Act-Limited-Consequentialism
The central question for any system of normative ethics is the determination of the moral force of the course and events of human life. This inquiry is traditionally divided into the question of actions and the question of omissions (i.e. the failure or refusal to act). It is within this distinction we find the core problem with traditional consequentialism: that the moral weight of the consequences of acts and non-acts are treated the same. It is a fundamental presupposition of pure consequentialism, and indeed most forms of consequentialist thought, that the moral force of a course of existence lies solely in the result. The consequentialist will judge acts and non-acts with equal impunity. This means that the moral actor who follows a consequentialist ethical framework is not only obligated to act to bring about the desired outcome, he is equally obligated to act in order to avoid an undesirable outcome — for both types of undesirable outcome carry the same moral weight. This is best illustrated by the “trolley problem” that pits deontological frameworks against consequentialist frameworks. Here, the consequentialist moral actor will inevitably pull the lever and trade 5-deaths for 1-death, since the outcome of “failing to act killing five” is “five dead” and “acting to kill one” is “one dead”. The fact that his own volition kills one person is of no importance in this analysis.
However, the equal valuation of acts and non-acts, as one of the central tenets of traditional consequentialism, is a form of practical (and thus philosophical) shortsightedness. We presume, by their equal valuation, that the results of acts and non-acts are equally discernible to the moral actor entrusted to make this decision. In reality, it is not. Of course, this problem does not exist in a constrained moral exercise such as the “trolley problem”, where the totality of results of each moral act and non-act is well defined by the problem itself. This presumption of well-definedness, however, cannot be further from the truth. The potential consequences of actions and omissions are fundamentally different in terms of foreseeability. That is to say, consequentialists are no more omniscient than other moral actors. If we act based on a consequentialist framework, we can only act on the outcomes that we are able to foresee.
The consequences of omissions, or non-acts, are speculative in nature as they result necessarily from the mind’s model in response to the question “what may happen if I do nothing”. On the other hand, the consequences of acts are causal in nature: we are much more suited to predict the outcome of a deliberate action than to weigh the branches of probability stemming from a natural sequence, or an unknown moral agent. Our causal minds prevent us from exploring the consequences of non-action with the same degree of accuracy as we can predict the consequences of action. That is, in simpler terms:
Without knowing the causal chain that leads up to the moment of moral decision-making, we do not have the tools to predict what will happen next.
The precept of consequentialism, then, falls apart upon considering this important reality-constraint. Consequentialism as a moral framework depends on the moral actor’s ability to reasonably foresee, at the end of a causal chain, the intended and unintended results. Consequentialism is thus limited to those scenarios that the human brain is adequately equipped to simulate. If, for example, the ordinary human moral actor is given the choice to save a human being or an alien space ship, the moral actor is left without any possible way to forecast the consequences of choosing the spaceship. The alien technology may cure cancer, or it may just as likely destroy humanity. Here, the moral actor must either step away from the consequentialist framework, or be forced to make a judgment call between two unknown and unquantifiable consequences: a guess. The crux of this example, as it happens, is the same as that of any omission (or non-act): the naive observer is inevitably confronted with situations outside his/her sphere of mental competence and foreseeability.
In other words, when we make moral decisions on “what would happen if we don’t act”, we are reasoning by speculation. Away from the realm of the directly-causal, our brains quickly fall upon a patchwork of biased guesses. Even in the realm of speculation, we cannot help but to use causal thinking: the woman who hitherto always aced exams after tapping on the table will continue to tap on the table. Is this the type of outcome forecasting we wish to apply consequentialism to? Should she be given the moral agency to decide between a table tap that injures her neighbour and the speculative possibility that she will ace her next exam? I think not.
Moreover, the nature of the moral question of omissions imply a pre-judgment call: that it is a scenario that warrants moral consideration in the first place. One does not stand at his bedroom window debating whether it is right to open the window, because one does not feel the moral drive to make a judgment call. If he does not foresee potential moral consequences stemming from his act of opening a window, then he is not a moral actor, he is merely a semi-autonomous being acting on desire and self-interest. It is not so obvious that it may be possible for a moral actor to initiate the process of adjudicating the morality of inaction without first appreciating, and thus passing judgment on, a likely consequence of said inaction. The consequence of a non-action at the stage of moral assessment, then, must be presupposed to be incorrect. If one is given a gun and told to shoot the masked fellow in front of him, he does not hesitate to say no — unless, for example, the masked stranger is revealed to be Hitler. Only at this latter stage does the moral actor begin to weigh the consequences of killing a man versus letting masked-Hitler go free.
We then come to the perfect illustration of the foreseeability divide between actions and non-actions. Killing the masked man produces a causal-deterministic outcome: a person will die. Letting him go, on the other hand, sets in motion a series of possible consequences: the masked fellow may or may not be Hitler, masked-Hitler may or may not kill again, the people he kills may or may not be deserving of death (i.e. lack moral innocence in the framework of the decision maker). That is, the consequential chain of killing masked-Hitler is speculative and probabilistic in nature. The only causal consequence of the non-action (i.e. not pulling the trigger and thus letting masked-Hitler go free), in this particular example, is that a man does not die. If our troubled moral actor is a consequentialist, then the very consequence that drove him to this decision point (as opposed to the initial refusal to shoot) must be the mere possibility that masked-Hitler will kill and/or destroy if he is allowed to live.
It is true, then, that the class of scenarios that require a decision between action and non-action is necessarily one that pits a causal consequence against a speculative consequence: we have no certainty of outcome unless we are the causal agent for it. There is no great moral controversy when deciding between swerving to hit a person versus doing nothing to hit a tree. There is controversy only when, for example, the decision lies between staying the course to hit two and swerving to hit one (or vice versa). The causal consequence is deliberately aiming your vehicle at a living person with the intent and result of killing him. The speculative consequence is that by letting your vehicle continue forward, it may indeed kill two as predicted, or it may not.
The troubling difference is not the naive probability of each consequence, but rather, the subjective bias that we necessarily acquire in favour of action even in the face of causal uncertainty. We fail to appreciate our inability to compute probability chains and treat both outcomes as certainties. We feel morally obligated to act to cause the more desirable outcome. Not only does the lack of distinction between acts and non-acts in pure consequentialism regularly force the moral actor to speculate, it also forces the moral actor to choose action whenever his simulations show less desirable outcomes. When all is said and done, and the choice to act produced a “less undesirable” outcome, the moral agent will pat himself on the back and move onto the next problem — not realizing that it may very well be the case that if not for his action, an even more desirable outcome might have materialized. In an infinitely complex universe, it cannot be a good thing that our moral systems default to action. To use an old investing adage: “the market is a game where the less you play, the more you win”. The logic behind this statement is that every time you make an investing decision (i.e. buy or sell) you are introducing the potential for error. This logic is, I think, equally applicable to moral decision making.
Through the pure consequentialist view of equal moral weight in action and non-action, we set up a system where causal and speculative possibilities are taken to be of equal value. It may indeed be possible that causal consequence X and speculative consequence Y are given the same likelihood of materialization. This subjective assignment of likelihood and value does not, however, mean that X and Y are equally probable in actuality: we must also account for the inherent uncertainty found in speculative consequences. In other words, the speculative consequences are the product of a mind not designed to compute them. Both the primitive man and the modern man have a causal model of the universe: where the primitive man attributes all unfamiliar events to supernatural agents, the modern man attributes all unfamiliar events to natural causal chains (i.e. if I am aware of the result, I am biased to believe that I must also be aware of the cause. We are bad at dealing with random chance). This means that for the practical problem of moral judgment, we are not only more inclined to attribute causality, we are naturally more proficient at computing the causal consequences than the probabilistic consequences. It is no secret that even the best physicist struggle to understand the nature of quantum mechanics, while the dumbest layman is able to effectively use classical mechanics in daily life. We would much prefer to see electrons as billiard balls rather than probabilistic clouds: the first we can see and model, the second we can do neither.
All the foregoing discussion leads to this conclusion: that the presuppositions of consequentialism cannot be appropriately applied to outcomes of non-action. We lack the mental capacity to effectively predict, or foresee, the natural consequences that stem from leaving the course of events as is. Where conditions are such that we are unfamiliar with, or simply do not know, the causal precedents of an event, we can only (and poorly for that matter) speculate as to the range of possible consequences. This type of indeterminacy destroys consequentialism. A set of intrinsically speculative outcomes is no more useful to a moral actor than a kaleidoscope is to an astronomer.
The only appropriate course of moral judgment, then, must be one where action is weighed differently from inaction. For, if nothing else, actions allow for the direct causation calculation that permits the moral actor to make more accurate choices. The foreseen outcomes of actions must be given higher moral force than those of inaction. The gist of act-limited-consequentialism is this: consequentialism requires the accurate and efficient computation of consequences in real time by the moral actor, and the moral actor is incapable of doing so; thus, consequentialism can only be limited to the sphere of deliberate actions. One’s chosen metrics for consequentialist decision making (e.g. suffering, quantity of life, happiness, etc) should only be used to measure the potential proximate outcomes of one’s deliberate actions. The moral actor will remain morally innocent by merely refusing to bring about negative consequences by his/her own hand. Whatever potential consequences that stem from one’s inaction may equally be subject to moral scrutiny, but consequentialism cannot be the tool we use to examine these outcomes. Our concern, in the act-limited-consequentialism framework, is limited to the results we actively and directly cause to materialize.
The practical consequence of act-limited-consequentialism is that outside of constrained system such as the “trolley problem”, one’s only concern should be whether one’s direct action caused the world to become better or worse. We are not imbued with the moral imperative to forecast probabilities that stem for inaction. For example, it is sufficient for a moral actor to stop all activities that cause global warming (i.e. to refrain from acting to bring about global warming). It is not necessary or appropriate to further our consequential scope by considering the results of not actively returning global temperatures to the level of the 1900s. Doing so is speculative in nature, and necessarily implores action on a global geoengineering scale to induce global cooling on the faulty moral authority of “doing is better than not-doing”. The true consequences of this type of action is dangerously wide in scope, of which we neither know nor appreciate.
Act-limited-consequentialism specifically prohibits moral judgment on speculative grounds. Thus, it avoids the pathological consequentialist calculations that allowed for the murder of the few for “the good of the many” — the root cause of the tragedies of the 20th century. Those moral actors under Stalin and Hitler, to the extent that they were ethical non-sociopaths, necessarily believed that killing of Kulaks or Jews led to a better future for their entire nation. The very reason why this type of calculation is pathological, — that it pits a causal-deterministic result (e.g. 1000 killed) against a speculative-probabilistic result (e.g. 1 million benefited) — is the exactly what act-limited consequentialism is designed to prevent.
The true cause of the pathologies of our era is that we believe we have the moral agency and right to inflict suffering on our fellow humans, on nothing more than our belief that it is for the greater good. On what moral authority do we make this judgment call? Who must be called to answer to the loss of human life and dignity of the 20th century? To this question, the utilitarians and consequentialists are pitifully silent. This fatal moral agency is no different from what out philosophers and ethicists have granted us: the power to make moral judgments on nothing more than a wisp of possibility. Pathological indeed, is our ability to guess our way into the bowels of hell.
It is imperative that we recognize and internalize the conceptual difference between the outcomes of acts and non-acts: that one is definite and causal, and the other is speculative and inchoate. I implore you, dear reader, when the time comes to bear this difference in your heart and trudge on knowing that your actions have real consequences.
All my articles are dedicated to the public domain under the terms of the Creative Commons Zero licence. Please translate, copy, excerpt, share, disseminate and otherwise spread it far and wide. You don’t need to ask me, you don’t need to tell me. Just do it!
