The prefix en- (from French) has a variant spelling em-. (This is also associated, although I believe imperfectly, with the use of the sound /m/ in the pronunciation of the prefix.) Although the general pattern of its use seems fairly clear (it shows up before labial consonants such as P and B), I've been finding it hard to give an exact description of its occurrence.

The Oxford English Dictionary says

In modern orthography and pronunciation en- becomes em- before b and p, and occasionally before m.

This rule was not fully established in spelling before the 17th cent.; in Middle English, as in Old French and Old Spanish, enb-, enp- are more frequent than emb-, emp-, though the latter may perhaps represent what was the actual pronunciation.

One thing that I find interesting is that even though this prefix is related to and often has been interchanged with the Latin-derived in- prefix, the rules for which forms to use are not entirely the same for both of these prefixes. The in- prefix regularly assimilates in English to ir- before R (as in irradiate) and to il- before L (as in illuminate), while the en- prefix does not (as in enrich, enlarge). And the rules for these prefixes are different from the rule for the native English prefix un-, which is regularly invariant, being spelled with N regardless of the identity of the following sound or letter.

One thing that isn't covered in the OED entry is possible exceptions or "edge cases". Can anyone give a more detailed description of the distribution of the two forms of this prefix? I'm hoping to have a comprehensive description of the use of en- vs. em- in present-day English; any further information about the use of these two spellings in earlier stages of English, and how the spelling of the prefix developed/changed over time, would also be welcome.

For this question, I'm only interested in the verb prefix, not the etymologically distinct negative prefix that is found in a few words like enmity.

  • Is the the ir- before 'R' words and il- before 'L' words an English thing. I found Latin "illustrare" (to light up). I couldn't find an irr- example in Latin, but then again the only irr- in English I found are "irradiate" and maybe "irrupt". All the others have irr- to make negative, "irreducible", "irrelevant". As you said, the exceptions seem to be few, so is this question basically asking why words like "enmesh", "enplane", "disemvowel" don't seem to follow the rule? If I had to guess about "disemvowel", I'd guess it's because it sounds like "disembowel". For the other I don't have a clue.
    – Zebrafish
    Commented Oct 26, 2018 at 5:22
  • @Zebrafish: My understanding is that the assimilated forms ir- and il- developed in Latin, but during the Classical period their use was optional. Here is a Latin SE question about -NL- vs -LL-. So spellings like inrito and inlumino were used in Latin, as well as spellings like irrito and illumino.
    – herisson
    Commented Oct 26, 2018 at 5:26
  • "... a more detailed description of the distribution of the two forms of this prefix? I'm hoping to have a comprehensive description ..." -- Too broad, too long, voting to close.
    – Kris
    Commented Oct 26, 2018 at 6:08
  • Let's just go by the dictionary definitions of more detailed description and comprehensive description. It was nothing about my opinion at all.
    – Kris
    Commented Oct 26, 2018 at 6:18
  • My gut instinct is that the likelihood of assimilation decreases as the transparency of the en- as a productive, English suffix increases. Enplane and enmesh ‘feel’ much more like productively formed prefixed verbs than envision or enable do. Assimilation is not usually a productive process in English prefixes, so newly formed en- verbs would probably not be expected to assimilate (or they’d be less expected to, anyway). Commented Oct 26, 2018 at 20:38

2 Answers 2


Here are the "complications" or exceptions that I know of so far. This is not a long list, so it certainly seems feasible to discuss all of the exceptions in the space of one answer post on this site.

  • en- before M: the OED quote in the question hints at the use of en- before M. The most common verb spelled with enm- seems to be enmesh, which the Google Ngram Viewer shows as being many times more common than the variant spelling emmesh. I don't know of any common prefixed verb that is usually spelled with emm-. I would say that the use of em- before M actually seems to be negligible in present-day English.

    The word enmarble seems to not be common enough to show up in the Google Ngram Viewer, but dictionaries indicate that en- and not em- is preferred here in present-day English. I'm not sure whether enmew, emmew or immew is more common, but this verb seems to be obsolete in present-day English, mainly occurring in modern sources only in the context of discussions of older texts.

  • em- before V: This occurs in the specific context of the word disemvowel, which is clearly built to resemble disembowel. It seems to be a special case: all other words with this prefix before V seem to use the spelling with N, as in envision and envenom.

  • en- before P: this doesn't seem to have entirely died out yet. The Google Ngram Viewer shows approximately equal usage of emplane and enplane, and many dictionaries such as the American Heritage Dictionary, Merriam-Webster, Oxford Dictionaries (US) and Collins have the enplane spelling as the main headword.


Etymonline reference on this topic may help fix some points:

En- (1):

word-forming element meaning "in; into," from French and Old French en-, from Latin in- "in, into" (from PIE root *en "in"). Typically assimilated to em- before -p-, -b-, -m-, -l-, and -r-. Latin in- became en- in French, Spanish, Portuguese, but remained in- in Italian.

Also used with native and imported elements to form verbs from nouns and adjectives, with a sense "put in or on" (encircle), also "cause to be, make into" (endear), and used as an intensive (enclose). Spelling variants in French that were brought over into Middle English account for parallels such as ensure/insure, and most en- words in English had at one time or another a variant in in-, and vice versa.

en- (2):

word-forming element meaning "near, at, in, on, within," from Greek en "in," cognate with Latin in (from PIE root *en "in"), and thus with en- (1). Typically assimilated em- before -p-, -b-, -m-, -l-, and -r-.


word-forming element meaning "put in or into, bring to a certain state," sometimes intensive, from French assimilation of en- "in, into" (see en- (1)) to following labial stop (-b-, -p-, and often -m-), or from the same development in later Latin in- (to im-).

"This rule was not fully established in spelling before the 17th c." [OED], but it is likely the pronunciation shift was in Old French and Middle English and spelling was slow to conform. Also a living prefix in English used to form verbs from adjectives and nouns (embitter, embody). In words such as emancipate, emerge, emit, emotion the e- is a reduced form of Latin ex- (see ex-) before -m-.

  • I think there may be a mistake in the Etymonline entries--neither L nor R is labial, so I don't see how en- could assimilate to em- before L or R (and it does not in fact appear to do this). The en- (1) entry seems to actually say "Typically assimilated before -p-, -b-, -m-, -l-, and -r-", which looks like a description of the rule for the prefix in-.
    – herisson
    Commented Oct 26, 2018 at 11:28

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.