Prijeđi na kretanje stranicama – Plan stranice
Contributions from the 1st Genova-Slavic Seminar in Legal Theory

Defeasibility, norms and exceptions: normalcy model

Vojko Strahovnik
p. 61-76
Uklonljivost, norme in izjeme: model normalnosti


The paper discusses the notion of defeasibility and focuses specifically on defeasible (moral and legal) norms. First, it delineates a robust notion of the phenomenon of defeasibility, which poses a serious problem for both moral and legal theory. It does this by laying out the conditions and desiderata that a model of defeasibility should be able to meet. It further focuses on a specific model of defeasibility that utilises the notion of normal conditions (normalcy) to expound the robust notion of defeasibility. It argues that this model fails in its attempt to do this, particularly since it presupposes further pertinent norms and we have reasons to doubt if these are defeasible. It thus does not allow defeasibility to go “all the way down” in the normative domain and limits it merely to a feature of some sort of mid-level norm. In conclusion, it draws lessons from this and positions defeasibility models within a more general pluralistic approach to norms.

Vrh stranice

Cjelovit tekst

1 Introduction

  • 1 Ferrer Beltrán & Ratti 2012; Guastini 2012; Hooker & Little 2000; Lance, Potrč & Strahovnik 2009.
  • 2 Chiassoni 2012: 162; Lance and Little 2007.

1The concept of defeasibility, especially in the legal and moral domains, has become an increasingly popular point of discussion in the last couple of decades.1 Defeasibility is a multi-faceted concept which is used in different senses and can be related to various subjects. There are several open questions or dimensions in relation to it. First, there are a number of candidates for being defeasible, amongst them concepts, norms, norm formulations, rules, standards, principles, laws, generalisations, ideals, reasoning, facts, opinions, statements, decisions, regulations, kinds, etc.2 For the purposes of this paper, I shall focus my attention on defeasible norms, with a prospect that what will be established will in general be transposable, in a more or less direct way, to other defeasible phenomena in close proximity (principles, rules). Next, there are several important open questions and diverging views about the origins, nature, and scope of defeasibility. And finally, there is the question about the consequences of defeasibility for the theoretical aspects of the given normative domain, as well as for normative practice. Answers to these questions vary in the debate with deep theoretical disagreements on almost all the said aspects.

2I shall proceed in the following way. After examining briefly some general considerations of defeasibility, I delineate a robust notion of defeasibility in relation to the notion of exception. This notion represents a serious problem for any theory of defeasibility and defeasible norms, including those formulated within moral and legal theory. In delineating the notion, I put forward a number of conditions and/or desiderata that a model of defeasibility should meet (§2). In §3, I focus specifically on a model of defeasibility that utilises notions of normal conditions and normalcy to expound my robust notion of defeasibility. After presenting examples of a normalcy-based model in the fields of both legal and moral norms, I go on to argue that this model fails in its attempt to do this, particularly since it presupposes further pertinent norms and we have reasons to doubt if these are defeasible. It thus does not allow defeasibility to go “all the way down” in the normative domain and limits it merely to a feature of some sort of mid-level norm. In conclusion, (§4) some lessons from this debate are discussed and briefly related to the traditional pluralistic model of norms.

2 Defeasible norms and exceptions: conditions and desiderata

  • 3 In Plato’s dialogue Statesman, we can follow a debate between Socrates and a young stranger from El (...)
  • 4 Dworkin (1977: 24–25), amongst others, maintains such an optimistic view: “Of course a rule may hav (...)
  • 5 I do not want to deny that there are some who would claim that what defeasibility boils down to is (...)

3The debate about defeasibility can be situated within a more general debate about the relationship between general principles and particular cases, which has been present in philosophy since its early beginnings.3 These discussions focus on the relationship between general norms on the one hand, and particularities or exceptions on the other, but frequently such understandings of exceptions are not radical enough since exceptions are understood to be mere side effects of underspecified or incomplete norms, which could in principle be somehow avoided.4 But, as I seek to claim, genuine, robust defeasibility understood to be a “serious problem” goes beyond this and includes genuine exceptions, which are not such that they could already be properly explicitly included in a general norm or fully specified in advance and thus in principle avoided.5

  • 6 Brożek 2014.

4To get an initial grip on the concept of defeasibility, it is common and useful to relate it to the concept of an exception in general or to the presence of (the possibility) of exceptional cases in particular, i.e. cases which, on the one hand, fall under a certain norm or rule, but which, at the same time, have unbefitting normative consequences which we tend to exclude these cases from falling under the mentioned norm or rule. Along these lines, Brożek claims that a “rule of the form A => B is defeasible iff it is possible that although A obtains, B does not follow.”6 The notion of an exception or an exceptional case, as opposed to normal cases, is thus one of the hallmarks of defeasibility.

5What do we in fact mean when we say that, e.g., a certain norm, rule, reasoning or concept is genuinely defeasible? We must add some further amendments to the initial grasp of the concept as described above. I put forward these amendments in the form or conditions or desiderata, which will, at the same time, serve as a guide in constructing or evaluating a theory of defeasibility. Proceeding this way, I do not claim that all of the conditions and desiderata must necessarily be met in a straightforward way; another option is to propose a suitable accommodation of a given condition or desideratum.

  • 7 Celano (2012: 281) labells such exceptions as “true exceptions” (as opposed to merely prima facie e (...)

6(I) First, an exception must in a sense be a “genuine” exception, meaning that the exception is not merely a consequence of an initially poorly specified norm.7 The defeasibility of norms is in this way not merely due to their incorrect, imprecise, or vague formulation that could in principle be resolved or more clearly spelled out on demand. A model must allow for genuine exceptions which cannot be specified in advance (genuine exceptions condition).

7(II) Second, defeasible norms are not to be associated merely with some kind of “rule of thumb” norms, which we can use most of the time, but which we are also able, if required, to specify and turn into exceptionless norms. If defeasible norms were associated with just this type of norms and exceptions to them, they would neither be a particularly interesting phenomenon nor would they pose a serious problem. Defeasible norms must in this sense be fundamental, full-fledged norms and such that, after them, judgment takes over (as opposed to other norms) (fundamentality condition).

8(III) Third, a set of possible exceptions must in principle be open, meaning that we can never specify all the possible exceptions to a norm and, in turn, round the set off. If that were the case, then these exhaustively specified exceptions could be built into the norm itself and the norm would cease to be defeasible (open-endedness condition).

9(IV) Fourth, a defeasible norm remains the same and retains its normative power even when we are able to find an exception to it. In this sense, it “survives” the exception and can hold in all further, non-exceptional cases. Being prepared to abandon or modify a norm when we encounter an exception would make the phenomenon of defeasibility fairly empty. Thus, a model of defeasibility must be such that it leaves the initial norm intact when we come across an exception to it. A defeasible norm must survive beyond the point of arriving at an exception and remain the same norm as before an exception was identified (identity condition).

10(V) Fifth, a defeasible norm must be able to remain in “normative space” even in the case of an exception and can shed light on the nature of the exceptional case or can indirectly influence the final normative solution. A defeasible norm must be such that it has a possible normative pull even in exceptional cases (possible relevance condition).

11(VI) Sixth, defeasibility must go “all the way down” in the normative field and include the most basic norms for a given domain (basicity condition).

  • 8 Cf. Strahovnik 2004.

12(VII) And lastly, a model of defeasibility must preferably be able to cover or accommodate the legal domain, morality and other normative domains, including – although not being limited to – epistemology, aesthetics,8 social conventions, and etiquette (generality desiderata). The paper focuses primarily on moral and legal norms, hoping that the discussion in hand will at least start paving the way for a more general discussion of defeasible norms that I will have to leave for some other occasion.

13Some of these points about the conditions stated above can briefly be demonstrated with a rather simplified example of a supposedly defeasible moral norm N expressed as: “N: Causing pain is morally wrong”. For this norm to be defeasible, (i) it must allow for exceptions, i.e., cases of causing pain which are not morally wrong, or in which pain is not wrongdoing (e.g., cases of justified medical treatment where pain is unavoidable or cases of causing pain as part of sports activities); (ii) these exceptions must be genuine exceptions in the sense that they are not merely a matter of an imprecise, rule of thumb formulation of the underlying fundamental norm N* (e.g., “The unwarranted causing of pain is morally wrong.”); (iii) the set of possible exceptions must be open in the sense that we cannot reformulate the norm otherwise so as to include all the exceptions (e.g., “N**: Causing pain is morally wrong except in cases in which this is part of a justified medical procedure or athletic achievement.”); (iv) the initial norm N must remain the same and must retain its normative power even after stumbling upon an exceptional case in the sense that the next time a paradigmatic case of pain-causing pops up it will still render our judgement about its wrongness warranted; and (v) the norm N must remain part of normative space and must influence our judgment indirectly (e.g., if there were two options available for the performance of a given medical procedure, both involving pain, although one substantially less than the other, then N would still be part of our judgment about which one is morally optimal).

14There are several models of defeasibility which try to capture the core of its facets related to exceptions as revealed above (e.g., utilising the notions of indeterminacy, vagueness, open-texture, etc.). One of the most prominent models is the one that focuses on the notion of normalcy. I shall now turn to this model to see how it helps to elaborate the understanding of defeasible norms and if it meets the said conditions and desiderata.

3 Defeasibility and normalcy

  • 9 Celano 2012: 285–287.
  • 10 Lance and Little 2007; 2008.
  • 11 Celano 2012: 284.
  • 12 Celano 2012: 286.

15One way to spell out the defeasible nature of a given norm is to state that it only holds in normal conditions or in normal circumstances.9 Besides utilising the notion of normalcy, authors sometimes talk of “privileged conditions”, “typical conditions”,10 “what standardly, or normally happens”,11 “paradigmatic”12 cases or central cases. The basic idea behind all of them is the same. A given norm applies only within a set of normal circumstances, which are such that they cannot be explicitly fully stated and included in the norm as such. Such a proposal seems to be well in line with the above highlighted connection between defeasible norms and genuine exceptions, since exceptions represent exactly those cases which fall outside the scope of normal conditions.

16In what follows, I shall first focus more closely on two particular models of defeasible norms that employ such a normalcy condition. The first model is proposed by Bruno Celano and is aimed at norms in general, although it concerns mostly legal norms and originates from within the debate about legal defeasibility. The second model is proposed by Mark Lance and Margaret Little primarily for the domain of moral norms, but this model could also be transposed to norms in general. These models are supposedly complementary and Celano explicitly appeals to the latter model as a supplement to his own proposal.

  • 13 For readers unfamiliar with the debate between generalism and particularism, this is a brief recapi (...)
  • 14 Celano 2012: 270.

17Celano puts forward his proposal as part of his defence of limited particularism13 concerning norms. He begins with considering potential response strategies to the possibility of conflicting norms and therefore the need to allow for an “exception” with respect to at least one norm involved in this conflict. First, he rejects the specificationist approach, which proposes that, when norms conflict with each other, all “we have to do is specify (that is, suitably restrict the domain of application of) at least one of the norms, or the relevant norm, so that, thanks to the inclusion of further conditions within its antecedent [...] the conflict – or the unsatisfactory verdict – eventually vanishes”.14 What we seem to be doing is enriching, refining, qualifying and grasping the subtleties of the initial norm, thus treating it as defeasible. But the problem lies, first, in the implausibility of insisting that we are still dealing with the same initial norm even after many amendments have been made and exceptions recognised. Second, there is the danger of an in-principle impossibility of specifying all the exceptions.

  • 15 Celano 2012: 276.

Achieving a fully specified ‘all things considered’ norm, thereby ruling out the possibility of further, unspecified exceptions (apart from those already built into the norm itself) would require us to be in a position to draw a list of all potentially relevant properties of the kind mentioned. And this, we have seen, is misconceived.15

18Celano also rejects a similar approach to defeasibility which regards exceptions as always already implicitly included or provided for by a norm. A given norm is thus just a sort of shorthand for a more complex norm that lies in the background and can – if needed – be brought to the foreground. But this approach fails for the same reasons since it understands exceptions not as true exceptions – not as real “holes” in the norm – but as some sort of prima facie exceptions that allow for the filling in of the holes.

  • 16 The “moral flatness” worry was raised in moral theory by several authors (McNaughton and Rawling 20 (...)

19According to Celano, one must thus accept a moderate sort of particularism in order to do justice to (the possibility of) conflicts of norms and genuine exceptions. In line with this, he proposes an understanding of norms as defeasible conditionals supplemented by a “normalcy condition”. The reason why this position is labelled as moderately particularistic – as opposed to radical particularism – is the so-called normative flatness worry. The leading idea is that radical particularism cannot properly account for the thought that some considerations are more central than others, in the sense that we recognise some reasons are “normally” relevant and more central than others.16 The normalcy condition is thus supposed to work both ways, i.e., allow and account for the possibility of genuine exceptions, as well as provide some sort of “basic patterns” in the normative domain. Here is the core of Celano’s proposal:

  • 17 Celano 2012: 285.

Norms are defeasible conditionals liable to true exceptions, i.e. conditionals such that the consequence follows, when the antecedent is satisfied, under normal circumstances only. The crux of the matter is, of course, how is the qualification ‘normally’ (‘in normal circumstances’, etc.), to be understood? Here, it seems, normalcy includes, but does not boil down to, the notion of what happens, or holds, ‘in most circumstances’.17

20He is well aware that this proposal is not without problems. Nevertheless, he leaves it open and appeals to the work of Lance and Little as it provides more details for such a model of normalcy.

  • 18 They defend the model in a series of papers, Lance & Little 2005; 2006; 2007; 2008.

21I shall now turn to the Lance and Little model (addressing specifically moral norms or principles, although similar considerations could be put forward for norms in general). Lance and Little are primarily concerned with the functioning of reasons in general, and with the variability of moral reasons in particular. They employ the notion of privileged conditions, in which a given reason ‘behaves normally’, as opposed to conditions which are not privileged and in which a reason can change its moral relevance.18 The model is committed to deep moral contextualism: the right- or wrong-doing, and good- or bad-making features of actions vary in contexts in ways that preclude codification by exceptionless principles. A full-fledged recognition of exceptions to moral generalisations does not mean that the picture of morality that one must accept is entirely free from any important kind of generalities. The sharp divide between generalism and particularism results in the following consequence: their respective views about the nature of explanation are either too strict or too narrow. According to these views, genuine explanatory reasons must be governed by universal exceptionless principles.

22An alternative model of explanation figuring defeasible norms and exceptions is offered, a model that covers non-moral ground as well. The features of such acts as promise-keeping, lying, inflicting pain or being kind are the building blocks of everyday morality that entertain an intimate connection with their moral import (as core moral reasons that can de-flatten the moral normative landscape). They are genuine explanatory features of the moral status of acts and may be captured within defeasible generalisations. Defeasible norms (e.g., “Defeasibly, lying is wrong”, Defeasibly, killing is wrong” or “Defeasibly, causing pain is wrong”) are introduced through the notion of privileged conditions. A more general formulation of such principles is:

“Defeasibly, for all actions x: if x is A, then x is wrong/you ought not to do x.”


“In privileged conditions, for all actions x: if x is A, then x is wrong/you ought not to do x.”

  • 19 The same holds in other areas as well. For example, the non-moral generalisation “Fish eggs develop (...)

23This model thus argues that some features of acts can entertain an intimate connection with their moral import and are genuinely explanatory of the moral status of acts, although it allows for exceptions.19 The singling out of a connection between a particular descriptive feature of such acts as “causing pain” and the negative moral import of this feature which is neither necessarily universal nor pervasive nor usual can be done by saying that, defeasibly, causing pain is wrongdoing.

  • 20 There are cases in which a defeasible generalisation tracks paradigm cases, which are, in this sens (...)
  • 21 Lance and Little 2008: 64–68.

24When a defeasible generalisation faces an exception, something has gone off course – the context has relevantly changed in respect to privileged conditions, and our moral understanding must track this. There are several types of such defeasibility dynamics, such as the paradigm/riff, justificatory dependence, and idealisation/approximation.20 Moral understanding is the understanding of the structure of moral privilege and exceptions. One must understand the nature of a certain feature in privileged conditions, and when outside such a context, the relation of the last context to the first one, the compensatory moves required, and the acceptability of various deviations.21

  • 22 McKeever and Ridge (2006: 60–72) put forward quite a forceful argument in this direction. First, on (...)

25It seems that this model of defeasible norms gives us what we are looking for, i.e., a model of norms that would allow for genuine exceptions and, at the same time, de-flatten a given normative landscape. Now, I want to mount a challenge to this model of normalcy. This proposal of defeasible norms either fails to de-flatten the normative landscape or succeeds in this task, but only at the expense of positing more basic and plausibly indefeasible norms. I shall not pursue the first horn of this dilemma, although there are arguments that cast serious doubts about the success of this proposal in this regard.22 What I want to do is to initially grant that the model succeeds in de-flattening the normative landscape, i.e., in making a distinction between core and other marginal reasons, and consequently finding a sufficient number of so-called defeasible reasons and norms that would function as basic normative building-blocks.

26But it seems that this is so only at a cost of reducing this distinction to a distinction between basic and derivative reasons, and thus seemingly limiting defeasibility to the level of mid-axiom norms. In particular, the proposed model seems to collapse to a model according to which a given consideration together with normal (privileged) conditions delimits a central and apparently invariable reason and appeals to a more basic norm which has full explanatory power, or in which normal (privileged) conditions include reference to further reasons and norms. For instance, it is not merely lying or telling someone something untrue which is normatively central from a moral standpoint, but honesty, sincerity and deception are. The intermediate conclusion is thus that the normalcy-based model of defeasible norms enables one to explain or justify an exception to the initial norm as an exception only by appealing to some further pertinent norm. If one looks closer at the examples proposed by Lance and Little, it does indeed seem that this is the case.

27For example, in relation to a norm that lying is defeasibly or in normal circumstances wrong the model refers to examples in which lying is not wrongdoing (e.g., in circumstances of playing the Diplomacy game or in circumstances when a Nazi officer bullies you into revealing the location of his next targeted victim). The way in which these cases can be understood is that privileged conditions are those that do a lot of both normative and explanatory work in respect of why lying is wrongdoing in “normal” cases, but not in the mentioned ones. This is so since the space of privileged conditions is further shaped with basic moral considerations, consisting of such notions as consent, fidelity, justice, honesty and the like. And it is this large chunk, including invariant moral reasons and norms, which pops up in an explanation of the moral status of a certain feature and, in turn, of the whole act. This then just transposes the question whether these background basic normative considerations are defeasible or not.

  • 23 Little 2001: 34, emphasis mine.
  • 24 Lance and Little 2007: 153, emphasis mine.

28If we look more closely at some of the mentioned examples, this worry becomes apparent. In the case of lying, the model claims that “intentionally telling a falsehood” is not wrong “when done to Nazi guards, to whom the truth is not owed”,23 or it is not wrong because a particular person is not worthy of the truth since “part of what it means to take something to be a person […] is to understand the creature as belonging to a kind that defeasibly has a claim on our honesty. Situations in which one takes something to be a person but not worthy of honesty are inherently riffs, as it were, on the standard theme of person.”24

29Thus, in such cases, one can plausibly claim that what actually functions as a reason in this and other cases is a combination of certain features of action and privileged conditions, which make reference to some further, more basic reasons and seemingly indefeasible norms. But are these pertinent norms which are appealed to in explaining or justifying exceptions to the original norms themselves defeasible? Admittedly, our appealing to norms has to stop somewhere, with judgment taking up the slack. But the normalcy model of defeasibility must then provide further reasons for these more basic norms to also be defeasible and for them to be part of the same normalcy model, i.e., for their defeasibility to also be explained and elaborated by appealing to normal conditions.

30Lance and Little anticipate the objection raised above. They reply that this rising to a higher level of abstraction (e.g., from lying to honesty, or form causing pain to cruelty) might seemingly offer us a more stable ground and order when it comes to the invariability and indefeasibility of reasons and norms. They respond by claiming that, (i) in making this move, one loses something important, namely the intimate connection that lying itself has to moral wrongness, that “being a lie” is the main driving force behind such an action being wrong; and (ii) that, even on a higher or thick moral level, the normative domain is full of exceptions, which are revealed in statements such as “it hurts so good” or “sometimes you must be cruel to be kind”. Therefore, even seemingly indefeasible considerations, such as cruelty or honesty, are not invariant moral reasons and might figure only in defeasible norms.

31This first point is crucial to all attempts that combine variability with moral generalities, since they must convince us that what functions as a moral reason in a given case is really variable and that the rest of what a moral principle refers to is not a part of this reason. As far as the intimate connection between simple moral reasons and the rightness or wrongness of acts is concerned, we must ask ourselves what does the explanatory work. Maybe we often cite such things as “telling a lie” or “keeping one’s word” as reasons, but if privileged conditions for such considerations encompass such things as honesty, sincerity and fidelity, the question of their role is justified. When privileged conditions change into non-privileged ones, these considerations are exactly those that we employ in our explanation of why a case is deviant or defective in relation to the standard one. And it further seems that they do not simply function as enablers and disablers of initial reasons, but are employed as basic reasons themselves.

32Regarding the second point, most of such talk must be understood to be metaphorical. If we must sometimes be “cruel to be kind”, then it is most probably not the cruelty itself that makes our action kind or be the ground of its moral rightness. Let us imagine a more detailed case. Let us say that I have to give my friend an honest opinion about her project or action, and that I know that it would be painful for her to hear, but, on the other hand, it would spare her the frustration in the long run. In this sense, I shall be cruel to her or brutally honest, but, at the same time, this is the only way of convincing her to give up some actions and maybe spare her the future disappointment and pain. In this sense, we can say that I have to be “cruel to be kind”. Nevertheless, the “cruelty” here involved is not a reason that contributes to the moral rightness of my action. If there was a way of convincing my friend that was not cruel but just “plainly kind”, then it would be morally wrong, even horrible to pick the first option. We can never be cruel just to be cruel and get away with it morally. Another way to respond to a case such as this would be to claim that the question of the role of cruelty simply does not arise at all since this is not a case of cruelty. The situation is similar with respect to the “it hurts so good” statement that Lance and Little appeal to and other similar cases.

  • 25 In this discussion, I have mostly focused on moral norms, but the same applies to defeasible legal (...)

33All this reduces the proposal to understand defeasibility in terms of normalcy to a point when the model allegedly must accept at least some basic, indefeasible norms. Given the examples above, one can conclude that the “normalcy” approach does not go “all the way down” in the normative domain. In the end, it allegedly presupposes at least some basic, invariable and indefeasible norms effective in the background of the normative domain, serving as a line of demarcation between normal circumstances or privileged conditions on the one hand, and exceptional or unprivileged ones on the other. In this way, the model does not meet either the fundamentality condition or the basicity condition.25

4 Defeasible norms and pluralism

  • 26 Tur 2001.

34Such models of defeasibility as those developed by, e.g., Celano, Tur,26 and Lance and Little are attractive models, but their limits are manifested in the fact that the defeasible norms which they propose have to be understood against a normative background of basic reasons that we appeal to in deciding what the relevant case is or in our interpretations of a given norm and the normalcy of the conditions. It seems that, irrespective of the way in which we work out the defeasibility structures of defeasible norms, their models must appeal to some wider set of basic (moral, legal, etc.) considerations related to norms which reside in the background and illuminate the exceptions. Normalcy-based defeasibility is thus limited to surface or mid-level norms only. In §2, I have put forward a number of conditions and desiderata for a model of genuine defeasibility. Then, I have presented a model based on the notion of normalcy, and have shown that it fails to meet some of them and that defeasibility within it does not reach “all the way down” in the normative domain, and nor does it include the most basic norms.

  • 27 Ross 1930; 1939.

35One lesson to learn from this is that one can plausibly proceed by developing an account of defeasibility which explicitly includes an appeal to a wider set of an evaluative, axiological or normative background of basic normative considerations. In the domain of moral normativity, such a model is proposed by moral pluralism understood in the following way. Traditional moral pluralism27 builds on the notion of basic moral reasons which can be captured by general, indefeasible norms (e.g., principles specifying prima facie duties). Together with what is morally relevant in a given situation prima facie duties jointly determine the moral status of an action. A final assessment of an actual duty requires careful consideration, the weighing of reasons and seeing how they fit together (judgment).

  • 28 Ross 1930: 27–28.
  • 29 Ross 1939: 190.

36There has been quite a lot of controversy over just how exactly we could understand the notion of prima facie duty proposed by W. D. Ross and the mentioned relation between prima facie duty and an actual duty. One of the more plausible suggestions or interpretations is that in terms of basic moral reasons that we shall employ here. According to this understanding, basic prima facie duties (Ross’s famous list of seven duties includes: fidelity, gratitude, reparation, beneficence, justice, self-improvement and non-maleficence) are best understood to be basic moral reasons or considerations that are always morally relevant and count for or against an action. Given their thick moral descriptions, they seem indefeasible. Besides this basic list, Ross also employs so-called derivative or complex prima facie duties. My suggestion is that we could understand these derivative prima facie duties to represent derivative moral reasons and to be prime candidates for mid-level defeasible norms that we often employ in our ordinary moral thought. Let us have a look at an example. Ross claims that duties, such as obeying the laws of one’s county, are derivative, comprising one or more basic prima facie duties; in the case of the mentioned duty these are a prima facie duty of gratitude for the goods that an individual receives from his or her country, a prima facie duty of fidelity emerging out of an implicit agreement of living with others in the same country by some shared rules and a prima facie duty of benevolence since obeying laws contributes to the common good.28 The same applies to our prima facie duty not to lie or to keep the promises we have given. Both are grounded in the basic prima facie duty of fidelity. Ross also refers to such derivative prima facie duties as media axiomata that represent a helpful way of applying general moral norms related to prima facie duties in particular types of situations.29 What we have here is a model of defeasible norms which is explicit about the normative background consisting of basic, indefeasible norms. The claim is not that these basic norms cannot come into conflict (allowing for conflict between norms that cannot be resolved by appealing to some further meta-principle is one of the key posits of pluralism) and that occasionally we are faced with exceptions to prima facie duties, but that these exceptions are related to a dimension of a given prima facie duty ending up also being an actual, final duty.

37This model is thus similar to the normalcy-based approaches that I have investigated, although it does differ from them in that it explicitly recognises basic indefeasible norms, which explicitly limits normalcy-based defeasibility to the mid-level of the normative domain. Accordingly, one way to look at the arguments presented above is that they aim to collapse normalcy-based approaches to a more general pluralist approach, claiming that defeasibility is a feature of basic or fundamental norms and that we do not need indefeasible norms within such a model.

38In more general terms, we can thus distinguish between two fundamentally different views about the possibility of “codification” of a given normative domain. On the one hand, there is generalism, which combines the possibility of codification and a deductive model of normative thought, and, on the other. particularism, which rejects the possibility of (the complete) codification of a given domain of normativity. The pluralist approach seems to sit between these two approaches. Within the latter, normalcy-based defeasibility can be interpreted as a consequence of the structure of normative pluralism and the norms within it, and is plausibly understood in relation to an indefeasible normative and axiological background which stands in relation to particular cases of judgment.

Acknowledgment. I would like to thank the audience at the Genoa-Slavic Seminar in Legal Theory (Genoa, 11–12 December 2014) for a lively discussion based on a draft of this paper. While preparing the paper, the remarks of two anonymous referees, as well as an ongoing discussion of these issues with Terry Horgan and Matjaž Potrč, have been very helpful.

Vrh stranice


ARISTOTLE, 1908: Nicomachean Ethics (trans. W. D. Ross). Oxford: Clarendon Press.

David BAKHURST, 2000: Ethical Particularism in Context. Moral Particularism. Eds. Brad Hooker & Maggie Little. Oxford: Clarendon Press. 157–177.

Bartosz BROŻEK, 2014. Law and Defeasibility. Revus. Journal for Constitutional Theory and Philosophy of Law, 23. 165–170. URL:

Bruno CELANO, 2012: True Exceptions: Defeasibility and Particularism. The Logic of Legal Requirements. Essays on Legal Defeasibility. Eds. Jordi Ferrer Beltrán & Giovanni Battista Ratti. Oxford: Oxford University Press. 268–287.

Bruce CHAPMAN, 1998: Law Games: Defeasible Rules and Revisable Rationality. Law and Philosophy 17(4). 443–480.

Pierluigi CHIASSONI, 2012: Defeasibility and Legal Indeterminacy. The Logic of Legal Requirements. Essays on Legal Defeasibility. Eds. Jordi Ferrer Beltrán & Giovanni Battista Ratti. Oxford: Oxford University Press. 151–181.

Roger CRISP, 2000: Particularizing Particularism. Moral Particularism. Eds. Brad Hooker & Maggie Little. Oxford: Clarendon Press. 23–47.

Jonathan DANCY, 2000: The Particularist’s Progress. Moral Particularism. Eds. Brad Hooker & Maggie Little. Oxford: Clarendon Press. 130–156.

Jonathan DANCY, 2004: Ethics without Principles. Oxford: Clarendon Press.

Jonathan DANCY, 2013: Moral Particularism. The Stanford Encyclopedia of Philosophy. Ed. Edward N. Zalta. URL:

Ronald DWORKIN, 1977: Taking Rights Seriously. London: Duckworth.

Jordi FERRER BELTRÁN & Giovanni Battista RATTI, 2012: Legal Defeasibility: An Introduction. The Logic of Legal Requirements. Essays on Legal Defeasibility. Eds. Jordi Ferrer Beltrán & Giovanni Battista Ratti. Oxford: Oxford University Press. 1–7.

Riccardo GUASTINI, 2012: Defeasibility, Axiological Gaps, and Interpretation. Legal Defeasibility: An Introduction. The Logic of Legal Requirements. Essays on Legal Defeasibility. Eds. Jordi Ferrer Beltrán & Giovanni Battista Ratti. Oxford: Oxford University Press. 182–192. [Slovenian translation: Riccardo GUASTINI, 2010: Uklonljivost, vrednostne praznine in razlaganje, Revus – Revija za evropsko ustavnost (2010) 14. 41–56. URL: DOI: 10.4000/revus.1332]

Brad HOOKER & Maggie LITTLE (Eds.) 2000: Moral Particularism. Oxford: Clarendon Press.

Mark LANCE & Maggie LITTLE, 2005: Particularism & Anti-Theory. The Oxford Handbook of Ethical Theory. Ed. David Copp. New York: Oxford University Press. 567–594.

Mark LANCE & Maggie LITTLE, 2006: Defending Moral Particularism. Contemporary Debates in Moral Theory. Ed. James Dreier. Oxford: Blackwell. 305–321.

Mark LANCE & Maggie LITTLE, 2007: Where the Laws Are. Oxford Studies in Metaethics, Vol. II. Ed. Russ Shafer-Landau. New York: Oxford University Press. 149–171.

Mark LANCE & Maggie LITTLE, 2008: From Particularism to Defeasibility in Ethics. Challenging Moral Particularism. Eds. Mark Lance, Matjaž Potrč & Vojko Strahovnik. New York: Routledge. 54–74.

Mark LANCE, Matjaž POTRČ & Vojko STRAHOVNIK (Eds.), 2009: Challenging Moral Particularism. New York: Routledge.

Sean MCKEEVER & Michael RIDGE, 2006: Principled Ethics: Generalism as a Regulative Ideal. New York: Oxford University Press.

Sean MCKEEVER & Michael RIDGE, 2007: Turning on Default Reasons. Journal of Moral Theory 4(1). 55–76.

David MCNAUGHTON & Piers RAWLING, 2000: Unprincipled Ethics. Moral Particularism. Eds. Brad Hooker & Maggie Little. Oxford: Clarendon Press. 256–275.

William David ROSS, 1930: The Right and the Good. Oxford: Clarendon Press.

William David ROSS, 1939: The Foundations of Ethics. Oxford: Clarendon Press.

Frederick SCHAUER, 2012: Is Defeasibility an Essential Property of Law? The Logic of Legal Requirements. Essays on Legal Defeasibility. Eds. Jordi Ferrer Beltrán & Giovanni Battista Ratti. Oxford: Oxford University Press. 77–88.

Vojko STRAHOVNIK, 2004: The Riddle of Aesthetic Principles. Acta analytica. 19(33). 189–207.

Vojko STRAHOVNIK, 2012: Defeasibility of Moral and Legal Norms. Dignitas 53/54. 101–115.

Vojko STRAHOVNIK, 2013: Legal positivism and defeasibility of legal norms. Dignitas. 59/60. 219–234.

Richard H. S. TUR, 2001: Defeasibilism. Oxford Journal of Legal Studies 21(2). 355–368.

Vrh stranice


1 Ferrer Beltrán & Ratti 2012; Guastini 2012; Hooker & Little 2000; Lance, Potrč & Strahovnik 2009.

2 Chiassoni 2012: 162; Lance and Little 2007.

3 In Plato’s dialogue Statesman, we can follow a debate between Socrates and a young stranger from Elea about what defines a good statesman, one that would regulate public affairs justly. The dialogue also moves to the question of whether it is possible to rule and govern without laws. The stranger, in trying to defend the affirmative answer to this question, proposes the idea that it is better that a “royal man” governs instead of laws, since “[l]aw can never issue an injunction binding on all which really embodies what is best for each: it cannot prescribe with perfect accuracy what is good and right for each member of the community at any one time. The differences of human personality, the variety of men’s activities and the inevitable unsettlement attending all human experience make it impossible for any art whatsoever to issue unqualified rules holding good on all questions at all times”. He continues by arguing that the one who governs will probably be unable to avoid any general law being put forward, and so one “will lay down laws in general form for the majority, roughly meeting the cases of individuals . . . under average circumstances”. Nonetheless, both Socrates and the stranger agree that, if exceptions to these general norms were to emerge, it would be unwise, unjust, or even ridiculous not to correct such cases (Plato Statesman: 294a–b, quoted in Schauer 2012: 78). A similar proposal can be found in Aristotle’s Nicomachean Ethics. “The reason [i.e., that justice and equity are not quite the same thing, and that equity can be seen as a correction of legal justice; n. VS] is that all law is universal but about some things it is not possible to make a universal statement which shall be correct. In those cases, then, in which it is necessary to speak universally, but not possible to do so correctly, the law takes the usual case, though it is not ignorant of the possibility of error. And it is none the less correct; for the error is not in the law nor in the legislator but in the nature of the thing, since the matter of the practical affairs is of this kind from the start. When the law speaks universally, then, and a case arises on it which is not covered by the universal statement, then it is right, where the legislator fails us and has erred by oversimplicity, to correct the omission – to say what the legislator himself would have said had he been present and would have put his law if he had known” (Aristotle NE: 1137a-b).

4 Dworkin (1977: 24–25), amongst others, maintains such an optimistic view: “Of course a rule may have exceptions. ... However, an accurate statement of the rule would take [these exceptions] into account, and any that [it] did not would be incomplete. If the list of exceptions is very large, it would be too clumsy to repeat them each time the rule is cited; there is, however, no reason in theory why they should not all be added on, and the more there are, the more accurate is the statement of the rule”.

5 I do not want to deny that there are some who would claim that what defeasibility boils down to is precisely such a phenomenon. My debate is framed in the discursive space of views that do allow for and in fact support a more robust notion of defeasibility and genuine exceptions. Cf. Chapman 1998: 448; Celano 2012.

6 Brożek 2014.

7 Celano (2012: 281) labells such exceptions as “true exceptions” (as opposed to merely prima facie exceptions) and elaborates this concept in the following way: “For a case to qualify as a true exception it must not be already provided for in a reasonably detailed and precise ‘unless ...’ clause attached to the norm”.

8 Cf. Strahovnik 2004.

9 Celano 2012: 285–287.

10 Lance and Little 2007; 2008.

11 Celano 2012: 284.

12 Celano 2012: 286.

13 For readers unfamiliar with the debate between generalism and particularism, this is a brief recapitulation of the main views and concepts. Particularism regarding a given normative domain (e.g., morality) is a view characterised by a negative attitude towards principles, norms and rules. Moral particularism can thus be associated with a simple thesis that there are no moral principles or with a more elaborate claim that ‘the possibility of moral thought and judgment does not depend on the provision of a suitable supply of moral principles’ (Dancy 2004: 73). Moral thought need not consist in the application of moral principles to cases, and the morally perfect person is not to be regarded as a person of principle. The opposite view is generalism claiming that the moral status of an action is determined by its falling under a general moral principle or rule. Parallel to this debate, atomism and holism (also labelled as contextualism) are views about reasons. Atomism claims that reasons are context insensitive, meaning that they always function (count for or against) in the same way. Holism, on the other hand, claims that reasons are context sensitive, that they can vary in their relevance and strength, and that they can additionally be combined with each other in ways that go beyond a mere addition model (e.g., two reasons which would separately count for a given action, can – if present together – represent a reason against it. Particularism is usually combined with holism and generalism with atomism.

14 Celano 2012: 270.

15 Celano 2012: 276.

16 The “moral flatness” worry was raised in moral theory by several authors (McNaughton and Rawling 2000: 273; Crisp 2000: 36; Bakhurst 2000: 167; McKeever and Ridge 2006: 4; Celano 2012: 283) and can be summarised in the following way. Given the holism of reasons, a set of morally relevant features of an action is open, which means that any feature could be morally relevant and can stand as a reason for or against an action. Furthermore – given the particularistic thesis – this set of features cannot be ordered by general principles. But why does morality seem to be nevertheless ordered? Why do we think that morally central features very often have to do with, e.g., causing pain and suffering, sincerity, honesty, keeping promises, benevolence, dignity, etc.? All that moral particularism can say is that some features are often more relevant than others and that is all. It cannot capture the idea of them being “central” in a way to morality. The moral landscape painted by particularism is flat. In order to avoid this charge, particularism must offer us “some way to distinguish those considerations which normally and regularly do provide reasons of a certain valence (e.g. pain) from those that normally and regularly do not provide reasons (e.g. shoelace colour). For absent some such distinction, particularism threatens to flatten the moral landscape by suggesting that insofar as they might provide reasons all considerations are on par” (McKeever and Ridge 2006: 45). If we try to isolate the fundamental worry, we can recapitulate it in the following manner. The moral non-flatness requirement says that any moral theory must somehow account for the fact that some considerations or features of acts are more central to morality than others. Amongst particularists, Dancy as its prominent defender put forward a proposal that introduces the notion of a default reason to deal with the flatness worry (cf. Dancy 2000: 137; Dancy 2004: 112–113).

17 Celano 2012: 285.

18 They defend the model in a series of papers, Lance & Little 2005; 2006; 2007; 2008.

19 The same holds in other areas as well. For example, the non-moral generalisation “Fish eggs develop into fish” is a defeasible generalisation. It is not that most fish eggs develop into fish (quite the opposite is true since most of them end up as food for other animals). Something else is captured by this particular generality. One should read it as follows: “Defeasibly, fish eggs develop into fish” or “In privileged conditions, fish eggs develop into fish”, where privileged conditions are defined as conditions that are particularly revealing of the nature of the thing in question or of a broader part of reality in which the thing is known (Lance and Little 2008: 62).

20 There are cases in which a defeasible generalisation tracks paradigm cases, which are, in this sense, privileged. This is the case with, e.g., “Defeasibly, chairs are things we sit on”, and there are riffs of this paradigm as in the case of ornamental chairs. The moral case would be the case of pain as defeasibly wrongdoing, but it would not be so in the case of athletic achievement. In other cases, there is a justificatory mutual dependence of privileged and non-privileged cases in the sense that we must appeal to privileged cases in order to explain and understand what is going on in a non-privileged case (e.g., lying and lying as part of the Diplomacy game; pleasure and sadistic pleasure). And lastly, there could be an idealisation-approximation relation, as in the case of the ideal gas law pV = nRT and the actual behaviour of actual gases. The example of similarly defeasible moral is the norm that in the Kingdom of Ends (full information, genuine autonomy, basic trust) people are owed the truth. (Lance and Little 2008: 64–73).

21 Lance and Little 2008: 64–68.

22 McKeever and Ridge (2006: 60–72) put forward quite a forceful argument in this direction. First, one can label reasons capturable in defeasible generalisations as paradigmatic reasons and others as non-paradigmatic. The argument assumes that the best way to capture the paradigmatic vs. non-paradigmatic distinction is via an explanatory asymmetry account of this distinction employed by Little and Lance. Cases that fall outside privileged conditions are explanatorily dependent on how a particular consideration functions in a normal case when the conditions are privileged. By contrast, there is no such dependence going the other way. E.g., pain is normally bad or a reason against an act, but this is not the case when, e.g., pain is constitutive of athletic challenge and accomplishment where it shifts its polarity (Lance and Little 2006: 319). One must understand that normally pain is something bad in order to understand how it functions in the case of athletics, and not the other way around. It seems that the proposed model is useful only for the reason that we have one valence in privileged conditions and the opposite valence in non-privileged conditions. But the proposed model is not successful with reasons that sometimes lack moral relevance. “For it will be true of any consideration whose status as a reason can sometimes be defeated that we can adequately understand why it is not a reason here only if we understand how it can be a reason elsewhere. For to understand why something is not F here we must in general have some idea of how it can be F elsewhere if it can. If this is enough for a consideration to qualify as an instance of asymmetric reasons then any consideration whose status as a reason can ever be defeated will qualify as an instance of asymmetric reasons, and that makes the distinction far less interesting than it first appeared.” (McKeever and Ridge 2007: 67). It further seems that some core reasons might be non-paradigmatic, and that not all peripheral reasons need to be non-paradigmatic.

23 Little 2001: 34, emphasis mine.

24 Lance and Little 2007: 153, emphasis mine.

25 In this discussion, I have mostly focused on moral norms, but the same applies to defeasible legal norms. One possible way out of this conundrum is proposed by Guastini (2012). For Guastini, both defeasibility and axiological gaps are related to axiological judgments made by interpreters of norms. Furthermore, defeasibility is not a special feature of legal principles or norms; defeasibility is not an objective property of norms which is already there before we start to interpret them. Our axiological judgments employed in interpretation are thus neither the consequence of some objective defeasibility of the rule itself nor a genuine, interpretation-independent normative gap, but the origin or cause of interpretative defeasibility. This, in the end, includes some basic, indefeasible reasons and norms, such as justice. Cf. Strahovnik 2012; 2013.

26 Tur 2001.

27 Ross 1930; 1939.

28 Ross 1930: 27–28.

29 Ross 1939: 190.

Vrh stranice

Preporučeni oblik citiranja:

Časopisni bibliografski podatci

Vojko Strahovnik, « Defeasibility, norms and exceptions: normalcy model », Revus, 29 | 2016, 61-76.

Elektronički bibliografski podatci

Vojko Strahovnik, « Defeasibility, norms and exceptions: normalcy model », Revus [Mrežno izdanje], 29 | 2016, Datum mrežne objave: 10 octobre 2016, pristupljeno 23 mars 2017. URL : ; DOI : 10.4000/revus.3571

Vrh stranice


Vojko Strahovnik

Assistant Professor and Research Associate, Faculty of Theology, University of Ljubljana and European Faculty of Law, Nova Gorica (Slovenia).

Address: Ulica Hermana Potočnika 21, SI-1000 Ljubljana, Slovenia.


Od istog autora

Vrh stranice

Autorska prava

All rights reserved

Vrh stranice