It occurred to me to wonder why the word nephew, which comes from French neveu, is written with –ph-, so I looked it up in the OED, which (though the entry was updated in September 2003) is uncharacteristically unhelpful — after listing over a hundred variant spellings (including neveaw, newowe, neuo, nephwoy, and nevvey) gives the following etymology:

< Anglo-Norman nevou, neveu, nevew, nevu, newu and Old French, Middle French neveu (also in Old French as nevou, nevo, nevu, nepveu, etc.; French neveu), originally the oblique case of Old French nies, niers (c1100; 2nd half of the 12th cent. in sense ‘grandson’, c1500 as nepveux (plural) in sense ‘descendants’) < classical Latin nepōt-, nepōs, grandson, descendant, a prodigal (see sense 2c), a secondary shoot (see sense 5), in post-classical Latin also nephew (4th cent.), niece (13th cent.), cognate with neve n.1. Compare also nepote n.

Which has some interesting information (I didn’t know about the OF nominative nies, niers, or the native Germanic form neve, parallel to German Neffe), but doesn’t address the spelling issue. Spellings with –p– go back way earlier than I would have guessed (?1456 Duke of York in Paston Lett. & Papers (2004) II. 100 “To take possession and saisine, in the name and to þe vse of our ful worshipful nepueu th’Erl of Warrewic”); I realize it must be Latinizing, after nepōs, but it seems very odd — we write river, not riper or ripher, even though again French –v– is from Latin –p-. Does anybody know anything more about the history of this spelling change, and the concomitant spelling pronunciation with /f/ which is universal in the US and exists in the UK as well? Come to think of it, that’s another thing I’m curious about — I’ve long been aware of the UK pronunciation /ˈnɛvjuː/, but for some reason I had the impression it was antiquated; the OED, however, implies it’s the more common one:
Pronunciation: Brit. /ˈnɛvjuː/, /ˈnɛfjuː/, U.S. /ˈnɛfju/

So I’ll ask you Brits: do you say it with /v/ or /f/, and do you think of the former as standard or old-fashioned?


  1. John Wells included “nephew” in his surveys of British English in the 1990s (reported here). “Nephew pronounced with /f/ has today clearly overtaken the older /v/ pronunciation, leaving it behind with only 21%. … An overwhelming majority of over 90% of those born after 1948 adhered to the spelling pronunciation /’nefju:/.”

  2. Consider your own name, which can be spelled with either a v or a ph, but is always pronounced /v/ due to the same sound change. The hypocoristic form nevvy landed in the 18C and is spelled as pronounced, even to the “vv” which normally doesn’t occur in English words (chivvy, civvies, divvy, flivver, navvy, revving, savvy, skivvies all have similar offbeat or unknown origins), as it would be read as “w”.

  3. An idea that I’ve got after about 1 minute of thinking. Maybe unusually strong influence of Latin -p- is because nephew should often have been written in wills?

  4. Paul (other Paul) says:

    Born before before 1948 in Australia, long UK-based, I usually say “nevu”, but on reflection, I find that “nefu” sounds fine too, and I may even use them interchangeably.

  5. Brit born in the 80s here. I can’t remember ever hearing the “v” pronunciation of nephew.

  6. Chris McG says:

    I only hear the pronunciation with /v/ from people who voice all fricatives anyway.

  7. Bathrobe says:

    I use /f/. Only upon reading your post did it occur to me that the pronunciation with /v/ sounds familiar. I might even have used it occasionally. I’d never actually thought about it.

  8. Like Bathrobe above, I’ve never paid close attention but I realize now that I pronounce both ways. /v/ usually occurs when I’m tired or rushed, /f/ when I’m more alert or saying just the word. After reading John Cowan’s comment above, I also realized I pronounce Steven and Stephen differently.

    For what it’s worth, I have a weird mish-mash accent. Born in America’s upper Midwest, raised in the Deep South in a close community of transplants, with close family ties in Britain and New England, and extensive French and Latin lessons.

  9. Huh, so it looks like my original impression was correct and the /v/ pronunciation is antiquated. I wonder why the OED lists it first in a 2003 revision?

    Consider your own name, which can be spelled with either a v or a ph, but is always pronounced /v/ due to the same sound change.

    I’m not following you. How can /sti:vn/ reflect the same sound change as /nefju(:)/? Also, people often do pronounce the -ph- spelling with /f/; see Joy’s comment above.

  10. Charles Perry says:

    Surprised to find the OED gives /fɑiəl/ as the pronunciation of phial. All my life I have assumed that it was pronounced the same as vial.

  11. Jim Parish says:

    Well, one of Randall Garrett’s Lord Darcy stories was titled “The Ipswich Phial”….

  12. Of course talking about “the same sound change” was nothing but a brain fart on my part. My intended point was that since Stephen is a really old borrowing, it came in before there was a phoneme /v/ in English. At that time [v] was solely due to intervocalic voicing of /f/, so Stephen and a few other Latin words of Greek origin got the English used to occasionally writing ph for /f/, while automatically voicing it between vowels just like native /f/.

    So when the Old French word for ‘grandson’ landed in Middle English, it must have seemed natural to write it indifferently nepheu, neveu, nefeu (and variants), the pronunciation in all cases being [v]. Then the arrival of foreign words with initial [v] led to the creation of the new phoneme /v/ (vane, vat, vixenare native, but come from dialects with pervasive initial voicing). The Great Spelling Shakeout eventually landed us with the spelling nephew, now with /v/ as well as [v]. Finally, spelling pronunciation did its evil work and gave us /f/, first in the U.S. and then in the UK, and now we hear that its baleful influence is starting to corrupt Stephen as well.

    The story of phial/vial must be much the same, except that /f/ is the original sound in English here, and vial must also reflect a dialect with initial voicing. In Randall Garrett’s Lord Darcy series, in which the Anglo-French Empire grew under the descendants of Richard Lionheart instead of breaking up (and magic is effective), there is a story called “The Ipswich Phial”, a parody of the title of Len Deighton’s spy novel The IPCRESS File (made into a Michael Caine movie).

  13. David Eddyshaw says:

    I (who am antiquated) say it with a v (as in Weller.)
    My wife, who is much less antiquated, says it with f and v.
    My children say it with f; mind you, they don’t yet have cause to use the word as much as I do.

  14. Jay sterling says:

    East End of London born. & bred. When reading your piece I’m thinking that I say it with a ‘v’ but I start questioning myself. I think it’s one of those words where I change the pronounciaton according to the social register of the other paticipant(s) in the conservation: so ‘v ‘ with family and ‘f ‘ when using register of a more standard English to fit in with speakers of other English varieties or RP.

  15. Fascinating — I’m glad I asked!

  16. marie-lucie says:

    The spelling of (older) French nepveu besides Standard neveu must be due to the same reasoning that has Lefebvre as an alternate spelling for Lefèvre: the pronunciation has the fricative v but the Latin original has a stop (nepos, faber), which a latinizing spelling reintegrates into the French word (just as Latin b was integrated into English debt because of Latin debita).

    For French, latinizing spellings are typical of the Renaissance (eg Rabelais’ original text). Many such spellings were simplified later, but some of them have remained, eg doigt ‘finger’ (Latin digitus) and vingt ‘twenty’ (Latin viginti), perhaps to differentiate these words from homophones (eg dois, doit ‘must, owe(s)’ or vin ‘wine’).

  17. marie-lucie says:

    As for nephew, I first learned this word in the fifties, with [f].

  18. Bathrobe says:

    spelling pronunciation did its evil work its baleful influence

    I gather you don’t like spelling pronunciations.

  19. David Marjanović says:

    I was taught to say it with [v], probably even without the alternative being mentioned. But of course L2 teaching is prone to archaisms.

    or vin ‘wine’

    More likely vint “came”; the t resurfaces in such common contexts as vingt-et-un.

  20. David Marjanović says:

    I gather you don’t like spelling pronunciations.

    Tell us what you really feel 🙂

  21. The worst spelling pronunciation perversion I’ve encountered is “HUSser” for “hussar” (should be “huzZAR”). I heard someone say it that way and I politely pointed out that it’s “huzZAR” and then we drew our phones like Billy the Kid and googled it and my nemesis pointed to one of those sites like or whatever where “HUSser” was given as an acceptable pronunciation.

    I told him that that only proved the internet was a wretched hive of scum and villainy and that he should trust me on this one.

  22. M-W, an impeccably American dictionary, gives only “huzZAR,” so I think your interlocutor didn’t have a leg to stand on. One might, in theory or in desperation, claim “HUSser” to be an American form that M-W somehow ignores, but we Yanks don’t have hussars, so that’s right out. Similarly, I say “can-TOON-ment” for the relevant sense of cantonment, since Brits have them and we don’t.

  23. And se-CON-ded, for the same reason.

  24. marie-lucie says:

    David: or vin ‘wine’ — More likely vint “came”; the t resurfaces in such common contexts as vingt-et-un.

    You must be right!

  25. CanTOONment is only the second pronunciation in either the OED or the ODO; the first uses the LOT vowel, and so I’d say can-TAHN-ment, using my own LOT vowel. I’d do the same with boffin, full stop, letterbox, and other LOT Briticisms.

    Wikipedia says we do have cantonments, anyway: “In United States military parlance, a cantonment is an essentially permanent residential (i.e. barracks) section of a fort or other military installation such as Fort Hood.”

  26. Athel Cornish-Bowden says:

    A curiously difficult question to answer, and I’m not really sure how I say it, but I think it’s with /f/ or perhaps a very weakly voiced /f/. I think I’ve always said it that way (born 1940s).

  27. A survey of myself reported 90% -f- 10% -v-

  28. Born 1957 in Abingdon, Berks: I say nephew with a /v/. The pronunciation with an /f/ doesn’t sound too bad though, certainly not as bad as pronouncing the t in often. Fighting a losing battle for forrid (forehead).

  29. CanTOONment is only the second pronunciation

    But it is the standard pronunciation in the sense “quartering of soldiers”; check your Daniel Jones or watch a British war movie.

    Wikipedia says we do have cantonments, anyway

    I stand corrected!

  30. I haven’t thought of “seconded” as specifically British. I’ve heard people here in Washington, DC, use it in referring to people being temporarily transferred from one government agency to another.

  31. Dammit, are all my UK/US shibboleths going to crumble to dust? So how do people in Washington, DC, say it: se-CON-ded or (as I would expect from Americans) SEC-onded? If the former, they must have picked it up from the Brits.

  32. Sec-CON-ded in Toronto.

  33. It’s “se-CON-ded” in DC too.

  34. But I don’t think these same people would talk about “taking” a decision or describe something as being a “one-off” or use other Britishisms.

  35. Speaking of “nephew”, does anyone use the word “nibling” seriously?

  36. Never heard of it.

  37. Jim Parish says:

    I use “nibling” from time to time; I have several of them, not all of the same gender.

  38. Aha, nibling: “Coined by linguist Samuel E. Martin in 1951 from nephew/niece by analogy with sibling.” I guess if it were going to catch on, it would have done so by now.

  39. Rodger C says:

    In the U.S. Army, 1969-71, the only pronunciation I ever heard for “cantonment” was “c’n-TONE-ment.”

  40. Rodger C says:

    And as for online dictionaries, I had occasion to look up “eschew” recently and was flummoxed to find that the first pronunciation given was “ee-shoo.”

  41. And now I too am flummoxed.

  42. It seems like everybody has a different pronunciation of “eschew.” I just checked an online dictionary, and it didn’t have that pronunciation, but it did list six others, half of which do not sound right to my ear.

  43. David Marjanović says:

    Fighting a losing battle for forrid (forehead).

    I’m still surprised this pronunciation isn’t simply considered part & parcel of H-dropping.

  44. I just checked an online dictionary, and it didn’t have that pronunciation, but it did list six others, half of which do not sound right to my ear.

    This is why I tend not to bother with online(-only) dictionaries. The AHD has only es-CHEW, which (conveniently) is my own pronunciation.

  45. David, it’s the stressed vowel of forehead that matters: traditionally LOT, but by spelling pronunciation FORCE, as in (be)fore. I say /fɑrəd/ myself.

  46. Bathrobe says:

    The spelling ‘forrid’ obviously based on ‘torrid’ 🙂

    When I was a kid, ‘forrid’ was the norm and ‘forehead’ was used by people who didn’t know better. Now it seems that ‘forrid’ is on its last legs.

  47. Chris McG says:

    I say ‘forrid’. But not ‘weskit’.

  48. AJP 'Des' Perado says:

    nibling: “Coined by linguist Samuel E. Martin in 1951 from nephew/niece by analogy with sibling.” I guess if it were going to catch on, it would have done so by now.

    Thanks for this. I know a woman in London who uses this word all the time for her siblings’ children. I’d always thought it was her own family’s coinage.

    I only say /f/ in nephew.

    With hussar the two pronunciations I’ve heard are həZAR, the more common one that I use, and hʊZAR with the u like the oo in foot. The latter was used by a man I knew who actually had been one, but he was a dreadful snob so it may be a bit frou-frou.

    I used to say forrid but now I say fore-head.

  49. AJP Perado says:

    Come to think, sibling itself is not used very often compared to its equivalent in other languages (søsken, geschwister). In English, I’d be more likely to say ‘brothers & sisters’.

  50. I was introduced to “nibling” by a friend some years ago; so far I think I have only managed to use it when talking to that same friend.

  51. @Christ McG: How about saucep’n?

  52. In English, I’d be more likely to say ‘brothers & sisters’.

    Me too.

  53. (Oh sorry, I didn’t mean to deify Chris.)

  54. Sibling was an Old English word meaning ‘member of the sib or kinship group’. Both words were revived around the beginning of the 20C by geneticists and anthropologists, the former in a narrower sense than before. The first person on record to use sibling ‘Geschwister’ was Karl Pearson of Pearson’s r, the correlation coefficient.

  55. (saved too soon)

    Consequently, it is an artificial and technical word, not much used in ordinary communication, as noted above.

  56. marie-lucie says:

    In French, it’s frères et soeurs, but in Spanish just hermanos, meaning either “brothers’ or ‘brothers and sisters’.

  57. Russian doesn’t even have a technical term as far as I know (though there’s дети одних родителей ‘children of one set of parents’). They say братья и сёстры [brat’ya i syostry] ‘brothers and sisters’ (as in the title of the 1958 Abramov novel).

  58. Rodger C says:

    Anyone here whose house still has clabberds?

  59. And do you have clabbered milk in your clapboard house?

  60. David Marjanović says:

    David, it’s the stressed vowel of forehead that matters: traditionally LOT


    Did it escape the NORTH-FORCE merger somehow?

  61. No, I think it shifted from /for+hɛd/ to /fo+rɛd/, followed by irregular shortening. It’s of OE age.

  62. Doesn’t it need to rhyme with “horrid” so that the rhyme about the little girl who had a little curl works? Or do you have /ɑ/ in “horrid” too?

  63. I do, yes. Some Americans have NORTH=FORCE, those who have changed all ambisyllabic /r/ to belong to the previous syllable only, but I’m not one of them. On the other hand, CLOTH=THOUGHT for Eastern Americans like me.

  64. What? HuZAR? Who the devil spelled it in English with two esses, then? That’s just mean.

  65. Scissors.

  66. Bathrobe says:


    And “chassis”, pronounced “shazzy” (but only in Australia).

  67. dissolve, possess, Missouri

  68. In Ireland, the -lace is “necklace” is often unreduced, as in “shoelace”. I still dislike the “neckless” pronunciation.

  69. ss can even be [ʒ], for some speakers, in “fission”.

  70. and [ʒ] for all(?) speakers in scission, re~, and ab~

  71. Similarly ti is [ʒ] in “equation”.

  72. I’ve noticed that I have two pronunciations for “equation” – when it means the act of equating, I use [ʃ], but in the more fully nominal senses, like in math or chemistry, I use [ʒ].

  73. @Lazar: I do that too.

  74. David Marjanović says:

    What? HuZAR? Who the devil spelled it in English with two esses, then? That’s just mean.

    And it’s not etymological or anything either. It’s completely random.

  75. It was borrowed from (Northern) German, where it’s spelled Husar and pronounced with [z]. So that’s all explicable, it’s only the doubled s that’s random. Hussar is an etymological doublet of corsair, but with a borrowing path of Italian > Serbo-Croat > Hungarian > German > English instead of Italian > Occitan > French > English. Both are ultimately < Latin cursare ‘journey (v.)’ < currere ‘run (v.)’. So from traveling we go to traveling for plunder, and then on the hussar side to being a light Hungarian horseman, doubtless notorious for plundering.

  76. David Marjanović says:

    it’s only the doubled s that’s random

    That’s what I mean.

Speak Your Mind