1. Substantive versus Nonsubstantive Modes of Answering the Question: Yaffe's “Transfer Principle”

Seeing the more general context of Yaffe's first question about attempts allows one to see also the need for what I call the standard, or substantive, way of answering it. On the standard route, one first gets clear why we have an area of law called “criminal law,” an area that is defined by its devotion to meting out harsh sanctions and moral condemnation. Usually the answer to this question is called a “theory of punishment.” Next, one derives a theory of the proper reach of criminal legislation by applying whatever one's theory of punishment might be to the role of the lawmaker.Footnote ⁴ Given that the rule of law (or the “principle of legality,” in criminal-law parlance) demands that the rules of criminal law be prospective, general, public, clear, and so forth, the form that criminal-lawmaking must use is dictated by values other than those in one's theory of punishment. Yet the content of such rules should be such that they serve the ends of having criminal law at all, which ends are dictated by one's theory of punishment.

I illustrate this standard (or substantive) approach to Yaffe's first question with my own answers to these subquestions. We punish because and only because people deserve it by their culpable wrongdoing.Footnote ⁵ Criminal law, in other words, exists to achieve retributive justice, which is achieved when those who deserve to suffer get what they deserve at the hands of the state. A retributivist theory of punishment, in turn, generates a legal-moralist theory of criminal legislation, which holds that prima facie, all and only acts that are morally wrong to do are to be prohibited by the criminal law, and that prima facie, all and only acts it is morally wrong not to do are to be required by the criminal law.Footnote ⁶ Trying to harm another in situations where there is no good reason to do so is morally wrong to do, and so, prima facie at least, attempts should be prohibited by the criminal law. Yaffe eschews any such substantive route to justifying the criminalization of attempts. Rather, he makes a comparative rather than a substantive move here. He defends what he calls the “transfer principle,” which he describes as follows: “If a particular form of conduct is legitimately criminalized, then the attempt to engage in that form of conduct is also legitimately criminalized” (21).

In such a way Yaffe seeks to piggyback the justification of attempt liability onto whatever justification is given for criminal liability as a principal for completed crimes. The transfer principle does not justify attempt liability tout court; rather, it justifies criminalizing attempts only if the action attempted is already legitimately criminalized.

Yaffe's motive for this circumspect mode of proceeding should be apparent: he hopes to economize on large and controversial claims in the theory of punishment and in the theory of criminal legislation. Yet in this ambition (to remain unambitious theoretically), he is bound to be disappointed. As Julie Andrews has it (in The Sound of Music), “nothing comes from nothing and nothing ever can.”

Notice that for the transfer principle to do its work, it must have two things: first, it needs its base (or “central,” or “paradigm”) case of justified penal liability to be established. It needs, in other words, criminal liability for completed crimes as a principal to be justified. Second it needs to justify the analogy of the conditions of attempt liability to the conditions that mark completed crime liability as a principal. For this latter task, it needs a criterion of relevant likenesses between the conditions making for completed crimes and the conditions making for attempts.

The problem for Yaffe's attempt at theoretical abstemiousness is that both of these two tasks require a substantive theory of punishment and a substantive theory of criminal legislation. Such theories are clearly needed to justify the central case of completed crime liability. And such theories are also needed to specify the properties by virtue of which attempts are relevantly like acts that are legitimately criminalized (viz., acts that are completed crimes).

So Yaffe needs to step up to the plate of substantive theory here. Once he does so, and once he substantively justifies attempt liability, then he might well find some employment for his “transfer principle.” Indeed, he could well be right when he says: “The Transfer Principle . . . is of great practical importance: it serves as a silent premise in the legislature's reasoning about what to criminalize” (21). Yet the principle can serve this heuristic function for legislatures only if it is first shown to be true; and that showing is made not in terms of the usefulness of the principle as a heuristic shortcut but rather by seeing substantively that the conditions that justify criminal liability anywhere are present in cases of attempts. Legitimate “transfers” will then abound. For example, when the legislature passes a statute whereby it prohibits causing state S, it will be taken (by several transfer principles, not just one) also to have prohibited anyone with a duty omitting to prevent S; anyone aiding another to cause S; anyone procuring another to cause S; anyone agreeing with another to cause S; and anyone knowingly risking S. Yet all such transfers will be legitimate only if the conditions of justified penal liability are present in each of such cases. There is no avoiding the task of establishing what those conditions are.

Fortunately, Yaffe does not lack substantive intuitions here. To begin with, Yaffe rightly sees that the theory of legislation is linked to the theory of punishment. As he puts it, “it is not justifiable for the state to criminalize a type of conduct no token of which can be justifiably punished” (31). This is not a trivial principle. Indeed, legislatures unfortunately flout it every time they pass strict liability (or “regulatory”) offenses with the intent that they will thereby prevent certain forms of conduct but without giving much thought to punishing those tokens of such conduct as are not prevented.

Second, Yaffe seems to be an intuitive retributivist in his (implicit) theory of punishment. This is because his central principle of criminalization holds that criminalization of a type of conduct is legitimate if the blame and suffering intrinsic to punishment are deserved (32). Yaffe admits that he does not “have an argument for the claim,” but he is in pretty good company there. Honest retributivists typically have quite short argument chains supporting their retributivism.Footnote ⁷

Third, Yaffe sees that wrongdoing is part of what makes one deserving of punishment (the other part being culpability). Prima facie wrongdoing is constituted by the types of acts that are usually wrong to do and that are the object of legislative prohibition; actual (or “all-out”) wrongdoing is prima facie wrongdoing where there are no exceptional circumstances justifying the doing of a prima facie wrongful act on this occasion. Yaffe rightly sees (32–33 n.12) that justifications only complicate the story about wrongdoing and criminalization; they do not change the basic story line, which is that wrongdoing is an essential component of desert and as such is “the leading feature” (25) in the conditions that must be “met by both completion and attempt” (25) in order to justify criminalization.

Fourth, Yaffe rightly worries that the wrongdoing present in attempts is different from the wrongdoing present in completions (25–27). This is a real worry, because there are differences here. Whether those differences should make a difference for criminalization can be seen only by seeing what these differences are and whether they touch on what makes one deserving of the blame and the suffering that is punishment. This leads Yaffe to a search that all retributivists have to undertake to complete their retributivism: What are the natural properties by virtue of which one becomes blameworthy and deserving of punishment? What are, in other words, the proper desert bases for blame and punishment? Yaffe's theoretical abstemiousness again leads him to try to avoid answering this question directly. But despite himself, he eventually comes to what I regard as a substantive answer. Let me trace his path.

Yaffe first considers a view that goes like this: whatever the differences between attempts and completions, there is no difference between them vis-à-vis the presence of a culpable mental state executed into at least a basic act (34). That is, attempts require an executed intention to do the wrongful act; completed wrongs require no more (although they may require less) by way of culpable mental state. And, some would say, such culpability of executed intention is a sufficient desert base; success (in the form of completing the attempt or causing the harm intended) is not required and indeed adds nothing to one's desert.

Yaffe does not out-and-out adopt the “culpability alone is sufficient” view just described, because it unabashedly takes a substantive position on the necessary and sufficient conditions of blameworthiness. Yaffe instead affects an agnosticism about whether causing the harm one intends affects one's just deserts. If it does, he says, then attempts cannot completely differ from completions, because the closeness (to success) of any attempt is only a matter of degree, and it is implausible to think that there is no moral wrongdoing in attempts when they share some (“close causings”) of what makes completions the wrongs they are.

One might respond to Yaffe that there is nothing irrational or contradictory about a morality that proclaims “a miss is as good as a mile.” Less metaphorically: on the response to Yaffe being imagined, if the bad state of affairs does not occur, it does not occur, and coming close to causing it to occur just does not count. Yet this is an inconclusive response; if I were Yaffe, I would substitute a scalar notion of causation for closeness of causing. Then the bad state of affairs does occur, only one's causal contribution to it drops below some threshold needed to be held the (or a) cause of that state of affairs. Then Yaffe can object to the arbitrariness of the threshold, urging that some attempts possess some of the feature (causing) that makes completions the wrongs that they are.

This is a pretty inconclusive line of argument, either in Yaffe's original version or in my amended version. He implicitly recognizes this by eventually going substantive here. Yaffe bluntly claims that the censure or blaming involved in punishment is deserved for conduct that is the “product of a faulty mode of recognition or response to reasons for action” (38). Yaffe nicely unpacks what he means by a “faulty mode” of reason recognition/response in terms of characterological items such as insufficient concern for others, and so forth (38). But again, Yaffe is embarrassed by his inability to defend his substantive view of a proper desert base, saying “I can't claim to have defended it” (38). Yet again, if Yaffe should be embarrassed about this, it is an embarrassment widely shared. Extended but still cogent arguments for why certain natural properties are or are not bases for moral desert are notoriously hard to find.Footnote ⁸ Mostly what one finds in the literature here are theorists who think they have an argument, such as the popular delusion that degree of control can isolate that for which we are properly to blame.Footnote ⁹

Yaffe goes on to show how his favored desert base is but a deeper form of the desert base mentioned above, that based on executed intention. Yaffe urges that mental states such as intention matter to our deserts but that they matter because of what they tell us “about the deeper facts constitutive of culpability, facts about the agent's ways of navigating the terrain of reasons” (43).

What Yaffe does not do (here or anywhere else in the book) is to take a definitive position on the other aspect of the executed-intention desert base mentioned above, namely, does success (in the form of causing the harm intended) matter to deserts or not? One rather suspects from isolated asides at various points in the book (221, 323–333) that Yaffe thinks not, in which event his substantive view is but a deeper version of the executed-intention view of many contemporary theorists.Footnote ¹⁰ In any case, whatever turns out to be his view on this bit of the executed-intention desert-base idea, Yaffe, at the end of the day, does have a substantive view about what deserves the censure and sanction of punishment. His transfer principle thus ends up as a makeweight in his justification of why we criminalize attempts. As he himself eventually recognizes, the “crucial premise in the argument for the Transfer Principle . . . is the claim that the person who attempts the crime is recognizing and responding to reasons in the same, or a worse way than the person who completes the crime” (101). This “crucial premise . . . for the Transfer Principle” actually contains within it the basis for dispensing with the transfer principle: if the person who attempts a criminal act is engaged in faulty recognition/response to reasons, then on Yaffe's own theory of the true roots of blameworthiness, attempts should be criminalized—period—no matter what the penal system is doing about completed crimes.

2. Why Not Criminalize Nonattempt Inchoate Wrongs?

Seeing the truly substantive basis on which Yaffe urges criminalization of attempts allows one to see another issue for Yaffe. If causation of a harm intended is not necessary for at least some punishment (leaving aside whether as much punishment as for the completed wrong or not), and if, therefore, the culpability of executed intention is sufficient (because of the faulty mode of recognizing and responding to reasons evinced by such intentions), then why is the culpability of knowingly or recklessly risking a harm to another not also deserving of some punishment, even in the absence of causation of the harm risked? After all, presumably Yaffe thinks that agents who are indifferent to the harms their actions cause others are faulty in their reasoning. That is presumably what makes it right, in Yaffe's view, to punish knowing or reckless versions of such completed wrongs. But then, by parity of reasoning, is it not also right to punish knowing or reckless risking by itself, without requiring causation of the harm risked? Yaffe's sufficient desert base seems present here as well as in the case of executed but unsuccessful intentions.

Let me set aside two issues that may seem to be at stake here but are not. One is the issue between the common law and the Model Penal Code on whether risking harm to another is the central idea behind inchoate liability generally in the criminal law. Arguably, common-law doctrines (such as those involving so-called “legal impossibility” or those relying on “dangerous proximity”) evidence a kind of “near-miss” view of inchoate liability, namely, what is generally sought to be punished by inchoate liability are dangerous attempts, that is, the near misses that come dangerously close to success. If our aim were the utilitarian one of preventing certain harms, the common law rightly sees that a plausible way to do that would be by punishing actors when they come close to causing those harms as well as punishing actors who actually cause them.

Even if Yaffe's principle was extended to cover knowing and reckless states of culpability in addition to intentions, that would not be to adopt the risk orientation of the common law just depicted. For the common-law view described is not concerned with the culpability of knowing risk creation; rather, it is concerned with the wrongdoing of objectively unreasonable risk imposition, knowing or not.

The second issue to be put aside as not being involved here has to do with the Alexander and Ferzan view of culpability.Footnote ¹¹ On their view, intentions are culpable because they are accompanied by perceived risk; and for them, perceived risk is the ultimate root of culpability. Even if Yaffe's view were extended in the manner contemplated, that would not be to validate the Alexander and Ferzan view of culpability. On the extension of Yaffe's view here considered, executed intentions and knowing risk impositions might both be culpable because they each involve faulty modes of reasoning; yet they could be faulty in their own distinct ways. The culpability of intending could lie just where Aquinas put it, in “setting one's will towards evil,” whereas the culpability of knowingly risking could lie in the attitude of indifference. Neither need be more basic than the other; they could be just different modes of faulty reasoning.

These two extraneous items aside, why is Yaffe not committed to extending inchoate liability in the criminal law to all levels of culpability, not just those involving intention? Yaffe initially sounds as though he in fact accepts the extension (44), but then he adds a huge caveat: knowing or reckless risk impositions that in fact have a low probability of harm are not justifiably criminalized (44). This cashes out later in the book to admitting only to the justifiability of criminalizing knowing risk impositions where the “actual likelihood” of the harm risked coming to pass is so high that “knowing what will happen if one acts is just as bad as intending it” (123), whereas Yaffe generally eschews treating the culpability sufficient for the completed crime (such as knowledge) as sufficient culpability for the inchoate version of such a crime (48–49). Yaffe thus evidences no acceptance of the true implications of his desert base, no acceptance of the legitimate extension of his transfer principle: “If attempts are to be criminalized through appeal to the Transfer Principle, then intent must be an element of every crime of attempt, even when that crime can be completed without intent” (49). And this is because “the Transfer Principle . . . depends on tethering what we criminalize as attempts to what attempts really are” (49), and attempts really are tryings and tryings really require intendings.

Contrary to Yaffe, I think his desert base (and the expression of it in his hands, his transfer principle) extends ineluctably to the culpability of knowing risk imposition no less than to the culpability of intending. This does not mean that Yaffe is wrong in his account of attempts; it does mean that attempts do not occupy the entire field of inchoate liability in a properly conceived penal code. Yaffe's error here, if error it be, is one of emphasis and breadth of topic selection. Such error by him is the mirror image of a like error by Ferzan and Alexander, who emphasize knowing risk imposition to be exclusion of intending. A complete exposition of the topic of inchoate liability would have both of these items in it.

3. Does Yaffe's Desert Base in Faulty Modes of Reasoning Justify Criminalizing Bare Intention (Rather than Executed Intention, or “Trying”)?

Seemingly Yaffe's desert base in faulty modes of reasoning implies a broader form of inchoate liability than what he recommends, in a dimension quite different from that concluded above about knowing risk imposition. I refer to the possibility of an inchoate liability to punishment for intending to do some wrong even when that intention is not an executed one; that is, one has not yet tried to bring about that which one intends.

To see whether this is so, attend to what Yaffe says in detail about his chosen basis for moral desert, faulty modes of recognition or response to reasons for action. Such faulty modes of recognition/response to the reasons for action that bind us all are instanced by Yaffe as follows:

Now the worry I see for Yaffe is as follows: Cannot all of these instances of faulty reasoning be present in the formation of an intention to do some wrong? The intention (in Bratman's theory of it, which Yaffe rightly adopts) adjudicates between conflicting reasons; it is a decision based on such reasons. As such, all of these defects of underweighting, overweighting, overconsidering, and so forth, if they are present, are present in the forming of the intention to do some wrongful act. Why, then, is not the intention alone a sufficient basis on which to blame and to punish? Why does Yaffe insist on an execution of that intention into what we call a trying, that is, the willing of some step toward completion?

Note that this is not the question that Yaffe does ask in chapter 8, which is why we require acts (in the sense of basic acts, or bodily movements) before we hold someone for trying. My question is prior to Yaffe's question: Why do we require trying in the first place, instead of intending? It is no good repairing to the conceptual truth that attempting requires trying. It does, but that proves as useful here as did an analogous conceptual truth (that attempting requires intending) in the previous section. In both cases Yaffe has used as the ground of his analysis a desert base of faulty modes of reasoning. That base, here as in the preceding section, renders irrelevant conceptual truths connecting trying (or intending) to attempting.

Because of Yaffe's theory abstemiousness, he does not see that he needs to answer this last question. Yaffe might attempt to use here the answer he gives to a different question. To the question: Why does trying require a basic action—or, alternatively worded: Why does trying have to “break the skin” in a willed bodily movement—Yaffe gives an evidentiary answer. As he puts it, “although it is possible to try without acting, we can know someone is trying only if he does” (235). Making this into an answer to my question would be: although it is possible to intend a wrong without trying to cause it to come about, we can know someone is intending to do that wrong only if he tries to do it.

I waive the obvious quibbles here about how trying has no evidentiary advantage over intending, given that trying (on Yaffe's quite correct conceptualization) is fully as much an internal state as is intending.Footnote ¹² I waive this otherwise cogent point because I doubt that Yaffe would make the answer I just put in his mouth. He thinks (I think) that it is no mere evidentiary consideration that disqualifies the criminalization of bare intending; he thinks that bare intending does not deserve punishment, whereas trying does. The trick for Yaffe is how to get that distinction out of his stated desert base (of faulty modes of reasoning).

I suspect that Yaffe feels, in some sense, the pressure of the problem. I suspect this because of his conjunctive wording of his desert-base principle: there need to be faulty modes of recognizing reasons for action and faulty modes of responding to those reasons. “Responding,” in Yaffe's mind, is not equated with intending or deciding; rather, responding to a conception of a balance of reasons for action is, for Yaffe, “translating that conception into action” (38).

So, formally, Yaffe might think that he has stated his principle of desert in a way that requires faulty execution of intentions that are themselves the product of faulty recognition of reasons. Yet the “Why” question goes unanswered in any such formal defense. It is plausible that weighting one's own interests too heavily or indifference to others constitutes faulty modes of reasons-recognition that make intending a plausible desert base; it is not plausible that executing that balance-of-reasons-based-intention by trying is itself a necessary part of that desert base. If one's idea of a desert base is characterological in the way that Yaffe's is, nonrecognition of reasons for actions seems relevantly faulty, whereas execution is faulty only in that it is based on such faultily formed intention. Such merely derivative faultiness reveals the true desert base, which is in faulty reason-recognition.

4. Why Stop at Intention as a Desert Base?

Yaffe at one point chides legal theoreticians for the shallowness of their analysis of culpability. As he puts it:

Let me here plead guilty, on behalf of myself and many of my fellow criminal-law theoreticians, to Yaffe's charge of shallowness. I confess it: my long-held view has been that prima facie culpability is constituted by the agent either wanting, or intending, or believing some wrongful action to be done by him.Footnote ¹³ That is why I sloganize prima facie culpability as “wrongdoing in the mind of the actor.” To know that what you are doing is a type of action that morality forbids, or to want or intend that type of action, is to choose that which morality forbids. Morality makes it wrongful to do certain acts, and because of this, it is culpable to choose (want, intend, believe) to do these same acts.

Shallow folks like myself find this satisfying as it stands. Yaffe, however, as we see above, indulges the philosopher's characteristic penchant here,Footnote ¹⁴ which is to attempt to justify something that on its face does not demand some deeper justification. He does this by justifying intention's role (as a marker of culpability) in terms of the faulty modes of reasoning behind intentions. My main objection to this lies in a substantive disagreement as to the correct desert base. I find doing wrong in one's own mind (by wanting or intending to, or believing one is) a more plausible basis for culpability than the characterological defects onto which Yaffe fastens as his desert base. For the latter desert base, if spelled out in detail, will probably suffer all the defects of aretaic accounts of deontic ethics generally.Footnote ¹⁵

My second objection here is a reductio that parallels the objection raised to Yaffe's defense of trying as a desert base. In the preceding section I query why Yaffe tacks on execution of intention, to the having of the intention in the first place, as a proper desert base. But continue that query: Why stop at intention, executed or not? If the locus of our blameworthy culpability lies in faulty modes of reason recognition, why use any proxy at all when we make judgments of responsibility? Why does Yaffe substitute the proxy of trying, that is, of executed intending? Why even a proxy of bare intending? If what really grounds culpability is faulty modes of the reasonings that produce culpable intentions, why do we not cut to the chase and use faulty modes of reasoning directly in our judgments of culpability?

Yaffe again could repair to some evidentiary rationale, as he does later in his book to justify why tryings must at least go so far as to “break the skin” with the bodily movements that are basic acts. Yet that will be no more successful here than it was with a like use of that rationale to justify tryings as a proxy. Intentions are in no obvious way superior (in terms of provability) to reasons and modes of reasoning; and in addition, Yaffe, if I understand him right, is not even tempted to make the evidentiary defense here.

This last point is worth dwelling on. My interpretation of Yaffe is that he thinks, plausibly enough, that the desert base for judgments of culpability is limited to human agency. What is bad is the setting of the will toward evil. Bare, unexecuted intending to do some wrongful act is a part of such agency; but it is not all of it, nor even the most important part. Such agency is mainly constituted by the subject moving from decision (intention) to execution (trying). True enough, for Yaffe, such agency is further constituted by that directness of causation (of both the bodily movements constitutive of basic acts and further real-world consequences of those bodily movements) that Yaffe believes is enshrined in our use of causative verbs of action, like “kill,” “hit,” “maim,” and so forth. Yet despite this last “agency-involvement factor” (as Yaffe calls it), the essence of agency for Yaffe lies in the execution of intentions into actions. It is this conception of agency that seemingly underwrites Yaffe's requirement that intentions be executed into tryings before one has the culpability needed for being an attempter. Such a conception is in no obvious way dependent on some “deeper” account in terms of faulty modes of reasoning. It can stand on its own as a desert base.

Going this route allows Yaffe to do exactly what he does in the succeeding chapters of his book, which is to take very seriously the constraints incident to intending (chap. 2) and trying (chap. 3) and using those constraints to rationalize or criticize the current law on attempts (chaps. 4–7). Indeed, by my lights, staying “shallow” here allows him to do what he does in his book with great commitment and seriousness; this is because he can justify his focus on trying and intending not as mere proxies but as the real thing when it comes to the true roots of culpability in human agency.

1. Yaffe's General Structure for Attempts

Yaffe transitions from the first question of why we blame and punish attempts to consider a second question, namely, the question of just what attempts are. Yaffe's answer to this second question is straightforward: blamable and punishable attempts to do some crime, C, are tryings to C. Consistent with his nonsubstantive answer to the first question—in terms of his transfer principle—Yaffe uses this principle to justify his “tryings” answer to his second question: “To criminalize attempts through . . . appeal to the Transfer Principle is to commit oneself to a construction of the elements of attempted crimes that respects the nature of trying to act” (71). This is because “the rationale for . . . the Transfer Principle . . . depends on tethering what we criminalize as attempt to what attempts really are” (48), and what attempts really are, are tryings.

Yaffe's answer to the question of the nature of criminal attempts—they are tryings—can also be justified more substantively. If the reason we criminalize attempts to do something wrong is because attempts themselves are blameworthy (even in the absence of the wrong attempted being completed), then the nature of attempts must be sought in those features of them that make them morally blameworthy in their own right. It is plausible to think that what makes attempts blameworthy is that they are tryings to do wrong. Tryings are, as Aquinas says, a setting of one's will and agency toward evil, and are blameworthy on that account.Footnote ¹⁶ Tryings are then rightly criminalized as attempts.

However one gets there, surely Yaffe is on solid ground in holding that the nature of attempts centrally focuses on tryings. Yaffe's next step is to connect trying to A, to intending to A. Yaffe describes this connection in terms of what he calls the “guiding commitment view” of trying. Yaffe initially describes this view intuitively as follows: “to try to do something is to be committed to each of the components of success and to be moved by those commitments” (73). Notice that “the guiding commitment view” has two components: commitment and guidance. This is as it should be. Intuitively, tryings are executed intentions, that is, intentions that one acted upon. “Guidance” is Yaffe's name for the execution part of trying: “there is guidance by one's intention-based commitment when the intention causes something on the way towards performance of an act that one believes to be a means to doing as one intends” (74).

As a first cut, “commitment” is Yaffe's term for intending (I qualify this below). Certainly the commitment essential to trying, according to the guiding commitment view, necessarily involves intending: “A person is committed to success in the way involved in trying just in case he has a certain intention” (73).

I focus on the commitment/intention part of Yaffe's account of trying. I do this because it is here that Yaffe is most subtle, most complicated, most original, most relevant to practical concerns, and (for all of these reasons) most interesting.

2. Yaffe on the Nature of Intending

Philosophical discussions of intending/intention usually distinguish two sorts of questions.Footnote ¹⁷ The first kind of question is about the nature of intending as a mental state: How does it differ from and how is it related to beliefs? To desires? To emotions? And more generally, what sorts of things are intentions, beliefs, desires, emotions, and mental states as such? Are they physical states of the brain? Behavioral constructs? Functional states? The second kind of question is about the content (or “object”) of intending as an Intentional mental state:Footnote ¹⁸ How is the content of an intention fixed? And more generally, what sorts of “things” are the objects of Intentional states such as intention, belief, and desire? Are they real-world events or objects? Sentences about these? Propositions about these? Or dispositions of their holders to avow certain sentences or propositions?

Yaffe does not divide his discussion of intention between nature and content. I think this is because Yaffe is guilty of some sleight of hand here, hoping that his fudging the nature of intention can let him fudge the content of intention. So in the hopes of keeping Yaffe's hands where we can all see them, I am forcing the traditional organization onto Yaffe's discussion. I start with the nature of intending.

Yaffe's view on the nature of intention has a history.Footnote ¹⁹ The modern part of that history starts with the view often attributed to early Donald Davidson:Footnote ²⁰ that intentions just are the belief/desire sets that cause as well as rationalize actions; through the view attributed to later Davidson: that intentions are judgments (beliefs?) about all-out desirability of certain courses of action;Footnote ²¹ and through Michael Bratman's view: that intentions are neither belief/desire sets simplicatur and neither are they to be identified as all-out beliefs (or Davidson's “judgments”) about all-out desirability, rather, intentions are a third and distinct kind of mental state involved in the production of rational action.Footnote ²²

What makes a mental state an intention, distinct from both belief and desire? On Yaffe's account it is the distinctive norms of rationality that are applicable to intentions but not to beliefs or desires. As Yaffe puts it, “assessibility by such norms is what makes an agent's mental state one of intending” (68),(See also page 64.)

This, to my ear, is a quite jarring claim. What is jarring is not so much that it seems wrong as that it seems wrongheaded (in the sense of a category mistake). Suppose one is puzzled about what makes a state of a person one of death—does it respond to one's curiosity to be told that “to be dead is to be subject to the appropriate concern of a mortician?” Or, closer to home, that “to be dead is to be no longer subject to the same norms of sociability?” Likewise, suppose one is puzzled (as we surely all are at least sometimes) about the nature of romantic love. Is one's curiosity sated by the (in)famous Erich Segal line in Love Story, namely, that to be in love means never having to say that you're sorry? Yet again, suppose one is puzzled about what makes a member of the species Homo sapiens into a person—is it responsive to be given an all-too-common response: persons are “any beings subject to certain norms of rationality and responsibility appropriate only to persons?”

It is not that such claims are false—they might be, but that is not the worry here. It is rather that something else, something more basic and more grounding of these “appropriate assessibility” claims, is wanted in answer to these questions. So, too, with the nature of intention. Intentions may indeed often (or even always) be assessable in ways that beliefs and desires are not; but wanting is some nature to intentions that grounds such distinctive kind of assessibility. Put another way, that intentions are assessable in certain specific and unique ways should be a contingent fact about them, not a constitutive fact of them.

Someone in Yaffe's position might fairly respond to the foregoing by asking, “but what does a nonwrongheaded answer look like to the question of intention's nature?” A fair enough question. I take it that the modern answer for intentions starts with their functional specification in terms of the causal roles they play in the etiology of rational action. If one elongates the history given a while ago, starting with Aristotle rather than Donald Davidson, recall that Aristotle famously proclaimed that the conclusion of a particular syllogism was an action and that when the agent possessed both premises of the practical syllogism, “straightaway” he acted.Footnote ²³

Not so fast, says the post-Bratman philosophy of mind. A causal intermediary between one's background motivations (belief/desire sets whose content forms “valid” practical syllogisms) and one's actions is one's intentions. The initial specification of such intermediate states of intention is in terms of this dual causal role: they are the effects of the belief/desire sets that constitute motivation and the cause of acts that serve such motivations.

As it stands, this causal-role, functionalist specification of intention is incomplete. Intentions are not just any state that causally mediates between action and motivation; intentions do their causal mediation in a way that rationalizes the action, as Davidson would put it. This means first, that the content of the causally intermediate mental state must be related both to the content of the actor's belief/desire sets, on the one hand, and to the tryings done in execution of those motivations, on the other, if that mental state is one of intention. And also required is that the causal relations between the putative intention and the trying, and between the motivations and the putative intentions, be “nondeviant.” A mental state has to be caused and to cause, “in the right way” (Alvin Goldman's description of the kind of causation involved in agential execution)Footnote ²⁴ for that causally intermediary mental state to be an intention.

Functionalists divide on how they regard the physical realization of functionally specified states like intention. Suppose the state that plays the functional roles specified for intentions highly correlates with certain physical goings-on (let's call them, “X”) in the parietal cortex of the brains of human beings. Does X constitute the deeper nature of intention? Or is the essence of intention to be limited to the functional specification? If it turned out that in a few isolated cases, a state playing the functional roles of intention was not physically realized by X but by some other brain goings-on, Y, would that state be an intention? Conversely, if some state was physically realized in X but did not play the functional roles (or at least, not all of them) assigned to intention, would that state be an intention? In these examples one should recognize the Kripke-Putnam-style thought experiments seeking the essence of natural kinds like gold, water—and intentions.

However one comes out on this last question (of function/structure priority), one should still be worlds from Yaffe's assessibility criterion of what is an intention. That is not the issue that Yaffe does explore—which is whether such norms of rational assessibility of intentions are omnipresent or merely usually present rules of thumb (62–64). One can concede to Yaffe his view on that issue, which is that such norms are omnipresent and apply without exception to anything that is properly called an intention. Even so, such rational assessibility is the tail of the dog, not the dog itself. Intention is a natural kind of state functionally specified by its distinctive causal roles (themselves distinctive by virtue of the content-relations and kind of causality) and (perhaps) physical realized in some kind(s) of distinctive brain states. That such states are rationally assessable in the ways that Bratman and Yaffe describe is a contingent (not essential) truth about them that holds because of the causal roles played by intentions in our rational actions. Put more simply: intentions have a nature that is deeper than assessibility; a mental state may be an intention if it is assessable in these distinctive, Bratman–Yaffe ways, but it will not be an intention because it is so assessable.

3. Yaffe on the Content of Intending

One of the features that intention shares with some other mental states (such as desire, belief, and emotion) and with some action descriptions (namely, the intentionally complex ones, such as “trying”) is that they are all Intentional states, that is, states with content. One does not just intend, period, any more than one just tries, desires, or believes, period. Rather, one intends that something be made the case, tries to make it such that something is the case, desires that it be the case, or believes that it is the case. What follows the “that” in such constructions is the Intentional object of intentions, or their content.

The criminal law generally cares about the content of intentions in two ways. These correspond to two “match” questions one must ask and answer in order to convict one of a completed crime having “specific” intent as its mens rea.Footnote ²⁵ First, one must ask whether the intention some defendant had on some particular occasion matches the type of intention the criminal law requires for conviction. Second, one must ask whether the action the defendant actually performed matches the type of action he intended to perform. Since only the first of these match questions is involved in attempts—because, as an attempt, the defendant did not do what he intended to do—let us focus on it alone.

Notice the first match question involves three subquestions. (1) The question of law: What type of intention does the criminal law require for conviction of this offense? (2) The question of psychological fact: What token-intention did the defendant have on this occasion? And (3) The question of interpretation: Does the intention the defendant had on this occasion match the type of intention required for conviction closely enough for him to be considered guilty of the crime (in its mens rea elements)?

Yaffe is rightly concerned with all three of these subquestions. Let us start with the first, the legal question of what intent should be required by law for attempt liability. Prima facie, the answer is simple: the actus reus of the crime attempted should be the content of the intention required for conviction of attempting such a crime. This needs a slight modification: some elements of the crime attempted may be irrelevant to the underlying moral wrong—such elements are added for jurisdictional or other reasons of “legal technicality.” So one eliminates these not-“material” elements from the content of the prohibited intention. Thus my slogan: “culpability is wrongdoing in the mind of the actor.”

Yet this simple answer gets more complicated in a hurry, for there is a tricky interpretive problem here that Yaffee glosses over. The problem can be approached by adverting to a familiar distinction that Yaffe does introduce, that between de re (or in the older language, “transparent”) and de dicto (or “opaque”) readings of Intentional content (143–146). If I intend to hit the president, on a de re reading I intend to hit whoever I am referring to as the president (say, Jones), even if it turns out that Jones is not the president; whereas in a de dicto reading, I intend to hit whoever it is that truly possesses the property of being the president (144). This is to construe the de dicto/de re distinction in terms of Keith Donnellan's more general distinction between attributive versus referential uses of descriptions.Footnote ²⁶ Yaffe without mentioning Donnellian nonetheless pretty clearly adopts this Donnellian construal of de dicto/de re: “When a descriptive phrase is interpreted de re, it contributes only its referent to the truth conditions of the sentence. When it is interpreted de dicto, by contrast, the properties referred to in the description contribute to the sentence's truth conditions independently” (144).

Notice that there is a crucial ambiguity in the usage of the de re/de dicto distinction as thus far described. As put just above, the distinction is between two kinds of intentions that people can have on different occasions. So used, the distinction is internal to the psychology of the holders of the intentions. Two people, each of whom form the content of their intention as, “I kill the President,” can nonetheless differ in that one intends de re to kill the President whereas the other intend de dicto to kill the President. The first intends to kill Obama, referring to Obama in the content of his intention through one of the properties Obama possesses, viz, that he is President; whereas the second intends to kill whoever is President (which, because Obama has that property, turns out to be Obama). If these two people differ in this internal de re/de dicto manner, they differ in the psychological states that they each possess.

Yet we also use the de re/de dicto distinction externally. So used, the distinction is between two ways of speaking about one and the same intention. We might say of some would-be Presidential assassin, Jones, that “Jones intends to kill the President”; by this, we might mean, de re, “Jones intends to kill Obama, and Obama you know is the President,” or, de dicto, “Jones intends to kill the President whom Jones knows is Obama.” Notice that on the externally de re expression, Jones may or may not refer to Obama under the representation, “the President.” The speaker of such an externally de re sentence is noncommittal about how the holder of the intention represents Obama because what justifies such external usage of the description, “the President,” is the objective fact known to the speaker that Obama is the President. On a de dicto usage, by contrast, Jones has to represent Obama as the President when he forms the intention to kill him. Jones may form that intention either de re or de dicto internally; either way of representing Obama as the President will do. Such externally de dicto description of Jones’ intention is noncommittal about whether Jones has a de re or de dicto kind of such an intention.

Yaffe confines his use of the de re/de dicto distinction to its internal use in distinguishing two kinds of intentions. He thus bypasses the tricky interpretive problem that is reached when one focuses on the externally de dicto usage of intention's content. Such a focus is demanded by the law and by morality because in the external sense they use “intention” in an exclusively de dicto manner.

This is not because ordinary usage is univocally de dicto externally. Indeed, ordinary speech is truly ambiguous as to whether it is using “intention” in an externally de re or de dicto manner. If I am a Secret Service officer, I presumably care very much that Obama is the President; so much so that when I learn that Jones intends to kill Obama, I will likely describe Jones’ intention as an intention “to kill the President” irrespective of whether Jones represents Obama as the President or even knows that Obama is the President. Legal and moral usages are not open to such externally de re usages, however. If Jones is accused of assault with intent to kill the President, for example, the facts that Obama is the President and that Jones intends to kill Obama are insufficient to find that Jones intended to kill the President. Needed is that Jones represented the person he intended to kill as the President in order for Jones to have such a wrong in mind and thus be culpable; needed, in other words, is for Jones to intend, in an externally de dicto way, to kill the President.

One can see this last point rather clearly with a time-worn example. Suppose that I am accused of intentionally stealing Smith's umbrella. Suppose that I do in fact take Smith's umbrella rather than my own. If I intend to take “that umbrella,” believing it to be my own, when in fact it belongs to Smith, I cannot be held to have intended to take the property of another. The content of my intent in this context must be externally de dicto, and this requires use of the properties I used to pick out the umbrella. If those properties do not include the property, “belonging to Smith,” I do not intend to take the umbrella of another.Footnote ²⁷

The externally de dicto reading of “intent” as used in moral and legal contexts is the source of the interpretive problem earlier alluded to. The problem is how we are to apply the descriptions of intention-types contained in criminal statutes to the intention-tokens possessed by some defendant on some particular occasion. In the umbrella example just given, the common law of theft requires an intention having as its content the actus reus of theft, namely, the “asportation of property of another.” No one, not even lawyers, forms an intent to “asport” anything. More typically, a thief might intend something like, “I grab Smith's umbrella.” If the thief does in fact grab Smith's umbrella, he has performed the actus reus of theft. For grabbing something moves that thing, the essence of common law asportation; an umbrella is property; the thief is not Smith; and that the umbrella belongs to Smith makes it “of” another.

If we could give an externally de re reading of the intent required for conviction of theft, the just recited facts (about the thief's act-token instantiating the type of act required for conviction) would suffice to ascribe a guilty intention to him as well, based on the fact that he did intend to grab Smith's umbrella. For on an externally de re reading of “intent,” the actual facts of the matter govern what someone intends. But this is not true for the externally de dicto reading of “intent” that we must use in legal and moral contexts. In such contexts the types or descriptions “in the head” of the intender governs what he intends, not the actual facts of the matter. And no one, as I say, ever intends in the exact language of criminal statutes.

The interpretive problem stems from the fact that we are here bereft of our normal, extensional tools in classifying particulars. That a bus is a vehicle in fact or law does not mean that if I intend to steal a bus I intend to steal a vehicle. Whether I in fact intend to steal a vehicle depends on how I represent what I intend to do, not the actual facts of the matter about what I do.

This matter is complicated by the fact that intentions, like predictive beliefs and some desires, do not take individual actions as their objects. Admittedly it is idiomatic in English to say, in the past tense after some action has been done or some harm has been caused, “she intended that act A,” or “she intended to cause that harm H.” Yet that act and that harm did not exist at the time the intention was formed. Instead of referring to the particular act that turns out to be the one the actor performed or to some particular harm that that act caused, a more accurate rendering of the content of the intention would be: “she intended that she do some instance of type of act A,” or “He intended to cause some instance of type of harm H.” Intentions have within their content existential quantifications over types, not reference to some particulars.Footnote ²⁸

It might seem that this complication helps rather than hurts, because types of acts or harms can be classified into larger types (like “asportings”) in the usual extensional ways. Yet the content of intention is not a type but a description or representation of a type. The upshot of this is that we are not asking, in our first match question, whether the type of action (some instance of which the defendant intended) is within the extension of the type of Intentional content the statute prohibits. Rather, we are required to classify descriptions or representations of types of actions or harms, because that is what defendants have as the content of their intentions.

The consequence of this, important in this context, is this: in classifying intention-tokens as being of some prohibited intention-type, it is unclear how much we get to use of our usual tools of instantiation and classification. Even though the representation of the act I intend to do in my mind is, “grab that umbrella,” and even though grabbing is a form of moving of personal property, the essence of asportation, and even though “that umbrella” in fact refers to the property of another, say Smith, none of this licenses us to conclude that I intended “asportation of the property of another,” the actus reus of common-law larceny. Grabbing Smith's umbrella is in fact asportation of the property of another—the act done is an instance of that legally prohibited type of action. Even so, given the de dicto nature of Intentional content, we cannot conclude I intended that type under any description. For the description of that content in my head may not be picking out the same type as that which is prohibited.

The problem is that we are bereft of our normal (extensional) classificatory tools in these circumstances. And yet we cannot require defendants to represent what they intend to do in the exact language of the statute. For as shown above, that would be far too narrow—even lawyers do not commit their thefts literally intending to asport anything. Needed is some middle ground between content identical in its description to that stated in the law (too narrow) and content that refers to acts in fact violative of the law (too broad). The problem is that it is far from obvious where such middle ground is to be found or even how one calibrates how “middle” it is between these extremes.

It is perhaps true that ethics and law face no more of (and no different) a problem here than is faced generally by the folk psychology of intention. For in our ordinary ascriptions of intentions to ourselves or others, we grant some but not too much leeway in redescribing the objects of our intentions. Quine's old analogy to indirect speech is here a fruitful one.Footnote ²⁹ In reporting what we or others have said, we often do not use direct quotation, as in: “Joe said, ‘Mary doesn't have much use for John.’” Rather, we paraphrase the content of what Joe said into expressions like: “Joe said that Mary doesn't think much of John,” or “Joe said that Mary doesn't like or respect John,” and so forth. In indirect speech, we give the gist but not the literal wording of what was said. The analogy to indirect speech yields a “two-level” analysis of the content of Intentional states such as intentions.Footnote ³⁰ The basic level directly quotes the canonical description in the mind of the actor; the second level paraphrases the gist of it.

It is plausible to construe what a defendant needs to intend in criminal-law contexts as operating at the second level, that of paraphrase of “the gist.” In which event, it does no violence to the language of such statutes to construe a defendant's intention that he “grab Smith's green umbrella,” as a instance of the type of intention, “to asport the property belonging to another.”Footnote ³¹

Such an analysis, however, only broadens the discomfort; it does not remedy it. Such an analysis shows that context of utterance must supply parameters of legitimate paraphrase of the objects of intention. The analysis does not even hint at what, in the context of ascribing moral responsibility for an attempt, those parameters are or even how they are to be located.

I am unclear whether Yaffe intended in his book to reach this problem. Certainly he does not do so directly, because nowhere in the book does he tell us what content is or how we fix it. What I am unsure about is whether his (what I later call) “content supplement” is intended to deal with the problem. Yaffe's “content supplement” is his view that we all intend not only “conditions” (his word) that are literally within the content of our intentions but also “those conditions that are true when an intention plays its proper causal role” (89). I think this content supplement is invented by Yaffe largely to deal with the problem of intention with respect to a circumstance element, and so I am dealing with it there. If it is also intended to deal with this classification-of-descriptions problem within literal content, I do not see how, but I could be missing something. As I read Yaffe, he thinks that “the account is very straightforward” and unproblematic when a particular act-token instantiates a type of act intended (74), so he sees no description-of-types problem.

4. Yaffe's Disassembling of Intentions’ Content into Different Pieces

However one solves the interpretive problem just discussed, there remain other problems about the content of intention required if one is to be guilty of trying (and thus of criminal attempt). One set of problems stems from the long-noticed fact that the content of intentions and other Intentional states has “parts” (Yaffe calls them “conditions,” or, in legal contexts, “elements”) (74–82), and that “intention” does not seem to apply univocally to all parts of its content. “Intent” with respect to lack of consent in rape, for example, prima facie seems to mean something different from what it does with respect to sexual contact with the victim.

One cannot even get started examining this issue without some taxonomy of the different kinds of elements there might be within an intention's content, a fact that Yaffe recognizes by devoting considerable attention to the issue (74–82). Yaffe taxonomizes the elements of content into the acts, results of acts, circumstances in which acts cause results, and mental states with which such acts are done. Before examining this fourfold classificatory scheme, follow Yaffe in asking what level or kind of robustness should be claimed for these distinctions.

This question of robustness is raised by the long tradition in the philosophy of action that regards these distinctions not to be “carving nature at the joints” but rather as justified by the practical purpose for which these distinctions are drawn. Yaffe agrees with this tradition, as do I.Footnote ³² Yaffe also sees that just because such distinctions are drawn in context-sensitive ways and for interest-relative reasons, does not mean that they cannot have sharp edges: “there is . . . a right and a wrong way to classify the elements” (75).

So far, so good. But Yaffe goes awry with the policy on which he rests his distinctions between kinds of elements. He says, “the question of how to classify a particular element just is the question of what the prosecution bears a burden to show beyond a reasonable doubt” (75). This may remind one of a familiar explication, favored by some torts scholars, as to who is right between Chief Judge Benjamin Cardozo and Judge William S. Andrews in Palsgraf v. Long Island Railroad Co.,Footnote ³³ on whether the problem for Mrs. Palsgraf's recovery is a problem of duty or one of proximate causation: the tort scholars’ view is that if a judge should decide the question at issue (of foreseeability of the harm to the defendant), then the issue is one of duty; whereas if the issue should go to the jury, the issue is one of proximate causation. This, because duty questions are for judges, causation questions for juries.

Yaffe sounds like he is making this kind of lawyerly, legal-process, procedural kind of move, whereby some procedural consequence of drawing some distinction is used as the only criteria guiding the distinction. But he is not. It turns out Yaffe's classifications of elements are not being guided by who should have the burden of proof (for on all four kinds of elements the prosecutor bears the burden of proof), nor does Yaffe's classification seek guidance from what kind of burden (production or risk of nonpersuasion) either party should have—because again, the prosecutor uniformly has both kinds of burdens. Rather, Yaffe is concerned with what the prosecutor has the burden of proving. For example, Yaffe says about the result/circumstance distinction, “For every result element of a crime, the prosecution is required to show . . . that a certain event occurred, and further, that the defendant caused that event. The existence of this second burden is what distinguishes result elements from circumstance elements” (76).

Yaffe is thus simply mistaken in his denial “that the classification is to be determined by the nature of the event or state in question” (76). A result element (on Yaffe's own classification) is an event that exists because it is caused to exist by the defendant's act, and a circumstance element exists when the defendant acts but it is not caused to exist by the defendant's act. It is only because of this difference in the nature of results versus circumstances that there is a difference in “the burdens that the prosecution bears with respect to that element” (76), Yaffe wrongly has it the other way around.

My own view of the right policy by which a rational penal code classifies elements of offences into two, three, or four kinds is the policy of aiding in the assessment of culpability by the kind of mental state that makes one culpable. Are there generalities to be had about the sorts of mental states that make one culpable (to a certain degree) for elements of type X, and do those generalities not apply (even if others do) to elements of type Y? If the answer to this question is in the affirmative, then we have reason enough to separate elements of type X from elements of type Y, even though X and Y may both appear in the actus reus and content of intentions prohibited by common criminal codes.

Without duplicating what I argue for elsewhere,Footnote ³⁴ I think Yaffe's mental state/circumstance and act/result distinctions to be untenable on the basis of this policy. It is not that we cannot distinguish such things; only that the relevant policy gives us no reason to do so. I affirm the result/circumstance distinction, however, and will thus divide my remaining discussion of content between these two kinds of elements only.

5. Yaffe on Intending the Result-Element Part of Content

It is time we give Yaffe his due on what his detailed view is as to the commitment part of his guiding commitment view of trying. Above I make do with Yaffe's intuitive approximation of the guiding commitment view and with equating commitment in the view with intention. Now we need to drop the intuitive approximation and to state more completely Yaffe's detailed view.

Yaffe speaks of “intention-based commitments,” rather than “intentions” simplicatur, as the commitment part of the guiding commitment view of attempts. He does this because of the disjunctive nature of what it is to be committed in the way one has to be to be an attempter of some criminal action A, according to Yaffe. The obvious way to be committed to each circumstance or result element E₁, E₂, . . . E_n making up A is by having that element in the content of the intention to do A. The less obvious way to have some element E _{n + 1} committed to by the attempter is when E _{n + 1} is not in the content of the intention, but nonetheless, if the intention to A is to play its proper causal role (in causing the result elements in A), then E _{n + 1} will necessarily exist (72–74, 82–89).

Remember that commitment, for Yaffe, is not a psychological idea. To commit oneself, for Yaffe, is not to steel oneself, grit one's teeth, make a resolution, and so forth. Rather, commitment is a logical relation in Yaffe's usage. To be committed in this sense is to be bound by the norms of rationality distinctive of intentions. One is so bound, according to the view, either because of the content of one's intention or to those things that must be true if the things referred to in the content of one's intention are caused by one's acts to be achieved.

Yaffe uses the content of these norms of rationality distinctive of intentions to spell out more precisely this second disjunct of how one is committed by an intention. Some attempter is committed by his intention:

Yaffe also believes that these three non-content-based commitments are also governing of what is intended (as is of course the content of an intention) and that thus, in some extended sense of “content,” one might say that what one is committed to in these ways is also part of the intention's content (88–89). Still, he recognizes the lumpiness of a notion of content so extended; his preferred locution is to speak of an intention's literal content, on the one hand, and on the other hand, of the things to which one is committed by virtue of having an intention with such a content (89).

This complicated machinery works well enough for the result elements in content. Take one of Yaffe's examples:

As long as the element death is not included in the intention's content, and as long as it is not “logically necessary” (89) if the intention to injure plays its proper causal role, then the agent has no intention-based commitment to the victim's death, has not tried to cause such a death, and is not guilty of attempted murder. Whereas if D intended to stop Victim's heart and lungs and to flatline his brain, then since there is no possible world where this can be done without killing Victim, D will be taken to intend Victim's death. Likewise, where D intends to inherit Victim's wealth and does an act likely to kill Victim because of that intent, D will also be held to intend Victim's death—this, either because D has two intentions, one for means and one for ends, and Victim's death is within the content of D's means-intention; or because it is not possible to inherit Victim's money without Victim being dead.

That, as Yaffe later shows (chap. 4), squares well with the common law of attempt, with its “specific intent” requirement. Things get stickier for the view once we move to what I am calling the “very certain foresight cases.”Footnote ³⁵ I refer to cases like this: Herod desperately wants to please Salome, who will be pleased only if Herod delivers John the Baptist's head on a platter to her. So Herod orders his underlings to cut off John's head. Herod sincerely reports that he did not intend John to die in the giving of his order; he only intended that John's body be separated from his head so that the latter could be delivered on a platter to Salome. Herod truly liked John and thus regretted what he foresaw with certainty would happen once John was decapitated, namely, that John would die. Herod truthfully claims that he would be delighted if he could achieve what he intended (to cut off John's head) and still have John survive (out of sight of Salome, presumably).

Such cases are not well handled by Yaffe's guiding commitment view. But then, such cases are not well handled by any other view of intentions (or of “intention-based commitments”). At least Yaffe is honest about this (119–124). Compare Yaffe's forthright confession of discomfort with the early attempt by Antony Duff to reconcile these cases with his own theory of intention, trying, and attempt. Duff also imagines a victim, Brown, whom D intends to decapitate. Duff at one time urges that there is a “logical connection” between decapitation and dying (or killing): “‘Brown is decapitated but survives’ does not specify an intelligible possibility, since it is part of the logic of our concept of ‘human beings’ that decapitation kills them: if we could imagine a being who was not killed by decapitation, that would not be a human being.”Footnote ³⁶

If Duff is right here, the second disjunct of Yaffe's commitment criterion (the causal-role disjunct, or what I call his “content supplement”) would classify both the death of John the Baptist and the killing of him as part of the content of what Herod intended.

This is the conclusion wanted, for it is counterintuitive (both to our folk psychology of intention and to the conclusions about moral culpability built on that folk psychology)Footnote ³⁷ to think that Herod did not intend John's death. Yet Duff's “logical connection” assertion is surely false (and this even granting his elevation of the semantics/pragmatics of expressions into “logic”). Just as in Yaffe's earlier example, D can intend to injure victim without intending to kill victim (because there are possible worlds where the injured do not die), so, too, Herod can intend to decapitate John without intending to kill John (because there are logically possible worlds where the decapitated do not die).

If this last gives you pause, ask whether the cryogenically economical amongst us—namely, the folks who, to save money, freeze only their separated heads for future revival when science advances sufficiently to do so—are making some logical mistake? And while they may be making some scientific mistake—the laws of our world may turn out so that we can never be “heads in a vat”—this is far from obvious either; and in any case, such scientific mistakes would render the scenario imagined (where John is decapitated but survives) only nomologically but not logically impossible, and Yaffe is clear that the guiding commitment view focuses only on the latter and not the former (120).

So Yaffe is like Duff and the rest of us: prima facie, our views on intentions commit us to classifying deaths and killings like that of John the Baptist as unintended. Given the necessity of intending a causing of death/killing in order for one to be guilty of either a (moral) trying or a (legal) attempt, we are all also committed to Herod not being guilty of attempted murder (or of aiding and abetting attempted murder) if Herod's underlings botch their attempt to cut off John's head.

Yaffe rightly sees that no plausible theory of intention, trying, or attempt can live with these conclusions. The question is how to avoid them. There are only three general approaches here, as I chart elsewhere in some detail.Footnote ³⁸ One is to bite the bullet on intention and admit that Herod does not intend to kill John. This is Yaffe's preferred strategy. Yaffe (like others who go this route) then seeks to avoid the moral/legal implications of this psychological conclusion by backtracking on his view that intention is required for legal or moral attempt liability. Yaffe rightly thinks that Herod is “just as bad” as someone who did intend John's death: “Sometimes knowing what will happen if one acts is just as bad as intending it,” recognizing that “usually it is not” (123). So we can blame Herod as if he were an attempter.

There are two problems with this first strategy as it is used by Yaffe. One is that it does not rescue the folk psychology from an implausible conclusion generated by the commitment criterion for intention: this strategy still admits that Herod did not intend to kill John. The other problem is that the strategy is incomplete at a crucial juncture, namely, how do we separate the “usual” cases where foresight is not “just as bad as intent” from cases like Herod's where it is? Yaffe sees this second problem and admits that “it is very hard to specify the precise conditions under which foresight is just as bad as intent” (123), Nonetheless, manfully enough, he has a go at it in terms of three factors: “All of the following factors, and perhaps more, probably matter to the calculation: the severity of the foreseen harm, the actual likelihood of its coming to pass, and the degree to which the intended act is itself wrong independently of the foreseen harm” (123).

I think that pretty much on their face, none of these factors is up to the work Yaffe assigns them here. The first and third, true, get more blood to our eyes as we judge such cases; yet we confuse apples and oranges if we let our judgment of heightened blame stemming from greater harm or number of wrongs done affect our judgment of the culpability of the actor who did such wrongs or caused such harm. Even though in attempts the “wrongdoing is in the mind of the actor” because the wrongs were not completed and the harms not caused, even so the greater culpability we do rightly sense to exist is due to wrongdoing-related factors, not the culpability of foresight being equal to the culpability of intention.

Suppose Herod orders the decapitation of a lowly snail darter; does the lesser severity of the harm foreseen—death of the snail darter—in any way change the question, “Did Herod intend the death of the snail darter or only foresee it?” Or suppose the snail darter belongs to someone else, so that causing decapitation and its death is not only a wrong to the darter but also an independent wrong to its owner's right of property. Does that in any way affect the question of whether Herod intended the death of the darter or only foresaw it?

Lastly, consider the second factor, the “actual likelihood” of the darter's death following upon its decapitation. Actual likelihood is to be distinguished from Herod's subjective assessment of the likelihood, and once one sees that distinction, surely only the latter is relevant to culpability. Objective risk-imposition on some theories is a kind of wrongdoing,Footnote ³⁹ but on no theories is it a culpability enhancer.

The problem is that the other two strategies available to deal with the “very certain foresight” cases have weaknesses of their own. The second general strategy is to redo radically what we mean by “intention.” On this view, the noun and verb “intention” and “intend” join the adjective and adverb “intentional” and “intentionally,” as being disjunctive in their meaning. To intend some result R, on this view, would be either to commit oneself to causing R, or to foresee that R will certainly come about if one achieves some other result, X, that one does commit oneself to causing. Bentham headed in this direction insofar as he allowed that there could be “oblique” intentions as well as “direct” intentions. Yaffe rejects this view (121–122), as do I,Footnote ⁴⁰ for this bloated notion of intention both leaves behind what seems essential to it (commitment and guidance, for Yaffe) and is far too broad in what it classifies as intended.

The third strategy is one that Yaffe does not consider. This is the strategy that enlarges the content of what is intended in the following way: if D intends one result R₁ (say, a decapitation), and another result R₂ (say, death) is so close to R₁ that it would be regarded as absurd (even though possible) to intend one without intending the other, then to intend R₁ is also to intend R₂. This is known in the trade as “the closeness doctrine.”

Notice that this third strategy does not tinker with the nature of intending, as does the second view by making foreseeing a kind of intending. Nor does this third strategy adopt a moral equivalence (of intention to certain foresight), as does the first strategy adopted by Yaffe. Indeed, this third strategy does not depend on actors like Herod foreseeing results of their actions like death in order to be said to intend those results. This “extender” of intention's content is not foresight but closeness.

The closeness strategy has a fifty-year history in recent ethical thought.Footnote ⁴¹ There are at least ten proposals for what is meant by “closeness” in this context, ranging from spatiotemporal proximity, a mereological or constitutive relation, an “invariably connected” relation, various subjective, logical, or moral relations, and so forth. Despite the number of philosophers who have taken up a lance on this question over these last fifty years, one could hardly claim consensus. I myself plump for the kind of “plain man” or “common sense” criterion of closeness also adapted by Jonathan Bennett:

Even those of us who adopt the third strategy are only moderately enthusiastic about it. As Bennett concedes, the idea of closeness is “rather loose.” Or as another proponent of the closeness strategy puts it, “sadly, the key notion (closeness) is vague and the salient principle (of intent consolidation) is a little ‘hinky.’”Footnote ⁴³

It would not take much of a stretch to modify Yaffe's view so as to accommodate the closeness strategy. For Yaffe's “commitment criterion” already has a content supplement in place in its causal role disjunct. All he need do is to expand this supplement so that “plain-man-regarded inconceivability” replaces logical impossibility as the test of what else is included in an intention's content besides what is literally there. Then D who intends injury is not held to intend death but Herod who intends decapitation is held to intend death. This transforms Yaffe's answer from being an instance of the first strategy to being an instance of the third. This has the advantage of all versions of the third strategy, which is to align the conclusions of the folk psychology of intention (e.g., “Herod intended to kill John”) with intuitive moral conclusions about culpability (e.g., “Herod has the culpability of one who intends death”). The cost, of course, is the unsatisfying “ad hoc-ness” of the “closeness” (or “plain-man-inconceivability”) test, a test that is not nearly as crisp as the line between logical and nomological impossibility.Footnote ⁴⁴

6. Intention with Respect to a Circumstance

Yaffe (like everyone else who has seen this issue) realizes that intentions with respect to circumstance elements of content look different from intentions with respect to result elements. Indeed, Bentham thought the differences to be so pronounced that he held that there was no such thing as an intention with respect to a circumstance.Footnote ⁴⁵ In this he was mistaken,Footnote ⁴⁶ but he was not wrong in intuiting that something is different in intentions with respect to circumstances.

One can see the intuitive difference of intentions with respect to circumstance (over and against consequences) by attending to the intention/foresight distinction. As shown in the last subsection, maintaining the line between intention and predictive belief, or “foresight,” is crucial to having a viable concept of intention. Yet this is just the line that threatens to erode in the case of circumstances.

The problem is not that we cannot draw a line psychologically between intention and belief vis-à-vis circumstances. We can; consider an example built loosely on the facts of an actual case arising in the federal prison at Marion, Illinois.Footnote ⁴⁷ In United States v. Fountain, the defendant, a lifer without possibility of parole in a jurisdiction lacking a (constitutional) death penalty, is engaged in a guard-killing contest. He is tied with another inmate at three guards apiece. He tries to kill a fourth guard, and is prosecuted for assault with intent to kill a correctional officer. That the would-be victim was a correctional officer is invariably classed as a circumstance element of this crime's actus reus. Defendant not only believed that his would-be victim was a correctional officer; he also was motivated to do what he did because the would-be victim was a correctional officer. Indeed, defendant would not have tried to kill this person if he were not a correctional officer. So we can rightly and easily say of this defendant that he intended (to kill a correctional officer). And such a defendant is plainly and easily distinguishable from the more typical defendant who tries to kill another merely knowing that the other is a correctional officer but being motivationally indifferent to that fact. This distinction is as clear as it is between two kinds of attempted rapists: those who merely know the victim is not consenting and those who want only forced sex, that is, they are motivated by the lack of the victim's consent.Footnote ⁴⁸

The problem is that hardly anyone wants to use this robust-enough psychological distinction for purposes of assessing the responsibility of attempters (and of others charged with other specific intent crimes). This is because our Aquinas-inspired culpability base in “setting one's will towards evil,” so clearly present in the case of intended consequences and so clearly absent vis-à-vis merely foreseen consequences, seems to lose its potency when applied to circumstances. As I often put it, circumstances are the poor relation when it comes to grading culpability, in that distinctions that matter morally vis-à-vis consequences do not seem to matter nearly so much vis-à-vis circumstances.

Criminal lawyers are left wondering what they are supposed to do with the codified distinction between intention (most serious legal culpability) and knowledge (less serious legal culpability) in light of this evaporation of a moral difference between the two with respect to circumstances. Criminal law—both the Model Penal Code and the common law that preceded it—takes a plausible enough position on this issue: treat belief about a circumstance as if it were intention with respect to that circumstance. If the only alternative to this standard legal answer were the restriction of attempts and other specific-intent crimes to cases like that of my motivated cop killer, then the standard legal answer looks pretty good.

Yet Yaffe wants to do better. He seeks to show us how one can preserve an intention/belief distinction with regard to circumstances without restricting attempts to those motivated by the circumstance in the way my imagined would-be cop killer and would-be rapist are. Yaffe seeks, in other words, to give sense to the idea that a circumstance can be intended (and not merely believed to exist) and yet that intention need not amount to its holder being motivated (“moved to act”) by the existence of such circumstance. Yaffe rejects the criminal law's abandonment of any intention/belief distinction in reference to circumstance elements. He does this because he does not think it impossible to articulate what an intention with respect to a circumstance might be in a way that does not collapse an intention (to cause R in C) into: (1) an intention (to cause R) when (2) the actor believes (C exists). Given the focus of the guiding commitment view on “intention-based commitments,” what Yaffe seeks in his own language is to “know the conditions under which a person can be rightly said to have an intention-based commitment to a circumstance . . . [for] it is that commitment that is needed for attempt” (142).

There are two criteria by which we should judge whether Yaffe succeeds in this task. One is a psychological criterion: Does Yaffe's idea of what intention with respect to a circumstance is square with a plausible conception of intention in philosophical psychology? The second is a moral criterion: Is intention with respect to a circumstance (as Yaffe conceives it) relevant to moral culpability? Does it, for example, heighten the culpability of actors who have it over the culpability of actors who only believe that a circumstance is present as they act so as to further some intention?

Consistent with his general theory, Yaffe here repairs to the commitment criterion part of the guiding commitment view with which he analyzes trying. Recall that the commitment criterion is disjunctive; there will be an intention-based commitment to C if:(1) that circumstance C is within the content of the actors intention; or (2) C must exist if that intention is to play its proper causal role. (132). That presumably means here that some circumstance C can be said to be intended by some defendant D if either C is within the literal content of D's intention, or C is within (what I earlier called) Yaffe's “content-supplement,” namely, C has to exist for D's intention to “play its proper causal role” (which is to cause, via appropriate actions, states of affairs matching the content of that intention).

As I read Yaffe, all the heavy lifting for his view (“The View”) about circumstance- elements, is done by the second disjunct of his commitment criterion. Before getting to that main show, however, let us pause to examine Yaffe's first disjunct: What does (would) Yaffe mean by a circumstance being within the literal content of an intention? He could mean just the kind of case adverted to above, where the cop killer is motivated by the identity (as a cop) of his victim or the rapist is motivated by the refusal (lack of consent) of his victim. These cases surely do present clear examples of intent with respect to a circumstance, so their inclusion here is unproblematic.

But are there not other kinds of cases where the circumstance is also included in the literal content of an attempter's intention? For example, the inmate intends (to kill the tallest correctional officer) using “correctional officer” as an incidental means of picking out Jones, his intended victim, but intends to kill Jones because Jones makes short-person jokes, not because Jones is a correctional officer. Is the circumstance (of the correction-officer status) not included within the literal content of the inmate's intention? Given that Yaffe never tells what he thinks content is or how it is fixed, it is hard to say what he thinks about this kind of case. Since who or what intentions are about can be referred to by definite descriptions as easily as by proper names, whatever one's view of content, are not incidental descriptions like “correctional officer” part of such content?

Psychologically this would not be an untoward result. Intentional states do take objects in these ways, and on some views of those objects, at least, the descriptions “in the actor's head” are the ones to use in formulating the (literal) content of what he intends. But morally this result seems problematic. For it is implausible to think that the inmate is less culpable if he refers to his victim as simply “Jones” than he is if he refers to that victim as “the tallest correctional officer.” Morality should be blind to such luck-driven considerations. How Jones is picked out on a given occasion by someone in forming an intention is surely something of a lottery, the outcome of which seems morally inert.Footnote ⁴⁹

Since Yaffe does not advert to this problem, I leave it and turn to the other disjunct of his analysis of intention's content. Let me begin where Yaffe ends his analysis of this issue, with his summary of his position:

Let us examine this analysis one piece at a time, starting with Yaffe's idea of an “intention-based commitment” vis-à-vis circumstance elements. He unpacks this in terms of two commitments specific to having an intention vis-à-vis a circumstance: (1) one is committed “to not reconsider one's intention on the grounds that one believes the condition to hold”; and (2) one is committed “to not complaining that the world fails to be as intended in light of the fact that the condition holds” (148). So the imagined inmate who is not in a guard-killing contest but who knows that his intended victim of homicide is a guard, will intend (to kill Jones who is a guard) if and only if the inmate's intention to kill Jones commits him: (1) not to reconsider (before he kills Jones) whether to do so in light of the inmate's belief that Jones is a guard; and (2) not to regret (after he has killed Jones) that he did not succeed in doing what he intended because Jones turns out to have been a guard.

I think Yaffe is on solid ground in thinking that one who intends (to do some act A to some person or object that has some property P) makes the two commitments he describes. Intentions are generally “sticky” against reconsideration of the overall desirability of what we intend, and one can hardly complain that he did not succeed in executing his intention on the grounds that the world is what he intended it to be. More interesting is whether these two commitments thread the needle in the way Yaffe recognizes they must. Clearly Yaffe meets one parameter of the “eye” of this needle: these commitments to a circumstance existing “fall short of [being] commitments to causing that circumstance's occurrence.” But are these commitments yet strong enough that they exceed like commitments already in place by virtue of an agent merely believing (but not intending) that the act A he intends to do will be to a person or object that possesses property P? This is crucial for Yaffe if he is to distinguish his view from the standard legal analysis that regards intention with respect to a circumstance to be no more than and no different from belief with regard to that circumstance.

I defer the issue until we complete our account of Yaffe's full analysis. Next, what does Yaffe mean by his thought that one who intends vis-à-vis some circumstance must be “moved by” the two commitments he describes? The etymology of “motivation” is from “moved,” yet Yaffe cannot mean that the actor must be motivated by the existence of a circumstance; for that is the narrow view that one intends in reference to a circumstance only when one is motivated to act by its presence. That is the inmate who is in a cop-killer contest and his ilk, but not anyone else.

So Yaffe cannot mean what he says here. What could he mean? Could he mean that the commitments he describes must be transferred from being merely the implications of norms of rationality distinctive of intentions to being psychological commitments of identical content? That at least would put the commitments “in one's head” (and not merely the logical implications of what is in one's head). Yaffe does want to make this move anyway, so long as he wishes to infer what is intended from what kind of commitments are being made. Commitment to nonreconsideration and nonregret can be merely the logical/rational implications if they are to be merely what follows from the content of one's intentions; they must themselves be psychological states if they are to be the basis of inferring what it is one intends. (The latter idea is that we, the observers, see these psychological commitments and infer that the only state that could rationally generate them must be a state of intention, not mere belief.) More on this below, too.

Now (and most crucially) consider what Yaffe means by “intention” in his phrase, “intention-based-commitment.” (Watch Yaffe's hands closely here, because this is where the “sleight of hand” I mention above occurs.) Yaffe believes that there are two ways to characterize the objects of intentions, de re and de dicto. He treats these as two kinds of intentions, “de re intentions” and “de dicto intentions” (142), which means he treats the distinction as an internal one, being between two ways of having an intention (and not as an external distinction, between two ways for an observer to regard one and the same intention in someone else).

Yaffe rightly sees that intentions can more often and more easily be internally de re regarding the objects and their properties that make up circumstance elements than they are internally de re regarding acts and results. Indeed, according to Yaffe, “it is so natural to give the de dicto interpretation of the descriptions of acts and results that appear in the contents of intentions that it sometimes takes some mental gymnastics to even see what it would be to interpret them de re” (147). Yaffe is right about this because of the fact adverted to above: the cross-world identity conditions are much more robust for objects such as persons than they are for events.Footnote ⁵⁰

In any case, an example of an internally de re intention for Yaffe would be: Jones intends (to steal one million dollars from the governor of California). Suppose Jones believes Sylvester Stallone to be the governor of California (Stallone does, after all, make movies just as bad as Arnold Schwarzenegger's). An intention that is internally de re with regard to the description of the circumstance “governor of California” would be an intention to steal from Stallone; for de re, it is the referent that counts, the description being but one of many ways usable by an intender to refer to Stallone. Whereas if the intention were internally de dicto regarding this element, then Jones would be intending to steal the monies from whomever is governor of California, which is not Stallone.

Yaffe's analysis focuses on the intention-based commitments that are based on intentions internally de re as to the circumstance elements in their objects. For, as he rightly sees, commitments arising out of intentions internally de dicto with regard to such elements would be too narrow: they would include only commitments to causing these conditions to exist (147, 149), and this is the strong commitment that Yaffe is trying to avoid in his analysis.

In truth, I thought a major fudge was going to occur in Yaffe's analysis at this point. I thought he was going to argue, for intentions internally de re as to circumstance elements (like the gubernatorial status of the victim of Jones's intended theft), that if the circumstance exists in fact, then the actor will be taken to intend that it exist (irrespective of what he believes about the existence of that circumstance). If Jones intends to steal from Schwarzenegger, and Schwarzenegger is in fact governor of California, then Jones necessarily intends to steal from the governor of California and can be convicted of a crime having, “intent to steal from the governor of California” as its mens rea.

Yet this would be to confuse the external sense of the de re/de dicto distinction with its internal sense, and this is far too crude and unsophisticated for YaffeFootnote ⁵¹ (if not for everyone who has written about attempts).Footnote ⁵² His move is much subtler and, correspondingly, much more interesting.Footnote ⁵³ Imagine two would-be thieves accused of attempting the crime whose actus reus is “stealing from the governor of California:”

• D₁ intends internally de re (to steal from the governor of California), and D₁ believes that the guy he intends to steal from, Arnold Schwarzenegger, is the governor of California, which he is.

• D₂ intends internally de re (to steal from Arnold Schwarzenegger) and D₂ believes that Arnold Schwarzenegger is the governor of California, which he is.

As a first cut (I discuss Yaffe's second cut below), Yaffe believes that there is a “psychological difference” (145) between D₁ and D₂ such that only D₁ has an intention with respect to the “governor of California” circumstance, only D₁ has intention-based commitments (of nonreconsideration and nonregret) with respect to that circumstance, only D₁ has tried to steal from the governor of California, and only D₁ is guilty in law of attempted stealing from the governor of California.

We need to assess two things here: first, whether there is any “psychological difference” between D₁ and D₂; and if there is, second, whether this psychological difference makes any moral difference in the culpability of D₁ as compared to D₂. Let me examine the psychology first.

Yaffe assures us that “there must be” a psychological difference between D₁ and D₂ because “the world does not match the intention” of D₁ unless Schwarzenegger really is governor of California, and that is not true of D₂ (145). D₁'s intention thus commits D₁ to not reconsidering his contemplated theft in light of his belief that Schwarzenegger is governor of California, whereas it is only D₂'s belief that commits him to not reconsidering his contemplated theft—for D₂'s intention is silent on the question of whether Arnie is governor of California. Likewise, it is D₁'s intention that commits him to not complaining after the theft just because Schwarzenegger is governor of California, whereas if D₂ is also so committed, it is only by his belief, not by his intention.

I find the analysis so far by Yaffe to be a bit strange. I thought we were working with the second disjunct of Yaffe's “commitment criterion,” and that this second disjunct functioned to show how an element, even though not within the literal content of a defendant's intention, might nonetheless be part of the “content supplement” if that was required for the intention to play its proper causal role. Yet Yaffe's psychological difference between defendants like D₁ and D₂ is based on their having different (literal) contents to their respective intentions. If this is so, we do not need the complicated story provided by the second disjunct of the commitment criterion—the simple story of the first disjunct will do. If the circumstance element is within the content of D₁'s intention, well, then, it is! End of story—at least, so long as we have some psychological account of what literal content is and how it is fixed, matters about which Yaffe is silent.

I should have thought that the logic of Yaffe's argument train would seek to provide us with a commitment-based account of when some element not otherwise part of the literal content of one's intentions could nonetheless be shown to be within the content of that person's intention. And this, by showing how the commitments distinctive of intention (nonreconsideration, nonregret) are present in certain cases where the defendant seemingly has only a belief about such elements; the presence of these two commitments distinctive of intention would then compel the inference that the elements with respect to which such commitments are made must be intended.

Despite the unpromising beginning that Yaffe here makes—that is, when he just asserts that the circumstance element (like being the governor of California) is in the content of D₁'s intention—Yaffe does seem to use the second disjunct of his commitment criterion in the way just sketched. That is, he uses it as a reason for concluding there is intent with respect to a circumstance and not just a charting of what follows once it is established that an intention includes such circumstance in its content. Unfortunately Yaffe puts the argument I see him as making here in terms of an evidentiary inference: we generally can infer that someone intends a circumstance from the fact that he believes the circumstance to exist, because the commitments distinctive of intention are present in such run-of-the mill cases. But in four situations, Yaffe tells us (150–151, 154–155) the commitments distinctive of intentions will not exist, in which event the actor only believes but does not intend with regard to the circumstance (151–152). If we reverse Yaffe's order here and drop the pretense that we are making a mere evidentiary inference, Yaffe is equivalently saying: intention with regard to a circumstance exists when: (a) belief with regard to that circumstance exists; and (b) the actor is not in any of the four circumstances Yaffe describes. Thus, Yaffe does (sort of) make the argument I sketch above.Footnote ⁵⁴

The first thing to notice is that a belief that a circumstance exists when one forms an intention usually does generate the two commitments Yaffe thinks are distinctive of intentions with regard to a circumstance. For brevity of exposition, let me focus only on the first of these commitments, that against reconsideration. If D₂ believes that he will be stealing from the governor of California, at a minimum he has no reason to reconsider his decision just because he believes later what he believed when he formed his intention, namely, that the person he intends to steal from is the governor of California. But beyond not giving any reason to reconsider, the belief also represents a resolution of competing tendencies-to-belief, just as an intention represents a resolution of competing component desires. As such, beliefs have their own “stickiness” with regard to reconsideration, in the sense that rational beings with limited time and resources are committed to nonreconsideration of their theoretical “resolutions,” that is, their beliefs. So an intention (to steal from Arnie) coupled with a belief (Arnie is the governor of California) will generate the commitment to not reconsidering (the decision to steal in light of a belief that Arnie is governor of California) just as will the intention (to steal from the governor of California).

Yaffe cannot accept this as stated, for if he did, mere belief that a circumstance exists would always suffice for there being an intention that a circumstance exist. So Yaffe needs some restriction(s) on the kinds of beliefs that can generate the commitments (of nonreconsideration and nonregret) that are uniquely possessed by intentions and that, when possessed by beliefs, transform those beliefs into intentions. One such restriction is to beliefs the content of which describes a circumstance element because it is salient. Yaffe adds “salience” here,Footnote ⁵⁵ as a requirement to move from belief that a circumstance exists to an intention whose object includes that circumstance. To see what Yaffe means by “salience,” continue with our present example. D₂ intends to steal from Arnie, and he believes that Arnie is governor of California. Suppose now that D₂ “could not care less that Arnold Schwarzenegger” is or is not governor of California; it is no more important to him than is the fact that Arnie has an Austrian accent (150). In referring to Arnie, D₂'s use of “the governor of California” is then like his use of “the Austrian-accented person”—in both cases, the circumstance is incidental, randomly selected, unconnected to any interest in D₂'s intended victim as either governor or Austrian in origin. As Yaffe puts it, in such cases being the governor of California is not salient, and, “the insalience” of this fact “blocks the inference” from the intention/belief pair—D₂'s intention to steal from Arnie coupled with D₂'s belief that Arnie is governor of California—to D₂ having the intention (to steal from the governor of California) (151). The reason Yaffe gives as to why this inference is blocked is because this is a case “in which the agent falls short of a commitment to not reconsider the intention in the face of a belief that the relevant condition holds” (151). And this is because the person for whom Arnie being the governor of California “is completely insalient can . . . reconsider an intention” to steal from him (151).

Let me see if I can put this in my own words here. Yaffe seems to share the thought initiated by David KaplanFootnote ⁵⁶ and later seconded by me, both some time ago:Footnote ⁵⁷

Yaffe's salience, as I would construe, is very much like Kaplan's vividity.

The work done by salience/vividity for Yaffe in this context is, then, weakly motivational: our use of descriptions that are salient/vivid to us is not incidental because such use evidences how we think of some individual. Even though D₂ would steal from Arnie even if he were not the governor, perhaps “the governor of California” is a salient/vivid way of referring to Arnie for D₂ because D₂ takes pride in theft from such eminent personalities, or perhaps he basks in the “connection” he establishes to such eminence by stealing from the governor (think of John Lennon's assassin).

Yaffe's own last example of salience sustains this interpretation of Yaffe. Yaffe compares two would-be statutory rapists, where the relevant circumstance is not lack of consent but the girl's age. R₁ is an eighteen-year-old defendant of comparable emotional maturity to his intended victim V, who is fifteen. As Yaffe puts it, “the question is whether he [R₁] takes the minor status of his would-be partner to be at all salient” (153). On these facts, Yaffe's conclusion is “that the victim's age just isn't salient at all and so prevents the content of the belief [as to V's age] from making its way into the content of his intention”Footnote ⁵⁹ (153). Different for Yaffe is R₂, who is fifty years old and thus thirty-five years senior to V. This is different, Yaffe tells us, because “it is hard for us to imagine that the minor status of one's partner would be entirely insalient to a person of that age” (154).

It is time we take stock. Grant Yaffe that some of the descriptions that we use to refer to individuals are salient, others insalient. Grant further that such descriptions are salient/insalient for the reasons Yaffe gives. Even so, does salience of a description like “governor of California” license movement of that description from the object of belief to the object of intent? And does it issue such license because salience shows a commitment to nonreconsideration that insalient descriptions do not?

This last might be so. One might think that salient descriptions stay in the forefront of one's thoughts (Yaffe's “occurrent” rather than merely “dispositional” beliefs) in a way that insalient ones do not. Because of this, one might further think that salient descriptions carry greater commitment to nonreconsideration because the properties they describe are important enough for one to have decided with the firmness and stickiness that makes reconsideration irrational.Footnote ⁶⁰

So perhaps one can tease out a psychological difference here, a difference between the D₂ who only (insaliently) believes that the man he intends to steal from is governor of California and the D₂ who intends (to steal from the governor of California) because he saliently believes his victim is the governor. In truth, this is slicing the baloney very thin; but given the elusiveness of the problem and the dearth of solutions competitive to Yaffe's, we should give Yaffe his due here.

There are two problems, however. One is that insalience can be true of descriptions used in the literal content of intentions, not just of descriptions used in the contents of beliefs. In our example, D₁ who intends (to steal from the governor of California), could have the description “governor of California” only insaliently in his mind. This is indeed a possibility that Yaffe seems to recognize (151). Should we withdraw our classification of D₁ as having the required intention with regard to a circumstance when his description of that circumstance is insalient? I think Yaffe's answer is (and should be) “no,” but I am unsure that that is in fact Yaffe's answer.Footnote ⁶¹

The second problem is a moral problem. Yaffe says, “Recall that under the Guiding Commitment View to try to do something is to be committed by one's intention to each of the components of success” (149); also recall that such trying is indispensable to being blameworthy and sanctionable for an attempt. This means that the psychological difference that perhaps Yaffe has isolated must carry a bit of moral freight. The D₂ who saliently believes that Arnie is the governor of California goes to jail for attempted theft from the governor of California, whereas the D₂ who only insaliently believes this goes free (or is convicted of some lesser crime of knowing risk creation, if there is one). It is implausible to me that the outer boundaries of attempt liability should be drawn here.

Where that line should be drawn is a closer question. There is something to be said for classifying as attempters only those who are motivated by the existence of some circumstance to do what they do. Those are the internally de dicto intenders, the intenders who are committed to causing a circumstance to be present in one of the two ways Yaffe depicts (147). Alternatively, there is a lot to be said for the line drawn by current criminal law, which lumps all those who merely believe that a circumstance exists in with those who are motivated to do what they do by that fact, and punishes all as attempters. Yaffe's middle ground looks morally tenuous compared to these two alternatives, because the breakpoint in culpability looks less dramatic for his line when compared with these other two lines.

I thus conclude that Yaffe's case for construing intentions with respect to circumstances as he would do so is, in the language of the old Scottish verdict, “not proven.” However, one should not let the negative conclusion on this point detract from one's admiration for the sophistication, ingenuity, originality, intricacy, and carefulness of Yaffe's analysis. There is nothing comparable to this in the literature on this topic, perhaps because the topic is itself so devilishly complex. Moreover, the sheer talent and ingenuity that Yaffe has devoted to this topic allows a modest induction here: if this kind of sustained and sophisticated effort by Yaffe has not isolated a plausible middle ground for locating intention with respect to a circumstance, that is reason to believe that there is no such middle ground. This would mean that common law and the Model Penal Code have long had it right on this issue, a worthwhile conclusion in its own right.