aloud by another. 
 
 We are now ready to discuss some concrete problems. 
 
 The easiest sort of error to eliminate is the single reading of 
 
 T (our base), when all the other MSS agree against it. One case 
 
 occurs in 1 . 49 : 
 
 Cesar fey seide we se wel ichone. 
 
 "pey] panne," T; "pey rest." 
 
 The reading peculiar to one sub-subgroup is the next to the 
 easiest to eliminate. One occurs in 5. 16: 
 
 Prries and plomtrees wern puffed to Pe erpe. 
 
 "plomtrees] plantes," TH2D. 
 
 Even the reading of a whole main group must be held to be of 
 
 no weight critically if the other main groups agree against it. A case 
 
 is in 1.37: 
 
 t'at is fe wrecchide world Pe to betray e. 
 
 "wrecchide," THsDRTsLWDil^; "wicked," HV (UAH3 defective). 
 These three problems are all simple. A more comphcated one 
 comes in Prol. 42 : 
 
 Fayteden for here foode foujten at Pe ale. 
 417 
 
158 Thomas A. Knott 
 
 "Fayteden H] Flite p>aime," T; "Faytowrs," Hz; "Flytteden &," 
 D; "pei fliten," URTa; "Faytours/' L; "They failed," W; "And 
 flyted fast," Di (Di is full of contaminations from the C-text in the 
 Pro!., but the C-text here reads " Faytynge ") ; "Fayted," I; "Fey- 
 neden hem," V; "Fayteden," z (AH3 defective). The reading of 
 TDURT2 supports "Fliten" for this subgroup of y. H2 must be 
 a contamination. But L, I, and W support "Fayteden." Di goes 
 strongly against its sister, W, in favor of "Fliten," for Di's reading, 
 which is in disagreement with that of the C-text, must be the original 
 reading of its A-text ancestor. "Fayteden" of H is supported by 
 "Feyneden hem" of V, which seems to be a substitution of a synonym 
 for H's reading, the reading of x. "Fayteden" in z supports H 
 (and re) and practically three of the subgroups of y. The CT must 
 therefore be "Fayteden." 
 
 A more complicated problem, or pair of problems, is to be found 
 in 5.221 (229). 
 
 panne sat sleuj^e vp & seynide hym faste. 
 
 "seynide hym faste," TH2U]; "semed hym faste," D; "shryned {or 
 shryued?) him faste," R; "signed him faste," L; "sayned hem 
 fast," W; "seynyd hym ofte," Di; "schraffe hym full fast," T2; 
 "syhed ful faste," A; "syhede faste," H3; "crowchid him fast," I; 
 "seidetohymsiluen," H; "sikede sore," V; "seyned hym swithe," z. 
 Three branches of y attest "seynide" — L, WDi, and I, which 
 has an obvious substitution of a synonym — "crossed" for "signed." 
 Of the fourth subgroup of y, four MSS (all belonging to one sub- 
 subgroup) support "seynide." They are TH2DU. R, the sister 
 of U, with "shryned," an error due to an imperfect auditory image, 
 illuminates the step that must have existed in the ancestor of T2, 
 which had "shryned," which in turn was misread "shryued," and 
 then changed to "schraffe"; "syhed" in AH3 looks Uke the result 
 of a careless visual image of "syned," or perhaps of some mechanical 
 carelessness at some stage of transmission. The reading of x can 
 hardly be reconstructed with certainty, but x was so careless that 
 the reading may have been that of H, "seide to hymsyluen"; for 
 "seide" may have resulted from the omission of the horizontal 
 nasal stroke from above the "i." V's "sykede sore" may be an 
 attempt to improve some such reading as that of H, or it may be for 
 
 418 
 
The A- Version of "Piers the Plowman" 159 
 
 "sihede," which may have arisen in the same way that it did in AH3. 
 We see then that two and a half of the four subgroups of y have 
 "seynide," while the other MSS of y have readings probably or 
 obviously derived from "seynide." z has "seyned." Both V and 
 H may come from " seynide." The CT must therefore be " seynide. " 
 
 As the last word in the hne, y has "faste," and z has "swipe." 
 X is not certain, but may have had V's reading, "sore." z and x 
 have four alliterating syllables in the line, an arrangement of course 
 not unknown to the A-text, but rather unusual. Further, z's incli- 
 nation to change readings rather whimsically casts a great deal of 
 doubt on "swithe." x has such a multitude of clear errors, devia- 
 tions and demonstrable substitutions that even if we could certainly 
 determine its reading, we could not rely on it as surely as we usually 
 can on that of y, which, we have seen, has very few errors. And, 
 further, when x and y differ, with readings between which there is 
 little or no choice, if z supports either, it almost invariably supports y. 
 All probabiHty therefore favors y's "faste." The CT for the half- 
 line therefore should read, "seynide hym faste." 
 
 Another interesting problem may be found in 6 . 67 (70) : 
 
 panne shalt Pou blenche at a bergh here no fals wytnesse. 
 
 "bergh WDi] berwh," L; "berwe," D; "bowrne," TH2; "brige," 
 T2; "brook," H3; "bowhe,"I; "brok,"V; " berghe," 2 (H defective); 
 "at a bergh] abak," UR; A omits (U inserts "see" before "blenche"). 
 The reading "bourne," "brook," of TH2H3V is rendered improb- 
 able by both the context and the genealogical evidence. The author 
 has already used a brook in his symbolical geography, and named it 
 "be buxum of speche" (line 53). And of the feature in our line, 
 whatever for the moment we may consider it likely to be, he says 
 in the two lines following: 
 
 He is friHd in wit> floreynes & ot'ere fees manye; 
 Loke fou plukke no plante Pare for peril of J^i soule. 
 
 This description is obviously unsuitable for a brook, but perfectly 
 appropriate for a hill. 
 
 The genealogical evidence of L and WDi, forming two branches 
 of y, favors "bergh"; "bowhe" of I may have an "o" for an "er"; 
 "berwe" of D supports that reading for the ancestor of TH2D, for 
 
 419 
 
160 Thomas A. Knott 
 
 D branches off collaterally with TH2, being by itself of as much 
 genealogical weight as TH2 combined; "brige" of T2 might have 
 resulted from an erroneous expansion of "b'ge," which even might 
 have been miswritten as "b'ge." It is quite conceivable that 
 "brok" of V and H3 may have been the result of a similar error. 
 The abbreviation for "er" might have been misread for that of 
 "ur," "ru," while "k" and "w" in fourteenth-century handwriting 
 look a great deal alike. The reading of UR, "abak," can hardly 
 be attributed to any classifiable sort of error, though it might be 
 a conjectural emendation of "abrok." z has "berghe." y and z 
 therefore support "bergh," and this is to be adopted into the CT, 
 "Other things being equal," the genealogical evidence must 
 determine the CT. But sometimes other things are not equal. We 
 may perhaps close with the discussion of a problem of this kind. 
 It occurs in 2.198 (212): 
 
 And ek wep & wrang whan heo was atachid. 
 
 ''wrang," THaDURWDiz]; "wrong hire hondes," LAIHV (T2 
 omits the line). 
 
 Here, so far as the MSS strictly of the A-text are concerned, 
 "wrong hire hondes" is critically attested in the Original. But 
 other elements must enter the problem. On the principle of the 
 lectio difficilior, "hire hondes" would naturally be rejected by the 
 text critic if the evidence were evenly balanced, for the inclusion of 
 these words in the phrase certainly makes it the "easier reading." 
 That is, its presence would be more probably the result of scribal 
 conjectural emendation than its absence would be due to intentional 
 omission. The absence of the words in TII2DUR and in WDi is 
 evidence in two of the four subgroups of y in favor of its omission. 
 On the other hand, L and I are the other two subgroups of y, and are 
 of equal genealogical weight with the former two subgroups. L, I, 
 and VH (group x), and also MS A, a member of the same subgroup 
 with TH2DUR, all attest the presence of the words in the CT. 
 Might not the omission of the words be due to the carelessness of 
 two scribes at some points of transmission of parts of group y? 
 On the other hand, the third main group z omits the words; one 
 and one-half main groups support each reading. Finally, in a line 
 
 420 
 
The A- Version of "Piers the Plowman" 161 
 
 outside of "Piers Plowman," fully attested by MS evidence and 
 meter, the verb "wringen" appears in this meaning without "the 
 handes": 
 
 And lat him care and wepe and wringe and waille.' 
 
 We must, I feel, attribute the presence of "hire hondes" to several 
 independent inclinations to emend, or, rather, to write the obvious 
 for the slightly more idiomatic phrase, on the part of the facile 
 editor-scribe. The CT therefore must omit "hire hondes." Here 
 "other things are not equal." 
 
 Thomas A. Knott 
 University of Chicago 
 
 1 "The Clerk's Tale," E 1212, Oxford Chaucer. Note also Le Morte Arthur (Harl. 
 MS 2252), Furnivall, line 3931, and line 3746: 
 
 Alle nyght gan he wepe and wrynge 
 And went aboute as he were wode. 
 
 Also Cursor Mundi, 23962: 
 
 I se him [Christ] hang, i se hir [Mary] wring, 
 {)e car all of I)at cumli king.— MSS GCE. 
 
 And observe the scribal editing in MS F: 
 
 hir loueli fingris ho did wringe, etc. 
 
 421 
 
14 DAY USE 
 
 RETURN TO DESK FROM WHICH BORROWED 
 
 LOAN DEPT. 
 
 This book is due on the last date stamped below, or 
 
 on the date to which renewed. 
 
 Renewed books are subject to immediate recall. 
 
 M^-l. 
 
 — w 1^ ULJ 
 
 HOI/2 4 '04-6 m 
 
 6■l^n' 
 
 6^Mfe 
 
 HEC'O LD 
 
 M 6"65-3pj| 
 
 LD 21A-60m-4,'64 
 (E4555sl0)476B 
 
 General Library 
 
 University of California 
 
 Berkeley 
 
THIS BOOK IS DTTE ON THE LAST DATE 
 STAMPED BELOW 
 
 AN INITIAL PlSFoP 25 CENTS 
 
 WILL BE ASSESSED FOR FAII no^\. ^ 
 
 THIS BOOK ON THE DATE DUE TmIV^""^ 
 WILL INCREASE TO SO CEN^i^^' J PENALTY 
 
 DAY AND TO $[ OO on tm ° ''°'"''^" 
 
 OVERDUE. '^ ^"^ SEVENTH DAY 
 
 jGTM 
 
 JUN~Tl969B 7 
 
 LD 21-l00m-7,'33 
 
RETURN CIRCULATION DEPARTMENT 
 
 TO— i#^ 202 Main Library 
 
 LOAN PERIOD 1 
 HOME USE 
 
 2 3 
 
 4 
 
 5 6 
 
 ALL BOOKS AAAY BE RECALLED AFTER 7 DAYS 
 
 1 -month loans may be renewed by calling 642-3405 
 
 6-month loans may be recharged by bringing books to Circulation Desk 
 
 Renewals end recharges may be made 4 days prior to due date 
 
 DUE AS STAMPED BELOW 
 
 
 7h 
 
 
 REC CIR M 
 
 15 '83 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 UNIVERSITY OF CALIFORNIA, BERKELEY 
 FORM NO. DD6, 60m, 12/80 BERKELEY CA 94720