I recently realized that it has been a year since I started writing regularly on Substack (can I now officially call myself a “Substacker”?), and one of the unexpected joys of sitting down week after week to write these articles is the insight it’s given me on how important editing can be.
I make no claim to be a great prose stylist. I claim only that the versions of the article I hit “send” on are much better than their first drafts.
Learning to write well for the internet has been a long process, though, and I’m nowhere near the end of it. But I’d never have got this far without an amazing editor. Through our work together, I’ve learned how even seemingly small changes to the prose can make accessible topics out of things which many would consider too dry or difficult for a general audience.
When the sentences flow smoothly, the reader’s attention flows along with them. When the prose stumbles, the reader stumbles too.
So I’ve become a little bit obsessed with the kinds of changes that have come out of the editing process. And I noticed that the changes I would make were most often at the ends of sentences or clauses. I found that a clumsy sentence could often be made much more graceful by changing nothing but the phrase at the finish.
But for a long time, I had no real theory about what I was changing or why.
This bothered me for a while — I like my theories, after all — until, some weeks later, with my mind steeped in theories of poetic rhythm, I realized that my editor and I had been doing by instinct something that has thousands of years of theory behind it. But it’s largely disappeared from the discussions we have about writing style today.
I’m talking about the study of prose rhythm, and more particularly, theories of cadence: the rhythm of the end of sentences (and other things which we’ll talk about soon).
Writers have cared about “sticking the landing” — that is, ending sentences on a satisfying rhythm — since the days of orators of Ancient Greece.1 They had very clear ideas about which rhythms worked at the end of sentences and which ones didn’t, as did the ancient Romans and the medieval churchmen. But none of them wrote in English.
So I set out to see if anyone had examined how these ancient and medieval theories of prose rhythm might apply to English. This article is the result of my investigations.
One caveat before we begin: I want to emphasize that I have found no recipe for good prose, no step-by-step list to follow that will guarantee a good result. But what I have found is a way of analysing and classifying some of the rhythmic ingredients of English prose, to give you a better idea of what exactly — to continue the metaphor — you’re cooking with.
You’re reading The Dead Language Society. I’m Colin Gorrie, linguist, ancient language teacher, and your guide through the history of the English language and its relatives.
Subscribe for a free issue every Wednesday, or upgrade to support my mission of bringing historical linguistics out of the ivory tower and receive two extra Saturday deep-dives per month.
If you upgrade, you’ll be able to join our ongoing Beowulf Book Club. You can also catch up by watching our discussion of the first 1962 lines (part 1, part 2, part 3, part 4) right away.
This too starts in ancient Greece
It seems to have been en vogue in the Edwardian period (1900–1919) for scholars to examine the rhythm of English prose.
They were drawing from recent studies that had elucidated first the rhythmic tools used by classical orators such as Cicero, and the subsequent development of those tools into the Latin of the Middle Ages.
To understand how all this applies to English, however, we’ll need to acquaint ourselves with some of the terminology that originally applied to the study of prose rhythm in Latin and Ancient Greek.
Today, we divide prose into paragraphs, sentences, and clauses. We mark these divisions typographically: paragraphs are signalled to us through the use of indentation and spacing, sentences by capitalization at the beginning and the use of the period or full stop (.) at the end, and clauses by the use of punctuation like semicolons (;), and colons (:) at the end, although not entirely consistently.
The ancients understood prose divisions differently. Their definitions were based on oratorical practice, that is, the art of public speaking. What these ancient concepts of prose organization care about is rhythm. And rhythm is determined by how the words hang together into larger units, and where it’s appropriate to take a breath.
Their largest division in this ancient system did not correspond to our modern paragraph,2 but to our modern sentence. They called this division a period, from the Greek περίὁδος períhodos ‘cycle.’
The period was a unit of organization which could stand on its own, just as we understand the modern sentence today. Those of us who call the punctuation at the end of a sentence a period have taken the name from the chunk of language that the punctuation mark concludes.
The period was composed of smaller units called cola (singular, colon). The word colon comes from the Greek κῶλον ‘limb’.3 A colon was a unit of prose which could be pronounced in a single breath, after which there would be a pause.
As we can tell from the etymology of the word, a colon was not a complete thought: it was considered as a part of the whole, like a limb of a body, where the body corresponds to the period.
We might think of the colon as equivalent to the modern concept of a clause, although the comparison is not exact. The clause is fundamentally a grammatical concept: a unit consisting of a subject and a predicate. The colon, on the other hand, is defined in terms of when a speaker would have to take a breath.
Like the period, the colon has given its name to a unit of punctuation, although the modern function of the punctuation mark we call a colon (:) is not simply to mark the end of a colon in the rhetorical sense.
Cola, too, were divided into smaller units. Each of these was called a comma (plural, commata), from the Greek κόμμα kómma ‘piece.’ The comma corresponds roughly to the modern concept of a phrase, that is, a group of words that act together as a single unit.
For example, a big matzoh ball is a noun phrase, really very nice and good is an adjective phrase, and put an end to maritime oil spills is a verb phrase.
Phrases in the modern sense can nest: maritime oil spills is a noun phrase which exists inside another noun phrase, an end to maritime oil spills, which is itself a part of the larger verb phrase put an end to maritime oil spills.
The classical comma doesn’t worry about nesting in this way, since it has its origin in oratory: for an orator, the relevant thing is not the grammatical structure of the phrase, which is indeed hierarchical, but rather its rhythmic structure.
In other words: where are the places where a small pause could be inserted to let the speaker take a breath? These are the comma divisions. And, as you’ve no doubt noticed by now, the ancient rhetorical comma has given a name to a modern punctuation mark (,), which is used in some places to mark pauses in speech, and in others to mark particular grammatical relationships.
Why no one knows how to use commas today
I’ll pause here to note that modern punctuation has a dual heritage: on the one hand, it descends from marks used to indicate the rhythm of the sentence, and therefore what kind of breath a speaker should take at various points; on the other, it seeks to express grammatical relationships.
Modern punctuation standards are a sometimes uncomfortable blend of both systems. For example, in the sentence, I had a good time and she did too, the use of punctuation to mark grammatical relationships forbids a comma (that is, the punctuation mark) between these two conjoined clauses, but the use of punctuation to signal rhythm and breathing points seems to tempt us to add a comma after time, which many writers do indeed do.
As we’ve seen, the way the ancients divided prose had much more to do with rhythm than with grammatical relationships. And this was not just a characteristic of oratory, which would be primarily delivered verbally, but also of writing. There was no great difference between the two modes of delivering prose in how the ancients thought of them: a letter or a work of historical narrative was divided into periods, cola, and commata just as a speech was.
For our purposes, these divisions were of special importance because of another feature of ancient rhetoric: writers and speakers paid special attention to the rhythm at the end of each period, and often of each colon as well. Each period was thought of as akin to a flight, rising at the beginning and falling at the end, and, as I’m sure you’ll all agree, the most important part of any flight is an orderly and controlled landing.
So in classical Greek and Roman practice, it was considered a matter of good style to end a period in one of only a few rhythmic patterns. These patterns were called by the Romans clausulae, but I’ll use the later term cadence to describe them, because our primary interest is English, where the term used for these period-ending patterns is cadence rather than clausula.
And, as it turns out, these cadences were crucial for understanding not just the style of the ancient Greeks and Romans, but of many of the great English writers as well, even if these writers were wholly unconscious of this system of rhythmic cadences.
“Shave and a haircut” is a classical cadence
But before we talk about individual English writers, we need to understand which cadences were considered appropriate for ending a period, and which were not.
The first thing to know is that the ancient Greeks and Romans defined rhythmic patterns in prose in the same way they did in poetry: in terms of long and short syllables.
These patterns of long and short syllables grew out of the nature of the Ancient Greek and Latin languages, which distinguished between long and short vowels (much like Old and Middle English did).
For someone who hasn’t studied Ancient Greek or Latin, these patterns of long and short syllables can be obscure. Fortunately, we can avoid talking about them because the Romans themselves eventually switched to a different way of determining rhythm.
That is, they started to use the stress pattern of the words to determine the rhythm, just like we do in English today — by the way, if you’re not familiar with the concept of stress and rhythm in English, you can brush up on that here. It’ll come in handy as we proceed.
Let’s look at some examples of what the late Romans — and their medieval descendants4 — considered acceptable cadences.
There were three main cadences considered acceptable, and each had a name: there was the cursus planus ‘level cadence,’ the cursus tardus ‘slow cadence,’ and the cursus velox ‘fast cadence.’ Each of these referred to a different arrangement of stressed and unstressed syllables, which I’ll notate with the symbols / for stressed and x for unstressed, respectively.5
Cursus planus: / x x / x, e.g. Latin vóces testántur ‘voices bear witness’
Cursus tardus: / x x / x x, e.g. Latin méa curátio ‘my guardianship’
Cursus velox: / x x / x / x, e.g. Latin gáudia pèrveníre ‘to arrive (at) the joys’6
If you don’t know Latin, listen to the rhythm of the stress as I read them here — note that I’m reading them in a reconstructed classical pronunciation, which is not how the medieval church would have read them. But what matters here is not the precise details of the vowels and the consonants, but the pattern of the stress accent, which is the same in all pronunciation schemes:
Along with these three basic cadences, there were some acceptable variations, but we’ll discuss these later on.
The use of these cadences in writing rhythmic Latin prose continued through the latter half of the Middle Ages, until the rediscovery of classical learning in the Renaissance — including the practice of basing rhythm on long and short syllables rather than the stress accent — brought an end to the practice.
But, for a long time, scholars have wondered whether the practice truly died out everywhere. In particular, many writers have wondered whether the system survived in the writing of English prose, even if the influence was unconscious. After all, English is a language which, like Medieval Latin, uses a stress-based system of accent.
One of the first places they looked for this influence is in the Book of Common Prayer, first published in 1549 as a product of the English Reformation. New versions followed in 1552, 1559, and 1604, amid (and as a result of) the religious and political turmoil of that period in English history, before it finally stabilized in the 1662 edition.
The language of the Book of Common Prayer is worth an issue unto itself, because it’s been one of the most influential works on English prose style, with a level of influence similar to the Authorized (i.e., King James) Version of the Bible.
But for now, I’ll just say that a set of prayers7 from the Book of Common Prayer was used as the data set for a 1912 paper by John Shelly, which examined the influence of the Latin cadences on English prose. These prayers had been translated into English from Latin by Thomas Cranmer, although Cranmer made many adaptations to fit his own ideas about the Church.
Shelly’s contention was that Cranmer was replicating in his translations the cadences that he found in the original Latin — Latin that worked according to the three cursus patterns of the Middle Ages.
Corresponding to the Latin cursus planus vóces testántur, we have the English hélp and defénd us. The Latin cursus tardus méa curátio is echoed by the English góverned and sánctified, and the cursus velox of Latin gáudia pèrveníre corresponds to the English púnished for òur offénces.
In Shelly’s analysis, around 50% of the English prayers in the data set ended in one of these three cadences. Other scholars, including Albert C. Clark and Oliver Elton, expanded the investigation to encompass English prose more generally, and extended the analysis to include the following variations of the three classical cadences:
variant cursus planus: / x x x / x, e.g. Latin dóna sentiámus ‘may we experience the gifts,’ English wrítten for our léarning
variant cursus tardus: / x x x / x x, e.g. Latin (vir)tútis operátio ‘the working of power,’ English dángers and advérsities
trispondaic (extended cursus velox): / x x / x / x / x, e.g. Latin (er)rántium córda rèsipíscant ‘may the hearts of the erring repent,’ English páss to our jóyful rèsurréction.
These scholars found that these classical cadences do indeed appear — and in great number — in some of the most classically influenced English authors: Edward Gibbon (1737–1794), the author of The Decline and Fall of the Roman Empire, comes to mind as an especially good example of such an author, as do Thomas Browne (1605–1682), author of Urn Burial, and Thomas Babington Macaulay (1800–1859), author of The History of England.8
Classical cadences in the hands of a master
Let’s look now at how Gibbon, to take the best-known of these authors as an example, manages his cadences to see how many of them are in accordance with the classical forms.
I’ve broken up the first paragraph of The Decline and Fall of the Roman Empire according to cola, and marked the stress pattern of the cadences, as I understand them:9
Keep reading with a 7-day free trial
Subscribe to Dead Language Society to keep reading this post and get 7 days of free access to the full post archives.