I wish I had had a tool like this twenty years ago. At about that time I was engaged on a stylistic analysis of the three-voice Mass in the later fifteenth and early sixteenth centuries as part of my doctoral studies.1 My aims were ambitious: to understand some fundamental changes in the way musical textures operated and were conceived across as broad a sample as possible of pieces with shared scoring and texts. It was painstaking work, involving the close study—aided only by pen and paper—of what I had selected as a corpus of “representative” scores. This produced some useful conclusions about changing approaches to musical texture: I was able to trace a steady if piecemeal shift toward an increasingly stratified and contrapuntally nonhierarchic texture, to plot some important steps in the germination of the imitative style and the conditions that stimulated it, and to witness some intriguing mismatches between empirically observable stylistic change and the lasting influence of the theoretical principle of the discant-tenor duo.

My heuristic tools were patterns of ranges, the nature, degree, and frequency of tenor-contratenor part crossing, the incidence of large leaps, and imitation—basic parameters of structure and local compositional building blocks. The same kinds of material, in other words, that form the focus of the Josquin Research Project. But whereas my efforts involved many hours of close observation and counting, and equipment no more sophisticated than an electronic calculator, this resource, at the click of a few buttons, provides researchers with an array of data that can open up a broad and—in its extent and diversity—currently unforeseeable range of research trajectories. If my study was perforce limited by the sheer number of man-hours involved and inevitable repertorial choices, this body of material should, as it continues to expand in its remit, suffer from fewer and fewer such constraints and engender much greater progress in attempts to understand the basic building blocks of musical style.

The Josquin Research Project began as a tool in debates about authenticity. As we read on the site's “About” page, by quantifying the stylistic parameters of everything firmly ascribed to Josquin, Jesse Rodin and his colleagues sought to assess the likelihood of his having composed other works ascribed to him. But the project quickly branched out into more general directions, as its architects realized that the tools they were fashioning had much broader application, encouraging everything from close analysis of individual works to comparative study across the corpus of composers, genres, manuscript traditions, and larger shifts of style over time.

To illustrate the scope of the project I shall begin by walking the reader through the resource and its tools. I shall then consider what these may permit in terms of investigatory directions, and what—from my own perspective, at least—might prove to be useful new additions, in terms of both repertory covered and analytical tools. I should say at the outset that the directors of the project view it in strongly interactive terms, intending it to offer a broad base of support to analytical studies across the scholarly community; for this reason they welcome contributions from users, whether in the form of transcriptions, suggestions for further developments, feedback on features that may strike “outsiders” as counterintuitive, notification of “bugs,” and so on.

The opening screen (accessed from http://josquin.stanford.edu/), which serves as a shop window for the project, informs users of the latest news and additions to the site, and offers a quick turnover of “Sample Works” (see Fig. 1). The sample work functions as an invitation to explore the project's tools. Clicking on the title brings up the image seen in Figure 2, the main area of which gives access to “raw materials.” These take the form of a copy of the score (in a “diplomatic” transcription, in which users may choose whether or not to view editorial accidentals), an mp3 file of the piece as a basic MIDI sound file, and a stave showing voice ranges and the incidence of component pitches. “Hovering” the cursor over these pitches brings up the number of each within the piece; clicking “Show by duration” switches to a view of the sums of their durations (figures for which are likewise accessible via the cursor). This data gives clear indications about tessitura, and about the parts of the range (which usually comprises a modal octave plus a note or two on either side) in which activity is greatest.

Figure 1

Josquin Research Project home page, accessed January 19, 2015, http://josquin.stanford.edu.

Figure 1

Josquin Research Project home page, accessed January 19, 2015, http://josquin.stanford.edu.

Figure 2

Josquin Research Project, work view for Johannes Martini, Ad Dominum cum tribularer, accessed January 19, 2015, http://josquin.stanford.edu/work/?id=Mar2001.

Figure 2

Josquin Research Project, work view for Johannes Martini, Ad Dominum cum tribularer, accessed January 19, 2015, http://josquin.stanford.edu/work/?id=Mar2001.

The real “business end” of the site, however, is found on the left-hand side. Data that in my study of the three-voice Mass repertory was the fruit of hours of manual counting is available here in seconds via the interactive boxes at the top of the page. Pitches, intervals, and rhythms can be searched both individually and in combination by following the encoding rules accessed through the “?” symbol. This makes light work, for example, of the search for melodic/rhythmic motifs, either within the chosen piece or across genres and works of individual or all composers in the database (selected via the two lower dialog boxes). This is a facility that could prove invaluable for those of us interested in motivic practice in this repertory, in which distinguishing between actual quotation and more general motivic small change has been a perennial sticking point. While there will always and inevitably be a grey area, as well as differences of perspective concerning the compositional implications raised, this tool will certainly refine the process of sifting the basic data.2 

The “Enter Interval” box makes it possible to search successions of intervals, specifying successive directions via “+” and “–.” A particularly useful feature is that “+” and “–” function alone, facilitating searches for leaps of a specified size followed by movement (of unspecified distance) either within or beyond the leap. This would allow the tracing, for example, of contrapuntal behavior according to the “classical” principle of “quitting within” intervals, or otherwise; but it also permits quick assessment of the spread of intervals across the parts of the polyphonic texture, and thus the extent of textural consistency.

Buttons on the lower left give access to a series of visualizations of data generated by the site architects. “Ribbon” brings up a schematic outline of the average ranges, actual pitches, and outlying notes of each voice (these can also be viewed in vertical alignment with one another), together with their average (the grey ribbon at the bottom; see Figure 3 for a vertical section from the Martini piece). This allows anyone concerned with analyzing the relative behavior of the voices to quickly perceive, for example, their degree or absence of conformity to modal octaves, and the speed and frequency of digressions. Like many of the site's tools, it is therefore primarily useful for quick sifting prior to more detailed analysis of vocal profiles. “Activity” plots the number of separate rhythmic attacks against time calibrated in five-measure increments, both as actual numbers of attacks (below, amalgamating all component voices; see Fig. 4) and as density of rhythmic activity (presented on the site in black/greyscale). Dissonances, suspensions, and parallels, color-coded into the score, are accessible from their respective buttons. These allow the user direct access to the kinds of detail concerning contrapuntal practice that I, twenty years ago, had to identify and quantify by reading through entire scores. The “Piano Roll” button allows a quick view of the melodic profiles and durations of the individual parts, enabling rapid recognition of imitative points, part crossings, and so on. A quick look at an example on the website will reveal the source of the name: the visualizations plot relative ranges/durations on vertical/horizontal axes. This tool too serves a general “sifting” purpose that would typically be followed up by more detailed analysis of scores; but here as elsewhere the “Gestalt” view that it permits should prompt closer analysis of the broad patterns that are thrown into relief.

Figure 3

Josquin Research Project, detail of ribbon function for Johannes Martini, Ad Dominum cum tribularer, accessed January 19, 2015, http://josquin.stanford.edu/work/?id=Mar2001.

Figure 3

Josquin Research Project, detail of ribbon function for Johannes Martini, Ad Dominum cum tribularer, accessed January 19, 2015, http://josquin.stanford.edu/work/?id=Mar2001.

Figure 4

Josquin Research Project, detail of activity plot for Johannes Martini, Ad Dominum cum tribularer, accessed January 19, 2015, http://josquin.stanford.edu/work/?id=Mar2001.

Figure 4

Josquin Research Project, detail of activity plot for Johannes Martini, Ad Dominum cum tribularer, accessed January 19, 2015, http://josquin.stanford.edu/work/?id=Mar2001.

Returning to the home page, both “Quick Browse” and “Browse” (the latter adding choice of number of voices to the former's “Composers,” “Genres,” and “Title”) provide access to any individual work or to lists of pieces. Particularly powerful, however, is the “Search/Analysis” tab, which facilitates quick comparison of various analytical parameters, with options to limit by particular composers and/or genres in different combinations. The results are presented both in words and in score, allowing for easy comparison. Figure 5 presents a sample of “Parallels,” using the four databased works by Isaac. What we see is a synoptic view of the parallel fifths and octaves in all the items by Isaac that have been added to the website to date, with names of works and information about the component voices highlighted in blue. The data also distinguishes between “hard” and “soft” parallels, the former being those that are directly struck while the latter involve an intervening (typically passing) note. Rhythmic patterns are similarly quantifiable, with lists easily generated that detail all available rhythmic configurations within perfections, together with numbers of occurrences. While this data derives from the supplied transcription and does not, of course, take account of source variants, it does provide useful basic information concerning the rhythmic profiles of different composers, allowing the plotting of changing approaches to rhythm across genre and time. There is no doubt, for example, that the music of the Josquin generation, with its growing emphasis on motivic cohesion, sees much closer control of the possibilities of these patterns than had earlier musics, a shift easily observable from these charts.

Figure 5

Josquin Research Project, results of search for parallels in four works by Henricus Isaac, accessed January 19, 2015, http://josquin.stanford.edu/analysis/parallel.

Figure 5

Josquin Research Project, results of search for parallels in four works by Henricus Isaac, accessed January 19, 2015, http://josquin.stanford.edu/analysis/parallel.

Of course it is in the nature of a research tool such as the Josquin Research Project that it can only be as useful as the questions that are asked of it, and that this will largely depend on the degree of foresight shown by its architects in envisaging their possible range. The beauty of a website of this sort is its adaptability in relation to both these conditions: as noted above, the intended aims and scope of the project have been shifting since its early stages, and will surely continue to do so as it unfolds—in response to hitherto unforeseen questions—in ever new repertorial and analytical directions, and in accordance with users’ stated priorities.

It would seem reasonable, at least prima facie, to suggest that the most propitious tools are those that allow purchase for the broadest possible range of questions. Those that are embedded in scores achieve this most obviously, by dint of the simple fact that they permit one to see, and hopefully to comprehend, each highlighted feature in the complete context in which it appears. To be able, for example, to quickly observe patterns of leaps, parallel intervals, and suspensions within a score is to be able to perceive their roles in contrapuntal practice more generally, and in turn to suggest further questions as to how such practice may operate. On the other hand, larger patterns of contrapuntal behavior may be more easily grasped by means of the broader viewpoint afforded by reductive visualizations such as the piano roll. Thus the combination of different kinds of tool creates a heuristic field of possibility larger than the sum of its parts.

In some areas, though, one can envisage improvements that could substantially refine the questions to which the site might be responsive. A particularly noteworthy example concerns the charts of measure-long rhythmic patterns that, as they stand, give little sense of how such patterns affect the contrapuntal surface. To have an iambic pattern in all voice parts, for example, means something very different from having it in one, set against an array of different configurations in the other sounding parts; similarly, rhythmic uniformity across all voice parts makes a very different impression from a pattern shared by only two while the others may be resting. It would be a tremendous boon if at some point the architects were able to show color-coded rhythmic patterns in scores at the same time as, for example, (differently) color-coded leaps of different sizes and/or suspensions, parallels, and patterns of imitation. Only in the context of complete textures can such patterns reveal what they have to tell us about the distinctiveness of, or distinctions between, contrapuntal idioms. The same is true of the graphs for “Attacks per measure,” which do not differentiate the relative activity of the parts, though the charts showing rhythmic density do permit a general impression of such difference/similarity. The ability to “switch on” different stylistic features in a score simultaneously would greatly enhance their individual interpretative value, making it possible, as such a format would, to study their many and various interactions.

Perhaps the biggest single question that the Josquin Research Project allows us to address concerns the fundamental shift that takes place in the later fifteenth century toward a generally de-hierarchized musical texture characterized most obviously by pervading imitation. The relative behavior of the individual parts is clearly key to this, most particularly the erosion of the primacy of the structurally fundamental duo between discantus and tenor that had for so long held hierarchical sway over compositional texture. Likewise, increased parity between the contratenor part(s) and the discantus and tenor in terms of intervallic and rhythmic usage and (particularly) motivic similarity is clearly a signpost toward a more contrapuntally equal texture.

It is evident that the ability afforded by the site to quickly highlight motivic and intervallic practice is of great value in attempting to track such a change. To take a simple example, using the interval tool I was able to plot the use of leaps of a sixth in the four mature Masses of Du Fay. Unsurprisingly, these predominated in the pair of Masses long viewed as the earlier of the four, Se la face ay pale containing fourteen sixths and L'homme armé twenty; equally unsurprisingly they strongly predominated in the contratenor altus, which had (respectively) eleven and fifteen sixths compared to three in the superius in Se la face and three in the bassus and one each in the superius and tenor in L'homme armé. By contrast, Ecce ancilla domini has only one sixth (in the bassus) while Ave regina celorum has seven, scattered as follows: one in the superius, two in the bassus, and four in the contratenor.3 Basic though this tally is, it is symptomatic of more general shifts across these works that can be enumerated equally quickly with the help of the interval and pitch tools.

Clearly the usefulness of the Josquin Research Project is contingent upon users’ familiarity with the works as sounding totalities. Local details and “events” are of interest because of their contribution to the character of the larger tapestries they weave, and it is in turn awareness of these larger impressions that engenders queries at the local level. There will never be any substitute for the close hearing, singing, and directing of the repertory in question. It should also be noted that printed scores will similarly retain their importance, for two principal reasons: first, because the scores presented here are based not on primary sources (with their many incidental but sometimes substantive variants) but on already-published editions; and second, because they lack text underlay, a feature that conditions and interacts with other compositional decisions (such as imitation), increasingly so in the latter part of the period covered. But the more familiar a user is with the repertory the more opportunities he or she will find the project opens up, especially since its contents are constantly on the move.

Everyone will have his or her own “wish list” regarding hoped-for future developments. Chief among my own pleas, especially in relation to the earlier part of the period under scrutiny, for which the bulk of surviving music is anonymous, would be to have more coverage of unascribed music. In this context the architects’ solicitation of transcribed material is germane: given the size of the repertory the degree of usefulness of the site must in the end be a reflection of the extent of its users’ direct involvement in its growth. Certainly, access to the vast anonymous repertory seems a precondition for any kind of comprehensive understanding of what it is that makes the work of particular composers distinctive, as does being able to identify anonymous pieces that stand in close proximity to ascribed ones. Perhaps we may ultimately look forward to a day when the taxonomy of this repertory is littered with such art-historical-type pseudo-identities as “the Master of the Brussels anonymous Mass.” Further in terms of broader taxonomies, it would also be helpful to build in methods for assessing repertory by manuscript rather than just by composer, and to be able to compare different readings of the “same” piece, especially when, as in the case of some Masses by Martini, for example, they present radical differences. The principle whereby scores both with and without accidentals are accessible could clearly be extended in this direction, though a better model would be some directly visible, perhaps color-coded “master score.”4 

But it is only to be expected of such a broad-based project that every user will have his or her hopes for the future; and things are shifting at a pace whereby I can predict that many bug-fixes, requests for change, and broader developments will already have occurred by the time this report goes to print. That process can only be helped by use of the site as a teaching tool—a use to which I, and I am sure many others, have already put it. As we push our students (and ourselves) toward questions about musical construction that, years ago, could be approached only with pen, paper, and a liberal supply of fortitude, many more insights (and hopefully transcriptions) will emerge to drive the agenda forward. The central and abiding point for now, however, is that all of us in the field owe the architects of the Josquin Research Project a tremendous debt of gratitude: what they have taken on is ambitious to the point of heroism, and one hopes and trusts that it will attract an ever-growing community of users and aficionados of a repertory that still, half a millennium after its composition, holds so much to attract and absorb us.

1.
Later published in revised form as The Three-Voice Mass in the Later Fifteenth and Early Sixteenth Centuries: Style, Distribution and Case Studies (New York: Garland, 1995).
2.
The Josquin Research Project uses the open-source Humdrum standard as the primary format for the musical data. These files are in turn used with various Humdrum tools for search, analysis, and transformation into graphical notation. See the explanations and tools available at http://www.humdrum.org, http://extras.humdrum.org, and https://github.com/humdrum-tools.
3.
These figures are slightly skewed in the responses generated by including “dead” intervals occurring across double bar lines, a minor “bug” that would be worth fixing.
4.
Open-source projects such as the Josquin Research Project make it possible for others to learn from and adapt the methods and software architecture. Documentation for the project can be found at http://wiki.ccarh.org/wiki/josquin and https://github.com/josquin-research-project/jrp-website.