Talk:Proto-Finnic language

Notation of voiced obstruents

Is the difference between b and β (and analogously for the others) really useful? The difference between p and b wasn't even phonemic, let alone the difference between b and β. I think it would be better to just write b everywhere. CodeCat (talk) 15:20, 1 June 2013 (UTC)

If we're going to be marking allophones (and it's definitely necessary for a discussion of gradation), I see no reason to not also record the stop/spirant distinction. This also would make difficult discussing the distinction between the Late PF stage and older ones; the original situation likely still involved [b d g] in most positions.

Also, if you want to argue for a single symbol, most of the literature uses β δ γ. I don't have Laakso (2001) at hand to check, but b d g sounds like some kind of an idiosyncrasy. --Trɔpʏliʊm • blah 15:44, 1 June 2013 (UTC)

Proto-Germanic uses a similar convention of denoting both stops and fricatives allophonically with the same letter, even though they are distinguished in some (but not all) literature. CodeCat (talk) 15:52, 1 June 2013 (UTC)

True. The difference is that voiced obstruents are phonemic in PGmc compared to their voiceless counterparts. The cover symbol */b/ is thus used to cover two allophones *[b], *[β]. The analogous level in PF would be the cover symbol */p/ to cover three allophones *[p], *[b], *[β]. I don't see why we would need to add a third, imtermediate level of transcription to cover [b] and [β] but not [p]? --Trɔpʏliʊm • blah 16:32, 1 June 2013 (UTC)

On the other hand, the difference between β and b is much more easy to recognise, because b only occurs after a nasal, just like in Proto-Germanic. The difference between p and b on the other hand, while still predictable, takes a lot more work for the reader to decypher because they have to figure out whether the syllable is closed or open. So for me, the difference between plosive and fricative feels a lot more redundant than the difference between voiced and voiceless. It may not be phonemic synchronically, but from the point of view of a casual reader it would seem that way. CodeCat (talk) 16:36, 1 June 2013 (UTC)

Vowel syncope and apocope

In Finnish and many other Finnic languages, there are words that syncopate stem-final vowels under some conditions. For example, Finnish has:

• vesi ~ veden ~ vettä (expected *veteä)
• lapsi ~ lapsen ~ lasta (expected *lapsea)
• nähdä ~ näen ~ näkevä (expected *näkeä)
• Superlative obsolete essive singular -inna (< presumably -impna, compare modern -impana)

Finnish even has a few words where this contraction in the partitive results in a break of vowel harmony: meri ~ merissä but merta. As far as I know, Proto-Uralic roots always ended in a vowel, which was inherently part of the root. So the only way to account for this that I know of is a regular sound change that syncopated the vowel under some circumstances, some time in Pre-Finnic times (presumably). Are there any sources that describe this phenomenon and the conditions in which it applies? It seems to happen only with stem-final -e, but never with -a or any other vowel (I think Uralic roots could only end in -e or -a).

It seems that final vowels were also apocopated in some situations. In Finnish there is a rather large word class with the suffix -us that goes like: -us ~ genitive -uden ~ essive -utena ~ plural inessive -uksissa. As far as the singular is concerned, this very clearly resembles uusi ~ uuden ~ uutena, except that the final -i of the nominative has been dropped. But it must have been there at some point because that's the only way to account for the final -s in the pattern of alternations. There's also the superlative which goes like -in ~ -imman ~ -impana, where the apocopated vowel seems to have been -a. So why was it dropped? Is there a regular sound change that accounts for this? I suppose there's also the more general question of the distinction between vowel stems and consonant stems in Uralic. Roots should have all been vowel stems, but what about derived stems? Were there true consonant stems in Uralic? CodeCat (talk) 14:36, 26 February 2014 (UTC)

Consonant stems in Proto-Uralic are pretty much an open question. I don't think any general analysis has been made yet, only observations on the "existence of tendencies" to drop a non-close stem vowel from inflected forms — which turns up widely in those Uralic languages that do retain such vowels at all. (My best guess? There does not seem to have been any distinction between zero and an unstressed non-close vowel in PU. The situation may have been something like in contemporary Mansi and Khanty, with an epenthetic vowel *[ə] inserted in various places to resolve awkward phonotactic structures.)

I also haven't seen any general treatment of this for Proto-Finnic, though here there could well be something I'm not aware of. FWIW one recent paper to brush on the topic in English would be Kallio's The non-initial-syllable vowel reductions from Proto-Uralic to Proto-Finnic if you want to start digging up references. Most of the relevant literature is in Finnish though.

Oh wait, I see you've already cited him in the article. Still worth noting for anyone else reading, I guess. --Trɔpʏliʊm • blah 11:23, 27 February 2014 (UTC)

What we do know to have happened in Proto-Finnic or earlier is at least:

• apocope of *ə occured in open syllables, in all cases before coronal consonants, and in most cases only after coronal consonants too. Though some cases have other contractions: *-mət- > *-mt- > *-nt- is fairly regular, e.g. lumi : lunta); *-kət- > *-kt- > *-ht- is found in nähdä tehdä : näke- teke- (but still mäkeä lukea pukea etc. rather than ˣmähdä ˣluhda ˣpuhda).
• *ə was also lost after semivowels, creating diphthong stems like soi-, täi, or long vowel stems like syö-, vie-
• syncope of *ə was regular beyond the 2nd syllable. Hence in the inherited lexicon of the Finnic languages there are only words like askel : askele- and avain : avaime-, none like ˣaskeli or ˣavaimi. OTOH, this probably was not direct loss, but went thru the *i stage (as indicated by your example of -us adjectives, as well as ordinal numbers: e.g. neljäs : neljännen.)

Citations might have to be pieced together from disparate sources, though.

The superlative, then, is not as different a story as it looks on the basis of Finnish. This originally ended in just *-in : *-ime- and is inflected in Finnish with -impa- by analogy to the comparative. Several Finnic varieties only have analogized this as far as -impe-, IIRC. --Trɔpʏliʊm • blah 11:17, 27 February 2014 (UTC)

From what I can see, regardless of whether consonant stems were different from ə-stems at an early time, they had a strong tendency to be merged right from the start. The epenthetic -e- in Finnic is one way that could have happened, and as you say perhaps Proto-Uralic also knew such an epenthetic ə. So it becomes hard to tell from the few remaining bits of information, whether ə was original and then syncopated, or whether ə was not original but added by epenthesis. There are some loanwords where we can be pretty sure they were loaned as consonant stems originally, like *kuningas. But I don't think *mees is a loanword, and because it's monosyllabic there could hardly have been any apocope. So this might be the only remaining true consonantal root noun.

About the comparative, I wonder why it has *-mpi in the nominative but *-mpa- elsewhere. The nominative should have been *-mpa, so where does the -i come from, and why was that -i not apocopated if it is old enough?

Concerning meri in Finnish, do you think this could have been loaned as *mëri originally (it must be an IE loanword after all), and therefore have had back harmony? If that's true, then the partitive merta is an archaism (because it's a frequently used form), while the front vowel ä was introduced to the other endings by analogy, once ë and e had merged. On the other hand, Võro has "meri" too, not "mõri", so maybe not. CodeCat (talk) 13:56, 27 February 2014 (UTC)

Is the back unrounded vowel an archaism or an innovation?

Proto-Uralic had a back unrounded vowel (variously noted ï or ë), and so do the southern Finnic languages (written õ). Those that preserve vowel harmony still retain it as a back counterpart of e, while those that don't use it only in initial syllables. So I wonder, is this a retention that should be projected back into Proto-Finnic, or was this feature lost on the way to Proto-Finnic and then regained later? CodeCat (talk) 19:08, 26 February 2014 (UTC)

This is a debated point. *ë was considered an archaism up to the mid-20th C. Later on it was determined that in the first syllable, this vowel occurs just about solely in IE loanwords that originally had *e (cf. e.g. rõngas "ring"), which led to a theory that an assimilation *e-a > *ë-a had occurred in the South Finnic languages. This is what most current general-level references state.

On the other hand: the Southern Finnic group has by now turned out to be areal; and we also know that Votic and South Estonian distinguish õ from e even in unstressed syllables, including in some words after the neutral i where 2ndary harmonization is not an option. This seems to require an original phonemic status for *ë after all.

So there has been in recent times (the last 20 years or so) some support for a more complex history. Something along these lines:

1. The Proto-Uralic non-open stem vowel becomes a full vowel *e/*ë depending on vowel harmony
2. Loanwords of a shape *e-a assimilate to *ë-a in Proto-Finnic times
3. Later on, *ë shifts back to *e in the North Finnic group (which may have a better claim for not being areal)
4. Even later, vowel harmony is lost in Livonian and Standard Estonian

However, I don't think any papers to argue on this topic in detail have come out so far (though I know there are a few forthcoming) so putting this kind of a summary up in the article would end up WP:SYNTH…

— PU *ë, though, develops to PF *a in all cases (e.g. *mëksa "liver" > Estonian maks, Votic mahsa) and is not related to this issue.--Trɔpʏliʊm • blah 09:18, 27 February 2014 (UTC)

Then it seems to me that the article is ok as it is now, with a cited bit stating that ë in unstressed syllables is an archaism, but without saying anything on stressed syllables. The part that says PU *ë developed into *a is interesting though... maybe we should try to include a list of changes leading up to Proto-Finnic as well? CodeCat (talk) 13:47, 27 February 2014 (UTC)

The 2nd person plural verb ending

There is something strange about this ending. It's reconstructed in the table as *-tte/Ak, but this is a long consonant in a closed syllable, so it violates consonant gradation. There probably was a final consonant, because final -e becomes -i otherwise. So how is that possible? CodeCat (talk) 20:20, 4 March 2014 (UTC)

Hey guys, need to use {{citation}} not {{cite book}} etc.

To CodeCat and Tropylium: Your {{sfn}} links don't work because you're using the wrong kind of citation template. The {{citation}} template automatically adds the appropriate argument so that Harvard-style citations work; the other's don't. Benwing (talk) 20:21, 4 April 2014 (UTC)

Middle vs. Late

We seem to be running here two different standards on where to cover Middle Proto-Finnic changes.

Problems would begin if we were to add *č > *t, *čč > tš into the PU > PF section. This is the point where South Estonian already diverges, as *čk develops to *ck rather than *tk there. Note also the following points of chronology:

1. *č > *t does feed into *ti > *ci: e.g. *künči : *künče- > *künci : *künte- > Finnish kynsi : kynte- 'nail'.
2. Contra Kallio (2007), most other research has continued to hold that *ti > *ci did pass thru a stage *ći, and that this changed to *ci only during the general Finnic loss of palatalization. The fact that *-δ´i also develops to *-ci does turn out to then require two series of depalatalization, but the merger of *δ´ and *δ is found also in Samic and Mordvinic, and has been considered by various researchers a pre-Finnic change that had nothing to do with the shifts *ś > *s or *ń > *n.
3. The change *š > h also happened later than *ti > *ci, as *lehti 'leaf' shows.

Hence much of this has at least as good claim for the "Later developments" section as the cluster reductions like *kt > *ht/*tt. The stage of Proto-Finnic that was common to all three Finnic main groups still closely resembled Proto-Uralic: it featured e.g. postalveolars and palatalization, and had no *h or assibilation of *t.

Now several protolanguages have similar phenomena as well: dialectal divisions appear much before the last changes common to all attested languages were completed. Wikipedia has worked around this issue by e.g. forking all Late Proto-Slavic developments from Proto-Slavic to a separate article History of the Slavic languages. Likewise Proto-Norse and Old Norse are usually recognized as distinct entities. But here the convention in the source literature remains to usually work with Late Proto-Finnic forms, common only to Gulf Finnic (as this is a much more prominent branch than South Estonian or Livonian). This is what our current phonology section does as well. Perhaps we should hence discuss the late Proto-Finnic period still in this same article. Yet, I belive it would be best off explicitly separated from the late changes specific to individual languages.

--Trɔpʏliʊm • blah 21:04, 23 August 2014 (UTC)

Maybe the "later developments" section should be named "dialectal developments"? This is really quite a common problem, and it's something that even modern languages have to deal with. For example, how do you describe a single "English" language when there are really several dialects that are rather divergent? The variety within modern English is probably still more than what can be reconstructed for Late Proto-Finnic. Something like kt versus tt or tk versus ck is a very minor thing and would not have hampered intelligibility in the slightest, it would have just given the inland Proto-Finnic speakers a slight accent. Compare for example the Ekavian and Ijekavian dialects of Serbo-Croatian, which are distinguished by the outcome of the Slavic vowel ě (yat), having either e or (i)je. This does not cause any serious problems in understanding for speakers today. CodeCat (talk) 21:16, 23 August 2014 (UTC)

The traditional terminology is simply misleading. Proto-Finnic is, by definition, the latest common ancestor of all Finnic languages, not just North Finnic. Therefore, we should go with Viitso and call "Middle Proto-Finnic" plain Proto-Finnic and "Late Proto-Finnic" Proto-North-Finnic. (Viitso writes Fennic, but that's immaterial.) Reconstructed proto-languages are strictly speaking proto-dialects, that's why they don't have any appreciable internal diversity (except "vertically", i. e., in registers), at least no more than the dialect of any small town or even mid-sized city, or some other close-knit community. Proto-Finnic was part of the Uralic dialect continuum, there is no need at all for it to have any internal diversity.

There's an alternate option as well: to interpret only Gulf Finnic / Core Finnic (not "North Finnic", which is usually taken as Finnish+Karelian+Ingrian+Veps) as "Baltic Finnic" sensu stricto, and demote Livonian and South Estonian to closely related outgroups. I think this approach actually has a better degree of backing in the literature. Back in the late 1800s Livonian, Samic and Core Finnic were treated as three equally coordinate branches within "Finnic" (= Finno-Samic; frequently the term "Finnic" was even extended to everything now called Finno-Permic). I've seen authors even in recent times also speak of Finnic arising by the "convergence" of two originally separate branches (though "parallel evolution of closely related dialects" would be more appropriate).

Now here is a further problem: the amount of explicit sources on Middle Proto-Finnic is just about nil — even "Early Proto-Finnic" i.e. Proto-Finno-Samic has better coverage — while Late Proto-Finnic is fairly well-described. So to not OR ourselves in the foot, for now I would favor the strategy of just sketching out the MPF era and then focusing on LPF. --Trɔpʏliʊm • blah 13:27, 25 August 2014 (UTC)

I don't agree with the conventional idea of proto-languages covering areas comparable to an entire country at all.

Clearly true, yes. The problem is people who treat a proto-language as an Ausbausprache or even a Dachsprache, instead of a single language variety at a limited point in space and time — and hence the period following the splitting of a proto-language would still involve people speaking "the" proto-language, just "dialects of it". --Trɔpʏliʊm • blah 13:27, 25 August 2014 (UTC)

For example, Latin wasn't spoken (natively at least) across all of Italy in the 2nd century BC, far from it; it was essentially limited to Rome (even the surrounding countryside, like Tibur and Praeneste, spoke slightly differing "Latin" dialects, where you had things like lōsna instead of lūna, and some irregularities like the prematurely monophthongised Romance cōda or the occasional failure of the littera/Iuppiter rule may well be ruralisms). When it actually was spoken across the whole peninsula from about 100 years later on, it began to develop internal diversity. Or think of the way Koine Greek was essentially based on the Attic dialect of Greek, despite some Ionic (and Boeotic?) influences and a few occasional Dorisms possibly: Attic was not spoken in a particularly large area originally, that's why it was so homogeneous at the beginning, and what did spread was probably originally based on the prestigious Athens dialect of the 5th century BC, not some rural form of Attic. So the problem "why do our reconstructs not have any dialectal variety and are so uniform?" is a non-problem: because they're themselves only dialects in a larger continuum, duh. It may be an artifact of the method, but that doesn't mean it doesn't square with historic reality: spreading languages start out fairly homogeneous, because they're prestigious dialects (even if standard languages don't exist in the prehistoric period, prestigious dialects do exist). --Florian Blaschke (talk) 01:23, 25 August 2014 (UTC)

But no language has no diversity whatsoever. Every speaker has their own way of speaking, which is always going to be slightly different from how others speak. However, in small communities, the language of each speaker influences that of the others, so that keeps them from drifting too far apart. When speakers have less contact with each other, their speeches can diverge further. That doesn't stop changes from propagating through the community, but it's no longer guaranteed that changes will always be evened out throughout. That's how dialects probably start off, and this is what led to the first differences in Proto-Finnic. I'm not going to conjecture too much about how speakers were located geographically, but the Inland Finnic speakers must have been slightly separated from the larger community, even if at first the separation was no more than a river or something like that. CodeCat (talk) 01:41, 25 August 2014 (UTC)

That's what I said. A language (as in "dialect continuum") has diversity, but a dialect not necessarily. There is diversity even in a local dialect of course, but on the vertical, social level, not on the regional level: in a small town, people keep talking to each other on a daily basis, so they don't drift apart from each other too much. When a prestige dialect spreads, it comes into contact with more or less closely related (and also, but less frequently, quite unrelated) dialects and languages, which is enough for regional dialectal diversity to develop. --Florian Blaschke (talk) 01:49, 25 August 2014 (UTC)

The article could at least be more clear from the outset that due to reasons that, if they are obscure even to me, are certainly puzzling to the non-Uralicist reader, this article does not cover the most recent common ancestor of all Finnic languages, but only of the North-Central group, excluding Livonian and South Estonian. (I'd say it's simply because Finnish and Estonian scholars are more interested in the common ancestor of Finnish and Standard Estonian alone, and less interested in the southern branches. Therefore, they prefer to include the changes that lead to both standard languages, even though they are missing in the southern branches. Proto-Finnic proper still has its strengths, didactically, and it would be wise to employ both.) This state of affairs would be easier to tolerate if the article showed clearly the differences between Proto-Finnic proper and Proto-North-Central-Finnic, so that the reader can easily derive Proto-Finnic proper from the data given here. --Florian Blaschke (talk) 01:48, 30 June 2020 (UTC)

As for the idea that Livonian and South Estonian are somehow no more closely related to North-Central Finnic than Sami is, I've never seen the slightest evidence for it, and Finnic languages blatantly contradicts it. The three-branch classification is simply outdated, as well as the whole idea of a Finno-Samic branch, and it is now accepted that Finnic and Sami form two separate branches of Uralic whose further relationship with each other and to the other seven branches accepted by every Uralicist (Mordvin, Mari, Permic, Hungarian, Mansi, Khanty and Samoyedic) remains unclear. --Florian Blaschke (talk) 01:59, 30 June 2020 (UTC)

separate discussion of PF from discussion of history?

The Later Developments section is getting a bit long really… Maybe it is time to summarize a bit more, and to fork the full details to a separate article History of the Finnic languages? --Trɔpʏliʊm • blah 11:47, 5 November 2014 (UTC)

I would be ok with that. CodeCat (talk) 16:32, 5 November 2014 (UTC)

Elaborate on the outcomes of lost *ŋ and *x, and also *w and *j?

@Tropylium: Right now the article is a bit vague on what happens with these sounds, but the changes are an important characteristic of Finnic, as they are what led to the creation of monosyllabic stems. So I would like it if the article could elaborate on it more, but I don't have much information available. Ideally there should be a table of each combination and the outcome. CodeCat (talk) 20:17, 15 November 2015 (UTC)

I made a small table with the reflexes I could find for a following *e. You can hover over the reflexes to see the words that I found that had it.

	-xe	-ŋe	-we	-je
a-	aa	?	awe	?
e-	?	öö	öö	?
i-	?	ii	iwe	?
o-	oo	oo	?	oi
u-	uu	uu, uwe	uu	ui
ä-	?	ää	äw	äj
ü-	?	üü, üwe	?	üü

CodeCat (talk) 21:40, 15 November 2015 (UTC)

There is no up-to-date research published on this topic: the only overview paper to have ever investigated this matter in detail is from 1949. That in mind, it might be premature to put much about it on Wikipedia yet.

I will probably publish an article revisiting this within the next 5 years, though. And I do have just about all the data collected by now. See wikt:User:Tropylium/Proto-Finnic/notes. --Trɔpʏliʊm • blah 15:40, 12 July 2017 (UTC)

Second-syllable contractions

On closer thought, these are probably better treated separately in the articles for the individual languages. Their development greatly diverges, and is conditioned on the very language-specific consonant losses — there is nothing especially Proto-Finnic about them. --Trɔpʏliʊm • blah 21:15, 15 November 2015 (UTC)

Time to implement a periodization update

Do we still need the obsolete periodizations from 1960s to 1980s in the article? A new scholarly consensus regarding a more shallow timeframe emerged in the 2010s, the most recent work in this area I know being Schalin, 2019 who uses different nomenclature (Coastal Finnic for the Late Proto-Finnic here) https://www.academia.edu/40362178/Scandinavian_Finnic_Language_Contact_and_Problems_of_Periodisation 5.178.188.143 (talk) 19:05, 14 February 2024 (UTC)