Re: Morimarusa

From: dgkilday57
Message: 65657
Date: 2010-01-15

--- In, "Torsten" <tgpedersen@...> wrote:
> --- In, "Brian M. Scott" <BMScott@> wrote:
> >
> > At 9:17:36 PM on Wednesday, January 13, 2010, dgkilday57
> > wrote:
> >
> > [...]
> >
> > > On an unrelated matter, does Danish <pande> 'pan' require
> > > a protoform (inherited or borrowed) with -nd-, or can it
> > > be borrowed from a garden-variety Low German <panne>?
> >
> > It's <pænnæ> in the oldest Danish mss. (ca. 1275-1350),
> > according to Ordbog over det Danske Sprog
> > (
> >
> > As <panna> it's in a Norw. document of 1366 from Tolga in
> > Hedmark, and SAOB lists fornsvenska <panna>; it's an old
> > borrowing in WGmc., but it looks as if NGmc. probably got it
> > from MLG.
> >
> The -nd- is an artifact of Danish spelling rules. All Germanic languages have a problem with representing long vs. short vowels and solves it by writing double either the vowel or the auslaut consonant; Danish, because of the lenition of stops after long vowel, does not have that problem with syllables with stops in auslaut; a large part of the remainder of short vowel roots have -l- and -n- in auslaut, and since the pronounciation of original -ld- and -nd- has merged (via -ly- and -ny-) with original -ll- and -nd-, that spelling tended to spread to roots where it was unoriginal, since that way Danish spelling could avoid a general rule to write double the auslaut consonant, eg. 'mand' /man?/ "man", like 'land' /lan?/ "land". In other words, never trust a modern Danish -ld- or -nd- to mean anything other than -ll- or -nn-.

Thanks to you and Brian for this information. My guess is that the WGmc word comes immediately from Gaulish, since <panna> (apparently acc. pl.) is attested at La Graufesenque. But I speculate further that Gaulish may have borrowed NWBlock *panda: 'balance-pan', a deverbative related to Latin <pendo:> 'I hang, weigh', in the generic sense 'pan'. It could then be connected with WGmc *pandam 'pledge, security, guaranty' (OFris <pand>, OHG <pfant>, etc.) if the latter is borrowed from the NWB for '(fixed) weight, (fixed) value' vel sim. The form *pandingaz 'coin having fixed weight, standard coin' would then have been created in WGmc using the productive coin-suffix *-ingaz. Rather than postulating additional protoforms *panningaz, *paningaz (cf. the OED s.v. <penny>) I prefer to take the /d/-less variants as due to folk-etymological influence from 'pan' in the individual WGmc languages. OE <pening>, the favorite form in West Saxon, creates some difficulty since 'pan' itself is only cited as <panne>, <ponne>. But the compound <panmete>, <ponmete> 'cooked food' suggests an old variant *pann (like <binn> beside <binne>, wk. f. from Gaul. *benna), with the geminate regularly reduced from *pannmete. I am not sure that I can argue that *pan- would have been extracted from such a compound early enough to antedate OE /i/-umlaut, however, which is required to get <pening>.

If WGmc *pandam and Gaul. *panna both come from NWB deverbatives of *pend-, the larger issue is whether NWB shifted inherited */o/ to */a/ and vocalized syllabic resonants with */u/, like Gmc. (which Kuhn seems to favor implicitly), or retained */o/ and vocalized the resonants with */a/, like Celtic. In principle NWB *pandam, *panda: could have either zero-grade or /o/-grade. If zero-grade, we might expect some Gaulish loanwords from NWB to have /o/. This suggests a very tentative explanation of Treviran Latin <pottus> 'potter'. North Gaulish *pottos could perhaps continue a NWB *podnos, with the assimilation occurring within Gaulish, and the NWB protoform built regularly on *pod- (OE <faet> etc.). This would avoid the rather uncharacteristically irregular shenanigans proposed by Hubschmid to link 'pot' with 'vat'. But more detailed examination of NWB phonology is required.

Another relevant word here is OE <palstr> 'spiked shepherd's staff' vel sim. which I consider borrowed from a NWB deverbative of *pel- 'to drive (as sheep)'; the literal sense would be 'driving implement (for sheep), shepherd's goad'. It is unclear at this point whether we have /o/-grade or zero-grade, but I lean toward the latter. If that is correct, the possibility that <Morimarusa> is NWB instead of unshifted Proto-Gmc. (or Celtic) must also be considered.