Between the devil and the deep blue sea
My discussion of idioms does not rest on a solid foundation. In examining the etymology of a word, I can rely on the evidence of numerous dictionaries and on my rich database. The linguists interested in the origin of idiomatic phrases wade through a swamp. My database of such phrases is rather rich, but the notes I have amassed are usually “opinions,” whose value is hard to assess. Sometimes the origin of a word is at stake. Then I find myself “at home.” For example, Walter Skeat and his coauthor and occasional opponent A. L. Mayhew disagreed about the etymology of the exclamation dear me. But the polemic centered on the origin of dear (Mayhew seems to have been right), and the arguments were familiar to me. However, if I read a note on the occasion that allegedly gave rise to the phrase between the devil and the deep sea, I cannot check the reliability of the source. I am sure Stephen Goranson is right, and perhaps the story is partly apocryphal, but, there may be a grain of truth in it. As for deep, rather than deep blue sea, the rhyme is perhaps wrong, but the alliteration devil/ deep saves the whole. Incidentally, knowing nothing about the circumstances in which the phrase was coined, I suspected its authenticity just because of the alliteration: the whole struck me as too good to be true. Whatever the original comment was, it must have been given a few last touches before it became proverbial (see Etymology Gleanings for June 2018).
Pilgarlic
No doubt, pulling garlic, while bemoaning one’s fate, cannot be an expected agricultural activity (see the post on the amorous pilgarlic). But, if the phrase was current as an idiom, I see no objection to it. Likewise, if we read that some event got a person’s goat or ruffled his feathers, we hardly conjure up the image of an irate horned animal or of a pinioned individual. But I must confess with great embarrassment that I cannot find the comment in which the authenticity of Chaucer’s phrase was called into question. When and where was it posted?
Now pillock. This is a widely-known British word for a despicable idiot. From a semantic point of view a pillock is the same as a jerk. And, yes, there is every reason to believe that it goes back to Middle and early Modern Engl. pillicock “penis; boy.” The OED refers to Norwegian dialectal pill “penis,” which, I am sure, is the same word as pille “pillar,” its apocopated form (so the reference is to an erect penis). The word may perhaps be a loan from northern English dialects, but there is no certainty. If pill is a variant of peel (as discussed in the post on pilgarlic) and an occasional synonym of pull, peel a cock acquires the familiar sense! Inserted i and a are common, as in handicraft, cock-a-doodle-doo, and the like (I discussed such words in my etymological dictionary in the entries ragamuffin and cockney; posts on those two words also exist).
Some math
The origin of number and arithmetic is not problematic. The ultimate source of arithmetic is Greek “the art of counting” (tékhnē, as in technique), but the word came to English from Old French, where it sounded as arismetique, which, through folk etymology, yielded ars metrica. In the sixteenth century, arithmetic was made to conform to its Greek source, with th substituted for s. Number goes back to Anglo-Norman and Old French, but ultimately to Latin. Old French already had the excrescent b, and Modern French retained it (nombre). Rather probably, Latin numerus is related to Greek némein “to deal out, distribute.”
With regard to handful, most of my eight informants preferred to use a singular verb with it. Yes, number does take the plural (a number of people were present), but nothing follows from this fact about handful, especially in American English, in which collective nouns are usually followed by singular verbs. The problem is that American speakers may say a handful of judges are not because handful is a collective noun but because they tend to make the verb agree with the noun closest to it. We’ll never know why the journalist used the plural verb in the sentence I quoted, and at the moment, this is the least of our concerns. To make my point quite clear, here is a recent quotation from Newt Gingrich: “We learned that every one of those agencies have interest groups that desperately want to survive.” (Nowadays, I read newspapers mainly for linguistic purposes.) The mood of the tales are gloomy will remain a mistake until we agree that such concord is correct (as simple as that). “It may be wrong,” a student once said to me, “but it flows.” And it certainly does.
Greek and English words
Returning to this issue is probably a waste of time, but let me repeat: etymology is based on the achievements of historical linguistics. Comparing look-alikes was the method practiced in the eighteenth century and earlier, and there is no need to return to it. I’ll confine myself to one example. Engl. comb does not resemble Russian zub. Yet they are related to each other and to Greek gómphos. This conclusion is not a postscript to an imaginative story but the result of a study of old forms. There still are people who attempt to trace all words of a given language to one source. The gamut goes all the way from Hebrew to Scottish Gaelic, with Greek, Latin, and (deplorably) Russian included. Ernest Weekley called such enthusiasts monomaniacs. Language history allows us to connect dissimilar-looking words. Thus, related to Old Engl. comb was the verb cemban (from kambjan). Someone who was “uncombed” is now “unkempt.” Without recourse to the history of English, no one would have made this connection. Let me repeat: stringing together look-alikes is a waste of time, and why should hundreds of Old English words have been borrowed from Greek, seeing that the illiterate (“preliterate”) speakers of North Sea Germanic had no contacts with the speakers of Classical Greek? As to Greek honi “horn,” I can only say that the word was not recorded in Classical Greek, so it must be recent in the modern language and, by definition, it has no old connections with Engl. horn: a mere coincidence, of which there are many, unless it is a recent borrowing from English.
The origin of fake
This is what Professor Maher wrote to me: “Old Engl. fæc is the coiled rope. A rope that is not whipped at the end comes undone…. The end of the fæc is the fag end. The primary sense of faking for sailors and hairdressers was skillful arrangement of sisal fibers and hairs. The sense ‘spurious, counterfeit, faux’ inevitably followed and far outnumbers that nautical and tonsorial uses today.” I know only Old Engl. fæc “space of time; division, interval,” but, regardless of the Old English word, I have trouble with fake, because it surfaced so extremely late (in the 19th century). Be that as it may, any attack on such a hard problem is welcome.
Two Russian digressions
Russian uzhas “terror” is indeed related to anguish and the rest, but I wonder where I found the information that uzkii “narrow” is related to it. Indeed, not in Vasmer or any dictionary, though I am sure that what I wrote was not my invention.
Yes, the Russian noun kolkhoz is an abbreviated compound made up of two roots: kollektivnoe khoziaistvo, that is, collective ownership or a collective economic unit (collective business, as it were); compare English words like Caltech. After the revolution, such words were coined by the dozen; some of them stayed in the language. Collective farm is the way that awkward Russian coinage was rendered in English. As to ferma “farm,” Russian seems to have borrowed it from French (and added the feminine ending a). In dictionaries, it does not predate 1837. English also took over ferme in its Romance guise, but in late Middle English, the group er was rather regularly broadened to ar; hence Berkeley (which is pronounced as Barkley in the name of Bishop Berkeley and elsewhere; compare clerk ~ Clark); Derby (= Darby), parson, an etymological doublet of person; varsity (that is, University), and many other common words, for instance, far, carve, star, tar, and farthing, in which ar goes back to er.
The spelling of mattress
Why tt and ss? Double letters are the bane of English spelling, even though some people believe we have not got enough of them. For instance, in the United States, the plural of bus very often appears as busses. Old French and Middle English had materas. Double letters in this word began to appear as early as the 16th century (OED). Possibly, after the loss (syncope) of e in the middle, tt was supposed to indicate the phonetic value of the letter a, though, as we can see from matrimony and matrix, there is no system here. The letter s was probably doubled, to make the word look more impressive (a common phenomenon in the Middle Ages and later; some scribes repeated the last letter even three times). Conclusion: join the ranks of Spelling Reformers (Carthago delenda est).
On dogs and men
A correspondent informed me that Persian folk etymology derives sag “dog” from seh–yak “one third,” because one third of the dog’s essence is human. It would have been a pity not to share this piece of information with a larger audience. I’d rather say that at least one third of the human being’s essence is canine.
Spanish wisdom
Another correspondent wrote me about a tile inscribed in Andalusian Spanish, which reads (in translation): “Lord, bless anyone who does not waste my time.” In his friend’s interpretation, it means “God give me patience with time wasters.” This sounds like a good equivalent.
Many thanks for the numerous letters and comments and forgive me for a post of inordinate length: I hope I did not waste too much of your time.
Featured Image: A handful. A collective noun? Featured Image Credit: “Strawberries Fruits Close-Up Food Fresh Hands” by Pexels. CC0 via Pixabay.
Anatoly Liberman is the author of 
Dear Anatoly,
“Latin numerus is related to Greek némein “to deal out, distribute.” ”
Not to Greek “numero”?!
I don’t know how much you know about numbers, but there are two types of numbers. “Cardinals” (counting) and “ordinals” (ordering).
The word for “cardinals” is indeed the Greek “numero” (number). Which derives from the Greek “nu” (mind) and “mero” (place).
The Greek word “numero” actually means to mentally hold a place. A mental place-holding activity. Which is the essence of counting.
While the Greek “arithmos” derives from “rithmos” (order, rhythm). In the sense: first, second, etc. Which describes ordinal numbers.
Your etymology blind spots trully astound me!
You write, “etymology is based on the achievements of historical linguistics”.
Maybe! Depends on who writes that history! As the above quote amply attests!
Kostas
Poodle-faker is dated British slang for a young man who provided attention to older women for the sake of professional or social advancement. I wonder if it once meant someone who arranges the hair of a poodle?
DOUBLED CONSONANTS are supposed to mark short, stressed a, e, i, o and u – to distinguish them from their longer versions. If done consistently, it would be very helpful to native young reader and all foreign students, as it is in: ‘mate – matter, era – error, diner – dinner, pole – pollen, cupid – scupper’. Young children’s ‘misspellings’ (verry, reddy, annimal) show us that the basic idea is easy to grasp.
Unfortunately, Samuel Johnson exempted over 500 mostly Latinate words from this system (cabin, habit, agony ..), but not all (horror, terror, mammal).
He made matters even worse by using doubling in over 200 words to mark Latin spelling changes (accuse, account, arrive), rather than for indicating a short, stressed vowel. – U can see the whole mess under ‘Long and short vowels’ on my ‘ImprovingEnglishSpelling’ blog.
Cocking a snoot at Johnson and permitting sensible, rule-governed consonant doubling to become as acceptable as the current unpredictable, dictionary-based usage, would be one of the simplest ways to make English spelling more fathomable. It would save a huge amount of word-by-word memorisation for all students.
The supposed “hundreds of Old English words have been borrowed from Greek” is admissable if, as you of all people should know, by the time the earliest attested evidence of Old English appears, the bible, was already in circulation, translated to Germanic languages from the ancient languages which might as well have been Old Greek next to Latin and Hebrew. To be fair both questions have to be consiered, “Why?” and “Why not?”. Using historical arguments as basis for linguistics is not exactly backwards, but fraught with lack of and outright misinformation.
Now, if you’ll excuse me, I will be comparing words based on superficial etymology, because your lament against that doesn’t stand up to reason, because surely the fact that dissimilar looking words may be related says nothing about the chance of similar looking words. Either way, whether angle and angel look similar is begetting the question. It is kind of easy to imagine semantic links, so it’s fraught with mistakes and folk etymology, I give you that, but folk etymology is part of history. That’s the most important note here.
I do embrace a remark of yours, but with a caveat: “The less we know, the easier it is to indulge in breathtakingly bold speculation. Scholars often have clear notions of what happened five thousand years ago but are at a loss when they have to explain some process in Middle or early Modern English.”
It puts emphasis on the difficulty of reaching back into time. But it kind of denigrates the importance of speculation, as time grows forward, not backward, so the opposite case, to have a clear notion of recent language without understanding of it’s origin is likewise impossible. To know everything is but an ideal. To know something with sufficient certainty requires model building, verification and falsification. But there are only so many concurring mental models one can hold in ones mind. So it’s only correct to react with rejection to foreign proposals, if that’s an instinct to falsify those and rather verify ones own, or otherwise with ignorance.
On the question of number, a single comment is of course not convincing, but after a long thought I found that μέρος does make sense in a number of ways (no pun intended).
Everyone will object that for numero the Proto-Italic *numeso was reconstructed, and further Proto-Indo-European *nem. But, if you look at the cognates which mean take, distribute, meadow, custom-law and sanctuary, holy-place, we can infer some sense of title, claim, plot of land. This fit’s further with my assumption that number and name must be connected, a number one for example being but the name of an abstract concept. That leads me to “word”, from *werH (speech, talk), and Ger. “Wert” (value, numeric value) from *wert- (turn, exchange) where “exchange (of words)” is one obvious link between *wert and *werH. The other *we- roots are ver-y ver-satile, too.
So, what does that tell us? German Wiese (meadow) is from *weHs- (flowing), and has a sense of glowing, as does Aue. Exchange would often go along a river. The sense of flowing is implied in language too, and reason is not too far from rheo (OK that last one is a crutch).
Look up *(s)mer- whence meros, it means shiny, pale (whence a meaning dying, to fade — murder, maybe to let the blood flow?) and so does the meadow root and the Wiese root (whence then flowing with clear water as opposed to mud, I guess) and even “Aue” looks similar to arurum.
But that’s certainly not the end of it. Looking up “place” I get “simeio”, too, related to “sema”, sign, mark, spot …
I’ll leave the conclusion alone, as far as I am concerned it opens more questions than it asks. I know how to tackel numbers in the decimal place system, that should suffice.
I hope you can see what I’m getting it. I don’t know the turn roots too well (they make me dizzy). I will stop here and look up sum words, to sum them up.
Thanks for checking on the Russian word and pointing out how the e became an a.