Ian Benjamin Hollenbaugh
Ancient Greek prohibitions show an asymmetrical paradigm compared to their corresponding imperatives:
Positive Negative
present imperative μή + present imperative
aorist imperative μή + aorist subjunctive
This paradigm raises two major problems that have received no satisfactory solution. The first is formal: Why is the aorist subjunctive used in prohibitions rather than the aorist imperative? The second is functional: It is usually said (e.g., Smyth 1956:410–1) that the present imperative construction (PIC) regularly refers to present time (“inhibitive,” type stop doing that!), while the aorist subjunctive construction (ASC) refers to future time (“preventive,” type don’t do that!). Yet this distinction has many counterexamples (Louw 1959). Thus a functional question must also be addressed: Why is aspectual contrast between the present and aorist not maintained in the prohibitive construction?
From an Indo-European perspective, the prohibitive marker *meh₁ (> μή) seems not to have occurred with modal forms but was restricted to the augmentless preterite forms (called “injunctive”) used prohibitively and, in Vedic Sanskrit, the aorist was strongly preferred to the present in prohibitions (Avery 1885). In Greek, however, there was a functional merger between the indicative preterite forms and the “injunctives.” When this happened, the original way of marking prohibition would no longer have been admissible, as it would now involve what was functionally an indicative preterite. To remedy this, the prohibitive aorist injunctives were replaced by subjunctives wherever this would not affect the meter (Stephens 1983). In all 10 examples of the ASC in Homer the corresponding injunctive would not affect the meter. In the three cases where the meter would have been affected by changing to the subjunctive, the injunctive remains (formally identical to the imperative).
Putting this together, I answer the first question posed as follows: The aorist was already the form for making prohibitions before historic Greek, so when the injunctive was lost it was the natural locus for substitution with the injunctive, to the exclusion of the present. The aorist with μή is thus an archaism with a morphological update. The newer—and accordingly much more productive—formation is the PIC. Once μή became free to combine with modal forms in Greek, a natural way to express negative commands would be to simply negate the imperative with μή. Because the archaic ASC already occupied the aorist “slot,” the present was the only imperative form actually used—the aorist imperative being categorically blocked by the ASC. Once the ASC and PIC coexisted, they were reinterpreted as the present and aorist counterparts to the positive imperatives (Willmott 2007:107).
As for the functional issues, comparison with Sanskrit again reveals that the aorist in inhibitive use was an archaism rather than an anomaly. Hollenbaugh (fthcm.) has demonstrated that the aorist injunctive in Sanskrit was used inhibitively and that no categorical functional difference between the present and aorist injunctive existed (contra Hoffmann 1967). This aspectual neutralization of the aorist in prohibition was simply carried over into Greek.
Thus, Greek’s apparently bizarre prohibitive system, while historically accidental, is not inexplicable.