Creating a Circus of Words, Music, and Sound:
From Roaratorio to Owenvarragh

Martin Dowling and Úna Monaghan
School of Creative Arts
Queen’s University of Belfast


This essay documents an artistic collaboration that draws on a variety of sources, juxtaposed in the particular context in which we find ourselves in Belfast. The sources are Irish traditional music, Irish literature, American artist John Cage’s use of indeterminacy in the composition of music, and contemporary concepts and technologies for sound diffusion and spatialization. The context of the collaboration is a university setting which enabled these various sources to flow together into a performance. For a number of years, we have been working and teaching in the Sonic Lab of the Sonic Arts Research Centre (SARC) in Queen’s University of Belfast. Audiences entering the Lab walk out onto an acoustically transparent, modular grid floor suspended four meters above the structural floor, and find themselves in the middle of a three dimensional space with a 48 channel sound diffusion system arrayed around them. Approximately seven meters above is a technical gantry spanning the perimeter of the Lab from which loudspeakers, stage lights and/or microphones can be suspended. The Lab also has a series of 48 acoustic absorbers which can be raised or lowered to expose or cover each of the walls, affording control over a wide range of acoustic responses. The 48 speakers are arranged in five arrays of eight channels at different heights in the room. One of these is below the audience on the structured floor and another is five meters above the audience. There are additional layers at eye level and around the ceiling.[i] We have heard a wide range of music in the room, much of it electroacoustic in composition or using live electronics in performance. We listened from our perspective as Irish traditional musicians, playing with the possibilities of bringing traditional music into this environment until the ideas motivating the present work arrived.

We were both familiar with Cage’s use of traditional musicians in Roaratorio, an Irish Circus on Finnegan’s Wake. Úna was familiar with a tape version of Roaratorio presented at the Belfast Festival at Queen’s in 1997.[ii] Martin had longed possessed, more as a novelty item in his collection of Irish music than anything else, the double compact disc Cage: Roaratorio; Laughtears—Cage, Shöning, Heaney, Ennis, et. al; Writing for the Second Time Through Finnegan’s Wake.[iii] In one of the two booklets included in this recording Cage published a score with the following title:


_______________________ , ________  ___________ Circus On ________________

   (title of composition)               (article)       (adjective)                            (title of book)


Diagram of The Sonic Lab in SARC

Immediately following this interesting title, and before the beginning of the score, Cage described the piece as a “means for translating a book into a performance without actors, a performance which is both literary and musical or one or the other.” Cage’s score comprises seven numbered paragraphs of instructions. These can be summarised as follows: (1) choose a book and obtain permission for its use; (2) create a chance determined text based on the book and record and time a recitation of this text; make lists (3) of the places and (4) of the sounds mentioned in the book; (5) collect recordings of these places and sounds and, using time points in the recitation and page/line locations in the original text as guides, mix the recorded sounds with the recitation into a multi-track recording; (6) record or perform “relevant musics” by soloists and superimpose these onto the multi-track recording; (7) realise the piece as a stereo recording, a multi-channel recording, or a performance with the layers created in (2) and (6) performed live.

We each independently recognised that the Sonic Lab would be a wonderful place to perform this type of “circus.” When Úna left her position as Technical Assistant in SARC to embark on PhD research on uses of technology in Irish traditional music performance, we found ourselves together as supervisor and supervisee, and in our first brainstorming session Martin told Úna he had been harbouring an ambition to perform this Cage piece in the Lab, but not Cage’s Finnegan’s Wake version. He wanted the piece to be grounded in Belfast, where we live and work, and he felt that the obvious choice of author was fellow traditional musician Ciaran Carson, Professor of Poetry in Queen’s University. The ideal work of Ciaran’s for the project was The Star Factory, the enigmatic, stream-of-consciousness memoir of his Belfast childhood. [iv] Úna was inspired by the idea and wasdetermined to make it happen, so she did, by securing funding from a scheme run by Enterprise SU in Queen’s University of Belfast Students’ Union.[v] This allowed us to set off on a fascinating and totally absorbing journey. We are not scholars of Cage or of the history of the interpretation of his work, and generally we did not take account of this history when interpreting the score. We endeavoured to follow the score as closely and truly as possible. We encountered some instructions that were ambiguous in their meaning, that invited our own interpretation, or that have been outdated by the advance of recording technology since 1979. We occasionally looked to Roaratorio for clarification. In what follows, we record all of our decisions, and show in detail what we did in response to the score. We hope that this document might provide inspiration and reference for more realisations of this score in the future.

Choosing a Book

We met with Ciaran Carson on the 25th of March 2011 and asked for his blessing and whether he would take part in the performance. He agreed. Having thus satisfied the first of the seven instructions, we embarked on our journey to produce a performance of Cage’s score in 2012, the centenary year of his birth. We gave it the title Owenvarragh, a Belfast Circus on The Star Factory. “Owenvarragh” is a place in west Belfast next to the road where Úna was born and raised. It is also the title of the central chapter of the book. In Ciaran’s childhood, there was an “undeveloped half-acre” there which served as a “foyer or rehearsal space for that rural hinterland beyond us” (p.97). Owenvarragh is an Anglicization of the Irish abhainn bharrach, barred river. As Ciaran explains in a footnote, “the bars being stakes or staves; and the Owenvarragh, in its lower reaches, from Stockman’s Lane to where it joins the Lagan, becomes the Blackstaff; barra can also mean a sand-bar, or fearsad, so these waters—Blackstaff, Owenvarragh, Farset—form an etymological confluence” (p. 97). These little rivers continue to flow, largely unseen, beneath the urban infrastructure of Belfast. They provide an apt metaphor for the fluid ground of etymology, narrative, and memory on which Ciaran constructed The Star Factory.

Creating Mesostics

We agreed that Martin would address instruction (2) of the score and Úna instructions (3), (4) and (5), leaving decisions about collaboration with Ciaran and other musicians until after these initial materials had been assembled. Instruction (2) begins with a procedure for reducing the text to a poem: Taking the name of the author and/or the title of the book as their subject (the row), write a series of mesostics beginning on the first page and continuing to the last.

We chose the title, hoping for a greater alphabetical variety from the title of the book over Ciaran’s name. Also, there happen to be fourteen letters in the title of the work, so the possibility of making “sonnet mesostics” appealed to us. We took the first page of the book to be the page marked ‘1’, which is the first page of chapter 1. The score does not provide for disregarding any words, so we included the chapter titles. The book contains very lengthy quotations from other published works in the body of the text, and these were also included. The words “and continuing to the last” we interpreted as going onward from the first page and not turning back, doing this throughout the text so that the words in the mesostics appear in the order they do in the book.[vi]

Instruction (2) continues: Mesostics means a row down the middle. In this circumstance a mesostic is written by finding the first word in the book that contains the first letter of the row that is not followed in the same word by the second letter of the row. 

Following Cage’s own realisation in Roaratorio, we took this to mean “a column down the middle.”
















At the start of the first mesostic, we were looking for a word that contains a T, which has no H anywhere after the T in that same word. We considered a word that has an H somewhere before the T to be eligible (so the word ‘shallot’ would be acceptable, but not the word ‘teach’). We interpreted “is not followed. . .” to mean “is not followed anywhere in the remainder of the same word,” not “is not followed immediately.” The latter interpretation would have allowed ‘teach’ but not ‘with,’ but in our interpretation both are excluded. Nor did we interpret this to mean “is not followed immediately at the beginning of the next word in the text.” We ignore situations such as “. . .be applied to Belfast. Here, the kitchen. . . ” (p. 78).

What happens when a word contains multiple instances of the desired letter, some of which may be followed by the next letter in the row, and one of which may not? For example, even though the word ‘thought’ contains a T which is followed by an H, it subsequently contains another T which is not followed by an H and so might appear as ‘thoughT’ in a mesostic. We allowed this, but regarded the word ‘teeth’ to be ineligible since both Ts are followed by a H in the word. Hyphenated compound words abound in Ciaran’s book, and when they were encountered the instructions were applied to the entire compound word.

Instruction (2) continues: The second letter belongs on the second line and is to be found in the next word that contains it that is not followed in the same word by the third letter of the row. Etc.

We noted that ‘the next word’ includes the word which in the text comes directly after the one just used in the mesostic,[vii] meaning that successive words in the book might, and occasionally did, form successive rows of the mesostics. The “Etc.” does not help us choose the last word in each mesostic. We considered using the first letter of the next mesostic as the constraint on choosing the word for the last line of each mesostic, so that when we came to search for a word in the text which contains a Y, that Y could not be followed in the same word by a T. We consulted Cage’s Finnegan’s Wake mesostics, which are based on the name “James Joyce.” We could not find cases where a word with an E was followed in the same word by a J, and so were unable to ascertain whether Cage adopted this constraint. After creating a couple of dozen mesostics, it became apparent that disregarding words which had in them a Y before a T would not significantly affect the words encountered in the text with a Y. The relative rarity of the letter Y in the English language, and its frequent occurrence at the end of words, combined with our decision to adopt the index of syllables constraint (see below), helped us to decide to select the next word that contains the last letter of the mesostic without worrying about what letters follow it in the same word.

Instruction 2 continues: If a shorter rather than a longer text is desired, keep an index of the syllables used to represent a given letter. Do not permit for a single appearance of a given letter the repetition of a particular syllable. Distinguish between subsequent appearances of the same letter.

We always intended to perform this piece, so we considered that its length should not be too ambitious for the audience. Therefore a “shorter rather than longer text” was desired. The chance operations did not allow us to predict with any certainty what length would result. We took ‘index’ to mean a list of the syllables used for each letter of the fourteen letters in THESTARFACTORY. This required the making of separate and independent list of syllables for each occurrence of each letter, including each of the three Ts, two As and two Rs, because the occurrence of a syllable in any column restricts its subsequent use only in that column, and does not prevent its subsequent use in another column.

The question of what defines a syllable then arose. Martin consulted the Winston Simplified Dictionary: Encyclopaedic Edition.[viii] Each entry in the dictionary is divided by dashes into syllables. The definition of ‘syllable’ in this dictionary is: “1. a sound, or set of sounds, consisting of a vowel, or of a vowel and one or more consonants, uttered by a single effort of the voice and forming a word or element of a word; also, 2. The letters or characters representing such a sound.” This definition presented us with a “chicken-and-egg” problem regarding the vocal and etymological aspects of a syllable. Both aspects are required: a single effort of vocalisation and an element of a word. It is possible therefore to have a single effort of vocalisation which is not a syllable, and it is possible to mispronounce and misspell a syllable, because what is uttered is neither a word nor an (etymologically correct) “element” of a word. Interestingly, this particular dictionary is inconsistent. For example “mechanic” is divided “me-chan-ic” but “mechanism” is divided “mech-a-nism.” The editors of the dictionary seem to prioritise spoken speech over etymology, though it is also entirely possible that they did not have a coherent approach, as none is specified in the introductory sections of the dictionary. Martin found the more linguistically specific definition of a syllable found on Wikipedia useful. According to this definition, a syllable has three components: onset, nucleus and coda. The onset and coda are usually consonants or groups of consonants, and the nucleus a vowel or vowel sound. Clearly “it”, “bit”, “whit” are all different syllables. As with the case of “mechanic” and “mechanism,” Martin found the occasional ambiguity regarding whether some consonants are the coda of one syllable or the onset of the next, and whether these vary in different occurrences and context. When uncertainty arose, he consulted the Winston Simplified and followed its syllabic divisions exactly in each case.[ix] While working through in this way, a number of practical rules evolved which appeared to be consistent with the instructions and helped to achieve the goal of making a shorter rather than a longer text. For example:

  • Wherever possible, Martin gave the etymology of the word priority when deciding between possible syllabic divisions within speech, separating prefixes, suffixes, and root words. “Everlasting” is divided “last-ing” not “las-ting” for example.
  • Martin did not distinguish between word-syllables like “raise” and variants of that word like “raised”, or between “street” and “streets,” because these were syllables derived from a common root. If one of these was entered, all were then excluded if encountered later. In other words, Martin did not consider the addition of a suffix or pluralisation as creating a new syllable, even if pronunciation warranted it.
  • Two utterances that are homonyms, sounding the same, but with different spellings and etymologies, were considered to be distinct syllables. “Fare” did not exclude the subsequent use of “fair.” However, homonyms which were in fact the same word or root, though perhaps spelled differently, were not so considered. For example, Martin, having entered “foot” into his index of syllables, subsequently encountered “foote”, an archaic spelling in a quotation, and excluded it.
  • When abbreviations were encountered where in normal speech the abbreviated word would be pronounced (i.e. “St.”, “etc.”), then the fully pronounced word, and its syllables, were considered for eligibility. When abbreviations or acronyms were encountered where the abbreviation itself is normally pronounced (i.e. “BBC”) that normal pronunciation was taken.

In the “index of the syllables used to represent a given letter,” we took “a given letter” to mean the letter of the row of the mesostic, contained in a word we were searching for in the text, according to the instructions of the score. The syllable of the word that contains the letter of the mesostic was recorded, never to be used again for that instance of the letter in THESTARFACTORY. Our first mesostic begins with “sTreet.” This is one syllable, and Martin made a note of it, never to use it again in the first row of a mesostic. Where the letter appears twice in a syllable or in separate syllable, Cage does not say which instance to use. If both instances of the letter satisfy the rules, then we used the first one first. Cage appears to have done the same in his realisation of Finnegan’s Wake. So “sTreet” is used, not “streeT.” Keeping with the first line of our first mesostic, the use of “street” rules out the word “streetlamp” but not “streetlight” because “light” is a different syllable, thus we might have “streetlighT.” It is possible to use the word “streetlight” twice, once as sTreetlight” and later as “streetlighT,” consistent with the rules (subsequently, the syllable “light” would be disallowed). In fact, it would be possible to use it six times, twice for each T in THESTARFACTORY. Martin created a table with 14 columns which then grew to have as many rows as were mesostics generated by the instructions. The syllable index grew to a total of 1883 unique syllables. We are contemplating its potential for generating new texts.

Instruction 2 continues: Omit punctuation and capitalise the row, reducing all other capitals to lower case. If at the end of the book or a chapter of it a mesostic is not complete, leave it incomplete or complete it by returning to the first page of the book or chapter and continuing your search for words containing the necessary lettersHaving completed the mesostics, identify each line by page and line of the original from which it came.

Martin left the incomplete mesostics unfinished at the end of each chapter. Having decided not to complete the last mesostic, Martin decided that he should not start a new one if it was highly likely that it would not comprise more than a few rows. This was a clear departure from Cage’s instructions and his intentions, but Martin felt that it would result in a more elegant and concise set of mesostics, and be in keeping with the goal of a shorter rather than longer work. Where a mesostic was completed on the last or penultimate page of a chapter, a new one was not started in that chapter, but with the title of the following chapter. Chapter titles were considered to be line 1 of the first page of each chapter. In the end, 138 mesostics were created, with 26 incomplete mesostics at the end of chapters.

This was an agonising, tedious, but in the end exhilarating process. At about mesostic 115 Martin discovered that he had inadvertently used a couple of syllables twice. This provoked a thorough checking of his syllable index which uncovered a handful of mistaken words in need of correction, which involved returning to the text where the syllable was used the second time, and finding the next eligible word that did not use a syllable which had been previously used. If this syllable was found in a word occurring before the next used word in the text, and it was not used later on, there were no further troubles. If not, a cascading series of revisions ensued, rippling down the mesostic in question and deep into the series of mesostics at the affected letters. It was one thing to correct syllables not properly excluded, but another to ensure that words were not missed that should have been included earlier than they were eventually to be. Martin accomplished the former, after many hours, but abandoned the latter as futile, or perhaps as an exercise in “purposeless purposelessness” beyond the limits of his endurance. In the end, we included human error in the grand strategy of chance determination.[x]

The strategy of keeping a syllable index more than doubled the time it took to find and record the next word in a mesostic. Having made more than 80 mesostics in the first hundred pages of the book, it was not yet apparent when, or even if, the strategy would significantly shorten the text. It gradually became apparent that first the Y, and then the F, would give the desired result. The process produced mesostics drawn very unevenly from the text, with the lines THESTAR using words quite close to each other in the text, followed by increasingly lengthy gaps between R and F. Another “bunching” effect occurred with the letters ACTOR, followed by another increasingly lengthy gap before Y, the last letter of the mesostic. Keeping an index of syllables using S proved quite early on to be pointless, because of all the unique syllables in pluralized words. In contrast, the letter F proved to be relatively rare, and the words using an F which also have an A following that F were encountered (and skipped over) in a high percentage of instances. Martin typically coasted through a paragraph or more looking for an F without an A following and excluding all the syllables that had piled up in the index, before alighting on something unusual and eligible like “doffers” (p.101) or “effulgence” (p. 184). Then he faced more grinding through letter by letter until the Y, which appeared even more rarely than F, and rarely in places other than a “-y” suffix. After collecting in the index a good few of the internal syllables as well (“psy”, “cy”, etc), a larger and larger number of lines were passed in search of a unique use of Y. This search took on the form of a quest, and Martin foolishly began to think about half way through that all the Y syllables were used, so that the end of each chapter would be quickly reached, and the job thereby reduced. The mesostics later in the piece feature some quite unusual, and for Martin mildly infuriating, Y words: “yakking” (p. 127), “chrysanthemum” (p. 136), “beryl” (p. 161), “brylcreem” (p.164). Nearing the end, Martin was skimming through Ciaran’s description of Kelly’s Cellars, the famous city centre bar well-known for traditional music when Ciaran tells the reader that Bank Lane was formerly known as “Bryce’s Lane” (p. 248). “Bryce” is the bottom row of the 129th mesostic, and it was something of a turning point. Martin only had a handful more to do, and became fascinated by the possibilities that remained. “Analyse” came very late (p.255), and then “pyramids” and “hypsophobia.” The finish was beautiful, and we are still astounded at the serendipity of it. The penultimate Y word, on the third last mesostic, is “rhythm”, and the last Y word, on the penultimate mesostic, “rhymes.” The last word of the last mesostic is “pleasure.” Rhythm, rhymes, and pleasure: it finished up wonderfully.

In our edition of the mesostics, we insert the page and line number of the first line of every mesostic, and also indicate at what row the mesostic moves to a new page. Cage acted similarly in his publication of the mesostics on Finnegan’s Wake. We discovered eventually that this was not enough information to effectively follow subsequent instructions, as the reader will soon see.

It is possible to stop here, having made mesostics with one word per line, with punctuation removed and all letters except the central column in lower case. However, instruction (2) also includes the following: Other adjacent words from the original text (before and/or after the middle word, the word including a letter of the row) may be used according to taste, limited, say to forty-three characters to the left and forty-three to the right, providing the appearance of the letters of the row occurs in the way described above. 

Our mesostics include adjacent words, according to our “taste.” The Winston Simplified defines taste as “the faculty of nice discrimination as to what is beautiful, refined, elegant, or fitting; sound critical judgement.” In our judgement, more than a single word for each line of the mesostics was required to make manifest the rich poetry in Ciaran’s prose. At the same time, what appeared most beautiful, elegant, and fitting to us were mesostics produced by using as few additional words as possible. The mesostics have an aesthetic value and meaning that is independent of the prose out of which they are lifted. But because of the “bunching” effect of the procedure described above, many of the mesostics contain words that are located very close to each other in the book. Indeed, those from the first half of the book are often drawn from within a single paragraph. By adding a few adjacent words, entire phrases, complete lists, and complete images from the text can be rendered in the mesostics. These “denser” mesostics inevitably capture more of the meaning and the message of the passage from which they were lifted. Later, these “bunches” are more thinly spread across the page, and the gaps between them stretch across more pages. The later mesostics move farther away from the narrative of the text, and sometimes have their own enigmatic elegance.

Another matter of taste involved with this adding of adjacent words was whether to treat each mesostic as a separate unit of meaning—as a separate poem—or whether to imagine units of meaning stretching across successive mesostics, beginning in the middle of one and finishing in the middle of another. Occasionally, for example, the gap in the book between the occurrences of words might skip over paragraphs or pages of text, only to resume in a dense pattern of selection, marking the mesostics with clearly different batches of related vocabularies, with different potentials of meaning. The word belonging to the last Y of a mesostic was often much closer in the text to the bunch of words belonging to THESTAR of the next mesostic than to the previous R word. We were willing, within the confines of the mesostics of a particular chapter of the book, to occasionally redraw the boundaries between mesostics. This has possible ramifications for the recitation of the text.

We allowed the font “perpetua” size 16 and the left and right margins on an A4 page of 2.54 centimetres to limit the number of characters to the left and the right. This was always less than 43 characters, so it appeared to be in keeping with Cage’s instruction. We used only whole words that fitted. This instruction might be read to imply that the word containing the next letter of the row be allowed as part of the added words, if it were to fall within the limit. However such repetitive use of words in a mesostic fit neither our taste nor, apparently, Cage’s own in the Finnegan’s Wake realisation. Accordingly we have interpreted this instruction to have a further clause, “and providing that the word from the adjacent lines of the mesostic is not used.” In the end, the space limit was rarely reached before a pleasing mesostic had been made.

It gradually became apparent that concerns we may have had about randomly chopping up Ciaran’s text and constructing something indeterminate and nonsensical were unfounded. The mesostics have their own fluidity, but they also evoke the text in many elegant ways. The thirty-two chapters of The Star Factory, many taking the name of a street, building or place in the city, are connected by Ciaran’s fabulous hook-and-eye links. One of the two eponymous “Star Factory” chapters begins by describing not the building itself (a now derelict shirt factory in West Belfast) but the author’s earliest misremembering of it, which soon diverts to speculations on the building’s “great secret.” “According to another narrative”—and in the book many are offered for a given place—there was within the building “The Zone.” This is not a place but a story one enters, which itself “comprised many stories, yet their diverse personal narratives and many-layered timescales evinced glimpses of an underlying structure, like a traffic flow-chart with its arteries and veins and capillaries” (p. 62). Many of these stories come from Ciaran’s father, and the chapter moves towards a reflection on his mnemonic techniques. We are on a “terrain honeycombed with oxymoron and diversion, and the tiny ancillary moments of your life assume an almost legendary status. There are holes within holes, and the main protagonists are wont to disappear at any time, as in my father’s story, which follows:” This ends the chapter, and leads to the hilarious story in the next chapter about the man from Mullaghbawn, County Armagh, who sees an advertisement of a hole for sale in the Newry Frontier Sentinel newspaper, buys the hole for a fiver, and endeavours to bring it home.

We mention this passage to suggest how well the structure and narrative style of The Star Factory lends itself to the technique of the poetic reduction Cage invented to translate Finnegan’s Wake. While working on the piece we found many aesthetic resonances between Cage and Carson. In one uncanny moment of our work on the book, we found Ciaran reminiscing on the bookshops of the long since destroyed Smithfield Market in Belfast’s City Centre. Hitting one day upon Vol. VI of The Works of Henry Fielding, Ciaran notices at the bottom of each page

the first word of the next page is reproduced, or anticipated, indented in the right hand margin. I like this 1771 device, which simultaneously stops you and carries you forward, in a knit-purl fashion, and you could make a mini-narrative or arbitrary précis of the matter by conjugating these dropped stitches, a yarn made all the more mysterious because it contains hyphenated bits, broken by whatever way the typesetter thought fit to justify them, for example, ‘pray, deed, they, too, upon, martyr, -sently, -ship, O love, -gar, -ness, more, you, passion, returned, but , fellows, you, Mr., that, was, they, -tainty, or, all, unless, mistake, time, -wouse, Joseph, poor, The, thief, CHAP.’, which give a tolerably fair, if etiolated, encryption of Chapters V-VIII (p. 226).

This text, the result of a 1771 device, is in Owenvarragh cut across again by Cage’s 1979 device, spelled out in the second instruction of the score, resulting in the following fragment of a mesostic:

the Typesetter                          226/04 tHought fit to justify them pray dEed they –Ship o love Thief chAp a tolerably faiR if etiolated encryption baFfled                  229/05 she trAnslated a pattern

Making a “Tape Part”[xi]

Instruction (2) continues: Make a tape recording of the recital of the text using speech, song, chant, or sprechstimme, or a mixture or combination of these. Ascertain its time-length. Subtract that from a total program length, and distribute the thus determined silence between large parts and chapters of parts and at the beginning and end of the tape.

A first draft of the mesostics was complete in October 2011, and Ciaran recorded a recital of the text the following month. After editing out some pauses taken by Ciaran to clarify the handwritten mesostics, we had a sound file of a recitation of the mesostics lasting one hour, 13 minutes and 45 seconds. This was a pleasing result because, knowing that additional silence was required by the score, we would in the end have a piece near the typical concert length. We took our “total program length” to be 90 minutes. It seemed appropriate to divide a ninety-minute piece into two parts to provide the audience with a short interval, with the first half slightly longer than the second. We were pleased that the central chapter of the book (from which we took our title) occurred at just about the 50-minute mark of Ciaran’s recitation. Úna separated the audio file here to make two parts of the concert, the second half opening with the Owenvarragh chapter. We then had one file of around 40 minutes of recitation and another file of around 33 minutes. After the required subtraction of the length of Ciaran’s recitation, 16 minutes and 15 seconds of silence remained to be distributed. Throughout the process of inserting the silence, Úna kept in mind that in the live performance of the piece, Ciaran’s recording of the mesostics would be replaced with live recitation. Úna wanted to leave enough silence at the end to give Ciarán some flexibility if he were to take the live recitation at a slower pace than in the recording. She began by inserting silence “at the beginning and end of the tape.” She left 45 seconds of silence for the end of the piece. She also allocated 30 seconds to the beginning of the piece, so that some of the sounds (see below) would be heard before the speaking began. As it turned out, the first words of the book (“Raglan Street”) are both a place and the first line of the first mesostic, so the tape part starts at exactly the same point as the recitation. She therefore incorporated the 30 seconds she had set aside for the start into the breaks between chapters in the first half, so as to make available eight minutes of silence for the first half and seven and a half for the second. She chose to distribute the allocated time in equal parts after each chapter (including the last one, which already had 45 seconds afterward).

Since the book has no “large parts or chapters of parts” we used the 32 separate chapters of The Star Factory as points for distributing the required amount of silence into the recitation. Úna allowed one minute for the extra time at the end of the first half, which left seven minutes to be divided between the 12 chapters of the first half (35 seconds after each, and 35 + 60 seconds after the last chapter of the first half). Seven minutes and 30 seconds of silence were allocated to the second half, plus the 45 seconds of extra time at the end. Because recordings were made which were to be placed before the first word of the first mesostic of the Owenvarragh chapter, she was able to allocate 30 seconds at the start of the second half (something we could not do at the beginning of the piece). She tacked on a further 30 seconds to the end. The remaining six minutes and 30 seconds were divided by 20, the number of chapters in the second half, allocating 19.5 seconds to be inserted after each chapter. She rounded this down to 19 seconds, which left 29 seconds after the last one, in addition to the 30 + 45 seconds we had already allocated as spare time at the end of the piece. These decisions regarding the distribution of time in whole seconds followed naturally from Úna’s use of digital audio editing technology.

Once the above structure was conceived with its periods of silence between chapters, it was transferred to the timeline of Pro Tools. Úna edited the recording she had made of Ciaran reading the mesostics into chapters, producing 32 separate regions, and positioned them separately in time in Pro Tools according to the structure in the diagram below. As she worked, Úna inserted markers on the Pro Tools timeline labelled with the page and line numbers where each chapter region began. The first chapter was positioned at time 0.00, and ended at 4 minutes and 49 seconds. This figure was rounded to the nearest whole second before the 35 seconds of silence was added. This rounding made the calculation easier and we presumed the combination of rounding down and rounding up would balance over time.




Diagram of piece structure, including amounts of inserted time. Total length is 90 minutes.


With this work completed, we arrived at the end of the second instruction, which concludes: You then have a ruler in the form of a typed or printed text and in the form of a recited text, both of them measurable in terms of space (page and line) and time (minute and second), by means of which the proper position (see 5 below) of sounds (see 3 and 4 below) may be determined.

In our case, of course, the two forms of the ruler were the printed mesostics on the one hand and the Pro Tools file on the other.

Sounds and Places

At this point, the literary work of the realisation was complete, but the immense project of creating a circus of sound to accompany the mesostics had just begun. Úna had been for months marking her copy of The Star Factory and filling up a spread sheet according to Instructions (3) and (4). Instruction (3) begins: Make a list of places mentioned in the book and a list for each of the pages and lines where the mention is made for each. If the list once made is unmanageably long, reduce in some chance-determined way, e.g. to a number equal to the number of pages in the book.

We took ‘place’ to be a noun, a name, somewhere you could find on a map and possibly visit. Any thing which cannot be moved easily was recognised as a place, like the Albert Clock (p. 237) and Westminster Abbey (p. 149). We recognised places that were used as part of a thing, a modifier or descriptor of a thing (“Wellington” in “Wellington Bomber”, p. 15; “Mississippi” in “Mississippi Steamboats”, p. 28), but not adjectives like Egyptian. We included real addresses, landmarks, cities and towns, and countries, even those that no longer exist but might still be visited (Soviet Union, p. 64; Raglan Street, p. 1; various defunct businesses which had premises whose addresses we could find). We excluded all ideal or abstract places like “the park”, “the world”, “the lavatory” or “a library” (unless referring to a specific address), as well as mythical places. When a place was nested in a larger place, such as “the UTA depot in Newry” (p.74), it was recognised as one entry in the list, not two. If a place and a sound were mentioned together (“the tolling of the Albert Clock” p. 244) they were put into both lists. Our list includes every mention of a particular place, even though some places (Belfast for example) are mentioned many times. For any reader of this book, there will be some uncertainty about the actuality of some of the places mentioned. Often the first mention of a place did not give us enough information to identify it as such, though it may be corroborated as a place later in the text. So for example the first mention of “The Star Factory” (p. 61) did not identify it as a place, and was not included, but in subsequent mentions it was clear that its location could be determined, and these were included. Later, “the biggest shipyard in the world” was included because we identified it as Harland and Wolff, but in the same paragraph “the biggest linen spinning mill” and “the longest rope walk” were not included (p. 179).

The places are very unevenly distributed throughout the book, with numerous places on some pages, and none on many others. Ciaran’s descriptive style frequently makes use of lists, and occasionally these are lists of places. On page 8, he is “reminded how the arbitrary power of the alphabet juxtaposes impossibly remote locations”, and 23 street names and other places follow. The chapter “The General Post Office” describes Ciaran’s youthful obsession with stamp collecting. “I was initially drawn to collect, sort, mount, and annotate the stamps of the British Empire,” he writes (p. 32), and over the next eight pages over 60 places are mentioned.

Úna counted a total of 1,052 places mentioned in the book. We trust the reader will agree with us that this was “unmanageably long.” We therefore followed Cage’s recommendation to reduce the number of places to the same number of pages as in the book, 291, and adopted the following procedure as our “chance-determined way” of selecting these. Úna attempted to choose one place per page, beginning with page 1 and moving to the end of the book. For pages with more than one place, she rolled dice to pick a line, using a 10-sided and a 6-sided die.[xii] If there were multiple places on the chosen line, the first was used. In the event that she did not find a place on the selected line, she chose the nearest place to that line. Finding two places equidistant from the line, she rolled a die again to choose between them. Úna took a note of pages on which there were no places, and when next encountering a page with more than one place she allowed for the selection of additional places, up to the number of pages without places previously accumulated.  There were no places on Page 291, the last page. She therefore rolled dice to select a page at random, and then rolled again for the line on that page. In this way she arrived at line 5 of page 91, which gave the 291st entry chosen for inclusion in the reduced place list. At the end of this process, the reduced list still included multiple mentions of the same place. Unique sound recordings were used for each instance of these.

Instruction (4) is: Make a list of sounds mentioned in the book and a list for each of the pages and lines where the mention is made. If the list once made is unmanageably long and baffling because of the large number and kind of sounds, establish families of sounds and extract from the whole list those related to certain of these.

To follow this instruction, we needed to decide what constitutes the mentioning of a sound.[xiii] Obviously we include words that signify only a sound. If the word has other significations in addition to a sound, then for it to qualify, the context of its use must directly signify a sound. The context may be the mention of words such as “heard” or “listened” or “sound of.” This was not as straightforward a procedure as it may seem, as the following examples demonstrate, and there were a great many difficult borderline cases. We might not have taken “my father’s voice” (p. 66) because this does not signify the sound of the father’s voice. However, the full context is: “At such points my father’s voice would elevate and quicken.” This “elevating” and “quickening” of the voice makes it a sound. It requires that the recording have this particular feature: the elevating and quickening of his father’s voice. In contrast, “compressed mnemonic musical devices” (p. 66) suggests a sound, but as there was no mention in the paragraph of these devices sounding or being heard, it was excluded. “Footsteps” were included only because they “echoed each other in the frosted air,” and would not otherwise have qualified as a mention of a sound (p. 277). A “kettle steaming on the hob” (p. 67) is not a sound, but “the bugle of an army” is taken to mean the sound of the army bugle rather than the instrument of the army (p. 67) and was included. Elsewhere Ciaran writes about a radio and the “comfort of its disembodied voice” (p. 104). “Voice” would not be included if it was mentioned on its own. But because the voice was disembodied in this way, it cannot be recognised in any way other than as a sound. The different ways people pronounce “Belfast” (p. 45) were regarded as sounds, and similarly for the pronunciation of “bonefires” (p. 22). These are references to accents which exist primarily through sound.

.Examples underpinning the visual as opposed to sonic orientation of Ciaran’s imagination proliferate in the book, Lacking an indication of hearing or listening, or a description of the sound, these passages do not qualify as sounds. For example the passage “beaded bubbles soaring through a glass of Lucozade to spit and blink within its rim” (p. 256) immediately conjures a sonic image. To record the sound of a bottle of Lucozade being poured into a glass would have been a typical task. But the description is visual, not sonic. “A mini choreography of struck matches would flare up throughout the auditorium” was also excluded, even though “flare up” produces in the mind a distinctive image of performance with its associative sounds. In contrast, “bubbles escape with a hiss from champagne glasses” (p. 273) is included. Unlike the beaded bubbles of Lucozade, these bubbles “hiss,” and are heard, rather than merely visually “soaring,” “blinking” and “spitting.”

The subject of silence recurs in the book. Is silence a sound? Do the mentions of the experience of silence in the book qualify as the mention of a sound? We are aware that these are questions central to Cage’s art. Practically, it would be difficult to record silence, and to make appropriate decisions about its duration. And in the context of the piece, we did not see how the audience would be able to appreciate the difference between recorded silences and, say, the sounds of a hard drive revolving, or the silences in the tape part where there were no recordings. So we decided that only sounds that sound qualify. When “the others went silent” (p. 133), we did not regard this as a sound, nor did we regard “the noise of life itself” (p. 133) as a particular sound, though Cage might well have. There were some intriguing exceptions. “An inaudible smooch” (p. 95) refers to the sound of one cigarette being lit by another, and so it was recorded. “The quiet of the introit” (p. 142) refers—very much in the spirit of Cage’s 4’ 33”—to the ambient sound of a church full of people sitting quietly, and so it was recorded. Ciaran describes standing before a painting to “eavesdrop on its silent narrative” (p. 51), and Úna stood before the painting in question (in Ciaran’s home in Belfast) and recorded it. We are still wondering whether we should have recorded “the silence” of the bell given to St Ciaran of Clear Island by St Patrick, which, according to one of the tales in the book, he carried until it finally sounded (p. 114).

We found the number of sounds mentioned in the book (256) to be manageable, and disregarded Cage’s instructions for reducing the list.

Instruction (5) begins: Collect as many recordings as possible made in the places mentioned (3) and of sounds mentioned in the book (4).

Úna made as many recordings in their places as she reasonably could before our time ran out. Many recordings captured generic, ambient sounds. Because we wished to make a piece in which the sound of the wind, traffic, and motionless interiors did not predominate, Úna looked for opportunities to capture an interesting or unique sound in each place. On many occasions it was not possible to gain access to buildings, for example private houses, closed shops, or businesses not open to the public. For example, security at the Short Brothers Factory (now Bombardier Aerospace) in east Belfast was so strict that Úna was only allowed to record from across the street from the main entrance. Because so many outdoor recordings were forced upon us, she tried to balance this by making as many indoor recordings as possible, a task which was quite adventurous at times. Úna found “St Finian’s Church” in Clonard, County Meath, to be derelict, but was able to climb through an open door and make a recording. She found the cobbled back entrance to “the Glass Factory” in Belfast where delivery drivers used to arrive with their Clydesdale horses and carts. The building was gone but these cobblestones now form part of the private back yards of a street of terraced houses in the city centre. When a car came out of a gate, she ran in and was able to speak with locals about their memories of the original building.

Recording the sounds was a different kind of adventure. While in rural England, Úna was awakened by an owl one night only to realise that it was needed for the piece and she scrambled in the dark for her phone to record it. She held an impromptu New Year’s Eve party in the deserted, rainy, and freezing docks of Belfast in an attempt to capture “the slow majestic note of a ship’s siren” (p. 133), because she had heard, correctly as it turned out, that all the ships then in dock sounded their horns at midnight on New Year’s Eve. This produced one of the most powerful recordings in the piece. Úna collected field recordings from various places in and around Belfast. We travelled to Dublin, Tara, Clonard, Kerry, Derry, London, and elsewhere. We noted (with relief at times) Cage’s use of the term “collect” rather than “make” when referring to the recordings. This meant that we were able to feature historic recordings and sounds from the further away places our budget did not permit us to visit. We corresponded with people from many far-flung places who kindly recorded or collected and sent us sounds, including “Persia”, “the Vatican”, “the Mississippi.”[xiv] This direction to collect rather than necessarily make recordings also meant we could feature more of those mentioned in the list of sounds. For example, it is not straightforward to capture “an almighty clap of thunder” (p. 146) or “monkeys whooping” in Ireland. On these rare occasions Úna was not averse to using publicly or commercially available sound libraries if it produced the best example. The recordings were therefore collected from a variety of sources, and in a variety of formats. Úna had more flexibility with larger places (e.g. ‘Belfast’ or ‘London’), which provoked thought about how to represent in a short continuous sound recording large places, with their wide variety of associations, long history and possibly rich sonic landscapes. In preparation for the sound collecting trip to London, she consulted people via social media by asking them to suggest their response to the question “If I had to sum up London in a sound, what would it be? What is your most iconic London sound?” “Ireland” appeared several times in the reduced place list. We were fortunate to have access to some recordings from the RTÉ[xv] archive, which enabled Úna to include iconic historic recordings in an attempt to portray some of the main associations with “Ireland,” and capture in sound the era of Ciaran’s childhood.

The collective nature of gathering the sounds necessarily produced a wide range of recording formats and quality. Those who contributed recordings had a wide range of experience and equipment, from professional sound engineers and artists with high quality microphones, to bemused friends who agreed to capture a place recording on their mobile phone. Some of the RTÉ recordings were quite old and distorted, but as previously mentioned this variation in quality was thought to be a natural result of the collecting process. We felt this variation in quality and character of the recordings added to the chance determinations and was in keeping with the ethos of the piece. Content was more important to us than quality, and we felt it was better to have a recording of lower quality than none at all. Only on one occasion did Úna exclude a recording based on its quality. We aimed for 16 bit files with a sample rate of 44.1kHz. Where this was not possible, for example in many donated recordings, we asked for the highest quality available. Whatever the original format, all sound files were converted to 16 bit / 44.1 kHz in Pro Tools or Logic Pro. We felt this standard to be a good compromise between the limitations of some equipment and the standard required to take advantage of the high quality sound reproduction system we used in the Sonic Lab. This choice of bit depth and sample rate also resulted in a manageable amount of data for moving between studios the hundreds of sound files collated in one Pro Tools session.

Instruction (5) continues: Re-record them in stereo one at a time on a multi-track tape at proper points in time (see 2 above) following chance determinations (a) of stereo position (using a large number of positions as can be conveniently distinguished), (b) of relative duration (short, medium, long), (c) of attack (roll on, fade in, switch on), (d) of successions of loudness or loudness, and (e) of decay (roll off, fade off, switch off).

Here the ramifications of not having followed instruction 2 completely by recording a page and line for each row of each mesostic are revealed. Lacking this information, Úna was faced with an extraordinarily unmanageable task. In order to find every “proper point in time”, Martin used Úna’s spreadsheet of sounds and places to assign to each entry a mesostic row after which that recording was to be placed.

Our recordings were made digitally and could easily be transferred to Pro Tools software. Úna positioned each of the recordings so that the beginning of the sound file aligned with the appropriate page and line position in the mesostic timeline. She placed a Pro Tools “marker” at the position of each newly added recording to label each with its page and line number. The accumulation of markers allowed Úna to navigate the Pro Tools session with increasing facility. The sound file, not the actual sound, was aligned with its place in the text. So for example the recording of a kettle beginning to whistle as its water comes to a boil begins when the heat is turned on under the kettle, not when the whistling starts. The whistling begins in the piece long after the point where “whistle” occurs in the recitation.

Cage gave no instructions for dealing with sounds that occur on the pages of the book between the end of the last mesostic of a chapter and the beginning of the first mesostic of the next (recall that the score gives instructions to insert periods of silence in the recitation at these divisions). Where to put these sounds? Úna was free to arrange these recordings according to her taste, while keeping to the order in which they appeared in the book. The collection of sound and place recordings was an on-going task and positioning was therefore done in stages as more were collected. The recordings going in first constrained her freedom to arrange later sounds, as did the amount of silence remaining in particular segments after these earlier placements of recordings.

We then confront the ramifications of the massive changes in recording technology since Cage wrote his score. Accompanying the process of positioning the sounds are a number of chance-determined decisions about (a) where those recordings should be positioned in the stereo field, (b) how long those recordings were, (c) how the start of each recording was to be presented in the final piece, (d) how loud they were, and (e) how the end of each recording was to be presented in the final piece. With regard to (a), we interpret this as a result of the final tape part being intended as a stereo production. Almost all the recordings were made in stereo, usually using a Zoom H4 portable stereo recorder. Some vocal sounds Úna recorded in mono were bounced to stereo. The Sonic Lab provides multiple channels for projection of sound, and we decided to interpret “a large number of positions as can be conveniently distinguished” in the stereo field to refer to the eight speakers arrayed at the various levels of the Lab. We translated Cage’s instructions regarding a stereo field into the three-dimensional context of the Sonic Lab.

We decided to distribute the left and right channels of each recording randomly among all eight speakers of the ground level array, and then duplicate this on the ceiling level. The result of this was that the perceived stereo field of the recordings varied greatly in their width and orientation, and in addition to that, they varied according to the listener’s position in the room. At the time we imagined that the audience might perhaps be mobile, as they were in some of Cage’s performances of Roaratorio, further varying the ways in which the piece might be heard. Úna found a pair of differently coloured eight-sided dice, and rolled to see which of the eight speakers would be assigned the left and right channel of each stereo recording, rolling again if both sides of the stereo file were assigned to the same speaker. If the speaker already had a sound file in it and there were free speakers in the ring, she did not send multiple sound files to each speaker, taking instead the free speaker which was closest in number to it. If there were two equally close, the dice were thrown again to decide which of those two to choose. The later chapters of the piece contained fewer mesostics, and so the number of simultaneous recordings increased as the piece went on. In these and other dense areas, where there were more than four simultaneous stereo files, Úna repeated the process for allocating e the left and right parts of each stereo file as before, reusing the speakers. In exceptional cases, we decided that a particular sound in combination with others worked better projected only from above (choral music recorded in a church, for example, which was positioned simultaneously with the sound of children in conversation). Also, we felt that some sounds were better projected in a more immersive pattern, surrounding the audience (the previously mentioned “slow-majestic note of a ship’s siren” (p. 133), for example). On these occasions the recording might be sent to additional pairs of speakers on the same level, to provide a more enveloping, less directional experience. In this way Úna made aesthetic decisions which built upon the chance-determined process of sound positioning.

We found (b) very difficult to interpret. In the case of the sound of a sound, we felt that no recording should be so short as to be inadequate to its description. In the case of the sound of a place, it seemed that if we were able to capture an interesting sound happening at that place then it should be preserved completely in the piece and not cut short. Where possible we took the duration of the content of the field recording, determined by the circumstances bearing on that recording, to be the chance determination of its length in the piece. In circumstances where no particular or unique content determined the length of the recording, the recording was stopped at a random point, usually between 30 and 60 seconds. This length was long enough for a recording to have an identity, but short enough to permit Úna to practically render a coherent overall sound featuring all the recordings without excessive overlap.

Cage’s language in (c) and (e), relates to the actions required to transfer a stereo tape recording to a multi-track tape, determining how the beginning and end of each recording is presented in the final piece.[xvi] In our realisation the whole sound file could be copied to its correct position in time without having to “roll on,” “fade in,” or “switch on,” (which we took to mean the playing of the original tape of each recording, the fading up of this output to the multi-track tape, and the start of recording on the multi-track tape). We note that if, in the earlier realisation of this piece as Roaratorio, chance determinations of these aspects took place after the editing of each recording it is possible that they affected how much of the original recording was heard. It is possible also that Cage based his decisions about “roll on,” “fade in,” and “switch on” upon chance determined timings without concern for the sounds themselves. Among the chance determinations of the duration of fades in our realisation were the distance between the start or end of the sound file and the sound we wanted to capture in that file (that is, sometimes there was little choice about the duration of the fade). Also, the indeterminate use of different zoom levels in Logic Pro, which affected the length of the fade in real time. Different shapes of fade were also used. Our general interpretation of (c) and (e) was that Cage wished for a range of presentations of the onset and coda of each sound file. In our case this instruction related to the fade digitally assigned to the sound file when bouncing it for inclusion on the timeline, and the fade applied when mixing in the Sonic Lab (see below).

With regard to (d) it seemed to us that there may be a typographical error in Cage’s score. Should it say “successions of loudness or quietness?” As with duration, we felt that there was a constraint on how quiet a sound could be. We wanted all the recordings to be audible. The level of loudness in the file would depend to an extent on the loudness of the original recording. Whenever possible when making field recordings, Úna set the microphone gain at a level which captured the sound at the strongest signal beneath the threshold of clipping. The loudness of each recording was revisited at the mixing stage.

Instruction (5) continues: Placing transparent graph paper over the typed or printed text, inscribe the placement of each tape horizontally with respect to words in the text and vertically with respect to available tracks. To do the work described in this paragraph, two people, one of them a sound engineer, must work together. When a multi-track tape is filled up, proceed to another.

We understood ‘the typed or printed text’ to be a reference to the mesostics. It is not at all clear what the remainder of this sentence means today. The use of a digital audio workstation with no relevant “vertical” (i.e. number of simultaneous tracks) or “horizontal” (i.e. length) constraints appears to make these instructions superfluous. It was also obvious to us that the work described in instruction 5 was better done by one person rather than two, and that that person be Úna and not Martin.

Instruction (5) continues: Given the circa half hour length of multi-track tapes, divide the available studio time (reserving several days at the end for final mixing and assembling tapes) first by the number of half hours and then by two (place sounds and text sounds). That will permit the planning of a work schedule so that an equal amount of time may be devoted to each aspect and each part of the composition. The work is finished when the available studio time runs out.

Having exhausted ourselves realising this piece, we recognise Cage’s wisdom in devising a method to “permit the planning of a work schedule” with a time constraint. Lacking tapes, we could not follow his method of dividing the labour. We also had virtually unlimited access to Logic Pro and Pro Tools software. We tried to adhere to some makeshift strategies of dividing our time, and determining when that time “runs out”, but were not able to do so. The date of first performance of the piece, on March 27th 2012, was the only truly effective external determinant. We also realised that editing the duration of sounds and positioning them should optimally be separate tasks, so that decisions about the former are not affected by knowledge of where they will be positioned in the piece. Úna had decided to stop collecting, editing, and positioning sound recordings before beginning the mixing stage, and she allocated five evenings, March 12th to March 16th 2012, for the latter task. This time constraint determined how much time Úna could spend mixing each part of the piece in the Sonic Lab.

Cage’s instructions do not reveal much about how Roaratorio was mixed or what he intended us to do at this stage. Cage may not have envisioned any further mixing after instruction (5) was completed, because for him no further mixing was possible. After the assembly of the recordings in time, he might not have been able to independently adjust or mix them, because they were committed to tape. For us, the potential for exploiting the full complexity of the sound reproduction system in the Sonic Lab called for further manipulations. It was still possible to adjust almost all aspects of the sound of each recording after it had been positioned on the timeline. The recordings could be separately adjusted and edited up to and even during the performance. Úna treated this mixing stage as a separate step, even though it is not mentioned in the score. Although the aspects of the recordings discussed in 5 (a) to 5 (e) were chance determined to begin with, they were further affected at the mixing stage by decisions to improve the balance and maximise the impact of the sounds, especially with regard to their loudness (d) and their position (a). Úna set the loudspeaker levels at unity and used detailed track-based volume automation in Pro Tools to adjust the relative level of each sound recording. We wanted each recording to be recognisably audible and also to blend with other overlapping sounds. These sometimes involved gradual increases in volume automation, which effectively introduced fades in addition to the chance determined fades already prescribed by Cage in instruction (5) above. The perceived level of loudness of each file in the room was also controlled by copying sound files to varying numbers of speaker pairs. Úna achieved the desired level and direction of sound through a combination of volume automation and positioning of each sound recording in the room.

Adding “Relevant Musics”

Instruction (6) begins: Using recordings of relevant musics (in performances by soloists) or composed variations of such solos, make a chance determined total program for each having at least twice as much silence as music. Superimpose these on a multi-track tape to make a circus of relevant musics.

We were inspired by Cage’s use of Irish traditional musicians in Roaratorio, but also intrigued that this avant-garde composer took Irish traditional music to be “relevant” to Finnegan’s Wake. The same question might be raised about the The Star Factory. The book is much more about narratives, images, scenes, photographs, and sounds, than about music. It seems to us that a wide variety of styles and approaches might constitute “relevant” music for either of these literary works. But for the present collaborators, all Irish traditional musicians, the choice was obvious. We decided that our “relevant musics” would be drawn from the existing repertoire of Irish traditional musicians based near Belfast. Four instrumentalists, including ourselves, plus a traditional singer, participated.[xvii]

Our chance determined program is modelled on an Irish session, where the music is spontaneous and drawn from the collective repertoire of tunes of those present, and where musicians play together in unison, but in a very loosely coordinated way which we felt corresponds well to Cage’s call for “soloists.” The performance by the five musicians in our realisation is meant to be spontaneous, with each musician free to play at any point in the piece for any length of time they deem appropriate or “relevant.” We neither encouraged nor disallowed unison playing. The total programme length of 90 minutes required that each soloist, in order to reserve “at least twice as much silence as music,” not exceed thirty minutes of playing time. Each musician was given a stop-watch to track their accumulated playing time as the piece progressed.

How might a traditional musician choose tunes from her repertoire that were relevant to The Star Factory? Martin devised the following procedure for generating a repertoire from the words of the mesostics. The series of 138 one-word columns built around THESTARFACTORY, before the adjacent words were added to complete the mesostic poems, was used as a musical template. We circulated this document asking each musician to write on each page the name of tunes they would like to play that contained a word on that page, using their own colour ink. Only those tunes for which a word in the title matched a word, or part of a word, on a mesostic page qualified. If a previous musician had already written down the tune title and they liked it too, they added a check mark next to the title, thus highlighting the common repertoire. This sampling of the collective repertoire of the five musicians generated a massive number of tunes and songs, over sixty on the first twelve pages (representing the first of 32 chapters of The Star Factory), with a significant proportion of this repertoire shared by the musicians. For each chapter, a mesostic was made using only the words from the mesostics of that chapter for which tunes were associated. For any letter, the first assignment of a tune to that letter was used, and later tunes for that letter discarded. In the case of a tune qualifying in this way more than once, only the first instance qualified. In the case of multiple tunes matching a particular mesostic word (for example “mountain”, “house”, “road”, etc.) [xviii] the tune which was more popular was selected, to give each musician more options of pieces to play (hopefully without simultaneously encouraging too much unison playing). In the case of multiple occurrences in the mesostics of words being associated with tunes (for example “dream”, “street”, “house”, “road”), only tunes which had not yet been selected qualified. So for example at the first occurrence of “dream,” chance determined that “Paddy Ryan’s Dream” be selected, and on the next occurrence “Sergeant Early’s Dream.” By this method, 32 “tune mesostics,” one for each chapter, were created. Here is the first of the 32 “tune mesostics”:

peter sTreet fisHing for eels E wheelS of the world dublin porTer the dAwn R F paddy ryan’s dreAm C sporT the irOn man loRd mayo sargeant earlY’s dream

This method created from the musicians’ respective repertoires tunes whose names were “relevant” to the text of The Star Factory. We recognised that this did not necessarily mean that they were relevant, for each musician, to the piece called Owenvarragh. Ciaran’s words are central to the piece, but in the spontaneity of the performance, many other things are happening: the sounds, the decisions of the other musicians, and the combination of all these. Each musician had a different relationship to the “tune mesostic” tunes, as we differ greatly in our memory of the names of tunes in our respective repertoires. This was particularly relevant to a traditional singer: the titles of obviously relevant repertoire (Belfast street songs, songs in praise of Belfast in the Irish language, etc.) might not overlap at all with the words of the mesostics. So we agreed that the “tune mesostic” strategy should not be the only way for musicians to be musically “relevant” in the performance of Owenvarragh. In the end, the tune mesostics served as a guide to the unfolding of the piece, but each musician remained free to react to the moment and play what seemed best, according to their own individual taste as a soloist.

We were even more concerned with how much each musician played than with what they played. While reflecting in rehearsal on the close proximity of the musicians, Ciaran, the singer, and the audience, we agreed that the piece might easily overwhelm the audience with sound. We endeavoured not to play for very long when two or more others were playing. Each musician was nevertheless still free to join another in unison if the tune they were playing suited, or to introduce a tune inspired by what was previously played. We found the piece to be more effective with less than the maximum possible sound. The resulting mixture, or circus, of relevant musics still left significant stretches of silence allowing attention to focus on the sounds and the words.

Putting it all together

Finally, we arrived at Instruction (7): Without erasing any reduce the collection of multi-track tapes to a single one. The material is then in a plurality of forms. At one extreme, each pair of tracks can be given its own sound system. At the other (suitable for radio transmission) all tracks can be heard through a single stereo system. Other uses of the material are also available, so that the various “layers” (2, 3, 4, 6) can be heard alone or in any combinations, and some of them (2, 6) as recorded, or, when some or all the performers are available, live.

Ciaran’s live recitation of the mesostics introduced into the performance the question to what extent the timing of the original recitation be preserved. As we reported, Ciaran read the mesostics for the first time in a recording session in November 2011. This first reading, like any reading, was unique. Ciaran read with only Úna as an audience. There were a number of pauses for clarification of the hand-written mesostics. As the reading progressed, his voice became livelier, introducing new accents and inflections, his pace quickened, and the recitation began to take on a rhythm he found in the mesostics. Over subsequent months Úna used the sound file of this original recording as a spine upon which she added over 300 recordings, with the silences between chapters also introduced. When in March 2012 Ciaran would read again through the mesostics, in order to move with the piece he would have to replicate not only the variable rhythm of his first reading months earlier, but also introduce the appropriate silences into the reading. We felt that it was neither desirable nor possible for Ciaran to render this replication-with-silences. As we noted above, we had added durations of silence at the end of each half of the piece to reduce the chance of the sound files coming to an end before Ciaran’s reading of the mesostics was complete. It also gave him the flexibility to give a reading which was different from his original recording, which he did in great style on March 27th.

Even with these buffers of silence, we felt that a live performance might still go so far astray that the connections between sounds and text might be lost. In our only significant addition to the score, we decided to introduce a visual element to the piece. We were aware that Cage had subsequently developed Roaratorio into a visual as well as sonic spectacle in collaboration with dance choreographer Merce Cunningham.[xix] We were particularly inspired to follow this precedent, because of the wealth of imagery, particularly photography, upon which Ciaran dwells in The Star Factory. Later we realised that we could exploit these images as cues to mark the time points of the beginning of new chapters in the sound file. We constructed a slide show, using imagery from the text, photography and other images from books mentioned in the text, and moving images of films and events discussed in the text. The slide show was constructed using the same principle Cage prescribed for the live music, that there be at least twice as much darkness as imagery. We engaged Belfast based visual artist Ita Monaghan as set designer, and decorated the room with lamps, bookshelves, books,“Star Factory” shirts, etc. We wished that the visual element of the experience underpin but not overwhelm the sounds and the narrative, while at the same time giving Ciaran and the musicians cues showing when successive chapters of the piece were commencing in the sound file.

The spatial organisation of the performers and audience was unique to the Sonic Lab. Ciaran was seated at a table in the centre of the ground floor of the room, facing a large screen at the front wall on which the slide show was projected, with the audience of 150 arranged in a U shape around him and oriented toward the screen. The singer was seated on this same level in one corner of the room, while the four musicians were arranged in a square on the structural floor beneath the audience. The musicians were amplified using only the speakers of the bottom array placed very near them, in order to highlight the unique effect of their position below the audience. With this identification of their sound and their position we intended to reinforce their role in the piece as “soloists.” The ground and ceiling-level arrays of eight were deployed to project the audio file sounds with Ciaran and the singer projected through the mid-level.

Only when the piece began to be put together in this way in the week before the performance did the relationships between the “bunching” effect within the mesostics and the actual sound of the piece become apparent. We have mentioned that the syllable index process produced mesostics drawn very unevenly from the text. Also, as Martin moved through the text and the gaps between letters lengthened, fewer mesostics were produced for each successive chapter (twelve for the first chapter Raglan Street, but only one or two for some of the later chapters). As one progresses through the book, and therefore the piece, the number of eligible words was artificially reduced by the syllable index but the number of places and sounds was not. Recall that the method for making the number of places more manageable was to reduce them randomly across the entire text, and required a complete list with many repetitions of places to be made first. This combination of procedures dramatically affected the contour of the piece. The F and Y gaps called for a disproportionate number of sounds to be positioned between lines 7 and 8 and between lines 13 and 14 of each mesostic, because of the number of sounds found on the pages of these increasingly large gaps. In contrast, constructing parts of the mesostics from a pack of closely located words reduced the number of sounds located within the “bunches.” The shrinking number of mesostics per chapter also resulted in a greater density of sounds in the second half of the piece. The piece therefore has a unique oscillation of noisiness and quietness, and it becomes increasingly boisterous as it progresses. We imagine it might have been realised much differently had we chosen Ciaran Carson’s name for the spine of the mesostics, which might have produced a much different textual rhythm. Also, the piece would be much different had Cage’s separate instructions for making manageable the collection of words, places and sounds cohered differently.

It was also only at this final stage that we could appreciate how these various elements of the “circus” would sound together. When editing Úna treated each recoding as a separate creative piece, made with consideration both of its description in the book and its ability to evoke a response. Multiple recordings were usually made and the best chosen. As we have noted, numerous other aesthetic decisions were made to make the whole tape part to into an integral piece of electroacoustic music that might stand on its own, in the same way that the mesostics might stand on their own as poetry. The most poignant moments in each were the result of a combination of chance determination and our constrained use of our taste. This affected our thinking about the arrangement of relevant musics. We found ourselves in complete agreement with Cage’s instruction to introduce periods of silence in the recitiation and to have at least twice as much silence as music from the soloists, because with regard to both the tape part and the mesostic recitation, we were forced to accept that some of these poignant moments might be lost. However, we recognised that the nature of the piece would leave the actual outcome to chance. The tape part plays the same way in every performance, but the rhythm of the recitation and the playing of the soloists would be different every time. We were also hugely indebted to our live sound engineer, Chris Corrigan, who was responsible for balancing the three main elements of the circus (tape part, soloists, and recitation) as the performance unfolded. We agreed that these three elements should be treated with equal importance. On March 27th 2012, we found this oscillating and forward moving quality of the piece beautiful and engaging. Punctuated by the occasional visual images, accompanied by the attentive give-and-take of four instrumentalists and a singer, with a recitation by Ciaran which grasped the rhythmic potential of the mesostics themselves and their coherence with the sounds and images, the piece moved continually, rising and falling, absorbing the audience into the world of The Star Factory, the world of Belfast.


[ii] This installation was designed by Michael Alcorn and Chris Corrigan in collaboration with John David Fullemann, using an eight channel tape version of Roaratorio realised at WDR Köln Studio Acoustic Art by Benedikt Bitzenhofer, John David Fullemann, and Klaus Schöning in 1997.

[iii] Mode Records, 28/29, 1992.

[iv] London, Granta, 1997. All further citations of The Star Factory are given in parentheses in the text.

[vi] We did so even though this seemed to be at odds with a later direction on how to treat unfinished mesostics at the ends of chapters and the end of the book. See below.

[vii] The book has occasional and sometimes lengthy footnotes. We chose to ignore these in the creation of the mesostics, because the presence of the words “next word” suggested to us an order of reading of the narrative which was interrupted by, rather than inclusive of, the footnotes.

[viii] Edited by William Dodge Lewis, Henry Seidel Canby, and Thomas Kite Brown (Chicago: John C. Winston Company, 1932). This work was done in a cabin near New Auburn Wisconsin, the home of this dictionary, known in the Plochman family (Martin’s in-laws) as the “trusty dusty.”

[ix] There were numerous troubling words (“Kalahari”, “Egyptian”) which did not appear in the dictionary.

[x] At an early stage we considered and rejected the idea of developing an automated process for selecting words, writing a program which follows Cage’s instructions using optical character recognition of the text. We much preferred the tactile and personal experience offered by Cage’s way of “reading through.”

[xi] We use the term “tape part” to refer to the pre-recorded element of the live performance. In its final form in our realization this consisted of a Pro Tools session with 300 soundfiles correctly positioned and mixed. This term is not to be confused with the word “tape” which is frequently used in the score and elsewhere in this document, referring to the recording media in use when Cage produced Roaratorio.

[xii] The number of lines on a page never exceeded 29. Úna used the six sided die to decide which decade (1 and 2 on the die = 0, 3 and 4 on the die = 1, and 5 and 6 on the die = 2) and the ten-sided die to choose which unit of the decade. For example, rolling a 5 on the six-sided die and a 7 on the ten-sided die meant she chose a place found on line 27.

[xiii] Ciaran at one point discusses the meaning of the word sound in Irish, fuaim: “sound, noise, clamour, report, echo” (p. 193).

[xiv] We would like to thank the following people for collecting, donating, or facilitating our acquisition of sounds: Abdullah Jamal Ashraf (“Persia”), Dionysis Athinaios (monkeys), Phil d’Alton (the Albert Clock), Rui Chaves (Belfast), Patrick Davy (trains), Aidan Deery (Mullaghbawn and much else), David Drury (many Belfast recordings), Massimo Greco (the Vatican), Martin Green (Fair Isle), Ciaran Gorman (New York), Florian Hollerweger (Wellington, NZ), Javier Jaimovich (Chile), Orestis Karamanlis (Athens), Conor Kennedy (a whistling kettle), Alana Kerr (Hollywood, LA), Caoimhe McAlister (Berlin), Brendan McCann (Clonard and Tara) Tomás McCann (The King’s Hall, Belfast), Charlie Monaghan (Black Mountain), Ita Monaghan (‘fake milkbottles’ clinking), Nóirín Nic Alastair (London) Daithi Sproule (the Mississippi), Jim Skelly (Derry), Paul Stapleton (Los Angeles) . We also thank Peter Carson and Robbie Hannan and the Ulster Folk and Transport Museum, Cultra, County Down, Malachy Moran and RTÉ Audio Services and Archive, Maria Ponce and the Dirección Técnica de Producción Estudios Churubusco México.

[xv] Raidió Teilifís Éireann, Ireland’s national public service broadcaster

[xvi] We are grateful to Chris Corrigan, Technical Manager of the Sonic Lab at SARC, for advice on the matters discussed in this paragraph.

[xvii] Úna played harp and concertina, and Martin played fiddle. We were joined by Tiarnán Ó Duinchuinn on uilleann pipes, Patrick Davy on flute, whistle, and fife, with Éamonn Ó Faogáin singing.

[xviii] There are many well-known Irish tunes using these words in their titles, for example: The Kid on the Mountain, The Mountain Road, The Green Mountain, The Mountain Lark, The Last House in Connacht, The First House in Connacht, The Woman of the House, The Corner House, The Tar Road to Sligo, Over the Bog Road, The Road to Ballymac, etc.

[xix] It appears that Roaratorio has, in recent years, become been transformed into a dance piece in the public mind. See for example:


Owenvarragh, A Belfast Circus On The Star Factory

Part 1

Part 2

Part 3

Part 4

Part 5

Part 6

Part 7


Will McConnell, film maker

Conor Gillespie, additional camera work