A few remarks on Michael Bratman, "A Desire of One's Own" (2003)
S.W.
VI.2010
Bratman takes on the Platonic problem (e.g., Republic, Phaedrus) of what makes a desire one's own in an ethically relevant sense stronger than the desire's just being among one's psychological states. One problem with the article is that Bratman moves readily among several related but clearly non-equivalent explananda. His main idea is to take a desire's being one's own as a necessary condition for an action pursuant to it to be free, to be a case of acting on one's free will. (This closely follows Plato as well as being Frankfurt's preferred formulation). However this is not clearly the distinction needed if we are, for example, trying to mark off akratic actions, for it may be held that I act freely when I take that sixth chocolate truffle, even while claiming "I couldn't resist." Also it's not obvious that I do not act freely on a desire which I "in some sense disown" (Bratman 221). —All these are notorious tangles on which hardly anyone could claim to be clearer than Bratman. But he begins inauspiciously by stating as a rather definite (if in itself difficult) analysandum what is probably more than one target. —In what follows I'll lean toward the "free will" interpretation unless otherwise noted.
RW: I agree with this assessment. My view is that—on the notion of acting freely that I would endorse—one does not act freely when one acts against one’s all things considered judgment that one ought not to do A. But I completely agree that this has to be established. And there are vexing questions about evaluating counterfactuals (“if Jones had just tried harder, he would not have given into his fear of sharks”).
Bratman's fundamental opposition is between what he calls an hierarchical model of desire ownership and the Platonic challenge.
Basically, the Platonic idea is that a desire is one's own if and only if it is endorsed by one's reflectively held system of values and evaluative judgments.
RW: This seems to me to overlook a number of cases that we distinguish.
(1) Between reflectively held and unreflectively held values.
(2) Between “endorsing all things considered” and “endorsing (not all things considered)”.
(3) Between believing the desire (or more properly the belief/desire pair because the means matters) is endorsed, and its being endorsed but not believed to be so.
(4) Between various sorts of cases of conflict of desires, and cases of no conflict.
(5) Between endorsement and being willing to offer the desire as justification in the absence of any clear endorsement.
The lack of these distinctions makes it hard for me to know what Bratman means (below) by “concerned with the future functioning,” “part of their function to guide in this way,” and “commitments on the part of the agent concerning the role of D in her agency.” These all fall apart into different things in different types of cases.
For now we follow the article in taking this to be clear enough.[1]
RW: Despite the above, I agree; clear enough for what you do below.
The hierarchical model defines ownership in terms of five conditions given on 224-5, plus a sixth given on 240. Thus, the higher-order attitudes that determine whether a first-order[2] desire D is one's own or alien are:
(1) Higher-order attitudes about D;
(2) Themselves desires (or desire-like motivating attitudes);
(3) Concerned with the future functioning of D (or of other tokens of the same type)[3]—
(4) —Indeed it is part of their function to guide in this way;
(5) Commitments on the part of the agent concerning the role of D in her agency;
(6) Reasonably stable, that is, tending to persist unless the agent finds definite reason to drop them.
Bratman's initial presentation, in terms of an opposition between the hierarchical and the Platonic models, is recurrently misleading, sometimes just in error (e.g. 226). It is obvious that any desire that is one's own on Platonic criteria is one's own by hierarchical (or, as Bratman puts it from 226 on, Frankfurtian) criteria. (Go thorough (1)-(6) and see.) The only serious questions of the article are whether the Platonic condition is necessary, and if not, whether Frankfurt fills the gap. Might a desire be one's own without Platonic evaluative endorsement? If so, will the hierarchical model explain the ownership?
RW: What about this sort of case?
Imagine Gouge’s intensely religious upbringing instilled in him the conviction that a man should not feel erotic desire for another man. Over the years, Gouge has lost both his faith and his objection to homosexuality. He does not, however, conceive of erotic desire for another man as a possibility for him; he thinks, “I am not the kind of person who desires another man.” He does, however, eventually find himself sexually attracted to another man, and when he looks at him, he enjoys doing so under an array of features A that includes several features indicative of his sexual attraction. Thus, where those features comprise the array A, he occurrently believes, of that experience, under A, that it is occurring, and he has the felt desire, of the experience, under A, that it should occur for its own sake. It is patently obvious to him that his feeling of desire is charged with intense erotic attraction, and to his surprise, he does not how to regard the belief/desire pair. In his adolescence, he would have regarded his belief as Satan’s making him see his friend in a sexual light, and he would have regard the desire as Satan’s work, making motivating him toward evil instead of good. He would never have even dreamed of offering the belief/desire pair as a justification for looking at his friend in a sexual way. He would have acknowledged the existence of the belief/desire pair, but firmly convinced that the pair should serve as a personal reason, he would have denied its petition for citizenship in the realm of justifying belief/desire pairs, and he would never have offered the belief/desire pair as a justification for looking in a sexual way at his friend. The adult Gouge reacts differently. He finds himself willing to offer the belief/desire pair to himself and others as a justification for allowing a sexual gaze to settle on his friend, and the discovery of this willingness seems like a revelation of a long denied part of himself. Three points bear emphasis. First, he finds that he does not think that, all things considered, the pair should not function as a personal reason. Second, he believes that the pair provides at least some degree of justification for gazing at his friend in a sexual way. This follows from the fact that he is willing to offer the belief/desire pair to himself and others as a justification. While one cannot infer the existence of this conviction from the willingness to offer the pair as a justification to others (one may be lying, for example), one can infer it from the willingness to offer the pair to oneself as a justification. To be so willing is to believe (although the converse is not true). Third, Gouge also finds that also does not believe that that, all things considered, the pair should function as a personal reason. He grants the petition for citizenship in the realm of justifying belief/desire pairs, but only provisionally. His former understanding of himself argues persuasively for the ultimate denial of the petition, for the conclusion that the pair, and any similar such pairs, should not serve as personal reason, but the unexpected discovery of his willingness to offer the pair as a justification suggests that his former understanding of himself may be mistaken. He opts for provisional status of the belief/desire pair, and similar such pairs as justifying pairs, pairs he is, at least temporarily, willing to offer as justifications.
Are the relevant homoerotic desires Gouge’s own? If so, is it “endorsed”? If it is endorsed, is it endorsed in a “Platonic” way? (As I understand it, no.) Is there a Franfurt-like higher order desire? If so, does it do any explanatory work? Or is all the work done by what would explain/entail the existence of the higher order desire? (As I would argue.)
***
The present note does not focus on the Platonic idea, but briefly:
First, someone (perhaps Bratman) might doubt that Platonic valuing satisfies condition (2) above—that finding-good is as such motivating. I would disagree, but anyway this is one place to insist on the "Plato" part. Plato crafted his theory to show that apprehending the good, at least if one comes to this in the right way, entails desiring it (indeed strongly and stably).[4]
RW: Entirely with you on this one.
Second, Bratman's suggested elaboration of the Platonic position is dubious.[5] Thus (226-7):
One might respond by emphasizing that issues of ownership also arise concerning value judgments.[19] It seems, though, that the Platonic challenger can grant this point. She claims that ownership of desire consists, at bottom, in value judgment; to this she can add that the relevant value judgments need to be, in an appropriate sense, the agent’s own. But she can claim that for a value judgment to be an agent’s own is not for it to be supported by a higher-order Frankfurtian attitude. After all, even a Frankfurtian theory will at some point appeal to features of a Frankfurtian attitude that do not involve a yet-higher order attitude but help give it agential authority. This is where appeals to, for example, satisfaction and/or a cross-temporal organizing role can take center stage.[20] The Platonic challenger can herself draw on similar resources but apply them to value judgment rather than Frankfurtian higher-order attitude.[21][6]
(i) In his cited [2001] article Bratman expresses doubts about "satisfaction and/or a cross-temporal organizing role." It's a little odd to recommend to the Platonist resources he does not himself consider effective (but maybe that's because he thinks the Platonist will fail in any case?). Moreover when you look at the details, the notion of satisfaction which Bratman there borrows from Frankfurt appears hopeless for the purpose—not to mention that it seems to appeal to the notion of valuing ("evaluation", ibid. quoting Frankfurt [1999])! Meanwhile "cross-temporal organizing" is on its own far too weak a condition to shed much light on agential authority (whatever exactly we take that to be). "Never give a sucker an even break" is highly cross-temporally organizing. But it might be supremely authoritative for agent A, while B regards it as no more than a wicked, persistent (and in some ways very satisfying!) whim. The difference in authority must lie elsewhere.[7]
RW: Absolutely! Whence the “authority” of Gouge’s desires?
(ii) Bratman falsely assumes that higher-order iteration is not in the Platonist's repertoire. Stability under actual or potential iteration—valuing X, valuing that I value X, etc. —is a plausible component of the Platonic position and could well figure in accounts of authority or ownership.
Here we leave the truth of Platonic positions open. I would join Bratman in rejecting the version, arguably also Plato's, that one acts freely only if one is aiming at the (morally) good (as one understands it).[8] But Bratman's intent is to show the advantage of a Frankfurtian position over the Platonic challenge. That requires being conscientious about the challenger's resources. I don't think he meets this standard.
RW: Agree. I do think that one has the ability to act freely (in some favored sense) only if one has the ability to act on those desires one thinks provide justifications all things considered. But this obviously requires a lot of explaining and motivation.
...
Here are some moves that set up the rest of Bratman's discussion:
An initial move appeals to cases of weakness of will and the like—cases
in which value judgments do not bring with them relevant commitments,
and relevant commitments go against one’s value judgments.
Perhaps I think it strictly better to be a person who forgives and turns
the other cheek but nevertheless, in a kind of self-indulgence, allow
into my life a willingness to express reactive anger. Though this role
of my desire to express my anger diverges from my relevant evaluative
judgments, it is not a desire I reject or disown. Indeed, in some
versions of such a case—ones Watson calls “perverse cases”—I really
am fully behind the expression of such reactive anger. As Watson
says, “There is no estrangement here.”22 So, value judgment is one
thing, ownership another.
I think this is right as far as it goes; but it does not yet go far enough.
After all, it seems that these are not cases of a well-functioning agent.
A defender of the Platonic challenge might still offer his view as a
view about a well-functioning agent: at least for such an agent, desire
ownership and rejection are constituted by value judgment rather
than Frankfurtian higher-order attitude. I think, however, that even
this qualified Platonic challenge does not work. (227)
The first paragraph seems to assume that my doing ~F shows I have no commitment to F ("value judgments do not bring with them relevant commitments"). Although routine violation does at least raise a question about the presence of a commitment, weakness of will ("and the like") is precisely a counterexample to this assumption.
RW: Agree! I just could not make heads or tails of this: “Perhaps I think it strictly better to be a person who forgives and turns the other cheek but nevertheless, in a kind of self-indulgence, allow into my life a willingness to express reactive anger. Though this role of my desire to express my anger diverges from my relevant evaluative judgments, it is not a desire I reject or disown.” The seemingly patent equivocation on italicized phrases is just so obvious, that I kept thinking, “He can’t be making that mistake!”
The paragraph also holds that I can be "fully behind" my expression(s) of reactive anger while counter-valuing them. From a Platonic viewpoint this cannot be: if I value something, part of my soul strives for it. (Plausible, no?)
RW: Plausible and true.
Oddly, the first of these assumptions gives valuing too much credit (it precludes all contrary commitments), the second too little (it's not part of my motivating or motivational self). So we see Bratman trying to argue against "the Platonic challenge" while lacking a credible formulation.
RW: Agree! And with the excellent points below.
Bratman's manner of offering the Platonist a tactical concession (second paragraph) is also misleading. If desire ownership is to be a necessary condition for free will, then conceding that it occurs only in well-functioning agents might rule nearly all of us out. —It is possible to regard freedom proper as an ideal to which such fragile and blinkered creatures as ourselves can only aspire and approximate. But even then any Platonic position would have to explain how a given, fragile and blinkered, agent's configuration of values and desires does permit some approximation. —Moreover in excusing the Platonist from this task Bratman also excuses himself. As he may choose to, except that it is hard to see why the divided agents of the first paragraph should be much more tractable for him than for his (casually sketched) opponent. The Platonist's problem is supposed to be that certain desires are (in the relevant sense) owned while being counter-valued. Dismissing valuing as relevant to ownership does avoid that problem. But first, a flat dismissal does not appear open to Bratman ("... such Frankfurtian attitudeswill normally be to some extent grounded in and constrained by reflection on what one takes to be of value", 226). Second, for the same cases Bratman's own devices will not likely serve. "Satisfaction" is a poor candidate, since if anything's clear about these cases it's the agent's dissatisfaction of some serious kind! "Cross-temporal organizing" can't do, since the values—which Bratman too grants are there, even if he differs on their psychic role and moral significance—are as cross-temporally organizing as any state or disposition could be.
So the larger dialectic with the Platonic challenger is poorly executed. We will still look at what happens once Bratman sets the weakness(-like) cases aside.
Section IV—"Three examples [alcohol, anger at Green, and the military draft] and an objection." ...
—This section is muddled by Bratman's false exclusion between value considerations and higher-order considerations.
—The basic idea is: "(i) X and Y can differ in respect of their will of the desires they own, even though they agree on all matters of value, by differing in commitments. (ii) When X has and Y lacks a commitment, the difference shows up in Frankfurtian attitudes. X has forward-looking and guiding higher-order desire(s) (conative attitude(s)) that Y lacks."
—The argument in the alcohol case is bad. It's given that each of X and Y values the (appropriate) enjoyment of alcohol and values abstinence. When X abstains (as a policy or practice) and Y drinks (ditto), Bratman infers that the difference lies in Frankfurtian desires, which are not valuings, which X (say) has and Y lacks. But pending further information, the Platonist or anyone else can just suggest that the two differ in how much value each places on the respective practices. Similarly for the anger-at-Green case.