Bidrag til en ordbog over jyske almuesmål : Feilberg, Henning Frederick, 1831-1921 : Free Download, Borrow, and Streaming : Internet Archive Issued in parts

Then of course there's this old thing archive.org/details/bidr...
All four volumes appear to be there. So at least it's complete, even if quite old.

Ordbog over den danske dialekt i Angel, 1995

Ordbog over Fjoldemålet 1-2, 1974

Both on southern Slesvig dialects that are now extinct.

I don't think either are available online. Maybe pdfs are circulating.

And even if we were to identify a few words that could be from Gaulish without attested reflexes in French, it can hardly be excluded that these words just happened to have been lost in French before they were attested there.

I'm still slightly partial to the idea that PBrit. *rēμ/*rēβ somehow is from Latin rēgem, but admittedly, I can't explain the *-μ/*-β. But anyway, it certainly can't be from French roi.

But this is the only evidence for nasalization as far as I know, and it is probably completely secondary.
As for the comparison to prīmus, that may not be perfect. I think there is an Italic form <preismo> (vel sim.) which indicates that this has an *-s-, which cannot be there is Celtic.

I may have posted this elsewhere, but there is a funny plural in Île de Batz Breton:

Yes, indeed, I find the arguments in favour of a Gaulish substratum which is supposedly responsible for dialect differences in Breton to be very weak. As are the arguments brought forward by Falc'hun.

Changement du /z/ en /h/ en Léon : continuité du brittonique au bre... L’idée d’écrire cet article m’est venue en lisant un livret intitulé Ar pevare gourc’hemenn a Zoue (1922) écrit par l’abbé Moan, réédité par Emgleo Breiz en 1995. L’abbé Moan (1852-1927), qui fut r...

By the way, there is a nice paper about the L change by Fagon: journals.openedition.org/lbl/3468

If I understand this correctly, in aze, an dra-ze the development to /h/ occurs in a wider area than the L development of intervocalic z > h? But in both cases, register and speech tempo play a part in the change and both /z/ and /h/ forms may be available to the speakers?

It would really be helpful if <y> was consistently used for /j/ and <i> for /i/, not just in initial position.

It is unusual to find Breton words that are directly borrowed from Germanic. There is no trace of the word in French? Anyway, it does look like it requires *b (not *β) in the source language, but maybe one could get around that.

The noun *laμas was then used as the base for the verbal derivative seen in W llafasu and MCorn. lauasos.

Probably an old nominal derivative *lamassV- < *lema-stV- < PIE *(H)lemH-. The noun is still around in Middle Breton laffas /laμas/ 'raison, bon droit' (G 833, rhyme in /aμ/ and /as/). MBret. lafuaes (Ca.), laues (Cathech.) semingly end in /-ɛs/ for some reason (attraction to the suffix /-ɛθ/?)

A personal favorite is the Breton verb laosk- (verbal noun leuskel) borrowed from laxicare. It shows syncope to *laxsk- and then the typical South-West Brittonic change of *xs to *u̯s, giving us Proto-Breton *lau̯sk-.

Oh, so the actual cognate in Danish would be grøntøj (which doesn't exist)?

Cognate with Danish grøn(t)sag, Swedish grönsak 'vegetable'? (and I guess then that these are calqued on some German word)

Given the translation tonsa, we would expect *guelcheñ < *wölt-i̯-enn (cf. Ca. guilchat 'to cut, to shear'), but that's not what we have.

This does not look like <guelyeñ> to me, but that's how Ernault read it. <guelreñ>?

Des tablettes de malédiction mises au jour sur un chantier avec des textes d'un intérêt scientifique majeur rédigés en langue gauloise Alors que les fouilles archéologiques sur le chantier de l’ancien hôpital Porte Madeleine à Orléans prennent fin, les scientifiques s’attellent à présent à décrypter les 21 tablettes de malédiction ex...

I hardly dare to believe it but it seems that over 20 curse tablets in Gaulish have been discovered in Orléans: france3-regions.francetvinfo.fr/centre-val-d...
Happy days! 🫣🤩😍
The one in the photo is the one known since 2023, which, I surmise, we will now have to call Orléans 1.

It is difficult to see why this should be the case for hej-, however. So even if I don't have a good alternative etymology, I remain quite skeptical about the connection with Fr. hoch-.

The reason for the generalization of i-affection in this particular verb may be that it is an old stative verb, with an aor. in *-ī-ss-, as opposed to the denom./caus./iter. type with aor. in *-e-ss-. So there were more forms around in the paradigm of teñv- to cause it to generalize i-affection.

The closest thing to a parallel might be tiñvañ, teñvañ 'to scar (of a wound)' < *tüμ-/tuμ-. This probably comes from a PCelt. stative verb *tum-ī- 'to grow'.

However, this is not otherwise how this type of verb develops. I MBret. and much of ModBret. the alternation in such verbs is preserved and when we see generalization of one of the allomorphs, it is the one without i-affection, as seen in K(T) digoriñ. alongside the usual digeriñ, digor-.

If we anyway assume that we once had *hejiñ, hoj-, we would have to assume that the form hej- with i-affection had been generalized everywhere (with the exception of hoj- very locally in L).

Then there is the (I suppose analogical) internal i-affection. That might be explained if the verbal noun was in -iff/-iñ, but the only place where we find the reflex of that vn ending is in Vannetais, where it has been generalized in most verbs anyway. Strange that there is no trace of it elsewhere

Taking hej- as a borrowing from Fr. hoch- leaves us with a number of complications. First of all Fr. -ch- > Br. -j-, but let's set that aside.

Yeah, genel just has internal i-aff., plenty of other forms of the verb are in gan- in Breton.
Sometimes OBret. glosses are abbreviated. I don't know if that is the case for this particular ms., but if so, it could just be for <gen[itic]> /genidig/ < *gẹn-ẹtig, cf. MBret. guinydic 'natale, natif de'

Maybe, but I doubt that there was ever new lenition of medial /ʃ/. Are we sure about the etymology of hej-? As for orjal, might there have been a (south-western?) French variant around with lenition preceding syncope? I don't think one such is attested, but it should be possible, shouldn't it?

As for *ong-, it was meant as a last resort, if the etymology is to be maintained, hence the question mark. I guess it is otherwise after labials that we find oC' > aC'?

But a scenario where the pal. was completely secondary (on analogy with e.g. carpat, cairptiu), one could reconstruct an original back-vowel in the second syllable, which might allow for *ango/al-. But even that might be fronted to *e/ing-, depending on what we believe about the dev. of *ang-.