At the ACPA conference last month, Harrison Lee of Baylor delivered a paper (“Two Objections to a Broad Scope Theory of Intention,” which will appear in the Proceedings of the American Catholic Philosophical Association) raising objections to some Thomistic theories of intention, including my own. Lee argues that my claim that essential consequences of an intended action fall under the scope of the intention ends up reducing to the claim that statistically likely and thus expected consequences of an action fall under the scope of the intention. In this way the principle of double effect becomes useless, because one can no longer distinguish between intended and foreseen bad effects. (My account of intention was presented at a book panel discussion of Larry Masek’s Intention, Character, and Double Effect at the 2019 ACPA conference–available on the “Writings” page of this site–later revised as a book review in the ACPQ.)
I would resist the claim that on my account, the essential consequences of an action (those consequences which are per se in the 4th mode) reduce to those consequences that are statistically likely. Correlation does not imply causation, and a fortiori it does not imply essential causation. However, when a correlation is observed, it is reasonable to suspect causation. Correlation prompts one to investigate and look for the connection between two items. Perhaps the one is the cause of the other, or vice versa; perhaps they are both effects of the same cause; perhaps the one item is somehow concomitant with a third item which is the true cause of the other; etc. In the case of an action and its consequences, to establish essential causation (that the consequence is predicated of the action per se in the 4th mode), some sort of conceptual insight is required, by which one sees that the action, in virtue of what it is, produces that effect, even if the effect can be impeded by external factors.
At the time of my panel presentation, however, I had not sufficiently clarified in my mind the relationship between statistical likelihood and essentiality. Lee recognizes an inconsistency/weakness in my answer to one of Masek’s objections. Masek had brought forward an example of an adult son with a stutter who must testify on behalf of his father. He knows that he will almost certainly stutter as he testifies and thus weaken the persuasiveness of his testimony. Masek claims that it is unreasonable to suppose that the son intends to stutter just because he knows that stuttering is closely and inevitably joined to his intended act of testifying. In my presentation I claimed, unreasonably, that the son does intend to stutter, because “he knows that for him stuttering is essential to the act of speaking, for he always or usually stutters when he speaks.”
I see now that the son does not intend to stutter, nor is stuttering an essential consequence of his intended action of speaking on behalf of his father. One can see that the connection between speaking and stuttering is not essential even for him, for other people do not stutter when they speak, and it is clearly a result of an unusual condition he has. So the connection between his speaking and stuttering is mediated: it is a consequence of two things, his intention to speak coupled with his condition. It is a mere accident that the condition and the intention are coupled in his case, because in most people the condition is not present although the ability to speak is. This accidental conjunction of the ability to speak with the psychological or physiological condition produces a strong correlation between his speaking and his stuttering. But since the stuttering son does not will his own condition, the stuttering that it co-causes does not fall under his intention; it is not an essential consequence of his intended action. The stutter is a clear case of what Thomists would consider a natural defect, in which the natural effect is not well achieved. Thus the proper moral analysis of his situation is that he knows that he is risking stuttering. In fact he knows that it is highly likely that he will stutter, and he must perform a cost-benefit analysis of this unintended effect compared to the benefits that could be achieved by acting. And the cost-benefit analysis comes out in favor of speaking.
A similar analysis, with the opposite conclusion, would apply to another example Lee brings up in this context: that of a man who commits manslaughter by driving home drunk. He does not intend to kill when he gets in the car, nor is killing an essential consequence of his action. The action is one of driving, and driving does not in and of itself tend to kill human beings. However, the circumstance of drunkenness significantly increases the likelihood of accidentally killing, and thus the drunk man is knowingly taking a big risk of harming people. The cost-benefit analysis comes out strongly against taking the risk, and thus it is highly irresponsible to take the risk. If the drunk man fails to fulfill his intention of driving home safely, he has committed manslaughter-unintentional but reprehensible killing–but not murder–intentional killing.
I am grateful to Lee for his criticisms, as they have helped me to see what was wrong with my presentation. What I previously said about the stuttering son example likely produced a misunderstanding in the minds of many about the nature of my position. I am also interested to see where Lee is going with his project.