:: :: University of Virginia Library

ALTHOUGH THE BEST MANUSCRIPTS OF LAK of Stedfastnesse are in substantial agreement on most readings, they divide into two groups at lines 5, 10, and 28. One group reads ben no thyng on (l. 5), ffor now adayes, (l. 10), and dryue thi peple (l. 28); the other, Is no thing lyke, For amonge vs now, and wed thi folk. The choice between the first two pairs of readings—everyday phrases, practically synony-mous—is relatively inconsequential. With the third pair, however, the variation between 'drive' and 'wed' considerably affects the interpretation of the poem, for each indicates a different attitude toward kingship.[1]

Both sets of readings are in passably Chaucerian English; and previously, so far as manuscript evidence was concerned, editors have virtually been forced to toss a coin when considering these lines. Hence one finds Furnivall, Holt, MacCracken, and Koch supporting the first set; Skeat, Brusendorff, and Robinson, the second; Heath, both (in the belief that they derived from 'separate originals').[2] With the discovery of a new manuscript, however, it is now possible to show that one set of readings is the work of scribes.

This new text—an eighteenth-century transcription of the burnt Cotton manuscript Otho A. XVIII [3]—reads the lines in question as follows: Is no thynge oon, Amonges vs nowe, and dreve

106

thy folke. As a moment's inspection will show, these readings are almost beyond doubt a transition between the two groups of variants given above. It follows, of course, that only one of the two groups can be Chaucerian. But unfortunately the readings from Otho A. XVIII do not reveal which group that is.

Even so, the knowledge that one set of variants is of scribal origin is sufficient, I believe, to allow one to solve the problem, and in the end to arrive at a firmly established text for the poem. To accomplish this aim, however, one must analyze anew the Lak of Stedfastnesse manuscripts—a desirable project in any event, for the published studies of the text of this poem are conflicting and incomplete.[4]

The authorities for Lak of Stedfastnesse, with their sigils, are:[5]

A Additional 22139. British Museum. Fol. 138a.
B Advocates Library i. 1. 6. (the 'Bannatyne Ms.'). Edinburgh. Fol. 67a.
C Cotton Cleopatra D. VII. British Museum. Fol. 188b.
Co Cotton Otho A. XVIII (transcription). British Museum. After p. 548 in a copy of Urry's Chaucer (643. m. 4).
D Trinity College, Dublin, No. 432. Fol. 59a.
F Fairfax 16. Bodleian. Fol. 194a.
H1 Harley 7333. British Museum. Fol. 147b.
H2 Harley 7578. British Museum. Fol. 17a.
Ht Hatton 73. Bodleian. Fol. 119a.
L Lambeth Palace Library No. 344. London. Fol. 11a.
M Pepys 2553 (the 'Maitland Folio Ms.'). Magdalene College, Cambridge. Page 329.

107

R1 Trinity College, Cambridge, R. 3. 20. Page 356.
R2 Trinity College, Cambridge, R. 3. 21. Fols. 245b and 319a (identical copies; the envoy only).
R3 Trinity College, Cambridge, R. 14. 51. Fol. iia.
Th Thynne's printed version, derived from a lost Ms., in The Workes of Geffray Chaucer (London, 1532). Sig. Vvv4a.

If my reasoning is correct, these authorities arrange as follows:

I. In explaining the tree, it will be of help to establish first the relationships which depend for their proof solely upon the distribution of the variants, since in so doing we can not only keep distinct the kinds of evidence with which we have to deal but can also reduce the number of texts which need careful consideration from fifteen to seven. These relationships are as follows:

(1) H1 and R1 share three readings found in no other Ms. and hence must be considered as closely related. Neither, however, can be regarded as descended from the other, for each text reads uniquely in places where the other agrees with the rest of the Mss.

108

First, H1's unique readings (R1's readings are supported by the other Mss):

3. is fals; R1 is so fals.
17. wyght; R1 man.
27. thorow all goodnesse; R1 trouth and worthynesse.

In addition, H1 reads (with C),[6] in line 19, permutacioun; R1 a parmutacone, the reading of all other Mss.

Now, this unique reading in R1 (H1's reading is supported by other Mss):

6. by wikked wilfulnesse; H1 thorowe mede & wylfullnesse.[7]

In addition, R1 reads (with R3), in line 8, the (H1 and the rest, this), and (with L), in line 9, folkes (H1 and the rest, less R3, folke).

These readings are uniquely common to H1 and R1:

2. was holde (R1 was holde for). The others read was.
4. werke. The others read dede.
8. H1 made, R1 mape. The others read maketh, causep.

On the basis of these readings we may assume ex2H1 R1 (the a of the tree). The significant readings for a are as follows:[8]

1. þis worlde was so stedfast and stable.
2. was holde (or was holde for).[9]
3. And . . . so fals and disceyvable.
4. worde and werke as in conclusyone.
5. Beon no thing oon for tourned vp so doune.
6. thorowe mede and wylfulnesse.
8. maþe (or made) this worlde to be so variable.
9. folk . . . in discencioun.
10. ffor nowe adayes.
11. But if . . . by sum collusyone.
12. Do.
13. What causeþe þis but wilful wrechednesse.
16. Vertue haþe nowe no domynacone.

109

17. Pytee exyled no man is mercyable.
18. is blent.
19. þe worlde haþe made a parmutacone.
20. frome trouth to fikulnesse.
22. O prynce desyre for to beo honurable.
23. Cherisshe þy folke.
24. may beo.
25. To þyne estate doone in þy Regyoune.
26. Shewe forpe þy swerde of castigacioun.
27. trouth and worþynesse.
28. And dryve þy people ageine to stidfastnesse.

(2) As MacCracken has observed,[10] L is apparently an actual copy of Ht, although not a very faithful one. The two texts share three readings not found elsewhere:

3. But (L Byt) now. The others (less R3, Now) read And now.
12. Do to. The others (less D, To do) read Do.
13. that. The others (less R3, And alle causyht) read this.

Ht has no unique readings. L, however, has the following two, which establish the descent as Ht>L:[11]

2. That man ys word; Ht (and the rest) that mannes word.
13. wrechednesse; Ht (and the rest) wilful wrecchednesse.

The significant readings of Ht, so far as they differ from those of a, are as follows:

2. was.
3. But . . . so fals And disceyvable.
4. word and dede as in conclusioun.
6. for mede and wilfulnesse.
8. makith this world to be so variable.
12. Do to.
13. What causeth that but wilful wrecchednesse.
22. O prince desyre to be honurable.
27. trouthe and rightwesnesse.

(3) Except for reading uniquely In for To in line 25, R2, a fragment consisting only of the envoy, is identical with Co. The

110

significant readings for Co, so far as they differ from those of a, are as follows:

2. was.
4. wurde and dede as yn conclusion.
5. Ys no thynge oon.
6. for mede and wilfulnesse.
8. Whate makith this worled to be varyable.
16. As line 17.
17. As line 16: Pitey is so gyled no man is mercyable.
28. dreve thy folke.

(4) The two late sixteenth-century texts B and M share nine readings found in no other Ms. Because of the following unique readings in B (where M agrees with the rest of the Mss), we cannot, however, assume the descent B>M:

9. quhilk; M pat.
16. nane at hir devotioun; M now na dominatioun.
25. B writes line 26 here.
26. That vertew may rigne within thy regioun; M Schaw furth þi Sworde of castigatioun (D R3 Th yerde for sworde).

Nor, because of the following unique readings in M (where B agrees with the rest of the Mss), can we assume the descent B<M:

4. deidis; B deid.
11. Except; B Bot gif.
19. is; B hes (= hath).

The following readings are uniquely common to B and M:

4. in conclusioun. The rest read as in conclusioun.
5. bot. The rest read for.
9. B bot discretioun; M of indescretioun. the rest read in discencion.
12. Doing The rest (less A, Done) read Do.
13. makis . . . wofull. The rest (less D, causep . . . sotel) read causeth . . . wilfull.
17. and na man meretabill. The rest (less H1, wyght for man, and A, man merciable) read no man is mercyable.
20. ressone to wilfulnes. The rest read trouth to fikulnesse.
24. be. The rest read may be.

In addition, B and M insert a stanza which seems plainly non-Chaucerian immediately before the envoy, and reverse stanzas 2 and 3.

111

On the basis of these readings, we may assume ex2B M (the d of the tree). The significant readings for d, so far as they differ from those of a, are as follows:

1. this (M ye) warld so steidfast was & stabill (order).
2. was.
4. word and deid (B adds discordis) in conclusioun.
5. Ar no thing lyke bot turnd up syd (B up and) doun.
6. for neid (B greid) and wilfulnes.
8. makis pis warld to be so variabill.
9. folk . . . of indiscretioun (B bot discretioun)
10. Amang ws now.
12. Doing.
13. Quhat makis this bot wofull wretchitnes.
17. Petie exylit and na man meritabill.
18. Blind is.
19. The warld hes maid a permutatioun.
20. fra rasoun to wilfulnes. Non-Chaucerian stanza.
24. be (B bene)
27. treuth and rychtuousnes.
28. And leid (B bring) þi folk agane to stedfastnes. Stanzas 2 and 3 reversed.

(5) R3 and Th share two readings found in no other Ms. Because of these unique readings in Th (where R3 and the other Mss agree), we cannot, however, assume the descent R3<Th:[12]

6. fykelnesse; R3 wylfulnes.
12. and; R3 or.

As corroborating evidence, one should note that Th reads uniquely an, line 2 and is, line 5, whereas R3 and the other texts agree in omitting these words.

Similarly, we cannot assume the descent R3>Th because of unique readings in R3 (where Th and the rest of the Mss agree):

1. Wylum; Th Somtyme.

112

3. Now hyt ysse; Th And nowe it is.
5. els butt; Th lyke.[13]
13. And alle causyht; Th What causeth this but.
19. Thys; Th The.
28. ayeyn thy folke; Th thy folke ayen.

In addition, R3 reads uniquely, in line 11, maner, which Th omits with all other Mss, and shares with a scattering of texts certain other readings which, since Th's are widely supported, prevent the descent R3>Th:

1. was stedefast (with H2); Th so stedfast was.
5. up & don (with B A); Th up so doun.
8. causyht (with D); Th maketh.
27. Ryhtwysnis (with Ht L B M); Th worthynes.

These readings are uniquely common to R3 and Th:

9. men. All other Mss read folke.
10. For amonge vs. The other Mss read ffor among vs now, Among ws now, ffor now adayes.

In addition, the following readings, although not uniquely common to R3 and Th, are evidence for grouping them together:

17. Pyte is exyled (with D). The others (less Co, Pitey is so gyled) read Pite exiled.
22. Prynce (with A). The others (less D, illegible) read O prince.
26. yerd (with D). The others read sword.

On the basis of these readings, we may assume ex2R3 Th (the f of the tree). The significant readings for f, so far as they differ from those of a, are as follows:

1. Somtyme þe world so stedfast was and stable.
2. was.
4. worde and dede as in conclusioun.
5. Is nothyng lyke for turned vp so doun.
8. maketh the worlde to be so variable.
9. men . . . in discensyon.
10. For amonge vs.
17. Pety ys exylyd no man ysse mercyable.
22. Prince desyre to be honourable.

113

26. Schow forthe thy yerd of castygacyun.
27. trouthe and worthynes (R3 t. a. Ryhtwysnis.)[14]
28. And wedde thy folke ayen to stedfastnesse.

(6) H2 and F share one reading not found elsewhere. Because of the following readings in H2 (where F and the other Mss agree), we cannot, however, assume the descent H2>F:

1. Somme tyme worlde; F Some tyme the worlde.
5. torneth; F turned.
16. none; F noo.

In addition, H2 reads (with A R3), in line 1, was stedfast. F and the other Mss read was so stedfast (B M so steidfast was).

Nor, because of the following unique reading in F (where H2 and the other Mss agree), can we assume the descent H2<F:

10. holde; H2 is holde.

The following reading is uniquely common to H2 and F:

3. so fals and so disceyuable. All other Mss omit the second so.

Furthermore, except for the readings cited for lines 1, 5, 10, and 16, the text of F and H2 are identical, frequently even in spelling.

On the basis of the above evidence, we may assume ex2H2 F (the g of the tree). The significant readings for g, so far as they differ from those of a, are as follows:

1. Some tyme the worlde was so stedfast and stable.
2. was.
3. And . . . so fals and so disceyuable.
4. word and dede as in conclusyon.
5. Ys noo thing lyke for turned vp so don.
6. for mede and wilfulnesse.
8. maketh this worlde to be so variable.
10. ffor amonge vs nowe.
11. conclusioun.
22. O prince desire to be honourable.
28. And wedde thy folke ayeyne to stedfastnesse.

114

A, C, and g share one reading which seems valid for grouping them together: the obvious error conclusion for collusion in line 11. None of these texts, however, can be the parent of any of the others, g because of its reading in line 3, A and C because of readings elsewhere. The readings which prevent descent from A are:

12. Done. C, g, and the rest (less B M Doing) read Do.
23. Speke with folke. C, g., and the rest read Cherice þi folk.

The following reading for C, while not unique, prevents one from considering it the parent of any Ms.:

19. permutacioun. A, g, and the rest (less H1) read a parmutacioun.

Finally, no two of these Mss can have a common ancestor which excludes the third, since none shares readings with one that it does not also share with the other. We must hence assume ex2 C A g (the h of the tree). This Ms. differs from g as follows:

3. And . . . so false and disceyuable.
10. ffor among vs now.

The reading for line 10, confined to h and its descendants, has become in effect a 'unique reading.'

As for the remaining Ms., D, one can deduce from the distribution of the variants alone only that it cannot be the parent of any of the other texts. This we can show by listing the unique readings which it contains:

1-7. Omits the first stanza.
8-28. Begins with the envoy and reverses stanzas 2 and 3.
11. But he can com be sum ymaginacioun. The rest read But yf he can by som collusioun (A C F H2 L conclusion).
13. sotel dowblenes. The rest (less B M, wofull wretchitnes) read wilful wrechidnesse.
15. Troupe is rebuked & reson is hold but fable. The rest read Trouthe is putte doune resoun is holden fable.
25. To your astate wher ye have correctioun. The rest (less B, which reads uniquely) read To thine estaitt doen in þy regioune.
26. your. The rest read thy.
28. & knyt to gydre your peple with stedfastnes. The rest read And wed (Co R2 Ht L R1 H1 drive, B leid, M bring) thi folke (Ht L R1 H1 peple) ayen to stedfastnesse.

Although we cannot relate D to any of the other texts at this

115

point, we can remark that it agrees with Ht a in reading ffor now on dayes (I. 10) and peple (I. 28). We can also note that the general badness of D's text—its beginning with the envoy and proceeding backwards through the next two stanzas, its many unique and obviously incorrect readings—make it unthinkable that D was copied from a Ms. Almost certainly D's scribe was writing from an imperfect memory of the poem.

A glance at the italicized lines (indicating unique variants) in the 'significant readings' given above for Ht, Co, D, and the reconstructed parents of H1-R1, B-M, R3-Th, and H2-F-A-C will show that none of these can be the parent of any extant text. Hence we have seven authorities, none of which can be directly descended from any of the others.

We have carried our study of the relationships of the Mss about as far as we can solely on the basis of the distribution of the variants. We have established certain relationships, and in so doing have reduced our original fifteen authorities to seven—three actual Mss and four postulated ones. If we continue our procedure of determining kinship by shared variants, we soon arrive at the impasse mentioned in the first paragraph of this paper. On the one hand, we have Ht a D reading, in lines 5, 10, and 28, ben no thyng on, ffor now adayes, and dryue thi peple; on the other, d f h reading Is no thing lyke, For amonge vs now, and wed thi folk. In between lies Co, with its transition readings, Ys no thing oon, Amonges vs nowe, and dreve thy folk, showing that only one of the above sets can be Chaucerian. But the only readings uniquely common to Ht a D are their variants for lines 5, 10, and 28; and, similarly, the only readings uniquely common to d f h are their variants for lines 5, 10, and 28. Distributional study, then, leaves us in a dilemma. If we are to relate the seven authorities to each other, as shown in the tree on page 107, we must seek other means.

II. So far we have not examined the headings for the poem. Only five Mss have titles. These are Co Ht L R1 H1—with the exception of Co, the Mss which read ben no thyng on, ffor now adayes, and dryue thi peple. If we study these titles, we find a similarity between them which suggests that the five texts are related closely.

Co Balade Ryalle made by Poetecall Chaucyer a Gaufrede.

116

Ht Old title: Chaucier send (no more visible); new title: These balladis were send to the kynge. (L has Ht's second title.)
a R1 Balade Royal made by oure laureall poete of Albyoun in hees laste yeeres. H1 This balade made Geffrey Chaunciers the laureall poete of Albion and sent it to his souerain lorde kynge Richarde the seconde pane being in his castell of Windesore.

We note that the three texts which mention the poet's name spell it -ier, not -er. We also note that Co and R1 are similar in that each calls the poem a ballad royal, and that Ht and H1 are alike in asserting that Chaucer sent the poem to the king. Further, since H1 and R1 are sister texts on strong evidence, we are justified in assuming that their parent, a, read much as does H1 but began, say, 'Ballad Royal . . .,' as do R1 and Co. If we attempt to classify these texts on the basis of their titles, we have the following arrangement:

The readings ben no thyng on, ffor now adayes, and dryue thi peple thus appear to have been begun by the scribe who wrote c and then 'improved upon' by the scribe who wrote b.

But before accepting this explanation, we should certainly like additional, and stronger, evidence. Nor have we far to go to find it. In the manuscript books of Ht (and L), Co, and H1, Lak of Stedfastnesse immediately follows Chaucer's Truth; in R1 it immediately precedes it. Again we find a similarity in the titles.[15] More important, however, we find from a purely distributional study of

117

the text that the copies of Truth are related in one or the other of two ways:[16]

As one can see, the second tree is identical with that proposed above for Lak of Stedfastnesse. The coincidence between the trees virtually proves that the two poems, side by side in the manuscripts, have the same ancestry.[17]

III. There is no evidence whatsoever for the alternate assumption, that the readings Is no thing lyke, For amonge vs now, and wed thi folk are of scribal origin. We may hence assume that Ht and a have a common parent (the b of the tree). Because D also reads ffor now on dayes and peple, I have graphed it as descended from b.[18] The significant readings for b, so far as they differ from those of a, are as follows:

2. was.
4. word and dede as in conclusioun.
6. for mede and wilfulnesse.
8. makith this world to be so variable.

Further, for reasons already stated we may assume that b and the transition Ms. Co have an exclusive common ancestor (the c

118

of the tree). Except for the following, the significant readings for c are the same as those for b:

5. Is no thynge oon for turned vppe so down.
10. Amonges vs nowe.[19]
28. And dreve thy folke ageyn to stedefastnesse.

The readings of c for lines 10 and 22 (for to be honorable) are found also in d. In addition, both texts have had trouble with line 28 (d reads either leid þi folk [M] or bring thy folk [B]). Because of the unique readings listed above for d, however, we cannot assume the descent d>c. Nor can we assume the descent d<c, since c reads, in line 5, oon, which has been shown to be of scribal origin, and d reads lyke, which has been shown to be the reading of the original. Hence we must assume ex2 c d (the e of the tree). The significant readings for e, so far as they differ from those of c, are these:

5. Is no thynge lyke.[20]
28. **** thy folke.

V. In the tree I have graphed e, f, and h as descended inde-dently from O. That this conclusion is inescapable the following table shows. The table lists only the readings on which the three texts disagree.

	e	f	h
1.	this world was so stedfast	þe world so stedfast was	the worlde was so stedfast
8.	this world	the worlde	this worlde
9.	folk	men	folke
10.	Among ws now	For amonge vs	ffor among vs now
11.	collusioun	collusioun	conclusion
17.	exiled	is exyled	exiled
22.	O prince for to be honorable	Prince to be honourable	O prince to be honurable
26.	swerd	yerd	swerde
28.	**** thy folke	wedde thy folke	wed thi folk

As one can see, neither e nor f can be descended from h because of

119

its reading in line 11; neither e nor h can be descended from f because of its readings in lines 1, 2, 9, 10, 22, and 26; and neither f nor h can be descended from e because of its readings in lines 1, 10, and 28. Hence we must assume that all three texts derive independently from 0.

VI. We have now completed the explanation of the tree.[21] We have four Mss at one remove from the original—Th R3 A C. Of these, C is decidedly the best. It is the basis of the text which I now print.

[Lack of Stedfastnesse]
Sumtyme the worlde was so stedfast and stable
That mannes worde was obligacioun
And nowe it is so false and disceyuable
That worde and dede as in conclusioun
[5] Is no thing lyke for turned vp so doun
Is all this worlde for mede and wilfulnesse
That al is lost for lak of stedfastnesse

120

[2]
What maketh this worlde to be so variable
But lust that folke haue in discencioun
[10] For amonge vs now a man is holde vnable
But yf he can by som collusioun
Do his neyghbour wrong or oppressioun
What causeth this but wilful wrechednesse
That al is lost for lake of stedfastnesse

[3]
[15] Trouthe is putte doun resoun is holden fable
Vertu hath now no dominacioun
Pite exiled no man is merciable
Thorugh couetyse is blent discrecioun
The worlde hath made a permutacioun
[20] Fro right to wrong fro trouth to fikelnesse
That al is lost for lak of stedfastnesse

[Lenvoy]
O prince desyre to be honurable
Cherisshe þi folke and hate extorcioun
Suffre no thing that may be reprouable
[25] To thine estate don in þi regioun
Shewe forth thy swerde of castigacioun
Drede god do law loue trouthe and worthynesse
And wed thi folk ayen to stedfastnesse

[Explicit]

Although this text is not markedly different from Skeat's or Robinson's,[22] there are a few differences worth noting:

1. the worlde. Skeat and Robinson, this world.
5. Is no thing lyke. Skeat and Robinson, Ben nothing lyk (this reading occurs in no Ms.).

121

10. For amonge vs now. Skeat, Among us now.
12. Do. Skeat and Robinson, Don (this reading occurs only in A).

With the other group of texts—Heath's, Holt's, Koch's, and Mac-Cracken's—the differences are of course numerous, but there seems little point in cataloging them here since the manuscripts on which they based their texts have been shown to be poor authorities. It is interesting, however, and perhaps instructive, to examine the reasoning which led them into error.

The last important difference is to be found in the final line of the poem, where Skeat has 'wed thy folk' and I have adopted the reading of the H S Hat [H1 R1 Ht] group, 'Dryve thy people.' . . . The obvious reason why I have adopted the H S Hat reading is that the verb 'Dryve' makes sense where the other does not seem to do so. I quote from line 26:

Shew forthe thy swerde of castigacion,
Drede god, do lawe, loue trouthe and worthynesse,
And dryve thy peple ageyne to stedfastnesse.

He is implored to draw his sword and drive the people to stedfastnesse; is this not more sensible than to draw his sword and wed his folk to stedfastnesse?[23]

The weakness of this reasoning is apparent if one asks oneself the question, Why should a scribe emend a reading which 'makes sense' to one which 'does not seem to do so'?

If one reads the line in its true context, however, 'wed' makes much better sense poetically than 'dryve.' I quote from the beginning of the envoy:

O prince desyre to be honurable
Cherisshe þi folke and hate extorcioun
Suffre no thing that may be reprouable
To thine estate don in þi regioun
Shewe forth thy swerde of castigacioun
Drede god do law loue trouthe and worthynesse
And wed thi folk ayen to stedfastnesse

Since three scholars have misinterpreted this passage, I hope I may be forgiven for obtruding literary criticism upon a discussion of manuscript relationships. Chaucer asks the king to 'shewe forth'

122

his 'swerde of castigacioun' against those who do things 'that may be reprouable'—the extortioners, the godless, the lawbreakers: the enemies of steadfastness. He can hardly ask the king to cherish his folk in line 23 and five lines later to drive them with the sword. Nor does he, if the analysis in this paper is correct.

Notes

[1]

Cf. F. J. Furnivall, A Parallel-Text Edition of Chaucer's Minor Poems, p. 433.

[2]

Furnivall, ibid.; L. H. Holt, JEGP, IV (1906), 419-431; H. N. MacCracken, The College Chaucer, p. 557; J. Koch, Anglia, IV (1881), anz. 109 and Geoffrey Chaucers Kleinere Dichtungen, (Heidelberg, 1928), p. 35; W. W. Skeat, Oxford Chaucer, I, 394-395; A. Brusendorff, The Chaucer Tradition, pp. 275-276; F. N. Robinson, Chaucer, pp. 1037-38; H. F. Heath, Globe Chaucer, p. xlix.

[3]

Printed in my article, "Otho A. XVIII," Speculum, XXVI (1951), 306-316.

[4]

The following texts were unclassified until the present study: Co, D, M, R2, and the Black Letter prints subsequent to Thynne's (for the meaning of the sigils, see text below). Nine of the remaining authorities were arranged by Koch, and later by Heath, as follows: (1) R1 H1 (2) A C F H2 B R3 Th. Holt published a similar bifid tree. Brusendorff, who criticized Holt's article severely (and justly), offered a trifid arrangement: (1) A C F H2; (2) H1 R1 Ht L; (3) Th R3 B. (Robinson's brief textual note agrees with Brusendorff's classification. Skeat did not attempt a classification, but his text is similar to Robinson's.) Only Koch, Heath, and Holt published actual trees; only Holt argued in detail.

[5]

Except for B and D (of which I have only seen photostats), I have examined all of the Mss and have based my study upon my transcriptions of the poem. All the Mss, including the Ms of Co (see fn. 3), however, have been published (for references, see the Brown-Robbins Index of Middle English Verse). The most important difference between the published transcriptions and my own is in A, the final (envoy) stanza of which the Chaucer Society failed to print. In consequence, the Ms. is always listed as lacking the envoy (cf., e. g., Robinson, p. 1037). Except for B (1568) and M (c. 1570), the Mss are from the 15th century; A C F R1 are from the first half.

[6]

In view of the evidence just given, the agreement with C seems non-significant. For a further discussion of this agreement, and of all other agreements considered non-significant, see fn. 21 below.

[7]

R3 thro mede & wylfulnes; the rest fore mede and wilfulnesse.

[8]

'Significant' in that at least one Ms. reads differently. (Where no reading is given, the Mss are in agreement.)

[9]

The italic type indicates that the reading is not found outside the hypothetical Ms. and its descendants.

[10]

MLN, XXIII (1908), 212-214.

[11]

In addition, L reads, in 6, In alle thys worle (with A M), and in 11, conclusyoun (with A C F H2); Ht and the rest read Is al this world and collusioun. Although the agreements seem clearly the result of chance (see fn. 21 and 22), they make the descent Ht<L even more unlikely.

[12]

From Th, however, are ultimately descended the versions of Lak of Stedfastnesse in the Black Letter Chaucers of the 16th and 17th centuries (and also the copy in the Urry Chaucer of 1721). Since these texts are practically identical, they are of no value for this study (for the order of descent, see my article, "The Text of Chaucer's Purse," [Studies in Bibliography], I (1948), 111; the printed versions of the two poems have the same history.) Likewise, the copy of the envoy in Meredith Hanmer's The Auncient Ecclesiasticall Histories (1577) which Miss Spurgeon mentions (Chaucer Criticism and Allusion, I, 112) is of no value textually. Except for changes in spelling, it is identical with Th.

[13]

R3's reading is virtually illegible (butt, the most likely guess, is also the reading of the Chaucer Society).

[14]

The evidence for this line is as follows: Ht d R3 trouthe and rightwesnesse. The rest (less H1) trouthe and worthynesse. H1 all goodnesse. As a glance at the tree will show, there is every reason to believe that Chaucer's text read worthynesse; I hence assume that this was the reading of f. (Brusendorff and Robinson, however, group R3 Th with B, which reads rychtuousness. While this arrangement may seem to explain the presence of the same variant in R3 and B, it is in fact not an explanation, for it merely shifts the problem from R3's Ryhtwysnis to Th's worthynes. It is now contradicted by the readings which B shares with Co.)

[15]

Co A balade by Geffrey Chaucier uppon his dethe bedde lyinge in his grete Anguysse. Ht (Old title) Chauncier balade up on his deth bed (new title: Good conseylle; so L also). R1 Balade þat Chaucier made on his deeth bedde. H1's title bears no resemblance to the others (A Moral Balade of Chaucyre).

[16]

Discussed in detail in my article cited above, fn. 3.

[17]

Although the coincidence seems to be too striking to be due to chance, we cannot positively reject the first tree. Both trees explain the evidence reasonably well, but the first, which requires us to make fewer assumptions, would ordinarily be preferred on the ground of simplicity. Whichever tree we accept, however, these Mss. are obviously very closely related in Truth; and, in view of the fact that the two poems seem to go as a pair, we are safe in concluding that they are closely related in Lak of Stedfastnesse.

[18]

But D also shares three readings with R3: 8, causep; 17, is exciled (so Th also); 26, yerde (so Th also); hence D might just as well be graphed, perhaps, as a sister of R3.

[19]

Despite the for in b's reading, we can hardly assume that c read, compositely, For amonges vs nowe, since we should then have to assume that Co and d edited identically. Moreover, it is just as easy to believe that b derived For nowe adayes from Amonges vs nowe as from For amonges vs nowe (one automatically supplies for or an equivalent connective).

[20]

Here a composite reading seems to me to be justified. Ar no thing lyke is most easily explained as a 16th-century 'correction.'

[21]

But strong contradictory evidence has forced us, from time to time, to regard some readings as non-significant; and since there is no absolute certainty in textual matters, I collect these readings here. Most of them are easily explained as the result of careless copying, of dialect differences, or of time. (1) Careless copying (small changes, roughly synonymous): 1, this for the (A e), wasse stedfast for w. so s. (A H2 R3); 6, In for Is (A d), thorowe for for (H1 f); 8, pe for this (R1 f): 9, han for haue (A F-H4 Ht); 12, and for or (Co Th); 19, permutacioun for a permutacioun (C H1); 22, Prince for O prince (A f). Here probably also belongs conclusioun for collusioun (L h), 11. (2) Dialect or time: 5, Ar for Is (A d; a 'correction' analogous to the Beon of H1 R1 Ht L D); up and doun for up so doun (A B R3); the effect of time upon idiom—cf. M's up syd doun). The remaining variants are possibly the result of contamination: 27, rightwesnesse for worthynesse (d Ht R3); 8, causep (D R3); 17, is exciled (D R3 Th); 26, yerde for swerde (D R3 Th). I have not, however, assumed contamination to explain the first pair, since the only evidence is the one reading and it is capable of another explanation, that of independent editing. Although one may hesitate to adopt this explanation, he has, I feel, good grounds for making it here. First, the nature of the variant itself. The number of metrically satisfactory substitutes for worthynesse is extremely small. Hence if any number of scribes decide to emend the reading (because, say, their copy is illegible), the chances are that a large proportion of them will hit upon rightwesnesse. Further, 'truth and righteousness' has a kind of obvious force which 'truth and worthyness' lacks, and thus might seem to a scribe an improvement worth making. Second, the nature of the Mss reading rightwesnesse. Ht has every appearance of being a faithful copy. R3, however, is quite corrupt and is clearly not the work of a professional scribe (it is written on a flyleaf of a book of medical receipts and is preceded by a stanza, likewise corrupt, from Gentilesse). Such variants as Wylum for Sumtyme (1), no thyng els butt for no thing lyke (5), and by sum maner colusyon for by som collusioun (11) show that R3's scribe was either having difficulty reading his copy or was writing from memory. As for the third Ms, d, which survives in the late 16 century texts B and M, emendation seems almost to have been the rule (see p. 110 above). At least in R3 and d, then, it is not surprising to find some other reading than worthynesse, and, since the substitutes are few, not especially surprising to find rightwesnesse in both. Nor have I assumed contamination to explain the agreements between D and R3 (Th). It is quite possible that D's scribe was familiar with two versions of the poem and that these had fused in his very faulty memory. But with two such corrupt texts as D and R3, a question mark, I feel, more exactly expresses the situation than a dotted line. Whether or not my reasoning in this footnote is correct, I should like to express indebtedness, for the method I have employed, to two articles, one by Sir Walter Greg, "The Rationale of Copy-text," and the other by A. A. Hill, "Some Postulates for Distributional Study of Texts," Studies in Bibliography, III (1950), 19-36, and 63-95.

[22]

Although he did not classify the Mss, Skeat correctly chose C as his basic text. Robinson concluded that neither R1-H1 nor C-A-F-H2 is consistently superior, but in most cases gave preference to the latter group.

[23]

I quote from Holt (op. cit., p. 431). Heath and Koch reason similarly, but Mac-Cracken does not say why he chose H1 for his text.