Talk:Indo-Uralic languages
From Wikipedia, the free encyclopedia
 |
This article is within the scope of the WikiProject Languages, an attempt at creating a standardized, informative, and easy-to-use resource about languages. If you would like to participate, please visit the project page, where you can join the project and see a list of open tasks. | |
| ??? | This article has not yet received a rating on the quality scale. |
Contents |
[edit] Cognate table
I decided to add a meaty list of possible cognates for budding Indo-Uralicists since this article looked a little empty.--Glengordon01 18:27, 17 April 2006 (UTC)
I took a link out of the entry of 'I, me'. I didn't think that a site written entirely in Finnish is going to be helpful to the average non-Finnish speaking Anglophone, quite frankly. Secondly, it will get godawful-nasty if we add links like this to every word in this table. Kinda pointless. If people want to track individual words down, they can get a dictionary. This article is about Indo-Uralic -- not Finnish, or Hungarian, or Samoyedic, etc. I think it makes more sense to restrict the topic here to IE and Uralic reconstruction, void of the specifics. The specifics will be found in their respective articles afterall, n'est-ce pas? --Glengordon01 04:12, 18 April 2006 (UTC)
[edit] Assibilation
Added by Vuo
Word-final assibilation of -ti occurs in modern Finnish, a Uralic language.
Irrelevant. This IE-Uralic correspondance works for both the plural (IE *-es, Ur *-t) and the 2ps (IE *-s, Ur *-t). Finnic assibilation (as in Ur *weti > vesi) has absolutely nothing whatsoever to do with the preceding sentences in that paragraph. --Glengordon01 04:06, 18 April 2006 (UTC)
[edit] The third person and the stem *i-
i for thir person? Proto-Uralic had -βa. -- anonymous
Yes, Proto-Uralic *sa- (Finnish hän). I'm afraid you're right: There is no expression of the equivalent of IE *ʔi- in Uralic. I'm thinking Altaic and mixing diachrony too. Horrible! Personally, I still think there is an early stem *i for 3rd person oblique here but my views on Indo-Uralic are different and I have trouble calling this stage we're talking about "Indo-Uralic". I consider Uralic, IE and Altaic representatives of three seperate branches (so I don't quite accept the Indo-Uralic hypothesis in its purest form). Indo-Aegean (IE plus Aegean) and Altaic shared special isoglosses since they were both located in the southern regions of this main proto-group, situated somewhere in Central Asia circa 9,000 BCE. With the three language families, *i could be better established. Internally in Indo-Aegean, both *ʔi- and *ʔe- are relatable due to the Centralization event when *i and *u were replaced with central vowels early on in pre-IE: merging to schwa *ɜ (> IE *e) in accented closed syllables, becoming diphthongs *ei and *eu respectively in accented open syllables, or becoming *ɜ and *a respectively in unaccented syllables. This was probably due to areal influence with an early form of Proto-Abkhaz-Adyghe. I reconstruct Indo-Aegean 3ps animate *i, inanimate *in and general 3ps oblique *e, all from earlier 3ps *i / *in and therefore relatable to Altaic *an. It gets more complicated because of split ergativity in the 3ps inanimate 'n' stuff. But that's just my shtick and Wikipedia isn't interested in that because it's too "POV". --Glengordon01 14:41, 6 July 2006 (UTC)
- That's all very interesting because you seem to reach the same conclusions as the Nostratic hypothesis! According to the latter, PIE went through a Great Vowel Steamrollering (as it should IMHO be called) that turned all vowels into *e, except when preceded by **/ʔ/ or when **/i/ and **/u/ became *ai/ei/au/eu. That seems to be the same as your centralization. I only wonder how you arrive at the precise value [ɜ]. Isn't that so precise as to be untestable? – Split ergativity occurs in Kartvelian, another Nostratic branch.
- Have you published any of this? I'd like to read more… and of course, if it's published, it's no longer Original Research…
- David Marjanović | david.marjanovic_at_gmx.at | 00:13 CEST | 2006/10/29
[edit] Deleted sentences
I deleted a couple of sentences which included somewhat unclear or dubious claims:
- "Indo-European and Uralic languages look remarkably similar for neighbouring languages traditionally considered unrelated." -- This is a subjective impression, and as such hardly relevant; moreover, Proto-Uralic and Proto-Indo-European look rather different to me at least.
- "One challenge to this research is that it is often assumed that similar words in Uralic and Indo-European languages are simply loans from IE to Uralic, even if this is not chronologically possible" -- I deleted the part after the comma - this could be reintroduced to the article if explained and referenced.
- "We know now that IE and Uralic, both dated to approximately 4000 BCE, were probably very distant from each other at the time. Uralic would have been northward up the Volga River while, if the Ukrainian thesis is correct, PIE would have been centered in the north-west Pontic region." -- While I'm inclined to agree with the idea, I think that the verb know is too strong in claims of speaking areas of proto-languages in remote prehistory; at least, a reference should be provided. --AAikio 20:06, 3 September 2006 (UTC)
This seems very reasonable to me, actually. --Glengordon01 05:01, 5 September 2006 (UTC)
[edit] Vetää
(quote: Ante Aikio) "rv - this is regarded an IE loan by standard etymological dictionaries"
Isn't "nimi" and "myydä" also regarded IE loans by standard etymological dictionaries?
The word "vetää" is in the Illich-Svitych's Nostratic poem: k̥aλai palhʌ-k̥ʌ na wetä --Muhaha 13:34, 9 November 2006 (UTC)
[edit] On kuningas etc.
A user with the IP address 82.181.182.83 has reverted a large number of edits back to a previous version of his. Might I offer a few suggestions:
(1) If you wish to remove extensive sections of correctly referenced and formatted text, it might be desirable to get a username so we can track your edits and you can participate with maximum credibility in the Wikipedia process.
(2) You have reverted several formatting errors that I had (at the cost of some labor) corrected, for example in italicization and in use of ś rather than š as per standard Indo-Europeanist usage. (Indo-Iranian šatam gets š as per Indo-European, as does Old Iranian when this sound occurs, e.g. in xšwaš '6'; Sanskrit śatam gets ś because of the standard transcription adopted in the late 19th century. Thus there is inconsistency in the standard transcriptions, but these both stand for the same sound.)
(3) You have eliminated the example kuningas, but this is a standard example in comparing Indo-European and Uralic, used for instance by Henry Sweet in The History of Language (1900) and by Raimo Anttila in Historical and Comparative Linguistics (rev. ed. 1989). In addition, the phonology of this word happens to be particularly clear and easy to convey. Readers deserve to be informed of it. I grant that kuningas is a more recent loan than porćas (also a standard example), but this usefully emphasizes the very long period during which Indo-European and Uralic languages have been in contact and, consequently, have had the opportunity to interact. Let’s give readers both.
Please note that I did not eliminate your addition of porćas but strove to integrate it with the existing text of the article. It is a real contribution.
(4) As to the information you provide on Koivulehto’s arguments, I welcome this. He is clearly a major authority in the field.
(5) You are obviously in a position to make substantial contributions to this and other Wikipedia articles. Please, let’s not get into an editing war (the bane of the wiki format). If you think there are points of view which are left out or need to be emphasized more, I’m sure I would be sympathetic.
(6) There is no article on Koivulehto which is why his name appears in red in the “References” list. Perhaps you could contribute one?
Regards, VikSol 04:12, 22 February 2008 (UTC)
Thanks VikSol,
I take your point and I did overreact to the numerous editing needs in your text by reverting it completely back to an older version. I now see that there is a better way with this discussion site. I tried to make a more careful edit, just adressing the problematic issues as limited as possible and leaving as much as possible intact. I hope we can continue in this constructive way.
(2) (numbers refer to your points above) For the time being (I remain waiting for you to give a more thorough explanation) I left your *š in the text for Indo-Iranian *šatam even though it is a most confusing transliteration. It is generally used for a later Proto-Baltic sound which is already depalatalized and therefore substituted by non-palatalized Finnic *š becoming later Finnish /h/. In contrast the Indo-Iranian sound was still palatalized, as attested by Indo-Aryan and by many Finno-Ugrian loanwords. Only in Proto-Iranian was the sound depalatalized, whereas Indo-Aryan remained with the palatalized *ś. In fact the possibly Indo-Iranian original for *porćas ‘piglet’ is a case in point, it would still have contained a palatal PIE *ḱ or Indo-Iranian *ć (the word is not attested in Indo-Aryan). So right now the text is not internally coherent. How shall we go about? Shall we explain to the reader that in this particular case of *šatam, *š stands for a palatalized sound, or shall we just change it to the unambiguously palatalized *ś?
(As for *ć/*ś it has recently (1999 see Koivulehto 2001, p.255 and references) been shown, that the Proto-Iranian articulation must still have been affricate when the Pre-Iranian palatalization was lost because in Nuristani the representation has conserved a depalatalized affricate /c/ = [ts]. Logically the affricate manner of articulation, as opposed to a fricative one, is a more archaic stage of development intermediate from the PIE palatovelar, and must be reconstructed into Proto-Indo-Iranian as well, but with palatalisation. The original for *śata on the other hand, may well be a younger borrowing from the neighbouring Indo-Aryan branch)
The problems your text highlighted for the etymology of *porćas ‘piglet’ are not really problems. I adjusted the text to explain briefly how the substitutions would have worked. What grounds do you have to put forward an alternative reconstruction **porkos? All satem languages bears testimony of palatalisation, jus as Finno-Permic does! (By the way: Permic has been argued to be a borrowing from Mordvinic, therefore I changed to Finno-Mordvinic)
(3) The example of *kuningas is of course very illustrative but due to its very young age (may be as late as Proto-Nordic, considering its three syllable structure in Proto-Finnic) utterly irrelevant for the historic period considered. In the good spirit of cooperation that you have shown also, I found a way to leave it in the text ignoring one additional problem: Note that from a Uralic point of view, to be precise, the nonsense suffix is not *-as but *-ingas, because **kuning- is not a stem in accordance with phonotactic limitations for Uralic stems.
(6) An article on Koivulehto would take some time. There is one in German in the prologue of Verba Mutuata but I am note sure whether copyright allows us to bring a wild translation of it here? Ay least I would favour caution.
Finally I take issue with the highly problematic mentioning that Proto-Uralic would have been spoken 6700 BCE throughout its historic area. I substituted it with a more mainstream reference bringing the dating a couple of millenia foreword. In fact a very strong case has been made as well by Petri Kallio to bring the datings even more foreward http://www.kotikielenseura.fi/virittaja/hakemistot/jutut/kallio1_2006.html by one or two more millenia or so. Unfortunately only the abstract is available in English. If you insist on Mithens radical datings (not to mention his wide Urheimat!!) we would need to mention Kallios article as well to balance it. But is it necesary for the article at hand, would a short mentioning of the mainstream view be enough? I admit there is a strong logical link to the Indo-Uralic hyphothesis, but the whole subject would be a pandoras box expanding the article disproportionally into archeology.
—Preceding unsigned comment added by Dinji (talk • contribs) 22:45, 24 March 2008 (UTC)
Dinji,
(1) Regarding ś versus š:
It’s a classic problem in linguistics for different signs to be used for the same sounds when representing different languages. For example English sh and German sch both represent the voiceless postalveolar sibilant, IPA [ʃ], but use different combinations of letters to represent it. Similarly, in the scholarly transcriptions adopted to represent non-Roman writing systems, sometimes this symbol, sometimes that symbol is used, often depending on what was in fashion at the time the language first came under intensive study. Thus it happens that in Indic studies [ʃ] is represented by ś (earlier often by ç), but in Indo-European studies by š, including in reconstructions of Proto-Indo-Iranian. š also represents [ʃ] in Iranian studies.
The regular outcomes of Indo-Iranian [s] and [ʃ] are [s] and [ʃ] (transcribed s and ś) in Old Indic (i.e. no change), but [h] and [s] in Old Iranian (i.e. [s > h] and [ʃ > s]). However, sandhi phenomena preserve many [ʃ]’s in Old Iranian. So both languages had [s] and [ʃ], but most often not in the same positions.
In short, [ʃ] is represented by š when discussing reconstructed Proto-Indo-European, Proto-Indo-Iranian, Proto-Iranian, and Iranian, but by ś when discussing Indic languages (including modern Indo-Aryan languages). Iranian š = Indic ś = IPA [ʃ]. This is a mere convention and has no absolute justification.
My understanding is that in Uralic studies, where you obviously have much more extensive knowledge than I do, ś ordinarily represents the palatal sibilant [ç] and ć the palatal stop [kʲ].
(2) I am aware of the view that the Nuristani languages have an affricate from an earlier č [tʃ] where Sanskrit and Old Iranian have [ʃ], but the matter is a bit complicated and has not really been resolved to my knowledge (I haven’t looked at the material by Koivulehto you mention). Some claim the Nuristani c [ts] is of uncertain origin and that consequently no claims can be based on it. (Perhaps what they have in mind is a development [s > ts].) Others see it as attesting a pre-Indo-Iranian state of Indo-European, when [kʲ] had become the affricate č [tʃ] but not yet the sibilant š [ʃ]. The view you mention, namely that Proto-Indo-Iranian still had č, is probably the dominant one currently, but I see several problems with it. For one thing, Old Iranian and Sanskrit both reconstruct back to a common point [ʃ], not [tʃ], and this [ʃ] is an essential sound in a coherent, integrated sound system. I see no necessity for [tʃ] from [kʲ] to have still existed at this point.
Another problem, which you touch on, is that Finno-Ugric *śata [çata] evidently continues a *[ʃata-] from one of the Indo-European successor languages. This cannot be Baltic, since the PIE m̥ in *km̥tóm would lead to im, not a, and (for various reasons) is not likely to be Slavic. Indo-Aryan is a possible source phonologically but is ruled out by geography as a practical matter. *śata [çata] thus comes from an Indo-Iranian *[ʃata-] or, presuming it had not yet become [sata-], an early Iranian *[ʃata-] (but I am skeptical the language prior to the [s > h] and [ʃ > s] changes can be linguistically defined as Iranian). To me it needlessly complicates the chain of events to suppose Indo-Iranian still had [tʃ], not yet [ʃ].
This suggests to me that *śata comes from Indo-Iranian *[ʃata-], not any earlier or later state of the language.
(3) Your point on the location of Uralic is well taken. Actually, Mithen only claims one site of 6700 BC and later as quite possibly Uralic (based on continuity in a peculiar burial custom). This plausibly locates Uralic in its "general" zone, i.e. to the north of Indo-European. I did not mean to imply Mithen claimed to have located Uralic in the full extent of its historical area.
As you suggest, I’m not sure how much we need to get into a detailed chronology of IE and Uralic in the context of this article. After reading David Anthony’s book, though, it seems to me that the early chronology of IE and Uralic is going to become increasingly relevant. If so, it might become desirable to add a section on chronology.
(4) The defense of using *kuningas is not that it is the most linguistically relevant example but that it is the easiest example for the uninitiated to grasp. So, logically, it could not be in first position, but, pedagogically, there is a case that it should be. However, I am willing to let this go as non-essential.
(5) The reason for reconstructing the Indo-European antecedent of *porćas as either *porḱos or *porkos is that Indo-Europeanists are not in agreement on whether the palatovelar series, ḱ ǵ ǵʱ, IPA [kʲ gʲ gʲʱ], was part of the original language or a later development from k g gʱ, such that k g gʱ calved into two separate series, k g gʱ and ḱ ǵ ǵʱ. The classic view since Brugmann’s time has been that the two series were original, but the trend at present (as I read it) is to view only the plain velars as original, and the palatovelars as a development from them. Still, both views have currency. The dual reconstruction *porḱos or *porkos is meant to be agnostic (a) as to which the original form was, and (b) as to which form was the direct antecedent of Finno-Ugric *porćas.
Possibly the answer to (b) can be determined. I have very little experience in how Finno-Ugric ć (which I presume represents [kʲ]) correlates with Indo-European ḱ [kʲ]. Since ć is phonetically identical to ḱ, it is intuitively attractive to suppose that it was used for IE ḱ, but I don’t know this for a fact; I can also imagine that ć might replace IE k on occasion. Is there a simple or a complicated answer to this?
(6) I have reworked the current version a bit in light of considerations raised, nothing radical.
(7) I don’t know whether copyright restrictions would apply but even a sentence or two providing biographical information would be useful and could always be elaborated.
Regards, VikSol 01:12, 28 March 2008 (UTC)
Hello VikSol,
From what you write I see that the problem is not really about transliteration but we really do not share the understanding of Indo-Iranian phonology. I am the first to admit that I am reliant on litterature mainly dealing with historic Uralic contacts. But I am confident that this research is bound to use mainstream views of Indo-Iranian soundhistory. The Wikipedia article on Classical_Sanskrit supports my view. Moreover the phonological shape of lexical borrowings in Finno-Ugric may contribution significantly as primary evidence as well.
Proto-Uralic, Finno-Ugric and later Pre-Finnic made a very rigorous destinction between so called palatalized sibilants and non-palatalized sibilants. The palatalised sibilants were:
1) /ć/ = IPA [ʨ] voiceless alveolo-palatal affricate
2) /ś/ = IPA [ɕ] voiceless palatalized postalveolar (alveolo-palatal) fricative,
The non-palatalised were:
3) /č/ = IPA [ʧ] voiceless postalveolar affricate,
4) /š/ = IPA [ʃ] voiceless postalveolar fricative,
5) /s/ = IPA [s] voiceless alveolar grooved fricative,
This richness in sibilants makes Finno-Ugric an excellent indicator of the process of depalatalisation in Indo-European. The most known depalatalisation was that of centum languages. Before the centumisation PIE ḱ was substituted by FU /ć/ = IPA [ʨ]. After the centumisation a velar stop was applied.
In the light of FU borrowings a depalatalisation of Proto-Baltic also occurred early. Before that depalatalisation satemized Balto-Slavic /ś/ = IPA [ɕ] (from PIE ḱ) would have rendered Pre-Finnic /ś/ = IPA [ɕ]. Later Proto-Baltic was however depalatalized because the substitution in Finnic is consistantly non-palatal post-alveolar /š/ = IPA [ʃ]
Archaic Indo-Iranian borrowings always has a substitution of historical PIE palatovelars with palatalised FU /ć/ = IPA [ʨ] or palatalised FU /ś/ = IPA [ɕ]. Borrowings from Indo-Aryan likewise. There is no other explanation possible than the fact that PIE ḱ has developed in the palato-alveolar position from Indo-Iranian /ć/ = IPA [ʨ] and further to Indo-Aryan /ś/, which definitely cannot be IPA [ʃ]. Even after developing into Classical_Sanskrit no IPA [ʃ] has appeared in the phonemic system: /ś/ is still represented by IPA [ɕ].
The level of reconstruction common for Avestan, Old Persian and Nuristani on the other hand had already depalatalised sibilants, meaning that alveolo-palatal sounds had become alveolar (perhaps through a post-alveolar intermediate stage). Thus Indo-Iranian /ć/ = IPA [ʨ] had become Proto-Iranian [ʦ] (voiceless alveolar affricate). In Proto-Iranian borrowings the Pre-Finnic substitute of this affricate is intervocalically [-ks-] because Pre-Finnic had neither that affricate or the sequence **ts. Later, after the Iranian affricate had become plain fricative it was substituted by plain [s].
Based on this discussion I will change “Indo-Iranian *šatám” into “Indo-Iranian *ćatám”
On point (5) the controversy is in effect whether there should be reconstructed three points of articulation or just two, taking account the labio-velar series as the third. You are right that there seems to be a tendency to assume, that orignally ther were just two series, the labiovelar one and the palatovelar one. You are perfectly right that as the middle series (plain velar in classic description) is questioned, the palatovelar series has been thought to have aquired its palatalisation as a secondary fronting. Nevertheless, in transliteration it is the standard to keep up the three way convention. And as transliteration is heavily phonemic-based (as opposed to phonetic based) it would certainly not be logical to give up the marked transliteration for the palato-velar series. —Preceding unsigned comment added by Dinji (talk • contribs) 21:06, 17 April 2008 (UTC)
Regards, Dinji —Preceding unsigned comment added by Dinji (talk • contribs) 20:29, 17 April 2008 (UTC)
[edit] Bad quality article
The section Arguments for relationship between Indo-European and Uralic mainly gives arguments against. Somewhere in the history, I believe, someone actually wrote a section of arguments for the theory. Then others developed the text so that each paragraph also held counterarguments, and so, without keeping an overview, the paras became kind of a negotiation consensus of the opinions proponed by the editors. This is wrong and in disaccord with the principes of wikipedia, which intends to be a reflection of the debate outside, not of the opinions inside. The article should be layouted approximately:
- the arguments for should contain a discourse, a coherent set of arguments leading to the conclusion that PIU is reasonable to believe,
- the arguments against should as systematically as possible try to disprove each argument for, and probably also refer to scientific principles used within linguistics and general philosophy.
Now they're mixed, in such a way that for and against cannot be understood by a non-linguist: for example para 2 in "Arguments for...":
- Numerous words in Indo-European and Uralic resemble each other. The problem is to weed out words due to borrowing...
Now, is "the problem..." an argument against, or is it a sweeping reference to a set of methods used for "weeding out" true cognates? Then a specific method should be referred to, or the "weeding out" problem is a counterargument that shouldn't be the second sentence in the second para in a for-section. Said: Rursus ☻ 09:00, 10 May 2008 (UTC)
- I think I'm inclined to write a short method for writing a compact article. Said: Rursus ☻ 09:02, 10 May 2008 (UTC)
[edit] Not if you read it more carefully
From a scientific standpoint, the methods which serve to identify borrowed words leave genetic cognates as their residue, should there be any, so it makes no difference whether they are applied from an "anti-" or a "pro-" perspective, in principle. There may of course be tendentiousness on the part of some applying them, but if correctly applied the methods should work. It is just as important to weed out borrowings for those who hope to establish genetic relationship as for those who hope to disprove it. With all due respect, it's a little over the top to condemn an entire article on this ground. Nevertheless, I may give some thought to clarifying some of the phonological material which in my opinion is a bit abstrusely presented. Sincerely, VikSol 13:50, 10 May 2008 (UTC)
[edit] Section removed
I have removed the section “Some example words” on the ground that it focuses on words in modern languages, whereas to establish genetic relationship between languages it is ordinarily preferable to use the earliest forms available, e.g. not Russian but Old Russian, or better still Old Church Slavic, likewise not English but Old English or better still Proto-Germanic, since it can be reliably reconstructed and is nearly identical with the earliest runic inscriptions. For example, we can more easily see that Proto-Germanic ek is related to Indo-European *egom than that English I or even Old English ic is, especially once we know the sound law that Indo-European g became Proto-Germanic k. Also, this section was largely superseded when the table "Possible cognates" was added. On the other hand, I think this section retains enough interest that I have moved it here rather than simply deleting it. VikSol 18:06, 26 May 2008 (UTC)
[edit] Some example words
Indo-European: French 'moi', English 'me', Russian 'menja'; Finno-Ugric: Estonian 'ma', 'mina', Finnish 'minä'.
IE: English 'water', 'wet', Russian 'vodá'; FU: Finnish and Estonian 'vesi' (oblique 'vete-'), Hungarian 'víz'
IE: Russian 'vodít' (to lead), FU: Finnish 'vetää' (to pull, to lead), Hungarian 'vezetni' (to lead)
Some Indo-European roots (I.E.) and their equivalents in Finno-Ugric languages (F.U.)(examples are added from different languages too):
I.E. mey-, to exchange (derivatives include Latin 'mutare' (to change), German 'mischen' (to mix)), , F.U. meqi-, to sell, give > Estonian 'müük' I.E. mesg-, to wash, F.U. moski-, to wash > Estonian 'mõskma', Hungarian 'mosni'
