Monteverdi: onwards and downwards

Fresh from directing a performance of Monteverdi's 1610 Vespers in New York, I was a little surprised to read in Roger Bowers's article in last November's Early music that downward transposition by a 4th of its high-clef movements was supposedly no more than 'theoretically' possible.¹ While he argues that 'the properties of the music itself deny its applicability',² my own recent experience (with both Lauda Jerusalem and the Magnificat a7 down a 4th) had been of an ease and naturalness that seemed to render further theoretical justification wholly unnecessary.³ Bowers wholeheartedly accepts the need for downward transposition; it is the appropriate interval of transposition that he calls into question.⁴ Where I have long advocated downward transposition of a fourth (↓4th), Bowers now proposes a whole tone (↓2nd).

This turn of events-somewhat unexpected after two decades-has had the entirely healthy effect of sending me back to my original Early music article (hereinafter 1984),⁵ as well as to a pile of notes on subsequent findings. Rather than delay in order to accumulate yet more material (while readers mislay their copies of Bowers's article, forget the details of its contents or simply lose interest), I have chosen to respond swiftly and as succinctly as the subject permits, focusing on the Magnificat a7. The present contribution therefore reiterates very little of 1984 and should be regarded as supplementary to it. New material is marked with an asterisk (*).

Revisiting the subject in this way and with the benefit of others' more recent research into related matters of mode⁶ has left me with a much clearer understanding of certain critical points. As a consequence I am now distinctly more confident than before that the only plausible transposition for these high-clef movements really is ↓4th. Attractive though Bowers's suggestion may appear, I can discern no basis for it in the practice of Monteverdi's time. Even though I argue for its rejection, my hope is that the reasons for doing so will help clarify an issue which has ramifications far beyond this one publication and its composer.

Bowers's approach

Acknowledging that music variously written in high clefs and 'normal' clefs (= Bowers's 'low' clefs)⁷ represents 'not two distinct pitch-levels, but a single pitch-level diversely notated' (see table 1), Bowers observes that in Monteverdi's publication (hereinafter 1610) transposition ↓4th 'fails to accomplish a perfect or all-but-perfect reconciliation' of the two pitch-levels.⁸ His analysis of the vocal ranges aims to show that there is a differential of only a 2nd between the two configurations.⁹ Since transposition by this interval is said to 'receive some support from contemporary local theory',¹⁰ and since the resultant instrumental writing presents 'no problem that was actually insurmountable',¹¹ the intended 'solution' must be to transpose 1610's high-clef movements ↓2nd.

Table 1.

Voice parts in high and normal clefs

Bowers does not seek to introduce new historical information (theoretical or musical), nor does he choose to re-examine any source other than 1610. Instead he sets out to isolate 1610 from the available contemporary material, discounting an 'array' of comparative vocal ranges as 'exceptional' and dismissing a 'cornucopia' of evidence on transposition as inapplicable to 'North Italian practice on the Mantua-Venice axis' (exeunt all Romans, plus Praetorius, Schütz et al.).¹² Even the testimony of Monteverdi's own high-clef Laudate pueri a5 (1650)-unequivocally marked to be performed ↓4th-is disregarded.¹³ As a consequence, Bowers's case rests not on historical foundations of the usual sort, but on his own 'additional body of data', a new analysis of 1610's vocal (and instrumental) ranges. Indeed, it is no exaggeration to say that the case he presents depends utterly on the presumed absence of historical evidence 'relevant' to 1610.

Any conclusion based on such an approach naturally needs to be treated with extreme caution. We may begin therefore by asking whether there is precedent, first, for the specific transposition proposed (↓2nd) and, second, for the resultant 'key'. (Bowers provides instances of neither.)

Clefs and transposition down a 2nd

Fluency in transposing was cultivated by most professional instrumentalists.¹⁴ Such is the sheer abundance of evidence for the practice that we may wish to imagine there to have been little restriction on what could be transposed and by how much. Thus, in support of his proposal, Bowers merely cites my earlier 'copious' examples, and reports that 'Italian theorists of the late 16th and early 17th centuries ... specifically mention downward transposition by one tone.'¹⁵

But matters are not quite so simple. High clefs and low clefs were in fact each associated with their own set of transpositions, as Virgiliano (c.1590) shows perhaps most clearly;¹⁶ see illus.1 and table 2, which indicates how cornettists and trombonists may transpose.

Here, as with Virgiliano's other instruments, transposition ↓2nd does not occur with high clefs, and is instead linked exclusively to normal-clef writing, just as it is in Diruta's explicit instructions (1609) for organists.¹⁷ Equally, both writers specifically associate music written in high clefs (usually with one flat) with the wider transposition ↓4th. Zacconi (1592) expresses it thus:

you can generally play the musical Tones given in their natural positions [i.e. in normal clefs] a 2nd lower and [those in high clefs] a 4th or 5th [lower] ...¹⁸

This is not abstract theory but practical information for instrumentalists. It therefore comes as no surprise to find these principles confirmed in the musical sources themselves. One of the very earliest publications to include an organ bass-and therefore one of the very earliest sources in which one might hope to find evidence of such transposition-is *Croce (1594), a collection of motets a8 'comodi per le voci, e per cantar con ogni stromento' by the then vicemaestro di cappella of S. Marco in Venice.¹⁹ All but two of its 17 items are supplied with instructions for transposition, mostly as a pair of options such as Alla quarta, e quinta bassa (↓4th & ↓5th); see table 3.

Table 2.

Transpositions for wind players, after Virgiliano

Table 3.

Transpositions in Giovanni Croce, Motetti (1594)

1.

Aurelio Virgiliano, Il dolcimelo (c.1590), p.[105]: 'Modi tutti da sonar il cornetto'

The two main points to observe here are that transposition ↓2nd again occurs only with normal clefs, and that high-clef writing in each case goes ↓4th at least.²⁰ The additional flexibility of Croce's collection does not in any way conflict with these broader principles. As Agazzari (1607) puts it:

one must see which [transposition] is most appropriate and suitable to the given Tone, not as some do who pretend to play every Tone at every level, for if I could argue at length, I would show them their impropriety and error.²¹

[My italics]

In short, while the general practice of transposing is handsomely documented, the specific transposition which Bowers proposes-of high-clef music ↓2nd-is conspicuously absent not only from Zacconi, Croce, Virgiliano and Diruta, but from all the relevant contemporary sources I have so far encountered.²²

If the Magnificat a7 is to be transposed ↓2nd from its written level, it moves-in instrumental terms-from G minor into an outlandish F minor (with a signature of three flats but plenty of additional D♭s). This is barely mentioned by Bowers.²³ If there is any string or wind music from this period- yet alone any comparably sophisticated and virtuoso writing-specifically intended for this remote and problematic key, I have yet to see it. (It does not appear even amongst Diruta's 23 written-out examples of organ transposition.)²⁴ On the other hand, in D minor (i.e. ↓4th) Monteverdi's writing for violins and cornetts not only lies completely naturally and idiomatically under the fingers but emerges as a strikingly exact counterpart in range, tessitura and character to passages in his Orfeo (performed 1607, published 1609); see ex.1, which also shows some parallels of vocal writing.²⁵ Bowers offers no comment on this phenomenon.

Experienced organists and others probably did find themselves in F minor from time to time. If so, they are more likely to have arrived there by a different route from the one Bowers proposes, perhaps by taking a normal-clef G minor piece ↓2nd.²⁶ And, just as we are more accustomed to transposing hymns or simple anthems than intricate cantatas, it is conservative vocal polyphony rather than elaborate concerted music that is most likely to have attracted such transposition. For any composer of Monteverdi's day to write a virtuoso vocal and instrumental showpiece specifically intended to be played in F minor (whatever its notated level) would presuppose some very special circumstance. To write such a work without any apparent concession to the particular nature of this highly unusual key would also be quite remarkable.²⁷ And, not least, to do so without including any instruction to alert performers to the need for a wholly exceptional form of transposition-un tuono più basso from high-clef notation-would be to court disaster, causing performers instead to adopt a conventional one.

Transposition down a 4th

According to Bowers there is only one unequivocal source for the degree of transposition appropriate to high clefs: namely, Praetorius (1619). A key passage begins thus:

Every vocal piece in high clefs ... must be transposed ... as follows: if it has a flat, down a 4th ...²⁸

This major source is then rejected as 'peripheral' to Monteverdi's practice on the grounds that 'Praetorius was no Italian': only the music and writings of Northern Italy should be taken into consideration- and these are said to be silent on the matter. 'If ever it turns out that some writing of Michael Praetorius happens to match and coincide with this prime evidence, then well and good.'²⁹

By these means Bowers argues himself into a position where the only evidence that counts is his own analysis of 1610's vocal ranges. The vast body of information enshrined in the works of Heinrich Schütz, for example, is therefore disregarded in its entirety³⁰-despite the composer's five or so years of study in Italy (mostly in Venice), first with Giovanni Gabrieli (whose high-clef Lieto godea he reworked, ↓5th)³¹ and later perhaps with Monteverdi himself.³² Ruled out similarly for geographical reasons is the testimony of Roman publications such as G. F. Anerio's (1613), with its 37 high-clef pieces, each and every one bearing an instruction for transposition either ↓4th or ↓5th.³³ But is modal theory, which lies at the heart of transposition practice,³⁴ really so particular about geography?³⁵ Roman practice demonstrably accords with that of Germany, for example.³⁶ Bowers (beyond his analysis of 1610's vocal ranges)

Ex1.

Passages from Orfeo (a, c, e) and the 1640 Gloria a7 (g), compared with 1610, Magnificat a7 (↓4th) (b, d, f, h)

offers no evidence to suggest that Northern Italy followed a separate path.

Fortunately the matter need not remain purely conjectural, as 'prime' evidence from Northern Italy turns out not to be lacking. Asked by his discepolo about high clefs, the maestro in *Banchieri (Bologna, 1601), for example, answers in terms that correspond exactly to those of Praetorius (1619):

when the G2 clef has a flat, then ... the notes are to be taken a 4th lower.³⁷

(From Naples *Picerli (1631) says the very same thing: 'when the composition uses ... the G2 clef, it is to be transposed, mentally or in writing, down a 4th or 5th ... with B♭ it is to be transposed down a 4th'.)³⁸ And for two examples of comparable evidence from the musical sources we may turn first to the instructions given in *Croce (1594), where- as we have seen above-all five high-clef pieces are marked to go ↓4th at least. (Even if Croce's practice here is seen as somewhat more flexible than Banchieri's and Praetorius's theory, there is no disguising the fact that it also flatly contradicts Bowers's proposal.) A more orthodox example is found in *Palestrina (1608), a Venetian reissue of his fourth book of motets a5 (1584): complete with organ continuo, all 23 original high-clef items have here been transposed and renotated in normal clefs, ↓4th in the case of each of the 15 motets with a signature of one flat.³⁹

The 'reconciliation of inconsistent clef-systems'

In Bowers's view, when Monteverdi's high-clef movements are set ↓4th, the 'tessituras of all the voices emerge as gratingly low compared with those of the generality of the rest of the movements'.⁴⁰ With this we come to the heart of the matter. It is this perception-shared perhaps by many-that generates the search for 'an alternative solution'. But what degree of uniformity is it reasonable to expect between one piece and the next?

Monteverdi's publication is a dazzling catalogue of styles and techniques, and the individual characteristics of its component parts naturally give rise to-and result from-differing tessituras. Thus, when the normal-clef movements are compared with each other, we find that some lie significantly higher than others (compare Laetatus sum and Ave maris stella),⁴¹ while the tessituras of individual voice parts can vary by a good minor 3rd; see ex.2. To think purely in terms of composite tessituras-the aggregates of several movements-is therefore potentially misleading. And while 1610's high-clef writing (when .4th) may well end up lower than the 'generality' of its normal-clef pieces, the Magnificat a7 nevertheless does align revealingly well with two other individual items, namely Audi column and Ave maris stella; see ex.3. If, even after transposition ↓4th, the Magnificat a7 proves so remarkably similar in tessitura to these two movements, are they, too, to be considered implausibly low? (And, if so, are we to

Ex2.

Divergent ranges within 1610

start transposing all such movements at will in order to bring them in line with a supposed norm?) Or is it merely the Magnificat's wider-and lower-bass range that puts it beyond the pale?

Before looking at vocal scoring in general, and at these bass parts in particular, we may query the underlying assumption of Bowers's article, that 'the perfect or all-but-perfect reconciliation of pitch-levels ... is the objective sought by the application of some degree of transposition'.⁴² Are we really justified in demanding that high and normal clefs always yield virtually identical sounding ranges?

Clefs and clef-combinations served, in part, as compositional tools to regulate vocal (and instrumental) ranges. Having chosen one or other set of clefs, a composer was assured that workable ranges for each voice (and appropriate relative ranges for the voices as a whole) could be achieved through the simple expedient of not writing above or below a stave; any need for ledger lines acted as a warning that normal modest limits were being exceeded. Individual and overall ranges from Monteverdi's time are therefore much more intimately related to clefs than in most later repertories. As ex.4 shows, if two hypothetical compositions are compared-one in high clefs, the other in normal clefs, each employing the full extent of its staves-we find that to transpose the former ↓4th necessarily produces a lower tessitura (by one degree) than that of the latter, for the simple reason that the two systems themselves do not lie a 4th apart but only a 3rd.⁴³ It follows that conventional transposition ↓4th is no guarantee of an exact alignment of ranges and that the sort of

Ex3.

Voice ranges in 1610 compared: (a) Audi coelum; (b) Ave maris stella; (c) Magnificat a7 (↓4th)

Ex4.

In (a) º = written pitch; in (c) these pitches are shown transposed ↓4th following the normal-clef range of each voice.

'mismatch' (of roughly a tone) that may seem puzzling in 1610 is almost intrinsic to the process.⁴⁴ In short, high clefs tend quite naturally-if somewhat paradoxically-to yield lower tessituras than normal clefs.

Core vocal ranges

'The starting-point for a fresh evaluation of the evidence', writes Bowers, is the 'identification of the core ranges' (as opposed to total ranges) of 1610's vocal writing; this 'can usefully be undertaken in respect only of the movements composed for chorus or ensemble of non-specialist voices'.⁴⁵

In the context of the Vespers music this use of the word 'chorus' is not intended (I trust) to carry any necessary implication of more than one singer per part. While the idiom of 1610's Mass does suggest such an ensemble, the remainder of the music gives not the slightest indication that Monteverdi had anything other than solo voices in mind.⁴⁶ What is more striking, though, is the notion that 'non-specialist voices' ('standard voices') might have had a role in courtly concerted music-making of this sort. Certainly some sections demand less virtuosity than others (just as some lie higher than others), but there are no serious grounds for imagining that Monteverdi is writing for two distinct groups of singers, 'virtuoso voices' and a 'non-virtuoso ensemble'.

Unfortunately, this spurious distinction seriously undermines the value of Bowers's analysis: in order to achieve 'the most secure conclusions', all virtuoso writing is excluded from consideration. The reader, however, is left in considerable doubt as to what remains in and what is out; mention is made only of 'the solo motets (including the first section of Audi coelum) and certain Magnificat sections'.⁴⁷ Are the duetting tenors in Laudate pueri ('Excelsus super omnes gentes dominus') and all the cascading voices of Laetatus sum ('Illuc enim') therefore presumed to be 'non-specialist'? And if the tenor's solo opening of Audi coelum has been discounted, why does its single fleeting top f' remain to define the upper limit of tenor 'core range' for that movement?

Core range, rather than 'the rare liberties represented by the extremes of range', is justifiably considered to be the critical indicator of a composer's understanding of voice-types,⁴⁸ but Bowers's analysis contains noticeable inconsistencies. With Dixit Dominus a unique and quite short f#' (in S I only) is allowed to represent the upper end of soprano core range; similarly, a single (semiquaver) bass high c' is retained for Laetatus sum.⁴⁹ As it happens, the effect of both oversights, and of that in Audi coelum, is to create the misleading impression of a somewhat higher tessitura than is actually the case. (All three are normal-clef movements.)

With the Magnificat a7 the picture is more seriously distorted. If its 'non-virtuoso ensemble movements' have been omitted from consideration, we need to know which they are; are we left with seven of the 12 sections, or just three?⁵⁰ Clearly, the duet-ting basses of 'Quia fecit' have been ruled out; the conspicuous absence of their repeated high written f's (↓4th = c' ) at 'fecit mihi magna' causes an entire minor 3rd to be chopped off the top of the movement's bass core range. Here, then (in this high-clef movement), the impression is given of a lower tessitura than is actually the case.

Vocal scoring and ranges

Transposition was not a free-for-all, nor was vocal scoring. Uncertainty about pitch standards and (as we have seen) a lack of familiarity with the practical implications of the different clef-systems have conspired to cloud the issue for us. But another factor complicates matters further: the simple assumption that Monteverdi's vocal categories equate directly with our own.

Instead of needing to scrutinize each vocal line in the time-consuming way that both Bowers and I have done, musicians of Monteverdi's time understood that clefs were a succinct indicator both of core ranges and of vocal scoring. In each system a particular clef identifies a specific voice-type - soprano, alto, tenor, bass (as in table 1)-and also broadly prescribes its potential range, which is roughly that of the stave itself (as in ex.4). (As the distances between clefs also show the relationships of the various parts, we may note that alto and tenor lie closer to each other than other voices.)

*Banchieri (1614) is rather more specific about vocal ranges: for each voice he gives both a (core) range of a 10th-for the Cantore perfetto-(see ex.5, range (a)) and one of a 12th-for the Cantore perfettissimo- extending a degree further in each direction (range (b)).⁵¹ Naturally enough, it is in solo writing that we can most often expect to encounter these wider ranges. Particularly convenient for comparison with Banchieri are the Cento concerti ecclesiastici by *Viadana (1602), a collection which opens with short solo items for each voice-type in turn (ten apiece, plus falsobordoni).⁵² While Viadana's soprano and alto conform fairly closely to Banchieri's wider limits, his tenor reaches down as far as 'bass' F in two separate items (albeit briefly) and his bass extends both higher and lower (touching low D on two separate occasions);⁵³ see ex.5, range (c). Not surprisingly these composite ranges from Viadana turn out to be wider for each voice than for any individual Vespers item in 1610, including the Magnificat a7. Indeed, both tenor and bass possess what Bowers describes as 'gigantic overall ranges of two complete octaves', something he regards as intrinsically 'neither very practical nor very probable' for 1610 as a whole.⁵⁴ Although Monteverdi calls more frequently for bass low D (with the movement set ↓4th), all his ranges lie comfortably within those of Viadana; see range (d). By contrast, the effect of Bowers's proposed narrower transposition ↓2nd- range (e)-is to nudge each of Viadana's (already wide) ranges discernibly upwards, consistently creating new upper limits and scarcely approaching the lower ones.

From today's perspective normal-clef music from the late 16th and early 17th centuries may generally appear to lie rather low for voices., and anything lower is certainly liable to be regarded with scepticism. ⁵⁵ Clearly, differences in pitch standard need to be taken into account (see below), but the matter is further complicated by a natural tendency to assume that the terms 'soprano, alto, tenor, bass' carry more or less the same connotations as they do for us today. In particular, while the alto in church music of

Ex5.

Voice ranges compared: (a) Banchieri (Cantore perfetto); (b) Banchieri (Cantore perfettissimo); (c) Viadana, Cento concerti ecclesiastici; (d) Monteverdi, 1610, Magnificat a7 (↓4th); (e) Monteverdi, 1610, Magnificat a7 (↓2nd); (f) Palestrina, Vergine saggia (º = written note, · = sounding note, after Doni (1640))

Monteverdi's time is commonly taken to have been either a falsettist or a castrato (as in some later traditions), the weight of evidence points instead to a voice that we would call a (high) tenor.⁵⁶ (Bowers, who probably has the falsetto voice in mind, notes that 'for some reason as yet undetermined' Monteverdi's normal-clef alto parts are relatively narrow, with a usual upper limit of a' ;⁵⁷ while their close proximity to tenor parts removes the need for much use of the lower range, it is easy enough to see why the non-falsetto alto's upper range would be restricted in this way.) Falsettists and castratos were both still associated with soprano (rather than alto) parts, while boys who could hold their own in elaborate concerted music were probably a rarity.⁵⁸ The remaining male voices were divided not into two categories (tenor and bass) but into three (high, medium/high and low), with the range of the middle voice perhaps best described as that of a baritonal tenor.⁵⁹ This understanding of early 17th-century practice in Italy is compared in table 4 with the more familiar distribution of voices in two types of present-day choir-the mixed-voice choir and the Anglican cathedral or collegiate choir.

The implications of this are considerable. As an illustration we may take Palestrina's setting of Petrarch's Vergine saggia; see ex.5, range (f). Apart from a rather high-lying bass part the madrigal's written ranges conform tolerably well to modern expectations of SATB scoring, such that today's editors would probably be loath to transpose it more than, say, ↓2nd. (And at such a level it would neatly match Bowers's proposed transposition of the 1610 Magnificat a7; see range (e).) However, *Giovanni Battista Doni (1640) is adamant not only that Palestrina's madrigal should be transposed-'because it is notated high, as they say, and would prove too uncomfortable if it were sung as it stands'-but that the 'most usual type of transposition' (in this instance) is of altogether another order: namely, ↓5th.⁶⁰ Our present SATB paradigms, it seems, are very unreliable guides to early 17th-century vocal scoring.

Basses and pitch standards

A specific effect-or aim-of Bowers's proposal (ex.5, range (e)) is to eliminate the bass low Ds that result from transposition ↓4th (range (d)). These notes, which many still find perplexing, are all doubled by organ and are almost wholly confined to the Bassus part (B I) of 'Et misericordia', a slow-moving dialogue between three upper and three lower voices.⁶¹ (It may be helpful to remind readers that the music under consideration here is concerted music and thus primarily, if not exclusively, the preserve of select solo voices; see n.46.) In 1984 I pointed to no fewer than ten other sacred works by Monteverdi himself that unambiguously call for this note.⁶²Ex.6 reproduces a telling extract from his 1640 Gloria a7 ('pax hominibus'), where the seven voices, each in its lowest register, combine to produce a dark, sober texture unmistakably similar to that of 'Et misericordia' (↓4th), complete with low Ds⁶³-and thus, despite the composer's full authority, 'gratingly low' in Bowers's terms.

In questioning the significance of these and other instances of low vocal ranges, Bowers suggests that 'each represents not conventional but exceptional practice'.⁶⁴ (Interestingly, the very opposite argument is employed in defence of some unprecedented instrumental writing: the instrumental demands of the Magnificat a7 are, he claims, 'not routine but genuinely exceptional'.)⁶⁵ Let me nevertheless briefly offer two further instances of low bass writing

Table 4.

Varying understandings of SATB

Ex6.

Monteverdi, Gloria a7, 'pax hominibus'

in Venetian publications. In *Ercole Porta's Missa secundi toni (1620) the Kyrie, Gloria and Credo each require the bass voice, in their very opening bars, to descend to low D ('elei-son', 'Et in terra pax', 'et ter-rae'),⁶⁶ while in *Giovanni Valentini's Missae quatuor (1621) two specified transpositions for optional organ (↓4th and ↓5th) take the bass part respectively to low D (Missa Susanna) and, further still, to C (Missa Stabat mater).⁶⁷

Little is known, sadly, of Mantua's bass singers in the early 1600s, but some decades later *Doni (1640) commented that Monteverdi had had 'certain very low basses' at his disposal there.⁶⁸ Moreover, in 1614, shortly after the composer's departure for Venice, *Ferdinando Gonzaga's agent in Rome reported having found there un Basso per camera, a new bass singer for the court's chamber singing: 'va basso sino al ottava di D sol re' ('he goes down to low D').⁶⁹

Exactly how low might Monteverdi have expected this D to sound? The question has been deferred until this point because Bowers's subject-and mine-is the relation in pitch between high- and low-clef movements in 1610, a matter which is indeed 'preliminary to, and entirely independent of, identification' of an appropriate pitch standard.⁷⁰ Suffice it to say that all salient information on Italian pitch standards for the period has helpfully been brought together by Bruce Haynes⁷¹ and that the two strongest candidates in the case of 1610 appear to be, in approximate terms, a' = 440 and one semitone above a' = 440. (For what it is worth, the 1565 Graziadio Antegnati organ in Mantua's court chapel of Santa Barbara has been restored in recent years to its original condition, reportedly at the higher pitch.)⁷² Having performed the Vespers music at both pitch levels, but always with the same principles of vocal scoring (see above), I merely offer two observations: that, while on balance the higher pitch is probably somewhat easier for most voices, the overall width of ranges has never proved problematic, and that the relative pitch levels of high- and normal-clef movements (with the former ↓4th) have always felt utterly convincing.

Instruments

If I have scarcely touched on the obbligato instrumental writing in the Magnificat a7, this is primarily because I have dealt with the matter at length elsewhere. ⁷³ But it is also the least persuasive part of Bowers's argument: his suggestion that the very real problems caused by transposition ↓2nd can be solved by 'the simple engagement of instruments of a size identified as capable of realizing the sounding of these particular parts transposed down by one tone' side-steps too many issues and is in any case far from 'simple'.⁷⁴ And, while the very basis of his analysis of ranges is that 1610 is written for 'a single unchanging vocal ensemble' (of 'non-specialist' singers?),⁷⁵ the reader is required to imagine that by contrast the corresponding instrumental ensemble has one set of instruments for Domine ad adiuvandum, Dixit Dominus, Ave maris stella and the Sonata sopra Sancta Maria but a second set especially for the Magnificat a7. (This would certainly put a damper on another-and, in my view, more plausible- hypothesis, that the Sonata, despite its printed position in 1610, was intended to function as an antiphon substitute at the Magnificat's conclusion;⁷⁶ with half a dozen or so virtuoso players needing to switch instrument, it would be hard to avoid either an unseemly delay or an unholy scramble.)

Almost as if seeking refuge from these imponderables, Bowers proceeds to justify the presence in 1610 of the second Magnificat (a6)-a related work without obbligato instruments-on the grounds that 'for many ensembles, and perhaps for most,' the more elaborate setting was 'simply out of reach' on account of Monteverdi's 'unconventional' instrumental demands.⁷⁷ A more satisfactory explanation of 'the ostensible enigma' of the two Magnificats⁷⁸ is, in my view, that the composer's intention was the wholly practical one of providing a complete Vespers that could be performed either with or without instrumental ensemble: instead of the grand toccata-like setting, the usual simple chant could be reinstated for the opening Response, the ritornellos in Dixit Dominus and Ave maris stella could be omitted (as sanctioned by the composer himself in the case of the former),⁷⁹ and the liturgically inessential Sonata could safely be dropped altogether.

Neither the flaunting of extreme (if 'playable') top notes nor the mere avoidance of 'insurmountable' problems are hallmarks of Monteverdi's other known instrumental writing, not least the virtuoso Sonata sopra Sancta Maria.⁸⁰ Yet to perform the Magnificat a7 ↓4th would, Bowers suggests, reduce 'the capacities required of the players of violin and cornett to those of entirely conventional performers'.⁸¹ Given its striking similarities (at this lower pitch) with passages for the same instruments in Orfeo (see ex.1), the claim would also require us to imagine that Monteverdi had been content to supply his Orpheus with undemanding writing for 'entirely conventional' instrumentalists at the very moment (in Act 3) when music's powers are being tested to the full. The wholesale transfer of idiomatic figuration to a higher and distinctly less grateful key scarcely seems how a composer of Monteverdi's subtlety would choose to demonstrate 'the brilliance of his artistic imagination'.⁸²

Despite its modest role, the third cornett part in the Magnificat a7, which lies exactly a 5th beneath the others, has attracted attention on account of what Bowers calls its 'anomalous low pitches' (↓4th).⁸³ While the part can be played on a treble instrument,⁸⁴ a low range of this sort may well be unexpected. However, a glance at the double-choir 'echo' canzona a10 in *G. Gabrieli (1597) demonstrates that our expectations may need some revision. ⁸⁵ Here, no fewer than eight cornetts are seen at work-four in each choir, with four different clefs (G2-C1-C2-C3) and four distinct ranges, the top one playing a full octave higher than the lowest (but-we may note-only up to a" )⁸⁶. All are simply labelled 'cornetto', as are Monteverdi's, although it seems very unlikely that all are intended for treble cornetts. An *inventory from Stuttgart (1589) lists a cornett 'two tones lower than the treble cornett',⁸⁷ and among the musical details of a *Florentine wedding celebration (1608) is reference to a 'Cornetto Torto p[er] un Contralto', apparently a tenor instrument playing an alto part.⁸⁸ But perhaps even more important than the size of the instrument is the player's embouchure and ability to play in the particular register (see n.84). Certain players undoubtedly specialized in inner/lower parts, and Bologna's municipal wind ensemble (the Concerto Palatino), for example, included a position for contralto di cornetto. ⁸⁹ An intriguing detail of 1610's Sextus part-book may support the idea that Monteverdi was thinking along similar lines. At 'Quia respexit' the cornett III part shares a stave with trombone I, making it entirely feasible in this verse for one and the same player first to play cornett (of one size or another), then tenor sackbut, and finally cornett again, using the intervening woodwind duets to switch instruments.

From around the time of 1610's composition and compilation there is overwhelming evidence not only that vocal music notated in high clefs was generally understood by Italian musicians to demand downward transposition, but also that the expected interval of transposition was most frequently a 4th and sometimes a 5th.⁹⁰ Thus, when high-clef madrigals by Monteverdi turn up in a collection of spiritual contrafacta by *Aquilino Coppini (1607), it is no surprise to find that-without comment-the Partito (score) sets two of them ↓4th (and not less) from the pitch of the vocal part-books. ⁹¹ Only later in the century did a narrower downward transposition (↓3rd) begin to be associated with high-clef pieces.⁹² By contrast, transposition .2nd was seen not as a primary transposition required by the notation and essential for correct vocal (and instrumental) scoring, but as a secondary one which, at the discretion of the performer, could in some circumstances effect a helpful small adjustment (where, for example, an organ's pitch was unusually high). For all its originality, Monteverdi's Magnificat a7 possesses no exceptional features that would have obliged, encouraged or even allowed contemporary performers to override these fundamental principles.

What of 1610's three other high-clef movements? Despite its lack of obbligato instruments, the less well-known Magnificat a6 is in notational terms identical with its sibling a7; consequently, the same conventional transposition is almost certainly to be understood. In both the Mass and Lauda Jerusalem, however, the absence of a flat signature would, according to theory, indicate a wider transposition (↓5th ),⁹³ though in practice this particular rule was increasingly ignored in favour of transposition ↓4th (as is seen, for example, in much of the Vesperi by the Milanese composer *Serafino Cantone, 1602).⁹⁴ In Monteverdi's case, and especially for the Mass, the voice ranges suggest that performers of the composer's time are most likely to have taken the higher option-as does the later Brescian organ score of the Mass.⁹⁵

Innumerable untransposed performances over the past 70 years have doubtless helped establish the view that 'extrovert brilliance' constitutes the 'particular characteristics' of the Magnificat a7.⁹⁶ The perceived threat to these supposedly intrinsic characteristics is perhaps the biggest obstacle to any objective re-evaluation of the question of its transposition.⁹⁷ We therefore do well to remind ourselves that specific expectations of this sort derive in most cases not from acquaintance with the notated music or with the performance conventions on which the composer's notation depended, but from the infinitely more potent experience of hearing it (perhaps repeatedly) as performed by musicians of our own time. But this questionable familiarity is by no means the only barrier to understanding Monteverdi's Magnificat a7 more fully: while the niceties of 16th- and 17th-century transposition practice are simply too remote from current modes of thinking to attract widespread attention,⁹⁸ critical differences between earlier and current conventions of vocal scoring have scarcely begun to be acknowledged. Small wonder then if the idea of transposing the movement downward (especially by as much as a 4th) appears, to some, 'still controversial'.⁹⁹

The present discussion will, I hope, contribute to a somewhat clearer understanding of high-clef notation. If so, its usefulness need not be thought to be confined to a single composition (the Magnificat a7), to a single collection (1610) or to a single composer's output (Monteverdi's): vast quantities of high-clef Renaissance and early Baroque vocal music-a good two-thirds of Palestrina's output, for example¹⁰⁰-await re-exploration in fuller awareness of the notation's practical significance.