Sunday, February 26, 2017

On Olathe

A few days ago, two unarmed young engineers from India were shot in a bar in Olathe, Kansas by a man yelling "Get out of my country!", as was a heroic bystander who tried to stop the shooter. As this contemptible crime put a normally quiet suburb of Kansas City into the international news, journalists and readers worldwide must have been wondering, as I wondered the first time I heard of it a couple of years ago: "How do you pronounce Olathe, and what sort of a name is that anyway?"

The way the locals pronounce it is /ou'leɪθʌ/, as you can hear early in the Mayor's speech. This is remarkably irregular: I can't think offhand of any other word in the English language in which a final e is pronounced /ʌ/, except occasionally "the". You might expect the etymology to provide an explanation, but it turns out to complicate the story further.

The town of Olathe was founded in 1857 by one John Barton, a doctor from Virginia, who - by his own account - got it into his head that "beautiful" would be a good name for the town he envisaged, and:

... meeting Capt. Joseph Parks, head chief of the Shawnees, he said: 'Captain, what in the Shawnee language would you call two quarters of land, all covered with wild flowers? In English we would say it was beautiful." Parks replied: "We would say it was 'Olathe,' "giving it the Indian pronunciation Olaythe, with an explosive accent on the last syllable. Barton made the same inquiry of the official interpreter, an educated Indian, who made the same reply, adding that for English use it would be best to pronounce it "Olathe," with the accent on the second syllable. So it came to pass that the new town was named "Olathe," the city beautiful. (History of Johnson County, Kansas)

In Shawnee, an Algonquian language, (h)oleθí is indeed documented as meaning "pretty" (Gatschet II:2, II:6, III:5); the root also seems to mean "good", judging from its occurrences (spelled <lafi>) in Alford's Shawnee New Testament translation, eg in Matthew 5:45, 19:16, 20:15. One might assume the Shawnees had their own name for the place, but that is not necessarily true, considering they had gotten there barely a generation earlier. Originally from Ohio, they were induced to sign a treaty to move to Kansas in 1831, onto land originally belonging to the Kaws (Kanzas). A few years after the foundation of Olathe, they were pushed out again, to Oklahoma.

It thus seems pretty clear that the original pronunciation of the town's name was /ou'leɪθi/, corresponding better with the spelling (cp. "synecdoche"). How did that turn into /ou'leɪθʌ/? I think the answer lies in English sociolinguistic variation. In the 19th century, standard English word-final /ʌ/ was often pronounced dialectally as /i/, yielding forms like "Americkee" for America or "Canadee" for Canada. In more recent times this pronuciation seems to show up mainly in caricatures of rural or Appalachian speech. The current pronunciation of Olathe as if it were Olatha can thus best be understood as a hypercorrection by people who didn't want to sound uneducated.

Update: A very helpful article linked by Y below, The Pronunciation of Missouri, reveals that the phenomenon is more systematic in the area than I had realised: it extends not only to placenames like Missouri, but even to words like spaghetti, macaroni, or prairie. This makes hypercorrection seem a less likely explanation. Instead, it looks as though final /ɪ/, which becomes /i/ in standard American English, was instead reduced to schwa in parts of the Midwest, including the area surrounding Kansas City. Andrews' (1994) Shawnee Grammar indicates that Shawnee /i/ was often realised as [ɪ], so this fits together nicely.

17 comments:

David Marjanović said...

Fascinating.

Guillaume Jacques said...


The final -eθi in holeθi seems to be from the VAI *-esi- final; I have not checked any Shawnee dictionary, but I surmise that this VAI verb means "pretty, beautiful" of an animate entity, probably not the most appropriate term for a piece of land. The initial stem is probably from *wal-, the same you have in the name of Oregon: http://www.ohs.org/research-and-library/oregon-historical-quarterly/upload/OHQ_105_2-Goddard-Love-Oregon-the-Beautiful.pdf

Those interested in Shawnee historical phonology can have a look at http://research.library.mun.ca/920/3/Muzaffar_TowhidB.pdf

Y said...

Wouldn't someone wanting to sound educated have used a spelling pronunciation?

For more on this final lenition, specifically in the context of Missouri, see
http://missourifolkloresociety.truman.edu/Missouri%20Folklore%20Studies/THE%20PRONONCIATION%20OF%20MISSOURI.htm . The author seems skeptical of hypecorrection explanations for words like sody, but I haven't seen his paper on that subject.

Chillicothe, Ohio (pronounced with a final /i/, not /ǝ/), is from a the name of a Shawnee group which I have seen spelled as Chalahgawtha or Chalakatha. Since the final vowel is not ǝ, I'm not sure if this can be explained as another instance of final /i/ lenition.

Y said...

(Forget the last sentence.)

Alexander said...

In the 19th century, standard English word-final /ʌ/ was often pronounced dialectally as /i/

Incidentally there is a similar phenomenon in some dialects of Yiddish (in some Lithuanian and Polish cities) where word-final schwa is pronounced as /i/.

Y said...

The town of Leoti and Miami county, both in Kansas, are supposedly pronounced /liˈoʊtʌ/ and /mǝˈjæmʌ/.

Anonymous said...

I think some words from German ending in "e" may be anglicized with the /ʌ/ vowel (even though that's more like German "er"). I'm thinking of the usual pronunciation of "Goethe" as "Gerta", and the two-syllable pronunciation of "Porsche".

Lameen Souag الأمين سواق said...

Thanks everyone for some great comments!

Guillaume: That seems plausible; -eθi does seem to be a Shawnee VAI final, and that would be compatible with Gatschet's Huthámi huléthi "she was too pretty". On the other hand, Gatschet also records olethí "pretty, tumblers, glass (tissues)"; I know Algonquian has a number of irregular animates, but I don't remember glasses being on the list.

Y: After looking at the data presented in that article, I agree that hypercorrection is not an adequate explanation for a phenomenon that turns out to be quite systematic. Rather, it seems that American English, like British RP, used to allow final [ɪ], and that some dialects - notably, ones in a large circle around Kansas City - reduced the latter to a schwa.

Anonymous: Good point. I had actually wondered whether German-American influence might be involved here, but I rejected that hypothesis because I figured the θ wouldn't have survived in that case.

Lameen Souag الأمين سواق said...

Guillaume: Interestingly, I notice that in a late 18th-century Shawnee dialogue recorded by Butler we find the spelling "wil-li-thie" (glossed as "handsome") for later holeθi. That seems to fit the *wal- etymology.

Lameen Souag الأمين سواق said...

Also, it appears that the pronunciation with final [ʌ] was already in place by 1893, when we find the name spelled "Oletha": http://shawnee-bluejacket.com/uploads/3/4/8/4/34847868/volume_3_page_5.htm .

John Cowan said...

I think the Missoura-accent is accounted for by phonological rule ordering. In many AmE varieties (including mine), final [ɪ] > [i] (happy-tensing) and then all remaining unstressed [ɪ] > [ə] (the abbot-rabbit merger). But in this accent, the merger applies first, bleeding happy-tensing.

Y said...

When Leoti was founded in 1885, it was apparently not allowed to have its chosen name at first, because there was a Leota post office elsewhere in Kansas at that time. Perhaps the pronunciation had an [ʌ] right from the beginning.

Leota is a woman's name, which I suspect was pronounced Leoti by some final-a-raisers, so the story is more akin to that of Chillicothe than that of Olathe.

petre said...

"standard English word-final /ʌ/"

I'm not aware of any such in British English, except as a conscious imitation of Southern US or other foreign speech. We have [ə] of course, in many contexts an allophone of /ɪ/. I can't work out whether my failure to understand what you're getting at is just a matter of different transcriptions, or whether there's a substantial point that eludes me. Elucidate!

Lameen Souag الأمين سواق said...

JC: That sounds good, but I'm not sure it works out: surely they don't go so far as to say "happa" for "happy"? Or has the change been reversed on common words by the influence of mass media?

Petre: Yes, RP does not have final /ʌ/ as far as I recall. In American Standard English, however, the two "a"s of Ithaca are not pronounced the same: the first "a" is [ə], the last is [ʌ].

protouralic said...

Re the updated position: hypercorrection and phonetic regularity surely do not need to rule each other out! This would not the the first time when a dialect, upon increasing exposure to a prestige variety, ends up "inverting" some of its characteristic features. I can't claim to have detailed statistics on this, but my rough feeling is that this is particularly common when a basilect represents phonetic structures A and B of an acrolect the same, as A. If realizing acrolectal B as A (such as /-ə/ as /-ɪ/ ~ /-i/) ends up as a stigmatized feature, speakers may end up adopting an "inverse" sound change A > B (/-i/ > /-ʌ/)… while still not actually making the contrast (and now thus having B also for acrolectal A).

(I happen to have just read earlier today of a neat exact example of this, from Torne Valley Finnish as described by Martti Airila in 1912. Here A = word-initial uo-, yö-; B = word-initial vuo-, vyö-. The contrast is marginal but established in most varieties of Finnish: the only native examples of type A are 'night' and yökätä 'to gack, to vomit', while a few additional examples such as the personal name Uolevi 'Olaf' are found in loanwords in some dialects. Examples of type B are more common, e.g. vuosi 'year', vuotaa 'to leak', vyöryä 'to roll'. Around 1910, some older speakers of Torne Valley Finnish had uniformly no v- in either category — but most speakers seemed to have shifted to uniformly v- in both. Airila remarks that this brings the dialect closer to standard Finnish, although it does have the effect that now it's instead the few exceptional forms like vyö 'night' that stand out as a dialectal shibboleth.)

John Cowan said...

I think Protouralic is right, and hypercorrection can become so pervasive that it is a regular rule rather than an irregular change. Initial /h/ in the English of England (not Scotland or Ireland) was pervasively lost by 1550 or so, and then restored almost everywhere by hypercorrection from the spelling, so that the few unhypercorrected words like heir, hour, honor, honest stand out. AmE adds herb to this, but hypercorrection has given it /h/ in England today. If we didn't know this history, we'd assume that initial /h/ > zero was a rare sporadic change.

As for "happa for happy", they may in fact have said that — and then the change was missed by abbot-rabbit-merging field researchers before happy-tensing! I've noticed that it is much harder to notice how you or others pronounce reduced vowels: they are out of psychological focus.

Lazar Taxon said...

@petre @Lameen: From what I've read and observed, a lowered value of final /ǝ/ occurs in BrEng too – even all the way down to [a] in traditional Cockney.