This paper endeavors to formalize imperatives that convey actions. Imperatives, unlike propositions, do not hold the value of true or false. Peter Vranas proposed an alternate logical formalism in the literature of imperative logic with three values, namely: Satisfaction (S ), Violation (V ) and Avoidance (A). Although this formalism takes into account the conditional imperatives, it does not address imperatives from the perspective of actions. According to Mīmāṁsā (one of the six systems of Indian philosophy), the prime motive of an imperative is to carry out action so as to reach the goal. This paper modifies the notion of Vranas' imperative logic using Mīmāṁsā principles. The resulting imperative formalism, focusing on action representation, is presented with syntax, deduction rules and semantic tables. The effectiveness of this representation is demonstrated in the area of artificial intelligence planning.