Chuck is a self-made monster. He spent years hardening himself, on purpose, into someone who enjoys killing, and one afternoon he murders his neighbor George without a flicker of reservation. Beth is the sweetest person you could meet, until a team of neurobiologically sophisticated intruders performs a "personality swap" while she sleeps. She wakes with her memories intact and Chuck's values fully integrated, and that afternoon she kills George with the same untroubled resolve (the case comes from the indomitable Al Mele). Stipulate that Chuck and Beth are psychologically identical at the moment of the killing. Most of us say Chuck is responsible if anyone is, and hesitate about Beth. That hesitation is the wedge between two camps. Structuralists hold that responsibility is fixed by your psychological structure at the time of action: your values, cares, and responsiveness to reasons. Historicists hold that structure isn't enough; it also matters how the structure came about. Chris Cho takes the hard line in a new PPR paper : Beth is responsible, full stop. The defense has two steps. First, deflation: the kind of manipulation that best serves the historicist leaves Beth's new values stable and propagated and an agent like that already has everything historicists wanted. Second, debunking: whatever intuition survives is an anchoring bias that fixates on pre-manipulation Beth, "without a normative justification."
The deflation meets its hardest case when Beth wakes already stable and propagated, with no conscious rationalization anywhere in the story. Where, then, was her "reasonable and fair opportunity to filter" the new values through her character, as historicism demands? Cho's answer is that the filtering happened overnight: the integration could only have been accomplished by Beth's own evaluative system, cognitive control needn't be conscious, so "she nonetheless exercises the requisite control during her sleep." Everything turns on that sentence, and I think it faces a dilemma. Any control theory needs a line between a mechanism merely operating and an agent exercising control over what it does. Draw the line thinly, so that whatever your cognitive system produces counts as your exercise of control, and far too much gets in, such as intrusive thoughts and stray associations. Draw it thickly enough and the process must answer to commitments or a practical standpoint that are already the agent's own, independently of the process. But that is just what the manipulation removed. Cho himself concedes that the evaluative framework doing the overnight filtering was "engineered or conditioned by the very process under question." So either the integration is a subpersonal happening that cannot deliver ownership, or it counts as Beth's doing only by borrowing the agential history the deflation was supposed to expel.
Suppose the deflation worked anyway. The intuition that Beth isn't culpable persists, and Cho's diagnosis is anchoring: we fixate on pre-manipulation Beth the way an arbitrary number distorts estimates of the Mississippi's length. Two problems. Being anchoring-like doesn’t support debunking. In work with Peter Finocchiaro, I've argued that everyday confabulations, the kind that idealize a marriage, hold a judgment fixed against an inconvenient baseline and can still be apt, because they track goods of commitment rather than nothing at all. Cho needs the Beth anchor to be the bad kind, and his river and house cases can't show that; they were chosen because their anchors track nothing, which is what makes them disanalogous to anchoring on an agent's entire prior practical identity. The second problem comes from running Cho's own test in reverse. If anchoring produces the intuition, then removing the anchor should make it vanish. It doesn't. For Ernie, created as a zygote with no prior self to anchor on, Cho concedes that "intuitions about this case diverge." For Serana, who is barely intervened upon at all, the intuition persists, and Cho's reply is that "as long as there is a radical action, our anchoring bias kicks in." A diagnosis that fires for every out-of-character action, manipulated or not, is no longer being tested by the cases. It has been immunized against them.
Structuralism's deepest commitment is that how a structure arose is irrelevant to responsibility, yet Cho's strongest defenses of that commitment can't be stated without appeal to processes: an overnight integration to supply the control, a cognitive mechanism to supply the bias. Once the defense moves from structure to process, it inherits the question of when a process is an agent's exercise of control rather than something that merely happens in her. At every point where Cho's argument touches that question, the answer either fails to deliver ownership or quietly reintroduces history. The hard-liner can still bite the bullet on Beth. It is a larger bullet than the deflate-and-debunk strategy makes it look.