As noted above (I.7) one scribe was responsible for the transcription of both the main body of the text and the rubricated sections. Occasionally a correction to the main text will be made in the rubricating ink or a correction to the rubricated section will be made in the original grey ink. So at G.20.439 he is added above the line with a caret mark but then crossed out in red ink, while at G.15.309 sanitas is written in red ink and ti (to give sanitatis) added above the line in grey ink. The scribe who made the original transcription also made, in fact, a considerable number of corrections. Where these are in the original ink and thus appear to have been carried out at the time of the original transcription, they have been recorded as corrections by hand1. Later corrections, in the same hand but in brown ink, have been recorded as corrections by hand1.1.The identification of the brown ink corrector with the original scribe is based mainly on comparison of scripts where such corrections are substantial (see, e.g., the - admittedly cropped - line added at the top of f.31v), but there is other suggestive evidence such as the treatment of superscript <a>: compare, for instance, the use of superscript <a> by the original scribe at G.8.189 (f.32 l.33) with that of the brown ink corrector at G.6.342 (f.20v l.36). Moreover, as will become clear in the course of this discussion, the types of correction carried out in the two inks are often very similar. The use of a single rather than a double-lobed <a> for corrections such as that at G.2.86 (added all) further suggests that this particular corrector is the original scribe, rather than WH (for whom see Introduction I.10 and I.12). WH seems, in fact, to have carried out corrections which post dated the <u> to <v> corrections (see Introduction II.1.2).

A number of changes are also made by the original scribe in a different (brown) ink (hand1.1). Such changes appear to have been made at a time when the scribe, having completed the original transcription, was able to concentrate more closely on details such as spelling, as well as on other editorial matters.

In a considerable number of cases, unique G readings are shared with A and/or C and such readings are listed at the end of this section. Not all such correspondences suggest a relationship. A high proportion are at least conceivable coincidental readings. Sometimes such readings result naturally from language change.For Kane and Donaldson's treatment of ac variants, see their Introduction, , 18, note 13. Instances where Bx ac appears in G and one or more A and/or C manuscripts as but have not been included in the lists given below for this reason: ac died out during the sixteenth century and, as a result, G regularly (and often uniquely) alters to but even in places where neither A nor C has such a reading (see, e.g., G.6.309, G.10.58, G.13.29, G.13.84, G.13.126, G.13.240). For further examples of shared readings likely to result from language change, see G leyue for Bx  lete (G.7.278), G loked for Bx waited (G.8.155), G symplenes for Bx symplete (G.11.175).Let = "to quit, abandon, forsake" is infrequent after the early fifteenth century ( let v.1 I.5); wait = "to keep watch, to look intently" does not survive beyond the end of the of the fifteenth century ( wait v.1 I.4); while simplety = 'simplicity" dies out at the beginning of the fifteenth. Other instances where date, as well as the use of an obvious easier reading may have resulted in coincidental GA readings include men for Bx ledes (G.4.98), blabber for Bx baber (G.6.192), nowe for nouthe (G.11.51), and synnes for coupe (G.6.308).  Moreover, many shared readings involve substitutions common elsewhere in G where there is no corresponding A or C reading. This is the case, for instance, with G's use of neuene for Bx nempne (G.2.21), a variant which at this point also occurs in AE but which can also be found independently in G at G.3.181, G.6.331, G.6.340 and G.17.19.  The loss of initial And, too, is common in the shared variants but also occurs, without any equivalent AC reading, at G.2.163, G.2.169, G.6.11, G.6.87, G.10.42 etc.  The choice between other and or (= "or" or "either") is probably a matter of dialect (see 4, items 111 and 198), and in any case the shared GA and/or C readings are invariably the commoner form (or = "or", other = "either"; see G.2.177, G.7.254, G.8.103, G.14.329).  The spelling errors G shares with A and C (items II.2.3.2.9 and II.2.3.3.8) are perfectly plausible coincidental misreadings (misreading of <k> and long <s>, misordering of letters, misnumbering of minims).

Nevertheless, in the case of A in particular, the evidence does seem to suggest a relationship.  Kane and Donaldson invoke the concept of shared GF memory of A to account for the GF reading at G.3.180 (poeple for Bx pryues),Kane and Donaldson, , 31. and there is certainly evidence to suggest a link between A and G which is independent of other B manuscripts, in particular, a number of shared GA readings involving major category words (items II.2.3.2.5 and II.2.3.2.6 below): G.2.8 kepe for Bx wilne, G.2.101 asketh þe for  Bx wolden al, G.3.187 thys meyny for þise men/ þes oþere etc., G.4.46 baud after  for brokour als, G.4.84, rysen for rychen, G.5.143 helpe for saue, G.8.148 better for moche more, etc.Some other types of reading can also suggest a relationship. See, e.g., the shared GARVJA reading couple (subjunctive) for couplest (G.11.170). The matter is less clear, however, in the case of C, since the vast majority of shared GC readings, including those involving major category words could be coincidental.So at G.4.347 the reading leefe could result from shared anticipatory error, at G.6.468 the use of heythen for hennes could be a matter of dialect (see 4, item 148), at G.6.512 mother is a likely easier reading for dame as are lose for tyne at G.12.36 and lyght for leye at G.18.280, while coincidental misreadings could well account for the variants at G.7.66, G.14.64, G.18.220 (pylgrymes for pylgrymage, moned for morned, walken for waken). Still, there are a few shared GC readings which might perhaps suggest a relationship: Inowe for at wille (G.11.10), looue for laude (G.12.109), nother for manere (G.14.397).  No two of these appear in the same C version manuscript or manuscripts, so what these readings seem to suggest is a succession of scribes consulting different C manuscripts.

As far as the relationship between G and A is concerned, the evidence for a relationship with a particular A version manuscript is also conflicting. At first sight a relationship with AD seems likely; see readings at G.2.190 (sheued for Bx cheyned), G.7.311 (G cleremeyne, AD chermayn for Bx clerematyn), G.9.46 (chapman for champioun). In fact, however, all these could be interpreted as coincidental misreadings, especially so in the case of the last because of its echo of G.1.64. ARa or AHa is suggested by the reading at G.7.243 (bygge for Bx bidde), with Ha, given readings elsewhere, a more likely candidate than Ra.Ra lacks line G.3.11 where G shares the A reading rynges (Bx golde wyre) and at G.8.148 Ra shares the Bx reading moche more (cf. the GAx reading better). The other possibility is a relationship with the manuscript group ATChH2 as suggested by the G reading at G.4.84 (rysen for Bx rychen).For this grouping, see George Kane, ed., , rev.edn. (London: Athlone Press; Berkeley and Los Anglese: University of California Press, 1988) 85-6.  There is something appealing about this particular possibility since these three manuscripts are all AC combinations and the use of one or other of them might explain the paucity of any sort of overlap between G and C in the early sections of the poem.In G passus 2, for instance there are only 5 shared GC readings, compared with 23 shared GA readings.  It is true that there are places in this passus where B is closer to A than to C thus making overlap with the former somewhat more likely, but the difference is insufficient to explain the discrepancy. In the first 210 lines of G passus 12, by contrast, there are 14 shared GC readings. No significant shared GC readings occur in the earlier passus. The shared GCTChH2 reading looue for laude at G.12.109 supports this suggestion (though note that H2 alters to laude). The evidence is, however, so slight that no firm conclusion is possible.The possibility of a relationship with one or other of two further AC manuscripts (Wa and N) also arises: at G.3.219, G reads  many mery, a combination of B manye and AWa meri, while between G.15.126 and G.15.127 a line has been omitted (see KD.14.120), and this omission is also shared with CN.  In neither case, however, is there much in the way of evidence for a relationship both with A and with C, while in the case of the shared GCN omission, the fact that CN omits two lines at this point (the equivalents of KD.14.119 and 14.120) provides an added complication, suggesting that G's relationship, if any, with C would have to be with an ancestor of this manuscript (the loss of the equivalent of KD.14.119 probably occurred after the loss of KD.14.120 since the first line does not really make sense without the second).

Given the far stronger evidence in the earlier part of the text for a relationship with A rather than C, it seems likely that any significant unique G readings from this earlier section which appear in both A and in C probably reflect a relationship with the former rather than with the latter. If we accept this, then it becomes clear that there are places where significant correspondences between A and G occur in clusters of two or more. Thus there is a cluster of such agreements in G passus 2 (at G.2.8, G.2.97, G.2.133, G.2.193) and a further cluster in passus 3 (G.3.4, G.3.8, G.3.11, G.3.187; note especially the first three of these), together with two correspondences in G passus 4 (at G.4.46 and G.4.84).  Of course, the G scribe may simply have been particularly well acquainted with the A version of these particular passus, but the possibility arises that he actually had the A text open for comparison at this stage of his transcription (these are, after all, passus where comparison of A and B would not be particularly problematic).

In any case, whether or not he actually checked against the A text, the G scribe was clearly acquainted with this particular version. Knowledge of C on the other hand, though it seems possible, is less certain.

The G readings given here normally only include corrections when these were made at the time of writing. Given that hand1.1's corrections can result in error (where the letter <n> is read as a <u> and altered to <v>, for instance), their inclusion would only obscure the relationships between G and the A and C versions. Where, exceptionally, such a correction is cited, this is made explicit.

In nouns the plural inflexion is normally -<es>, less commonly -<ys>, with a limited number of forms in -<n>: oxen (G.20.251), peasone (G.7.305; but see peyse at G.7.199), shoone (G.15.343) and, most commonly, eyne, eyene (G.1.74, G.3.92, G.6.111, G.6.136 etc.). Some  uninflected plurals have been lost (see, e.g., G thynges for remaining manuscripts' þing at G.10.30, G yeres for most manuscripts' ʒere at G.14.3, G wynters for remaining manuscripts' wyntre at G.13.3).

The use of final -e on adjectives appears to be random, i.e. it does not appear to be used to indicate either a plural or a definite inflexion (thus good men G.17.9, a goode pope G.6.168).

Pronouns show the following developments: III.1.3.1 The form of the third person singular feminine is normally she.  Forms in h-, presumably found in G's exemplar, are occasionally retained, as at G.6.315 (with CrYCBL), G.10.56 (general, possibly alliterating),There is some confusion in the tradition as to the sex of Anima, but see G's use of the possessive hyr ("her") in the following line. G.12.112 (with YOC2CBF), G.12.353 (with YOC2CB).  Their rarity, however, suggests that these were not the G scribe's preferred form.Note also the hyper-correction at G.15.322.  In spite of the reading at G.10.56, alliteration on /h/ does not in general appear to influence the G scribe (see, e.g., readings at G.2.74, G.4.29). III.1.3.2 The genitive of the third person singular masculine pronoun is normally hys; forms without <h> occur occasionally (see, e.g., G.13.257). III.1.3.3 The form of "it" is normally ytt, though It occurs once at the beginning of the line (at G.4.55),Line initial capitals (like all capitals) are unusual in G, and the expected form would be ytt as at G.2.38, G.2.88, G.2.143. while at G.1.10 an attempt is apparently made to alter In to Itt, but this is then crossed out and relaced with yt.  Forms with initial h- occur only rarely (at G.7.306, G.8.51 and G.17.151); at G.2.136, initial <h> is written and then apparently crossed out at the time of first writing.According to the (s.v. it, pron.), forms in <h-> died out in standard English during the sixteenth century. III.1.3.4 The nominative of the second person plural is normally ye but you, though most frequently used for the oblique case, does also appear as a nominative (as at G.2.180, G.5.148, G.6.580, G.7.10, G.7.132) and the percentage of nominatives in you gradually increases as the text progresses. Although the cites examples from as early as the fourteenth century, nominative use of you did not become common until the sixteenth, and Barber suggests that its encroachment was not rapid until the 1540s.Charles Barber, (London: André Deutsch, 1976), 204. By Shakespeare's time, you was the normal form.Barber, , 205. As Görlach observes (, 85), the change may have been partly influenced by the reverse vowel patterns of thee and thou, as well as by the fact that ye and you shared a weak form, /jə/. III.1.3.5 Forms of the third person plural pronoun regularly appear with initial þ- or th- rather than h-, (see G they for most manuscripts hij or hy at G.2.58; G theym for most manuscripts' hem, him at G.2.70; G theyr for remaining manuscripts her(e), hir(e) at G.1.30).  Where the G scribe does use a form in h-, it seems possible that this may simply be an error (see, e.g., G.6.254, where the use of hyr may, given the material omitted in G, reflect the scribe's failure to understand the passage). At this date, the use of forms in þ- or th- is not significant from the point of view of dialect. Note also what appear to be occasional weak forms in the/þe (e.g., at G.6.150, G.6.195, G.12.235, G.17.58).See also the back spelling "they" for "the" at G.11.415. III.1.3.6 Though forms of the reflexive pronoun in "-self" are normally singular (hym‑selffe G.1.70, thy‑selffe G.2.144 etc.) the distinction between singular and plural is not consistent, thus theym-selfe (G.10.123), theym-selfen (G.1.117), my‑selue (G.3.23), theym‑selue (G.1.163), theym‑seluen (G.1.59), thy‑seluen (G.2.24).  There is one isolated example of what would later become the standard plural form in -s: vs‑selue[es] (G.8.142), where the abbreviation for -es has apparently been added by the original scribe as part of his programme of spelling corrections (see II.1.1.3.2).

As far as verbs are concerned, the following points are of interest:

A number of forms suggest a Northern connection:

However, as the example of charyss[h]yng suggests, there is evidence that not all these forms were acceptable to the G scribe. A number, such as bees, gyffe, It(t, thrught, are either isolated examples or occur extremely rarely. Words with <sr> for / ʃr/ do not persist after G passus 6 and, even where they do occur, are usually corrected by the scribe either at the time of writing (s[h]ryue G.1.64, G.1.89, sr- > s[h]ryuen G.6.312) or later (s[h]rewe G.1.196 and G.3.41).See II.1.1.2.1 and II.1.1.3.2.  Forms in ar(e for "are" are more common than forms in er(e, and the former are used in the Table of Contents.  The use of neuen- for "name" persists throughout the text, but the scribe appears to have copied this form without full understanding, since, when he makes his later spelling corrections, he invariably alters it to neue[v]-, nyue[v]- etc.See G.2.21, G.3.181, G.6.340 etc. A similar spelling error affects renk(e ("man"), which is invariably altered to re[v]k(e. The use of vowel plus <y> to indicate a long vowel (see III.2) may possibly suggest a Northern connection,See Görlach, , 59. As Görlach observes, the early loss of dipthongs in the North freed i/y to be used to indicate vowel length. A spelling which might suggest such a loss of a dipthongs is pleden for "played" at G.1.20, but this is an isolated example and may well reflect either the retention (at what is, after all, a very early stage in the text) of a Northern form from the scribe's exemplar, or a misunderstanding: if putten earlier in the line is read as a present rather than a preterite, then it would be natural to read pleden as the present tense of the verb "to plead" ("wrangle"?). Note also the AV reading Pleden and the CUcDc reading Pledoun. but it is worth noting that this normally only occurs in the combination <ey> (it is very unusual to find instances of <ay> or <oy> where the <y> might be present to indicate length)In words such as doyth="does," the <y> is simply part of the inflexion (cf. the regular G form doethe). The one instance of doyethe (G.10.98) probably just reflects a merger of the two forms.  Note, however, laythe for "loath"(G.13.248) together with  sayterday (G.6.14, and for possible long a see Dobson, , §23), and the alteration of poundemale to povndemayle at G.3.224 (see II.1.1.3.2). and that <ey> forms in words such as "meat" have been recorded by for a number of Southern counties. 4.317.  The counties in question are Berkshire, Buckinghamshire, Hampshire, Hereford, Hertfordshire, Kent, Warwickshire and Worcestershire. Note also forms of "hear" in hey- and hei- in Kent, Warwickshire, Essex and Somerset ( 4, item 144) and heyuon ("heaven") in Warwickshire ( 4, item 145).  <ey> spellings are also common in those of the Cely letters which were written by Richard Cely the younger, who lived in London, but, since he also makes use of a number of Northern forms (such as <qw(h)> for <wh> and at for that), it difficult to be certain what to make of this. See Alison Hanham, ed., , EETS OS 273 (1975), xxiii, Asta Kihlbom, , (Uppsala: Uppsala Universitets Årsskrift, 1926), 66-7.  As Dobson observes, the use of <ei> or <ey> for ME /ę:/ can be satisfactorily explained as an example of inverted spelling due to the common use of the spelling <ei> and the pronunciation /ę:/ in words adopted from French (receive etc.), since in Anglo-Norman the diphthong ei was monopthongised to /ę:/ and both pronunciations passed into English (Dobson, , §115, Note 4 and §131). For the use of the digraph <ea> in Anglo-Norman words, see above III.2 In general, the distribution of Northern forms and the fact that many are clearly not acceptable to the scribe suggests that, while G may have had an ancestor with Northern connection, the G scribe himself was not a Northerner.For further examples of corrections made to Northern forms by the original scribe, see II.1.1.2.1.

In fact, the evidence suggests a Southern or Midlands rather than a Northern provenance:

The evidence thus once again suggests that the G scribe was not a Northerner.The fact that the third person singular present indicative inflexion is always -<th(e>, never -<s> has not been offered as evidence here, since, in Early Modern English, the choice of inflexion was normally made on grounds other than dialect (-<eth>, for instance, was the more formal variant; see Görlach, , 88-9). Nevertheless, the fact that no -<s> endings appear is, at the very least, consistent with a southern provenance.

There are occasional traces of what appear to be West Midlands relicts (though these are not necessarily all Langland's own).  Many are isolated or infrequent examples: III.4.3.1 One example of the preterite singular of the verb "to drink" in <o> (dronke at G.14.65), beside usual forms of OE and ON /a/ or /o/ before a nasal in <a> (can, man, ran, began, wanne, and, many). III.4.3.2 Forms of class II weak verbs and related nouns etc. in <y>:Such forms do, however, also appear in other areas in the South, see 4.324. louye (G.12.331),But note G forms without <y> where most manuscripts do have <y> spellings at G.12.19, G.12.55, G.12.113, G.12.175 etc. lyuyers ("people who live") (G.13.128), tholye(d (G.12.395, G.14.268, G.19.74), shony (G.6.171). That these are not the scribe's preferred forms is clear from the correction at G.16.543 where wonyen is altered to wonnen at the time of transcription (see II.1.1.2.1). III.4.3.3 Forms in <y> for "thus" ( 4.315). It is difficult to be certain that some of these have not simply been misread as examples of "this" but in a number of cases (e.g. at G.16.148, G.16.302, G.19.415, G.20.156) such misinterpretation seems unlikely.Note also the back spellings "thus" for "this" at G.14.85, G.20.173. At G.11.259, the scribe clearly interprets his exemplar form in <i> as "this" with consequent omission of the pronoun "it." III.4.3.4 Forms of "little" with <ul> (lytull(e beside more usual lytle) ( 1, Dot Map 464). III.4.3.5 Forms of "first" in <u> (furst(e; 46x beside 5 examples with <y>).As Lass observes, it seems unlikely that sixteenth-century spellings of words like "first" with <u> are genuine examples of the NURSE merger (by which "stir," "turn" and "earth" all ended up with the same vowel), since examples are limited to those words which had /u/ <OE y. See Roger Lass, "Phonology and Morphology," in Roger Lass, ed., (Cambridge: CUP, 1999), 112, and also Dobson, , § 82.  These persist throughout the text, and there are two examples at the very beginning of the Table of Contents. After this, however, the Table has forms in <y>, and it therefore seems likely that this was the G scribe's preferred spelling. Note also one example of "birth" with <u> beside one with <y>, murthe (5x) beside myrthe(s (14x), one example of spure ("enquire"), and one of the verb "to shut" in <u> beside one in <e>. The word "church" normally has <u> (occasionally <e>), but this is the regular form by this date.See Dobson, , §82. With these exceptions, however, forms with OE /y/ and /y:/ appear with <y> or, occasionally <e>.Thus OE, ON /y/ appears as <y>/<e> in brygge, brydges, bygg- and begg- ("to buy"; and note the correction of bugger to bygger G.11.315, apparently made by the original scribe at a later date), dyd(e, kysse(n, kyssyng, kyst, kynne(s, kyns, synne(s, fylled, dynne, dynt(es, hylles and hillys, kyng(es, kechyn(n)e, kernell, lyst (from lysten "to please"), lystenethe, left(e (adj.), mylner, pytt, rygge and ryggbones, styrre and styrryd, thynke (seem), thynne, merye (9x) and meryest (1x) beside one example of myrrye. OE, ON /y/ before lengthening clusters appears as <y> in byrde ("lady"), kynd(e, mynde.  OE, ON /y:/ appears with <y> in "fire" (and note fure altered to fyre as part of the original transcription at G.13.288), brydall and brydeale, myse ("mice"), hyde(n ("to hide"), lyes ("lice"), pryde, hyre ("hire"), fyst ("fist"; fust at G.14.319 is an error for furst), lythe ("to listen"). III.4.3.6 Forms of the verbs "to lie" and "to say" and "to buy" in <gg> probably also reflect the original ( 1, Dot Maps 506 and 507).  Where such forms appear in other B manuscripts, G normally retains them, though note G sey for the majority reading in <gg> at G.12.301, G ley for most manuscripts legge ("wager") at G.3.35.Forms of the verb "to live" with <bb>, by contrast, cause problems. At G.1.222 the scribe simply omits the word libbyng, while at G.6.151 he misreads and writes <lyw>, attempts to correct by overwriting and finally deletes the whole thing and writes lyven. III.4.3.7 Forms of the conjunction ere ("before") in ar(e (G.6.20, G.6.53, G.6.461, G.6.468).See Samuels, "Langland's Dialect," 79. III.4.3.8 Occasional forms of "she" in h- (see III.1.3.1 above). Where such forms are retained, they appear as he rather than heo, although the latter was probably Langland's form.Samuels, "Langland's Dialect," 72 and 81, note 16. III.4.3.9 One example of they for "the" (G.4.10; see 4, item 1 and p. 315).

Such forms as these are clearly relicts.For the alterations made to West Midlands forms by the original scribe, see II.1.1.2.1.  Many may have been present in Langland's original, but, if Langland's form for "she" was indeed heo, the G relict form he suggests an intermediate layer of West Midlands copying preceding the copy which introduced the Northern dialect forms.

The G scribe, in any case, appears to have come from an area south of the Wash.  Given the number of forms which change during the course of the transcription, it is further clear that he is only happy to retain forms which are at the very least part of his passive repertoire.  As a result, a form such as "first" with <u>, which, though gradually replaced by forms in <y> in the Table of Contents, is nevertheless retained throughout the transcription of the poem itself, must be assumed to be familiar to the scribe, and this suggests that he is unlikely to have come from the East Midlands or from the western part of Suffolk ( 1, Dot Map 417).  Forms of French loan words in <v-> with initial <w-> (see III.4.2.9 above) suggest the east rather than the west of the remaining area, while forms of the verb "to work" in worch- suggest that Norfolk is unlikely ( 1, Dot Map 315). Within the remaining eastern area, the use of forms in segh- for "saw," combined with forms of "hundred" in -th and forms of "earth" with initial y- ( 4, item 211; 1, Dot Map 454; and 4.319) suggest an area slightly to the North of London, a placement which is supported by the retention of forms of "say" and "lie" in <gg> ( 1, Dot Map 506),Such forms, however, were presumably by now outdated and therefore part of the scribe's passive rather than active repertoire. and by the use of forms in thogh- for "though" ( 4, item 32).  Note, however, that the form soych- for "such" (which was presumably familiar to the scribe (this is the form which he adopts in his spelling corrections, replacing suych-) occurs just south of London in Surrey ( 4, item 10).   

The following abbreviations have been expanded: A loop with a tail for <es> (G.1.3). A forward facing curve for <er> (G.1.12), occasionally for <re> (G.10.85). Superscript <t> in þ(a)t, w(y)t(h) and in words ending in -m(en)t (G.1.28, G.1.34, G.2.162). Superscript <u> in þou (G.2.5). Superscript <ch> in wych="which" (G.3.110). A mark curving backwards over the final <u> etc. for final <m> or <n> (G.1.39, G.1.64). A superscript horizontal line over the middle of a word and sometimes at the end for <n> or <m> (G.1.41, G.1.61, G.1.141). A bar through the descender of <p> for <par> or <per> (G.1.68, G.1.151). A slanting line through the tail of long <s> for <ser> and <syr> (G.1.95, G.3.116). A curved line through the descender of <p> for <pro> (G.4.148). <q> plus superscript <d> for <quod> (G.1.167)

All these are standard and no record of the form of the abbreviation is offered in the transcription or the tagging. However, the scribe's use of superscript <a> as an abbreviation mark is idiosyncratic and is therefore treated differently. There is no example in G of superscript <a> being used to represent <ra> (as is often the case in the other manuscripts). Instead, in G, except in the case of error, <ra> is always written out in full (grace G.2.80; trauell G.1.120 etc.) It seems likely, however, that superscript <a> was present in at least one of the G scribe's exemplars, but that he failed to understand its significance; note, for instance the occurrence of guntethe for grauntethe at G.12.100 and grede with superscript <a> for gradde in most manuscripts (G.17.80),G's use of portren for purtraye at G.4.62 and of sharpe for schrape at G.6.126 may also result from such a misunderstanding, though it should be noted that the first is present in one, the second in several C version manuscripts and that in any case the fact that the records portred as a possible past participle form suggests that G's may have been an acceptable infinitive. as well as the regular use of otiose superscript <a> in conjunction with written <ra> (thus graauntyth at G.3.157, bytraaye at G.11.136, graace at G.20.28).According to Petti, the use of superscript <a> for <ra> was rare in the Renaissance. See Anthony G. Petti, (London: Edward Arnold, 1977), 24. The fact that such words have written <ra> in spite of the scribe's apparent failure to understand the abbreviation suggests that he may have had access to more than one exemplar. In such cases the superscript letter is recorded as part of the transcription on the grounds that the scribe presumably found it in one of his exemplars and must have wished to record it, even though (or perhaps because) he did not fully grasp its function. The scribe does, in addition, use superscript <a> as an abbreviation mark for a number of other letters or combinations of letters in accordance with Latin practice. Thus at G.4.156 "Iesu" is written Ihua, at G.6.150 "spiritualte" is written spaualte. In such cases, the form of the abbreviation is recorded in the tagging.See also the traditional abbreviation quemqam for quemquam at G.15.297, but note miaam for misericordiam at G.6.293. A problem arises with the words "grant" and "covenant," which frequently end in <a>+2 minims+<t>, with a superscript <a> over the minims. It is possible that the superscript <a> here is otiose and that, like the examples given above, it simply records an abbreviation (whether for <ra> or <au>) which was present in the scribe's exemplar but which is no longer necessary. However, it is noticeable that, when superscript <a> is not used, the scribe's spelling of "grant" is almost invariably graunt, The one exception is granted at G.4.173. i.e. he provides four minims, whereas in cases where superscript <a> is present, he normally writes only two. This suggests that the scribe may have interpreted superscript <a> here as an abbreviation mark for <n> and it has been therefore expanded as such in the transcription both of "grant" and of "covenant" (for which there is a similar pattern, although the numbers are smaller).

Special problems also arise with flourishes and bars which could possibly be taken to indicate final -<e>

a) There are three cases of long <r> plus flourish (at G.3.68, G.4.158, and G.6.269). Since this is a very small number, it is difficult to draw any very firm conclusions as to scribal intention, but this combination has nevertheless in each case been expanded to <re> on the grounds that "their" and "her" with final -<e> are certainly possible (the latter is unusual in this manuscript, but see G.4.127). It should be noted, however, that it may well be the case that none of these examples was in fact the work of the original scribe. In the first two cases the <r> and the flourish are very definitely later additions and, although the third case is not as clear, it seems possible that here too there may well have been later correction: the <r>+flourish appears to have been squashed in, leaving no space between words (see notes to these lines).

b) In the case of <d> with a tail, the evidence suggests that the flourish is otiose. Not only does such a flourish regularly appear in positions where it would not be expected (almost invariably, for instance, on "quod"), but elsewhere the expansion to <de> would destroy the distinction between long and short vowels which the scribe goes to some pains to maintain: "god" appears as <god>+flourish, or as <god>, "good" appears as <gode>, <good> or <goode>. In each of the following cases, "god" appears with a flourish on final <d>: but god to all gode men suyche wrytyng defendyth (G.4.64) more to good þen to god þe gome hys loue cast (G.14.358).

c) The tail on <ff> occurs most frequently on off and yff, i.e in positions where final <e> would not normally be expected,There are two examples of þer-offe at G.17.5 and G.17.243, but these are the only other instances. and since <ff> almost invariably occurs with either final <e> or with a flourish,There are exceptions to this (see, e.g., Off at G.10.76 and off at G.11.50) but these are unusual. it seems clear that the flourish is otiose.

d) The bar through final <ll>, on the other hand, though frequent, is not invariable and does often occur in places where a final <e> would be historically appropriate, as on the noun wyll< OE willa (G.1.150), on the plural adjective small (G.1.147), and on the infinitive tyll (G.1.120 ). It is difficult, nevertheless, to argue in favour of expansion. Even in cases where such an <e> would be appropriate, evidence concerning the G scribe's practice is mixed: the noun "will" does sometimes appear with written final <e> but it can also appear with final <ll> without bar (as at G.6.607, G.9.94, G.9.125); plural adjectives such as "all" can appear both with written <e> and with neither bar nor <e> (as at G.1.58, G.1.103); infinitives, though they frequently do appear with final <e>, can also appear without inflexional ending (to know G.1.122, to lett G.1.186 and note especially to kyll at G.2.67, ending in <ll> without a bar). In addition, such bars regularly occur in places where final <e> would have no historical justification: see the preterite singular befell+bar (G.1.6), the singular indefinite adjective all+bar (G.1.178, G.1.185), and the singular verb shall+bar (G.1.184, G1. 202, G.3.36, G.3.174 etc.). Neither the preterite singular verb "fell" nor singular "shall" ever appears with written final <e>. Finally, it is worth noting that it is possible for barred <ll> to be followed by a written <e>, as at G.6.65 (Peronelle). Consistent expansion to <lle> would not therefore accord with the G scribe's practice, and the bar has been treated as otiose.

As far as the forms of individual letters are concerned, allograph forms such as long <s> and sigma <s> or long <r> and two-shaped <r> have not been distinguished. In the case of yogh, where one letter is used to represent more than one sound, the velar spirant and /j/ have been recorded as yogh (<ʒ>) and the sibiliant has been recorded as <z>. Thus myʒte (reading of manuscripts other than G, G.2.165), ʒerne (G.7.304), ʒemen (G.10.214), but dozene (G.5.39), forbyzyne (G.9.29), sarazenes (G.12.123).

Capitalization is that used by the scribe and by whoever provided the litterae notabiliores (see I.10 and notes). Such capitalization is, in any case, minimal. Lines, for instance, do not normally begin with a capital letter. The digraph <ff> as been recorded as <F> where it occurs at the beginning of a word. Capital <I> has been recorded as <I> whether or not it indicates an <I> or <J>/<j>.

Scribal punctuation (consisting mainly of virgules and, in the Latin sections, double points, like a colon, with the occasional single point) has been retained. It is not always easy to distinguish a point from a mark caused by lifting the pen from the paper (this is particularly so in the rubricated sections) and where there is any doubt the matter is discussed in a paleographic note.

The word divisions found in the manuscript have been followed as far as possible. Where two elements of a single word are separated by a space, they are linked by a shadow hyphen <->. Thus for-wandred (G.1.7). Where two words are run together, both the original and the regularized form are supplied, as in <orig>shalbe<orig><reg>shal be</reg> (G.2.178). Proper nouns have been treated as English unless they have Latin inflexion. Thus cesar at G.2.49 is treated as English but reddite cesari at G.2.53 is tagged as Latin. The problems which arise in other manuscripts at G.11.329 and G.14.55 are not present in G, since G lacks the English inflexion which other manuscripts give to beatus vir.

Scribal misspellings have been recorded with a <sic> tag and corrected with a <corr> tag.

Forms of highlighting which reflect the scribe's interpretation of the text (rubrication of Latin, boxing of important words) have been tagged so that they appear as part of the transcription. Boxing is not, however, tagged when it results from the fact that the line is too long and wraps round, i.e. where it is simply being used to indicate which words belong to which line. In such cases, the boxing is recorded in a codicological note. G's boxing is usually incomplete. The display will show a complete box, but in fact G's boxes rarely have more than three sides.

Scribal corrections made at the time of writing have, wherever possible, been distinguished from those made by the original scribe at a later date and those made by other hands (hands 2 and 3).For corrections made by hands 2 and 3, see Introduction II.1.2. This is not always an exact art, and where it is impossible to determine the corrector, such corrections are recorded as being by handx. Even where identification of the scribe seems possible, corrections are normally minor (the change of a letter, the addition of a short word) and it is quite possible that some have been misattributed. The original scribe clearly made a considerable number of later corrections in brown ink (see discussion at II.1.1. and II.1.1.3) and since these persist throughout the text it has been assumed, unless there is contradictory evidence, that brown ink corrections where the hand is difficult to distinguish (for instance, the addition of a large number of virgules) have also been made by the original scribe as part of this programme of corrections. The alternative is to assume two very persistent correctors, employing very similar shades of ink. It is also the case that the later insertion of virgules makes good a lack in the earlier stages of the original transcription which does not persist into the last few passus, i.e. the corrections result in a more consistent transcription, and it is easy to imagine that the scribe, having realised towards the end of his transcription that he had been in the habit of omitting virgules, went back later to correct this fault in the earlier passus. Virgules which have been added to separate words (e.g., after a correction which involves added letters), and which have no significance as metrical or punctuation marks, have, however, been recorded, not as part of the text, but in paleographic notes. As far as other corrections are concerned, where erased text is legible, this is recorded within the deletion tags. Where it is illegible, this has been indicated with one punctus per deleted character.

Four sets of annotations are provided — codicological, lexical and linguistic, paleographic and textual.

Codicological notes draw attention to physical features of the manuscript and to later additions in the margins. These notes are marked with a red, superscript <>.

Paleographic notes comment on letter forms, particularly where these are ambiguous. These notes are marked with a red, superscript <>.

Lexical notes comment on lexical ambiguities and are marked with a red superscript <>.

Linguistic notes comment on items of linguistic interest. They are marked with a red superscript <>.

Textual notes supplement the information on textual relationships provided in the apparatus tags (see Treatment of Textual Variants below) and may, in addition, provide information about the relationship with the A and C texts. These notes are marked with a red, superscript <>.

All instances of textual variants in which three or fewer B manuscripts share G's reading are recorded in the apparatus tags. Instances where four B manuscripts share G's reading are recorded only where these manuscripts do not belong to the β4 section of the stemma. Unique and original G readings showing agreement with A and/or C are listed in the Introduction (at II.2.2), and such readings are therefore not normally recorded in the notes. However, where corrected readings or readings which G shares with one or more B manuscripts show agreement with A and/or C which is particularly interesting, such correspondences are recorded in the textual notes. In general, B text variants are not normally recorded unless they are substantive. However, spellings which may or may not indicate a substantive variant are recorded (see, e.g., G.17.246, where, given the G scribe's inconsistent use of double and single consonants, the G reading bleden may or may not indicate substantive variation from the Bx reading Bledden) together with some non-substantive variants which are of linguistic interest (the use or otherwise of the y- past participle prefix, extension of an -es ending to previously uninflected plurals). On the other hand, dialect variants which are very persistent (such as those involving variation between <s> and <sh> spellings (and vice versa) are only recorded where it seems likely that there may have been a genuine misunderstanding. Manuscripts are cited in the following order: G L M Cr W Hm C G O C2 Y B R F. Where a shared reading in a particular manuscript (say L) results from later correction the reading cited is normally taken from the manuscript next in order.