Grice writes: "My purpose in this section is to give a little thought to the question 'What are the general principles exemplified, in creature-construction, in progressing from one type of pirot to a higher type? What KINDS of steps are being made? The kinds of step with which I shall deal here are those which culminate in a licence to include, within the specification of the content of the psychological state of certain pirots, a range of expressions which would be inappropriate with respect to lower pirots; such expressions include connectives, quantifiers, temporal modifiers, mood indicators, modal operators, and (importantly) names of psychological states like "judge" and "will". Expressions, the availability of which leads to the structural enrichment of specifications of content. ... In general, these steps will be ones by which items or ideas which have, initially, a legitimate place outside the scope of psychological instantiables (or, if you will, the expressions for which occur legitimately outside the scope of psychological verbs) come to have a legitimate place within the scope of such instantiables: steps by which (one might say) such items or ideas come to be internalised. I am disposed to regard as prototypical the sort of natural disposition which Hume attributes to us, and which is very important to him; name, the tendency of the mind 'to spread itself upon objects' to project into the world items which, properly (or primitively) considered, are really features of our states of mind. I shall set out in stages the application of aspects of the genitorial programme.” We then start with a zero-order, with pirots equipped to satisfy unnested judging and willing (i.e. whose contents do _not_ involve judging or willing). We soon reach

Pirot-1. "It would be advantageous to pirots-0 if they could have judging and willing, which relate to their own judging or willing." Such pirots (pirots-1) could be equipped to control or regulate their own judgings and willings. They will presumably be already constituted so as to conform to the law that caeteris paribus if they will that p and judge that ~p, if they can, they make it the case that p in their 'minds'. To give them some control over their judgings and willings, we need only extend the application of this law to their judging and willing. We equip them so that caeteris paribus IF they will that they do not will that p and judge that they do will that p, (if they can) they make it the case that they do NOT will that p. And we somehow ensure that sometimes they CAN do this. It may be that the installation of this kind of control would go hand in had with the installation of the capacity for evaluation."

Pirot-2. Unlike it is the case with a pirot-1, a pirot-2's intentional efforts depend on the motivational strength of its considered desires at the time of action. We have been seeing the process by which conflicting considered desires motivate action as a broadly causal process, a process that reveals motivational strength. But a pirot-2 might itself try to weigh considerations provided by such conflicting desires in deliberation about the pros and cons of various alternatives. In the simplest case, such weighing treats each of the things desired as a prima facie justifying end. In the face of conflict it weighs such desired ends, where the weights correspond to the motivational strength of the associated considered desire. The outcome of such deliberation will match the outcome of the causal motivational process envisioned in our description of a pirot-2. But since the weights it invokes in such deliberation correspond to the motivational strength of the relevant considered desires (though perhaps not to the motivational strength of the relevant considered desires), the resultant activities will match those of a corresponding pirot-2 (*all* of whose desires, we are assuming, are considered). To be more realistic we might limit ourselves to saying that a pirot-2 has the capacity to make the transition from unconsidered to considered desires but does not always do this. But it will keep the discussion more manageable to simplify and to suppose that *all* its desires are considered.

Pirot-3. We shall not want these pirots-2 to depend, in each will and act in ways that reveal the motivational strength of considered desires at the time of action, but for a pirot-3 it will also be true that in some (though not all) cases it acts on the basis of how it weights the ends favoured by its conflicting considered desires. Pirot-3's considered desires will concern matters that cannot be achieved simply by action at a single time. Pirot-3 may, for example, want to nurture a vegetable garden, or build a house. Such matters will require organized and coordinated action that extends over time. What the pirot-3 does now will depend not only on what it now desires but also on what it now expects it will do later given what it does now. It needs a way of settling now what it will do later given what it does now. The point is even clearer when we remind ourselves that pirot-3 is not alone. It is, we may assume, one of some number of pirots-3; and in many cases it needs to coordinate what it does with what other pirots-3 do so as to achieve ends desired by all participants, itself included.

Pirots-4. These costs are magnified for a pirot-4 whose various plans are interwoven so that a change in one element can have significant ripple effects that will need to be considered. Let us suppose that the general strategies pirot-4 has for responding to new information about its circumstances are sensitive to these kinds of costs. Promoting in the long run the satisfaction of its considered desires and preferences. Pirot-4 is a somewhat sophisticated planning agent but it has a problem. It can expect that its desires and preferences may well change over time and undermine its efforts at organizing and coordinating its activities over time. Perhaps in many cases this is due to the kind of temporal discounting. So for example pirot-4 may have a plan to exercise every day but may tend to prefer a sequence of not exercising on the present day but exercising all days in the future, to a uniform sequence the present day included. At the end of the day it returns to its earlier considered preference in favour of exercising on each and every day. Though pirot-4, unlike pirot-3, has the capacity to settle on prior plans or plaices concerning exercise, this capacity does not yet help in such a case. A creature whose plans were stable in ways in part shaped by such a no-regret principle would be more likely than pirot-4 to resist temporary temptations.

Pirot-5. So let us build such a principle into the stability of the plans of a pirot-5, whose plans and policies are not derived solely from facts about its limits of time, attention, and the like. It is also grounded in the central concerns of a planning agent with its own future, concerns that lend special significance to anticipated future regret. So let us add to pirot-5 the capacity and disposition to arrive at such hierarchies of higher-order desires concerning its "will".

Pirot-6. This gives us a new creature, pirot-6. There is a problem with pirot-6, one that has been much discussed. It is not clear why a higher-order desire -- even a higher-order desire that a certain desire be one's "will" -- is not simply one more desire in the pool of desires (Berkeley God's will problem). Why does it have the authority to constitute or ensure the agent's (that is, the creature's) endorsement or rejection of a first-order desire? Applied to pirot-6 this is the question of whether, by virtue solely of its hierarchies of desires, it really does succeed in taking its own stand of endorsement or rejection of various first-order desires. Since it was the ability to take its own stand that we are trying to provide in the move to pirot-6, we need some response to this challenge. The basic point is that pirot-6 is not merely a time-slice agent. It is, rather, and understands itself to be, a temporally persisting planning agent, one who begins, and continues, and completes temporally extended projects. On a broadly Lockean view, its persistence over time consists in relevant psychological continuities (e.g., the persistence of attitudes of belief and intention) and connections (e.g., memory of a past event, or the later intentional execution of an intention formed earlier). Certain attitudes have as a primary role the constitution and support of such Lockean continuities and connections. In particular, policies that favour or reject various desires have it as their role to constitute and support various continuities both of ordinary desires and of the politicos themselves. For this reason such policies are not merely additional wiggles in the psychic stew. Instead, these policies have a claim to help determine where the agent -- i.e., the temporally persisting agent -- stands with respect to its desires. Or so it seems to me reasonable to say.

Pirot-7. So the psychology of pirot-7 continues to have the hierarchical structure of pro-attitudes introduced with pirot-6. The difference is that the higher-order pro-attitudes of pirot-6 were simply characterized as desires in a broad, generic sense, and no appeal was made to the distinctive species of pro-attitude constituted by plan-like attitudes. That is the sense in which the psychology of pirot-7 is an extension of the psychology of pirot-6. Let us then give pirot-7 such higher-order policies with the capacity to take a stand with respect to its desires by arriving at relevant higher-order policies concerning the functioning of those desires over time. Pirot-7 exhibits a merger of hierarchical and planning structures. Appealing to planning theory and ground in connection to the temporally extended structure of agency to be one's "will". Pirot-7 has higher-order policies that favour or challenge motivational roles of its considered desires. When Pirot-7 engages in deliberative weighing of conflicting, desired ends it seems that the assigned weights should reflect the policies that determine where it stands with respect to relevant desires. But the policies we have so far appealed to -- policies concerning what desires are to be one's will -- do not quite address this concern. The problem is that one can in certain cases have policies concerning which desires are to motivate and yet these not be policies that accord what those desires are for a corresponding justifying role in deliberation.

Pirots-8. A solution is to give our creature -- call it pirot-8 -- the capacity to arrive at policies that express its commitment to be motivated by a desire by way of its treatment of that desire as providing, in deliberation, a justifying end for action. Pirot-8 has policies for treating (or not treating) certain desires as providing justifying ends -- as, in this way, reason-providing -- in motivationally effective deliberation. Let us call such policies self-governing policies. We will suppose that these policies are mutually compatible and do not challenge each other. In this way pirot-8 involves an extension of structures already present in pirot-7. The grounds on which pirot-8 arrives at (and on occasion revises) such self-governing policies will be many and varied. We can see these policies as crystallizing complex pressures and concerns, some of which are grounded in other policies or desires. These self-governing policies may be tentative and will normally not be immune to change. If we ask what pirot-8 values in this case, the answer seems to be: what it values is constituted in part by its higher-order self-governing policies. In particular, it values exercise over nonexercise even right now, and even given that it has a considered (though temporary) preference to the contrary. Unlike lower pirots, what pirot-8 now values is not simply a matter of its present, considered desires and preferences. Now this model of pirot-8 seems in relevant aspects to be a (partial) model of us. (in our better moments, of course). So we arrive at the conjecture that one important kind of valuing of which we are capable involves, in the cited ways, both our first-order desires and our higher order self-governing policies. In an important sub-class of cases our valuing involves reflexive polices that are both first-order policies of action and higher-order policies to treat the first-order policy as reason providing in motivationally effective deliberation. This may seem odd. Valuing seems normally to be a first-order attitude. One values honesty, say. The proposal is that an important kind of valuing involves higher-order policies. Does this mean that, strictly speaking, what one values (in this sense) is itself a desire -- not honesty, say, but a desire for honesty? No, it does not. What I value in the present case is honesty; but, on the theory, my valuing honesty in art consists in certain higher-order self-governing policies. An agent's reflective valuing involves a kind of higher-order willing.