Current Status
tempus fugit ...

22 August 2004

This website has been neglected, because I have been deeply involved in the NEUMES Project during the past few years. When I can find some free time, it is my intention to revisit my work on this site and make improvements or updates to it. In the meantime, it continues to provide an accessible introduction to neumatic notations and a still-valid overview of my long-term objectives in this work.


13 May 2001

I have been remiss in maintaining the Web site for the Neume Notation Project, because (a) I was engaged for two years as Assistant Professor of Mathematics and Computer Science at Suffolk University, and (b) since then I've been preoccupied with developing a theoretical framework of data representation for "nondeterminate" symbol systems (of which neumes are an example).

Recently I've had very productive discussions with Prof. Tom Kelly concerning the functional requirements of the Neume Notation Editor. He points out that typically it will be graduate students in musicology who will do the data entry for manuscript transcriptions. It is likely that most such students will not have the expertise in medieval musicology required for doing an "intensional" transcription; instead, they will likely be capable only of "alphabetic" transcriptions (i.e., simply identifying neume forms on a document without interpreting their meanings). After discussing this with Prof. Gagliardi and Dr. Jeavons, the proposed solution is as follows:
  1. the 'alpha' version of the Editor will be intended for use and evaluation by professional musicologists, who can be expected to do the interpretation necessary for an intensional transcription;
  2. the 'beta' version will include interpretive software working in the background to translate an "alphabetic" transcription to an intensional transcription;
  3. the ambitious task of writing the interpretive layer of software will not be wasted effort, as the OCR program will essentially provide the same thing a graduate student can do--an "alphabetic" transcription--and so, the interpretive layer should be reusable as a back-end for the OCR program.
My scheme for the interpretive layer is to use a knowledge-based system (or, "expert system"), which always includes an inference engine and a rule base. For compatibility with the other software and for platform independence, I'll likely use the Jess (Java Expert Systems Shell) inference engine (which is written in Java and for which the source code is available). Jess can read rule sets written in CLIPS format. The idea will be for me to interview an expert in each supported neume-notation style, and encode as rules the heuristics s/he uses for interpreting neume forms. There would, then, be a separate rule base for each notational style. Transcriptions entered in "alphabetic" form would be interpreted by the rule base, and the corresponding intensional data would be written out to the data stream. Currently Jess supports only forward-chaining inference (not backward-chaining); this will be okay, since the basic pattern of reasoning during transcription is to draw diagnostic conclusions from the evidence. Support for certainty factors (CFs) in Jess is quite limited, and so I'll likely have to modify the inference engine for CF computation.

In February 2001, Prof. Cheatham and I went to Oxford, where we had a number of excellent meetings. Of special interest was the evening we spent with Dr. Julia Craig-McFeely. She is doing most of the technical work for the DIAMM project (Digital Image Archive of Medieval Music). She has been able to achieve some remarkable results in image enhancement of degraded manuscripts by means of high-resolution digital photography and layered, color-separation filtering in Adobe PhotoShop. She has already photographed some unheighted neume material at Worcester Cathedral (ms. F173). Also, some lower-resolution images, but perhaps having a wider variety of neumes, have been made of the Wincester Troper (housed at the Parker Library, Corpus Christi College, Cambridge). Prof. Caldwell will inquire as to whether we can use some of these images for OCR tests.

In November, 2000, Prof. Cheatham, Dr. Bob Walton, and I met with Dr. David Stuart (Peabody Museum and the Department of Anthropology, Harvard University) to discuss whether other writing systems exhibit Y-axis semantic significance, as is found in neume notation. Stuart is an expert on ancient scripts and one of world's leading authorities on decipherment of Maya hieroglyphs. Our understanding from the conversation is that Western music writing is perhaps unique in this regard. We did, however, find common ground with Maya hieroglyphs regarding uncertainty of interpretation; outside the bounds of the Neume Notation Project, I might use the Maya writing system to test the generality of my data representation theory for nondeterminate symbol systems. (Incidentally, I found a very informative, high-level treatment of early writing while in England: Andrew Robinson, The Story of Writing; Alphabets, Hieroglyphs & Pictographs, [London: Thames & Hudson, 1995].)

Most of my efforts at this juncture are concerned with finalizing for publication some papers about my theoretical model. Prof. Kelly, Prof. Cheatham, and I are trying to get research funding to support production work on constructing an exhaustive taxonomy of neume forms and proceeding with the Neume Notation Editor program.


Revision date: 13 May 2001
Copyright © 2001, Louis W. G. Barton