Rein and reign

« previous post | next post »

The word rein, which the OED glosses as "A long narrow strap, frequently of leather, attached to the bridle or bit of a horse or other animal on either side of the head and used by a rider or driver to control and guide the animal", was apparently borrowed into English from French a millennium ago. The Wiktionary entry gives the etymology in detail:

From Middle English rein, reyne, borrowed from Anglo-Norman reyne, resne, from early Medieval Latin retina, ultimately from Classical Latin retineō (“hold back”), from re- + teneō (“keep, hold”). Compare modern French rêne.

Displaced native Old English ġewealdleþer (literally “control leather”).

But the OED entry makes an interesting (apparent) mistake in this case — the full etymology seems good, but the "Origin" line gives the French etymon as regne — and règne (in the modern spelling) is actually French for "kingdom", from Latin rēgnum, which is the origin of a different English word, namely reign.

Rein and reign have been pronounced the same way in English for some time — perhaps always? — and their meanings overlap in extended or figurative uses having to do with control. This led to some early eggcorns, even back in the days when most people had personal experience with physical reins. Thus the OED entry for free rein, glossed as "Freedom of action or expression. Chiefly in to give (a) free rein (to)", includes "free reign" citations going back to 1834:

1834 J. Eberle Treat. Dis. & Physical Educ. Children (ed. 2) i. i. 6 She, who giving a free reign [1833 (ed. 1) free rein] to her appetite, indulges it to excess.
1924 Times 26 Sept. 11/5 Others thought themselves above the law, and gave free reign to their passions.
1993 Outdoor Canada Mar. 33/3 So few pike survive to a large size that the ones who do have virtual free reign to raid the pantry.

Now that most people's experience of physical reins is limited to books and movies, "reign" is taking over rein-connected metaphors like "free rein" and "rein in". As a random example, there are at least two cases so far this month in Forbes.

"Another Nail In The Coffin Of The Russian Economy", Forbes 11/10/2022:

In an antidumping case, Commerce determines the dumping margin, or the extent to which the product is being sold at less than fair value. To determine this, it uses prices in the exporting country as a benchmark, if that country has a market economy. But if prices in that exporting country are not set by market forces, then Commerce has free reign to use prices from another country. This usually results in a very high dumping margin, and very high duty.

"Why Are Tech Companies Laying Off So Many Workers?", Forbes 11/16/2022:

With these headwinds on the horizon, tech companies need to reign in their spending to ensure they get through the volatile period safely.

The NOW corpus ("News On the Web") shows 7136 instances of "free rein" vs. 3738 for "free reign", for  reign/rein = 0.52. The iWeb corpus, from less formal sources, has 3264 instances of "free rein" vs. 5043 for "free reign", for reign/rein = 1.55.

The Google Books Ngram Viewer shows "free reign" rising to around 40% of the count for "free rein" as of 2002, with some subsequent decline probably due to changes in the books making up the dataset:

It's harder to quantify "reign/rein in", since there are many cases where reign belongs in "reign in", like Milton's famous "Better to reign in Hell, then serve in Heav'n." But in a random sample of 100 from the 11104 instances of "reign in" from the NOW corpus, 31 are spellings of the idiom "rein in".

[h/t Linda Seebach]


  1. Robert Coren said,

    November 24, 2022 @ 10:39 am

    Well, I sure got caught by the Recency Illusion; I assumed that "free reign" had arisen during my lifetime, as first-hand knowledge of reins faded out. I don't think I knew about "reign in", which is puzzling; it's hard to imagine what people think its literal meaning is.

  2. Cervantes said,

    November 24, 2022 @ 11:50 am

    Well, the idea of "free reign" actually makes sense, especially if given to a government agency or really any sort of authority. "Free rein" makes more sense in the case of someone who is normally subject to authority, such as a child. I wouldn't necessarily classify this is a mistake, it seems to be an intended meaning.

  3. Geoff McL. said,

    November 24, 2022 @ 3:24 pm

    I did once see a commercial on local cable for a dietary supplement that "reigns in appetite" as the titles read. When I shared the anecdote on Facebook, someone commented "Like Henry VIII?"

  4. JPL said,

    November 24, 2022 @ 8:08 pm

    What about, e.g., "take/took the reigns"? Is 'reign' increasing there as well?

  5. Mark Young said,

    November 24, 2022 @ 10:10 pm

    @JPL, Google Books Ngram Viewer reported a maximum of 0.25% in 2007 for this version over the "reins" version, but the metaphorical phrase isn't hogging all the action.

    Harry H Batsford · 2019 · ‎Preview
    >> “Take the reigns,” he handed the leather straps to her, swapping them for the gun. “Ammo?” questioned Seth. “On the roof!” blurted Constance, as she wrestled with the reigns, trying to keep the horse in as straight a line as possible

  6. Mark Young said,

    November 24, 2022 @ 10:13 pm

    Sorry — 8.9%. Forgot about case sensitive….

  7. Michael Watts said,

    November 25, 2022 @ 1:39 am

    “Take the reigns,” he handed the leather straps to her, swapping them for the gun. “Ammo?” questioned Seth. “On the roof!” blurted Constance, as she wrestled with the reigns, trying to keep the horse in as straight a line as possible

    As Mark Young notes, this isn't evidence for mistaken use of reign where rein would be expected. It is transparently a case of misspelling rein in a context where the word unambiguously refers to the straps that control the direction of a horse.

    We can then ask how many people are doing the same thing when they write "free reign". That is difficult to answer.

  8. ktschwarz said,

    November 25, 2022 @ 3:21 am

    It's not clear that the OED's "Etymon: French regne" is a mistake. Anglo-Norman did not have remotely standardized spelling, and regne was just one of many attested spellings for the word, several of which are shown in the OED etymology linked in the post, and even more at the Anglo-Norman Dictionary. Which spelling is given as "the" etymon is therefore rather arbitrary.

    Conversely, the Anglo-Norman spelling regne could correspond not only to Modern French rêne 'rein' or règne 'reign', but to reine 'queen' as well.

  9. John Swindle said,

    November 25, 2022 @ 3:40 am

    There's another word waiting to come out and play. Compare Latter Reign Ministries, the Christian musical group Latter Reign—and the Church of the Latter Rain. Of these I've only encountered the last, in the form of a lovely little church building, but I suppose each is intended as written, the first two referring perhaps to prophesied End Times and the last to God sending real or metaphorical showers.

  10. languagehat said,

    November 25, 2022 @ 8:53 am

    It's not clear that the OED's "Etymon: French regne" is a mistake.

    In fact, it's quite clear that it's not. Here's the start of the etymology:

    < Anglo-Norman raigne, raine, rein, reine, reisne, ren, reyn, reyne, Anglo-Norman and Old French, Middle French resne, Anglo-Norman and Middle French rene, Old French reigne, Old French, Middle French regne, Middle French rayne, renge (French rêne, †resne; in Anglo-Norman and Old French also as redne (mid 12th cent.)) […] < post-classical Latin retina rein (8th cent.).

    The OF/MidF form seems to have had a -g- in it, for whatever reason; the modern spelling is irrelevant to the origins of the English word. I personally would hesitate a long time before suggesting that the OED was mistaken, especially in an entry revised as recently as 2009. (Note that this Late Latin word is etymologically distinct from retina 'retina,' which is from rete 'net.')

  11. Robert Coren said,

    November 25, 2022 @ 10:25 am

    @Cervantes: Yes, it does make a kind of sense, but I don't think it's deliberate; I think, rather, that it's an alternative understanding of what the phrase means, based on having heard "free rein" without seeing it written down. (There's no way to prove this either way, of course.)

  12. Cervantes said,

    November 25, 2022 @ 11:22 am

    RC — yes, it could have originated from someone hearing "free rein" and interpreting it as "free reign," but since it makes sense either way, no harm in writing it down that way either. Actually it enriches the language since we now have two different nuances of meaning available.

  13. ktschwarz said,

    November 25, 2022 @ 1:11 pm

    languagehat: The OF/MidF form seems to have had a -g- in it, for whatever reason

    There's no such thing as "the" OF/MidF form, is what I'm saying; regne was *one* of the attested forms, as the OED's etymology shows. The Anglo-Norman Dictionary chose the spelling reisne as the headword, and also tells us that it can be found in other historical French dictionaries under the headword rene or resne; maybe these dictionaries picked whatever was closest to the modern French spelling as a convenience to the reader, I don't know, but they all list multiple spellings.

    The OED introduced abbreviated summaries like "Etymon: French regne" in a 2015 update; before then they just listed multiple forms. Most often the spelling they use for the summary etymon is the same as the AND headword, no idea why it's different here.

  14. Bloix said,

    November 25, 2022 @ 1:23 pm

    "Free rein" and "rein in" are of course idioms that reflect the experience of riding. The two reins are attached to each side of the bit in the horse's mouth, and tied together at the far ends. They are long and can be held at various positions. When you hold them with your hands low and toward the ends so that they dangle loosely against each side of the horse's neck, you are "free reining," meaning that you are not making contact with the horse's mouth through the bit. At the same time you relax your legs so that you are not putting pressure on the horse's flanks. The combined lack of pressure and absence of contact with the mouth signals to the horse that it can move in any direction it chooses. You might free rein your horse after a hard canter, or while chatting with another rider in the ring, or simply while waiting around for some reason.

    If your horse starts to wander off in direction you don't want to go, or when it's time to get going, you shorten the reins by sliding your hands up them. You lift your hands and pull them back into your body, until you make contact with the bit. At the same time you apply light pressure to the horse's flanks with your legs. By "reining in" the horse in this way, you communicate, through the application of light pressure, that you are in control and will direct the horse's movements now on. A trained horse responds to these signals promptly and without protest.

    To anyone who rides – which at one time meant everyone of a high enough social class – the pair of idioms is expressive, describing the application and relaxation of mild pressure to communicate the assertion and release of control without inflicting pain or implying disapproval.

    Without the experience of riding, the idioms are just dead cliches and are ripe for mishearing and misunderstanding.

  15. ktschwarz said,

    November 25, 2022 @ 1:46 pm

    languagehat: yes, we should be very careful before concluding that the OED has a mistake in a recently revised etymology, but it does happen. Sometimes there are obvious typos: under yoke, n. you can currently read about a connection to "Sanskrit yoga act of joking", and under pekan (furry mammal) a language name is misspelled as "Passamaquody". And sometimes there are real errors in the spelling of foreign source words: the etymology of kayak used to give the source as "Greenlandic Inuit qajaq and its cognate Eastern Canadian Inuit qayaq", but only the -j- spelling is used in any Latin-alphabet Inuit orthography (this entry has recently been corrected). The etymology of posh, adj. still spells a suggested Persian source word as pōś, but ś is not used in any romanization of Persian; it should be pōš, and in fact it is pōš or poš in other entries.

  16. Andrew Usher said,

    November 25, 2022 @ 2:33 pm

    Even if regne and other spellings with a 'g' are attested, they don't deserve to be the main choice, and by modern standards would be called misspellings. The word has never had a 'g' etymologically or phonetically, and in modern French and English is never written with one. The OED would at least have made a poor choice had they intentionally decided to write that.

    And I doubt that anyone really uses 'free rein' and 'free reign' for 'two different nuances'. Nor can I see any reason to believe that they will start doing so.

    k_over_hbarc at yahoo dot com

  17. languagehat said,

    November 25, 2022 @ 2:56 pm

    maybe these dictionaries picked whatever was closest to the modern French spelling as a convenience to the reader

    I suspect so, and I admire the OED for choosing one that rubs the reader's nose in the fact that English does not derive from Modern French (except, of course, for modern borrowings). It seems to me from the way they set up the entry that they consider the form with -g- representative of the relevant earlier stage of the language, and the fact it isn't in the MF spelling is irrelevant, as is the modern reader's shocked reaction.

    but it does happen.

    As you know, I am well aware of that. As you also know, it is rare, and as I said, we should be very careful before assuming it has happened. In this case, I strongly doubt that is the case.

    Even if regne and other spellings with a 'g' are attested, they don't deserve to be the main choice, and by modern standards would be called misspellings.

    This is exactly the kind of presentist response I deprecate and am glad they avoided.

  18. J.W. Brewer said,

    November 25, 2022 @ 7:48 pm

    Even though direct personal experience of reins may have waned as the percentage of the population with personal experience on horseback has declined, it also seems like direct personal experience of reigns has also waned as the percentage of the population with personal experience of living in a monarchy rather than a place with elected (however nominally) rulers has declined.*

    It is almost 50 years now since The Who played with the separate potential ambiguity between "reign" and "rain," in the concluding track of the Quadrophenia album.

    *I am American but spent three years of my childhood living in Japan during the reign of the Shōwa Emperor (as he is posthumously known), so I guess I have *some* personal experience, including that of years being numbered in a regnal calculation rather than numbered Anno Domini.

  19. Jonathan Smith said,

    November 25, 2022 @ 8:46 pm

    Well, re: Chinese at least, OED is to be commended for being far better than any Chinese dictionary I know of in being interested *at all* in etymology + the regional languages, but in my experience is far from infallible; how could it be given scope. I wouldn't cite it uncritically and admit I wouldn't hesitate to describe many spellings as "apparent mistakes", gulp… in terms of what is online now, a quick sample re: "Cantonese" and "Hokkien" is

    chop suey has the misspelling "shap suì", hyson has the unsystematic or misspelled "hei-ch'un" (Yale hēi chēun), many loaned headwords don’t have Cant. info so kumquat says "[t]he Cantonese dialectal form of the Chinese [?!] name kin kü" but no Cant. (Yale) gām gwāt, Sze Yap no sei yāp, kwai-lo no gwái lóu (and no etymology); mee has the misspelling mĩ (should be mī), under kiasu Hok. kiasu is misspelled unless "Singapore Hokkien" were specified… and so on.

    though most frustrating is the frequently-used form "Etymon: Chinese 'X'" where "Chinese" means "any Chinese language" :D

  20. ktschwarz said,

    November 26, 2022 @ 5:02 am

    I agree with languagehat about avoiding presentism; I just don't find it "quite clear" whether the OED chose regne at random (since there's no perfect solution), or by mistake as Mark Liberman thought, or because they thought it was representative. What evidence is there, other than "errors are rare"? If it was representative, why does the Anglo-Norman Dictionary show no -gn- spellings in 14 citations, but 7 with -sn-? I know citations are chosen to be representative of meanings, not spellings, but if the -gn- spelling was typical, wouldn't you expect some to show up?

    It's also possible that the choice wasn't wrong given the information available at the time, but has been superseded: the Anglo-Norman Dictionary's entry has been revised more recently than the OED's, so it will be more complete.

  21. ktschwarz said,

    November 26, 2022 @ 6:01 am

    Jonathan Smith: some of the OED entries you mention are old (hyson is from 1899) and can't be expected to meet today's standards, but yes, even in revised entries OED is a lot weaker on Chinese loanwords than it is on Anglo-Norman.

    As you probably know, the American Heritage Dictionary is much better than the OED on the origins of Chinese loanwords, but it covers a smaller range of them.

  22. Taylor, Philip said,

    November 26, 2022 @ 6:17 am

    I have, from time to time, reported apparent errors in the OED to its lexicographic/editorial team, but such reports are rarely acknowledged nor do they necessarily result in an emendation being made. A few examples will suffice —

    [1] The current OED has an "Origin" field which was, I believe, not present in most if not all previous editions. However, under the entry for "Word", one finds: "Origin: A word inherited from Germanic.". MIght it not have been better to avoid the circular reference by simply writing "Origin: Inherited from Germanic." ?

    [2] At, the etymology concludes with the words "see earlier French cathédrale." However, as far as I have been able to ascertain, "cathédrale" does not appear as a headword within the online OED, and it is therefore unclear as to where or to what the reader is being referred with this "see earlier" field. Furthermore, entering "cathédrale" into the Quick Search box and invoking Search results in the rather odd message "No dictionary entries found for ‘cathédrale’." — why is "cathédrale" being re-mapped to "cathédrale" as a part of the search process ?

    [3] May I respectfully query the IPA transcription of the Irish pronunciation of the phrase deoch an doris, n. [1] in the Online OED ? You have /ˌdʒʌx (ə)n ˈðvrəʃ/, but the third element does not appear to correspond with what your oral informant says at all. For /ˈðvrəʃ/, I hear /ˈdɔrəʃ/, or perhaps /ˈdɒrəʃ/, but nothing like /ðvrəʃ/. I would be most grateful for your comments on this apparent discrepancy.

    [1] "deoch an doris, n.". OED Online. June 2016. Oxford University Press. (accessed August 31, 2016).

    [4] A search for "spraint" leads the reader only to "sprent, v." (187687), yet there exists within the dictionary the headword "spraints" (187589, with plural concord) which admits of the singular usage. Should not a search for "spraint" lead the reader first
    to "spraints", and only secondarily to "sprent" ?

  23. Victor Mair said,

    November 26, 2022 @ 9:42 am

    "The American Heritage Dictionary of the English Language, 5th edition" (11/14/12)

    As I have upon many occasions stated, AHD is my favorite dictionary of the English language, though I still treasure the Merriam-Webster I grew up with in grade school, junior high school, and high school (the terms that we used for them in Osnaburg Township, Stark County, Ohio).

    Some of the reasons:

    IE and Semitic roots

    good etymologies across the board

    esp. good etymologies for East Asian languages

    succinct, clear definitions

    very good coverage of terms from historical, recent, and contemporary periods

    nice pictures!

    etc., etc., etc.

  24. Philip Anderson said,

    November 26, 2022 @ 2:15 pm

    @Philip Taylor
    Your reference to the Irish pronunciation of a (Scottish) Gaelic phrase confused me, since I took this to mean the pronunciation in Irish, until I followed the link and realised you meant the pronunciation in Ireland (that is in Irish English). Welsh has the advantage of having different adjectives for language and country, Saesneg (English language) vs Seisnig (English in other contexts).

  25. Taylor, Philip said,

    November 26, 2022 @ 3:42 pm

    Yes, sorry, the OED [now] reads "Irish English", not just "Irish" at this point. Whether it read "Irish English" when I first pointed out this inconsistency I regret that I cannot say.

RSS feed for comments on this post