Ishtiyaque Haji has developed an enormous number of Frankfurt-style cases that exhibit puzzles about moral responsibility. He presents the standard argument against free will and moral responsibility and then suggests that this "age-old grand puzzle" might be more tractable if we consider a more restrictive form of "responsibility" that he calls "appraisability".

Appraisability apparently grows out of Peter Strawson's notion of the "reactive attitudes" of blame and praise that humans would feel whether or not determinism or free will is the case.

Haji explains "moral appraisability" by identifying actions that 1) we normally regard as free (he calls this "volitional control"), 2) actions that are performed autonomously (which he defines as issuing from an "authentic evaluative scheme"), and 3) actions that the agent regards as morally obligatory, right, or wrong. (Moral Appraisability, 1998, p. 237)

In his later work Deontic Morality and Control, Haji identifies these last three attributes as "deontic" properties. They make the acts deontic acts.

Haji says he has identified three conditions that are necessary and sufficient for moral appraisability for intentional actions.

The conditions are recorded in principle Appraisability, which has three central constituents: a control constituent, which says that the sort of control required for moral appraisability is volitional control; an epistemic one, which — stripped down to its core — says that in order for an agent to be morally appraisable for an action, the agent must believe she is doing something wrong or morally amiss, or she is either executing her moral duty or at least doing what is morally permissible by performing that action; and finally, an authenticity constituent, which says that the agent's action for which she is appraisable must issue from actional springs that are "truly her own."
(Moral Appraisability, 1998, p. 237)

Like many of his modern colleagues, Haji is agnostic on the truth of determinism.

My last objective in this work is to motivate the suggestion that the three conditions laid down in the analysis of appraisability I defend are threatened neither by determinism nor by (certain varieties) of indeterminism. (p.viii)

While Haji focuses on morally appraisable actions, his work is not a form of restrictivism. He says it applies also to non-moral actions.

I suggest that the conditions registered in the analysis, with suitable amendments, capture conditions for nonmoral but other normative varieties of appraisability like etiquettical or prudential appraisability. One result of this exploration into varieties of appraisability that will undoubtedly strike many as controversial is that, even if appraisability is not undermined either by determinism or indeterminism, most of us most of the time are not morally appraisable for what we do. (p.viii)

His condition that the agent must believe she is doing something wrong, right, or obligatory implies that others not sharing that belief will make mistakes assigning praise or blame.

I discuss the implications of the analysis for what I dub "intersocietal" attributions of blameworthiness: people who are not part of a particular culture — "outsiders" relative to a culture — often attribute blame to a person within that culture for doing something regarded as morally repulsive by the outsiders. Relying on the epistemic element of my analysis — that blameworthiness requires belief in what is wrong (and not "objective" wrongness) — I argue that outsiders' attributions of blame are probably frequently erroneous. (p.viii)

Haji accepts that agents lacking control for reasons of addiction, hypnosis, manipulation, etc. are not morally appraisable. But then, in a surprise, he thinks agents are morally appraisable for their thoughts and even dreams. Note that in our two-stage model of free will, thoughts are the result of alternative possibilities generated in part indeterministically.

Falling back partly on conclusions I draw regarding appraisability for unconscious thoughts, I argue that we can be appraisable for some of the thoughts of our dream selves (at least if dreams are experiences involving mental activity). (p.viii)

In his book Deontic Morality and Control, Haji extensively examines Alfred Mele's proposal for a Modest Libertarianism.

In Mele's theory indeterminacy in the actional pathway leading to an agent's decision or other sort of action occurs relatively early at the juncture between the agent's deliberations about what to do and the formation of a best judgment regarding what to do (Mele 1995: ch. 12). In standard nondeviant cases of intentional action, the agent decides on the basis of such judgment, say, to perform some action and then acts accordingly, intentionally performing the action. It is the best judgments, on this view, that are undetermined. I argue that this sort of modest libertarianism, even if it accommodates some deontic morality, accommodates far too little. (p.89)

Indeterminism at the decision stage invites the randomness objection of the standard argument, which Mele calls the Luck Objection

Next, I turn to a more robust form of modest R-libertarianism in which indeterminacy is located relatively late in the pathway, culminating in action at the juncture between, roughly, the consideration of reasons and the formation of intentions or the making of decisions. Here, it is the intentions or decisions that are undetermined. My verdict will be that this more robust form of libertarianism is hospitable to deontic morality.

However, prior to being in a position to draw this verdict, a serious hurdle I previously thought could not be overcome (see Haji 1999a) will have to be cleared. It has been charged that the more robust version of R-libertarianism that I have distinguished succumbs to a powerful objection: In brief, if our choices are indeterministically caused, and it is the very intentions or decisions that are undetermined, then our choices are a matter of luck.

Even in a two-stage deliberation-decision process, Mele notes that random alternative possibilities are a matter of luck

For holding constant the conditions of the past that include the agent's powers, values, deliberations, and character, the agent could have made one choice just as easily as she could have made another. Moreover, with nothing about the agent's capacities, states of mind, reasons for action, and so on rationalizing this difference in outcome —as would be the case if the past were held fixed — the difference in outcome does seem to be a matter of luck. Luck of this sort appears to be incompatible with responsibility.

I will argue that this luck objection does in fact have bite. This concession may seem suicidal to the project of accommodating deontic morality by appealing to elements of a robust modest R-libertarian theory. For luck, it appears, is not only incompatible with responsibility, it is incompatible with deontic morality as well. How could one, for example, do moral wrong if all of one's actions were luck infected and so, it would seem, out of one's control?

John Martin Fischer criticized Daniel Dennett's "Valerian" proposal which adds indeterminism to the early stages of deliberation in order to enhance control. Fischer called it alchemy. He clearly thought it reduced control. Haji notes the same criticism applies to Mele, but he denies Fischer's claim of alchemy.

To respond to Fischer's charge of alchemy — the charge that adding indeterminacy as to which belief states will come to mind in an otherwise deterministic sequence that, without the added indeterminacy, would be considered by the libertarian to be one in which an agent lacks control — it will be helpful to remind ourselves, once again, of Mele's distinction between proximal and ultimate control. The former, as we have seen, concerns the direct causal production of agent-involving events. The latter, in contrast, involves the causal influence of agent-external events. In order for an agent S to have ultimate control over an event, say, S's x-ing at a time t, where x-ing might, for example, be the making of a decision, there must be no time at which there are minimally causally sufficient conditions that do not include events or states internal to S, for S's x-ing at t. In other words, to have ultimate control over S's making some decision, roughly, there shouldn't be conditions "external" to S that are minimally causally sufficient for S's making that decision. Hence, agents could have ultimate control over their actions only if determinism is false. But proximal control, according to Mele, is compatible with determinism (1995: 211).

Modest libertarians should not deny that an entirely deterministic "actional process" is one in which an agent lacks proximal control but need only deny that it is one in which the agent lacks ultimate control. Indeed, Mele insists that satisfaction of his compatibilist conditions for autonomous or responsible action, along with the proximal control that that involves, together with doxastic indeterminism, suffices for the agent's having ultimate control over the pertinent action. On his view, then, as he says, ultimate control, rather than requiring the possession of any special "control power" beyond the powers required for satisfaction of compatibilist conditions for responsible action, "is something one has in virtue of satisfying the compatibilist conditions and being suitably internally indeterministic" (1995: 213). Transformation of a deterministic actional process from one of lack of ultimate control to one containing such control by installation of the sort of internal indeterminacy that Mele recommends, should, consequently, not smack of alchemy.

---

Chapter 1.4 - The Philosophy	Chapter 1.6 - The Scientists
Home	Part Two - Knowledge

Normal | Teacher | Scholar