Christopher Culver’s Linguistic Weblog

One of the peculiar parallel developments in Mari and Chuvash is that they both employ a different verb in basic go expressions than the rest of their respective language families.

For the Uralic family, the central verb of source- and path-oriented motion in the proto-language was *mene-, based on comparison of Khanty, Hungarian, and North Saami. Söder (2001) gives the following examples:

• Khanty: Ма манлам юш хувам ‘I walk along the road’ (Söder 2001: 47).
• Hungarian: Péter a Laci házába megy ‘Peter walks into Laci’s house’ (Söder 2001: 80).
• North Saami: Ándde ja Biera manaiga biillain ‘Ándde and Biera went away by car’ (Söder 2001: 61)

The root has etymological counterparts in all Uralic languages besides Mordvin. Mari has also inherited this root from the proto-language, in the form mijaš. However, in elicitating basic motion verbs from a native speaker using a tool developed by Wilkins (1993), this verb does not appear. Instead, the central verb of motion appears to be kajaš:

Pojan-vlak telə̑m jugə̑ško kajat. ‘Rich people go south in winter.’

Kevə̑tə̑š kajet. ‘You will go to the store.’

The situation in the Turkic languages is similar. A Proto-Turkic root reconstructed as *bar- is found in nearly all Turkic languages (Tenišev 1961: 233). Chuvash has also inherited this root, in the form pyr-, but it does not appear when eliciting basic motion verbs from a native speaker with the tool developed by Wilkins (1993). Instead, the basic verb of motion is kaj-:

Văl kuntan bibliotekăna kajat’ ‘He goes from here to the library.’

Văl xĕveltuxăśalla kajat’ ‘He goes east.’

Chuvash kaj- is generally believed to be inherited from Proto-Turkic, as the motion verb *kajt- ‘return’ found throughout the rest of the Turkic family probably combines this root with a derivational suffix. Mari thus borrowed its verb kajaš from Chuvash.

As noted above, in both languages the original central motion roots are retained, but their meanings have changed. Mari mijaš, as far as I can tell, emphasizes completed motion, while Chuvash pyr- seems exclusively path-oriented or without a deitic centre (I guess). Linking the restructuring of both languages to language contact is an intriguing possibility. Further study is necessary to determine the source and motivations of this restructuring.

References

1. Söder, Torbjörn (2001). “Walk This Way”: Verbs of Motion in Three Finno-Ugric Languages. Studia Uralica Upsaliensia 33. Uppsala: Acta Universtitatis Upsaliensis.
2. Tenišev, E. R. (1961). “Глаголы движения в тюркских языках”. In: Историческое развитие лексики тюркских языков. Москва: Издательство академии наук СССР, 232–293.
3. Wilkins, David P (1993). “Preliminary ‘come’ and ‘go’ Questionnaire”. In: Cognition and Space Kit (version 1.0). Ed. by E. Danziger. Nijmegen: Cognitive Anthropology Research Group at the Max Planck Institute for Psycholinguistics, pp. 29–46.

The Crestomație de literatură română veche edited by I. C. Chițimia and Stela Toma (Cluj-Napoca: Editura Dacia, 1984) that I picked up for cheap last year in a Cluj antiquary has so much trivia on Slavonic hangers-on in the early modern Romanian lexicon that I could do an endless series of posts here. Less visible but often even more intriguing are the signs of contact with Hungarian. A liturgy book produced in Brașov in Transylvania in 1570 features the following lines of translation of Psalm 50: Ție unuia greșiiu și hiclenșug înaintea ta feciu, ca să dereptezi-te întru cuvintele tale și pîrî-veri cînd veri judeca ‘Against thee, thee only, have I sinned, and done this evil in thy sight: that thou mightest be justified when thou speakest, and be clear when thou judgest.’

Here the word hiclenșug ‘betrayal’ is clearly a borrowing of Hungarian hitlenség ‘infidelity’, a perfectly Magyar formation formed by hit ‘belief’ plus the caritive suffix lan/len and then the abstract formation ság/ség. I don’t know how to explain the final vowel in the Romanian form other then by wondering if the Hungarian ending, like many Hungarian endings, originally had an invariable back vowel before it was made to conform to vowel harmony.

The dissimilation before /l/ of /t/ to /k/ is a bit odd too, but things get stranger still when one considers that hiclenșug is archaic in Romanian, with not even an entry in the Dicționarul explicativ al limbii române, but the loan survives still today in the further altered form vicleșug. Now, I cannot help but wonder if that change of /h/ to /v/ is related to the word-final change of /h/ to /v/ in loanwords that I previously wrote about here.

Turning the pages of Timofej Jevsevjev’s Mari song collection as edited by Saarinen (1994) reveals a large amount of stock imagery and phrases with which the Mari liked to make up songs.

Take for example the folksong epigram from Sergei Chavain’s novel Elnet: Кеҥежым кайышым лышташ йымач, / Телым кайышым лум ӱмбач. (‘In summer my course is under leaves, / In winter my course is over snow.’). We find something similar in this song popular among army recruits (Saarinen 1994: 196):

Telə̑m kajə̑šna lum ümbač
keηežə̑m tolna lə̑štaš ümbač
Otə̑ šȇηgal oš luδə̑žə̑m
lüjalal lüjalal iδa nal.
Azjal marijə̑n kačə̑žə̑m
lüjalal lüjalal nə̑ηgajat.

Im Winter gingen wir über den Schnee,
im Sommer gingen wir über die Blätter.
Die hinter dem Hain lebenden (eigtl. seienden) weißen Enten
schießen schießen nehmt nicht.
Die Tscheremissenjungen von Azjal
schießen schießen holt man (zur Armee).

And other material from this song is met with again in a wedding song from some pages later (Saarinen 1994, 216):

Otə̑ šegal oš luδə̑žə̑m
lüjalal nalaš iδa šono.
Polanur marijə̑n üδə̑ržə̑m
lüjalal lüjalal iδa ončo

Die hinter dem Hain (lebende) weiße Ente
gedenkt nicht durch Schießen zu nehmen.
Die Tscheremissenmädchen von Pulanur
seht schießen nicht an.

References

1. Saarinen, Sirkka, ed. (1994). Timofej Jevsevjevs Folklore-Sammlungen aus dem Tscheremissischen. Vol. IV: Lieder. Mémoires de la Société Finno-Ougrienne 219. Helsinki: Société Finno-Ougrienne.

There has been discussion here previously about Tatar’s flip-flop of mid and high vowels, where Common Turkic high vowels /u/, /ü/ and /ı/ were reduced and lowered, and the Common Turkic mid vowels /e/, /o/ and /ö/ rose up to fill their place. Poppe (1968: 9) gives the following vowel inventory for standard Tatar:

Now, Hattori (1993) calls this system peculiar because it has more mid vowels than high vowels. His solution is to interpret the so-called high vowels as diphthongs with a mid vowel onset, basically doing away with high vowels at all in order to avoid the contending between high and mid inventories. For example, the vowel that in the Cyrillic orthography is represented <у>, suggesting /u/, functions structurally as if it were /əw/ (or perhaps /ow/).

Since when is it unlikely cross-linguistically for a language to have more mid vowels than high vowels? Hattori offers no citation for this statement, and I’d like to read more about it.

References

1. Hattori, Shirô (1993). “Phonological Interpretation of Tatar High Vowels”. In: Studies in Altaic Languages. Vol. 4. Selected Papers by Shirô Hattori. Tokyo: Sanseido, pp. 256–264.
2. Poppe, Nicholas (1968). Tatar Manual. Second, Revised Edition. Indiana University Publications Uralic and Altaic Series 25. The Hague: Mouton.

It’s remarkably common cross-linguistically for the word ‘head’ to be found either in source-oriented expressions of motion, or in expressions meaning ‘begin’, both of which might be joined into the same semantic sphere. For the former, consider the English verbal phrases ‘head off to’ and ‘head down to’. In Veps, this phenomenon appears in nominal declension. Here the Proto-Finnic elative and ablative cases have been replaced with a new construction consisting of the inessive or adessive plus the suffix -pAi < Proto-Finnic pää ‘head‘, e.g. perti-špäi ‘out of a house’ and perti-lpäi ‘off a house’ (Grünthal 2003: 123).

For the sense ‘begin’, one can compare such distant languages as Latin and Turkic. In Latin, principio ‘beginning’ is formed with the root for ‘head’, cf. caput. In Tatar, ‘begin (tr.)’ is expressed by the verb bašla- and ‘begin (intr.)’ by bašlan-. These consist of baš ‘head’ followed by derivational suffixes for factitive and intransitive verbs (Poppe 1968: 96–97).

References

1. Grünthal, Riho (2003). Finnic adpositions and cases in change. Mémoires de la Société Finno-Ougrienne 244. Helsinki: Sociéte Finno-Ougrienne.
2. Poppe, Nicholas (1968). Tatar Manual. Second, Revised Edition. Indiana University Publications Uralic and Altaic Series 25. The Hague: Mouton.

You may have noticed that the site has a new name and URL. I am feeling more confident in my research, having nearly completed my M.A. and thinking ahead to my application for doctoral studies, so I think the site deserves a new look. I also want to write in a more serious manner than before, including references to publications cited, so that I can generate both weblog entries and occasionally formal publications out of the same source. However, much of the content here will remain the same: whatever I come across in my day to day reading.

In his textbook Teach Yourself Beginner’s Hindi Script Rupert Snell offers the following charming anecdote:

Legend has it that in the days of the Raj the British memsahibs, indifferent to real Hindi, would learn simple Hindi commands by assimilating them to English phrases: ‘There was a banker’ was to be interpreted by servants as representing दरवाज़ा बंद कर darvāzā band kar ‘Close the door’, and ‘There was a cold day’ meant दरवाज़ा खोल दे darvāzā khol de, ‘Open the door’. Thankfully, those days are long gone

While a Google search on the first phrase gives no results, a search on the second comes up with numerous references to this phenomenon, but only as an urban myth. If this were really a commonplace among Brits in India, one would expect it to appear in some contemporary written form, perhaps something penned as advice to recent arrivals.

While the poetry Gennady Aigi wrote in Russian from the 1950s through the 1980s was published in three large collections after perestroika, his Chuvash-language poems were much smaller in number and appeared in the individual collection Сурхи йӗпхӳ (Springtime Drizzle) put out by Chuvashkoe knizhnoe izdatel’stvo in 1990. These poems date from 1948 to 1989 and reveal that at the same time Aigi was writing in an unusual modernist style in Russian, the poetry in his native language remained more traditional.

This poem is the first in the volume and was written when the poet was 14 years old.

Continuing my exploration of Tatar ortographical ambiguity, I might mention the fairly common words сәгать /sægæt/ ‘hour’ and мәкаль /mæqæl/ ‘riddle’. Typically such words would be straightforwardly spelt with two front low vowel signs, e.g. һәйкәл ‘monument’. However, Tatar speakers here directly represent the frontness only of the first vowel in сәгать. The second vowel is written with a letter typically representing a back low vowel, with only the final Cyrillic soft sign indicating that this vowel too is front. This is all necessary to show that the velar consonants involved are back instead of front.

A further weakness of modern Tatar orthography is an amount of legacy spellings, which makes it impossible to know how to write certain words heard, or pronounce certain words written. Among such spellings which appeared at the beginning of the Cyrillic era and somehow persisted through subsequent standardization, one might use as an example the derivational suffix -ханә ‘place for X’ in e.g. даруханә ‘pharmacy’ (< дару ‘drug’ + -ханә). This suffix, even when attached to words with back vowels, is in fact pronounced with two front vowels: /xænæ/. In seeing the written form, the learner must wing it and assume that both vowels have front articulation based on principles of vowel harmony.

But one cannot always assume vowel harmony is at play, because there are plenty of loanwords where vowel harmony does not apply, e.g. гаилә which plays by today’s rules in being precisely /ɣajilæ/.

While its modified Cyrillic script in general serves Tatar quite well, and I don’t understand the urgency with which some call for a switch to the Latin alphabet,there are some cases where Tatar’s Cyrillic system is frustratingly ambiguous.

In Proto-Turkic the velar stops had back allomorphs [q] and [ɣ] before back vowels, and front allomorphs [k] and [g] before front vowels. If articulation depends on context, then Tatar needs only one symbol each for the voiced and unvoiced velar stops and it ostensibly makes sense to use only the Cyrillic letters <к> and <г>. In squarely Turkic words like кар ‘snow’ and туган ‘birth’ one can expect the back allomorphs, while кил ‘come’ and бүгeн ‘today’ must contain the front allomorphs.

The flood of Persian-Arabic loanwords with freely occuring /q/ and /ɣ/ has now split those allomorphs into two distinct phonemes, and it has also broken the dependable vowel harmony of the Turkic languages. Now, determining the quality of the velars and/or frontness or backness of vowels is a laborious process. Consider the word шагырь ‘poet’. This Persian loanword (< شاعر) has a voiced velar fricative proceeding a front vowel, but the Cyrillic system employed by Tatar has no means of straightforwardly representing that, unlike e.g. Kazakh which chose to use the standard Cyrillic letters everywhere to represent the front allomorphs and introduced <қ> and <ғ> for the back forms. In шагырь, the Cyrillic soft sign at the end of the word is necessary to show that the preceding vowel has front articulation, but the back vowel letter ы is then employed to show that the velar is pronounced as [ɣ]. If it were taken at face vowel, one might think the word were pronounced [ʃaɣɨrʲ], but one has to read between the lines to get at the correct pronunciation [ʃaɣir].