I think it’s confusing that we name the non-indicative tense-forms with the same terms as indicative tense-forms. For example “present indicative” and “present infinitive”. The word “present” doesn’t mean the same thing in both cases.
When there is a past/non-past alternation in Greek (e.g. imperfect/present or pluperfect/perfect), only one of the pair is possible in non-indicatives.
The reason for this is simple: only the indicative mood makes a past/non-past distinction. In other cases, only aspect is conveyed.
But this is undermined when we then go and choose for the non-indicative, “aspect only” forms the same terms that, in the indicative mood, are specifically conveying a non-past tense.
It would be far better to use a term with the non-indicatives that conveys only the aspect.
“Imperfective” and “perfective” are obvious choices instead of “present” and “aorist” respectively (although it’s not clear what we’d use for the perfect or future non-indicatives).
The same issue arises in discussion of “systems” and “stems”. Rather than the “present system” or the “present stem” should we instead talk about the “imperfective system” and “imperfective stem” in Greek?
If we use “perfective stem” rather than “aorist stem” we avoid the asymmetry of talking about an augmented/un-augmented aorist stem but not (or at least not without some awkwardness) an augmented/un-augmented present stem. (One might be forgiven for thinking Greek involves a morphological process of removing an augment if some descriptions of the aorist/perfective system are to be believed.)
Of course even in the above, there is the confusing use of terminology for what to call the bundle of aspect and tense.
Sometimes the bundles themselves are called “tenses” and the tense axis (as opposed to aspect) is referred to as “time”.
Sometimes the bundles are called “tense-forms”, which I think is better but still slightly confusing as that should really be “tense-aspect-forms” or, perhaps, “aspect-tense-forms”.
As an aside: the use of “form” is interesting as it places the bundling squarely in the realm of form, not meaning. In other words, even though the realization involves cumulative exponence (to adopt the terminology of Matthews), the meaning is just the union of the tense and aspect.
All of this plays into morphological tagging as well. I’ve suggested for the rethink of the parse codes in MorphGNT 7 that tense and aspect be split into two features.