The ancient poetic law that explains basically everything

From Sanskrit grammar to nursery rhymes

Rye field in the Jura (1863), Frank Buchser

There’s a children’s song that I remember hearing ad nauseam when I was a kid: it was called “Oats, Peas, Beans, and Barley Grow.” Later in life, when reminiscing about the song, I learned that not many people had heard it.¹

Perhaps they didn’t have a copy of Raffi’s 1980 smash hit record Baby Beluga, which featured a version of the tune as the third track on the album.

Little did I know at the time, but my young ears were in the presence of what linguist Calvert Watkins called “a masterpiece of the Indo-European poet’s formulaic verbal art.”²

Now, lyrically speaking, “Oats, Peas, Beans, and Barley Grow” doesn’t have too much going on. About 40% of the song’s lyrics are repetitions of the title. But, embedded in those six words is an illustration of many of the principles of the poetic tradition that extends from Ireland to India, and spans the centuries to link Homer (8th century BC) to Hávamal (AD 1000) and beyond.

These poetic principles include alliteration (beans and barley) and ring composition (oats… grow),³ but they also feature a technique that I find even more interesting.

This technique shows up in ancient epic poems and children’s folk music, but even more so in the language of everyday life. In fact, once you start to see it, you see it everywhere.

It’s called Behaghel’s Law (of Increasing Members).⁴ And it’s the principle at work in the fact that, in “Oats, Peas, Beans, and Barley Grow”, barley comes last, after oats, peas, and beans.

This is because the word barley is more structurally complex than the others: it has two syllables: bar-ley, while oats, peas, and beans each have only one.

This Law of Increasing Members was formulated by the German linguist Otto Behaghel as follows:

> Given two phrases, when possible, the shorter precedes the longer.

Behaghel’s Law isn’t the kind of immutable natural law that the very existence of the universe depends on. (Cookies and cream. See, there, I broke it and nothing bad happened!) Instead, it’s a tendency that explains many disparate phenomena extended over time, space, and subject matter.

Let’s take a look at some of them!

---

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 also be able to join our ongoing Beowulf Book Club and watch our discussion of the first 915 lines (part 1, part 2) right away.

---

One, two, and Homer makes three

Behaghel’s Law is phrased in a very particular way: “the shorter precedes the longer.”

Shorter how? The wording is general by design, and allows us to capture different ways in which we can measure “short” and “long.”

For example, a phrase could be long or short in terms of how many words are in the phrase: so a phrase consisting of the single word barley isn’t as long as the phrase the barley that my uncle Jack planted last year.⁵

But you could also count syllables, rather than words: counting in this way, oats isn’t as long a phrase as barley, despite the fact that both are composed of a single word. Behaghel’s Law is flexible enough to handle both of these definitions of “short” and “long.”

When we apply Behaghel’s Law to the number of words in a phrase, we find a phenomenon that will be familiar to readers of Homer’s Iliad and Odyssey. Consider the following line, which is of a type that recurs over and over again in Homer:

Ἄργός τε Σπάρτη τε καὶ [εὐρυάγυια Μυκήνη] (brackets show the phrase boundaries)
Argós te Spártē te kaì [euryáguia Mycḗnē]
‘Argos and Sparta and [broad-streeted Mycenae]’ (Iliad 4.52)

In this line, we see three names in succession, the first two bare of any adornment and the third supplied with an epithet or other description.⁶

It occurs throughout the writings of Homer, usually in the catalogue sections, namely, the long lists of people and nations that pepper the epics. Some further examples:

Ἶλός τ’ Ἀσσάρακός τε καὶ [ἀντίθεος Γανυμήδης],
Îlós t’Assárakós te kaì [antítheos Ganymḗdēs],
‘Ilus and Assaracus and [godlike Ganymede]’ (Iliad 20.232)

Τυρώ τ’ Ἀλκμήνη τε [ἐυστέφανός τε Μυκήνη]
Tyrṓ t’Alkmḗnē te [eustephanós te Mykḗnē],
‘Tyro and Alcmene and [well-crowned Mycenae]’ (Odyssey 2.120)

The use of epithets also provides an easy way of making a particular line conform to the requirements of the metre, as all the conventional name + epithet combinations are guaranteed to fit the metre.⁷

The classicist M. L. West gave this construction a name: the augmented triad. It appears to be very old, and not just because it’s found in Homer. It’s also found in Vedic poetry (that is, composed in an early stage of the Sanskrit language) as well as Celtic and Germanic verse. Some examples, courtesy of M. L. West’s book Indo-European Poetry and Myth (pp 118–119).

Vedic Sanskrit:

Tváṣṭā, Savitā́, [suyáma Sárasvatī]
‘Tvashtr, Savitr, and [easily-guided Sarasvati]’ (Rigveda 9.81.4)

Old Irish (Celtic):

trí meic Nóe nair cech neirt:
Sem, Cam, [Iafet aurdairc.]
‘Three sons of Noah, of every (kind of) strength:
Shem, Ham, [Japheth the glorious]’ (Lebor Gabála Érenn, 189–190)

Old Norse (Germanic):

Vara sandr né sær né [svalar unnir].
‘There was not sand nor sea nor [the cool waves].’ (Vǫluspá 3)

The presence of the augmented triad in not only Greek, but in Celtic, Germanic, and Vedic Sanskrit (see what I did there?) means that it was likely inherited from the common ancestor of all four.

Just as Greek and Sanskrit, not to mention the Celtic and Germanic families, all descend from the ancestral Proto-Indo-European language, so too do the poetic traditions of these cultures descend from a Proto-Indo-European form of poetry, as it was practised by the speakers of our ancestral language, some 4000–6000 years ago.

---

The thing is, Hālga wasn’t even good…

Given that the augmented triad is so common in Homer, we wouldn’t be shocked to see it in Modern English verse, much of which was written by poets weaned on Homer and his imitators.

But we also see it — albeit rarely — in Old English poetry, where it appears to be an Indo-European inheritance.

Here are two examples from Beowulf:

Heorogār ond Hrōðgar ond [Hālga til].
‘Heorogar and Hrothgar and [Halga the good]’ (Beowulf 61)

Herebeald ond Hæðcyn oððe Hyġelāc mīn
‘Herebeald and Hæthcyn or [my Hygelac]’ (Beowulf 2434)

Other than these two examples, the augmented triad occurs in one other place in Old English poetry, in the poem Genesis:⁸

Aner and Manre, [Escol þriddan].
‘Aner, Mamre, and [Eshcol third],’ (Genesis 2045)

Although the use of the augmented triad is relatively rare in Old English poetry, it persists into Middle English poetry, and, in fact, flourishes there. Here are some examples:

Sadoc and Samiel and [Symeon þene alde]
‘Sadoc and Samuel and [Symeon the old]’ (Laȝamons Brut 9150)

Afirike and Arraby and [Egipt the noble].
‘Africa and Arabia and Egypt the noble’ (Parlement of the Thre Ages 418)

Launcelot and Lyonel and [Lucan þe gode].
‘Lancelot and Lionel and [Lucan the good]’ (Sir Gawain and the Green Knight 553)

Are these Middle English triads true survivals of the old Indo-European technique in English? Or are they literary imitations of classical epic?

J. P. Oakden — famous, incidentally, as a scholar of Staffordshire place names — considered this question in a 1933 paper, which, incidentally, is the source of the selection of triads I quoted above.

The Middle English period (1100–1450) was a time when the influence of French on English was at its strongest. Given that the French language descends from Latin, and Latin writers were, in Roman times at least, intimately familiar with the Greek epics, it stands to reason that the augmented triads of Middle English literature could have come indirectly from Greek.

But what Oakden found was that these Middle English triads were probably genuine examples of the survival of the old Indo-European augmented triad. He reasoned that, if French were the vehicle by which augmented triads were carried into Middle English, we’d surely expect to see some in (Old) French literature. But we don’t seem to see any!

So the augmented triad as a live poetic device in English must have survived, passed on from poet to poet, until the late Middle Ages. Beyond that, however, it’s hard to say what became of it, at least in the world of poetry.

As I mentioned earlier, in Modern English poetry, it becomes harder to distinguish between authentic survivals of the augmented triad formula and poetic imports from Homer, filtered through his admirers.

But poetry isn’t the only place we can find applications of Behaghel’s Law. It applies in prose too: witness the popularity of the ascending tricolon, in which three parallel structures are used, increasing in some way from beginning to end.

For example, Caesar’s famous phrase I came, I saw, I conquered is certified Behaghel-friendly, in English at least. This device of the ascending tricolon is much beloved both of the masters of classical rhetoric, and, as you may have noticed, ChatGPT.

But when we broaden our horizons even further, away from the elevated language of poetry and rhetoric, we see that Behaghel’s Law is alive and well even in the very earthy English of everyday life that we speak today.

---

Why it’s not “gentlemen and ladies”

In 1975, the linguists William Cooper and Haj Ross opened a paper⁹ with the question:

Why do we say kit and caboodle but not ❌caboodle and kit? (here, the preceding ❌ emoji indicates that the phrase is ill-formed when given in that order.)

You might answer, “It’s simply a fixed phrase; we shouldn’t expect any logic in its ordering.”

But kit and caboodle is far from the only fixed-order phrase in English. Cooper and Ross assembled a long list of them, including: bigger and better, fore and aft, ladies and gentlemen, but crucially not ❌better and bigger, ❌aft and fore, ❌gentlemen and ladies.

What Cooper and Ross noticed is that, across the English language, these fixed-order phrases, which they called freezes, do seem to have a logic to them. Now, the logic governing these freezes is complicated, which is why the paper is 48 pages long.

But one crucial element of the logic, as you might guess from the ordering of kit and caboodle and ladies and gentlemen, is Behaghel’s Law: the word with the greater number of syllables comes second.¹⁰

With this one rule, suddenly the characteristic shape of many English idioms makes sense:

• vim and vigour

• hot and heavy

• wild and woolly

• hale and hearty

• rough and ready

• bread and butter

It’s probably no accident that these are also alliterative, and, in fact, would be perfectly well-formed half-lines in an Old English-style poem.¹¹

Both Behaghel’s Law and the use of alliteration in poetry likely date back to Proto-Indo-European times. As pointed out by M. L. West in his Indo-European Poetry and Myth, alliteration is found not only in Germanic poetry, but also in Irish poetry and early Latin poetry (although it was later discarded after Latin poets started writing according to Greek models), not to mention in Vedic poetry and other early poetic traditions.

Other freezes in English are non-alliterative but still adhere to Behaghel’s Law:

• rough and tumble

• bread and water

• free and easy

• lock, stock, and barrel

In the hands of Guy Ritchie, the last of these became a true augmented triad: Lock, Stock, and Two Smoking Barrels.

Now, Behaghel’s Law is not the only rule governing English freezes. Meaning also seems to play a role in many cases. For example, closer things tend to precede farther things, whether in time or space: now and then, here and there, in and out, sooner or later. Positive things tend to precede negative things as well: friend or foe, all or none, confirm or deny. So too does the male tend to precede the female, the living to precede the dead, and the singular to precede the plural.

These meaning-based principles sometimes line up with Behaghel’s Law: for example, now or never increases in syllables while keeping the positive before the negative, as does plus or minus. But sometimes they contradict Behaghel’s Law, and we get violations of either Behaghel’s Law or the meaning-based principle.

An example of a freeze where the semantic principle wins out is the living (2 syllables) and the dead (1 syllable), as opposed to the Behaghel-friendly ❌ the dead and the living.

So too, alongside the Behaghel-friendly ladies and gentlemen, we also have the Behaghel-unfriendly husband and wife, which patterns with other male-before-female freezes such as men and women, boys and girls, and king and queen.

There are also other considerations which Cooper and Ross (a Behaghel-unfriendly ordering!) identified, including one which involves the vowels and consonants used in each of the elements of the freeze. This allows them to explain things like pitter-patter and hickory dickory dock, but to go fully into that story will take us too far from Behaghel’s Law.

The point of showing you all these examples and counterexamples is to show how Behaghel’s Law is one of those laws that gets broken all the time when there’s reason for it. In other words, it’s one principle among many which exist in competition with one another.

Cooper and Ross spend the latter half of their paper connecting the various principles they identified as governing freezes with broader psychological phenomena. In the end, they conclude by proposing a principle which they call the “Me First Principle”, which tries to unite many of the meaning-based constraints into one:

“First conjuncts refer to those factors which describe the prototypical speaker.” (Cooper and Ross 1975: 67)

In other words, Cooper and Ross set up an abstract concept “Me” which represents “Here, Now, Adult, Male, Positive, Singular, Living, Friendly, Solid, Agentive, Powerful, At Home, Patriotic, among other things.” (67)

The more like “Me” something is, the more likely it is to come first in a freeze. Note that the actual person speaking need not be all of these things — it’s rather as if the language itself aims to put these things first.

So much for the exceptions to Behaghel’s Law. But what about Behaghel’s Law itself? Where does it come from? We’ve traced it back in the history of Indo-European poetry, but does it have an existence of its own, independent of its use in poetry?

As it turns out, yes, it does. But to see how, we’ll need to go back to the very birthplace of linguistics itself, ancient India.