New observations on the quilisma: occurrence, position and function in northern Low Countries sources

Abstract

The quilisma is a sign found in the earliest surviving notations of Gregorian chant. Since the ‘chant restoration’ of the nineteenth and twentieth centuries, the sign is included once again in all chant books, and poses an inevitable question of interpretation to those studying and performing chant. However, since medieval times, there has been considerable debate whether the sign denotes a particular method of voice production, a particular rhythmic value and/or an element in a melody's modal orientation. Chant manuscripts of the northern Low Countries (the northern half of the modern-day Netherlands), although not well known for their melodic content, point to highly interesting developments in response to a continually changing musical aesthetic. Likewise, these manuscripts offer new insights into the quality of the quilisma: the sign was widely used in the region up to the sixteenth-century Dutch Revolt, and for much longer than in the better-known square notation. Through analytical and comparative study, the sign's occurrence, position, development, relation to other signs and functions are clarified and presented in this article along with insights into its unique notational development.

Introduction

The quilisma is a sign found in the earliest surviving chant notations. It is most often notated as a wavy shape, ending in a longer ascending stroke. It never occurs in isolation but always as part of a larger neume group; for example, a quilisma-pes or as the middle element of a quilisma-scandicus. With the rise of diastematic notation in the twelfth century, the sign gradually disappeared from a number of notational systems, such as the square notation used in the southern Low Countries. In the hufnagel notations of Germany and the northern Low Countries – named so for the resemblance of its notation to the nails (hufnagel) in horseshoes – the quilisma retained more or less the same shape as in the oldest notations, only gradually disappearing in the sixteenth century with the rise of printed music and the religious wars that devastated the region. Since the ‘chant restoration’ of the nineteenth and twentieth centuries, the sign is included again in all chant books.

Since the time of Aurelianus Reomensis (fl. c.840–50) and his description of the sign in Musica disciplina, sharply contradicting interpretations regarding the sign's quality and musical meaning have been proposed. A recent and comprehensive study of its notation history and the interchangeability with other neumes, as apparent in medieval theoretical descriptions, can be found in a publication by Constantin Floros.¹ An overview of theories of interpretation has been made by Franz Tack² and more completely by Walter Wiesli.³ In summary, four groups of theories can be discerned: the quilisma as an ‘ornamental sign’ (trill, tremolo), a gliding tone (glissando, portamento), an indication of particular vocal quality, and a light transitional tone. Since the quilisma occurs throughout Gregorian chant of all ages and genres, the sign poses an inevitable question of interpretation – and consistency of interpretation – to those performing chant.

Northern Low Countries sources

As mentioned, the quilisma retained more or less the same shape in hufnagel notation, a type of chant notation that was common in different forms in central, northern and eastern Europe (with the exception of Hungary) and also in the northern Low Countries. This area, north of the river Rhine, coincided roughly with the historic (arch)diocese of Utrecht (695–1580); the region south of the Rhine belonged to several dioceses, in particular Liège and Cambrai. The far north around Groningen was part of the diocese of Münster, though the city of Groningen belonged to Utrecht. The see of Utrecht was located at St Martin's Cathedral (the Dom) in the city of Utrecht. The ‘Utrecht connection’ between manuscripts from the region can be found in the ecclesiastical organisation of their locations of use (e.g., churches founded by and/or in possession of Utrecht chapter churches, or chapter churches in a close relationship with Utrecht), their liturgical content (e.g., the Utrecht calendar of saints), musical notation (hufnagel notation on black lines) and/or styles of penwork decoration.⁴ Some sources have headings such as Graduale Trajectense (‘Utrecht Gradual’) on their first pages, making clear the ecclesiastical focal point of the region.

Besides geographical and liturgical demarcations, manuscripts from the region share a chronological demarcation: the oldest (adiastematic) notations originating from the northern Low Countries date from the twelfth century. Chant performance and manuscript production came to a standstill some centuries later during the Dutch Protestant Reformation, which coincided with the Dutch War of Independence. These demarcations offer an attractive – and manageable – window for the study of chant and its possible developments within one region, the quilisma in particular. It occurs in all chant genres and all types of manuscripts, from the oldest surviving notations to those of the sixteenth century.

The quilisma in northern Low Countries sources

The quilisma in manuscripts originating from the region knows only one shape: that with two ‘curls’ (Figure 1).⁵ In all sources, the quilisma is reached as a unison or higher pitch.

Figure 1. Utrecht, Universiteitsbibliotheek, 407, fol. 38r: quilisma on exultabis, Domini, veni and Syon.

To gain a precise notion of the quilisma's nature in sources from the region, a comparative analysis of twenty-four manuscripts featuring quilismas was made (Table 1). Criteria for selection of the sources included in this analysis are that they were used in the northern Low Countries, that there is clear evidence of the location(s) where they were in use,⁶ and that they contain a reasonable amount of chants for comparative analysis. The sample offers a cross-section of the entire region, with locations of use ranging from parish to chapter churches and monastic communities to knightly orders. Manuscripts are referred to here by their respective number in the first column of Table 1.

Table 1. Manuscripts selected for analysis

The broadest possible comparative analysis of these sources is possible through the chants of Eastertide: the majority of antiphoners are so-called summer antiphoners and a fair number of Mass Proper chants of Eastertide are included in such books. A total of 138 chants were chosen for the sample: fifty-three Mass Proper and sixty-six Office chants from Eastertide, together with seventeen Mass Ordinary chants. They are listed in Table A1 in the Appendix.⁷

Extent of quilismas

One way of gaining an idea of the extent of quilismas per source would be to simply count all the source's quilismas and divide this number by the number of chants involved. However, this calculation would overlook the possibility that the sign occurs at differing rates within each chant category. Moreover, it would not make clear to what degree one source matches with or diverges from other sources in this study. So, a calculation of variation among and between manuscripts for the same melodic positions was added to the search in Excel sheets: a particular note (group) is only counted if at least one other source shows a variant for that particular position. To make clear the variation of a quilisma, a search was made for the concurrence of two variants:

• • Variant 1 (concurring at the same positions as variant 2): two successive ‘normal’ notes of unspecified pitch; this two-note group may be part of a larger neume group.

• • Variant 2 (concurring at the same positions as variant 1): two successive notes of unspecified pitch, of which the first is a quilisma; this two-note group may be part of a larger neume group.

The demarcation allows for the inclusion of the two-note variation pes/quilisma-pes as well as the larger note group variations scandicus/quilisma-scandicus and torculus/quilisma-torculus. An example of such variation between sources can be seen in Figures 2 and 3: in the highly ‘quilismatic’ ms. 9, quilismas are found on Ecce virgo and eius E-ma-nuel; in ms. 10, these are ‘normal’ notes. Ms. 10 only has the quilisma on pariet (quilisma-scandicus), which is set to ‘normal’ notes in ms. 9.

Figure 2. Utrecht, Universiteitsbibliotheek, 405 (Ms. 9), fol. 10v: communion chant Ecce virgo.

Figure 3. Deventer, Athenaeumbibliotheek, 11 L 2 KL (Ms. 10), fol. 14r: communion chant Ecce virgo.

As illustrated by Figures 2 and 3, the melodic variation between manuscripts from the region always points to the ‘substitution’ of a quilisma for another note; no examples of variation with a quilisma's position left ‘unfilled’ could be found.

The results of the Excel-based analysis are shown in Tables 2–5. The number of variants per source is translated into a percentage of the total number of positions found in that source. These percentages offer a direct overview of the ‘quilismatic’ nature of sources in question; they also make clear that some manuscripts are markedly more so than others. For example, in the Mass Proper category, ms. 3 is a clear ‘unquilismatic’ manuscript (97% variant 1), while ms. 9 is a relatively high ‘quilismatic’ source compared with contemporaneous manuscripts.

Table 2. Mass Proper chants

Table 3. Mass Ordinary chants

Table 4. Office chants

Table 5. Sum of manuscripts featuring multiple categories

The analytical results show that even in the use of a single sign, manuscripts from the same era and region can differ greatly, and different melodic traditions can be discerned. The gradual disappearance of the quilisma is apparent when comparing the oldest with the youngest sources (lowest to highest ms. nos. respectively). The differences in the manuscripts originating from Utrecht's influential chapter church of St Mary's (nos. 1, 2, 4, 11, 12, 13, 24) underline the possibility of chant evolution, even in the same location. Also note the differences in the quilisma's frequency per chant category.

Position of the quilisma in northern Low Countries sources

A remarkable feature of the quilisma in the sources analysed is that its melodic position contrasts sharply with that known from the modern-day Graduale Romanum and other publications of Solesmes Abbey. In the Graduale Romanum, but likewise in Lorraine notations, the quilisma always occurs at the lower note of a semitonal position: mi, si and la below si♭. This position seems to add to the idea of a passing note leading to a stronger note, almost as a leading-note. In the manuscripts of the northern Low Countries, it is very rarely found at this position; the quilisma mostly occurs as the lowest note of a minor third position (re, sol (with si♭) and la). This is an important observation from the point of modal theory: in contrast to the ‘passing notes’ of the Graduale Romanum, the quilisma is just as often found on modally important pitches.

To gain a statistically grounded notion of the quilisma's position with regards to pitch, all quilismas in the fifty-three Mass Proper chants were listed, comprising a total of 366 positions (Table A2 in the Appendix). (Proper chants were chosen to include two sources discussed in the next paragraph.) By numbering the frequencies of the quilisma's pitch combinations, the quilisma's position becomes clear (Figure 4).

Figure 4. Frequency of the quilisma per pitch combination in fifty-three Mass Proper chants.

As can be observed, the pitch combinations most often found are la-do, re-fa (minor thirds in the higher register of melodies) and the whole tone position do-re, which is echoed in other whole tone positions: sol-la, fa-sol, re-mi and sol-la. The quilisma is very rarely found on minor third positions with a semitone at its basis: mi, si and la below si♭.

The quilisma and melodic traditions

The story of chant in the northern Low Countries is one of many co-existing chant traditions. Two examples of such traditions (also included in Table A2) make clear that the differentiation also holds for the use of the quilisma.

One of the oldest manuscripts kept in the Netherlands is the missal from the Hellum parish church (Utrecht, Catharijneconvent, BMH h7) dating from the second half of the twelfth century or c.1200. Hellum, in the northern region, belonged to the exclave of the Münster diocese. The missal features detailed adiastematic notation of Mass Proper chants, which suggests a melodic tradition very much in line with that of the Utrecht diocese. Although it contains many quilismas, it also lacks at least ninety-four quilismas shared by the sources from the Utrecht diocese (see Table A2). The difference, which does not seem correlated to any particular type of quilisma, can be observed in almost every chant transcription table.

Another example of differentiation can be found in a fourteenth-century gradual belonging to the Commanderij Jan Baptist of the Knights Hospitaller of Haarlem (Haarlem, Noord-Hollands Archief, 184 C7). The knightly order, an ‘import tradition’, followed the so-called liturgy of the Holy Sepulchre, which is notable for a specific liturgical ordering and unique combinations of order-related and local feasts. Moreover, all the commandry's manuscripts feature different melodic traditions than those of other contemporaneous Low Countries sources. A comparison of quilismas makes clear that a further distinction can also be observed in this particular sign: the selection of Eastertide Propers contains comparatively few quilismas (only ten such moments in fifty-three chants) and apart from one unique quilisma on mi-sol, these all concern whole tone quilismas. This contrasts sharply with the degree of quilismas and the preferred minor third position in contemporary sources from the Utrecht diocese. A further exploration of the Hospitaller manuscript confirms this first impression – in fact, no further minor third quilismas were found at all.

More study could perhaps shed light on the quilisma's favoured position per melodic tradition throughout Europe.

The whole tone quilisma: the exchange of oriscus and quilisma

It should be noted that a preference for specific positions of the quilisma does not wholly exclude other positions. As mentioned, all ‘Utrecht’ sources also contain examples of the sign as the lowest note of a whole tone – but only specific positions of whole tones. In the transcription tables of the 138 Proper chants under scrutiny, this became especially apparent in cadences ending with do-re, such as in Haec dies chants. In the oldest known notations of Gregorian chant, this is a position usually reserved for the pes quassus (two ascending notes of which the first is an oriscus).⁸

The phenomenon can partly be explained by a notational development of the apparently similar oriscus and quilisma. As pointed out by chant scholar Franz Karl Praßl, compared with the oldest surviving notations such as in the Hartker Antiphoner (St Gall, Stiftsbibliothek, 390–391), the neume called oriscus is often ‘replaced’ with a quilisma in younger notations.⁹ This only seems to concern ascending neume groups that feature the oriscus, such the pes quassus and the salicus (three ascending notes of which the second note is an oriscus). The oriscus as part of a descending note group (the pressus) is never ‘replaced’. A reverse development, with a quilisma as found in older sources being notated as an oriscus in younger sources, has not yet been observed.

This development is corroborated in northern Low Countries sources: ms. 1 is a good example. As chant scholar Ike de Loos pointed out in her dissertation on the manuscript,¹⁰ the ‘replacement’ may be explained in part by developments in notational technique: however sophisticated the notational system of the manuscript, it does not contain the pes quassus; therefore, it seems ‘logical’ to replace this neume with the structurally similar quilisma-pes. The salicus, however, is nearly always found ‘replaced’ by a simple scandicus (three ascending ‘normal’ notes). De Loos explained the interchangeability by the suggestion that the difference in performance practice between oriscus and quilisma had disappeared during the centuries separating Hartker and the Utrecht antiphoners. It could therefore explain the quilisma's occurrence outside its ‘preferred’ minor third position. Two phrases with an exceptional high quantity of the pes quassus may serve as example, as shown in Figures 5–8.

Figure 5. St Gall, Stiftsbibliothek, 390, p. 47: ending of Beata Dei genitrix.

Figure 6. Utrecht, Universiteitsbibliotheek, 406, fol. 23r.

Figure 7. St Gall, Stiftsbibliothek, 390, p. 174: responsory Contumelias et terrores.

Figure 8. Utrecht, Universiteitsbibliotheek, 406, fol. 89v.

Function

The interchangeability of the oriscus for a quilisma draws attention to the quilisma's function, which remains debated up to today. As described in great detail by Dom Eugène Cardine in his Gregorian Semiology, the oriscus of ascending note groups indicates a melodic tension towards the (higher) note ‘of special importance’ which follows.¹¹ The same interpretation is given for the quilisma of the quilisma-pes and quilisma-scandicus with compelling argument. The similarity of both signs’ function seems corroborated by the development described in the previous paragraph.

Unknown so far is that a similar function is also suggested in some northern Low Countries sources, in particular ms. 9: the missal (with notation) Utrecht, Universiteitsbibliotheek, 405, once belonging to Lopik's parish church (c.1450) (Figure 9).¹² As could already be observed in Tables 2, 3 and 5, it features an abundance of quilismas: the sign is included in every possible position, ranging from the ‘regular’ minor third position to the more ‘exotic’ ascent of a fifth, and also in regular connection with liquescent neumes. Remarkably, its quilismas are found on positions where none of the oldest known chant sources (e.g., Einsiedeln, Stiftsbibliothek, 121) give such a neume.

Figure 9. Utrecht, Universiteitsbibliotheek, 405, fol. 17v: introit, Puer natus est. Note that quilismas are written with upward strokes (‘puer natus’) and strophae with downward strokes (‘est’).

Part of the missal's more uncommon quilismas can be explained on the basis of the aforementioned interchangeability of oriscus and quilisma in ascending neume groups. In line with Praßl's observations, the oriscus is still present in descending note groups (e.g., nobis, angelus). But a great deal of Lopik's quilismas seem odd in comparison with contemporaneous and older manuscripts, and cannot be explained on the basis of this interchangeability alone. After examining different scenarios, one type of neumatic phenomenon showed the most similarity to the quilisma's use in the Lopik Missal: initio debilis.

Initio debilis is best discernible at (and most associated with) the ‘torculus initio debilis’: a torculus ‘with a weak initial note’, ‘with a relative predominance of the last two notes’.¹³ The phenomenon is only associated with the oldest neumatic notations. It can be detected very clearly in the famous Einsiedeln 121 (c.960–70) by its use of the ‘excessive’ torculus and the addition of litterae significativae (most often ‘tenete’) (Figure 10).

Figure 10. Einsiedeln, Stiftsbibliothek, 121, p. 258: two ‘normal’ torculi followed by a ‘torculus initio debilis’ with ‘tenete’ in the introit Accipite iucunditatem.

The earliest surviving neumatic sources from the northern Low Countries (such as the aforementioned Hellum Missal) give no suggestion of initio debilis. Owing to its clarity of notation, Einsiedeln 121 was used for comparison with the Lopik Missal (with only the clearest examples of ‘torculus initio debilis’ selected for comparison). The results can be found in Table A3 in the Appendix. They show that every initio debilis in Einsiedeln on an intonation or mid-word syllable is given shape as a quilisma in the Lopik Missal; in short, all torculi that ‘lead’ to ensuing notes.

The positions concerning final syllables point to a markedly more mixed use of quilisma, the missal featuring a quilisma-torculus, a pressus (major), or a normal torculus. This can be explained by the phrases in question, since these are all melodic and/or textual endings where the melodic flow is rounded off instead of continued. At such points, the cadential pressus is usually the neume of preference.

The remarkable feature of the Lopik Missal is that the lower debilis note of first and mid-word syllables is given shape not by a regular note, but by a quilisma. Even more remarkable is the fact that this occurs in a fifteenth-century manuscript in hufnagel notation, while the phenomenon is only associated with the oldest known adiastematic neumatic sources. As far as it is known, hufnagel notations never include initio debilis, nor is there any sign associated with this phenomenon.

As a slight digression from the topic, the question could be asked if the ‘Lopik style’ should be considered a one-off flourish by an eccentric scribe, or a remnant of an older tradition, or perhaps a notational attempt to preserve something of an older performance tradition. It should be noted that the missals of Elden (ms. 5) and Almkerk (ms. 6) – two churches about 80 kilometres apart – are often in agreement with the quilismas of Lopik, especially those at less regular positions. This similarity is already apparent in Table 2. It is hoped more sources can be found to further study this phenomenon.

Initio debilis continued: the ‘quilisma-tristropha’

The idea of a quilisma indicating initio debilis may also explain the nature of two symbols of unknown significance encountered in a number of Low Countries sources.¹⁴ Both symbols always appear as the first element of a strophae group. In the oldest Utrecht sources,¹⁵ the symbol is a wavy horizontal line of two strokes, similar to a tilde (Figure 11).

Figure 11. Utrecht, Universiteitsbibliotheek, 417, fol. 47r: three examples of the ‘tilde’, each time followed by two strophas.

Although its shape suggests a tractulus (usually a horizontal line indicating a lower note in a phrase) of some sort, the tractulus is not included in the neumatic alphabets of any of these sources, its role already being taken by the punctum. Moreover, tristrophas opening with a punctum can be observed in all manuscripts in question.

At similar positions as the unknown sign in Mass chants, the fifteenth-century graduals ms. nos. 5 and 6 feature a neume that looks like an isolated quilisma or a distropha.

So far, it has been unclear how many notes the sign indicates, and why it often occurs as the first lower element of a tristropha – a position reserved for either a tractulus or stropha in manuscripts throughout Europe. The idea of quilisma as an indication for initio debilis offers an explanation: the sign could be a quilisma, indicating a note of a passing nature, leading towards the notes that are emphasized by their repercussion. This idea is given argument by the fact that in late medieval sources, the symbol is notated in the same manner as the quilisma (as part of a larger neume), namely with two upward strokes. It thus results in a ‘quilisma-tristropha’ (my terminology), which is not only remarkable for its composition, but also for its notation: although obviously part of a neume group, the quilisma is notated as an isolated element (Table 6).

Table 6. Examples of the ‘quilisma-tristropha’ in the Christmas introit Dominus dixit

The ‘quilisma-tristropha’ is relatively rare in sources, and is most often observed in more melismatic chants. Table 7 shows a comparison of its use. As can be observed, the ‘quilisma-tristropha’ does not occur in all sources: sometimes, only a quilisma-pes fills this position. However, the element important to the comparison – the quilisma – is nearly always present. This strongly suggests that both signs indicate a quilisma.

Table 7. Comparison of manuscripts with regards to the ‘quilisma-tristropha’

Note: q-tr = quilisma-tristropha; q-pes = quilisma-pes.

The ‘quilisma-tristropha’ is only found in manuscripts nos. 1, 2, 4, 5, 6 and 9 on ascents of a whole tone, minor third and fourth; it is suspected in the (somewhat crudely written) gradual Utrecht, Universiteitsbibliotheek, 413. It could not be found in any of the other manuscripts studied.

Conclusions

As stated by David Hiley, ‘it is true that sources are in general surprisingly uniform … even in the placing of such special features as quilismas. … But the agreement is general, not exact.’¹⁶ This also holds for manuscripts from the northern Low Countries, even sources from the same location, as demonstrated by those from St Mary's, Utrecht. The quilisma's gradual disappearance from chant notation over the ages – a development that differed per notation system – could account for some of that variation. But the manuscripts from the northern Low Countries draw attention to other factors that influence quilismatic variation.

First, as could be observed in the quilisma's occurrence in the Hellum Missal (of the Münster diocese) and the gradual of the Haarlem Knights Hospitaller, a further distinction in melodic traditions can also be observed in the use of this particular sign.

Second, the interchangeability of the oriscus for a quilisma in ascending note groups over the ages has added to the signs’ variation: a development also observed in northern Low Countries sources that corroborates Cardine's observations on the function shared by both signs: ‘leading towards’ a following note of importance.

Third, comparative analysis has demonstrated the possibility of the quilisma's use to indicate initio debilis (a phenomenon so far unknown in hufnagel notation, let alone fifteenth-century chant notation). Consistent with this function, two signs of so far unknown significance, encountered as the first lower element of a tristropha, could be explained as a quilisma, thus realising a ‘quilisma-tristropha’ with a first quilisma-note ‘leading towards’ notes that are emphasized by their repercussion. Both of these observations add further argument to the sign's function mentioned.

Lastly, this short study of the quilisma's occurrence in northern Low Countries manuscripts is a demonstration of the musical wealth contained in these sources. It is hoped that many more discoveries may follow.