Disscussion agentive

Agentive vs assocative
The historic Germanic languages were at a dearth for marking overt verbal agency, commonly marking verbs with the n-stem individualizing suffixes which also marked general association.

The "primary" and most productive way to mark a verbal agent was with the preterite zero-stem (where available) form of the verb with the individualizing n-stem, therefore: In some verbs, this might even be enlarged with an associative yod-stem element, giving the suffix *-j-an-, cf. Gothic arbinumja vs Old English irfenuma, both "inheritance-taker; heir". This could cause confusion with secondary verbs and homophonic stems where the suffix merely mark association, compare: It is therefore desirable to distinguish these semantic roles in some way, either by finding an alternative agentive suffix, or by more overtly marking association with an alternative suffix or suffixes.
 * *bītaną, "bite" > *bit-an-,"biter"
 * *nemaną, "take" > *num-an-, "taker"
 * *burg-an-, which could be understood to mean both "preserver" < *berganą, "to preserve" and/or "burgher, resident of a city" < *burgz, "city"
 * *burg-j-an-, which could be understood to mean "preserver" < *berganą; "one who keeps safe; protector" < *burgjaną, "to keep safe, protect"; "burgher, resident of a city" < *burgz, "city"

Transpositive or inherited suffixes
The Germanic language family seems to have lost the PIE suffixes *-tḗr and *-tōr, which were the primary markers of verbal agency. The only unequivocal remnant is the Samic loanword (Northern Sami) áittar, by which we can reconstruct a North Germanic *aihtār > Proto-Germanic *aihtēr, "owner, possessor". From this we can judge that Germanic preserved at least *-tḗr, and at least in the paradigm of a preterite-present verb.

Less clear are a group of formations in *-þrijan-: *burþrijan-, "bearer"; *murþrijan-, "murderer"; *rōþrijan-, "rower"; *spinþrijan-, "spinner; spider"; possibly also *wurhtijan-, "worker, wright", if dissimilated from *wurhtrijan-. If these aren't simply n-stem derivatives of deinstrumental verbs, *burþrą > *burþrijaną > *burþrijan-, then they could have been original tor-stems transformed and transferred to a more productive class: *-þōr, oblique stem *-þr- > *-þrijan-. Although the fact that the first three verbs in this group all have attested instrumental formations, *berþrą, *murþrą, rōþrą, rather must lead us to the conclusion that these should be conservatively interpreted as such. However, small counterpoints to keep in mind for this notion are that:
 * The instrumental formation of *beraną, "to bear", is *berþrą, not *burþrą, and a derived verb would be *birþrijaną, not *burþrijaną, a verb which isn't attested anyway.
 * The attested occurrences and semantics of the word-family consisting of *murþą, *murþijaną, *murþrą, *murþrijaną are hazy in understanding at best, the root means simply "to die", not "to kill":
 * *murþą < *mŕ̥-tom, "death; murder" is best explained etymologically as derived from *mr̥-tós, "dead", with contrasting accent, which gives an original meaning "that which is dead" > "(dead) body".
 * *murþrą is possibly the only instrumental *-þra- formation in Germanic which refers to an act, the seemingly only other occurrence is *berþrą, "child-birth; fetus", although in this case "fetus" would have to be presumed to be the primary meaning, "that which is carried".

Transposed IE *-tōr/*-tor-
As generally reconstructed for PIE, Proto-Germanic should have inherited the agentive tor-stem declined as follows:

If Germanic had done as Italic, or at least Latin, and leveled the nominative stem -tōr- throughout the paradigm, we would expect:

On the other hand, if Germanic had done as Ancient Greek, it would have simply leveled the strong case stem -tor- throughout. In this case, the forms would match our regularized declension of an-stems near perfectly, leading us to additionally change the nominative -þōr to -þār to match the vowel color of the scheme, as done with the n-stems:

The latter is possibly the most appealing, as the o > a change and our specific "regularization" of -ō- to -ā- give this suffix a rather peculiar and unique "Germanic-ness" to the suffix.

Do or *-dōndz
One innovative option for an alternative suffix, and one with which there is at least some realistic basis, is the present participle of *dōną, "to do". As Germanic present participle formations were very late somewhat productively used to form agent nouns, it is perhaps appropriate that "doing" > "doer" could become an attractive marker for agency. In compound with the zero-grade, Verner-voiced action nouns of the type: *dribą, "driving", *flugą, "flying"; ... this would give the forms driba-dōndz, "he who does driving > driver"; fluga-dōndz, "he who does flying > flyer".

This presents a simple mono- and bi-syllabic element to overtly mark the agent of a verb. The feminine form of the present participle would be *dōndī, plural *dōndijōz, giving eg. numa-dōndī, "she who takes; taker".

In turn, taking the thematic vowel present in this formation into consideration as if akin to the thematic vowel in the past participle of a normal strong verb paradigm, that is: numadōndz vs *numanaz, one could see a logical stepping stone to introducing the suffix into the weak verb paradigm: *tawjaną, "to make", past participle *tawidaz = tawidōndz ... which gives us the following stem + suffix formations:
 * Class 1: -idōndz, -idōndī
 * Class 2: -ōdōndz, -ōdōndī
 * Class 3: -aidōndz, -aidōndī