A little recap of history: Freelance Reconstruction, the blog you’re currently reading, [1] was originally started as a Tumblr microblog. It turned out though that my blogging style needs a sturdier framework, and for several years now, I’ve been happy to be based on WordPress instead.
This much some old readers may recall. However I never have gotten much into doing quick-paced community engagement blogging on here, in part indeed due to the heavier-duty software. And since I still hang out on Tumblr for unrelated reasons, I’ve also found it useful to have an outlet to comment on things related to linguistics that come up in there.
Thus, enter a new, more casual linguistics sideblog: possessivesuffix.tumblr.com. This has been running for a bit over a year by now, but I don’t think I’ve mentioned anything about it earlier on here. Perhaps I should also request that anon asks be redirected there instead of the old defunct version of this blog?
Here is also a list of some posts on there that might be of interest to the readers on here as well.
1. Original blog posts and commentary on topics:
— on the structure and history of Finnish:
- On the rendering of ts in (my) colloquial Finnish
- Sociolinguistic archeology of gendered terminology in Finnish
- the Finnish final gemination and its levelling-out
- Finnic etc. loans in Germanic (and whether path might be one)
- varieties of kaali
- Some Finnish minimal multiplets
- Suomen monikon genetiivin historiaa
- the ending -elO in Finnish
- phonemic syllable breaks in Finnish
— on Uralic linguistics in general:
— on phonological fun facts and typology:
- Spirantization in Iranian, part 1 • part 2
- Classical Armenian r-metathesis (+ follow-up)
- interdentalization in Central and Eastern Asia
- weird consonant cluster patterns, part 1 • part 2
- does root structure link to consonant inventory sizes
- on the Swedish sj-sound (“/ɧ/”)
- typology of sibilant systems across Eurasia
— other stuff:
- contra Atkinson & co. 2011 (the “phonological diversity decreases by distance from Africa” study)
- cognitive differences in the degree of generativeness of grammar
- ‘moon’, ‘horse’ and lexical replacement in Indo-European
- linguistic universals as dependant on linguistic theory
- Typical Endangered Language Fallacy (+ follow-up)
- Maybe some isolates are creoles?
- some opinions on Altaic
- language homelands outside of the center of gravity of diversity
2. Links to other blogs, articles etc. without much additional insights of my own:
- phonotypology of Khoisan languages
- Iranian Dialectology and Dialectometry (PhD thesis)
- Reflexes of PIE ablaut in modern Greek
- Two thousand years of history that were completely forgotten
- ＡＳＹＮＣＨＲＯＮＩＳＭ ＯＦ ＰＲＩＭＡＬ ＦＵＳＩＯＮＳ
- Initial *gʷʰ in Germanic
- linguistic diversity in eastern Tibet
- The Sino-Tibetan Languages (Routledge)
- Pronoun borrowing in colloquial Vietnamese
[1] I’ve seen this blog occasionally linked under the name “Protouralic”, but to be exact, that is only my blog’s URL, not the title. The discrepancy is mainly since I can foresee maintaining this blog long enough that I will no longer be doing freelance reconstruction… It remains to be seen what the blog will be renamed at that point, though.
Freelance reconstruction 
Internal reconstruction of Burushaski phonology: very interesting, but:
1) I’m surprised stress isn’t indicated in the transcriptions. Is it omitted in the database? Berger’s work ( > Wikipedia) makes clear that stress is phonemic (or at least unpredictable) and always spells it out. I’d expect stress to have a major impact on aspiration, and perhaps on voice, too.
2) “Pickax” is /tɕak/? That’s too good to be true. :-) It’s even better than Polish kilof!
The StarLing DB does indicate stress; just omitted by me as it doesn’t appear to be relevant for any of this. E.g. the words with medial aspirates, with ND/N correspondences, or with /ld/ are all about evenly split between first and second syllable stress.
Science as “truth-seeking procedure”: I find it better to think of science as the quest to identify all falsehood. There’s only a vague hope that at some point the truth (or, rather, reality*) may be the only thing left standing.
That’s why only creationists and historical linguists use words like “proof” and “prove” anymore when talking about science. Sure, most of the time that’s just a ritual, when we write “strongly suggests” we really mean “proves beyond reasonable doubt” and when we write “shows” or “demonstrates” we often mean “proves beyond any doubt we can think of”, but at least it’s a reminder.
* I like to distinguish truth from reality in epistemology. Reality is that in which the argumentum ad lapidem is not a logical fallacy; it is what science is about. Truth could be the same as reality (physicalism), or the truth could be that reality is some kind of illusion that is inescapable for Puny Humans, e.g. I could be the solipsist or God could be the solipsist or whatever.
Verbing personal names: merkeln “do nothing and say nothing unambiguous when you’re supposed to be making an important decision”. I think it never got very far, though, and may already be dying out again.
Just came across scholzen: “communicating good intentions and then finding, using or inventing every possible reason to delay and/or prevent them” according to this blog post. No clue how widespread it is (or will become).
fuck’s sake, are we no longer allowed to write our own HTML in WordPress
Looks like we at least have some buttons for basic formatting now (and, fixed your previous comment). Was there a more worrisome phase there without either for commenters?
Thanks! I’m pretty sure the buttons were there, I just didn’t think we had to use them.
Natural languages: “Why noun when you can verb?”
Bureaucratic German: “Why verbiage in the face of the possibility of nouns?”
Indo-Uralic: I was very excited when I first read about the Anatolian possessive suffixes, but… they’re clitics in the sense that they’re declined separately. For example, Hittite nom. sg. atta-s=ma-s “my father”.
(Spelled at-ta-aš-ma-aš whenever it isn’t spelled a-bu-ya.)
Not counting reduplication, prefixes are generally recent developments in IE, except apparently *n- “un-“. The verb prefixes are still rather loosely bound “preverbs” in Sanskrit and even Gothic.
The works on IU I find most exciting are in Russian (and they’re slides of a conference presentation and a list published as a paper in a Gedenkschrift), except for this conference handout with two examples. Put together, they dramatically increase the number of comparanda as far as I know.
Heh, I had dramatically underestimated how far the “preverbs” had progressed toward Germanic-/Slavic-style prefixes in Hittite!
The same post reminds me that the “Indo-Iranian as the trigger of Uralic expansion” paper proposed that the “name” word was a cultural loan from IE into Uralic because of the IE cultural association of name and fame – not so much “what is your name” as “what glorious epithet has been bestowed upon you”. Importantly, this kind of explanation remains unavailable for things like “water”.
Either way it does not help that the “name” word is a horrible mess to reconstruct on the IE side. How many laryngeals and which ones? Which accent & ablaut grades? Is it a *n-stem or a *men-stem, the latter meaning the root is actually *HnVH- and not some kind of Uralic-looking *(H)nV(H)m- at all? Or is it one of these and later got reanalyzed as the other? Here’s a fifty-page treatment (most of it in the footnotes, in Italian, from 2005), and if you don’t like its conclusions, just pick another, because there are others of comparable length out there.
Anyway, to explain the Uralic form with its *i as a loan, we need to squeeze *ē into either Pre-Indo-Iranian or Pre-Tocharian; that’s doable but decidedly nontrivial.
Thurneysen, Kortlandt & friends: I agree the Latin ‹ngn› spellings aren’t decisive, but while ‹gn› was possible root-initially, word-initially it seems to be limited to the praenomen Gnaeus. Compare cognatus with natus, where *gn- is etymological.
Pommes de terre are shortened to pommes in all compounds, e.g. pommes duchesse or pommes noisette. In other words, apples and potatoes are treated as accidental homophones, with de terre only added for disambiguation or nearly so. It is admittedly unusual for French homophones to be homographs.
*headdesk* *headdesk* *headdesk*
Pommes frites. Imported as such, or as Pommes, into German even.
Conservative languages: max1461’s description of stabilizing selection is accurate (source: am evolutionary biologist). So is esoanem’s description of selective pressures that are known to act on languages.
I agree that “is conservative” is always a gross oversimplification. Loss of word-final material (or, I guess, word-initial in prefixing languages) seems to be caused by stress-timing. What the vertical gene transfer causes languages to become stress- or mora-timed (not sure if syllable-timed really exists) is beyond me. In many cases substrate effects are obvious, but in many others it’s obvious that no such effects have existed: European Portuguese is becoming stress-timed in front of our ears, and my colleague Fernandes is already [fr̩ˈnãd̥ʑ̥]; Russian is stress-timed, Ukrainian is not, neither is Polish, and neither is anything Russian has been in contact with…
Troll hypothesis: French is stress-timed. The stresses are just so far apart it’s hard to notice.
Argh, and, could Estonian nael be reborrowed from a Low German version of Nagel, with [ɣ] or even [ɰ]?
The short V1 and the a-stem (naela-) don’t look like good fits though, and *akR > *aiR > aeR is regular enough (also e.g. kael ‘neck’, naera- ‘to laugh’, naeris ‘turnip’). A LG re-loan could be well the case for an occasional dialect variant naal however…
Thanks! Definitely silly of me to forget about vowel length (in Estonian of all).
Phonotypology:
I’m quite sure that French /gʁ/ actually comes out as [gʁ], not as [ɢʁ].
That may be typologically normal. Tyrolean, like Swiss German, has turned /x/ into [χ] unconditionally, and /k͡x/ into [g̊͡χ]: it, too, changes places of articulation halfway through.
Other such mobile affricates are the good old Upper German /p͡f/, which really does consist of a bilabial stop that’s released into a labiodental fricative (unlike the all-bilabial or all-labiodental affricates found in a few other languages), and one of the two realizations of the Polish trz, which is a laminal alveolar [t] released into a postalveolar [ʃ], famously contrasting with cz which is postalveolar throughout: [t͡ʃ] vs. [t̠͡ʃ]. If you pronounce this version of trz too slowly, you get [t͡sʃ].
(…The other version of trz differs from cz only in length – but, remarkably, in the length of the fricative, not of the stop.)
Theoretically, BTW, Polish has the same contrast for drz vs. dż, but dż is such a rare loanword phoneme that it usually comes out as drz even though that’s the one that’s harder to articulate.
Just as a curiosity, the superficial similarity of “mother” and “cow” in Dahalo reminds me of all those Turkic languages where “horse” is /at/ and “father” is /ata/…
Alas, it is German where “blood” and “flower” are related, not Nahuatl.
“Blossom” actually; “flower” looks like some sort of root cognate, but not with a remotely productive suffix.
Map of northern Africa 8 ka ago: That may be the biggest version of Lake Megachad I’ve ever seen… insert desperate attempt at Gigachad joke here! I had no idea of the other lake that was even bigger, though. And the lack of a Nile delta is intriguing, though it could well be correct.
Crossposted from “the platform known as Twitter”: wow, that workshop looks fascinating.
Makes sense to me a priori. “Year” = “period between two Norwegian visits”? Late July to late July?
One issue here are different definitions of “Proto-Germanic”: “the last common ancestor of all attested Germanic languages” is younger and probably still less often used than “the first stage of the Germanic branch that was ‘recognizably Germanic'”, sometimes explicitly defined by Grimm’s law. These two definitions may be 500 years apart.
Definitely pre-PG by the stricter definition, because it was followed by the creation of the PG *ā from *aja > *ā and *azr > *ār (both of them general sound laws that also worked on other vowels).
As you’re well aware, Finnic also has loans that lack *en. > *in.; even though Tacitus recorded Fenni without this change, and even though Gothic has at first glance erased the evidence for whether it had undergone it, Gothic does have ih for *enx > *inx > PG *ĩːx > Gothic *iːx, so it does look like this was not even the very last change before PG (though close).
The loans into Finnic and Samic, and maybe the few out of Finnic that have been identified, seem to be the strongest evidence that PG was spoken in or around southern Scandinavia. This is, however, followed closely by the archeological and genetic continuity there since the Corded Ware culture was established, and by the evidence for linguistic contact with Baltic. There’s an intriguing argument that the well-known Celtic superstrate was actually imported across the sea by a seafaring elite that traded up & down the Atlantic coast and into the Mediterranean, not by terrestrial contact in central Germany (which happened around Roman times). Kroonen has also pointed out that the grammatical ramifications of Kluge’s law ended up looking a lot like consonant gradation, so the linguistic contact may not be limited to words. Baltic definitely shares not only words with Germanic, but also grammatical features (marking of definiteness on adjectives, by quite different means than Germanic, comes to mind).
There’s also an absence of evidence that PG was spoken anywhere else. There’s an onomastician who says all hydronyms in Thuringia-or-so are Germanic and therefore puts the Urheimat there, but he’s not a historical linguist, and apparently it shows… other than the big isles of Denmark, there don’t seem to be any non-Germanic place names in or around southern Scandinavia.
Concerning the other definition of PG, it’s too bad that both Grimm’s and Verner’s law are (as you’ve pointed out previously) invisible in Finnic; the change of initial *x to [h] may be visible, but that’s most likely a post-PG change that may have spread from East Germanic during and even after Roman times (Frankish *x- words in early-medieval Latin sources all seem to be spelled with ch-).
You’ve previously noted loanwords in Finnic that reflect Germanic *w *j as *p *ć, and I’ve mentioned a paper that argues Gothic had turned *w *j into [v ʝ] (plus the well-known outcome of *jj as ddj), so I speculated this Germanic variety was real and belonged to East Germanic… so we’d finally have stronger evidence for EG in Scandinavia than that one lance tip with tilarids (**-az) written on it.
Au contraire. Sievers’s law may still have been active in PG, but it was definitely a thing of the past by PNWG times. It’s not just the famous katsoa; WG is full of words that have undergone WG consonant stretching but couldn’t possibly have if Sievers had still been working. Contrary to Ringe’s book, this is not limited to a handful of extremely puzzling exceptions, it’s all over the place: heizen “heat” < PWG *haittja- < PNWG *hajtja-, **hajtija-; Weizen “wheat” > *hwaittja- < *hwajtja, **hwajtija-; MHG wülpe “she-wolf” < *wulbbja < *wulβjō, **wulβijō; Schnauze “snout”, schnäuzen “clean one’s nose” < *snauttj- > **snawtj-, **snawtij-… see from here on for niesen “sneeze”, which has a long /sː/ in my dialect…
Fascinating. If the evidence isn’t linguistic, what is it?
As you’re well aware, Finnic also has loans that lack *en. > *in.
Yes, though I no longer even count that as a PG marker… I keep meaning to blog at some point about recent results that show maintenance of this contrast even into Proto-Norse, which seems to me like important Germanistic news, but which does not seem to have gone around widely so far.
evidence for EG in Scandinavia
Alas loans in Samic show still instead the Proto-Norse skærping *jj > *ggj; and then on top of that the second round to ? *ggɟ, to yield Samic *kć.
If the evidence isn’t linguistic, what is it?
Ethnographic, with distinctively un-Nenets practices like little sign of reindeer hunting (and none of herding, though that’s thought to have come later anyway); living in permanent peat-covered dwellings dug into the earth, instead of any kind of a teepee analogue; unfamiliarity with salt even as a trade item. The apparent Proto-Nenets–Enets term for ‘salt’ could be a ‘firewater’ type calque though, it’s the same word as ‘ice’, ‘white’.
Interestingly it’s also possible that linguistic evidence might still exist: the expedition mentions even capturing a few “Zemblians” and bringing them back to Copenhagen, and maybe checking the Danish Royal Archives about that could turn up something more.
(All this does not match up though with how the recorded name of one locale these people were met at has a good Nenets etymology: Boranday, at the lower Pechora = ? modern War°nday° (Russian Варандей) ‘utmost, bordermost’.)
You’ve previously mentioned Johan Schalin’s work (which I’ve meanwhile read) showing that the merger of *e into *i happened before P(NW)G only outside the stressed foot, not everywhere outside the stressed syllable as everyone had believed. This is about a different sound change, the merger of *e into *i in the stressed syllable if a nasal follows in the same syllable. Is there evidence this hadn’t happened in Proto-Norse, too?
Wow.
Amazing.
I found Aikio’s presentation. Fascinating stuff in there:
– Two loans that lack *en. > *in., so should be older than PGmc;
– two that lack *ē > *ā, so should not be NWGmc;
– of the ones that lack Sievers’s law, one has an *ā; I don’t know what the source of that *ā is, but if it’s *ē, that means the variety that modified postconsonantal *j was not East Germanic.
I don’t think borrowing *[ŋx] > PGmc *[ː̃x] as *ŋ is necessarily an archaism; it could simply be a way to deal with nasal vowels. *ð as *t is fascinating, though, if there’s no Saami-internal explanation for it (e.g. was *rð allowed?).
Even more fascinating is *[ŋʷgʷj] borrowed as *mć.
Also, I forgot to mention that Gutnish is not NWGmc, but its sister-group: it lacks *u…a > *o…a (but has meanwhile acquired a short /o/ from other sources).
Austria and the east: the Germanic root is very much *awst-, preserved unchanged in West Norse and with the regular outcomes of *aw everywhere else. Of course it’s cognate with the Latin one and a reflex of the PIE “dawn” root. The trick is the OHG monophthongization which turned *au into [ɔː] when followed by alveolars or /x/ (or something), followed much later by Standard German being so inconsistent in shortening long vowels in overlong syllables that the spelling hasn’t kept up: Osten, Ost-, östlich have short vowels nowadays unless used on the North Sea (the Low German and Dutch reflex of *au is [oː], unconditioned), but Ostern “easter” (same root) and Österreich retain their long ones, as does Kloster “monastery” < Latin claustrum.
Meanwhile, T-shirts with the Australian traffic sign for “animals crossing” and the text “NO KANGAROOS IN AUSTRIA” have been sold in Austria for over 30 years. Google it. ^_^
…and as usual I forgot the arguably most important part: Austria is a calque of, well, ôstarrîhhi that mistook the Germanic meaning for the Latin meaning of the root.
Might fall under a corner-case type of what’s already called “phonosemantic matching”, or as what I call “etymological misnativization”.
And yes in retrospect of course *au, as also seen e.g. from how English would have **eest and not east if it were *i-umlaut from *ō. Swedish öst or, for that matter, Danish øst are though as nondiagnostic as German looks off the cuff. Unless it should’ve turned into **uost > **ust if it was already PG *ō? but I really cannot keep track of at what relative chronology precisely vowels shorten in German, when I don’t think I’ve even learned consistently the extant not-graphically-obvious exceptions (I would’ve been happy to go with a short [O]stern or [Ö]sterreich).
Oh yes!
Yes.
After MHG was over and the NHG mono- and diphthongizations had run their courses. Standard Licht has a short vowel, following monophthongization of the diphthong that’s retained in Upper German and written in Liechtenstein.
Husten “cough” comes to mind, another case with st, and another that retains its diphthong in Upper dialects. Conversely, northern pusten “blow”, which probably never had a diphthong, and duster ~ düster “gloomy” (southern finster…). Then there’s Mond “moon”, which got an excrescent enhancement of the word boundary (like the short-voweled Saft “juice, sap”, Axt “ax” and I think one or two others). Rache “revenge” can be heard long and short, and the short version should probably be considered a spelling-pronunciation. Oh, and hoch “high”, where the irregularity is not the long vowel but the preserved short /x/ – the endingless form must be borrowed from an Upper dialect while the others are all regular (hohe, hoher, hohes, hohem, hohen, höher(-)) – but Hochzeit “wedding” somehow got shortened (while the superlative höchst(-) did not).
That may be the complete list, apart from placenames probably.
Kuchen “cake”, unlike Küche “kitchen”.
Apparently Hüfte “hip”; the -e, and likely the feminine gender with it, is evidently backformed from the -en of the plural.
A slightly longer discussion of not-graphically-obvious long vowels in German starts here.