This is an essay on the word hag, but let me first thank those who have commented on the latest posts, corrected the mistakes, and made suggestions. I would also like to explain why I translated elf-shot as “lumbago”: this is how Elfschuß is glossed in at least some German dictionaries. In the Middle Ages, “Germanic” elves (not to be confused with the “Celtic” elves of Midsummer Night’s Dream) were believed to inflict all kinds of diseases, and German Alptraum “nightmare” (Traum “dream”) points in the same direction (-mare in nightmare is the name of another evildoer, prominent in old superstitions). I do not know why German authorities associated elf-shot with lumbago. Our information on elves is scanty. Once upon a time, as follows from the evidence of Old Icelandic poetry, they were equal to the pagan gods, known as Æsir (Thor, for instance, was one of them). This follows from the Old Icelandic alliterating formula: “How is it with the Æsir, how is it with the elves?” (The backbone of Old Icelandic and of the rest of Old Germanic poetry was alliteration, not rhyme, and all vowels were allowed to alliterate with one another. Hence the pairing off of æ and e).

Image via Thomas Quine on Flickr; CC by 2.0.
Snorri Sturluson, our greatest authority on Scandinavian mythology from the thirteenth century, knew nothing about elves beyond the existence of two groups: light elves and dark elves. He was, however, fully aware of the popularity of elves in folk belief. They seem to have been a group of tutelary (guardian) spirits, kind, when treated with respect, and vengeful if offended.
Next to nothing is known about the etymology of elf, and the same holds for the “proto-hag,” the heroine of this post. The word surfaced as hegge ~ hagge in the thirteenth century but was rare for the next three hundred years. As suggested by the OED, “witch” (or “a female evil spirit”), the gloss, usually given as the earliest sense of our word, may be a secondary sense. Indeed, nothing is easier than to call a repulsive (cantankerous, querulous, vituperative—add more synonyms if you care) female a witch, and I would like to propose that such was indeed the case. I would even like to strengthen this suggestion and say (pointblank, as it were) that hag never meant “witch” in English. The Oxford Dictionary of English Etymology (1966) gave as the earliest sense of hag “female evil spirit.” Very probably, this sense did not exist.

From Irish Fairy Tales, 1920, illustrated by Arthur Rackham. Public domain via Wikimedia Commons.
The main theme of the extremely rich literature on the origin of the word hag resolves itself into the idea that hag is related to German Hexe and Dutch hecs “witch.” In written texts, Hexe emerged as hagazissa, obviously, a compound word (haga-zissa). It must have been very common if it lost half of its bulk with time. (The more frequent a word is, the easier it is for it to contract.) The Old English cognate of hagazissa is hægtesse ~ hegtes. Likewise, Dutch hecs “witch” goes back to haghetisse.
In a nutshell, the traditional conclusions about the etymology of hag are as follows. In the compounds cited above, the first half is related to the nearly forgotten English noun haw “fence” (from haga) but recognizable in hawthorn (which, too, would have been unknown to many, but for the last name Hawthorn) and hedge (from hegg ~ hecg). The second part of the compound (-tisse ~ –zissa) is almost hopelessly impenetrable, and most efforts have focused on its meaning and origin. However, the oldest recorded English forms of hag (hegge ~ hagge) have no traces of it, and we might have ignored it, but for one circumstance.
It is usually believed that hag also goes back to a compound like the Dutch and German words cited above, but that in English, this second part disappeared without a trace. The inspiration for the etymology of the Dutch-German word is Old Icelandic tún-riða (ð has the value of th in English the), with an exact analog in Middle High German. The word meant “fence-rider” (tún is of course a cognate of English town) and “witch,” because witches were supposed to rule fences, perhaps in the capacity of spirits from the wood, rather than “witches.” For millennia, separating inhabited spaces from the surrounding wilderness (forests) was a most important part of civilized life.

The witch outside Witches Galore at Newchurch in Pendle. © Marathon, CC BY-SA 2.0 via Geograph.org.uk.
In the linguistic literature on hag, beliefs and superstitions have been discussed in an exemplary way, but all attempts to solve the riddle of hag traditionally turn around the meaning of the second component of the old compound. However, as noted above, we may perhaps ignore that part and the rich scholarship connected with it, because the English word has never been recorded with it. Its presence is a mere guess. I would risk suggesting that hag has nothing to do with German Hexe and its Dutch cognate. (Hag, “broken moss ground, etc.” is a different word, related to hew, and need not bother us.)
Hag, as we have seen, surfaced in English comparatively late and for a long time occurred very rarely. When it did turn up, it was hagge, and final –e is a mere ending (perhaps a letter not reflecting any pronounced sound), certainly, not a trace of a discarded component. The Dutch and the German words have also shrunk, but not so dramatically. In principle, shedding the entire second part of a compound is not impossible. Thus, hussy ~ huzzy, which is another term of abuse for women, goes back (surprisingly!) to the perfectly respectable word hūswīf “housewife.” But usually something is left of the discarded stump, as in husband (from hūs-bonda “master of the house”) becoming hubby.
If hag is not a stub of an old compound, what can it be? English words ending in g (like big, bag, bug, bog, dig, egg, cog, log, flog, fog, grog, and many others) are rarely old, and if they are old (like pig, hog, frog, and dog), they pose great difficulties to etymologists. Here are a few words rhyming with hag (the numbers refer to the century of the first attestation in texts): bag (XIII, borrowed), brag (see my post for October 2, 2013), crag (XIII, borrowed), fag (late, much discussed, including in An Analytic Dictionary of English Etymology), flag (all homonyms are relatively late and of dubious origin), nag (very late, of unknown origin), rag (several homonyms, late or borrowed), snag (late, of unclear origin), stag (apparently native; obscure, like dog, frog, and a few other animal names of the same type), and tag (late, of unknown origin). Only shag (as in shaggy) is old and native. Also wag (verb, XIV) is native and may be a sound–symbolic formation. I suspect that hag belongs with fag, ragtag, and other words denoting undignified creatures. A word like hag seems to have grown on the rich soil of informal, unbuttoned English slang. It would be a relief to stop bothering about German and Dutch witches, hedges, and other matters that increase the bulk of publications on the etymology of hag but do not advance the subject at hand. Obviously, I am not so arrogant as to believe that I have solved an almost insoluble riddle, but perhaps my approach has some merit. It would be interesting to hear from our readers what they think about this etymological heresy.
Featured image: Illustration by Arthur Rackham from Shakespeare’s A Midsummer Night’s Dream. Public domain via Wikimedia Commons.
I figured that my comment on the “hemlock” issue wouldn’t be visible to NG Middlemiss, so I repeat it here.
Nat Coler7th March 2025
N M G Middlemiss:
Probably you will not reply, but the take on “one day” as exclusively used for future time is false!
“He used to walk around the forest, and ONE DAY he saw . . “
Nat Coler:
Thank you for chiming in.
My fault for not addressing the issue of “one day”, and my “Nice” was referring only to ‘puss/pus’!
I thought Middlemiss was making a joke!
So, I say it now: If it’s not a joke, then it’s very wrong/false!
Dan Beran, Nat Coler:
I hope Anatoly sees your comments!
I add mine here:
I seemed to me as well that it was a joke on the part of Middlemiss. I just don’t find any reason for such action!
Linguists – if he/she/they are that – can only take a spoken language and structure it such that it can be taught to the children in that language, or to the foreigners!
If a language survives, and the users survive, it means that they understand one another!
It is beyond ridiculous to come in and posit non-sensical “rules”. The language has been doing just fine without such interventions – e.g., “no split infinites”, “not ending sentences with prepositions”, etc.
When I started reading Anatoly Liberman’s posts, I commented that – coming from scientific background, where there are innumerable suppositions, distortions, and outright lies – it was refreshing and soothing to find someone doing just a sincere research, without the arrogance and heedlessness that I found in sciences.
I urged: “Keep going!”
Now I realized that in linguistics as well the phenomena I evaded are occurring!
“One day” to be used only for future!!
Oh, no-no-no!
I also, like the other commenters, hope that Anatoly sees these comments of ours.
Sorry, in the previous comment a typo:
It should be “It seemed to me”, not “I seemed to me”
The hag’s revenge and vindication—what a fascinating exploration of etymology and folklore! Anatoly Liberman’s deep dive into the origins of the word hag and its connections to Germanic and Dutch roots is both enlightening and thought-provoking. It’s intriguing how words evolve and carry layers of cultural meaning, often shedding their original forms entirely. This essay reminds me of how music, like language, evolves and carries stories within it. Speaking of music, if you’re looking for a platform to explore the best tunes, I highly recommend checking out https://spotich.com/. It’s a fantastic site for discovering music that resonates with your soul, much like how etymology reveals the soul of words. What do you think about the idea that hag might have grown from informal English slang rather than being tied to older compounds? And have you found any musical parallels to these linguistic mysteries?
The comment from Middlemiss, at the previous blog (Hemlock), concerning “puss” and “pus” is not affecting Anatoly’s erudition. An unnoticed autocorrect event could have easily caused the typo.
As for the second comment, “one day”, I can label it “astonishing”. For the fun of it, I related the matter to a linguist friend of mine. He labeled it “superstition”.
ch amin said:
“This essay reminds me of how music, like language, evolves and carries stories within it.”
Yes.
It also gave me the feeling of excursing with one objective, but experiencing successive, unexpected, seemingly unrelated vistas!
Anatoly Liberman’s style is witty and pleasant. At times, even tongue-in-cheek:
“Such a time machine is not yet available, and this is a blessing in disguise, because if it existed, the few etymologists who are still employed would have lost their jobs, and there would have been no need for this blog (the latter event would have been a catastrophe of global dimensions).”!!
From:
https://blog.oup.com/2025/01/the-year-is-new-and-young-everything-else-is-old/