Ornamentation in Dowland's lute music

Martin Shepherd

You should have some rules for the sweet relishes and shakes if they could be expressed here, as they are on the Lute: but seeing they cannot by speach or writing be expressed, thou wert best to imitate some cunning player, or get them by thine owne practise, onely take heed, least in making too many shakes thou hinder the perfection of the Notes. In somme, if you affect biting sounds, as some men call them, which may very well be used, yet use them not in your running, and use them not at all but when you judge them decent. (Jean-Baptiste Besard, 1603, as translated by John Dowland, 1610)

This is a quick guide to the subject, with many expressions of opinion and few backup references - all the background information can be found in my article "The interpretation of signs for graces in English lute music", The Lute, XXXVI, 37-84. (1996).

Ornaments in lute music are of two sorts: divisions and graces. Divisions are the extra notes added in the repeat of a strain, for example, where long notes have been replaced by running passages. These may be written out, or improvised, but they are a common feature of English music of this period. Graces are notes added by the action of the left hand only. Usually the extra notes are not shown in the tablature at all, or are indicated only by a sign next to the main note which is to be graced. These are the hammer-ons, pull-offs, trills and so on which are native to plucked instruments. The following notes are only about graces.

Modern performances of this music often omit graces altogether, especially if the source (like Varietie of Lute Lessons) is a printed source where there are no signs for graces. We are so accustomed to hearing this music played without graces that we may take a while to feel that graces are "correct" - as with other aspects of historically informed performance, a bit of acclimatization might be required. There is little doubt, however, that Dowland and his contemporaries played many graces - manuscripts like the Margaret Board lute book, and the Folger lute book, in both of which Dowland wrote, contain numerous signs for graces. The fact that some other manuscripts contain no signs almost certainly indicates that the conventions of how to apply graces were widely understood but their actual application was a matter of improvisation and personal taste. With printed books the addition of these signs would considerably complicate the job of the printer.

Most modern authors have concluded that there is little logic or consistency in the use of these signs: different copies of the same piece have different ornaments, different scribes use different signs to mean the same thing (Diana Poulton even claimed that Dowland's system was different from everyone else's). But this is only true to a limited degree: in fact there is considerable agreement among sources, as we would expect if there was a conventional interpretation for these signs which was widely understood. It is certainly not the case that "anything goes". Our problem is that we have only limited information about how the signs were interpreted. From the information we do have, I think it is possible to apply a consistent interpretation, which is explained in as few words as possible below:

English manuscripts typically contain two signs, # meaning a "shake", and + meaning a "fall". Sometimes a dot is used instead of one or other of these signs, but it is clear from context which is which.

A shake starts on the main note, moves to the auxiliary note (which can be above or below the main note) and back to the main note. It can be repeated any number of times. So the following are all shakes:

Typical contexts: shakes can occur more or less anywhere but are particularly common on the third of a chord.

A fall starts on the auxiliary note (which is usually, but not always, below the main note) and moves to the main note. If the main note is on the third fret or higher, the fall may start two notes below the main note (Mace called this a "whole fall", as opposed to a "half fall"). In some musical and fingering contexts, the auxiliary note may be the upper note (called a "backfall"). So the following are all falls:

Typical contexts: falls usually occur on the tonic note, quite commonly in the bass. When they occur on the third fret or higher, this is frequently on the (usually minor) third of a chord.

The two signs occasionally occur together on the same note, in which case there are various possible interpretations. It could mean a "fall, shaked" where the shake sign indicates repetition of the fall, or a fall starting on the lower note followed by a shake on the upper note or vice-versa.

From this summary it can be seen that there are many options open to the performer - how many repetitions in a shake, whether it uses the upper or lower note, whether a fall should be a half fall or a whole fall, etc. In making these decisions, considerations of musical context and fingering possibilities loom large. As to whether to use the upper or lower note for a shake, a rule of thumb (given in various keyboard treatises) is that in stepwise passages, the grace note should not anticipate the next main note - so in an ascending scale use the lower note, in a descending scale use the upper note:

Such rules will always have exceptions, especially when left-hand fingering possibilities are taken into account.

Another consideration is whether the shake should be "hard" or "soft". Mace (1676) describes both kinds: the hard shake is one where the shaking finger plucks the string by pulling it sideways whereas in the soft shake the finger lifts up, sounding the lower note more softly. Lest we should dismiss this as irrelevant to Dowland because Mace was writing 50 years after Dowland's death, we find the same distinction in Robinson (1603) who speaks of "either a strong relysh for loudnesse, or a milde relysh for passionate attencion".

Finally, we should consider vibrato. Vibrato as a grace on single notes (as opposed to a general habit) is clearly described in a number of treatises and is clearly applicable to Dowland's music. Vallet uses a special sign for it in one of his books (1620). Mersenne, writing in 1636, says that vibrato was overused in the past (in the early 1600s?) but it was equally wrong to underuse it. Since vibrato is more effective on notes on high frets, this is where we would expect to find it, and this is true of Vallet's music. As for Dowland, two pieces in the Margaret Board lute book (Lachremæ on f.11v. and, in Dowland's own hand, the Almande by Robert Dowland on f.12v.) have a sign, a cluster of little dots, which I believe is a sign for vibrato:


Examples of gracing by Dowland

In the Margaret Board lute book on f.12v. Dowland wrote out an Almande by his son Robert, complete with signs for graces, which I have interpreted following the principles outlined above. The ++ sign I interpret as an indication for a whole fall, as opposed to + meaning a half fall. Note that several of the falls must be backfalls, starting on the note above, because they are on an open string - others I have interpreted as either forefalls or backfalls, depending on context. This piece is the most direct evidence we have about the way Dowland used graces. On the next page Margaret Board has written an Almand by John Dowland with similar patterns of graces.

Mp3 Soundfiles:

Almande by Robert Dowland

Almand by John Dowland

PDF Scores:

Almande by Robert Dowland

Almand by John Dowland

I leave the final word to Thomas Mace:

For you must know this, that whatever your grace be, you must, in your farewell, express the true note perfectly, or else your pretended grace, will prove a disgrace. ( Musick's Monument, 1676, p.105)

Martin Shepherd, July, 2008