I recently got word of a new Manding-language publication in Brill's "African Sources for African History" series: "The Epic of Sumanguru Kante" edited by Stephen Bulman and Valentin Vydrine, INALCO, Paris.
I won't go into any details of the tale here, but the book is an annotated bilingual typesetting-translation of a 1997 recording of Abdoulaye Sako. The typesetting, annotations and linguistic commentary are first-rate. Linguistically, the oral performance is represented in a semi-idiosyncratic orthography that fully marks tone. It should be readable for anyone familiar with standard Bamanan orthography, though the editors in no way attempted to "standardize" the transcription. As such, one finds a range of dialectal forms typical of the Banamba variety of Bamanan that appear sometimes regularly, sometimes interchangeably throughout the performance (e.g., twá : tɔ́gɔ 'name'). Most impressively the text also includes a large number of conventions for marking "discourse features" related to intonation and prosody:
- final-rising pattern (marked by \)
- sentence-initial tonal bounce: low tone for first element, sharp rise on second element (marked by double-underlining)
- partial sentence-initial tonal bounce: lowering of the first element's underlying tone
- anti-downdrift: an entire line is pronounced at essentially one tonal register (marked by a square root symbol)
- elevated register: above typical frequency range of 130-250 Hz (marked in bold)
- low register: below typical frequency range, often low volume (marked with a dotted underline)
- downdrift sequencing: distinct resetting phases of downdrift that occur in one long line (marked with an up-arrow)
- up-tempo: fast speech with high degree of phonetic reduction (marked in italics)
- false start: (marked by an equals sign following the segment)
- hesitation: (marked by double angle brackets)
The amount of phonetic analysis that had to go into marking all of this is quite impressive. Unfortunately, in practice it isn't something that most readers will pay much attention to; in fact, I imagine that most of these features would not recognizable to people unless they heard them (even then it might be difficult). On that note though, you are in luck!
One of the best parts of the book's release is that it includes links to an open-source repository of the recordings in m4a audio format. For many purposes, the book's typesetting and translation cannot truly be appreciated without recourse to this recording, so the fact that it is easily accessible and is freely available is awesome. On the other hand, the files being stored on the open source data website figshare makes one wonder -- is a book with separate audio files really the best format for this project at all? Of course, there might not be a good way to include footnote annotations and commentary, but with the work that went into the book we could easily create a black-screen video with separate subtitle files (one for the transcription and one for the translation) that users could select or switch between while watching. In fact, if the transcription and translation were done with an open-source tool like ELAN and put onto the figshare repository for the book than I could put such a product on YouTube in the next five minutes. Of course, people might be less interested in purchasing the book in that case, because they'd have free access to the transcription and translation. In that case, I guess one argument for the book still existing or being desirable would be the annotations, commentary and introduction etc. And of course, published books get people jobs, tenure and the like. Unfortunately though, annotated and translated books that fully mark discourse features are neither the best medium for a post-facto audience nor do they give researchers access to the open-source data that would allow them to perform analyses on things like "sentence-initial tonal bounce".
So for now, I am listening to the recordings on my Mac using InqScribe while following along with the digital version of the book from my tablet. Not going to lead to open-source data for all but it does allow me to quickly jump back 8 seconds (for repeats) and to add linked cues for page-turning...