Capital Improvements: The Initial-Caps Wars

/ /

Like all specialists, poets and critics can become greatly exercised over matters that look small to everyone else. Here is one such matter: whether a poet may observe the old rule of using a capital letter at the beginning of every line in a poem. I’ve been conversing with fellow poets online for almost two decades, asking about this every so often, and the answers always include a certain noisy angst. Among poets in the US (things in the UK seem less polarized) and especially among poets who work in form and meter, the warring camps cry aloud angrily:  You mustn’t capitalize—it’s reactionary! You must capitalize—it declares your allegiance with the tradition! Don’t capitalize—it interrupts the flow of the meter! Capitalize—it asserts the integrity of the meter and the line!

One reason the war grinds on is that combatants on both sides make huge generalizations that lack the support of evidence. And the generalizations harden into editorial biases and house styles. It’s time to probe those generalizations, see what there is to back them up, and where evidence is lacking, say so.

I’ll examine some typical claims, with the caveat that I’m talking primarily about poetry in English.

Claim 1: Line-initial capitals are just a printer’s convention.

This claim neglects a great deal of history. The use of capitals (or majuscules) to set off poetic units is much, much older than printing. It seems to date back to Roman antiquity and shows up in many examples in medieval Europe. To simplify—and probably to oversimplify—I’ll stick to examples of its evolution in England.

The wealth of manuscript material now available online makes it easy to see that, over the history of book-making in England, devices were added to poems to make reading easier. Imagine that we start with scriptio continua, in Latin documents, in which there is no space even between words.  Later, in manuscripts like the Lindisfarne Gospels at the turn of the eighth century, we can see the beginnings of those spaces, and the use of larger letters that mark the openings of the gospel chapters. Our words initial (from initium, beginning) and capital (from caput, head, and so related to capitulum, chapter) descend from those early uses. Those letters are not only larger but also decorated, and they certainly draw the eye.

Moving on: in the tenth-and-eleventh-century book known as the Junius Manuscript, famous for heroic Old English renditions of biblical stories, we see capitals in more places. Enlarged, decorative capitals still mark the beginnings of poems, but small capitals also mark the beginnings of many paragraphs—and meter is marked, too, with dots between one half-line and the next. As Old English evolves into early Middle English and meters become more often accentual-syllabic, we start to see books in English using a changing assortment of tools to clarify meter. For example, the twelfth-century scribe (who was probably also the poet) of the metrical poem called the Ormulum uses initial capitals along with punctuation to set off metrical units: there are no line breaks, but colon-like marks appear at the ends of tetrameter units and periods at the ends of the trimeter units that follow them. Another book, a thirteenth-century manuscript of the early Arthurian poem called Layamon’s Brut, uses those same marks, but on rough-accentual half-lines. It uses capitals and red highlighting, too, but apparently for grammar rather than meter. Other thirteenth-century books, like this manuscript of the rhyming poem The Owl and the Nightingale, use not only line-ending (or sentence-ending) dots and red highlighting but also line breaks, with initial capitals on most lines. By the fifteenth century, when the making of books outside religious houses becomes a well-organized business, scribal practice becomes more settled, and we usually find that line-end dots are dispensed with, lines of poetry are broken, and capitals appear on every line, as in Corpus Christi College’s manuscript of Chaucer’s Canterbury Tales.

Not all late-medieval English books were so carefully made. Quicker jobs done in a cursive hand, rather than a book hand with detached letter forms, did not always have initial caps. One such down-market product—one I worked with for my own long-ago dissertation—reverts to red highlighting, with a sloppy red stripe drawn down the page through all the first letters of lines, actually making them less readable. But the high-quality manuscripts were the ones the first printers tried to imitate;  William Caxton’s first printed edition of Chaucer, set next to any of the famous Chaucer manuscripts, makes the likeness clear.

There is more, and earlier, evidence of line capitals in books made outside England, and a different timeline can be drawn. But this much is clear: Initial capitals in poetry came into printing because they had been used for centuries already to show metrical structure. In short: Caps as a convention came first; printing later.

Claim 2: Poets stopped using capitals on every line starting with the Modernists. Caps scream nineteenth-century “POESY.”

I regularly hear from people who have studied creative writing that they were told absolutely not to capitalize, on the grounds that line-initial caps are of a piece with Victorian poetic diction, inverted syntax, archaisms, and sentimentality. (Granted, those people were often told to abandon meter and rhyme, too.) But check that anthology you might have on a shelf, and you’ll find that not all poets from the Modernists on can be lumped together as abandoning capitals.

Quick-and-dirty literary history teaches that the groundbreaking noncapitalizers were William Carlos Williams and his contemporaries, and that this innovation burst upon the world in the little magazine Others, when Williams became its assistant editor in 1916. In fact, the trend away from capitalization starts earlier, with F. S. Flint, under the influence of French vers libre. For this information I’m indebted to Timothy Steele, whose current work in progress is a much more detailed and scholarly history of line capitalization than I can attempt here.  Flint, as Steele reports, used noncapped lines in a poem in The New Age in 1908 and became, in 1913, the first poet to appear in Poetry magazine without line caps.

Among the Modernists, though, there are also plenty of capitalizers. Some poets made their moves haltingly, settling eventually on lowercase. In H. D.’ s poems we can find a mix of approaches, and Marianne Moore started with caps and underwent a conversion. But many others stuck with line capitals.  Pick up T. S. Eliot’s Collected Poems, 1909–1962, flip its pages, and you won’t find a single line that begins with a lowercase letter. Skim through the Poetry Foundation’s offerings of Ezra Pound and Gertrude Stein: yet more capitals.

A look through the various Norton anthologies might reveal whether, and when, there is a clear trend toward fewer capitals. Perusing my oldest one (of English Literature, second edition 1968) I see not one lowercase line-initial letter, not in Robert Graves, W. H. Auden, Louis MacNiece, Dylan Thomas, Donald Davie, Thom Gunn, or Ted Hughes, none of them guilty of “poesy.” Another Norton on my shelves (of Poetry, dated 1996) tells me more: poems that start some lines with lowercase letters appear only scatteringly until one reaches the poets born at about midcentury: Louise Gluck, Michael Ondaatje, Michael Palmer, Eavan Boland, Craig Raine, Jane Shore, Leslie Silko, Richard Kenney, Agha Shahid Ali, Deborah Greger—a long list that ends with Cynthia Zarin, but has some capped interruptions. Mary Jo Salter would be added to this non-cap list, but she is one of the anthology’s editors and has not included her own poems. So, line-initial mixed case looks less like a Modernist cause and more like a Boomer reflex. (Full disclosure: I am a Boomer.)

And what about poets writing in the twenty-first century? A. E. Stallings, the American living in Greece who was recently elected Oxford Professor of Poetry, has been using line-initial capitals almost exclusively from her first book in 1999 onward. Emily Grosholz, with an even longer career, used mixed case in her early books—but in her recent new-and-selected has gone back to line-initial caps. Flipping through Alfred Corn’s recent collected volume, which sums up fifty years of publication, I find from the outset an eclectic mix. In Dana Gioia’s collected 99 Poems, same. And Kim Bridgford, and Greg Williamson. Among poets of a younger generation, there are James Matthew Wilson, Ryan Wilson, Morri Creech, Matthew Buckley Smith, and (most of the time) Ernest Hilbert and Stephen Kampa. In the less formalist world, there are Lucie Brock-Broido, Tracy K. Smith, Jericho Brown, and (sometimes) Amit Majmudar.

In short: Any generalization about what “poets have stopped” doing will meet a hundred counterexamples.

Claim 3: Line-initial capitals in poetry interfere with reading comprehension.

This is the loudest claim of the cap-haters, who often grumble that cap-users must be too dumb or lazy to override the feature of their word processors that capitalizes after every hard return.

There are reasons to argue that capitals might make a reader pause. We’ve seen that their original purpose was apparently to do something very near that—to stand out to the eye. And readability research asserts that blocks of full capitals, LIKE THESE, are harder to read than words in mixed case. That dictum is now a commonplace of page design, especially of web pages. Might a vertical stack of capitals slow reading in the same way because they’re massed at the left margin? Might they misdirect because they don’t convey the usual information about sentence beginnings or about naming? Or does the brain tune them out (as quite a few readers report) because they’re irrelevant to the information? These are questions worth asking but unasked, as far as I can discover, as research questions in the psychology of reading.

Some readers claim that the visual bulk of capitals distracts them; Simon Armitage is said to have compared their use to bolting a fire escape to the side of a beautiful building. Others claim they don’t notice. But how well can we trust either claim? We’re all familiar with certain optical illusions, like the face/vase graphic, that demonstrate how tricky human perception is and how much our seeing depends on the way our attention has been directed. If you’re studying the effect of caps on comprehension and you ask readers directly how capitals affect them, their statements won’t carry much weight.

To investigate whether readers really are distracted or misdirected by capitals per se, a researcher would have to be careful of a great many matters: deciding what audience to test (experienced or naïve readers of poetry?),  deciding how to find and recruit members of that audience and select them randomly, deciding how simple or complicated or familiar or unfamiliar a text to use as a test piece, writing a script to explain the investigation, and presenting the text to each subject individually. Then how will distraction be measured? If it’s by eye movements, there must be eye tracking devices. If it’s by comprehension questions, a set of questions must be designed that includes enough distractors not to reveal that the effect of capitals is the real subject being studied.

Are we ever likely to get such research? It may be possible, but it would need to be done with the methods of cognitive psychology and linguistics.  In the next section I review two studies that at least graze the questions we’re interested in.

Claims 4 and 5:
Capitals interrupt the flow of the meter.
Capitals helpfully indicate a new unit of meter.

We have here two opposing claims. Do capitals support meter, or do they undermine it?

One would expect the fans of traditional form to argue that line-initial capitals help us see meter. So it’s odd to find traditionalist critics reporting that the contemporary student of poetry perceives meter—in poems with those capitals—poorly, and performs that poetry in ways that reflect no metrical understanding. William Logan, for instance, in The Undiscovered Country, describes a rather frightening example:

A colleague of mine once turned to me in the middle of a thesis defense and said, “Did you know that many of Wallace Stevens’s poems are syllabic? All the lines have ten or eleven syllables.” When I pointed out to her that the poems were in blank verse, she complained that I was trying to impose an academic scansion on these beautiful, rhythmic, syllabic lines. And I could not convince her of her folly.

The line-initial capitals in Stevens—he uses them both in blank verse and in free verse— seem to have had no influence on the colleague in question.

Garrick Davis’s observations (in the essay “The Innocent Ear”) of MFA students reading aloud certain translations by Spender lead the same way:

I was . . . astonished by two facts. The first was that the nine students who read had, not a variety of speaking voices, but a remarkably uniform delivery: they mumbled out the lines as rapidly as they could read them, oblivious to line breaks, or rhyme, or rhythm. Garcia Lorca was read, essentially, in the same monotone accorded an office memo.

The second fact was that the teacher, by invariably interrupting each student after a few lines to correct the speed and intonation of their faulty recital, was aware that these near-graduates of a prestigious writing program, these most promising of our young poets, had still to learn how to read poetry.

So according to these two, the long tradition of line-initial capitals has not apparently helped present-day readers to perceive meter, or to reproduce it orally.

By contrast, the actual empirical research—or at least the two articles I can find—suggests at least that line-initial capitals do something of benefit to the reader. When capitals are present, test subjects recall texts better, or they read passages aloud more slowly and carefully. Unfortunately, neither study made the exact comparison we want: comparing a text with both caps and line breaks against the same words, but with line breaks only.

David Hanauer’s 1998 paper was aimed at testing certain theories about how students read poetry. To test the truth of those theories, Hanauer wanted to know what features of a text subjects were actually paying attention to, and he used “verbatim recall” as a measure of that attention. Hanauer tested undergraduate subjects’ recall of four texts: (1) a short metrical poem by James Joyce with line-initial capitals, line breaks, and xaxa quatrains; (2) the words of the poem, but without capitals or line breaks; (3) a text with capitals and line breaks, but without the original rhymes and with a number of other word changes that disrupted meter; (4) the same text as in (3), without capitals and line breaks.

What Hanauer found was that subjects’ verbatim recall was worst for (2) and (4), the samples without caps or line breaks. It was better for (3), the sample with caps and line breaks but with word changes that disrupted meter and rhyme, and it was best by far for (1), the original poem, with its rhymes, meter, and formatting—a result that will hardly surprise lovers of meter and rhyme, although plain words like “meter” and “rhyme” appear not at all in the paper.

A more recent study by Stefan Blohm and others (2022) does include words like “verse” and “meter” and “prosody,” although its goal was to test theories of “discourse comprehension.” It was designed to probe the hypothesis that “readers adjust their reading behavior and their oral text performance to the literary genre.” The experimenters tracked readers’ eye movements and recorded their oral readings of forty-eight different constructed texts, none more than two lines long. All were in German, with the same set of words described to subjects as poetry (and formatted with line-initial capitals and line breaks) or as prose (formatted without capitals or line breaks).

Blohm’s results that speak to our initial-caps debate are these: Readers articulated poetry more slowly, with longer durations and more silent pauses. Eye movements, too, were slower with poetry. But evidence that subjects were perceiving meter was not clear: “ . . . we did not observe increasingly rhythmic articulation in oral poetry reading. Thus, our results do not lend support to the idea that strategic poetry comprehension per default involves constructing representations of systematic prosodic regularities (i.e., the metrical pattern).” The authors suggest several reasons why awareness of meter might not have shown up in their experiments, but the bottom line is that it didn’t.

So, do line-initial capitals obscure meter or support it? In short: What line-initial caps do, by themselves, to readers’ perception of meter remains a puzzle. Neither of the studies above tests what we want tested: caps and line breaks versus line breaks alone. The case is still unproven, the right research still to be done.

This leaves us with the sorts of claims about capitals that can’t be demonstrated by experiment, because they are simple assertions of opinion:

Claim 6: Formalist poets who don’t use capitals are silly. They’re just trying to ape free-verse poets so they can look avant-garde and hang out with the best people.

In the claim above, I’m paraphrasing a fairly prominent formalist critic, who shall remain nameless. Think about this statement a moment. How would you classify poets who are regularly published in The New Yorker and have been receiving major awards for many decades? Those would be some of “the best people,” wouldn’t they? I’m referring to Frederick Seidel and to John Ashbery, in his work before 1991. They used capitals, both for formal and free verse. And look also at Ted Hughes’s “Pike,” Gary Snyder’s “A Walk,” or Yusuf Komunyakaa’s “Facing It,” to name a few random examples. Or U. A. Fanthorpe (just on the edge of meter but not quite). Free versers and capitalizers all. Flipping through Sylvia Plath’s collected poems reveals not a single lowercase initial letter. The idea that metrical poets are being beaten down, oppressed by the hegemony of the free versers into using mixed case at the start of lines, is hard to square with the evidence broadly viewed.

Moreover, some formal poets choose not to capitalize in specific forms, or in translations of those forms. One example is Charles Martin. In his book Signs and Wonders, most of the poems do use line-initial capitals, but the poems in sapphics and alcaics have capitals at the beginnings of stanzas rather than lines. The translated sections of Ovid’s Tristia also lack line-initial capitals, even though that translation is in heroic couplets. There is no free verse in the book; the dividing line is between the poems related in some way to the classics and everything else. “Trying to look avant-garde” is clearly not an issue here, and mixed-case initial letters are used with older poetry, not newer.

Consider also Joshua Mehigan’s poem “Fire Safety,” a good demonstration that a formal poet can convey meter without the help of capitals or even of equal-measured lines. Stanza breaks and rhyme do the work instead. This can hardly be called aping free verse.

So where are we?

I’ve reviewed two claims that are, I contend, inaccurate, three that are unproven, and one that is just plain silly. I move now from claims about how readers will read the poems, to claims about how the poet feels about capitals while writing the poems. These are claims about associations inside the poet’s working brain, and it may not be possible either to confirm or to refute them. It is possible, though, to assert that there are better and worse reasons for defending one’s choice. The better ones, I think, make a connection between the choice and the poetry it produces.

Here, for example, are some of the reasons Alberto Rios gives for using capitals:

to remind myself that I am writing a poem; to connect myself to history for a very brief moment before I go on to say what I myself have to say now; to give each line—however subtle—its own authority; to suggest that, although I may be telling a story, it is not a regular story, and certainly not prose; to make my enjambment have to work honestly, and to give my end-stopped lines greater Moment; to build up thoughtful pacing in a poem, suggesting or invoking a little more strongly all the reasons we break lines to begin with—breath, heartbeat, dramatic intention; to recognize this use of the shift key as a self-conscious act, which raises the stakes for everyone and everything—the poem, the poet, and the reader; to do more work in this small moment, knowing that work makes more things happen. . . .

In case you assume that Rios is speaking of metrical poems, let me present his example:

Teodoro Luna’s Two Kisses

Mr. Teodoro Luna in his later years had taken to kissing His wife Not so much with his lips as with his brows. This is not to say he put his forehead Against her mouth – Rather, he would lift his eyebrows, once, quickly: Not so vigorously he might be confused with the villain Famous in the theaters, but not so little as to be thought A slight movement, one of accident. This way He kissed her Often and quietly, across tables and through doorways, Sometimes in photographs, and so through the years themselves. This was his passion, that only she might see. The chance He might feel some movement on her lips Toward laughter.

About any of Rios’s reasons for capitalizing, another poet might say, “Well, it wouldn’t work that way for me!” And the point would be beyond argument, a matter dependent on each poet’s fickle muse. But who would want to deny to a poet any method that might have such benefits? Surely there is a lot to be said for learning tolerance of caps in reading, even if we choose not to use them, if the poets who do use caps believe so strongly that caps help produce better poems. There is even more to be said for teaching tolerance of caps in reading, so that we don’t cut readers off from centuries of poetry in English.

And for the capitalizers, surely an equal tolerance of the mixed-case metrical poet is in order. I haven’t found a manifesto that matches Rios’s statement but speaks to the other side, so let me try to put the case myself. I will be continuing to begin the lines of (almost all) my poems (which are almost always metrical) with mixed-case letters to remind myself, while I write, of these principles:

that sound, and not typography, must make clear that this is a metrical poem;

that the line break must be validly chosen for the syntax and must do its work by itself;

that the choice of a place to break should stay fluid during composition, and not be set up at the outset with a concrete monument;

that the poem should not be boxed into a meter it resists, but should be helped, gently, to find a form.

What then has my list of claims accomplished? Is it likely to change your mind about your own reaction to those line-initial caps? Perhaps not. But what I do hope for is less reliance on inaccurate or unproven claims and a more generous live-and-let-live attitude. A lowering of the noise level. A little more tolerance, in this most intolerant of times.