LINGUIST List 15.1540

Thu May 13 2004

Sum: Latin / English Plural of Virus

Editor for this issue: Steve Moran <>


  • Ninja Looter, Summary: Latin / English Plural of Virus

    Message 1: Summary: Latin / English Plural of Virus

    Date: Thu, 13 May 2004 13:21:01 -0400 (EDT)
    From: Ninja Looter <>
    Subject: Summary: Latin / English Plural of Virus

    Dear Linguistlist,

    I'd like to thank you all for the replies to my question on the pluralization of a borrowed Latin word, the word in question being "Virus" (Linguist 15.1271).

    One curious tendency that I noticed was of many to shy away from even claiming that VIRUS would, could, or did have a plural form, as it was completely unattested in the existent literature. I was merely wondering what would be the plural of VIRUS, if it were a countable noun. Doesn't linguistic reconstruction quite often rely on unattested forms? Why such a tendency to not even postulate a plural?

    I'd like to thank the following people for their replies.

    Michael A. Covington Bruce Despain Maarten Jansonius Roger Lass Anne Mahoney Sanford Steever R�my Viredaz Nicholas Widdows Martin Wynne

    A summary of the replies follows:

    The form 'virus' is an English word. While it may ultimately descend from a form in Latin, it is commonly used by English speakers whose otherwise have no access to the form's etymology. This is as true for the Anglo-Saxon core of the language: no one thinks to use Anglo-Saxon plurals or case forms in Modern English even though a ME word may have a perfectly good AS etymology. ''Respect'' is not a linguistic notion, but an ethical one; applying this term outside its domain will not help you answer a linguistic question.

    English 'virus' and Latin 'virus' have different pronunciations. The vowels are different and the initial consonant is most likely different. I would also hazzard the guess that the 'r' and 's' are different as well.

    Their grammar is different. The nominative, dative, and accusative forms in English are all 'virus'; not so in Latin. The English possessive (if it is a case) is 'virus's', the Latin is not.

    I once heard a speaker of English use 'viri' as a plural, but the audience thought it was priggish for him to do so. The only case in which such a plural might be used is a case of code-switching.

    Oh yes. The meaning of the two forms is different. It seems most likely that speakers of classical Latin did not use the form 'virus' to refer to the class of entities that speakers of Modern English do.

    Your question may well rest on an assumption that you have not considered. The question assumes that a language is merely a collection of words. Modern linguists view language as a system of rules and processes that generate forms, phonological, lexical, morphological, semantic. A word ''belongs'' to a language by virtue of its participation in the network of those rules. There is no such thing as a Latin, English, Chinese, Hindi word that exists apart from its participation in the rules of a natural language. That being the case, it is natural that the form virus, when borrowed into English, should conform to English rules of pronunication and to English rules of plural formation.

    The view of this Latin scholar is that ''virus'' has no attested plural in Latin. It was an unusual, rather rare, indeclinable mass noun. I don't know the actual origin, but most Latin speakers probably thought it was Greek or some other neighboring language.

    (You should look it up in the Oxford Latin Dictionary, which I don't have handy, in order to confirm this.)

    So I use no plural for ''virus'' except ''viruses.''

    Among computer enthusiasts I see ''virii'' moderately often. This smacks of pseudo-pedantry because of course the regular plural of -us is -i, not -ii.

    A little more information:

    corpus/corpora, genus/genera, tempus/tempora are real Latin singular/plural pairs, 3rd declension neuter.

    -us is the regular ending of 2nd and 4th declension nouns which you are already aware of.

    virus has no attested plural in Latin. The reason we don't use a Latin plural for ''virus'' is that there is none.

    If ''virus'' were 2nd declension masculine or feminine, its plural would be ''viri''.

    ''Virii'' would be the 2nd declension plural of ''virius'' which is not a word.

    If ''virus'' were 4th declension masculine or feminine, its plural would be ''virus'' with a long u in the plural form.

    But ''virus'' is neuter. The only Latin neuter nouns that end in -us are 3rd declension (corpus, tempus, genus); these are contracted from unattested preclassical *corpors, *tempors, *geners and did not historically have --us (or its predecessor -os) at all.

    ''Virus'' does not even seem to be one of these. As I recall, no case or number forms other than the nominative/accusative singular are attested at all. The Latins seem to have treated it as a foreign word with an incomplete paradigm.

    As a Latin word, 'virus' does not have an attested plural. It has a second declension genitive singular but it is quite rare. The word isn't common in general. As a result, the appropriate English plural really should be 'viruses'.

    This question comes up a lot; it's something of a FAQ on the Classics list, where you can find a relatively recent discussion at (and the most current archives at if you're curious).

    Hoorah for our side! It's hard to change the course of a railroad train barreling down the tracks, but one small step for a man can be a leap for mankind. Keep pushing for the correct form. Maybe it will catch on. The use of ''locus'' with its proper plural of ''loca'' is seldom seen. In Mathematics they say ''loci'' and think nothing of it. It's the norm. Nowadays the most popular usage will win out, I'm afraid. I always try to use the Latin plurals for ''census,'' ''apparatus,'' and ''excursus'' but it is not always easy.

    The best, most thorough discussion of the plural of 'virus' was at The conclusion was that not only did the Latin word not have a plural, but there was no firm analogy by which to form one. So the plural is simply English 'viruses': none of the Latinate ones is supported.

    Unfortunately that page is no longer there, though Google has it cached. As my homepage had a link to it, I've temporarily taken a copy of the Google cache and stored it as while I check with the author.

    I disagree with your posting on several points.

    Firstly, I have never heard or seen any of the 'Latin' plurals of virus in use, only 'viruses'. A search of the major reference corpora of modern English find no citations of the forms which you suggest ('viri', 'virii', 'vira'). The only occurrences I can find on the web are discussions which cite them as incorrect forms. Furthermore authoritative reference sources such as the OED have only viruses as the plural form.

    I also think it is incorrect to argue that words of Latin origin should have Latin plurals, out of ''respect''. The reasons why some latinate plurals are in use in technical scientific usage are rather more complex and varied than ''respect''.

    And finally, I think your Latin is wrong, but I'm sure the experts will respond to that one

    Latin uirus means 'poison' and has no plural i.e. it is uncountable. It is neuter but inflected like a masculine (or feminine) of the second declension. A few Latin masculines are neuter in the plural, e.g. locus 'place', pl. loci m. and loca n. , the latter with a collective value. However, if uirus had a plural, it would not be of collective meaning, at least not in the modern sense of the word. I think that we should not do linguistics-fiction and that the only legitimate plural of the modern word virus is viruses in English (and virus in French, where plural -s is not added to words already ending in -s).

    I've asked myself this question before and did some digging, so it doesn't take me too long to come up with something:

    The plural of the modern English word ''virus'' a much debated issue on the web because of some varieties circulating (virii, ...) - but the dictionary sources are consistent: ''viruses''. Apparently, the Latin ''virus'' is a 2nd declension word of the *neutral* gender, a fairly rare thing in Latin. Too rare even, to be sure about a possible plural form. The heart of the problem in finding the ''original'' Latin plural, is that the word ''virus'' in Latin (meaning ''venom'', ''snake poison'') is a mass noun: it has no plural. ( I got this from )

    As far as I can tell, Latin doesn't use plurals of mass nouns to mean ''different sorts of X'', like English does: oil - oils : oil being a mass noun, oils means ''more than one kind of oil''. wine - wines : same thing (Another plural use of mass nouns, ''two beers'' meaning ''two glasses of beer'' doesn't quite seem to apply to ''viruses''.) Apart from this, I suppose the modern-day English word ''virus'' is a countable noun, not a mass noun.

    The most Latin-true educated guess to a plural form might be ''vira'', seeing as the stem is ''vir-'' and neutral plurals are always in -a. Still, I think it makes sense not to ''create'' a Latin plural for a modern English word with different meanings (both computing and medical terms), using uncertain morphology!

    There's one more thing I'd like to mention, just for comparison: In Dutch, the use of latinate plurals is similar to that in English, and perhaps even more widespread (''musicus - musici'' for ''musician - musicians'', ''musica - musicae'' for female musicians, etc). Still, there's some cases where this doesn't work. A nice example is the word ''casus'', meaning ''case'' (legal and medical context). The Latin plural (4th declension, masculine) is also ''casus'', which is not at all convenient. So the standard Dutch plural-forming in ''-en'' is used: ''casussen''. (the ''s'' is reduplicated by an orthographic rule), even though the legalese and medical jargon is traditionally a stronghold of latinate terms. (By the way, the Dutch plural of ''virus'' is ''virussen'' ...)

    Well - I hope you don't mind my lengthy reply :)

    This is an interesting query. I've never heard any plural except viruses, and the others would seem to be me rather pedantic and counter-trend productions. (Note increasingly common 'indexes', etc.)

    Virus 'poison' is in fact Latin 4th declension, so one would (abstractly) expect L [wiru:s] as the plural. But L 4th declension nouns don't come into English with the plurals expected, simply because English doesn't use vowel length as a pluraliser. the only parallel case I know is 'punctus' y"ull stop in palaeography, which is also 4th declension, and comes into technical usage unchanged, one simply says 'punctus' for sg and pl. At least I've never heard a palaeographer say *[pUnktu:s] except as a joke.

    The word itself in its present usage in any case does not come into English from Latin but French (as far as I know it was coined in the late 19th c. by Pasteur). There was a Latin borrowing used in the 18th c., but our use (a pathogen that is not a bacterium) comes from Pasteur's use for an invisible pathogen that cannot be seen (hence it's a kind of 'poison', the original sense) but passes through ceramic filters that block bacteria.