In a publicity stunt, Toyota took out a New York Times ad, put out a YouTube video, and distributed a survey at the 2011 Detroit Auto Show, asking the public what the plural of Prius should be, in a campaign announcing that there is going to be a family of Prius models. I hesitate to reward them with more publicity for such a willfully dumb question. After all, Jan Freeman already covered the topic back in 2007 in her Boston Globe column, as has Nancy Friedman on her Fritinancy blog. But I can't help myself. This is too good an excuse to talk about the wider topic of phony Latinate plurals. Well-played, Toyota.

I called this a willfully dumb question because the Prius has existed since around 2000, so there has been plenty of time for a plural form to have been created and used, and that form is the normal English plural you'd expect: Priuses. Even when there was only one kind of Prius, there were still individual units of it to talk about. Now this is where it's customary to make a joke along the lines of, "The question is moot, because who would own more than one Prius?" I will stipulate this point because it's irrelevant: What about all the Priuses in Toyota factories and showrooms, or in an entire city or state? On Toyota's own website, there's a PDF file on the history of the Prius, which says:

Between the start of the pre-sale and its arrival in dealerships, 1,800 Priuses were sold with an MSRP of $19,995.

In the "Pressroom" section of the site there is a document talking about when you should expect to have to change a Prius's battery:

At an average of 15,000 miles per year for seven years, that suggests that these 2001 Priuses could have a bit more than 100,000 miles on them.

Also on their website is a page for stories from Prius owners. One talks about a road-trip game of counting Priuses, which includes the rule: "Pictures of Priuses don't count!"

If you truly believe that Prius deserves a Latin plural, what you want is Priora, the same answer given by a reader of Jan Freeman's column, by a Latin scholar quoted in a Detroit Free Press article about the campaign, and by Ben Zimmer in the same article.

What, not Prii? No, and this brings us to the main topic of bogus Latin plurals. It's true that many Latin nouns whose singular ends in -us have plurals ending in -i; for example, there's the singular alumnus, and its plural alumni. (In fact, this is an oversimplification: These are just the singular and plural endings for the nominative case—the form used for subjects of sentences. The four or five other cases have other singular and plural endings.) The trouble is that this pattern holds only for one family of Latin nouns and adjectives, known to Latin scholars as the second declension. (For our purposes, declension just means a family of nouns or adjectives that all follow the same pattern for making singular or plural forms.) There are five declensions in Latin, two more of which contain singulars that end in -us.

In the fourth declension, all the masculine singulars, such as fructus "fruit" and lacus "lake," end in -us. But fourth-declension plurals don't end in -i; they end in -ūs. This declension is where nouns derived from verbs tend to end up, such as status (from the verb meaning "stand"), which is why the plural of status is not stati. If you're speaking Latin, it's statūs, but since you're not, it's statuses.

Prius actually belongs to the third declension. It's a neuter adjective meaning "earlier." (We borrowed the masculine/feminine form into English as prior.) There are also third-declension neuter nouns ending in -us, such as opus "work of art" and corpus "body." Their respective plurals are opera and corpora, and with this context the plural form priōra makes a bit more sense.

Of course, there's also the option of saying that Prius is a newly created Latin word (Why not? They do it at the Vatican all the time!), a homonym of the third-declension prius, but actually belonging to the second declension. In that case, yes, the plural would be formed by replacing the -us with -i, giving us Prii. This is the same way we get the plural radii from the singular radius.

This brings us to another mistake made by overenthusiastic Latin pluralizers: Assuming that every plural that ends in -i actually ends in -ii. In Latin, that only happens with second-declension words ending in -ius (the so-called i-stems). For some reason, this error is especially common when people are playfully taking words that aren't Latin borrowings at all, and treating them as Latin second-declension nouns because they end in -us. Or rather, they end phonetically with [əs], whether that's spelled -us or not. This is how the old-fashioned feminine word stewardess gets a Latin masculine plural ending in stewardii, and how Elvis impersonators (or for that matter, people who are actually named Elvis) are dubbed Elvii.

The biggest stretch I've seen for the second-declension pattern is for a word that didn't even end phonetically with an unstressed [əs] syllable, but with a stressed [ɔs]: Winklevoss, the last name of the twin brothers who sued Mark Zuckerberg for stealing the idea that became Facebook. They're sometimes referred to as Winklevi, and more often as Winklevii. The confusion over whether to use one i or two in these words probably exists because with phony Latin plurals, -ii is usually pronounced the same as plain -i: "eye." In contrast, real Latin plurals ending in -ii have both vowels pronounced: "ee-ee" or sometimes "ee-eye".

Returning to Elvis, this word exemplifies another mistake involving the third declension. In addition to neuter nouns like corpus, the third declension also contains masculine and feminine nouns ending in -is, whose plural forms end in -ēs. So if you had to have a Latin plural for Elvis, it would be Elvēs (pronounced "el-veez," analogous to crises or analyses, not like Santa's helpers). We've borrowed real third-declension masculine nouns into English, and they get mistaken for second-declensions, too. Case in point: penis, whose Latin plural is not peni(i), but penēs.

In addition to these classes of mistakes with -i plurals, there are also a few one-offs. There's the singular word omnibus, which was borrowed from a Latin third-declension dative plural (meaning "for everyone"). The suffix here isn't -us, but -ibus. But the word is still a Latin borrowing ending in -us, so it's almost inevitable that English speakers would create the plural omnibi alongside omnibuses. Ignoramus isn't even a noun in Latin: It means "we don't know," and -amus is the first person plural present tense ending for ignorare. All the same, you can find ignorami where speakers intend to refer to ignoramuses.

And then there's virus, which means "slime" or "poison" in Latin (the Romans didn't know about viruses). By now you know why its plural isn't virii, but you might be surprised (as I was) to learn that the plural isn't even viri (which actually means "men" in Latin). Not because virus belongs to the third or fourth declension—it belongs to the second. It's because virus is that strangest of nouns, a second-declension noun ending in -us that is neuter instead of masculine. (Almost all second-declension neuters end in -um.) In classical Latin, there are simply no plurals attested for nouns like these. Luckily, English is equipped to handle this gap with some pluralization tools of its own: viruses. (For more details on virus, see this thread in the forum pages on a website devoted to the Latin language.)

So far, every bogus Latin plural we've discussed has been a result of words that don't belong there getting pulled into the vortex of the Latin second declension. However, the first declension causes some confusion, too. This is the home of mostly feminine nouns and adjectives with singulars ending in -a; for example, the pair alumna/alumnae. Aside from first-declension words, the most common words to end in -a are neuter plurals, and in English, they'll sometimes get re-pluralized with the first-declension -ae suffix, as in operae and phenomenae.

Actually, phenomena is a Greek, not Latin, neuter plural. In addition to neuter plurals, Greek also contributes a fair number of neuter singulars that end in -a, such as stigma and schema. The Greek plural forms are stigmata and schemata, but after a trip through the Latin first-declensionator, they become stigmae and schemae.

But now that we've reached Greek, this seems like a good place to stop. If you want to find out more about Latin plurals stapled to Greek words, check out this Language Log post on syllabi, and Ben H. Winters's Visual Thesaurus column all about octopi.

As one participant wrote in the virus thread on the Latin website, "It is indeed strange how people feel compelled to use a plural form from a language they don't understand when there's a perfectly good plural form ready at hand in their own language!" The only reasons I can think of are to appear intelligent or to be funny. Either way, the intended effect relies on the assumption that the speaker knows the rules of Latin pluralization. When that assumption turns out to be false, the effect is nullified.

Click here to read more articles from Behind the Dictionary.

Neal Whitman blogs at Literal-Minded, where he writes about linguistics in everyday life from the point of view of a husband and father. He taught English as a second language while earning his degree at Ohio State University; has published articles in Language, Journal of Linguistics, and other publications; and writes occasional scripts for the podcast "Grammar Girl's Quick and Dirty Tips for Better Writing." Click here to read more articles by Neal Whitman.

The Detroit Free Press reported on the Prius-pluralizing P.R. stunt.
The author of "Sense and Sensibility and Sea Monsters" got stuck on the plural of "octopus."