I'm afraid there's no one resource available. There is Lehmann's introduction to PG which is useful for some things but it can differ quite a bit from other reconstructions.tezcatlip0ca wrote:Is there an online source where I can find Proto-Germanic endings reconstructed with final *-n and *-t and medial *-j- conserved, like we are using here? I found Orel’s "A handbook of Germanic etymology" on Gutenberg, but that just shows the lexemes, not the endings.
http://www.utexas.edu/cola/centers/lrc/ ... gmc01.html
The best thing to do is probably to use the Wiktionary reconstructions (mostly based on Ringe) and work backwards. For the stems, you can look them up in Orel and for the endings, comparison with PIE (also available at Wiktionary and Wikipedia) takes you a long way. The version of PG I've used above (which I call Proto Germanic Proper, PGP) is very close to Wiktionary/Ringe (Wkn/Ringe) except it lacks the following changes:
- Loss of final *n in unstressed syllables, leaving only nasalization of the vowel. In Wkn/Ringe nasalization is indicated by an ogonek on the vowel (ą etc.), PGP has final *n instead. Note that PIE final *m had become *n at an earlier time (before PGP), already before the loss of final *e and *a (often from *o). After the vowel loss, new final *m was created which remained as such, even to the present day in some languages.
- Loss of final *t in unstressed syllables. This introduced new final *n from previous *nt. In PGP, final *t and final *nt remains. Note that at some point before Grimm's Law, final *t had become *d (perhaps already in late PIE), which again became *t due to Grimm's Law. This is the reason for some alternation between *þ/*d and *t, for example in the the noun *alut (later *alu) but gen.sg *aluþiz.
- Loss of unstressed intervocalic *j, except in *ijV. However, *iji become *ī. This causes alternations such as *ija:*ī in the verbal paradigm. The loss is also notably found in the weak class II verbs with infinitives in *ōjanan. PGP retains *j in all these instances.
- Loss of *j before *i. This causes some alternation between *i and *ja. For example that Wkn/Ringe has *farjanþi:*fariþi for PGP *farjanþi:*farjiþi.
- I-umlaut of *e in stressed syllables, before a following nasal and before *i or *j in the following syllable. Due to umlaut, the diphthong *ei always became *ī in Wkn/Ringe. PGP does not include any of these umlaut effects (but it does include the shift *e > *i in unstressed syllables).
Other instances of final *t require a bit more thought. Comparison with PIE is sometimes necessary. But there aren't a lot of instances of final *t. The most important cases, I think, are the strong neuter nominative and accusative singular of adjectives and pronouns, which is *-at. There's also the third person singular of verb forms that stem from the secondary endings in PIE, like the present subjunctive *-ait, past subjunctive *īt and weak past *-dēt. There may also be the adverbal endings *-ôt and *-êt, if they are from PIE ablative endings as they are frequently thought to be. There are also a few *þ-stem nouns that end in *t > 0 in the nominative singular but *þ in other forms, like *alut, *aluþiz and *ilit, *iliþiz.
The loss of *j is even trickier. Comparison with PIE is of help for the nominal i-stem, where PIE *eye corresponds to PGP *iji, Wkn/Ringe *ī. In the verbal inflection, it's often easy to tell where the *j should be if you look at forms followed by a vowel other than *i. Also, Orel is of great help here as the *j is actually part of the stem. For the weak class 2 verbs, he writes *-ōjanan for example. Wiktionary has *-ōną for these verbs, which is actually different from Ringe who has *-ôną (or <ō̄ną> in his notation). I think all weak verbs of this classes used to have *-ōjanan so this shouldn't be a problem.
Non-umlauted words are found in Orel and Lehmann. Also, comparison with PIE is helpful. Generally, PIE did not have *i before nasals so Wkn/Ringe *in almost always corresponds to earlier *en. Also, since we're mostly using logograms anyway, and maybe not writing short vowels, this might not be particularly important except for glossing.
Apart from online resources, Don Ringe, From Proto-Indo-European to Proto-Germanic does tell a lot about where the lost consonants used to be in the descriptions of the development of PG. They're just not part of Ringes version of Proto-Germanic (which appears to me to be a bit anachronistic), so they're not found in the tables.
Ringe has a chart of the relative chronology of sound changes which is quite useful. It can be matched against Heikkilä's attempt at an absolute dating of some sound changes:
https://helda.helsinki.fi/bitstream/han ... dragti.pdf
(In Swedish, but even if you can't read it the list of sound changes should be understandable)
Another good source for the endings is Dirk Boutkan, The Germanic ‘Auslautgesetze’.
Both Ringe and Boutkan may be available on-line at places where you can find e-books.
Also, it may mostly be me who's using this slightly more archaic (and I think less anachronistic) version of Proto-Germanic. But since this is a writing system that's meant to be able to write all Germanic languages, it should actually be able to handle using different versions of PG.
The idea behind using <ᴛ> and <ɴ> is in part to make it easier to just use the Wkn/Ringe reconstruction. The rune <ɴ> can be thought of as just marking nasalization. The rune <ᴛ> still requires some thought but I imagine it can be dropped in the later languages, so it is mostly of interest for early Germanic.
Dealing with *j is trickier but I think there are som major advantages to writing *j that's been lost in Wkn/Ringe, because it makes paradigms more consistent, and it means that languages that analogically reintroduce *j don't have to change their spelling.