Horace’s “Ars Poetica” (or “horse’s arse” poetica, as some certain academic wits and half-wits have joked over the centuries, and who can blame them, poetry often needing a good kick in the pants to lighten up a bit) is a poem about poetry itself, where the poet reflects on their craft and answers the questions of what, how and why. I am making a similar attempt in this short article, though where Horace spoke of poetry’s affiliation with drama, I concern myself with poetry’s affiliation with music, most specifically, song writing.
In 2025 I had a collection of poems published by ECW Editions (a Misfit book) titled …more songs the radio won’t play… (the title inspired by Canadian singer-song writer Kathleen Edwards’ tune, “One More Song the Radio Won’t Like,” which got me to thinking, yeah, it’s one thing for an artist of any description to create a work, quite another to figure out how to get it out into the world and market it) in which I took popular songs from varying genres and through a series of techniques, i.e.: mixed discourse (including allusions and verbiage from the arts and sciences, philosophy, pop culture, literary theory), plus self-referentiality, plus the insertion and bending of biographical and historical facts (both of the songs and the performers) and tried to alter the lyrics so as to make the recycled songs outwardly unrecognizable while retaining some lingering sense or ‘feel’ of the original. Whether I succeeded or not on any substantive level is up to each individual reader, if and where there are any. What I did accomplish (I think) was to create pieces that were virtually impossible to exist as radio-playable songs. And if it all sounds too heavy and formalistic, I assure you, it isn’t. Wasn’t. It was a joy for me, with often humorous results.
Fine, I won’t bore you with examples. Suffice to say, I put my poetry writing hat back on the hook until I stumbled upon a Molly Johnson CD, titled Molly, where I noticed that the songs in the liner notes were written in block form as opposed to the usual stanza/chorus fashion that I was used to. Later still, I found that the group Chumbawamba had done the same thing with their CD, WYSIWYG, though adding side-note anti-establishment commentary rants alongside their already anti-establishment lyrics aimed at consumer society, vanity culture, corrupt politics and the soul-less entertainment industry: “all dressed up in drag inside a Gucci body bag,” as they so sweetly put it. Very cool, I thought. The songs were written more prose-poem-like, which made me imagine it might be fun to write my own prose poems, but use some song lyric techniques, such as including obvious hooks, bridges, rhythms, and nonsense phrases like “sha-na-na,” “oh me, oh my,” and “hey, hey, yeah, yeah, yeah…” in the mix. I proceeded to write seventy pieces (to date) under the title Borderblur Songspiel, “borderblur” issuing from the notion that I’m blurring the line between fiction, poetry and song, “songspiel” issuing from the Brechtian practice of using actors rather than trained singers to deliver stage songs in a manner more spoken than sung.
The manuscript is currently out in the ether, knock, knock, knocking on publisher’s doors. Wish me (and the manuscript) luck.
Now, most recently, I happened to be watching the local news where two broadcasters discussed using an AI generator program that enabled them to take their words (or lyrics) and by pasting and hitting a few choice buttons as to genre, mood, and gender, the program spat out a song, no knowledge or expertise of musical notation or instrumentation required. Which was perfect for me, since even though I enjoy listening to music, I have no professional training nor natural gift for the category. To wit, I recall having taken ballroom dance lessons with my ex-wife and still had to ask: Is this a fox trot or what? I still had to count the steps in my head. The concepts of 4/4 time and just ‘feel’ the music and let it move you were alien to me. I don’t hear the beat nor does the beat have any dance-like influence on my body, especially my two left feet. Needless to say, the only musical instrument I’ve ever been able to play with any sort of proficiency is the radio. But, I digress. At any rate, it struck me that maybe I could take some of my latest songspiel prose poems and convert them to actual songs. I searched for a program — the simplest I could find, natch, me being a confirmed Luddite — paid my money, and got down to business.
What discoveries did I make, you might ask? Well, to begin, I learned that much of my language — which I felt was perfectly adequate, interesting and presentable in its poetic state — often came across as too highfalutin and pretentious as a song. Add to this the fact that the AI generator had its own limitations, such as difficulty pronouncing certain proper names (‘Ingeborg’ with a soft rather than a hard first ‘g’), difficulty with long words (‘technological’ as one example, which was more-or-less slurred in some versions), a censor that made it unable to generate a particular text, though no reason given (I had to re-read a piece to see if I could discover what might be causing the difficulty and thought maybe the fact I’d used ‘fag’ instead of ‘cigarette’ might be the guilty culprit, and, sure enough, when I made the switch, everything was fine), and most obvious, I could tell that the program didn’t always understand the text’s meaning or intent, so sections became a wash, the program more concerned with making a line scan according to a built-in algorithm as opposed to having the line (or lines) make logical (never mind emotional) sense: the program didn’t really understand or feel the lyrics to any great degree. All of this suggesting in no uncertain terms that I had to edit my work both to suit the generator’s programming as well as to meet my own needs. And let’s be honest here, I mainly wanted to hear my words come out of someone else’s mouth as a song, so I was willing to compromise and even sacrifice a level (or two) of poetic erudition and creativity. As for the question, “to rhyme or not to rhyme,” there really wasn’t much argument from me: Make it so, number one!
Bottom line, it’s been kind of a fun experiment and I’ve now managed to generate a couple of songs that I’m reasonably happy with. Sure, the music itself is generally simplistic and recognizable, though, at the same time, I was sometimes pleasantly surprised by the inclusion of back-up vocals and instrumental riffs as a sort-of bonus. Anyway, beyond the music, I appear to have come full-circle with this endeavour, first mutilating songs to become poems, secondly mutilating poems to become songs, to the point where now I’m unsure where one form ends and the other begins.
Thanks to Kathleen Edwards, Molly Johnson, and Chumbawamba for their inspiration, albeit unaware. And if all I’ve ended up with is “one more song the radio won’t like,” nothing new there! It isn’t the first time for me and I’m sure it won’t be the last, so, I’ll just continue to futz about and play happily alone in my own little sand box. Rock on!
Stan Rogal lives and writes in Toronto along with his artist partner Jacquie Jacobs. His work has appeared almost magically in numerous magazines and anthologies. The author of several books, plus a handful of chapbooks, a 13th poetry collection was published in March 2025 with ecw press. Co-founder of Bald Ego Theatre and former coordinator of the popular Idler Pub Reading Series.







