The sources for the Old Hispanic Rite are striking for their paucity.Footnote ¹ The Third Council of Toledo, meeting in 589, was attended by about seventy bishops or their representatives from all across the Spanish peninsula and southern France. The Fourth Council of Toledo, meeting in 633, provided that ‘when presbyters are ordained in the parishes they should receive from their bishop an official book so that they may go to their churches instructed and not give offence in the divine sacraments through ignorance’.Footnote ² The existence in c.600 of seventy bishoprics and an indeterminate number of dependent parishes with books implies the existence of a great many manuscripts widely distributed. We cannot be certain what these books might have been like or in what kind of detail they might have presented the liturgy. Since the goal, however, was to ensure uniformity of liturgical practice across a broad region, there was surely considerable detail rather than simply an outline upon which to improvise. In any case, by c.700 and the date of the Oracional visigótico (the Veronal Orational), the liturgy certainly existed in a form like the one preserved in the first notated sources of the tenth century, complete with texts for the antiphons and responsories of the office for much of the liturgical year.Footnote ³ Yet what remains are a few dozen sources, many incomplete, and from a limited number of locations. The part of the peninsula south of Toledo is not represented at all by anything like a complete source. Both the provenance and the dating of many of the surviving sources remain in considerable doubt furthermore.

As far back as the First Council of Braga in 561 it was ordered that monastic practice should be kept separate from public worship, which included the Mass and the offices of matins and vespers. A recently published volume edited by Emma Hornby and others, entitled Understanding the Old Hispanic Office: Texts, Melodies, and Devotion in Early Medieval Iberia, distinguishes between ‘public’ worship and ‘cloistered’ worship and gives an excellent summary of what we know about the office, taking account of the substantial bibliography on the topic, which I will not rehearse here.Footnote ⁴ But what we know does not add up to a complete picture. There are significant voids, and what we do have in the small number of relevant sources does not present a picture of uniformity or even the strict distinction between public and cloistered liturgy that the Hornby volume sets out. The manuscripts BM51 and S7, for example, do include the public services of matins and vespers for the commemoration of general classes of saints (e.g., virgins, bishops, confessors) as well as cloistered services. And AL, the pre-eminent source for public worship, includes music for some of the cloistered services. A few nuggets can be gleaned, however, that cast light on (1) the formation of the liturgy and (2) the relationship among the sources that provide elements of the office.

What follows in addressing these two points limits itself to the study of responsories for those parts of the Night Office that are termed ad medium noctis and ad nocturnos. Hornby's book lists all the cloistered services and the manuscripts that preserve them. There are only five, BM51, Sant, Sal, S7 and T3, and no single manuscript preserves them all. Typical of the inconsistencies among sources, AL, which is a source for public worship, nevertheless includes some of the responsories for some of these services. Table 1 lists the responsories occurring in AL, Sant, Sal and BM51.Footnote ⁵ T3 does not include the services that I will discuss, and S7 includes only some of them, which are listed in Table 2 with their responsories. Furthermore, S7 does not provide notation for most of its pieces and thus cannot be brought to bear on the arguments for grouping the manuscripts on notational grounds.

Table 1. Responsories ad medium noctis and ad nocturnos in AL, Sant, Sal and BM51

Table 2. Responsories ad nocturnos dominicales in S7

To begin with the responsories sung ad medium noctis, the lists in Sal, Sant and BM51 have a good deal in common, but Sant and BM51 have two or three unica each, Sal specifies that most of its pieces are assigned to Lent, and BM51 assigns its pieces to various seasons and the common of saints. One might say that Sal and Sant are more closely related to one another than either is to BM51, but the picture is decidedly mixed.

For responsories to be sung ad nocturnos on Sundays, AL provides a list of six that is identical to the list in Sal. In Sant, the first four are identical in order to the first four in AL and Sal and are followed by two unica. BM51 includes four that are shared among the other manuscripts, but not in the same order, and two unica. Most important here is that AL and Sal are identical and closely related to Sant, whereas BM51 stands somewhat apart from the others.

BM51 then continues with nine further responsories ad nocturnos on Sundays that are not found in Sal or Sant but that do appear in AL assigned variously to matins on ordinary Sundays, to the commemoration of the dead and to nominally cloistered services. What follows is a set of six pieces shared between BM51 and Sant but not in the same order and assigned in Sant to post nocturnos and in BM51 to ad nocturnos for the Sundays in Lent. This is the closest correspondence between BM51 and any of the other sources, though once again there are considerable differences in the order in which the pieces appear and the places in the liturgy to which they are assigned.

Sal, Sant and BM51 then provide pieces to be sung ad nocturnos for weekdays, and here a very different picture emerges. There are four pieces for each day, and thus a total of twenty-four, and they are identical in all three sources in the order in which they are presented (Table 3). Especially striking is that they are all psalmic, and they proceed through the psalms in order. Furthermore, BM51 takes this list and assigns its pieces in a systematic way to the Sundays in Lent. Here each of the four Sundays in Lent has six responsories. The four pieces for Monday are distributed across the four Sundays as the first in the series of six. The four pieces for Tuesday are distributed across the four Sundays as the second in the series of six. And so forth, as described in Table 4, where the numbers are the order-numbers in the psalmically arranged list of the twenty-four assigned to weekdays. This arrangement suggests strongly that the list for the weekdays precedes the arrangement for Sundays, which evidently derives from it, since it is very difficult to imagine that the arrangement for Sundays in Lent in BM51, which obscures or at least complicates the orderly progression through the Psalms, can have been the source for the very straightforward distribution of these pieces across the weekdays in all three manuscripts.

Table 3. Order of Reponsories ad nocturnos for weekdays in Sal, Sant and BM51

Table 4. Order of Responsories ad nocturnos for Sundays in Sal, Sant and BM51

From all of this we can conclude that our four sources have common roots and share a loose collection of responsories for the Night Office but that this collection was also employed in some degree in public worship. The list of pieces for the Ferial Office, because of the way in which it uses the psalms and the fact that it appears identically in Sal, Sant and BM51, must antedate the shared loose collection and the divergences of these sources from one another. This list must therefore be part of the oldest surviving core of the Night Office. The evolution of the Night Office thus began with a set of pieces for the Ferial Office, drawn from the psalms in order and all closely related in appealing to God to listen and to have mercy. Only later were pieces assigned to or composed for Sundays and other occasions, by which time the traditions embodied in these manuscripts (as defined especially by their musical notation, as we will see) had begun to diverge. The systematic distribution of pieces for the Ferial Office across the Sundays in Lent speaks to the ways in which the structure of Lenten worship evolved, for other sources suggest that the two halves of Lent emerged separately, whereas BM51 distributes the list of psalmic responsories across the whole of Lent as a single structure.Footnote ⁶ In particular, this list must antedate the separation between BM51, on the one hand, and AL, Sal and Sant, on the other, that has been hinted at in several ways in what we have seen so far. To this separation we now turn in examining the musical notation of these responsories.

Figure 1 transcribes in parallel and in alphabetical order all the responsories shared among AL, Sant, Sal and BM51.Footnote ⁷ In my earlier study of the responsorial psalm tones, I showed how AL, Sant and Sal share a version of these tones that differs from the version in BM51 and related manuscripts. I called these the León and Rioja traditions, respectively.Footnote ⁸ Figure 1 makes possible a comparison of versions of the whole of each piece and not just the verse. The responsory refrains are, of course, much more elaborate than the verses and employ a much wider array of notational symbols. It is nevertheless clear that AL, Sant and Sal form a consistently similar group that regularly differs from BM51 in notational detail, even though there is no doubt that all four manuscripts give versions of the same melodies and surely have a common ancestor. For example, Alleluia deduc me, the second piece in Figure 1 and the first piece that is present in all four sources, sets the word ‘Alleluia’ at the opening in BM51 with a series of puncta and podatus. AL, Sant and Sal, on the other hand, set this word with a much more elaborate melody and are virtually identical to one another. Careful comparison of the richer notation of the refrains thus shows how the refrains too confirm the grouping of the manuscripts made on the basis of their verse formulas.

Figure 1. Responsories shared between AL, Sant, Sal and BM51 in alphabetical order.

Figure 2 similarly transcribes in parallel and in the order in which they appear in the manuscripts the twenty-four responsories assigned to the Ferial Office. AL does not include these pieces, and Sal gives each one twice – once for the Ferial Office and once for the Sundays in Lent. Here, again, Sal and Sant clearly form a pair that differs in detail from BM51 while nevertheless presenting the same melodies. That Sal gives each piece twice offers a unique opportunity to study this notation as employed by a single scribe and thus to learn something about the variability and lack of it between two examples of the same piece. In general, the notation within Sal and between Sal and Sant is remarkably consistent. They were both, like AL and BM51, copied from well-established written traditions in which scribes did not exercise any significant individual freedoms.

Figure 2. Responsories for the Ferial Office in Sal, Sant and BM51.

Much remains to be learned about the provenance and dating of Old Hispanic sources, and even the manuscripts studied here have occasioned differing views. Indeed, there is not a unanimous view about any of the manuscripts under study here. The following summary account relies on the list of sources given in the appendix to Hornby and Maloy's ‘Melodic Dialects in Old Hispanic Chant’ (see fn. 1).

AL has been dated in both the tenth and the eleventh centuries, and its provenance has been most often said to be in or near León but perhaps copied from a model from as far away as Toledo or the southern part of the peninsula.Footnote ⁹ One item in this manuscript has not yet been satisfactorily explained, however, and that is the appearance at the end of the Mass for St Leocadia (fol. 49v) of two pieces with the rubric Ad sepulcrum. St Leocadia was from Toledo, and there are early references to her sepulchre as a site of devotion there. It is not clear why a manuscript copied anywhere else but Toledo should include such a rubric, even if it was being copied from a model that did indeed come from Toledo.

Miguel Vivancos views Sal, on the basis of writing and decoration, as similar to the products of Silos and reports that it was copied by Cristóbal in 1059.Footnote ¹⁰ Similarly, the scribe of Sant, one Pedro, is described by Manuel C. Díaz y Díaz as having finished his work on the manuscript in León in 1055 but to display clear evidence of training in or near Silos and to display a debt to the style of Florencio of Valeránica. He notes that Sal could be thought of as a ‘twin’ of Sant but for its quality. Of Pedro he says that ‘at no time can he be considered to be from León on the basis of his writing’.Footnote ¹¹ BM51 is most often said to be from Silos, but some have expressed doubts.Footnote ¹²

The musical notation of these sources does, nevertheless, offer some evidence that is clearer than arguments based on the palaeography of the texts and leads to conclusions different from those just cited. The comparison of manuscripts based on the handwriting of the texts ultimately rests on questions of style. To compare them based on their musical notation, however, rests in the first instance on whether they use the same or different neumes in any given position. Different scribes might of course write one and the same neume in slightly different ways. But in comparing two manuscripts on the basis of their neumes, one can determine the frequency with which the two do or do not use the same neume in the same place. This is not a matter of style, about which distinguished palaeographers might disagree, but rather the sum of a series of questions to which the answer is either yes or no.

What is clearest is that Sant and BM51 cannot be from the same place. The dominant opinion about BM51 is that it is from Silos. If Pedro, the copyist of Sant, shows palaeographical signs of being from Silos or nearby, though he was active in León, then he must not have copied the music, for this music unequivocally belongs in the same family as Sal and AL and not in the family of BM51 and other manuscripts from Silos and nearby. Similarly Sal cannot be from Silos and must instead be from León or wherever AL and Sant were copied.

AL, Sal and Sant clearly belong together somewhere, and that is not in Silos. Where then? The one thing of which we can be surest is that Sant is from León, for it was commissioned for King Ferdinand I by his wife, Queen Sancha. Hence, on the basis of the musical notation, AL and Sal must be from somewhere in the Leonese orbit. BM51 clearly belongs somewhere else. That somewhere else may well be Silos, but it is certainly somewhere to the east of León, given its relationship (in regard to the responsorial psalm tones) with other manuscripts thought to be from Silos and the Rioja.

What then of the origins of AL?Footnote ¹³ Beja, to the west, and a hypothetical manuscript from 806, has been proposed, as have Toledo and the southern part of the peninsula, from which refugees would have brought manuscripts to the north. The rubric ad sepulcrum at the feast of St Leocadia in AL points to Toledo, but there is no other manuscript from Toledo that is anything like it. T6 is perhaps a partial exception. Although its musical texts were clearly intended to receive notation, only a few texts were actually so supplied. This notation is in at least three different hands and is of the type described as northern – that is, the type of notation found in AL and related sources and in the manuscripts from Silos and the Rioja. One of these hands bears a strong resemblance to those of the León group and to the hand of BM51. The notation of these manuscripts is in general more upright and could be said to be more elegant than the other northern notations in T6 and in BM45, a manuscript thought to be from Silos. The first part of Figure 3 enables a comparison between AL and one of the hands in T6 and shows their notation for these pieces to be virtually identical. The following pages of Figure 3 enable a comparison of AL, T6 and BM45. Here again, AL and T6 are quite close, and BM45 clearly stands apart in its choice of notational symbols, though one could reasonably conclude that all three manuscripts transmit versions of the same melodies. BM45, furthermore, is instantly recognizable as the work of a different and rather cruder hand, like one of the other hands in T6 and the hand of the few examples of notation in BM46, also usually thought to be from Silos. Unfortunately, the melodies shared by AL, T6 and BM45 are not found in any other manuscripts. Hence, one cannot bring other witnesses to bear on whether the hand of the shared pieces in T6 is more like, say, BM51 than like AL. Furthermore, the shared pieces do not include any responsory verses, which could point clearly in one direction or another.

Figure 3. Pieces shared between T6, BM45 and AL.

From the southern part of the peninsula, that is from south of Toledo, we have no sources of consequence at all.Footnote ¹⁴ Do the similarities between AL and T6 suggest a relationship of AL to Toledo? That would require a belief that T6 is from Toledo, as has usually been thought. But the unusual character of T6, with its several ‘northern’ hands and its sparsity of notation, suggests that its origins need to be rethought. At the same time, the origins of all the manuscripts usually labelled as being from Silos should probably be rethought, as there are at least two quite different styles of notation present in them – one represented by BM51 (if indeed it is from Silos) and one by BM45. Perhaps the best we can do, then, is to say that given its strong similarity to Sant, which we know to be from León, AL is likely to be from the north and in a region around León or to the west and not to the east, where we enter the terrain of BM51. This could include Beja, but we have no examples of musical notation from there.

Since there is the possibility that AL was copied from a model that came from somewhere other than León, we might ask what this model could have been like. It would surely have been an antiphoner, like AL, with texts and musical notation. Such a collection of texts must have existed by the time of the Oracional visigótico (before 711), but AL includes material not present in the Oracional, and its model must therefore be from well after the beginning of the eighth century.Footnote ¹⁵ Can we imagine an antiphoner with the elaborate set of texts implied by the Oracional and expanded upon in AL without musical notation?

I find this hard to imagine. AL and other sources preserve an elaborate and intricate liturgical structure in which the relationships of text to melody, both within individual pieces and among related pieces, was carefully thought out. It is hard to imagine that text and melody were not worked out simultaneously. It is almost as hard to imagine that the elaborate structure of the texts, having been conceived in relation to melodies that would set them, circulated for very long before someone thought to find a way to write down the melodies and the texts together. One might object that the Sextuplex manuscripts in the Gregorian tradition provide a counter-argument in that they are the earliest sources, and they leave no physical space for notation. But missals and breviaries with complete texts and no notation have existed down to the present. The absence of notation in these early manuscripts, then, cannot be thought to mean that there was no notation at the time of their copying.

Only one conclusion is certain. We know very little about the dates and provenance of the Old Hispanic sources, and the whole complex requires renewed study. One of the most promising tools for establishing relationships among these sources, however, has not yet been fully exploited. That is musical notation. Exploiting it requires the tedious work of copying all the pieces in all the sources in parallel and comparing them symbol by symbol. For example, there are several different forms of the scandicus that are clearly distinguishable from one another to a much greater degree than different versions of the letters of the alphabet. This notation thus provides a much more concrete basis for judging similarities and differences among sources than does palaeographic study of the texts.