Project Description

Contents

1. Introduction
2. Research context
3. The sources
4. Project description
   1. The problems
   2. The solution
   3. Transcription and primary enrichment
   4. Target group and importance

1. Introduction

The TMI is an initiative of the Department of Computer and Humanities at Utrecht University, the aim of which is to provide a tool for the study of Italian music treatises from the second half of the sixteenth and the early seventeenth century. This tool will consist of a digital corpus of these treatises on an Internet server. The TMI will be an extensible system, to which users will be able to add source materials and knowledge.

The TMI answers a need which is strongly felt not only in Netherlands musicology but also world-wide. The music theoretical sources that will be included in the TMI are central to the study of Renaissance music, and their limited availability and accessibility at present is widely regarded as an obstacle to research. The TMI will eliminate this obstacle.

The TMI is closely related to other departmental initiatives in the field of electronic publication of reference materials and sources for the humanities, such as the 'Image and Word' project centring around a number of 16th-century emblem books, which will be published on CD-ROM. In addition to images and texts, the TMI will also make the music examples in the treatises accessible, both aurally and visually.

The TMI project has importance for the following fields:

• musicology: the selected treatises are among the most important of a much-studied period. The TMI will allow all these sources to be studied for the first time together, in the normal working environment;
• performance: the sources contain a great amount of information about the musical repertoire, instruments and performance practice of the period;
• humanities computing: the establishment of a de facto standard for the edition of enriched tri-medial source materials (i.e. containing text, images and music) based on the ISO standard SGML, and the querying of such materials on the Internet.

The innovative character of the TMI is twofold. First, it is innovative as regards contents, providing access to tri-medial music historical sources and a means for enriching these with knowledge acquired by the researcher. Secondly, it is innovative as regards technology because of the employment and further development of the most recent methods for electronic publication and retrieval.

Because of these innovative aspects, the project group IWI (Innovation of Scientific Information Provision) of the SURF foundation (Co-operating Academic Computing Facilities) granted a considerable subsidy for the first one-year phase of the construction of the TMI. At the end of this phase we produced a stand-alone demo prototype of the TMI, a CD-ROM containing four treatises, which was published in August 1997. The second and third phases (one year each) of the project were be sponsored by the Netherlands Organisation for Scientific Research (NWO). Thereafter the TMI will be extended by user contributions and maintained by exploitation revenues.

Return to top of page

2. Research context

Musicians from the Low Countries dominated music history in the greater part of the fifteenth and sixteenth century (the musical Renaissance). However, many of these musicians spent a considerable part of their creative life in Italy, however. The most famous case is that of Josquin Desprez (c. 1440-1521), whose music displays a new ideal of both musical expression and stylistic elegance. The next generations of Northern musicians working in Italy, among whom were Adrian Willaert (c. 1490-1562), Cipriano de Rore (1515/16-1565) and Giaches de Wert (1535-1596), continued to refine these ideals. Many Italian musicians studied with them and took part in a quest for new and more forceful forms of musical expression. This led to a stylistic diversification towards the end of the sixteenth century, the emergence of new genres such as opera, and also a vigorous debate as to which musical means were acceptable and which unacceptable. The so-called Artusi-Monteverdi controversy (ca. 1605), a turning point in the history of music, is a case in point.

Music theory played a significant role in these developments. Music theory has a venerable and rich tradition (both in terms of quantity and of quality) in the history of Western music, from its earliest beginnings in Greek Antiquity onwards. Two separate traditions of music theory developed during the Middle Ages. One was concerned with the philosophical and mathematical aspects of music, the other with its notational, rhythmic, melodic, and contrapuntal sides. From the end of the fifteenth century onwards, the two traditions became more and more intertwined. Moreover, music theory underwent a strong influence from — then often only recently rediscovered — classical writings about music, rhetoric, philosophy and mathematics. Many sixteenth-century theorists, especially those working in Italy, tried to formulate an integrated view of the technique, ethics, philosophy and expression of music. Such efforts are the intellectual counterpart to the compositional developments that took place at the same time. It is obvious that theory did much more than merely record the latter: considerable cross-fertilisation took place between theory and composition.

During the last century, the music theory of the Renaissance has been studied intensely. At the same time this study has generally been selective in its subject matter. The focus has been on the 'practical' side of theory, which concerns performance and composition. By contrast, non-technical and speculative parts of theory were usually neglected, so that considerable parts of well-known texts have remained virtually unstudied. Areas extensively studied include notation, ornamentation, counterpoint and (to a lesser extent) modality. Since practical application of findings — for example in education, and in the construction of musical grammars or computational models — has been a major consideration, the degree of truthfulness and consistency of the sources has received much comment, often of a less than favourable nature.

However, a contrasting approach has received increasing attention over the last few years, namely to study music theory in its own right from a historical point of view. To the prime examples of such work belongs Claude Palisca's Humanism in Italian Renaissance Musical Thought (1985). It does not ignore 'practical theory', but he presents music theory in the context of the entire document or the intellectual tradition to which it belongs, not isolated. Generally, subject areas that receive considerable attention include the variation of theoretical approaches in treatises, patterns of argumentation that shape entire treatises, the place of speculative music and arcane issues such as musical magic, the reception and reinterpretation of ancient learning (both musical and non-musical), the relationship to the history of ideas (as in Palisca's book), and to the history of mentality (Gary Tomlinson, Music in Renaissance Magic, 1993).

Such research is no longer confined to selected passages from a relatively small number of 'original' music treatises: very many sources, hardly explored so far, are potentially relevant in their entirety. This point is illustrated in Frans Wiering's dissertation The Language of the Modes (1995), in which considerable new insight in the history of polyphonic modality was acquired by the study of many hitherto neglected sources. This raises two issues in the area of information management: the restricted availability and poor accessibility of many such treatises, especially those in Italian. For Latin treatises these problems have been greatly diminished by the creation of a computer-readable corpus of treatises, the Thesaurus musicarum latinarum. The proposed research tool, the Thesaurus musicarum italicarum, will do the same for Italian-language texts, so that materials in the two major languages of Renaissance music theory will be covered.

The Department of Musicology of Utrecht University has an international reputation for research into the music from the Renaissance. One of the individual research projects within this Department that will benefit in particular from the TMI is entitled 'Zarlino's ideal of the musico perfetto'; financial support is currently being sought for it. The following example may show what kind of knowledge could be generated from the TMI for this project.

Gioseffo Zarlino's Sopplimenti (1588) is a lengthy reply to Vincenzo Galilei's criticism of Zarlino's earlier writings. The book appears to be a point-for-point refutation — and not a very successful one — of Galilei's criticism. On the surface, Zarlino maintains an icy courtesy towards his opponent; however, an examination of Zarlino's terminology with the help of the computer shows that below the surface he is continually accusing Galilei of being a heretic not only in musical but especially in religious matters, thus foreshadowing the famous conviction of Vincenzo's son Galileo for heresy.

Return to top of page

3. The sources

From the numerous Italian music treatises dating from the 16th and 17th century, a 'core corpus' has been selected, which can be extended by contributions from others. The criteria for the selection of this corpus were importance, coherence and diversity:

1. Gioseffo Zarlino (1517-1590) was undoubtedly the most influential Italian theorist of the period; his Le istitutioni harmoniche (1558, 1573, 1589) is one of the most important music treatises of all times. Zarlino's works are discussed in more than 60 publications that have appeared since 1986. However, in almost all of these the authors discuss only tiny portions of Zarlino's writings and do not attempt to integrate the results into a broader view of his musical thought. All authorised editions of Zarlino's three treatises will be included in the TMI.
   A second group of works stems from the 'radical movement' in Italian music theory. Vicentino's L'antica musica ridotta alla moderna prattica (1555) is the oldest of these. A third group documents the reception of Zarlino's writings, and includes treatises by Aiguino, Pontio and Tigrini. Finally, the fourth group of works consists of documents surrounding the Artusi-Monteverdi controversy mentioned above in section 2.
2. The selected works display a notable coherence, by means of references and allusions to each other, and by emphasising certain subject areas (reception of ancient writings, theory of modes, compositional theory, musical expression) and musical repertory (vocal music of Monteverdi, Palestrina, Rore, and Willaert).
3. In order to construct a model of representation which is as comprehensible and flexible as possible, a maximum diversity of form has been aimed at. The sources are both manuscripts and printed works; the genres are quite divergent (letter, dialogue, textbook, axiomatic treatise, 'musical cookbook' of composition). In addition to text, the treatises contain tables, mathematical diagrams, music examples, complete compositions with commentary, or construction plans for musical instruments.

These are the treatises which form the core corpus.

The sources of this corpus total around 4600 pages, more than 1300 of which have already been digitised in phase 1 of the project. In addition, a number of short, mostly manuscript, documents that are closely linked to the sources listed above, will be included. These include letters by Zarlino, short treatises by Bottrigari, fragments of otherwise lost works by Artusi, and letters and prefaces by Monteverdi.

Return to top of page

4. Project description

4.1. The problems

In spite of their widely recognised significance, the sources listed above are not nearly as often used in research as they deserve. As stated above, there are two reasons for this: their restricted availability and poor accessibility.

1. The original sources are rare and vulnerable. They are usually not allowed to leave a library and their use is often subject to considerable restrictions. The problem of the availability can be overcome by reproduction (microfilm, facsimile). But in order to be able to use a microfilm, one must have access to a microfilm reader. Facsimile editions of a number of the listed treatises do indeed exist, but
2. such reproductions do not solve the problem of poor accessibility. If the original sources contain any indexes at all, these are not up to modern requirements. Modern editions — of which only few exist of Italian sources, however — often improve accessibility by the addition of indexes, but the improvement remains a 'passive' one, which the user must take as it stands and cannot adapt to his or her own wishes. Digitising the source text provides an answer: an electronic text can be indexed and enriched by the user, either manually, semi-automatically or automatically, and can be queried in any manner one wishes.

Digitised corpora already exist for several disciplines; theology and classical and modern literature are but two. In the area of musicology, the Thesaurus musicarum latinarum (TML) has already been mentioned. It contains ASCII texts of Latin music-theoretical sources. Most of these corpora are encoded as 'flat text', to which only 'milestones' are added for orientation. This has the advantages of simplifying text searches and easy exchange between different platforms. But there are also a number of disadvantages to this manner of encoding:

• non-standard characters are difficult or impossible to render (for example accented letters and music notation);
• details of text layout and structure, which often support or even clarify the meaning, are not encoded;
• only simple searches can be performed on the texts, since mark-up (an indication of the structure and content of a text) is minimal or non-existent;
• images that illustrate the text are not included or are separated from their context;
• there is no access to a (digitised) facsimile of the original, and thus no possibility to check the correctness of the transcription.

Return to top of page

4.2. The solution

The goal of the TMI is to improve the availability and accessibility of the sources by rendering them in electronic form and by presenting them as a complete research tool. First, each source will be fully digitised, resulting in a digital facsimile. Next, a machine-readable edition will be made of each source, which will preserve the entire document's content in its original context, enriched by extensive mark-up. A powerful, user-friendly software environment will be developed for the consultation of these materials. The World Wide Web will be employed for the distribution of the TMI.

Music treatises are 'tri-medial'. They contain text, images, and music. Each medium requires its own method of representation. Text will be transcribed and enriched by mark-up. Images will be represented as bitmaps, and will be made accessible through a standard iconographic classification system. Music notation will be stored as an image, but also as notation code so that it can be printed, analysed and played with the help of suitable software. These three media will be integrated into one hypermedial environment.

Such a hypermedial representation eliminates the disadvantages of more primitive systems while retaining their advantages. Specifically, the advantages are:

• the materials can easily be distributed and consulted without specialised instruments such as microfilm readers;
• facsimile and modern edition are available simultaneously;
• the ability to query the source 'actively', according to the user's specification;
• the traditional 'linear' reading of the full text with images and music examples remains possible;
• the presentation conforms closely to the usual reading strategy of users, which is 'non-linear': Treatises are most often read along 'paths' determined by internal and external references, quotations and allusions in the text, and by the reader's questions and associations. Hypermedial representation offers a model in which these reading paths are easily recorded, followed and adapted;
• some hypermedial applications allow the user to add annotations, links and bookmarks to the documents. The TMI CD-ROM has a facility for these, organising them in so-called 'personal webs'. It is technically possible — and an aim of the TMI project — to allow such enrichment by the user to be carried out on the TMI server. In this manner research results can be immediately shared by the scholarly community.

Return to top of page

4.3. Transcription and primary enrichment

Transcription and primary enrichment of the documents is the greater part of the work. The result will be an electronic edition in the form of an SGML application. SGML (Standard Generalized Markup Language) is the ISO standard for system-independent encoding of documents. SGML is a set of rules which allows the user to define a specific mark-up system for each document type in order to record the structure and content of a text in as much detail as needed.

Although the World Wide Web will be used for the distribution of the material, HTML (the langage for encoding 'pages' on the World Wide Web) will not be used for the mark-up of the documents for several reasons. HTML is currently too primitive to encode source information in sufficient detail. Extensions are being continually proposed and implemented in browsers, but these reflect Internet fashion rather than research requirements. And contrary to SGML's original purpose, HTML concentrates more on document lay-out than on document structure.

Instead, the TMI will employ a specialised SGML mark-up system for the representation of historical source materials developed by the international Text Encoding Initiative (TEI). In addition to indicating its structure, this system allows the recording of layout and physical properties of a source, which are also of great importance for research. This system also allows primary enrichment through the mark-up of all kinds of content elements. For example:

• names (composers, authors);
• internal references: to pages, chapters, music examples or illustrations;
• external references: to compositions or writings (explicit links can be provided if these are in the TMI);
• terminology and definitions;
• variants in different editions, or manuscript notes in the documents made by other readers.

One can have access to both the original text of the source as well as an 'edition' with corrections and additions by a modern editor simultaneously. It is also possible to link the results of modern research to the text, and to allow the user to add comments or self-discovered links between text parts. By these means the static electronic text develops into a dynamic and open hypermedial information system.

The user may switch between different views of the source with only minimal effort: between facsimile and edition, or between 'plain text' and text with comments and annotations. The addition of TEI-mark-up makes the answering of complex questions a simple matter. One can for example search for:

• quotations from the works of a specific composer or author;
• names or terms which occur in a particular context;
• common printing errors;
• terms, either defined or undefined;
• available knowledge about a certain passage of text, music example or concept.

A universally accepted standard for the electronic representation of musical information does not yet exist. The most consistent and most widely-used system is DARMS (Digital Alternate Representation of Musical Scores), which however is dated in many respects. A newcomer to this field is the recently-approved ISO-standard for music representation: SMDL (Standard Music Description Language). SMDL is an SGML architecture, and since the TMI is designed around SGML, the logical consequence is to use SMDL as well. However, there is little expertise world-wide with SMDL as yet, and the standard seems to provide too few specific means for the encoding of musical notation in the same detail that the TEI allows. Proprietary formats, as are used by must music publishing programs, are clearly not an alternative. Therefore, DARMS code has been selected as a music representation in the first phase of the project. In the second and third phase either SMDL will be employed or (what seems the most likely solution at present) a music-notation module for the TEI will be created. Because of the consistency of DARMS code, it will be relatively easy to convert DARMS examples to a new code. Encoded music of almost all kinds can be converted to playing instructions (MIDI, Musical Instrument Digital Interface), for which software is widely available. It will thus be possible both to see and to hear the music examples, which is especially useful for didactic purposes.

Images will remain available in their original form as a facsimile. Diagrams will be 'transcribed' whenever this is useful. In addition, the image materials will be made accessible through the standard iconographic classification system ICONCLASS, so that they can be queried. Images and diagrams can easily be integrated into the SGML application, as can music notation code and MIDI.

Return to top of page

4.4. Target group and importance

The target group is 'professional' by nature and includes the following individuals and institutions:

• libraries of universities, conservatories and research institutions;
• researchers and teachers of musicology and music theory at universities;
• graduate and postgraduate students (AIOs, OIOs);
• musicians interested in historic performance practice;
• teachers and students of music theory in conservatories.

The TMI will be an important tool for research at the Department of Musicology of Utrecht University, since one of the focal areas of that department is the music of the Renaissance. Music from this period is studied world-wide, however, and it is certain that the TMI will attract many users from elsewhere. Many people have communicated their interest on the basis of the information on the website alone. Leading specialists in Renaissance music have given their support. Forty-five copies of the CD-ROM containing the demo prototype of the TMI were sold within two days at the Conference of the International Musicological Society in London (August 1997). The most recent information from the (Latin) TML shows that ca. 20,000 files were retrieved over a six-month period. It seems reasonable to assume that the use of the TMI will be in the same magnitude when it will be available on-line in its full proposed form. Since TMI will be used on a world-wide scale, English has been selected as the language for the software environment and documentation.

The project has however another important function, namely as an example for other projects for digital publication of source materials. It has recently been pointed out in the report De computer in het alfaonderzoek. Advies van de Commissie Geesteswetenschappen over de toepassing van de informatietechnologie bij het onderzoek op het gebied van de geesteswetenschappen (KNAW, May 1997) that the application of information technology to the humanities in the Netherlands is still too rare. The TMI will provide an example of how such application can be realised in a scrupulous and fundamental manner; and its methodology can be followed in other projects. But to be able to fulfil an example function it is essential to prepare a generous quantity of materials, for only in this manner it can one test whether the model is sophisticated and flexible enough.

Return to top of page