According to (online) Merriam-Webster, "may" has the following two distinct definitions, among others

1 b: have permission to

4: SHALL, MUST —used in law where the sense, purpose, or policy requires this interpretation

Oxford dictionary does not seem to have the definition of may = shall, must.

What is the history of may being used to mean "must"?

  • 1
    The mother of Grok.
    – Hot Licks
    Commented Mar 26, 2020 at 16:03
  • 3
    Yes, which 'Oxford dictionary'? Commented Mar 26, 2020 at 16:09
  • 4
    The full (subscription-only) defn I-6b OED says In the interpretation of statutes, it has sometimes been ruled that may is to be understood as equivalent to shall or must. But their first recorded instance of this sense is 1715, which is nearly five centuries later than immediately-preceding defn I-6a Expressing permission or sanction: be allowed (to do something) by authority, law, rule, morality, reason, etc. Commented Mar 26, 2020 at 16:25
  • 3
    What's the only word that means mandatory? Here's what law and policy say about "shall, will, may, and must." - faa.gov/about/initiatives/plain_language/articles/mandatory
    – user 66974
    Commented Mar 26, 2020 at 16:37
  • 1
    Interesting, I was not aware of such (ab)use of these words. That finally explains to me why RFCs always explicitly reference tools.ietf.org/html/rfc2119 for definitions of words I would have considered self-explanatory ... Commented Mar 27, 2020 at 18:55

4 Answers 4


The OED dates this sense of the word back to 1715:

Where a Statute directs the doing of a Thing for the sake of Justice or the publick Good, the Word may is the same as the Word shall; thus 23 H.6. says, the Sheriff may take Bail; this is construed, he shall; for he is compellable so to do.

Reports of cases adjudg'd in the Court of King's bench

  • 2
    Good find! Such laws that have this use of may are usually ones where several conditions are required for something being granted/issued. The laws often start with "X will be issued by <an official> if the following conditions are met..." and end "if the above conditions are met then X may be granted/issued."
    – Greybeard
    Commented Mar 26, 2020 at 16:30

As a native UK speaker, I'd simply say that Merriam Webster is wrong.

The 1715 example just seems to be noting that when a statute says 'may', that - in the particular circumstances relating to the Sheriff - is the same as 'shall', because "he is compellable so to do".

If he wasn't compellable to do so, 'may' would not be the same as 'shall' or 'must'.

Even legally speaking, if someone tells me I 'may' use their swimming pool, that does not mean I 'must' or 'shall' do so - because I'm not under any obligation to go swimming whenever I can.

  • But this isn't asking about the everyday usage. Is OED also wrong (FumbleFingers' comment above)? Commented Mar 29, 2020 at 11:22
  • 1
    OED is not wrong insofar as it says that it has sometimes been ruled that a document containing may should be interpreted as if it had shall or must instead. A court would, however, so rule only rarely and hesitantly, when looking at the document as a whole, justice so requires. In doing so, the court would be aware that it is departing from the normal understanding of the word, in order to produce a fair result. It would be wrong to take the OED entry as implying that anybody deliberately uses the word that way in drafting legal documents.
    – jsw29
    Commented Mar 29, 2020 at 15:55

I am not a lawyer, and as such wont have anything to say about how may for shall struck me as odd when reading it in this context.

But I am a native German speaker and we distinguish the verb mögen in indicative ich mag (I like, I may want to), subjunctive (Konjunktiv I) ich möge and, I guess you would also call it subjunctive (Konjunktiv II) ich möchte. The precise details of these modi are immaterial, here, but it's interesting that möchte (ca. "might") has come to be understood as a separate stem, not anymore related to mögen, but without a form for, or without a practical use for, the infinitive, probably for thousands of years now. It's common consensus that the stem was the same though, which is what I will go with, though I'm incubating doubts.

The point being, "Er möge Kaution annehmen" has the obligatory sense as in "He may take bail". This does sound a bit old fashioned. The apparently sarcastic tone is used colloquially as well, Du möchtest den Müll raus bringen "you may bring out the trash", perhaps as a variant of an archaic, high council like mögest, but built from the preterite mochte, thus implying precedent. This is showing how the different moods and aspects are deeply intertwined.

Proto-Indo-European as reconstructed in the articles of wikipedia had subjunctive and optative mood. One of these was built with an *-iH, *-ey suffix, and then some, notwithstanding Schwebelaut. Before or in Middle High German, terminal -i caused Umlautung in the nucleus (Sturm, stürmisch -- storm, stormy), depending on several factors.

Clearly, in English this ending would be subsumed by y when the velar g, still visible in Old English magan, became however lenited to may. Henceforth, the grammatical difference was lost as in many other cases of erroding English morphology. Whether the meaning was nevertheless kept in good memory, as in you may go now, or reinvented by statute in the 18th c. as other opinions suggest, I don't know.

The precise interpretation of the different modi and aspects is not always clear. If you said the Sherrif can take bail, this might be construed as a statement of fact, irregardless of the actuality, whereas Germann kann sein chiefly means "maybe" in colloquial speech, largely indistinct from könnte sein "could be", though in legalese, as far as I'm aware, the former leaves the decision to the discretion of the administrator subject only to moral obligations, whereas the latter would strictly imply modus irrealis and as such cannot confer obligation. Similarly ich will "I want", ich werde "I will, I become" and English you will! show's a striking difference. mag sein is literally "may be", too, but das möchte wohl sein (doesn't translate) again plays on the preterite, ca. "that's well expected, anything else would be far from acceptable" (as if that must be, but where to stick the well?). I mean to say, there were inherent uncertainties that lead to the errosion of this grammatical system. The optative is perhaps still seen in I wanna, while the usual example of optative aspect is lemme (or rather let's), but as a grammatical function it was already lost in almost all descendants from Indo-European long ago.

I want to say this has a desiderative aspect, which, when coming from an authority, implies obligation to the subordinate, when success of the order is in their might at all (in der Macht stehen; also cp. Möglichkeit "possibility", möglich "maybe", possibly conflated with machbar "make-able, feasable"). But the difference between "I want to go" or "Don't you want to go now!?!" is quite obvious, so I don't quite know what to call either of both aspects.

  • 1
    It's very interesting, but all in the 'comment' bracket: not directly answering the actual question. Which FumbleFingers' comment does. Commented Mar 29, 2020 at 11:25
  • @EdwinAshworth there is an answer in paragraphs 4 and five. mood suffix *-iH being subsumed in sound change under indicative, without however completely loosing the connotation. Did you miss that in the wall of text, criticizing my writing as borring, or do you consider that not an answer, perhaps wrong even? FF explains jack, merely hints at a possibly well founded answer in the considerations of some rulings; the dating from the OED is unreliable anyhow, as SvenYargs frequently shows.
    – vectory
    Commented Mar 29, 2020 at 11:39
  • And, if you imply that because I said so (of God, the wigs, the caretaker) is a valid answer, that's highly fallacious. If everyone, or a significant share of the population (the wigs, the caretaker, sarcastic mothers) said so, that remains to be shown, I agree. But I don't accept de novo ex machina (?) as an answer.
    – vectory
    Commented Mar 29, 2020 at 11:43
  • Outlining all the German may be circumstantial and mostly irrelevant, but it's a proof of work, meaning it allows those who are knowledgable enough to point out mistakes in the argument. You are just being salty because you cannot deny nor approve it :P
    – vectory
    Commented Mar 29, 2020 at 11:51
  • 'Clearly, in English this ending would be subsumed by y when the velar *g became however lenited.' But how do we know that the loaning didn't occur with immediate semantic change? 'Henceforth, the grammatical difference was lost as in many other cases of erroding English morphology. Whether the meaning was nevertheless kept in good memory, as in you may go now, or reinvented by statute in the 18th c. as other opinions suggest, I don't know.' This just echoes / comments on existing answers. OED is considered pretty good on ELU, and way better than nothing. Commented Mar 29, 2020 at 14:00

Loved reading all the comments 💙🙌 I was told back in 2011 that legalese was a language created by ancient scholars to trick/manipulate/con the average layman out of his hard earned very small and newish industrial revolution wage packet.

I was told that the language was very similar to English but with a handful of words with slight changes to their meaning, one of those words was MUST. I was told that in legalese MUST is equal to MAY, and therefore when you receive a bailiffs letter stating that you MUST pay £X by a certain date, what it’s really saying is you MAY pay £X on that particular date in question, the letters in legalese offers the layman a choice he didn’t know he had…

I have done as mentioned to bailiffs in the past, simply pointed out what I knew and told them I have decided to decline and believe me or not it works, or put it this way they’ve never been back! So is my train of thought correct or have I just got lucky? Or am I chatting 💩


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.