We were listening to the radio this evening and a woman was being interviewed about a doughnut recipe that involved arrowroot. The interviewer asked jovially “So is it the root of the arrow, then?” and they had a good laugh; I, of course, headed for the dictionary. The American Heritage Dictionary had a particularly good “word history” sidebar, which I will now pass on to you:

The arrowroot is just one of many plants that the European settlers and explorers discovered in the New World. The Arawak, a people who formerly lived on the Caribbean islands and continue to inhabit certain regions of Guiana, named this plant aru-aru, meaning “meal of meals,” so called because they thought very highly of the starchy, nutritious meal made from the arrowroot. The plant also had medicinal value because its tubers could be used to draw poison from wounds inflicted by poison arrows. The medicinal application of the roots provided the impetus for English speakers to remake aru-aru into arrowroot, first recorded in English in 1696. Folk etymology—the process by which an unfamiliar element in a word is changed to resemble a more familiar word, often one that is semantically associated with the word being refashioned—has triumphed once again, giving us arrowroot instead of the direct borrowing of aru-aru.

So it’s like “sparrowgrass” for asparagus, except that it’s become the normal term. Who’d have guessed?
Update. Ian Preston, in the first comment, links to the relevant section in William C. Sturtevant’s “History and Ethnography of Some West Indian Starches” (in The Domestication and Exploitation of Plants and Animals, ed. Peter Ucko and G. Dimbleby, Chicago: Aldine, 1969), which pretty convincingly demolishes the aru-aru theory: “According to Barham, a Jamaica physician writing before 1711, the plant Sloane labelled Canna Indica was called ‘arrow root’ because it was first known as an Indian antidote for poisoned arrow wounds, for which the juice was taken internally and the bruised root was used as a poultice on the wound.” Sturtevant’s conclusion:

The historical development in the West Indies seems clear: an Indian antidote for poisoned arrow wounds, adopted by non-Indians as an antidote for other poisons then extended to other medicinal uses, then used as a food for the sick at about the same time as it became a source of starch for other purposes with starch gaining techniques very likely transferred from those used with manioc.


  1. This lengthy account seems to be sceptical of the link to aru-aru, pointing, for example, to earlier references in French to “l’herbe aux flèches.”

  2. Jane Austen has Emma Woodhouse sending some arrowroot “of very superior quality” to the ailing Jane Fairfax (who rebuffs her kind advance). That’s the first I ever heard of its being useful for more than thickening sauces. Apparently in Austen’s England it was a newly available import, prized for its digestibility and considered to be just what the doctor ordered for certain complaints.

  3. What is the grammatical structure of “aru-aru” ? Is it an intensifying form ? Is the English rendering “meal of meals” supposed to mean “the finest eats in town”, along the lines of “gentleman’s gentleman” ?

  4. I was wondering the same thing as Stu. Its repetition reminded me about Berlusconi and his bunga bunga parties. Bunga bunga, according to Wikipedia & thence The Observer, “has become an instant, supposedly hilarious, household expression”. No one seems to be sure where it originated, but Australia, the Philippines & “Africa” are mentioned. It in turn reminds me of “umpah, umpah, stick it up your jumpah”, the English children’s phrase made famous by John Lennon that we, at least, imagined was sung by dancing red-indians. What’s the social and linguistic background, if any, to this?

  5. Well, Ian’s link is pretty convincing; I’ll update the post accordingly.

  6. If aru=meal, presumably that’s in the ground-up-starchy-substance sense of “meal” (rather than Mahlzeit). Another site I looked at translated aru-aru as “starch of starches”.

  7. With “starch of starches” we’re back to square one. Does that mean the best starch of the bunch (“Holy of Holies”), or a starch that goes down well with other starches (“gentleman’s gentleman”), or does it mean “very strong starch” (that’s what I was getting at with “intensifying form”) ?
    The expression “X of X” is not exactly a commonly used idiom. I would never use it myself because, like “Y qua Y”, it’s just not clear what it means. All I would like to know is what these weird expressions “meal of meals” and “starch of starches” are trying to tell us.

  8. I’m thinking best of the bunch: King Starch, Most Digestible and Pharmaceutically Most Valuable of Starches, Most Pellucid of Inspissating Agents.
    I think that “qua” if used judiciously can convey a clear meaning, but maybe I’m imagining things.

  9. You may be right, although the job of judges is not to imagine things qua things, but to convey them as legal constructs.

  10. Hat, isn’t there an intensify-by-repeating form in Hindi, something like “hot-hot” to mean “very hot” ?

  11. Don’t look at me, I wasn’t there, I didn’t mean to do it, and it was two other guys. But maybe someone who actually knows something about Hindi will show up.

  12. Can you say “things qua legal constructs”? Or can you only say “X qua X”?

  13. Wiktionary gives these quotations of the form “X qua Y”: “animal and human bodies qua percipient”, “It was qua poet that Byron …”, “the military qua organization”, and “blame qua attitude”. I must say these all sound weird to me.

  14. marie-lucie says:

    I think that “king of kings” (a translation from old Eastern languages) = ‘paramount among kings’, ‘having kings as his vassals’, but “Solomon qua king” would mean ‘Solomon, considered only as a king’ or “Solomon, because he was a king’ (apart from other attributes he might possess). Grumbly’s examples sound OK to me, except “blame qua attitude” which sounds weird to me too.

  15. OK, John’s examples of “Y qua Z” make sense to me as reformulated by marie-lucie: “Y, considered only as a Z”. But “Y qua Y” makes no sense.
    I suspect “blame qua attitude” is a pop psychobabble attempt to address the notion of censoriousness. Blame is something that can be assigned based on evidence, but “blame qua attitude” is a disposition to blame people right and left, evidence or not. In German you could say Vorwerflichkeit ;-), but the standard expression is Tadelsucht (addiction to censure).

  16. Grumbly: Gentleman’s gentleman does not mean what I think you think it means: it refers to the high-end personal servant of a gentleman, the analogue of lady’s maid. It does not mean someone who even among gentleman stands out as a gentleman, as in the remark about the king being able to make a nobleman but not a gentleman.
    Here are the OED3’s quotations for qua, which by the way is pronounced to rhyme with say in the U.K., at least by some people, whereas Americans rhyme it with ah. (I darkly wonder if Hat says “kway” too.)
    1647 N. Ward Simple Cobler Aggawam 56 Every man was as good a man as your Selfe, qua man.
    1649 A. Ascham Bounds Publique Obed. 21 The Apostle commands Wives to submit to their Husbands, surely quà Husbands, not quà men.
    1776 Claim Roy Rada Churn 17/1 A body corporate, quà corporate, cannot make an affidavit.
    1847 M. F. Tupper in W. C. Armstrong Compl. Prose Wks. (1851) 490 The man, quà man … was nearer to his Creator, than the woman; who, quà woman, proceeded out of man.
    1867 J. A. Froude Spinoza in Short Stud. (ed. 2) 232 Because things modally distinguished do not quâ substance differ from one another there cannot be more than one substance of the same attribute.
    1885 Manch. Examiner 4 Apr. 4/6 Their censures are not directed against the Church quà Church, but against the Church quâ Establishment.
    1965 G. Grant Lament for Nation ii. 21 This failure to recognize the rights of French Canadians, qua community, was inconsistent with the roots of Canadian nationalism.
    1993 Guardian 21 Aug. (Weekend Suppl.) 6/2 Philip Larkin was unquestionably … better loved, qua poet, than John Betjeman, who was loved also for his charm.
    The 1645, 1885, and (indirectly) 1993 examples are good examples of contrast between “X qua X” and “X qua Y”.

  17. Thanks for the reminder about “gentleman’s gentleman”. I knew that from 19C novels, but had forgotten because (I think) I have more often encountered the expression in 20C texts (American advertising ?) with more or less the sense: “a gentleman whose manners are superior and yet do not excite jealousy in other gentlemen”.
    I agree with your list of good examples, except that possibly you meant to write 1649 instead of 1645. “The Apostle commands Wives to submit to their Husbands, surely quà Husbands, not quà men”: that is a subtle but telling remark.

  18. 1649, yes.

  19. I darkly wonder if Hat says “kway” too.
    Yes, I say “kway”; I fessed up to it in this post (in which thread you were the second commenter, so your dark wonderment was founded on a vague memory).

  20. I quah.
    Anyway, we used to say “”Oompah, oompah, stick it up your joompah” as ridicule of how it would have been said in England, in t’North.

  21. I had completely forgotten about that; I re-created it independently from your known proclivities. Indeed, much of the commentary on this post is re-creating that one, including Grumbly’s semantic issues with “X qua X”.
    (What’s your Hatly take on “re-create”? I say it always needs a hyphen, and “re-creation” even more so. When Cat Stevens, or whatever he is called nowadays, sings the line from “Morning Is Broken” as “God’s recreation of a new day,” I always have to resist bursting out laughing.)

  22. I know someone who unaccountably pronounces “representation” as if it were “presentation” with the prefix “re-“. Likewise “recommendation”.

  23. @Stu: “The expression “X of X” is not exactly a commonly used idiom.” It actually is, even if only in certain registers. Its genealogy can be traced through the KJV to Hebrew and Aramaic; in them, as in all Semitic languages, “X of Xs” is the standard way to construct the superlative. The model can be seen in the other name for the Song of Solomon, the Song of Songs (שיר השירים), as well as in “holy of holies”, both calques of the original Hebrew.
    As for “qua”, meaning “regarded as” or “in its capacity of”, it’s common philosopher-speak. It was one of the usual turns of phrase of the Scholastics, and it’s quite useful when making the kind of subtle distinctions philosophers are wont to. Using it in nonspecialised contexts smacks of pedantry to me.

  24. “Morning Is Broken”

    Of course it’s “Morning Has Broken”; that was a brain fart, not a jest.

    If I remember correctly, the first clue to reading Old Persian cuneiform was recognizing the known formula for referring to kings, “So-and-so, Great King, King of Kings, son of So-and-So” in royal genealogical inscriptions.

    I just finished reading an article today about the mysterious syntax of the Shema, the Hebrew prayer that Jews are to utter twice a day and with their dying breaths: שְׁמַע יִשְׂרָאֵל יְהוָה אֱלֹהֵינוּ יְהוָה אֶחָֽד׃ Sh’ma Yisra’el, YHWH ‘eloheinu, YHWH ‘eḥad lit. ‘Hear Israel YHWH God YHWH one’ (Deut. 6:4). There are many possible readings of this, but the article interprets it as something called “staircase parallelism”, whereby the meaning ABC is expressed by AB//AC. This trope is used mostly in Biblical poetry, but here appears in prose, says the author, and the meaning intended is simply “Hear, Israel, YHWH our God is one.”

  25. What’s the reference for the Shema article?

  26. I went off to find it but got distracted. Here it is: “Deciphering the Shema: Staircase Parallelism and the Syntax of Deuteronomy 6:4”

    While I am at it, it was Georg Friedrich Grotefend who worked out (with some errors) the Old Persian. Here’s WP’s summary of his conclusions:

    1. that the Persian inscriptions contain three different forms of cuneiform writing, so that the decipherment of the one would give the key to the decipherment of the others

    2. that the characters of the Persian column are alphabetic and not syllabic

    3. confirmed Niebuhr’s observation that they must be read from left to right

    4. that the alphabet consists of forty letters, including signs for long and short vowels

    5. that the Persepolitan inscriptions are written in Zend (which, however, is not the case), and must be ascribed to the age of the Achaemenian princes

    6. that a specific frequent word could refer to the Persian word for “king”

    7. that the inscriptions satisfy the two following schemes: A) X king, great king of king, son of Y king; B) Y king, great king of king, son of Z;

    8. that the presence of the two schemes A) and B) gives an opportunity to identify the people involved; it is necessary that X was a Persian king, his father was a Persian king too, but his grandfather was not king

    9. according to this idea Grotefend was able to identify X for Xerxes, Y for Darius and Z with Hystaspes.

  27. Thanks for the paper. I find it absolutely convincing. I’m surprised people didn’t figure it out a century ago.

  28. It’s funny, that’s how I always assumed it worked without knowing much if anything about Hebrew syntax. Chalk one up for ignorance!

  29. I don’t think his conclusion is wrong, really, but it does fall into the fallacy of assuming that a poetic sentence can have only one relevant grammatical parsing.

Speak Your Mind