The Poem as Engineered Structure: A Technical Introduction to Modern Poetics

The Crisis of Reading: Why "Liking" a Poem Is Not Enough

Every undergraduate who has survived a poetry seminar knows the small, embarrassed pause that follows the question, "What did you think of the poem?" If the answer is "I liked it," the seminar dies. If the answer is "I didn't get it," the seminar dies in a different way. The reason both responses kill the conversation is the same: neither offers a *technical account* of what the poem is doing. The first substitutes affect for analysis; the second confuses unfamiliarity with absence. Between them lies the entire problem this book proposes to solve—how to convert our private, inarticulate responses to modern verse into public, reproducible descriptions of how poems work.

The difficulty is not new. In 1934, F. S. Flint, recalling the early manifestos of imagism, complained that the new poetry had been "explained mostly by people who felt it without understanding it, or understood it without feeling it." Eight decades later, Flint's complaint still circulates in classrooms and critical reviews. We admire, or we dissect; we rarely do both in the same breath, and almost never do we explain our admiration in language that another reader could verify, replicate, or refute. The result is a critical culture that talks fluently about whether poems are "good" and very poorly about why they are made the way they are.

This book is an attempt to repair that defect. Its premise is unromantic but, I think, defensible: a poem is a designed artifact. Like a bridge, a sonata, or a printed circuit, it has been put together under constraints, using parts that can be named, measured, and compared. Once we grant that premise, a different kind of reading becomes possible—not the reading that asks "how does this poem make me feel?" but the reading that asks "what has the poet built, out of what materials, against what resistances, and to what end?" The first question is not forbidden; it is merely incomplete. The second question is the one this book is designed to answer.

The move from "liking" to "specifying" is a move from consumer to engineer. To like a poem is to enjoy its outputs. To specify it is to enumerate its inputs. Once enumerated, the inputs can be reordered, replaced, or removed, and the consequences observed. That procedure—isolate a variable, alter it, measure the change—is the engineer's procedure, and it is the procedure we will use throughout the book. We will treat poems as engineered structures not because metaphor is charming but because the engineering analogy yields specific, testable claims about how verse functions.

A note on scope. We are concerned almost exclusively with modern verse, by which I mean the body of poetry written in English roughly between 1910 and the present day that has consciously defined itself against nineteenth-century models. The opening of that period coincides with the founding moments of imagism and the early vers libre experiments of Pound, H.D., Williams, and Eliot; its closing edge is the present moment, including the conceptual and procedural poetics of the last forty years. The technical problems of these poets—how to render an image without Romantic effusion, how to make free verse rhythmic without meter, how to renew form without regressing to inherited patterns—will be the recurring problems of every chapter that follows.

Romantic Expressivism and Its Limits

Before we can justify a technical reading of modern poems, we must understand the dominant tradition they reacted against. That tradition is what we can loosely call *Romantic expressivism*: the view, codified most influentially by Wordsworth in the 1800 Preface to *Lyrical Ballads* and elaborated throughout the nineteenth century, that poetry is the overflow of powerful feeling, that the poem originates inside the poet and is then transmitted to the reader through a transparent medium. On this view, the technical features of verse—meter, rhyme, image, syntax—are *conduits* rather than *contents*. They exist to deliver a pre-existing emotional state; they are not themselves the object of attention.

Three consequences follow from the expressive model. First, technical analysis is subordinated to biographical or psychological interpretation: to understand the poem is to understand the poet's feeling. Second, formal features are evaluated by their transparency or invisibility: the best meter is meter the reader does not notice, the best image is one that disappears into the feeling it carries. Third, and most consequentially for our purposes, the poem is treated as a one-way pipeline: it originates in the poet's interior and arrives at the reader's interior, but the conduit itself—the technical machinery—is, in principle, dispensable.

The expressive model was not wrong about everything. It correctly identified that poems often carry strong affect, and that readers respond to that affect. It was, however, inadequate as an account of *how poems produce their effects*. By treating technique as a transparent carrier, it left readers unable to explain why one technical choice succeeded and another failed. Worse, it left nineteenth-century poets themselves unable to defend their innovations on anything except biographical grounds: "I felt strongly, therefore I wrote freely." That argument was unavailable the moment a later poet also felt strongly but chose to write in strict meter.

By the early twentieth century, the expressive model had begun to buckle under the pressure of two contradictory demands. On one side, late Romantic poets like Swinburne and the early Yeats had stretched the model toward ever more elaborate formal controls; on the other side, the symbolist and post-symbolist poets of France had begun to treat the technical apparatus of the poem as itself a kind of content. The stage was set for a wholesale revaluation.

The Modernist Turn: From Voice to Structure

The revaluation arrived under the sign of *construction*. In a 1913 letter to Harriet Monroe, editor of *Poetry*, Ezra Pound crystallized the new doctrine: "I believe in technique as the test of a man's sincerity." A year later, in "A Few Don'ts by an Imagiste," Pound sharpened the claim: "Use no superfluous word, no adjective which does not reveal something ... Don't imagine that the art of poetry is any simpler than the art of music, or than the art of painting." Notice that the test is no longer the poet's feeling but the poet's craft. The expressive model had asked, "Does the poet feel this?" The constructivist model asks, "Has the poet built this well?"

The vocabulary of construction saturates the early modernist manifestos. Pound writes of "logopoeia" and "melopoeia," of the poem as a "sort of algebra." T. S. Eliot, in "The Metaphysical Poets" (1921), praises Donne and Marvell for a "direct sensuous apprehension of thought" and a "mechanism" capable of "fusing" disparate experiences. William Carlos Williams, perhaps most consequentially for this book, declares in the introduction to *The Wedge* (1944) that "a poem is a small (or large) machine made of words." The Williams formulation is the one we will return to most often, because it makes the engineering analogy explicit: a machine is something with parts, something that does work, something whose performance can be specified.

The modernist move had three large consequences for how poems were written and read.

First, technical features became *topics* in their own right. Meter, lineation, syntax, and image were no longer transparent carriers but visible, decidable choices that could be evaluated. Williams's triadic line ("The red wheel-/barrow"), Eliot's fragmented syntax in *The Waste Land*, Pound's ideogrammic method in the *Cantos*—each is a deliberate technical decision that invites analysis on its own terms.

Second, the expressive hierarchy was inverted. Instead of feeling producing form, form now produced feeling—or, more precisely, feeling emerged from th