Skip to main content


Literary Archives in the 21st Century

Al Filreis, Jason Camlot, Steve Evans

Between April 26-27, 2013, the Whitney Humanities Centre at Yale University brought together a group of literary and information science scholars, historians, curators, archivists, writers, and publishers for a two-day symposium in the Beinecke Library, where they listened to papers and panel discussions and engaged in intensive discussion about the status of the literary archive in the 21st Century. Drawing on the specific perspective of their respective institutional and disciplinary affiliations, panelists explored the collaborations between library professionals and scholars around the use of manuscript material in teaching and research, the intersections between archival and literary theory, and the impact of the changing shape of archives on institutional stewardship and scholarship.

For the “Beyond the Text” symposium, Al Filreis (University of Pennsylvania), Jason Camlot (Concordia University), and Steve Evans (University of Maine) were invited to conduct a panel discussion that considered the literary Sound Archive. What follows is a textual translation of the original discussion held on Friday, April 26, 2013, 11:30-12:45, in the Whitney Humanities Center at Yale University. By textual translation we mean something beyond just transcription; we mean a collective re-writing (with corrections, additions, and transformations) of the original discussion, a sound recording of which can be heard here.


Al Filreis: We’ll begin by introducing our projects and ourselves. I’m the co-founder and co-director, with Charles Bernstein, of PennSound, an audio archive housed at the University of Pennsylvania. PennSound has two important corollary projects associated with it: the Kelly Writers House, a physical presentation space that’s not inside the university’s library and is not archival; and Jacket2 magazine, an archival publication through which many of the PennSound materials are annotated.

Jason Camlot: I have been publishing articles and delivering conference papers on early spoken recordings, by which I mean acoustic recordings made in the early era of the perfected phonograph (starting in 1888), as well as early flat disc (Berliner) and aluminum disc recordings made up until the 1940s. This research initially brought me to many different sound archives, including the British Library National Sound Archive, the Recorded Sound Division of the Library of Congress, and the quirky, and largely undocumented collection known as the Vincent Voice Library at the University of Michigan in Lansing. When seeing to hear rare recordings a decade ago I would often find myself in a listening room or booth at one of these archives, speaking to an archival engineer on a different floor of the library through an intercom system, asking him to play the tape dub that was made of the original wax cylinder. I would listen intently, scribble notes based on what I was hearing, but was ultimately constrained by the fact that I could not take the sound with me. I could ask the engineer to play it again, and again (and sometimes I did that), but in the end, what I left with were the silent, written notes I had taken about what I had just heard.

Things have changed significantly over the past several years. I have been working with sound archives for so long that I can now Google and stream digital versions of many, I’d even venture to say, most of the recordings I once listened to on site in the manner just described. So my relationship to sound archives has changed in interesting ways. I visit the archives that hold the recordings less for the purpose of listening to the recordings themselves. When possible, I have the recordings digitized and work with the audio signal in my own home city, on my own computer, on my own time. I continue to visit the archives though, increasingly to examine the material artifacts that hold the sound, for the metadata that sometimes is inscribed upon those artifacts, for the sleeves and related documents that are sometimes kept with the recordings, as well as for the sound recorders that are sometimes kept in the archives along with the recordings they generated. For example, the Columbia University Archives that hold the aluminum discs that comprise the audio collection of William Cabell Greet (the audio itself can be heard at the Library of Congress) also holds two aluminum instantaneous disc recorders that were used to make the actual recordings. The sound archive has become increasingly a site I visit for the perusal of silent material artifacts as much as for the sounds themselves.

More recently, I have been involved in a material to digital archival sound project of my own. The process has involved working with a corpus of reel-to-reel magnetic tape materials that document a poetry reading series held in Montreal between 1965 and 1974 presently housed in the Concordia University Archive, having these tapes digitized, and then thinking about how best to present the digitized audible materials for use by literary scholars, students and listeners at large. The project has led to the development of a Digital Audio Poetry Archive called SpokenWeb, which is something more along the lines of what Matthew Kirschenbaum describes as a kind of “grassroots” website, than something worthy of the magisterial name of Archive. We are using the development of the SpokenWeb site as a case study to explore questions around digital presentation (navigation, visualization) of audio materials, the documentation and historicization of literary events, the integration of oral historical methodologies into such acts of historicization, the curation of public events that mobilize the contents of the audible archive, and, of course, the opportunities and implications of digitizing an analogue audio collection. This process has raised myriad questions about the ways In which we engage with literary audio recordings, artifacts that are, by their very nature as temporal media, challenging to engage with critically.

Steve Evans: I’m currently working on a book project on phonotextuality – “The Phonotextual Braid: Timbre, Text, and Technology in Recorded Poetry” – in which I attempt to gather my thoughts on a topic that I’ve been writing and talking about, off and on, since the late 1980s. I also did an online project, through my website Third Factory that I called Lipstick of Noise, which I used as a kind of way of curating sound materials and giving its users a newsletter of what’s out there. “Lipstick of noise” is a phrase that I first heard in the beautifully French-accented English of the philosopher Jean-Luc Nancy when I was an undergraduate at UC San Diego in the 1980s. Back then I was under headphones preserving – thanks to a Department of Education grant, I think it was – audio files that were decaying, including a large cache of reel-to-reel recordings that Paul Blackburn had recorded in the 1960s. So I was under headphones at that time doing preservation and cataloguing, basically creating a database, and when I took up again this issue of phonotextuality, or how timbre, text, and technology braid together in the phonotext, I was reminded of sitting under the San Diego skies listening to Jean-Luc Nancy try to negotiate American English, then new to him, and give a talk on Eros and other issues, and he dropped that phrase, “poetry is the lipstick of noise,” and all of the poets fell in love with him instantly. He’s very good on this issue – what do we hear in the voice that does not particularly belong to the symbolic order, to the code – what are we listening for? This project has been integral to my interests and research for a long time and it just keeps changing. How do we frame and call attention to the principles of inclusion and exclusion that create the phonotextual archives that we’re familiar with? There are a lot of issues about what information travels with sound files. In 2007, when Al and I first talked, we were getting used to modes of listening and the phonotextual experience. Now, increasingly, people are seeing the edge of the PennSound project – they wonder about the principle of inclusion and exclusion, what kind of data is travelling around with these files.

Filreis: Exactly. Metadata is one of the issues that we need to discuss. But let me begin by asking you a question, Steve. You allude to how the experience of a “listener” is distinct from that of a “reader.” Can you point to three differences?

Evans: Yes. I think Jason’s experience with the vinyl culture around poetry and with the record is so fascinating – but that’s just one of the technologies. Most of the archive material that we deal with that wasn’t born digital was born magnetic tape, so you’re hearing the timbre of a human animal somewhere in their trajectory towards death, and you’re hearing a technology that is on a similar trajectory towards supersession. Those of us interested in recorded poetry are always looking at a parasitical use of a dominant, usually a capital-driven process, so these technologies are being rapidly phased in and out. Poets are making innovative uses of them that might last a window of time of five years or fewer, sometimes more. When we listen to audio recordings, we’re hearing timbre (the person speaking) and making assumptions about their state of ability or disability, their state of comfort with the event, the amount of what Erving Goffman might call their “fit to frame.” The question becomes: What about the person speaking? Is he or she getting the ceremony that they’re part of and if not, are you hearing that in his or her voice? Now we’re generating a new text. I’ve been thinking a lot about the poetry reading as a unit of publication. It’s a syntagm that the poet usually selects down from the wider paradigm of their work-to-date. In the phonotext, we are often hearing work being “published” many years in advance of its print manifestation. There was a period where Rae Armantrout’s books were stacked like planes over O’Hare – and the recorded readings available through PennSound provided a way of hearing her work three, four years in advance of the print record. I think it’s caught up now.

Filreis: Because of PennSound’s relationship with Rae Armantrout, I think we’re getting her books out there as recordings and we are segmenting whole readings by the poem, with poem titles that have been confirmed, at times months or even years before the book is published. So Armantrout fans can come to the archive and get a sense of what her poetry is about as it evolves in the present. Because of the usual delay of print publication, there is a certain understandable frustration that somewhat abates when you have access to an archive like SpokenWeb or PennSound, where one can get a much more contemporary or co-temporaneous sense of the sound of voice.

Evans: Vocal chords “season” with time, recording technologies – these things change. So how do we utilize some of those fan-based, crowd-sourced options that Matt Kirschebaum was talking about, in terms of feeding that kind of information back into something like PennSound or SpokenWeb?

Filreis: Jason, we were seated together during the previous panel, and I think both of us were thinking a million things about the ways in which those fan-based, crowd-sourced options are analogous to the kind of work we’re doing, outside of libraries. I’m sure during Q&A we’ll talk about our relationship with our respective libraries, but I would consider myself and Charles Bernstein and the staff at PennSound as doing our work outside of a library. Can you tell us one or two things you were thinking of while you listened to the previous panel?1

Camlot: I have been working with literary recordings or spoken recordings in ways that can be identified as both literary historical and media historical. As a Victorianist, and as a researcher who has studied and spent time listening and re-listening to early acoustic cylinder recordings, I think timbre is very much linked to the medium of recording itself. For example, if you are in an acoustic recording environment, the audio recording technology and respective media format that is used to capture and preserve an audible signal is most certainly going to inform timbre of the voice that we experience upon playback. The same can be said for recordings made and experienced within a digital environment. I am interested in thinking of creative ways to perform acts of comparative listening, apart, of course, from thinking about what it meant to record and engage with a recording technology in the 1890s or 1930s, and the impact on reading culture in those periods. A key question whenever you encounter an audio recording or spoken recording is, What exactly are we listening to? It seems to me that recordings, especially when you encounter them on the web or hear them via YouTube, are often stripped out of context in many ways, and I am very interested in trying to think of the different methodologies we can employ to help re-contextualizing these audio signals, to ground them again in the material, aesthetic and social structures that preserved them and informed their production and use. This might include an understanding of the symbolic import and practical constraints of the recording technology; it might involve research about the elocutionary, rhetorical or conceptual poetic protocols that informed how a poet decided to read his or her poetry in a particular way; it might be imagined in relation to an author’s life and the changes that are revealed in his or her recordings; it might involve deciphering the social environment in which a reading occurred, or that led to its preservation in a sound recording medium, et cetera, et cetera. In the SpokenWeb project, I have had the opportunity to work with librarians and archivists who have been a great help to me in exploring such questions, in developing approaches to the study of sound recordings, and, indeed, to defining what questions need to be asked in the first place. I was hired at my university in 1999. When I was called into the Chair’s office for my first meeting, I noticed a collection of reel-to-reel tapes on the top of his bookshelf. I was curious. I pointed to them and said, “Well what’s that?”

“Oh,” he replied, “those are some tapes of a reading series that happened here in the ‘60s and ‘70s.”

They were just sitting there on his shelf. They’d been passed down to him over the years, like the gold watch in that famous scene from Pulp Fiction. The question arises: Whose obligation is it to do something with these things? These tapes sitting on a dusty shelf in a department chair’s office, those Edison wax cylinders of Alfred Tennyson reciting his verse found decades after they were recorded, somehow, in a barn in South Africa, or the hundreds of tapes (reel-to-reel, cassette) sitting in shoeboxes in attics and basements across North America. Whose obligation is it to take custody of these cultural artifacts? How do we locate them, and what do we do with them once they are located. As part of the SpokenWeb project, the collaborating librarians composed and conducted a survey to get a sense of how many hidden collections of literary recordings there were in Canada, what state these materials were in and what the institutions and individuals who possessed them hoped to do with them. Enough to say for now that there are lots of shoe box treasures out there.

As far as the story of these particular reel-to-reel tapes goes, my department chair decided to deposit them in the University Archive. The archive received a grant to digitize them and then put them onto archival compact discs in WAV form. But, as far as I was concerned, this was still a perfectly useless collection – hundreds of hours of audio that someone would actually have to listen to in order to find out what is there, and that would be just the first step in an attempt to determine what it all means. So I embarked on a project of trying to develop a catalogue for these materials. I had an old fashioned, printed catalogue in mind (not unlike the pages of transcripts produced by oral historians from the hundreds of hours of interviews they perform). The idea was to create a guide so that someone could begin to navigate this audio collection and see what it was all about. Initially, then, the SpokenWeb project was a transcription project, the aim of which was to transcribe all the extra-poetic speech (the poetry banter, introductions, etc.) heard on the tapes, and to time-stamp all of the reading sections, so that we could have a map of what these hundreds of hours of sound looked like.

Filreis: How did you get the transcripts?

Camlot: Graduate students.

Filreis: Oh my goodness. Can we please have reliable speech recognition software – now?

Camlot: I have been experimenting with it. But for the time being, my approach has been very DIY. We have used transcription software that allows the transcriber to time the audio to his or her own typing speed (without altering the pitch of the speaker’s voice), and that automatically timestamps a segment every time one presses Return. While real-time (or slightly altered time) transcription is work intensive and time consuming, it is also extremely informative for the listener. Typing down what you listen to is an interesting exercise in intensive listening and it has created a small pool (a typing pool, actually) of experts on the audio archive in our possession. Our overall aim has been to develop a way to work with these materials so that people – and this leads back to the point about all of these pockets of similar sorts of archives that are out there in various forms – so that people can begin to make them available, and so that the materials may ultimately be aggregated and jointly searched, and thought of in relation to each other.

Filreis: Let me follow up on this point and tell you a little bit about PennSound’s process. You have two archives here – there are perhaps five or six such in the poetry world – but these are two good representatives. So indeed if we’re typical – and I think we are – many people in the room will be horrified to hear that our approach at PennSound, big as it is, is also DIY. Charles Bernstein and I promulgated a set of principles in the beginning: there were a lot of recordings out there, in many formats and various stages of decomposition, and we felt we needed to get these as fast as we can before they fully degraded and to digitize them and flatten files into mp3 even at the risk of not saving the higher-quality .wav file – because at first we didn’t have a big enough server to hold on to the .wav files – and just get these out there and use this handy thing called the World Wide Web. We moved quickly and were willing to risk losing or at least deferring the making of a preservation copy. We were in a hurry. In this specific context – the urgency of the situation – I have a lot to say about the inevitable slowness of libraries, perhaps during Q&A.

Evans: I hope there are no librarians . . .

Filreis: So we went out there on the operating principle: let’s just do it. We put recordings up – first whole readings, sometimes even unidentified. We might indicate, for example, “Blackburn recording, unknown date, unknown place,” and instead of hiding that ignorance we made it quite visible on the site so that others who might have time and knowledge of Blackburn’s life and work could help us make changes as we went along. We put some recordings on the site – made them available to all for free, with no barrier to entry – before we had anything close to full audiographical information. One of the things about getting PennSound up on the web there early was that people began to know about it and began showing up with materials they’d saved or stored. We found out about basements full of reel-to-reel tapes. We acquired entire poetry series on cassette tape, which had been sitting rotting in someone’s office somewhere. People began sending us boxes of unsorted materials and our goal was to just get through them as quick as we could, and to make it all available. And when Robert Creeley’s son Will came up the stairs of my third-floor office of the Kelly Writers House lugging four boxes of reel-to-reel tapes that his father had made, I sat down with Will, and I began asking him how much he knew about what was on those tapes. He knew little. He just knew that his father wanted those tapes to be in his hands. Four boxes. It took us only three months to go through all of them, to digitize them, to have a lot of different ears of colleagues listen to them, identify them as best we could, and, once again, put them up. And now, a few years later, we’re finally refining them, segmenting them. There are gems in there. You can only find the gems when you segment the whole readings. Five different performances of the most anthologized Creeley poem (“I Know A Man”) emerged from those boxes, in addition to the six others that we had already known about. Now listeners to PennSound can listen to Creeley perform that poem through his entire career, and can begin to look at the changes in end-stopping, in stuttered pausing, in the way different audiences interacted with him, and on and on.

So, to repeat: we just wanted to get it all out there, to see what the community of poets and scholars and veterans attendees of readings might make of the individual recordings. I suspect that Jason has had a similar experience. And when Steve puts his discoveries up on his site Lipstick of Noise, the implied sentiment is: let’s get it out there and have the wisdom in the room, the large internet room, figure out what to do with this stuff. Am I right? Are we agreeing on that principle?

Evans: Absolutely. But at this point a lot is up there. I like the PennSound Manifesto. Make it MP3, make it available, make it downloadable. But there are all kinds of consequences that follow from each of those decisions. The DIY nature that kept the presentation from being connected to a historicization is starting to feel problematic. How do we supply some contextualizing frame for these materials? I think that with spoken word it’s particularly vivid, since in order for our spoken words to operate efficiently they have to fade, they have to die out of the way of the next word. Is that peculiarity of spoken language what also makes us so good at shedding information about voices? As long as we can retrieve that signal, we’re comfortable ditching the transmission source. If I played a sound file of William Carlos Williams to my students—something I’m constantly subjecting them to—it’s important to say: you are not hearing William Carlos Williams, you are hearing a digitized version of a recording, the context for which has usually been completely lost. You’re hearing through four or five layers of suppressed information. And I feel like that drive to historicize needs to – we need to find at some institution a way to bring that in and braid that in to the files a little more.

Filreis: I don’t disagree with you, but I want to press back a bit. I hope to hear a more explicit form of the criticism that was in the first part of what you said. I think perhaps you were being too polite. I had said let’s just get it out there – and I think we’re still in that phase. I think what you’re saying is: we may not be in that phase but in the next.

Evans: The one thing I’ve heard for a couple years now is: PennSound needs to articulate its principle of exclusion better. It’s a curated site, it’s an edited site, and yet there really is no statement of editorial practice. It’s not wide open, that much now seems clear. If you did a network analysis of it, it would be fascinating. There are several qualities that overlap and are captured there really robustly, but what about the ones that barely peek in? Getting everything up in mp3 was a genius move. It’s astonishing. But it may be that mp3 is heading toward eclipse as a format as well. So we need to start providing some information while we’ve got it.

Filreis: Yes! But let me please respond further to that, and look forward to Jason’s response. Probably the most radical decision that Charles Bernstein and I made when we started PennSound – which has now grown to some 80 thousand recordings – it’s a big archive – the most radical decision we made was to make our material downloadable. At the time, all the sound archives, especially at libraries, were streaming-only – no downloads. Now the mp3 file can be put on all kinds of different players. It can be mixed and sampled. It gets out of control. The next thing you know, nobody knows who’s providing the poems. We embraced that risk of nomadism and chaos because we wanted the recordings to be available to anyone, anywhere. My favourite moment in all of this experience occurred when I was standing in line a bit impatiently at a grocery store in Seattle, behind someone who wasn’t paying attention because she was listening to her mp3 player, and I mildly chastised her: “Can you please check out and stop listening to your thing?” And then I noticed on her display that she was listening to a poem from PennSound.

Evans: “A supermarket in California”?

Filreis: In this mode, notwithstanding potential incoherences, poetry will get heard across and outside institutions normally associated with the dissemination of poetry. This distribution is like the anthologized printed poem that’s been carried around in the wallet of the soldier who’s on the Eastern Front, which is a famous story told of a Claude McKay poem, “If We Must Die,” a radical poem that inspired someone of a generation different from the one addressed by the poet originally, in a far-off place. Now the poem is in the pocket of the guy stationed in Afghanistan. He’s listening to Creeley on the darkness that surrounds us, because it’s easily available to him. I’m sorry! That was a little sentimental. Jason, I hope you will give us your response.

Camlot: I want to come back to the question of context. It’s really important to me in the project I’m working on, and so are the editorial principles that inform our different approaches to the digital presentation of archival audio. I think PennSound is more of a curatorial project than SpokenWeb, although principles and concepts of curation certainly inform our work, as well. But a core principal that underlies our work is in line with what Jerome McGann has called “philology in a new key,” that is to say, in line with an interest in identifying the ground of an event or artifact, even as we acknowledge that the historical/archival artifacts we engaged with are open to new contexts as a result of their new media environments. One may curate and thus frame artifacts—sound recordings, say—aesthetically, in a manner that has little to do with the historical contexts or frames that originally informed them. Look at UbuWeb, for example. Just the design, the look of the site– the curatorial principles are articulated, defined, just there, in the visual design of the site. UbuWeb is completely un-permissioned, totally rogue, totally pirated, and avant garde in its visual style. So to a great extent one knows what UbuWeb’s principles of selection are just by looking at it. I love UbuWeb and think it’s a great example of how distinctive design can reframe and redefine historical artifacts. I have been more interested in archives as they have been conceived historically, or as they may exist implicitly in the body of materials that have been presented to me as objects of study. So starting, for example, with Edison’s imagination of an archive of the great voices of mankind preserved for posterity’s sake, as he articulated this idea in articles he wrote in 1877 and 1888, upon the release of his tinfoil phonograph and then the perfected phonograph – that is one early idea of an archive that is conceptually linked to a nascent technological imaginary. Another example of a coherent archival collection would be the one articulated by Bennett Maxwell in an unpublished catalogue he wrote for a series of early, pre-commercial cylinder recordings, called the “Incunabula of Recorded Sound.” These are recordings that were made in the 1890s by George Gouraud, who was an agent for Edison in London in 1889/1890. Gouraud was basically trying to market the phonograph before they had the rights to sell it in England, and he invited famous people to his house, got them tipsy and then had them record things into an Edison phonograph. He would then use some of these recordings to display and promote the phonograph in public venues. That cluster of recordings made by George Gouraud in 1890 is a coherent collection of sorts. It was defined or imagined by Gouraud for particular instrumental purposes. Some of the recordings were scripted testimonials, and the entire lot of these cylinder recordings, when considered together, provide an audible document of a particular moment in the early history of sound recording technology, and what a certain class of people felt it might mean for the future of the human voice, or in even more gravely Victorian terms, the human race.

William Cabell Greet, whom I mentioned above, made a series of recordings of T.S. Eliot (among several other poet’s like Robert Frost, Gertrude Stein and Vachel Lindsay). He was a linguist who had an audio lab, and he was the editor of the journal American Speech. He was recording poets’ voices, but he was primarily interested in collecting dialects, American dialects. The result of his work was a very large collection of American dialect recordings with a few remarkable recordings of poets reading their works scattered amidst the dialect records. There are multiple ways to contextualize these Eliot or Lidsay recordings, but one important way to do so entails an understanding of the historical context that led them to be produced in the first place, some residue of which can be felt and found in the collection of which they are a part. So, these examples illustrate one interesting way we may think about grounding audio recordings, historically. The corpus of recordings that are presented via SpokenWeb have their own historical grounding, or, rather, layers of grounding, and a good part of the SpokenWeb project has aimed to make these layers visible. The materials that I have inherited, those tapes that were on the Chair’s shelf and then deposited into the university archive, are documentary: a set of recordings from a reading series that existed at Sir George Williams University in Montreal between 1965 and 1974. The voices that were recorded consist mostly of Black Mountain poets, San Francisco Renaissance poets, West Coast Canadian poets, and local Montreal poets. So, while I am interested in engaging with the poem as an audiotext—and thinking about how a particular poet reads a specific poem—I am also interested in disengaging from that hermeneutical enterprise of close listening, and, instead, in pursuing other ways of imagining contexts for these recordings and events. I am interested in asking questions such as: What is a poetry series, how is it imagined, and how does the archive allow us to understand a little bit more about what a poetry reading series is? What is a poetry reading – the whole reading, not just the reading of a single poem clipped out of context? The poetry reading often consists not just of poems being read but of what I have been calling “extra-poetic speech”, that is to say, the voice of a poet lecturing, or meditating out loud upon his or her own poetry. Such poetry-event commentary represents a kind of exercise in contextual presence that invites literary historians to think about how we go about contextualizing the documentary media that have preserved such existential performances for rehearing in the present?

Even the question of transcription – I come back to those graduate students transcribing the audio for the creation of a SpokenWeb catalogue – can be understood as a question of re-contextualization. How do you transcribe something that is spoken? How do you script it, and how far do you go in attempting to script it “accurately” or “meaningfully”? Given that transcription represents a process of remediation, the mere act of writing down what one hears in a sound recording represents a complex act of cross-media interpretation. The transcription of a poem as we hear it read in a sound recording is not necessarily going to look like it does on the page. Oral historians are very familiar with questions about transcription and what it means; it raises all kinds of interesting theoretical questions about the relationship between text and audio. And there are many poets who can be heard via the SpokenWeb site – Jackson Mac Low, bill bissett, to name just a couple – who were deeply interested in questions of scripting sound as part of their poetics.

Filreis: I think we’re in similar positions. Differences are probably just a matter of scale. Our decision about PennSound was mostly to not decide. If we have a series, for example, we add as much information as possible. At the same time, if we have an author page, a “single author page” for someone who performed in that series, the series reading will link to the author page as well. If we’re segmenting the poems performed in that poet’s part of the reading in the series, we will segment them and copy all the links to all those three pages. So we’re trying to do what the web does so well, which is to allow links to be copied and crosslinked to create different contexts for access. In the example I just gave, listeners can enter the site from an interest in the poet or, equally, from an interest in the series. The virtue of the World Wide Web obtains: the web archive permits multiple points of entry.

Evans: We think we’re recording poetry, but it’s a bit like the anecdote about how you might turn to ‘60s and ‘70s hardcore porn to study the kinds of furniture that petit bourgeois people of that epoch had in their homes. We think that poetry is in the foreground and that’s what we’re looking at – I’m fascinated by what other information we’re capturing, perhaps unintentionally, and figuring.



Emile Rubino [audience member]: Thanks to the three of you for all your work, it’s really invaluable. I have two quick questions. First, can you provide a bit of a timeline? In 2001, through my then involvement in politics, I managed to write an Op-Ed page for the New York Times about sound archives, and I’m just wondering, what do you think has happened to the sound archive since 2001, and where were things prior? Second, at Lost and Found we’ve been doing some transcriptions of PennSound materials. I’d like to know how you feel about them and how they relate.

Filreis: I’ll start with the second question and then maybe Jason and Steve can handle the first. I want to say, Emile, that you also deserve thanks for the work that you’re doing, which is really a sister project to ours. When we launched Jacket2, inheriting it as “Jacket” from Australia, our goal was to use it in part as a venue for transcribing, for creating critical material, including audiographical material, and then using it as a commentary site for the PennSound archive. So yes, you know, we ask all the same questions about what a transcription is. (We’ve become very interested in the introductions that people give before a poetry reading, for instance, so we made an anthology of introductions to Robert Creeley from 1955 to 1994). Really, Charles Bernstein and I see PennSound as a site of launching, a way of getting materials out there so that the scholars and critics and librarians and archivists and audiographers can begin doing the descriptive and analytical and theoretical work, and you’ve done a lot of that, you and your colleagues. Steve, do you want to handle the first part?

Evans: In 2001 we would’ve been having a very different conversation. The audio archive was inert, decaying, hard to get at. The preservation project that I worked on as an undergraduate was, as it turned out, ill-timed. It was right before DAT had been generalized. We were essentially preserving audio onto a video band of a VHS cassette. Now we’re converting reel-to-reel audio to mp3 and we have that same problem all over again. And not only that but we’re probably destroying the reels also – one more play and you’re out. But thanks to UbuWeb and PennSound, and to the work Christina Davis is doing at the Woodbury Poetry Room, that Steve Dickison is doing at The Poetry Center at San Francisco State, the list goes on– there’s so much more poetry phonotext out there. I love a west coast project called “A Voice Box.” It’s got a smaller scale, and has more of a ‘house party’ kind of feel, but it probably wouldn’t exist without having seen what PennSound could do with a little more institutional juice. But let’s talk about DIY. I think there are scales of DIY. There’s DIY within an institutional frame, for example.

Camlot: There are a lot of people involved, I think. In a way what you’re describing and encouraging is an informal kind of usability study, because we’re still feeling out what people want to do with this stuff and what we want to do with it. And so, in opening up, making the files available, just getting stuff out there and having people send posts back, or corrections, etc., we’re beginning to collect information about the user, what people want to do with these recordings, but maybe too informally. I wonder what you think, Al, about this usability question. Can it actually inform our design?

Filreis: Let me say something somewhat provocative to induce some of the librarians to go to the microphones with their questions: We try to work with libraries, and they’re ready to go and it would be great – but it’s too slow – they’re interested in preservation copies rather than making copies ideal for distribution.

Camlot: Is that such a bad thing? Is preservation against circulation, here? I mean, is it preservation vs. circulation that we’re talking about?

Filreis: No, making preservation copies isn’t a bad thing – I said it badly. The goal of preservation should not inhibit distribution. We wanted something user friendly and immediate. Years ago it would have taken hours for someone on a dial-up connection to listen to a WAV file. We were all about distribution and access. So, before libraries getting grants, and agencies giving out grants to libraries that were digitizing audio, these materials had to stay in the library. We wanted to get them out there, to be streamed.

Evans: Streamable but not downloadable.

Filreis: I travelled around to libraries. I went first to a library and it had many recordings of really important poetry readings from the 1940s and ‘50s. I had been emailing and asking, “Can we do make these available? Can we do this?”

“Well,” they replied, “we’ve got a kind of bottleneck.” Even thrwarted in such ways within their own institutions, the archivists won’t let you borrow the vinyl so that we can make a digital copy and then return it to their preservation-first bottleneck. But within these under-resourced libraries, it’s going to be five years before important recordings make it through the bottleneck, because, for instance, the music library is the only one that has a single staff member who can do the digitizing of the vinyl at the quality institutionally insisted. “So come back in five years.”

I didn’t want to wait five years, so I got on a plane and traveled to the city and walked into the library, somewhat unannounced. The vinyl was sitting there, on top of the radiator of the curator’s office. I said, “Can we get the people around who can make a decision about some way in which we can digitize these materials?” I stayed all day until finally I had a conversation with the person in charge, and it was not entirely a friendly conversation. I said, “You have just one staff person who does this work to your standards, and it’ll be five years before it gets done. What’s wrong with this picture? Can we help? We’re trustworthy – can we work together to get this done before the old medium degrades further?” The answer was a polite “No,” but it allowed us to borrow the time of the one staffer in the basement of another division within the university’s library system who could do a little digitizing in a key collection. Because of that, recordings of a major modernist were suddenly, for the first time, available to scholars, students, teachers at high schools, and now the proliferation of critical material of the sort that Steve’s been advocating for years – the student’s sophomore papers, dissertations, honours theses, newspaper articles – many of which include a reference to the recording of this poet’s work. I felt it to be something of an emergency. I really wanted to get it done.

Camlot: Well yeah, you did get it done. And the vinyl still exists and wasn’t destroyed.

Filreis: That’s true. In fact, our collaborative effort led to the preservation of the vinyl by getting it off the radiator. But there was a sacrifice we made – putting it in MP3.

Unidentified Audience Member #1: I’m interested in the copyright issues in what you’re putting out there. Are you just waiting to see if somebody complains?

Filreis: PennSound is entirely permissioned. Everything we put up is permissioned.

Evans: But Al, maybe you should acknowledge that your meaning of “fair use” is distinctive. Think of a highly sueable entity like the Poetry Foundation in Chicago. They don’t dare to go downloadable, not without a lot more probable legal documentation.

Filreis: We have all the documentation.

Evans: I don’t doubt that.

Filreis: So what are you doubting? It’s all permissioned.

Unidentified Audience Member #2: Some of the recordings made on campuses back in the ‘60s and ‘70s did not have permissions to record, so that might be an issue.

Evans: The New Writing Series at UMaine is twelve years old and unfortunately we’ve not properly collected permissions along the way. We’ll have to do everything in retrospect.

Filreis: But even in retrospect, is that good enough?

Unidentified Audience Member #2: Probably. Maybe. I don’t know.

Camlot: In the SpokenWeb project, permissions means “research access,” so we actually have some things that are password protected. Of course, we’ve made some available, for which we’ve received permissions. But I think the question of who carried out the actual recording is an interesting and important question. Partly it tells an interesting story about technology in the university at the time. Our recordings were made by a newly funded and built up AV department, which was connected to the language labs – which explains the high quality of many of our recordings. The AV department also had a mobile recording unit. But it also has implications for who holds the rights to the recordings and their use.

Unidentified Audience Member #2: The other question I had was about the preservation issue: What are you doing to preserve PennSound?

So we are now preserving. We bought a big server, so we’re preserving WAVs now of everything we can, and we’ve backed it up several times, which is very important. I mean several times in the history of the project. There are several backups constantly.

Camlot: What do you do with the original tapes? They go back?

Filreis: The original tapes go to the library. Or go back to the owner, depending on what the agreement is.

Unidentified Audience Member #3: When you were talking about library quality, and slowness, and bottleneck, I wasn’t so much thinking of the guy in the music library, but I was seeing dollar signs because, at Columbia for instance, any number of books and manuscripts can go to our preservation department and be lovingly treated and everything, but when I have recordings of either music or performance of some kind, all of a sudden it’s got to come out of my acquisitions budget to pay the amount that it would take to actually digitize the reel-to-reel or whatever, and it always seems to be at least five-hundred dollars a reel. So I wondered if you would talk a bit about the economics.

Filreis: First of all, we’ve never paid anything for receiving the recordings. It’s a rule of ours. The poets enter into an agreement.. I believe they feel that this will give them new exposure and that they will sell copies of their books. For a lot of them it’s speculative, but we’ve had some evidence, looking at sales based on the release of a PennSound author page.

Then, of course, there’s the cost of professional digitization. We have a couple of third-party digitizers we deal with, in some special cases, but when we can do it in-house we will, even if there’s danger that what we’re producing is not quite at the right speed, which has been a problem with a couple of our recordings. We will keep the original and later, when we get some extra funding, we will go out and have it digitized again and replace the original.

Our project funding comes from individual faculty research funds. We also receive funding from a private donor and from staff within the Center for Programs in Contemporary Writing (CPCW) at the University of Pennsylvania and the Kelly Writers House. Charles and I have donated as well. And, of course, volunteers and student interns contribute their time.

Unidentified Audience Member #4: I’m curious about successful collaborations between faculty, libraries, and university archives. One example of a very successful collaboration that I know of is at the University of Virginia, where the university archives housed substantial recordings of William Faulkner. A faculty member proposed and carried out the project, and the university was able to claim property ownership of this material because of its status as a university recording, so its place in the archives is what made it really possible for the project to go forward. And I’m curious whether you or any members of the audience can give some related examples.

Camlot: I can speak very positively about our collaboration with the University Archive. Partly it’s because the director of the archives at the time, when I approached them to begin working on this project, was someone who was very deeply involved in the Montreal literature/poetry scene since the 1960s, so there was a grassroots element to it even there. She was actually very strict about what we would be doing with the files. She was the one who made the application to digitize them.

The archives have supported us entirely. We have been provided with archival materials, contextualizing photographs – the Concordia archivists have done an enormous amount of research for us to help us piece together the metadata for each of the recordings. They had the money to digitize – it was their domain to apply for that grant to digitize the materials and to store the magnetic reels and the WAV files on archival compact discs. But they stopped there. The matter of “making it useable for scholars” became an issue that we needed to explore ourselves. The University Archive has given us aid in using and understanding these primary source materials as a case study for this further development. Our own work has since enhanced the usability of their collection. It has brought scholars into the archive to examine the collection, and, it has brought people from all over the place into their archive via the website that we’ve designed. So it has been a really positive collaboration.

One of the other things that we’re trying to do with this material is, through public events, bring the archival materials out there and have sound artists work with these materials and create new works out of them. It is another way of bringing these materials out of the material archive and keeping them alive in different ways.

Evans: In one of the National Poetry Foundation’s decades-based conferences, I sat on a ‘70s panel. We really wanted to find out about Bernadette Mayer’s legendary project called “Memory.” Everybody knows it as a pretty inadequate North Atlantic book. Well, what I should say is that it’s an amazing book, but it’s a very partial representation of what was a complex installation piece, and the Archive for New Poetry at UC San Diego set us up with digital files. It really took some doing. We got all the photographs digitized, figured out a format for a digital reenactment of some components of the original show, got the reel-to-reels preserved and did some documenting of them along the way. This was a piece that involved a recorded voice as a primary compositional tool, but then also an installation of images. And that came together, I’m sure at some expense to them, that they delightfully didn’t let us hear about.

Filreis: PennSound’s work represents probably thirty or forty collaborations with entities, including libraries. Sometimes the library will do the digitizing work and then reflect on both sites. We will name a page based on the collaboration, to give everybody the acknowledgements that they need. I would say most of the new material that’s out there comes with an institutional collaboration of some kind. I was only telling the frustrating stories about libraries before because it – I’m hoping that librarians will be sympathetic to the faculty who wants to get this stuff done for the sake of moving through that pedagogical impulse of wanting the students and teachers out there in the world to get this stuff. I teach a MOOC, it’s called “ModPo,” for Coursera (thirty-six-thousand students). I see it as an outreach. Most of the students in the course are not of the age to go to college. They’re not doing this as opposed to paying tuition at these universities – they’re just out there wanting the poetry. And PennSound’s recordings are a very important part of that distributed teaching. The students around the world who are now listening to the materials have a sense of the institutional collaborations and the importance of libraries and preservation simply in that experience of listening to a recording that’s decades old. That’s imprecise, but the fact is that these thousands of students have access to something they didn’t have, and they realize that libraries are a fundamental part of it.

Camlot: I just want to add one more note on this collaboration we have at SpokenWeb. Because we’re doing new events using the materials from the archive, and we’re bringing some of the poets who read in the series to read alongside their former or archival selves, so to speak,, and then we’re recording those events as well, and because we’re dong oral history interviews that are related to the archive, the archive has now asked us to create a new cache of the materials that we’re producing that will function as a separate fonds, alongside this other one. We are in the middle of negotiating what this means for the materials we have been creating in supplement to the original analogue recordings, because I don’t want any restrictions on what we can do with the materials that we’ve been working with and making, and archives are cautious in a lot of ways. It is an interesting outcome in a way – a new deposit will result from collaborative work upon an old deposit.

Lori Emerson [audience member]: I actually wondered if we could turn the negativity away from archives and point it toward the DIY aspect, because to me there are more up and coming young people who are doing this work, and this comes from my experience of running, which is basically the PennSound version of an archive but just dedicated to the Canadian poet, bpNichol, and I haven’t been able to find anybody to take it over, for years, and so finally his publisher has now taken it over – Coach House. So there’s that and, when I was in Banff in the spring I had a really odd experience of sitting down for beers with Fred Wah and Steven Ross Smith and those guys, and they’re handing me over envelopes with recordings and pamphlets and records and all kinds of stuff, and they’re like, “Do something with this.” And the only place to put it that I could think of was PennSound, which really disturbs me actually, I mean, what would we do without PennSound?

Filreis: So it’s not disturbing that PennSound is there but that there’s nothing else?

Emerson: Yeah. There needs to be more.

Evans: Against monoculture, huh?

Emerson: Yeah. No one’s stepping up to do the work, which I understand – they’re usually not compensated.

Evans: This is where the fan-base issue comes in. If you could transfer some of what goes along with any good band website, for example, especially a Grateful Dead-ish band, where you’ve got a lot of bootlegs and stuff– there’s all kinds of folk taxonomy that goes with that that would be really cool to bring in. I kind of agree with Laurie’s position that it’d be nice if there were seven, ten, fifteen proliferating sites, some of which overlapped with PennSound – because we don’t want to put all our valuables in that one space. I mean, I think that you want to encourage people to come up with new uses and new spaces, and to guard against a scenario where the reliance on PennSound is so pervasive that it limits the imagination

Filreis: Well it didn’t limit it for Jason, who created his own project, and it doesn’t limit it for Ammiel, who’s not specifically just dealing with sound recordings. Maybe you should have lunch with Ammiel and find out what he’s done to gather people around him who do this kind of work, no?

Emerson: But do you know what I mean, Al? Doesn’t it put a lot of pressure on you to be the purveyor of poetry recordings for North America?

Camlot: Well the force of sustainability, that seems to be what you’re really talking about. So I’ve done my project, I’ve been working at it for two years, now. The bpNichol site has been up for a short period of time. So the question is, Who are we going to pass these projects on to? What’s going to happen to them? And it’s an interesting question. It comes back to the preservation vs. the digital environment question – where these things are living right now. Is the original audio preserved somewhere, is that what we want to protect more than anything, or how do we then protect these new environments, these DIY environments or other kinds of environments in which the sound is being presented? There is a lot of thinking still to be done about how digital humanities and digital archives projects are funded. They tend to be funded for development, but it is more difficult to fund these projects for long-term sustainability. So many of the so-called archives that have emerged and are in the process of popping up may be quite ephemeral. Maybe that’s ok. Maybe they serve their purpose by inciting art, thinking, critical exploration, or whatever, at a particular moment, within a small window of time, and that’s it. And then the links are dead, or the page doesn’t load anymore. Maybe that’s enough. But if that’s going to be the case then they’re not really digital “archives”. They’re more like digital samizdat.

Jason Scott [audience member]: I work for the Internet Archive, I don’t even know where to begin. I’m excited to hear your take because it’s brutal and I like brutal.

Filreis: Is it brutal?

Scott: Brutal. Things are wrong and we should fix them. My title at is Free-Range Archivist, which means I go around and kind of talk to people. If you look at a lot of this digital world there are a lot of people saying, “I have no plates to put my food on,” yet two miles away somebody else is saying, “What am I going to do with all these plates.” Doesn’t this sound familiar? Laurie’s saying, “I don’t have volunteers” – and I’ve got volunteers going crazy. So that’s not the problem. I find sometimes it’s just a problem of some people don’t want to see their job shared by others, or some people are worried about liability as some sort of black cloud that may mean that they won’t have a parking space in about a month.

Filreis: I agree with you about distribution. In fact if you want to, automatically or by hand, have any of your folks take everything from PennSound and put it out there again, it’s all there for download.

Scott: Okay.

Filreis: I think the problem is getting the stuff out of archives, basements of bookstores, the heirs and executors of poets – the bottleneck is really in finding it, then building a relationship with people who provide this material. So I’m with you on the distribution and storage side, but I think the main road is what happens prior to that.

Scott: But it sounds to me like what I’m hearing here is some people saying, “Well no, we have everything but who’s going to do this? Yeah – nobody,” and I’m like, “I will find you a twenty year old who would love to do this.” People like nothing more than feeling like they helped bring immortality to something important, if it’s painted that way.

Filreis: Right. Thank you. Why don’t wrap up with one brief comment from each of us? I want to say something that returns us to something that Steve said at the very beginning, which was fundamental. Recently, I had the pleasure of teaching a special project course at the University of Pennsylvania. My students were fifteen Philadelphia public school K-12 English teachers.,. We worked with PennSound audio files, some of which, such as a performance piece by Tracie Morris, have no textual equivalent. To see what happens when the teachers get hold of that kind of material has been a complete revelation for me. These teacher-students began combing through the archive looking for other materials that didn’t have a text, so that they can then return to their schools, these under-resourced and in some cases quite desperate schools, and in their short class sessions play the recordings and encourage their students to appreciate the fact that there are works of sound art actually being called poetry now. Teachers and students can work with poetic material that sounds to them like music, to which they respond in the sophisticated way in which they respond to music, and begin to feel that that’s an intellectually legitimate enterprise to be doing with their teachers.

Evans: Born audio.

Filreis: Exactly. And the response of the teachers now writing new curriculum units based on this material: they feel that they are empowered to implement the use of the poetry sound archive into their schools. And so to me, having these materials available, being able to make it freely available to them, and then to encourage new pedagogical techniques: this is what it’s all about.

Evans: There’s an outreach component, there’s also a big data component, that I’m still sorting through and I can’t quite figure it out. That is, basically, “let’s figure out what happens when we slurp up all of PennSound and start running algorithms to help find x, y, z. . . (laughter, applause, pitch ranges).” There are things that are going to come with that that are going to be really helpful. But I also kind of want to lodge a preliminary plea to stay with close listening as a practice—to not let the big data model displace that work—and maybe not only a classroom practice. One of the great things about having so much audio out there is it creates an ambient environment where a lot of people are just listening to something on PennSound and going about their day. Stephen Fredman has done some interesting thinking about this. People are hearing a lot of things in a semi-distracted state. The practice of close listening that Charles Bernstein advocated early on as a way to try to create some value around these things – I still think there’s value to listening closely to a particular phonotext of a particular poet at a particular time. You can hear literary history in the way a person handles his or her voice, in the way that they interact with a recitation strategy that’s inherited or emergent – there’s just something that I still find kind of unquenchably interesting about zeroing in on that one thing. And the other part that you can’t talk about in big rooms like this is that there’s a psychoanalytical dimension to the voice as object that hasn’t yet been adequately accounted for . I think that some people who are good Lacanians could help us with “the voice as an object” and the way that it’s treated when it’s called poetry but also the many other ways, whether it’s folk culture, speech pathology tapes – there’s all kinds of things that we haven’t sorted out yet.

Camlot: I guess I’ll continue the point about close listening that you’re making. I’ve been talking more about historicization and contextualization but I’m really interested in close listening, not only just as a critical activity but as a design question, a digital design question. There are ways in which we can we design tools that will allow us to listen differently in a multi-model environment. How does seeing things or producing visualizations of sound alter the way we engage with the audiotext that we’re trying to analyze, appreciate or understand? There is a lot of work to be done that’s very interesting, that can really bring together digital design and questions about what we’re doing as critics with these audio materials. Consider Tanya Clement and the HiPSTAS project that explores questions of sound searching. There’s a lot to explore there – but you need a kind of powerful university engine and institutional support, I think, to do some of that work. She just so happens to have this huge room of massive computers and a team of software engineers who are helping her adapt what was originally a birdcall detection program to do sound searching through speech recordings and then see what can come of that, what interesting things might be found by searching the audio itself.

Filreis: Steve, Jason, thank you very much.

  1. The symposium’s first session, “Born Digital,” included Gabriela Redwine (Digital Archivist, Beinecke Library), Lori Emerson (Assistant Professor, Department of English, University of Colorado Boulder), Matthew Kirschenbaum (Associate Professor of English, Department of English, Director, MITH, University of Maryland), and Fran Baker (Assistant Archivist, John Rylands Library, University of Manchester). 

Article: Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.