Speech rhythm

[Dafydd y Garreg Wen]

Quote

I was tempted to quote more of Campbell's introduction, but feared it might send my remaining reader to sleep!  Clearly he didn't like people trying to sing in the rhythm of "conversational English" and deplored the excision of passing notes in Anglican chants. Whether he was still advocating the old measured style of chanting he doesn't say. If he was, he must have changed his mind on going to Windsor, for only a couple of years later he broadcast a Choral Evensong from there in which the chanting is most definitely speech-rhythm (it's on YouTube); however, he still thought omitting passing notes was wrong.

You were way ahead of me ... my next step was going to be to look for recordings!

As you say, speech rhythm at Windsor. But it's hard to see what else he can have meant by the rhythm of "conversational" English (tho' if it's speech rhythm he was aiming at he was using a straw man, for s.r. is supposed to be modelled on formal diction, as of one reading a lesson in a fairly resonant acoustic (without amplification!) - hardly "conversational").

Did he mean that speech rhythm was all very well for the professionals? Or did he find the style in Windsor so entrenched that he couldn't change it (or hadn't yet done so)?

I've never been convinced that speech rhythm requires the sort of drastic pruning of passing notes seen in e.g. the Parish Psalter. There is a case for taming the exuberance of passing notes in particular chants, but not across-the-board deforestation.

[Zimbelstern]

Bairstow gives directions for chanting at the beginning of “The Lamentation” as follows:

”The Lamentation should be chanted quite slowly, but in speech rhythm. The syllables apportioned to the bars following rhe reciting bars must not be sung slower than the recitation (Bairstow’s italics).”

There are certainly passing notes in the chants. We’re including this wonderful setting in a Lent concert. It will be interesting to see how it works out in practice.

 

 

[Dafydd y Garreg Wen]

2 hours ago, Zimbelstern said:

Bairstow gives directions for chanting at the beginning of “The Lamentation” as follows:

”The Lamentation should be chanted quite slowly, but in speech rhythm. The syllables apportioned to the bars following rhe reciting bars must not be sung slower than the recitation (Bairstow’s italics).”

It's ironic that this is precisely what e.g. Pluscarden Abbey does not do in singing psalms to Gregorian chant!

https://youtu.be/2K6DOXrGWZ8

Not that that's a objection to Bairstow's principles, but it is interesting.

The Lamentation is a fine thing.

 

[Zimbelstern]

.

[Zimbelstern]

Forgive me, I’m not an expert in these matters, but the monks are singing in Latin, albeit in a very fast, staccato style. A couple of years ago I attended a course on Gregorian chant in the Benedictine monastery in the Valle de los Caídos near Madrid in Spain. We sang Vespers every day with the monks. It did not sound anything like this, but then the speech rhythm of modern Spanish and Italian is not so distant from that of Latin. The speech rhythm of English is very different from that of Latin. One reason why Gregorian chant should be sung in Latin is because of the stress and accentual patterns of the language. The accent in English often falls on the last syllable of a line (hard ending), whilst in Latin the accent is normally on the penultimate syllable (soft ending). If you think of the words of, say, “There is a Green Hill Far Away” you will see what I mean. Gregorian chant can sound unnatural and stilted sung to English (essentially a Germanic language). which is no doubt the reason why we have Anglican chant. Reformation composers saw this immediately and could work in both idioms - thus Tallis’ responses. Tallis’ Canon would not work with a Latin text without modification (compare with Byrd’s canon “Non Nobis Domine”.  (A modern example of a chant written specifically for English would be Martin Shaw’s Anglican Folk Mass). It can be done - a good example is J. H. Arnold’s Compline using “traditional language”. One of the interesting things about the Bairstow is that he softens some hard endings by using accented passing notes in some of the voices!

[Dafydd y Garreg Wen]

3 hours ago, Zimbelstern said:

Forgive me, I’m not an expert in these matters, but the monks are singing in Latin, albeit in a very fast, staccato style. A couple of years ago I attended a course on Gregorian chant in the Benedictine monastery in the Valle de los Caídos near Madrid in Spain. We sang Vespers every day with the monks. It did not sound anything like this, but then the speech rhythm of modern Spanish and Italian is not so distant from that of Latin. The speech rhythm of English is very different from that of Latin. One reason why Gregorian chant should be sung in Latin is because of the stress and accentual patterns of the language. The accent in English often falls on the last syllable of a line (hard ending), whilst in Latin the accent is normally on the penultimate syllable (soft ending). If you think of the words of, say, “There is a Green Hill Far Away” you will see what I mean. Gregorian chant can sound unnatural and stilted sung to English (essentially a Germanic language). which is no doubt the reason why we have Anglican chant. Reformation composers saw this immediately and could work in both idioms - thus Tallis’ responses. Tallis’ Canon would not work with a Latin text without modification (compare with Byrd’s canon “Non Nobis Domine”.  (A modern example of a chant written specifically for English would be Martin Shaw’s Anglican Folk Mass). It can be done - a good example is J. H. Arnold’s Compline using “traditional language”. One of the interesting things about the Bairstow is that he softens some hard endings by using accented passing notes in some of the voices!

These are deep waters ....

My point (such as it was) was that it was interesting to find in one school of rendering Gregorian chant something not dissimilar to the stigmatised old way of singing Anglican chant, viz. a rapid reciting note followed by a more measured mediation or ending (as in the Cathedral Psalter, which in this respect, if not in others, seems to reflect the way that Anglican chant historically was performed). By contrast the "speech rhythm" school (nicely illustrated by Bairstow's comments) eschewed any distinction between reciting note and mediation/cadence.

I chose Pluscarden because they illustrate this style so plainly. Solesmes does the same thing but not in such a pronounced way. Solemes pre-eminence in Gregorian chant of course is not unchallenged, but that is for another discussion ....

Whether English can be sung satisfactorily to Gregorian chant is a vexed question, but the position of the (stress) accent, tho' often cited, is a bit of a red herring. Final-syllable stress is not unknown in Latin, and the chant accordingly makes provision for it; that provision can be applied just as well to final-syllable stress in English. It is true that final-syllable stress is commoner in English: J.H. Arnold reckons that the proportion is 8% as against half that in Latin. In both languages, however, stress falls elsewhere in the great majority of cases ("Where the dear Lord was cru-ci-fied"). (There are of course ambiguous cases: is it "There is a green hill far a-way" or "... far a-way"??!!)

Anglican chant seems to have developed from (improvised??) fa-burden harmonisations of Gregorian psalm tones, very probably a pre-Reformation practice which simply continued amid the various liturgical disruptions; these harmonisations seem to have taken on a life of their own only in the post-Restoration period when dropping the psalm tone element produced the (single) Anglican chants that we know and love to this very day. One can compare how Tallis harmonised the traditional plainsong tones in his responses, as did his immediate successors, but later the tone was dropped and all parts were freely composed. The change of language probably had little to do with it. (It's odd, however, that this seems happened much earlier with responses than psalm chants.)

Martin Shaw states that his Anglican Folk Mass is based on traditional (pre-Reformation) plainsong melodies, tho' I've never seen them identified (or been able to do so myself!).

Curiously Cranmer, whether deliberately or unconsciously, seems to have echoed the rhythm of the Latin in some places in his Book of Common Prayer (which he actually composed in Latin and then translated into English - some of his working notes survive). Thus in the Litany the extra words "miserable sinners" after "Have mercy upon us" allow the English to fit the cadence of the chant for "Miserere nobis".

Deep waters ... but fascinating.

[Vox Humana]

14 hours ago, Dafydd y Garreg Wen said:

As you say, speech rhythm at Windsor. But it's hard to see what else he can have meant by the rhythm of "conversational" English (tho' if it's speech rhythm he was aiming at he was using a straw man, for s.r. is supposed to be modelled on formal diction, as of one reading a lesson in a fairly resonant acoustic (without amplification!) - hardly "conversational").

Did he mean that speech rhythm was all very well for the professionals? Or did he find the style in Windsor so entrenched that he couldn't change it (or hadn't yet done so)?

He may have just meant that it was apt to be overdone. Here's what he said:

"It has been well argued that English in conversational rhythm has no place in singing in church. Without labouring the point, we may assert that when delivered upon a chord in four-part harmony it sounds musically fantastic and unstable. Few will deny that the quest for so-called speech-rhythm has carried the chanting of the psalms to ludicrous extremes besides encouraging the merciless pruning of dozens of perfectly good anglican chants. A careful reassessment of values from time to time is advisable."

Nearly all of his own chants (not that there are many) have passing notes.

He once spoke to me about the matching of chants to the psalms. I can't remember his exact words, but the gist was as follows. One had one's favourite chant/psalm matches and that, on taking up a new post, one's instinct was to make lots of changes; however it was much better to hold back and absorb the status quo properly before introducing any changes. On the other hand, you may have a point about the choir.  The boys loved him, but he frequently complained that the lay clerks "couldn't or wouldn't" do what he wanted. One should bear in mind that in Campbell's time the lay clerk-ships were jobs for life, with all the disadvantages that that entails. After Campbell's death, Christopher Robinson expanded the number of lay clerks from nine to twelve, which helped to improve the quality, and it wasn't long before most of the older men had gone. The standard shot up.

[Vox Humana]

8 hours ago, Zimbelstern said:

Tallis’ Canon would not work with a Latin text without modification (compare with Byrd’s canon “Non Nobis Domine”.

Really? Tallis's "Canon" (in the truncated form found in modern hymns books) fits Nunc sancte nobis Spiritus well enough, as well as other hymns in this metre. (O salutaris hostia might be a more familiar possibility.)

Incidentally, I don't think any scholars now believe that Byrd wrote Non nobis Domine. See Philip Brett, "Did Byrd Write 'Non Nobis Domine'?", Musical Times cxiii (Sept 1972), p.855.

[Vox Humana]

3 hours ago, Dafydd y Garreg Wen said:

Anglican chant seems to have developed from (improvised??) fa-burden harmonisations of Gregorian psalm tones, very probably a pre-Reformation practice which simply continued amid the various liturgical disruptions; these harmonisations seem to have taken on a life of their own only in the post-Restoration period when dropping the psalm tone element produced the (single) Anglican chants that we know and love to this very day.

I'm not sure that's entirely correct. It's true that an anonymous Scottish treatise describes how to improvise a four-part faburden upon a plainsong and Thomas Morley in his A Plaine and Easie Introduction to Practicall Musicke provides harmonisations for the different tones, so the technique was certainly known. Nevertheless, so far as I know, there  was no tradition in England before the Reformation of singing the psalms of the office in polyphony, faburden or otherwise. The only psalms that were sung thus were Laudate pueri Doninum and In exitu Israel, when sung during the procession to the font and back after Vespers on Easter Day - a special case. Polyphony during the office was, in any case, confined to feast days and then to a limited selection of items (e.g. at Vespers the responsory, hymn and Magnificat). There is no hint of festal psalm settings in Edward VI's reign (although it is also true that there are very few sources); they seem to be a product of the more musically tolerant reign of Elizabeth I. It would have been natural to have resorted, as Tallis did, to harmonising the plainsong tones traditionally used for chanting, but that doesn't necessarily point to an unbroken tradition. It is perfectly true that these settings evolved after the Restoration into the more rigid form of Anglican chant.

[Zimbelstern]

51 minutes ago, Vox Humana said:

Really? Tallis's "Canon" (in the truncated form found in modern hymns books) fits Nunc sancte nobis Spiritus well enough as well as other hymns in this metre.

Apart from the inappropriate nature and foursquare rhythm of the melody for such a hymn, how does that work with the second line of the first verse (Unum Patri cum Filio), given that the the first syllable of unum is accented, not the second?

[Vox Humana]

8 minutes ago, Zimbelstern said:

Apart from the inappropriate nature and foursquare rhythm of the melody for such a hymn, how does that work with the second line of the first verse (Unum Patri cum Filio), given that the the first syllable of unum is accented, not the second?

If I accede your point I would have to accept that the plainsong tune was accented differently in different verses (viz. Confessionem personent in verse 2), but is there any evidence for such treatement?

[Zimbelstern]

Again I have to say I am no expert in these matters, but the way I see it is this: plainsong has no regular metre, and so can accommodate different accent patterns (providing, of course, the singers know how to pronounce Latin correctly!). 

[Vox Humana]

But the point is that the text is metrical.* Let's turn this around. Many musicians argue that the absence of barlines in Medieval and Renaissance music, considered together with the micro-rhythms of the individual polyphonic lines, proves that these lines were played or sung not with regular stresses such as barlines impose, but in free rhythm that followed the music. The mensuration symbol (equivalent to our time signature) in which the piece was written governed the music only at the macro level. You, however, appear to agree with Peter le Huray, who points out somewhere in his book Music and the Reformation in England that performers would have felt the tactus and the microrhythms would have been felt against this regular pulse (an effect not that different from barlines). If the former view were correct then we wouldn't be debating whether or not Latin fits Tallis's Canon, because the stresses in the music would then follow the text as necessary, not the "time signature". (In fact Tallis's psalm tunes were printed without any mensuration symbols, but that's John Day for you). Le Huray was surely right. There is just no way anyone could have performed the complex proportions ('tuplets' in modern transcriptions) found in Morley's Introduction, John Baldwin's Commonplace Book and elsewhere without feeling the regular beat of tactus and prolation very firmly. Morley's book makes it very clear that the ability to maintain an unyielding, rock-steady tempo was crucial in performance. (His Christ's cross be my speed is so horrendously complex that I seriously wonder whether anyone has ever managed to perform it.) So, if we accept that the music was subject to a regular pulse, how do you account for the clear misaccentuation of "Discretor" in bar 3 and "Gloria" at bars 45-7 of Tallis's Jesu salvator saeculi? Similar examples can be found throughout the Tudor hymn repertoire - and the Tudors so routinely accented Alleluia on the second syllable in their responsories that I seriously wonder whether they didn't pronounce it that way.

* I find it interesting that Latin poetry in medieval services (hymns, sequences, rhymed offices) mostly obeys regular metres (Stabat mater dolorosa and Ave verum corpus natum are familiar examples) while English poetry of the Middle Ages is typically much more flexible in this respect.

[Zimbelstern]

I’m afraid I cannot debate with you in detail on this question because my knowledge and understanding are not great enough. My point was really in relation to Gregorian chant. Inasmuch as I have a basic understanding of the relationship between the rhythm and accents of Latin words and the melodies composed to accompany them, I stand in awe of the experts in the field, none perhaps more so than Dom Joseph Gajard of Solesmes. In his book “The Rhythm of Plainsong” he sets out the nature of the relationship between the pronunciation of Latin and the chant. This is (for me a least!) a highly complex subject, and I would not wish to try to summarise in a few sentences what he achieved so convincingly in such a short book. 

 

[Dafydd y Garreg Wen]

8 hours ago, Vox Humana said:

I'm not sure that's entirely correct. It's true that an anonymous Scottish treatise describes how to improvise a four-part faburden upon a plainsong and Thomas Morley in his A Plaine and Easie Introduction to Practicall Musicke provides harmonisations for the different tones, so the technique was certainly known. Nevertheless, so far as I know, there  was no tradition in England before the Reformation of singing the psalms of the office in polyphony, faburden or otherwise. The only psalms that were sung thus were Laudate pueri Doninum and In exitu Israel, when sung during the procession to the font and back after Vespers on Easter Day - a special case. Polyphony during the office was, in any case, confined to feast days and then to a limited selection of items (e.g. at Vespers the responsory, hymn and Magnificat). There is no hint of festal psalm settings in Edward VI's reign (although it is also true that there are very few sources); they seem to be a product of the more musically tolerant reign of Elizabeth I. It would have been natural to have resorted, as Tallis did, to harmonising the plainsong tones traditionally used for chanting, but that doesn't necessarily point to an unbroken tradition. It is perfectly true that these settings evolved after the Restoration into the more rigid form of Anglican chant.

It's hard to tell, isn't it?, given the lack of evidence.

At the most extreme harmonised chanting could have been a very late development and the post-Restoration examples of proto-Anglican chant (incorporating a psalm tone) almost a novelty. On that hypothesis psalms in English would have been chanted to unison psalm tones (leavened by the occasional festal psalm) for getting on for a hundred years before people began to harmoise them.

At the other a practice of informal harmonisation (I doubt it would have been literally improvised on the hoof!) could be pre-Reformation. You're right of course about the limited use of polyphony in the office, but I'm not sure that informally harmonised chant would count as polyphony and could easily escape note.

But the point is that the development of Anglican chant (early, late or somewhere in between) probably doesn't have much to do with the change from Latin to English.

[Dafydd y Garreg Wen]

51 minutes ago, Zimbelstern said:

I’m afraid I cannot debate with you in detail on this question because my knowledge and understanding are not great enough. My point was really in relation to Gregorian chant. Inasmuch as I have a basic understanding of the relationship between the rhythm and accents of Latin words and the melodies composed to accompany them, I stand in awe of the experts in the field, none perhaps more so than Dom Joseph Gajard of Solesmes. In his book “The Rhythm of Plainsong” he sets out the nature of the relationship between the pronunciation of Latin and the chant. This is (for me a least!) a highly complex subject, and I would not wish to try to summarise in a few sentences what he achieved so convincingly in such a short book. 

 

I think the confusion may arise because the nature of chant changed. Solesmes-style chanting aims to recover a relatively early style.

By the sixteenth century chant had become much more measured, and in the case of metrical texts (hymns etc.) the music itself may have morphed into something much more metrical. Tallis, for instance, turns Te lucis into triple time in his polyphonic settings. This may well reflect how the chant itself would have been sung. Alternatively it may have been sung in equal notes but to the same tactus as the polyphony. (At any rate, it wouldn't have been sung in the light flowing Solesmes style that one usually hears when this repertoire is performed alternatim - that's an anachronism.)

One penalty of this development would indeed be false accentuation in places, but people seem to have had a fair tolerance of this in singing English metrical texts (as we do to this day in hymns), so perhaps by this date they weren't that worried by it (however careful composers may have been about accentuation when setting prose texts polyphonically).

[Zimbelstern]

1 hour ago, Dafydd y Garreg Wen said:

I think the confusion may arise because the nature of chant changed. Solesmes-style chanting aims to recover a relatively early style.

By the sixteenth century chant had become much more measured, and in the case of metrical texts (hymns etc.) the music itself may have morphed into something much more metrical. Tallis, for instance, turns Te lucis into triple time in his polyphonic settings. This may well reflect how the chant itself would have been sung. Alternatively it may have been sung in equal notes but to the same tactus as the polyphony. (At any rate, it wouldn't have been sung in the light flowing Solesmes style that one usually hears when this repertoire is performed alternatim - that's an anachronism.)

One penalty of this development would indeed be false accentuation in places, but people seem to have had a fair tolerance of this in singing English metrical texts (as we do to this day in hymns), so perhaps by this date they weren't that worried by it (however careful composers may have been about accentuation when setting prose texts polyphonically).

Perhaps that’s why Pope John XXII declared in 1324:

”But certain practitioners of the new school, who think only of the laws of measured time, are composing new melodies of their own creation, with a new system of note values, that they prefer to the ancient, traditional music. The melodies of the Church are sung in semibreves and minims and with grace notes of repercussion. Some break up their melodies with hockets or rob them of their virility with discant, three-voice music, and motets, with a dangerous element produced by certain parts sung on text in the vernacular; all these abuses have brought into disrepute the basic melodies of the Antiphonal and Gradual. These composers, knowing nothing of the true foundation upon which they must build, are ignorant of the church modes, incapable of distinguishing between them, and cause great confusion. The great number of notes in their compositions conceals from us the plainchant melody, with its simple well-regulated rises and falls that indicate the character of the church mode. These musicians run without pausing. They intoxicate the ear without satisfying it; they dramatize the text with gestures; and, instead of promoting devotion, they prevent it by creating a sensuous and indecent atmosphere. . . . However, we do not intend to forbid the occasional use—principally on solemn feasts at Mass and at Divine Offi ce—of certain consonant intervals superposed upon the simple ecclesiastical chant, provided these harmonies are in the spirit and character of the melodies themselves, as, for instance, the consonance of the octave, the fifth, the fourth, and others of this nature; . . . for such consonances are pleasing to the ear and arouse devotion, and they prevent torpor among those who sing in honor of God.”

 

[John Furse]

I have followed this thread with interest and proffer my two-penn’orth. I posit that Morley’s “unyielding, rock-steady tempo” does not preclude a flexibility in the accentuation of syllables. The English language itself was still evolving at the time, was not standardised (particularly as regards spelling) and their pronunciation/s of it was not as now.

 

Thus, “Discretor” and, as Vox Humana says, countless other examples in the Tudor composers’ corpus, can be easily accommodated. This can be seen, joyfully, in Weelkes’ (albeit slightly later than Tudor) Alleluia, I heard a voice, where there can be multiple variations in the performance of the setting/s of the first word. (If that makes sense.) I’ve heard performances where this has been ‘ironed out’ - they sound somehow homogenised. Weelkes was a masterful exponent of ‘playing with words’; I’m sure he relished in this interplay of the differences.

 

Incredibly (in its original sense) to me, and as Vox Humana hints, there does seem to be no performance of Morley’s Christes crosse be my speede. I did a quick, then a more prolonged, search ? ! How can this be ? Please, someone, prove me wrong. I am more than tempted to revive my transcription tools and do a computer simulation.

 

Finally, I’m also struck by the last line of Zimbelstern’s quote from John XXII, where he promotes consonances  as a possible remedy for lassitude in the choir - what we might call ‘choral torportude’.  

[Vox Humana]

4 hours ago, Dafydd y Garreg Wen said:

It's hard to tell, isn't it?, given the lack of evidence.

At the most extreme harmonised chanting could have been a very late development and the post-Restoration examples of proto-Anglican chant (incorporating a psalm tone) almost a novelty.

There's plenty of evidence, just none for office psalms. Faburden (a form of harmonised chanting consisting at its simplest of chains of 6/3 chords with 5/8 chords at beginning and end) is at least as old as the second quarter of the fifteenth century and there are older pieces that are not so very different in concept. The many faburdens that have survived, whether as single voices or incorporated into polyphonic compositions, include Magnificats (which used the psalm tones), hymns (over 100 of these survive), Te Deum, processional items (Salve festa dies, litany refrains, the two psalms I mentioned above), antiphons, an offertory and a communion. It's always possible that new evidence will come along to upset received wisdom, but on the current information available I would have to say that the office psalms were not sung polyphonically until Tallis & Co began to write festal settings. For all that Edward VI was a staunch Protestant (if that word is not an anachronism at that time) he evidently did allow very elaborate music in his chapel. John Sheppard's  vernacular services could hardly have been written at any other time since he died almost at the same time as Mary Tudor. Recent research has shown how unwise it is to make assumptions about the chronology of Tallis's music, so I wouldn't entirely discount the possibility of Tallis's festal responses having been written for Edward's chapel, although they do seem to sit more comfortably in Elizabeth's. From this later example by Gibbons we can easily see how formalised Anglican chant later came about.

5 hours ago, Dafydd y Garreg Wen said:

At the other a practice of informal harmonisation (I doubt it would have been literally improvised on the hoof!) could be pre-Reformation. You're right of course about the limited use of polyphony in the office, but I'm not sure that informally harmonised chant would count as polyphony and could easily escape note.

The instructions that Scottish Anonymous gives for faburden do include rules for improvising it in four parts. I can't say I have internalised them, but there's an old article in Music and Letters that discusses them. In the Latin services all polyphony was a mark of ceremonial - although not officially sanctioned in the customary, it became to be regarded as a highly desirable tool in the same basket as incense, candles, silk copes and everything else that was brought out to add greater solemnity to the ceremonial of special occasions and feast days (and the more important the feast the more elaborate the ceremonial). The forms that exist in Faburden are exactly the forms for which composed polyphony also survives, so the distinction was just one of complexity.  Faburden would certainly have been viewed as polyphony.

4 hours ago, Zimbelstern said:

Perhaps that’s why Pope John XXII declared in 1324:

”But certain practitioners of the new school, who think only of the laws of measured time, are composing new melodies of their own creation, with a new system of note values, that they prefer to the ancient, traditional music...”

Flippin' clerics.

[Vox Humana]

5 hours ago, John Furse said:

I posit that Morley’s “unyielding, rock-steady tempo” does not preclude a flexibility in the accentuation of syllables.

I don't suppose that this is susceptible to proof either way. However, I find John Sheppard's hymns interesting. Rather like Morley and his endlessly inventive variations on the same "plainsong" phrase, so Sheppard seems to have had a bit of a thing about a point of imitation consisting of (in solfa) sol, sol sol, mi, ut, often flexed as sol, sol, sol, fa, re (along with several further variants thereof). Alas, few of these hymns are available on CPDL, but there are a few there (not all under the "Hymns" heading). Mostly the uploaded settings are the more interesting ones that avoid this figure, but Beata nobis gaudia uses it extensively. So, to a lesser extent does his Jesu salvator saeculi. From these and other similar treatments, one could certainly argue that what interested Sheppard was the musical phrase and that he didn't mind the occasional misaccentuation here. It's worth pondering because, as a general rule, it is quite obvious that Sheppard normally did go out of his way to accentuate words correctly so that strong syllables fell "on the beat". I might add that, so far as I know (and I have asked people who should know), no one has yet discovered any evidence of Tudor choirmen rehearsing and I'm not at all convinced that their standards were as high as we might like to think.