The Middle English poem of Erthe upon Erthe is one which occurs fairly frequently in fifteenth-century MSS. and even later. It was a favourite theme for Commonplace Books, and was frequently inserted on the spare leaves at the beginning or end of a manuscript. From the many texts of the poem which have survived, and from the fact that portions of it continued to be inscribed on walls and tombstones up to the beginning of the nineteenth century, a wide popularity may be deduced. The extant versions, moreover, point to a knowledge of the poem throughout the greater part of England, as well as in the south of Scotland. The grimness of the motive, based on the words Memento homo quod cinis es et in cinerem reverteris, allies the text both with the earlier group of poems relating to The Soul and the Body, and with the more or less contemporary Dance of Death, but whereas the two latter groups can claim a popularity which extended over western Europe, Erthe upon Erthe exists only in Middle English texts, and in one parallel Latin version.1 It is, indeed, difficult to see how the play upon the word earth on which the poem depends could have been reproduced with equal success in any language outside English, and the Latin version is distinctly inferior in this respect. There would seem, therefore, to be good reason for the assumption that Erthe upon Erthe is of English origin, belonging to the same class of literature as the English versions of the Soul and Body poems.

The earliest texts of the poem known to be extant are found in MSS. Harleian 2253 and 913, both dated about the beginning of the fourteenth century. The two texts vary greatly in length—MS. x Harl. 2253 consists of four lines as against seven six-lined stanzas in MS. Harl. 913—and the latter text has the parallel Latin rendering mentioned above, but they coincide so far as they go, and appear to represent a thirteenth or fourteenth-century type of the poem, which may be called the A version.2

Another poem of the same kind, which differs considerably from the A version, but is, in all probability, closely connected with it in origin, is common in fifteenth-century MSS. I have traced eighteen texts of this version, dating from the fifteenth to the seventeenth century, all of which represent or are based upon the same common type, though individual transcribers appear to have expanded the theme according to their own taste. Such additions may easily be distinguished, since they seldom succeed in maintaining either the grim simplicity, or the fundamental play upon the word earth, which characterize the genuine portions of the poem. This common fifteenth-century type may be called the B version.

Lastly, a single fifteenth-century MS. (Cambridge University Library, Ii, 4. 9) has preserved a text of the poem in which some attempt seems to have been made to combine the A with the B version. This text may be called the C version, or Cambridge text.

In the following pages an attempt has been made to justify the premises in part laid down already, and to show that the A and B versions may be traced back to a common source, and that this source was not only confined to England, but was itself English.

The following is a list of the manuscripts in which the poem occurs:—

MSS. of the A Version:

MSS. of the B Version:

MS. of the C Version:

The Cambridge Text. Cambr. Univ. Libr. Ii. 4. 9, fol. 67, r^o (fifteenth century). Eighty-two lines comprising twenty-two or twenty-three stanzas. The text is followed by a coloured picture of a young knight, standing on a hill with a skeleton below. A scroll proceeding from the knight has the words: Festina tempus et memento finis, while one proceeding from the skeleton runs: In omni opere memorare nouissima et in eternum non peccabis. Printed by Heuser, Kildare-Gedichte, p. 213.

The A version exists in two forms, one a short popular stanza of four lines (MS. Harl. 2253), apparently of the nature of a riddle, the other a longer poem of seven English and seven Latin stanzas (MS. Harl. 913), each English verse being followed by its xv Latin equivalent. The metrical form of the Latin verses is one often used in Latin poems of the twelfth and thirteenth centuries, a six-lined stanza, rimed aaaabb, with the rhythm of the well-known

méum ést propósitúm   |   ín tabérna móri.

The English verses are also in the form of a six-lined stanza aaaabb, but the first four lines have the same loose four-stress rhythm as the lines in MS. Harl. 2253, and the concluding couplet is on the principle of the septenarius. Both the English and the Latin lines rime at the caesura as well as at the end of the line, but this is less uniformly the rule in the English verses. There is close verbal connexion between the four lines in MS. Harl. 2253, and the opening lines of the longer poem, as will appear from a comparison of the two:—

MS. Harl. 2253.

MS. Harl. 913.

The connexion between these two versions might be explained in two ways. The short version of MS. Harl. 2253 may be the beginning of a transcript of the longer poem in which the scribe broke off because his memory failed him, or because he was only acquainted with a popular version of the opening lines. On the other hand, the short version may be the older, and the more learned composer of the poem in MS. Harl. 913 may have been elaborating this and other such riddling stanzas current at the time. But any attempt to decide between these two possibilities must necessarily depend upon the conclusion formed as to the relation of the Latin stanzas in MS. Harl. 913 to their English equivalents, and this question will be more conveniently discussed in connexion with the general origin of the Erthe upon Erthe poems. As regards the date of the two MSS., MS. Harl. 2253 is generally ascribed to the beginning of the fourteenth century, xvi and the Kildare MS. (MS. Harl. 913) is dated c. 1308 by Crofton Croker, c. 1308 to 1330 by Heuser, while Paul Meyer is of opinion that it may belong to an earlier period still. The dialect of both poems is South Midland, probably of the western part of the district. MS. Harl. 2253, which is commonly associated with Leominster, has heuede (4). MS. Harl. 913 has lutil, schrud, muntid, heo, mon, lond, and S. Midl. forms of verbs. We have therefore two types of the A version, standing in close verbal relation to each other, of much the same date and dialect, and representing in all probability the kind of Erthe poem current at the end of the thirteenth century in the South-west Midland district.

As will appear from the foregoing account of the MSS., the eighteen texts of the B version vary considerably in length, many of them introducing stanzas which do not recur elsewhere. A comparison of the number and arrangement of the stanzas in each text is given on the next page, the stanzas being numbered according to the order of their arrangement in the text to which they belong, and the corresponding stanzas in the various texts grouped under columns. MSS. Thornton, Selden, and Egerton have no mark of strophic division, but fall naturally into mono-rimed stanzas of four lines. All the remaining texts are arranged in four-lined stanzas with mono-rime,9 with the exception of MS. Porkington, which represents an evident expansion of the original metrical scheme, an additional long line being attached to each stanza by means of a short bob-line, giving a six-lined stanza, aaaabb. In MS. Rawl. Poet. each long line is written as two short lines, so that the usual four-lined stanza appears in this text as two stanzas, each consisting of four half-lines. This arrangement is facilitated by the regular internal rime on the word erthe. The order of the fifteenth-century MSS. of the B version observed in the table corresponds to that in the foregoing list of MSS., and in the printed text, and is not always strictly chronological, it being more convenient for purposes of comparison to group the texts according to their length. It will be seen that the three late texts (MSS. Harl. 984, Maitland, and Reidpeth) revert to the normal seven-stanza type, and that this appears to have been the form of the poem known to the compiler of the Cambridge text, a comparison of which is added.

The “Independent Stanzas”, here shown as a separate table, were printed as the last column of the main table, following “Common Stanzas”.

It will be seen from the table that eleven of these texts have seven stanzas in common, and that fifteen of them have five in common. Of the three remaining texts, MS. Harl. 984 has a missing leaf, but would clearly appear to belong to the seven-stanza type, raising the above numbers to twelve texts of seven stanzas, and sixteen of five. MS. Selden again obviously represents the usual seven-stanza type with the accidental omission of verse 5. MS. Titus has four of the customary five verses, breaks off to follow the arrangement of the Lambeth MS., and comes to an end after copying two of the additional verses in the Lambeth text before reaching the usual fifth verse. Assuming that it represents a transcription of the Lambeth text, MS. Titus might be classed with the five-stanza type, or possibly, like MS. Lambeth, with the seven-stanza type. It may therefore be assumed that all eighteen of the B texts have five stanzas in common, or are based upon such a common type, and that thirteen, or possibly fourteen of them, represent a common type with seven stanzas, six of which are further found in the Cambridge text. These common stanzas vary very little in the different MSS. as regards either the actual text or the order of lines and stanzas, and it seems probable that the normal B version consisted of seven stanzas, ending with a personal exhortation which has been omitted, or possibly not yet added, in five of the texts. In four MSS.—Lambeth, Laud, Rawl. P., and Harl. 4486—an interesting final stanza, containing a prayer, has been added. Three of these texts, MSS. Lamb., Laud, and Rawl. P., correspond in three other additional stanzas, which seems to point to some closer relationship between them, and two, or more strictly one and a half, of these additional stanzas are also found in MS. Titus, which appears to be a transcript of the Lambeth text. The scribe of MS. Titus followed the Lambeth text until he reached the middle of verse 6, when he apparently wearied of the task, and broke off with a new couplet of his own, entirely foreign in idea and metre to the Erthe upon Erthe poems:—

The text, as a whole, is badly written with many erasures, and points to a careless hand.

The additional stanzas cited in the table as independent contain xix mere variations on the main theme, and it is highly probable that the more expanded texts are the later, and represent individual additions to a popular poem, since they generally fail to maintain the internal rime on the word erthe which is an evident characteristic of the genuine verses. In the case of the five MSS. in question, MS. Harl. 4486 might be taken to represent the original type, and MSS. Lamb., Laud11, and Titus an expansion of this, while the author of Rawl. P. was obviously acquainted with the Lambeth text, or its original, and added to it certain stanzas of his own, leaving out three of the verses in Lambeth to make room for these. Whether the eighth stanza which MSS. Harl. 4486, Lamb., Laud, and Rawl. P. have in common belongs to the original type of the B version, or was itself a later addition, can scarcely be determined, but as it seems to be confined to these four texts, the latter view is perhaps the more probable. It must, however, have been added early, as it occurs already in MSS. Lamb. and Laud before 1450, and preserves the principle of the internal rime on erthe. The relative dates of MSS. Lambeth and Rawl. P. as fixed by Furnivall and Madden (MS. Lamb. 1430-1450, R. P. after 1450) would bear out this theory of the relationship between these two texts, and it may further be noted that both have the same prefatory De terra plasmasti me, otherwise found only in MS. Harl. 1671, and that both exhibit the same tendency to employ a direct personal mode of address, and to lengthen out the original text by superfluous words.

Cf. for example, MS. Harl. 4486, verse 5 (so MS. Laud, verse 8)—

with MS. Lamb. verse 8—

and MS. Rawl. P. verse 11—

The exact date of the text in MS. Titus is indeterminate, but, as stated above, it is evidently based on MS. Lambeth or its original, and might be ascribed to c. 1450 or later. The text in MS. Harl. 4486 has been added by some later owner of the MS. on the last leaves of a fifteenth-century transcript of Le Livre de Sydrac. The handwriting of Erthe upon Erthe is also fifteenth century, but the exact date again cannot be determined. The text, however, is far simpler and nearer to the original than that of the other four MSS., and evidently represents an earlier type than these, though the actual transcript may be later.

With the exception of these five MSS., it is not easy to group the eighteen texts of the B version on any system based upon the additional stanzas, since these fail to bear out any theory as to closer relationship between individual MSS., though the connexion of ideas is often close owing to the similarity of the theme. Thus the nine additional stanzas in MS. Balliol contain a digression upon the nine worthies with an interesting reference in verse 12 to the Dance of Powlis, i.e. the Dance of Death formerly depicted outside St. Paul’s Cathedral (v. Notes, p. 36). It is in the Cambridge text alone that the additional stanzas supply an interesting connexion with the A version, which places this text, unfortunately corrupt and difficult to decipher, in an important position as a link between A and B.

With regard to possible relationships dependent upon variations in the order or arrangement of the lines in the seven common stanzas, it may be pointed out that the first verse in MS. Egerton consists of three lines only, the usual second line being omitted, and that both MS. Harl. 1671 and MS. Porkington omit the same line, though each of these supplies a new and independent fourth line to fill the gap:—

(MS. Egerton 1995)

(MS. Harl. 1671)

(MS. Porkington 10)

It is obvious that these new lines are an afterthought, especially in the case of MS. Porkington, where the rime-word þouȝte has to be repeated. Possibly these three texts depend upon a common original in which the usual second line Erth hath gotyn vppon erth a dygnyte of noght was lacking, or MS. Egerton may have been the original of the other two. But MS. Harl. 1671 varies from the other two in the first line also, using a version which is otherwise confined to the Cambridge text—

Erthe apon erthe ys waxyne and wrought—

and both it and MS. Porkington begin erthe upon erthe like the later texts, as opposed to the more usual erthe owte of erthe, so that there is no clear evidence of a closer relationship between these three texts.

In verse 4, again, an inversion of the customary order of the second or third lines is common to MSS. Rawl. C., Porkington, Maitland, Reidpeth, and the Stratford-on-Avon inscription, but the verse easily lends itself to transposition of the kind, and in MS. Rawl. C. the usual first line is also put third, so that the order of lines as compared with the normal arrangement becomes 2. 3. 1. 4. Beyond the self-evident fact that the Maitland and Reidpeth MSS. must be grouped together, no relationship of the MSS. can be deduced from this transposition, though it may point to a second popular version with inversion of lines 2 and 3.

One of the most important differences of reading in the common stanzas occurs in the first line of the poem, where twelve of the eighteen MSS. read erthe out of erthe, while the remaining six, as well as the Cambridge text, have erthe upon erthe. Three of these six are definitely later transcripts: MS. Porkington is obviously a later modification of the original four-lined stanza, and MSS. Maitland and Reidpeth belong to the late sixteenth and early seventeenth centuries respectively; the beginning of MS. Harl. 984 xxii is not preserved, and the remaining two texts, MSS. Selden and Harl. 1671, belong to c. 1450, while the Cambridge text, as will be shown later, cannot be regarded as original. Evidently erthe owt of erthe was the original reading, but the version erthe upon erthe was introduced early, and appears to have survived the other. A similar change occurs in the last line of verse 2, where MS. Harl. 1671 and the Stratford text substitute erth upon erth for out of, from, of, of the other texts, and again in the third line of verse 4 (l. 2 in the texts mentioned above as transposing these lines) where the same two MSS. read erth upon erth for the normal erth unto (into, to) erthe; also in the fourth line of verse 7, where MSS. Harl. 4486, Lamb., Laud, Maitland, and Reidpeth read upon for owte of. Now the last two lines of the first verse of the poem invariably use the phrase erth upon erthe, and it occurs repeatedly throughout the poem as a synonym for man: e.g. verse 2, line 1; 3, ll. 1, 3; 4, ll. 1, 2 (or 3); 5, l. 3; 6, ll. 1, 3; 7, l. 1. It was very natural that the common phrase, and the one best adapted to serve as a title to the poem, should tend to replace others, but it seems probable that wherever the substitution occurs it may be taken as due to a later tradition, and consequently as a proof of non-originality or comparative lateness in the text in which it is found. A similar change, and one to be explained in a similar way, is the introduction of wonderly for wyckydly in the first line of verse 7 on the analogy of the first line of the poem, which occurs in MSS. Harl. 1671 and Stratford, and also in the late MSS. Maitland and Reidpeth.

Other variations of reading are less noteworthy. In the second line of verse 1, ten MSS., ranging from the early Thornton and Lambeth to the late Maitland and Reidpeth, read dignite, while the others vary between nobley (MS. Brighton, cf. the Cambridge text), nobul þyng (Billyng), worschyp (Selden), and an abbey, perhaps an error for nobley (Harl. 4486). The remaining three MSS. omit the line. In the fourth line of verse 2, the alliterative piteous parting of MSS. Billyng, Egerton, Brighton, Harl. 4486, Lamb., Laud, Titus, and Rawl. P., is replaced by hard parting not only in the Stratford text and in the later MSS. (Porkington, Balliol, Maitland, Reidpeth), but also in MSS. Thornton and Rawl. C., while other readings are dolful (MS. Selden, cf. the Cambridge text) and heuy (MS. Harl. 1671). It is difficult here to decide between piteous xxiii and hard, but the preference should probably rest with the alliterative phrase. In the fourth line of verse 3, the alliterative scharpe schowres is evidently the original reading, and it occurs in all texts except Stratford, Rawl. P., and Balliol.

In the first line of verse 4, erthe goeth upon erthe as moulde upon moulde occurs in thirteen texts, and two others (Stratford and Balliol, cf. also the Cambridge text) keep the rime mould while altering the line. The other two readings found, colde opon colde (Rawl. C.), and golde appone golde (Thornton), are obviously non-original, particularly the latter, which repeats the rime-word gold in two successive lines.

Other variations and occasional transpositions of lines occur in individual MSS., but are unimportant.

It will thus be seen that the popular traditional version of the poem tended to become modified, and even corrupt, already in the fifteenth century, and that such modifications are usually more apparent in the later texts. It is also evident that individual transcribers felt themselves at liberty to expand the traditional version, and that many tried their hand at such variations on the original theme, but the striking absence of proof of relationship outside the seven stanzas of the normal version, as well as the frequent unimportant variations found in the common stanzas, seems to point clearly to the conclusion that the original was a popular poem of seven, or possibly only five, stanzas, widely known over England, and that the more simple and naïve of the seventeen texts extant are also more genuine, and nearer to the original.

Many of the texts are accompanied by a short prefatory or concluding verse in English or Latin. The English verse—

occurs as a preface in MSS. Harl. 4486 and 1671, Lambeth, Laud, Rawl. P., and Egerton, and as a conclusion in Billyng’s text. The Latin Memento homo quod cinis es et in cinerem reverteris occurs, in full or in part, in MSS. Harl. 4486, Egerton, Rawl. C., Lambeth, and Billyng, and De terra plasmasti me in MSS. Harl. 1671, Lambeth, and Rawl. P. The two stanzas in rime royal on the Procese of Dethe which immediately precede Erthe upon Erthe in the Porkington MS. are transcribed as a separate poem, and if not separate, would rather belong to the preceding text, a translation xxiv of the Latin Visio Philiberti in rime royal, than to Erthe upon Erthe. The latter poem often accompanies either a Dance of Death or one of the numerous Soul and Body dialogues, no doubt because of the similarity of the theme, but it is not necessary to regard these kindred poems as forming an essential part of each other. So in the Balliol MS., Erthe upon Erthe is preceded by an eight-lined Latin stanza on the theme vado mori, which is probably part of a Dance of Death. Here again no basis for a grouping of the MSS. can be found.

The two late texts—MSS. Maitland and Reidpeth—represent a Lowland Scots version of the poem, and are obviously copies of the same original. Probably the Reidpeth text is a transcription of the Maitland, but it contains some obvious misreadings of it, as in verse 3, line 3, bowris (Maitl.), towris (Reidpeth) repeating the rime-word; 5, l. 20, within (Maitl.), with (Reidpeth). The Maitland MS., compiled c. 1555-1585, adds the colophon quod Marsar. The later Reidpeth MS., 1622-1623, concludes with the words quod Dumbar. Mersar, or Marsar, is mentioned in Dunbar’s Lament for the Makaris, and is usually identified with a William Mersar of the household of James IV, mentioned 1500 to 1503. In any case, if he were a contemporary of Dunbar, he could scarcely be assigned to a sufficiently early date to account for the widespread popularity of Erthe upon Erthe all over England in 1450, and the fact that the two MSS. assign the poem to different authors, of whom Dunbar is manifestly impossible, and Mersar at least improbable, may be explained as an instance of that readiness of posterity to attach a known name to a work of unknown origin, of which other examples are not wanting. It is, however, of interest to find that the poem had made its way to Scotland by 1550 or thereabouts.

As regards dialect, the majority of the MSS. of the B version show traces of Northern dialect, most of them preserving the Nth. plural in -is in the rimes touris, schowrys, &c. In verse 3 also the majority of the texts have the Nth. bigged or biggid, but six (MSS. Billyng, Egerton, Rawl. P., Porkington, Balliol, and the Stratford text) use the Midl. or Sth. bilded or billed. In verse 4 the rime requires the form wold rather than the common Nth. wald, and even the Maitland MS. retains wold for the sake of the rime, whereas MS. Reidpeth substitutes wald, sacrificing the rime. xxv MSS. Thornton and Rawl. C. show distinct Nth. features, such as the verb-endings -is (pres. ind. 3 sg.), -and (pres. part.), -id, -it, -in (past part.), and MS. Rawl. C. has the Nth. whate gates at þu gase riming with fase (foes). But few of the MSS. represent pure dialect-forms, and an investigation of the dialect of the texts is of little assistance towards determining that of the original poem. Such evidence as exists points, on the whole, to the North Midland district, and a widespread popularity in the North, which led to the later knowledge of the poem across the Border, but the popularity was evidently not confined to the North, and Southern as well as Northern forms may be traced in both early and late transcripts.

The Cambridge MS., as has been already stated, combines portions of both the A and the B version with several independent stanzas. At first sight it might appear to represent a transitional stage in the development of the B from the A type, but closer examination shows that this is not the case, and that the text is merely a later compilation from the two. The writer must have had some knowledge both of the longer A version represented by MS. Harl. 913, and of the common seven-stanza B type, and seems to have tried to combine his recollections in one poem, halting between the four-lined and six-lined stanza, repeating himself here and there, and adding certain new verses of his own. There is no grouping into stanzas in the MS., but a division is easily made by the rimes, and these give mono-rimed stanzas of four lines chiefly, with one of six lines, and some fragmentary ones of two or three. In one case a stanza has been broken up and the two couplets inserted at different points (ll. 9-10, 27-28). As has been shown in the table of MSS. of the B version, six verses of the B type may be traced, while four verses show distinct correspondence with A, and eleven are independent of either. A comparison of the similar lines follows:—

The additional verses in MS. Cambr. bear some slight resemblance to other additional lines found in MSS. of the B type, and this is interesting as showing that the writer worked on the same lines in expanding his text, and was perhaps acquainted with some of the longer B texts. On the other hand characteristic differences in the treatment of the theme would seem to support the view that these verses are really individual additions and not derived from any of the other texts. The lines in question are given below:—

It is therefore evident that the Cambridge text shows knowledge of both the A and the B versions, but the text in its existing form must represent either a corrupt copy of the original with frequent dislocation of lines, or, what is perhaps more likely from the instances of repetition of the same words or ideas which occur, a clumsy compilation from the two made by some one who perhaps had B before him and remembered portions of A imperfectly. xxix Such repetitions occur in verses 2 and 18, the latter repeating three of the rime-words of the former verse, as well as the phrase scharpe schowris; and again in verses 4 and 19, and in verses 6, 7, and 13. In any case the text must be regarded as later than the A and B versions, and not as forming a link between them. The dialect is Northern, but not uniformly so.

The question as to the source of the poem Erthe upon Erthe, and the relationship of the A and B versions to the original, and to each other, is a difficult one. The existence of a parallel Latin version in one of the oldest MSS. is clearly an important point to be taken into consideration in any attempt at an investigation of the origin of the poem, and it will be well before proceeding further to form some conclusion as to the relation in which the English and Latin stanzas in MS. Harl. 913 stand to each other. The correspondence of the two versions is not strictly verbal, but it is evident that either the English or the Latin stanzas represent a rather free rendering of the verses which accompany them. In favour of a Latin origin it may be pointed out that the metrical form of the Latin stanzas is one frequently employed in Latin poems of the time, that the subject is a favourite monastic theme, and that the manner of the poem is in keeping with contemporary Anglo-Latin compositions, such as the well-known Cur mundus militat sub vana gloria. The natural tendency would be to attribute a poem of the kind to Latin origin, especially if, as in this case, a Latin version were forthcoming.

On the other hand, it may be pointed out that the Latin text is not known to exist in any other MS., and appears indeed to have no separate existence from the English stanzas which accompany it, whereas English texts of the poem without trace of a Latin rendering or original are very common.14 The text was one frequently used in epitaphs, but no Latin epitaph of the kind is known to have existed, although Latin was commonly used in epitaphs at the time when the poem was most widely popular.

Further, word-plays of the kind found here upon the word erthe are certainly not common in Latin verse of the time, and the Latin xxx text does not render the play as effectively as the English does, employing alternately the three terms terra, vesta, humus, in place of the English erthe, and failing to maintain these consistently. The play on the word earth, which is the most essential feature of the poem, could not have been given with the same effect as in English either in Latin or in any mediaeval language.15

Thirdly, in support of an English origin it may be urged that close verbal connexion can be traced between the English text of both versions, but more especially of the earlier (A), and other poems dating from the twelfth to the fifteenth century, particularly the various Dialogues of The Soul and the Body:—

The play upon the word earth recurs in other English poems. Cf. A Song on the Times (MS. Harl. 913), early fourteenth century—

where the idea and the two rime-words are the same as in MS. Harl. 2253—

It will be remembered that these two MSS. (Harl. 913 and 2253) are the two which preserve texts of the A version, and the opening lines of the Song on the Times would appear to give further proof of a connexion between the two A texts.

Further, in MS. Lansdowne 762 (v. Reliquiae Antiquae I. 260), under the heading Terram terra tegat, occur these lines:—

In this case the English words are evidently based upon the Latin phrase, but this does not disprove an English origin for the poem Erthe upon Erthe, since any verses of the kind must ultimately have been based on the idea that man is dust, and the idea itself must have been first presented and have become widely known through such Latin elegiac phrases as Memento homo quod cinis es et in cinerem reverteris, or De terra plasmasti me, both of which so frequently accompany Erthe upon Erthe, or as the above cited xxxiii Terram terra tegat. The verse in MS. Lansdowne might rather be considered as supplying further proof of the popular tendency to replace such phrases by English verses, expressing the same idea, but themselves English, not Latin in origin, and making the most of the possible word-play. Such word-plays were evidently popular between the thirteenth and fifteenth centuries. Cf. the well-known passage in Piers Plowman, c. xxi. 389.

In view of this evidence, I am inclined to think that the Latin version in MS. Harl. 913 is the translation, and the English the original, and that the oldest form of Erthe upon Erthe which has been preserved is that found in the four lines in MS. Harl. 2253:—

Erþe toc of erþe erþe wyþ woh &c.

Short riddling stanzas of the kind, based upon the Latin phrases mentioned above, may have been popular in the thirteenth century, and this particular one was evidently known and used by the author of the Song on the Times.21 The writer of the version preserved in MS. Harl. 913 seems to have been a more learned man, acquainted with poems like the Dialogues between the Soul and the Body, who elaborated the four lines of MS. Harl. 2253, and perhaps other verses of the same kind, into a poem of seven six-lined stanzas, the additional couplet often introducing a new idea precisely as in the case of the similarly expanded verse-form in MS. Porkington. Either this man or a later transcriber appears to have added the Latin rendering which accompanies the poem, and to have further exercised himself in varying the word-play. Heuser22 points out that the mistakes in the MS. would support the view that the English text is a copy of an original in another dialect, and it is possible that the Latin version belongs to this MS. alone, since a second poem in the same MS. is accompanied by an unfinished translation into Latin.

This theory as to the origin of the two texts of the A version receives further support from the fact that it also accounts most satisfactorily for the development and popularity of the B version. Apart from the play on the word erthe and the similarity of the xxxiv theme, there is only one point of close verbal connexion between the two versions. In MS. Harl. 913 (A) the sixth stanza runs as follows:—

In the B version, the rimes gold : mold, toures : boures, regularly recur in the third and fourth stanzas, and line 5 of the A text is preserved in slightly modified form in the first line of verse 3:— (MS. Harl. 4486, vv. 3 and 4)

In the Cambridge text the rime-words towres : bours are introduced twice over, representing both the versions given above:—

as in the A version;

as in the B version.

The two stanzas of the B version which contain these rime-words are the two which recur most frequently on tombstones and mural inscriptions, and it seems possible that they represent a second early form of the Erthe poems. It is evident that the rime-words gold : mold, bowres : towres, depend upon an early tradition. Probably verses similar to the short stanza in MS. Harl. 2253, and containing these words, were in existence before the learned writer of the longer A text in MS. Harl. 913 introduced them in his poem, and, becoming widely known, formed the nucleus of the B version. Both the A and the B versions might therefore be held to depend upon popular stanzas of this kind, xxxv which gave rise about the end of the thirteenth century to the long poem of MS. Harl. 913, and during the fourteenth century to the original of the B version, a poem in seven four-lined stanzas. The earlier version is connected more particularly with the Southwest Midland district; the later seems to have originated rather in the North or North Midlands, but it soon became known all over England, and is found in the South of Scotland shortly after 1500. Only one fifteenth-century writer, the author of the Cambridge text, shows direct knowledge of the A text, but the B version was evidently widely known, and a favourite theme for additions and modifications. On tombstones and mural inscriptions it survived up to the nineteenth century.

As has been already pointed out, the Middle English texts of Erthe upon Erthe occur for the most part in the Commonplace Books of the day, often on the spare leaves at the beginning or end of the MS., as if the collector or some later owner had been struck by the poem and anxious to preserve it. That this interest was not confined to the fifteenth century is shown by the occurrence of the text in the Maitland and Reidpeth MSS. A still later instance of it occurs in the Pillerton Hersey Registers, dating from 1559 onwards, where the following verse has been scribbled on the last leaf, probably by some seventeenth-century clerk (cf. C. C. Stopes, Athenaeum, Sept. 19, 1908):—

Here the first two lines represent a corrupt type of the same lines in verse 3 of the B version, while the rimes wroght : nought recall verse 1.

Another interesting trace of a late popular version is mentioned in the Gentleman’s Magazine for March, 1824, where a certain Mr. J. Lawrence tells how he was invited, during a visit to Beaumont Hall, Essex, to see the following inscription, written and decorated by a cow-boy on an attic wall:—

Here portions of verses 3 and 4 of the B version have been combined as in the epitaphs at Melrose and Clerkenwell cited below, pointing either to a corrupt popular version of the B text, or possibly to an earlier type23 in which the rimes gold : mold, &c. were immediately associated with the rimes towres : bowres as in A (MS. Harl. 913, v. 6). The former assumption is the more probable, since the verse appears to be directly based upon stanzas 3 and 4 of the usual B version.

The majority of the later instances of the text occur on tombstones or memorial tablets. The poem was peculiarly adapted for this purpose, based as it was on the very words of the Burial Service. Indeed, the short verses from which it is here assumed to have originated might well be supposed to have been written in the first place as epitaphs, if evidence of the use of English epitaphs in the thirteenth century24 were forthcoming. As has been already stated, the seven verses of the normal B version occurred in full among the mural paintings in the Chapel of the Holy Trinity at Stratford-on-Avon, belonging to the Guild of the Holy Cross, where they appear to have been used as a monumental inscription already in the latter part of the fifteenth century.

A well-known late instance of the text is the inscription on a tombstone in the parish churchyard which surrounds Melrose Abbey, mentioned by Scott. The stone is headed as follows:—

Memento Mori.
Here lyes James Ramsay, portioner of Melrose, who died
July 15th, 1761.

On the back is the following verse:—

This was translated into German by Theodor Fontane (Poems, 4th edit., Berlin, 1892, p. 447). Cf. Fiedler, Mod. Lang. Review, April 1908.

Other inscriptions are as follows:—

On an old brass, quoted by W. Williams, Notes and Queries, I. vii. 577, and thought by him to belong to the Church of St. Helen’s, London25:—

‘Here lyeth y^e bodyes of
James Pomley, y^e sonne of ould
Dominick Pomley and Jane his
wyfe: y^e said James deceased y^e 7th
day of Januarie Anno Domini 1592
he beyng of y^e age of 88 years, &
y^e sayd Jane deceased y^e — day
of — D —

On a tomb at Edmonton of unknown date (possibly sixteenth century), mentioned by Weever (Ancient Funerall Monuments) in 1631, and by Pettigrew (Chronicles of the Tombs, p. 67) in 1857:—

Formerly on a headstone in St. James’s Churchyard, Clerkenwell, deciphered about 1812, but already lost in 1851, probably owing to the dismantling of the churchyard. (Cf. Notes and Queries, III. i. 389):—

Formerly on a tombstone at St. Martin’s, Ludgate, to Florens xxxviii Caldwell esq. of London & Ann Mary Wilde, his wife (Pettigrew, p. 67)26:—

This interesting monument has unfortunately disappeared. Doubtless there are many other traces of the poem to be found, but it appears to have been rarely used on tombstones after 1700,27 and earlier monuments, unless specially preserved, are rarely decipherable at the present day.

Erthe upon Erthe cannot be said to possess great literary value in itself. The interest of the poem lies chiefly in its evident popularity, and in the insight it gives into the kind of literature which became popular in the Middle Ages. It belongs essentially to the same class as the Soul and Body Poems, and the Dance of Death. In the early days of its introduction into Western Europe, Christianity made great use in its appeal to the mass of the people of the fear of death and dread of the Judgement. The early monastic writers dwelt upon the idea of man’s mortality and decay, and the transitoriness of human rank and pleasure. Hence the frequency with which such themes as the Dance of Death were treated in literature and in art. Closely allied with this idea of the fleeting nature of earthly things, and to some extent a result of it, was the xxxix conception of the separation of man’s bodily from his spiritual self which pervades all mediaeval post-Christian literature. In Old English times already, this sense of a sharp division between the two is embodied in No. xliv of the O.E. Riddles:—

This sets forth the same conception of the duality in man as is represented in the O.E. Speech of the Soul to the Body, and in the whole group of Soul and Body poems, and the idea recurs constantly in other monastic texts, cf. Morris, O. E. Miscellany, iii (Sinners Beware), p. 83:—

MS. Harl. 2253, fol. 106, v^o, l. 7: þe fleysh stont aȝeyn þe gost.

These two fundamental ideas of the transitoriness and hence xl worthlessness of man’s earthly part, and the cleavage between it and his spiritual part, lie at the root of much of the mediaeval literature, and represent the two not incompatible extremes to which the monastic ideal of life, from its very one-sidedness, was capable of leading: on the one hand a certain morbid materialism, on the other an ascetic mysticism. Nor can it be denied that the mediaeval mind took a certain grim pleasure in dwelling upon the more grotesque aspect of these things. The O.E. poet found the same enjoyment in describing his ‘Ȝifer’—

as the painters of the Dance of Death in the drawing of their skeletons and emblems of mortality, or the Gothic carver in his gargoyles. Perhaps, too, some satisfaction in dwelling upon the hollowness of earthly joys, and the bitter fate of those who took their fill of them, was not lacking to a few of those who had turned their backs upon them.

Erthe upon Erthe is perhaps more especially concerned with the first of the two conceptions mentioned above, man’s mortality, but, as has already been shown, a close connexion exists between it and the Soul and Body poems, and though the idea of the duality in man is not mentioned, it is certainly present. The poem is more popular in form than either the Dance of Death or the various Soul and Body Dialogues, perhaps because of its purely English origin, and seems to represent a later and more popular product of the ideas which gave rise to the other two groups. Its short mono-rimed stanza, its jingling internal rime, and its half-riddling, half-punning character, appear to have especially commended it to popular favour, and it is significant that it became most widely-known in its simpler forms.

In preparing the text of this edition, all the available MSS. have been consulted, the only two not examined being William Billyng’s MS. and the Brighton MS., which were formerly in the possession of private owners, and have eluded all search for them. As exhaustive a search as was possible has been made for other texts of the poem, but it has often escaped cataloguing, and it is probable that other copies of the B version, at least, exist.

The punctuation, inverted commas, and regular use of initial capitals in the text are the Editor’s. The MSS. vary in their use of capitals, the same MS. being often inconsistent with itself, while the Cambridge text frequently employs them for unimportant words in the middle of the line, as p. 33, l. 45, Ar, &c. Capitals have been added in the case of all proper names. Letters and words which are obscure or illegible in the MS., or which appear to have been accidentally omitted, are enclosed in square brackets, and a hyphen has been inserted where the MS. separates a prefix or particle from the rest of the word. The MS. writings ff, þ, ȝ, v for u and vice versa, have been retained in the text, and ɫɫ, ŧħ, expanded to lle, the, but it was not thought advisable to expand m~, n~, to me, ne, nor other letters such as d, r, g, when written with a final flourish. Fifteenth-century scribes appear to have used such flourishes at the end of the word rather as a matter of habit than with any particular meaning, and the forms to which expansion of them would lead, such as one, onne for on, are frequently most improbable. It was therefore thought better to ignore such flourishes, or to indicate the persistent use of them by a footnote.

As the conclusions arrived at in the Introduction with regard to the relationship of the English and Latin versions in MS. Harl. 913, and the verbal connexion with the Soul and Body Dialogues, agree, to some extent, with those indicated by Heuser, Die Kildare-Gedichte, pp. 176-80, it is only reasonable to state that the greater part of the work upon the subject had been done, and a projected article upon it written in reply to Professor Fiedler’s in the Modern Language Review, before I had any knowledge of Heuser’s text, and that my conclusions had been formed independently of his, though his have helped to strengthen and confirm them. Moreover I owe his work xlii a very real debt, since I first learned from it of the existence of the Cambridge Text, which has been a most important link in the building up of the general theory as to the connexion between the different versions of the poem.

In conclusion, it is a pleasure to express thanks for kind and courteous assistance to the authorities of the British Museum, the Public Record Office, the Bodleian, Cambridge University Library and Lincoln Cathedral Library; to the librarian of Lambeth Palace Library, to whom I am indebted for the collation of the Lambeth text; to the authorities of Magdalene College, Cambridge, for permission to copy and print the Maitland text; to Lord Harlech for the loan of the Porkington MS.; to Professor Fiedler for permission to use the Brighton text; to Professor Priebsch, who pointed out the text in MS. Harl. 4486; to Miss Helen Sandison, of Bryn Mawr College, U.S.A., for the discovery of the text in the Appendix and for two of the Analogues, and to Professor Skeat for valuable advice and suggestions. In particular this text owes much to my Father, Sir James Murray of the Oxford Dictionary, who has read the proofs, and in the midst of his own arduous work has always been ready with help and advice, to my friend Miss K. S. Block, Lecturer in English at the Royal Holloway College, and, above all, to Dr. Furnivall, in whom all scholars and students of English mourn to-day the loss of a great pioneer, and an ever-ready friend and adviser.

Oxford,

July 1910.

Since this was sent to press two other copies of the B version have come to light at Cambridge, and have by kind permission been inserted on pp. 47, 48 as Appendix II:—