I'm not going to attempt to provide a comprehensive guide to medieval mensural notation here—that'd be a major undertaking and I don't expect to get around to it any time soon. I hope though that this page will at least familiarize readers with the workings of my transcriptions of the Cantiga music, explain my editorial symbols, and generally give the performer enough information to get on with interpreting the original notation, and perhaps performing directly from it. It should also make it possible to study published versions in fully modern notation with a better informed, more critical eye.
My transcriptions are based solely on [E] for the 403 cantigas where that manuscript has musical notation. For the nine cantigas that appear twice in [E], I have transcribed the music only once. In the case of CSM 340 (= 412) have I used the later, higher-numbered version, in order to match the lyric structure. For the other eight, I have used the lower-numbered version, and pointed out any interesting differences in the annotations, although all are minor.
There are eleven cantigas which only have music in [To], and I have transcribed these too for the sake of completeness. The original notation for these is rather different from [E], mostly using o and e as the basic long and short mensural units rather than on and o, and generally being less strict in terms of the mensural values of the ligatures. For this set, the Transcription view (see below) uses slightly wider spacing for the stave lines, which gives a better approximation of the [To] style, as well as allowing the presence or absence of internal joining lines between note bodies to be seen more clearly.
Prologue A is just a poetic introduction without music, but five other cantigas are missing their musical notation: CSM 298 and 365 have ruled but empty staves in [E]; CSM 402 has unruled gaps in [E] where the staves should be; and CSM 408 and 409 only appear in the unfinished [F] manuscript, which has empty staves for all of its 104 cantigas.
I have no transcriptions from [T], since all of its cantigas are also found in [E]. Although this is a possible future project along with doing the rest of [To], I'm afraid neither is very high on my To Do list, which nevers gets any shorter however many things I finish. For the performer, I believe that having a complete set of transcriptions from [E] is by far the most important thing, and for the moment I'm content to have completed that work.
For each cantiga, three different presentations or ‘views’ of the music are available from the drop-down menu at the top of the Music tab. Each of these is generated automatically, being just a distinct rendering in the web browser of exactly the same internal Neumat transcription data, but each is designed for a slightly different purpose.
2.1 Transcription view
Until January 2014, just a single presentation of the musical notation was available on this website, which was the one now called the Transcription view. This was always intended as a ‘diplomatic’ transcription from the manuscript¹, i.e. one in which all of the original information is preserved as directly as possible, including features such as decorated capitals on the stave, variable numbers of stave lines, variable clef positions, original crossings-out by the scribes, and a complete representation of manuscript stave breaks, as well as explicit marking of my own editorial changes to fix obvious scribal errors and omissions.
 This is true despite the fact that the text underlay in the music transcriptions has always been, and will continue to be, my own edited text (based on Mettmann with corrections and modified spelling), rather than the original text from the manuscript in question. That's not a cop-out but a necessity: editorial changes to the musical notation only make sense in combination with the edited lyrics, and trying to preserve the original underlay would serve no purpose.
2.2 Square notation (normalized)
Whilst the Transcription view continues to be very useful when initially studying a piece, as well as being especially valuable to paleographers and musicologists, there is no doubt that many of its features are just distractions when it comes to actual performance, and it became clear that a simplified version would be a very useful addition, even without going all the way to modern notation. To address this need, a ‘normalized’ view of the square notation was added (and for a brief period labelled the Performance view), with the following features:
- All staves have five lines.
- A single clef at a fixed position is used throughout a single cantiga. Perhaps surprisingly, this is possible in every case without the use of leger lines, and only using clef positions that are common in the original manuscripts, i.e. the C-clef on any of the top three lines or the F-clef on the middle line. Note that C on the top line and F on the middle line are completely equivalent, but the F-clef is used unless there are more than twice as many C-clefs in the source. Also, the F-clef is used on the middle line, as long as this does not require leger lines, in all cases where the source exclusively uses the F-clef, even if using the C-clef would give a more centred vertical distribution of notes.
- Decorated capitals are omitted, being replaced by Refrain and Stanza markers as appropriate, with double bars at the end of sections.
- Annotations and editing marks are omitted, but editorial changes are kept and still highlighted in red.
- Original scribal crossings-out are applied, such that the affected note bodies and stems are not shown at all, where in the Transcription view they are shown struck through.
- Custodes are omitted. (There is only one in the whole [E] manuscript anyway.)
- Manuscript stave breaks are only shown when they are necessary to terminate the scope of an accidental. They are displayed at a smaller, less intrusive size than in the Transcription view, and never used in groups (see the tables below for details).
- To save screen space and clarify structure, reprises of refrains are omitted in the simple cases, i.e. where there is a complete or partial restatement of the refrain after the stanza. However, in all more complex cases where multiple stanzas are written out in the original source (e.g. CSM 87, 160), these are kept in the normalized view. Note that this hides any notational variations that might exist between the inital refrain and the reprise, but these can still be studied in the Transcription view.
Last, but not least, the normalized view of the square notation uses a more elegant, rounder rendering style based on modern editions of plainchant, in particular the Editio Vaticana, rather than on the CSM manuscripts. I hope will be familiar to medieval musicians and perhaps a bit easier on the eyes.
2.3 Non-mensural round notation
Although I would heartily recommend that all performers of medieval music study and even play from square notation as much as possible, I do realize that many will come to the Cantigas de Santa Maria with little or no experience of it, and I certainly do not want this to be a barrier at any level. To address this need (in part, at least) I added a third view of the music in March 2014: a transformation into non-mensural round notation, on a treble clef, with all ligatures expanded out to their component notes, so that the melodic contours can be read more easily.
The ‘non-mensural’ aspect here is important and worth explaining. What it means is that the notes on the modern stave represent pitch only, without rhythm. The mensural information is only present in the original square notes, ligatures and divisions, the shapes of which are laid out above the modern staves, and the rhythm must still be read from these. I cannot stress too highly that the absence of rhythmic information in the round notes is definitely not intended to invite a rhythmically free interpretation of the melodies: for the vast majority of the Cantigas de Santa Maria, which clearly have solid and regular rhythms, this would be entirely inappropriate.²
Here's a run-down of the features of the round notation view:
- The treble (G) clef is employed throughout. This should be familiar to the widest range of musicians, and suits the pitch range of the Cantigas very well. Leger lines are used when required.
- Accidentals in the original manuscript are converted automatically into key signatures wherever this produces the simplest reading. In fact, due to the melodic nature of the Cantigas, the only key signature ever used (apart from none at all) is that with a single B flat. In nine cases, the mudanza (the first part of the stanza music) has a different key signature from the refrain and the vuelta (the last part of the stanza), and here a B natural is used in the key signature to mark the change as appropriate (see CSM 85, 118, 124, 145, 176, 237, 317, 322 and 376). In a few other cantigas, just B flat accidentals are used, rather than a key signature, and the handful of E flats that occur in the manuscripts are only ever marked as accidentals.
- Regarding both key signatures and accidentals: please note that these are intended to accurately reflect what was explicitly notated in the manuscripts. In several cases where the manuscript has a clear scribal error that produces inconsistencies between accidentals in repeated phrases, I have corrected and annotated these. I have not, however, gone all out with the editorial marking of musica ficta, i.e. accidentals that are implied by the melodic mode but generally left unwritten in the original sources. I leave that to the performer's judgement.
- As mentioned above, the original square notes, ligatures and divisions are laid above the modern stave, keeping the mensural information all in one place. (Placing the divisions on the staves could make them very difficult to distinguish from the bar lines—see the next point.)
- The bar lines that I have added are actually generated from the metrical structure of the lyrics, although I expect them to be a very useful aid in determining the
rhythm of the music:
- A single bar line appears immediately before the syllable in each lyric line or hemistich that carries the metrical stress (i.e. the stress that matters for the rhyme). The initial assumption in working out a mensural interpretation of the music should always be that this coincides with a rhythmic accent: in other words, what would actually be the start of a bar in fully modern notation.³ So, rather than working from the beginning of each line of music to determine the mensuration, I would recommend starting from the single bar lines and working backwards and forwards. This may not lead ultimately to the best solution; I am merely stating that it is the best place to begin.
- A double bar line appears at the end of each lyric line or hemistich. Although I think this is a less reliable rhythmic indicator than the single bar lines representing stress, in most cases this will correspond to the end of a bar in fully modern notation, and it should be quite easy to determine how many rhythmic beats there are between single and double bar lines, since they are never very far apart (generally one or two modern bars). Exceptions to this are most likely to be the result of upbeats, where the first one or two syllables of the next line or hemistich fall at the end of the bar.
- Following the common convention for non-mensural notation, all notes are shown black, with no stems. Notes that come from plicas in the original notation (i.e. those written as stems, rather than bodies) are shown smaller. If you're working on your own mensural interpretation and want to print out the scores and fill in bodies and stems yourself, trying clicking on the score and typing the letter 'o', which turns the black notes into empty outlines.
- The individual round notes comprising a single ligature in the square notation are joined by a slur (which should always correspond to a single underlaid syllable).
- Finally, a checkbox option is provided to expand or collapse long plicas (oron onron onod oon ood and obod) and augmented (‘dotted’) notes (oe one and ono). When collapsed (the default) these use as few round notes as possible (two for plicas, one for augmented notes), avoiding repeated notes on the same pitch and giving the most natural melodic reading. When expanded, the number of round notes corresponds directly to the square note bodies in the original notation, thereby matching the mensural reading more closely. The latter system is commonly seen in modern transcriptions of the troubadour and trouvère repertoire, and some of you may be more comfortable with it.
 I have no plans to put up full rhythmic versions of the Cantiga music on the website, but if you would like them I'd recommend Chris Elmes' editions, published by Gaïta Medieval Music. As of March 2014 his first two volumes are out of print, but new editions are being worked on, with the first one expected to be on sale again over the next few months.
 In more medieval terms (for those used to thinking in them) this position is most likely to mark a division between perfections, i.e. a group of three beats in cases where the cantiga as a whole seems to have a ternary (modal) rhythm.
3. Editorial symbols
I'll start the detailed descriptions of my symbols with a brief run-down of the stave elements that have editorial meaning but no direct musical value, before moving on to more important things like clefs, divisions and, of course, the notes and ligatures. Most of these are only used in the Transcription view, the only exception being manuscript stave breaks, some of which need to be retained after accidentals in the normalized square notation in order to show where the accidental scope ends.
|^||decorated capital||Transcription view only. This box shows the location of an original decorated capital that sits within the score in the [E] manuscript. These always come at the start of a section—refrain, stanza or (sometimes) the reprise of the refrain after the stanza—so I've kept them in stylized form as a useful visual indication of the overall structure.|
|$||missing decorated capital||Transcription view only. Rare; I only use this symbol in the handful of cases where a space was clearly left on the stave for a decorated capital, and the required letter was not included in the normal underlaid text, but either the capital was never drawn or (possibly) it was erased later. This may be of interest mainly to paleographers, perhaps, but I do find that showing the location of such gaps is a useful additional indicator of overall structure (see Cantigas 250, 258 and 317 for examples).|
|>||manuscript stave break||
A wavy line shows the location of a break between staves in the original manuscript score. In the normalized Square Notation view these are shown at reduced size and only after accidentals, where they serve to show where the scope of the accidental ends. (If a flat has already been cancelled by a natural, no break symbol is necessary.)
In the Transcription view manuscript stave breaks are more prominent. A single wavy line > indicates a break between staves in the same column. Two wavy lines together > > are used to indicate where the score starts a new column on the same side of the folio, three > > > show where the score continues onto the reverse of the same folio, and four > > > > show where the score continues onto a new folio.
|X||deletion||Transcription view only. This mark shows where I have removed one or more clearly incorrect musical symbols in the original manuscript. In many cases this is an extra note or ligature that has no underlaid text. Checking the Expand edits box at the top of the Transcription view will display the deleted elements, using the bracketed notation described next.|
|[ old : new ]||edit||
Transcription view only. These symbols only appear if you check the Expand edits box. (If the latter is absent, it means there are no edits to expand.) They indicate places where I have made an editorial correction to the music in order to fix unambiguous scribal errors and to ensure correct fit with the lyrics for every stanza. Every edit is accompanied by an explanatory note. In some cases there is no old content—which therefore indicates an insertion; and in others there is no new content—which indicates a deletion. In the default view when edits are not expanded, only the new content is displayed (in red), or the X symbol in the case of a deletion (see above).
Please note that at this stage of my work on transcribing the music of the Cantigas my chief concern is to ensure that all of the music fits all of the lyrics, so as to make it possible to underlay every stanza without a problem. I also try to fix genuine scribal mistakes wherever possible, such as stretches of music being written at the wrong pitch (perhaps because the scribe misplaced the clef). I have, however, given scant attention for the moment to mensural problems, for example when a long on has been written in a context where a short o would seem more correct, through comparison with other musical phrases. By the way, if you see a crossed-out note body or stem in the Transcription view, for example ox or onx, remember that this is a reproduction of an original scribal correction in the manuscript, not one of my edits. In the Performance view, these crossings-out are applied so that the affected body or stem disappears completely.
4. Clefs, accidentals and divisions in square notation
In reading the descriptions of clefs and accidentals in the table below, remember that the lines and spaces of the medieval stave basically represent the notes of a diatonic scale on C—just like a modern stave regardless of the chosen clef—with the same natural sequence of tones and semitones as in C major, except when ‘flat’ symbols are written. The actual physical pitch identified with C is not important, however, and it is common practice to transpose the tunes to any modern key in order to suit the voices and instruments present.
|0c or 0c||C-clef||
The C-clef brackets the stave line that is to be read as the note C in the diatonic scale. In principle it can appear on any stave line, although in practice it keeps to the middle line or above. Rather than using leger lines to extend the stave, standard scribal procedure at the time that the Cantigas were written down was to move the clef position (or change clef) within a piece of music as necessary in order to keep the notes inside the usual five-line stave. My Transcription view reproduces this faithfully—along with occasional variations in the number of stave lines—so if you read from this directly (or indeed, from the manuscript itself) be on the alert for moving clefs, otherwise you can easily end up playing bits of the melody a third or more too high or too low. In my normalized Square Notation view, a single clef in a fixed position on a five-line stave is used throughout, as explained above.
|0f or 0f||F-clef||
The F-clef brackets the stave line corresponding to the note F in the diatonic scale. It is almost always on the middle line of the stave, notwithstanding the occasional four- or six-line stave in [E], some apparent scribal errors, and a very few apparently deliberate exceptions due to an exceptional ambitus (melodic pitch range) such as the run of four staves with the F-clef on the second line in CSM 411. Care still needs to be taken to keep track of clefs in a single piece, however, as there are frequent switches from c to f and vice versa.
Note that the F-clef on the middle line of the usual five-line stave is completely equivalent to a C-clef on the top line. More generally, whilst the C- and F-clefs cannot be rigidly identified with any particular octave (in modern transcriptions the C-clef is usually treated as marking the C above middle C, but middle C itself would do just as well), when the two are combined in one piece, the F-clef is always a fifth below the C-clef, never a fourth above.
|@ or @||flat||
The ‘flat’ symbol behaves more or less like the modern ♭ sign: it lowers, by a semitone, the note of the diatonic scale corresponding to the symbol's position on the stave. In practice, it is almost applies to B, turning it into B♭. There are also a few cases of E♭, for example in CSM 189.
Note that the flat sign normally appears immediately after a clef in the [E] manuscript, and remains in effect (flattening the B or E) until the end of the manuscript stave, which is marked in my edition with a wavy line > as discussed above. There are, however, plenty of examples where the flat is introduced mid-stave, so be on the lookout for it, and remember that even in this case the flat lasts until to the end of the manuscript stave and does not merely affect the next applicable note as an accidental.
The ‘natural’ symbol also behaves like its modern counterpart ♮, that is, it cancels the effect of a preceding flat on the same stave, restoring the corresponding note of the diatonic scale to its normal relative pitch—in practice, B♭ back to B♮ and E♭ to E♮.
The use of this symbol is rare in the [E] manuscript, since staves are generally short and the effect of a flat expires at the end of the stave anyway. Its appearance—at least as far as it is possible to determine from Anglés' facsimiles—is rather variable, in some places looking like just the top crossbar and lower right stem of the modern symbol (e.g. in CSM 217), and in others like the modern symbol but with no vertical space between the crossbars (e.g. in CSM 371). I have therefore used the modern shape for clarity, even in the Transcription view.
||1 |2 |3 |4 |5||division||
Division lines come in a continuous range of lengths in the [E] manuscript, from the shortest that crosses a single stave line, to the longest that crosses the whole stave and beyond. Rather than trying to reproduce their variety with pinpoint precision (which could single-handedly slow the transcription process down to a crawl) I have restricted the divisions to just five lengths, each of which can be centred at any vertical stave position. This, I believe, gives a good compromise between efficiency and notational accuracy. In fact, it is rather likely that fewer than five lengths would be sufficient to capture the ‘musical meaning‘ of the divisions, but I have preferred not to prejudge that particular issue, instead capturing a fair representation of the manuscript which can be analysed properly later.
What the actual ‘musical meaning‘ of the divisions is, in each case, is a matter of interpretation and debate. It's clear that the longer lines generally appear at metrically significant points—the end of a hemistich or line of verse—and therefore correspond to some kind of rhythmic cadence or caesura. The shorter division lines occur more freely, sometimes doubled (as in CSM 1), and are more akin to dots and rests in modern notation. As with the duration of the notes and ligatures, however, I am not going to attempt to give definite mensural values for the divisions here, and that is unlikely to be possible anyway.
Note that in the normalized Square Notation view, the longer divisions (lengths 3, 4 and 5) are always centred vertically on the stave, regardless of their original position, but their lengths are not changed. The shorter divisions keep their original variable positions. (Adjusting the shorter divisions to centre them at the final vertical position of the preceding note body would also be acceptable, and I may implement this in future.)
Transcription view only. The custos (pl. custodes) is very common in [To], and I have included it in my eleven transcriptions from that manuscript. Although it looks a bit like a note, it is easily distinguishable in practice. It always appears at the end of a stave, and is aligned so as to show the pitch of the first note on the next stave. In manuscripts this is useful for continuity across column and page breaks, and is particularly important when the clef on the next stave is in a different position: the custos makes the jump much easier to take in at a glance.
5. Notes and ligatures
For those of you who would like to play or sing directly from the square notation, reading the melodic contours isn't too difficult, and is mostly just a matter of identifying the stave pitches of the component symbols in turn, working from left to right as you would expect. It's important though to know that pitches may be represented by stems as well as body shapes, as shown in the following table:
|single pitch||e o on (in isolation)||0e 0o 0on||0c 0e 0o 0on|
|two pitches, step down||
on (in some ligatures)
bo (in rare cases, listed below)
|0ra 0raa 0ron 0owon 0boron||0c 0ra 0raa 0ron 0owon 0boron|
|two pitches, step up||oso bod od (in ligatures)||0oso 0oswo 0bod 0rad 0odeyeyed||0c 0oso 0oswo 0bod 0rad 0odeyeyed|
The descending oblique a represents a sequence of two pitches going down the scale, determined by where the upper and lower ends of the shape are placed on the stave. In the Cantigas this is never more than an interval of a second or a third down. In terms of pitch contour alone, a is the same as oyo, but they are not generally equivalent in terms of rhythm and duration (see the next section).
Similarly, the ascending pair oso (known as the pes or podatus) may represent one or two steps up the scale, depending on the note spacing, i.e. an interval of a second or a third. (Cantiga 422 is unique in having a fifth on oswo at the start of the stanza.) Again, oso is equivalent to owo with regard to pitch, but not rhythm and duration. For those completely unfamiliar with square notation, it's perhaps worth clarifying that oso always represents two pitches in sequence, like owo. There are no chords in the Cantiga music!
A stem that represents a separate pitch—transcribed in the table above as a smaller round note—is known as a plica, and notes or ligatures that have these are said to be 'plicated'. In the case of bod and ron the whole shape may be also called a 'plica'. The plica ascendens bod represents two notes separated by a single pitch step up the scale (an interval of a second) starting on the pitch where the square o body is placed. The upward stem on the right may also be attached to any note body in a ligature (as in od, ad and even ed) and seems always to indicate a plica—i.e. a step up—in every Cantiga that I have transcribed (including all of [E]).
Likewise, the plica descendens ron indicates two notes separated by an interval of a second down from the pitch where the square o body is placed. In ligatures the downward stem on the right is never found attached to oblique a or rhomboid e bodies, but in on it does seem to represent a plica except:
- when initial (e.g. in onron and onwa);
- when it forms part of any variant on the climacus onyeye, whether initial or not (e.g. in royoonyeye); and
- in a few other rare and debatable cases that are not worth enumerating in a brief overview.
An upward stem on the left-hand side of a note generally has no pitch value of its own. Either it is a purely conventional prop with no separable meaning as in bod, or it is a mensural indicator (not always reliable) of shortened duration, halving the length of the first two notes of a ligature (e.g. owo is long but bowo is short). There are a few cases however in which it makes more musical sense to read this stem as a ‘pre-plica’, i.e. a pitch step down to the body note, specifically in the shapes bo (CSM 36), bon (CSM 7), bood (CSM 18) and boron (CSM 4, 10 and others).
A downward stem on the left of a note body is never a plica and has no pitch value.
Finally, be aware that internal joining lines between note bodies have no musical value at all, so for example owwo and owwjo should be treated identically, as should oswo and oswjo.
If you have managed to follow all that, you will have realized that owo, oso and bod have identical pitch contours, as do oyo, ra and ron. Mensurally, however, they are quite distinct, as outlined in the next section.
There are well over two hundred different note and ligature shapes in the [E] manuscript alone, and as I said at the top of the page, a comprehensive guide to medieval mensural notation isn't really feasible here. However, what I can do to point you firmly in the right direction is reproduce—with kind permission—the data given by Chris Elmes in the introduction to the first volume of his performing editions, where he lists his ‘initial assumptions’ about length when transcribing into modern notation.
Length categories below are S = semibreve, B = breve, L = long, L+ = ‘perfected’ long. Remember that in ligatures of more than one note, these apply to the ligature as a whole, so the fact that bowo, for example, is B (breve) means that the pair of component notes is effectively SS. (Whether the division is actually binary (½ + ½) or ternary (⅓ + ⅔) is a matter of academic debate, but I'm in the binary camp.)
Single notes, plicas and two-note ligatures
|S||e (very rare as isolated notes in [E])|
|B||o ron bod bowo ba|
|B or L||or||owo ra|
|L or L+||or||on oron ood oso royo oyo|
The most important thing here is that interpretation as L or L+ cannot be fully determined by the shape alone; rather it depends on the ‘rhythmic mode’ and a lot of contextual cues. See Elmes' introduction for more on this important aspect of the mensural notation of the Cantigas, or Anglés or Pla if you can find them and can read Spanish confidently.
|L||onyeye eyeye bowowo bowoso rawo bowoyo|
|L+||oyoyo royoyo owowo rowowo oyowo owoyo|
Again, the mensural reading L or L+ is dependent on context: these are just Elmes' starting points.
Ligatures of more than three notes tend to increase in duration in a more-or-less sensible and predictable manner according to the added note shapes, but the broader musical context is always crucial to determining the best interpretation.
In future I hope to add some examples to this page to help explain the interpretation of mensural square notation more fully. In the meantime, the best thing you can do to familiarize yourself with how it all works is to get hold of an edition of the Cantiga music in modern notation—either Elmes, which is readily available, or the volumes by Higinio Anglés or Roberto Pla if you have access to a good library—and start comparing these with my transcriptions or the original manuscripts.
6. Current limitations
Lastly, although I am confident that the music transcriptions will serve their purpose very well in their current format, both as a solid base for performance and as rich data for future detailed analysis, processing and conversion, I wouldn't call them perfect just yet. So here are a couple of possible criticisms that I'd like to make before anybody else does.
- With only five division line lengths, and the same range of vertical positions as for notes, it is not possible to capture the exact appearance of divisions in the manuscript, and there are bound to be borderline cases where different readers would disagree. But even if Neumat supported a continuous range of lengths and positions, this would still not truly reflect the manuscript appearance, where all sorts of other factors come into play: the variable distance between stave lines (sometimes considerable); the fact that some divisions were drawn longer but only inked for part of their length; carelessness leading to slanted and curved lines; and so on. Ultimately, one has to leave facsimilar precision to the facsimiles themselves, and strike a sensible balance that gives us a machine-readable representation which is succinct and practical, but doesn't discard information that may be important. By the way, when in doubt, I have tried to give precedence to division length over position, since it seems unlikely that the latter has any inherent meaning, and also to consistency within a single Cantiga over needless variation.
- There is no explicit representation of the horizontal separation between division lines and the adjacent notes. The current Neumat rendering merely places the shorter lines (lengths 1 and 2) a bit closer to the preceding note than the longer lines (lengths 3 to 5) since in the majority of cases this matches the manuscript appearance and helps clarify the structure.