This past week, in thinking about the soundscape in terms of the poetry reading and the poetry phonotext, I went back to look at some of the initial information I compiled during the High Performance Sound Technology For Access and Scholarship (HiPSTAS) Institute at the University of Texas at Austin at the end of May. An NEH-sponsored gathering initiated by Tanya Clement, the HiPSTAS Institute brought together groups focusing on historical recordings, indigenous language preservation, poetry and phonopoetics. I wanted to take part in the institute simply to see what work people were doing across disciplines to critically approach the phonotext. Also, I was interested in simply getting together with the other poets and poet-scholars there to have the time and space to talk about these things face to face.

My initial application for the program was to focus on tactics for phonocritical practice. Over these last years, with the development of several digital sound archives and repositories – PennSound and SpokenWeb being the two with which I’m most familiar – there has been extended discussion on the archival and pedagogical implications of the phonotext, yet a critical vocabulary to approach the phonotextual object has only begun to be articulated. I wanted focus on certain strains of already emerged phonocritical practice that I have begun to write about, while learning as much as I could from those gathered at the institute.

With the introduction to a new software in development called ARLO (Adaptive Recognition with Layered Optimization) that would be experimented with during the institute, my interests shifted in an unexpected direction. ARLO was developed for classifying bird calls and using visualizations to help scholars classify pollen grains. ARLO has the ability to extract basic prosodic features such as pitch, rhythm and timbre for discovery (clustering) and automated classification (prediction or supervised learning), as well as visualizations. One example of ARLO was particularly interesting: an excerpt of William Carlos Williams reading from a poem was matched as having a tonal/frequency affinity with a short excerpt of Allen Ginsberg reading. A WCW poem and a Ginsberg poem have little in common in terms of their appearance and organization on the page, but sounded affinity makes sense knowing that both WCW and Ginsberg were both from the same area of northern New Jersey and that WCW was a major influence and mentor as a poet for Ginsberg.

This introduction enticed me to focus on one particular facet of phonocriticality: through the search and matching function of the ARLO software, I wondered if I might be to think about community poetics through sound, through a concept of sounded affinity. A few guiding questions to this were: If poets who share a common locale, a common dialect, place of community formation, syntactical style or rhetorical mode, might I be able to trace this out through ARLO? How might I use ARLO find affinities outside of schools or tendencies in poetic practice? Would ARLO be able to track affinities across gender lines? Finally, would I be able to use ARLO to perhaps map specific sounded features of distinct poetic modes or practices?

I have begun to compile test sets. A single analysis through the entire PennSound database can take over a day, and there are a number of kinks presently being worked out on the software in Austin. So, I will hopefully have more than an introduction to present in the coming weeks. Additionally, I would like to address the epistemological problem of utilizing an ocularcentric analysis of sound. For now, I’d like to present just a few shots of ARLO visualizations to present some of the phonotexts and their features I’ve been examining.

Visualization of speech from Spicer’s 1st lecture: “Yeats is probably the first modern to take the idea of dictation seriously.” This is an example of spoken language not in a lyric poem.

Visualization of Fred Moten’s poem “gary fisher“: “rodvan took me for doubles and doubles. / doubles and doubles and double and doubles in alleys.” Noting the visualized sonic slide between words, one that does not exist in the spoken language of Spicer’s lecture. Whereas the discrete unit in Spicer is visualized at the level of the word or phrase, here, the discrete unit is the line, the black unsounded gap being the (line)break.

N.H. Pritchard’s poem “Gyre’s Galax,” like Moten’s has that similar sonic slide across the line.

For the moment, I am noting the visualized sonic motion that is similar in terms of how the line carries forward in the Moten and the Pritchard examples. Might this sonic diffusion from word to word, phrase to phrase – as opposed to the shorter discrete units – across the line be a qualitative affinity between the two? Might I find the shorter discrete units of the Spicer lecture have a resemblance to, say, a David Antin talk?

This research continues.