Do you say “in the circumstances” or “under the circumstances”? Never mind, I wouldn’t believe you if you told me—people are notoriously bad at analyzing their own language use. Arnold Zwicky has been investigating the alternatives, and has posted his results, which are surprising and interesting:

In summary: the Google data suggest that “under” is preferred to “in”
with determiners “the” and “these”
(more strongly)
with determiner “which”
(very strongly)
with determiner “what”
(almost categorically)
with quantity determiner “no”
but that “in” is preferred to “under”
with quantity determiners “all” and “some”
with determiner “those” in general
with quantity determiner “many”
(almost categorically)
when “circumstances” means ‘personal situation’
(almost categorically)
with determiner “those” plus certain following relatives
(almost categorically)
with quantity determiner “a few”

See his post for the details (I’ve rearranged the results for clarity); it makes clear both the complexity of usage and the value of the internet for sifting it.


  1. Well, I’m a firm believer in the value of cautiously regimented and interpreted Google searches for such investigations. They’re especially useful for tracking current shifts in English usage by register and by geographical provenance. But for long historical trends a different and more settled corpus is called for.
    I’m sure there are many options for the more dedicated and the more professional; but I myself have found the humble old CD-ROM Library of the Future (World Library, 1991, 1994) convenient and copious enough, with its full texts of 1770 works by 205 “standard” authors, almost all out of copyright. One can search for whole phrases, and by period and country of origin. That’s enough for a start. I recall using it mapping the relative prevalences of while and whilst, for example. (I can’t abide whilst as current English.)
    Preliminary and unsystematically chosen searches suggested by some of the usages LH lists yield these data:
    the 76/42 [=titles hit/authors hit]
    these 47/30
    no 12/10
    such 67/41
    the 35/24
    these 21/19
    no 0/0
    such 33/26
    under within 5 words of circumstances
    in within 5 words of circumstances
    circumstances tout court
    You get the idea.
    By the way, LH: I can assure you that I for one can analyse [sic] my own usage sufficiently well to report that I wince inwardly if ever I find myself using under with circumstances, which I do from time to time, and from circumstances to circumstances. A mere atavistic reflex.

  2. Gene Fellner says:

    This illustrates a revelation that came to me when I began studying Chinese: Prepositions are useless. No, actually they’re worse than useless. By having a grammar that insists on using prepositions (and conjunctions) to express relationships, we’re severely handicapped when cultural changes bring about new kinds of relationships. We’re still using a set of no more than twenty prepositions (and even fewer conjunctions) that are left over from the Neolithic Era!
    They have been overused so much that they are now practically meaningless. Your example is a perfect one. For another, try explaining to a foreign student the difference between arriving at a class “in time” and “on time.”
    It’s almost to the point that the only real purpose prepositions serve is to help us identify native speakers!
    Is it any wonder that during the last couple of generations, when the Paradigm Shift from the Industrial Era to the Information Age began, we also revolted against our Stone Age grammatical paradigm and invented a new way to express relationships with constructions like cable-ready and user-friendly?
    I have always maintained that the Chinese language is far more adaptable because it essentially has only two parts of speech: nouns and verbs. (A handful of particles like “de” and “ge” really only serve to help parse spoken sentences which can be ambiguous due to homophones and would not be missed in writing.) Relationships are expressed by nouns and verbs–and since there are thousands of each, the potential for describing a new relationship accurately is limitless–and by Chinese’s strict rules of word order. Instead of, “The dog is IN the box,” you say, “Dog occupy box interior.” Instead of, “AFTER breakfast I went TO school ON the bus,” you say, “I eat breakfast ride bus attend school.”
    If China truly becomes the first Former Great Civilization to have a second chance at power, I believe it will be largely due to the advantage of a language that is perfectly suited for changing times.

  3. That’s funny, I always envisioned “in time” as being just that–within the space of time delineated on one end by some particular event–whereas “on time” meant right “on” some particular predefined spot. If students pressed me too far, though, I’d resort to the old “It’s an idiomatic usage” bit–teacherese for “just memorize it and quit bugging me.”
    English prepositions do all sorts of handy things. Look at “wait” plus “for,” “at,” “to,” “until,” etc. Chinese would just employ different strategies for these. 在…等, 等X, 等到X
    One day my old Chinese professor complained to me about students translating “on the phone directly into Mandarin, saying “電話上” She said she wanted to bring a telephone to class, stand on it, and say “In Chinese, ON the phone means this!” Then I asked her what she’d do when it came up that the correct preposition is “in.” (電話中) There’s no particular logic to either one.

  4. Oh, I forgot to mention, I am for “under” 100%. I was sad to see I was in the minority. Maybe I have become influenced by Chinese 情況下.

  5. I find that article reflects my instinctive usage 100%. Especially the note that we are in personal circumstances and under impersonal (work, political, economic, etc) circumstances.
    Why? I have no idea but one could suggest that impersonal circumstances are imposed from above, whereas personal circumstances surrond us and we have the power to change them. I have trouble carrying that metaphor out to other usages but possibly it could be done.

  6. Gene Fellner says:

    Actually, “wait” is another illustration of my own thesis. The distinction between “wait on” and “wait for” is rapidly being lost, at least in the US. When someone says, “I’m waiting on the FedEx man,” it makes me cringe, yet I have no trouble understanding that he’s not planning on serving the Fed Ex man lunch. Context is everything when it comes to prepositions.
    As for “dian hua” (sorry I don’t have a Chinese font but since I only know about 300 characters it’s not a problem), it helps to remember its root meaning: electricity speech. Phrases built around it refer to the speech itself, not the instrument. So when one is “on the phone” one does not “occupy electric speech top” — “dzai dian hua shang.” One “ignites electric speech” — “da dian hua.” “Da” has no consistent mapping to one English verb, its most common translation is “hit” but it also means to “play” a parlor game or a musical instrument or to “ignite” a phone call.

  7. “This illustrates a revelation that came to me when I began studying Chinese: Prepositions are useless. ”
    What are co-verbs if not prepostions? What do they show if not relations between the main verb and the object of the co-verb? Gen, ba, gui, xiang are all relational.
    “da” has the more general sense of “do something forcefully or intentionally” as in “da gong” = work, hold a job, or “da suan” = form a plan , suan – count, reckon. Really useful and probably pretty productuve.

  8. Furthermore, if dianhua actually meant ‘electricity speech’ it would not be used to refer to the physical object, which it is.

  9. “Dianhua” is pretty elastic. It refers to the appartus and to the call itself, as well as the technology. “Da ge dianhua” does not mean to beat the telephone.

  10. Yes, of course, I was just pointing out the problems with trying to analyze it as “electricity speech.”

  11. Also, if it’s “electric speech,” then English “telephone” is “distant speech.” The etymology is just more obvious in Chinese than in English. Of course, nobody thinks of it in that way.

  12. Exactly. Etymology is almost always irrelevant to meaning.

  13. Andrew Dunbar says:

    A better example than Chinese would be Japanese which has a tiny number of postpositions but の “no” is the jack of all trades covering a large number of English prepositions.
    Another interesting example is Swedish which has a “normal” number of prepositions but also a jack of all trades, “som”, which can mean just about anything prepositionwise.

  14. The word Circumstance comes from the Latin root “Circum”, meaning “around”.
    Something surrounded is within, not “under”.
    So it is “in the circumstances”, not “under the circumstances”
    I cringe every time I hear “under the circumstances”

  15. marie-lucie says:

    Ian Webb, I think your explanation falls into the category of “the etymological fallacy”.

    Speaking as a former ESL student, French speakers have a lot of trouble with English prepositions, which are usually more precise than the ubiquitous French à, which needs to be translated as at, to, in, with, and probably others, depending on the exact circumstances.

  16. But for long historical trends a different and more settled corpus is called for.

    Nine years later, of course, we have that corpus: Google Ngrams, where we see that in and under began almost neck-and-neck in 1800 and were almost neck-and-neck in 2000. From about 1830 to 1960, under surged way ahead, peaking in 1885 at about four times as frequent as in. Of course, we also know a lot more about the unreliability of ghits (as opposed to Ngram hits) than we did then.

    It’s easy to re-run Zwicky’s other tests too. These looks much like the, except that the period in which under beats in is 1805-1935, and at its peak under is six times as frequent. For which, it’s 1830-1930 and four to one, with much more irregularity 1800-1830. For what, things are different: in and under start almost even, and under rises to a 9:1 peak in 1860, but even when its wild popularity ends in 1939 it stands at 3:1, and then rises again to almost 5:1 by 2000. For no, under beats in 20:1 in 1800, rising to 38:1 by 1900, but then in begins a modest rise of its own, so that when under reaches its absolute peak in 1917, it is only leading by 11:1. When in peaks in 1950, under is only leading 3:1, and both decline thereafter to only 6:1 in 2000.

    I took my as a proxy for Zwicky’s “personal situation” case. Here in is always dominant. Both phrases peak by 1808, with in at about 8:1, and by 1869 both are on the skids, with in leading only 5:1. By 1820, though under is practically gone, and in was beating it 13:1 (with very low absolute numbers) in 2000. I had to turn off case-sensitivity to get meaningful results here, for some reason, so basically instances at the beginning of sentences are arbitrarily excluded.

    Those is very strange: in starts out more popular and goes up to 2:1 by 1807, but then they reach parity by 1815 and keep it until 1840. Then under has a modest surge and ina more dramatic crash, leading to almost 5:1 in 1881. They eventually converge by 1960 and remain so until 1990, after which in becomes a little more popular, but nowhere near 2:1 by 2000.

    For all the prepositions start equal, with under takes off fast, peaking at 5:1 by 1955, but the curves converge in 1980 and remain so. Some is nothing like all: it starts at 4:1, declines to parity by 1830, and continues to drop as under rises until by 1904 the ratio is 4:1 the other way. However, i>in begins a rapid and steady rise around that time, so that parity is reached again in 1960 and by 2000 in is winning more than 2:1 with much higher absolute frequencies.

    Many has in starting at 8:1 and dropping to parity by 1865, which is retained till 1945, when in takes off again, reaching 5:1 in 2000. The absolute frequency of under remains more or less constant throughout. A few is much the same as many, but under remains very rare throughout, with in plummeting by 1815 and not beginning to rise until 1940.

  17. David Marjanović says:

    I cringe every time I hear “under the circumstances”

    In German, unter diesen Umständen is decidedly the more common option, where Um-stand is an obvious calque of circum-stantia. I can’t remember having ever encountered in diesen Umständen; Google can, but finds 9.77 million results for in vs. 13.4 million for unter.

    (In anderen Umständen does exist, but is an obsolescent euphemism for “pregnant”.)

  18. Stefan Holm says:

    And north of the Baltic Sea we faithfully follow big brother in the south:

    Unter diesen Umständen = under dessa omständigheter.

    In (anderen) Umständen = i omständigheter.

  19. I suspect if Ian Webb were living in 1885 he would cringe every time he heard “in the circumstances” — and find reasons equally satisfying to himself for stating with confidence that “under” was the only acceptable option.

  20. To keep it rather simple, the correct terminology (according to my English tutor) is ‘in these circumstances’ this arises from the meaning of the word. Circum ‘around’ and stance ‘a position’ one is therefore sorrounded by (and therefore within) as opposed to under, meaning beneath. However the importance of language to make oneself understood….

  21. marie-lucie says:

    You must mean “beneathstood”.

  22. Stephen Bruce says:

    We really ought to say “surrounded by the stances”…

  23. Stefan Holm says:

    Whoever, Bill, said that our surroundings are only horizontal? Aren’t circumstances those very things we feel unable to change – like given by the heavens above? And under the firmament we stand (except those guys on board the ISS).

Speak Your Mind