Author(s): Guglielmo Feis
Source: European Journal of Legal Studies, Volume Issue ()


Ought Implies Can: Counter-examples and Intentions

Guglielmo Feis*


The paper tackles Ought Implies Can (OIC) from a slightly different angle compared to the one that is often adopted in the contemporary OIC sub-debates. I am mainly concerned with the thought according to which, even if the action we ought to do is impossible, it is still possible to intend to do it. This possibility of intending is used as a strategy to rescue OIC from possible counter-examples. I explore two different ways to rescue the principle: (i) OIC rescue by implicating intending in action and (ii) OIC rescue by separation, and show they both face problems.

Keywords: Ought Implies Can, intentions, impossible obligations


I. Introduction

This paper deals with supposed counter-examples to the Ought Implies Can (OIC) thesis (in the literature they are the addicted example, 'role-oughts' and, in general, 'ought but cannot' cases)[1] and challenges two possible replies to them, i.e. to ways to defend. A possible defense is arguing that it is possible to intend the thing that one ought to do, even in cases in which doing it might be really hard or even impossible. I call this move the Ought Implies Can Rescue (OIC-R).[2]

Such a rescue may happen in two different ways, either (1) rescuing by implicating-intending-in-action or (2) rescuing by separation. The first OIC rescuer – the rescuer by implicating-intending-in-action – states that the possibility to intend the action requires (and this implies) the possibility to somehow do the action. The second OIC rescuer – the rescuer by separation – states that the mere possibility to intend the action is enough to save OIC.

If we want to stick to one of the first inquiries on OIC, that by Georg Henrik von Wright, we delve into the OIC research in which 'one may discuss whether the idea, when understood in a certain way, is true or not'.[3]

I first offer a brief sketch on the vast usages and debates on OIC (section 2). The reader acquainted with OIC eager to know my argument may skip this paragraph and go to section 3, in which I lay down the structure of the counter-examples to OIC and what is at stake in this paper. Then, in the next two sections, I object to both these rescue moves: section 4 deals with rescuer by implicating-intending-in-action, section 5 deals with rescuer by separation. In section 6, I show that the thesis emerging from these attacks to OIC rescuers – intending to do A does not imply the possibility to do A – is compatible with another OIC related thesis, namely the thesis from the philosophy of action, according to which you can intend to do the impossible.[4] In the last section (section 7), I draw some conclusions.

II. Historical Preliminaries: OIC Context

The leading idea behind Ought Implies Can (OIC) is one according to which what we can do, what we can realize, what is (physically and logically) possible and what is feasible are all relevant in assessing what we ought to do. The idea is better cashed out moving from the direct OIC formula to its contraposition, i.e. as no Ought Implies no Can (Cont OIC).

(Cont OIC) is probably the way in which the OIC thesis is most often discussed and intuitively grasped. If we recall the old Latin brocarda in the Justinian Digest 'inpossibilium nulla obligatio' it is quite easy to find an ancestor of OIC.

This old Roman origin and the fact that the thesis is often associated with the name of Kant, made it the case that OIC gain traction in the beginning of the 20th century as the 'Kantian principle' or 'Kantian Maxim' or even 'Kantian Axiom'. As soon as deontic logic evolved, OIC was used as an axiom to ensure the logic has no contradiction. In the debate on moral dilemmas, OIC was used once more to ensure there was no moral dilemma.

Almost everything related to OIC is nowadays being placed under renewed critical scrutiny. All its components are questionable: which sort of ought, duty, obligation or normative requirement is expressed by 'Ought'? What is the relationship expressed by 'Implies'? Is it some logical relationship or a pragmatic and conversational one? Why not opt for a conceptual one? And, most delicate of all, what is the relevant 'Can' we have to measure and evaluate?

Not only do we have the issue of constructing a suitable OIC by way of picking up the correct elements for its three components, but the OIC ancestors are, in themselves, controversial. The OIC wording is not present in the Kantian corpus and having our contemporary OIC proposal fit into the old Roman adage is complicated. Further, it is controversial to state the relationships between OIC and the so-called Hume's Law. OIC is assumed to be a bridge-principle between Is and Ought when it is formulated contrapositionally as (Cont OIC): in fact, here, we are using what we can(not) do (i.e. Is) to say that an Ought should cease to be in force. Is that compatible with the gap between Is and Ought (assuming there is such a gap)?

Another highly debated issue connects OIC to free will, via the discussion of the principle of alternate possibilities that has been criticized by Frankfurt. As the debates in logics and moral dilemmas evolved, we developed frameworks that are able to tolerate or accept conflicts. Hence, OIC became more a choice than a necessity. Another issue is that of doxastic voluntarism, i.e. the thesis according to which we have control over our beliefs, and this connects to epistemology, as far as the issue of infinite regress in justifications is concerned. Referencing all these debates is complicated and may well need a review article.[5]

This quick overview was necessary to provide some broader context of the philosophical significance and the main problems of OIC. In the legal domain, OIC is probably less criticized and analyzed than it is in the philosophical literature. For example, in his 2007 contribution Ferrajoli is almost ready to swear on OIC, without taking into account any of the criticism that has been raised against OIC, and Guastini is similarly inclined.[6] In a recent paper Stef Feyen used OIC as a bridge principle to persuade legal dogmatics to include more empirical knowledge and data in their analysis.[7]

This paper does not engage directly with some of the issues of the vast and diverse OIC sub-literatures. I investigate some of the reasons we may have to support OIC when we are faced with some putative OIC-counter-examples (see section 3 for how these counter-examples work). There are at least two possible defenses for OIC: (i) rescue by implicating-intending-in-action (section 4) and (ii) rescue by separation (section 5). I explore and criticize these.

Emphasizing the connection between OIC and intentions points out a further connection between OIC and the legal domain, given the importance of intentions in the evaluating agency.

Before moving on, I think it is worth disclosing my overall attitude towards OIC. I believe OIC involves a plurality of theses within it. This is evidenced by any recent paper on OIC in which more than one OIC formulation is proposed and discussed. What comes out of this recognition of a plurality of OIC theses is that it is difficult to establish the existence of a non-controversial OIC simpliciter which is able to cover all the different fields and aspects of it.

III. Conceptual Preliminaries: OIC formalization and Counter-Examples

Consider the Ought Implies Can thesis:

(OIC): OA →◊A.[8]

For a counter-example to OIC, we need a situation in which:

(OIC countr): OA ˄~◊A,

i.e. a situation in which there is an obligation to do A and A is not possible/feasible.

Now, several examples have been provided in the literature for (OIC countr) cases.[9] I do not want to discuss whether the supposed OIC counter-examples are real counter-examples or not. For the purpose of this paper, I will simply take these counter-examples from the literature as something that whoever wants to defend OIC has to deal with.

A possible reply to the OIC counter-examples could be the following one: despite A being impossible, it is still possible to intend to do A, to have reasons to do A or to have some possible and feasible compensatory obligation(s) to compensate our not doing A.[10]

I call this move the OIC rescue through intending thesis, that can be spelt out as follows:

(OIC rescue): OA →◊(Int A) or, with reasons, OA →◊(Reasons to A).[11]

In order to move from (OIC rescue) to (OIC), we would need a thesis that says that the possibility of intending A somehow requires, states or implies that A is possible. The thesis that we need is the following; let's call it intending-in-action thesis:

(IIA): ◊(Int A) →◊A.

(IIA) states that the possibility of intending A implies that it is possible for us to do A.

Nonetheless, even in cases of a failure of the preceding (IIA) thesis, a different kind of OIC rescuers may claim that OIC still holds, even in cases of OIC counter-examples. Their reason would be that the 'true' OIC is not the one hinging on the possibility of A (i.e. (OIC): OA →◊A) but, rather, on the possibility of intending A (i.e. (rev OIC): OA →◊(Int A)). This second group of rescuers does not need an additional thesis such as (IIA), but has the burden to prove that what we have in mind when we discuss Ought Implies Can is not the (OIC) formulated above, but a revisionary intention based OIC:

(rev OIC): OA →◊(Int A).

Despite (rev OIC) and (OIC rescue) being the same formal thesis, their status and function in the OIC debate is different. (OIC rescue) is a functional thesis to defend (OIC) that requires a further argument for (IIA) to establish some relevant connection between (OIC rescue) and (OIC). (Rev OIC), instead, is a revisionary thesis. It says that the real (OIC) is in fact (rev OIC) and assumes no commitment to (IIA).

In the following two sections, I deal with both these parties. I call the first group of OIC rescuers the rescuers by implicating-intending-in-action: they hold that (IIA) is true and rescue OIC relying on this thesis. The second group is the one of OIC rescuers by separation: they claim that OIC is not the usual thesis we have in mind but rather something different, i.e. OIC ranges on intentions. Thus, they separate the two OICs, the usual one on the possibility to A, and their new and true one on the possibility of intending to A, and go to defend the second one, without specifying any relationship between the two OICs thesis – i.e. (rev OIC) and (OIC).

IV. Against OIC Rescuers by Implication-Intending-in-Action

OIC rescuer by implicating-intending-in-action states that the possibility to intend the action requires (and this implies) the possibility to somehow do the action. Given a counter-example to OIC, they will separate the components of action and try to go 'one level down': if doing A is impossible, they say, we go back and evaluate whether intending to do A is possible or not.[12]

The specific move of this rescue is: (i) go down a level – i.e. shifting from doing to intending; (ii) then, in order to save OIC, we need to go up again to the standard OIC level where we show that, despite A being impossible prima facie we have to reconsider its (im)possibility, because we have seen, at a lower level, that it was possible to intend to do A.[13]

Now, in order for this down-up move to work (down from doing to intending, then up again to doing),[14] what the rescuers need is a thesis that states the connection between the down and up, namely between intending and doing. Let us call this intending-in-action:

(IIA): ◊(Int A) →◊A.

The possibility of intending A implies that it is possible for us to do A. This amounts to saying that we can always do what we intend.

The (IIA) thesis seems to express a definition of 'intending' that may be justified upon particular rationality constraints (if you value consistency you would like to avoid the chance for an agent to intend something he cannot do), but in this contextthis seems unjustified.[15] Here, we are dealing with ways in which OIC can be rescued from its counter-examples by way of using the concept of intention and, in particular, by way of providing a better understanding of the concept of action involved in the discussion of OIC counter-examples.[16]

Thus, it seems possible to provide counter-examples to (IIA), i.e. cases in which it is possible to intend A but it is not possible to do it. Consider for example the case of Ponce de León, who intends to chase the source of eternal youth (and not only does he intend to do so, but he also starts a quest to find it), which is however not possible for him to do, because it does not exist.[17] Consider another example of failure of (IIA),[18] such as the statement 'I intend to square the circle, but I cannot do it' (because it's impossible).[19]

I think that this counter-example holds for both faces of intentions outlined by Michael Bratman in his 'Two Faces of Intentions'.[20] If we consider intention as intentional action, then intentional acting to (try to) square the circle will not make it possible to square the circle. On the other hand, if we consider intentions as plans, we are free to intend and make plans about squaring the circle and also embed it into bigger plans such as 'once we square the circle, we will be able to craft the philosophical stone'. Nonetheless, not even this ability will grant that our intention will become feasible.[21]

An objection to be anticipated is the one that appeals to consistency in order to disqualify from the start proposed examples as cases we should consider in a debate on OIC. I think consistency is somehow overrated and excessively desired and struggled, especially in philosophy of action.[22] Basically, we cannot get rid of the phenomenon of inconsistency when taking into account real-life agency. It is easier to model behavior with classical logic that adopts ideal rationality constraints and fears impossibility as its worst enemy, but a perfectly working model based on oversimplified assumptions will not work, no matter how formally beautiful it is (ask classical economists after a subprime crises).[23] Nonetheless, I think that, especially when dealing with agency, we need adaptive theories that will probably need a more complex technical machinery (which should look inelegant from point of view of the classical rationality constraints) but, somehow, will manage to do the work.[24]

V. Against OIC Rescuers by Separation

OIC rescuers by separation claim that in case you ought to do A and A is impossible, it is possible to show that you can intend to A (even if doing A will still be impossible). For them, that allows the rescue of OIC.

I think that OIC rescuers by separation commit a fallacy of relevance: OIC and their (rev OIC) are too far apart. Even if the case they are making for (rev OIC) were successful, it would not be relevant for the main topic of that discussion, i.e. the rescue of OIC. As long as (rev OIC) is not related to the starting OIC, such a rescue is just a straw man. In order for it not to be a straw man, two options are available to the OIC rescuers by separation:

1.       They can show that there is a link between (rev OIC) and OIC, proving that the separation is just a prima facie separation;

2.      They can claim that, despite there being a real gap between OIC and (rev OIC), what we mean by (OIC) is truly (rev OIC).

Option (1) amounts to prove that (IIA) holds, that is to say that there is a connection between (rev OIC) and (OIC) due to the fact that intentions lead to acting. The problem, here, is that (IIA) is exactly the thesis endorsed by the first kind of OIC rescuers, a thesis that I argued against in the previous section (section 4).

Option (2) requires an extremely detailed work of reconstruction and redefinition of all the uses and different claims involving OIC leading to a revisionary prescriptive proposal that can roughly be asserted as 'all we used to think about OIC was wrong, the True OIC is (rev OIC) and has to do with intentions'. I think that all the comparative and reconstructive work that needs to be done before stating the revisionary prescriptive proposal would be really useful. Nonetheless, I am afraid that this proposal will fail because of its self-ascription to be the only right one. Further, given my OIC attitude, I think that it is difficult to reduce all the issues of OIC as an axiom of deontic logic, as a Kantian principle or as the Latin old but gold set of brocarda impossibilium nulla obligatio and ad impossibilia nemo tenetur to a single True OIC as we have seen above in section 2.

VI. Does Intending the Impossible Offer a Last Chance to Rescue OIC?

One last chance to rescue OIC could be the rather theatrical move to endorse the thesis that it is possible to intend the impossible:

(): ◊Int ø

(I use the empty set symbol – ø –to represent something impossible).

This thesis, strictly speaking, is not much appreciated by OIC supporters, because it admits impossible actions – i.e. something they are opposed to – as something we need to discuss when intending is concerned. (Intending is often used with reference to things one ought to do, so accepting () is the first step to admit that there are impossible actions that are nonetheless obligatory, contra OIC).

Nonetheless, swallowing this harsh bite may help in rescuing OIC. The argument would run as follows: if you can intend to do the impossible, then OIC counter-examples (OA ˄~◊A or OA ˄øA) get neutralized because, despite A's being impossible (~◊A), you can still intend it (◊Int øA) and thus (OIC) is saved. Accepting () will allow both kinds of rescuers to include into their rescuing examples of what might be considered the hardest cases for them, i.e. cases in which we are intending the impossible.[25]

My reply is the following: what I showed before for intending in general (section 4) also holds true for intending the impossible. Intending the impossible does not imply doing the impossible, which, in turn, does not imply succeeding in doing the impossible. What happens, in the case of rescuers by intending-in-action, is that intending the impossible leads you to try to do the impossible, that is, a kind of doing that does not succeed. This doing an attempt (try) that fails is something possible. In particular, given that you are trying to do the impossible, your trying to do it will necessarily fail because you cannot succeed in doing the impossible.

On the other rescuers hand, allowing intending the impossible should make it easier for the rescuers by separation to resist the OIC counter-examples. Nonetheless, as previously shown (section 5), the game they have to win is on the metalevel, where they have to prove the relevance of the link between (OIC) and (rev OIC) in rescue, or justify the revision in favor, of (rev OIC).

VII. Conclusion

In the paper, I clarified what is at stake in bringing intentions into play when we are discussing OIC counter-examples and trying to defeat them. I identified two ways of using intentions to argue for OIC (rescue by intending-in-action and rescue by separation) and tried to develop arguments against both.

Making these rescue moves explicit and challenging them, allows us to progress in the OIC debate. We have a chance to focus on intentions and we can investigate OIC in relationship to attempts and impossible doings which are important topics for legal theory.

I later showed how the appealing thesis (at least for those who wish to unravel the OIC aura) that you can intend to do the impossible is not incompatible with the arguments laid down before.

*    Postdoctoral researcher at the State University of Milan.

[1]  The most impressive collection of OIC counter-examples (as well as of many original replies) is Peter B M. Vranas, 'I Ought, Therefore I Can' (2007) 136 Philosophical Studies 167. On the addicted case see Vranas at 183-185 and fn 18 for a bibliography on the addicted and OIC. On ought but cannot see Wayne Martin, 'Ought but Cannot' (2009) 109 Proceedings of the Aristotelian Society 103; Christopher Jay, Impossible Obligations are not Necessarily Deliberatively Pointless' (2013) 113 Proceedings of the Aristotelian Society 381; Alex King, 'Things we Ought, but Can't' (2014) 27 Ratio 316 and, for another opposition to OIC, see Peter Graham, 'Ought' and Ability' (2011) 120 The Philosophical Review 337. For empirical data on OIC see Moti Mizrahi, 'Ought, Can and Presupposition: An Experimental Study' (2015) 5 Methode 232.

[2]  This kind of intuitive strategy was firstly indicated to me by Wojciech Żełaniec whom I thank. According to Georg Henrik von Wright, Norm and Action (Clarendon Press 1963), 111 the 'can' is the can of ability, not of the success of each individual case. A more complex account that can be seen as a refinement of this idea is provided by Vranas (n 1); Jonny Anomaly, 'Internal reasons and the ought-implies-can principle' (2008) 39 Philosophical Forum 469; Clayton Littlejohn, ''Ought', 'Can', and Practical Reasons' (2009) 46 American Philosophical Quarterly 363 and Ralph Wedgewood, 'Rational 'Ought' Implies 'Can' (2013) 23 Philosophical Issue 70 evaluate the possibilities in terms of having reasons rather than a factual ability to perform the action. My argument against this rescue is probably not enough to be an argument against these defenses of OIC in terms of reasons. Nonetheless, I think this argument is useful to set the stage for such an argumentation.

[3]  Georg Henrik von Wright (n 2), sect. 2, 108. In his later researches von Wright (Id.), An Essay in Deontic Logic and the General Theory of Action. North Holland Publishing Company, 1968; Id. "Norms, truth and logic." In Id. Practical Reason. Blackwell, 1983, 130–209) develops at least three different positions about OIC. On this, see Guglielmo Feis, '"Ought Implies Can" in von Wrightian Deontic Logic' (m.s.).

[4]  An interesting debate on this issue is Gideon Yaffe, Attempts (OUP 2010), 51-64 critique of Michael Bratman, 'Two Faces of Intentions' (1984) 93 The Philosophical Review 375 argument for the thesis that it is possible to try without intending. See also Kirk Ludwig, 'Impossible Doings' (1992) 65 Philosophical Studies 257.

[5]  The Digest reference is Celso's formulation of inpossibilium nulla obligatio, included in the Justinean Roman Digest (50:17:185) as impossibilium nulla obligation est (this time with the m and not with the n as in Celso’s formulation). See also Bloomfield, 'Two Dogmas of Metaethics' (2007) 132 Philosophical Studies 439, Martin (fn. 1), Charles R. Pigden, 'Ought-Implies-Can: Erasmus, Luther and R.M. Hare' (1990) 29 Sophia 2 for references to the discussion in Augustine, Luther, Erasmus. On OIC as a way to cut down norms see R.M. Hare, Freedom and Reason, (Claredon Press, 1963), 59.

One of the earliest and most important linkages between Kant and OIC is found in George E. Moore, 'The Nature of Moral Philosophy'. In Philosophical Papers (Routledge & Kegan Paul, 1922), 317.

As soon as 1946 this Kantian OIC is depicted as a legendary quotation, see David Baumgardner, 'Legendary Quotations and Lack of References' (1946) Journal of the History of Ideas. For some examples of this 'Kantian glorification of OIC' as something that need not be questioned, here are some labels that attests such a glorification 'Kantian principle': David W. Ross, The Right, and the Good (Claredon Press 1930), 5; Arthur Prior, Time and Modality (Claredon 1957) 142; David Widerker, 'Frankfurt on "Ought Implies Can" and Alternative Possibilities' (1991) 51 Analysis 222, 223. 'Kantian "ought implies can" principle': Fred Feldman, Doing the Best We Can. An Essay in Informal Deontic Logic (Reidel 1986), 189. ‘Kant's principle’: Dale Jacquette 'Moral Dilemmas, Disjunticive Obligations, and Kant's Principle that "Ought" Implies "Can''' (1991) 88 Synthese 43, 43. 'The Kantian principle that "ought implies can"': Dirk Baltzly, 'Moral Dilemmas are not a Local Issue' (2000) 75 Philosophy 245, 245. 'Kantian injunction': Alex Blum, 'The Kantian versus Frankurt' (2000) 60 Analysis 287, 287. 'Kantian doctrine': Fred Feldman, Doing the Best that We Can. An Essay in Informal Deontic Logic (Reidel, 1986), 221. 'Kantian dogma': Bloomfield, (above), 439.

On Kantian roots of the principle see at least Robert Stern, 'Does Ought Imply Can? And Did Kant Think It Does?' (2004) 16 Utilitas 42.

For OIC in the literature on moral dilemmas, see at least Bernard Williams, 'Ethical consistency' (1965) 39 Proceedings of the Aristotelian Society, Supplementary Volumes 103; Bastian Cornelius van Fraassen, 'Values and the Heart's Command' (1973) 70 Journal of Philosophy 5; Ruth Barcan Marcus, 'Moral Dilemmas and Consistency' (1980) 77 Journal of Philosophy 121; Terrence McConnell, 'Moral dilemmas', in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (2014).

For the dispute on 'Implies' in OIC see at least Walter Sinnott-Armstrong, 'Ought' conversationally implies 'can'' (1984) 43 Philosophical Review 249; Bart Streumer, 'Does "Ought" Conversationally Implicate "Can"?' (2003) 11 European Journal of Philosophy 219.

On the OIC and Hume's Law see D.C. Collingridge, 'Ought-Implies-Can and Hume's Law' (1977) 52 Philosophy 348; Bloomfield (above).

On OIC as an axiom in deontic logic, see Examples of OIC as an axiom are found in Edward J. Lemmon, 'Deontic Logic and the Logic of Imperatives' (1965) 8 Logique et Analyse39 and Norman O. Dahl, ''Ought" Implies "Can" and Deontic Logic' (1974) 4 Philosophia 485; Jacquette (above).

On OIC and PAP see Widerker (above) and Blum (above), Gideon Yaffe, '"Ought" implies "can" and the principle of alternate possibilities' (1999), 59 Analysis 218; David Copp, '"Ought" Implies "can" and the derivation of the principle of alternate possibility' (2008) 68 Analysis 67; Guglielmo Feis, 'The OIC/PAP Dispute: Two Ways of Interpreting 'Ought' Implies 'Can' in Sofia Bonicalzi, Leonardo Caffo, Mattia Sorgon (Eds.) Naturalism and Constructivism in Metaethics (Cambridge Scholars, 2014).

On doxastic voluntarism and its role in epistemic justification see at least, Feldman (above); William Alston, 'The Deontological Conception of Epistemic Justification' (1988) 2 Philosophical Perspectives 257, and Scott F. Aikin, 'Who is afraid of epistemology's regress problem?' (2005) 126 Philosophical Studies 191.

On von Wright see (n 3). For recent counter-examples to OIC and defences see fn. 1.

[6]  Luigi Ferrajoli, Principia Juris (Laterza 2007), 253. Riccardo Guastini, La sintassi del diritto (Giappichelli 2011), 42 fn. 4.

[7]  Stef Feyen, 'Ought Implies Can and dogmatic enquiry: Some reflections on the methodology of legal scholarship? (2015) 46 Rechtstheorie 425. Feyen provides examples at p. 433 fn. 19 to 22 and pp. 434-435. In general, he offers an overview of OIC issues with a closer application to the law in the first two sections of the paper (pp. 425-439). For a proposal that is more focused on theoretical problems of OIC but deploys some applications as far as pragmatic and interpretations are concerned, see Guglielmo Feis & Chris Fox, ''Ought Implies Can' and the Law' (forth.) Inquiry.

[8]  This is the most used way to write down OIC and expresses the fact that we are mixing two different kinds of modalities. It is quite uncontroversial that OA stands for a deontic modality, whereas it is more controversial how to interpret the diamond (◊) of alethic modal logic (physical possibility, material possibility, …).

[9]  Again, see Vranas (n 1), 173-196, Martin (n. 1) and Graham (n 1) for new challenges.

[10] I shall not deal with the last option. A compensatory obligation to do B if you cannot do A recognizes that A is impossible. The fact that OIC holds for B does not cancel that it did not work for A. For a use of compensatory obligation see von Wright (n 2), 115 when he discusses being punished for the disobedience of an impossible order. Rem Blanchard Edwards, Freedom, Obligation and Responsibility (Springer 1969), 110-111 discusses compensatory obligations linking them to ability.

[11] In this paper, I shall analyze only the formulation with intentions. For some examples of a reason-based debate see Ulrike Heuer, 'Reasons and impossibility' (2010) 147 Philosophical Studies 235 and Bart Streumer, 'Reasons, impossibility and efficient steps: reply to Heuer' (2010) 151 Philosophical Studies 79. See also (n 2).

[12] In that way, they can expand on the von Wrightian idea of 'can' being that of ability. Moving to the level of intentions opens up more conceptual possibilities. It is in fact possible to claim that we can also intend what we have no ability (yet) to do. For example: I can intend to speak Japanese when I have no ability to do it now. That intention actually leads me to start learning Japanese. I am not saying that that option fits easily the rescuer's bill, but nonetheless it is an option.

[13] Vranas (n 1) holds that OIC is true if we construct it as a link of conceptual necessity between an obligation understood as a reason for action and 'can' as the sum of opportunity and ability; also von Wright (n 2) reads can as 'ability' and not as success.

[14] Given that we have the accordion effect in the philosophy of action (see Joel Feinberg, 'Action and Responsibility' in Max Black (ed.) Philosophy in America (Cornell University Press, 1965)), we can call this "the OIC elevator".

[15] One of the first accounts in the philosophy of action that valued consistency and coherence as requirements is Bratman (n 4).

[16] As a whole, the rescuers' researches are valuable because they help us to better understand what's inside the 'can' of OIC.

[17] A similar example is 'kidnapping Santa Claus' as Santa, does not exist as the source of the eternal youth in the other example.

[18] I think it is worth to investigate the different kinds of situations that may cause (IIA) to fail, nonetheless this research – that might be generalized into a research project on OIC starting from its (supposed) failures – is not part of the present work.

[19] This example might appear trivial but cognitively it is not. Innovation and discoveries are often due to people that want to go against the limits of what is considered impossible. Think about any Apple commercial or Google anecdotes. A more philosophical insight on this is provided by David Lewis, Counterfactuals (Blackwell 1973), 91: 'We will certainly construct ersatz worlds that disobey currently accepted physical laws; for instance, ersatz worlds where mass-energy is not conserved. Still, we cannot be sure of getting all possible worlds, since we cannot be sure that we have constructed our ersatz worlds at a high enough level of generality. If we knew only the physics of 1871, we would fail to cover some of the possibilities that we recognize today. Perhaps we fail today to cover possibilities that will be recognized in 2071. Our modal opinions do change, and physicists do a lot to change them. But this is not to say that we can argue from the contingent results of empirical investigation to conclusions about what possibilities there are. It is only to say that when we find it hard to locate our actual world among the possibilities that we recognize, we may reasonably be stimulated to reconsider our modal opinions. We may try to think of credible possibilities hitherto overlooked, and we may consider whether we are still as sure as we were about those of our modal opinions that have turned out to be restrictive. It is this reconsideration of modal opinions that may influence our construction of ersatz worlds, not the results of empirical investigation itself. We are concerned not with physics proper, but with the preliminary metaphysics done by physicists.'

[20] Bratman (n  4).

[21] A way to obtain this could be to load the concept of intention, plan, intentional action and so on with strong consistency requirements. Nonetheless, the point of the debate here is that of testing counter-examples to OIC – that is already assumed as a coherence principle, cf. Dahl (n 5) and Jacquette (n 5) for its role in deontic logic or in the moral philosophy issues of moral dilemmas. If we really want to do a fair test, I think we do not have to set the terms of debate using intention, plan, etc. in a way that is OIC-dependent (i.e. with strong consistency requirements). In the OIC debate there is this tendency to make it the case that possible strong counter-examples against OIC are not even worth to enter into the philosophical arena. Consider what Vranas (n 1), 184 says on a more extreme version of the addicted example I may call the 'conceptual addicted', i.e. the mere conceptual possibility that there may be an addicted who is necessarily addicted and cannot refrain from her addiction in every possible situation. According to Vranas' reply, even admitting such a case for the sake of the argument, then the zombie addicted will have no obligation. Such a move seems quite a sort of petitio principiis to me. Given that he wants to defend OIC he has to say that the conceptual addicted has no obligation because he cannot resist her addiction. With such a preliminary setting, it will turn out that the strongest OIC counter-example that the opponent of OIC may produce is not recognized as threatening by Vranas because he has disqualified it from the beginning.

[22] See Bratman (n 4), 381 in which consistency matters a lot and has a key role in his argument: 'The argument for strong consistency provides the basis for my argument against the Simple View'. Now we live in 'days of rampant pluralism' not only in ontology, but also in logic. We have many more logical tools and options (e.g. defeasible logics, conflict tolerant (deontic) logics, relevance logics, paraconsistent logics, dialetheistic logics) to deal with incoherence, inconsistency, irrationality and conflicts. Further, experimental economics and experimental game theory helped us to realize that we do not always act rationally or follow standard game-theoretic strategies.

[23] A similar point consisting in a general warning on using models that are formally simple and elegant but, once tested on real cases, fail to grasp the complexities we are interested in can be found in Gideon Yaffe, 'Reply to Jan Broersen, Thomas Nadelhoffer and Steven Sverdlik' (2012) 3 Jurisprudence 483 in replying to Jan Broersen, 'Three Points of Disagreement with Gideon Yaffe on Attempts' (2012) 3 Jurisprudence 467 comment of Yaffe (n  4).

[24] Formal theories allow for the use of more than one model at time that take into account different features (that may clash with one another). We have dynamics and non-monotonic logics as well as defeasible logics: given than even formal theories are no longer afraid to go deep down into inconsistencies and incoherence, why should philosophies willing to explain and understand what goes on when we act do so?

[25] At first glance, these cases could be conceived as a sort of 'double' OIC counter-examples: intending A is impossible and thus is doing A.


Back to Top