Oxford University Press's
Academic Insights for the Thinking World

Bolder than a boulder and other stumps and stones of English orthography

One good thing about English spelling is that, when you look for some oddity in it, you don’t have to search long. So why do we have the letter u in boulder (and of course in Boulder, the name of a town in Colorado)?  If my information is reliable, Boulder was called after Boulder Creek. A boulder near a small stream won’t surprise anyone, but the letter u in the word and the place name may, as journalists like to say, raise some eyebrows. Bolder (the comparative degree of bold), older, colder, folder, and holder do without u, but shoulder, unexpectedly, sides with boulder. American spelling has mold in all its meanings and the verb molder, while the British norm requires ou before l. What is going on here?

Perhaps the huge size of an average boulder contributed to the preservation of an extra letter in its name?
Perhaps the huge size of an average boulder contributed to the preservation of an extra letter in its name?

The easiest case is old, cold, gold. The regular predecessor of –old in such words was either –ald or a short diphthong. Before ld, the vowel a (short, that is, ă) was lengthened in Old English, and in the group āl the vowel ā, obedient to the general law, became ō and later, by the Great Vowel Shift, yielded the diphthong [ou] or /ou/, if you prefer phonemic transcription. In such words, –old– is never spelled ould. But boulder and shoulder once sounded as bulder and sculder. Bulder turned up only in Middle English and is probably a word of Scandinavian origin, while sculder goes back to Germanic antiquity, as evidenced by German Schulter and Dutch schouder. To understand what happened here, it is useful to remember that there are (and have always been!) two l’s in English, called dark (or hard) l and light l.

Phonetic metaphors (light l, dark l, thick l—the latter sound exists in Norwegian dialects,—soft l, hard l—those varieties are distinguished, for example, in Russian) say nothing to the uninitiated, for what is, for example, thick l? How “thick” is it? But these terms do reflect the impression such sounds make on some people. Rather than being lost in terminological details, I’ll refer to the experience of those who can compare British and American speakers. The difference can be heard a mile off: British l is “light” (and so is German and Swedish l, as opposed to Norwegian). American l is “dark,” and even those Americans who speak otherwise impeccable German have great trouble with their l. The curious thing is that all of us hear the slightest trace of accent in others but not in our own speech—an almost perfect analog of the mote and the beam paradox, but transferred from the eye to the ear.

The parable of the mote and the beam is fully applicable to our view of accents.
The parable of the mote and the beam is fully applicable to our view of accents.

The l called light shares some features with the front vowel i, while the closest neighbor of dark l is back (velar) u. Dark l is what we may hear if we pronounce it separately with strongly protruded lips. It will make the impression of the group ul, though u is, naturally, less audible than l. One can perhaps transcribe light l as [il], and dark l as [ul]. The history of English shows that [ul] is much more than an elegant (or whimsical) transcription by a professional linguist, because speakers did insert the vowel u before dark l in their speech. But for the influence of l, Modern English fall would have retained the vowel indistinguishable from the one in Modern German fallen. However, in English, [a] before l became [a-u], and this diphthong yielded the sound we now have. Such is the origin of the modern pronunciation of all, ball, call, hall, stall, and so forth. Exceptions are few: compare fallow, hallow, sallow, shallow, tallow, to say nothing of shall. Their history deserves special discussion.

For the same reason, shoulder has u in modern spelling. As we have seen, boulder and shoulder were once bulder and sculder. In principle, such words might have preserved their vowel intact, for they already had u before dark l! Yet for some reason, this did not happen, and ul also yielded a diphthong; hence boulder, shoulder, and others like them, in which ou, unlike au in ball, call, fall continued into Modern English unchanged. The consonant after l did not participate in this game: though d is voiced, while t is voiceless, coulter “part of the plowshare” and shoulder ended up with the same diphthong. This also holds for words like all: in falter, halter, Walter, voiceless t follows l, but it did not affect the main change. This fact is worth mentioning because in words like cold and old the ancient lengthening depended not on l but on the group -ld. Since in salt, a preceded –lt, salt and sold do not “rhyme.”

A short postscript to what has been said above may add a finishing touch to the phonetic history of the group represented by shoulder. Dutch has shouder, and this is not an unexpected form. Dark l, having performed its dark deed, disappeared, as behooves a criminal in a thriller. In English, such forms are rare, nearly nonexistent. However, the noun solder “the fuse used for uniting metal surface” has been recorded with three pronunciations:  with a diphthong, as in sold; with a monophthong, as in saw; and with a short vowel without l, thus rhyming with fodder. Historical phonetics is full of such tricks: a sound causes a change and goes away, but the harm stays behind. If I may refer to my previous post, the model for such changes is Lewis Carroll’s Cheshire cat: the beast vanishes, but the grin remains. Later, it too can be lost, but this is another chapter in our story.

The phonetic leaps described above may be partly unexplained but, on the whole, they usually make sense. It should only be remembered that, though people “generate” sound changes, they have no control over them. By contrast, spelling depends on our will. For better or worse, the letter group –old– regularly designates the pronunciation, as in bold, cold, fold, and old. Consequently, inserting u after o is unnecessary. Eliminating it in mold and a few others (as in American spelling) was a reasonable measure, but shoulder and coulter are spelled alike everywhere. I don’t know why mold ~ molder, rather than shoulder and boulder, fell victim to the simplification. Perhaps it was decided that bolder and boulder should look different on paper.

Orthography is based on several principles, the phonetic one being only the most obvious among them. Morphology, that is, grammar, not pronunciation, makes us spell lacked / lagged and lacks / lags with the same endings pairwise (whether it is wise to do so need not bother us at the moment). Still another principle can be called iconic. Thus, it can be argued that bolder and boulder should be distinguished in writing, even though they are homophones. History too is often taken into consideration. Most likely, names will probably withstand any spelling reform. We all know that no English place or family name should be pronounced according to its visual image. If you are not sure how to pronounce Strachan or Willamette, just find out (Strachan rhymes with dawn, and for Willamette, should you put the stress on the first syllable, you may hear the gentle reprimand everybody knows in Oregon: “Willamette, dam’ it”). If boulder is ever changed to bolder, the place name should probably remain Boulder.

All this is very sad, but for the time being enjoy bolder, boulder, colder, coulter, shoulder, should, molder  ~ moulder, solder, soldier, and poultry. If, however, you come away wounded, apply a poultice to the place that aches the most.

If after this exposure to the vagaries of English spelling, you need a poultice, help yourself.
If after this exposure to the vagaries of English spelling, you need a poultice, help yourself.

Image credits: (1) Rock Dolerite by Ron Porter, Public Domain via Pixabay. (2) Zorab mote and beam by Minus (Minas) M. Zorab, 1880, Public Domain via Wikimedia Commons. (3) “Poultice easily illustrated” by Snuvnia, CC BY-SA 4.0 via Wikimedia Commons. Featured image credit: Boulder, Colorado by marcelopmat, Public Domain via Pixabay.

Recent Comments

  1. Rudy Troike

    Enjoyable as usual. I always learn something. I would suggest that the convoluted explanation of the two [l]’s in English can be simplified by using the word “little” as an illustration. It is certainly impossible to reverse the two allophones.
    The connection of the [l]’s with the other resonants [y] and [w] is worth noting. In language acquisition, children usually produce /y/ for the first (light) [l] and /w/ for the dark [l]. The similarities with postvocalic [w] are responsible for the substitution of [w] for velar [l] in some varieties of English, and for the vocalization of the [l] in words like “would” and “should”. For some British speakers, “warden’ and “walden” are homophonous. [And why do we have an “ou” here?]

  2. Masha Bell

    Spelling, as u say, depends on our will. Sadly, when it stops being based on phonetic principles, when it no longer adheres to the alphabetic principle of using letters to represent speech sounds in a regular manner, as in English, it incurs enormous costs. – It makes learning to read and write excruciatingly slow and tedious, is harder and more expensive to teach and results in high rates of functional illiteracy. The latter entail entails many other personal, social and fiscal costs. (The EnglishSpellingProblems blog provides fuller explanations of them.)
    A messy orthography like the English one does offer great scope for etymological speculation, but I would prefer more people to be literate and better educated.

  3. Paul Nance

    There is also the spelling bowlder, which raises the question of whether the same process has affected the spelling of bowl (1) “round vessel.” OED suggests that bowl (2) “sphere, ball” is pronounced as it is because of a graphic confusion with bowl (1).

Leave a Comment

Your email address will not be published. Required fields are marked *