Self-governance Over Time

Thanks to Thomas Nadelhoffer for his kind invitation to participate in this blog.

I’d be pleased to discuss any questions/challenges to work I have done in the philosophy of action. But to get the ball rolling let me frame a question that concerns an agent’s self-governance over time.

In the background is my Frankfurt-inspired approach to a planning agent’s self-governance at a time (or anyway, during a short temporal interval). [Details are in my 2007 collection of essays, my 2010 APA Address, and a forthcoming reply (available on my web page) to a recent critique from Christopher Evan Franklin. Some of the first part of this post is drawn from that reply.]

Following Frankfurt, I suppose that an agent’s self-governance involves a relevant practical standpoint, one that consists of attitudes that help constitute where she stands on certain practical issues. When this standpoint appropriately guides, the agent governs. In central cases of self-governance (1) certain attitudes play roles in the agent’s psychic economy such that their guidance constitutes the agent’s direction of action; and (2) this guidance involves practical thinking and deliberation in a way that qualifies this agential direction of action as, in particular, a form of agential governance. There is self-governance when both these conditions are satisfied in a coordinated way.

I then propose that we specify the roles in (1) by appeal to attitudes that, as a matter of function, help constitute and support the continuities and referential connections highlighted in broadly Lockean theories of personal identity over time. What is needed to satisfy both (1) and (2) are structures of attitudes that play, as a matter of function and in a coordinated way, both Lockean diachronic roles and central roles in deliberation and practical reasoning.

The next point is that relevant plan states satisfy these dual design specifications. Plan states have it as a characteristic function to constitute and support Lockean cross-temporal ties. And plan states are potential inputs to practical reasoning. This will include deliberation in which one weighs various considerations. Such weighing is commonly shaped by policy-like attitudes in favor of giving certain considerations certain weights in deliberation. And such policies about weights will frequently be in part higher-order and concern how to treat relevant desires in relevant practical reasoning.

Such policies of weights are plan states that are suitably general. Their content involves the idea of treating a consideration as a reason/as having weight; and they may be associated with certain evaluative judgments. But they need not in general involve corresponding inter-subjectively accountable evaluative judgments.

This gives us a preliminary model of a planning agent’s synchronic self governance, one that highlights guidance of thought and action by a standpoint that involves relevant plan states, including policies about weights. Such plan-infused guidance is guidance by plan attitudes that play, as a matter of function and in a coordinated way, both Lockean cross-temporal roles and roles in deliberation and practical reasoning. Finally, in order to ensure that the agent is not conflicted about relevant policies in a way that baffles self-governance, we appeal to another Frankfurtian idea, the idea of wholeheartedness or satisfaction.

Putting these ideas together, we say that when relevant policies about weights -- policies that are frequently in part higher-order, and policies with which the agent is satisfied – guide, and the agent’s relevant structure of plans is appropriately consistent and coherent, then the agent governs in the relevant domain. These are, roughly, non-homuncular sufficient conditions for a planning agent’s self-governance at a time.  

And my question is: how can we best extend this to a model of a planning agent’s self-governance over time?

My initial proposal (sketched in a forthcoming paper and a forth-existing essay for a Routledge Handbook on Practical Reason) is that a planning agent’s self-governance over time involves her self-governance at times along the way together with relevant inter-connections between these forms of synchronic self-governance. What inter-connections? Here I think that we can appeal to the inter-relations between planning attitudes that are characteristic of planned temporally extended activity. These will include relevant continuity of intentions over time, relevant cross-references between intentions at different times, relevant mesh between sub-plans at different times, and relevant interdependence between intentions at a given time and past intentions and/or expectations of future intentions.

Further, I think that these cross-temporal intra-personal continuities, interconnections, inter-dependencies, and mesh between plan states over time will be importantly analogous to the inter-personal inter-connections of plan states that are characteristic of inter-personal shared intentional activity, as I understand it in my 2014 book. In this way we support the suggestive metaphor that in diachronic self-governance a planning agent is ‘acting together’ with herself over time. On this approach, then, a planning agent’s self-governance over time involves her self-governance at times along the way together with forms of intra-personal diachronic intention continuity, inter-connectedness, interdependence, and mesh that are analogous to inter-personal inter-connections characteristic of shared agency. (A fuller formulation of this idea would need to take into account that, given the hierarchical structure of plans, there can be such interconnections at a higher level despite a breakdown in interconnection at a lower level.)

An apt example is along lines discussed by John Broome (2001), though I will use an example from Sartre. Sartre’s young man decides to stay with his mother, though he sees the conflicting considerations of family loyalty and loyalty to the Free French as non-comparable. A little later he reconsiders. Since the considerations remain non-comparable, there seems no barrier to his newly deciding in favor of the Free French. But yet a little later he reconsiders and … Does nothing stand in the way of his shuffling between his mother’s house and the Free French office?

While this is not a complete answer, we can note that on the proposed, initial model of a planning agent’s self-governance over time, a case in which the young man retains his prior intention is a candidate for being a case of diachronic self-governance. In contrast, a case in which the young man abandons his prior intention and newly decides in favor of the Free French will violate a condition of diachronic self-governance. And these differing verdicts about diachronic self-governance are compatible with the recognition that the relevant considerations continue to be non-comparable.

To deepen this we would need to consider cases of potential willpower in the face of temptation. (Holton 2009) Might the continuity of intention involved in such willpower help constitute a form of diachronic self-governance? I plan to turn to this issue in my next post.

Comments

[Thomas Nadelhoffer]

Michael,

Thanks for joining us and thanks for sharing your recent thoughts on self-governance both at a time and over time. I had a few quick clarificatory questions to get the ball rolling here in the comment thread:

Along the way, you mention that you're providing "non-homuncular sufficient conditions for a planning agent’s self-governance at a time." On this view, all that is required is, "relevant policies about weights -- policies that are frequently in part higher-order, and policies with which the agent is satisfied – guide, and the agent’s relevant structure of plans is appropriately consistent and coherent."

First, is there a distinction to be drawn here between the agent and the self or are these terms used interchangeably? I ask only because I am curious whether you think all acts of self-governance at a time are free and voluntary acts. Or, instead, can agency be compromised while self-governance preserved? Relatedly, is saying an agent is self-governing at a given time the same as saying the agent is free (or has free will) or are self-governance and free will distinct as well?

Second, is the self something above and beyond these processes of weighting, deliberating, being guided, etc. or is the self constituted by these processes? It seems like we should want to be able to distinguish the self from the plans made by the self, the weights the self assigns to certain plans, etc. But then it starts looking like we end up with a homuncular self that is steering the proverbial ship.

Third, are the standards of consistency and coherency internal or external to the agent? Here I am thinking of agents who take themselves to be consistent and coherent in the face of those who claim the contrary. Take Trump and whatever plans he takes himself to have for moving forward. I would question the consistency and coherency of these plans. I suspect, alas, he would not. Is there a non-relative way of adjudicating between the two competing claims here?

That's it for now. I am quite confident these are all questions you have already addressed in your written work. But since not everyone will be familiar with some of your very recent (or forthcoming) work, I thought I would float a few questions to get things started. Thanks again for playing along.

[David Shoemaker]

Michael: Fascinating, rich stuff. I have many questions, but I will restrict myself to just one for now.

Given your adoption of a Frankfurtian, hierarchical view of agency, what is the right level at which we are to locate the agent for purposes of determining self-governance over time? You've described one version of Sartre's Frenchman in which it seems he's not a (good?) diachronic self-governor in virtue of his waffling. This locates the agent at the second order level, reflecting on the coherence of his various first-order intentions (I may not be putting this as precisely as you'd like, but you get the idea, I'm assuming.) But he might well be a more robust Sartrean, reveling in his freedom to reconsider various intentions at any time. So that the second-order agent "permits" reconsideration of the first-order intentions could reflect a *third*-order intention/desire to not be "settled" at the second-order about which first-order intentions to stick with. So this "waffler" would actually be a diachronic self-governor after all. Or would he?

[David Shoemaker]

OK, another question. The only "Lockean" constituent elements of diachronic (self-govering agential) identity you're focused on are intentions. But for most Lockean theorists these days, there are several other relevant connections and continuities for diachronic identity, including (a) memories (the current agent remembers or is continuous with a rememberer of some past experience), (b) persistence of desires/beliefs, and (c) character resemblance. Obviously, some of these can obtain where others don't. It's possible, e.g., that an agent could have strong intention-continuity but little of the other continuities. Do you then want to prize apart identity of the "self-governing agent" from the identity of "person," or do you want to equate agent and person and allow that, in light of the massive number of disconnections that obtain in such a case (among all the relevant forms of connection/continuity), the agent/person is just no longer the the same (despite strong intentional connections/continuities over time)? Hope this makes sense.

[Ben M-Y]

Michael,

Thank you for this post and for welcoming questions about your view.

I wanted to ask about the connection between three things that you mention in your post and have explored in your published work: (1) the normative content of self-governing policies, (2) the lack of intersubjective accountability for self-governing policies, and (3) the metaphor of acting with oneself over time.

As I understand your view (and please correct me if I’m mistaken in any way), the normative content of self-governing policies does not commit one to intersubjective accountability because the relevant commitments -- to treat certain desires as reason-giving in practical deliberation -- are only meant to be subjectively authoritative. For instance, my satisfaction with a policy of treating my desire to watch basketball as having a certain weight in my deliberations about what to do on a Wednesday night does not commit me to the claim that ideally rational agents would agree. This is my summary of (1) and (2), above.

But if this is correct, then I lose my grip on (3); I have a hard time seeing how “the suggestive metaphor that in diachronic self-governance a planning agent is ‘acting together’ with herself over time” is supposed to work. This is especially so since you suggest that a promising way of thinking about intrapersonal diachronic self-governance is along the lines of interpersonal shared activity. This suggests that the way to think about acting together with oneself over time is along the lines of acting with others. But if, as above, my self-governing policies do not claim intersubjective accountability, I lose my grip on how the details are supposed to work. Do my current policies bind my future self? Your brief discussion of the Sartre case suggests that perhaps you think they do. But how, especially if we are to think of my future selves along the lines of how we think of others, is this supposed to work in the absence of any claims to intersubjective accountability on the part of the relevant policies?

One response might be to claim that there is some intersubjective authority claimed by the normative content of self-governing policies, but nothing as strong as, say, the intersubjective accountability claimed by evaluative judgments. But this might end up weakening the bind between my current and future selves in problematic ways. I’m not sure.

But perhaps I’m missing the point entirely here. If so, I’d really appreciate your pointing this out to me (or simply pointing me to published work of yours that might clear things up).

Thanks!

[Michael Brent]

Hi Michael,

Thanks for the great post. I’d like to echo what I take to be one point in the comments by Thomas and David, regarding the causal role of the agent in her own acts of deliberation and practical reasoning. When we think carefully about your above description of what happens when the agent is weighing, deliberating, and being guided by the relevant features of her practical standpoint, it seems that the problem of the absent agent rears its head here, too, in the domain of mental action. As a general point, then, what sorts of relations obtain between the agent to whom we ascribe mental actions such as deliberation and practical reasoning, and the mental actions in question? Are you assuming, for instance, that the agent is *identical with* her acts of deliberation and practical reasoning?

[Michael Bratman]

Thanks to Thomas, David, and Ben for these thoughtful queries. I respond in broad and tentative strokes.

In response to Thomas:
I do not mean to provide an overall theory of all the myriad ideas that have been in the background of talk about “the self”. I just want to provide a broadly naturalistic model of an important phenomenon that is plausibly understood as a kind of self-governance on the part of a planning agent. (Note that I do not assume that all agents are planning agents.) I assume that the kind of guidance at issue insures that relevant actions are performed intentionally. But I take it that in cases of coercion by some other person one might act intentionally and still not act voluntarily, in one sense of the latter notion. And it seems to me one can also act in a self-governed way in response to such coercion. In such cases, one’s agency might be in a way “compromised” and yet self-governed. (This leaves open the hard question of what influences from others count as coercive – what are threats, what are offers?) Again, I take it that such self-governance is a kind of freedom – one that I see (to borrow from Daniel Dennett) as worth wanting. But I do not claim there are not other kinds of freedom to be investigated. Those with strong libertarian inclinations will enter the fray here, though it is my hope that the richness of the phenomena that can be captured by the framework I am sketching might help us here avoid a “panicky” metaphysics.
About the self as “above and beyond”: Given that one of our core practical capacities is a kind of reflectiveness, there is, as it were, the permanent possibility of stepping back and reflecting on oneself. But when we try to say what self-governance is, what we should look at is the agent’s practical standpoint as it is, though this may, of course, be shaped by such processes of reflection. In this way we can recognize the permanent possibility of stepping back and reflecting on oneself while holding onto a non-homuncular model of self-governance. Or so it seems to me.
The relevant forms of coherence and consistency can be, I take it, expressible in terms of associated “wide scope” norms. For example, an intention to A and an intention to refrain from A-ing will, taken together, violate one such norm. It will violate that norm even if the agent somehow thinks it doesn’t. That being said, we need to be alert to Davidsonian insights about the limits to such inconsistency: if the agent thinks his intentions are consistent then perhaps we have misinterpreted what he intends.

In response to David:
Consider a “robust” Sartrean agent whose most general commitment is to an ideal that at each moment he decides anew in light of the considerations currently present and without any concern at all for ‘acting together’ with himself over time. There seem to be two cases. A fully robust Sartrean agent does not even care about retaining this ideal over time. A less fully robust Sartrean agent does care about retaining this ideal over time. In each case there may nevertheless be forms of cross-temporal continuity as a result, in part, of what in my 1987 book I called snowball effects and the costs of reconsideration. But is the agent governing his life over time?
This returns us to the very idea of diachronic self-governance. And here I have been impressed with the metaphor that it is a kind of shared agency – where the sharing is across one’s own life. And this is what the fully robust Sartrean agent simply does not care about or have as an end, as I understand the example. So here I am inclined to dig in my heels and insist that this is not diachronic self-governance. (I return to the idea of having diachronic self-governance as an end in my third post.)
The less fully robust Sartrean agent is more complicated. (The issue here is an extreme version of what I alluded to when I spoke of the “need to take into account that, given the hierarchical structure of plans, there can be such interconnections at a higher level despite a breakdown in interconnection at a lower level.”) Given my effort to provide timely responses let me just say that I need to think more about this case.
About personal identity: I assume that if A is diachronically self-governing over a given temporal interval then the agent at the times along the way is one and the same person, A. So in the background there is a condition of personal identity over time; and I am supposing that we can work with a broadly Lockean model of that – one that involves (a)-(c) in David’s comment. But the condition that the agent at the different times along the way is one and the same person is not enough to ensure diachronic self-governance since it does not preclude extreme forms of shuffling. To articulate further conditions I proceed in two steps. First, there is the idea of guidance at a time by attitudes that both have it as a characteristic role to support relevant cross-temporal Lockean ties, and appropriately shape deliberation. Second, there is an idea of appropriate cross-temporal ties across these instances of self-governance at times along the way. Both steps are somewhat Lockean in spirit, since they both appeal to relevant continuities and interconnections. As noted, the second step also highlights the metaphor of acting together with oneself. And the conclusion I have come to is that the ties characteristic of such intra-personal, cross-temporal ‘shared’ agency are broadly Lockean ties across, primarily, relevant intentions and plans. As I said: “a planning agent’s self-governance over time involves her self-governance at times along the way together with forms of intra-personal diachronic intention continuity, inter-connectedness, interdependence, and mesh that are analogous to inter-personal inter-connections characteristic of shared agency.” But this not to say that there is a kind of cross-temporal identity of the self-governing agent that is to be distinguished from the cross-temporal identity of the person. There is, as I see it, just the sameness of the person over time. But when we try to spell out conditions for such a persisting person to be self-governing over time we can turn to these other uses of broadly Lockean ideas in the context of the parallel between intra-personal diachronic self-governance and inter-personal shared agency.

In response to Ben:
Your summary of my version of (1)-(2) is fine except that I also want to make room for self-governing policies that directly favor giving weight to a certain consideration. But this does not address your very interesting worry that (2) will block (3). Here my thought is that the intra-personal ‘sharing’ involved in diachronic self-governance is constituted by ties of continuity, referential connectedness, interdependence and mesh in ways that parallel the interpersonal case. If there are such ties across self-governing policies then there will tend to be agreement across those policies. But I don’t think that this means that the agent must see her policies at a time as ones on which ideal rational agents would converge. This is compatible with saying that if she cares about her diachronic self-governance (something I will discuss in my third post) then there will be pressure for intra-personal, diachronic convergence in her policies over time.
Consider the inter-personal case. You and I paint a house together. I do it because I can’t stand the color, you do it because you can’t stand the mildew. For us to act together I need not see my concern with color as one you ought to have; and vice versa. What is needed for us to act together is not sameness of reasons for which we participate in the shared activity but relevant ties across our intentions and plans, as I discuss in my 2014 book. (I briefly discuss these ideas about diachronic self-governance, and the parallel with shared agency, in my forthcoming “Rational Planning Agency,” available on my web page. I discuss it in much more detail in my forth-existing, but not yet available, “A Planning Agent’s Diachronic Self Governance”. Many of these ideas are in my 2016 “Pufendorf Lectures” – links to videos of the four lectures are on my web page and are recommended for insomnia.)

[Ben M-Y]

Michael,

Thank you for the thoughtful reply. One quick follow up, if you don’t mind:

I think I agree that “ties of continuity, referential connectedness, interdependence and mesh” would tend to result in intrapersonal agreement and continuity of policies over time. But I thought that part of your response to Wallace (2014) (which you discuss in your forthcoming response to Franklin, posted on your site) was that the relevant Lockean ties are not sufficient for self-governance. We also need the normative content of the relevant policies in order for the continuity and connection to amount to governance. (Do I have this right?) This suggests to me that there is important work being done by the normative content of self-governing policies, and not just the Lockean connections. My worry, however, is that, in the context of diachronic self-governance, this puts pressure on your theory to increase the sense in which one is interpersonally accountable for one’s self-governing policies.

I am inclined to think that in order for my self-governing policies to support my diachronic self-governance, they have to both exhibit the relevant Lockean ties and the relevant normative ties between my current and future selves. In a phrase, one might say that diachronic self-governance requires both Lockean (or psychological) authority and normative authority. And I’m not sure how the relevant normative authority is supposed to stop short of intersubjective accountability (perhaps to be analyzed in terms of ideal rational agents, or in some other way).

In your response to my earlier comment, you mention the notion of caring about diachronic self-governance, and maybe this will provide a different way of handling the issue. But I wonder what you think about all of this.