Moral considerations often affect reasoning about facts in the real world, clouding the judgments of both scientists and non-scientists. The elementary psychological processes that underlie this phenomenon are important to uncover. The experimental evidence presented in Knobe's target article, however, does not illuminate these underlying judgments.
Consider first the scenario in which a profit-maximizing individual A chooses an action that harms versus helps the environment, and a majority of subjects say the harm was intentional but the help was unintentional. Is there a disagreement concerning the facts among decision-makers? Almost certainly not. For instance, all subjects might agree with the assertion that A foresaw the effect of his decision on the environment and did not factor in this effect in deciding upon his action. All subjects must agree with this, in fact, because the description of the situation says precisely this. It follows that attributing intentionality in one case and not the other is not a judgment of fact, but rather a moral judgment. The experiment then shows that moral judgments affect other moral judgments, which is not a contested assertion.
One might object that attribution of intentionality is a factual statement concerning an individual's mental state, and sometimes indeed this is the case. For instance, we might conclude that after copulation, an insect may “intentionally” feed himself to his mate, or that the prey may “intentionally” reveal his awareness of the predator to the predator. In such cases, we are saying that it is a normal part of the behavioral repertoire of the organism to engage in this act even when the organism has the capacity and the information to behave otherwise. But intentionality has a distinct second meaning that lies clearly in the moral realm. We say a undesirable result of an individual decision is “intentional” if the individual foresaw the result and could have prevented the result and achieved all other effects of the decision, except that doing so would have incurred additional personal cost.
Intentionality has yet a distinct third meaning, also in the moral realm. We say a welcome result of an individual decision is “intentional” if the individual foresaw the result and acted to bring about the result at personal cost.
When a subject says that the harm was “intentional,” it is most plausibly the second meaning that is being invoked. When a subject says that the help was “unintentional,” it is most plausibly the third meaning that is being invoked. We rule out the first meaning of “intentional” in these cases – because this meaning is strictly factual, whereas the context of the situation calls for a moral evaluation.
The interpretation of this evidence is complicated by the fact that there are several other commonly used meanings of intentionally, one being “foresaw the result and acted in order to achieve this result.” In this sense, profit-maximizer A did not intentionally harm in the first scenario and did not intentionally help in the second. Very likely, many subjects chose to use this definition, despite the fact that it renders the choice completely trivial, as the statement of the problem includes non-intentionality overtly in the description of the situation.
The Gricean analysis of meaningful communication is relevant here. According to Grice (Reference Grice, Davidson and Harman1975b), in normal conversation, a listener assumes that when a speaker solicits information, the speaker expects the information to be useful to the speaker. Thus, if someone asks, “Is there a washroom on this floor?” acceptable answers include “Yes, down the hall on the right,” or “Yes, but it is out of order; there is a working washroom the next flight up,” or “You'll have to go across the street.” A simple yes or no would be considered a somewhat bizarre answer. In the current case, some of the common usages of the word “intentionally” are explicitly assumed in the statement of the situation, so a Gricean subject can supply useful information only by referring to those usages of the terms that require some sort of substantive inference. These usages are the second and third ones defined above.
Related problems of the multiple meaning of words beset Knobe's causation analysis. Consider the scenario of the philosophy department receptionist and the taking of pens. The question as to whether the professor, the administrator, or both caused the problem is not a matter of fact. The facts are laid out quite clearly in the statement of the scenario, and would be agreed upon by all. The notion of “cause” in question is not that of Newtonian mechanics, but rather systems theory or product design. To see this, let us change the scenario a bit, to a machine that needs a certain level of motor oil to prevent seizure:
Half the oil is devoted to a mechanism that burns 10% of its oil allotment each day, the lost oil being replenished at the start of each day. The other half of the oil is devoted to a mechanism designed to burn no oil at all. At the end of one day, the machine seizes up and it is determined that the first mechanism consumed its allotted 10% of oil, but the second mechanism consumed an additional 10% through a malfunction.;>If asked whether the first mechanism, the second mechanism, or both “caused” the failure, the correct answer is the second.
There is here, of course, no factual dispute and the inspectors are making no moral judgments in placing blame on the second mechanism. In general, when a complex mechanism fails, blame is placed on elements that failed their designed tasks, even if in some sense their behavior according to Newton's laws was exactly the same as other elements that performed as designed.
You have
Access
Moral considerations often affect reasoning about facts in the real world, clouding the judgments of both scientists and non-scientists. The elementary psychological processes that underlie this phenomenon are important to uncover. The experimental evidence presented in Knobe's target article, however, does not illuminate these underlying judgments.
Consider first the scenario in which a profit-maximizing individual A chooses an action that harms versus helps the environment, and a majority of subjects say the harm was intentional but the help was unintentional. Is there a disagreement concerning the facts among decision-makers? Almost certainly not. For instance, all subjects might agree with the assertion that A foresaw the effect of his decision on the environment and did not factor in this effect in deciding upon his action. All subjects must agree with this, in fact, because the description of the situation says precisely this. It follows that attributing intentionality in one case and not the other is not a judgment of fact, but rather a moral judgment. The experiment then shows that moral judgments affect other moral judgments, which is not a contested assertion.
One might object that attribution of intentionality is a factual statement concerning an individual's mental state, and sometimes indeed this is the case. For instance, we might conclude that after copulation, an insect may “intentionally” feed himself to his mate, or that the prey may “intentionally” reveal his awareness of the predator to the predator. In such cases, we are saying that it is a normal part of the behavioral repertoire of the organism to engage in this act even when the organism has the capacity and the information to behave otherwise. But intentionality has a distinct second meaning that lies clearly in the moral realm. We say a undesirable result of an individual decision is “intentional” if the individual foresaw the result and could have prevented the result and achieved all other effects of the decision, except that doing so would have incurred additional personal cost.
Intentionality has yet a distinct third meaning, also in the moral realm. We say a welcome result of an individual decision is “intentional” if the individual foresaw the result and acted to bring about the result at personal cost.
When a subject says that the harm was “intentional,” it is most plausibly the second meaning that is being invoked. When a subject says that the help was “unintentional,” it is most plausibly the third meaning that is being invoked. We rule out the first meaning of “intentional” in these cases – because this meaning is strictly factual, whereas the context of the situation calls for a moral evaluation.
The interpretation of this evidence is complicated by the fact that there are several other commonly used meanings of intentionally, one being “foresaw the result and acted in order to achieve this result.” In this sense, profit-maximizer A did not intentionally harm in the first scenario and did not intentionally help in the second. Very likely, many subjects chose to use this definition, despite the fact that it renders the choice completely trivial, as the statement of the problem includes non-intentionality overtly in the description of the situation.
The Gricean analysis of meaningful communication is relevant here. According to Grice (Reference Grice, Davidson and Harman1975b), in normal conversation, a listener assumes that when a speaker solicits information, the speaker expects the information to be useful to the speaker. Thus, if someone asks, “Is there a washroom on this floor?” acceptable answers include “Yes, down the hall on the right,” or “Yes, but it is out of order; there is a working washroom the next flight up,” or “You'll have to go across the street.” A simple yes or no would be considered a somewhat bizarre answer. In the current case, some of the common usages of the word “intentionally” are explicitly assumed in the statement of the situation, so a Gricean subject can supply useful information only by referring to those usages of the terms that require some sort of substantive inference. These usages are the second and third ones defined above.
Related problems of the multiple meaning of words beset Knobe's causation analysis. Consider the scenario of the philosophy department receptionist and the taking of pens. The question as to whether the professor, the administrator, or both caused the problem is not a matter of fact. The facts are laid out quite clearly in the statement of the scenario, and would be agreed upon by all. The notion of “cause” in question is not that of Newtonian mechanics, but rather systems theory or product design. To see this, let us change the scenario a bit, to a machine that needs a certain level of motor oil to prevent seizure:
Half the oil is devoted to a mechanism that burns 10% of its oil allotment each day, the lost oil being replenished at the start of each day. The other half of the oil is devoted to a mechanism designed to burn no oil at all. At the end of one day, the machine seizes up and it is determined that the first mechanism consumed its allotted 10% of oil, but the second mechanism consumed an additional 10% through a malfunction.;>If asked whether the first mechanism, the second mechanism, or both “caused” the failure, the correct answer is the second.
There is here, of course, no factual dispute and the inspectors are making no moral judgments in placing blame on the second mechanism. In general, when a complex mechanism fails, blame is placed on elements that failed their designed tasks, even if in some sense their behavior according to Newton's laws was exactly the same as other elements that performed as designed.
ACKNOWLEDGMENT
I would like to thank the European Science Foundation for financial support.