Reading time: 10 – 15 minutes
Introduction: Nouns, Verbs and Variable Vowels
In present-day English, the plural of mouse is usually mice, and one man plus another equals two men. While most English nouns are made plural simply by adding -s, making one cat into multiple cats, there is a sizeable minority that become plural through the process that we see with mouse and man: they change their vowel.
These vowel-changing nouns are very old; they go back to the Old English language and further back still. Their systems of alternation have endured, although the specific vowels involved may have changed. For instance, Old English fōt and fēt are recognisably the ancestors of modern foot and feet, although they were not pronounced in the same way.
This group of nouns is something of an inexplicable annoyance for second-language learners of English; approaching the language as it is, you can’t really predict which nouns are in it, and also how their vowels will change. It is at least a closed set, since new members are not being added to it (although really there’s nothing stopping us). This is the reason behind that meme that has done the online rounds in various incarnations since an archaic age of the Internet, the one that wonders why goose becomes geese, but moose does not become meese. While goose has been part of the lexicon since before English was English, moose entered English (borrowed from an Algonquian language in North America) far too late to join the vowel-alternating gang.
There is another category of word in English that does all this too — a category that, for ancient reasons, has a small, closed set of its words that convey grammatical information by changing their vowels. That category is the English verb.
The regular way to put an English verb into the past tense is to add -ed (or sometimes just -t) to the end, turning I play into I played. Yet, as any learner will tell you, English is more complicated than this one rule. While adding -ed to verbs is extremely common and considered typical and regular, some verbs add nothing, but rather use a two- or three-way alternation of vowels.
Sing, steal and write are three such ‘strong’ verbs. Rather than singed, stealed and writed, they become sang, stole and wrote in the past tense, and also sung, stolen and written as past participles. As with nouns, these patterns are very old, although the particular vowels that instantiate them today may not be. In Old English, present-tense iċ rīde on horse (‘I ride on a horse’) would be iċ rād on horse in the past tense, which look and sounded different to today’s words, but still correspond to modern ride – rode.
The general trend over the history of English has been the decline of strong verbs and their vowel changes. Many ‘weak’ verbs today (i.e., those that add -ed) used to be strong. The verb help had the pattern of helpan – healp/hulpon – holpen in Old English, but has been regularised since then into help – helped, although you may still come across holpen. Strong verbs are slowly losing out, although I swear I recently heard someone say “I’ve always just wung it” while talking about exams.
A Historical Marvel
Now, here’s the thing that just amazes me.
We have two categories of word (nouns and verbs) that both make use of vowel changes, with patterns that have prehistoric origins, and have been diminishing over time. You’d think therefore that there would be one historical source behind all this vowel swapping?
It is the very matter of its origin that just boggles my mind — or rather, its origins. While many verbs and nouns alike change their vowels, there is more than one reason for why English does this. The vowel changes of nouns developed separately from the systems of strong verbs.
In fact, despite the similarities, the two processes don’t seem to have had anything to do with each other!
Here’s why and how:
The Nouns
As mentioned, the vowel changes for nouns have prehistoric origins. They go back to a time before the written record, when what would become English was much closer (geographically and linguistically) to the rest of the Germanic language family.
The distinct vowel for plurals like mice, men, geese and teeth is due to the important process of umlaut. This word comes from German, and it’s widely used to refer to the dots above Ä, Ö and Ü, but its precise meaning is the sound change that created the sounds that those German letters represent. It was umlaut (specifically ‘i-umlaut’ or ‘i-mutation’) that was responsible for noun patterns like mouse – mice, and it can be simply defined as follows:
A back vowel becomes a front vowel when the following syllable contains the vowel /i/ (or similar sounds)
This needs unpacking. A back vowel is one pronounced with the tongue positioned back in the mouth, like the vowel in car. A front vowel is the opposite, pronounced with a forward tongue, like the vowel in sheep. Umlaut turned a back vowel into its front equivalent.
You can see the different vowels on the IPA vowel chart below (click here for the sounds), which respresents the possible positions of the tongue. A back vowel /u/ would become the front vowel /y/ through umlaut, and back /o/ would end up as front /e/.
Umlaut did this in English when the following syllable contains the vowel /i/ (like in sheep) or similar sound like /j/ (as in yes). Why? Because /i/ is a front vowel, as you can see at the top-left corner of the chart. The presence of the front vowel in the following syllable causes the previous vowel to become more like it, like a kind of harmonization between the vowels in a word.
Umlaut had a far-reaching effect on the Germanic languages, and there’s so much that can be said about its importance for English alone — it’s the reason why in English we tell a tale, gild with gold and bleed blood.
Why it specifically matters here is because some1 nouns at this pre-Old English stage formed the plural (in the nominative case) by means of the ending *-iz. The plural of *mann (‘man’) was *manniz. The plural of *gans (‘goose’) was *gansiz. That plural ending was all that was necessary for umlaut to occur.
Thus the newly fronted vowel of the plural form split off from the vowel of the singular, which remained a back vowel. No longer were there singular and plural pairs like *mann – *manniz, but rather *mann – *menniz. This difference was in turn ‘grammaticalized’ — it became a part of the grammar, a means of distinguishing singular from plural.
This meant that the ending *-iz was redundant, and could disappear along with so many of the old Germanic endings. When Old English first enters the written record, distinctions like mūs – mȳs are up and running. Having gone through the later Great Vowel Shift, these then gave us our modern words mouse and mice.
The Verbs
The system for nouns may be very old, but it’s a linguistic toddler compared to the vowel changes of verbs. These are really old. We’re talking multiple millennia.
These alternations go all the way to Proto-Indo-European itself. They have their origins in the important and ancient process of ablaut, a system of vowel changes that can still be seen across the Indo-European family. Sanskrit preserves the system well, but vestiges and relics of it are also there in Latin, Ancient Greek and indeed Old English. Ablaut is all about swapping vowels.
A given root originally could have three ablaut ‘grades’:
- The e-grade
– when the root of the word contains an *e vowel - The o-grade
– when the root of the word contains an *o vowel - The zero grade
– when the root of the word contains neither *e nor *o, and may have no vowel at all
Proto-Indo-European then used these three grades to construct its verbs. For instance, the present tense was typically built on the e-grade, while past participles (like English given and eaten) were formed by adding a dedicated suffix to the zero grade.
It’s this three-way distinction that English has inherited and maintains to this day in verbs like sing – sang – sung. Consequently, it’s an element of English grammar that gets linguists rather excited. True, the specific vowels that strong verbs use today are different to the vowels and ablaut grades of the ancient ancestor, but the system is still there and the ways in which the vowels now differ are regular and predictable. In the case of sing:
- Present-tense sing continues the e-grade
– with the regular change of *e to Germanic *i. - Past-tense sang continues the o-grade
– with the regular change of *o to Germanic *a. - The past participle sung continues the zero grade
– with the usual *u added in the absence of a vowel in the original root.
Let’s take write as another example. Its pre-Germanic story is unclear, but since it’s a strong verb, we can assume that it’s very old.
We can reconstruct its pre-Germanic ancestor as the root *wreyd-. Specifically, *wreyd- is the e-grade, while *wroyd- is the o-grade and *wrid- is the zero grade, in which *y steps in to be the vowel *i in the absence of *e and *o. This is what happens to this three-way distinction over time:
| Pre-Germanic | Proto-Germanic | Old English | Modern English |
|---|---|---|---|
| *wreyd- (e-grade) | *wrīt- (with a long vowel) | wrītan (the infinitive form) | write (after the Great Vowel Shift) |
| *wroyd- (o-grade) | *wrait- (with the regular change of *o to *a) | wrāt (with the regular change of *ai to ā) | wrote (after the Great Vowel Shift) |
| *wrid- (zero grade) | *writ- (with a short vowel) | writen (the past participle) | written |
This also works nicely with the verb ride, which has the similar-looking reconstructed ancestor *Hreydʰ-.
| Pre-Germanic | Proto-Germanic | Old English | Modern English |
|---|---|---|---|
| *Hreydʰ- (e-grade) | *rīd- (with a long vowel) | rīdan (the infinitive form) | ride (after the Great Vowel Shift) |
| *Hroydʰ- (o-grade) | *raid- (with the regular change of *o to *a) | rād (with the regular change of *ai to ā) | rode (after the Great Vowel Shift) |
| *Hridʰ- (zero grade) | *rid- (with a short vowel) | riden (the past participle) | ridden |
Some strong verbs in modern English only have a two-way vowel distinction, but often this is the result of a later merger, and we can see in the written record how they too used to have three. For instance, steal – stole – stolen works with two different vowels today, but these go back to a three-way pattern of stelan – stæl/stǣlon – stolen in Old English.
This is just a casual introduction to a lot of English linguistic history, but I hope it’s clear enough. I also dare to hope that it might inspire your own linguistic thinking! Umlaut and ablaut have been very important processes for the development of English as it stands today, and the patterns and connections are all out there, at the tip of your tongue and ready to be found.
If I have one takeaway point, it’s that it’s just fascinating how two systems of vowel alternations can arise seemingly independently of one another in a language. Of course, it’s possible and plausible that the existence of the older verbal system meant that the later vowel alternations for nouns weren’t an entirely new and alien phenomenon, but nevertheless, the motivations and mechanics of the two were quite distinct. That’s pretty cool.
END.
- Bonus etymological nerdery: the nouns that used the plural ending *-iz weren’t random — they are actually predictable on the basis of other Indo-European languages like Latin and Ancient Greek. They were ‘athematic’ nouns, also known as nouns with consonant stems, which originally had no linking vowel between their root and the added grammatical endings. The closer proximity between root and endings therefore made this class of nouns particularly susceptible to umlaut in Germanic. We can compare the vowel-shifting English nouns tooth, mouse, goose and foot with their Latin cognates dēns, mūs, ānser and pēs; these all belong to the third declension, where Latin keeps its consonant-stem nouns, and they use the similar plural ending -ēs in the nominative. Honestly, I just love this fact so much. It’s hardly the most accessible part of this whole story of course, but it’s thrilling for me. Using Latin to predict English? It’s the stuff etymological dreams are made of.
Image of the Penguin print of Of Mice and Men from here.
For a comprehensive and in-depth discussion of umlaut, I can recommend: Minkova, D. (2014). A Historical Phonology of English. Edinburgh University Press. 157-165.
Danny L. Bate 
The umlaut plurals are the example which I always use when explaining to my students why protolanguage reconstructions tend to get the more regular and agglutinative, the deeper into the past we go. Like this:
Imagine that all Germanic languages would die out (OK, not very probable) and all written documentation of them would disappear as well (even less probable, but let’s imagine). And the only language variety surviving would be an extremely simplified and pidginized form of English, in which only regular plurals like “mouses” and “mans” would exist (or perhaps plural forms with a novel suffix, such as -pela in Tok Pisin). If this really were so, we would have zero chance to reconstruct a plural marking involving vowel alternations. Or even if we tried it, it would remain a speculation hanging in thin air.
The rise of umlaut plurals shows us how “irregularity” and fusional marking (stem alternation instead of suffixation) arise from “regular” suffixation. That is, we can reconstruct regularity (as in *mūs- : *mūs-iz) behind irregularity (mouse : mice) – but not necessarily the other way round. Especially in the more remote past of languages, where there are no written documents and perhaps not even sister languages for comparison, our picture of the proto-language very quickly turns into a simplified drawing with less and less resemblance to what a real living language was like.
LikeLiked by 1 person