Medical Treatment, Genetic Selection, and Gene Editing: Beyond the Distinction Between Person-Affecting and Impersonal Reasons

Abstract

According to what McMahan and Savulescu (2024) call the “popular position”, embryo selection is less ethically problematic than gene editing (other things being equal). The Two-Tier View, defended by McMahan and Savulescu, implies that the popular position is mistaken. The authors treat gene editing of embryos similarly to standard cases of medical treatments that promise expected benefits for the (subsequent) person even though gene editing also may create risks of harmful side effects for her. McMahan and Savulescu assume that if gene editing is (successfully) done, it is better for the person who developed from the beneficently edited embryo. And, if the editing had not been done, although it was possible, that would have been worse for the same person in question. Thus, the comparator must always be a possible, even if unlikely, world in which she would have existed. That is why gene editing, in their view, resembles medical treatments. Therefore, assuming that standard medical treatments are not more ethically problematic than embryo selection, they conclude that (in general) gene editing should also be treated as not more problematic than embryo selection.

In this paper, I am taking one step further than McMahan and Savulescu in questioning the central assumptions in the bioethical debates about new reproductive technologies. In the previous work, we argued that the very distinction between person-affecting and identity-affecting interventions is based on a questionable form of material-origin essentialism (Żuradzki and Dranseika 2022). Here, I argue that the bipartite distinction between person-affecting and impersonal reasons, as used by McMahan and Savulescu, insufficiently represents different possible counterfactual comparisons between real and merely possible outcomes. In particular, I argue that the evaluation of gene editing requires counterfactual comparisons to situations in which the benefited person might or might not exist. Such evaluation requires an additional class of reasons, let me call them “semi-person-affecting,” that cannot be reduced to either person-affecting or impersonal ones. This aspect of gene editing makes it dissimilar to standard cases of medical treatments.

Indefinitely Many Alternatives

The main objection against the McMahan and Savulescu view, discussed in section 4 of their paper, derives from a simple observation that when an act is done, there are usually many (or even indefinitely many) ways in which it could have been done otherwise, and thus many alternative outcomes of events that might have occurred instead of the actual course of events (see discussions in Sparrow 2022; Douglas and Devolder 2022 in the context of reproductive choices, and for a more general approach in the context of consequentialism, see Hájek 2024). Moreover, in some of these possible worlds, the allegedly benefited person may exist and in some, she may not exist (say, because no person develops from the embryo or—given identity-affecting interventions—a different person develops from the same embryo). This is the main problematic factor for the disjunctive distinction between person-affecting and impersonal reasons. Thus, it is not clear what appropriate counterfactual comparator should be used to check if, in a particular situation of gene editing, we deal with person-affecting reasons or not. In contrast to standard cases of medical treatments in which a patient exists as a person (at least for some time) no matter if treatment is conducted or not, this is particularly problematic in cases of possible gene editing scenarios, when had gene editing not occurred, a defective embryo could have been, for example, discarded and thereby not have developed into a person. Moreover, as noticed by Douglas and Devolder (2022) the more serious the disorder, the more plausible it is that the afflicted embryo would be discarded, and thus there would be, in their view, no person-affecting reasons for gene editing.

I will discuss two answers to this objection. Both of them implicitly assume that there are modal facts about what would (not simply could) have happened in any given situation instead of the actual course of events. First, the above assumption is accepted by Douglas and Devolder (2022) who claim that it is sometimes possible to find the relevant baseline counterfactual scenario for determining the presence (or absence) of person-affecting reasons: “in some cases, editing out disease does [italics in original] benefit the edited child, because the child would [italics added] otherwise have existed and been afflicted by the disease” (Douglas and Devolder 2022) (even if the embryo could have been discarded). They conclude that only in situations when a person would otherwise have existed we may assume the existence of some person-affecting benefits from gene editing.

The second answer, pursued by the authors of the target article, assumes that what matters is not what would have been otherwise, but what could have been otherwise. They exemplify their approach with the hypothetical situation in which an embryo has a genetic disorder, but it was implanted without gene editing, even though the procedure was possible. However, if it had not been implanted, it would have been discarded by the parents who had an objection to gene editing, but not to discarding an embryo. In such a situation, the child who developed from such an (unedited) embryo may have a justified complaint against her parents: “Who cares… what they would have done if they had not implanted the unedited embryo from which I developed. What matters is that I have got a terrible disorder and they could have enabled me to be without it” (McMahan and Savulescu 2024, italics original). And in the opposite case, the authors claim that “whenever gene editing has successfully been done, that was better for the person who has developed from the edited embryo even if the embryo would [italics-TŻ] have otherwise been discarded rather than implanted without editing” (McMahan and Savulescu 2024).

Gene Editing and Counterfactuals

However, both above-discussed answers to the question of what is the relevant baseline counterfactual scenario for determining the presence (or absence) of person-affecting reasons are problematic.

From the ontological perspective, both views seem to assume “counterfacts realism” (Hájek 2024) which states that “There exist primitive modal facts that serve as truth-makers for all counterfactual claims”. Moreover, the target article assumes that some of these modal facts are represented by what would and some by what merely could have happened in any given situation instead of the actual course of events. This is visible in the above-mentioned example of the parents who had an objection to gene editing, but not to discarding an embryo. In this case, on the one hand, the authors of the target article clearly state what would have happened, if the parents had not implanted the unedited embryo (“they would have discarded the embryo”). On the other, they assume that a truth-maker that provides the relevant baseline counterfactual scenario for determining the presence of person-affecting reasons is a modal fact that parents could have edited this particular embryo that developed into a child with a genetic disorder, and they could have implanted it after editing for better.

Nevertheless, the distinction between what would and what merely could have happened is problematic, and counterfactual questions like “What would have happened had the parents not employed gene editing?” generate answers (e.g.,” The embryo would have been implanted as it was;” “The embryo would have been discarded;” etc.) which are often treated as false (Lewis 1973; Hájek 2024) or at least indeterminate (Stalnaker 1968). Under this interpretation, such answers resemble sentences like “Alice had an odd number of hairs when she first spotted a White Rabbit” in Alice’s Adventures in Wonderland (cf. Hare 2011). They do not have any truth-makers at all, because they are not in the business of having a truth value. Therefore, in contrast to what D&D assume, in the case of gene editing, there may be no fact of the matter about the comparator to person-affecting benefits or harms. McMahan and Savulescu are in this matter in a better position because their more liberal view assumes that it is enough that the comparator is represented by mere possibility (“Parents could have edited this particular embryo…”).

However, from the epistemological perspective, if Douglas and Devolder and McMahan and Savulescu’s arguments are to have any practical implications, they must also assume that an agent who evaluates the state of the affairs from an ex-post perspective has knowledge about counterfactual scenarios. (i) Either” what would otherwise have been at a particular time”, i.e., she has knowledge about the particular possible world that would have existed if the actual course of events had not occurred (Douglas and Devolder 2022). (ii) Or “what could otherwise have been at a particular time”, i.e., she has knowledge about all not actualized worlds, or at least those worlds that are relevant from the evaluator’s perspective, that could have existed if the actual course of events had not occurred (McMahan and Savulescu 2024). However, neither Douglas and Devolder nor McMahan and Savulescu provide a detailed description of evidence relevant in the context of knowledge about counterfactual situations. Douglas and Devolder, in order to establish what would have happened, limit their remarks to the degree of severity of the genetic disorder which may be “edited out” and to the evident characteristics of prospective parents who are “generally more motivated to avoid severe genetic disorders in their child than less severe disorders” (Douglas and Devolder 2022). The authors of the target article, who are in an easier position since they need to assume merely what could have happened, do not mention the problem of evidence, just assuming some possible ways potential parents could have acted if gene editing were possible.

Beyond The Distinction Between Person-Affecting And Impersonal Reasons

Both these points are crucial for the ethical evaluation because the Two-Tier View defended by McMahan and Savulescu accepts that there is a difference in strength between person-affecting and impersonal reasons. Therefore, it must also assume that the strength of all things considered reasons in the case of gene editing of embryos depends on events that otherwise merely could have happened at some particular time. In other words, editing out disease can benefit the edited child (i.e., there are person-affecting reasons to do it), only if the child could have otherwise existed, no matter how unlikely it would have been. The practical conclusion from this analysis is that gene editing of embryos (when the very existence of a person depends on the type of intervention) is not as similar to standard medical treatments (when the patient exists as a person no matter if intervention is undertaken or no) as the target article assumes. The moral evaluation of gene editing requires an additional class of reasons, semi-person-affecting, that cannot be reduced to either person-affecting or impersonal. This aspect makes gene editing dissimilar to standard cases of medical treatments. Consequently, McMahan and Savulescu’s attack on the “popular position” is not successful in this regard.