My wife asked me how many languages use the word for ‘tongue’ to mean ‘language’ — as do, for instance, the Romance languages — and furthermore, why ‘tongue’? Obviously the tongue is involved in speech, but so are various other parts of the mouth, and ‘mouth’ itself might seem a more obvious metonym. I gave her the easy answer, which is that a lot of modern languages do so because Greek and Latin did, but that just pushes the problem back (and did Latin imitate Greek, or is it an Indo-European thing?), so I thought I’d toss it out there for consideration by the Hattery.
A quick look at dictionaries of ancient Semitic languages shows the reflex of the triliteral root lšn (e.g. already Akkadian lišān) was used for both ‘tongue’ and ‘language’, so if anything, both Latin and Greek may have simply taken this from a common Mediterranean culture, if the polysemy isn’t simply attractive to any human society.
In Modern Hebrew, שָׂפָה safa lit. ‘lip’ is unmarked, לָשׁוֹן lashon lit. ‘tongue’ is more literary (both are used in BH). נִיב niv is ‘dialect’, ‘idiomatic expression’, which happens to have the homonym ‘canine tooth’.
both Latin and Greek may have simply taken this from a common Mediterranean culture
Ah, that would make sense.
It’s not a thing in Oti-Volta languages. “Tongue” in derivation goes with “taste” and “lick”, whereas the various words for “language” are usually just generalised from “word, speech.” “Mouth” does figure sometimes, though.
All the languages make names of specific languages by using the ethnonym in a different noun class, either the abstract-mass-noun “water” class.as with “Nawdm”, “Gulmancema”, “Akaselem”, “Yom”, “Mbelime”, or a language-specific class with the proto-Oti-Volta suffix *lɪ. as in “Mampruli”, “Nateni”, “Buli” … “Kusaal.”
I don’t think the languages reify languages in the way that modern SAE does. For example M pian’ad Kʋsaal “I speak Kusaal” is really basically “I speak Kusaasi-fashion.”
(Kusaal pian’ “speak” actually goes back to a proto-Oti-Volta root meaning “praise, congratulate.” Evidently the Kusaasi took on board the maxim that if you can’t say anything nice, you shouldn’t say anything at all.)
For this and similar questions, CLICS (Database of Cross-Linguistic Colexifications) is your friend. Hover over the graph edge of interest for a list and a map. The quality of data is variable, as is common with these databases.
Just browsing through the list of translations on Wiktionary it looks like the list includes
– most Finno-Ugric, Turkic and Mongolic languages
– Albanian
– Chechen
– Georgian
– Zulu
…so mostly an areal feature of central/western Eurasia, but not only.
For this and similar questions, CLICS (Database of Cross-Linguistic Colexifications) is your friend. Hover over the graph edge of interest for a list and a map.
Wow, that’s terrific — I knew the Hattery would come up with something!
Interestingly, Hausa is one of the tongue/language languages.
Unlikely as it looks, the Hausa harshe “tongue, language” is one of the very few words truly traceable to proto-Afroasiatic (“truly” as opposed to Ehret-style free-association-as-comparative-study.) It’s cognate with Arabic lisaan and Coptic las. I wonder if the semantic overlap actually goes all the way back to proto-AA too?
(Not impossible for such things to be remarkably enduring: a “send/work” overlap is found from Kusaal to Swahili and beyond: definitely proto-Volta-Congo. Hausa has it too, perhaps picked up from the neighbours/conquered subjects.)
Five languages (three Chadic, one Gyalrongic, one Dagestanian) use ‘neck’.
If it’s not pan-Chadic, that undermines my proto-AA proposal. And with Hausa, you always have to reckon with Arabic influence, too: the Hausawa could quite well have taken over the metaphor from Arabic.
I wonder if “neck” overlaps with “throat” in the Chadic languages in question? That would make it a lot more comprehensible.
(Throat/voice is another venerable proto-Volta-Congo one: or at least, Oti-Volta and Bantu both have it.)
The Sumerian word eme 𒅴 ‘tongue’ (see here) is used for ‘language’ as well. It matches the use of lišānum in Akkadian, of course. Sumerian for ‘Sumerian’ was emeŋir (see here), for example, but this is apparently not recorded before the Old Akkadian period. I couldn’t find any uses of eme in the meaning ‘language’ in the Early Dynastic Period through a quick search in the ePSD2, but it was only a quick search.
Five languages (three Chadic, one Gyalrongic, one Dagestanian) use ‘neck’.
That struck me as well. (Very guttural languages, the Chadic and the Gyalrongic…)
In Turkish, dil is both ‘tongue’ and ‘language’, while ağız ‘mouth’ can also be ‘dialect; local form of colloquial speech’, as in Diyarbakır ağzı ‘Diyarbakır dialect; Diyarbakır colloquial speech’.
This Turkish use of dil and ağız reminded me of the phrase lingua toscana in bocca romana, dearly loved by non-Italians. I wonder, to a person who has grown up in Italy, does an odor of fascism cling to it?
Xerib: lingua toscana in bocca romana
I thought of that too.
In Turkish, dil is both ‘tongue’ and ‘language’
Clearly, Tetun Dili, spoken on Timor, is Turkic. But what isn’t?
Chadic comes up a lot also on the ‘language’ = ‘mouth’ edge: I count 20 out of their 27 examples, unless one of them is an obscure synonym for something Cushitic. Meanwhile they don’t seem to have caught any of the fairly common Cushitic (and also Omotic I think) cases of forms like afa, afo etc. ‘language, mouth’ though!
The sample is loaded by the use of one particular comparative Chadic wordlist.
In Senthečen a language is referred to as a “throat”. For obvious reasons. I don’t know about how much more widely that’s used in other Salishan language. On the other hand in Lushootseed the term for language is derived from “speak”, though the lexical affix for languages ‘-ucid” (as in “Lushootseed”) means'”opening, mouth”.
German arguably does this, too – and all of Slavic definitely does.
Why would it?
In Persian, language and tongue are the same word (‘zabaan’).
It’s the same in central Victoria, Australia. The Aboriginal word for language and tongue is ‘wurrung’.
Fulfulde has it, I see: ɗemngal “tongue, language.”
(Though I was trying to think of another West African language where this could be plausibly blamed on Arabic.)
It’s interesting, now my attention has been drawn to it, that this is so much not a feature of Oti-Volta, as that’s obviously not a regional pan-West-African thing. I wonder about Volta-Congo in general?
Incidentally, ɗemngal is one of those words that lumpers like to use to “prove” that Atlantic is related to Volta-Congo; POV *lém-, proto-Bantu *-dímè … I remain sceptical. Who hasn’t got a word for “tongue” beginning with some random alveolar?
Talking of which, the definitely-not-Niger-Congo Tondi Songway Kiini has dɛ̀lɛ̀ and Homburi Songhay has dè:nè, both helpfully glossed “langue (partie du corps)”; so it looks like Songhay is like Oti-Volta in this.
Greek γλῶττα / γλῶσσα, apparently ‘pointy’ (and related to γλῶχες ‘awn of grain’), while not cognate with other IE ‘tongue’ words, still means both.
Chinook Jargon lelang took both sense from French, except that the name of that language itself is wawa ‘speech’.
English is listed among languages who don’t have it (but definitely, “language” is one of meanings of English “tongue”).
Also Lithuanian, Belorusian and Ukrainian and whole of Germanic, so it is Finns who look unusual now.
“zabaan” – To my horror, Wiktionary says, in Yemeni Arabic it means “bum”.
It’s not an all-or-nothing thing: in a lot of languages, “tongue” can mean “language”, but there are other perfectly good words for “language” too. English, Arabic, Welsh …
In other languages, “tongue” can’t mean “language” at all (like Kusaal), and in others it’s the normal word for it (like French.)
I dare say that there are languages in which using “tongue” to mean “language” is comprehensible but comes across as a strikingly poetic metaphor, and others in which it’s still felt as a metaphor but is a tired old trope.
Why should it?
Indeed. After all, there is nothing fascist about lingua tedesca in bocca irlandese, meaning the most widely spoken language in America the Nord.
In Irish, teanga “tongue” has displaced béalra>béarla “mouth-ery”, which is now elliptic for “Saxon mouth-ery” aka English
“Mouth”/”edge” is actually common in Ghana. I took a long time realising that the “mouth” of a river in Kusaal is in fact its bank. (I’m actually not sure how to render “mouth of a river” in Kusaal. The matter doesn’t really arise. I think you’d just have to paraphrase: “place where the river flows into the sea”; “sea”, naturally enough, is a loanword. From Hausa. Which seems in turn to have borrowed it from somewhere else,)
(I’m actually not sure how to render “mouth of a river” in Kusaal. The matter doesn’t really arise.)
Isn’t it the same as ‘tongue of a glacier’?
No doubt. These are deep waters.
Kanuri (a language which I know from nothing) seems to be another one: tə́lam “tongue; language.”
That also seems to obligingly fit two of my theories: (a) it’s a Muslim thing in West Africa and (b) everyone likes words for “tongue” that start with alveolars.
Not Samba Leko (an Adamawa language): mɛ́l is “tongue” as in “body part” and w̰ɔ̄ŋ is “language.”
(Pity about the m- of mɛ́l. The word is obviously from lɛ́m by metathesis. Work those epicycles!)
@DE, they just speak backwards. It is mɛ́l, but underlyingly lɛ́m.
. I wonder, to a person who has grown up in Italy, does an odor of fascism cling to it?
No, that would be lingua toscana in bocca romagnola
most Finno-Ugric, Turkic and Mongolic languages
Yes, in Mongolian it is хэл ᠬᠡᠯᠡ (hel) for both.
I should add that I’ve heard Монгол үг / ᠮᠣᠩᠭᠣᠯ ᠦᠭᠡ (Mongol üg) “Mongolian word” in Inner Mongolia but I don’t know if it’s regarded as strictly correct, and I know it would certainly not be regarded as correct in Mongolia.
I dare say that there are languages in which using “tongue” to mean “language” is comprehensible but comes across as a strikingly poetic metaphor, and others in which it’s still felt as a metaphor but is a tired old trope.
In German it’s part of the traditional poetic register and can also be found as a mildly exoticising device in historical novels and adventure / travel literature, with natives saying things like “he speaks the White man’s tongue”. I guess that is similar to English usage. In any case, more well-known trope than striking metaphor.
… and ‘mouth’ itself might seem a more obvious metonym
All of this is more metonym than metaphor, for that matter.
Paul Newman did an article about how “tongue” for “language” is basically foreign to Chadic, which prefers “mouth” or the like; even today, they say “Hausar Baka”, Hausa mouth.
Neither Korandje nor Siwi use anatomical metaphors for “language”, preferring “speech”. I think that’s typical for North Africa; it certainly is for Songhay.
Where does the Hausa teku “sea” come from? I think it has to be a loanword, for phonological reasons (short final vowel.) I can’t think of an obvious source, though.
All of this is more metonym than metaphor, for that matter
But metonymy is a metaphor, so it’s back to square one.
Given that defining “language” almost universally implies contact between languages and cultures, I would expect a good deal of borrowing and malleability in words used to designate languages, dialects, and language families. But chopping off the tongues of people who spoke wrong things was the thing of the old ages, so I also wouldn’t be surprised that tangues (rather than lips, teeth, or throats) repeatedly emerged as metaphors for speaking and languages
I mean, yes, chopping off tongues was a thing, but I don’t think it was common enough to influence this kind of metonymic usage.
But perhaps the combination – cutting off tongues, sewing lips together, knocking teeth in and cutting throats – contributed to this kind of metonymic usage. The tongue is still able to function when lips are sewn together, and without any teeth – and a cut throat merely changes the subject to death toot court. While life is still present, the tongue must be the root of the problems – whatever these may be.
Tocharian B kantwo means both ‘tongue’ and ‘language’, so I’d be more inclined to look either to IE inheritance or repeatable metaphor than to excessive Hellenistic influence. You could, I guess, explain the double meaning in Germanic, at least some Slavic languages, and Armenian (and maybe others) as due to Greek influence via Christianity, but I think it might be pushing it to insist on that.
I was reluctant to use to the Germanic double meaning as evidence not because of Greek influence through Christianity, but rather due to the Latin influence from contact with the Roman Empire. The first substantial attestations of Northwest Germanic already show an array of Latin borrowings.
MMcM, that Greek reference you cite is ancient. The standard reference now for Greek etymology is Beekes’ Etymological Dictionary of Greek, which is freely consultable at LibGen or Anna’s Archive. (Though in this case Beekes does uphold the same etymology.)
Thor sewed together Loki’s lips to shut him up. (He had wanted decapitate Loki, but while Thor had won Loki’s head in their bet, he did not own any of Loki’s neck.) I would guess that the reason the myth has Loki’s lips sewn up, rather than his tongue removed (which would be more effective), is that the sewing is at least plausibly reversible, so Loki could regain his power of speech in later stories.
There are some Latin borrowings in West Germanic and North Germanic, but far fewer that can be pushed back to Northwest Germanic. In West Germanic, which is attested in substantial amounts much earlier and which mediated some Latin words into North Germanic, the early loans mostly fall into predictable semantic spheres (trade, military, infrastructure, and for presumably later borrowings, religion) — I shouldn’t think a semantic calque on lingua would be very likely at all.
‘Twas Brokkr, not Thor, who won Loki’s head (in Snorri’s telling, of course).
@Xerîb et al.
As one person who grew up in Italy, I see nothing fascist about the phrase, and I doubt other Italians would either. The phrase predates fascism, and I can find no evidence it was picked up as a fascist slogan.
There are plenty of actual fascist mottoes that remain memorable and recognizable — and not always original, as we had recently occasion to note here — but their fascism is much more transparent: Credere, obbedire, combattere. È l’aratro che traccia il solco, ma è la spada che lo difende. Molti nemici, molto onore. Noi siamo contro la vita comoda. Tutto nello Stato, niente al di fuori dello Stato, nulla contro lo Stato …
Not only the phrase doesn’t sound fascist, but neither does the idea. Attempts to define a normative Italian pronunciation have always danced around the definition of non-dialectal Tuscan pronunciation. My knowledge is very limited, but I understand Rome always played a major part in that definition. Not only is it the only major city outside Tuscany natively speaking a central Italian dialect, but its educated speech had been heavily Tuscanized since the Renaissance (especially in the second half of the XVI century). I’m unsure lingua toscana in bocca romana means something unambiguously different from fiorentino emendato.
It’s true that Bertoni and Ugolini’s (1939) Prontuario di pronunzia e di ortografia for fascist radio (the national public radio, then EIAR) systematically picked a Roman pronunciation in the (rare) cases of conflict with the Florentine alternative. However, I presume those cases are mostly differences in vowel openness (such as Florentine colónna vs. Roman colònna), which I suspect a majority of Italians, like me, do not notice anyway.
Even for those who may notice, that’s certainly not what makes fascist radio announcers sound fascist. What does is that their intonation is “solenne, marziale, vibrante di romano orgoglio,” as Scola and his coauthors memorably put it in A Special Day.
I don’t know if an analysis has been performed, but I wouldn’t be surprised if public radio and TV actually sounded more discernibly Roman today than they did during fascism. I’m pretty sure attempts to impose any normative pronunciation have declined over the decades, and plenty of broadcasters naturally have something of a Roman accent of the kind that never was normative.
Does the gorgia toscana imply ‘throat’ as a metaph—, I mean, metonym for ‘dialect’? Or is it some perception of it as “guttural”? (I can’t see how, but other people are other people.)
Sanskrit primarily and famously uses भाषा (bhā́ṣā), from PIE *bʰeh₂- (“speak”) and so cognate with φωνή (“voice”) and much more. We easily find derivatives in Malay, etc. It has no connection with the anatomical tongue as far as I can determine.
Sanskrit has जिह्वा (jihvā́) for the literal tongue: from PIE dn̥ǵʰwéh₂s (“tongue”), so cognate with tongue, lingua, and jezik (Croatian, compare other Slavic). The venerable Carl Darling Buck held early that it was not known to mean “language”: “This use of jihvā́ ‘tongue’ is unknown, but it is represented in the Gypsy chib, jib, and in Afghan jiba, which is a loan word from some Indic dialect.” (Words of speaking and saying in the Indo-European languages: Second paper, American Journal of Philology 36(2) 1915). But it is attested in that role, as well as metonymically for the sense of taste.
As mentioned earlier for Farsi, Hindi uses zabaan for tongue and language. But the Indic word for tongue, jeebh (cf Sanskrit above), is just tongue and bhaashaa is just language.
@Y:
The latter. The name-giving phenomenon is debuccalization of [k] to [h] or [ɦ]. Since standard Italian has no glottal consonants, Florentines producing them have struck other Italians as speaking in their throat.
But the Indic word for tongue, jeebh (cf Sanskrit above), is just tongue and bhaashaa is just language.
Courageous words, Minister.
Monier-Williams includes this at जिह्वा (jihvā́), presumably referring us to the naighaṇṭuka kāṇḍa as commented on by Yāska i, 11:
And the entry in Dictionnaire sanskrit-français (Gérard Huet) supplies this information:
In Tamazight (Berber), Central and Southeast Morocco variants, /ils/ means “tongue” both organ and language. How it relates to Arabic “lisān” Heb. “leshon” is a question for the historical linguists.
Finnish has kieli, tongue, language and also a (violin etc) string. I’d like to think that tongue is a better metaphor than the mouth, as everyone has a visibly similar mouth. Only when people open it to speak, it turns out that many different sounds and languages can come out, this obviously has something to do with that not so visible flexible and sensitive hunk of flesh within.
@Giacomo, wow, thanks!
It seems, WP has an article https://en.wikipedia.org/wiki/Tuscan_gorgia
Sadly, I am not aware of a Russian name for fricative [ɦ~ɣ] in Slavic.
Other than ɦekat’ “to ɦ/ɣak”, and “Ukrainian ɦe”.
Very guttural ideology, the Fascist … Turns out btw that the Italian edition of wikiquote has helpfully compiled >250 allegedly Fascist catchphrases for your edification (equally useful if you want to use them or if you want to self-consciously avoid them, I should think?), although I expect they vary in ideological specificity and the extent to which they were also in pre-Fascist (and perhaps post-Fascist) use as well. https://it.wikiquote.org/wiki/Slogan_fascisti
“Me ne frego”?? Come on now, you’re just trying to boost the count.
Google translate of the footnote to “Me no frego”: “Although attributed to Gabriele D’Annunzio, the slogan probably spread among the Arditi during the First World War and the subsequent Fiume Enterprise. It originates from the writing that a wounded soldier had placed on his bandages, as a sign of total self-sacrifice for his homeland.” I assume the Ramones probably picked it up (in translation) from D’Annunzio …
In many languages in Papua New Guinea, ‘language’ = ‘talk’, as in Tok Pisin (PNG Pidgin/Creole) or tok ples (village language).
Speaking of all the theories of calquing ‘tongue’ to also mean ‘language’, it might be fun to mention here that I theorize the opposite of this to have happened for Hungarian nyelv and its Ob-Ugric and also Mari cognates. These mean again ‘tongue, language’, but etymologically it’s an action noun derivative of Proto-Uralic *ńälə- ‘to swallow’ (Hu nyel) , which might better seem to suggest a meaning like ‘mouth’ or ‘throat’ (‘a swallowing, a mouthful, a throatful’). The meaning ‘tongue’ in Proto-Uralic seems to be taken up by the family of Fi. kieli (lost in Hungarian, reflexes in Ob-Ugric fully abstracted as ‘word, language’). Also, ‘mouth’ is even what the Sami cognates (North njálbmi, etc.) still mean. So I think it started off as anatomic ‘mouth / throat’, then shifted directly from that to meaning ‘language’; and then later on was de-metaphorized to be also ‘tongue’ after the example of well, everyone else; perhaps set off e.g. by contact with Turkic.
@languagehat:
Not at all! Me ne frego was a prominent fascist motto. The entry for Fascismo in the 1932 Enciclopedia Italiana, allegedly authored by Mussolini, cites it twice:
The second citation refers explicitly to the unambiguous pre-fascist adopter of the motto, Gabriele D’Annunzio in his Fiume expedition. As a consummated propagandist and especially self-propagandist, he naturally wrote about it in his 1920 pamphlet Il Sacco di Fiume:
The first citation references instead the claim that during World War I the Arditi special forces would write the motto on their bandages. This is an enduring claim, which Capurso (2011) reports at face value. I’m more skeptical, if only because I suspect the regular army wouldn’t have shared D’Annunzio’s appreciation for the propaganda value of of crudeness (“But their commanders won’t allow them to write fuck on their airplanes because it’s obscene.”).
Understanding that “Me ne frego” was a squadrist motto remains widespread, as a Google image search conclusively demonstrates. I don’t know how to check if this lingering memory reduces contemporary usage of the phrase, which subjectively sounds much less common to me than to other conjugations (Chi se ne frega? Fregatene!).
Re, earliest attestations of “eme” as “language”. The majority of the pre–Old Akkadian attestations are neither literal tongues nor languages, but ploughshares from economic texts dealing with wood – tamarisk seems to have been the wood of choice for making eme. (Akkadian even borrowed this sense of eme as “emû”, ploughshare.) This is probably a function of what genres of texts have survived, since it isn’t until the second millennium that literary texts become somewhat common.
There are mid-third-millennium literary and lexical texts that contain ambiguous uses of eme. One incantation describes the god Enki as “eme-gal2 inim mul(x)”: literally “eme-having, [with] brilliant speech”, which could either mean that he has a tongue that’s good at speaking or he possesses good language.
Not at all! Me ne frego was a prominent fascist motto.
The things I learn around here…
Re, earliest attestations of “eme” as “language”.
I was hoping you’d show up, ə!
Menefreghismo at Language Log, sparked by Melania Trump’s “I really don’t care” jacket.
That’s a very enlightening discussion, thanks.
So I think it started off as anatomic ‘mouth / throat’, then shifted directly from that to meaning ‘language’; and then later on was de-metaphorized to be also ‘tongue’ after the example of well, everyone else; perhaps set off e.g. by contact with Turkic
Can’t it have gone from “mouth, throat” to “tongue” and then to “language”?
@@: Ploughshares could be metaphorical tongues. Are there other “tongue” words that are attested earlier?
I see I also invoked the Ramones in that 2018 Log thread, so I am clearly at this point in my life working with a finite set of possible free-associations …
Eme is also used to mean ‘blade’, as in daggers or spears (IIRC, axes have lips, which I don’t think are used for language-adjacent metaphors). So ploughshares and blades both seem to be the same metaphorical extension of the body-part word.
Kag, ‘mouth,’ is written with the same sign as the words for ‘word/speech’ and ‘to say’ (often nominalized into ‘sayings’ or something along those lines), so it’s not always clear if ‘mouth’ is being used metaphorically or if the sign should be read as one of its more abstract values. In ritual texts at least, mouths are usually more culturally significant as the locus of eating than of speech per se.
I don’t know if anyone has tried to make any etymological links between gu2, ‘throat, neck’, and gu3, ‘voice, shout.’ They’re very consistently spelled with different signs, but they are suspiciously homophonous (or maybe they weren’t, vowel length was almost certainly present in Sumerian, and the vowel ‘u’ is so common that it’s likely there’s at least a /o/ hiding behind some of the graphical ‘u’).
In theory yes, but looks unlikely since I’ve yet to find examples of a direct ‘mouth’ or ‘throat’ ↔ ‘tongue’ or similar semantic shift (none collected at CLICS, for example).
A striking difference being that “mother tongue” is just Muttersprache, so Zunge for “language” really is restricted to poetic domains.
Nelson Goering & Christopher Culver (and other interested hatters)-
In answer to Nelson Goering’s 4:25 posting yesterday, a “semantic calque on lingua” may or may not have been that likely in the early days of Latin-(West) Germanic contact (but see my next paragraph), but considering how ubiquitous reflexes of “lingua” in both meanings are in written Latin and spoken Romance alike (including, we must assume, in the now extinct Romance varieties replaced by West Germanic between the eighth and tenth centuries), I see no reason to assume this semantic calque must have entered West Germanic once: on the contrary, this calque could have entered various West Germanic varieties at multiple points in history (long after it had broken up into the ancestral dialects of the attested West Germanic languages), possibly from quite different sources (Latin as well as different Romance substrate/adstrate languages).
Mark you, I could well believe this calque to date back to Proto-West Germanic: if Don Ringe is to be believed, Proto-West Germanic (A LINGUISTIC HISTORY OF ENGLISH, volume 1, page 138) had already borrowed the ending *-arius from Late Latin/Early Romance (becoming *-ari in Proto-West Germanic), making productive use of it (Sorry, no indication of length: when I tried to do so it seemed garbled in the preview, and I decided not to take a chance).
It seems to me that if language contact gets intimate enough to allow a derivational suffix to be borrowed and used productively, there is no good reason to disbelieve that a calque such as the dual meaning of “lingua” could also date back to Proto-West Germanic times (“could” is the operative word here!)
In Croatian, apart from meaning both tongue and language, “jezik” also meant people, as in the 15th century quote:
“Nalegoše na jezik hrvatski” = [the Turks] attacked the Croatian nation.
zyxt, in Russian this verb firmly belongs to conversational langauge and brings about an image of a person who’s pulling something and leans forward to pull it with a greater effort.*
We do it to work or food, and of course to Croatian language, in which case it means: to start working on it seriously.
*Well, actually I’m thinking of a person who has her “work” before her rather than behind. So maybe pushing not pulling (and actually not even pushing). But I think pulling was meant here.
Do Arabs (otside of Egypt*) find English “guttural” because it has /g/?
*and dialects with g for q:(
I don’t think the spread of -ārī in West Germanic helps, since, as Ringe makes clear, this is just an example of a suffix being extracted in Germanic from words that were borrowed in their entirety — and the kinds of words that were borrowed are of the expected kinds, culturally: monetarius, molinarius, tolonarius, etc. I don’t see this as reflecting anything particularly ‘intimate’ (though people being people, there was presumably plenty of intimacy going on in other ways).
For what it’s worth, I’ve just been checking some of the earlier attestations of ‘tongue’ in various Germanic languages, and I’m now wondering if ‘speech’ rather than ‘(a) language’ might have been the more usual metaphorical sense. Not sure if that will hold up under a more careful review, though, or if it matters much to the current discussion.
In many languages in Papua New Guinea, ‘language’ = ‘talk’, as in Tok Pisin (PNG Pidgin/Creole) or tok ples (village language)
“Talk” is also the usual way of saying “language” in Western Oti-Volta; thus the respective versions of Acts 21:37 “You speak Greek?” in Kusaal and Mooré go:
Fʋ wʋm Girik dim pian’adɛɛ?
you hear Greek persons speech.POLAR-QUESTION
(pian’ “speak”, pian’ad, formally the plural of piaunk “word”)
Fo mii Gɛrk goamã?
you know Greek words
(gome “speak”)
The “place/village” semantic overlap is interesting. That one is prominent in Ghanaian English too (“your place” = “your village”), though Oti-Volta instead has “village” = “ground”, and “place” is usually derived from “sit.”
The Byali version goes A you Gərekədyəbə twagəsəm? “You hear (imperfective) Greeks’ speaking?”, where twagəsəm is an abstract noun from a verb which is the exact formal cognate of Kusaal tu’as “talk.”
(The proto-Oti-Volta root is *tok-, which is evidently the source of the “tok” in “Tok Pisin.” Merritt Ruhlen, thou shouldst be living at this hour!)
drasvi: Well, actually I’m thinking of a person who has her “work” before her rather than behind. So maybe pushing not pulling (and actually not even pushing).
I can’t help getting the impression that this is awkwardly talking around the type of work this self-employed woman is engaged in.
(At the risk of invoking YAEPT…)
Surely the most guttural consonant in English is “/l/” (I’ve seen people claim it can be about [ʟˤ] or [ʕ] in some innovative American accents by now), followed by “/r/”.
@Brett, well, the verb means “to lay on”.
He (and the long tradition behind him) is not to be believed according to these two papers, which have a better idea. Latin/Romance may have helped make the suffix more productive, but it is much more easily explained as native than as borrowed wholesale, and the explanation ties up a few other loose ends, too.
Oh, you can ignore that. It will look right for everyone else, and for you as well as soon as you refresh the page.
I once heard midline with [ʀ] or an approximant [ʁ̞].
@DE: Tok Pisin ples may have two sources: Engl. place and Engl. village. Many languages lack /v/ and replace /f/ with /p/, as in /pis/ for ‘fish’, and English sibilants are not distinguished, as in /pisin/ for ‘pidgin’.
That’s an interesting paper by Gąsiorowski, which I hadn’t seen before (I mean the 2017 one; I did know the 2012 already). I’m fully convinced by his defence of Sihler’s etymology of the -estre suffix, but I’m not sure I’m as persuaded by his attempt to bring -ere into things (using the OE forms to stand in for the larger sets).
I really don’t think he adequately addresses the issue of the apparent social prestige and semantic associations apparently present in the Gothic -āreis words (incidentally, he might have mentioned that Sievers’ law provides an additional confirmation of the vowel length there). He has a slightly rhetorical objection to Lowe’s description of them as ‘learned formations’, which I agree doesn’t capture it (Lowe is presumably just echoing Kluge’s ‘gelehrten begriffen’, p. 6 of ‘Nominale Stammbildungslehre der altgermanischen Dialekte’). But wullāreis is the only one of the Gothic words that doesn’t have an obvious association with learning, so is G’s objection really to the point? Maybe it really is true that the connotations of ‘social prestige’ only come in later, under Latin influence, but this is very far from obvious from the he provides.
More generally, as he notes, he’s telling a fairly complicated and non-obvious story in terms of derivational and semantic shifts. I’m not saying he’s definitely wrong, but it’s a lot to assume on what seems to me rather slender evidence. I’m not saying he’s clearly wrong, just that it’s hard to say he’s clearly right either.
Many languages lack /v/ and replace /f/ with /p/, as in /pis/ for ‘fish’, and English sibilants are not distinguished, as in /pisin/ for ‘pidgin’.
That would still leave us with /piles/ (or even /pilis/) rather than /ples/, but I can see the two words getting mixed up once they became confusingly similar…
Joel, would the vowel elision in village > *piles > ples be regular and expected?
David Marjanović, Nelson Goering and other interested hatters-
An indirect argument in favor of Proto-West Germanic *-ārī being borrowed from Late Latin/Early Romance is the fact that a productively used reflex of the very same suffix is found in…(INSERT DRAMATIC STAR WARS-LIKE MUSIC HERE)… Ogam Irish (See Ziegler, Sabine. 1994. “Die Sprache der altirischen Ogam-Inschriften”. Historische Sprachforschung, 36, Göttingen: Vandenhoeck & Ruprecht. Sorry, no page number, I do not have a copy at the moment).
Since Ogam Irish is nearly contemporaneous with Proto-West Germanic and, unlike Proto-West Germanic, was almost certainly not spoken in close contact with Late Latin/Early Romance (Even if British Romance existed, which is far from certain, incidentally, it must have been spoken as a vernacular in the Eastern part (AKA Lowlands) of Roman Britain, i.e. not in direct contact with any stage of Early Irish: tellingly, phonological data indicate that most if not all Latin/Romance early loanwords in Irish were mediated through Brythonic), there is nothing outlandish about assuming the Proto-West Germanic suffix to be a borrowing.
Joel: I must echo Y’s question: From what (little) I know of Tok Pisin I do not believe a stressed vowel from an English etymon is regularly lost in Tok Pisin. Hmm. Could */piles/ (from /village/) have become /ples/ (homophonous with the reflex of English “place”) through a kind of folk etymology, due to the similarity in meaning of two words which in Tok Pisin had become too similar phonologically?
“Die Sprache der altirischen Ogam-Inschriften”. Historische Sprachforschung, 36, Göttingen: Vandenhoeck & Ruprecht.
Online here.
S. A. Wurm & P. Mühlhäusler (eds.), Handbook of Tok Pisin (New Guinea Pidgin), p. 310 (from the chapter Intonation in Tok Pisin, by Wurm). Bold added:
[The l’s in the original have a háček over them, which for some reason is not rendering correctly for me.]
And, on p. 303, Don Laycock writes,
For the latter, he adds in footnote 9, p. 306:
The Ogham inscriptions in Britain do mostly have Latin versions alongside them, though, so people in those parts at least knew some Latin, despite not actually speaking it in everyday life. (On the other hand, I don’t suppose you need to know an awful lot of Latin to compose a sepulchral inscription in it, especially the bare bones type of the Ogham inscriptions.)
through a kind of folk etymology, due to the similarity in meaning of two words which in Tok Pisin had become too similar phonologically?
That’s essentially what I meant by “getting mixed up once they became confusingly similar”. IIRC there are actual etymologies where two originally distinct words that coincidentally converged into similar forms with similar meanings fused into one word with both meanings, though I can’t think of a specific example offhand.
[EDIT: come to think of it, the case of *-ārī could be one of those… though in the case where exactly one of the words is a borrowing, this is just straight-up phonosemantic matching.]
ples [pǝˈles] = village
…oh, so the regular outcomes would be [pǝˈles]~[piˈles] for “place” versus [ˈpiles] for “village”. I can see how they could get confused.
That would require the Gothic version to be borrowed from West Germanic (or directly from Latin), which may be a bit much to ask of a derivational suffix.
The Irish versions – two, borrowed on separate occasions, the second from Brythonic – are mentioned on p. 91 of the thesis; there’s no mention of how productive they were.
“But the Indic word for tongue, jeebh (cf Sanskrit above), is just tongue and bhaashaa is just language.”
“Courageous words, Minister.
Monier-Williams includes this at जिह्वा (jihvā́), presumably referring us to the naighaṇṭuka kāṇḍa as commented on by Yāska i, 11:
speech (cf. Naigh. i, 11) RV. iii, 57, 5
And the entry in Dictionnaire sanskrit-français (Gérard Huet) supplies this information:
… langue … le sens du goût; … lat. lingua; ang. language, tongue; fr. langue.”
@ Noetica — sorry I wasn’t more clear. My comment refers to the words in a Hindi context, not Sanskrit.
David Marjanović: The author gives two examples of the older reflex of the Latin/Romance suffix in Ogham, one of which was borrowed as part of a Latin word, and the other is attached to an Irish word: the latter example seems to indicate that the ending was indeed productively used in Ogham Irish.
And why should the Gothic reflex not be borrowed from West Germanic? If (pre-) Gothic and West Germanic at the time were endpoints of a dialect continuum rather than sharply differentiated languages, it would be trivially easy for such things as derivational morphemes to spread from one part of the continuum to the other.
Well, in that case that would have been a U-shaped dialect continuum because the Goths had already crossed the Baltic Sea and missed out on the whole list of Northwest Germanic innovations. I’m not aware of other evidence for such intense contact between West and East Germanic.
FWIW, adjectivizers are one of the commonest categories of borrowed affixes.
I notice that the Biblical expression “speaking in tongues” (gift of tongues) is Zungenrede in German and klanktaal, tongentaal in Dutch. Otherwise it appears to be translated into other languages more literally as speaking in languages (e.g. Japanese 異言を語る igen o kataru ‘speak another language’). Chinese 說方言 shuō fāngyán was a surprise; it literally means speak in dialects, but I assume this derives from the mutual lack of intelligibility between dialects. (Info is from Wikipedia.)
“Speaking in Tongues” is rendered with lāšôn ‘tongue’ in four Hebrew translations of the NT (The Hutter Polyglot, Delitzsch, Zalkinson-Ginsburg, Bible Society in Israel), for the three occurrences (Mark 16:17, Acts 2:4, 1 Corinthians 13:1), except śāp̄â ‘lip’ for the 1Cor. in the Z-G translation.
(It’s γλῶσσα in the original in the three sources.)
Borrowed in 19 languages, but less relevant than “nominalizer: agent”, which has been borrowed by… 26 languages.
But – as the younger paper points out – not only does PGmc *-ārijaz have a phonetically impeccable native etymology, it doesn’t look borrowed in that very few Latin nouns in -ārius were actually borrowed, and they’re mostly not deverbal agent nouns, but denominal…
Victor Mair loves to point out that the word was never restricted to Sinitic; it can mean “local languages” even if “Chinese” is treated as a unit.
The various Oti-Volta languages I have Bible translations for all just say the equivalent of “strange language”, again using “speech, words” for “language” and mostly using the same stem for “strange” as in “stranger/guest.” Presumably they would also be used of an earthly foreign language, not just of glossolalia.
Kusaal and Mooré have a different stem, which seems to mean more specifically “alien, unknown, other”, though in Kusaal, at any rate, the same adjective is used for foreign countries. In fact, it seems only to occur with “speech” and “land, country.” (The latter is teŋ, which can refer to pretty much any human habitation from a village upwards; it means “ground” too.)
Neil:
sorry I wasn’t more clear. My comment refers to the words in a Hindi context, not Sanskrit
The term you used was “Indic”, which is traditionally applied to the Indo-Aryan languages. That includes everything from the language of the Vedas to Hindi, Sinhalese, Romani, etc. etc. Some of these modern languages have the “tongue” word – exemplified by Sanskrit जिह्वा (jihvā́) – available for “language”. Romani calls itself Romani ćhib (with variations in spelling).
Audacia, impeto, non rifuggire dalla beffa agli avversarî, affrontare con indifferenza la morte, ecco i caratteri di questa milizia.
Certainly these are the features of our present crop of American fascist militias, with the notable exception of “indifferenza la morte”.
“But their commanders won’t allow them to write fuck on their airplanes because it’s obscene.”
The U.S. Army historically considered an obscene tattoo an obstacle to conscription, back when we had conscription. An old boyfriend of Gale’s, indeed, refined this to an art: he had FUCK THE ARMY tattooed on the outside of the fifth finger of his right hand, with the F at the proximal end.
That would require the Gothic version to be borrowed from West Germanic (or directly from Latin), which may be a bit much to ask of a derivational suffix.
The way in which derivational affixes get borrowed is through enough words containing the affix being borrowed, as m-l pointed out a long time ago in connection with re-, whose Latin/Romance provenance is beyond doubt. In English it can be attached to native stems like write > rewrite as opposed to Latin/Romance ones like print > reprint, whereas Dutch has many of those same borrowings, but has never generalized re- to native stems.
“Tongue” in derivation goes with “taste” and “lick”
Just like Latin, huh? Especially if the /l/ is pronounced lamino-dentally.
We were discussing the phonaesthetic quality of “tongue” a while back in connection with a somewhat dodgy paper that spoilt a good idea by overgeneralising it, IIRC.
I do think it’s probably true that “tongue” words are more likely than chance to start with alveolar stops or /l/, just as words for “blow, puff; whistle” (which feature a lot, along with “mama/papa” words, on lists purporting to show that e,g. Mande is too related to Volta-Congo) are more likely than chance to begin with /p f/.
“The term you used was “Indic”, which is traditionally applied to the Indo-Aryan languages. That includes everything from the language of the Vedas to Hindi, Sinhalese, Romani, etc. etc. Some of these modern languages have the “tongue” word – exemplified by Sanskrit जिह्वा (jihvā́) – available for “language”. Romani calls itself Romani ćhib (with variations in spelling).”
@ Noetica— The entire comment was on Hindi. The second part was contrasting the Indic Hindi word Jeebh has with the Persianate Hindi word zabaan. Both are legitimate and common Hindi words, just from different sources
Victor Mair loves to point out that the word was never restricted to Sinitic; it can mean “local languages” even if “Chinese” is treated as a unit.
Yes, I’m aware of the meaning of 方言 fāngyán in China, but I’m not sure I’ve heard Mongolian or Tibetan referred to as a 方言 fāngyán in China — although I wouldn’t totally rule it out.
The use of 方言 fāngyán to mean “dialect” in the English sense (local, non-standard variety of a unitary language) is possibly a result of the refashioning of Chinese-language concepts to fit Western models. There is also possibly Japanese influence, since the same word 方言 hōgen in Japanese is (now, at least) more closely aligned with English “dialect” than with Mair’s characterisation of the Chinese concept. (The Japanese usage is also possibly a result of influence from Western languages. It would be interesting to see whether or not the meaning of 方言 hōgen in pre-modern Japanese more closely mirrored the traditional Chinese meaning.)
Neil:
Both are legitimate and common Hindi words, just from different sources
Just Hindi? Then it would have been helpful to say that instead of “Indic”.
Yes, Wiktionary has an entry for Hindi ज़बान (zabān), and it gives two meanings: 1. tongue; 2. language. In any case, zabān appears to be etymologically connected with जीभ (jībh), way back.
Note, among other parallels, the Kashmiri word akin to jībh: \ज़्यव (zyav). It can mean either tongue or language.
Noetica: When Neil wrote “the Indic word for tongue, jeebh” he was contrasting it, etymologically, with the Persian loan “zabaan” in the previous sentence. This is clear once one understands that his whole post is about various words in Hindi (zabaan, jeebh, and bhaashaa). That context could perhaps have been more clearly established, but it would not have made any sense to replace “Indic” with “Hindi”.
“word of Indic origin”
Is there a way to refer to Indonesian in Indonesian other than “bahasa indonesia”?
Yes; and that seems not to have happened in this case.
I don’t imagine that the total number of borrowings is the sole deteminant: all it would take would be a few borrowings that were (a) very commonly used and (b) transparently contained the suffix in question.
Analogy is not to be underestimated. Welsh has a 3rd sg preterite flexion-odd which is now the regular ending, but seems to have begun in mistaking the final syllable of the (actually suffixless, ablauting) form goddiwawdd, from goddiweddaf “I overtake”, for a flexion and recreating the form as goddiweddodd. There were only ever a handful of verbs that made their preterites like this (though one of them was “say”) and goddiweddaf seems to be the only one with a stem ending in dd.
It’s as if English weak verbs went “I walked, you walked, he/she walkang” (following the analogy of “I sang, you sang, he/she sangang.”)
One can only surmise that the mediaeval Welsh had a real thing about overtaking, and talked about it all the time …
MInd you, Welsh must be the undisputed champion when it comes to analogical remodelling of all of Indo-European, if not Teh World.
One can only surmise that the mediaeval Welsh had a real thing about overtaking, and talked about it all the time …
On reflection, this would be due to the notorious introduction of an all-Wales twenty-mile-an-hour speed limit by Hywel Dda.
(Proving yet again that language and culture are intimately linked.)
Drasvi: If there indeed is no other way to expression “Indonesian language” in Indonesian except by means of the Indic loanward “bahasa” (Incidentally, does anyone know whether “bahasa” can be unambiguously traced back to Sanskrit, Pali or perhaps some more obscure Indo-Aryan variety?), this would make Indonesian very similar to English: “tongue”, in English, with the meaning “language”, is a very archaic/bookish word (Indeed, I have had a few L1 anglophone students who were unaware that “tongue” could be used with this meaning).
David Marjanović: Since 1-Ogham Irish does make productive use of a reflex of Latin/Romance *-arius, despite having just as few if not fewer Latin/Romance loanwords than Proto-West Germanic/Early West Germanic did, it must be concluded 2-That there is no reason in principle to exclude the possibility that the West Germanic suffix might likewise be a Latin/Romance borrowing, and 3-That the Gothic suffix may have been borrowed from West Germanic, despite there being few to no unambiguous borrowed West Germanic elements in Gothic.
David Eddyshaw: Hmm, Welsh as the champion of analogical remodelling in ALL of Indo-European? Them’s fightin’ words…with part of the problem being that all too many analogical changes are unclear in origin. For instance, in (some varieties of) Late Latin/Early Romance a significant change relating to the imperfect indicative involves the loss of the /b/ marker for verbs belonging to conjugation classes other than the first.
ONE theory holds that this first arose in a single verb: “habere”. It is argued that reflexes of Classical-like HABEBA- imperfect forms lost the second /b” through dissimilation, and the new */ea/ ending then spread to ALL verbs NOT of the first conjugation. Because “habere” was the verb expressing possession AND the future (The Romance future + conditional was grammaticized much later) AND the (emerging) Romance HAVE-perfect, it has been argued that it was frequent enough for its ending to spread the way it did.
It gets better: the French endings of the imperfect (-ais-ais-ait -ions -iez -aient) derive (with some analogical influence of the morphology of other tenses, notably the present indicative) from what historically were the endings of verbs not of the first group. So the imperfect-marking endings (All, not just the third person singular one!) of ALL French verbs (This set of endings is not the regular set, please note, it is the universal set!) today may go back to the irregular endings of the imperfect…of a SINGLE Late Latin/Early Romance verb.
Tim May:
That context could perhaps have been more clearly established, but it would not have made any sense to replace “Indic” with “Hindi”.
Thank you. I wasn’t able to construe Neil’s contribution that way, but I see it now.
Them’s fightin’ words…
I accept your challenge.
Welsh has collapsed all its regular verb conjugations into just one, except for the 3rd sg present indicative, which shows three different formations reflecting the three conjugations-as-were. Everything else has been levelled out.
The form of the 3rd person plural pronoun after prepositions has been remodelled after the 3pl flexion of verbs. So has the corresponding independent personal pronoun. Sometimes.
Welsh has the soft mutation of the initial consonant of objects after all finite verbs now, regardless of what (if anything) intervenes: a fairly radical reshuffling of the situation from Middle Welsh, in which the mutations at least still pretended to be based on the nature of the word actually preceding the object. It is actually a lot simpler as a system now, and has the great virtue of annoying Chomskyites no end, as they have great difficulty fitting the phenomenon into their pretty frameworks and are unfamiliar with the concept “counterexample.”
[I must admit, though, that I had an uneasy suspicion that French might prove to be a Worthy Opponent even when I made that former comment.]
Of course not. This is all a parsimony-based argument.
Sections 2.2 (starting on p. 129) and 3.3 (starting on p. 141) of the second paper show how few Latin nouns in -ārius could have been borrowed early enough; a larger number may be calques, but they imply the suffix was already productive on the Germanic side. The number of resulting (mostly East) Germanic words that were passed on into Slavic is considerably higher, so the Slavic productivity of the suffix (considerably lower than the Germanic one!) is much more easily compatible with extracting the suffix from borrowed words.
Contrast the total chaos that is the, uh, let’s call it conditional in my German dialect: it’s historically derived from the simple past by umlaut in the strong verbs and identical to the simple past in weak verbs, but the simple past is gone, unmooring the entire system… and don’t get me started on the preterito-presents at 1:17 am… 🙂
Back home we’ve replaced the 2nd-person singular clitic by the 2sg verb ending…
and don’t get me started on the preterito-presents at 1:17 am
Ooh, a verbal system that varies by time of day!
@Etienne, well, the qeustion is can they say it simpler?
in English you can say “in English”, in Russian you have more bookish (maybe calqued) “on Russian” and more conversational adverb of manner with the meaning of à la.
You don’t have to say “in the English language” or na russkom yazyke all the time unless you’re one of those children (and maybe teachers) who believe that excessive formality is what we all strive for.
Ooh, a verbal system that varies by time of day!
Reminds me of the aliens in Babel-17 whose personal pronouns vary by temperature.
I suspect that my own verbal system undergoes some simplification on Friday nights. Tends to shift from polysynthetic to isolating.
Seifart wrote a good paper based on his dissertation research, which also resulted in the affix borrowing database. He distinguishes between direct borrowing, based on knowledge of the donor language, and indirect borrowing, with analogy (reverse engineering if you will) of a small set of borrowed words containing the affix.
@ David E,
If they were hydra, that might simplify their social life.
I suspect that my own verbal system undergoes some simplification on Friday nights.
For me it’s Thursday-morning clarity, after the tortured verbal vicissitudes of Wednesday. Tense, aspect as opposed to Aktionsart, mood, voice, ergativity … now timeofdaytivity as opposed to timeanddativity? A mirative innovation if ever there was[12:31AEDT] one.
“Listen to them, Children of the Night: what music do they speak” (from memory?)
It’s “What music they make.“
Yes, I’m aware of the meaning of 方言 fāngyán in China, but I’m not sure I’ve heard Mongolian or Tibetan referred to as a 方言 fāngyán in China — although I wouldn’t totally rule it out.
Certainly true in Imperial times, when the central translation bureau translated all decrees into the various fangyan spoken on the border regions of the Empire.
Reminds me of the aliens in Babel-17 whose personal pronouns vary by temperature.
Peter Medawar talks about a variety of shrimp whose sexes vary by either water temperature or pressure. Delany might well have had this in mind. Medawar’s context was about what is and isn’t “heritable”, which varies according to the context of interpretation: in the West myopia is highly heritable, whereas in third-world countries it is strongly linked to social class, because it requires a confluence of both heredity and exposure to close-up work such as reading or sewing at an early age.
I’ll let Eddyshaw be the final arbiter here, but from what I recall, myopia is caused by under-exposure to sunlight in childhood, not by looking at things up close. Of course, sitting at home and reading or sewing go together.
‘but they imply the suffix was already productive on the Germanic side’
Or that the few early loans were prestigious and provided a model for a trendy suffix. It only takes one word to kick things off. Again, I think Gąsiorowski’s paper underplays the semantic coherence to the Gothic words. Surely bōkāreis and laisāreis are precisely the sort of word that a Latinate suffix might first spread to (especially if the spreading was done by the educated scribes and teachers who knew Latin themselves)? I’m not even sure that the wullāreis is much of a problem, since once a suffix begins to spread, it can easily keep doing so. (I have no idea what the socioeconomic status of fullers would have been in Gothic-speaking societies.)
As for parsimony, G’s account involves a rather complicated formal prehistory. Nothing impossible or really outlandish, but it’s not at all obvious to me that he ends up with a simpler account than the traditional one. This is the problem with an appeal to ‘parsimony’, generally speaking: it’s often relative to exactly which part of the problem you’re concentrating on, and when you try and weigh two explanations, each more parsimonious in its own subdomain, against each other, it’s not always very obvious what to prioritize. Sometimes it is very obvious, of course — but this seems like one of the non-obvious cases.
Well, yes. One’s career is not generally advanced by publishing papers which state the bleeding obvious (which is not an excuse for those who publish patent nonsense in the apparently well-founded hope that it will get them into Nature). Piotr’s paper on the origins of dog is equally unprovable, but it doesn’t have to run faster than the bear.
@ng
If the fuller was a lowly profession (something which requires contact with urine sometimes is), they could still add a posh suffix, either the fullers to make themselves feel better or the clerics to display their learning or sarcasm. I seem to remember it was a feature of German student speech not only to replace common words with Latin (some of these like Gaudi and Servus are still used) but also to attach Latin suffixes or prefixes to German words (but I can’t immediately come up with an example).
There is a, dare I say obvious distinction between an explanation being obvious, and it being obviously correct once proposed. G’s suggestion isn’t obvious, but I can’t imagine anyone criticizing it on those grounds (I didn’t). But DM shared it with the idea that it was clearly a superior (more parsimonious) explanation, and that’s the part that’s not obvious — and the part that should be.
(Provability is neither here nor there with this kind of thing.)
Heh. Kinda – there’s a lot of intra-individual variation. 🙂
Also, in fields like this, all the easy problems were most likely recognized and solved 100 to 150 years ago; all that’s left are developments with few examples that require more complex explanations which often had to wait for developments in related fields. I really don’t think anything remotely as blatant as Grimm’s law, or even Verner’s, awaits discovery in Germanic.
“Obvious” strikes me as a very high expectation. The parsimony advantage, as far as I can see, is:
– both -er and -ster, and much of their various usages in space & time, are explained at once;
– most of the oldest documented examples of -er are attached to native verb roots rather than Latin noun roots;
– molinarius, which made it into Northwest Germanic and was passed on to Slavic, isn’t attested in Latin until the 6th century, and its base noun molina itself only from the 4th onwards;
– tolonarius made it into West Germanic, but the corresponding East Germanic word has a native root (and was passed on to Slavic);
– that seems to leave only monetarius as a model that could have been borrowed early enough. (It’s not reflected in Gothic, but I don’t know if the passage with the money-changers is even preserved.)
As the paper says, Gothic bokareis, motareis, wullareis may well be calqued – but that requires the suffix was already productive; it can’t have been introduced this way.
I don’t know if the passage with the money-changers is even preserved
It is, but not helpfully so.
Mark 11:15 says “Jah iddjedun du Iairusaulwmai. jah atgaggands Iesus in alh dugann uswairpan þans frabugjandans jah bugjandans in alh jah mesa skattjane jah sitlans þize frabugjandane ahakim uswaltida” (emphasis added) using skattja = skatt ‘cattle, money, treasure'[1] plus the unquestionably native affix -ja, which Piotr also discusses. Note that in Gothic orthography ai, au, ei mean /ɛ/, /ɔ/, /i:/; the letters transliterated e, o are used only for their long tense equivalents.
Luke 19:45 speaks only of frabugjandans in izai jah bugjandans, ‘those that sold therein and those that bought’, words that also appear in Mark.
“Ho, frijond, allai skattjans sind waidedjans!”
[1] Which is everywhere in Germanic and also passed directly into OCS > Russian скот ‘cattle, livestock’.
‘both -er and -ster, and much of their various usages in space & time, are explained at once’
But the usage of -er is not really explained. It’s made to fit the etymology only with quite a bit of shoehorning. That’s not explanation, and is of course the main sticking point in all this.
‘most of the oldest documented examples of -er are attached to native verb roots rather than Latin noun roots’
I’m not sure that’s true. In Old English, the two oldest occurrences I’ve found attach one to a native verbal root, and one to a Latin noun. Gothic, the oldest attestations overall, fits well with the traditional explanation. The Gothic data is easy, but one thing that’s needed is an actual look at the data in North and West Germanic that has some philological rigour to it, to get a clearer picture of the dates of attestations of different formations. The lack of clarity there is a large part of why I’m reserving judgement. How can we say what explains the data better, if we haven’t properly examined the data?
‘its base noun molina itself only from the 4th onwards’
Attestation is, of course, not the same thing as the actual of the word. In any case, I’m rather agnostic about the dating of all this here, and have no real problem with the West Germanic examples largely getting going only in the later Roman period. Note also that words attested in Gothic and Old High German need not point to Proto-Germanic, or to West Germanic influence on Gothic (which is very unlikely), but could well represent Gothic influence on Old High German. The plausibility of this depends on details of what’s first attested where within West Germanic, but one example can show the dangers: one of the three examples G cites for a Proto-Germanic date of the suffix is laisareis, but the only cognate is OHG lērari, first attested in Alemannic (though later found up into Franconian). It is not used in either Old Saxon (which has lēriand, to the same root) or Old English (lārēow). Only liuþareis supposedly has a non-OHG cognate, but I can’t find the supposed lēoþere in the DOE Corpus. If I have a moment later, I’ll follow it up more. None of this is particularly definitive, but it certainly doesn’t readily support the case for an inherited suffix of the sort G wants.
‘but that requires the suffix was already productive; it can’t have been introduced this way’
That’s very much not true. When a suffix begins to spread, there must always be a few first words to show it. If we insisted on prior productivity, nothing could ever become productive! A better way to put it is to say that there must be some model for such formations to be analogical to. We don’t have that in Gothic (but our records are very limited). We do have examples for West Germanic. The question is whether the data gaps and available models are a big problem for us or not (or more accurately, a bigger problem than the derivational chains G has to go through to get the -er suffix to have the meanings and functions he wants from a starting point of ‘woman’).
Again, I’m not arguing against G. It’s an interesting idea I’m taking seriously. I’m just explaining why I’m agnostic rather than converted.
On OE *lēoþere, I’ve just found this in the DOE, headword hlēoþrere:
‘CollGl 12 4: rethor se hloðere (ms alt. from loðere in contemporary hand; from beda. Hist.eccl. 1.10, 38 quod pulchre uersibus heroicis Prosper rethor insinuat; an alternative interpretation of the gloss as a form of otherwise unattested *lēoþere ‘poet’ [cf. lēoþ] has been tentatively suggested).’
So it’s a ghost word. That leaves all three of the Gothic-WGmc pairings with this suffix limited to Old High German. None of these have matches in North Germanic.
That’s exactly what I mean: there must be loans of Latin -arius words – the more, the better – before the suffix can be abstracted from them and become productive. Very few such words seem to be candidates for this; we see the productive use as a suffix, but very few loans that could be old enough.
Molina started as an adjective to mola, which is attested much earlier and was also borrowed into Germanic.
That’s not a good way to think of science in the first place. It’s only science. It’s all a matter of probabilities all the way down.
That is definitely interesting.
‘the more, the better’
Rather: the better placed to be a model for other words, the better. Again, one would be enough, especially if the process were mostly driven by bilinguals (learnedly so or otherwise). A large number of the wrong sort wouldn’t be much good! For West Germanic, considered independently, things are probably actually fine. And if we allow for some of the OHG formations to be calques from Gothic, it only gets easier.
Actually the small number of ‘old’ examples is kind of interesting for another reason. If the suffix were genuinely part of Proto-Germanic morphology, we might fairly ask how come there are so few exact matches between the different languages (and what there is mostly takes place in geographically adjacent languages — in the case of OHG and Gothic, this being a secondary contact after a longish period of non-contact). On the other hand, this is natural enough if it only entered the various languages later on, through contact with Latin.
I was being tongue in cheek about conversion, but isn’t it a useful informal shorthand? You clearly are ‘converted’: you think the connection with -stre is a more probable explanation for -ere than the borrowing hypothesis. I certainly don’t mean to imply by that that you’re hanging onto this proposal out of stubborn dogmatism or something. By ‘agnostic’, I just mean that I’m currently willing to see both hypotheses has having enough explanatory value to be worth considering, but neither as so problem-free that I’d definitely go with one or the other. I don’t think there’s anything ‘unscientific’ about principled uncertainty between hypotheses that have roughly equal probabilities (not that this is quantifiable in a context like this…) of being true. Maybe more data would change that, if someone cared to put in the work to assemble a proper dataset (which wouldn’t be all that hard, but would still take a bit of time to do right).
I was being tongue in cheek about conversion, but isn’t it a useful informal shorthand?
Yes, it is. Scientists are people too, and people have beliefs, not just probability weights, even if they choose to deny it.
‘That leaves all three of the Gothic-WGmc pairings with this suffix limited to Old High German.’
I somehow overlooked bōkareis as a fourth example — and that one does occur in Old English, including in early poetry (both Daniel and Exodus). In Old Saxon there’s buokari as a gloss (but not in the poetry: paraphrases like buokspāha uueros are used instead), and OHG has puachāri (with the usual range of variants). This would look like the best candidate for an old word with -ere, except of course that the attested meanings aren’t very likely to have had a word in Proto-Germanic.
@ng
Does Ge. Zauberer, Dutch toveraar have an attested older form? That would be the sort of thing that might go back to PG.
There is a zaubarāri in OHG, but I don’t think this is supported by cognates outside of Continental West Germanic, and its base word continued to be readily available. It could have been formed pretty much any time between the spread of the -ari suffix (whenever that was!) and its first attestation (in the early 9th century, going by the Chronologisches Wörterbuch des deutschen Wortschatzes).
Fair enough, I overinterpreted your wording. It reminded me too much of that one time (at least ten years ago) I participated in a discussion with someone who, apparently, only ever changed his mind if he had a euphoric conversion experience of religious proportions.
one time (at least ten years ago) I participated in a discussion with someone who, apparently, only ever changed his mind if he had a euphoric conversion experience of religious proportions.
That’s a high bar which naturally requires a leap of faith. But at least change was not rejected outright. I deal with many people, more than my fair share, who are dead set against changing their mind. They change it, if they do, only after falling on their face because they didn’t reconsider in time.
I have come to the conclusion that words are not a particularly reliable means to convey ideas, or even to prevent the worst (apart from discussions on this blog, of course). The only degree worth having is awarded by the school of hard knocks, in which there are no classrooms.
Had I known this years ago I would have administered kicks right and left, and today would be revered as the Sam’l Johnson of effective cognition transfer. To put it another way, instead of shaking my cane I would have applied it with good effect.
the school of hard knocks
There is a Hausa proverb: Jiki ya fi kunne ji “the body hears better than the ears.”
[It works better in Hausa, in which ji means “feel” as well as “hear”: “perceive by any non-visual modality.”]
“perceive by any non-visual modality.”
That’s exactly what I mean. I find myself warming to the proverbial Hausa.
David Eddyshaw:
Okay, French versus Welsh:
1-“Welsh has collapsed all its regular verb conjugations into just one, except for the 3rd sg present indicative, which shows three different formations reflecting the three conjugations-as-were. Everything else has been levelled out”.
In French the inherited Latin verb classes are no longer relevant for finite verbs: all verbs in the imparfait have the same set of endings, all present indicative and present subjunctive forms are morphologically identical in the spoken standard (the irregular verbs cannot be classified on the basis of the verb class they belonged to), and synchronically the link between the infinitive (where the stems DO go back to verbs differentiated according to the classes inherited from Latin) and the future + conditional stem is now wholly obscure: synchronically French speakers treat the regular future + conditional stem as deriving directly from the present indicative stem, and not from the infinitive.
So, to my mind: Welsh and French are pretty much on an equal footing today -interestingly, NONE of the above statements would be applicable to Old French.
2-“The form of the 3rd person plural pronoun after prepositions has been remodelled after the 3pl flexion of verbs. So has the corresponding independent personal pronoun. Sometimes.”
In French the third person (clitic) nominative masculine singular form, “il”, does NOT go back directly to Latin “ille” or “illum”: either would have yielded a form /el/, which is attested in Old French: “il” was originally the PLURAL nominative singular masculine third person pronoun, regularly deriving from Latin ILLI. So: how did a plural become a singular? Well, the plural attracted a plural /s/ by analogy with the many nominal and pronominal elements pluralized with /s/ (notably the feminine forms of these pronouns, “elle” and “elles”, both of which do go back to Latin etyma: ILLA(M) and ILLAS, respectively).
After “il” became “ils” a new singular pronoun, however, “il”, was back-formed from the plural, and *wholly replaced* the older, inherited “el” pronoun.
In this instance French beats Welsh: the analogical form does not coexist with the inherited form, the inherited form is extinct.
3-“Welsh has the soft mutation of the initial consonant of objects after all finite verbs now, regardless of what (if anything) intervenes: a fairly radical reshuffling of the situation from Middle Welsh, in which the mutations at least still pretended to be based on the nature of the word actually preceding the object.”
Sounds quite mild compared to French liaison, which only arose in 1500-1600 as a consequence of the loss of final consonants, and while in the standard liaison consonants are still supposed to be triggered by the previous word (with the relevant final consonant still being present in the spelling), in practically all colloquial registers of French liaison now involves floating consonants prefixed to the first vowel-initial element in an NP. Most native speakers of French would spontaneously treat “C’est quelle sorte de surhomme”/C’est quelle sorte de surhommes” as being realized differently (whereas in the standard they are both supposed to be pronounced identically): singular /sekɛlsɔʁtdəsyʁɔm/ versus plural /sekɛlsɔʁtdəsyʁzɔm/.
The case for treating this /z/ as a prefixed plural-marking morpheme (which, as in the above example, can actually be infixed! Hmph!) is made more powerful by the fact that instances such as /gʁozɔm/, which in the standard could be singular or plural (kept apart in spelling: “gros homme” versus “gros hommes”), tend in colloquial French to be interpreted as plural only, with a corresponding singular form /gʁotɔm/ or /gʁonɔm/ (no semantic distinction between these two singulars…yet), where the /t/ and /n/ singular-marking morphemes derive from the liaison consonants of such common singular adjectives as “grand” or “petit” on the one hand and of the indefinite article “un” on the other.
So, whereas Welsh has extended some instances of mutation beyond their historically-justified range, in French we have created a whole mutation-like system over less than half a millennium. I think French wins out here.
Speaking of Welsh:
Sounds quite mild compared to French liaison,
Ah, but I only mentioned the easiest case of Welsh initial mutarion. We have two more mutations. All arose as “lisison”, essentially, and all three have been remodelled by analogy to the point that this origin is now unrecognisable to anyone but a historical linguist and is irrelevant synchronically.
It actually rather reminds me of the Kusaal rules for external tone sandhi, most of which clearly originated as straightforward phonological tone spreading, but can no longet sensibly be described that way synchronically owing to the loss of word-final vowels. (You can describe them that way by brute force and considerable abstraction from surface forms, but doing so achieves nothing except showing off your perverse ingenuity,)
(I did toy with the industry-standard way of doing this, which is to say, for example: “All imperfective verb forms are followed by a floating mid tone, which links to an initial low tone in a following word to create an underlying falling tone, realised, as always in Kusaal, as a high tone.” I came to the conclusion that this achieved nothing that couldn’t be better put as “all imperfective verb forms cause an initial low tone in a following word to become high.”)
Yes, a lot of phonological theory tries to ‘recapitulate phylogeny in ontogeny’, to say it in Old High Falutin. I think there’s been some real progress made in highlighting the issues with those approaches, but there’s a long way to go. One uncomfortable implication is that a lot of explanations in linguistics (at least in phonology) should be historical rather than synchronic…
“Everything is the way it is because it got that way.”
You can describe them that way by brute force and considerable abstraction from surface forms, but doing so achieves nothing except showing off your perverse ingenuity
…isn’t this kind of thing pretty much the requirement for something to be worthy of being called “sandhi” in the first place?
(In both directions: if things are too unpredictable/irregular, such that you’d need special rules for too many individual roots, it probably doesn’t count as sandhi either.)
One uncomfortable implication
For me anyway, quite comfortable.
David Eddyshaw: do please bear in mind that Welsh has had a system of mutations since Proto-Brythonic times (i.e. fifteen centuries or thereabouts): as soon as pre-Proto-Brythonic final syllables were lost the system became phonologically opaque (i.e. words triggered given mutations for reasons that were not synchronically deducible from the (surface!) form of said words). French liaison is much younger (just a few centuries old!), and yet in colloquial French can already be said to be in many contexts a system of morphological marking wholly unmoored from the “triggering” effect of a preceding word (indeed, this is so ubiquitous in all spoken French varieties that I could all too easily imagine these new prefixes -nominal plural /z/- first and foremost- spreading to new contexts and ultimately outliving much of the inherited morphology of French).
I am delighted that French is finally catching up. I think we still have a commanding lead, though.
Yes, a lot of phonological theory tries to ‘recapitulate phylogeny in ontogeny’
A lot of autosegmental approaches to tone seem to me to do exactly that. It’s perhaps a bit ungracious of me to say that, given that African terracing tone systems were the great impetus for the development of such things in the first place, and they do tend to be very pretty and often quite good for accounting for the synchronic facts. But it’s all too easy to lose sight of how very abstract they can be, if considered as actual synchronic mechanisms. “Floating tones”, forever inaudible and detectable only by their supposed effects on real tones … hmm.
In the specific case of Kusaal, tone sandhi behaviour has often demonstrably been altered by analogy: so while autosegmental theories are intuitive for cases where segments have been lost and their tones live on, forever seeking a new segment to call home, Kusaal quite often does the opposite.
For example, the combining form of nouns, used as the first part of a compound, is historically a bare stem and has just the tone sandhi effects on the following stem that you’d expect from purely phonological processes. But Kusaal has remodelled many combining forms after the singular free forms segmentally, which consist of a stem plus a class suffix. However, this remodelling never affects the tone sandhi, so the added materlal is just ignored tonally. You can capture that in an autosegmental account. But why should you even try? It just becomes a matter of adding epicycles to save the theory.
I don’t think that’s unusual. In my dialect, the diminutive suffix -/ɐl/ triggers stem-final consonant fortition, and this applies even to *-/mb/ which no longer exists as such (but did a few centuries ago), but the whole phenomenon can hardly be older than, again, a few centuries.
If anyone’s interested, I’ll explain tomorrow what historical sense I make of this.
If anyone’s interested, I’ll explain tomorrow what historical sense I make of this
Yes, please.
He knows perfectly well we’re interested — he just wants to make us beg.
Thirded.
But it’s all too easy to lose sight of how very abstract they can be, …
Deep Structure to the rescue!
Never. I was just tired.
So… there are two diminutive suffixes: syllabic /l̩/ and /ɐl/. Apparently they were simple and double diminutives 200 years ago, but I haven’t encountered that; most nouns take one or the other, not both.
Adding /l̩/ to a word in -/n/ straightforwardly causes epenthesis of /d/: /hoɐ̯n/, /koɐ̯n/ “horn, grain” → /hɛɐ̯ndl̩/, /kɛɐ̯ndl̩/. It looks like this /d/ was morphologized as part of a “diminutive stem” and copied into the other (the double?) diminutive: /ʃtɛɐ̯n/ “star” → /ʃtɛɐ̯ndɐl/ “little star; asterisk”.
But what do you do if a word already ends in /d/? Then you interpret /n/ → /nd/ as an example of a more general process of fortition: /p͡fɛɐ̯d/ “horse” → /p͡fɛɐ̯tːɐl/.
(Intervocalic short /t/ doesn’t occur word-internally. Or morpheme-internally or something.)
And on it goes: /lɒmː/ “lamb” → /lampɐl/.
This process is not applicable any further if there’s already a fortis at the end of the stem: /lɒmpm̩/ “lamp” → /lampɐl/, homophone of the above.
It may have been helped along by the words that show consonant lengthening in the diminutive: /fɪʃ/, /tɪʃ/, /froʃ/ “fish, table/desk, frog” → /fɪʃːɐl/, /tɪʃːɐl/, /freʃːɐl/. That’s historically the inverse: a case of word-final consonant shortening that was prevented not only by suffixes, but also by the full-size plural ending -e which has since been lost: full-size plurals /fɪʃː/, /tɪʃː/, /freʃː/. (Indeed that’s why a short /ʃ/ exists at all and is available for French loans that have [ʒ] in the original. We talked about the multifarious representations of garage in English recently; at home it’s straightforwardly /gaˈraʃ/, pl. /gaˈraʃn̩/.)
‘Mary had a little lamp, little lamp, little lamp…’
Hausa, in which ji means “feel” as well as “hear”: “perceive by any non-visual modality.”
just jumping in from my own parish to note that while the verbs for “hear” and “feel” in yiddish can both encompass “smell”, only “feel” also extends to “taste” (though a bit shakily).
Clywed in Welsh covers the same range as Hausa ji: hear, feel, taste and smell.
The underlying PIE root can be seen to have a similar semantic range, for example in Greek names like Androcles “Smells like a Man”, Pericles “Very Smelly.” (They had different aesthetic sensibilities in the Olden Days.)
Thread won.
the late robert farris thompson would like me to point out that this is parallel to the olofactory-positivity of “funky” (which i believe he derived from a kikongo etymon).
Where does the Hausa teku “sea” come from?
Neil Skinner (1996) Hausa Comparative Dictionary, p. 256, has the following:
Sng. stands for Songhay, Mnd. for Mandara (or Wandala, I gather). For kogi, Skinner gives the following:
I am not qualified to evaluate Skinner’s work, but my general impression of Skinner’s book accords with Newman’s far more informed judgments.
Now I am left wondering about the origin of Songhay tēko. Can an earlier referent be recovered? ‘Inundated floodplain’, as for instance the Inland Delta of the Niger? No possible relation to Hausa tafki, tabki ‘lake, pond’, as in Tafkin Chadi ‘Lake Chad’, I suppose.
The question of the origin of Songhay tēko interests me because so many words for ‘sea, ocean’ present interesting semantic developments (Welsh gweilgi !) or are otherwise of difficult or disputed etymology (Luwian alaššamma/i-; Vedic ságara-; Greek θάλασσα, ὠκεανός, some even dispute πέλαγος; Old Irish fairrge…). And also, what happens when peoples who have lived in the interior of a continent, who have only known rivers and relatively small lakes, come to meet great bodies of water and must find a name for them? More specifically, I have been looking for parallels to the case of Turkic talūy, as explained in G. Clauson, An Etymological Dictionary of Pre-Thirteenth-Century Turkish, p. 502:
The Chinese characters in question are 大 dà ‘great’ and 㶟 lěi, old name of the river Sanggan River (and perhaps other sections of the Yongding River system?). Mongol dalay forms part of the title Dalai Lama.
OCR imitating Луркоморье?
Die alttürkischen Inschriften der Mongolei
The lürkers support me in email.
Thanks, Xerîb! I thought that you could help, if anybody.
I agree about Skinner. It still amazes me how people can spend years of sustained hard work on comparative linguistics without ever realising that their work is almost completely vitiated by the fact that they never understood the basic principles. Skinner is, of course, hardly unique. Alas. His etymology for kogi is frankly ludicrous. Ruhlen-level stuff. (Why not Kusaal kɔlig “river”, too, while we’re playing lookalikes? And Mooré koom “water”, stem ko-?* It’s embarrassing to see this sort of display from a far from negligible respected Hausaist.)
Songhay tēko looks very plausible, and “inundated floodplain” would certainly do very nicely as the local next-best-thing to an actual sea. Lameen is the person who will know whether it’s actually real: it doesn’t appear in the few Songhay dictionaries I have, and references to “Songhay” as if it were just one language do not inspire confidence.
* This etymon is confined to Western Oti-Volta, and does not represent proto-Oti-Volta. Even.
it doesn’t appear in the few Songhay dictionaries I have
For some indications that this is real form—on this this website, there is a lexicon of Gao Songhay available for download (link here), and it has the entry teeko ‘mer’ on p. 83. Also in this online dictionary beside the obvious loanword albaharu. And this article analysing two oral narratives in Zarma recorded by the author in Niger also mentions teeko ‘mer’.
tafki, tabki
Robinson’s old dictionary of Hausa (1899) offers an Arabic origin for Hausa tafki, tabki: an Arabic طبـڧ ṭabq (that is, طبق , Eastern style) ‘rain over the whole country’ (an obscure meaning that Robinson plucked from a dictionary?). I can find no other mention of this etymology.
Thanks, Xerîb! Very interesting (and thanks for the links.)
Gao is certainly a fair way away from the zone where a loan from Hausa is all that likely, too. (Zarma, very much less so, of course.)
Still the possibility that Hausa and various Songhays have all borrowed it from a third party, I suppose. After all, a word for “sea” is going to be a prime candidate for borrowing on first principles up there in the Sahel/Savanna.
I wonder what Lameen thinks? This is very much his bailiwick.
I can’t see any way to get to teeku from tabki within Hausa itself (not that I’m any kind of Hausaist, though.) Robinson was rather too fond of ascribing Arabic origins to Hausa words (a fatally easy game, given the vast vocabulary of Classical Arabic: if you can’t find some lookalike, you just haven’t looked at enough dictionaries yet. I can do much the same with “Gur.” And of course, a lot of Hausa words really are derived from Arabic; just not as many as Robinson thought …)
Come to think of it, if the Western Oti-Volta “water” stem had been inherited from proto-Oti-Volta (which it can’t have been) it would have represented proto-Oti-Volta *kôɰ-, which obviously bears an uncanny resemblance to the Hausa kogi “river.” Anyone who believes that the WOV is borrowed from Hausa may be interested in buying a bridge that I happen to own. Anyone who believes that the Hausa is borrowed from WOV may be interested in ten bridges that I happen to own, and anybody who believes that the words are cognate is the reincarnation of Merritt Ruhlen and I claim my $25.*
* Oh, it’s you, Roger Blench. You certainly had me going for a bit there. Nice one!
I had occasion today to look up a certain Hittite ritual today, and it reminded me of this thread, so I thought I would share a relevant bit of the ritual here. See the translation in number [14] and the original Hittite text in footnote 39 on p. 25 in Gary Beckman (1986) ‘Proverbs and Proverbial Allusions in Hittite’, Journal of Near Eastern Studies vol. 45, available here. The translation (slightly edited and reference numbers removed, boldface mine) follows:
The Hittite phrase ‘The tongue is a bridge’ is written as follows:
To be read in Hittite, lālas-wa armizzi. (The determinative GIŠ should be superscript, but no superscript Š is available in Unicode, and I can’t get superscript html tagging to work. For the general LH readership who are unfamilar the principles of the Hittite writing system—it works somewhat like Japanese: Sumerian logograms (transliterated in uppercase roman, like EME ‘tongue’) are read as Hittite words, and their inflections are written with phonetic complements in the Mesopotamian cuneiform syllabary (transliterated in lowercase italic). Unpronounced Sumerian determinatives (transliterated with superscripts) are also used, like GIŠ for wooden objects such as armizzi ‘bridge’. The quotative particle that Beckman mentions is the -wa. To see a drawing of the tablet fragment with the characteristic ductus (sign-shapes) used by the Hittites, click here. It’s the second half of line 17 of column III.)
A trope thus officially Older Than Dirt.