Is Sumerican a Uralic Language?

[BettyCross]

For 150 years, Sumerian has been cited as a classic example of a language isolate. Yet, in recent years, a Sumerian-Uralic hypothesis has been proposed. For the link, click here.

I'm not saying I agree with this and I'm not saying I disagree. I just want people's opinions.

Betty Cross

[Count Iblis]

I don't really know much about Sumerian but I'm pretty sure that it will never be found to be a Uralic language. Sumerian may be a sister to it though. Alan Bomhard has tentatively included it in the proposed Nostratic macro-family (which includes Uralic). A couple thinkgs make me suspicious of this article though. First, Proto-Uralic is not reconstructed with labio-velar stops. The brain word is *ojwa. The Hungarian word agy comes from *ajNi "temple" (it may be that *ojwa and *ajNi are somehow related though). Second, the author should be comparing Sumerian words to Proto-Uralic rather than Finnish and Hungarian. However, the author may wish to stick with comparing Sumerian to attested languages, which is perfectly reasonable, in which case he should be using more than just Finnish and Hungarian.

[WeepingElf]

Mushrooms. Weird ones. Too many of them. Smoking them.

Is there any language family of which nobody says that Sumerian was related to it?

[Bob Johnson]

I thought Sumerian was descended from Sanskrit, just like Basque.

[Acid Badger]

I thought Sumerian was descended from Sanskrit, just like Basque.

And Quechua!

[David McCann]

Yet, in recent years, a Sumerian-Uralic hypothesis has been proposed.

Piotr Michalowski wrote
"Over the years various unsuccessful attempts have been made to link it with ... Chinese, Tibetan, Hungarian, Turkish, and Indo-European. These attemps have sometimes been flavored with nationalalist fervor. More recently, some scholars have tried to include Sumerian within the hypothetical Nostratic ... while others have excluded it..."
"The phonology of the language is not well understood, and it is fair to say it will never be fully recovered."
"... it is extremely difficult to establish a reliable grammar or lexicon ... there is still much disagreement about basic grammatical facts..."

This just about sums it up! I can't see any elements in Sumerian that would suggest a Nostratic connection. The adjacent language families were Afro-Asiatic, Dravidian, Caucasian, and Kartvelian. Sumerian pronouns don't suggest any connections with any of them.

[zompist]

Is there any language family of which nobody says that Sumerian was related to it?

Technically, no, as Ruhlen at least has long claimed that all languages descend from Proto-World. (Kind of surprisingly, though, he throws it in the Unclassified bin.)

But my impression is that the long-range comparers and cranks alike seem to avoid Niger-Congo and Australian languages.

[Ser]

But my impression is that the long-range comparers and cranks alike seem to avoid Niger-Congo and Australian languages.

Joseph Greenberg worked on the former!

[Bedelato]

I don't really understand all the flak that the long-range comparers are getting.

Other than the shortage of material to work with, I see nothing inherently flawed with the idea of long-range comparison. In theory, you should be able to apply the comparative method indefinitely and end up at the beginning.

But, real life isn't theory. Random, sporadic mutations, which the comparative method can't deal with, happen all the time, and we can only recover so much of those. The rest are gone forever. Over time these build up and obscure relationships even further.

Because of this little bit of non-ideality, the comparative method has the chronological limits that it does, and going beyond 5000 years or so is a stretch. That's not to say we shouldn't try, though. Just don't get your hopes up.

[Cedh]

But, real life isn't theory. Random, sporadic mutations, which the comparative method can't deal with, happen all the time, and we can only recover so much of those. The rest are gone forever. Over time these build up and obscure relationships even further.

Also, the comparative method presupposes that proto-languages are largely homogeneous entities, failing to take into account the role of sociolinguistic and situational variation even within a single speaker properly. Much of this variation evens out statistically if the population is sufficiently large, but still, the development of dialectal divisions in reality seems to be best described not with the classic "tree model", but with a "bush model" with a lot of undergrowth. A large percentage of those "random, sporadic mutations" that Bedelato mentions are most easily explained with dialectal and/or sociolectal mixing.

[Bob Johnson]

Other than the shortage of material to work with, I see nothing inherently flawed with the idea of long-range comparison.

"Other than having only one data point, I see nothing inherently flawed with extrapolation..." The lack of data is the whole problem. People can't even agree on the exact details of PIE (prove what the laryngeals are, win a cookie), despite having decent data from... what, eight? more? languages, including huge corpora from Latin, Greek, and Sanskrit. Reconstructing less-well-known language families is an exercise in futility.

In theory, you should be able to apply the comparative method indefinitely and end up at the beginning.

Beginning of what? Only one beginning? Which beginning? At some point it's just more reasonable to stop and say that two families are unrelated.

[Xonen]

I actually took Parpola's Sumerian course a couple of years back, and he had a habit of suggesting Finnish cognates for some of the Sumerian words we encountered. At first I thought that would be interesting, but soon it became all to clear that he hadn't really even tried to establish systematic sound correspondences, not to mention the fact that quite a few of his "cognates" were semantically rather, uh, unfalsifiable, to put it nicely. For instance, according to him, Sumerian dur 'to bake bricks' "just might be" cognate with Finnish töhry 'smudge'. I think I was pretty much on the verge of tears at that particular moment… For all his expertise in assyriology, he just doesn't really seem to understand the comparative method, nor does he seem to have a problem with rewriting quite a lot of what we know about the history of the Uralic languages in order to make it fit his hypothesis. I'd take this with a truckload of salt.

However, at some point I met a student who had, inspired by Parpola's hypothesis, done some of his own research into the matter. He had (reasonably, IMO), gotten rid of Parpola's idea of Sumerian being a daughter of Proto-Uralic and instead looked at a possible sibling relationship, and he at least claimed to have found about two dozen phonetically systematic and semantically reasonable word pairs to support this idea. As he even admitted himself, though, that amount of data just isn't enough to prove anything conclusively; furthermore, I later realized that the guy's understanding of Proto-Uralic phonology was several decades out of date (specifically, I recall him having a couple of cognates where Sumerian /a/ corresponds to Finnish /A/, even where the latter actually descends from Proto-Uralic */M/).

So, to boil it down: There just might be a connection, maybe - but even if there is, I very much doubt we can find enough evidence to prove it.

[Count Iblis]

I don't really understand all the flak that the long-range comparers are getting.

Largely because of shoddy work by Ruhlen, Greenberg, and Starostin. It's tainted the field. Also there's the bizarre idea floating around that you can't demonstrate family relationships beyong 8000 years (sometimes this number is higher or lower). Oddly, this number is derived from lexicostatistics, a la Swadesh, which has long been discredited. The meme remains powerful though.

Other than the shortage of material to work with, I see nothing inherently flawed with the idea of long-range comparison. In theory, you should be able to apply the comparative method indefinitely and end up at the beginning.

The data isn't really all that scarce though. Part of the problem is that there is no standard by which we can say that a relationship has been demonstrated. It's easy to dismiss something as being unconvincing when the goalposts can be so easily moved.

But, real life isn't theory. Random, sporadic mutations, which the comparative method can't deal with, happen all the time, and we can only recover so much of those. The rest are gone forever. Over time these build up and obscure relationships even further.

Sometimes words are lost in all daughter languages which also creates a limit.

Because of this little bit of non-ideality, the comparative method has the chronological limits that it does, and going beyond 5000 years or so is a stretch. That's not to say we shouldn't try, though. Just don't get your hopes up.

The problem is that 5000 is an arbitrary number. For all we know the limit is 50,000 years.

[brandrinn]

Having actually read a little bit of Ruhlen and Greenberg, I think the problem people have with them is not the basic idea of long-term comparison, but rather the particular methodology they use. This includes selective sampling, using multiple languages in the same family as multiple points of evidence, and positing ludicrous sound changes without applying them universally.

[Åge Kruger]

The problem is that 5000 is an arbitrary number. For all we know the limit is 50,000 years.

I don't think it's all that arbitrary. Latin is well attested in written sources, and yet we can't reconstruct it perfectly or purely from it's descendants. We can get close to Vulgar Latin, but there is a certain amount of error in our reconstruction. If we then take several reconstructions, and compare them, then we're just multiplying the error. We can use this to make a good estimate of the margin of error over time and comparisons.

[merijn]

Is there any language family of which nobody says that Sumerian was related to it?

Technically, no, as Ruhlen at least has long claimed that all languages descend from Proto-World. (Kind of surprisingly, though, he throws it in the Unclassified bin.)

But my impression is that the long-range comparers and cranks alike seem to avoid Niger-Congo and Australian languages.

There is a crank whose name is not Glenn Campbell, but something similar whose speciality it is to claim that language X is a Niger-Congo language. I have also seen cranks who believe that Sumerian is a Bantu language because 1) they are both agglutinating and b) they think noun prefixes and determinants are the same thing.
There are also some non-cranks working on linking Niger-Congo with other language groups. Specifically, there is one scholar who argues that Niger-Congo is a subgroup of Nilo-Saharan. The African phyla themselves are pretty long-range, and I once read an off-hand remark of a historical linguist of I thought Asian languages who said that Africanists are generally lumpers compared to other historical linguists and the evidence for the African phyla would never convince most people outside Africanist circles.

[Niedokonany]

Isn't Sumerican what those Mormon lost tribes used to speak?

[Miekko]

The problem is that 5000 is an arbitrary number. For all we know the limit is 50,000 years.

I don't think it's all that arbitrary. Latin is well attested in written sources, and yet we can't reconstruct it perfectly or purely from it's descendants. We can get close to Vulgar Latin, but there is a certain amount of error in our reconstruction. If we then take several reconstructions, and compare them, then we're just multiplying the error. We can use this to make a good estimate of the margin of error over time and comparisons.

And it's likely the error grows exponentially, right?

[Grigor]

Is there any language family of which nobody says that Sumerian was related to it?

Technically, no, as Ruhlen at least has long claimed that all languages descend from Proto-World. (Kind of surprisingly, though, he throws it in the Unclassified bin.)

But my impression is that the long-range comparers and cranks alike seem to avoid Niger-Congo and Australian languages.

There is a crank whose name is not Glenn Campbell, but something similar whose speciality it is to claim that language X is a Niger-Congo language. I have also seen cranks who believe that Sumerian is a Bantu language because 1) they are both agglutinating and b) they think noun prefixes and determinants are the same thing.
There are also some non-cranks working on linking Niger-Congo with other language groups. Specifically, there is one scholar who argues that Niger-Congo is a subgroup of Nilo-Saharan. The African phyla themselves are pretty long-range, and I once read an off-hand remark of a historical linguist of I thought Asian languages who said that Africanists are generally lumpers compared to other historical linguists and the evidence for the African phyla would never convince most people outside Africanist circles.

Indeed. Niger-Congo, Nilo-Saharan and certainly Khoi-San are probably not valid language families because their relationship has never been demonstrated to a satisfying degree. [citation needed]

[zompist]

Indeed. Niger-Congo, Nilo-Saharan and certainly Khoi-San are probably not valid language families because their relationship has never been demonstrated to a satisfying degree. [citation needed]

You can cite R.M.W. Dixon, who went back to Greenberg's African work after his Amerind proposal got roundly criticized, and found it of the same (low) quality.

[Salmoneus]

On the other hand, it mostly seems to be accepted by the experts. Not Khoisan, admittedly. But my understanding was that Niger-Congo was well-established, albeit vague (by which I mean that the exact internal structure and external limits are not well-defined, but that the core is generally agreed on).

I think the splitters are overly critical of the lumpers, because they assume the lumpers are just splitters who aren't good at being splitters. Whereas it seems more likely that they're actually trying to do something else entirely.

[Count Iblis]

The problem is that 5000 is an arbitrary number. For all we know the limit is 50,000 years.

I don't think it's all that arbitrary. Latin is well attested in written sources, and yet we can't reconstruct it perfectly or purely from it's descendants. We can get close to Vulgar Latin, but there is a certain amount of error in our reconstruction. If we then take several reconstructions, and compare them, then we're just multiplying the error. We can use this to make a good estimate of the margin of error over time and comparisons.

How did you decide that 5000 was a better estimate than, say, 10,000?

[Tropylium⁺]

There are also some non-cranks working on linking Niger-Congo with other language groups. Specifically, there is one scholar who argues that Niger-Congo is a subgroup of Nilo-Saharan.

That'd be Roger Blench, and what I get from it is the impression that the standard of comparision used to demonstrate NC and NS in the first place can be applied to demonstrate them being related to each other just as well.

(NC and NS at their widest, that is… that stuff like Kru, Kwa, Benue-Congo etc. belong together is fully accepted IIUC.)

The problem is that 5000 is an arbitrary number. For all we know the limit is 50,000 years.

I don't think it's all that arbitrary. Latin is well attested in written sources, and yet we can't reconstruct it perfectly or purely from it's descendants. We can get close to Vulgar Latin, but there is a certain amount of error in our reconstruction. If we then take several reconstructions, and compare them, then we're just multiplying the error. We can use this to make a good estimate of the margin of error over time and comparisons.

And it's likely the error grows exponentially, right?

That would be assuming that all evidence of relationship is equally likely to be lost or muddled, which is not true. *m is more likely to stay put than *o is, a noun class system is more likely to stay put than a set of locative cases is, etc. A relationship 10K years old is going to be harder to recover than one 5K years old, but not squared-as-hard.

[zompist]

*m is more likely to stay put than *o is, [...]

Is it? So does the incidence of m increase over time? If not, why are you confident that m is so stable?

[Tropylium⁺]

*m is more likely to stay put than *o is, [...]

Is it? So does the incidence of m increase over time?

No, because sound changes that create /o/, or most other "more volatile" segments, are also common.

If not, why are you confident that m is so stable?

List me three languages or language groups that have a sound change affecting original word-initial *m in any way? Or original intervocalic *m, for that matter.

Next try listing a few languages that have a sound change affecting original initial-syllable *o?

(Or if you really need the point driven home, languages that have a sound change affecting word-initial *k — in any way, ie. palatalization, spirantization, uvularization, etc.)