Plan Rationality

I have been discussing a planning agent’s self-governance, both at a time and over time. To begin this third and final post (though of course I’d be pleased to respond to further queries/challenges) I’d like to back up a bit. The planning theory behind these models of self-governance supposes that a planning agent’s practical thinking is guided by norms of synchronic plan consistency and means-end coherence, as well as by a diachronic norm of stability of intention. And it sees these as norms of practical rationality. But this can seem puzzling. First, we don’t think ordinary desires need to be consistent or means-end coherent. What is special about intentions and plans? Second, these norms may seem simply to favor forms of, in the words of Niko Kolodny, psychic tidiness. But why think such tidiness is a big deal?

It seems plausible that a general disposition to conform to these norms as an aspect of one’s planning agency will have significant benefits, especially given the important coordinating roles of our plans. But what we have learned from J.J.C. Smart is that an inference from a pragmatic defense of a general disposition of thought to a claim of irrationality in the particular case is fraught. So we need at least to supplement such a two-tier pragmatic strategy.

These questions in normative philosophy have implications for descriptive/explanatory philosophy of action. The thought (championed by Joseph Raz and Niko Kolodny, and anticipated by Hugh McCann in the 1980s) that these norms are a “myth” would tend to undermine our confidence that structures of planning – as understood within the planning theory – are basic for our descriptive/explanatory philosophy of action.

My response begins with two ideas: (i) Intentions and plans have distinctive coordinating, organizing roles in our individual and social lives. In part because of these roles, they help constitute where the agent stands, and so play an important role in self-governance (as discussed in my first post). (ii) The rationality norms at issue track conditions of a planning agent’s self-governance: plan inconsistency or means-end incoherence normally baffle a planning agent’s synchronic self-governance; and certain kinds of plan instability normally baffle a planning agent’s self-governance over time.

I argued for this claim about the synchronic norms in a 2009 essay. The idea that a diachronic norm of intention stability tracks conditions of diachronic self-governance is more difficult. Here I think we need to be prepared to adjust our formulation of the norm in the light of our account of a planning agent’s diachronic self-governance. Nevertheless, one idea that will emerge, given our approach to diachronic self-governance, is that the shuffling between non-comparable options discussed in the first post will be in violation of a diachronic norm that tracks conditions of diachronic self-governance.

But now we need to ask: How does the claim that these norms track conditions of self-governance – call this the tracking thesis -- show that they are norms of practical rationality?

Here again some might be tempted by talk of the constitutive aim of agency and try to argue that this aim is self-governance and that this explains why norms that track self-governance are norms of practical rationality. However, as I said earlier, this seems to overburden our descriptive and explanatory understanding of our agency. But then how does the tracking thesis explain why these are norms of plan rationality?

My tentative answer (sketched in a forthcoming paper and a forth-existing essay) is that it does this as part of an overall and reflectively stable understanding of these norms, an understanding that is available to a reflective planning agent. A planning agent who reflects on the basic norms that guide her plan-infused practical thinking would see the tracking thesis as part of the best explanation of those norms.

This explanation would involve the following ideas: (1) The tracking thesis shows that these norms do not merely track mere mental tidiness. (2) The tracking thesis articulates an overarching commonality across these norms, a commonality in light of which they make more sense. (3) The tracking thesis shows that if (a) one has a normative reason in favor of governing one’s own life, then (b) if one has the capacity for relevant self-governance, this reason transfers to a reason in favor of conformity to the norms in the particular case. (4) A conclusion I reached in my second post is that the end of diachronic self-governance is essential to the general exercise of a planning agent’s capacity for diachronic self-governance. So, suitably generalized, the tracking thesis supports the claim that the presence of the end of diachronic self-governance is itself enjoined by diachronic plan rationality. (5) Given this end, and the plausible assumption that diachronic self-governance is a good thing, a planning agent with the capacity for diachronic self-governance will have a normative reason in favor of diachronic self-governance, and so in favor of the involved synchronic self-governance. (Here I assume that the idea of a normative reason that is germane to these reflections of a planning agent will involve both a connection to her ends and the desirability of those ends.) (6) So given this end, (3)(a) is true. (7) In this way the tracking thesis supports (3)(b), and thereby the claim of a reason of self-governance to conform to the cited norms. (This is a response to John Broome’s and Niko Kolodny's question about whether there is reason to be rational.)

My idea, then, is that a planning agent who reflects on the basic norms that guide her plan-infused practical thinking would see the tracking thesis, together with (1)-(7), as part of the best explanation of these norms. It remains possible for her to reject this entire package. But then she would not be in a position to defend in this way central modes of practical thinking characteristic of her planning agency. In the absence of some other defense, she would then be under pressure to give up these modes of thinking -- though this is something she has pragmatic reasons not to do, and is also unlikely to be something she can do simply at will.   Short of this, however, she is in a position to see that her plan-infused practical thinking can be embedded within a framework that is available to her and that provides support for the application of the cited norms to particular cases, support that is over and above (but compatible with) pragmatic support for relevant general modes of thinking. So it will be reasonable – even if not strictly inescapable -- for her to retain her plan-infused practical thinking and its associated norms as part of this framework. So we need not worry that these norms are reflectively unstable in a way that would challenge the planning theory.

And – in closing – let me again thank Thomas Nadelhoffer for making it possible for me to try out these ideas on this blog.

Related articles

Self-governance Over Time - Flickers of Freedom

Doxastic Self-Control

Comments

[David Duffy]

Reading this I was reminded of examples of planning in nonhumans - eg chimpanzees will travel for 30 minutes to find the right type of grass to bring back to fish for ants in a nest. This implies a lesser difference between a norm and a practical policy or extended chain of behaviours.

[Thomas Nadelhoffer]

David raises an interesting question (or set of questions). If we adopt these accounts of diachronic and synchronic self-governance and tie them to plan rationality, how low does this set the bar for agency? It seems that not only children (at least of a certain age of maturity) will clear the hurdle. The same can arguably be said about other species--e.g., chimpanzies, gorillas, dolphins, elephants, dogs, and the corvids (to name a few). While Frankfurt was quick to deny that any other species on Earth had, for instance, the capacity for higher order reasoning, I am not sure this denial was grounded in close attention to the data from comparative psychology on just how rich and sophisticated the reasoning can be in these animals rather than simply a speciesist assumption on Frankfurt's part. I am not claiming that these animals positively do possess the capacity for second order desires and the like, but I am suggesting that perhaps it ought to be an open question. If so, this also broadens the scope of the investigation when it comes to moral agency and the like. Just a thought...

[Michael Bratman]

Thanks to David and Thomas for these interesting, thoughtful comments.

I agree with the spirit of both comments: agency comes in different forms, and it is an empirical issue which forms are instantiated by which animals. Indeed, this is in the spirit of the Gricean creature construction methodology that I have drawn on in my work. What we can do in philosophy is provide well-articulated, philosophically attuned models of different kinds of agency and let researchers – a prime example is Michael Tomasello and his colleagues – explore related empirical issues.

That said, I would want to highlight several ideas:

1. The kind of planning agency I have tried to describe involves (to use an idea from Allan Gibbard) a kind of normative guidance – where accepting a norm is tied, roughly, to a disposition to react to violations with a kind of “Darn it!” reaction. This allows for a kind of planning agency that is less demanding, as David suggests. And this is why I think that our descriptive/explanatory theorizing about human agency needs to be coordinated with our normative theorizing about practical rationality.

2. I do think that you can be a norm-guided planning agent without having the capacity for self-governance (an example might be young human children), since I take it that self-governance does involve some sort of capacity for effective reflection. This is why guidance by an end that is not modifiable under reflection (an example that comes up in my exchange with Kieran Setiya is a non-modifiable intention to smoke) will normally not be self-governance. (Though Frankfurt’s idea of volitional necessities complicates this.)

3. An important question – central, for example, to Tomasello’s work -- is the relation between the capacity for planning agency and the capacity for shared agency. Again, what we philosophers can contribute here are well-articulated, systematic and philosophically attuned models of these different forms of agency. (That is what I tried to do in my 2014 book.)

[David Duffy]

Just looking through Chrisman on "Epistemic Normativity and Cognitive Agency", who mentions beliefs may well be non-agentially acquired. I suppose there are empirical reasons why individuals who cleave to a norm of consistency and following through a plan ("good or bad") might be more likely to be practically successful, whether they follow the norm because of instinct (in animals, preverbal children), unthinkingly following societal rules, or because of reflection on previous successes by themselves or others.