Bonus Material 2017

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: 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 Uralic linguistics in general:

— on phonological fun facts and typology:

— other stuff:

2. Links to other blogs, articles etc. without much additional insights of my own:

[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.

Tagged with: , , ,
Posted in Links, Meta
61 comments on “Bonus Material 2017
  1. j. says:

    Additionally, people who read comments on blogs might be interested in checking out an ongoing mini-debate on linguistic classification that I also participate in, over at the Languagehat post Origins of the Japanese language.

  2. David Marjanović says:

    Hey! It’s 1 am, and I stayed up late yesterday!

  3. David Marjanović says:

    phonemic syllable breaks in Finnish – I find nothing mysterious about vowel clusters, including sequences of identical vowels that don’t contract to a long vowel. All over (at least) Central and South Bavarian, /aˈɛi/, /iˈɛa/ and /aˈɛia/ – each with as many syllables as vowels, none with inserted glottal stops or anything – are grammatical near-sentences.

    sisaruksetGeschwister! “Siblings” (rarely also as a backformed singular) formed as a collective from “sister”. Gebrüder is obsolete and only ever meant specifically “brothers” AFAICT.

    flap = one-contact trill – no! The American flap sounds rather different from the one-contact trills scattered over the world, even the one used for intervocalic /r/ in early-20th-century RP. It isn’t articulated the same way either.

    Pronoun borrowing in colloquial Vietnamese – also Thai (me, you) for the same reason.

  4. j. says:

    flap = one-contact trill – no! The American flap sounds rather different from the one-contact trills scattered over the world,

    That’s the point: the American English allophone of /t/ is not a flap, but rather a tap (if we define “flap” as a single-contant trill vs. “tap” as an overshort plosive — apparently confusion on which would be which is common).

    • David Marjanović says:

      I’d say these are three different things, not two; in the American /t d/ allophone, the tongue flaps past the alveolar ridge instead of staying there and building pressure. Admittedly, I don’t know if overshort plosives really exist.

  5. David Marjanović says:

    Forgot yesterday: deriving the Proto-Iranian prevocalic *f *θ *x directly from *pH *tH *kH by the usual preconsonantal fricativization is phonetically implausible, no matter whether *H was something like [x] or something like [h]. Rather, voiceless aspirates becoming fricatives may be an argument for dating the loss of voiced aspiration earlier than the fricativization of the voiceless aspirates.

  6. Crom daba says:

    Another problem is that Saka doesn’t fricativize *TH clusters, instead it has voiceless aspirates, Balochi and Wakhi also have voiceless stops in this position, although it might be later spirant hardening in the case of Balochi (and Wakhi is possibly related to Saka).

    • David Marjanović says:

      How were the fricatives written in Saka?

      • j. says:

        As Brahmi script aspirates, as far as I know (also, /tsʰ tʂʰ/ as ‹ts kṣ›).

        Kümmel has a proposal that, before the general loss of voiced aspirates, there was an “aspiration throwback” change where *T-NDʰ > *Tʰ-ND. This at least would require the fricativization of *Tʰ, which would be surely also the best routing for *TH.

  7. Crom daba says:

    On further review of Bailey’s dictionary, it seems unlikely that Saka aspirates are retentions as Lubomir Novak states. I’m not sure how to explain words with /ch/ though, maybe Indic loans or secondary aspiration.

  8. David Marjanović says:

    I forgot my favorite example of a phonemic syllable break! It’s a minimal pair in most Standard German accents: Karl /ˈkaːl̩/, kahl /kaːl/.

    Unfortunately Charles the Bald doesn’t exhibit it, due to adjective declension: Karl der Kahle.

  9. David Marjanović says:

    kaali – talking about German opens a can of Pieris caterpillars! There’s a lot of regional variation. In Austria, Kohl is specifically Savoy cabbage, which is called Wirsing in northern Germany, plus lots of other names elsewhere. Brussels sprouts are Kohlsprossen (pl.) or sometimes Sprossenkohl (sg.), cauliflower is Karfiol (final stress), and kale remains largely unknown. White and red cabbage end in -kraut rather than -kohl, same pattern as for Sauerkraut (Sauerkohl far enough north).

    Also, sour cherries are Sauerkirschen down north, but Weichseln in Austria.

  10. David Marjanović says:

    Preadaptation post: well, that’s exactly why we biologists have followed S. J. Gould’s advice and switched from saying “A was a preadaptation for B” to “B is an exaptation of A” – the use of A for a new function to which it wasn’t originally an adaptation.

    The ghits for exaptation number “about 177,000”.

  11. David Marjanović says:

    Post on random disappearance of rare phonemes & clusters: rarity can apparently trigger sound change. Words beginning with */dw/ are rare throughout Germanic. In Old and Middle High German that regularly became /tw/, but more and more of the few words were lost. Finally, almost all of the remainder were moved over to the most similar more common cluster, /tsv/, while the rest joined /kv/. There’s an etymological doublet: quer “transverse”, Zwerchfell “diaphragm” (in anatomy).

    I’m sure the northern word for cottage cheese, Quark, is from (something like) Polish twaróg [ˈtvaruk] (same meaning).

  12. David Marjanović says:

    Post on rock-paper-scissors (I should start adding links): in German it’s Schere, Stein, Papier, probably to keep a trochaic rhythm. – Compare further: rock-paper-scissors-lizard-Spock.

    Post on voiced and voiceless plosives switching places: “the comparatively rare voicinɡ of ɡlottalised consonants, apparently attested by the Nakh languages” – in the Caucasus area, ejectives are lenes with barely audible glottalization, so it’s not surprising that they sometimes become voiced pulmonic plosives in voiced environments.

    Post on Mitian: I’ll have to look up what the EDAL proposes about Turkic *b-, but it’s interesting in this context that the Inuit-Yupik 1sg pronoun has *v- (no idea about Aleutian). The 2sg *s- is in any case unexpected, but then most of Greek has sy which is irregular enough that we can only wave our hands about dû à la fréquence or something. – Good to know that shifts of meaning between “spirit, soul” and “brain” are attested, and not everyone has been looking in the heart or the spleen or somewhere!

    • David Marjanović says:

      Silly me. The EDAL says nothing about where Altaic comes from, so it simply takes the 1sg *b- for granted and doesn’t try to explain it. PA *m- > PT *b- is considered regular, with a pretty long list of examples. Within Turkic, we get (p. 145): “In most languages (except Tur. and Gag.) > m- before a following nasal, with slightly differing rules. The same is true for (*-p-) > *-b- > *-m- in the second syllable.”

  13. David Marjanović says:

    Also Mitian post:

    Apparently there are a few Spanish words that are, in some dialects, undergoing semantic drift toward their English cognates…

    For rare words that happens all the time. German realisieren has within living memory picked up the meaning “getting it” from English.

  14. David Marjanović says:

    In SUSA 96, the link to “New evidence on Para-Mongolic numerals” is broken. I figured out what’s wrong: this link works.

    • David Marjanović says:

      And a fascinating paper it is! The labels “Manchu” and “Jurchen” were hiding three languages instead of one!

      From footnote 2:

      […] kǝrdǝm is a Mongolic loanword ultimately of Turkic origin. Among modern Turkic languages only Khalaj här ‘man’ shows that there must have been an initial h- which may have a connection to Alchuka k-. Problematically, the consonant does not appear in Mongolic and the initial h- in Turkic is thought to go back to an older *p- (Doerfer 1985: 99, 1998: 280f.).

      I wonder if this word went in the other direction and was spread by the Avars.

      • j. says:

        I could suggest that the Turkic word is perhaps in turn borrowed from a preform of Hungarian férj ‘man’. The usual explanation is that this is an entirely native compound of two unproductive terms, fi- (beside fiú ‘boy’ < *pojə-ka) plus erj ‘man’.

        The latter component though is poorly attested elsewhere in Uralic (a clear cognate in Mari plus an unclear one in Finnish), and borrowing from Turkic has been suggested too, which seems like the better way to resolve this loop.

        • David Marjanović says:

          So, the word made a Turkic-Hungarian-Turkic roundtrip just to pick up a h-? :-) Why not, there are English words of ultimately Frankish origin that did similar things.

          I happened to find Khalaj här on p. 27 of the EDAL:

          “However, there is a significant number of cases where Khalaj has initial h- which appears to be an innovation (prothesis), cf.: Khal. hil- ‘to die’ < PT *öl < PA *oli; Khal. här ‘man’ < PT *ēr < PA *ā́ri;” and 26 further examples. “One may note that this prothetic h- is very frequent before long vowels and before the following -j-, -v-. However, the rules are not strict, and in general the emergence of h- in Khalaj is unpredictable. Absence of h- in Khalaj is therefore an almost certain sign of *0- (or *ŋ-) in Altaic, but its presence may be original or secondary. We shall thus continue to use PT forms without initial *h- (keeping in mind though that it was probably present in the system) – given that the reconstruction of *h- can be made only on Khalaj data, and the latter is often quite ambiguous.”

          Incidentally, hil- “to die” is also found, in exactly identical form, in Basque – an eerie reminder that coincidences happen.

          • j. says:

            No, I mean, it’s not possible for Turkic *(h)er to be both a loan from Hu. férj, and for that to originate as a compound where the latter part is a loan from the same Turkic word, so at least one of these etymologies has to be off. A supposed Uralic etymon that only has reliable reflexes in Mari and Hungarian smells very much like something that came from Turkic, too.

            The alleged Finnish cognate is a rare dialect term yrkä ‘young man’, but /y/ doesn’t match (Mari /erɣə/ and the Hu. point to *är(k)V) and this seems more likely to be a local variant of ylkä ‘groom’.

  15. David Marjanović says:

    The link to the paper on vatsa and vatsá- is also broken; fortunately, the correct URL is obvious.

    In that interesting paper, BTW, Parpola should have stayed away from the technobabble. “With 19 radiocarbon dates, the duration of the Sejma-Turbino network was fixed to 2150–1600 BCE (with sigma 1 calibration)” – reverse hull polarity and set the shields to an oscillating frequency! Rather, σ is the standard deviation, and all dates from 2150 to 1600 BCE are within 1 σ of the mean estimate. In other words, the probability that the true date lies outside that range is close to a third.

    (Except it’s not anymore, because now – as the rest of the sentence makes clear – there are 38 dates, so that the range – presumably 1 σ again – has narrowed down to “c. 2200–1900 BCE”.)

  16. David Marjanović says:

    Post on West Germanic diphthongizations:

    My suspicion — just a hunch so far — is that what is going on here is that West Germanic vowel phonology is organized differently from the other languages in Europe, and it is this that drives all these high close diphthongizations. Something along the lines of closing diphthongs being single phonemes sharing the feature [+high] with close vowels, maybe, instead of closing diphthongs being absent (Slavic) or remaining analyzable as vowel + glide (NGmc, Romance, Baltic, Hungarian, Samic, Finnic, Iranian…)? This would mean that something like /əi əu/ have for long been “conceptually possible” in West Germanic phonology, and therefore available as something that /iː uː/ could change into without too much disturbance.

    Specifically, phonologists and historical linguists tend to treat closing diphthongs and closed syllables in -Vj, -Vw as the same thing, and likewise for opening dipthongs and (C)jV-, (C)wV-. Indeed, closing diphthongs and closed syllables in -j, -w aren’t known to contrast anywhere in the world… except in Central Franconian, e.g. the dialect of Maastricht, which is about as central as you can get in West Germanic.

    West Germanic is also particularly rich in centering diphthongs. Those created by non-rhoticity have mostly been monophthongized in English and (except word-finally) northern German, but southern German keeps those (except that Switzerland remains rhotic) plus, in the dialects, the ones created from breaking of */eː oː/ over 1000 years earlier! Any interpretation of such things as [ɪɐ̯ ʊɐ̯] as beginning or ending with a consonant when etymological /r/ is not available would have to be very abstract indeed.

    So, perhaps the trick is that West Germanic really has diphthongs all the way down, while just about everyone else just has vowels next to /j w/.

    BTW, Zeh m. is not quite standard; unambiguously standard is Zehe f.. Interestingly, my dialect seems to mix those: [ˈt͡sɛçŋ̩] m. (vowel length not phonemic, feminine -e often corresponding to -/N/ probably by generalization from everywhere else to the nom. sg.).

    • j. says:

      Closing diphthongs and closed syllables in -j, -w aren’t known to contrast anywhere in the world… except in Central Franconian

      I know a guy whose idiolect of French would allegedly contrast these ( /ai/ versus aille /aj/, etc.), though I am not convinced it isn’t simply a syllabification contrast.

      BTW, Zeh m. is not quite standard; unambiguously standard is Zehe

      I initially wrote Zehe, then checked Wiktionary to be sure I’m not misremembering the spelling, and their slightly confusingly worded usage note claims that Zeh would be the more widespread variant.

      • David Marjanović says:

        Ah, the source the English Wiktionary used is the Atlas der deutschen Alltagssprache, where 1) many respondents probably gave forms they themselves wouldn’t necessarily accept as standard, and 2) many respondents were probably confused whether they were asked for standard or dialect forms, because the standard doesn’t have a colloquial register everywhere, yet that seems to be a basic assumption behind asking for Alltagssprache “everyday language”.

        The German Wiktionary calls Zeh “as a tendency mostly northern” and Zehe “as a tendency mostly central and southern”. Apocope is a northern and southern thing, so maybe that was correct at some point.

    • David Marjanović says:

      Terminus ante quem for underlying diphthongs: the reinterpretation of *Vjj, *Vww as *Vij, *Vuw.

    • David Marjanović says:

      Oh, also…

      I assume this is not a very satisfying answer though. If areal connections do not seem to be the explanation, then we would surely want to have some other reason for why unusually many West Germanic languages have ended up resolving their crowded long vowel systems specifically by the breaking of long close vowels. This is in contrast to how there is also a second well-attested resolution strategy: shunting long close vowels into the short vowel system. To me this appears to be much more common than the West Germanic approach. Something along the lines of /iː i uː u/ > /i ɪ u ʊ/ > /i e u o/ happens in all of Scandinavian, Romance, Slavic, Hungarian, Samic, Persian…

      Part of the explanation for why this didn’t happen in West Germanic may be that /e/ and /o/ already existed there, at least in High German.

      Even before umlaut, [e] existed and belonged to the same phoneme as [i], not as [ɛ]. After all umlaut products had become phonemic, the short front vowels were /æ ɛ e ø ɪ ʏ/ (/æ/ from “secondary umlaut” of *a, /ɛ/ from *e and AFAIK from *i…a, /e/ from “primary umlaut” of *a and from *i under the conditions in the linked paper). On the back side, given the origin of *o from *u…a in Northwest Germanic, maybe it was [o] rather than [ɔ] for some time.

      On the front side, this six-fold distinction is preserved in the more hardcore Swiss dialects. Even in my Central Bavarian dialect, where vowel length and rounding of front vowels is lost*, the contrast between /ɛ/ and /e/ is largely** unchanged, and the old /æ/ is (with complications) one of the sources for the new /a/ rather than merging with /ɛ/. On the back side, *o is [o] at least throughout Bavarian, but don’t ask me if that’s a retention or due to pushing from *a(ː) > /ɒ/ (plus /ɔ/ as a recent loanword phoneme in unassimilated Standard German words); evidence for a push shift may be that the OHG/MHG /ɔː/ (conditionally from *au) has merged into /o/ with perhaps an exception or two – or perhaps /ɔː/ became [oː] before it lost its length, I have no idea.

      So, when the long-vowel system and the short-vowel system are both crowded, the choice may be reduced to diphthongs or megamergers.

      BTW, Polabian participated in the ENHG diphthongization. I wonder if that made it the only Slavic language not mutually intelligible with all the others.

      * Length works like in Russian now: stressed vowels tend to be longer than unstressed ones, with little influence from how many consonants follow. There are new rounded front vowels from L-umlaut, which happened surprisingly early.
      ** mittelbairische e-Verwirrung

      • j. says:

        That’s a partial point, but probably not the full explanation. Long-close-to-short vowel rotation often enough seems to leads to full or partial mergers with pre-existing short mid vowels somewhere down the line, e.g. in Romance (/ĭ ŭ/ and /ē ō/ merged as /e o/), Slavic (“strong” yers widely > /e o/), standard Hungarian (*i > /e/ > /ɛ/, cleanly merging with *æ > /ɛ/), parts of Scandinavian (at least in Swedish, lowered *i > *e, original *ɛ where not broken, and even *a-i > *æ all merge as a single /e/; though lengthening environments usually only merge these two-ways as [eː] versus [ɛː]).

        • David Marjanović says:

          lengthening environments usually only merge these two-ways as [eː] versus [ɛː]

          Oh, cool! I didn’t know there was any evidence for an /ɛ/-/e/ distinction having ever existed in Germanic other than High German.

          • j. says:

            Danish of course has all of /e eː ɛ ɛː/ and also /ø øː œ œː/ even today, plus [æː] as the default realization of /aː/. I have no idea though how heavily exactly this has been reshuffled since Old Norse, but IIUC the big picture is that length is always secondary, as also in the other standard North Germanic varieties, and Old Norse quantity contrasts are mostly continued as quality contrasts.

  17. David Marjanović says:

    Map of grammatical gender in west-central Europe: Danish has three different systems (m/f/n, c/n, Ø) geographically distributed at a scale that would be visible on the map.

  18. David Marjanović says:

    How to acquire phonemic stop voicing: [r] > [d] is another option, as seen in Dakhota as opposed to Lakhota & Nakhota. (In the latter two, [b d g] exist only as allophones before [l] – interestingly, an epenthetic [ə] is inserted between the voiced plosive and the [l], though.)

  19. David Marjanović says:

    Paasonen post: “in Mator, his *ć is normally reflected as k” – oh dear. Almost everything else makes (varying amounts of) sense if you don’t reconstruct farther than *ć, but this is a loud beeping alarm.

    It reminds me, though. I once read (years ago on an archived LinguistList post) that some part of Samoyedic has turned *si into [ki]. Is this based on Paasonen’s work?

    KL > TL: I had no idea this had happened anywhere in German, and barely remembered the English one – amazing to see it in Webster! I did know that Rhaeto-Romance had KL > TL, though: the one given example was dlatscha “glacier” (German Gletscher). Interesting that the German occurrences are nowhere near Switzerland.

    Webster vowels post:

    I have no idea what ‹ɛ› is supposed to be; the first two examples suggest something like [ɛ] ~ [ɑ] variation, but novel doesn’t really fit into this.

    Judging from the examples, ‹ɛ› is /ə/ spelled e, ‹α› is /ə/ spelled a. Likewise, ‹ṳ› is /uː/ spelled u, ‹ōō› is /uː/ spelled oo. Rather than trying to use one symbol per phoneme or per sound, Webster evidently added diacritics to the existing spellings to explain how the spellings were pronounced. The only exception is the “vowel glide” apostrophe for syllabic consonants.

    • j. says:

      There’s a Paasonen post forthcoming on this blog too, but yes, Mator also turns (PU *ś >) PSmy *s into k before front vowels, something that he remarks on as well. The actual correspondence is not doubted by anyone, but the phonetics have been puzzled over by many generations of Samoyedologists starting from him (and occasionally prompting adjusted reconstructions that project this issue also to Proto-Uralic).

      • Crom Daba says:

        Looks like another point for the affricate or even stop realization of *ś. What’s the last word on the distinction between *ś and *ć anyways?

        • j. says:

          It sorely needs a new overview, really. The etymologies posited earlier have been generally simply declared “uncertain” without any new analyses by the skeptical front, which I think is a somewhat unsatisfying stituation.

  20. David Marjanović says:

    Post on IE languages with a single stop series from mergers: the best example probably remains Tocharian, where only *d escaped (by turning into ts). I also wonder about the Roman accent of Italian, however, where my impression is that all plosives are voiced all the time (but /ts/ is not); I don’t know if the length contrast survives. Next best may be the area of the Inderior German Gonsonant Weagening, which has a single series realized as fortes word-finally and before /r l n/, as voiceless lenes otherwise (apparently with one exception that has developed intervocalic voicing). Next best after that are probably the pure length contrast in apparently all positions (even word-initially; no fortis allophones that I know of) in High Alemannic or thereabouts, and the pure length contrast in some positions (not word-initially or in clusters; less common elsewhere than in High Alemannic; fortis allophones before /n/ and I think in a few other positions) in eastern Austria.

    • j. says:

      David, I have to say I’m not sure what you’re generally aiming for by commenting here on my posts elsewhere. If it’s just for the fun of it, then sure, feel free to carry on — but if you’d like me to pass any of the info on to the main discussion, you’ll have to actually say that.

      (Also fwiw, a tumblr account is not that big of a hassle to set up and keep around, at least if you aim to limit yourself to a particular topic such as linguistics commentary.)

      • David Marjanović says:

        My aim is the same as everywhere else: to add information I think might be helpful. What you do with it is up to you.

        I still hope I’ll blog one day, but I’m not going to start a tumblr account. The concept of having to republish a post just to be able to add a comment is just wrong-headed. I’ve seen several discussion threads branch, so that even scrolling through all the “liked this” and “reblogged this” notices to find the comments does not in fact find all comments.

  21. David Marjanović says:

    On the next Germanic sound shift: it already happened around the 6th century and has mostly stalled since then. A good short summary paper, in German, is here; the most important part is the aspiration-free belt with reliably voiced lenis plosives that includes not only most of Dutch, but goes all the way east from there, containing all of southern Low and northern Central German. South of that belt, the unshifted fortes of Central German dialects are aspirated where they haven’t succumbed to the megamerger called binnendeutsche Konsonantenschwächung or, as I prefer in English, Inderior German Gonsonant Weagening. The western end of the aspiration-free belt could of course be blamed on Romance, and its eastern end on Slavic, but that leaves the part in between unexplained. – Given the fact that the aspiration-free belt once was the southern fringe of the Germanic-speaking area, it’s easily possible that aspiration has a single origin in Germanic, spread areally until it reached the belt, and was then carried south of it by migrations that crossed the belt.

    Vennemann was fixated on the supposed absence of analogues to the High German shift, to the extent that he claimed aspirates don’t ever become affricates (so that their common ancestors müssen seltsame Laute gewesen sein). They do all the time. It has happened in Liverp[ɸ]ool and is spreading; it has happened in the Otherwise Low German dialects of Dithmarschen; it’s happening in Danish – where it’s often, e.g. by Vennemann, said to be limited to [tʰ] becoming [ts], which isn’t true – and more slowly in Vietnamese (where different dialects have [f tʰ kʰ] or [f tʰ x] for ph th kh); and Burushaski has apparently free variation of [pʰ ~ pf ~ f] and [qʰ ~ qχ ~ χ] (while [tʰ, ʈʰ and kʰ aren’t said to be going anywhere). The whole bifurcation theory thus has no basis.

    Evidence that voiced fricatives other than just *z are old (in most positions – not all) comes for example from “Bahder’s law”, which is a phenomenon interpreted as voiced fricatives devoicing next to resonants in undefinable but large parts of High German before the West Germanic *[ð] > *[d] shift, merging with the short voiceless fricatives that came out of Grimm’s law. Examples off the top of my head (there are more): elf “11”, zwölf “12”, Schaufel “shovel” (*β); Boden “floor, bottom” (*ð, forming a Kluge mess with bottom from nominative *buðmēn, genitive *buttaz). That was enough to convince Guus Kroonen that voiced fricatives are old in Germanic and that the glottalic theory (or at least Kortlandt’s version) is therefore untenable. It’s explained at more length in Kroonen’s 2011 book on the n-stems. (Only the first page, i.e. the first half-paragraph, of v. Bahder 1903 seems to be online at all. It’s in Idg. Forschungen; I haven’t read it.)

    Voiced aspirates becoming voiced fricatives has happened a lot in India. Positional restrictions on this are not surprising either.

    The West Germanic *[ð] > *[d] shift looks like a substrate effect to me. I’m going to blame Gaulish until further notice.

    The only part that’s not clear is whether word-initial [ɣ] is a West Germanic innovation. “Greek” shows up in all Old Germanic languages and even Middle High German with initial */k/, but its */eː/ – *ē₂ – makes it highly likely that at least Gothic kreks is a West Germanic loanword.

    Maybe also worth noting: Sami has had a *t | *ð contrast ever since Proto-Uralic, but *ð only turns up in loans from Proto-Scandinavian on, not in the earliest ones from PGmc or pre-PGmc.

    Maybe that’s evidence that *d₁ and *d₂ weren’t [ð]-like in earlier times. …Is the *ð intervocalic on the Germanic side?

    Korean-esque “tense” stops

    The actual Korean ones are ordinary unaspirated voiceless fortes, no different from the Dutch ones except for the “faucalized voice” they impart on the following vowel.

    • j. says:

      I’m definitely not defending a “Bifurcation Theory” of the PGmc tenues, only of the mediae. Kroonen (2011) doesn’t seem to disagree with this part: “Since PGm. *b, *d and *g were affected by Kluge’s law, whereas the fricatives *s, *z, *f, and *h were not (see § 3.3.1 and § 3.4.2), it is defensible to assume that *b, *d and *g had a plosive rather than a fricative articulation.” (p. 53)

      Not familiar with a Bahder’s law, but from the reconstructions I’ve seen, ’11’, ’12’ and ‘shovel’ all had just *f, even though they also have some related Verner doublets with *b.

      Given the fact that the aspiration-free belt once was the southern fringe of the Germanic-speaking area, it’s easily possible that aspiration has a single origin in Germanic, spread areally until it reached the belt

      …Or that it was that belt that originated by contact influence, perhaps by language shift, versus High German spreading into less firmly held territory as a real migration. I’m not really following what this belt would have even consisted of before a supposed areal spread of aspiration?

      Is the *ð intervocalic on the Germanic side?

      Yep, e.g. PS *mētə > NS miehta ‘honey’, preS *putə > PS *pətə > NS bahta ‘arse’, preS *rawtə- > PS *rōwtə- > SS roevtedh ‘to be sunburnt’. (This would be otherwise a concern on the Uralic side too: there are no **lð, **nð, **rð.)

      • David Marjanović says:

        “Since PGm. *b, *d and *g were affected by Kluge’s law, whereas the fricatives *s, *z, *f, *þ and *h were not (see § 3.3.1 and § 3.4.2), it is defensible to assume that *b, *d and *g had a plosive rather than a fricative articulation.” (p. 53)

        Oh, that. As Kroonen himself says in the same chapter, you could equally easily say that Kluge’s law only worked in Verner environments ( = after unstressed syllables) and therefore didn’t have access to the voiceless fricatives or their ancestors; as he also points out, it’s hard to turn [z] into a plosive in the first place (though it has happened to isn’t in some Englishes – in a Grimm instead of Verner environment). Further, lots of fun can be had with ordering Grimm, Verner, Kluge and Kümmel; a discussion on that started here five days after you left that thread. :-)

        I have no idea whether Verner doublets can or cannot be postulated to account for Bahder’s law, though it’s perhaps strange that “11” and “12” have b specifically in Gothic.

        …Or that it was that belt that originated by contact influence, perhaps by language shift, versus High German spreading into less firmly held territory as a real migration.

        Good point: before the migrations of the Marcomanni, Quadi & friends, the aspiration-free belt was simply the southern fringe and could therefore be blamed on Celtic, assuming that Celtic didn’t aspirate despite having lost its *p.

        But if Gothic didn’t aspirate, its sound system makes sense as the Spanish type: plain plosives vs. voiced fricatives/approximants, with rare voiced plosive allophones of the latter. If so, aspiration is at most a Northwest Germanic phenomenon.

        I’m not really following what this belt would have even consisted of before a supposed areal spread of aspiration?

        It would simply be the area where aspiration, coming from somewhere north, hadn’t (and still hasn’t) reached.

        • David Marjanović says:

          I have no idea whether Verner doublets can or cannot be postulated to account for Bahder’s law

          I do: the “bottom” word has *dʰ, so any doublet would have to be very analogical indeed.

          Conversely, “11” and “12” have *p, so they’re probably not examples of Bahder’s law at all.

        • David Marjanović says:

          Stupid me: voiced fricative in Caesar’s time – Suevi.

      • David Marjanović says:

        Oh, on the loans into Sami – could they have come in before the relevant part of Grimm’s law, so that *dʰ was still intact?

    • David Marjanović says:

      Ha! Bahder’s law is being repeated in some Englishes. Emma Watson has it – search for lovely on this page.

      • David Marjanović says:

        A bit more material on Bahder’s law is in this paper by Kroonen (2006): first, if it’s real, it wasn’t limited to High German or part thereof, because bothme is found in Old Saxon and Middle English; second, there’s a beautiful parallel in OHG wedamo vs. Old Frisian wettma “dowry”, both masculine like Boden.

  22. David Marjanović says:

    Periods behind names on gravestones: all headlines ended in periods. Newspapers, book chapters, everywhere. Even today, in reference lists in scientific works, titles and the whole citations usually end in periods despite not being sentences.

    “Dagger”: it’s not a dagger, it’s a cross, or was at least soon reinterpreted as such. A name and a cross make a complete German newspaper headline, meaning that the person is dead and giving them a typographic Christian funeral. Ichthyologists, even in English, put it before the name of every extinct fish group.

  23. David Marjanović says:

    Qiangic languages: this very recent paper says the “closest kin” of Tangut are “modern Qiangic languages”, followed by footnote 2, which is:

    “Qiangic classification is less unproblematic than it seems, as the original criteria of a Qiangic language were of a typological character. Katia Chirkova (2012) challenged the notion of Qiangic as a genetic branch, by showing ways in which a language can areally acquire common Qiangic characteristics. Nevertheless, there is a core group, shown to be indeed closely related in Jacques (2014), which is called Northern Qiangic by Sūn Hóngkāi (2001) and Macro-Rgyalrongic by Jacques (2014). This group includes at least Rgyalrongic, Qiang, Minyag and Pumi. In this study, I retain the label “Qiangic” to minimize confusion, and cite only data from Rgyalrongic languages.”

  24. David Marjanović says:

    Title deflation: Mycenean qa-si-re-o “wezir, administrator”, wa-na-ka “king”. In Homer, ánax is more like “high king”, and then it drops out of use, and eventually the Eastern Roman Emperor ends up as a basiléus.

    Also: Karl “name of the emperor”, #/kuning/ “king” > Slavic “king”, “prince”, respectively.

    The only case of a 19th/20th-century Balkan state digging up “czar” is Bulgaria. Serbia/Yugoslavia stuck with kralj, even though in the 14th century a few Serbian rulers had called themselves car, held ambitions to inheriting Constantinople, and one of them almost got there.

    • David Marjanović says:

      Uh, sorry. The Slavic double-example of course belongs to title inflation (during borrowing), not deflation. I wasn’t quite awake yet.

  25. David Marjanović says:

    Phonological interaction between things not expected to interact:

    Calling it Nayraca’s law is a stroke of genius. From now on I’ll remember both instead of getting hopelessly confused. ^_^

    Gemination triggered by schwa: seems unsurprising to me. Schwa is reduced, consonants lengthen in compensation. But then, my native idioms don’t allow light stressed syllables…

    ʕʷ > ɥ in Abkhaz: makes perfect sense if *ʕ(ʷ) was not pharyngeal, but epiglottal [ʢ(ʷ)]! Epiglottal consonants are (comparatively) widespread in the Caucasus, and while pharyngeals pull vowels toward [ɑ], epiglottals pull vowels toward [æ]. An approximant that both fronts and rounds should easily end up as [ɥ].

    Interactions between tone and vowel quality if they exist: they totally do. Slovene has [e o] in one tone, [ɛ ɔ] in the two others. Ket has [e ɤ o] in one tone, [ɛ ʌ ɔ] in the three others. Middle Franconian not only has such things across the vowel inventory, it even has a secondarily toneless area in the middle, where the already large vowel system has doubled in size to 14 long and 14 short unreduced monophthong phonemes plus a bunch of diphthongs, a weird inventory of vowel + /j/ sequences distinct from diphthongs in -/ɪ̯/, and a schwa. I’ll look up the paper sometime later this week.

    As interactions that really don’t seem to make sense, I can offer the Saussure effect and rhotoglottophilia. I suspect the latter, found in Sino-Tibetan, is a byproduct of something else, though.

  26. David Marjanović says:

    Etymology of Finn: as I learned from Mikko Heikkilä’s thesis, the word is Fenni in Tacitus (98 AD), which is why you can find geologists talking about “Fennoscandia”, and Φιννοί in Ptolemy (150 AD). That means three things. First, φ was already suitable for representing [f] in 150; which makes sense, because the oldest FILIPPVS is about that old. (In Pompeii there’s a PILIPPHVS on a wall somewhere.) Second, the Germanic *enC > *inC shift either happened in the first half of the second century, so Tacitus got the name from Pre-Germanic, or he got it from a Para-Germanic dialect. Third, the etymology from “find” is out of the question, because it requires the specifically North Germanic *nd > *nn shift.

    A source of PGmc. *nn is *zn from Verner’s law. And so, Heikkilä derives *finnaz from a Pre-Gmc. *feznaz, which he leaves unexplained. Piotr Gąsiorowski derives that from PIE *pes-nó-s, which is also what the derivationally circular word pēnis is derived from.

    Perhaps the people who brought the bronze axes to Mälar were particularly manly men. Or something.

  27. David Marjanović says:

    Winged pseudo-Proto-Uralic: Oh come on. Now that I learn it looks possible, I demand a full-sized fable like Schleicher’s!

    (It could be actually useful by highlighting underresearched areas.)

  28. David Marjanović says:

    Rare affricates: The map seems to be intended to depict phonemes, not random consonant clusters. If so, it’s wrong to paint the northern half of Germany brown for /p͡f/: morpheme-initially, pf is just pronounced [f] there, and yes, I am talking about Standard German. Elsewhere in a word, [p͡f] exists because there it can be interpreted as a weirdly common consonant cluster – it exists in the same sense that it exists in English words like hopeful.

    In a note, easternestablishmentarian wonders about [p͡ɸ]. I have never heard that in German, definitely not from native speakers. The trick may be that easternestablishmentarian is from Britain – it seems to me that Britons lack an overbite much more often than European mainlanders do. Experimentally, when I put my incisors on edge, [p͡ɸ] becomes easier than [p͡f].

    That said, I do think that the spelling ph that was so common in Old High German was meant to represent [p͡ɸ] in most or all cases.

  29. David Marjanović says:

    Phylogenetic analysis of Sino-Tibetan: I read that paper right when it came out, and then asked in the Facebook group “phylogenetic analysis ruined my life” how that program roots its trees, because the paper only says no outgroup was specified.

    Turns out it’s a slightly more sophisticated version of midpoint rooting. Basically, the root goes on the longest internal branch. What that really tells us is that Sinitic doesn’t have any documented very close relatives. A basic split into Sinitic and Tibeto-Burman is only one of several possible reasons for that fact.

    The exclusive use of lexical presence/absence data, with no regard for morphological or phonological innovations (very few of which have even been suggested for ST…), comes on top of that.

  30. David Marjanović says:

    The mystery of babies’ first words:

    not to diminism too much from the language acquisition part, but: “proto–land animals crawling from Cretaceous seas”? this is basically equivalent to “the heavy snowfall of May”, and as far as I know this is not even a well-known enough name of an era to be expected as metonymy (“mesozoic” I could understand maybe)

    Maybe they confused the Cretaceous (in the Mesozoic) with the Carboniferous (in the Pal(a)eozoic). But given this level of knowledge, I’d expect them to think not of the Carboniferous, but of the preceding Devonian, when Ichthyostega lived. (Long considered the first to walk on land, Ichthyostega has been known for 25 years now to have been incapable of walking; it does seem to have moved on land some of the time, but more like a seal or the Godzilla version of a mudskipper. It is not closely related to any terrestrial or amphibious animals; we’re descended from a separate transition to an amphibious lifestyle for full-grown adults that seems to have happened early in the Carboniferous.)

    On the language acquisition part, the first word of prospectively Spanish-speaking babies is traditionally expected to be “garlic” – ajo, another one of those that easily happen by chance like a mama-papa word.

  31. David Marjanović says:

    Winkler (1885): Your translation is good, but:

    deutlich geschieden I’d render as “clearly separate”. Incidentally, this is a good example for how much Standard German has changed since then – today, the only meaning of geschieden that is in common use is “divorced”, and for “separate” I’d use getrennt. Except I’d probably reword the whole thing and use deutlich unterscheidbar, “clearly distinguishable”.

    z. T. is zum Teil, “in part”. That’s hard to render in an idiomatic English sentence, so I’d go for “into a number of languages and dialects, which are for the most part substantially different”.

    Alone the Western Finnish

    Just “West Finnish alone”.

    not even a derived adjective like “Finnic”

    – that’s because German doesn’t have a way of forming such things. Only the Turkologists have managed to establish Turksprachen (while they themselves use Türksprachen), and the Turkish language is still often called Türkeitürkisch, “Turkey Turkish”, in linguistic contexts.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Enter your email address to follow this blog and receive notifications of new posts by email.

%d bloggers like this: