The text for the day comes from Paul Brians's Common Errors in English Usage, in the entry Artic/Arctic:
Although some brand names have incorporated this popular error, remember that the Arctic Circle is an arc. By the way, Ralph Vaughan Williams called his suite drawn from the score of the film Scott of the Antarctic, the Sinfonia Antartica, but that’s Italian, not English.
Brians's advice is specifically about spelling, but the spelling Artic is simply a reproduction of a very common pronunciation of the word, and it's the pronunciation that's the root issue.
My interest in this case comes from my interest in fashions in prescriptions: certain usages are widely proscribed, often with extravagant condemnation, while other, similar, usages escape attention. In the case at hand (and another I hope to post about soon), it's "simplified" pronunciation that is at issue. For Arctic/Artic, there's also a complex history (one that Brians might have misunderstood).
Brians's reference to arcs is just wrong as an account of the structure or etymology of the word. To be generous to him, I'll assume that he meant that reference to be merely a mnemonic. (Though by now there may be people who believe, eggcornishly, that arc is part of Arctic.)
The ultimate source of the word is Greek arktos (in transliteration) 'bear', a reference to the constellation Ursa Major, the Great Bear. The adjective derived from this noun was borrowed from Greek into Latin, where it was spelled with two c's. But then the [kt] sequence was simplified in the transition to Old French (and Italian, as you can see above), and the word was spelled artique in OF, and artic in English when English borrowed the OF word. So the artic version was the original in English.
According to the OED, starting in the 17th century the word was refashioned to make its spelling match the Latin (the history of English spelling is full of such oddities), and then people began pronouncing the first k to match the spelling. Yes, spelling pronunciation. (Something similar happened in French, leading to Modern French arctique.) Meanwhile, lots of people who spelled the word according to the new norm nevertheless simplified the [kt] sequence in pronunciation (a second shot at simplification) — I do so myself — and the [kt] pronunciation became the focus of intense prescriptive pressure.
Last September, the American Dialect Society mailing list took up the prescription question, set off by a posting from Wilson Gray:
Some time in the early to middle 'Nineties, a brief analysis of English consonant clusters, IIRC, was published in Linguistic Inquiry. The author noted that her analysis had one major flaw: it predicted that "Arctic" [arktIk] would be pronounced as though spelled "Artic" [artIk]. When I saw this, I "jumped straight up," as we say in Los Angeles BE.
As children in Saint Louis, we were specifically taught, in fourth-grade "georgaphy" - another pronunciation that the nuns labored to eliminate - that "Arctic" was to be pronounced as though spelled "Artic" [artIk] and *not* as [arktIk]. As a consequence, for the past sixty years or so, I've been incredibly annoyed by the seemingly-universal use of the spelling-pronunciation, [ar_k_tIk].
[The article Wilson Gray was recalling is
Borowsky, Toni. 1989. Structure Preservation and the syllable coda in English. NLLT 7.2.145-66.
On p. 154, Borowsky notes Antarctica and arctic as "especially problematic" for her analysis, and on p. 158 she concludes:
Cases like arctic should be ruled out but are not. I assume that they are entered in the lexicon with their aberrant syllabification. (But, notice that arctic is often pronounced [artik].)
(Thanks to Ellen Kaisse for finding the bibliographic details.)]
Jon Lighter then reported the opposite experience:
I grew up saying "Artic/antartic." Then one day in grade school (musta been 1956 or '57) a teacher told us emphatically that only losers failed to pronounce the "k." (I'm paraphrasing).
So I switched. Had I known Wilson was doing the opposite, however, I'd have stuck to my ways.
I then did some searches of sources:
Like Jon Lighter, and no doubt many others here, I had exactly the reverse experience [from Wilson's]. The facts are more complex than either of these teachings would suggest. From the American Heritage Book of English Usage (1996): "Arctic was originally spelled in English without the first c, which was later reintroduced after the original spelling in Greek. Both [pronunciations] are equally acceptable."
The Wikipedia page on "words of disputed pronunciation" shows a very complex pattern of advice (in this summary, (1) is the [k]-less pronunciation, (2) the pronunciation with [k]): "The debate is whether or not the ct cluster is pronounced [kt] or just [t]. M-W lists both, with (1) first, but OED only lists (2) while noting that the oldest spelling (dating from the 14th century) is Artik, implying that (1) is the older pronunciation. EEPD lists only (2). LPD lists both for both British and American English, but marks (1) as "considered incorrect" for British. K&K list both but mark (2) as "now rare". Generally, the same pronunciation for the ct cluster is used for both arctic and antarctic. However, M-W lists (2) first for antarctic."
M-W: the Merriam-Webster Dictionary; OED: Oxford English Dictionary; EEPD: Everyman's English Pronouncing Dictionary (Gimson rev., 1977); LPD: Longman Pronunciation Dictionary (1990); K&K: Kenyon & Knott, A Pronouncing Dictionary of American English
(NOAD2 and AHD4 list both pronunciations.) It's hard to know what to make of these conflicting lines of advice. In any case, dictionaries attempt to report actual usages, rather than opinions or beliefs about these usages. So Dennis Preston did a little bit of empirical work about attitudes:
A survey this semester [autumn 2007] of over 700 MSU undergrads (none had had any linguistics) revealed the following:
TABLE OF VALUES FOR ARCTIC
1.000 = /arktIk/ is correct and I pronounce it that way
2.000 = /artIk/ is correct and I pronounce it that way
3.000 = /arktik/ is correct but I don't pronounce it that way
4.000 = /artik/ is correct but I don't pronounce it that way
[results] … 65% believe /arktik/ is "correct" (and over 70% are "secure" in their pronunciation, regardless of which form they think is correct).
(Of course, these are self-reports on attitudes and practices; we don't know about the students' actual usage.) I found the 65% vote for the [kt] version surprisingly low, given the classroom and advice-manual pressure in its favor. This might be another case in which many speakers, especially younger ones, are simply disregarding prescriptions that run against their own usage preferences.
What is the basis for the prescription in favor of the [kt] variant? Two sources, so far as I can see: (a) the belief that this is the "original" variant (but see my discussion above), and that original variants should be preferred to innovations (no matter how venerable these innovations are, an idea that we have criticized here many times); and (b) the belief that words "should be pronounced as they are spelled", in particular that all the letters in them should be pronounced.
This second position gets the spelling-pronunciation relationship backwards — especially given the fact that for a great many words children learn a pronunciation well before they learn a spelling (my four-year-old granddaughter, for instance, has the [t] variant for Antarctica, and this is a word that comes up a lot, since we are a penguin-oriented family), and given the presence of a large number of "silent letters" in English spelling, letters that could in principle be pronounced (receipt could be pronounced to rhyme with reaped, indict to rhyme with predict, and so on).
As in many other cases we've talked about here, the general, "umbrella", principle (in this case, "pronounce words as they are spelled") is not actually functioning as a guide to practice, but instead serves as a justification for recommendations about very particular bits of practice. The advice is to pronounce particular sounds in particular words, and the advice has to stipulate these one by one; the umbrella principle adds nothing to this advice.
Now to the issue I began this posting with: fashions in prescriptions. When you look at objections to "simplified" pronunciations, you see that a small number of items come up again and again: Arctic missing the first [k], government missing the first [n] (I'll post on this one in a while), February missing the first [r], and a few others. Meanwhile, lots of other simplified pronunciations escape censure. In particular, tons of instances of "final t/d deletion" — last discussed here in connection with pronunciations of closed (in closed circuit, closed-minded, and closed-captioned) as [kloz], with no final [d], and with a link to an earlier discussion of different cases by Mark Liberman. (Then there's the extremely common pronunciation of last in last night without a final [t], and many other cases.)
All these examples really are simplifications, in that the number of articulatory gestures required is reduced. But for a few of them, there is prescriptive pressure against the simplifications — a kind of tradition handed down in the prescriptive literature which treats them as especially heinous. As a cartoon extra, here's a strip from Alex Hallatt's Arctic Circle that I've been saving for a while. It's about pantomime and dialogue, not about pronunciation and spelling, but at least it's suitably Arctic: