Love for the other is manifest in the degree to which the consideration of the other’s good (and by “good” I mean the subjective “good-to-that-other”) factors into one’s decision-making. One can think of a decision as an optimization process: maximize Objective(d) subject to Constraints(d), where d is the decision variable of interest. The objective may be decomposed into additive terms, such as objective_self(d) for oneself and objective_other_i(d) for the ith other. Each term in the sum can in general be weighted by some scalar, reflecting the importance of the term to the overall objective. Without loss of generality, suppose that the weights are normalized by the absolute value of the weight with the largest magnitude. They will then lie in the interval [-1, 1]. We can calculate the entropy of the positive weights by normalizing the positive weights by the sum of all positive weights and applying the formula for Shannon entropy.

A measure of the love manifested in a decision-making strategy is a monotonically increasing function of the sum of the weights of sub-objectives considered in one’s decision-making and of the entropy of the positive weights. Note that the maximal measure of love in a particular decision-making strategy occurs when a weight of 1 is applied to the sub-objectives of all persons who ought reasonably to be expected to be affected by the decision. The relative evil of two decision-making strategies is the difference between the measure of love for each decision-making strategy.

This conception of love and evil is of course a model, and a general one at that. For example, who “ought reasonably to be” expected to be affected by an action is ultimately itself a moral judgment and depends on the information available to the subject. How exactly the measure of love increases with the sum of the weights and with entropy of the positive weights is also unspecified (beyond monotonicity in each). This model is meant to influence action rather than to rigorously exhaust all the nuances of love and evil. 

Putting aside its limitations for a moment, this model has some promising features. First, evil and love are manifest in the decision-making strategy, not in the decision itself or its consequences. Second, evil is inversely related to the sum of any positive weights. All else being equal, the more persons whose goods are considered and weighted positively in contemplating an action, the lesser is the evil. Third, evil is inversely related to the sum of any negative weights. The more influential the malevolent intentions, all else being equal, the greater the evil. Finally, evil is inversely related to the entropy of the positive weights considered. The more partial the distribution of good intent (lower entropy of positive weights), all else being equal, the greater the evil.

Note that decisions are not in general made consciously in the manner described above, but implicitly. To make a decision is not to obtain a single mathematically optimal solution to a well-formulated problem but is rather a process of optimization with incomplete information and possibilities of recourse, where the objectives and constraints shift fluidly in understanding. Nonetheless, this seems to provide a good enough model for love and evil that expresses some key factors: (1) love is manifest in the consideration of one’s understanding of others’ goods in the course of deciding action; (2) evil is a spectrum from impartial love for all who are expected to be affected by an action to total malevolence toward the same in decision-making; (3) evil may be born of malevolence (negatively weighting someone’s good in the overall objective), negligence (not reasonably considering who will be affected by an action), or partiality (favoring the good of oneself or particular others).

The foremost moral imperative is to seek moral imperatives. The value of adopting a provisional moral imperative is ultimately in the questioning that it prompts during the process of decision-making. The provisional moral imperative offered here, to be continuously undertaken over the course of a lifetime, is to maximize the sum across instances of some measure of love in one’s own decision-making. Adoption of this provisional moral imperative should entail self-questioning along the lines of, “Does my current decision-making strategy take into account the good of all people who will reasonably be affected? Do I really understand what is good-to-that-other, or should I seek more information in order to reformulate my objective? Am I unduly favoring particular people’s goods over others’? Am I acting out of malevolence toward any others?”

Suffering is an experience brought about by an inability to bring to fruition a desired state of affairs. Evil is a response to suffering insofar as it is to make a decision that is intended to increase the likelihood of the fruition of a desire without maximal and maximally entropic consideration of the good of others who are (or ought reasonably to be) expected to be affected. Perfect love is an ideal of universal consideration of all others’ good as one’s own; evil is a spectrum from minimally imperfect love to perfect malevolence. Equivalently, perfect evil is an ideal of universal malevolence, and love is a spectrum from minimal mercy to universal consideration of others’ goods as one’s own.