An Argument from Reactive Attitudes for the Existence of God
March 22, 2011 — 14:12

Author: Kenny Pearce  Category: Existence of God  Tags: , , , ,   Comments: 16

In The Second-Person Standpoint, Stephen Darwall notes the fact that “we speak of being grateful for good weather” as a possible objection to his view that reactive attitudes are ‘second-personal’. He goes on to dismiss the objection on grounds that such gratitude “evidently involves the conceit that the weather is a free gift, as if from God” (p. 73). This remark struck me because I have known people who feel a sort of psychological need to believe in God in order to have someone to be grateful to (or, in other cases, angry at) for events beyond human (or animal, or presumably space alien) control. At first glance, it certainly appears (to me, at least) that belief based on this kind of psychological need would be irrational. Perhaps, however, the matter is not so simple. Consider the following argument:

  1. Human beings are so constituted as to generally feel reactive attitudes in appropriate circumstances.
  2. Many human beings feel reactive attitudes about, e.g., the weather.
  3. Therefore,

  4. It is appropriate to feel reactive attitudes about the weather (inductively, from 1 and 2).
  5. It is only appropriate to feel reactive attitudes about events which are actions of some agent.
  6. Therefore,

  7. Weather events are actions of some agent (from 3 and 4).

A lot of moral theories seem to be committed to (1). (2) is empirically verified. The strength of the inductive inference will depend on how reliable our tendencies to feel reactive attitudes in appropriate circumstances are, and also on how common it is to feel reactive attitudes about the weather (and other similar things). So that might be a weak point. (4) is pretty widely held and intuitively plausible, and (5) follows deductively from (3) and (4).
Personally, I think the kind of reliability we have in (1) is pretty limited (we get things wrong a lot), so I don’t think the argument is very compelling. Still, it does seem that there might be people whose epistemic situation is such that their credence in the existence of God can be justifiably boosted by an argument along these lines.
(Cross-posted at

  • Mike Almeida

    Hi Kenny,
    I don’t think (3) follows from (1) and (2). Humans might, with perfect generality, feel reactive attitudes in appropriate circumstances. That is consistent with feeling, also with perfect generality, reactive attitudes in inappropriate circumstances. So, I may be so constituted to feel grateful on every occasion on which it is appropriate, but that’s only because I’m just always feeling grateful whether it’s appropriate or not.
    Suppose we modify (1):
    1′. Human beings are so constituted as to generally feel the right reactive attitudes in appropriate circumstances and rarely to feel reactive attitudes at all in inappropriate circumstances.
    I think I’d urge that (1′) is false. I don’t think feeling apt reactive atittudes is a matter of constitution, and certainly not constitution alone. Feeling gratitude in the right circumstances (the the rght degree, at the right objects, etc.) takes training, and so do other apt feelings.

    March 22, 2011 — 17:56
  • Kenny Pearce

    Mike, that’s a good point. I guess there was an implicit ‘only’ in my original formulation. I agree that (1′) is false. (As I said in the post, I also think my original (1) is false.)

    March 23, 2011 — 1:48
  • Kenny:
    I am very sympathetic to arguments of this form and I particularly like your formulation.
    Note that you can also use arguments along these lines as arguments for the Fall or the existence of demons. People often feel resentment for events like earthquakes. If this resentment is appropriate, these events must have a cause in an agent who acted wrongly. But the best theories we have that make these events have such a cause are the Fall and the existence of demons active behind the scenes, and hence we have reason to accept at least one of these two theories.

    March 23, 2011 — 8:41
  • John Alexander

    We have reactive attitudes to our own existence. If we are constituted to have reactive attitudes in appropriate circumstances as 1 suggests and generally have this reaction to our own existence then, if 4 is warranted, we have to be constituted by an agent, or they are not occurring in appropriate circumstances. But this seems to beg the question.

    March 26, 2011 — 11:59
  • Kenny Pearce

    John, what sorts of reactive attitudes to our own existence do we have? Just the same kind of gratitude we have about the weather?
    I don’t accept (1) myself, but it seems to me that a lot of secular moral theorists are either explicitly or methodologically committed to it. For instance, for the method of ‘reflective equilibrium’ to make sense, our moral reactions to things need to be pretty reliable. If there are atheists who are committed to the premise, then the argument can hardly be question begging.

    March 26, 2011 — 12:34
  • John Alexander

    Some people end up feeling grateful for the suffering (a cancer for example) they endure and overcome. They see this suffering in terms of a test from God, which if passed will bring them closer to God. (We can substitute ‘resentment’ for those that do not overcome their suffering and/or see it as a gift from God. I had forgotten Peter Strawson on this issue:-))
    If (1) is warranted then if these feeling of gratitude or resentment are appropriate because of how we are constituted then we react as we do as a result of how we are constituted. Both the atheist and non-atheist can agree to this.
    However, the atheist cannot accept 4 if the agent is God. It seems plausible that the occurrence of e.g., cancer can be explained without reference to God, or any being for that matter, so one already has to accept the idea that events that generate reactive attitudes are caused by an agent in order for (4) to be warranted. Because they are already committed to the events that generate appropriate reactive attitudes as being caused by an agent then being constituted to have these appropriate reactive attitudes must be the result of being so constituted by an agent, God.
    This being the case, if (4) is warranted, then the position of the atheist is unwarranted as regards (1). As you suggest (4) is ‘widely held and intuitively plausible.’ But this can only be the case if one already accepts the idea (intuition) that events that generate appropriate reactive attitude are caused by agents. Hence the question is begged.
    Another issue is how to understand the function of ‘appropriateness’ regarding circumstances. If it is understood to mean that the circumstances are only appropriate if they are caused by an agent, then this too begs the question. (2) does not specific whether or not the circumstance is appropriate independent of our having the reactive attitude only that if we have the reactive attitude then it is appropriate. Does this not also beg the question because as Mike points out people can have reactive attitudes in inappropriate circumstances, e.g. making fun of a gift at a party in front of the person who gave the gift. It seems that it is the reactive attitude that makes the circumstance appropriate or not. If this is the case then we do not need the idea of agency behind circumstances, the concept of appropriateness will do just fine. It is appropriate to have a reactive attitude because the attitude is generated in an appropriate circumstances. But it is an appropriate circumstance because it generates appropriate reactive attitudes. Maybe this does not beg the question, but it does seem circular.

    March 27, 2011 — 11:55
  • Kenny Pearce

    So, my idea was something like this: a lot of moral theorizing relies on the assumption that our moral intuitions are more or less reliable. These moral intuitions seem to be connected to our tendency to experience reactive attitudes. For instance, we ‘intuit’ that we have been wronged in those cases in which we feel resentment, and we ‘intuit’ that an action toward us was supererogatory in those cases in which we feel gratitude. Now, these intuitions are about the evaluation of agents, so if the reactive attitude is evidence that an agent should be evaluated in a particular way, then it is evidence of that an agent is there to be evaluated. It seems like there are a fair number of atheist ethicists who are going to accept this much. But then it seems like the argument goes through.
    Now, I don’t see how this is question-begging in any sense other than the sense in which, as it is often said, every interesting philosophical argument is either question-begging or ad hominem. I mean, (3) and (4) do logically entail (5), so the atheist cannot accept both (3) and (4) (while remaining a logically consistent atheist). And if the atheist wants to accept (1) and (2) without permitting the induction to (3), she’s got some explaining to do.
    Here’s another way of looking at it. I’m betting that prior to considering this argument, there are atheist philosophers out there who accept (1) and (4). Upon considering this argument, they need to either block the induction to (3) or revise their views by rejecting one of (1), (4), and their atheism. So there are responses open to them other than accepting the existence of God, but that’s just the way philosophical argument works.

    March 28, 2011 — 10:18
  • John Alexander

    Hi Kenny, thanks for taking the time to respond to my comments. Please know that I understand how ‘philosophical arguments work’ although I may not understand all philosophical argumentsJ.
    “Human beings are so constituted as to generally feel reactive attitudes in appropriate circumstances.”
    My problem rests with understanding what is meant by ‘so constituted’ relative to how we became so constituted. I take it that there are many ‘yin-yang’ examples of reactive attitudes (RA): gratitude-resentment, joy-sorrow, happiness – sadness, etc. I also take it that, given premise 4, there can be appropriate responses only given certain circumstances (events), namely those caused by an agent. Does this mean that if there is an event that is not caused by an agent then a RA that I have to this event is inappropriate, or can I have appropriate RA’s in events not caused by agents? It seems that the latter is rules out by 4. If it is ruled out by 4, and 4 is warranted, then if I have an appropriate RA to how I am constituted – ‘I like who I am’ – then I must be so constituted to have this RA by an agent. Am I then responsible for my RA’s? If I have an inappropriate RA am I to blame for this or is it the result of faulty construction? Is any RA appropriate if it is a response to an agent caused event? Would this not mean that people who turn away from God because of the suffering they experience and the resentment they feel for having to endure this suffering are warranted in doing so? I know this is going beyond your post, but they are issues that seem to arise given your argument.
    Atheists do not have to find a problem with inferring 3 from 1-2. Atheists are going to have a problem only with 4, or they would not be atheists. It is only in 4 that the idea of agent causation regarding events comes into the argument. An atheist can accept 1-3, by providing some explantion, e.g., evolutionary theory that explains why we are constituted as we are, that does not rely on an agent. Because 4 is the controversial premise it seems to me that more then relying on our intuitions is needed to provide justification for thinking that 4 is warranted. An intuition that x is warranted may be evidence that x is warranted but it is not sufficient to make x warranted. If it were then we would not need the argument 1-5. So, other then your intuition that 4 is warranted what reason do you have to think that 4 is warranted. If I do not share you intuition that 4 is warranted, am I wrong to think that 4 is not warranted? I do not think that you can give a reason that does not beg the question. Now this may very well be the ‘Achilles heel’ of philosophy, but then it just reinforces the idea of van Inwagen that philosophical arguments are failures, but not necessarily a waste of time.
    This is certainly another side issue, but are you maintaining that all events within which we feel gratitude for the actions of an agent indicate that the agent’s action was supererogatory? Does this mean that if we feel resentment that the agents actions were intentional? What about events within which two agents have two different RA’s? How is this explained?

    March 28, 2011 — 11:42
  • Kenny Pearce

    According to the standard picture, reactive attitudes involve a sort of moral evaluation of agents. So it is a necessary but insufficient condition for the appropriateness of a reactive attitude that it be directed toward an agent. (One can morally evaluate an agent either rightly or wrongly, but if one is trying to evaluate a moral agent and there isn’t any agent there, then one is clearly wrong.)
    Now, one way an atheist could respond to this argument would be to reject this standard picture. This would amount to following the argument up to (3) and then rejecting (4). But this would require adopting a non-standard picture of reactive attitudes, and I would want to know what that picture would be.
    The notion of ‘appropriateness’ in (1) is supposed to be objective appropriateness. That is, reactive attitudes involve a sort of moral evaluation and to have reactive attitudes in appropriate situations is to have them when that evaluation is objectively correct. Now I think everyone can agree that we sometimes have inappropriate reactive attitudes, for reasons you note in your last paragraph. We may or may not be culpable for this. (Compare: we often have false beliefs, and this is sometimes because we do something wrong, but other times we’re doing the best we can and we end up with false beliefs anyway.) But some philosophers seem to think that the correlation between our reactive attitudes and appropriate circumstances for them is strong enough to be a good guide to certain morally relevant facts. That’s what (1) says. Now, the cost of rejecting (1) is that it is harder to do ethics, but I think ethics is hard, so I don’t mind rejecting (1). I think Darwall would also reject (1). So this is something else the atheist could do.
    Note that if an atheist rejects (1), then there is no reason why she can’t accept (4). She just has to say that it is inappropriate to feel gratitude regarding the weather. I think a lot of atheists do say this. By the way, liking is not a reactive attitude, so the atheist could still say that it is appropriate for her to like who she is/how she is constituted (or to like the weather). She would just have to say that it would be inappropriate for her to feel gratitude about this.

    March 28, 2011 — 16:30
  • John Alexander

    “According to the standard picture, reactive attitudes involve a sort of moral evaluation of agents.”
    Is the standard picture the result of defining ‘reactive attitudes’ to include a reaction to agency or is it based on empirical data that when people have a RA they are doing so in regard to agency/intentionality (Knobe type experiments for example)? If the latter, where is the data. Also, do you not need an error theory to explain those that do not support the trend? If the former, why should I accept the standard view – it seems to me to simply define something in such a way that we can get what we want from it, namely (4). If the standard view of RA is correct then it would follow that any time we experience an RA we are in relationship to an agent-caused event, by definition, and that is what I find problematic. So I guess I am taking a non-standard view where people can have RA’s to events like e.g., the weather, that are not agent caused, but are such that given the way we are constituted we will have RA’s to these types of non-agent caused events.
    Regarding the objectivity of appropriate RA’s in (1), are you suggesting that for S and S’ in circumstance C that both S and S’ will (should) have the same RA relative to C? The issue has to do with what is meant by being so constituted that we have RA in appropriate circumstances. Imagine that both S and S’ have terminal cancer and are suffering. S see this experience in terms of a test from God (agent caused) while S’ sees it simply as the result of a serious problem within his cells that is not agent-caused, etc. If the standard view is correct should we not maintain that S and S’ should have the same RA? But, clearly people do not, so how do we explain the divergence of the RA’s that are actually experienced.
    Wait a minute – I think I see your point. If S has an RA of gratitude or resentment, this type of RA only makes sense if there is an agent responsible for the circumstance that S find herself in. It is silly to resent something that cannot be other then it is. This makes sense regarding gratitude or resentment and seems to hold for other types of RA’s. HMMM – interesting, back to the ‘drawing board.:-)

    March 29, 2011 — 12:42
  • Kenny Pearce

    So, I think there are two ways that the concept of a reactive attitude gets introduced in the literature. One is to give some kind of definition, which usually will say something about the attitude involving a moral evaluation of an agent, and so (4) will get built in by definition. The other is to list examples, the most important being gratitude and resentment. On the non-definition route, (4) is still plausibly true, but it is possible to deny it.
    Note that even if we take the definition approach, it does not follow that whenever I have an RA I am experience an agent-caused event. (This is why it is open to atheists to accept (4), despite the empirical fact that people often feel RAs toward the weather.) First (just to make sure we’re totally clear), the notion of an RA doesn’t involve any kind of metaphysically heavy notion of agent causation (but I don’t imagine you intended to suggest that). Second, it might be that the attitude I experience is inappropriate, like if I get angry at my chair when I stub my toe on it. If there was some agent responsible for my stubbing my toe, then it might be appropriate to feel anger or resentment toward that agent, but surely it’s inappropriate to feel that way toward my chair.
    No one (that I know of) thinks that all of our RAs are always appropriate, and one piece of evidence that they are sometimes inappropriate is that, as you say, people sometimes experience radically different ones in the same circumstance. On the other hand, it is possible that there could be more than one RA appropriate in the circumstances, or that the circumstances have some important difference which is not apparent. (For instance, it could be that, for a particular individual, a particular instance of suffering is a good precisely because she regards it as a test from God, whereas for another individual a very similar instance of suffering is an evil because of the different attitude he takes toward it.)

    March 29, 2011 — 12:56
  • John Alexander

    Let us agree that 1-3 are plausible and acceptable. My problem is with (4). (4)does not follow from 1-3 because it introduces the idea that it is only appropriate to have RA’s to events that are actions that are agent-caused. If there is a moral evaluation and in order for there to be a moral evaluation there must be an agent towards whom this evaluation is directed then I think there are two senses of RA at play that are different in (1) and (4).
    Can we not feel gratitude or resentment at being in a circumstance that is not caused by an agent, or minimally does not need the idea of an agent to explain the circumstance we are in? Take the weather for example. Certainly there are plausible explanations for weather qua weather that do not involve the idea that weather is agent caused (at some point). (ID would have to be argued for, not assumed. I suspect that is where (4) leads us too:-))
    I woke up this morning and saw the blue sky and sun shinning and heard the birds singing and felt grateful that I was here and not in MI. This seems to be an appropriate RA within the context of (1) without me having to direct the feeling of gratitude to an agent. Now, this RA is not a moral evaluation of being where I am, or need not be one, it may be a aesthetic evaluation, or a psychological one (assuming there is a difference).
    But under (4) this would not be an appropriate RA because we can ONLY have appropriate RA’s regarding events that are actions of agents. So what seems appropriate in (1) is not, or need not be, appropriate in (4). You claim that “Note that even if we take the definition approach, it does not follow that whenever I have an RA I am experience an agent-caused event.” But (4) is worded so that it must be the case that the RA is appropriate only if it is a response to an event that is the action of an agent.
    It seems to me that your argument need only premises(4) and and a modified (2) to get (5). (2) should be (2′)Many human beings feel appropriate reactive attitudes about, e.g., the weather. Now, we need an argument for (4) that rules out my reaction to the weather as being an RA and claim that it is something other then an RA. Are feelings attitudes? Or maybe I have misdirected my RA. I am in AZ because we relocated for employment reasons so the move to AZ was an event that was caused by an agent and now that I am in AZ I am grateful to that agent for the move to AZ where I can enjoy AZ weather.
    Anyway, I am enjoying our conversation. It is making me think. Thanks.

    March 30, 2011 — 11:34
  • Kenny Pearce

    (4) is a premise about when RAs are appropriate, not about when they happen. That’s why the atheist can, and many atheists do, accept it. People are often inconsistent, especially when it comes to feelings. I might get angry at my chair when I stub my toe despite the fact that I am perfectly well aware that my chair is not an agent. I’m not sure why you want to insist that it is appropriate to feel gratitude about the weather in the absence of an agent responsible for the weather. At the end, I think you take this back when you talk about an agent being responsible for your experiencing that weather. That could be fine. But you talk about the attitude being ‘misdirected’. A misdirected attitude is an inappropriate attitude.
    Philosophers generally use ‘attitude’ only for things that have objects, so some feelings, like anger, are attitudes, because you need someone or something to be angry at, but others, like sadness, aren’t. (Sometimes we are angry without being able to identify what we are angry at, but in these cases we become confused and try to identify an object for our anger.)

    March 30, 2011 — 18:04
  • Gratitude is a kind of acknowledgment of a debt (“a debt of gratitude”, to put it circularly) owed to someone. But it is only appropriate to acknowledge a debt owed to someone when one owes a debt to someone. So gratitude is only appropriate when the sort of being to whom one can owe something is behind the event.
    Interestingly, it seems to me that gratitude belongs to a class of actions and attitudes that are appropriate only when morally required. Gratitude is never optional–it is always either required or inappropriate.

    March 31, 2011 — 20:21
  • John Alexander

    I agree now that gratitude is only appropriate when it is in response to events caused by agents:-) Therefore I can accept (4). But now, I reject(2)if the agent thinks that she is having a RA or she is misdirecting a RA. What I mean by ‘misdirecting’ is that I may feel gratitude for being in good weather, but if I am feeling this about the weather I am mistaken. I am actually feeling gratitude to the agent responsible for me being in a position to have a positive reaction to being in the weather I am experiencing. Certainly one can like, enjoy, feel good about, being in certain types of weather. But, I take it, given our discussion, that these are not RA’s, but some other type of psychological reaction that I am having, maybe just a simple emotional reaction.
    One possible explanation about having a RA type experience in (2) is that we ascribe agency to what is behind the existence of natural phenomena like weather because it has explanatory force for us in explaining why things are as they are. If S is inclined to believe in ID, then S would accept (2). If S’ is inclined to discount ID then S’ would reject (2). But, both S and S’ could accept (1) and (4). but only S would accept (5).

    April 1, 2011 — 10:43
  • Kenny Pearce

    You mean to reject (1), rather than (2), don’t you? That’s why you are talking about ‘misdirecting’? If so, I agree with you: (1) is the weak point. But one of the reasons the argument is interesting is that some atheist ethicists seem committed to (1).

    April 1, 2011 — 11:08