In most disciplines, research precedes execution—a hypothesis is formed, sources are gathered, and the work unfolds according to a plan whose logic is available to the maker from the start. The circumstances under which poems get written tend to resist this kind of advance organization. Poems arise in moments when accumulated experience and emotion suddenly condense into form, and the understanding of what mechanically produced that condensation arrives much later, through the slower work of revision and retrospection. In my own practice, revision has frequently functioned as a kind of archaeology—an attempt to excavate the puzzles that the mind had already constructed and left in the work without leaving any explanatory notes. The inquiry moves circularly rather than forward, doubling back through earlier decisions whose logic only clarifies under the pressure of sustained return.

This circularity helps explain why a first book of verse can continue functioning as a springboard decades after its composition. Many of the ideas that now occupy me began there in embryonic form, visible in retrospect but invisible at the time of writing, when I was too busy generating new work to perform any sustained examination of the work already completed. Only with the vantage of time—now being well into my fifties—have I begun unpacking the impulses that originally produced those early poems. What once appeared to be casual interests have, over the years, deepened into persistent fascinations: recursion in particular, the structural logic of reversal and cyclical return, the way palindromes and word squares and optical illusions share a common grammar with certain kinds of narrative. These structures surfaced repeatedly in the writing long before any conscious examination of them, forming a through-line in the work whose consistency I could only perceive from a sufficient distance.

The only time I consciously attempted to begin a poetic project with research was a long-abandoned sonnet cycle titled The Forum, conceived loosely as a modern analogue to Dante’s Divine Comedy—one hundred Shakespearean sonnets drawn from Roman historiography, from Julius Caesar’s Gallic Wars and the histories of Suetonius and Plutarch. The poems that resulted read less like poems than like lectures, the rhetorical scaffolding so thoroughly overwhelming the emotional core that the project was eventually abandoned for nearly twenty years. During those years, however, the writing continued, quietly informed by the rhetorical rhythms and idioms of Roman prose that the research had deposited. That influence surfaced indirectly in later works—a series of sestinas, a sonnet cycle, a set of persona poems—none of which required active research at the moment of composition because the research had already been absorbed into the available technical vocabulary. The years spent writing and eventually setting aside The Forum functioned as a long apprenticeship whose returns only became legible much later, when the internalized material began generating forms that could not have been reached by any more direct route.

This distinction between overt and internalized research matters enormously, and it is the one most frequently misunderstood by artists early in their development. Research is the process by which a maker’s understanding of a subject becomes specific enough to resist the gravitational pull of the generic—those default mechanisms by which narratives move through familiar grooves without generating friction or surprise. Every creative field has accumulated its own version of these mechanisms: the tropes and conventions that allow a work to feel legible without requiring genuine encounter with its ostensible subject. Sustained research into that subject opens it toward particularity, toward the specific detail or structural logic or historical circumstance that could not have been arrived at through the application of existing templates, however skillfully applied.

The careers of artists who made research foundational to their practice demonstrate what that particularity can produce at the highest level of execution. Albrecht Dürer spent two extended periods in Italy—first in 1494, then returning between 1505 and 1507—absorbing perspective theory, anatomical drawing, and the Venetian handling of color and light. He returned to Nuremberg carrying a technical vocabulary no German artist of his generation had access to, and that vocabulary restructured his draftsmanship from the inside, giving it a spatial authority whose sources were invisible on the surface of the work but whose effects were immediately legible to anyone looking. The research had been completely metabolized by the time it appeared in the prints and paintings, producing conviction in the work that could not have been manufactured through stylistic imitation alone. Keats’s On First Looking into Chapman’s Homer emerged from a single night of reading—he and his friend Charles Cowden Clarke working through the translation until dawn, Keats walking home and sending the finished sonnet by ten in the morning—and the poem registers what sustained attentive reading does to a body when it encounters something for which no existing category adequately prepares it, the hours of absorption making the poem’s central experience of sudden vast expansion available as felt sensation rather than reported event.

During my career in feature animation, these values were inculcated every day, as the production pipeline built this principle into the institutional structure of every project. A feature film of the scale Disney produced in that period—six or seven years from concept to release—devoted the first two-thirds of its timeline to investigation and development, long before any polished execution began. Research teams traveled to locations, studied animal behavior, consulted historians and cultural specialists, built physical maquettes, and generated thousands of drawings that would never appear in the final film. The work that reached the screen carried the weight of that invisible substrate, and the ratio of research time to production time reflected a hard-won understanding within the institution that execution without prior investigation produces imagery without depth, surfaces that audiences can receive without being genuinely moved or changed by what they contain.

This three-phase architecture—investigation, inspiration, execution—holds across every discipline where sustained creative work is attempted, though its phases are never strictly sequential and do not proceed in a clean forward line. Inspiration draws on what investigation has deposited; execution discovers structural problems that send the maker back into investigation; the phases interpenetrate across the full length of a project, each one conditioning the others in ways that cannot be anticipated from the beginning. The phase most frequently abbreviated—most readily sacrificed to impatience or deadline pressure—is the first, and the consequences show up in the work with a consistency that extended experience makes easy to recognize. A painting executed without sufficient research into its subject produces imagery that hovers above its referents, gesturing toward them without making genuine contact. A poem written without the slow accumulation of reading that builds a working knowledge of what a tradition has already accomplished tends to produce effects that exist elsewhere in that tradition, often in stronger form, in poems the maker has not yet encountered.

The history of artistic revision is substantially the history of research arriving late and restructuring work already in progress. Pound’s Cantos grew across five decades as he absorbed Chinese ideograms, Confucian ethics, economic history, and Provençal poetry—each new body of material pulling the ongoing work toward forms it could not have anticipated when the project began. Coltrane’s late period, culminating in A Love Supreme in 1964, drew on years of engagement with Indian classical music, African rhythmic systems, and the modal possibilities that Ravi Shankar’s practice had opened to him—research conducted through listening and practice rather than reading, but reshaping the harmonic language available to him at the moment of performance with the same structural thoroughness that textual research reshapes a writer’s available vocabulary. In both cases the research was folded continuously into the work as it developed, each new absorption altering the direction and scope of what the work was becoming.

In my own more recent practice, the retroactive discovery that a poem has been doing something I did not consciously intend—that a form has been operating beneath the surface of a draft, or that a set of images has been organizing itself around a structural principle I could not have named at the time of writing—has repeatedly generated new rounds of research years or decades after initial composition. The poems in Hallucinations have produced their own archaeology in this way. Forms I developed through practice before I had any name for them turned out to have partial precedents that only became visible through the research the poems themselves prompted. The Ghost Blazon emerged from accumulated compositional decisions before it existed as a named form; the research that followed gave it a conceptual framework, and that framework made subsequent formal experiments possible that could not have been imagined without the retroactive understanding the research produced. In this sense investigation does not conclude when composition begins—it shifts direction, moving from preparation into the ongoing excavation of work already made.

There is a version of this understanding that every serious maker arrives at eventually, usually through the experience of work that failed because the investigation was insufficient. The first idea—the one that arrives with apparent inevitability, carrying the subjective force of genuine discovery—tends to be the idea most thoroughly shaped by prior exposure to existing work in the same territory. Its sense of inevitability is frequently a sign of familiarity rather than originality, the feeling that accompanies recognition of a pattern the culture has already produced in numerous variations and filed under a recognizable category. Extended research into the actual subject of a work—its historical circumstances, its material conditions, its relationship to forms and traditions the maker may not yet have encountered—displaces that first idea by confronting the maker with the genuine complexity and particularity of what they are attempting to engage. The maker who remains with that complexity long enough to be genuinely surprised by what it contains is working in conditions where the work has some possibility of producing surprise in anyone else who encounters it.

---

What the preceding section describes in broad terms, the actual practice of writing makes granular and frequently bewildering. Nearly every invented form in this collection required building a lattice before touching content—working out spines, rhyme schemes, and hidden caudates, sometimes across several days of pure structural configuration before a single line of the poem could be attempted. The Ghost Caudate page on this site illustrates the scope of what that configuration can demand: Dante’s Inferno, the documented history of the Torre della Fame and Count Ugolino, a sustained analysis of Carpeaux’s sculptural rendering of the same material, and finally the lattice work required to tell the story again—the word “again” carrying weight here, because the attempt had already been made in The Forum across numerous sonnets, and the result had been rhetorical, didactic, and by any honest measure, unreadable. This raises a question worth sitting with: given material as viscerally extreme as a medieval count allegedly consuming his own children in a sealed tower, how does a poem become a slog? The answer, in retrospect, is that the research in 2005 was producing scaffolding for a poem that would not be written until 2026. The most important structural and historical work had been done; what had not yet occurred was the compression of that work into a form adequate to carry it.

Even after the compression became possible, one sonnet felt insufficient to hold the full argument. The recently completed Sestonnet series Precedents had produced a formal solution: the Acta Iterata, a six-line accumulation at the close of each Sestonnet sequence functioning as a Greek-chorus distillation of the dialectic between characters—refracted through a third-party perspective into non-adjudicating commentary, observation pitched at the archetypal rather than the particular. In the case of After Carpeaux, however, six additional lines would not constitute a tail in the traditional sense of the caudate; they would amount to an entirely new poem. The solution was to hide a condensed version of that argument collapsed into three lines, readable vertically in a structure of nested rhyme schemes: the initial words carrying a Petrarchan scheme, the fifth word of each line carrying a Spenserian scheme, and the terminal words carrying a Shakespearean scheme. Embedded as structural easter eggs, invisible to the casual reader but audible to anyone listening for them—this is the ghost resonance the form operates on, the sense that something is sounding beneath the surface of what the eye follows. The total arc from the first Forum sonnets to the finished After Carpeaux ran to twenty-two years, each phase of research becoming the raw material for the next.

Some research arrives without prior intention, in the form of an experience that only later reveals its necessity. A walk along the Iron Goat Trail in the Cascades—through the snow sheds and past the wreckage of the 1910 Wellington avalanche disaster, in which two trains and nearly one hundred passengers were swept into the canyon below—produced the material for a three-sonnet cycle, Dead Man’s Slide. The poem required names, dates, the specific geography of the slide path, and the documented sequence of the disaster in enough detail that no amount of imagined plausibility could have substituted for them; the research was simply the condition under which the poem could be honest about what it was describing. St. Catherine’s Head presented a different kind of problem: the poem was written in a fugue state inside her cathedral in Siena, the experience arriving faster than any research could have prepared for, and the visual confirmation of what her disembodied head could actually see from its reliquary required looking afterward rather than before. Honeymoon was assembled from notes made in transit, compressed and sharpened with historical scaffolding added later—though “later” in this case meant a library visit, since the poem predates Google by three months and Wikipedia by three years. The library, for all the condescension that attaches to recommending it now, remains the correct answer for a substantial category of research problems: the internet has absorbed a great deal of human knowledge, but has not absorbed all of it, and what lies behind academic firewalls or exists only in out-of-print monographs is not retrievable by any search engine currently operating.

First-person research, when available, produces a quality of knowledge that no secondary source can replicate, and the body registers it differently—in three dimensions, with all the senses operating simultaneously, including the ones that have no name. Before production began on Mulan, the animation team knew the film would require large-scale explosions from mortars and fireworks, and the resident expert on the film—having been given the honor of destroying the villain at the climax—had no practical reference beyond the Fourth of July and whatever fragmentary footage existed on an internet that, in 1996, had neither adequate search engines nor anything resembling YouTube. The studio solved this with characteristic directness: five animators were taken to a field to watch unexploded fireworks destroyed in a controlled detonation, where they put on protective glasses and settled into a trench dug at two hundred meters from the blast site, with no very clear idea of what two hundred meters of distance would actually mean.

Four lessons arrived in less than four seconds. The impact blast is invisible—the pressure wave hits the body before the eye has registered anything. The heat travels with it: every hair on both cheeks singed off at two hundred meters, which is not a detail any written description of an explosion had previously conveyed. The sound arrives late, delayed by the gap between the speed of light and the speed of sound, which the body understands viscerally in a way that no physics explanation produces. And the cloud that rose from the detonation was a mushroom—a fact that overturned the assumption, held until that moment, that mushroom clouds were the exclusive phenomenon of nuclear explosions. These four pieces of knowledge restructured every subsequent piece of work that involved fire, explosion, or the physics of destruction, and they did so permanently, in the way that only direct physical encounter restructures understanding. This is what first-person research produces that no archive can replicate: the body as instrument, registering what the secondary literature does not think to record because it does not know the body is listening.

The deeper implication of all this is that research is continuous rather than periodic—something the serious maker is always doing, even when no project is currently in view. Experience accumulates and gestates; what presents itself as sudden inspiration is frequently the delayed emergence of material that was deposited weeks or years earlier and has been processing out of sight. A leather journal beside the bed, an extensive shelf of rhyming dictionaries and history books and volumes of criticism, the discipline of ending the day with whatever the day had left behind—these are not romantic affectations but working methods, the infrastructure of a practice organized around the understanding that the material arrives on its own schedule and requires a place to land when it does. The Aristotelian peripatoi—the walking lectures, the ambulatory philosophy—understood this at the level of pedagogy: thought clarifies in motion, the body’s movement loosening what the stationary mind holds too tightly. Many of the distillations and moments of formal clarity in this collection arrived during walks, and arrived specifically because no phone was present to intercept them.

The cell phone is the most consequential research tool most makers currently carry, and also the most consequential obstacle to the kind of absorption that produces usable material. Its virtues are genuine and not worth dismissing: cameras with night sensors and depth-of-field simulation that would have required substantial equipment a decade ago, video with slow-motion and compression built in, instant access to geolocation data and historical records and fact-checking at the point of encounter, notes that do not get left in coat pockets or soaked in rain. These are not trivial capabilities, and they are woven into a regular working practice without apology. The liability is equally real: a mind connected to the feed at every available moment loses the interval between experience and reflection in which the most useful processing occurs. The experience goes directly into the phone and stays there—documented, retrievable, and inert. The experience that goes unrecorded, that the mind holds and turns without external assistance, is the one that tends to generate the formal and linguistic pressure that poems require. The walk without the phone is not a rejection of the tool; it is the condition under which the tool becomes useful, because the material it will eventually help to organize has first been allowed to move through the body without interruption.