After reading Neil Cohn’s essay entitled “Visual Lexicon” (he’s studying visual language via comics and speaking in Toronto this Thursday), I had basically no brain cells left to really reflect critically on his research, although I tend to have a soft spot for work that tries to bring cognition into the realm of the creative arts. In terms of my musical interests, Cohn brings up the point that music appears at first to have all the qualities of a Natural Language (instead of a lower case language) with the key exception that it does not convey concepts. The music advocate in me wants to object (what can’t music accomplish?!), but upon reflection, the only viable angle I really have is that while music does not convey concepts, we do have concepts about music in order for us to talk about it.
Music cannot be transmitted alone devoid of all concepts in that sense, but it is a far way from being a coherent argument for a “musical language” as Cohn argues for a “visual language” in human cognition. Ethnomusicology may be the perfect domain in which to seek answers regarding this issue, because they study music not only from the western tradition, but adapt a variety of ethnographic approaches to the study of all manifestations of music. In fact, now that I think about it, that may have very well become my topic had I pursued a masters in ethnomusicology. Now, since I am not doing a musicology degree, what exactly then, is my interest in VL and related research in the context of music information retrieval? I’m not sure, but here are some thoughts that crossed my mind.
For one, we do need to have conceptual models of “translating” music into a “language” that serves for easy retrieval from the systems perspective, as well as from the human perspective. There may not be a formal “musical language” that is identifiable or qualifiable, but is there a universal way (or a few broadly used forms) of music cognition that exists and can be applied in IR system design to improve access and retrieval of musical information?
More brainstorming later. However, for now, I have a more nuanced understanding of the challenges that come with trying to map “linguistic” features on top of something that only fits the layman definition of language.
