Ce texte reprend la communication que j’ai prononcée à la -Journée philosophique de Gordes, organisée par Jean-Maurice Monnoyer, I.H.P., Aix-en-Provence, Aix Marseille Univ, 05 novembre 2016
In "He could have done otherwise", Roderick Chisholm elucidates the meaning of this quite meaningfull sentence that gives the article its title ; he tries to provide a frame to this counterfactual claim, but his main concern is not a moral one. So I will try here to import the results of his analysis of its meaning in the ethical area, in the general framework provided by John Skorupski as following : "The dilemma, then, is this : either spontaneity gives one no account of normative knowledge (even if it yields an account of normative warrant) – in which case it is not an adequate account of the epistemology of the normative, since we rightly think that we know some pure normative truths – or if it yields an account of normative knowledge it has to be backed up by the constructivist metaphysics which I reject. Since I think that we know some normative truths I am forced to constructivism by my own epistemology, according to which warrant for and knowledge of the truth of normative propositions rests solely on the interplay of spontaneous propensities to judge, and not at all on any form of receptivity".
My main focus, here, will be to fix the meaning of "he could have done otherwise" in a moral context, that is to say, in the kantian context of the third antinomy, on freedom, where Kant elucidates the meaning of an other counterfactual statement "He ought not to have done so", in order to grasp the moral judgement. For, if he ought not to have done so, it can not be the case that he could not have done otherwise. There is a link between both of these statements, which may be a foundational one, the question being whether "he ought not to have done so" is a grounding for "he could have done otherwise", or "he could have done otherwise" is a grounding for "he ought not to have done so". The hypothesis is that, from a kantian point of view, that is to say, from a moral point of view, the "ought statement" is grounding of the "could statement". This is the main stream interpretation about the kantian thought, a main stream that I will criticize in this talk. I will be mainly concerned with elucidating what one means, in moral thinking, by "he ought not have done this way" which is not only "he could have done otherwise and he ought have to".
The logical connection between these both statements, "He ought not have done so" and "he could have done otherwise", once cleared up, will provide us a counterfactual scheme for the moral claim. I will elucidate the meaning of "he could have done otherwise" according to Chisholm, then I will try to specify the specificity of a moral context in which such a statement can be made.
I. The first counterfactual : "he could have done otherwise"
The first concern is : what does mean "he could have done otherwise" since he has not done otherwise ? According to Chisholm, this statement, that is to say the could statement can be interpreted as involving either a logical possibility, or an epistemic possibility, or as being iffy. Let us consider those possible interpretations, knowing that he will criticize each of them :
1. "He could have done otherwise" as not merely a logical possibility
If the could statement involves a logical possibility, we might say "he could have arranged things so that he would be in Boston now" as well as "he could have arranged things so that he would be on the moon now"have arranged things so that he is in Boston now", we may know that he has not arranged things so that he could be in Boston now, so that his being in Boston now is not consistent with everything we know. So "he could have done otherwise" is a statement that needs not to be consistent with everything we know.
3. The third hypothesis is the iffy one (according to Chisholm)
According to the iffy hypothesis "he could have done otherwise", instantiated in "he could have arranged things so that he would be in Boston now" means "there are certain things such that, if this morning you have undertaken to bring them about, then you would be in Boston now". But there are things that are consistent with the if statement, as we may call it, that are not consistent with the could statement. For, according to Chisholm, "consider those things which are such that, if this morning our agent had undertaken to bring them about, then he would be in Boston now. And let us suppose (i) that he could not have undertaken to bring any of those things about and (ii) that he would be in Boston now only if he had undertaken to bring them about" (p. 411). In this case, one could assume "there are certain things such that, if this morning you have undertaken to bring them about, then you would be in Boston now" but one can no more assume that he could be in Boston now.
So Chisholm can claim that all the translations from "could statement" to another have failed ; none of the "if statements" has exactly the same meaning than the could statement, and the "could" is neither a logical could, nor an epistemic could, nor a constitutionnally iffy one. In order to make its meaning clear, one needs to know what we exactly mean while assuming "he could have done otherwise". After having criticized all of them, he can assume that "he could have done otherwise" means : "His making A happen at t’ is directly in his power at t, provided only : (a) there is an event X such that his undertaking at t to make X happen would cause A to happen at t’ ; and (b) there is no sufficient causal condition at t for his not undertaking at t to make X happen" (p. 414). Where "something being in one’s power" means "His making a certain thing A happen is in his power at a given time if, and only if, either (i) by making happen what is directly or indirectly in his power at that time, he would make A happen, or (ii) A is an undertaking that is directly in his power at that time" (p. 417).
II. The second counterfacual : "he ought to have done otherwise"
Kant, troisième antinomie de la Critique de la raison pure, Gallimard, 1980, p. 489 :
"Pour éclairer le principe régulateur de la raison par un exemple tiré de l’exemple empirique de ce principe, je ne pas pour le confirmer (car des preuves de ce genre sont sans valeur pour des affirmations transcendantales), que l’on prenne un acte volontaire, par exemple un mensonge de nature maligne par lequel un homme a introduit un certain désordre dans la société ; qu’on recherche d’abord les causes déterminantes d’où il est sorti, et que l’on juge ensuite comment il lui peut être imputé avec toutes ses conséquences. Sous le premier point de vue, on pénètre le caractère empirique de cet homme jusque dans ses sources, que l’on recherche dans la mauvaise éducation, dans une détestable société, en partie aussi dans la méchanceté d’un naturel insensible à la honte, ou qu’on rejette sur le compte de la légèreté ou de l’irréflexion, sans perdre de vue les causes occasionnelles et leur incitation. Dans tout cela, on procède comme on le fait en général dans la recherche de la série des causes déterminantes pour un effet donné de la nature. Or, bien que l’on croie que [A 555 / B 583] l’action est déterminée par là, on n’en blâme pas moins l’auteur, et cela non pas à cause de son funeste naturel, non pas à cause des circonstances qui ont influé sur lui, non pas même à cause de sa conduite antérieure, car on présuppose que l’on peut mettre tout à fait de côté ce qu’a été cette conduite, regarder la série des conditions écoulées comme n’étant pas arrivée, et cette action, au contraire, comme entièrement inconditionnée par rapport à l’état antérieur, comme si, par là, l’auteur commençait entièrement de lui-même une série de conséquences. Ce blâme se fonde sur une loi de la raison, où l’on regarde celle-ci comme un cause qui a pu et dû déterminer autrement la conduite de l’homme, nonobstant toutes les conditions empiriques qu’on a citées".
By elucidating this sentence, that will be my second point, one will be able to elucidate the connexion between "he could have done otherwise" and "he ought to have done otherwise", which is the kantian sentence about the lack of moral behavior in the third antinomy. He lied, anyone can understand why he lied, which means anyone can understand the reasons why he lied, and nevertheless, anyone will maintain that "he ought have done otherwise". Both of them are counterfactuals claims about of what happened for he lied, but he ought to have done otherwise, and, since what one ought to do, one can do it, he could have done otherwise, that is to say, in a moral context, he could have done what he ought to do :
He ought have done otherwise
So there is at least one action he ought to have done and he had not done
Since what one ought to do, one can do it
There is at least one action he could have done in order to to what it ought to
This is a quite important point about reasons, for we are able to understand not only the reasons he had to do what he has done, but the reasons he had to do what he has not done, but he ought to have done. He had reasons to do otherwise and even if those reasons were not the highest, nevertheless, he had them too.
Let us now keep in mind the example of the liar in the third antinomy, on freedom. One can understand the reasons why the liar has lied. His reasons are in a way constituing his motivational set (he needs money, he can not get it back, he knows that point, no one will help him in such circumstances, he has not received moral instruction … etc … Nevertheless, according to Kant, one may claim about him that "he ought not to have done so" in spite of all the reasons he has to lie. He has a lot of reasons to lie, some of them quite convincing, so convincing that one can not be sure that, in the same circomstances, he would not lie too. And actually, he lied. What does it mean to assume that he ought not to have done so, for he has done what he has reasons to do, even though these reasons were not moral reasons ?
So my main focus here will be the following one : in order to assume that "he ought not have done so", does one need to have a moral preconception he expresses in this statement, or is it possible to assume this statement with no moral preconception ? In the first hypothesis, there is a kind of vicious circle here, which is not the case in the second hypothesis, that is, according to me, either Kant’s hypothesis or nevertheless, if not Kant’s one, the good one. But it this case, we need to examine furtermore the connexion between ought and could : are (1) "what one ought to do, one can do it", and (2) "he ought to have not done it because he could have done otherwise", coherent with one another ?
III. The third counterfactual : counterfactual reasons (?)
Let us now come back to John Skorupski’s thesis according to which "The only condition that can legitimately be placed on reference (I say) is that we should know, and be able to communicate to each other, what we are talking about – what our topic is". I will not discuss here the quite obvious contradiction with Quine’s ontological thesis, but I will focus on an other point. For this thesis might be too strongly connected to the reasons an agent can give for what he has done. For the reasons an agent had to do what he has done might be strongly counterbalanced by the reasons an agent had to do what he has not done, or what he should have done but has not done, or, moreover, what he ought to have done, for, unless strong determinism, he could have done otherwise. Actually, all the reasons an agent might give of what he has done are not silencing the reasons he will not give for what he has not done.
The question, here, can be expressed in terms of sharp cut distinction between first-person reasons and third-person reasons, sharp cut distinction that I claim not to be the case. As claimed by John Skorupski, "truths of a given kind are reasons for you only if you can potentially tell in a first-person way (not just on testimony) that truths of that kind are indeed reasons – only if you can potentially recognize for yourself their reason-giving force". This can be understood as an internalist thesis for an agent having reasons to act in a way is, for this agent, the same as being able to express these reasons in a first person way, that is to say, the same as being motivated by them. Nevertheless, this quite strong internalist thesis allows us not to recognize these reasons but to potentially recognize them. I guess this is quite a suggesting adverb here. I will propose a weaker form of internalism, according to which an agent can recognize reasons to do otherwise than he has done, and this weaker internalist thesis is compatible with John Skorupski’s claim.
That is to say, an agent may have chosen to follow reasons to do what he has done, by expressing them in a first person way, but, even in those circumstances, he may nevertheless recognize reasons to do otherwise, although he has not expressed those reasons to do otherwise in a first person way, but he might have expressed them in a first person way, even if he did not. This is the lever by which one can move from a particular point of view on reasons to do something expressed by first person reasons, to a universal point of view on reasons to do otherwise than he did, potentially expressed in a first person way.
So this is the third counterfactual we needed : he could have done otherwise, for he ought to have done otherwise, and he would have done what he ought to if he had weighed reasons in a different way that could have been a first person way.
So let us rely on Chisholm’s analysis of "he could have done otherwise" and let us come back to the translation he gave of this sentence, according to which "he could have done otherwise" means the same as "his making A happen at t’ is directly in his power at t, provided only : (a) there is an event X such that his undertaking at t to make X happen would cause A to happen at t’ ; and (b) there is no sufficient causal condition at t for his not undertaking at t to make X happen".
There might have been one sufficient causal condition at t for his not undertaking at t to make X happen, that is to say if, at t, he could not have done otherwise and this condition is the following one : he could not have done otherwise at t if, at t, he could not have had reasons to do otherwise, reasons that could have overwhelmed the reasons he followed, even though they have not overwhelmed them.
Actually, by saying "he ought to have done otherwise", we are although saying "he could have done otherwise", since "ought implies could", and by saying "he could have done otherwise", we are saying "he could have recognized reasons to do otherwise" for, could he not have recognized reasons to do otherwise as overwhelming the reasons he has followed, then, he could not have done otherwise. That is to say, "he ought have done otherwise" implies "he could have recognized reasons to do otherwise", even if he has not.
Conclusion : an ontological problem
But, if my hypothesis is true, then I have to focus on the ontology of reasons it presupposes and to solve an other problem about tthe ontology of reasons. As Pascal Engel puts it : "Pour Skorupski aussi les raisons sont des faits, mais il adopte une conception minimaliste des faits : un fait n’est rien d’autre qu’une proposition vraie. Si c’est un fait que Marie aime Jean, cela ne veut pas dire autre que chose que : il est vrai que le Marie aime Jean, ou que Marie aime Jean est un fait. Les faits n’ont pas d’autre structure que celle des propositions. Ils n’ont pas non plus d’impact causal. Cela revient en fait à défendre une conception tout aussi peu« ontologique » que celle de Parfit. La relation de raison pour X de F-ier est elle-même normative. Mais elle n’est pas une certaine sorte de chose dans le monde".
These are the questions I have now to ask in order to maintain my hypothesis : if reasons are facts, that is to say, no more no less than true propositions, according to John Skorupski according to Pascal Engel, but if reasons are not "a certain kind of things in the world", how could there be reasons for doing otherwise that it has been done ? My analysis is a counterfactual one and obviously this is the main source of the difficulty I encounter here. For, in a non-counterfactual analysis, reasons can be normative, and nevertheless they can be not factual. But, in a counterfactual analysis, how can there be reasons that do not pertain, that are not relevant and that, nevertheless, could have pertained even though they have not. Is this a kind of counterfactual reason, that is to say, a reason to do otherwise than we have done, that is to say, a reason to do what we ought to have done.
I may give an example of such a reason : in any circumstances, there is a reason to act morally, according to Kant. This reason might be expressed in the following way : the maxim describing such an action is universalizable. Even though he has lied, as in the third antinomy, he could have done otherwise, he could has said the truth, and if he had said the truth, that he has not said, he would have had a reason to do so, for doing so is universalizable. He could also have had quite different reasons, but this particular reason is the only one according to which he would have acted morally. It is quite interesting that the moral reasons are assumed by Kant to be counterfactual. And, morevover, it is quite important, for we can never be sure that a moral action has been done in this world.
But, if there are counterfactual reasons, as I do think, does it mean that reasons are facts ? Or how reasons could not be facts, if there are counterfactual reasons ? The problem is, as we know, that facts and reasons are strongly distinguished by John Skorupski :
"So the very point I have in mind here is connected with JohnSkorupski’s claim about the difference between norms and facts : (quotation) "It starts from what has always seemed to me to be an utterly simple and plausible idea : namely that difference between a factual and a purely normative proposition is just that. In other words, to assert a factual proposition is to claim that some fact obtains, whereas to assert a purely normative proposition is not to assert that some fact obtains. Any assertion, factual or normative, purports to inform ; information, however, can be factual or it may be normative. Thus whereas the claim that a fact obtains can - obviously - be true only if the fact in question does obtain, the truth of a purely normative assertion does not depend on the obtaining of any facts. This simple thought - that purely normative information is information but not factual information is the essence of what I call ’cognitivism irrealism’. " John Skorupski, "Précis of the Domain of raisons", Philosophy and Phenomenological Research, Vol. LXXXV, n°1, July 2012, p. 176.
So, if my hypothesis is true, reasons can be counterfactual ; either there is a sharp cut distinction between counterfactual reasons and other reasons, or they are the same kind of reasons. I guess that, if they are not facts, as John Skorupski puts it, they are not counterfactual. They are neither facts nor counterfactuals. But, if they are not facts, how could it possible to speak of reasons one, including me, had to do what he has not done, and even reasons he did not have in mind to do what he had not done. My main point is the following : either there is not a counterfactual kind of reasons, that is to say the reasons one had not in mind to do what he has not done, or we have to make this kind of reasons coherent with the claim that reasons are not fact, for actually I maintain there is such kind of reasons. To put it in a few words, reasons may be facts if they may be counterfactuals.
Roderick Chisholm, "He could have done otherwise", The Journal of Philosophy, Vol. 64, No. 13 (Jul. 6, 1967), pp. 409-417 http://www.jstor.org.lama.univ-amu.... Pascal Engel, Recension par P.Engel de On what matters de Parfit et de The domain of reasons de Skorupski. https://www.academia.edu/29087472/E... Emmanuel Kant, Critique de la raison pure, troisième antinomie (sur la liberté), A555/B583. John Skorupski, "Précis of the Domain of raisons", Philosophy and Phenomenological Research, Vol. LXXXV, n°1, July 2012, p. 176. https://www.academia.edu/20955391/P... John Skorupski, "Normative spontaneity", http://semaihp.blogspot.fr/2016_10_...
1ère mise en ligne et dernière modification le 2 décembre 2016.