An errant leaf and a divided poem: The Lay of Juan de Torres in SA7

Few extant manuscript cancioneros are original compilations. Most are fair copies, sometimes copies of copies, of single lost exemplars. An original compilation is, of course, also based on lost exemplars, but the difference is that its exemplars are, by definition, multiple and heterogeneous. The Cancionero de Palacio (Salamanca University Library MS 2653), better known as SA7, is one of these valuable exceptions, valuable because their physical composition contains a wealth of clues to the sources that lie behind them.¹ In the words of Barry Ife, discussing another such volume, the erstwhile Cancionero of the British Museum, or Rennert (LB1), 'A great deal of evidence about the process of compilation can be gleaned from a close examination of the physical characteristics of the manuscript' (2002: 64). Unlike the mere copy (Baena, Herberay, or Stúñiga, for example), which throws up a screen of varying penetrability between the textual critic and the trail of transmission, the original compilation offers a relatively accessible array of clues about its sources, such as changes to the mise-en-page, false starts, and duplicate texts.

Fundamental to this sort of analysis are the internal divisions of a collection, the way the texts are organized, and the sequences in which they are copied. In this respect SA7 is a disaster: the manuscript has come down to us in a state of apparently intractable disorganization. Sixty years on we can only endorse the verdict of its first editor, Francisca Vendrell: 'Múltiples y difíciles de resolver en este manuscrito son los problemas relativos a lagunas de texto y a transposiciones de folios' (1945: 10).² SA7 is a folio volume of 178 numbered leaves, containing upwards of 370 poems. Internal textual and material evidence (watermarks, writing style) suggest that the cancionero was assembled around 1440 (Tato 2003: 517). The modern bound volume preserves the foliation of an earlier binding, which may have been close-to-contemporary with the compilation. Traces of this foliation (Roman numerals in a 15th-century hand) reveal that when it was first bound and foliated the manuscript was already in disorder (see, for example, Tato 2003: 498; 2005: 64). Discrepancies between the old and the modern foliations testify to a loss of at least eight leaves since then, but greater losses were evidently suffered by the manuscript in its pre-bound state (Tato 2003: 501–02; 2005: 60).

The problem with SA7 is not so much the losses as the combined effect of loss and transposition of folios. There are coherent runs of leaves, but the breaks in continuity surrounding them suggest we may never establish the order in which the discrete blocks were originally arranged. However, the crux of the matter is that as medieval compilations go SA7 is peculiarly prone to disorder, for the simple reason that the poems it contains are predominantly short. Approximately three-fifths of its 370-odd texts are lyrics of between 10 and 20 lines, fitting typically two to a page.³ As a result, the end of numerous verso pages coincides with the end of a text, creating a natural break in continuity. Significantly, those stretches of the manuscript that present a reliable sequence of texts contain poems long enough to extend over pages, if not leaves. So too the transpositions that have been identified by modern scholars (and in one case signposted by an early reader) involve interruptions to longish poems or shorter poems belonging to a series.⁴ Efforts to reconstruct the original sequence of texts are further hampered by our ignorance: no other witness exists for the overwhelming majority of poems in SA7, or indeed for most of the poets. Moreover, as I hope to demonstrate, we cannot trust the quiring.⁵

The commonest mark of disorder detected so far involves the displacement of single leaves. This disruption seems particularly to affect a run of 63 leaves in the first half of the manuscript, from fols 24 to 86 inclusive, towards the middle of the volume.⁶ The present ordering requires the reader to jump from fols 24–28 to fol. 34, from 33 to 65, and from 64 back to the starting point at 24. The following diagram gives an idea of the leapfrogging action and the circularity.⁷

Figure 1.

Links between poems disrupt the folio sequence 24–86

It is hard to envisage a quiring arrangement that could accommodate these folio acrobatics. Unsurprisingly, recent attempts to establish the codicological structure of SA7 have proved inconclusive.⁸ To judge from the prevailing anarchy in the disposition of leaves, I suspect that a full codicological examination of the manuscript will only be practicable once we know what we are looking for. The obvious inference is that folios so far adrift from their original moorings, such as 33 and 64, became detached at some point in the prehistory of the bound volume and migrated as singletons.

The identification of another such case allows me to reconstruct three sets of poem sequences that have been skewed by the migration of a single leaf. The most immediate evidence is palaeographic. Folio 2 is copied in a different hand from the hand that copied adjacent leaves 1 and 3 and the 20 consecutive leaves which contain an unbroken run of poems to fol. 23 (see Figures 2 and 3).⁹ This hand, which may be regarded as the mano principal, or Hand 1, is the most ornate of the 4 or 5 hands which intervened in the copy.¹⁰ The less distinctive hand which copied fol. 2 is well represented later in the manuscript. I shall refer to it as Hand 2. (See the sample of parallel texts in Figure 4 for a comparison between Hands 1 and 2. The difference is most marked in the treatment of ascenders, which

Figure 2.

Folios 1v and 2r

Images of MS 2653 (SA7) are reproduced by kind permission of the Universidad de Salamanca. I am grateful to Doña Margarita Becedas González, Directora of the Biblioteca General Histórica, for securing this permission and for supplying me with digital versions of the images.

Figure 3.

Folios 2v and 3r

Figure 4.

Hand 1 (fol. 8v, detail) contrasted with Hand 2 (fol. 83r, detail)

are taller and looped in Hand 1. Note especially the 'b' in 'pobre', 'bienes', and 'buscado'.) Somehow the intruder status of folio 2 has been overlooked, perhaps at least partly because a spurious uniformity has been imposed on the 1v–2r opening by the addition in a later hand of running heads erroneously announcing: 'obras de diego furtado de men doça' (1v) and 'obras de \do/n albaro de luna con destable de castilla' (2r) (see Figure 2).¹¹ But clues have not been wanting. As well as the external evidence there are textual indications.

Table 1.

Opening texts in SA7 with contents of fol. 2r-v shown in bold

MS no.	fol.	rubric	fol.	incipit	lines of text	ID
SA7-1	1r	Perque diego furtado de mendoça	1r-v	Pues no quiero andar en corte	59 ll.	2395
SA7-2	1v	Cancion luna condestable	1v	Si dios nuestro saluador	3, 8	2396
SA7-3	1v	Cancion luna condestable	2r-v	Porque de llorar	8 x 3	2397
SA7-4	2v	Montoro	2v	Amor que yo vi	3, 3 x 5	2398
SA7-5	2v	Enyego lopeç hermano De mendoça	3r	Mis oxos fueron a veer	4, 8	2399
SA7-6	3r	Cancion diego furtado de mendoça	3r	Ya con tanta fermosura	4, 8	2400

Table 1 sets out the rubrics and first lines of the first six texts in SA7 as presently constituted, displaying the contents of fol. 2r–v in bold. The first thing to note is the mismatch between the rubric at the foot of 1v and the form of the poem on 2r, 'Porque de llorar'. The rubric, 'Cancion luna condestable', refers to a formal category of lyric that cannot by any definition of the term be stretched to apply to the text that follows: eight hexasyllabic tercets rhyming aab aab bbc bbc ccd ccd dde dde.¹² Secondly, neither of the two authors named in the rubrics on fol. 2v, 'Montoro' and 'Enyego lopeç hermano De mendoça', figures in this part of the MS or in the other blocks of text copied by Hand 1.¹³

Coherence returns if we remove the interloper (see Figures 5 and 6 and Table 2): the rubric at the foot of fol. 1v comes good, because the text following at the head of 3r, 'Mis oxos fueron a veer', is indeed a canción, and very plausibly attributed to Álvaro de Luna. Discounting the unlikely eventuality that fol. 2 was grafted into the first quire to replace a missing leaf, fols 1v and 3r make a perfect match. It is possible that the run of leaves 1 + 3.23 once constituted a complete single quire, through-copied by Hand 1.¹⁴ However, this act of virtual surgery is

Figure 5.

Folio 1v

Figure 6.

Folio 3r

motivated primarily by the lure of a better destination for fol. 2. The true home of the migrant leaf is at some distance in terms of codicological plausibility, but in every other respect entirely appropriate.

Table 2.

Reconstruction of sequence of texts in fols 1r-v and 3r after the removal of fol. 2

MS no.	fol.	rubric	fol.	incipit	lines of text	ID
SA7-1	1r	Perque diego furtado de mendoça	1r-v	Pues no quiero andar en corte	59 ll.	2395
SA7-2	1v	Cancion luna condestable	1v	Si dios nuestro saluador	3, 8	2396
	1v	Cancion luna condestable
SA7-5			3r	Mis oxos fueron a veer	4, 8	2399
SA7-6	3r	Cancion diego furtado de mendoça	3r	Ya con tanta fermosura	4, 8	2400

The unusual strophic form of 'Porque de llorar' (SA7–3) recurs only once elsewhere in the cancionero, on 32v. Folio 32 is the last of a short run of leaves, 29–30–31–32, stranded in no man's land by the desertion of adjacent folios, and copied by the same hand – Hand 2 – as the stray leaf wedged between fols 1 and 3.¹⁵ The group is bounded on one side – 29r – by an unidentified fragment (SA7–76^bis) and on the other – 32v – by four tercets of Torres's Lay, 'Ay triste de mi' (SA7–86), which Dutton records as truncated in his headings to the transcription: '4 x 3 … ' (1990–91: iv, 101b). Folio 2 can be repositioned with absolute confidence as the leaf following 32, because what we have in 'Porque de llorar' is the longer second half of 'Ay triste de mi', the Lay that breaks off after only four stanzas. Figures 7, 8 and 9 show the sequence of fols 32v, 2r, and 2v, containing the complete poem, which I edit in the Appendix.

The affinity between the two texts SA7–3 and SA7–86 is unmistakable, and has not gone unremarked. Only a few years after the appearance of Vendrell's edition, Pierre Le Gentil included 'Porque de llorar' in his discussion of 'Ay de mi triste' as a Castilian example of the French lai, and he printed both texts one after the other, saying rather plaintively: 'Le poème suivant ['Porque de llorar'], emprunté au même recueil et attribué par lui au fameux connétable Alvaro de Luna, ne mérite-t-il pas d'être rapproché du précédent ['Ay de mi triste'], bien qu'il soit qualifié de canción?' (1952: 198). On this hint from Le Gentil, and on the understanding that there were two extant examples of this metric type, Tomás Navarro was led to make some extraordinary generalizations about the nature and the form of the lay in Castile, such as 'Su estrofa usual era la sextilla simétrica con dos rimas, aab:aab' (1956: 140). Following the sequence of the texts in SA7, he gives precedence to Álvaro de Luna: 'El lay de don Álvaro de Luna, Canc. Palacio, núm. 3, se distingue por su ritmo trocaico y por sus rimas agudas'; whereas Juan de Torres, the author of 'Otro lay' ('en el citado cancionero, núm. 86'), gets rapped on the knuckles for defective technique: 'muestra flexibilidad semejante en su trabada construcción gramatical, si bien procede con menos uniformidad en cuanto al tipo rítmico del verso' (141). Álvarez Pellitero also classifies 'Porque de llorar' as a lay in a note to her edition of the poem (no. iii), where she explains her decision to recast it (as also 'Ay triste de mi', no. lxxxvii) in the form of sextillas.¹⁶

It would have been sufficient for the purpose of rebuilding Torres's Lay to establish the link between fols 32 and 2. However, it is possible to provide a fuller setting for fol. 2 by recruiting 57 as the follow-on leaf. This is a single folio, adrift from its neighbours, and also copied by Hand 2 (see Figure 10). The preceding leaf, 56, is the last in a straight run of leaves (45–56), copied in yet another hand (Hand 3), and there is no rubric at the foot of the verso to herald the text that heads 57r.¹⁷ Nor is there any rubric at the end of 57v, yet the following recto, 58r, also starts with an unrubricated poem.¹⁸ The repositioning of fol. 57 is for the moment provisional, given the current state of our knowledge about the quiring, but the fit is good: 2v ends with a rubric and 57r begins – like fol. 3r – with a rubric-less text, in this case, 'Adios quedeys linda corte'.

Table 3 reconstructs the configuration of texts arising from an original sequence of leaves 31–32–2–57, all of them copied by Hand 2. There are significant continuities in the style of rubrication (for example, the erratic capitals in 'alfonso De montoro' on 31r, 'hermano De mendoça' on 2v), and in the contiguity of authors: fols 31r–v and 2v both contain rubrics which refer to a Montoro by surname alone, presumably the Montoro who has already been identified as Alfonso at SA7–81, the poem heading 31r.¹⁹ Likewise, the conjunction of fols 2 and 57 brings together the only pieces attributed to the two obscurer Íñigo López homonyms in SA7, 'hermano De mendoça' and 'fixo de Johan furtado', so labelled no doubt to distinguish them from each other, as well as from the future Marqués de Santillana, who is designated 'senyor De buytrago' in another part of the manuscript (fols 84r–85r) copied by Hand 2. The sequence 32–2–57 seems a desperate remedy, maybe, but this manuscript presents no ordinary problems.

It would be premature to comment on the codicological implications of this

Figure 7.

Folio 32v

Figure 8.

Folio 2r

Figure 9.

Folio 2v

Figure 10.

Folio 57r

anarchic displacement of leaves: the problems of SA7 will have to be tackled piecemeal as we become more familiar with the texts and their authors, with the habits of the individual scribes, the house style of the copyshop they were employed in, and the nature of the sources they were using. I am keeping an open mind about the possibility of the volume being factitious; there is a good case to be made for its integrity on grounds of consistency of contents and layout. A more likely explanation is that the quiring is factitious; that the disordered bundle of inside-out quires, stray bifolia, and single leaves was reassembled into improvised gatherings preparatory to binding in the late medieval period. The sporadic rescue of single leaves by stitching them into alien quires would account for the anomalous displacements of fols 33 and 64, as well as of fol. 2.

Table 3.

Restoring fol. 2 (in bold) to its place in the sequence 31—32—2—57

MS no.	fol.	rubric	fol.	incipit	lines of text	ID
SA7-81	31r	Cancion alfonso De montoro	31r	pues por ti peno amor	5, 10	2471
SA7-82	31r	Montoro	31r	apartar pueden a mi	3, 7	2472
SA7-83	31r	Cançion de Johan de torres	31v	Si uos plaçe que mantenga	4, 8	2473
SA7-84	31v	Montoro	31v	Si el coraçon me basta	4, 8	2474
SA7-85	31v	Otro dezir de suero de Ribera	32r-v	En vna linda floresta	4 x 8, 2	2475
SA7-86	32v	Lay de Johan de torres	32v	Ay triste de mi …	4 x 3	2480
SA7-3			2r-v	…Porque de llorar	8 x 3	2397
SA7-4	2v	Montoro	2v	Amor que yo vi	3, 3 x 5	2398
SA7-121	2v	Enyego lopeç hermano De mendoça	57r	Adios quedeys linda corte	4, 8	2512
SA7-122	57r	Eneyego [sic] lopeç fixo de Johan furtado	57r	Muy de grado seruiria	4, 8	2513
SA7-123	57r	El mesmo	57v	Amor pues ya non veo	4, 8	2514
SA7-124	57v	Gonçalbo de quadros	57v	De uos seruir et loar	4, 8	2515

One consequence of the rehousing of fol. 2 is an adjustment to the small corpus of poems attributable to Álvaro de Luna. In exchange for the unsatisfactory half-lay, he is credited with 'Mis oxos fueron a ver', a canción which shows pleasing signs of affinity with his other verse. The (admittedly only provisional) recompense offered to the otherwise unknown Íñigo López, 'hermano de Mendoza', for this loss is the acquisition of a suitably archaic alternative, 'Adios quedeys linda corte', which has to date suffered from the double stigma of orphanhood and suspected fragmentary status.²⁰ In fact it has the standard structure of a canción of the era: 4, 8 lines, rhyming abba cdcdbb[b]a. In Vendrell's view, this Íñigo López was the brother of Admiral Diego Hurtado de Mendoza, and Santillana's uncle (1945: 27–28). Certainly the use of the senhal 'estrella do norte' in line 12 puts its author squarely into the generation of poets better represented in the Cancio-nero de Baena.²¹ But the main effect of the transfer has been the recovery of a unique lyric of the Spanish 15th century, split by misadventure into two unequal halves at a distance of 30 leaves from one another and attributed to two different authors. The reunified text of Juan de Torres's poem is the only surviving lay in the Castilian repertory recorded by Dutton.²²