Revising OED Etymologies.

Recently I lamented the confusing wording of OED updates, and ktschwarz commented:

Could be anything from minor formatting changes, to correcting errors in quotation dates and sources, to restoring spelling and punctuation as they appeared in the original editions, to adding or dropping quotations (potato, for example), to adding and updating etymologies ahead of full revision (see December 2020 blog post), maybe even changes to the definition.

The Life of Words is annoyed too: “The impression is of a very up-to-date dictionary, which at least half of is very much not.” […] This is not as bad as making all these changes without indicating them at all, which is what they did up until this year (see Examining the OED for some blatant examples), but it’s not as good as it should be.

As I said then, I love the OED and they do great work, but they need to do better in this regard. Here’s the first paragraph of that OED December blog post, so you can see their explanation of what they’re up to:

Over the past eighteen months we have begun a new initiative as part of the ongoing revision of the OED: revising the etymology and variant forms section in entries that have yet to be revised in full. We are doing this in order to remedy deficiencies in entries that hitherto lacked an etymology, or where we have been aware that the etymology and variant forms sections already offered could be significantly improved, ahead of revision of the full OED entry. We have done this by making use of spare moments in our schedules, where members of the etymology team had completed their work on OED revision ranges slightly ahead of expectations. By working in this way, we have now managed to publish over 1500 revised etymologies and variant forms sections in this new stream of work. For each of these entries, a note appears in the “Entry history” window, “Etymology and variant forms provisionally revised”, together with the date when these revisions were published.

But I recommend reading all the links if you’re interested in this stuff.


  1. There are more details on short term priorities, plus a major change in practice here: The major recent practice shift has been away from what was primarily an entry-by-entry revision workflow (though various things were always being tweaked in the background) towards more deliberate targeted partial revisions:

    …we diversified our approach to revising and updating the OED to include cross-text improvements and lexicographical projects, as well as entry-by-entry editing … Fundamentally, we believe it makes the dictionary more useful sooner.

    OED is “useful” in many ways, of course. Some of us are concerned that the shift will delay the systematic updating of pre-2000 material, perhaps indefinitely in areas deemed less immediately “useful”, but which in the aggregate make it the great resource it is.

  2. Yes indeed.

  3. OED is great. But. E.g.

    The OED entry on Hoosier, n. still gives a first use as

    1826 in Chicago Tribune (1949) 2 June 20/3 The Indiana hoosiers that came out last fall is settled from 2 to 4 milds of us.

    Jeffrey C. Graf, of Indiana U, Library (as noted by Fred Shapirio) wrote many years ago:

    “A research worker in the Indiana State Library discovered the
    letter, and the Library reported the discovery in the April 1949 issue of
    the Indiana Bulletin of History. The Chicago Tribune picked up the story
    and ran it on June 2, 1949, but improved the spelling and turned
    “hoesiers” into “hoosiers.”

    The letter is real enough, but for some reason the date is wrong,
    1846 rather than 1826. 1826 is not smudged or marred in any way on the
    manuscript, but a curator or librarian has indicated in pencil [1846] on
    it to correct the writer’s error. Holt County, Missouri, was not created
    until 1841, named for a man who died in 1840. Census records of 1830 and
    1840 place James Curtis in Indiana. The letter itself refers to the
    marriage of C. J. Beeler and Margaret Vondy (born on the Isle of Man) on
    Thursday, February 5th. February 5, 1846 was a Thursday, and the Holt
    County Historical Society has published a list of Marriages Recorded in
    the Holt County Missouri Courthouse. One entry reads: G. I. Beeler.
    Margaret Vandy. 5 March 1846.”

    While on the subject, here’s a little-noted use:

    Baltimore Republican, page 2, col. 2, September 29, 1831 [America’s Historical Newspapers]. It reports (after the Louisville Advertiser) on a Jackson party election victory celebration in Louisville, Kentucky held on the 17th. After speeches
    “The following regular toasts were drunk, with appropriate music:
    1. Our Country.–Her soil is consecrated to liberty by the blood of our forefathers [Hail Columbia….
    10. Gen. John Adair.–In his return to the ensuing Congress, Kentucky exhibits her lively recollection of and gratitude for his eminent public services [Hoosier March by A. M…..”

    Gen John Adair’s service included northwest territory fighting.

  4. ktschwarz says

    Something I noticed recently about the revised etymologies in general: they’re consistent about tracing anything Indo-European as far back as possible, and when they come to something with no PIE source, they say so. For example, the revised ocean, geo-, anthrax, narcosis are from ancient Greek and then “of unknown/uncertain origin”, whereas brachy-, pachy-, techno- are from Greek and then from an “Indo-European base”. The old edition just said “Greek” for all of these and left it there.

    Also, they do suggest pre-Greek as a possible origin, but very rarely and cautiously, in only eight entries so far and qualified with “perhaps”: aegilops, anthericum, empusa, esalon, Ethiop, Europe, simpulum, thalasso-. Plus absinthium, where they say “of uncertain origin” but add a note:

    < classical Latin absinthium … < ancient Greek ἀψίνθιον wormwood < Hellenistic Greek ἄψινθος wormwood (although this is apparently first attested later), of uncertain origin + -ιον , suffix forming nouns. Compare later absinth n. 1.
    The element -νθ- in Hellenistic Greek ἄψινθος indicates that the word is of pre-Hellenic origin.

    Is “pre-Hellenic” correct there? I thought “pre-Greek” was the usual term for the hypothetical substrate language, but I know nothing about this. Also, “ancient Greek < Hellenistic Greek” seems backwards: isn’t Hellenistic Greek later? Should it say “< ancient Greek ἀψίνθιον < a presumed earlier *ἄψινθος (although this is attested only later, in Hellenistic Greek)”? Or am I just confused?

    They *are* willing to say definitely that Latin or Greek words came from an unknown language in some cases, e.g. the just-revised vaccinium (genus of cranberries, blueberries, etc.):

    < classical Latin vaccīnium whortleberry or bilberry, in post-classical Latin also a plant with purple flowers, perhaps the violet …, a loan from a non-Indo-European language, which was also borrowed as ancient Greek ὑάκινθος hyacinth n.

    This update was a long time coming; that origin is already in Merriam-Webster online, and Meillet all the way back in 1908 argued that the Latin and Greek words were parallel borrowings from a third language, rather than Latin borrowing from Greek.

  5. isn’t Hellenistic Greek later?

    Yes indeed. You’re not confused, they are, and they’re confusing the public. Wake up, OED!

  6. I imagine they are taking Beekes as an authority. His is the most up-to-date Greek etymological dictionary, but his enthusiasm for a Pre-Greek substrate is not universal, from the little I know.

  7. ktschwarz says

    They aren’t taking Beekes as an unquestioned authority: e.g. ocean, geo-, anthrax, narcosis are all counted as pre-Greek by Beekes but “unknown” by the OED, and likewise for labyrinth, obelus, and many others. I’m guessing when they just say “unknown” it means they think the evidence is insufficient even to say whether it’s a loanword or not, since there are many other words that they now say were “perhaps” or “probably” loanwords into ancient Greek: abacus, acacia, agate, adamant, alabaster, angel, etc.

    All of the above examples came into English via Latin (except narcosis directly from Greek), which got them from Greek. There are some others like vaccinium that the OED judges as parallel borrowings into Latin and Greek from an unknown language, e.g.:

    plumb: “classical Latin plumbum lead, ball of lead … a loanword, apparently from the same source as ancient Greek μόλυβδος lead, probably from an Iberian language (lead came to the Greeks and Romans originally from Spain).”

    mint (plant): “inherited from Germanic … < classical Latin menta, mentha, probably borrowed, like ancient Greek μίνθη (in Hellenistic Greek also μίνθος), from an unidentified source.”

    Also, I’m bothered that they use “pre-Greek” as a technical term in etymologies but they haven’t entered a definition for it. They do have a sub-entry for it under “pre-“, but only as an adjective referring to time before (“pre-Greek Anatolia”), not in the noun sense referring to a substrate language.

Speak Your Mind