This is one of the more simpleminded questions I’ve posted, but I can’t find an answer to it, so I turn to the Varied Reader. Wikipedia sez:

The Lozi people are an ethnic group primarily of western Zambia, inhabiting the region of Barotseland. […] The Lozi are also known as the Malozi, Silozi, Kololo, Barotose, Rotse, Rozi, Rutse, or Tozvi. The Lozi speak Silozi, a central Bantu language.

The word Lozi means ‘plain’ in the Makololo language, in reference to the Barotse Floodplain of the Zambezi on and around which most Lozi live. It may also be spelt Lotse or Rotse, the spelling Lozi having originated with German missionaries in what is now Namibia. Mu- and Ba- are corresponding singular and plural prefixes for certain nouns in the Silozi language, so Murotse means ‘person of the plain’ while Barotse means ‘people of the plain.’

OK, so if “the spelling Lozi […] originated with German missionaries” and the forms referring to the people are Murotse and Barotse, then it stands to reason that “Lozi” = Lotsi (or Lotse?). And indeed, “The Lozi are also known as the […] Rotse.” But both Oxford and Collins give the pronunciation as /ˈləʊzi/, with a -z-. Now, you can say that’s just anglicization, but that doesn’t happen with other German-origin words, like Nazi (/ˈnɑːtsi/). So how is it pronounced by those who actually deal with the language? And if it’s /ts/, why do most of the “also known as” forms have -z-? Is there a dialect difference (as presumably with the l/r variation)? What’s going on here?

Update. I’ve just found an extended discussion in Paul S. Landau’s Popular Politics in the History of South Africa, 1400-1948, pp. 54-56. He’s talking about the Rozvi:

We do not know much about the next few generations on the highveld; but much later, at the end of the nineteenth century north of the Molopo and Limpopo, a Rozvi elite still claimed the right to install new chiefs over a wide area. A presiding official proclaimed, “I am the Rozvi who stands here today.” This sounds very like the ceremonial stature recognized for the “BaHurutshe” of the highveld. Often called “the senior Tswana tribe” in the ethnic CP model, their history also extends at least back to the 1600s. South of the Limpopo, the first part of the Rozvi’s statement, above, would most likely be “Ke morotzi,” or perhaps (if far enough south) “ke mogarotse” (cf. mohu-rutshe), “I am he (or they are they) of the place of Roz[v]i.” Words differed in their pronunciation across speech communities, even nearly adjacent ones. Thus “the place of Roz[v]i” appeared as Ga-rotse (alt. Ha-rotse) and Fhoo-Rotshe or Khoo-Rotshe, the latter two signifying distance (foo and koo). One finds Harootsi, Kurutse, and Fhurutshe recorded on the highveld before the synthesized and standardized name of the twentieth century, “Hurutshe” or “BaHurutshe,” emerged as the “correct” spelling and became known as the parent tribe of the others.

In a list of thirty highveld “tribes” provided by Thomas Hodgson in 1823, the number 10 is “Bamarotsi.” Is this the same as “Hurutshe”? Yes, it is. Burchell in 1812 offered Mõrútzies, and Chief Mothibe, in 1813, told Campbell of the Marootze. Later, Andrew Smith in 1839 offered “Baarootzie,” and elsewhere Smith says he heard “that all the Bechuanas were [once] govered by the Barootzi king” (ba-rootzi) and that circumcision began with the “Barootzie” [sic]. Rotse spoken in one place is Rootzi in another. Lichtenstein, a naturalist who prided himself on precise representation, wrote “Muchuruhzi” on his map, probably a rendition of mo (person) goo rozi. He located these people to the far north and they, with hardly more specificity possible, said their forebears came from north of that.

On the highveld, ga– or ha-rotse (rotse-place) people instantiated the centralizing trend of the 1600s. [There follows a table with the Shona form varozvi and a whole variety of Sechuana forms, including barotse and marotse as well as some of the ones mentioned above.]

Most likely Rozvi (rozvi) is the initial template for this rutzi, ruhzi, rotse, and so on, on the highveld. There may also, however, be a loan-term behind all three that we do not know. Folk etymology on the highveld points to the melons of the first fruits ceremonies, marotse. In some areas rotse is a verb in the perfect form meaning to have unloaded a burden or a present. In Shona, “Rozvi” may have an early connotation as “defrauder” (conjugated from the verb, roza). Perhaps the stimulation of Indian Ocean commerce early on bequeathed a spoken morpheme to Africans; perhaps the source was even “Shirazi,” a word heard along inland routes from the Swahili coast and so appropriatable like any other — but this cannot be considered likely.

Whew! That’s more than you ever wanted to know about this obscure term, but at least I’ve collected a bunch of data and speculation in one place.


  1. Nazi is the exception, not the rule, I think, and even it is/was pronounced with a [z] by some (notably Col. Hill, Hank’s father in King of the Hill). The names Zimmer and Zimmerman as well are pronounced with a [z] by US anglophones.

  2. True, but those are family names, which are bound to get worn down over the years by frequent contact with non-German-speakers. A specialized term like this is presumably used mainly by those who know something about the tribe/language, and if z = /ts/ you’d expect them to say it that way. And maybe they do! That’s why I’m asking.

  3. George Fortune’s “Outline of Silozi Grammar” makes it look like the language does have a /z/ (described as a ‘voiced alveolar spirant’, and written with the letter ‘z’), and lacks a /ts/. He spells other words with ‘z’, as well as words like Silozi (which refers to the language), Balozi (which is the plural word for Lozi people), and Malozi (which refers to the nation)–this is standard Bantu noun-class stuff, I think. I guess it’s possible that ‘Lozi’ is just a misleading spelling established by German missionaries before the rest of the orthography got nailed down–but given that he describes the language as lacking a ‘ts’ sound, I sort of doubt it. Seems more likely that ‘Lozi’ really is their name for themselves, and that ‘Rotse’ is somebody else’s (maybe etymologically related) word for them.

    aha! moreover, here is a thesis about the current state of the language:


    …in which, on p. 6, the author speculates about the origin of the name ‘Silozi’, in a way that seems to make it clear that the z is a z: “Another possible explanation is that the “r” in Barotse became an “l”, a “ts” a “z”, and the “e” an “i”, in keeping with common phonetic processes.”

  4. Those family names have the confounding issue that the Z comes first. English speakers seem to have trouble with /ts/ starting words, even in examples like tsunami. OTOH other German and Italian names of Americans also seem to undergo this transformation.

  5. On other “z” words:

    Zeitgeit is usually pronounced |zaitgait|

    Luzern is usually pronounced |lusɜːn|

    All this points to the usual (lack of) logic and consistency that we have come to expect of our gloriously wonderful English language.

  6. marie-lucie says:

    As in Frank Zappa.

  7. @Y: That might just be the natural anglophone resistance to /ts/ in initial position. The names Schwarz and Schulz reliably take /ts/ in AmEng – though that might be because we’re accustomed to the spellings Schwartz and Schultz (which are more common here, but less common in Germany). Leibniz takes /ts/ as well – though names like Belzer or Danzig will typically take /z/. So… a mixed bag, I suppose.

  8. PuzzleCraig says:

    @zyxt, I think you mean that Zeitgeist is usually pronounced |zaitgaist| in English. Looking around online, I find many dictionaries that say that the “ts” pronunciation is common, and also that capitalization (Zeitgeist vs. zeitgeist) isn’t consistent.

    Luzern is usually spelled in the French way as Lucerne which accounts for the |lusɜːn| pronunciation.

  9. Norvin: Thanks very much, that seems to clear things up nicely! I suppose I should edit the Wikipedia article to get rid of the misleading “the spelling Lozi originated with German missionaries” bit, but I’m feeling lazy.

  10. ə de vivre says:

    Oh Lord, stuck in Lozi again…

  11. @ Y Nazi is the exception, not the rule, I think, and even it is/was pronounced with a [z] by some

    Yes, I noticed in the movie “Churchill”, Churchill himself (Brian Cox) consistently used [z], whereas everybody else /ts/ (as is usual in the UK). Presumably this follows Churchill’s recorded speeches.

  12. Alex M. says:

    Pre-war Nazi with [z] was the standard UK pronunciation as can be found in most sound recordings, not just Churchill. The [ts] pronunciation came to dominate during the war and especially post-war. I have always assumed this was due to the prominence of US news reels which only used [ts].

  13. “He couldn’t even say Nazi. He would say Noses, Noses!

  14. Marianne Stromschläger says:

    FWIW, Dutch has [z] in nazi.

    PuzzleCraig: “Luzern is usually spelled in the French way as Lucerne which accounts for the |lusɜːn| pronunciation.”

    That is all well, but then we have oddities like the non-Frenchifried spelling Leipzig for which Wiktionary gives an “anglicized” pronunciation of /ˈlaɪpsɪɡ/, in which the /s/ nicely matches Leipsic (“an older spelling of the name in English”), which nicely matches French Leipsick and (post-Medieval?) Latin Lipsia, while the /g/ matches how speakers unfamiliar with German phonology would pronounce the last letter; I don’t know whether any variety of German ever used /g/ except in Leipziger etc. Apparently, the /s/ pronunciation has survived despite the matching spelling having fallen into disuse.

  15. marie-lucie says:

    French Leipsick ????

  16. From Google Books:

    “Les trois foires qui se tiennent à Leipsick, au nouvel an, à pâques et à la St. Michel, sont célèbres” [1811]

    “Cette prospérité, Leipsick la doit à son industrie et surtout à son commerce.” [1862]

    “LEIPSICK (500.000 habitants) Leipsick, ville saxonne, où siège le tribunal le plus élevé de l’empire allemand, le Reichsgericht, présente de multiples intérêts musicaux que nous examinerons les uns après les autres.” [1908]

    “Balzac devait revenir chercher Mme Hanska à Leipsick le 6 décembre, pour l’amener incognito à Paris.” [1933]

  17. Perhaps I missed it, but I haven’t noticed anyone ask zyxt how his nom-de-plume is pronounced, if ever. Not to mention meaning or source.

  18. I always just said /zikst/ and never gave it a second thought. But now that you’ve brought it up, I’m curious too.

  19. Tseekst, I vould sink.

  20. It’s the last word in the printed OED. Old Kentish for seest (2sg. pres. of see).

  21. Ah, of course! And now that you say that, I realize I knew it at one point. Sigh.

  22. It’s now under see in the OED Online, without a separate entry. The forms of the 2SG present given are OE seohst, syhst, sihst, sixst, siist, síst, ME sihst, sichst seh[s]te, seost, sext, syst, sucst, sikst, sixst, sest, sist, sixt, suxst, sixte, syxt, ( sys), Kent. zixt, zyxt, zist, suxt, sexst, sestt, seist, ME–15 seyst, ME– seest.”

  23. Presumably the y is a tense [i].
    In what varieties of English is a post-vocalic OE *h regularly hardened to something else?

  24. Isn’t the hardening here due to the following /s/ (as in “six”)?

  25. marie-lucie says:


    I am not surprised to see this spelling in older documents but I am surprised by the 1933 one. I guess that this quotation, which mentions an episode in the life of Balzac, must have reproduced a spelling from his time rather than a modern one. Personally I have never seen the city called anything but Leipzig in French documents I have read.

  26. Lars (the original one) says:

    FWIW, Danish has [næːsɪ] for Nazi. And they had the country occupied for 5 years so it’s not like Danes didn’t know how the Germans said it, but willingness to use a foreign sound was probably low. Now wondering if Norwegian even spells it with an {s}.

  27. Google’s Ngram viewer shows -sick, -sic and -zig all competitive in French until around 1830, then about 2/3 -zig to 1/3 -sick through the mid 19th, then a total dominance of -zig thereafter. On the other hand, Brunswick still appears to be current in French. Is the -ick in these is a sort of faux-Germanism?

  28. J.W. Brewer says:

    At least three different U.S. states have a town named Leipsic, spelled thusly. The most recently-named of the three was (according to the internet) platted circa 1857. Their founders, if asked, might well have thought “Leipzig” pretentious. I mean, it’s Florence, Alabama, not “Firenze,” innit?

    I grew up barely 50 miles from one of these Leipsics, but I’m not sure I ever set foot there — it’s a small place and off the main roads. On the other hand I must have crossed the Leipsic River quite a few times in my youth (the bridge where Route 13 goes over it is maybe 5 miles upstream from the town of the same name).

  29. Trond Engen says:

    Lars 1.0: Now wondering if Norwegian even spells it with an {s}.

    No. Nazi- [‘na:si-] as a prefix and nazisme [na’sisme] and nazist [na’sist] as nouns. But nasjonalsosialisme.

  30. Trond Engen says:

    I should perhaps have marked those as phonemic transcriptions, not phonetic. But the difference is only in the quality of unstressed vowels and rarely cared about.

  31. David Marjanović says:

    I don’t know whether any variety of German ever used /g/ except in Leipziger etc.

    Those without Auslautverhärtung (syllable-final fortition) do, so basically those south of the White-Sausage Equator that haven’t collapsed /k/ and /g/ completely. Both of mine are examples.

    In what varieties of English is a post-vocalic OE *h regularly hardened to something else?

    I’m sure we’re looking at Verner’s law instead. Dialects like mine preserve the *x ~ alternation in this word, and at least some people in Vienna seem to have generalized the /g/, including 2sg /sigsd/ in the mesolect.

    Is the -ick in these is a sort of faux-Germanism?

    Actually, Braunschweig is a High German folk etymology (“brown” + “not speak”…) of Low German /bru:nsvi:k/ *”Bruno’s village”…

  32. Lars (the original one) says:

    Brunsviger — however the baked good seems to have been invented in Copenhagen and the name picked to match the color. The ODS ignores the LG form and lists the word as danified from HG. which may well be true given late attestation in Copenhagen. What goes around, comes around.

  33. January First-of-May says:

    Their founders, if asked, might well have thought “Leipzig” pretentious.

    IIRC there’s a place somewhere in the US (I forgot where) named Rolla, whose founders supposedly wanted to name it for Raleigh, North Carolina, but thought that the spelling “Raleigh” was too pretentious.

  34. Rolla, Missouri. Quoth WP:

    Two stories account for how Rolla was named. One story, widely regarded as a folk legend, and acknowledged as such by the Phelps County Historical Society, arises from the competition between Rolla and neighboring Dillon, Missouri, to be designated the county seat. When Rolla was made the county seat in 1861, the residents of Dillon, having lost a round, were allowed to choose the name of the new city and named it Rolla, after a good-for-nothing hunting dog.

    The more widely accepted story came from a citizens’ meeting about naming the town. Webber was said to prefer the name Hardscrabble (which was used to describe the soils in the region) and Bishop [the founder, a railroad contractor] pushed for the name Phelps Center. New settlers from North Carolina voted to name the community after Raleigh, their hometown, but chose to spell the Missouri version phonetically.

    (FWIW, my wife, who grew up in Raleigh, says /rɑli/, but the Missouri form may be the same “Weak Vowel Merger before happy-tensing” rule ordering that gives us /mɪzurə/.)

  35. Marianne Stromschläger says:

    DM: Thank you, I didn’t know there were any without AV.

  36. David Marjanović says:

    I didn’t know there were any without AV

    Just to be sure – this doesn’t mean /b d g/ are voiced under any conditions. They’re voiceless; they’re just not fortis. Talk “talcum” and Talg “tallow” are a minimal pair for me, but you’re perhaps not likely to notice from hearing me speak if your ear is only tuned to English, and certainly not if you’re used to Russian instead. (Indeed, a Russian once understood Nathaniel when I said Daniel.)

  37. January First-of-May says:

    I think we’ve discussed, at some point, whether it is possible for someone to produce phonemic differences that they don’t notice themselves… can’t recall if we ever got an answer.

    My own idiolect of English distinguishes between /θ~t͡θ/, usually spelled th (the usual English dental/interdental consonant), and /t͡s/, usually spelled c (originally a holdover from Russian, in international words such as cent, but found its way into native English words, such as scent; slowly normalizing to /s/).
    The traditional minimal pair I give is think and zinc – the latter due to Russian цинк; for a while I proposed thimble and cymbal, but by now the latter had apparently ended up in the “normalized to /s/” category.

    However, I had no idea they were even different – believing that I simply pronounced English /θ/ as [t͡s] – until, about a year ago, I happened to notice that my tongue wasn’t doing the same thing in those consonants.
    (The rendering [t͡θ] – a sound that is, in fact, attested in some American English dialects – is my interpretation of the reason for this perceived similarity to [t͡s].)

  38. Very interesting!

    By the way, the other day I heard someone from Angola interviewed and she was speaking Lozi (I actually heard a couple of words before the interpreter’s voice overrode it); considering how rarely I’ve ever encountered the name, let alone the language, I thought that was a remarkable coincidence.

  39. David Marjanović says:

    whether it is possible for someone to produce phonemic differences that they don’t notice themselves…

    I think so, if they’re used to a spelling system that ignores them and if minimal pairs or rhymes don’t come up often. I was a bit surprised to discover that the “no” word of my own dialect (and many other Bavarian-Austrian ones) isn’t /na/, but /nã/…


  1. […] Hat investigates the deeper etymology of “Lozi”, a people of […]

Speak Your Mind