Oxford University Press's
Academic Insights for the Thinking World

The blunt edge of “knife”

The word knife came up in one of the recent comments. I have spent so much time discussing sharp objects (adz, ax, and sword) that one more will fit in quite naturally. The word that interests us today turned up in late Old English (cnīf) and is usually believed to be a borrowing of Old Norse knífr (both ī and í designate a long vowel, as in Modern Engl. knee). The word occurred in Old Frisian and Old Low (that is, northern) German in nearly identical or slightly different forms (for example, gnippe). Today, the cognates of knife are Icelandic hnífur (along with similar forms in all the modern Scandinavian languages) and Dutch knijf. We have seen that cutting tools may denote both house utensils and weapons. The same holds for knife. The Old Icelandic idiom deila kníf ok kjötstykki “to share knife and meat (pieces)” implies that knives were used for cutting food, but knífr also denoted a dirk, “such as the ancients wore fastened to their belts; and so a knife with a belt is frequently mentioned as a gift; the handles of these knives or dirks were neatly carved of walrus’ tusks” (this is a quotation from the great dictionary of Old Icelandic by Cleasby-Vigfusson). Slavic nozh- “knife” was sometimes glossed as gladius (“sword”) in old texts. Old Icelandic knífr also meant “penis,” obviously, a secondary meaning.

This is the object they called knífr. From the Swedish History Museum. CC by 2.0 via Wikimedia Commons.

As could be expected, the origin of knife remains a matter of speculation. Lithuanian knêžas “knife” looks like Slavic nozh-, but with the enigmatic initial k-, as in knife. True to our rule not to explain obscurum per obscurius, that is, one word of undiscovered origin by referring to another “dark” word, we will, for the time being, remain on Germanic soil, though the temptation to find some cognates in Sanskrit, Greek, Latin, Celtic, and so forth is always great.

At first sight, the hypothesis I’ll cite below has little potential, but we will see that it is not fanciful. About a century ago, Edwin W. Fay, at that time a well-known specialist in Sanskrit, Greek, and Latin linguistics, was very active in the area of Indo-European etymology. He seems to be almost forgotten. Wikipedia does have an entry about him, but it is very short and gives no idea of the range of his activities. Those interested in a rather representative list of his publications, at least to the extent it concerns Germanic, will find it, among other places, in my Bibliography of English Etymology. Obviously, if knife has a non-Germanic cognate, it should begin with gn– (by the First Consonant Shift). Fay cited Greek genus “lower jaw,” allegedly from a root meaning “to cut.” He pointed out that the jaw with the teeth was the first cutting instrument and that Samson made use of the jawbone of a donkey for bloody slaughter. Several other scholars also tried to derive knife form a gn-root meaning “to cut.” Perhaps knife, or rather knífr, did at one time mean “cutter,” but those who have read the previous posts on sword, axe, and adz (see, for instance, the latest one for June 17, 2020) will remember the many rather uninspiring attempts to explain all of them as cutters.

Samson at the height of his fame. Samson slaying the Philistines by Hilaire Pader. Photo by Didier Descouens. CC By-SA 4.0 via Wikimedia Commons.

We should rather concentrate on the phonetic shape of knife. A glance at an array of kn-words will take us to knob and its twin nob “head”, knop “bud,” knub ~ nub “a small lump,” knot, German Knopf “button,” and Knospe “bud,” among very many others (thus, various protuberances). It is hard to tell how many relatively recent words spelled with an n might once begin with kn-. Perhaps the underlying sense of knob and the rest was “swelling.” Close to them are Engl. knee and German Knochen “knee.” Unlike knot, knee has respectable relatives, with Greek gónu and Latin genu among them (genu is familiar to English speakers from genuflection). It seems natural to reconstruct the meaning of this root as “to bend,” especially because Greek gōníā means “angle.” However, a knee is also a kind of protuberance. Not incredibly, the kn-root in Knochen and elsewhere is sound-imitative, as in German knack “crack.” Nor is Engl. knack “skill at performing a difficult task” too far from knack “a sharp blow or sound.” Fay’s jaw, knee, German Knochen “bone,” and their kin are members of an amorphous mass, containing the names of vaguely similar objects.

The German for “nutcracker” is Nussknacker. Image by Monika1607. Public domain via NeedPix.

Knife looks different, but it too belongs with several kn-words, and the goal of the previous excursus was to point out that the multitudes with which we are dealing have porous borders.  Two German words spring to mind: kneifen “to pinch” and its regional variant kneipen. Other kn-words have different forms but mean approximately the same: Engl. knead, German knatschen “to crumple,” German knicken “to snap; fold, bend,” and so forth. We wade through this morass (knack, knick-knack, knock, knop ~ knob) and begin to realize that the words are eerily alike: they denote lumps, bones, small hard objects, blows, cracks, and so forth. We hardly detect a root there in the strict sense of this term. In such words, one rather senses an impulse, perhaps sound-imitative, perhaps sound symbolic: kn– and more or less arbitrary “supplements.” Say kn– and add whatever you want; people will understand you. This process goes back to antiquity. Outside Germanic, gn– performed the same function. Is knife one of such words? Perhaps. If so, it is, from an etymological point of view, not a cutter, but rather a crusher (Dutch knappen means “to crush”). Walter W. Skeat thought that knife is related to nip and nibble: compare Low German knibbeln and Dutch knabbeln (the same meaning).

Other suggestions about the origin of knife are not radically different from those we have seen—except one. It is usually believed that French canif “knife” is a borrowing from Germanic (possibly from Old English). The twelfth-century form was canivet, which to the unprejudiced observer looks like a diminutive form of the same word, with a inserted to facilitate the pronunciation of the unfamiliar group kn. Theo Vennemann, a distinguished German philologist and linguist, concentrated on Basque ganibet ~ kanibet “knife.” According to him, the word consists of two parts: kani and bet, both of which he analyzed in detail. This word, he believes, traveled to Catalonian and French and from the Basques to Germanic, including Old Icelandic. Knowing nothing about Basque, I’ll leave kanibet alone, though, I suspect that the proposed etymology is too complicated, and complicated etymologies seldom survive. Finally, to repeat, I am not sure why canivet could not be a diminutive of canif. Vennemann’s hypothesis is part of a much larger whole, for he believes in the existence of a significant Basque substrate in the modern European languages. This aspect of his reconstruction cannot be discussed here.

The Basque country. Map by Panonian. Public domain via Wikimedia Commons.

In my opinion, knife should not be separated from a host of other Germanic kn-words. Occam’s razor works very well in etymology. Knife, as it seems, was coined in Germanic, most likely, in Old Norse. It belongs with many words designating objects and actions that suggested knocking, pushing, bending, pressing, and so forth. Not improbably, all such words were “expressive.” From the Vikings the word became known in English, French, and Basque; Basque borrowed it in its diminutive form. We are in the dark about the attraction of the Germanic knife, but, given the movements of great masses of soldiers at that time, the names of various weapons could and did easily become common property.

Feature image credit: photo by Nicolas Monasterio. Public domain via Pixabay.

Recent Comments

  1. Stan smith


  2. Constantinos Ragazas

    Anatoly you write,

    “It is hard to tell how many relatively recent words spelled with an n might once begin with kn-.”

    This illustrates my reasons opposing ‘spelling reform’! Much to lose and nothing to gain!

    Such ‘double consonants’ in the spelling of English words (and other European languages) may signify a transliteration of a Greek consonant with no equivalents in the Latin alphabet. Especially in the beginning of a word.

    Such Greek consonants as ξ (ks), ψ (ps), χ (ch) and γ (ng), for example. Note further these may transliterate slightly differently in different languages. Depending on the phonology needs of that language. [“Outside Germanic, gn– performed the same function”]

    In my view, “knife” comes from the Greek word for knife “ξίφος”. Among other things this would explain the “k” in “knife”.

    Note also “ξιφος” is well attested in Greek. With words like “ξιριζω” (shave), “ξεραφι” (blade), “ξινω” (scratch), etc. No need to make up stories to justify various “academic etymologies” for it.

    Not all English words that begin with a double consonant “kn” need to derive from “ξ”, however. Or all transliterations of “ξ” to have lead to “kn”.

    The word “knee” likely derives from the Greek “γόνα” . With the contraction of “ο” to form a monosyllable “γνα” we can easily see the “knee”.

    As I have often mentioned here in the past, ‘primitive’ English words often are contractions of typically polysyllabic Greek words.

    The polysyllabic nature of Greek all the way back to prehistoric times words may seem to some similar to an agglutinative language.

    [ “What can be said about Linear A is that it renders an agglutinative language containing both suffixes and (a large number of) prefixes[3]. This is not typical of Indo-European languages.” Philip Ingvaldsson Kitselis]



  3. A

    I remember when i was yung, sum of my frends and i would jokingly call it “kin-iff’-ee”! Seeds of a desire to upgrade spelling?

  4. Rudy Troike


    If you lived in Spain or Mexico, you would find words like ‘psychology’ or ‘philosophy’ spelled ‘sicologia’ and ‘filosofia’, and no one protests. Most English speakers don’t know that these words come from Greek, so the etymological beauty argument holds little water for 99% of English speakers. Transliteratting from Russian, do we use ‘czar’, which preserves some semblance to ‘caesar’ (which German happily respelled as ‘Kaiser’ in titling their ruler), or as ‘tsar’?


  5. Stephen Goranson

    The Mason and Dixon Line was well-known in early US history and is often suggested as the origin of Dixie Land. About the proposed move from “Dixon’s Line” to “Dixie’s Land,” two related collocations may be of interest.

    “Dixon’s land” appears in June 15, 1835 (Monday) Evening Star [New York, NY] p.2, col. 2 [AmHistNewspapers] :

    Query—What would be the punishment of a negro flogging an alderman, south of Mason and Dixon’s land?

    And “Dixon’s land” also appears in many July, 1861 accounts about politician John Bell of Tennessee.

    “Dixey’s line” appears in the Feb. 10, 1861 Sunday Dispatch [Philadelphia], p.1 col.7 [AmHistN]:

    …for two months, there hasn’t been a paragraph in any paper north or south of Mason & Dixon’s line, or on Dixon’s, or Dixey’s line itself that hasn’t been as reeking with blood ….

  6. Constantinos Ragazas


    “…the etymological beauty argument holds little water for 99% …”

    It’s the 1% I am appealing too! And that hopefully includes linguists!


  7. Allan Campbell

    Kostas, u rite: “It’s the 1% I am appealing too (sic)! And that hopefully includes linguists!”

    U probbably dont hav to convince them, so why bother?

  8. Constantinos Ragazas

    Allan, … but I do unfortunately! Take Anatoly, for example! Who thinks simplifying spelling by erasing the past is fine with him!


  9. Allan Campbell

    Kostas: If u ar so enamored of the past, may i presume that u travel by horse and buggy, bathe in a tub, wear hirsute clothing, and communicate with letters ritten with a quill pen (except here, of course!)?

    Or maybe u move with the times?

  10. Constantinos Ragazas


    We can be modern and still respect the past. When we cut loose from our roots and truth, we are lost left drifting aimlessly in the current…


  11. Allan Campbell

    Kostas We agree! At least with your first sentence.

    Upgraded spelling does not necessarily cut loose from its roots. Etymologists, as here, can usually connect them.

    As for truth: What is the truth of an electric saw other than a development from the hand saw? Technology doesnt stand still. Tools, such as saws and spelling, ar free to develop!

Comments are closed.