Modifications for metrical regularity

This page lists and categorizes all cases of textual modification and markup that I have applied to Mettmann's texts (over and above the basic corrections) in order to produce a version that the performer can rely upon to be 100% metrically regular. The goal is that there should never be any doubt as to how the words fit the music, and that every stanza can be underlaid, syllable by syllable, in exactly the same way.

I should stress that Mettmann already did a lot of work in this regard, and that all of the changes listed here are relative to his edition, not to the manuscripts from which his version was carefully and selectively constructed. There are many cases where Mettmann elides words, restores elided words to their full form, inserts or deletes words and otherwise ‘mixes and matches’ between manuscript sources in order to make lines of the right length. But Mettmann had to do everything manually, and by eye, with nothing like my software to aid him. It is also generally accepted in the present-day academic community studying the Cantigas that Mettmann was unduly biased toward the [E] manuscript in his edition, often at the expense of metrical regularity. Even more importantly, though, it must be remembered that his intended readers were language historians chiefly interested in faithfulness to the sources, with very different needs from performers. This latter consideration would certainly have limited Mettmann's options for achieving metrical regularity. In effect, the 581 changes listed on this page quantify the difference between his goals and my own.

Each category of modification or markup is described in more detail in the sections below. In the 248 cases where I have changed Mettmann's text, rather than merely marking it up, the third column in the main tables shows the change in isolation, otherwise keeping Mettmann's original spelling. The last column shows the correction highlighted in the context of the whole line, but with my modified spelling and your spelling preferences applied. Line numbers in the first column are my own, but as on the corrections page, these are cross-linked to the main text, which will give enough context to help the reader find the corresponding line in Mettmann's edition if necessary.

The second column, "Sources", comes with a small caveat: at present this is a merely list of the manuscripts that the cantiga in question originally appeared in, in some form. However, due to the loss of folios, or incomplete work (mainly in [F]) the specific referenced line may not have survived. So in a very small number of cases it may appear as though an alternative source exists for a line when it does not. I hope to make this more precise in future, but in the meantime I think the scarcity of such cases makes the inclusion of the "Sources" column worthwhile in its current form.

By the way, you might look at the Elision (forms of ũu) table below and worry that it has been accidentally truncated. But it is quite accurate: cases where I have had to elide ũu to un (without finding un in another manuscript) really do start from CSM 209. In fact, although the transition is less stark, you can see that the main Elision table has a similarly uneven distribution, with just 2 cases in the first hundred cantigas, 5 in the second hundred, 15 in the third hundred and the remaining 51 from CSM 301 to the end. Synalepha and diaeresis show the same pattern as well. One of the many tasks on my ever-expanding To Do list is to undertake some statistical analysis of this interesting feature. However, the general message (which won't be news to most CSM scholars) is quite clear: the earlier-composed cantigas found in [To] and [T] are metrically far superior to the later ones that only appear in [F] and [E], and need much less editorial attention.

Cantigas processed: 420 (including 2 prologues but excluding the nine repeated Cantigas)

Total lines: 21656

Textual modifications: 248 cases (11.5 per 1000 lines)

Type (click for details) Cases
Alternative form taken from another manuscript 69
Elision 73
Elision (forms of ũu) 59
Restore elided vowel 2
Alternative word form 17
Letter names 10
Insertions 9
Deletions 5
Moves 4

Metrical mark-up: 333 cases (15.4 per 1000 lines)

Type (click for details) Cases
Synalepha 105
Synaeresis 122
Diaeresis 59
Melisma 6
Breaking 41

Alternative form taken from another manuscript (69)

In this first category of modifications I collect together all those cases where metrical problems in Mettmann's edition can be resolved simply by reading from a different manuscript. You'll see from the totals that this accounts for about one-eighth of all fixes listed on this page.

As in the other tables below, the "Sources" column in this first table lists the manuscripts in which the cantiga in question is found, but for this special category I've highlighted in red the manuscript (or manuscripts) from which I took the metrically superior form. I have also given all footnotes for the referenced lines a standard format that begins with a list of the forms found in all manuscripts, along with Mettmann's own editorial choice. (Place the mouse cursor over the † symbol to see the notes.)

In the seven cases where [E] is the only manuscript listed, this means that the cantiga in question is one of the nine which are included there twice, and I have drawn from the second version. See the footnotes for details.

Other cases in which Mettmann's reading corresponds to none of the manuscripts, but has a metrical problem for which the manuscripts are able to supply an acceptable fix (often just restoring a missing word), are generally treated as editorial errors. You'll find these listed on the corrections page instead.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
1:59 To, T, E juygarjulgar dizend': “Assí verrá julgar, CSM 1:59Mett. juygar; [To] julgar; [E] juygar; [T] juigar (corrected from julgar).
11:47 ToT, E QuandoQuand' Quand' est' a compa' oía CSM 11:47Mettmann quando; [To] quand; [T] quand; [E] quando.
12:7 To, T, E ArcebispoArçebisp' O Arcebisp' aquel día | a gra missa be cantou; CSM 12:7Mettmann Arcebispo; [To] arçebisp; [T] arcebispo with partial erasure of o; [E] arcebispo.
25:45 To, T, E omagẽesomages e as omages e mostrou CSM 25:45Mettmann omagẽes; [To] omagẽes; [T] omages; [E] omagẽes.
27:19 To, T, E omeesomes Os Apóstolos, com' mes de bo se, CSM 27:19Mettmann omees; [To] omees; [T] omes; [E] omees.
28:9 To, T, E poboopoblo foss' o blo que guardar CSM 28:9Mettmann poboo; [To] poblo; [T] poboo; [E] poboo. Compare lines 24 and 41 where the three-syllable form poboo is necessary.
33:26 To, T, E ũuun mẽes. E u saltar CSM 33:26Mettmann ũu; [To] ũu; [T] ũu; [E] un. Given Mettmann's general preference for [E], it is very odd that in this case he discards the form un, which scans correctly, in favour of ũu from [To] and [T], which does not.
35:39 To, T, E cossarioscossairos de cossairos que faa | e aquel mar mal assaz. CSM 35:39Mettmann cossarios; [To] cossarios; [T] cossairos; [E] cossarios. Given the scarcity of Mettmann's form cossarios (CSM 379:17 has the only other in-verse example) it is difficult to know whether to read it as having three natural syllables (cossarios), in which case it would be unproblematic in this position, or four with antepenultimate stress (cossários, paralleling words like hóstïa, espécïas and empérïo). However, since [T] has the unambiguous three-syllable form cossairos I have chosen to use it here (cf. cossairo in CSM 236:8, also with three syllables).
46:4 To, T, E omeesomes vai ant' mes descreúdos. CSM 46:4Mettmann omees; [To] omes; [T] omees .o + .ron; [E] omees .o + .oron. Beware that Elmes has made a mistake in transcribing this line as vai ant' *omees *descreudos, where the last two starred words are each divided wrongly, and the rhyme with sabudos is broken.
50:11 ToT, E mẽesmomeesmo Ca Déus e si meesmo ele mingua no há, CSM 50:11Mettmann mẽesmo; [To] meesmo; [T] meesmo (direct underlay to one ligature + one note); [E] mẽesmo. The form found in [To] and [T] is more amenable to the synaeresis required here. Note that mesmo doesn't occur in that exact form in the cantigas, although the feminine mesma crops up in CSM 194:15 and 360:9.
63:11 ToT, E delete E De bõos costumes haa assaz CSM 63:11Mettmann E de; [To] De; [T] De; [E] E de. The extra word E is in keeping with the usual narrative style of the CSM, but is quite unnecessary to the sense, and it may well be the result of confusion: the [E] lyric scribe perhaps wrote de when he should have left the D for the illuminator, and then the illuminator assumed that another E was needed.
64:64 To, T, E peepe que a çapata ao assí se ' aprês CSM 64:64Mettmann pee; [To] pe; [T] pee; [E] pee.
65:9 ToT, E sobervi'soberv' houv' u séu feegrês soberv' e loução, CSM 65:9Mettmann sobervi' e; [To] soberve; [T] soberv e; [E] sobervio &. Although Mettmann's sobervi' e can be read as having the right number of syllables, I think the [To] and [T] form is clearer, and it matches CSM 152:5 and 369:13.
65:132 To, T, E perigoosaperigosa est' hóm' e sentença jaz mui perigosa; CSM 65:132Mettmann perigoosa; [To] perigoosa; [T] perigoosa; [E] perigosa.
69:50 ToT, E galogal' que ja oía o gal' e a rãa. CSM 69:50Mettmann galo; [To] gal; [T] gal; [E] galo.
69:58 To, T, E çizillãaçiziliãa ou de Salérna, a ciziãa.” CSM 69:58Mettmann çizillãa; [To] ciziliãa; [T] cezilãa; [E] çizillãa.
72:26 ToT, E casacas' de sa cas' entô saiu; CSM 72:26Mettmann casa; [To] cas; [T] cas; [E] cassa (sic).
75:10 ToT, E orgullos' e | sobervi'orgulloso | e soberv' mui riqu' e muit' orguoso | e soberv' e torticeiro; CSM 75:10Mettmann orgullos' e | sobervi' e; [To] orgulloso e sobervie; [T] orgulloso e soberve; [E] orgullose sobervie. This is a rather clear case where Mettmann's overriding preference for [E] gets him into a complete mess.
76:23 T, E sãosã' o méu sã' e vivo vĩir se lijô ne mal.” CSM 76:23Mettmann são e; [T] sane; [E] são e. Here I have used the elision from [T] but regularized the spelling.
78:63 ToT, E aquelaquele Quando viu el conde aquele que chegara CSM 78:63Mettmann aquel; [To] aquele; [T] aquele; [E] aquel.
96:38 T, E ũuun per que cada u seja ê sabedor CSM 96:38Mettmann ũu; [T] un; [E] ũu.
96:40 T, E almaa alma e come dïabres a alma cuidara CSM 96:40Mettmann alma; [T] a alma; [E] alma (possibly with a small a inserted later into the preceding space, though this is unclear from Anglés' facsimile). In effect, [E] has the definite article a elided out of existence, as indeed Mettmann observes in a footnote. But [T] keeps it, and it is syntactically necessary, so I have used this form and marked the required synalepha.
97:30 To, T, E omẽesomees séus mees por el lógu' enou; CSM 97:30Mettmann omẽes; [To] omees; [T] omẽes; [E] omẽes. Here I have used the form omees from [To] in order to be able to mark synaeresis to fix the metre.
113:3 T, E insert ben que quando morreu por nós be sei CSM 113:3Mettmann suggests that the "primitive" metrical schema for the refrain may have been 9-9-9-9, but his own version comes out as 9-8-8-9. Given the words and music as they appear in [E], however, I have chosen to diverge at this point and go for a 9-8-9-8 refrain, and have thus restored the word be from [E] in this line, and reduced Mettmann's porende to porê (= [E] porend, [T] por en) in the next.
113:4 T, E porend[e]poren que porê se foro fender. CSM 113:4See note to line 3.
113:23 TE insert E E começaro Déus beinzer CSM 113:23[E] has & começaron Deus a bẽeizer, and [T] has & começar Deus a bẽizer. Mettmann has just começaron Deus a bẽizer, apparently deleting the conjunction to mitigate the metrical surplus, even though it looks stylistically necessary. The following word começar in [T] is very strange and must be a scribal error: neither infinitive nor future subjunctive make any grammatical sense in this position. The [T] form bẽizer (or beĩzer, with the tilde floating between e and i in typical manuscript fashion) is also odd. My assumption is that it is intended as a two-syllable word, pronounced [bei̯nʣeɾ] with the same stressed sound as treinta, and functionally equivalent to synaeresis in the normal three-syllable bẽeizer. All in all, it's a tricky line to sort out. However, Stephen Parkinson notes in [SP2007] that the preposition a is not essential between the verb começar and its complement, citing CSM 84:53 Enton [...] começaron bẽeizer // a Virgen Santa Maria as evidence. I have therefore decided to reject Mettmann's deletion of E on stylistic grounds, and to sacrifice the a instead. This, in combination with the two-syllable bẽizer, which I have respelled beinzer to clarify the pronunciation, gives the desired syllable count, and rhymes, and makes sense.
115:53 To, T, E ũuun cada u e séu leito CSM 115:53Mettmann ũu; [To] un; [T] ũu; [E] ũu.
119:22 TE omẽesomees mandou fiar a séus mees lanças e escudos, CSM 119:22Mettmann (I, II) omẽes; [E] omees; [T] omees. Although forms of this word with the tilde do, of course, occur in the manuscripts, Mettmann does not normally add it, and to do so in this case seems particularly odd when the word must clearly be compressed to just two metrical syllables, in order to maintain the consistent 8 + 6 rhythm of the lines without a break in the middle of the word.
119:25 T, E diaboosdiabos Mas ta tóste de abos compaa sobeja CSM 119:25Mettmann diaboos; [T] diabos; [E] diaboos. Compare line 5, where diabos is found in [E], and occupies the same metrical position in the line as here.
125:60 To, T, E gaanneygannei mui grand' algo que éu teo, | que gaei se trichaa.” CSM 125:60Mettmann gaanney; [To] gãei; [T] gaãnei; [E] gaãney.
125:84 To, T, E escaecemos'scaecemos te 'scaecemos mui tóste, | louca, maa e sana?” CSM 125:84Mettmann te escaecemos; [To] te scaecemos; [T] te escaecemos; [E] te escaecemos.
125:93 To, T, E ll'lle E o bispo chegou lógo, | e disso-e o novio: “Fól CSM 125:93Mettmann disso-ll'o; [To] disse-lle o; [T] dissoll o; [E] dissoll o.
127:40 T, E peepe fia o de téu fio, | e no esres a cras, CSM 127:40Mettmann pee; [T] pe; [E] pee.
132:61 To, T, E seerser e que houv' a ser grãado CSM 132:61Mettmann seer; [To] ser; [T] seer; [E] seer.
176:5 T, E tĩiantĩan a tia, por u crischão | que foi ontr' eles caer CSM 176:5Mettmann tĩian; [T] tĩan; [E] tĩjan. Synaeresis is clearly intended in [E], which has tĩjan laid under just two notes/ligatures .on + .bowowa, implying two syllables.

This line is a fine example of the music scribes, both in [E] and [T], blithely copying the refrain music as the vuelta even though the metrics are different and the words don't fit. Specifically, the music for the second hemistich has eight individual notes/ligatures, which fit the 7' metre of the refrain with its feminine rhyme, but not the plain 7 with masculine rhyme in the stanza. The fact that the lyrics were amended by a later hand to que foi ben ontr' eles caer is of no use to us whatsoever: this painfully naïve "correction" might allow us to sing the first stanza to the music, but it messes up the poetic metre and does nothing to fix the following stanzas—we can't insert ben in all of them! So, to make all stanza lyrics work with the minimum of rhythmic disruption, it seems that the only defensible option is to join up .ood and .on at the end of this line into one ligature, such that the final word stress falls at the same musical point. This is what I've done at E1 in the music transcription. The resulting shape .ohodwwwon is not one that I've seen elsewhere, but the intention is of course that it should sound exactly like the .ood + .on combination that it replaces, only with a single syllable sung to it.
192:119 T, E leeley d' a lei dos crischãos CSM 192:119Mettmann lee; T lee; E 192 lee; E 397 ley.
194:32 T, E ũuun catando-se u co outro | e que ja no se falava, CSM 194:32Mettmann ũu; [T] un; [E] ũu. In Mettmann's version of this line, the first hemistich has one syllable too many, whilst the second is missing one. Keeping ũu from [E] and carrying the syllable -tro of outro over into the second hemistich might work in theory (compare CSM 12:23 fe- zera and CSM 419:78 Jo- safas), but rhythmically and musically it sounds dreadful. I have therefore used the [T] form in the first hemistich and, in the absence of any scope for diaeresis in the second, inserted the filler word ja, which doesn't greatly affect the meaning.
213:53 F, E comigomigo E o démo contra eles | disse: “Que havedes migo? CSM 213:53Mettmann comigo; [F] migo; [E] comigo.
213:81 F, E delete o pela sa grand' homildade | nos déu pera o ida CSM 213:81Mettmann pera o ceo; [F] pera ceo; [E] pera o ceo. The [F] reading without the definite article is supported by CSM 313:51 os ollos a ceo alçou in [E], though in fact there the roles are reversed and [F] has a o ceo.
224:34 FE delete [e] de levar i a meni·a, | firo atal postura. CSM 224:34Mettmann I and II de levar la menynna. Given that the first hemistich of this line ends with an unstressed syllable in both manuscripts, it still needs another syllable to fix the metre. I have inserted the word i on the model of CSM 176:26 e levou i a omagen, and changed la to a accordingly. The original word la appears in both manuscripts but wouldn't normally be expected in this position (levar a or levá-la would be more regular), so it could be a Castilian-influenced copying error for i a or ia anyway.

Mettmann I fizeron; Mettmann II e fizeron; [F] fezeron; [E] fizeron. The e inserted by Mettmann in his second edition makes some sense, stylistically, but I don't believe it to be essential. If you prefer to keep it, then reading the second hemistich here as e fizeron tal postura would be an acceptable compromise.
224:39 F, E seerser houro de a levar | alá por ser soterrada CSM 224:39Mettmann seer; [F] ser; [E] seer.
228:34 F, E comocom' rono com' entrou dentro, | mostrando grand' homildade. CSM 228:34Mettmann como entrou; [F] com ẽtrou; [E] como ẽtrou.
238:8 F, E ChristoChrist' Jesú-Crist' e Guimaes | du jograr mao rafez, CSM 238:8Mettmann Christo en; [F] crist en; [E] christo en (using the common Greek-based abbreviation xp̃o).
238:32 F, E enono e noa do joízo | estarás a séu poder, CSM 238:32Mettmann eno; [F] no; [E] eno.
241:27 F, E menĩomeninn' e o meni' u vaso | e sa mão fiava CSM 241:27Mettmann menĩo un; [F] menỹn un; [E] menĩo un.
242:11 F, E EsteEst' Est' éra mui bo maestre | de pédra põer co cal, CSM 242:11Mettmann Este; [F] est; [E] este.
249:28 F, E aosos aa Virge groosa, | que os séus val e mantê. CSM 249:28Mettmann que aos seus; [F] que os seus; [E] que a os seus.
267:2 F, E perigoadoperiguado be póde valer a todo periguado. CSM 267:2Mettmann perigoado; [F] periguado; [E] perigoado. The spelling periguado in [F], whilst not the most common form, clearly indicates that just four syllables are required, and is backed up by the appearance of periguada in CSM 313:15 in both [F] and [E]. In [E] the synaeresis in perigoado is also clearly intended from the underlay to the music, with the metrical syllable -goa- given a single ligature .oweyeye.
270:19 F, E merceemerce' Gra merce' ao mundo | fez por esta Seor CSM 270:19Mettmann mercee; [F] merce; [E] mercee.
273:30 F, E ũuun de sacar end' u enteiro, | ne que e dissésse deo. CSM 273:30Mettmann ũu; [F] un; [E] ũu.
274:39 F, E insert el Quando ' esto diss' a Virge, |go xe-' el espediu. CSM 274:39Mettmann xe-ll' espedyu; [F] xell el espediu; [E] xell espedyu.
279:14 To, E fezfezo que me fezo mais verde mia coor CSM 279:14Mettmann fez; [To] fezo; [E] fez.
289:6 E gaannegãe que vos e de séu Fio | dos pecados perdoança. CSM 289:6Mettmann gaanne; [E] 289 gaanne; [E] 396 gãe. I have used the form from [E] 396, but synaeresis in gaanne is clearly indicated in [E] 289, with a single ligature .ra written above the compressed metrical syllable gaa-.
306:8 F, E ũuun a u hege e Roma, | e conta que foi assí, CSM 306:8Mettmann ũu; [F] un; [E] ũu.
306:18 F, E neũuneun ela de Spírito Santo |go se neú lezer. CSM 306:18Mettmann neũu; [F] neun; [E] neũu.
321:40 F, E corp'corpo a ela e a séu Fio, | tod' o córpo e a cara, CSM 321:40Mettmann tod' o corp' e a cara; [F] tod o corpo & a cara; [E] todo o corpea cara. Mettmann's version of this hemistich elides too many syllables, whereas both manuscripts are metrically correct, but in different ways. I have chosen the [F] version as it's easier to pronounce and fits the rhythm of the music better, and because corpo is a more important content word than todo.
335:13 F, E ũuun eno tempo dos genis | a ume e Cezia, CSM 335:13Mettmann ũu; [F] un; [E] ũu.
335:20 F, E ũuun quanto de comer haa |ra e u mao ano CSM 335:20Mettmann ũu; [F] un; [E] ũu.
337:4 F, E ũuun haveredes poi-l' oírdes, | que conteu a u barô CSM 337:4Mettmann ũu; [F] un; [E] ũu.
340:48 E tẽenteen a Déus, e que por mal tee CSM 340:48In this line Mettmann has tẽe from [E] 340. I prefer to use the form tee from the Festas version (see the main note above) in order to preserve the rhyme with cree and vee.
349:16 E omeesomes and' mes e moéres, | e séus fios e sas fias. CSM 349:16Mettmann omees; [E] 349 omees; [E] 387 omes.
349:27 E omeesomes ou d' mes e moéres | enfermos e mal coitados CSM 349:27Mettmann omees; [E] 349 omees; [E] 387 omes.
384:1 E insert muy A que por mui gra fremosura | éste chamada Fror das frores,
384:1 E ééste A que por mui gra fremosura | éste chamada Fror das frores,
384:2 E insert lle mui mais e praz quando e loa | séu nome que d' outras loores.
419:51 To, E escritura'scritura E lógo que chegaro, | com' a 'scritura diz, CSM 419:51Mettmann a escritura; [To] a scritura; [E] a escritura.
422:11 To, E fugistifogist' verrá, di-i como | co el fogist' a Egito. CSM 422:11Mettmann fugisti; [To] fogist; [E] fugisti.
422:33 To, E u | [foi]| u di-' o que sentiste | u pósto no monumento. CSM 422:33Mettmann has u | [foi] posto here, moving u to the first hemistich and inserting the word foi, which is not in [E], but is in [To]. Whilst this undoubtedly improves the sense, it breaks the scansion, spoils the repeated rhyme -iste in the first hemistich, and ignores the recurring pattern whereby the word u begins the second hemistich.
424:23 To, T ũuun porê cada u e levou CSM 424:23Mettmann ũu; [T] un; [E] ũu. [To] has llevou at the end of this line, an obvious Castilianism.

Elision (73)

‘Elision’, as I define it here, means the omission of a vowel, when a word that ends in a vowel meets another that starts with one. Normally it is the first vowel of the pair that is lost, but in a few cases it may be the second, especially where the first word is just the vowel itself.

Elision is a fundamental and pervasive feature of the language of the Cantigas de Santa Maria, but it has no reliable explicit marking in the manuscripts. The elided vowel is simply omitted, and the words are commonly run together without a space, although in many cases the remaining vowel is marked with a plica¹ (an angled line above, looking like a long and very fine acute accent) in order to show that this vowel starts a new word. However, these same plicas are used for other purposes than indicating word separation, so they're not necessarily a marker of elision. One of the main tasks for any modern-day editor of the cantigas is therefore to pick apart these cases and insert apostrophes to mark the ‘missing’ vowel, which is something that cannot be done mechanically—it requires a thorough knowledge of the language and comprehension of what the text is saying.

Given that elision is so common in the texts as we find them, the editorial use of supplementary elision to fix metrical problems is entirely defensible. My own use of this option, and of synalepha in which both vowels are kept but pronounced as a diphthong, are often alternative solutions when there is a surplus vowel, and therefore an extra syllable. My general approach here has been to use elision where it commonly occurs in the word in question, especially when the resulting conflated syllable will fall on a single note in the music, rather than a ligature. Conversely, I mark synalepha (see below) when it seems more natural to preserve both vowels as a diphthong, especially when this will fall on a ligature, and both vowels will therefore be pronounced and heard easily in singing.

The special case of elision of the word ũu to un is separated out and dealt with in the following section.

[1] Not to be confused with ‘plica’ in the musical sense, e.g. the short plica .ron or any other ‘plicated’ ligature.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
5:93 To, T, E enana O marieiro, poi-la na barca meteu, be come fól CSM 5:93Synalepha in poi-la ena here would leave the combined metrical syllable -la e- on a single long note .on, which may be a bit awkward, especially given the generally tricky rhythm of the music. Instead, I suggest the simpler solution of replacing ena with its much more common short form na. Note that this line does not survive in [To], as three folios are missing, starting from (my) line 12.
35:56 To, T, E tĩiatĩi' e o que tií' a arca | da Virge, Madre de Déus, CSM 35:56Mettmann has e o que tĩia a arca here, which seems to require synaeresis or synalepha, but this phrase is identical in [E] with the first hemistich of line 53 above, and therefore it is more appropriate to apply the same resolution—i.e. elision—to the extra syllable.
102:51 T, E contracontr' contr' a luz déu apelido,
119:19 T, E AgỹaAgỹ' e dizend': “Agi', agia, mui tóste fiade
137:38 T, E podessepodess' de pecado, que no podéss' seer e tal feira; CSM 137:38Synaeresis in seer would be the "obvious" resolution to the extra metrical syllable in this line, but elision of podésse is actually better in this case since, as discussed in the note to line 35, it avoids having a word that spans the boundary between rhythmic sub-units. By the way, if you are doubtful as to whether podéss' seer is actually pronounceable, observe that line 25 above has perdess' sa alma, which is found in that exact form in both Mettmann and [E].
196:6 T, E eraer' dela oír no quea, | tanto ' ér' avorreçuda. CSM 196:6The word éra is neatly laid under a single virga .on in the music in [E], and another single note follows, so I have chosen elision over synalepha here. Elmes also has ér' in his edition.
197:14 T, E fillasefillass' mandou-e que séus gãados | fiass' e guardasse be; CSM 197:14Mettmann has fillase here, whereas [E] has fillasse; this error is not counted on the Corrections page since I cannot currently mark one of my own edits as being both a correction and a scansion fix (i.e. elision).
238:41 F, E crerigocrerig' Quand' est' o crérig' oiü, | diss': “Ai, Groosa, val! CSM 238:41See note to line 16, and compare line 27. Here both manuscripts have Quand' est' o crerigo oyu, again with the right number of syllables but the wrong stress, so oiu must be expanded to three syllables by diaeresis with crérigo elided to compensate. (I have chosen elision over synalepha since the compressed syllable falls on a single note, and elision of this word is quite common elsewhere.)
249:23 F, E en'n ne sentiu sól se caera | ne recebeu ' neú mal;
251:10 E ded' ũa que e nacera | d' e orde-na meter.
251:21 E logolog' gu' ela s' esfurtava | e ía-se parar
253:44 F, E quequ' qu' eas péças caeü | ja feit', e porê sinou-se CSM 253:44As in line 15, I have fixed the first hemistich to ensure that it ends on an unstressed syllable. In this case, elision of que at the start is required to compensate for the diaeresis in caeu.
255:83 To, E acorr''corr' 'córr' a mesquia CSM 255:83[To] and [E] both have acorr' a mesquinna que en ti fia (or variant spellings thereof) with the extra vowel at the start and without the word se, which is Mettmann's addition. But this is no good metrically (or musically) as it displaces the rhyming word mesquinna. Instead I've marked elision, indicating that the final -a of Reínna absorbs the initial vowel of acorr', and accepted Mettmann's insertion of se (fiar-se en ... and fiar en ... mean the same, though the latter is more common.)
256:7 F, E quandoquand' Esto foi e aquel ano | quand' o mui bo Rei gãou,
262:33 E comocom' a Virge Santa Maa | e com' a todos oiu
263:5 F, E sempresempr' Ca ela sempr' a nós dá(a) | que façamos o meor CSM 263:5The text for the first hemistich of this line, as written by the lyric scribes in both [F] and [E], and copied by Mettmann, is Ca ela sempre a nos dá. This has the correct number of syllables, and Mettmann lets it pass without comment. Things are not quite so simple for those of us who care about the music, unfortunately: the hemistich is supposed to end with an unstressed syllable, to match every other line, but the verb here naturally attracts the stress. The music scribe in [E] obviously latched onto this, and wrote just a single note for -pre a, in order to force synalepha (or elision) and thereby bring forward in the melody to a point where the rhythm matches the stress, leaving it with a long note at the end of the phrase. Of course, the seven notes/ligatures that remain then completely fail to fit the other stanzas. To fix this, I have just added an extra punctum at the end of the phrase for the unstressed syllable, at E2 on the Music tab, matching the music for d' erro in the next line; and for the first stanza I have kept the desired elision to sempr' a and drawn out at the end with a small melisma, so as to keep the stress and rhythm in agreement. With reference to Elmes' edition, this means that the fourth bar of line 5, which contains just a dotted minim on D, should be changed to match the fourth bar of line 7, which has D minim + D crotchet.
264:8 E defendedefend' e defend' a si misma | quand' algú tórto prende;
265:59 F, E dissediss' E diss' a u séu hóme: “Vai-te, seér,
270:34 F, E yryayry' que e aquel no creess' | irí' a perdiçô.
277:2 E en'n aquele que as sas féstas | no guarda e ' pouco te. CSM 277:2Mettmann, following [E], has non guarda e en pouco ten in the second hemistich here. Elmes deals with the surplus syllable (which does not have an extra note in the music) by eliding guarda, giving non guard' e en pouco ten, but I would suggest elision of e as a more natural-sounding resolution that better preserves the important content words (since guard' e sounds too much like the subjunctive guarde).
281:12 E agora'gora “Mesquio desamparado, | que se'gora de mi? CSM 281:12As an alternative to the elision of agora here, it would also be perfectly acceptable to substitute the synonymous short form óra.
284:9 F, E esperitalsperital e se no foss' a Virge, | Reía sperital, CSM 284:9Synaeresis in the word reía does not occur elsewhere, despite the high frequency of the word, and doesn't fit well with the music, so I don't recommend it as a resolution to the extra metrical syllable in the original Reía esperital here. However, the phrase Reía sperital occurs in CSM 124:17 and 287:29, and therefore elision seems to be a good choice.
306:33 F, E comocom' juso com' a moér virge, | e lógo e descreceu
315:7 E hũahũ' E ũ' aldea préto | de Madride morava CSM 315:7Elision of ũa before a- is fairly common—e.g. ũ' aldea in CSM 215:16 and 351:5, ũ' arca in CSM 148:7 and 362:10, and elsewhere—so I have used it here in preference to marking synalepha on a single punctum.
319:10 F, E serviçoserviç' o que faz serviç' a esta pïadosa.
319:65 F, E guaridaguarid' a e guarid' e da virge contaro
334:27 E mancebomanceb' O manceb' oiu aquesto | e foi lógo sospeitar CSM 334:27See e.g. CSM 228:13 and 228:19 for examples of elision of the word mancebo in the manuscripts. As usual, I have chosen elision over marking synalepha here as the vowels are identical and the compressed metrical syllable falls on a single short note in the music.
341:44 E correndocorrend' sobiu encima da pena, | corrend' esforçadamente,
341:52 E avendoavend' havend' e Santa Maa, | que é verdade, creença,
342:8 E devedev' no dev' aquesto nu' hóme | por maravia tẽer,
343:25 E QuandoQuand' Quand' o crérig' oiu esto, | puou de seer del quito,
343:25 E crerigocrerig' Quand' o crérig' oiu esto, | puou de seer del quito,
345:13 E comocom' pero a vila dos mouros | com' i estava leixara,
345:26 E querendoquerend' querend' ir aquela noite | a Guadeíra mãer. CSM 345:26Mettmann, unusually, points out the hiatus in Guadeíra in his footnotes.
345:79 E LogoLog' gu' el Rei e a Reía | por aquesto entendudo
350:38 E nossanoss' Madre e ss' avogada, CSM 350:38In this line it is very hard to choose between Madr' e nossa avogada and Madre e noss' avogada as the better resolution for the surplus metrical syllable. Each works well with the music, and both elisions are well attested elsewhere.
352:22 E cavaleirocavaleir' e o cavaleir' haa | desto pesar e tristura. CSM 352:22The word cavaleiro is commonly elided elsewhere, e.g. CSM 16:19, CSM 59:15; see the Concordance for more.
355:20 E esteveestev' e estev' e atendeu-a; | e pois chegou, preguntou-e
355:52 E digodig' nunca me casarei vusco, | no vo-lo digu' e trebeo,
355:58 E ouveouv' meteu mui fórt' apelido | e houv' o rostro rascado.
355:102 E feitofeit' daqueste feit' a verdade, | que no houv' ê re negado.
363:8 F, E dissediss' e diss' a u séu hóme, que no foi preguiçoso: CSM 363:8In this case, the very common elision of the word disse is musically more elegant than synalepha, as the elided syllable falls on a single punctum .o.
364:19 E serviçoserviç' ca a Virge gloosa, | e cujo serviç' estava CSM 364:19See CSM 271:39, CSM 305:13 and CSM 308:37 for attested cases of elision in the word serviço.
365:26 E EstadeEstad' que contar no sabea. | E diss' ela: “Estad', estade, CSM 365:26Compare CSM 45:34 and CSM 175:22, both with the exclamation “Estad' estade!”.
366:45 E estoest' e pois est' houro dito, | chamaro quanto poa
366:54 E campocamp' foi, ca lóg' a ele vẽo | e u camp' u aradores CSM 366:54The word campo is not common in the Cantigas: there is only one other instance, in CSM 309:48. Elision seems nonetheless to be the best resolution for the metrical surplus here as the elided syllable falls on a single punctum .o, and the word is still easily identifiable without its final vowel.
367:20 E enfermidadeenfermidad' de grand' enfermidad' a maravia,
367:45 E estaest' e co est' enfermidade sobeja
367:47 E quandoquand' E quand' os madudios começaro
369:2 E pexepex' u pex' e que achou ouro | que por si e el peitasse, CSM 369:2The word pexe gets a single ligature .owo in the music in [E], implying a single syllable, so elision (or synalepha) is assumed to be intended.
369:54 E llell' e deceu-se pera Tejo, | por dar-' a bever no río | e o topete lavá-lo.
369:55 E en'n E ' lavando-o de rejo, | quis Déus que ' escorregasse
369:66 E avedeaved' dizendo: “Madre, comede | e haved' algú conórto; | ca sea maravia,
369:77 E aqueleaquel Des que ' houv' assí comprado | aquel pexe, a menia | foi-s' a sa madre correndo CSM 369:77In the second hemistich here, elision of aquele fits the natural rhythm of the words much better than elision of pexe.
374:25 E logolog' e g' oferer-o foro | e déro queno posse
376:23 E llell' Do Maél e dá-' este | anél que ' hei prometudo.”
376:46 E ssess' tornou-s' al Rei. E pois ' houve | esta cousa retraúda,
377:26 E meadademeadad' ca séu amig' o outr' éra | que a meadad' haa.
381:1 E ComoCom' Com' a vóz de Jesú-Cristo | faz aos mórtos viver, CSM 381:1The music in [E] provides just a single long note for the first word Como, hence the elision here.
381:17 E espiritalspirital lerono ao Pórto | da Raía spirital; CSM 381:17Compare CSM 197:4 and CSM 284:9.
382:8 E miragremiragr' gra miragr' a Santa Virge | do Pórto, cabe Xerez,
384:7 E clerigoclerig' Este mui bo clérigu' éra | e mui de grado liía CSM 384:7Compare CSM 65:7 and the many instances of the alternative form crérig' (see the Concordance).
384:40 E fradefrad' E o frad' espertou lógo | e foi ao leit' agia;
385:8 E porendeporend' E porend' e aquel lógo | mostrou miragres ceros
389:12 E doentedoent' por u menio que éra | mui mal doent' e verdade,
392:30 E ssess' e el foi-s' e fez u furto, | e que passou a postura.
398:7 E [O]nde[O]nd' Ond' e atal semeança | demostrou Santa Maa
399:5 E porendeporend' e porend' o poder dela | e u del recude CSM 399:5In the music in [E] there are just two separate notes for the word porende, so this elision (or possibly synalepha) seems to be intended.
420:23 E bẽeytabẽeyt' e bẽeit' a ta carne | e que jouv' enserrada
420:29 E virgĩidadevirgĩidad e bẽeita a túa | viridad sagrada
420:60 E fustefust' e tu co pïedade | sobr' el fust' acostada.
423:23 To estrelasstrelas E u as strelas fez, de cérto sei CSM 423:23This is a perhaps a rather odd case of elision, but it is the only resolution for the extra syllable in this line that leaves the words E u... at the start intact, which I believe is important as the same formula begins all stanzas after the first two. Although the form strela normally only occurs after a vowel (see CSM 1:37, 94:113, 100:2, 325:1), the -s of as and the s- of strelas will merge quite naturally here, resulting in [as·tre·las] which can be sung quite comfortably.
423:35 To escrito'scrito E u bestias, per com' o 'scrito diz, CSM 423:35Compare CSM 139:43 where escrito is elided in the manuscripts.

Elision (forms of ũu) (59)

A special case of elision is that of the indefinite article ũu, which I write as un when a single syllable is required, rather than ũ'. (The form ũ' does actually appear occasionally, but only for elided ũa.)

To reassure speakers of modern Castilian (Spanish) who might feel uneasy when seeing something like cada un: although this would have to be cada uno in Spanish, the medieval Galician-Portuguese forms ũu and un do not have distinct, fixed grammatical roles. For example, CSM 148:9 has cada un sa tea leva in the original manuscripts. In modern Galician there is only one form, un.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
209:34 F, E ũuun cada uendo e térra sa faz.
214:34 F, E ũuun fez que u daqueles dados | se fendeu per meiadade.
215:39 F, E ũuun E ta tóste a deitaro | e uego redondo; CSM 215:39An alternative resolution for the extra syllable in the second hemistich of this line might be to keep the original ũu and employ synaeresis in péego. However, contraction of ũu is very common, whereas this is the only occurrence of péego in all of the Cantigas, so I think it is better to let the latter word keep its three natural syllables.
223:11 F, E ũuun E porend' ume bõo, Do Matéus,
245:52 F, E ũuun per meógo do paaço, | e cada u a catou;
245:102 F, E ũuun cada u e neú deles | no foi depois malfeitor.
273:16 F, E ũuun que entrou ume bõo, | e viu estar desaposto CSM 273:16The words desaposto in this line and dẽosto in the next would be expected, etymologically, to have the open stressed vowel [ɔ], and that is indeed how I have generally treated these words, as well as all forms of posto and aposto. However, agosto, from Latin AUGŬSTUS, should have close [o], so for the rhyme to work in this stanza it is necessary to assume that the process of metaphony, which later closed many cases of stressed [ɔ] to [o] under the influence of [o] in the following syllable, was already having an effect in the language of the CSM. Compare stanza XV of CSM 411 in which dẽosto and posto must also rhyme with agosto, and contrast stanza III of CSM 59 in which dẽosto and aposto rhyme naturally without the need to assume metaphony.
274:46 F, E ũuun e cada u parou mentes | ena Virge se verría.
292:42 F, E ũuun obrados mui ricamente | cada u a séu sinal.
292:52 F, E ũuun u anél d' ouro co pédra | mui fremosa, com' achei
299:25 F, E ũuun est', e cada u deles respona:
309:7 F, E ũuun E aquel tempo e Roma | u Papa santo haa,
309:8 F, E ũuun e u Emperador bõo | per quant' ele mais poa CSM 309:8[F] has the wonderful 'scribo' emperadõo in this line, obviously anticipating the following word.
311:12 F, E ũuun no ano, e amizade | haa co u borgês;
311:21 F, E ũuun assí que feriu u deles | aquel hóme, que morrer
312:15 F, E ũuun E rogou a u maestre | que mui be a entaasse;
312:24 F, E ũuun e a u bo mõesteiro | que haa i levou-a,
313:29 F, E ũuun mais u crérigo que éra | i, pois viu a coita tal
313:61 F, E ũuun e cada u dos da nave, | assí como prometeu
313:63 F, E ũuun E da oférta fero | ulez mui grand' assaz
314:31 F, E ũuun e demais que foi fazê-la | e u ávol muradal
321:21 F, E ũuun mas entô ume bõo | consseou-e que dissésse
324:32 F, E ũuun u mudo que dentr' estava | per sinas empreguntou
325:39 F, E ũuun Pois a cativa espérta | foi, achou-s' e u camio
328:19 E ũuun Guadalquivir é u deles, | que éste mui nóbre río
328:22 E ũuun e e cada u daquestes | há muito bõo pescado.
329:26 E ũuun e cada u do que teve | foi sobr' o altar põer.
329:45 E hũuun gu' e atal com' ũa pédra. | Diz u deles: “Que farei? CSM 329:45Synalepha in the first hemistich is the only option here, and although the combined metrical syllable -guea- falls on a single long virga .on, the result sounds natural enough.
329:47 E ũuun E u deles foi a ele | e falou-' e o tangeu,
332:13 E ũuun menio, ontre séus braços, | e sé e u chapitél
332:33 E ũuun e poder os elementos; | ca ne u no vai ne ve,
333:9 E ũuun e u mesquio que éra | de todos nembros contreito
334:5 E ũuun que houv' ũa vez guarido | u mancebo lavrador
334:8 E ũuun morou co uo hóme | que el mui de coraçô
334:31 E ũuun E assí andou u ano | toeit' era de se,
335:48 F, E ũuun Eles se partiro lógo, | e cada u a buscava
335:49 F, E neũuneun quanto mais buscar poa, | mais neú nona achava.
337:16 F, E ũuun E este barô haa | u menio fio séu
338:15 E ũuun Onde ' avẽo ua | a aquel manceb' andando
339:21 F, E ũuun mais ume bõo es disse: “No
341:16 E ũuun e mui coitada chorando | co ele falou ua
341:47 E ũuun Pois sa razô acabada | houv', e u brïal be feito
341:56 E niũuniun ne e ú outro nembro, | ant' houv' o córpo be são;
342:17 E ũuun E u serrava u deles, | viro dentro de coor
345:3 E ũuun E desta gra maravia | u chanto mui doorido
345:19 E ũuun ta fèramente, e guisa | que u ric-hóme honrrado
345:55 E ũuun E traa e séus braços | u ta fremoso minio
347:3 E ũuun Desto direi u miragre | que e Tua avẽo,
348:24 E ũuun ficou muito daquel soo, | e chamou u séu privado
348:32 E ũuun Mais depois be a u ano | fez hóste sobre Grãada,
351:28 F, E ũuun mas chegou ume bõo, | que es diss' esta razô:
352:9 E ũuun e que u açor mui bõo | ũa vegada e déra
353:19 E ũuun E co mui gra pesar desto, | u deles, que e ficou,
353:20 E ũuun a u abade mui santo | du mõesteir' enou
353:32 E neũuneun o ía muitas vegadas, | ca e al neú prazer
354:4 E ũuun Madre de Déus Jesú-Cristo, | a u rei que muito tiía CSM 354:4In this line, the compressed syllable tii- of tiía sits below a convenient long plica .ood, so the natural syllables are preserved in singing.
354:19 E ũuun Onde ' avẽo ua, | indo per ũa carreira,
355:9 E ũuun e ume de Manssea, | mancebo barva pungente;
369:23 E ũuun E enou, dous, dizendo | a cada u que puasse

Restore elided vowel (2)

This is the opposite of elision, but I've used it much less frequently, as an alternative to diaeresis when a line contains too few syllables.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
137:35 T, E perdess'perdesse E fez-e que no perdesse oos, pées ne mãos CSM 137:35In this line I've taken the rather unusual step of restoring an elided vowel in one place and then marking synaeresis in another. The reason is that in almost every other line of this cantiga there are rhythmic units of 8 and 6 syllables, corresponding closely to the musical phrasing, with no words split across units. In this line, however, oos spans the unit boundary if perdesse is elided. It is better, therefore, to use the full form of this word, and compensate with synaeresis in pées, which is easily justifiable since it occurs elsewhere (e.g. CSM 213:66) and the shorter form pés is quite common. (When I say "almost every other line", this unfortunately excludes line 4, which is a bit of a fly in the ointment: the rhythmic boundary there is in the middle of the word éra and cannot be fixed so easily. See also the note to line 38.)
138:55 T, E tod'todo Pois Sa Joane todo aquesto viu, CSM 138:55Expanding tod' to its full form todo to make up for the missing metrical syllable in this line is not a particularly satisfying resolution, since the phrase phrase todo aquesto does not appear anywhere else: it is invariably tod' aquesto. However, it is perhaps the least of several evils. Firstly, although diaeresis in viu occurs many times elsewhere, it is of no use here, as it would break the rhyme and scansion: an unstressed final [u] in viü cannot rhyme with the stressed final [iu̯] in faliu and saiu. Diaeresis in pois, on the other hand, is not found elsewhere, and doesn't work well musically in this case. Finally, expanding Sa to Santo might just about be acceptable, except that the latter form, when it serves as a title, is generally reserved for Santo Tomas (since "San Tomas" might be heard as "santo mas").

Alternative word form (17)

This category of modifications covers cases that cannot convincingly be described as elision or the removal of elision, but still involve changing words in the text directly for metrical reasons. Note that CSM 146:41 and 185:37 are slightly different: these changes were made to fix the rhyme, not the scansion.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
20:44 To, T, E PorenP'ren P're m' acomendo CSM 20:44The surplus syllable, which features in all three manuscripts, has no easy resolution here. [E] has pore macomendo—not porẽ as would be expected—but this is just scribal error and offers no clues. It is clear though that, despite the presence of the long plica .oron at the end of the musical phrase, we should not break it to fit -mendo, as this would be completely at odds with the rhyme and rhythm. The only reasonable option seems to be to squeeze the whole word porê onto the first note .on of the phrase of music, as indicated here. Fortunately, pr is a common syllable onset in other words, so the result sounds quite natural.
96:54 T, E oraagora agora jazer mórto e se coor.” CSM 96:54Mettmann is true to [E] and [T] in this line, but it has a missing syllable and no scope for diaeresis to create one. In the absence of any other hints I suggest replacing óra (= "now") with its longer synonym agora to make up the shortfall.
113:21 T, E Dizend' estoDit' esto Dit' esto, saíro fóra | da eigrej' e viro estar CSM 113:21This whole stanza is a bit of a metrical disaster in Mettmann's text. For this line, I'd previously suggested synaeresis in saíron to fix the metrical surplus, but on further consideration this seems awkward and implausible with the vowels a and stressed í. Instead, I now prefer just to replace the original first word Dizend' (saying) with the past participle Dit' (= dito said). This barely affects the meaning, especially given that the phrase is rather formulaic anyway, and the result is just a variation on Esto dito and Esto dit' which are found elsewhere.
113:22 T, E fezofezera o penedo que caera, | que Déus fera desar; CSM 113:22To fix the metrical shortfall in the second hemistich of this line, I had previously marked diaeresis in Deus to give it two syllables. However, whilst the close diphthong [eu̯] is frequently subject to diaeresis elsewhere (most often in preterite verb forms) this is never the case for the open diphthong [ɛu̯] which occurs in Déus, and I now believe that diaeresis is best avoided. Instead, I have taken a suggestion from the footnotes to the draft version of this cantiga in the Oxford CSM Database (as of mid-October 2012) and replaced the preterite fezo with the pluperfect fezéra, something that the original poets appear to have done quite readily when an extra syllable was required, regardless of the (minor) effect on meaning.
127:4 T, E PoePoi a Virge Santa Maa | na eigreja do Poi fez CSM 127:4Mettmann has Poe here, following [E] and [T]. I have substituted the form Poi which is found elsewhere (e.g. CSM 172:22, CSM 262:5), since oe does not normally represent a diphthong, and yet marking synaeresis would clearly be inappropriate in what is basically a foreign name (the town Puy-en-Velay in France, where the lentils come from). Beware though that Elmes has divided the syllables incorrectly in his underlaid text, conflating na ei- as if by synalepha into a single syllable at the start of line 8 of the music, and giving two syllables to Poe (split lentils?) The singer will therefore need to realign the words and notes so that eigreja starts on the second note of the line (E), do is on the first note of the third bar (C) and Poe (= Poi) follows on the long falling ligature.
146:41 T, E sandeus, sandeu de téus ẽemigos, sandeu, CSM 146:41Both manuscripts have the plural sandeus here, and Mettmann reproduces it, obviously taking it as an attributive adjective modifying ẽemigos (he writes de teus ẽemigos sandeus with no comma). José Filgueira Valverde's translation into Spanish is also based on the same interpretation, as he writes "de tus sandios enemigos". However, the resulting non-rhyme defendeu / contendeu / sandeus would be quite unique in the Cantigas de Santa Maria, which is sufficient in itself to rule out this reading. Instead, I think the originally intended word must have been the singular noun sandeu used as a vocative by the desperate mother to her son, i.e. "There are bad people there, (of) your enemies, you fool, and I know well that nothing will save you from death." This must then have been misinterpreted a dozy scribe somewhere along the copying chain, who wrongly "corrected" it to the form that has survived.
171:45 T, E NoEno Eno camio se meter CSM 171:45Here it is possible to resolve the missing metrical syllable by expanding no to eno, on the model of CSM 138:20 fórono meter // eno camio que devia tẽer.
185:37 T, E -nos, nos o porend' ao caslo | nos queremos tornar, nós.” CSM 185:37Mettmann (I and II) edits the end of this line as nos queremos tornar-nos, with a hyphen, presumably interpreting the first nos as subject (= nós) and the second as a reflexive pronoun. However, the rhyme requires that the final word be nós with the open vowel [ɔ], so it seems that this must be the subject word (as a sort of emphatic afterthought) and the first nos the reflexive. Grammatically weird, perhaps, but then grammatical weirdness for the sake of a rhyme is more the rule than the exception in the CSM.
204:32 F, E virgẽesvirgens se dalí conas sas vírgens | e aos céos sobiu, CSM 204:32The original word vírgẽes in the first hemistich here is naturally stressed on the first syllable, but a word with antepenultimate stress is clearly not permissible metrically. Rather than require the stress to move unnaturally to the second syllable, therefore, I have substituted virgens [βiɾʤeŋs] (functionally equivalent to synaeresis) and inserted the definite article as before the possessive to compensate.
224:34 F, E lai a de levar i a meni·a, | firo atal postura. CSM 224:34Mettmann I and II de levar la menynna. Given that the first hemistich of this line ends with an unstressed syllable in both manuscripts, it still needs another syllable to fix the metre. I have inserted the word i on the model of CSM 176:26 e levou i a omagen, and changed la to a accordingly. The original word la appears in both manuscripts but wouldn't normally be expected in this position (levar a or levá-la would be more regular), so it could be a Castilian-influenced copying error for i a or ia anyway.

Mettmann I fizeron; Mettmann II e fizeron; [F] fezeron; [E] fizeron. The e inserted by Mettmann in his second edition makes some sense, stylistically, but I don't believe it to be essential. If you prefer to keep it, then reading the second hemistich here as e fizeron tal postura would be an acceptable compromise.
228:22 F, E àaa e foi ê maraviado | e diss' aa gent': “Uviade CSM 228:22The second hemistich in Mettmann's version of this line (which matches both [F] and [E]) is short of a syllable, and there are no less than three possible resolutions, all involving the removal of contractions, i.e. diss' → disse, à → aa or gent' → gente. I have chosen the second of these, not just because it seems to fit the rhythm of the words, but also because the reality of the contraction of aa (or a a) to a single syllable à in the natural speech of the time (as opposed to a metrical artefact) is the subject of some debate (Stephen Parkinson calls it "highly unlikely" in [SP2006]).
253:15 F, E ome bõobon ome O bo hóme entendeü | que andava e pecado, CSM 253:15The first hemistich of this line must end on an unstressed syllable, so I have marked diaeresis on entendeu. However, this means that a syllable has to be lost earlier in Mettmann's line (O ome bõo entendeu), and the usual options are of little use: synalepha at the start in O ome is horribly clumsy, synaeresis in bõo is technically blocked by the nasal [ŋ], and elision to bõ' before entendeu has no precedent. (The only occurrences of bõ' are where the feminine form bõa is reduced before a word beginning with a-, which is much more natural.) Since home bon is not an option either as bon must precede the noun, I have taken the liberty of swapping the words around. Fortunately, bon home is a common phrase elsewhere, occurring around a dozen times.
346:22 E muimuito por mostrar a sa vertude | que muito tóste lara. CSM 346:22Diaeresis in mui is a possible solution here (cf. CSM 139:23), but it isn't as convincing musically as the case in the previous line. Instead I have substituted the longer form muito, as suggested by both [JMMS] and the draft version (as of 5th November 2012) at the Oxford CSM Database; the latter suggests that mui toste is a 'reduction' of muito toste, presumably meaning an accidental omission of the repeated syllable to.
371:42 E offereçeroffreçer e ela foi-s' ao Pórto | aa Virge ofrecer. CSM 371:42Following [E], Mettmann has offereçer here, which has four natural syllables. However, the syncopated form ofrecer occurs in several other cantigas, so I have substituted it in this line as a more natural-sounding alternative to synaeresis in aa.
391:8 E adeantedeante haa: o qu' é deante, | atrás, com' éu aprendí, CSM 391:8The semantic distinction between adeante and deante in the Cantigas is vague at best, so I have substituted the latter word here in order to fix the surplus metrical syllable. Compare CSM 108:61 Que o rostro e tornar fez Deus o deant' atras.
399:31 E porendeporen os maos feitos, dar-t-ía | porê maa mórte, CSM 399:31Mettmann, following [E], has porende in this line. I have substituted the shorter synonymous form porê to fit the metre.
407:41 To creicre' i E cre' i óra esto que ti digu' éu: CSM 407:41Mettmann (I and II) has E crei ora at the start of this line, which is lacking a syllable. However, [To] has a prominent plica over the i to indicate that it is a separate word, so I have reproduced the form cre' i as transcribed by [JMMS].

Letter names (10)

CSM 70 and CSM 410 talk about the letters in the name of Santa Maria. The vowels A and I are no problem, but M and R have in each case to be pronounced as their letter names eme and erre for proper scansion. I have placed all the letter names in single quotes anyway, to help distinguish ‘A’ and ‘I’ from the normal words a and i.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
70:3 To, T, E M‘EME’ ‘ÉME’ móstra MADR' e MAIOR CSM 70:3Mettmann just has M and R, as in [E]. The names of the letters are eme and erre, and they must be pronounced as such to fit the metre.
70:7 To, T, E A‘A’ ‘A’ demóstra AVOGADA,
70:11 To, T, E R‘ERRE’ ‘ÉRRE’ móstra RAM' e RAÍZ, CSM 70:11See note to line 3.
70:15 To, T, E I‘I’ ‘I’ nos móstra JESÚ-CRISTO,
70:19 To, T, E A‘A’ ‘A’ ar diz que AVEREMOS CSM 70:19This line rather throws a spanner in the works for the marking of etymological h on the word haveremos, so I have left it off just this once.
410:15 E M‘EME’ A primeira, que ‘ÉME’ é,
410:19 E A‘A’ ‘A’ móstra a saüdaciô
410:23 E R‘ERRE’ ‘ÉRRE’ móstra como reinou
410:27 E I‘I’ ‘I’, indo de be e meor
410:31 E A‘A’ ‘A’ ar mostrou carreira tal

Insertions (9)

Adding words to the Cantigas, when they appear in neither the original manuscripts nor Mettmann's edition, is not an activity to be undertaken lightly. I have only used it as a last resort when lines are too short, but diaeresis or the substitution of a longer form of an existing word are not options—or at least very unattractive options. In each case I have done my best to find a word that contributes as little as possible to the meaning, and could plausibly have been left out accidentally by the scribe.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
96:41 T, E insert lle levar que se confissô e achara. CSM 96:41Mettmann notes that [E] has lacharan here, and [T] llacharã. Presumably he intended to replace these with lle acharan to fix the metre—but in the end accidently omitted the lle.
113:26 T, E insert que Déus, sa Madre groosa, |eno quantos que i va CSM 113:26In the absence of any attractive options for fashioning an extra syllable from the existing words in Mettmann's version of this line, I have added the word que after quantos on the model of CSM 165:23 (ta mercee no fal // a quantos que a demanda) and CSM 386:30 (e dest' acordados // foro quantos que i era des Tui ata Ocaa).
138:58 T, E insert i sa alma foi i u é Sa Simeô. CSM 138:58Again, various resolutions suggested themselves for the missing metrical syllable in this line. Apart from possible diaeresis of foi to two syllables, alternatives include adding the definite article a at the start of the line, giving a sa alma foi, or expanding Sa to Santo. The resolution I've chosen though is to add the word i after foi, on the model of CSM 259:17 e foro i u es ela mandou, since the accidental scribal omission of i after the same letter in foi is quite plausible.
159:14 T, E insert o u o séu jantar tiía, | ond' éra cobiiçosos. CSM 159:14Diaeresis in séu might in theory be used to resolve the missing metrical syllable here, but it is not found elsewhere in this word (nor in related words like méu and téu), and is musically a bit clunky here. Instead, I suggest the insertion of the definite article o which is optional before the possessive, and could quite plausibly have been left out accidentally by the scribe.
171:25 T, E insert y i viro u río correr CSM 171:25Here I have added the word i (= "there") which seems to be the least intrusive word that could fill the metrical gap. I have placed it at the beginning of the line for two reasons. First, it seems the best fit for the natural spoken rhythm of the words. Second, the fact that it then follows an identical but stressed vowel in assí makes it perhaps more plausible that the scribe could have overlooked it.
171:32 T, E insert o como o contaro a mi, CSM 171:32In this line I have inserted the word o on the model of CSM 124:4 e o contaro a mi, CSM 183:25 que o contaro a mi, and similar formulaic expressions. Accidental scribal omission of this o after the same vowel in como seems quite likely.
171:43 T, E insert ora fomos, vaamos óra seer CSM 171:43óra seems to be the most neutral filler word here, and although adding two new syllables requires synaeresis in seer to compensate, the latter is very common and the result feels quite natural.
194:32 T, E insert ja catando-se u co outro | e que ja no se falava, CSM 194:32Mettmann ũu; [T] un; [E] ũu. In Mettmann's version of this line, the first hemistich has one syllable too many, whilst the second is missing one. Keeping ũu from [E] and carrying the syllable -tro of outro over into the second hemistich might work in theory (compare CSM 12:23 fe- zera and CSM 419:78 Jo- safas), but rhythmically and musically it sounds dreadful. I have therefore used the [T] form in the first hemistich and, in the absence of any scope for diaeresis in the second, inserted the filler word ja, which doesn't greatly affect the meaning.
204:32 F, E insert as se dalí conas sas vírgens | e aos céos sobiu, CSM 204:32The original word vírgẽes in the first hemistich here is naturally stressed on the first syllable, but a word with antepenultimate stress is clearly not permissible metrically. Rather than require the stress to move unnaturally to the second syllable, therefore, I have substituted virgens [βiɾʤeŋs] (functionally equivalent to synaeresis) and inserted the definite article as before the possessive to compensate.

Deletions (5)

This is another small but mixed bag of changes, in each of which I have deleted a short word, even though it appears in all of the manuscript versions. See the associated footnotes for reasons.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
65:202 To, T, E delete Amen. levó-o consigu', e el seja loado. CSM 65:202All three manuscripts have the word Amê tacked on the end of the last line, but it cannot be metrically a part of it. Curiously, [To] and [T] even follow this with the refrain A creer devemos..., so how the Amen is supposed to fit in is anybody's guess.
113:23 T, E delete a E começaro Déus beinzer CSM 113:23[E] has & começaron Deus a bẽeizer, and [T] has & começar Deus a bẽizer. Mettmann has just começaron Deus a bẽizer, apparently deleting the conjunction to mitigate the metrical surplus, even though it looks stylistically necessary. The following word começar in [T] is very strange and must be a scribal error: neither infinitive nor future subjunctive make any grammatical sense in this position. The [T] form bẽizer (or beĩzer, with the tilde floating between e and i in typical manuscript fashion) is also odd. My assumption is that it is intended as a two-syllable word, pronounced [bei̯nʣeɾ] with the same stressed sound as treinta, and functionally equivalent to synaeresis in the normal three-syllable bẽeizer. All in all, it's a tricky line to sort out. However, Stephen Parkinson notes in [SP2007] that the preposition a is not essential between the verb começar and its complement, citing CSM 84:53 Enton [...] começaron bẽeizer // a Virgen Santa Maria as evidence. I have therefore decided to reject Mettmann's deletion of E on stylistic grounds, and to sacrifice the a instead. This, in combination with the two-syllable bẽizer, which I have respelled beinzer to clarify the pronunciation, gives the desired syllable count, and rhymes, and makes sense.
315:15 E delete as E quand' entrou na messe | u outras espigava, CSM 315:15Synalepha in the original u as here would be awkward, leaving the compressed syllable on just a single semibreve .e. Instead I have opted to drop the unnecessary word as, following [JMMS]. A plausible alternative would be to drop the initial vowel of espigavan and pronounce the two s's as one: u as outras'spigavan.
384:5 E delete e que as hóras desta Virge | dia de mui bo grado, CSM 384:5See the main footnote to this cantiga.
392:26 E delete e que ambos lo cofondesse | se fera tal loucura CSM 392:26Mettmann, following [E], has the word e at the beginning of this line, but it fits neither the sense nor the metre, so I have deleted it.

Moves (4)

These four individual insertions and deletions represent just two cases where I have moved a word between adjacent lines. In the case of CSM 276, this is purely in order to be able to mark synalepha, since the syllabation software cannot currently handle synalepha across line breaks.

Line Sources Modification vis-à-vis Mettmann II Text in this edition
255:67 To, E insert a mas a aogo CSM 255:67Both [To] and [E] have mas ao fog' a levou que ardia here, masking the rhyme on fogo entirely. In order to restore it, I have moved the object pronoun a from the next line, and marked synaeresis here to compensate. The phrase mas a a o fogo levou que ardia (= 'but [he] took her to the burning fire') is, I think, a good candidate for being the text as originally intended, with the first a (= 'her') then having to drop out to fit the metre, by means of conflation with the following a (= 'to'). At some later point in the copying chain this was then 'corrected', restoring the first a but in a different place, which made the grammar more explicit but played havoc with the metre. By the way, I have marked synaeresis in ao rather than synalepha between the two a's, simply because assigning ao the long plica .oron is better musically.
255:68 To, E delete a levou que ara.
276:61 F, E insert e sempre é mui loa·da e CSM 276:61This line has two extra natural syllables, and therefore requires synalepha in loada e on top of the ligature breaking found in other stanzas.
276:62 F, E delete e será.

Synalepha (105)

Synalepha is a metrical process in which a vowel at the end of one word meets another vowel at the beginning of the next, and the two fuse into a single syllable. This can happen whether or not the two vowels would ever naturally form a diphthong within a single word.

For speakers of Spanish (Castilian) it's worth stressing that in the medieval Galician-Portuguese of the Cantigas, synalepha is not obligatory in every place where two vowels meet across a word boundary, although traditional Spanish versification rules generally treat it as such. In fact, compared with the number of actual instances of synalepha, there are easily a hundred times more cases where it could occur, but is clearly unwanted as it would reduce the syllable count too far. So it should only be applied where it is explicitly marked in my texts by the thick red underline.

Having said that, almost half of all synalepha cases that I have marked involve the personal pronouns mi and ti. This can indeed be considered obligatory: these words always form a diphthong with a following vowel when in their unstressed ‘proclitic’ form, and in the manuscripts they are invariably joined onto the next word, without a space. I have therefore excluded these from the main table below, as they are really not that interesting from a metrical point of view, but there is a full list of references below the table if you want to have a look at a few cases.

Note that although synalepha reduces the number of metrical syllables, the resulting combined syllable very often falls on a multi-note ligature in the music (as indicated in the fourth column of the table), which means that the natural syllables of the two words can be preserved clearly in singing. I therefore tend to opt for marking synalepha when this is the case, and to use elision to omit one of the vowels entirely when the compressed syllable falls on a single note.

Line Sources Text in this edition Music
9:40 To, T, E no e vẽo a mente, | que el prometera; CSM 9:40The combined metrical syllable -o a from vẽo a with synalepha falls on a single long note .on, which is musically tolerable though not ideal. Elision is not an attractive alternative however, as vẽ' is not found elsewhere. .on
26:79 To, T, E a alma onde a trouro, CSM 26:79Synalepha in onde a works well, with the combined metrical syllable -de a falling on a long plica .oron in the music. .oron
62:38 To, T, E per meia a vila, de todos vïúda, CSM 62:38For this line, the music is different from the other lines with metrical problems. The suggested synalepha works well though, with the combined metrical syllable -ia a falling on a very comfortable five-pitch plicated ligature .eyeyehod. .eyeyeod
96:40 T, E e come dïabres a alma cuidara CSM 96:40Mettmann alma; [T] a alma; [E] alma (possibly with a small a inserted later into the preceding space, though this is unclear from Anglés' facsimile). In effect, [E] has the definite article a elided out of existence, as indeed Mettmann observes in a footnote. But [T] keeps it, and it is syntactically necessary, so I have used this form and marked the required synalepha. .on
105:65 To, T, E E todo-los daquela vila ara CSM 105:65The combined syllable -la ar- resulting from synalepha here falls on a three-note ligature .onyeye in the music, so the natural length is preserved. .onyeye
107:4 T, E fez e Sevi' a cidade .royo
115:26 To, T, E E térra de Roma houv' i, CSM 115:26The synalepha here works well musically, as the combined metrical syllable -ma ou- sits on a long plica .ood. .ood
129:9 T, E e be até eno toutiço entrara, CSM 129:9The word até is almost always followed by eno/ena/enas so I have kept eno here and marked synalepha, rather than shortening it to no. However, CSM 33:32 has be ate nos fondamentos, so be ate no toutiço would be acceptable here too. Note also that [T], which has the music written out for the second stanza here, has atẽeno laid under three notes/ligatures .on + .oyo + .o, confirming that this is the right place in the line for the metrical fix. .oyyjo
131:71 T, E E eles o fero, pois todos e az CSM 131:71Synalepha seems to be the only option for shortening this line, even though the combined metrical syllable falls on a single short note in the music, effectively squeezing the first word E out of existence. .o
135:136 T, E novio a novia e solaz; CSM 135:136Synalepha in novio a fits the musical phrasing much better than synalepha in novia e, and the combined metrical syllable -vio a falls conveniently on a two-note .oso ligature, so the natural syllables are preserved in singing. .oso
155:10 T, E quena ama e a sérve | e sabe e ela fïar. .o
185:38 T, E E tornaro-se correndo | e foro-se e el meter. CSM 185:38Synalepha here works well with the music, as the combined metrical syllable se en falls on a comfortable three-note ligature .eyeye. .eyeye
185:49 T, E e mandou ao alcaide | que se o caslo pedir .on
194:41 T, E e quantos aquesto oíro, | os grandes e os mẽores, CSM 194:41Synalepha here works well musically: the combined metrical syllable -to o- falls on a two-note .ra ligature, so the natural hiatus is preserved in singing. .ra
220:3 E Ca muito é gra dereito, | que d' ángeos é servida .ood
220:4 E e nos todo mal tóe | no mundo e nos dá vida, CSM 220:4See main note to this cantiga. .owo
222:13 F, E há i u rico mõesteiro | de donas, e castidade CSM 222:13Though I don't make a habit of philological observations, it's interesting to note that há i at the start of this line—which clearly has to be pronounced with synalepha as a single syllable—is a nice clear example of the origins of modern Galician "hai" and Castilian (Spanish) "hay" in the combination of verb and adverb. .o
224:26 F, E juntado de sũu assí | que no éra de costura. CSM 224:26Note that the first hemistich here does not have a feminine rhyme: assí is stressed as always on the final syllable. A more normal resolution is therefore appropriate for the surplus syllable, and synalepha in sũu assi fits the bill, with the combined metrical syllable -u a- falling on a descending plica .ron. .ron
225:22 F, E mas ontr' o coir' e a carn' ía | aquel bestigo astroso. CSM 225:22[E] does not have the definite article a in the first hemistich of this line, and therefore there is no metrical surplus. The a comes from [F], and although it appears to be necessary for correct grammar, its inclusion does result in an awkward case of synalepha on a single short note. .o
226:26 F, E ne enfermea u cuidava sãar, CSM 226:26Synalepha works satisfactorily in this line, with the combined metrical syllable -a u falling on a short plica .bod. Note that the (necessary) word u is missing from [E] but present in [F]. .bod
235:4 F, E que avẽo a Do Afonsso, | de Casl' e de Leô CSM 235:4[E] has the avẽo a laid clearly under three individual notes/ligatures .ra + .on + .o so this synalepha seems to be required. Elision to avẽ' is best avoided as it sounds unnatural and does not occur elsewhere (neither is the base word vẽo ever elided to vẽ'). .o
245:51 F, E E el aquesto dizendo, | ũa dona entô entrou CSM 245:51Synalepha here works well musically, with a long ligature .raron for the combined metrical syllable -na en-. .raron
246:8 F, E puav' a ũa igreja |a e oraçô oír, CSM 246:8Synalepha in sua e works fine with the music, with the combined metrical syllable sitting on a long plica .oron. .oron
256:28 F, E de todos aqueles maes | guariu a Reía ta be CSM 256:28Synalepha in guariu a here is awkward, with the combined metrical syllable falling on just a single long note in the music. However, although synaeresis in Reía might at first sight look like a possible alternative, this in fact never occurs elsewhere despite the high frequency of the word (125 occurrences); it always keeps its three natural syllables. .on
261:48 F, E loando a Déus e apósto cantar, CSM 261:48Synalepha in loando a works well with the music, as the combined metrical syllable -do a falls on a long plica .oron. .oron
262:41 E a Virge Santa Maa, | senô porque a muitos val; CSM 262:41Synalepha in porque a is awkward musically, as the combined metrical syllable falls on a single short note. Elision to porqu' a is best avoided though, since the elision of que to qu' is very rare in general despite the extremely high frequency of the word, and it seems never to occur before the vowel a. See also CSM 312:45. .o
276:61 F, E sempre é mui loa·da e CSM 276:61This line has two extra natural syllables, and therefore requires synalepha in loada e on top of the ligature breaking found in other stanzas. .oron
282:8 F, E que mostrou grand' e Sevi' a, | com' éu e verdad' achei, .e
300:8 E e ela fia e crïada, CSM 300:8The synalepha in this line is clear from the music in [E], where the combined metrical syllable -a e is laid under a single ligature .eyeye. .eyeye
304:11 F, E d' outr' óio senô d' olivas | mui límpi' e muit' esmerado; .o
312:45 F, E a el, e fero tanto | que a sa casa ' a trouro. CSM 312:45Synalepha in que a is a bit awkward, as the combined syllable falls on a single punctum .o. Note though that elision of que to qu' is rare (around 0.15% of all occurrences of the word) and only happens before the vowels "e" and "i", so it is not an attractive option here. .o
312:50 F, E éra muit' ende a donla | e el muito se mesura. CSM 312:50Synalepha in ende a works well musically, with -de a getting a descending plica .ron. .ron
313:51 F, E O crérigo, pois diss' esto, | os oos a céo alçou CSM 313:51Synalepha in ceo al works well musically, the combined syllable -o al sitting on a two-note ligature .royo. .royo
315:27 E E cuidando que éra | de poçoa aquel feito .ron
315:37 E “Virge Santa Maa, | com' éu creo e entendo, CSM 315:37Synalepha in e entendo also comes out well, with the combined metrical syllable e en- also falling on a plica .ron. .ron
320:26 F, E per bõo entendemento. CSM 320:26Synalepha: the combined metrical syllable -o en- here sits comfortably on a two-note ligature .royo, preserving the natural syllables in singing. .royo
325:26 F, E e que co mouros mui ricos | a ambas casá-las faa. CSM 325:26Synalepha between the two identical vowels in a am- is quite natural and obviously intended here, even though the combined metrical syllable falls on just a single punctum .o in the music. .o
326:21 F, E derredor da eigreja, e, com' oí, CSM 326:21Synalepha in eigreja e works well musically, as the combined metrical syllable -ja e falls on a three-note ligature .eyeye. .eyeye
329:45 E gu' e atal com' ũa pédra. | Diz u deles: “Que farei? CSM 329:45Synalepha in the first hemistich is the only option here, and although the combined metrical syllable -guea- falls on a single long virga .on, the result sounds natural enough. .on
343:20 E e lóg' omo e disse: | “Ai, crérigo escomungado, CSM 343:20Synalepha here works well musically, the metrical syllable -go es- falling on a long plica .oron. .oron
345:14 E e avẽo que por esto | a houra pois a perder. CSM 345:14The combined metrical syllable a hou- resulting from synalepha in a houvera falls here on single punctum .o in the music, but elision is obviously not an option. .o
355:97 E que todo o mund' avonda | e nósso be sempre avanta, CSM 355:97The combined metrical syllable -pre a- falls nicely on a two-note ligature .owo. .owo
359:9 E e éra mui bõo vizio | a quantos morava i .on
359:33 E que Domingo éra vẽudo | a sa casa, e se ir CSM 359:33The combined syllable resulting from synalepha here falls on a descending plica .ron in the music, so the vowels can keep their separate identity when sung. .ron
361:19 F, E de pedir-as; e ta tóste | a Reía espirital CSM 361:19The synalepha here works fine musically, with the combined syllable -a es- falling on an ascending plica .bod. Note that synaeresis in Reía (giving reia) is not an alternative as it never occurs elsewhere; the word invariably keeps its three natural syllables. See also CSM 284:9. .bod
363:7 F, E El Co Simô éra conde rico e poderoso CSM 363:7The combined metrical syllable -co e here falls on a two-note ligature .oso in the music, so I have preferred this synalepha over the elision of rico, even though the latter is well attested elsewhere. .oso
365:26 E que contar no sabea. | E diss' ela: “Estad', estade, CSM 365:26Compare CSM 45:34 and CSM 175:22, both with the exclamation “Estad' estade!”.  
366:50 E o que o trage e toer-a, | pua d' haver séus sabores CSM 366:50The combined metrical syllable -ge e resulting from synalepha here falls on a long plica .ood, so the natural syllables are preserved when sung. .ood
369:83 E como aquel anél achara. | E ela que o mostrasse CSM 369:83The combined metrical syllable -mo a- falls on a two-note ligature .owo. .owo
371:35 E Ant' ía sobre la agua | como se fosse u batél, CSM 371:35The combined metrical syllable in fosse u with synalepha falls here on a descending plica .ron in the music, so the natural syllables are preserved. .ron
375:27 E Este escriva del Rei éra, CSM 375:27The combined syllable -te es- resulting from synalepha here falls on a two-note ligature .owo, preserving the hiatus. .owo
376:39 E E o outr' o anél lógo | e déu, dizendo-e: “Amigo, CSM 376:39The combined metrical syllable -e a- resulting from synalepha here falls on a comfortable three-note ligature .eyeye in the music, so the natural syllables are preserved when sung. .eyeye
382:9 E du ric-hóme que pea | al Rei herdade ũa vez, CSM 382:9The combined metrical syllable -de ũ- resulting from synalepha here falls on a four-note ligature .boyoron in the music, so the natural syllables can be preserved in singing. .boyoron
383:31 E d' i entrar, e e querendo | sobir per ũa escaeira CSM 383:31The combined metrical syllable -a es- from the synalepha here falls on a two-note ligature .ra, so the natural syllables are preserved when sung. .raa
386:26 E ca que a seor demanda | se guisa, é cousa estraa. CSM 386:26The metrical syllable -sa es- resulting from synalepha falls on a two-note ligature .ra in the music, so the natural syllables can be preserved when sung. .ra
419:26 To, E séu Fio Jesú-Cristo, | que e ela encarnou. .ron
420:16 E e outrossí a óbra | que end' houviste obrada, CSM 420:16The combined metrical syllable -te o- from synalepha here falls on a four-note ligature .onyeyeye, so the natural syllables are easily preserved when sung. .onyeyeye
422:36 To, E E u todo-los reïs | fore ant' el homildosos, CSM 422:36The first hemistich of this line should end in an unstressed syllable to be consistent with the metre of the other stanzas, so I have marked diaeresis in reis (which is common elsewhere) and compensated with synalepha in E u at the start, which falls on a convenient two-note ligature .oswjo in the music. .oswwwjo
424:26 To, T até ena térra dos judéus, CSM 424:26See the note to CSM 129:9. [T] has antena terra dos judeus (= ante na...?) which does scan correctly, but this doesn't make as much sense as the [To] version, and may be a scribal error. .e
427:47 To por el e encima mórte padecer; CSM 427:47The combined metrical syllable e en falls on a two-note ligature .oso in the music in [To], so the natural hiatus can be preserved in singing. .osjo

Synaeresis (122)

Synaeresis is another process whereby two normally independent vowels, forming two syllables, are fused into a diphthong, or if the original vowels were identical, into a short vowel. The result is one syllable fewer. In contrast to synalepha, synaeresis operates inside a single word, not across word boundaries. I mark synaeresis in cases where a line is too long metrically, and since it crops up repeatedly in the same small set of root words, it is generally preferable to synalepha or elision (if these are also an option on the line).

Here's a full list of word forms in which synaeresis is used, and the frequency:

aa (8), ángeo (5), ángeos (4), ao (4), cobiíça (2), cobiiçar (2), cobiíças (1), cobiiçasse (1), cobiiçava (1), cobiiçosa (1), cree (1), creed' (1), áboo (1), escoorido (1), escïas (1), gaa' (1), gafee (2), guaaa (1), mees (5), hóstïa (1), juïgado (1), juïgava (1), manteer (1), meesmo (1), mercee (9), pavoa (1), pées (3), pegoo (1), pérssïa (1), seelos (1), seer (12), seerás (1), seere (1), siía (3), soo (2), tiía (13), tiía (3), triígo (1), triínta (1), veê (1), vee (2), veede (1), veedes (1), veer (3), veea (1), vií' (1), viía (4), viía (1), viía (3), viía (2), vïíana (1)

It is quite significant that every word in the above list has a reflex in the modern languages in which the affected syllables are permanently conflated. The equivalent words in official Galician spelling are as follows (with syllables marked in a few cases for clarity):

á, anxo, anxos, ao or ó, cobiza, cobizar, cobizas, cobizase, cobizaba, cobizosa, cre, cred[e], diabo, cor (root of escoorido), es·pe·cias, gañ[a], gafeño, gaña, xulgado, xulgaba, manter, homes, mesmo, mercé, palabra, perigo, pes, Per·sia, selos, ser, serás, seren, – (no reflex of siian), son, hos·tia, tiña, tiñan, trigo, trinta, ve, vede, vedes, ver, verían, ví[a], vía, vían, viña, viñan, viñana

In Portuguese, the words are:

à, anjo, anjos, ao, cobiça, cobiçar, cobiças, cobiçasse, cobiçava, cobiçosa, crê, cred[e], diabo, côr, espécies, ganh[a], gafém, ganha, julgado, julgava, manter, homens, mesmo, mercê, palavra, perigo, pés, Pérsia, selos, ser, serás, serem, –, sou, hóstia, tinha, tinham, trigo, trinta, , vede, vedes, ver, veriam, vi[a], via, viam, vinha, vinham, vinham-na

This fact suggests that the poets of the CSM were not merely applying an artificial, purely metrical contraction in these cases, but rather taking advantage of natural variation in speech at a time when vowels that were originally in hiatus had begun the process of simplification that led to the modern forms.

Line Sources Text in this edition Music
4:33 To, T, E que viía resprandecer CSM 4:33Synaeresis, which is common in forms of the verb veer (= "to see"), is required here to reduce viía to two metrical syllables (contrast line 24 where it has its natural three). The compressed syllable vii- benefits from a long note .on in the music, however. .on
6:42 To, T, E e a quantos ela viía, | a todos preguntando CSM 6:42Synaeresis is commonly found in forms of the verb veer, including viía, and in this line works much better musically than synalepha in e a. .on
15:44 To, T, E rra de Pérssïa, quéro vĩir CSM 15:44The word Pérssïa naturally has three syllables with antepenultimate stress—compare CSM 265:23 and CSM 265:51 and also perssïãos in line 20. Here synaeresis is required to compress the two final unstressed syllables into one. .royo
16:22 To, T, E de querer séu be e de o mais d' al cobiiçar. CSM 16:22Synaeresis in cobiiçar works well with the music here, with -bii- falling on a three-note ligature .roya. The same contraction is often found in other forms of this verb too, e.g. CSM 67:11, CSM 157:8. .roya
16:23 To, T, E Mas co coita grande que tiía no coraçô, CSM 16:23The compressed metrical syllable tii- resulting from synaeresis here falls on a two-note ligature .ra in the music, so the natural syllables are preserved in singing. .ra
18:55 To, T, E dando aa Madre loor CSM 18:55Synaeresis in aa works fine musically, the compressed syllable falling on the four-pitch ligature .bohod. In [To], the first a of aa appears to have been erased, although it is still just visible. .bood
23:14 To, T, E ca no tiía senô pouco e u tonelcio. CSM 23:14Mettmann pouco; [To] pouc with erased o; [T] pouco; [E] pouco. Although [To] has obviously been modified in its own way to correct the syllable count, the result is completely awful from a music point of view, so I've chosen to mark synaeresis here instead. .o
28:26 To, T, E que tiía ascondudo. CSM 28:26Synaeresis in tiía is common, and is the only sensible resolution for the extra syllable here. .o
38:52 To, T, E séus dedos tiía foi lógo deitar. CSM 38:52Synaeresis in tiía works very well against the music here, as the compressed metrical syllable gets a long three-note ligature .bowowo, and the natural syllables are therefore preserved in singing. .bowowo
38:56 To, T, E que tiía a Madre fez be so as tetas decer, CSM 38:56Synaeresis in tiía is again the best option here too, rhythmically and musically, with the compressed syllable falling on a long plica .oron. .oron
41:19 To, T, E porque a sa mercee é mui maior CSM 41:19Synaeresis is common in the word mercee, and is confirmed in this case by [T] which has the music written out for all stanzas, and assigns a single plica .oron to the compressed syllable -cee. In [E] the corresponding ligature (in the first stanza) has the unique shape .orod, which looks like a possible scribal error. .oron
47:14 To, T, E muit', e mui be sa orde | tiía, com' aprendí; CSM 47:14Synaeresis in tiía is very common, and is confirmed here by [T] where the music for this stanza is written out, and this word (spelled tĩja) gets just one note and one ligature, .o + .ra. .o
50:11 To, T, E Ca Déus e si meesmo ele mingua no há, CSM 50:11Mettmann mẽesmo; [To] meesmo; [T] meesmo (direct underlay to one ligature + one note); [E] mẽesmo. The form found in [To] and [T] is more amenable to the synaeresis required here. Note that mesmo doesn't occur in that exact form in the cantigas, although the feminine mesma crops up in CSM 194:15 and 360:9. .onwaron
58:39 To, T, E Esto dit', uáboo a puxou CSM 58:39All three manuscripts have this line as shown here. Esto dito at the start in Mettmann II is an error not present in Mettmann I, but the form diaboo is faithful, and requires synaeresis to fit the metre. The shorter form dïabo is much more common in the CSM, and in fact the four-syllable variant dïáboo only occurs once in the singular (here) and six times in the plural dïáboos, with five of the latter being in the non-metrical epigraphs. .ood
63:16 To, T, E tiía o logar e aquela sazô, CSM 63:16Synaeresis in tiía is very common, and although the compressed syllable falls on a single short note .o here, the musical result is certainly superior to the alternative of synalepha with the following word o. .o
67:11 To, T, E E por meor fazer esto | que muit' ele cobiiçava, CSM 67:11The compressed metrical syllable -bii- resulting from synaeresis here falls on a two-note .royo ligature in the music, so the natural length is preserved in singing. .royo
67:62 To, T, E Diss' o bispo: “Vea lógo, | ca de veer-l' hei soïdade.” CSM 67:62Synaeresis is common in forms of the verb veer. Note that soïdade has four natural syllables and they should be maintained here. .o
75:109 To, T, E Entô diss' aaa: “Ve-te | ja comigo, ai amiga, CSM 75:109Synaeresis in aa is common, and in this instance the compressed syllable falls on a two-note ligature .ra in the music, so the natural length is preserved in singing. .ra
79:7 To, T, E e de o oír seer-vos-á saboroso, CSM 79:7This line is slightly problematic, because all 13 natural syllables get a separate note or ligature in the music, in all three manuscripts. However, from the musical structure, it is obvious that seer should occupy the slot that becomes one syllable in subsequent stanzas, and since the contraction of seer through synaeresis happens many times elsewhere (see the Concordance), this is the suggested resolution here. From the point of view of the music itself, this means joining the two separate puncta given to seer into a single .oyo ligature (AG in bar 2 of the fifth line of music in Elmes). .oyo
84:27 To, T, E Ond' avẽo pois u día | que siía a séu jantar; CSM 84:27Synaeresis in siía works very well musically, as the compressed metrical syllable sii- falls on a four-note .bowowa ligature, so the natural length (siia) is easily preserved in singing. .bowowa
88:52 To, T, E por aa eigreja passar CSM 88:52Here aa with synaeresis falls on a long ascending plica .ood in the music, so the natural length is preserved in singing. .ood
92:10 To, T, E que no vií' acá ne alá. CSM 92:10The compressed syllable vii- from synaeresis here falls on a descending plica .ron in the music, so the natural syllables come out in singing. .ron
93:14 T, E quiso Déus que caess' e el mui gra gafee, CSM 93:14The word gafee does not occur elsewhere, but it can be assumed from the spelling that it would have three natural syllables (at least some of the time). Here, and in line 31, gafee must have just two syllables however, so I mark synaeresis as usual. .on
93:31 T, E e ta tóst' a gafeego del se partiu, CSM 93:31See note to line 14. .on
95:24 T, E pera as gentes que viía i e romaa, CSM 95:24The compressed syllable resulting from synaeresis in viía falls on just a single short note .o in the music, but synaeresis does occur several times elsewhere in this word or closely-related forms (as well as tiía etc.) so it seems likely to be intended here. The possible alternative of synalepha in pera as at the start of the line isn't too terrible, but it does leave viía feeling rather overstretched instead of compressed. .o
97:30 To, T, E séus hómees por el lógu' enou; CSM 97:30Mettmann omẽes; [To] omees; [T] omẽes; [E] omẽes. Here I have used the form omees from [To] in order to be able to mark synaeresis to fix the metre. .bowowo
105:1 To, T, E Gra pïadad' e mercee e nobreza, CSM 105:1Mettmann has piadad' in the first line and piedad' in all of the reprises, following [E].

Here mercee must have just two metrical syllables, but the music provides a three-note ligature .brawo for -cee, so the compression of synaeresis is not apparent in singing. Note that Elmes has "merce" in his edition, having missed out an e.
.brawo
117:1 T, E Toda cousa que aa Virge seja prometuda, CSM 117:1[E] has a a laid quite clearly under a single long virga .on in the music, so synaeresis of Mettmann's aa is called for here. .on
119:22 T, E mandou fiar a séus hómees lanças e escudos, CSM 119:22Mettmann (I, II) omẽes; [E] omees; [T] omees. Although forms of this word with the tilde do, of course, occur in the manuscripts, Mettmann does not normally add it, and to do so in this case seems particularly odd when the word must clearly be compressed to just two metrical syllables, in order to maintain the consistent 8 + 6 rhythm of the lines without a break in the middle of the word. .o
125:86 To, T, E farei mui de bõa mente; | mais este, de que soo moér, .bawo
135:64 T, E que cuidara seer casado CSM 135:64seer is reduced by synaeresis to one metrical syllable here, but the word gets a comfortable three-note plicated ligature .rowon in the music. .rowon
136:14 T, E dados hómees e moéres, com' é sa usança. .o
137:35 T, E E fez-e que no perdesse oos, pées ne mãos CSM 137:35In this line I've taken the rather unusual step of restoring an elided vowel in one place and then marking synaeresis in another. The reason is that in almost every other line of this cantiga there are rhythmic units of 8 and 6 syllables, corresponding closely to the musical phrasing, with no words split across units. In this line, however, oos spans the unit boundary if perdesse is elided. It is better, therefore, to use the full form of this word, and compensate with synaeresis in pées, which is easily justifiable since it occurs elsewhere (e.g. CSM 213:66) and the shorter form pés is quite common. (When I say "almost every other line", this unfortunately excludes line 4, which is a bit of a fly in the ointment: the rhythmic boundary there is in the middle of the word éra and cannot be fixed so easily. See also the note to line 38.) .o
139:48 T, E houres, sempre pois migo seerás CSM 139:48The verb seerás is compressed by synaeresis to two metrical syllables here (as the infinitive seer commonly is elsewhere). .o
140:9 T, E sa mercee e sa franqueza CSM 140:9[T] has the music for the second stanza written out, and the word mercee is clearly laid under just two notes .o + .on, confirming the synaeresis that I have marked here. .on
141:42 T, E pois dos ángeos e dos santos sempr' é servida. CSM 141:42Synaeresis in ángeos (normally with three natural syllables) is required to fit the metre, but works fine musically as -geos falls on a three-note ligature .eyeye. .eyeye
153:44 T, E e por veê-la ía CSM 153:44Somebody involved in the preparation of the text of Mettmann II seems to have got a bit muddled at this point about what was supposed to rhyme with what, though the mistake is so daft that it seems rather doubtful that it was Mettmann himself. As his summary of the metrical and rhyming scheme correctly indicates, this line is in fact meant to rhyme with the penultimate line of every other stanza, the rhyme being [i.a] (romaría, perfía etc.) The text in [E] and [T], which is correct in Mettmann I and which I have restored here, does exactly that. In Mettmann II, however, the words have been juggled incorrectly such that the line reads e ya por veela (where ya = ía), presumably intending that this should rhyme with séla and donzéla in this same stanza, but unfortunately overlooking the non-rhyme between open [ɛ] in the latter words and close [e] in veê-la.

Whatever the word order, synaeresis is required here in veê-la, but this is unremarkable as it occurs rather often in forms of the verb veer.
.on
155:3 T, E Du o pecador promete | de seer amigo de Déus CSM 155:3Synaeresis is very clearly required in seer here, as the whole word is laid under a single .oso ligature in the music in [E] and [T]. However, the two notes of the ligature allow the word to keep its two natural syllables when sung. .oso
155:45 T, E que ao ermitá foi lógo | e diss': “A Seor de prez, CSM 155:45The word ao is compressed by synaeresis to one metrical syllable here, but it falls on a two-note ligature .oso in the music. .oso
157:8 T, E de faria que traa | tal cobiíça e creceu .o
158:6 T, E o metero e e férros, | como gente cobiiçosa, CSM 158:6The intended synaeresis in cobiiçosa here is quite clear from the text laid under the music in [E], where the compressed metrical syllable -bii- gets just a single long note .on. .on
159:9 T, E e vio; e entre tanto | foro aa Virge rogar CSM 159:9Synaeresis in aa is required to fit the metre here, and is quite common elsewhere. .on
167:21 T, E e pois que chegou a Salas, | diss' aa Virge: “Se no mente .on
170:15 T, E Loada deve seer mais d' outra re CSM 170:15Here seer is reduced to a single metrical syllable by synaeresis, as it commonly is elsewhere. [T], which has the music written out for all stanzas, confirms this by assigning the word just a single ligature .oyo. [E] has the musically equivalent .oron at the same position in the melody. .oron
171:43 T, E fomos, vaamos óra seer CSM 171:43óra seems to be the most neutral filler word here, and although adding two new syllables requires synaeresis in seer to compensate, the latter is very common and the result feels quite natural. .eyeye
177:21 T, E Diss' ele: “Ca éu be cuido | aínda veer destes méus, CSM 177:21Synaeresis is very common in forms of the verb veer. .on
180:50 T, E e santidade, u mercee achar CSM 180:50Synaeresis is very common in the word mercee. In this particular case the compressed metrical syllable -cee falls on a two-note ligature .oyo in the music, so the natural length is preserved in singing. .oyo
194:9 T, E aquel cuteif' avarento | tal cobiíça ' ê crea, CSM 194:9[T] has the music written out for this stanza, and cobiiça is laid under just three notes .on + .o + .o, indicating that synaeresis is necessary. .o
195:4 T, E seer--áadosa. CSM 195:4The word seer is laid under a single short punctum .o in the music in [E], so the synaeresis is clearly intended. .o
195:26 T, E que a cobiiçasse CSM 195:26The compressed metrical syllable -bii- resulting from synaeresis in cobiiçasse also falls on a short punctum in the music. However, synaeresis is required elsewhere in half-a-dozen other words with the stem cobiiç-, so it's obviously intended here. .o
198:6 T, E o terrá, e que metudo | dev' a seer ontr' os preçados. CSM 198:6Synaeresis in seer is clearly indicated in [E], where the whole word is laid under a single long plica .oron in the music, and therefore has just one metrical syllable. .oron
211:6 To, F, E cobiiçar d' haver a sa compaa. CSM 211:6The intended synaeresis in cobiiçar is clear from the musical underlay, where the word is assigned just three notes implying three syllables: .o + .ron + .o in [E] and .o + .o + .o in [To]. .ron
213:66 F, E ca ficou todo britado | dos pées tro ena moleira. CSM 213:66Synaeresis in pées also works well against the music, as the word benefits from a long plica .ood. The short form pés is quite common elsewhere. .ood
218:19 F, E Ca dũa parte viía | sa grand' enfermedade, CSM 218:19Synaeresis is very common in forms of the verb veer, and works well musically in the form viía here, as the compressed metrical syllable vii- falls on a three-note ligature .onyeye. .onyeye
220:3 E Ca muito é gra dereito, | que d' ángeos é servida .owo
226:5 F, E que Santa Maa fez, e creed' a mi CSM 226:5This is a very definite case of synaeresis, as creed' is clearly laid under a single virga .on in [E]. .on
226:13 F, E e tiía o lógo guardad' e e paz; .o
232:16 F, E e foi-se pera sa térra | e séus hómees enou CSM 232:16The compressed metrical syllable -mees from synaeresis in hómees falls on a single short note in the music, but the result is entirely natural since the shorter form hómes is very common elsewhere. .o
241:66 F, E e manteer castidade | nos días que durasse, CSM 241:66Synaeresis in manteer works well with the music, as -teer falls on a two-note ligature .royo. Note that both [F] and [E] actually have the more regular form mantẽer, and although this doesn't technically allow synaeresis without losing the nasal anyway, there's no particular reason not to sing the word that way when the music accommodates the syllables quite adequately. .royo
243:18 E e chegaro aa agua, | cada ũu como viía, CSM 243:18It's hard to choose between elision of ũu to u, and synaeresis in viía, as the better resolution for the surplus metrical syllable in this line. Both are common elsewhere. I have plumped for the latter though, as it feels slightly better rhythmically, and gives the compressed syllable vii- an ascending plica .bod, which brings out the natural syllables anyway. .bod
245:5 F, E ros dos ángeos reía; | e porend', amigos méus, CSM 245:5The intended synaeresis in ángeos is clear in [E], with -geos sitting neatly below a single virga .on. It also occurs several times in the same word elsewhere. Note, however, that Elmes has mistakenly assumed synaeresis in reía instead, though this is not found anywhere else. .on
245:42 F, E ca mais no soo atrevudo | de dar, par Nóstro Seor.” .royo
249:30 F, E e nos gaa' o de séu Fio, | que nos vẽo remiír. CSM 249:30Synaeresis in forms of the verb gaaar occurs a few times elsewhere, as do shorter forms from the infinitives gaar and gãar. .o
251:76 E e pois viü a Hóstïa | alçar, disso: “Ahá, CSM 251:76The word hóstïa and its plural hóstïas appear a total of 18 times in the Cantigas, and in 17 cases it clearly has three syllables. However, the modern words in Galician ("hostia") and Portuguese ("hóstia") have just two syllables. This, along with the fact that the word never participates in a rhyme in the Cantigas, and that this line is the only one where it appears at the end of a non-rhyming hemistich, is pretty conclusive evidence that the word has antepenultimate stress, on the first syllable, os-. In order to maintain the metrical scheme of this Cantiga, therefore, I have combined diaeresis in viu (which is very common elsewhere) with synaeresis in hóstïa, so that the natural stress falls in the right place. .on
255:42 To, E e escoorido, CSM 255:42[T] has mort o marid' e escoorido (all on one line), which elides away the rhyme on marido. I prefer synaeresis in escoorido to fix the metre, taking advantage of the ascending plica .bod onto which the compressed syllable falls. Note that the root coor (= 'colour') becomes monosyllabic cor in modern Portuguese and Galician. .bod
255:67 To, E mas a aogo CSM 255:67Both [To] and [E] have mas ao fog' a levou que ardia here, masking the rhyme on fogo entirely. In order to restore it, I have moved the object pronoun a from the next line, and marked synaeresis here to compensate. The phrase mas a a o fogo levou que ardia (= 'but [he] took her to the burning fire') is, I think, a good candidate for being the text as originally intended, with the first a (= 'her') then having to drop out to fit the metre, by means of conflation with the following a (= 'to'). At some later point in the copying chain this was then 'corrected', restoring the first a but in a different place, which made the grammar more explicit but played havoc with the metre. By the way, I have marked synaeresis in ao rather than synalepha between the two a's, simply because assigning ao the long plica .oron is better musically. .oron
256:17 F, E e haa ta gra féver, | que quena viía entô CSM 256:17Synaeresis in forms of the verb veer is common. .on
257:13 F, E nas relicas, pero siía envóltas e panos; CSM 257:13In this line and line 16, synaeresis is required in the same word, siía. Here the compressed metrical syllable falls on just a single short note, but the musical result is quite acceptable. See also CSM 84:27 which has synaeresis in the same word. .o
257:16 F, E e as arcas e que siía mal desbaratadas; CSM 257:16See the note to line 13. In this case the compressed metrical syllable happens to fall on a descending plica .ron so the three natural syllables come out in the singing. .ron
271:12 F, E todaa a cuidava | e vïíana combater; .ood
282:3 F, E Ca muito é gra vertude | e pïadad' e mercee CSM 282:3Synaeresis at the end of this line, and the next two, is required in order to correct the syllable count and give the hemistich final stress, thereby matching the masculine rhyme in all other stanzas. Unfortunately, the music scribe in [E] was rather inconsistent here, writing an extra note for the final -e of mercee and cree, but not vee. The fix, from the point of view of the manuscript, is simply to delete the extra notes, as I have done at E2 and E3 on the Music tab, so that all stanzas then fit once the synaeresis is applied in the first one. From the point of view of Elmes' transcription, it is necessary to replace the two crotchets on A at the ends of lines 3 and 4 of the music with a single minim (as is found at the ends of the other four lines). .o
282:4 F, E d' acorrer sól por u vérvo | a que e ela be cree; .o
282:5 F, E ca estando co séu Fio, | todo sab' e todo vee, .o
292:62 F, E guisado de seer ta alte | com' ela, ne ta igual. CSM 292:62Synaeresis is very common in the infinitive seer and other forms of the same verb. .o
293:24 F, E que e pées estar no pode | e lógu' erra caeu; CSM 293:24The word pées with synaeresis just gets a single long note in the music, but given that the shorter form pés occurs elsewhere (CSM 134:36), the result is entirely satisfactory. .on
293:32 F, E a Virge Santa Maa | quis dele mercee haver, CSM 293:32Synaeresis in mercee is very common elsewhere. .o
297:1 F, E Com' é mui bõ' a creença | do que no veem' e cree, CSM 297:1The synaeresis in vee is clearly indicated in [E], where a single plica .ron is written above the word. .ron
301:20 F, E dos céos co gra compaa | d' ángeos que sigo tiía, CSM 301:20In this line both ángeos and tiía are obvious candidates for synaeresis, each word commonly being compressed elsewhere, but I have chosen the latter as it fits the rhythm of the music in [E] much more convincingly. .on
302:22 F, E feitas cousas desguisadas | ne cobiíças per pobrezas. .o
305:13 F, E serviç' a Santa Maa | per que sa mercee housse, .on
306:17 F, E vẽo o ángeo do céo, | per que s' houve d' emprear CSM 306:17Synaeresis in ángeo is well attested elsewhere. .o
306:26 F, E e disse aos crischãos: |Veede que ides creer: CSM 306:26Synaeresis in forms of the verb veer is also very well attested in other cantigas. .o
313:27 F, E Todos e pegoo éra | e e gra coita mortal, CSM 313:27The compressed metrical syllable -goo in perígoo falls on just a single virga .on, but the synaeresis can easily be justified easily from the proparoxytone (antepenultimate) stress, and the appearance of the synonymous three-syllable form perigo in CSM 36:11 and CSM 235:82. .on
314:52 F, E sempr' a nós be e saúde, | que guaaa do que na cruz CSM 314:52Synaeresis in guaaa works well musically, as guaa- falls on a two-note ligature .bowo. Note that the verbs guaaar, gaaar (forms of which exhibit synaeresis in CSM 125:60, 249:30 and 289:6) and the two-syllable gaar (e.g. CSM 234:5) are all variants of the same word. .bowo
315:2 E per que séu santo nome | dev' a seer mui loado. CSM 315:2From the single overlaid punctum .o in [E], synaeresis in seer is very clearly intended. .o
315:21 E de triígo, que na boca | meteu e que passada CSM 315:21Synaeresis in triígo works well musically, as trii- gets a descending plica .ron. Note also that the shorter variant trigo appears in CSM 112:9. .ron
340:60 E t' éu veer co el, ca sea .on
345:58 E viía co el fugindo, | ca viía fóg' acender CSM 345:58Synaeresis is required in viía here, and is common elsewhere in forms of the verb veer. .on
349:1 E Muito praz aa Virge santa | que Déus fiou por parenta CSM 349:1The desired synaeresis in aa is clearly indicated by the music in both copies of this cantiga in [E], i.e. CSM 349 and CSM 387. .o
349:22 E o ángeo de Déus co ela, | u ela disse: “Sergenta CSM 349:22Synaeresis in ángeo is rather common, a clear example being CSM 245:5. A two-syllable plural form written anjos is also found in CSM 329:21. .on
351:19 F, E E pero que be coa, | no tiía que éra re, CSM 351:19Synaeresis in tiía works well with the music, as the compressed metrical syllable falls on a two-note ligature .oso, preserving the natural hiatus in singing. .oso
354:4 E Madre de Déus Jesú-Cristo, | a u rei que muito tiía CSM 354:4In this line, the compressed syllable tii- of tiía sits below a convenient long plica .ood, so the natural syllables are preserved in singing. .ood
355:41 E e dormiü na igreja | e aassa avogada CSM 355:41It is possible that the original spelling dormio in the first hemistich here was the scribe's attempt to indicate the necessary diaeresis. (On the other hand, pediu in the next line was originally spelled pedio, and no diaeresis is required there.)

In the second hemistich, aa with synaeresis falls on a two-note ligature .owo in the music, thus preserving the natural syllables in singing.
.owo
355:42 E pediu mercee e ajuda | que ' housse perado. CSM 355:42Synaeresis in mercee is common elsewhere, e.g. CSM 105:1, CSM 180:50. .on
358:10 E que a mẽor pédra delas | no poa seer movuda CSM 358:10Synaeresis is very common in seer, and in this case the word keeps its two natural syllables when sung, as it falls on a long ascending plica .ood. .ood
359:34 E quisse i, o veea. | E eles de gra randô CSM 359:34Synaeresis is found quite commonly in forms of the verb veer. .o
364:15 E Alí jaa cavando | u día triínta obreiros CSM 364:15The compressed syllable of the word triinta with synaeresis falls on an ascending plica .bod, preserving the natural hiatus. Modern Portuguese and Galician have trinta with two syllables. .bod
365:32 E e tornou-s' ao dormidoiro | e foi mui léd' aficado CSM 365:32Synaeresis in ao appears to be the only resolution here; cases are rare in this word, considering its frequency, but there are other examples in CSM 64:64 and 155:45.  
366:28 E e os va vender a furto | por no seere coosçudos. CSM 366:28Synaeresis is common in forms of the verb seer. .on
369:14 E que per el nïú dereito | nunca be éra juïgado; | demais éra orguoso CSM 369:14The metrical syllable jui- from juïgado with synaeresis falls on a four-note ligature .bowowa, so the natural syllables are easily preserved. Note also that the variant form julgado of the same word has just three syllables (see CSM 301:11). .bowowa
371:15 E Outros viía per lavrare | e gãar i séu jornal .on
377:4 E Santa Maa do Pórto | por u hóme que se tiía CSM 377:4The word tiía requires synaeresis here to fit the metre. Confusingly, though, the music in [E] clearly supplies nine separate notes and ligatures for the hemistich, and three of them are written above this word. On closer inspection, however, it must be the earlier punctum .o written above hóme that is the impostor, as it has no counterpart in the otherwise rhythmically identical phrase for fez u miragr' a Reía in the previous line. I recommend therefore that this note be struck out as shown on the Music tab, such that this line fits with the indicated synaeresis, and the corresponding lines in the remaining stanzas all fit as written. .on
377:17 E mais o que tiía os seelos | a houve mui mal parada CSM 377:17Both tiía and seelos (seelo = 'seal' < Lat. sigillum) require synaeresis in this line. .o .oron
378:6 E macar s' el muito trabaa | de nos seer destorvador. CSM 378:6The word seer gets just a single punctum .o in the [E] music, so this synaeresis is clearly intended. .o
379:22 E e cujo termino éra, | come hómees malfeitores. .on
380:49 E mercee, e rogar--ei .on
382:1 E Verdad' éste a pavoa | que disse Rei Salamô CSM 382:1The music in [E] gives the -voa of parávoa a single short note, and synaeresis is therefore clearly intended. Contrast CSM 69:78 and CSM 65:135 where parávoa and parávoas have four metrical syllables. Note though that the more common form of the word (meaning "word") is paravla with just three syllables, and palavra also occurs. .o
383:5 E por ũa moér que tiía | o coraçô desejoso CSM 383:5The word tiía gets just two notes .on + .o in the music in [E], so the synaeresis here is clearly intended. .on
384:25 E e pero assí jaa, | viía-e sempre a mente CSM 384:25It is rather hard to choose between synaeresis in viía and elision of sempre to resolve the surplus syllable here; neither would be remarkable in a different context. I've chosen the synaeresis, though, if only because the scribe could easily have written sempr if that had been intended. .o
384:36 E levar-t-ei suso aoo, | u verás as aposturas, .o
389:17 E de ta grand' enfermidade | que por mórto o juïgava CSM 389:17Compare the variant form julgava in CSM 245:30, and the related word juïgado with synaeresis in CSM 369:14. .o
389:22 E a sa casa o levasse, | e escïas e comios CSM 389:22The word espécïas normally has four syllables and antepenultimate stress, but synaeresis between the final two unstressed vowels (matching the modern pronunciation) seems the most natural way to resolve the extra syllable here. Compare hóstïa in CSM 251:76. .o
398:34 E ca o Bõo Pastor tiía | a sa Madre por caiado. CSM 398:34This is one of no less than 17 cases of synaeresis in the verb form tiía (or its plural tiía) that are found throughout the Cantigas. .on
403:21 To o tiía, e osmando .o
411:10 To, E do linaj' onde viía | esta Seor honrrada. CSM 411:10Synaeresis is commonly found in the word viía. In this particular case, the compressed metrical syllable vii- falls on a long ligature .onyeyeye in the music, so the natural length is preserved in singing. .onyeyeye
411:42 To, E quis Déus que do séu ángeo | foss' ela confortada. CSM 411:42Synaeresis in ángeo is rather common: it is also required in lines 81 and 97 here, along with several other cantigas, e.g. CSM 141:42, 306:17, 349:22. Pronunciation as anjo would be quite acceptable, especially as the plural form anjos is attested in CSM 329:21. .on
411:81 To, E atẽes que o ángeo | foï dalí partido, CSM 411:81Mettmann places the hemistich bar after ángeo in his main text, as here, but points out the metrical problem in a footnote and suggests moving the final -o of ángeo to the second hemistich as a possible resolution. However, angeo with synaeresis falls in the corresponding position in lines 42 and 97, and in fact it looks as though in [To] an attempt was made to erase the e of angeo in this line (though it is still just visible). It therefore seems most elegant to use synaeresis again in this case, with diaeresis in foi to compensate (cf. CSM 414:11). .on
411:97 To, E que o bẽeito ángeo | fera ja certeira CSM 411:97In this case, [To] has que o bẽeit angeo, which has the right number of syllables but doesn't fix the stress problem. I have therefore kept bẽeito from [E] and marked synaeresis in angeo as in lines 42 and 81. .on
422:16 To, E U verás os ángeos | estar ant' ele tremendo, .on
423:14 To u por seerme deles decendeu. CSM 423:14Synaeresis is commonly found in the verb seer. .ron
426:31 To ena maneira que o veedes daCSM 426:31Synaeresis is commonly required in forms of the verb veer. .o

Diaeresis (59)

Diaeresis is the opposite of synaeresis: a single-syllable diphthong is divided to become two vowels in hiatus, and therefore two syllables. As is traditional, I mark this with a trema (two dots) above the ‘semivowel’ which is not the nucleus of the diphthong, thereby promoting it to a full vowel, but this is also highlighted in red to distinguish it from the ordinary marking of natural hiatus (e.g. in dïabo). Here's a full list of the words in which I have marked diaeresis, and the frequency:

apareceü (1), bẽeto (1), beveü (1), caeü (2), coseü (1), creceü (1), dormiü (1), encobriü (1), entendeü (2), faïçõada (1), foï (2), guariü (1), maïs (3), meteü (1), muï (2), muïto (2), oiü (4), pascöa (1), pascüa (1), pediü (1), pidiü (1), prendeü (2), reïs (7), reliquas (1), resorgiü (1), saiü (3), sentiü (1), serviü (2), viü (11)

You can see that well over half of these cases involve 3rd person preterite verb forms ending in -eu or -iu, which were presumably somewhat unstable during the period, and indeed disyallabic forms in -eo and -io exist in some modern Galician dialects (see [NOMIG] pp. 110–111). Seven further cases involve the word reis (= kings), for which it is rather difficult to ascertain whether the basic pronunciation had one or two syllables (see the footnote to CSM 1:36).

In the table below, the note or ligature shown is the one that the promoted vowel falls on, and for diaeresis it is generally preferable that this be a short note, so as not to give unnatural prominence to the artificial syllable—but, of course, the broader musical context is very important.

Line Sources Text in this edition Music
6:39 To, T, E mas déu muï maa noite | a sa madre, a mesquia, CSM 6:39There are perhaps three potential sites for diaeresis in this line: déü, muï or noïte, but only in mui does diaeresis occur elsewhere. (It's very common in preterite verb forms ending in -eu pronounced with close e [eu̯] but apparently never necessary in déu which has open e [dɛu̯]). And although diaeresis in noite has by far the best musical result, with a very pleasing pattern of falling notes on the word (at E1 in the music), it's metrically rather suspect (since the artificial syllable i would have to be stressed) and so I have stuck with mui as the site for the fix here. .royo
35:9 To, T, E por ũas sas reliquas | que levaro ũa vez CSM 35:9The word reliquias has a rather odd-looking case of diaeresis here, but there is no other way to resolve the missing syllable. As it happens, the musical result is not too bad in a cantiga already full of long ligatures and stretched-out syllables, and the artificial syllable -qui- does at least fall on a single punctum .o. Contrast CSM 362 where reliquias occurs four times, always with the expected 3 syllables. .o
76:31 T, E Quand' a moér viü o gra miragre que fez CSM 76:31This is the first of no less than 11 cases of diaeresis in the word viu that are required in the Cantigas. .on
115:77 To, T, E dessa Pascöa santa; CSM 115:77Diaeresis: Pascoa requires three metrical syllables here, but only two in line 256 and generally elsewhere (also spelled Pasqua or Pascua). The only other case of diaeresis is in CSM 333:39 (Ca des Pasa i jouve). .oron
116:21 T, E muïto, que jaou CSM 116:21Diaeresis of muito to three metrical syllables is required here to fill out the line, and fits the music very well: mu- and -i- fall on short puncta .o + .o (the first two crotchets of line 5 line Elmes' edition). .o
122:1 T, E Miragres muitos pelos reïs faz CSM 122:1Diaeresis: reis requires two metrical syllables here, which is very clear from the text laid under the music in [E] and [T]. See also the note to CSM 1:36 for more on this word. .on
139:23 T, E déu aa Virge, muï pequenio, CSM 139:23Diaeresis in mui resolves the missing metrical syllable here. .bod
157:25 T, E i, e muïto chorando; | e pois fez sa confissô, CSM 157:25Diaeresis in muito is really the only available resolution for the metrical shortfall in the first hemistich here, but it is at least found in the same word in CSM 116:21. The musical result in this case is acceptable, if not exactly scintillating. .on
171:51 T, E viü séu fi' ant' o altar CSM 171:51Diaeresis occurs in viu more often than in any other word, and is therefore a good resolution for this line's metrical shortfall. .on
192:37 T, E Maïs no pora, CSM 192:37Diaeresis in mais here works satisfactorily with the musical structure. See also line 69, where it occurs again, with mais in the same position in an identical phrase of music. .oron
192:69 T, E maïs defendendo CSM 192:69See note to line 37. .oron
193:49 T, E e el dalí adeante | sempre serviü de grado CSM 193:49Diaeresis is fairly common in 3rd person single preterite verb forms, such as serviu here. In this example, the musical result is quite pleasant, with each of the three syllables falling on a single long note. Interestingly, [T] has the noun serviço in place of serviu, which does scan correctly, but makes no sense in isolation given that the rest of the stanza is identical with [E]. This suggests that another verb may have been lost at some point along the copying chain: for example, the text might originally have said: e fez dali adeante | sempre serviço de grado (or e_el fez... with synalepha) or perhaps e el dali adeante | sempre serviço de grado // fezo a Santa Maria | .... .on
214:33 F, E Mas a Virge, que de Reïs | ve de todo-los aos, CSM 214:33Compare CSM 122:1, which also requires diaeresis in reis. .eyeye
221:1 F, E Be per está aos reïs | d' amare Santa Maa, CSM 221:1This is a case where it is clear from the text laid under the music in [E] that reis has two syllables. See the note to CSM 1:36 for more on this word. .o
222:21 F, E caeü dentro no cáliz, | esto foi sabud' e visto, CSM 222:21Mettmann, following both [E] and [F], has caeo at the start of this line, and in this case the variant spelling may well represent an attempt by the scribe to indicate the necessary diaeresis. I have normalized the spelling and marked the diaeresis with a trema as usual. .ood
222:33 F, E pelo braço e saiü | viva aquela araa, CSM 222:33Compare line 21 above, and also CSM 171:51, 193:49 etc. This is another case where the preterite verb form requires diaeresis, turning the final diphthong into two syllables and thereby shifting the stress away from the end. .o
237:24 F, E e saiü-s' entô dalí, | e no muit' ascondudo. CSM 237:24In this line and in line 95 we have another pair of examples of the fairly common use of diaeresis in the preterite verb form. Both cases here come from the verb saír. .royo
237:95 F, E E a cleria da| saiü mantenente, CSM 237:95See note to line 24. .o
238:16 F, E na Virge e pois prendeü | por el mórte dos judéus, CSM 238:16There are four lines in this cantiga where a preterite verb form undergoes diaeresis in order for the first hemistich to end in an unstressed syllable; apart from prendeu here, the other cases are in lines 27 (oiu), 41 (oiu) and 54 (serviu). In this line [F] actually has prendeo, where the non-standard -o may be intended to indicate a separate syllable, although this is not done consistently. .royo
238:27 F, E O capelá, quand' oiü | dizer mal do Salvador CSM 238:27See note to line 16. Here [F] has quando oyu, which has the right number of syllables, but still with the wrong stress. .royo
238:41 F, E Quand' est' o crérig' oiü, | diss': “Ai, Groosa, val! CSM 238:41See note to line 16, and compare line 27. Here both manuscripts have Quand' est' o crerigo oyu, again with the right number of syllables but the wrong stress, so oiu must be expanded to three syllables by diaeresis with crérigo elided to compensate. (I have chosen elision over synalepha since the compressed syllable falls on a single note, and elision of this word is quite common elsewhere.) .on
238:54 F, E qual seor ele serviü, | assí o gualarou. CSM 238:54See note to line 16. .royo
251:76 E e pois viü a Hóstïa | alçar, disso: “Ahá, CSM 251:76The word hóstïa and its plural hóstïas appear a total of 18 times in the Cantigas, and in 17 cases it clearly has three syllables. However, the modern words in Galician ("hostia") and Portuguese ("hóstia") have just two syllables. This, along with the fact that the word never participates in a rhyme in the Cantigas, and that this line is the only one where it appears at the end of a non-rhyming hemistich, is pretty conclusive evidence that the word has antepenultimate stress, on the first syllable, os-. In order to maintain the metrical scheme of this Cantiga, therefore, I have combined diaeresis in viu (which is very common elsewhere) with synaeresis in hóstïa, so that the natural stress falls in the right place. .oron
253:15 F, E O bo hóme entendeü | que andava e pecado, CSM 253:15The first hemistich of this line must end on an unstressed syllable, so I have marked diaeresis on entendeu. However, this means that a syllable has to be lost earlier in Mettmann's line (O ome bõo entendeu), and the usual options are of little use: synalepha at the start in O ome is horribly clumsy, synaeresis in bõo is technically blocked by the nasal [ŋ], and elision to bõ' before entendeu has no precedent. (The only occurrences of bõ' are where the feminine form bõa is reduced before a word beginning with a-, which is much more natural.) Since home bon is not an option either as bon must precede the noun, I have taken the liberty of swapping the words around. Fortunately, bon home is a common phrase elsewhere, occurring around a dozen times. .royo
253:28 F, E a eigreja do camio | viügo mantẽente CSM 253:28Diaeresis in the ending of a preterite verb form is required here and in line 39 to fill out the metre. .on
253:39 F, E E perdô de séus pecados | pidiü be alí lógo .o
253:44 F, E qu' eas péças caeü | ja feit', e porê sinou-se CSM 253:44As in line 15, I have fixed the first hemistich to ensure that it ends on an unstressed syllable. In this case, elision of que at the start is required to compensate for the diaeresis in caeu. .royo
264:21 E como moér que fosse | mui be faïçõada, CSM 264:21Mettmann actually points out the diaeresis in the word faiçõada here, in an act of generosity to the potential performer that is without precedent in his first two volumes. .e
267:68 F, E que sól no sentiü coita ne rancura; CSM 267:68Compare e.g. CSM 238 and 253, each of which has several cases of diaeresis in the preterite verb form. .on
268:18 E os oiü esta dona, | fez o que nós vos diremos: .on
273:40 F, E todos a Santa Maa; | e el coseü os panos CSM 273:40[F] has coseo here, which might be intended to indicate the necessary diaeresis. However, in the next line [F] has the regular ending in encobriu, so we cannot draw any firm conclusions. Besides, consistency wouldn't be any fun, would it? .on
273:41 F, E mui be co aqueles fíos | e encobriü os danos, CSM 273:41See line 40. .on
282:27 F, E Quand' est' oiü o padre | e a madre, gra loor .o
291:36 F, E apareceü-' a Virge, | a que de todo be praz, CSM 291:36This is another preterite verb form with diaeresis; see the Modifications for metrical regularity page for a complete list. .bod
294:12 F, E e perdendo aos dados, | creceü-' ê ta gra saa CSM 294:12[F] creceo. This is another of many cases of diaeresis in a preterite verb form. The only unusual thing about it is that Mettmann actually comments on the pronunciation in his footnotes, observing that the word creceu must have three syllables. .bawon
304:7 F, E d' ôrdi de Sa Bẽeto. | E óra chus da eigreja CSM 304:7As an adjective, bẽeito normally has just three syllables, but Bẽeito here is a proper noun—the Latin name "Benedict(us)"—and a more conservative pronunciation is therefore quite natural. Note that the diaeresis which separates the e from the i corresponds to the position of the lost -d-. .o
305:24 F, E e viü seer na rúa, | com' éu achei por verdade, CSM 305:24Here and in line 51 we have another couple of examples of the frequent diaeresis in 3rd person singular preterite verb forms. .on
305:51 F, E que meteü na balança, | chẽos de prata e d' ouro; CSM 305:51See note to line 24. .o
310:19 F, E que por nós prendeü mórte .ra
312:59 F, E Entô aquel cavaleiro | entendeü que errara .ron
315:44 E com' éra tod' inchado; | maïs no e sentiro CSM 315:44Diaeresis in mais also occurs in CSM 192:37 and 192:69. .o
316:38 F, E viügu' e foi be são, | e começou de chorar CSM 316:38The first hemistich of this line could also, in theory, be fixed by restoring the unelided form lógo, but since diaeresis in viu and other 3sg preterite verb forms is so common, it is almost certainly intended here. .on
319:48 F, E viü a eigreja, lógo mantẽente CSM 319:48This is a clear-cut case of diaeresis being required in the word viu; there is no other simple resolution for the metrical shortfall. .o
321:24 F, E todo-los reïs crischãos | ha aquesto por vertude CSM 321:24Diaeresis occurs several times elsewhere in reis; see the note to CSM 1:36 for more on this word. .o
329:50 E a compaa que se fora, | no viü i re seer. .o
333:39 E Ca des Pascüa i jouve | assí como vos éu digo CSM 333:39Compare CSM 115:77 where the same word, with the alternative spelling Pascoa, also undergoes diaeresis. .o
334:38 E resorgiü e foi são | como soía seer. CSM 334:38Yet another example of diaeresis in a 3rd person singular preterite verb. .o
346:21 E a guariü be daquela | enfermidade sobeja CSM 346:21The word guariu shows another example of the common diaeresis in 3sg preterite verb forms. In this case the musical result is very good, as the extra syllable falls on a single punctum. .o
353:51 E O abad' ao menio | viü-e cambiar a faz .bod
355:19 E El catou e a manceba | viüir e pesou-e, .on
355:41 E e dormiü na igreja | e aassa avogada CSM 355:41It is possible that the original spelling dormio in the first hemistich here was the scribe's attempt to indicate the necessary diaeresis. (On the other hand, pediu in the next line was originally spelled pedio, and no diaeresis is required there.)

In the second hemistich, aa with synaeresis falls on a two-note ligature .owo in the music, thus preserving the natural syllables in singing.
.on
357:18 E os reïs aos séus presos | que no seja justiçados, .o
377:13 E mais o hóme por mercee | e pediü que e désse .o
384:32 E e viü que ao leito | se chegava passo indo, .bowa
393:26 E que beveü muita d' agua, | co que guareceu de mórte. CSM 393:26Here we have another example of the very common appearance of diaeresis in 3sg preterite verb forms. The only unusual thing about this particular case is that Mettmann actually points it out in his footnotes. .o
411:81 To, E atẽes que o ángeo | foï dalí partido, CSM 411:81Mettmann places the hemistich bar after ángeo in his main text, as here, but points out the metrical problem in a footnote and suggests moving the final -o of ángeo to the second hemistich as a possible resolution. However, angeo with synaeresis falls in the corresponding position in lines 42 and 97, and in fact it looks as though in [To] an attempt was made to erase the e of angeo in this line (though it is still just visible). It therefore seems most elegant to use synaeresis again in this case, with diaeresis in foi to compensate (cf. CSM 414:11). .on
414:11 E a Joséf, co que foï esposada, CSM 414:11Diaeresis in the word foi is not common—CSM 411:81 has the only other place I have marked it—but it works well enough against the music here, with the unstressed -i falling on a short plica .ron. .ron
422:36 To, E E u todo-los reïs | fore ant' el homildosos, CSM 422:36The first hemistich of this line should end in an unstressed syllable to be consistent with the metre of the other stanzas, so I have marked diaeresis in reis (which is common elsewhere) and compensated with synalepha in E u at the start, which falls on a convenient two-note ligature .oswjo in the music. .on
423:26 To dos Reïs foi aorado por Rei. CSM 423:26The word reis undergoes diaeresis in several other cantigas. See the note to CSM 1:36 for more on this word. .ron

Melisma (6)

Melisma, in the restricted sense that I use here, is the singing of a single syllable of text to more than one individual note or ligated group. Whilst this may be the rule in many kinds of Early Music, it is an exception of extreme rarity in the Cantigas de Santa Maria. I have found it to be necessary in just three cantigas, and in two of those it is only to fix genuine faults with the music as written. (Note that although the marking of diaeresis in diphthongs (see above) has, in practical terms, the same musical realization as a melisma, the former is a purely metrical solution which actually creates extra syllables, rather than extending existing ones.)

Line Sources Text in this edition Music
72:21 To, T, E E u(u) quis do ventre séu CSM 72:21See main note to this cantiga. Note that in a different context, the resolution for this line would probably be something a little more normal, such as replacing quis with its full form quiso. .o
72:25 To, T, E Séu pa(a)dre, quand' est' oiu, CSM 72:25See main note to this cantiga. Once again, in the absence of other problems, I'd be far more inclined to fix this line by expanding quand' to quando or est' to esto. .o
72:37 To, T, E te dou(u) que o acharás CSM 72:37See main note to this cantiga. .o
72:41 To, T, E da té(è)sta e a serviz. CSM 72:41See main note to this cantiga. [T] actually has E a testa here, rather than Da testa, which on the face of it would appear to scan properly. Given the evidence of the other two manuscripts, however, it seems likely that this is just the result of the illuminator adding the wrong decorated capital (a common mistake) after the scribe had written -a testa. .o
169:26 T, E quand' el Rei d' Ara(o), | Do James de gra prez, CSM 169:26The first hemistich here is short of a syllable in the original, and given that it also has final stress (in Aragón) which is at odds with the metrical scheme, it would appear that the syllable is missing from the end, technically speaking. The use of melisma here therefore seems likely to have been the intention, i.e. drawing out -gón over the two notes .on + .o, paralleling the syllable -rrar of honrrar on the two-note clivis .royo in the refrain. (This is in effect the opposite of the fix for line 8.) .on
263:5 F, E Ca ela sempr' a nós dá(a) | que façamos o meor CSM 263:5The text for the first hemistich of this line, as written by the lyric scribes in both [F] and [E], and copied by Mettmann, is Ca ela sempre a nos dá. This has the correct number of syllables, and Mettmann lets it pass without comment. Things are not quite so simple for those of us who care about the music, unfortunately: the hemistich is supposed to end with an unstressed syllable, to match every other line, but the verb here naturally attracts the stress. The music scribe in [E] obviously latched onto this, and wrote just a single note for -pre a, in order to force synalepha (or elision) and thereby bring forward in the melody to a point where the rhythm matches the stress, leaving it with a long note at the end of the phrase. Of course, the seven notes/ligatures that remain then completely fail to fit the other stanzas. To fix this, I have just added an extra punctum at the end of the phrase for the unstressed syllable, at E2 on the Music tab, matching the music for d' erro in the next line; and for the first stanza I have kept the desired elision to sempr' a and drawn out at the end with a small melisma, so as to keep the stress and rhythm in agreement. With reference to Elmes' edition, this means that the fourth bar of line 5, which contains just a dotted minim on D, should be changed to match the fourth bar of line 7, which has D minim + D crotchet. .on

Breaking (41)

Breaking is the opposite of melisma: two syllables are sung to a single note or ligature. Although there are many more cases of breaking than melisma, breaking is still very rare. I have only used it in seven cantigas, although it does tend to happen repeatedly within each of those, being generally indicative of a systematic metrical problem. (For example in CSM 162 and CSM 173 the metre of the stanzas changes wholesale part way through.) See the notes to the individual cantigas for the full justification in each case.

Note that in a handful of the cases where I've marked breaking, the purely metrical solutions of synalepha or synaeresis would work too, as there is no consonant at the syllable boundary. However, whilst synalepha would certainly be preferred out of context, I have used breaking instead for consistency with other stanzas in the same cantiga.

Except in CSM 173, the two syllables in question always fall on a two-note ligature, which means that the musical fit is simple and obvious. See the footnotes to CSM 173 for more on how to deal with that slightly more awkward case.

Line Sources Text in this edition Music
18:8 To, T, E e Estre·madura, CSM 18:8This is the first of a small number of awkward cantigas where the first stanza is metrically irregular relative to all later stanzas. In this case, the line E Estremadura has six natural syllables rather than the five required. Unfortunately, as often happens, the music appears to have been set down with only this first stanza in mind: in all three manuscripts, it includes the irregularity, providing a separate note or ligature for all six syllables (see E2 in the music, and click 'Expand edits' if necessary). However, comparing the musical phrasing for lines 5 and 6 with that of lines 7 and 8, it quickly becomes apparent that the two separate puncta .o .o given to -trema- (Elmes: D and C in the fifth bar of line 4) are in effect a broken .oyo ligature that must be joined up again in order to fit all subsequent stanzas. The most appropriate musical resolution for this hypermetric line is therefore to conflate the natural syllables tre and ma as indicated, although of course the pronunciation is unaffected. By the way, this is a very good example where ignoring the music would lead almost inevitably to the wrong conclusion: from a purely metrical point of view, the elision 'n Estremadura or En 'Stremadura would most likely be proposed, and that'd be the end of it. .oyo
62:16 To, T, E porque a usura e creceu atanto CSM 62:16See main note to this cantiga. .ood
62:36 To, T, E “ve acá, méu fio”, e poi-lo deitava CSM 62:36See main note to this cantiga. .ood
62:39 To, T, E que sól no disséro: “Dona, onde vẽes?” CSM 62:39See main note to this cantiga. .onwo
62:40 To, T, E ne “de que ovas, gra tórto nos tẽes.” CSM 62:40See main note to this cantiga. .ood
62:41 To, T, E Esto fez a Virge que ja outros bẽes CSM 62:41See main note to this cantiga. .onwo
162:20 To, T, E foi e pose-a no maior altar, .oyo .oron
162:26 To, T, E porque a no viu de bo semear. .oyo .oron
162:32 To, T, E a foi u x' an·te so·ía estar. .oyo .oron
162:38 To, T, E e a eigre·ja fo·sse be serrar. .oyo .oron
162:44 To, T, E a foi a quan·tos s' i fora juntar .oyo .oron
162:50 To, T, E muitas gentes e do séu ofertar. .oyo .oron
173:15 T, E [...] E que esto no dissésse | a outri, mas sa carrei·ra .on
173:16 T, E se foss'. E el espertou-se | entô e achou entei·ra .on
173:17 T, E a pédra sigo na cama, | ta grande que verdadei·ra- .on
173:19 T, E Entô el a Virge muito | loou. E nós a loe·mos .on
173:20 T, E por este ta gra miragre, | e gra dereito fare·mos, .on
173:21 T, E ca sempr' ela nos arre | enas coitas que have·mos, .on
224:6 F, E e que aquesto no cre·e, | faz torpidad' e loucura. CSM 224:6In this line we could assume ordinary synaeresis on cree with the same musical outcome. I have chosen to mark the ligature breaking, however, for consistency with the other hypermetric lines where synaeresis is not an option due to intervening consonants. .oyo
224:12 F, E e riba d' Aguaa·na | há u logar muit' honrrado CSM 224:12[F] en is missing at the start of this line, which makes no sense at all. .oyo
224:18 F, E muito e Santa Ma·a; | mais haa gra tristura CSM 224:18[F] muit' en Santa Maria. .oyo
224:33 F, E que faz grandes e Tere·na; | porend' ambos outorgaro .oyo
224:34 F, E de levar i a meni·a, | firo atal postura. CSM 224:34Mettmann I and II de levar la menynna. Given that the first hemistich of this line ends with an unstressed syllable in both manuscripts, it still needs another syllable to fix the metre. I have inserted the word i on the model of CSM 176:26 e levou i a omagen, and changed la to a accordingly. The original word la appears in both manuscripts but wouldn't normally be expected in this position (levar a or levá-la would be more regular), so it could be a Castilian-influenced copying error for i a or ia anyway.

Mettmann I fizeron; Mettmann II e fizeron; [F] fezeron; [E] fizeron. The e inserted by Mettmann in his second edition makes some sense, stylistically, but I don't believe it to be essential. If you prefer to keep it, then reading the second hemistich here as e fizeron tal postura would be an acceptable compromise.
.oyo
224:46 F, E graças a Santa Ma·a, | que é Seor d' apostura. .oyo
224:49 F, E e Béja e nos loga·res; | e pois que esto souro, CSM 224:49[F] en Bej' e nos logares. .oyo
224:50 F, E feito ta maravio·so, | loaro a Virge pura. .oyo
255:74 To, E ant' a Ma·gestade CSM 255:74In line 73 [E] keeps the full word piadade, but [To] once again elides away the rhyme with por piedad' ant' a Magestade. Unfortunately, neither the latter, nor piadade 'nt' with double elision, nor even piadade_ant' with synalepha, has any musical merit—the caesura between piadade and ant' is marked in the musical notation by a division which seems entirely justified rhythmically, and Anglés and Elmes both observe it. Since there is no obvious way to compress line 74 in isolation to 4' syllables, I think this is a rare case that justifies accepting the line as hypermetric, and applying a purely musical solution: namely, breaking the plica .bod to accommodate the two syllables Mages-. As it happens, this is the same musical shape which carries the synaeresis in escoorido in line 42, so a certain symmetry is at least maintained. .bod
276:6 F, E e u logar viço·so, .oron
276:11 F, E esta que nos cau·la .oron
276:16 F, E entrou mui deantei·ro .oron
276:21 F, E foi e caeu sobr' e·sse. .oron
276:36 F, E pois que a te ta ·le, .oron
276:41 F, E com' e verdad' acha·mos; .oron
276:56 F, E sa vertude fremo·sa .oron
276:61 F, E sempre é mui loa·da e CSM 276:61This line has two extra natural syllables, and therefore requires synalepha in loada e on top of the ligature breaking found in other stanzas. .oron