« previous post | next post »

Wrong ethically? Practically? Legally?

No — according to this t-shirt, polyamory is wrong morphologically:

The morphological picture is actually more complex — the OED gives

Etymology: < poly- comb. form + classical Latin amor (see amour n.) + -y suffix, after polyamorous adj.
In form polyamoury probably after French amour amour n.

And poly- is originally from Greek, but was enthusiastically borrowed into Latin, French, and German. English poly- words have been borrowed from those languages — "combining chiefly with second elements ultimately of Latin or Greek origin". As the OED explains:

Etymology: < ancient Greek πολυ-, combining form (in e.g. πολύγωνος polygon adj.) of πολύς , πολύ much (in plural, πολλοί , πολλαί , πολλά many) < an ablaut variant (o -grade) of the Indo-European base of fele adj. Compare classical Latin, post-classical Latin, and scientific Latin poly-, French poly- (formations in which are found from at least the late 17th cent.), German poly- (formations in which are found from at least the early 19th cent.).

Attested earliest in the Old English period in the classical Latin loan polytrichon n., and subsequently in the late 14th and early 15th centuries in loans from Latin and French (polyp n., polypus n., polymite adj.), and in the 16th and 17th centuries in adaptations of Latin, Greek, and French words (e.g. polygamy n., polysyllabe n.), which become very frequent from the late 18th cent. onwards, chiefly in scientific vocabulary. Occasional formations within English are found from the beginning of the 17th cent. (e.g. polytragic adj., polytopian n. at sense 1), but the bulk of English formations date from the 19th and 20th centuries.

Compare multi- comb. form, which shows considerable semantic overlap with poly- comb. form, although it does not show the specialized use in chemistry (see sense 2). Compare also mono- comb. form.

Combining chiefly with second elements ultimately of Greek or Latin origin.

Restrictions of that kind ("Combining chiefly with…") are one of the many quasi-regular aspects of morphology, which might stand as a metaphor for the quasi-regularity of morality.

[h/t Fritz Newmeyer — I'm not sure who gets credit for the t-shirt.]

Update — Ben Zimmer points us to Stan Carey's terrific 2011 blog post, "The monstrous indecency of hybrid etymology", as well as the source of the t-shirt.


  1. David Charlton said,

    December 2, 2018 @ 7:46 am

    This is a word with a possible sub meaning like "politics", which stems from two roots, "poli" which means "many" and "tics" which are blood sucking insects.

  2. Ricardo said,

    December 2, 2018 @ 7:52 am

    If Polyamory Is Wrong I Don't Want to Be Right

  3. Andrew (not the same one) said,

    December 2, 2018 @ 8:43 am

    Keeping Greek and Latin roots apart has been a losing battle since the word 'bicycle' was invented.

    But are there any historic examples of 'poly-' used with Latin roots? All the examples in the OED quote above seem to have Greek roots.

  4. outeast said,

    December 2, 2018 @ 9:08 am



    From a curious scrape of my dictionary I see most poly- words are scientific, such as polyamide, polycarbonate etc. In these examples, amide and carbonate both have Latin roots, albeit by a circuitous route, so they fit your "mixer" criterion. Or maybe there are earlier Greek roots that aren't listed.

    In general it seems many poly- words are based on Latin words that were loaners from Greek in the first place (eg polychromatic, where we got chromatic from French chromatique, in turn from L chromaticus, earlier Gr. khromatikos). Bicycle is similar: yes, bi- is Latin and kuklos Greek, but cyclus was Latin so it's not really a cross-language admixture.

  5. Q. Pheevr said,

    December 2, 2018 @ 9:26 am

    Chamberlain: We belong to a sort of secret society, the Order of Chaeronea, like the Sacred Band of Thebes. Actually it’s more like a discussion group. We discuss what we should call ourselves. ‘Homosexuals’ has been suggested.
    AEH: Homosexuals?
    Chamberlain: We aren’t anything till there’s a word for it.
    AEH: Homosexuals? Who is responsible for this barbarity?
    Chamberlain: What’s wrong with it?
    AEH: It’s half Greek and half Latin!
    Chamberlain: That sounds about right.

    —Tom Stoppard, The Invention of Love

  6. David Marjanović said,

    December 2, 2018 @ 9:54 am

    Ticks are mites, closer to spiders and horseshoe "crabs" than to crustaceans including insects. Just saying. :-)

  7. Richard Hershberger said,

    December 2, 2018 @ 10:03 am

    This discussion concedes too much. It tacitly accepts the claim that Latin and Greek elements can't, or at least oughtn't, be combined, but then argues that poly- is an exception to the general rule. Why are we to believe the premise? And if we do, is this a general or a specific rule? That is, is the rule that English words with elements borrowed from other languages much be entirely from one language, or does this only apply the Latin/Greek combinations? Either way, what evidence is there that such is actually a rule of English morphology? Quite the contrary, the claim mostly seems to me to be evidence that the claimant has never actually looked at a good English dictionary. I strongly suspect that the origin of this supposed rule is status anxiety, the point being to demonstrate one's knowledge of Greek and Latin. See also: the claim that "the hoi polloi" is bad English.

  8. Robert Coren said,

    December 2, 2018 @ 10:48 am

    You didn't happen to see that T-shirt on television, did you?

  9. outeast said,

    December 2, 2018 @ 11:01 am

    It occurs to me that the fundamental premise of the knock knock joke is flawed on multiple levels, not least that door knockers are present on a clear minority of doors and almost never to the exclusion of doorbells. These jokes above all demonstrate a failure to look even cursorily at doorstep interactions, despite the fact that accurately representing a normal interaction would be, ipso facto, effective priming for an incongruity response (Berger,
    1976; Deckers & Divine, 1981; McGhee, 1979). The corpus of doorway-initial interaction (hereafter DII) is sadly lacking (surely a research opportunity) but there is little evidence to support the premise that a typical DII follows even approximately the model "knock, knock – Who is there? – Firstname – Firstname who? – Firstname-surname".

  10. Claire Graf said,

    December 2, 2018 @ 11:20 am

    -Graeco-Latin terminology in medicine and academic English allows mixing. Appendectomy. So that isn't new. It happens once those morphemes are proper loans.

    -Those morphemes are loans and in everyday English not transparent regarding their etymology. People will recognise them as belonging to the academic/medical semantic field. Beyond that, they will build their rule set on examples they have learned. And the LGBTQ+/sexual orientation terms have a solid level of mixing. Heterosexual. Homosexual.

    -Given the semantics of -philia in sexuality terminology, polyphilia would mean something slightly different.

    -Semantic shift is normal. It happens. Morphemes aren't safe from it.

    -Within the sexual orientation semantic field certain morphemes have become the only morpheme for "many" or "one". That can easily happen. Medicine did the same.

    -Language is shared code a group of speakers agrees on. The community that uses those terms regularly has agreed on this term. So it's already over.

    -Morphemes can lose or loosen their etymology restrictions. See Yiddish morphology.

  11. Michèle Sharik Pituley said,

    December 2, 2018 @ 11:26 am

    @Richard: “is the rule that English words with elements borrowed from other languages much be entirely from one language, or does this only apply the Latin/Greek combinations?”

    Octopodes or octopuses, but never octopi! Lol!

  12. MikeA said,

    December 2, 2018 @ 11:44 am

    Mixing might be "ignorant" (or just "who could possibly care?"), but sometimes has an actual purpose. The numerical radix one bit larger than octal was referred to as (Latin) Sexadecimal until IBM started calling it (mixed) Hexadecimal. Some old-timers felt this was due to grey-suited prudes.

  13. Doug said,

    December 2, 2018 @ 11:52 am

    Another well known mixed word is "monolingual."

  14. Y said,

    December 2, 2018 @ 12:15 pm

    Is Democracy another word for Polyopoly?

  15. SlideSF said,

    December 2, 2018 @ 12:19 pm

    I think I saw that tee-shirt in Vacaville, CA

  16. Rodger C said,

    December 2, 2018 @ 12:30 pm

    Did you see it from a bicycle or an automobile?

  17. Ben Zimmer said,

    December 2, 2018 @ 1:55 pm

    See also Stan Carey's 2011 blog post, "The monstrous indecency of hybrid etymology."

    (The T-shirt was designed by "KiltInspector" on Zazzle.)

  18. Sili said,

    December 2, 2018 @ 5:33 pm

    I think this is a recycled joke about "homosexuality". The danish gay activist Helmer Fogedgaard was an advocate for the proper "homophilia"/"homophile".

  19. Max Wheeler said,

    December 2, 2018 @ 6:00 pm

    Epic troll, outeast!

  20. empty said,

    December 2, 2018 @ 9:06 pm

    I did not know that insects have been merged with crustaceans!

  21. /df said,

    December 2, 2018 @ 11:11 pm

    Let's not forget television. Owing to a lack of emoticons ca. 1930 we can't now tell whether CP Scott, a veteran British editor and politician, intended a witticism or an attack on hybrid etymology. The quotation appeals especially to the sort of cultural elitists who despise television and who may also applaud unilingual etymology.

    With reference to Q. Pheevr's Stoppard extract, it's likely that he would have been familiar with the linked quotation: he did use another Scott quotation in Night and Day.

  22. stephen said,

    December 2, 2018 @ 11:12 pm

    Apparently, autokinon or ipsemobile would be correct terms. But kinonmobile has a nice ring to it.

    From Wikipedia: The name Minneapolis is attributed to Charles Hoag, the city's first schoolmaster, who combined mni, a Dakota Sioux word for water, and polis, the Greek word for city.

    And Texas is from the Caddoan word Tejas, friend; Arkansas is from either Quapaw or Sioux. And so the city of Texarkana mixes words of different Native American languages.

    And would it be just as logical to criticize people whose first, middle and last names come from different languages? But then we can extend that to criticize people who belong to ethnic groups different from the ethnic groups/languages their names come from.

  23. Joshua K. said,

    December 2, 2018 @ 11:12 pm

    Isn't "television" also derived from a combination of Greek and Latin roots?

  24. Ben Jones said,

    December 3, 2018 @ 3:03 am

    The Japanese (almost) have a word for this – in fact two, dating back to the 15th century (ref: http://kanjibunka.com/kanji-faq/old-faq/q0182/). 湯桶読み yuTŌ-yomi means a Kanji compound where the first character is read the Japanese way, the second the Chinese way; and 重箱読み JŪbako-yomi is the reverse. Naturally, yuTŌ and JŪbako are examples themselves.

  25. Peter Erwin said,

    December 3, 2018 @ 6:23 am

    There's a passage in David Wootton's The Invention of Science where he discusses the introduction in the 1830s by William Whewell of the word "scientist", which met similar disapprobation from some, well, scientists, who objected quite strenuously to the term for years on the grounds that it was a blasphemous miscegenation of Latin and Greek. The geologist Adam Sedgwick wrote in the margin of a book by Whewell: "better die … than bestialize our tongue by such barbarisms."

    (Wootton speculates that Whewell might have been looking for a gender-neutral alternative to the then-common term "man of science", since when he introduced it he was reviewing a book by the science writer Mary Somerville.)

  26. Andy said,

    December 3, 2018 @ 7:17 am

    @Peter Erwin: Deliberate irony on the part of Adam Sedgwick, I wonder, or did he not realize that 'bestialize' is just such a hybrid?

  27. richardelguru said,

    December 3, 2018 @ 7:24 am


    What about the poor innocent parrot!!!!!!

  28. Rodger C said,

    December 3, 2018 @ 7:47 am

    I've already told the story about how St. Augustine tried to talk to St. Ambrose about the Donatists and the latter was like, "no, no! Donatians!"

  29. J.W. Brewer said,

    December 3, 2018 @ 1:42 pm

    Wiktionary's category "English words prefixed with poly-" currently has 799 entries, which seems like a large enough universe for someone wanting a research project to play around with looking for, if not rules and exceptions, at least statistical tendencies. Of course it seems likely that all 799 entries are not created equal, either in terms of frequency of use or of being or not being the sole or dominant lexeme for their referent. Indeed, I suspect that for some X's, both "poly-X" and "multi-X" and/or something else are Out There as synonyms, and it might be even more interesting to see if there are patterns in which coinage predominates when there's a selection to choose from.

  30. BZ said,

    December 6, 2018 @ 1:01 pm

    In my mind the "vision" part of "television" is unambiguously English, so I don't view it as a mixture of foreign roots. This is strengthened by the fact that Russian keeps the "tele" part, but uses a Russian word for vision (though it's cognate)

  31. Paul Kay said,

    December 6, 2018 @ 8:33 pm

    A true story. Some readers may have heard of the sociologist Talcott Parsons. He was the ultimate in sociological hot stuff, at least around Harvard, in 1960 or so when, as a lowly graduate student, I heard him recount the following. He had received from Oxford University an invitation to give a series of lectures; the letter included the suggestion that the lectures be on 'social theory' rather than 'sociology' because… you guessed it.

RSS feed for comments on this post