Nelson Goering continues to work on Old English (see this LH post), and in a recent Facebook post he reported the publication of his open-access paper “A sound change that never was: h-loss and vowel lengthening in Old English” (English Language and Linguistics, doi:10.1017/S1360674325000164), which looked interesting enough to post about here; the abstract:
In the 1880s, Sievers proposed that in Old English words such as *feorhes, the loss of the post-consonantal *h caused compensatory lengthening of the vowel: fēores. Since there are no unambiguous traces of this sound change in later English, widespread analogical restitution of the short vowels was assumed (e.g. from feorh). The evidence for this lengthening is largely metrical. I argue that while Sievers is correct that words like <feores> often need to scan with a heavy initial syllable, this need not be explained by a general lengthening in the language at large. Indeed, the distribution of where heavy scansions are required in verse is typical for metrical archaisms: late prehistoric metrical values of words preserved for poetic convenience. Just as wundor ‘marvel’ can continue to be scanned as monosyllabic *wundr, or contracted hēan can scan as disyllabic *hēahan, so can light-syllabled feores continue to scan as heavy *feorhes. The same sets of poems that prefer non-epenthesized or non-contracted forms also prefer the heavy scansions of feores-type words. If heavy scansions of feores-words are seen as a matter of poetic convention, then the hypothesis of compensatory lengthening in the language generally is left without evidence and should be rejected.
In the FB comments, Haukur Þorgeirsson quoted a passage and followed it with his own thoughts:
“I leave it open how poets, especially older poets, actually pronounced the forms in question. Did they simply retain the historical h-sounds (i.e. [x])? Did they substitute some other sound as a place holder, such as [ʔ] or [h]? Did they adopt an artificial syllabification, such as feor-es, for the sake of performance, as suggested by Hogg (Reference Hogg2011: 169)? Did they use the younger pronunciation, and simply understand the mismatch with metrical form to be a legitimate licence? Or, perhaps, did they actually lengthen the vowel of the root syllable, saying fēore, as a way of maintaining the proper syllable weight? I would emphasize that the last option is a reasonable possibility – but under the view argued for here it would be entirely limited to the recitation of poetry, a feature of performance, and never part of the ordinary language.”
Thank you for lining out such a juicy set of options. Some thoughts on each in turn:
* Pronouncing the historical h-sounds seems like a very natural option for which I think various parallels could be cited.
* Artificial syllabification seems sus to me, I’m just not sure syllabification is a ‘real’ enough thing for that to work out. Exactly what is it supposed to entail here in concrete audible detail? An unnatural pause in the middle of a word?
* Using the younger pronunciation and just understanding the mismatch: This is the sort of horror that keeps me up at night. People love to propose explanations like this but this just seems so hard to do in practice. To be sure, people composing traditional Latin poetry in the Middle Ages were doing this at scale – basing their syllabic weight on theory and history rather than on audible features. And the Chinese poets diligently based their rhyme on historical correspondence classes rather than on actual pronunciation. But these guys had handbooks and tables and stuff!
* Lengthening the vowel: Can’t rule it out, I suppose. But I find it hard to imagine how a tradition like that would originate if Sievers was not right about it being an actual sound change to begin with. I’d rather lengthen the consonant!
Nelson Goering responded:
I’d probably be inclined to agree that reproducing the historical sounds is the simplest option, and yes, I’m personally very skeptical about the mismatch explanation. But it’s a very common one among OE scholars (I had a reviewer for a different paper once bring it up, presenting it almost as if it were an unanswerably plausible default option), and I would want to take a good deal of consideration and space to properly argue against it, and didn’t want to distract particularly from the other parts of the argument here. I’d like to write an article specifically on the subject sometime, though.
Raphael Turrigiano adduced a parallel from Latin:
As far as lengthening the vowel goes, in Latin this is directly attested to by Aulus Gellius just after the classical period regarding compounds like ‘adjicere’ < ad + jacere where the first syllable is a closed long syllable. Evidently either these compounds had fallen out of speech (they leave no trace in romance) or the sequence /ji/ had simplified to /i/ or /iː/ leaving the first syllable short. As a result, Gellius notes that many people when reading classical verse would lengthen the first vowel ‘ādicere’, but he criticizes this and advocates instead for the etymological pronunciation.
It’s great to see this kind of knowledgeable back-and-forth (and there are further good comments in the thread). While I’m on the subject of Goering, he has a more recent post that’s also of interest:
I’m very glad I have decided to put together my own glossary for Beowulf, and that I took Irina Dumitrescu’s advice (which was in line with my own inclination, but it was good to have a timely nudge) to gloss elements of compounds separately (except when they’re fully univerbated: gārsecg “open sea” and fyrwet “interest, curiosity” aren’t treated as compounds).
There are at least three reasons for this. One is that it’s a nice chance to clear out a lot of the traditional but misleading glossing that’s piled up over the decades. Everyone glosses fǣhð(o) as “feud”, but this is an etymological gloss, and the word just means “enmity, hostility”. There just isn’t any implication of the tit-for-tat or prolonged nature of a feud inherent in the word. It *can* cover that kind of hostility, but in the same way that “enmity” or “hostility” do, not as a direct part of its meaning (“their enmity goes back 30 years” is not quite the same thing as “their feud goes back 30 years”, and fǣhð seems, to judge from the contexts when it’s used, to mean the former, not the latter). Not surprising, since it’s a completely transparent (within Old English) abstract noun to the adjective fāh “hostile” (think modern “foe”). This is just one example of a fairly common issue in Beowulf glossaries.
Another reason is that it’s a good opportunity to reconsider compounds. Ever since Grein parsed seld-guma as “hall-man”, this has been largely unquestioned in Beowulf glossaries, even as people struggled to make sense of the word in context. I’m pretty sure that this is to read the word with the wrong “seld”, but the idea never occurred to me until I came to gloss the compound element by element. And as the standard edition of Beowulf notes, many compounds obviously have a wide range of potential interpretations of exactly how their parts relate, and no glossary can reasonably point to them all.
And thirdly, it’s just been a good way of forcing myself to reconsider the text overall. Sometimes this has led me to abandon emendations I’d initially included (e.g. I’ve given up on ge·þrūen for MS geþuren, though it’s still a puzzle), sometimes it’s made me change my mind on well-known cruces (like just what the fyrgen-strēamas at Grendel’s mere are), and sometimes it led me to shore up and make more precise points I’d already had suspicions about (like how to sort out the gast, gæst, gest, gist complex).It’s certainly been an informative process for me. I hope the end result is useful to other people, once it’s done (I’m still working on it, especially the cruces and harder words — I’ve just decided that no, I, like everyone else, don’t really know what the “scennum” on the sword hilt are).
(Note: I’ve hesitated a bit on formatting, and things like “pronoun” will probably end up being abbreviated in the final version.)
I have often had occasion to grumble at bilingual dictionaries that gloss a foreign word by a cognate English one rather than using a more accurate equivalent.
Unrelated — I just heard about the passing of mathematician and regular LH commenter Jack Morava.
Damn, I’m very sorry to hear that. Do you know what happened? It seems like just a few days ago he was commenting vigorously. (Wikipedia hasn’t noticed yet.)
A great pity.
It just happened this morning, and I don’t have any further information. It was announced on a mailing list for algebraic topologists.
*pressing F to pay respects*
For this it’s enough if it was an actual sound change in just one not too obscure dialect: once people knew a solution to the problem existed, some of them would have used it even if it hadn’t occurred to them on their own.
That would be somewhat similar to the German convention that rounded and unrounded front vowels of the same openness are allowed to rhyme. The implied unrounding is a real change in over half the dialects, but it’s also all over the works of poets whose native dialects never did this. Like at the beginning of the basic epic poem:
Ach, was muss man oft von bösen
Kindern hören oder lesen!
A good comparison.
זכרונו לברכה
…and now I’ve read the paper, which says there’s evidence from placenames in southwestern England that this may actually have happened there.
Something else I’ve learned from it is a case of phonetic pedantry paying off: Middle English, or at least the part that includes Chaucer, had a three-way contrast of /ɒː ɔː oː/ – there’s an ABABB stanza from Chaucer that rhymes on /ɒː/ for A and /ɔː/ for B, with the second A and the second B being identically spelled words!
In the interest of further phonetic pedantry, however, I have to comment on this quote from the paper:
This is about how the e of swer could have come from an *i. In (Old) High German, *i…a gives e regularly, just as *a…i does, but elsewhere (well, I forgot about Old Saxon) this is rare or absent, with nest* and wer** the only cases that are found all over Northwest Germanic (in East Germanic the evidence is erased).
That’s where I miss a citation of this paper. Section 5 proposes that first-syllable *i became *[e], which stayed distinct from */ɛ/, shortly before Proto-Germanic whenever it was followed by *r, *st or, but only if in the same syllable, *z.
…and as if on cue, my dialect has /e/, as opposed to /ɛ/, in /nest/. Indeed, the only -/ɛst/- word I can think of is rösten “roast ~ fry”.
* < Roughly PIE *nizdó- “nest”, composed of *ni-sd-ó- “down-sit-”
** < Proto-West-IE *wiró- “man ~ warrior ~ hero” < Proto-Indo-Actually-European *wihró- “id.” < Proto-Indo-Tocharian *wihró- “young”
The last comment (didn’t do exhaustive search).
The angle brackets caused loss of a word in the first block quote. It should say “I argue that while Sievers is correct that words like <feores> often need to scan with a heavy initial syllable….”
Fixed, thanks!
@DM, I thought I’d generally kept up with Hill’s papers, since they’re usually very good, but I clearly missed that one! I’ll take a read through, but if he’s right about a-umlaut, that would be very interesting indeed.
“For this it’s enough if it was an actual sound change in just one not too obscure dialect: once people knew a solution to the problem existed, some of them would have used it even if it hadn’t occurred to them on their own.”
True in principle, though I think not the best explanation in this case. For one thing, the “not too obscure” part is at least questionable in this case, since the regions in question don’t feature terribly prominently in the political or cultural geography of English-speaking Britain during this period. For another, the distribution of forms within the corpus seems about backwards, if this were the explanation: they’re most common in older, often (where we can tell) Anglian poetry, and decline in later and southern poetry (the latter of which would have the best chance of any to be acquainted with dialectal peculiarities from the far Southwest). In any case, if the explanation here lay in dialect geography rather than in the conservatism of poetic tradition, then I’d wonder why it’s so parallel to things like epenthesis and contraction (which are surely not also to be explained as Southwesternisms!). As usual, I’m not sure one can conclusively rule out this explanation, but it doesn’t seem excessively promising to me.
(There are, to be sure, cases of dialect variants used for metrical convenience. Things like the geminated nominative fædder or the use of the weak past sceþede alongside the more usual strong scōd are examples. But they’re not actually all that common, especially in comparison to variant forms that probably had no synchronic existence outside of poetry. Outside of metrically relevant forms, later southern poetry, like The Metres of Boethius or The Battle of Brunanburh will Mercian spellings in items of poetic vocabulary, even when these make no difference to the scansion. This is probably a fairly specific context, with Wessex emerging as the new dominant power, but against a backdrop of Mercian prestige.)
Fair enough!
The interesting thing about Hill’s paper is that I misremembered what it has to say about a-umlaut (*i…a > *[e]). I thought it just exempts nest & wer and explains them differently; but no, it seems to deny the whole phenomenon, pointing out quite rightly that it didn’t happen in past participles (shown on p. 13) and directionals (p. 14) and analogical leveling seems to lack a model (p. 14).
So I looked it up in Ringe & Taylor (2014; on Google Books here). P. 34: first, “[i]n Norse there appears to have been a comparable lowering of *i to *e (Noreen 1923: 53–4)”, treated by Hill in a footnote (that points out spelling variation i ~ e in Old Norse) but otherwise implicitly denied by the past participles cited on p. 13 there; then, “there was a modestly extensive lowering of *i to e in the southern part of the WGmc area, usually before labial and velar obstruents which were in turn followed by nonhigh vowels”, citing the standard OHG dictionary from 2004. Sure enough, the 12 examples look pretty good, and they all have a labial, a velar, or in one case a labialized velar; and the 3–5 I can find in my dialect have /e/ as opposed to /ɛ/.
Hill actually mentions one of them (“lick”) in a footnote, lists more variation than Ringe & Taylor and says it’s actually two unrelated verbs. But if German schlecken “lick” is related to one of them, well, it’s got /e/ in my dialect.
One example looks like a blunder: “PNWGmc *spika- ‘bacon’ (ON spik) > OE spiċ, OHG spek (?→ OF, OS spek)”. Note the lack of HG consonant shift; Speck must be *spikka- and is indeed /ʃpekː/ in my dialect. This leads me to two words Ringe & Taylor didn’t cite, Zeck(e f.) m. “tick” and Dreck m. “dirt” (pejorative), which likewise end in -/ekː/ where I’m from.
Then comes a list of three exceptions, two with labials, one with a velar. In the first, “ship”, OHG has both i and e. In the second, “tremble”, only i is listed, but modern German only has the e form (beben “quake”; not available in my dialect). The velar example gets a relatively lengthy discussion of etymological uncertainty. So maybe it isn’t real; maybe the other two already show Low German influence, or perhaps the *j of “*bibja- (*bibā-?) ~ *bibai-” is to blame (but it is also present in 2 of the 12 examples and had no effect there).
The most annoying part is that Hill cites a past participle with *-ig- as an example of a-umlaut not applying (“step”, ON: inf. stíga, pp. stiginn; OE stīgan, gestigen; OHG stīgan, gistigan) while Ringe & Taylor cite practically the same as an example of it applying: “PWGmc *sti/egōn ‘to ascend’ > OE āstigian, OHG stegōn“. To this I can actually add the modern noun Steg n. “narrow walkway”. Hill’s other examples all have coronals, but if *g swings both ways that doesn’t help much.
Anyway, if a-umlaut is real, it would help explain why the /ɛ/-/e/ contrast seems to have become phonemic very early in Upper German despite never catching on elsewhere. /e/ would have the following sources:
1) former allophone of /ɪ/ following Hill;
2) *i…a with the above complications;
3) *a…i/j absent well-understood blocking factors;
4) early western Romance loans (Pech, Pfeffer, Semmel < Latin pica, piper, simile).
Finally, just to inject some more madness into the situation, unexpected /e/ isn’t limited to Nest in my dialect. It’s also found in essen + fressen and sechs; why, I have no clue at all. I mean, etymological nativization from Standard German could do it*, and for “6” that might be just barely imaginable, but why would “eat” be borrowed so widely?
* Following open-syllable lengthening, monosyllabic lengthening, closed-syllable shortening, and the Central/Standard German reshuffling of vowel qualities according to the new quantities, the most common correspondences of Standard /ɛ/ & /eː/ with dialectal /ɛ/ & /e/** are crosswise. Tee and OK have accordingly ended up with /ɛ/.
** Phonemic vowel length is lost entirely, phonetic vowel length depends on stress and syllable weight like in Russian. Phonemic consonant length remains at the ends of stressed syllables, with losses and partial restorations word-finally.
DWDS on Speck, citing Pfeifer’s Etymologisches Wörterbuch, italics omitted:
Looks like *kk on the mainland but not in England, and beyond Germanic there’s at best a root etymology.
There’s a verb spicken, basic meaning “pierce a piece of meat and stick little bacon pegs in so the whole thing becomes fatty enough to fry and/or tastes more of salt and smoke” or something like that, metaphorically “stud with”, e.g. writing studded with archaisms or whatever.
Sad to hear of Jack M’s passing (though I hadn’t known of him as a mathematician, only as a smart and thoughtful LH commenter).
On the problem of how to deal with sound changes messing up scansion, Homer is an obvious place to look for parallels. There are certainly cases where even very awkward mismatches are left unfixed (e.g. φίλε έκυρέ with second syllable scanned long, ἀνδρότητα with second syllable scanned short, and even ἀνδρειφόντῃ with first two syllables scanned short!), but also cases of repair by lengthening either a vowel or a consonant (δείδιμεν, ἔδδεισα). I don’t know if there’s evidence that e.g. Attic poets ever dealt with the difficulty by pronouncing lost digammas, which they might have known about from other dialects.
First, because it was *a.nr.tā.ta once.
Yes, thanks. Commented before coffee…
There’s a verb spicken, basic meaning “pierce a piece of meat and stick little bacon pegs in so the whole thing becomes fatty enough to fry and/or tastes more of salt and smoke” or something like that, metaphorically “stud with”, e.g. writing studded with archaisms or whatever.
English “lard”, which I think includes actual lard as well as bacon. “Larded with” has a similar figurative meaning. I don’t think it would be widely understood.
Sad to hear about Jack Morava.
“Larded with” has a similar figurative meaning. I don’t think it would be widely understood.
Aww the figurative sense seems pretty familiar to me. (And how it arises from the literal sense is clear, though I don’t think I’ve run across that.)
“Larded with” evidence usually connotes some claim is trying too hard, with a hint the weight of evidence doesn’t actually support the conclusion/quantity over quality.
I’ve had a look at Hill’s paper now. Despite having some quibbles (I think he treats lengthening too rigidly, at least, seeing matters large as specific sound changes at fixed moments in time, rather than ongoing alternations of stressed and unstressed variants), I think he’s probably right, both about lowering before *r (and more?) — it at least gives a satisfactory explanation to the long-standing mystery of *wer-, and the fact that it can also account for the lowering in *hē̆r (the lengthened reflexes are not necessarily of PGmc date) gives it a real plausibility.
Congratulations to Nelson. It’s exciting to watch scholarship move forward – and to see discussions in this community be part of it.
Jack Morava will be missed. To be so obviously brilliant, he had an admirably unassuming and uninsisting presence.
Amen to all of that.
But that is the simplest hypothesis, isn’t it? Is there evidence for short ones that don’t mean “hither” (*hera)?
That, however, is certainly a good point.
The simplest hypothesis, to my mind, is that this began as alternating *hĕr when unstressed and *hēr when stressed. So, on the one hand, yes, the lengthened reflexes would probably be of PGmc date. But what I should have said was that it’s not necessarily clear that the lengthened reflexes would have been uniformly generalized in Proto-Germanic. Note that in general, there’s a lot more length variation, often easily demonstrable, than Hill credits. Old English *be* must be short much of the time, far more often than in verbal prefixes (this is the normal spelling of the adposition, and it’s not very plausible to assume that the common adpositional use is entirely analogical to the verbal prefix). To this day, Dutch wij has orthographic evidence for length, but only appears as as a diphthong under emphasis, and is far more frequently a schwa. Old Norse hvar must, pace Hill, be short at least some of the time. Leaving aside the fact that some MSS are not nearly so bad at length marking as he claims is universally the case (such that he can disregard *all* orthographic evidence on this point), the shortness is shown by Jórunn Skáldmǽr’s Sendibítr 4.1 (early 10th c.), which has a hending between *hvar (vitu)* and *ǫrvir*. This speaks against Hill’s uniform early length in this adverb, and in general things are much easier to work with if we allow stress-conditioned length variations for a long period of time. The here-word may, I suspect, be more prone to emphatic use, allowing the lengthened form to be more persistent than in *there*, but that doesn’t make length in PGmc universal. And the fact that it didn’t lengthen to ˣhār in WGmc shows, as is well known, that we don’t have a normal Germanic *ē in this word (note Hill’s framing of it as a “mere allophone”: he wants the change phonetically, but knows he can’t claim it phonemically).
This is also easy to frame prosodically if final *r was extrametrical in PGmc. This would mean that *hĕr could surface as such whenever is was of low enough prominence to not need to be footed, but would automatically be lengthened to *hēr, heː<r>, when footed, to meet the bimoraic minimum. Probably the same explanation underlying the be/big variation in OE, etc.
“would automatically be lengthened to *hēr, heː, when footed”
That is meant to be *hēr, followed by heː + extrametrical r. The brackets I used to show extrametricality of course got taken as code and were gobbled up by the hungry internet demons.
Spække is what you do to meat loaves, and spækket med has the same metaphorical use as larded with. (I used to love snatching the crisp bits of smoked lard from the meat loaf as a snack. But my sisters and father left them uneaten because of an inexplicable fat-phobia).
But it can also mean “wrapped in bacon,” which has to be a later development since bacon as a product is a more recent thing. (We used to sell it to the English, but Denmark stuck to the gold standard after WWI long enough that the Krone appreciated wildly and they had to sell it domestically).
That is meant to be *hēr, followed by heː + extrametrical r. The brackets I used to show extrametricality of course got taken as code and were gobbled up by the hungry internet demons.
I managed to subdue the hungry internet demons and get it fixed.
Well, this problem disappears immediately if the normal Germanic *ē was [ɛː], but Hill’s sound change only produced [e] and, in stressed “here”, [eː]. That means *[e] and *[eː] remained allophones of */ɪ/ in Proto-Germanic, not of */ɛː/. In East Germanic this situation was destroyed by the general raising (discussion starts here) of *[ɛː ɔː] to [eː oː]; in Northwest Germanic, *[ɛː] moved in the opposite direction, and *[e] & *[eː] were variously lost or phonemicized later when the same sounds were produced from other sources.
[e] as an allophone of /ɪ/ is globally not rare; it’s right here in the Rgveda.
The charm against the hungry internet demons is < for < and > for >.
Thanks for the fix!
I understand Hill’s solution, and noted his use of a dot to indicate tense e. But this is just restating the same thing, that there was only one phonemic non-high front vowel in PGmc, and here didn’t have it. I’m only trying to emphasize that this means that the long vowel here must have been in phonologically secondary for a long time, and that I would suspect it’s phonetically not universal. I wonder how long PGmc hēr could have survived unmerged with normal *ē without support from a short vowelled form.
If it was a completely predictable allophone of */ɪ/ – lowered by the *r, lengthened by the *r in a monosyllable under stress – then I’m tempted to say “indefinitely, until another source of [eː] or a new source of -[ɪr] comes along”.
But of course it needn’t have survived for a long time at all. East Germanic eliminated the issue as mentioned, and Northwest Germanic created a new */ɛː/ that this *[eː] did apparently merge with immediately.
(The main source of the new */ɛː/ is described in excruciating detail, in German, in these two Hill papers that I’m sure you know, but if anyone else is interested…)
I tend to be skeptical about language learners recovering diachronic sound changes indefinitely as synchronic rules (Juliette Blevins has a whole book on why this is a dubious assumption). When the rule in question would have so very few inputs (perhaps a few more than we can now reconstruct, but this is hardly a frequent process!), a large portion of which would be pretty non-obvious (even assuming the connection to *hi- was recovered generation after generation — not implausible, but not to be taken for granted either — *hēr is subject to a further rule, and what grounds would any learner have for positing lowering in *weraz?), I’m extra doubtful.
Long-standing rules that get transmitted robustly tend to involve active morphophonemic alternations, which are arguably just morphology, not phonology at all. This ain’t that. (I realize this is heading dangerously towards messy theoretical issues. If you like Distributed Morphology, you won’t like what I’ve just said very much!)
Good points (I don’t know Distributed Morphology, though).
On the other side, both of Hill’s sound changes could have happened very shortly before Proto-Germanic as far as I can tell, so perhaps we simply are dealing with a short-lived phenomenon that left few traces.
“both of Hill’s sound changes could have happened very shortly before Proto-Germanic as far as I can tell, so perhaps we simply are dealing with a short-lived phenomenon that left few traces.”
Yes, that would also be pretty plausible (though less so when the evidence for more variable length in *hʷā̆r is taken into account).
“Where” is unstressed pretty often, as a relative pronoun or in questions like “huh, where’s that now?”. “Here” much less so, I’d expect.
If it’s accepted that the lengthening had something to do with prosodic prominence (again, I’d specifically connect it to the well-known requirement for stress-words — but not low-stress particles — in early Germanic to have at least one two-mora foot), then there are really only three options:
1) PGmc had stressed *hēr, unstressed *her.
2) PGmc generalized the length in *hēr even to unstressed instances.
3) PGmc only had stressed uses of *hēr.
I think 3 can be dismissed out of hand as a fairly absurd notion (certainly there are plenty of examples of low-stress hēr in early West Germanic verse, at least, and no reason to think there is anything innovative about them). 2 seems to me unlikely before a stage in which tense *ē had become phonologized (though once this stage was reached, independently in East and NW Germanic, it would be very natural for the lexeme to be phonologized as /heːr/). Since Hill explicitly says the lengthening is allophonic, to me the conclusion is that this lengthening was still conditioned by stress, and so varied synchronically by stress.
Thinking about this in light of this conversation, I’d be inclined to posit the following stages:
1) Pre-PGmc *þar, *hir
2) Lowering of *i before *r: */þar/, */her/
3) Extrametricality of final *r: */þar/ [þar], [‘þaːr]; */her/ (or /hir/, if you can justify lowering still being active) [her], [‘heːr] — this is PGmc proper
4) *ẹː becomes phonologized: */þar/ [þar], [‘þaːr]; */heːr/ [heːr], presumably still with shortening to [her] sometimes in particularly low-stress contexts, though I don’t think there’s any actual direct evidence for this
5) WGmc: *hēr remains the same; */þaːr/ also becomes phonologized with length
*hʷar should be fully parallel to *þar, unless I’m forgetting about something (which is possible: the 4-month-old was restless last night, and the brain is starting to feel a little molasses-like).
We seem to be in violent agreement, except of course that you know details I didn’t.
@Nelson: I finally got around to reading the paper. It’s very well written and easy to follow.
Looking up Myrvoll 2015, I see that he aims to demolish a very similar case for compensatory lengthening in Old Norse. Like you, he is discussing the evidence from metric poetry, but without any explicit reference to metrical archaism (which may have less tradition in the analysis of Old Norse poetry). In your view, would archaic reading make sense even in Old Norse?
[Sidenote: You quote without translation, even from one of the lesser languages in Germanic scholarship. Does that actually still work with your colleagues, or is it for the hell of it?]
the 4-month-old was restless last night, and the brain is starting to feel a little molasses-like
This calls for belated congratulations! You won’t get your brain back, but I hope there’s a decent paternal leave looming.
Norse poetry has less synchronic alternation of forms the way Old English does, but there are certainly cases where archaizing occurs. I don’t know if fýri is well explained this way. Ironically, its very isolation is a problem — not that we have all the Norse poetry there ever was, of course, but it’s a more ad hoc explanation in a Norse context.
“You quote without translation, even from one of the lesser languages in Germanic scholarship.”
I left German untranslated, but I assume that’s not what you mean, and I thought I translated everything else. The only other language I can remember quoting is from Myrvoll’s paper in Norwegian, and I do translate that. If there was something else, it was probably a slip on my part not to translate!
“This calls for belated congratulations! You won’t get your brain back, but I hope there’s a decent paternal leave looming.”
Thank you! And yes, a quite generous allotment of leave, beginning tomorrow actually.
You’re right, of course. I failed to notice the footnote. I presumably would have if I needed it.
Enjoy your leave. If it’s long enough, you’ll be longing back to work by the end of it (and feeling guilty about it). My young colleagues are like that, mothers and fathers alike.
I left German untranslated, but I assume that’s not what you mean, and I thought I translated everything else.
Returning to Trond’s question, is that due to publication requirements or just an act of generosity on your side? When I studied, it was expected that in historical linguistics you would be able to read scientific literature in all extant languages of the subfamily you studied (e.g., all Slavic languages in Slavistics).
If I remember right, I submitted the article with no translations, figuring I’d add in any needed later depending on what got changed in revision. One of the reviewers specifically requested the Norwegian be translated, but I probably would have done it anyway. I guess it wouldn’t be too challenging, but it is a tad less widely studied than German.
It may have contributed that Myrvoll’s language is on the far side of the Nynorsk spectrum when seen through the lens of another Continental Scandinavian language.
Nelson: Norse poetry has less synchronic alternation of forms the way Old English does, but there are certainly cases where archaizing occurs. I don’t know if fýri is well explained this way. Ironically, its very isolation is a problem — not that we have all the Norse poetry there ever was, of course, but it’s a more ad hoc explanation in a Norse context.
That’s more or less what I’ve picked up in bits and pieces, but it seems to me that at least some of Myrvoll’s examples of variant pronunciations might be seen as licensed archaisms. Perhaps my question is if the Norse and Anglo-Saxon poetic traditions were so intertwined that what can be safely shown for one can be assumed also to be the case for the other.
seen through the lens of another Continental Scandinavian language: Not impossibly so, though without consulting a dictionary much remains an educated guess. I don’t know what the man on the bus to Vanløse would make of it.
I’d never catch a gender mismatch, for instance. And already the second word in the title is challenging: uturvande must be ‘unusual’ (‘out of the usual’), but Danish doesn’t do that sort of prep+noun compounds. En usædvanlig forlængelse af vokaler før bortfaldet h efter r og l would be cromulent in Danish.
“Perhaps my question is if the Norse and Anglo-Saxon poetic traditions were so intertwined that what can be safely shown for one can be assumed also to be the case for the other.”
They’re certainly closely related, but not particularly obviously intertwined. It’s often very useful to compare them, but it’s not at all safe to simply project features from one onto the other — though they do often turn out to share features, particularly in terms of metrical workings. But even where the same scholarly terminology is used, as with “kennings”, the feature described may turn out to be more different than similar. With things like h-loss, contraction, and so on, that involve parallel innovations operating separately (and at different times) in the two traditions, things seem to play out rather differently.
Uturvande a. “Unnecessary”, negated prec. part. of turva v. “need”, cognate of Ger. dûrfen.
@Nelson: Thanks again. Adjusting my obvia accordingly.
@trond, lars
Danish seems to have lost the v in the verb stem due to conflation with another Old Norse verb (cognate to English dare?).
https://ordnet.dk/ddo/ordbog?query=turde
Yes, though I think the v was lost before the conflation happened. We still have a labial fricative in nødtørft = ‘what you gotta do when you gotta go,’ to be a bit delicate about it; the meaning of turde is only that of ‘dare’ now. It looks like the distinction with the other one was always a nice one, and the reorganization to give a (strangely spelled) infinitive to a near-modal can’t have helped.
(Infinitives have accreted to most Danish modals so they can be used in a synthetic future: Jeg ville ikke kunne/turde sige sådan. [Or possibly for other reasons, I didn’t look it up, but grammarians of the 18th didn’t think an infinitive existed or was needed]).
In fact the cognate of turva is so thoroughly lost that it rang no bells for me, hence my WAG at uturvande. The form makes all sorts of sense once the base verb is called to mind, but it illustrates very well that that kind of Nynorsk is very far from being part of a common language with Danish.
Nødtørft m. “(bare) necessities” is rare bur unremarkable in Norwegian, perhaps kept alive by the more common nødtørftig a. “down to the bare necessities, minimal in scope or quality”. I would think that -tørft is from LG.
A remnant of the native word is found in tarvelig “trashy, crappy”, also from an original sense “(minimum) necessary”.
I forgot this:
Lars: The form makes all sorts of sense once the base verb is called to mind, but it illustrates very well that that kind of Nynorsk is very far from being part of a common language with Danish.
Yes, apart from the international linguistic terminology, the register is purist and archaizing, meaning “avoiding the German layers and Danish forms”. This removes it from Danish but not always from Swedish.
Maybe that actually makes it easier when you’re coming from West Germanic; I’ve now read the whole thing and didn’t understand every sentence, but most. I wouldn’t have guessed uturvande, though.
Notdurft f., extinct outside seine Notdurft verrichten “go pee” (highest possible register).
seine Notdurft verrichten “go pee”
It’s actually ambiguous to whether you do the small or the big business, it serves as euphemism for both.
There’s also an adjective notdürftig “provisional, makeshift, scanty (of clothing)”.
Norw. gjøre sin nødtørft can also be a euphemism for going to the toilet, but that would sound not only high register but very archaic. Think 19th century Bible translation.
Norwegian also has the noun tarv, which I would gloss as a neuter plurale tantum, but Nynorskordboka says feminine and Bokmålsordboka neuter or masculine. This probably shows that it’s a literary word in Bokmål (and for me).
The meaning is a little broader than bare necessity, more like “need, basic interest”.
tarvelig, of course. Also tarv ~ ‘best interest’ when public institutions have to protect barnets tarv. These two are not nearly as obsolete as nødtørft, every child knows tarvelig = ‘mean’ but tarv is restricted to that collocation which is however used a lot. tarvelige kår = ‘mean circumstances’ is historical, though.
There is also nødtørftig as in en nødtørftig reparation ~ ‘an emergency/sloppy repair’, but I’m not risking any opinions on whether YPNAD knows it.
I somehow forgot to ask: is båe the same as både, and do they both mean “both”?
Oh, yes. That’s a good one. Both Bokmål and Nynorsk split “both” as a quantifier (“both legs”) from “both” as a conjunction/list prompter (“both this and that”). Bokmål uses begge as quantifier (“begge bein”), while Nynorsk prefers båe (“båe bein”), but both of them have både as conjunction (“både dette og hint”). Archaic Nynorsk should regularly write the etymological but generally silent d in båe (quantifier), but usually doesn’t, probably out of need to discern the two words.
Now that I think of it, både with pronounced d is probably a reading pronunciation of written Danish. Anyway, all three forms go back on different cases of ON báði “both”, the cognate or source of Eng. both.
I need a copy-editor!
Most importantly, formatting:
I forgot to italisize “båe” near the end of the first paragraph.
I misclosed the italics on “d” in the last paragraph.
The rest is just vanity:
I forgot an a in “as a quantifier” in the first sentence.
I edited the second sentence to pieces before posting and ended up with “both of them has”.
Fortunately, there’s a copyeditor with magic powers hanging around! I think I fixed everything; let me know if you need more magic spells.
Thanks. There’s always another error, but no. I have to stand for them.
Like og “and” vs. òg “also”? And I guess lét is a disambiguation, too?
Curiouser and curiouser – both fits both báði and G beide, but those two don’t agree with each other!
The dental element in these words is a later extension in Germanic, as shown by Gothic; originally the word inflected similar to “two”. Norse on one hand and West Germanic OTOH must have selected different case forms for the extended variant (Norse has different forms in different parts of the paradigm).
(May I use this opportunity to link to my oldish paper on Hittite humant- and Vedic ubhau, where I summarize my thinking of the development of the IE word for “both”? Warning, it’s a link to academia.edu)
There are difficulties when you look at the whole paradigm, but case forms are indeed the traditional explanation, with Norse báðir being seen as extended from acc. báða, from *band-aid. English both doesn’t match the vocalism here, and must come from the nominative, *bai-þai (same for German). Worth noting that unextended bā, later bō is common in older English.
*band-aid
I assume that’s autocorrupt?
“*band-aid
I assume that’s autocorrupt?”
Argh, yes. Should be *bans-þans. Grr.
Because of the spedbarn (since we seem to be doing Norwegian here, and I rather like this word), I’m checking things on my phone more. This is only reinforcing my long held view that proper computers are far superior.
The acknowledgment section in your paper made me think you might be in Norway now. If you are, it calls for a belated welcome as well. And congratulations to whatever private/professional coalition that got you here!
Spedbarn is a good word, though almost completely replaced colloquially by baby [bæibi]/[be:bi]. There’s also nor, which I’ve never met outside of a crossword puzzle.
We called ours håndbagasje, at least for travelling purposes.
@trond
I doubt you carried a handbag, as these, like belly tops, are still mainly for women. King Charles may have inherited his late mother’s handbag, but I do not believe he makes use of it in public.
No, bagage straight from French. Hand luggage. ^_^
Because of the spedbarn
For those following along at home, spedbarn means (per Einar Haugen) ‘babe in arms,’ and the d is silent.
This is only reinforcing my long held view that proper computers are far superior.
We are as one in that matter.
I heartily agree. This does not mean that I live by my conviction.
For those following along at home, spedbarn means (per Einar Haugen) ‘babe in arms,’ and the d is silent.
The compositional meaning is “thin, weak” + “child”
sped stemme “faint voice”
spede spirer “new sprouts”
spedgris, -kalv “suckling piglet, – calf (on a menu)”
den spede begynnelse “the first beginning”
etc.
A friend from out west has been trying to get us to call our son a nurk.
I’m pretty sure I learned spedbarn from the pamphlets and websites we’ve been bombarded with as new parents. Now that you mention it, I don’t think I’ve ever heard someone say the word out loud, though I hadn’t picked up on that on my own.
(My wife got a job in Oslo a couple of years ago, and it is, by a stroke of very good luck, a permanent academic position. I’m here on a postdoc, and hoping to find something after that. The odds of finding a position in a particular city are, I know, rather slim, but I was about done with moving every three years for a new postdoc anyway. I may well end up graduating from academic flotsam to jetsam.)
Nelson: A friend from out west has been trying to get us to call our son a nurk..
Cute. I forgot about that. Literally “baby, very small child”, but the connotations of “cute, harmless” means that you can use it of kittens, or of adults so sweet and open and maybe a little chubby that you just want to tickle them under their chin and talk baby-talk. Or a small and charming car, for that matter.
It doesn’t strike me as regionally western. If anything, I’d put it in the Dano-Norwegian part of the lexicon.
Me: Literally “baby, very small child”.
But “literally” is not the point. Your friend’s suggestion is funny because it invokes the image of an old aunt looking into the pram while the baby is sleeping, and saying Se på det nurket! “Look at that (word for sweet little thing)!”
Heh. Haugen (writing in 1965) defines it as “1 manikin (= small, stunted person); small child. 2 deficiency, lack.”
Why isn’t it in Wiktionary?
Maybe it’s time to move from introspective to descriptive lexicography.
It appears that the dictionaries at ordbokene no know things I don’t.
nurk 1 n./m. “small, densely built person; small child”
nurk 2. m. “lack of force to thrive (in plants)”
nurket(e) a. “small and misgrown”
Norsk Ordbok has more or less the same words and meanings, but from older sources.
Could the “lack of force” meaning indicate a ne+”work” derivation?
I can’t find it etymologized, so that leaves it open for speculation. i like your idea, but…
Based mainly on the variant forms listed in the Norsk Ordbok entry above, I’ll suggest that it belongs in the sound-symbolic complex of words starting with gn-, kn- or hn- in Old Norse. Under the verb nurka “squeek, make small noises”, Norsk Ordbok also lists knirka/i> and knurka. The variant nurkla can also mean “work with small things”. The adjective nurkeleg refers to the variant nukkeleg “with halted growth”. Mod. Norw, gni “rub” (< ON gníða) has a variant form and alliterative doublet gnu, which in turn has the extended forms gnure and gnukke. gni has the extended form gnisse
I won’t try to work out this sound symbolic mess in detail, but nurk could be a deverbal formation. The verb nurka (> nukka) could be twice extended from a *(h)nu-, parallel to the forms with gn-. If so, the basic idea is “make friction”, and e.g. “squeek” and “slow down” would be secondary.
DWDS has
“Knirps m. ‘kleiner Kerl’
knirschen Vb. ‘ein mahlendes, reibendes Geräusch von sich geben’ (besonders mit den Zähnen knirschen)”
but does not connect them etymologically (also the “squeak” word is quietschen, not knirschen). Would the loss of initial k in the kn cluster be a normal development in Norwegian?
Nope. My suggestion is rather that some forms got a k- after the phono-semantic contamination party* got going. Others got new senses added. It probably started with a couple of inherited words and regular formations, but sorting out which ones may be impossible now. That’s what I meant by not working out the mess in detail.
* or clusterfuck, Pick your metaphor,
Also in the past participle gesessen /gsesːn/. Maybe the sequences /ɛsː/, /ɛst/ (except from unrounding) and /ɛks/ just don’t exist anymore.
Another vanished /ð/ in non-Danish, then. Både retains it in spoken Danish, and we have the same as Bokmåk with begge ben and både det ene og det andet ben.
(Both are båda in Swedish).
Amazingly overlooked above: größer has /ɛsː/, größt- has /ɛst/, but again that’s unrounding.
knirke is the normal Danish word for creaking like a floorboard does. A mouse piber. nørkle is doing detail work (with a slightly obsessive tint), and there’s no hint in the Sprachgefühl that those two might be connected. (And gnide means ‘rub’ in all senses I can think of, but the Norwegian side forms ring no bells).
@lars
Mäuschen, piep!
…it’s not there in “16” and “60”, it only just dawned on me this morning; they both have -[ɛçt͡s]-. Looks like there really must have been a ban on /ɛks/ once.