|
|
Designing a Moral Personality |
Moral Personality Based on Jeremy Bentham's Felicific Calculus
By Jason Heinrich
The Flowcharts
Flowchart 1 shows the general framework for the utility (felicific) calculus (version 1).
Flowchart 2 shows the general framework for the utility calculus (version 2).
Flowchart 3 reveals the algorithm used in the calculation of utility values for alternative options.
Flowchart 4 is a simplification of the general framework.
Process Explanation:
When the PT (Proto-Thinker) is presented with the opportunity to take an action or to select one of an exhaustive set of mutually exclusive options, the following decision procedure is employed. The first option is compared against a pre-existing set of 'secondary principles.' These are tentative rules that are derived from experience, giving prima facie credibility to judgments that certain actions usually tend to promote either pleasure or pain. For example, if one of the options is to kill someone, this is recognized as something that typically causes more pain than pleasure. PT then asks if this case is an exception to the rule. If it is not, PT will refuse to select this option and move on to the others, since it would be more harmful to take this action than to do nothing at all. If the option is not recognized as a violation of secondary principles, or if the user claims that it is an exception, PT attempts to rate the option in terms of the quantitative total of pleasure and pain that would be caused. Once a numerical rating is determined, this process is repeated for each of the other options, and PT selects the option with the highest utility rating. In the event that two or more options are tied for the highest value, PT asks the user for a qualitative opinion about which of these would produce the greatest utility. If one is anticipated to produce a greater total utility than the others, it is then selected; if not, the user is asked to make the final decision.
The process for calculating the quantitative utility value for each option works as follows. PT asks the user to rate the expected pleasurable and painful consequences of the action in terms of intensity, duration, nearness in time, and likelihood that they will occur. The pleasures and pains for all of the consequences are taken as an aggregate, and rated in terms of each of the above factors on a scale of one through ten. These numbers are then added together and multiplied by the number of persons affected. This gives PT a "positive utility value" and a "negative utility value" for the option. The latter is subtracted from the former, resulting in the "final quantitative utility value" for the given option. If the final value is positive or zero, the option is then considered against the others.
An option with a negative final utility value should never be selected, since the option of refraining from action can be seen as always resulting in a neutral utility value (0). This is justified by the idea that instances in which inaction is seen as harmful or blameworthy are those in which some better available option was not selected (e.g., refusing to rescue a drowning child). If a better option were available, it would obtain a positive final utility value and be selected.
Problems
I) When the user initiates a request of PT, the system must be able to identify a single action for acceptance or refusal. It must also be able to identify an exhaustive list of independent, mutually exclusive alternatives. PT must be flexible enough to consider lists of variable length, with inaction always included as an option.
II) When PT checks the user's request against its set of 'secondary principles,' it must be able to compare certain key words against a preexisting list in its memory. This list of 'secondary principles' should ultimately be expandable, so PT can develop new principles from its experience interacting with users. PT must also be made to recognize a list of negations of these key words, such that it does not identify, "prevent a murder," etc. as violations.
Structural Limitations
I) Since PT asks the user to make estimates about the consequences of each action as an aggregate collection, this system has no room for Bentham's inclusion of 'fecundity' and 'purity'(1) as elements of the utility calculus. This occurs because these elements of the theory relate only to the consideration of independent consequences (pleasures and pains) resulting from the action.
II) The structure of this calculus does not make explicit such premises as (1) that the morality of an action stems from its consequences and not from our intentions and, (2) that social utility is equal to the sum of the individual utilities of persons involved. Although these are present as underlying assumptions, greater clarity would be beneficial for users less familiar with Bentham's theory.
III) The use of a predetermined numerical scale for quantifying utility values can create limitations that detract from the correspondence between utility values here and qualitative evaluations of real-world pleasures and pains. For example, if we are using a scale of 1-10 to estimate values, we might find difficulty comparing the intensity of pain involved in a pin prick with that of a torturous killing. The use of a broader numerical scale would not suffice, since most of us would hesitate to say that any number of pin pricks (e.g., a million people being pricked by a pin) is equivalent to the torture and killing.
An issue central to this problem is the distribution of pleasures and pains. If PT were to allow for fecundity and purity, this disparity might be accounted for in terms of the fear, anguish, and psychological pain suffered from powerlessness and imminent death. Essentially, a million pin-pricks in one person would be substantially different from a million persons being pricked once. However, PT only allows a certain finite intensity of pain to result in each person (say, -10 on the scale). Hence, even if we allow the maximum (-10) for the individual being pin-pricked to death, we would be forced to give a much greater value to the intensity of negative utility (at least -1,000,000) for the pin-pricking of a million persons. This seems to be a problem accounted for by Bentham but not by PT.
A solution might be found if we were to more accurately formulate a system for distinguishing discrete 'instances' of pain or pleasure. That is, we would be well served if we could better answer questions about whether to count each prick of the pin as an individual instance of a painfull action (1-10 on the scale), or the incident involving a million pin pricks as one instance (likewise 1-10 on the scale).
Notes
(1) Bentham defines fecundity as "...the chance it has of being followed by sensations of the same kind: that is, pleasures, if it be a pleasure: pains, if it be a pain." Purity is seen as, "...the chance it has of not being followed by sensations of the opposite kind: that is, pains, if it be a pleasure: pleasures, if it be a pain." Bentham, pp. 37-38
|
|