Why Some English Plurals Don't Just Add -S

Ask why "mice" isn't "mouses" and most people reach for the same answer: English is just inconsistent. It's a fair reaction, but it skips past something more interesting — every irregular plural in English traces back to a specific, datable historical event, and those events belong to at least three unrelated stories layered on top of each other over more than a thousand years.

Story one: a vowel shift that predates written English

The oldest layer of irregular plurals — foot/feet, mouse/mice, goose/geese, and their relatives — isn't the result of a foreign word muscling its way into English at all. It's a homegrown sound change called i-mutation: Old English once pluralized these nouns with an extra syllable tacked on the end, and that syllable contained a vowel sound that pulled the pronunciation of an earlier vowel forward in the mouth as speakers said the word aloud. Centuries later, the triggering syllable itself vanished from speech entirely — but the vowel shift it had caused stayed behind as the only remaining signal that the word was plural. Because this happened through ordinary sound change rather than borrowing, it was locked to a small set of everyday Old English nouns already in the language at the time, and the door on that set closed for good roughly a thousand years ago.

Story two: academic and scientific borrowing from Latin and Greek

A second, much larger and still-changing layer arrived by an entirely different route: wholesale borrowing of Latin and Greek nouns into English scholarship, medicine, and law, complete with their original plural endings attached — criterion/criteria, analysis/analyses, appendix/appendices, cactus/cacti among them. Unlike the frozen i-mutation set, this layer is actively in motion right now: as a borrowed word drifts from specialist usage into everyday speech, English speakers increasingly default to the ordinary -s ending instead of the inherited classical one, exactly the same frequency-driven pressure that slowly regularizes irregular verbs over time. "Formulas" now sits comfortably alongside "formulae"; "indexes" alongside "indices" — the choice mostly signals which register a writer is aiming for, not which one is correct.

Octopus makes the best case study in this layer precisely because its history gets misapplied so often. The word looks Latin, which invites people to guess the plural the same way they'd guess "cactus"/"cacti" — hence "octopi." But octopus actually came into English from Greek, so the classically "correct" plural is really "octopodes," a form that's all but extinct in ordinary use. The version that actually wins in practice, "octopuses," sidesteps the whole argument by simply following English's own regular rule instead of borrowing anyone else's.

Story three: an accident of Old English grammar, not a sound change at all

A third, entirely unrelated layer covers nouns that take no plural marker whatsoever — sheep, deer, moose, and a handful of others. These didn't lose a plural ending to sound change the way the i-mutation nouns did; Old English simply filed them under a grammatical class where the singular and plural forms were already identical from the start, long before any of the sound changes above ever occurred. Most of these are, tellingly, animals people once counted in herds or hunting parties rather than one at a time, where the surrounding sentence usually made the quantity obvious without needing a separate plural form to do it. Fish belongs to this group too, with one genuinely useful wrinkle: it takes a real plural, "fishes," but only when the sentence means multiple species rather than a body count of individual animals — a distinction that actually carries information, unlike most of the fossilized patterns covered above.

Why three unrelated causes look like one big mess

These three stories share no common ancestor and happened centuries apart from each other — an i-mutation plural, a classical-loanword plural, and a zero-plural noun class are three separate accidents of history that only look related because they all produce the same surface symptom: a plural that doesn't just add -s. Lumped together as "irregular plural exceptions," they read as arbitrary. Sorted by which of the three actually produced a given noun's plural, they resolve into three coherent, traceable narratives instead of a wall of memorization.

A fourth story worth knowing: a consonant that changed sound mid-word

One more pattern deserves its own mention because it comes from neither vowel-mutation nor foreign borrowing: knife/knives and its relatives (wife, wolf, shelf, calf) all descend from a plain phonetic rule — a word-final "f" sound automatically turned into a "v" sound in Old English whenever a vowel followed it, which is exactly the position it ended up in once the old, fuller plural ending was attached. That ending eventually wore away from pronunciation, but the consonant shift it had triggered stuck in the spelling regardless — visible today even though nothing in current pronunciation still explains why it's there. Some -f nouns skipped this pattern altogether (roof, chief, belief simply add -s), and a handful of borderline cases, like "hoofs"/"hooves," are still genuinely unresolved between dictionaries, a live-in-progress shift rather than a settled rule.

Compound nouns: a different kind of irregularity entirely

A related but structurally different category of plural mistake involves compound nouns, where the actual puzzle isn't which ending applies but which word inside the compound gets marked plural in the first place. "Attorney general" becomes "attorneys general," not "attorney generals," because "attorney" is the noun actually being counted and "general" is a modifier trailing behind it — a French-derived word order where the descriptive word follows the noun instead of leading it. "Passerby" follows the identical logic in reverse, pluralizing as "passersby" because "passer" is the noun and "by" is the trailing element. Some compounds are visibly losing their internal structure in speakers' minds, and pluralizing at the very end instead: "cupful" is increasingly written "cupfuls" in place of the historically correct "cupsful," a shift that tracks how strongly a given speaker still perceives the compound as noun-plus-modifier rather than as one solid word.