Research Narrative
ACADEMIC BACKGROUND
My academic preparation includes studies in Brazil, USA and France in Linguistics – especially in Phonetics and Phonology –, Literature, Second Language Acquisition, Language Pedagogy, Heritage Speakers, Network Science, Socio-Linguistics and Speech Processing. My Ph.D. dissertation adapted to Brazilian Portuguese Dennis Klatt’s model on segment duration of speech sounds for speech synthesis, developed for American English at MIT. My study of Klatt’s work showed me how models can reveal and predict speech patterns in discourse. Although Klatt’s work was developed for artificial speech, my experience with his work allowed me to have better educated guesses of how natural languages work.
CURRENT RESEARCH
My current research focuses on the transcription and analysis of speech prosody using musical notation. Prosody is a cover term for intonation, tone or pitch, duration, amplitude or loudness, timing or rhythm, sound quality, stress, accent, and phonological processes. Musical notation and musical theory have been used in the past in the transcription of speech. The pionneer work in this area of study is Joshua Steele’s (1775) An essay towards establishing the melody and measure of speech, to be expressed and perpetuated by peculiar symbols. London: Boyer and Nichols. 2nd. Edition 1779, Prosodia Rationalis. London: Nichols. Composers such as Bach, Beethoven, Schoenberg, Reich, and Janaček have represented speech using traditional scores, but there is nothing comparable to Joshua Steele’s essay. My research using musical notation, now joined by other co-workers in the US and abroad, is the result of a new and innovative methodology in speech analysis, aimed at a better understanding of how speech prosody works, as well as at the improvement of applications that can be developed with prosodic studies.
THEORETICAL ISSUES
One of the main issues with research in speech prosody is the different views of different Schools of Thought and the resulting methodology used for analyses. The development of prosodic studies reflects the elusive nature of speech prosody. The disparity of approaches to analyze prosody and the almost chaotic proliferation of a non-standardized terminology in prosodic studies to date confirm the difficulty that we find when studying any area of prosody, especially speech rhythm and intonation.
THE SOLUTION TO THESE ISSUES
My current work is intended to solve these issues. Musical notation, contrary to other transcription systems, is fairly universal, especially in the Western World. It can point out language behavior in a way that allows for effectively transcribing and searching speech patterns. Musical notation is designed to represent both the dynamic and static events of music and by extension, speech prosody. This is so because we can interpret notes as static events, given the existence of micro-movements in speech. Note durations, however, are dynamic, since they represent time with exactitude. Even the dynamic intensity can be inferred in MIDI data through velocity, namely the greater the intensity, the greater the MIDI velocity. Musical notation can bring us insights to predict and understand rhythmic and intonational patterns in speech.
MY THEORETICAL VIEW
I have an enthusiastic interest in linguistic studies. My view of human language systems is primarily dynamic. This is so because language is not only speech sounds. Language is also gestures. It connects to contexts and to our body to communicate ideas. Language is a dynamic system of intentions. As Wilhelm von Humboldt already stated, language is a live organism. Although language systems are mainly dynamic, some of its elements are quasi-static or quasi-linear. It follows naturally that I do not agree with the emphasis given to language as a primarily static system or a system of speech events in series. With this vision, I have the firm purpose of pushing my research to its edge, to look for solutions through unconventional approaches, and challenge prevailing paradigms.