Reconstruction talk:Proto-West Germanic/fali

PGmc *felaz

Hi Victar! When I added the etymology to (deprecated template usage) felon a long time ago, I didn't know as much about Pgmc as I do today. Thanks to a lot of help from the many knowledgeable editors/admins here, I reworked the original reconstruction *felaz to *faluz. As you can see, Old Frisian fal and the jumble of disassociated forms in Old English (fel, felo, fæle), could not have come from *felaz; but only from *faluz. Same for the Danish and MHG forms. I didn't know if you wanted to relink to the new form for PGmc or not. Leasnam (talk) 23:48, 27 May 2013 (UTC)

There is also the Old Dutch fel (“cruel, base”), which you might want to add : ) Leasnam (talk) 23:53, 27 May 2013 (UTC)

Oops, my bad! It appears I never fully created the PGmc entry! Let me do that real quick Leasnam (talk) 23:55, 27 May 2013 (UTC)

OK, done. Hope this is right. Any insight you can give would be very helpful. Koebler does have *felaz (or more appropriately, *fēlaz by the descendant ON and MHG forms: ON fála "witch"; MHG vālant "imp"; but I think these may only be distantly related to *faluz above. ? Leasnam (talk) 00:14, 28 May 2013 (UTC)

You're right, I'd say it was u-stem as well. Thanks Leasnam! --Victar (talk) 04:08, 28 May 2013 (UTC)

Latin *felus ?

@Victar I have no clue how one can reconstruct a Latin *felus while the Old French oblique case of 'fel' is originally 'felon'. Exarchus (talk) 09:37, 22 July 2024 (UTC)

And if you want to point to forms like 'fels'/'feus', these obviously were created by analogy when 'fel' and 'felon' started to become independent. From what I can see at Godefroie's dictionary, those forms first occur in Benoît de Sainte-Maure's Roman de Troie (1160-1170). Exarchus (talk) 10:10, 22 July 2024 (UTC)

A similar case is Old French ber, oblique baron: there also occurs a nominative singular 'bers', for example in Les quatres livres du roi, a bible translation from 1291-1294. No need to reconstruct a Latin *berus because of this.

Given that you want to reconstruct *falljō as ancestor of Latin fello (and I can't exclude this possibility), also reconstructing *felus simply complicates things.

And now I just realised that Latin *fĕlus would result in Old French nom.sg. 'fiels', oblique 'fiel'. Sorry, but I'm really baffled by mistakes like this. I think one would have to reconstruct *falus or *falis to get to 'fel' (/fæl/), disregarding the problems pointed out above. Exarchus (talk) 23:35, 22 July 2024 (UTC)

I think you're missing is that this is a reconstruction for a Vulgar Latin adjective, so your counterexample using baron is not relevant. Incidentally, there should also probably be a separate reconstructed entry for the noun, with Medieval Latin fellō being a reborrowing from Old French fellon. @Nicodene --{{victar|talk}} 02:58, 23 July 2024 (UTC)

It's a surprise to me as well, but the adjective does seem to have the -on inflection pattern, according to the data gathered by the DEAF, with nominative singular fel attested from ca. 1000 and nominative singulars with [-s] from the middle of the twelfth century onward.

Fello appears to be documented as early as the year 858, per {{R:Niermeyer|fello|412}}, which is slightly before the cut-off I use for inheritance into Romance, namely ca. 900. That choice of date is admittedly arbitrary (first Latin–Romance distinction attested in 813, and I leave the rest of the century as a grace period). Nicodene (talk) 04:40, 23 July 2024 (UTC)

Strangeness aside, only the noun is attested in Latin, not the adjective, whence Old French fel, Occitan fèl (and also Italian fello?). --{{victar|talk}} 06:49, 23 July 2024 (UTC)

According to {{R:it:TLIO|fello|024540}} the Italian fello was borrowed from French. That it was so is not clear to me since, according to the same source, fello is documented in Italian since the thirteenth century. That is when Italian itself was first documented, in any substantial way.

Perhaps the Latin record could be of some use. Here are the quotes from Niermeyer arranged in chronological order:

• Non tibi sit curae, rex, quae tibi referunt illi fellones atque ignobiles - epist. synod. Caris. a. 858.
  • “Do not concern yourself, oh king, with what those fellones and nobodies report to you.” – Epistola synodi Carisiacensis ad Ludovicum regem Germaniae directa (Quierzy, France, 858).
• Nolit ministrum vel capellanum habere, quem fillonem non audet appellare. - Rather., Qual conj.
  • “He would rather not have a minister or chaplain whom he dares not call a fillo.” - Ratherius' Qualitatis conjectura cujusdam (Verona, N. Italy, 965). Notably, the author came from what is now Belgium. I have excluded two other quotes from him.
• Fillones illi fugitivi. - Ekkehard., Cas. s. Galli
  • “Those fugitive fillones.” - Ekkehard's Casus sancti Galli (St. Gallen, Switzerland, ca. 1047–53).
• Dampnant fellones, cruciant furcisque latrones - Donizo, V. Mathildis
  • “They sentence the fellones and crucify the brigands.” - Donizo's Vita Mathildis (Canossa, N. Italy, ca. 1112–5).

That isn't much to go on at all but does seem consistent with a north–south trajectory. Nicodene (talk) 09:37, 23 July 2024 (UTC)

@Nicodene I am wondering whether the variation "fillonem" in the second quote might be etymologically significant. Ratherius does use 'fello' in the other two quotes though.

And now I see Ekkehard uses 'fillones' too. I don't exclude it's a hypercorrection though. Exarchus (talk) 12:06, 23 July 2024 (UTC)

"According to {{R:it:TLIO|fello|024540}} the Italian fello was borrowed from French. That it was so is not clear to me"

Where would it otherwise come from? You'd have to reconstruct some Latin *fellus, which doesn't explain the declension of Old French and Occitan (as Diez also adds "prov.", see quote below). Pianigiani's dictionary gives (for what it's worth) that it comes from the Provence. Exarchus (talk) 14:37, 23 July 2024 (UTC)

@Exarchus Italian is quite capable of inheriting Latin nominatives. See Appendix:Survivals of the Latin nominative in Romance for examples, including some doublets with or without a final -one. Nicodene (talk) 19:14, 23 July 2024 (UTC)

The page does mention the question of the -one suffix though. But then it could still be a 'de-derivation' from fellone, so not necessarily from French. I'd have liked to know the reasoning of the Dizionario etimologico italiano, but can't find it Exarchus (talk) 19:36, 23 July 2024 (UTC)

If you do manage to find a way to access the Dizionario Etimologico Italiano online I’d love to know. That’s proved something of a white whale for me. Nicodene (talk) 21:54, 23 July 2024 (UTC)

@Nicodene: So how might one reconstruct a Proto-Gallo-Romance n-stem adjective, if indeed that's what the OF and Occitan are, like a present active participle? --{{victar|talk}} 21:44, 27 July 2024 (UTC)

The noun and the adjective are declined the same way, so you want to propose one etymology for an adjective declined 'fel' (nom.), 'felon' (oblique), and then you propose another etymology for a noun declined 'fel' (nom.), 'felon' (oblique)?

Seriously? Exarchus (talk) 21:51, 27 July 2024 (UTC)

If you want a very simple description of things, look here: "felon adj. et n.m. cas rég., fel, cas sujet". Exarchus (talk) 22:05, 27 July 2024 (UTC)

And if you want to reconstruct for some later stage like proto-Gallo-Romance, what's the point of using an -us ending? Exarchus (talk) 21:57, 27 July 2024 (UTC)

@Victar I would probably lemmatize it as *fellonem or *fillonem, with the inflections based on early Old French fel^nom.sg/felon^obl.sg/felon^nom.pl/felons^obl.pl. Incidentally it would be interesting to see how it declined in Old Occitan and when it was first attested there.

The Old French felon was also (albeit less commonly) a noun, like the attested Latin forms. That being so, I'm not sure we'd actually need a PGR reconstruction here. Also, and I should probably have mentioned this earlier, the first Latin attestation has the word juxtaposed with an adjective:

• illi fellones atque ignobiles
• those^nom.pl fellones and ignoble^nom.pl

It wouldn't be enough for me to add “adjective” as a part of speech on the Latin entry, but it is perhaps worth considering. Nicodene (talk) 22:36, 27 July 2024 (UTC)

@Nicodene: According to Winter (1909), the Old Occitan is felon (obl.sg.), fel (nom.sg.), but it doesn't give any examples. --{{victar|talk}} 23:15, 27 July 2024 (UTC)

Interesting, thank you.

From a quick look at the DOM entry (its first cited source), it's attested as felh in the writings of Marcabru, which is quite early. Nicodene (talk) 23:41, 27 July 2024 (UTC)

A present active participle on -ans would result in Old French nom.sg. -anz, oblique sg. -ant. The idea that you're going to invent some etymology here which everyone the last few hundred years has missed, for such a basic word, I find that mindboggling. Exarchus (talk) 00:16, 28 July 2024 (UTC)

LMFAO, you're too much, my guy. And you wonder why you end up talking to yourself in discussions. --{{victar|talk}} 00:46, 28 July 2024 (UTC)

You suggested a derivation for fel/felon from something "like a present active participle". I'm pointing out to you that Latin active participles become -anz/ant in Old French, a far cry from fel/felon. You could have known that too by doing some very basic research. Latin participles end on -ans or -(i)ens, so how you would have gotten the vowel correct is already unclear. (Or did you perhaps mix up with Ancient Greek participles on -ων/-οντος?) Just to add: the -anz in the OF nominative developed by analogy with adjectives like grant; a paradigm with a regular phonetic outcome is the noun enfant, with nom.sg. enfes (< infans). Exarchus (talk) 12:46, 29 July 2024 (UTC)

Given the declension, the adjective came from the noun. Exarchus (talk) 11:49, 23 July 2024 (UTC)

By the way, and I think I'll leave it there, the noun ber/baron I mentioned also developed a use as adjective, with forms like 'ber' and 'barons' being used, see the Anglo-Norman dictionary Exarchus (talk) 18:40, 23 July 2024 (UTC)

Yes it isn't uncommon for a noun (at first used appositively) to slip into an adjective. This was the case for instance with various Latin nouns in -ator and still to some extent French nouns with -eur (cf. poseur). The oddity in felon is the dominance of the adjective. Nicodene (talk) 23:02, 27 July 2024 (UTC)

Diez clearly states that the original declension is: 'fel' nominative, 'felon' oblique: "Die ursprüngliche declination im prov. und altfranz. ist nom. sg. fel (fels), acc. felon (so durchaus in der Passion Christi und im Leodegar), der nom. felon ist selten und eine spätere verirrung."

But it's irrelevant as *felus simply doesn't work phonetically. Exarchus (talk) 10:23, 23 July 2024 (UTC)

Also, Godefroie states about 'fel': "adj. et subst." Exarchus (talk) 10:42, 23 July 2024 (UTC)

And about this hypothetical Latin *falus or *falis that I gave above to maybe get to Old French 'fel', it wouldn't work for Occitan. So only something with double -ll- would do. Exarchus (talk) 15:35, 23 July 2024 (UTC)

See also the Anglo-Norman dictionary entry for 'felun': "nom.sg. often: fel, fels, felx, feu, feul, feus, fiel"

Note that 'fel' (in the sense we're talking about) is simply a cross-link to 'felun'.

Yes, the entries for Old French 'fel' and 'felon' could probably use some further clarifications. But I'm not really sure at which point they become really independent (still in Old French period?). The reason why Godefroy treats 'fel' and 'felon' separately (most dictionaries don't) is probably because he's reasoning that 'fel' was sometimes [but not in the early period] also used as oblique case, justifying a separate entry. Exarchus (talk) 16:45, 23 July 2024 (UTC)

Latin 'fello' / reconstruction of Latin *felus

Given that maybe you have pings turned off (didn't know that was possible), I'll link to two discussion pages just in case you haven't seen them:

- fello about the misleading claim that a borrowing from Frankish '*falljō' is a common etymology

- *fali about the dubious reconstruction of Latin *felus Exarchus (talk) 10:57, 22 July 2024 (UTC)

PS: if you say *fillō or *fillijō is a bad reconstruction for Latin fello, that might be completely correct, but then I'm interested to hear your argumentation, e.g. as 'reconstruction notes' to *falljō Exarchus (talk) 13:45, 22 July 2024 (UTC)

I'd like to ask why you give *fellō as alternative reconstruction at *falljō, while the TLF reference clearly gives *fillo (whether this is likely or not is not my point). Exarchus (talk) 07:30, 24 July 2024 (UTC)

I think you might be missing my edit summaries. See diff. --{{victar|talk}} 07:32, 24 July 2024 (UTC)

I did miss that, but then I would add that as reconstruction notes. Exarchus (talk) 10:25, 24 July 2024 (UTC)

@Victar I want to come back to this reconstruction *fillō, which according to you should have become *fellō because of a-umlaut. At first I simply believed you on this and assumed you'd know what you are doing. But I've been looking at this more closely and it appears that i > e before nonhigh vowels was not so regular at all. The Wikipedia page shows that u > o was fairly common, but i > e was "rare" (last paragraph).

When you look at the reference given (Campbell's 'Old English grammar', page 43), you'll see he gives the three Old English examples that exist, namely nest, wer and speċ (a variant of spiċ), and 'speċ' might simply be an error, see Ringe, The Development of Old English, page 34. Then Campbell gives 5 examples from Old Frisian, but also indicating that there are variants with 'i', as you can check in the Altfriesisches Handwörterbuch: leth ~ lith, quek ~ quik, etc.

Further on the Wikipedia page, a theory by Lloyd is mentioned, who proposes an alternative explanation for these i > e cases.

But let's look at more recent sources. Firstly, in 'From Proto-Indo-European to Proto-Germanic' Ringe says on page 79: "OE wer exhibits an unexpected lowering of PGmc *i to e". And about OE nest: "the lowering of the vowel in OE is puzzling" (p.100). Suggesting that he doesn't see a-mutation of i to e to have occurred at all.

Then in 'The development of Old English' Ringe describes the a-mutation for u > o starting at page 27 (in the 'first proof' text that I could find on the web). At page 34 he mentions the lowering i > e: "In Norse there appears to have been a comparable lowering of *i to *e (Noreen 1923: 53-4), but no such general lowering occurred in WGmc. Only two words with inherited *i, both with solid PIE etymologies, appear with e throughout the WGmc area [i.e. PIE *wih₁rós and PIE *nisdós]" ... "The reasons why these two words consistently exhibit e remain unrecoverable."

He further says: "However, 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" (so *fillō would not qualify for this change). Then about Old Frisian (note on p.35): "The lowering of *i to e in OF exhibits a very different pattern: a majority of examples are found before coronal consonants, and most of those that are well attested actually exhibit variation between i and e [...]. It is difficult to escape the conclusion that this is a separate OF development".

So assuming a general a-mutation of i to e is completely unsupported. Maybe those dictionaries suggesting *fillō were not that ignorant after all...

I can only conclude (as how am I supposed to conclude anything else?) that your knowledge about all this (and now apparently also about Proto-Germanic, which I honestly assumed you would know rather well) is only superficial. Maybe it's time to considerably raise your standards.

pinging some Proto-Germanic folks: @Eruditum @Phillipm0703 @Wuduweard @Jaktfalk @JohnC5 @Mannrinn @Rua @TheSilverWolf98 @Caoimhin ceallach @DerRudymeister @Leasnam @Wiljahelmaz @Xenos melophilos See this specific post in the first place, but then also the rest of this page. To be clear: don't trust me, but verify for yourself (and preferably double-check). It shouldn't be my job to be pointing out these errors, as I don't pretend to be an expert in this anyway. Exarchus (talk) 12:52, 29 July 2024 (UTC)

Now I see that you linked to the *noun* felle in your edit comment, not to the adjective. But felle is simply the expected form of the nominalized adjective fel. That there's also a noun variant fel can be explained by being a very close borrowing from French 'li fel'. Given that the only early Middle Dutch attestation of this variant occurs in West Flanders (so within France back then) suggest this too.

edit: I have to correct myself here, the noun variant doesn't need to be related specifically to French, it can simply come from the nominative masculine singular in the indefinite form, so "een fel" (like in the attestation). In Modern Dutch the form would need to be with -e added. Exarchus (talk) 09:39, 25 July 2024 (UTC)

Looks as though I sent you the wrong link. See https://gtb.ivdnt.org/iWDB/search?actie=article&wdb=VMNW&id=ID72133. --{{victar|talk}} 17:40, 25 July 2024 (UTC)

Yes, but that was the link I at first thought you sent me and here there is no discussion, 'felle' is simply the definite form of the nom.masc.sg. Exarchus (talk) 17:47, 25 July 2024 (UTC)

Please explain what you think "definite form" means. --{{victar|talk}} 18:00, 25 July 2024 (UTC)

Like in German: after the definite article, "der große Mann". The definite form is any case how it's called here on Wiktionary, see fel (although 'feller' should be 'felre' but I have yet to find a grammar which gives a rule for this). Exarchus (talk) 18:59, 25 July 2024 (UTC)

I thought you would be familiar with this as Proto-Germanic also has strong and weak declension. Exarchus (talk) 19:02, 25 July 2024 (UTC)

The only remnant in Modern Dutch is for the neuter singular: een groot huis vs. het grote huis Exarchus (talk) 19:19, 25 July 2024 (UTC)

By the way, I've been looking at this grammar for Middle Dutch from the University of Ghent, but they're making a mistake: they give the example dat paert is clein, but clein should be cleine. The site of the Univ. of Ghent seems pretty sloppy, as I've seen them making an ignorant mistake about the Brabantian dialects (it's here, still not corrected), namely that 〈oe〉 is always pronounced long, while there's obviously a long and a short /u/. There has been one paper about this here, ironically from Ghent University. But Brabantian is generally underresearched. This as a rambling aside. Exarchus (talk) 20:05, 25 July 2024 (UTC)

Hmm, it can actually be "clein", but this is not the main form, so not a good example.

I'm improving the adjective declensions btw (using this). Exarchus (talk) 09:38, 26 July 2024 (UTC)

What you should basically look at in the VMNW for a real variant is the predicative form ('pred.'), for example at vroet you have 'pred. -t, -de', so 'vroede' can be given as variant of 'vroet'. Exarchus (talk) 07:08, 26 July 2024 (UTC)

Although even here 'vroede' can be interpreted as a declined form (in this case agreeing with a plural noun), as the reference given says (§141). But in principle the predicative is undeclined. Exarchus (talk) 07:42, 26 July 2024 (UTC)

@Exarchus: Yes, I am familiar with definitives, but not all forms built from the weak stem are definitive forms -- some are just simply variants as a result of leveling. See examples like *flugi and *lugi. Unless you can demonstrative that all those occurrences of felle are used in the definitive, those are examples of the indefinitive as well. --{{victar|talk}} 21:01, 27 July 2024 (UTC)

@Victar The main thing is that as the predicative, only 'fel' is given. But in theory, if 'felle' would be used after 'een' and before a masculine or neuter singular noun, yeah okay. But of course I checked this before, and there is no such example in the quotes at VMNW or MNW (hint: 'een felle strijd' is a modern translation of 'een battalie fel'). And MNW does not give 'felle' as variant, as there are no variants here. Exarchus (talk) 21:44, 27 July 2024 (UTC)

Bet.

Daer hi versach .. een casteel ..; daer woende een felle gast, een roese groet ende starc^[1]

Een felle strijt began daer doe^[2]

--{{victar|talk}} 22:28, 27 July 2024 (UTC)

Those quotes I hadn't seen, so on that point I stand corrected, but I have to add that these can perfectly be seen as declined forms as adding -e before a masculine noun is simply the standard form in later Dutch. That development has to start somewhere.

Stoett has this to say about it: "§112. De invloed der zwakke verbuiging is ook hier wederom merkbaar. Vandaar dat men nu en dan een nom. sing. masc. en nom. acc. neutr. op e ontmoet, doch meestal in het latere Middelnederlandsch; na het pron. poss. geschiedt dit eveneens en zelfs dikwijls in goed Mnl.: een gedaechde man; een bloedige strijt; een hebreeuwsche wijf; sijn goede halsberch; mijn jonge lijf; enz." Exarchus (talk) 23:03, 27 July 2024 (UTC)

Just adding that the tendency to add -e in indefinite position in later Middle Dutch is exactly not what you'd expect if this was some kind of reinterpreting the adjective as having a basic form on -e, because there the later tendency is to get rid of final -e. Exarchus (talk) 00:24, 28 July 2024 (UTC)

a speculation about the Old English compounds

I have a specific idea about the compounds ælfæle, ealfelo, wælfel, but it might be too speculative to put on the main page.

The thing is that the two occurrences of ælfæle/ealfelo are both with attor (“poison”) ("Ealfelo attor" / "Attor ælfæle"), suggesting that this was a fixed expression, while the simple 'fæle/felo' had died out (partly because of homonymy with other terms?).

The term 'wælfel' ("slaughter-cruel") strikes me as a rather poetic term, which could then be a kind of pun on the above ælfæle/ealfelo.

At least an interesting hypothesis I'd think. Exarchus (talk) 14:58, 24 July 2024 (UTC)

Dutch ± inheritance very doubtful

@Victar From what I now understand, PWG *fali would result in Middle Dutch 'vele' /veːlə/. That's three phonetical issues if you want to get to 'fel'.

Further, reconstructing the noun 'felle' (where there is indeed a variation 'fel') as being separate from the adjective strikes me as artificial, as adjectives can be nominalized in Dutch at any time. If you look at the first attestation of the noun 'felle' (with VMNW giving "Gesubstantiveerd bnw." btw), you'll see that it is juxtaposed to exactly that: a nominalized acjective. The quote is: "die ungetruwe end die felle".

But the most interesting thing about that first attestation is that the text (the Limburg Aiol) is a (free) translation of a French chanson de geste. The Middle Dutch text can be found here (line 454), the Old French text (from a later manuscript) here. The relevant episode (the bad guy Makaires wanting to drown two children in a river) is at about line 9198 in the French text, and then again retold by the saviour of the children at about line 10520. The Dutch 'die ungetruwe' apparently translates 'li traitres' (e.g. line 10519) and 'die felle' translates 'li fel' (e.g. line 10520). It's difficult to ask for a clearer indication of a borrowing. Exarchus (talk) 07:38, 25 July 2024 (UTC)

Middle Low + High German would normally also have kept a final -e in the adjective. So I think the professional thing to do is to just keep the English and Frisian branches and leave out the rest. Exarchus (talk) 11:05, 25 July 2024 (UTC)

By the way (to also say something positive), I find the link to *frafal(i) (Dutch wrevel etc.) genuinely interesting, I'd need to have a closer look to really evaluate it. Exarchus (talk) 11:35, 25 July 2024 (UTC)