A verb meaning “to wage war” (SD/439), one of two examples of an derived-verb and the only one with any attested inflections. Carl Hostetter and Patrick Wynne initially suggested it is composed of the words azga “war” and a causative suffix -râ- (VSH/24), but later suggested (AAD/12) it may be related zagar so that its literal meaning was “wield a weapon” or “put to the sword”, similar to ᴹQ. mahta- “wield a weapon, fight” < ᴹ√MAK “sword”. Andreas Moehn also suggested (EotAL/ZAG’R) the verb may be related to zagar “sword”.
This verb appears in slightly different forms in each version of the Lament of Akallabêth, changing as follows: azgaranādu (SD/311) >> azagrāra (SD/312) >> azaggara (SD/247). All three versions have similar glosses, “was waging war” or “was warring”, which seems to imply these are all instances of the continuative-past tense. As they are the only clear instances of this tense, it is difficult to decipher the conceptual development of this verb.
The first form azgaranādu belongs to the draft Adûnaic period and is probably grammatically distinct from the later forms. Carl Hostetter and Patrick Wynne thoroughly analyzed this draft form (VSH/35). The two later forms appeared in very similar versions of the Lament. Rather than representing a revision of the grammar, they could represent distinct verb tenses, as for example Tolkien’s vacillation between the forms hikalba and hikallaba in the sentence Anadûnê zîrân hikalba. The question is, to which tense does each form belong?
Carl Hostetter and Patrick Wynne (VSH/28) concluded that the azaggara represents the continuative-past without analyzing azagrāra. Andreas Moehn agreed with them in his grammatical article on Adûnaic (LGtAG) but changed his mind in his etymological analysis (EotAL/ZAG’R), concluding it is the past tense instead. I also believe azaggara is more likely to be a simple past tense.
The double-g of the form azaggara resembles other verb forms identified as the past tense in the theories used here, and it seems likelier to me that Tolkien changed the verb from the continuative-past to the simple past without revising the gloss. The second a in azaggara is likely inserted in the verb stem to break up the consonant cluster in az-ggara, since Adûnaic allows at most two consecutive medial-consonants (SD/418).
If the above reasoning is correct, this would make the form azagrāra the best candidate for the continuative-past. Perhaps the repetition of the final suffix -ra is used to mark the continuous nature of the action. All of this is highly speculative, however, since we are working only from a single example.