Open main menu

Chaucer's Works (ed. Skeat) Vol. I/Romaunt - Introduction

[ 1 ]

INTRODUCTION.

THE ROMAUNT OF THE ROSE.

§ 1. In the Third Edition of my volume of Chaucer Selections, containing the Prioress's Tale, &c., published by the Clarendon Press in 1880, I included an essay to shew 'why the Romaunt of the Rose is not Chaucer's,' meaning thereby the particular English version of Le Roman de la Rose which happens to be preserved. I have since seen reason to modify this opinion as regards a comparatively short portion of it at the beginning (here printed in large type), but the arguments then put forward remain as valid as ever as regards the main part of it (here printed in smaller type, and in double columns). Some of these arguments had been previously put forward by me in a letter to the Academy, Aug. 10, 1878, p. 143. I ought to add that the chief of them are not original, but borrowed from Mr. Henry Bradshaw, whose profound knowledge of all matters relating to Chaucer has been acknowledged by all students.

§ 2. That Chaucer translated the French poem called Le Roman de la Rose, or at least some part of it[1], no one doubts; for he tells us so himself in the Prologue of his Legend of Good Women (A 255, B 329), and the very frequent references to it, in many of his poems, shew that many parts of it were familiarly known to him. Nevertheless, it does not follow that the particular version of it which happens to be preserved, is the very one which he made; for it was a poem familiar to many others besides him, and it is [ 2 ] extremely probable that Middle English versions of it were numerous. In fact, it will presently appear that the English version printed in this volume actually consists of three separate fragments, all by different hands.

The English version, which I shall here, for brevity, call 'the translation,' has far less claim to be considered as Chaucer's than unthinking people imagine. Modern readers find it included in many editions of his Works, and fancy that such a fact is conclusive; but it is the merest prudence to enquire how it came there. The answer is, that it first appeared in Thynne's edition of 1532, a collection of Chaucer's (supposed) works made more than a hundred and thirty years after his death. Such an attribution is obviously valueless; we must examine the matter for ourselves, and on independent grounds.

§ 3. A critical examination of the internal evidence at once shews that by far the larger part of 'the translation' cannot possibly be Chaucer's; for the language of it contradicts most of his habits, and presents peculiarities such as we never find in his genuine poems. I shewed this in my 'Essay' by the use of several unfailing tests, the nature of which I shall explain presently. The only weak point in my argument was, that I then considered 'the translation' as being the production of one author, and thought it sufficient to draw my examples (as I unconsciously, for the most part, did) from the central portion of the whole.

§ 4. The next step in this investigation was made by Dr. Lindner. In a painstaking article printed in Englische Studien, xi. 163, he made it appear highly probable that at least two fragments of 'the translation' are by different hands. That there are two fragments, at least, is easily discerned; for after l. 5810 there is a great gap, equivalent to an omission of more than 5000 lines.

§ 5. Still more recently, Dr. Max Kaluza has pointed out that there is another distinct break in the poem near l. 1700. The style of translation, not to speak of its accuracy, is much better in the first 1700 lines than in the subsequent portions. We may notice, in particular, that the French word boutons is translated by knoppes in ll. 1675, 1683, 1685, 1691, 1702, whilst, in l. 1721 and subsequent passages, the same word is merely Englished by botoun or botouns. A closer study of the passage extending from l. 1702 to l. 1721 shews that there is a very marked break at the end of l. 1705. Here the French text has (ed. Méon, l. 1676):— [ 3 ]

'L'odor de lui entor s'espent;
La soatime qui en ist
Toute la place replenist.'

The English version has:—

'The swote smelle sprong so wyde
That it dide al the place aboute'—

followed by:—

'Whan I had smelled the savour swote,
No wille hadde I fro thens yit go'; &c.

It will be observed that the sentence in the two former lines is incomplete; dide is a mere auxiliary verb, and the real verb of the sentence is lost; whilst the two latter lines lead off with a new sentence altogether. It is still more interesting to observe that, at this very point, we come upon a false rime. The word aboute was then pronounced (abuu·tə), where (uu) denotes the sound of ou in soup, and (ə) denotes an obscure vowel, like the a in China. But the vowel o in swote was then pronounced like the German o in G. so (nearly E. o in so), so that it was quite unlike the M.E. ou; and the rime is no better than if we were to rime the mod. E. boot with the mod. E. goat. It is clear that there has been a join here, and a rather clumsy one. The supply of 'copy' of the first translation ran short, perhaps because the rest of it had been torn away and lost, and the missing matter was supplied from some other source. We thus obtain, as the result to be tested, the following arrangement:—

Fragment A.—Lines 1-1705. French text, 1-1678.

Fragment B.—Lines 1706-5810. French text, 1679-5169.

Fragment C.—Lines 5811-7698. French text, 10716-12564.

It should be noted, further, that l. 7698 by no means reaches to the end. It merely corresponds to l. 12564 of the French text, leaving 9510 lines untouched towards the end, besides the gap of 5547 lines between Fragments B and C. In fact, the three fragments, conjointly, only represent 7018 lines of the original, leaving 15056 lines (more than double that number) wholly untranslated.

§ 6. Discussion of Fragment B.

Test I.—Proportion of English to French.—As regards these fragments, one thing strikes us at once, viz. the much greater diffuseness of the translation in fragment B, as may be seen from the following table:— [ 4 ]

A.—English, 1705 lines; French, 1678; as 101.6 to 100.

B.—English, 4105 lines; French, 3491; as 117.5 to 100.

C.—English, 1888 lines; French, 1849; as 102.1 to 100.

Thus, in A and C, the translation runs nearly line for line; but in B, the translator employs, on an average, 11 lines and three-quarters for every 10 of the original.

§ 7. Test II.—Dialect.—But the striking characteristic of Fragment B is the use in it of a Northern dialect. That this is due to the author, and not merely to the scribe, is obvious from the employment of Northern forms in rimes, where any change would destroy the rime altogether. This may be called the Dialect-test. Examples abound, and I only mention some of the most striking.

1. Use of the Northern pres. part. in -and. In l. 2263, we have wel sittand (for wel sitting), riming with hand. In l. 2708, we have wel doand (for wel doing), riming with fand. Even fand is a Northern form. Chaucer uses fond, riming with hond (Cant. Ta. A 4116, 4221, &c.), lond (A 702, &c.); cf. the subj. form fond-e, riming with hond-e, lond-e, bond-e (B 3521).

2. In l. 1853, we have the rimes thar, mar (though miswritten thore, more in MS. G.), where the Chaucerian forms there, more, would not rime at all. These are well-known Northern forms, as in Barbour's Bruce. So again, in l. 2215, we find mar, ar (though mar is written as more in MS. G.). In l. 2397, we find stat, hat; where hat is the Northern form of Chaucer's hoot, adj., 'hot.' So also, in 5399, we have North. wat instead of Ch. wot or woot, riming with estat. In l. 5542, we find the Northern certis (in place of Chaucer's certes), riming with is.

3. Chaucer (or his scribes) admit the use of the Northern til, in place of the Southern to, very sparingly; it occurs, e.g. in Cant. Ta. A 1478, before a vowel. But it never occurs after its case, nor at the end of a line. Yet, in fragment B, we twice find him til used finally, 4594, 4852.

4. The use of ado (for at do), in the sense of 'to do,' is also Northern; see the New E. Dict. It occurs in l. 5080, riming with go.

5. The dropping of the inflexional e, in the infin. mood or gerund, is also Northern. In fragment B, this is very common; as examples, take the rimes lyf, dryf, 1873; feet, lete (= leet), 1981; sit, flit, 2371; may, convay, 2427; may, assay, 2453; set, get, 2615; spring, thing, 2627; ly, by, 2629; ly, erly, 2645; &c. The Chaucerian [ 5 ] forms are dryv-e, let-e, flit-te, convey-e, assay-e, get-e, spring-e, ly-e. That the Northern forms are not due to the scribe, is obvious; for he usually avoids them where he can. Thus in l. 2309, he writes sitting instead of sittand; but in l. 2263, he could not avoid the form sittand, because of the rime.

§ 8. Test III.—The riming of -y with -y-ë.—With two intentional exceptions (both in the ballad metre of Sir Thopas, see note to Cant. Ta. B 2092), Chaucer never allows such a word as trewely (which etymologically ends in -y) to rime with French substantives in -y-ë, such as fol-y-ë, Ielos-y-ë (Ital. follia, gelosia). But in fragment B, examples abound; e. g. I, malady(e)[2], 1849; hastily, company(e), 1861; generally, vilany(e), 2179; worthy, curtesy(e), 2209; foly(e), by, 2493, 2521; curtesy(e), gladly, 2985; foly(e), utterly, 3171; foly(e), hastily, 3241; and many more.

This famous test, first proposed by Mr. Bradshaw, is a very simple but effective one; it separates the spurious from the genuine works of Chaucer with ease and certainty in all but a few cases, viz. cases wherein a spurious poem happens to satisfy the test; and these are rare indeed.

§ 9. Test IV.—Assonant rimes. Those who know nothing about the pronunciation of Middle English, and require an easy test, appreciable by any child who has a good ear, may observe this. Chaucer does not employ mere assonances, i. e. rimes in which only the vowel-sounds correspond. He does not rime take with shape, nor fame with lane. But the author of fragment B had no ear for this. He actually has such rimes as these: kepe, eke, 2125; shape, make, 2259; escape, make, 2753; take, scape, 3165; storm, corn, 4343; doun, tourn, 5469.

Other strange rimes.—Other rimes which occur here, but not in Chaucer, are these and others like them: aboute, swote, 1705 (already noticed); desyre, nere, 1785, 2441; thar (Ch. there), to-shar, 1857; Ioynt, queynt[3], 2037; soon (Ch. son-e), doon, 2377; abrede, forweried, 2563; anney (Ch. annoy), awey, 2675; desyre, manere, 2779; Ioye, convoye (Ch. conveye), 2915, &c. It is needless to multiply instances. [ 6 ]

§ 10. It would be easy to employ further tests; we might, for example, make a minute critical examination of the method in which the final -e is grammatically employed. But the results are always the same. We shall always find irrefragable proof that fragment B exhibits usages far different from those which occur in the undoubted works of Chaucer, and cannot possibly have proceeded from his pen. Repeated investigations, made by me during the past thirteen years, have always come round to this result, and it is not possible for future criticism to alter it.

Hence our first result is this. Fragment B, consisting of ll. 1706-5810 (4105 lines), containing more than fragments A and C together, and therefore more than half of 'the translation,' is not Chaucer's, but was composed by an author who, to say the least, frequently employed Northern English forms and phrases. Moreover, his translation is too diffuse; and, though spirited, it is not always accurate.

§ 11. Discussion of Fragment C.

I shall now speak of fragment C. The first noticeable point about it is, that it does not exhibit many of the peculiarities of B. There is nothing to indicate, with any certainty, a Northern origin, nor to connect it with B. In fact, we may readily conclude that B and C are by different authors. The sole question that remains, as far as we are now concerned, is this. Can we attribute it to Chaucer?

The answer, in this case, is not quite so easily given, because the differences between it and Chaucer's genuine works are less glaring and obvious than in the case above. Nevertheless, we at once find some good reasons for refraining to attribute it to our author.

§ 12. Rime-tests.—If, for instance, we apply the simple but effective test of the rimes of words ending in -y with those ending in -y-e, we at once find that this fragment fails to satisfy the text.

Examples: covertly, Ipocrisy(e), 6112; company(e), outerly, 6301; loteby, company(e), 6339; why, tregetry(e), 6373; company(e), I, 6875; mekely, trechery(e), 7319. These six instances, in less than 1900 lines, ought to make us hesitate.

If we look a little more closely, we find other indications which should make us hesitate still more. At l. 5919, we find hors (horse) riming with wors (worse); but Chaucer rimes wors with curs (Cant. Ta. [ 7 ] A 4349), and with pervers (Book Duch. 813). At l. 6045, we find fare, are; but Chaucer never uses are at the end of a line; he always uses been. At l. 6105, we find atte last, agast; but Chaucer only has atte last-e (which is never monosyllabic). At l. 6429, we find paci-ence, venge-aunce, a false rime which it would be libellous to attribute to Chaucer; and, at l. 6469, we find force, croce, which is still worse, and makes it doubtful whether it is worth while to go on. However, if we go a little further, we find the pl. form wrought riming with nought, 6565; but Chaucer usually has wrought-e, which would destroy the rime. This, however, is not decisive, since Chaucer has bisought for bisoughte, Cant. Ta. A. 4117, and brought for broughte, id. F. 1273. But when, at l. 6679, we find preched riming with teched, we feel at once that this is nothing in which Chaucer had a hand, for he certainly uses the form taughte (Prologue, 497), and as certainly does not invent such a form as praughte to rime with it. Another unpleasant feature is the use of the form Abstinaunce in l. 7483, to gain a rime to penaunce, whilst in l. 7505, only 22 lines lower down, we find Abstinence, to rime with sentence; but the original has similar variations.

§ 13. I will just mention, in conclusion, one more peculiarity to be found in fragment C. In the Cant. Tales, B 480 (and elsewhere), Chaucer uses such rimes as clerkes, derk is, and the like; but not very frequently. The author of fragment C was evidently much taken with this peculiarity, and gives us plenty of examples of it. Such are: requestis, honést is, 6039; places, place is, 6119; nede is, dedis, 6659; apert is, certis, 6799; chaieris, dere is, 6915; enquestes, honést is, 6977; prophetis, prophete is, 7093; ypocritis, spite is, 7253. Here are eight instances in less than 1900 lines. However, there are five examples (at ll. 19, 75, 387, 621, 1349) in the Hous of Fame, which contains 2158 lines in the same metre as our 'translation'; and there are 19 instances in the Cant. Tales.

We should also notice that the character called Bialacoil throughout Fragment B is invariably called Fair-Welcoming in C.

We should also remark how Dr. Lindner (Engl. Studien, xi. 172) came to the conclusion that Chaucer certainly never wrote fragment C. As to the rest he doubted, and with some reason; for he had not before him the idea of splitting lines 1-5810 into two fragments.

§ 14. A consideration of the above-mentioned facts, and of others similar to them, leads us to our second result, which is this. Fragment C, containing 1888 lines, and corresponding to ll. [ 8 ] 12564 of the French original, is neither by the author of fragment B, nor by Chaucer, but is not so glaringly unlike Chaucer's work as in the case of fragment B.

§ 15. Discussion of Fragment A.

It remains to consider fragment A. The first test to apply is that of rimes in -y and -y-e; and, when we remember how indiscriminately these are used in fragments B and C, it is at least instructive to observe the perfect regularity with which they are employed in fragment A. The student who is unacquainted with the subtle distinctions which this test introduces, and who probably is, on that account, predisposed to ignore it, may learn something new by the mere perusal of the examples here given.

1. Words that should, etymologically, end in -y (and not in -y-e) are here found riming together, and never rime with a word of the other class.

Examples: covertly, openly, 19; redily, erly, 93; by, I, 111; bisily, redily, 143; by, I, 163; I, by, 207; povrely, courtepy[4], 219; beggarly, by, 223; enemy, hardily, 269; awry[5], baggingly, 291; certeinly, tenderly, 331; prively, sikerly, 371; redily, by, 379; Pope-holy, prively, 415; I, openly, 501; queyntely, fetisly, 569; fetisly, richely, 577; only, uncouthly, 583; I, namely, 595; sikerly, erthely, 647; lustily, semely, 747; parfitly, sotilly, 771; queyntely, prively, 783; fetisly, richely, 837; sotilly, I, 1119; enemy[6], tristely, 1165; sotilly, therby, 1183; newely, by, 1205; fetisly, trewely, 1235; I, by, 1273; trewely, comunly, 1307; lustily, sikerly, 1319; merily, hastely, 1329; I, sikerly, 1549; I, craftely, 1567; openly, therby, 1585; diversely, verily, 1629; openly, by, 1637. Thirty-eight examples.

We here notice how frequently words in -ly rime together; but this peculiarity is Chaucerian; cf. semely, fetisly, C. T. prol. A 123, &c.

2. Words that, etymologically, should end in -y-e, rime together. These are of two sorts: (a) French substantives; and (b) words in -y, with an inflexional -e added.

Examples: (a) felony-e, vilany-e, 165; envy-e, masonry-e, 301; [ 9 ] company-e, curtesy-e, 639; melody-e, reverdy-e, 719; curtesy-e, company-e, 957; vilany-e, felony-e, 977; envy-e, company-e, 1069; chivalry-e, maistry-e, 1207; villany-e, sukkeny-e, 1231; envye, Pavie, 1653.

(b) dy-e, infin. mood, dry-e, dissyllabic adj. (A. S. drȳge), 1565.

(a) and (b) mixed: melody-e, F. sb., dy-e, infin. mood, 675; espy-e, gerund, curtesy-e, F. sb., 795; hy-e, dat. adj., maistry-e, 841; dy-e, gerund, flatery-e, F. sb., 1063; curtesy-e, F. sb., hy-e, dat. case, pl. adj., 1251; dy-e, infin. mood, remedy-e, F. sb., 1479. Seventeen examples. (In all, fifty-five examples.)

Thus, in more than fifty cases, the Chaucerian habit is maintained, and there is no instance to the contrary. Even the least trained reader may now fairly begin to believe that there is some value in this proposed test, and may see one reason for supposing that fragment A may be genuine.

§ 16. A still closer examination of other rimes tends to confirm this. There are no Northern forms (as in B), no merely assonant rimes (as in B), nor any false or bad or un-Chaucerian rimes (as in both B and C), except such as can be accounted for. The last remark refers to the fact that the scribe or the printer of Thynne's edition frequently misspells words so as to obscure the rime, whereas they rime perfectly when properly spelt; a fact which tells remarkably in favour of the possible genuineness of the fragment. Thus, at l. 29, Thynne prints befal, and at l. 30, al. Both forms are wrong; read befalle, alle. Here Thynne has, however, preserved the rime by making a double mistake; as in several other places. A more important instance is at l. 249, where the Glasgow MS. has farede, herede, a bad rime; but Thynne correctly has ferde, herde, as in Chaucer, Cant. Ta. A 1371. So again, at ll. 499, 673, where the Glasgow MS. is right (except in putting herd for herde in l. 673).

At l. 505, there is a false rime; but it is clearly due to a misreading, as explained in the notes. A similar difficulty, at l. 1341, is explicable in the same way.

§ 17. So far, there is no reason why fragment A may not be Chaucer's; and the more closely we examine it, the more probable does this supposition become. Dr. Kaluza has noticed, for instance, that the style of translation in fragment A is distinctly better, clearer, and more accurate than in fragment B. I find also another significant fact, viz. that in my essay written to shew that 'the translation' is not Chaucer's (written at a time when I unfortunately [ 10 ] regarded the whole translation as being the work of one writer, a position which is no longer tenable), nearly all my arguments were drawn from certain peculiarities contained in fragments B and C, especially the former. I have therefore nothing, of any consequence, to retract; nor do I even now find that I made any serious mistake.

§ 18. The third result may, accordingly, be arrived at thus. Seeing that Chaucer really translated the 'Roman de la Rose,' and that three fragments of English translations have come down to us, of which two cannot be his, whilst the third may be, we may provisionally accept fragment A as genuine; and we find that, the more closely we examine it, the more probable does its genuineness become.

§ 19. Summary.—Having now discussed the three fragments A, B, C, successively and separately (though in a different order), we may conveniently sum up the three results as follows.

1. Fragment A appears to be a real portion of Chaucer's own translation. Its occurrence, at the beginning, is, after all, just what we should expect. The scribe or editor would naturally follow it as far as it was extant; and when it failed, would as naturally piece it out with any other translation or translations to which he could gain access. This fragment ceases suddenly, at the end of l. 1705, in the middle of an incomplete sentence. The junction with the succeeding portion is clumsily managed, for it falsely assumes that the previous sentence is complete, and leads off with a false rime.

2. Fragment B is obviously from some other source, and is at once dissociated from both the other fragments by the facts (a) that it was originally written in a Northumbrian dialect, though this is somewhat concealed by the manipulation of the spelling by a later scribe; (b) that it was written in a more diffuse style, the matter being expanded to the extent, on an average, of nearly twelve lines to ten; (c) that many licences appear in the rimes, which sometimes degenerate into mere assonances; and (d) that it is less exact and less correct in its method of rendering the original.

3. After fragment B, there is a large gap in the story, more than 5000 lines of the original being missing. Hence Fragment C is from yet a third source, not much of which seems to have been accessible. It neither joins on to Fragment B, nor carries the story much further; and it comes to an end somewhat suddenly, at a point more than 9000 lines from the end of the original. It is, [ 11 ] however, both more correct than Fragment B, and more in Chaucer's style; though, at the same time, I cannot accept it as his.

§ 20. There is little that is surprising in this result. That translations of this then famous and popular French poem should have been attempted by many hands, is just what we should expect. At the same time, the enormous length of the original may very well have deterred even the most persevering of the translators from ever arriving at the far end of it. Chaucer's translation was evidently the work of his younger years, and the frequent use which he made of the French poem in his later works may have made him careless of his own version, if indeed he ever finished it, which may be doubted. All this, however, is mere speculation, and all that concerns us now is the net result. It is clear, that, in the 1705 lines here printed in the larger type, we have recovered all of Chaucer's work that we can ever hope to recover. With this we must needs rest satisfied, and it is a great gain to have even so much of it; the more so, when we remember how much reason there was to fear that the whole of Chaucer's work was lost. It was not until Dr. Kaluza happily hit upon the resolution of lines 1-5810 into two fragments, that Chaucer's portion was at last discovered.

§ 21. The External Evidence.

In what has preceded, we have drawn our conclusions from the most helpful form of evidence—the internal evidence. It remains to look at the external form of the poem, and to enquire how it has come down to us.

The apparent sources are two, viz. Thynne's edition of 1532 (reprinted in 1542, 1550, 1561, and at later dates), and a MS. in the Hunterian collection at Glasgow. But a very slight examination shews that these are nearly duplicate copies, both borrowed from one and the same original, which is now no longer extant. I shall denote these sources, for convenience, by the symbols Th., G., and O., meaning, respectively, Thynne, Glasgow MS., and the (lost) Original.

The resemblance of Th. and G. is very close; however, each sometimes corrects small faults in the other, and the collation of them is, on this account, frequently helpful. Both are remarkable for an extraordinary misarrangement of the material, in which respect they closely agree; and we are enabled, from this [ 12 ] circumstance, to say, definitely, that the C-portion of O. (i. e. their common original) was written (doubtless on vellum) in quires containing 8 leaves (or 16 pages) each, there being, on an average, 24 lines upon every page. Of these quires, the fourth had its leaves transposed, by mistake, when the MS. was bound, in such a manner that the middle pair of leaves of this quire was displaced, so as to come next the two outer pair of leaves; and this displacement was never suspected till of late years, nor ever (so far as I am aware[7]) fully appreciated and explained till now[8]. This displacement of the material was first noticed in Bell's edition, where the editor found it out by the simple process of comparing the English 'translation' with the French 'Roman'; but he gives no account of how it came about. But a closer investigation is useful as showing how exactly 'Th.' and 'G.' agree in following an original displacement in 'O.', or rather in the still older MS. from which the C-portion of O. was copied.

In the fourth sheet (as said above), the pair of middle leaves, containing its 7th, 8th, 9th, and 10th pages (G, H, I, K, with the contents recorded in note 2 below) was subtracted from the middle of the quire, and placed so that the 7th page (G) followed the 2nd (B), whilst at the same time, the 10th page (K) came to precede the 15th page (P). The resulting order of pages was, necessarily, A, B, G, H, C, D, E, F, L, M, N, O, I, K, P, Q; as is easily seen by help of a small paper model. And the resulting order of the lines was, accordingly, 6965-6988, 6989-7012, 7109-7133, 7134-7158, 7013-7036, 7037-60, 7061-84, 7085-7108, 7209-7232, 7233-7256, 7257-7280, 7281-7304, 7159-7183, 7184-7208, 7305-7328, 7329-7352; or, collecting the successive numbers, ... -7012, 7109-7158, 7013-7108, 7209-7304, 7159-7208, 7305, &c. And this is precisely the order found, both in Th. and G. [ 13 ]

We see further that the fourth and last quire of this C-portion of O. consisted of 7 leaves only, the rest being torn away. For 7 leaves containing 48 lines apiece give a total of 336 lines, which, added to 7352, make up 7688 lines; and, as 10 of the pages seem to have had 25 lines, we thus obtain 7698 lines as the number found in O.

The A-portion of O. was probably copied from a MS. containing usually 25 lines on a page, and occasionally 26. Four quires at 50 lines to the leaf give 32 × 50, or 1600 lines; and 2 leaves more give 100 lines, or 1700 lines in all. If 5 of the pages had 26 lines, we should thus make up the number, viz. 1705. Of the B-portion we can tell nothing, as we do not know how it was made to join on.

As O. was necessarily older than G., and G. is judged by experts[9] to be hardly later than 1440, it is probable that O. was written out not much later than 1430; we cannot say how much earlier, if earlier it was.

§ 22. G. (the Glasgow MS.) is a well-written MS., on vellum; the size of each page being about 11 inches by 7½, with wide margins, especially at the bottom. Each page contains about 24 lines, and each quire contains 8 leaves. The first quire is imperfect, the 1st leaf (ll. 1-44) and the 8th (ll. 333-380) being lost. Nine other leaves are also lost, containing ll. 1387-1482, 2395-2442, 3595-3690, and 7385-7576; for the contents of which (as of the former two) Th. remains the sole authority. The date of the MS. is about 1440; and its class-mark is V. 3. 7.

It begins at l. 45—'So mochel pris,' &c. At the top of the first extant leaf is the name of Thomas Griggs, a former owner. On a slip of parchment at the beginning is a note by A. Askew (from whom Hunter bought the MS.) to this effect:—'Tho. Martinus. Ex dono dom' Iacobi Sturgeon de Bury scī Edmundi in agro Suffolc: Artis Chirurgicæ Periti. Nov. 9, 1720.' It ends very abruptly in the following manner:—

'Ne half so lettred as am I
I am licenced boldely
To Reden in diuinite
And longe haue red
Explicit.'

[ 14 ] The third of these lines is incorrect, and the fourth is corrupt and imperfect; moreover, Thynne's copy gives four more lines after them. It would thus appear that G. was copied from O. at a later period than the MS. used by Thynne and now lost, viz. at a period when O. was somewhat damaged or torn at the end of its last page. A careful and exact copy of this MS. is now (in 1891) being printed for the Chaucer Society, edited by Dr. Kaluza.

§ 23. Th.—The version printed in Thynne's edition, 1532, and reprinted in 1542, 1550, 1561, &c. The first four editions, at least, are very much alike. The particular edition at first used by me for constructing the present text is that which I call the edition of 1550. (It is really undated, but that is about the date of it.) Its variations from the earlier editions are trifling, and I afterwards reduced all the readings to the standard of the first edition (1532). The MS. used by Thynne was obviously a copy of 'O.', as explained above; and it shews indications of being copied at an earlier date than 'G.', i. e. before 1440. On the whole, 'Th.' appears to me more correct than 'G.', and I have found it very serviceable. We learn from it, for example, that the scribe of 'G.' frequently dropped the prefix y- in past participles, giving l. 890 in the form 'For nought clad in silk was he,' instead of y-clad. Cf. ll. 892, 897, 900, &c.; see the foot-notes.

'Th.' supplies the deficiencies in G., viz. ll. 1-44, 333-380, &c., as well as four lines at the end; and suggests numerous corrections.

§ 24. The various later reprints of the 'Romaunt,' as in Speght (1598) and other editions, are merely less correct copies of 'Th.', and are not worth consulting. The only exceptions are the editions by Bell and Morris. Bell's text was the first for which 'G.' was consulted, and he follows the MS. as his general guide, filling up the deficiencies from Speght's edition, which he describes as 'corrupt and half-modernised.' Why he chose Speght in preference to Thynne, he does not tell us. In consequence, he has left lines incomplete in a large number of instances, owing to putting too much faith in the MS., and neglecting the better printed sources. Thus, in l. 890, he gives us 'clad' instead of 'y-clad'; where any of the printed texts would have set him right.

Morris's edition is 'printed from the unique MS. in the Hunterian Museum, Glasgow'; but contains numerous corrections, apparently from Thynne. Thus, in l. 890, he reads 'y-clad'; the y- being printed in italics to shew that it is not in the MS. [ 15 ]

§ 25. The Present Edition.

The present edition principally follows 'G.', but it has been collated with 'Th.' throughout. Besides this, a large number of spellings in Fragment A. have been slightly amended on definite principles, the rejected spellings being given in the footnotes, whenever they are of the slightest interest or importance. Silent alterations are changes such as i for y in king for kyng (l. 10), and whylom for whilom (in the same line), to distinguish vowel-length; the use of v for consonantal u in avisioun for auisioun (l. 9); the use of ee for (long) e in Iolitee for Iolite (l. 52) for the sake of clearness; and a few other alterations of the like kind, which make the text easier to read without at all affecting its accuracy. I have also altered the suffix -is into -es in such words as hertes for hertis (l. 76); and changed the suffixes -id and -ith into the more usual -ed and -eth, both of which are common in the MS., usually giving notice; and in other similar minute ways have made the text more like the usual texts of Chaucer in appearance. But in Fragments B and C such changes have been made more sparingly.

I have also corrected numerous absolute blunders, especially in the use of the final e. For example, in l. 125, I have no hesitation in printing wissh for wysshe, because the use of final e at the end of a strong past tense, in the first person singular, is obviously absurd. Owing to the care with which the two authorities, 'G.' and 'Th.', have been collated, and my constant reference to the French original, I have no hesitation in saying that the present edition, if fairly judged, will be found to be more correct than its predecessors. For Dr. Kaluza's help I am most grateful.

§ 26. For example, in l. 1188, all the editions have sarlynysh, there being no such word. It is an obvious error for Sarsinesshe (riming with fresshe); for the F. text has Sarrazinesche, i. e. Saracenic.

In l. 1201, the authorities and Bell have gousfaucoun, which Morris alters to gounfaucoun in his text, and to gownfaucoun in his glossary. But all of these are 'ghost-words,' i. e. non-existent. Seeing that the original has gonfanon, it is clear that Chaucer wrote gonfanoun, riming with renoun.

In l. 1379, late editions have lorey; in l. 1313, Bell has loreryes, which Morris alters to loreyes. There is no such word as lorey. Thynne has laurer, laurelles. Considering that loreres rimes with [ 16 ] oliveres, it is obvious that the right forms are lorer and loreres (French, loriers); see laurer in Stratmann.

In l. 1420, where the authorities have veluet, the modern editions have velvet. But the u (also written ou) was at that time a vowel, and velu-et (or velou-et) was trisyllabic, as the rhythm shews. The modern velvet seems to have arisen from a mistake.

Several other restorations of the text are pointed out in the notes, and I need not say more about them here.

N.B. After l. 4658, the lines in Morris's edition are misnumbered. His l. 4670 is really l. 4667; and so on. Also, 5700 is printed in the wrong place; and so is 6010; but without throwing out the numbering. Also, 6210 is only nine lines after 6200, throwing out the subsequent numbering, so that his l. 6220 is really 6216. At his l. 6232, 6231 is printed, and so counted; thus, his 6240 is really 6237. His 6380 is eleven lines after 6370, and is really 6378. After l. 7172, I insert two lines by translation, to fill up a slight gap. This makes his l. 7180 agree with my l. 7180, and brings his numbering right again.

For a few of the Notes, I am indebted to Bell's edition; but most of the work in them is my own.

§ 27. The French Text.

For some account of the famous French poem entitled 'Le Roman de la Rose,' see Morley's English Writers, 1889, iv. 1. It was commenced by Guillaume de Lorris, born at Lorris, in the valley of the Loire, who wrote it at the age of five-and-twenty, probably between the years 1200 and 1230[10]. He must have died young, as he left the poem incomplete, though it then extended to 4070 lines. It was continued, a little more than 40 years after Guillaume's death, by Jean de Meun (or Meung), born (as he tells us) at Meung-sur-Loire, and surnamed le Clopinel (i. e. the hobbler, the lame). See, for these facts, the French text, ll. 10601, 10603, 10626. He added 18004 lines, so that the whole poem finally extended to the enormous length of 22074 lines.

Jean de Meun was a man of a very different temperament from his predecessor. Guillaume de Lorris merely planned a fanciful allegorical love-poem, in which the loved one was represented as a Rose in a beautiful garden, and the lover as one who desired [ 17 ] to pluck it, but was hindered by various allegorical personages, such as Danger, Shame, Jealousy, and Fear, though assisted by others, such as Bel Accueil (Fair Reception), Frankness, Pity, and the like. But Jean de Meun took up the subject in a keener and more earnest spirit, inserting some powerful pieces of satire against the degraded state of many women of the day and against various corruptions of the church. This infused a newer life into the poem, and made it extremely popular and successful. We may look upon the former part, down to l. 4432 of the translation, as a pretty and courtly description of a fanciful dream, whilst the remaining portion intersperses with the general description many forcible remarks, of a satirical nature, on the manners of the time, and affords numerous specimens of the author's erudition. Jean de Meun was the author of several other pieces, including a poem which he called his 'Testament.' He probably lived into the beginning of the fourteenth century, and died about 1318.

§ 28. Professor Morley gives a brief analysis of the whole poem, which will be found to be a useful guide through the labyrinth of this rambling poem. The chief points in it are the following.

The poet's dream begins, after a brief introduction, with a description of allegorical personages, as seen painted on the outside of the walls of a garden, viz. Hate and Felony, Covetousness, &c.; ll. 147-474 of the translation.

We may next note a description of Idleness, the young girl who opens the door of the garden (531-599); of Sir Mirth (600-644); of the garden itself (645-732); again, of Sir Mirth, the lady Gladness, Cupid, or the God of Love, with his two bows and ten arrows, and his bachelor, named Sweet-looking (733-998). Next comes a company of dancers, such as Beauty, Riches, Largesse (Bounty), Frankness, Courtesy, and Idleness again (999-1308). The poet next describes the trees in the garden (1349-1408), and the wells in the same (1409-1454); especially the well of Narcissus, whose story is duly told (1455-1648). The Rose-tree (1649-1690). The Rose-bud (1691-1714).

At 1. 1705, Fragment A ends.

§ 29. Just at this point, the descriptions cease for a while, and the action, so to speak, begins. The God of Love seeks to wound the poet, or lover, with his arrows, and succeeds in doing so; after which he calls upon the lover to yield himself up as a prisoner, which he does (1715-2086). Love locks up the lover's heart, and gives him [ 18 ] full instructions for his behaviour (2087-2950); after which Love vanishes (2951-2966). The Rose-tree is defended by a hedge; the lover seeks the assistance of Bialacoil or Belacoil (i. e. Fair-Reception), but is warned off by Danger, Wicked-Tongue, and Shame (2967-3166); and at last, Fair-Reception flees away (3167-3188). At this juncture, Reason comes to the lover, and gives him good advice; but he rejects it, and she leaves him to himself (3189-3334).

He now seeks the help of a Friend, and Danger allows him to come a little nearer, but tells him he must not pass within the hedge (3335-3498). Frankness and Pity now assist him, and he enters the garden, rejoined by Fair-Reception (3499-3626). The Rose appears more beautiful than ever, and the lover, aided by Venus, kisses it (3627-3772). This leads to trouble; Wicked-tongue and Jealousy raise opposition, Danger is reproved, and becomes more watchful than before (3773-4144). Jealousy builds a strong tower of stone, to guard the Rose-tree; the gates of the tower are guarded by Danger, Shame, Dread, and Wicked-tongue (4145-4276); and Fair-Reception is imprisoned within it (4277-4314). The lover mourns, and is inclined to despair (4315-4432).

§ 30. At this point, the work of G. de Lorris ceases, and Jean de Meun begins by echoing the word 'despair,' and declaring that he will have none of it. The lover reconsiders his position (4433-4614). Reason (in somewhat of a new character) revisits the lover, and again instructs him, declaring how love is made up of contrarieties, and discussing the folly of youth and the self-restraint of old-age (4615-5134). The lover again rejects Reason's advice, who continues her argument, gives a definition of Friendship, and discusses the variability of Fortune (5135-5560), the value of Poverty (5561-5696), and the vanity of Covetousness (5697-5810).

§ 31. Here ends Fragment B, and a large gap occurs in the translation. The omitted portion of the French text continues the discourse of Reason, with examples from the stories of Virginia, Nero, and Crœsus, and references to the fall of Manfred (conquered by Charles of Anjou) and the fate of Conradin. But all this is wasted on the lover, whom Reason quits once more. The lover applies a second time to his Friend, who recommends bounty or bribery. Here Jean de Meun discourses on prodigality, on women who take presents, on the Age of Gold, and on jealous husbands, with much satire interspersed, and many allusions, as for example, to Penelope, Lucretia, Abelard, Hercules, and others. [ 19 ]

At last Love pities the lover, and descends to help him; and, with the further assistance of Bounty, Honour, and other barons of Love's court, proceeds to lay siege to the castle in which Jealousy has imprisoned Fair-Reception.

§ 32. Here begins Fragment C; in which the ranks of the besiegers are joined by other assistants of a doubtful and treacherous character, viz. False-Semblant and Constrained-Abstinence (5811-5876). Love discusses buying and selling, and the use of bounty and riches (5877-6016). Love's Barons ask Love to take False-Semblant and Constrained-Abstinence into his service (6017-6057). Love consents, but bids False-Semblant confess his true character (6058-6081). False-Semblant replies by truly exposing his own hypocrisy, with keen attacks upon religious hypocrites (6082-7334). Love now begins the assault upon the castle of Jealousy (7335-7352). A digression follows, regarding the outward appearance of False-Semblant and Constrained-Abstinence (7353-7420). The assailants advance to the gate guarded by Wicked-Tongue, who is harangued by Constrained-Abstinence (7421-7605), and by False-Semblant (7606-7696). And here the English version ends.

The above sketch gives a sufficient notion of the general contents of the poem. Of course the lover is ultimately successful, and carries off the Rose in triumph.

§ 33. It deserves to be noted, in conclusion, that, as the three Fragments of the English version, all taken together, represent less than a third of the French poem, we must not be surprised to find, as we do, that Chaucer's numerous allusions to, and citations from, the French poem, usually lie outside that part of it that happens to be translated. Still more often, they lie outside the part of it translated in Fragment A. Hence it seldom happens that we can compare his quotations with his own translation. In the chief instances where we can do so, we find that he has not repeated his own version verbatim, but has somewhat varied his expressions. I refer, in particular, to the Book of the Duchess, 284-6, as compared with Rom. Rose, 7-10; the same, 340-1, beside R.R., 130-1; the same, 410-2, beside R.R., 61-2; and the same, 419-426, 429-432, beside R.R., 1391-1403.

§ 34. In the present edition I have supplied the original French text, in the lower part of each page, as far as the end of Fragment A, where Chaucer's work ends. This text is exactly copied from the edition by M. Méon, published at Paris in four volumes in [ 20 ] 1813[11]. I omit, however, the occasional versified headings, which appear as summaries and are of no consequence. Throughout the notes I refer to the lines as numbered in this edition. The later edition by M. Michel is practically useless for the purpose of reference, as the numbering of the lines in it is strangely incorrect. For example, line 3408 is called 4008, and the whole number of lines is made out to be 22817, which is largely in excess of the truth.

Fragments B and C are printed in smaller type[12], to mark their distinction from Fragment A; and the corresponding French text is omitted, to save space.


  1. It is not very likely that he ever finished his translation, when we consider his frequent habit of leaving his works incomplete, and the enormous length of the French text (22074 lines in Méon's edition).
  2. By the spelling malady(e), I mean that the word must be pronounced malady in the text, whereas the Chaucerian form is malady-ë in four syllables. And so in other cases.
  3. Doubtless the author meant to employ the form quoynt or coint; but Chaucer has queynt, Cant. Ta. A 2333, G 752.
  4. Courtepy rimes with sobrely; Cant. Ta. prol. 289.
  5. As to awry (or awry-e?), we have little evidence beyond the present passage.
  6. Enemy rimes with I, Cant. Ta. A 1643, royally, id. 1793; &c.
  7. As it is the natural instinct of many critics to claim for themselves even small discoveries, I note that this paragraph was written in July, 1891, and that the curious, but not very important fact above announced, was first noticed by me some three months previously.
  8. The calculation is as follows. A quire of 16 pages, at 24 lines a page, contains 384 lines. Three such quires contain about 1152 lines, which, added to 5810 (in A and B), bring us to l. 6962 (say, 6964). In the fourth quire, if A, B, C, &c., be successive pages, these pages contained the lines following. A, 6965-6988; B, 6989-7012; C, 7013-36; D, 7037-60; E, 7061-84; F, 7085-7108; G (25 lines), 7109-33; H (25 lines), 7134-7158; I (25 lines), 7159-7183; K (25 lines), 7184-7208; L, 7209-32; M, 7233-56; N, 7257-80; O, 7281-7304; P, 7305-28; Q, 7329-52.
  9. I have been greatly assisted in this matter by D. Donaldson, Esq., who gave me some beautifully executed photographic copies of three pages of the MS., which I have shewn to many friends, including Mr. Bond and Mr. Thompson at the British Museum.
  10. The allusion to prince Edward, 'son of the lord of Windsor' (see note to l. 1250), is not in all the copies; so it may have been added afterwards. Edward I. was not born till 1239.
  11. Some copies are dated 1814; but I can detect no difference in them, except that the later copies have an additional frontispiece.
  12. Not in Wikisource edition-Uploader.