« Should I Be Offended by Ann Coulter? | Main | Why Evolutionary Biology is Relevant to the Law: A Challenge to the Gruter Institute & to Brian Leiter/Michael Weisberg »

Friday, October 12, 2007

Why no deontic logic?

I’m devoting my blogtime this month to talking about some of the peculiarities and prejudices of philosophy of law in the United States. My first topic was why American philosophers of law don’t talk about Hans Kelsen (see my three posts  and Brian Leiter’s and Larry Solum’s responses).

My second – and, as we shall see, related – topic is the following: Why don’t American philosophers of law talk about deontic logic?

Deontic logic is the branch of logic that deals with normative terms such as obligation, permission, and prohibition. The original motivation for the field was the recognition that certain deductive inferences involving sentences with these normative terms are possible. For example, the sentence "Paying one’s taxes is obligatory" is true only if the sentence "Not paying one's taxes is forbidden" is also true.

There are American philosophers, logicians, and computer scientists who do work in deontic logic. Roderick Chisholm (a philosopher at Brown) did important work in the field. It is striking, however, that those who work in the philosophy of law in this country have little to say on the topic. Interest is somewhat greater among philosophers of law in the UK. But it really increases as one moves to non-Anglophone countries.

Compare, for example, Legal Theory (arguably the premier journal for Anglo-American philosophers of law) and Ratio Juris (an English-language journal in the philosophy of law that primarily publishes articles by non-Anglophone authors). Legal Theory has published nothing on deontic logic (at least since 1999, which is as far back as I can go on-line). In contrast, Ratio Juris has articles on deontic logic by Italians (Volpe, Guastini, Mazzarese, Martino), Spaniards (Alarcon-Cabrera, Marin), Argentinians (Alchourron, Vernengo), Finns (von Wright, Niiniluoto), Swedes (Spaak, Lindahl, Odelstad), Poles (Wolenski, Opalek), Germans (Lehman), Austrians (Weinberger), and Britons (Coyle).

This is all the more striking because many American philosophers of law have gone through graduate programs in philosophy. A course in first-order predicate logic is usually required in such programs, and formal logic, although perhaps no longer as central to philosophy as it once one, is still very much a part of the field.

How are we to explain this difference? Consider how Brian Leiter explained the relative lack of interest in Kelsen among Anglophone philosophers of law. The reason, he argued, was that Kelsen had been shown to be mistaken: “[H.L.A.] Hart persuasively undermined the two most distinctive features of Kelsen's jurisprudence.”

And what explained the continued interest in Kelsen among non-Anglophone philosophers of law? Why didn't they recognize that Hart had undermined Kelsen? Some of the possibilities Leiter offered are: 1) that they are not “well-informed about Anglophone jurisprudence,” 2) their low “level of philosophical sophistication,” and 3) “issues of national pride and loyalty (sometimes an issue, one senses, in some German-language jurisprudential work when it comes to Kelsen).”

Now deontic logic has serious challenges, some of which may stand in the way of its being of significant use in the philosophy of law. But it can hardly be said that non-Anglophone philosophers of law are interested in deontic logic because they lack philosophical sophistication. If anyone is lacking philosophical sophistication, it’s the Americans.

As I indicated in my earlier posts, I don't think interest in Kelsen can be attributed to lack of philosophical sophistication either. Indeed, it is a significant fact that, among philosophers of law, interest in Kelsen and interest in deontic logic are closely correlated. They are at their lowest in the US, increase somewhat in the UK, and are at their height in the non-Anglophone world, particularly in Italy and Spanish-speaking countries. Ota Weinberger (Austria), Eugenio Bulygin (Argentina), and Norberto Bobbio (Italy) are three prominent examples of philosophers of law with a strong interest both in Kelsen and deontic logic.

I have already written on the connections I see between Kelsen's legal theory and attempts to uncover the formal qualities of a legal system (attempts that include, but are not limited to, deontic logic). I won't repeat these arguments here. But American philosophers of law should be open to the idea that it is we who are missing out on something, both when we ignore deontic logic and when we ignore Hans Kelsen.

Posted by Michael S. Green on October 12, 2007 at 04:50 PM in Legal Theory | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Why no deontic logic?:


I'm pretty sure Legal Theory has never published an article in deontic logic, so your sense of the situation is correct.

But let me pose a related question: why are Anglophone moral philosophers almost entirely uninterested in deontic logic? Deontic logic exists in its own "ghetto," as it were, in the Anglophone philosophical community, having little or no influence on the main currents of moral philosophy.

The answer to this question may suggest an answer to yours. (I'm not sure of the answer, this is just an hypothesis that there is a connection.)

Posted by: Brian | Oct 12, 2007 7:45:30 PM

As a sidenote, I do wish you'd quote fully what I said about why Kelsen looms larger in the non-Anglophone world:

"The non-Anglophone world is not a monolith, and the extent to which the Hartian critique of Kelsenian views has made an impact varies quite a bit, often a function of factors like (1) the extent to which the jurisprudential communities in question are well-informed about Anglophone jurisprudence (as, for example, the Italian and Spanish-speaking communities tend to be), (2) the level of philosophical sophistication in these jurisprudential communities, and (3) issues of national pride and loyalty (sometimes an issue, one senses, in some German-language jurisprudential work when it comes to Kelsen). (2) may be particularly important, since Hart's work emerged out of developments in post-WWII Anglophone philosophy that are still either resisted or misunderstood in other parts of the world."

As someone who also shares an interest in German philosophy, you know full well how true (2) is, and not just with respect to jurisprudence.

Posted by: Brian | Oct 12, 2007 8:01:40 PM

Brian - I agree that the lack of interest in deontic logic extends to Anglophone moral philosophers. Whatever explains why they don't write on deontic logic might help explain why the Anglophone philosophers of law don't either.

On the other hand, it's my impression that in non-Anglophone countries, the moral philosophers don't write on deontic logic, but the philosophers of law do. The reason, I think, is Kelsen's influence on the latter.

As for you second post, I agree that much non-Anglophone philosophy is unsophisticated (although that could be said of much philosophy done in this country too). But I don't think interest in Kelsen is an example of this lack of sophistication.

Posted by: Michael Steven Green | Oct 15, 2007 11:15:04 AM

Isn't the main problem with Kelsen that his work is conceptual at such a high level of generality that it is worthless when deciding any practical questions about how legal systems do or should work?

Posted by: Ignorant American | Oct 15, 2007 11:03:19 PM

Your point about lack of practical applications might explain why non-philosophers are not interested in Kelsen. My question is why philosophers of law are not interested in him. After all, the philosophers talk about Hart, but it is not clear that Hart's approach has any practical applications either.

Posted by: Michael Steven Green | Oct 16, 2007 10:59:50 AM

I would be curious if you could identify one jurisprudential issue that deontic logic helps illuminate or resolve (aside, of course, from the proper representation and logic of deontic relations). If there aren't any, then perhaps the neglect of deontic logic would not be an instance of prejudice or parochialism.

Posted by: Scott Shapiro | Oct 16, 2007 1:53:40 PM

Some people insist that legal conflicts are possible, for example, that an act and its negation can both be obligatory in the same legal system. Under standard systems of deontic logic, the consequences of allowing such conflicts are dire, since it would follow that anything is obligatory. Under standard deontic logics, (p → q) → (Op → Oq), so we can derive Oq, no matter what q is, from O(p&-p).

Now I’m not sure whether you wouldn’t just describe this as a problem of the proper representation of deontic relations. Deontic explosion, you might argue, suggests that we should not represent deontic relations through a system under which obligations are closed under entailment. But it certainly looks as if deontic logic is illuminating something.

Posted by: Michael Steven Green | Oct 16, 2007 2:46:05 PM

Thanks. Let me rephrase my question. In order for research in deontic logic to be a worthwhile endeavor for anglophone legal philosophers, it must bear on issues that they really care about. Some might care about the possibility of legal conflicts, but it is hardly one of the burning issues of the day. And if one is interested in legal conflicts, one can study them without getting into the weeds of deontic logic.

Ultimately, it's a question of costs and benefits. Do the benefits of becoming proficient in deontic logic outweigh the investment costs? I am skeptical that they do but am open to persuasion.

Posted by: Scott Shapiro | Oct 16, 2007 3:40:07 PM

Although I think the weeds of deontic logic do reveal something interesting about legal conflicts (since one might not have recognized the problem of explosion otherwise), I have to admit that, given deontic logic’s current form and the current interests of Anglophone philosophers of law, the former does not have a huge amount to do with the latter.

One problem is that deontic logicians haven’t come up with a way of modeling legal change – eg constitutional amendment, legislation, and adjudication. It is the dynamic quality of legal systems that philosophers of law are really interested in – not static time-slices of legal systems. But perhaps these problems would be tackled if more philosophers of law got involved. (Part of the solution, I think, is to move away from an analysis of sentences about norms, which is what deontic logic does, to a direct analysis of the norms themselves.)

Another problem is that Anglophone philosophers of law, in my opinion, have too narrow a set of interests. Deontic logic might look more relevant if they were more interested in examining the structure of legal systems. Why isn’t there more work done in the philosophy of law that is analogous to the work done in the philosophy of language? There are countless modest and discrete problems (say, truth-functional accounts of adverbs) that people work on. Not everyone writes about the general question “what is language?” or whether linguistic meaning is based on social facts alone.

Some of these modest and discrete problems in the philosophy of law are immediately apparent. One is the paradox of self-amendment (that is, whether amendment clauses can be used to amend themselves). But I think that if we tried to model legal systems, many more puzzles would reveal themselves. But this is just a guess, of course.

Posted by: Michael Steven Green | Oct 16, 2007 5:12:17 PM

I think what Scott's question, and your answer, brings out is that you haven't really made any case at all establishing the "peculiarities" or "prejudices" of Anglophone philosophy of law; rather, you are identifying topics that Anglophone philosophy of law ignores, perhaps rightly (perhaps not). It would be more interesting if you would make the case for the relevance of deontic logic (or the relevance of Kelsen, as you tried to do in one of your later posts). That some philosophers of law somewhere discuss deontic logic is neither here nor there. Some philosophers somewhere discuss almost any topic one could imagine. More interesting would be the arguments for the relevance of a particular topic or approach.

Posted by: Brian | Oct 16, 2007 9:39:37 PM

I guess I would disagree with Michael's claim that deontic logicians haven't explored the dynamics of legal change. Carlos Alchourron and David Makinson worked on this topic during the 1980's (first in “Hierarchies of regulations and their logic” and then later in "The logic of theory change: Contraction functions and their associated revision functions" and "On the logic of theory change: Safe contraction") and their work formed the inspiration for much of the work in formal epistemology done today (legal codes are often represented as closed sets of propositions and hence have the same formal representation as belief sets: the thought, then, is that the revision of belief sets should obey the same principles as the revision of legal codes).

Posted by: Scott Shapiro | Oct 17, 2007 2:11:50 PM

Brian –

It would indeed be more interesting if I could make a case for deontic logic (or Kelsen) in some brief blog posts. But it is unreasonable to expect me to do so. After all, you don’t make a case for naturalism in legal philosophy on your blog. That’s what books and articles (one's own and others') are for.

But even if I don’t make a case for deontic logic, I can’t see how it isn’t at least *peculiar* that a good number of non-Anglophone philosophers of law write about deontic logic, but Anglophones ignore it almost completely in favor of writing yet more articles on the inclusive/exclusive legal positivism debate, or the proper interpretation of Hart, or the “methodology debate” about conceptual analysis.

The only reason this wouldn’t be peculiar is if it were well known among the Anglophones that there is something confused or worthless about deontic logic. But that’s not so. Indeed, there isn’t enough familiarity with deontic logic among Anglophone philosophers of law for them to make an informed judgment about its merits. And from all appearances, the non-Anglophones are addressing real philosophical problems that are tied to questions about the law. That virtually no Anglophones (and especially Americans) do the same is peculiar and, since it is not the result of informed judgment, a prejudice.

Posted by: Michael Steven Green | Oct 17, 2007 4:43:27 PM

Scott –

My understanding was that Alchourrón and Bulygin (both Kelsen scholars, I might add) attempt to account for change in the sense of how a set of propositions or norms would be altered in the light of the introduction of a new proposition or norm. What I was thinking of was accounting for how that new proposition or norm is introduced in the legal system in the first place. Why is it that a legislature’s pronouncements change the norms of a legal system but mine do not? This would mean modeling relationships of authorization (and limits on the scope of authorization). I may be wrong – but I thought that this had not been done.

Posted by: Michael Steven Green | Oct 17, 2007 4:57:17 PM

Michael, I didn't post on a blog that there is something "peculiar" or "parochial" about American philosophers of law because they're not naturalists! You, by contrast, have made that claim about American philosophers of law with respect to Kelsen and deontic logic, but you can't really support it. You can get away with saying such things on blogs precisely because they tend to be superficial and most of those reading are not experts. But by taking advantage of those facts you do a disservice both to those who work in philosophy of law (who you paint, without any justification, as indefensibly "parochial" and "peculiar") and to those who might be curious about the field, and who will now be misled.

Instead of adopting the posture that you are exposing "parochialness" and "peculiarities" of Anglophone philosophy of law--it's obvious by now that you aren't--why not just stick to the merits? You are, for example, unusual among Anglophone philosophers (not just philosophers of law) in thinking deontic logic is of interest at all. Why you hold this unusual view would be interesting to read about.

Posted by: Brian | Oct 17, 2007 9:19:26 PM

In response to the question of whether the benefits of becoming familiar with deontic logic outweigh the costs: sure they do. First, becoming familiar with deontic logic, to the extent of understanding most modern research, is not a huge investment, especially if you already understand the rudiments of predicate logic. Perhaps 20 hours/month for 5 or 6 months. Second, the value of deontic logic for legal thinkers approximates the value of the axioms of probability for psychologists: though it's not how we actually think, it's useful to have an ideal to compare our systems to. Importantly, even our moral intuitions can be wrong, and deontic logic points out where and why (given our assumptions). Third, it adds an interesting perspective where issues of moral fault--instead of causal fault--are grounds for guilt or liability. Finally, familiarity justifies itself if it allows for cross-Atlantic communication and scholarship.

Posted by: Tim | Oct 17, 2007 9:42:51 PM

Thanks Tim. You're right that formalization often highlights hidden confusions or ambiguities. It's something to be encouraged, which is why I find Brian's resistance here so strange. In addition, the philosophical problems that have arisen so far in deontic logic - such as normative contradictions and contrary to duty imperatives - are genuinely interesting in their own right. I find it hard to see how I am somehow doing the profession a disservice simply by noting that deontic logic is currently being ignored by philosophers of law in this country.

Posted by: Michael Steven Green | Oct 17, 2007 10:05:17 PM

A few more points:

Brian is wrong that I am unusual among Anglophone philosophers in thinking that deontic logic is of interest at all. Roderick Chisholm was no slouch and he thought it was interesting. Two current American philosophers who do work in the field are Paul McNamara http://www.unh.edu/philosophy/Faculty_Pages_info/mcnamara_page.htm and John Horty http://www.umiacs.umd.edu/~horty/biocv.html. In the UK there is Sean Coyle. http://www.ucl.ac.uk/laws/academics/profiles/index.shtml?coyle

Brian writes as if I have failed to identify anything interesting about deontic logic. I identified the problem of normative conflicts. If normative conflicts are impossible, this is a matter of profound importance for the law, since such conflicts appear to arise in the law. There is less law out there than we thought. If normative conflicts are possible, this also has profound implications for our understanding of what the law and legal obligations are. Deontic logic is a useful tool in coming to a decision on this matter, and in understanding the consequences of our decision.

Another interesting topic in deontic logic is contrary-to-duty obligations. Even though John has a duty not to have a fence, if he breaks that duty and sets up a fence, he might have a duty that his fence be painted white. There are difficulties representing such situations without contradiction.

This is in itself an interesting problem, but it also highlights an important quality of legal (and moral) obligations. In chess, breach of a duty takes one outside the normative space – one is no longer playing chess. The game is over. That’s not so in ethics or law. Indeed, as Kelsen was fond of noting, the person committing a wrong is in some respects like a legislator (or a person entering into a contract) in the sense that he exercises a power to create new obligations. Once again, deontic logic is a useful tool in exploring these issues. Indeed we largely have deontic logic to thank for the recognition that there is a problem here at all.

What I admitted to Scott Shapiro that I could not do was show that deontic logic in its current form has all that much to do with the *current* interests of US philosophers of law. But that was my point. There is no reason that there shouldn’t be at least a few US philosophers of law interested in normative conflicts or contrary to duty obligations.

I'm not saying that one can't be a good philosopher of law without knowing deontic logic. What I'm saying is that it is peculiar (and arguably the consequence of some hidden prejudice) that there is virtually no philosopher of law in this country who is interested in deontic logic, when there are so many abroad.

Finally, I have no doubt that Brian would have blogged that American philosophers of law were *peculiar* (or worse) for ignoring naturalism in legal philosophy - if naturalism were commonly discussed by philosophers of law on other continents. He would have so blogged, even if he could show the true merits of naturalism in legal philosophy only in books and articles.

Posted by: Michael Steven Green | Oct 18, 2007 3:52:08 AM

Michael's post reminds me: becoming familiar with deontic logic necessitates becoming familiar with standard modal logic, and standard modal logic (in its various incarnations) is a popular tool used to understand counterfactuals. As counterfactuals are integral to at least one dominant theory of causation, becoming familiar with deontic logic also buys you a more analytical perspective on causation for free.

Posted by: Tim | Oct 18, 2007 9:14:01 AM

Forgive me for taking up a matter that was raised early in the thread and then dropped. I only now stumbled onto this thread. I'm just wondering if either Brian or Michael would care to elaborate on their seeming agreement (although Michael adds an inmportant caveat) that German philosophy of law is "unsophisticated." Do Habermas and Luhmann qualify as German philosophers of law? Are they unsophisticated? I think this relates to the discussion of whether American philosophers of law are "parochial" or "peculiar".

Posted by: Jeremy Telman | Oct 18, 2007 9:14:34 AM

Michael, you have not been "simply...noting that deontic logic is currently being ignored by philosophers of law in this country."
You began (read what you wrote!) by saying that you're going to call attention to "peculiarities and prejudices of philosophy of law in the United States." It's that latter case--the one you set as your task--that you have failed to make. Making up what I might have written on a blog (but did not in fact write) is really a very weird response!

Even if one is interested in normative conflicts, one can study them, as Shapiro noted, without deontic logic.

I am aware, of course, that there are some Anglophone philosophers (alive and dead) interested in deontic logic; several of my colleagues here in Austin actually are. That doesn't change the fact that most Anglophone philosophy departments, and almost all moral philosophers, ignore it, that it is a very specialized niche, and so you are "unusual" in thinking it so important...even more so, in thinking it so important without a compelling explanation of why those of us not studying deontic logic are misguided.

Posted by: Brian | Oct 18, 2007 9:26:10 AM

Hi Michael:

You clearly think that anglophone legal philosophers are criticizable for ignoring deontic logic -- I don't think they are. Let me try to set out some considerations in their defense.

Most everyone would agree that deontic logic is both intrinsically and instrumentally valuable. It is intrinsically valuable because developing a formal tool for representing normative relations is an activity that is good in itself. It is instrumentally valuable in that it can shed light on some jurisprudential debates.

I don't think it would be fair to criticize anglophone legal philosophers for ignoring deontic logic based solely on deontic logic's intrinsic value. After all, there are many things in life that are intrinsically valuable that we don't engage with -- life is too short. True, if the issues that Anglophone legal philosophers dealt with were so much less valuable than the problem of, say, representing counter-to-duty imperatives, then you would be right. But I am not persuaded of this at all.

You might think, then, that deontic logic is instrumentally valuable. But for anglophone philosophers to be *criticizable* for not exploring the relevance of deontic logic, you would have to show that delving into the deontic logic literature and acquiring expertise in it is worth the costs (as compared to, say, acquiring expertise in decision theory, meta-ethics, the philosophy of mind or Kantian moral theory).

Your only example where deontic logic would be instrumentally valuable is the question of normative conflict. As you no doubt know, anglophone legal philosophers have debated this question (for example, re: Dworkin's distinction between rules and principles) but they have not thought that deontic logic is helpful here and I think they were not irrational for so thinking.

Regardless of why anglophone legal philosophers actually ignore traditional deontic logic, let me set out one reason why they ought to. One might genuinely doubt that traditional deontic logic has the power to represent adequately normative relations in law. If one is a Razian, for example, one thinks that the law purports to be an exclusionary reason. How is one to represent this in deontic logic? No doubt someone could develop such a logic, but to the best of my knowledge none exists. So if one is a Razian, then one will shy away from using deontic logic as a way of settling questions about normative conflicts. If one is a Dworkinian, one will see the law as imposing prima facie obligations to obey. But traditional deontic logic has problems representing prima facie obligations. Sure, there are non-monotonic logics that purport to do a good job, but this is new technology and hence much more technical and controversial than the simpler, traditional models. Philosophers should be excused from not appealing to these theories until more is known and some consensus forms among philosophical logicians.

Posted by: Scott Shapiro | Oct 18, 2007 10:16:07 AM

Thanks Scott. A couple points.

Let’s start with the intrinsic value of deontic logic (which I think is non-negligible). Of course, no individual US philosopher of law can be criticized for failing to talk about deontic logic based solely on its intrinsic value. Philosophers have a right to ignore even very interesting topics.

But it is a different matter when one looks to US philosophers of law in the aggregate. There is something odd about the fact that basically no one here talks about deontic logic, when there are many philosophers of law elsewhere who do.

Now we might just chalk this up to a geographic division of philosophical labor. (The division probably has its source in the influence of Kelsen on the non-Anglophones – and one of my primary reasons for talking about deontic logic in the first place was to show this influence of Kelsen abroad.)

But I think that there is something that can be criticized here (although no individual philosopher can be the target of the criticism). I am struck by how philosophers of law in this country mostly write about the same small circle of topics - e.g. the inclusive/exclusive legal positivism debate or the methodology debate about conceptual analysis. (I would set aside normative philosophy of law, which I think is much heathier.) It would, in my opinion, be a good idea for US philosophers of law to broaden the scope of their interests. And one way to do this is to look at what people abroad are doing.

Second, as far at the instrumental benefits of deontic logic are concerned, let me just engage in a little autobiography. I was exposed to deontic logic through Kelsen. I started thinking about the problem of legal conflicts and it occurred to me that the problem (or at least a good deal of it) might be solved through applying the deontic operators to sentences in predicate rather than only sentential logic.

When one says that p and not–p are obligatory (or O(p & -p)), one might in fact be saying something like the following. It is obligatory that if there exists a person x to whom legal duties apply, then x Ps and not-Ps. So understood, there is no contradiction, since one can be released from the contradictory obligations by not existing (or otherwise being released from the scope of legal duties). The solution is similar to Russell’s solution to the problem of the present king of France being neither bald nor not bald in On Denoting. Although this would, I think, solve the contradiction, it would also explain our strong aversion to contradictory laws, since they amount to a prohibition on the existence of legal actors.

Now I might be completely wrong about this solution. (I am reluctant to blog about it given that it is still in the thinking stages.) But I would never have come up with it without exposure to deontic logic (which in my case is very limited). So I think deontic logic has enough instrumental value for it to be a good idea that at least some philosophers of law in this country be familiar with it. But even here I don’t think any particular philosopher can be criticized for knowing nothing about it.

Posted by: Michael Steven Green | Oct 18, 2007 11:22:39 AM

And concerning your argument that philosophers of law ought not consider deontic logic because it cannot, as yet, represent exclusionary or prima facie reasons for action, I have two responses.

First of all, the problem you mention concerns representing moral, not legal norms. Or, more generally, it is a problem with representing how all practical reasons come together to recommend a course of action to an agent. It is not about representing what legal obligations an agent has.

This can be clearly seen in Raz’s case. Even if it is true, as Raz argues, that the law claims to create exclusionary reasons for action, it may not actually create such reasons. If it doesn’t, there are no exclusionary reasons to represent. But there still are legal norms that come together to recommend a legal course of action, and more standard deontic logics might be very useful in representing those legal norms.

I suppose you might be saying that deontic logic can’t represent the law’s *claim* to create exclusionary reasons for action – which Raz thinks exists whenever there is law – but a claim to create a norm is not a norm and so it is unreasonable to ask it to be represented by deontic logic.

The problem is more complicated for Dworkin, however, since he (probably) thinks that legal norms must create prima facie moral reasons for action. Their existence as legal norms is their existence as such moral reasons. For Dworkin, therefore, deontic logic’s capacity to represent legal norms would depend upon its capacity to represent prima facie reasons.

One response to this problem is to say – so much the worse for Dworkin. It was odd in the first place that the determination of a legally recommended course of action should depend upon how prima facie moral reasons come together.

But more generally (and this is my second response), why should philosophers of law - or moral philosophers - completely abdicate to the logicians the job of representing formally these exclusionary or prima facie reasons for action? I’m not saying that we all must be in this game – only that it is healthy when some of us are.

Indeed, one has to have some idea of the moral notions of an exclusionary or prima facie reason for action to represent them (the problems are particularly delicate with respect to exclusionary reasons). So one has to be a moral or legal philosopher to properly begin the job. Furthermore, the attempt to engage in the representation can have positive consequences for one’s job as a moral or legal philosopher. Philosophical problems (or solutions to such problems) can reveal themselves that would not have been evident before.

Posted by: Michael Steven Green | Oct 19, 2007 2:49:38 AM

Sorry, but I don't understand your response about Raz. Are you saying that standard deontic logics *can* represent propositions that state the existence of exclusionary reasons? This would have to be the case in order for deontic logics to have the power to represent legal propositions on the Razian account.

Posted by: Scott Shapiro | Oct 19, 2007 7:57:33 AM

My apologies. My argument was too compressed. I am assuming that standard deontic logics *can't* represent exclusionary reasons, but that this would not be a problem for representing legal norms, even for Raz.

Let's take a simple legal norm: A command to pay taxes. I take it that for Raz the legal content of that norm is that one should pay taxes. The legal content is *not* that one should take the command as an exclusionary reason, which then would be factored with other exclusionary reasons (and the primary reasons that were not excluded by these exclusionary reasons) to produce an all-things-considered best course of action. One would represent the legal norm as an exclusionary reason only if one wanted to represent all the reasons for action pertaining to that actor. And one would do that only if the legal norm *were indeed* an exclusionary reason (not if it merely claimed to be one).

So even though Raz thinks that the law always claims to create exclusionary reasons for action, that is not relevant to the job of representing legal norms. An adequate representation of the norms of a legal system would simply show how the various legal norms come together to produce a legally recommended course of action. An adequate representation of the norms of the game of chess would simply show how the various norms of chess come together to produce a recommended move in the game. In both cases it would not represent how law norms or chess norms might actually factor into a different normative system, namely ethics or, more generally, the system including all one's reasons for action. That would be relevant only if one wanted to represent that other normative system.

For Dworkin the situation is different of course, because he takes legal norms to be moral norms (or I am assuming that he does). So the prima facie quality of legal norms in the moral reasoning of an actor would have to be accounted for even in a representation of the legal norm itself.

Posted by: Michael Steven Green | Oct 19, 2007 9:23:49 AM

On Raz's view, legal norms purport to be exclusionary reasons (much as the name "Plato" purports to refer to Plato and the proposition that the cat is on the mat purports to represent truly the state of affairs where the cat is on the mat). If so, then deontic logic, as currently conceived, cannot represent legal norms (or statements about them) because they cannot represent valid inferential relationships between norms that purport to be exclusionary reasons.

Posted by: Scott Shapiro | Oct 19, 2007 9:51:38 AM

Scott – I don’t disagree with any statement that you made above, but what you say does not tell against the (limited) point I made about Raz in the previous two posts. I’ll try to make the point one more time.

First, some stage setting. Deontic logic seeks to represent normative systems. It does so indirectly, by representing statements about the norms of those systems. (That is part of the problem with deontic logic as currently conceived, but I won’t get into that here.) There are many normative systems that one might want to represent: law, ethics, chess. One might even want to represent hypothetical systems (the law of Gondor, Satan’s morals, a variation on the game of chess).

Now my point about Raz was merely the following: Let us assume, with Raz, that legal norms purport to be exclusionary reasons and let us assume your claim that exclusionary reasons cannot be represented by standard deontic logic. The question remains: with respect to *which* normative system is deontic logic having a problem?

My point is that the problem does not concern the representation of the legal system. Where the problem occurs is in representing the normative system of morality (or, more generally, the normative system that includes all one’s reasons for action). And that’s true only if the law *actually* creates an exclusionary reason for action.

Now in situations where the law does not create exclusionary reasons, but only purports to do so, I guess one could say there is still a problem with representing a normative system. The problem is representing the normative system that is claimed to exist when the law purports to create exclusionary reasons. There is a problem in representing this *hypothetical* normative system. [So you might rightly disagree with my statement earlier that "a claim to create a norm is not a norm and so it is unreasonable to ask it to be represented by deontic logic." The claim asserts the existence of a hypothetical normative system, and deontic logic should be able to represent that system.]

True enough. But even here the problem is not with representing the legal system. The hypothetical normative system that deontic logic can’t represent is a hypothetical system that includes all one's reasons for action. It is not a hypothetical legal system. That was my limited point.

Posted by: Michael Steven Green | Oct 19, 2007 11:41:44 AM

Michael, I want to reinforce Scott's criticisms of what you say, which yyour clarificatory posts do nothing to displace.

(1) In your reply to Scott, you have legal norms wrong. A mandatory norm (such as a norm of obligation) is an exclusionary reason and, barring exceptional cases, it is also a merely prima facie reason (i.e. it is capable of being defeated by countervailing reasons that it does not exclude). A mandatory legal norm (such as a norm of legal obligation) is therefore an exclusionary legal reason and, barring exceptional cases, a merely prima facie legal reason (i.e. it is capable of being defeated by countervailing legal reasons that it does not exclude). Formal deontic logic is unsuitable for the analysis of such a subject matter because, as Scott suggests, formal deontic logic is suited only for the analysis of situations from which all rational and normative conflict has been stripped out (and hence in which the notions of the exclusionary and the prima facie have been made redundant). In law, as in life more generally, such a logic has very limited application as most of the problems that call for a solution are problems about the nature and implications of rational and normative conflict. An example: Legal obligations often conflict with legal permissions. But so far as I know there is no version of formal deontic logic in which an obligation to phi does not entail the absence of a permission not to phi (such that a conflict between an obligation to phi and a permission not to phi cannot be consistently represented).

(2) You have Raz wrong. Raz devised the view of norms and reasons, including legal norms and reasons, that I have sketched above. He also claims, as it happens, that a legal norm is a claimed moral norm, that a legal reason is a claimed moral reason, and so on. But this does not in any way affect the application to legal norms and reasons of his analysis of norms and reasons in general, which, because of the centrality to it of an understanding of conflict, cannot be represented by any system of deontic logic known to me.

(3) In your plea for deontic logic I think that you also forget, more generally, the importance of Aristotle's dictum that we must not expect more precision [in philosophy] than the subject-matter admits. One lesson that moral, political and legal philosophers have learnt since the revival of these subjects in the 1960s and 70s is that practical rationality does not lend itself, except in a very narrow range of cases, to the kind of formalisation that systems of formal deontic logic aspire to. They are therefore right to resort only sparingly to such formalisations.

Posted by: John Gardner | Oct 19, 2007 12:03:25 PM

A certain irony has developed in this thread. The irony is that Raz, particularly in his early work, was very clearly under the influence (one might even argue the spell) of then recent work in deontic logic, especially Von Wright's Norm and Action. A quick glance through the indexes and endnotes of The Concept of a Legal System and Practical Reason and Norms would bear this out.

But more than this, in his discussion of Bentham and Kelsen in Chapters VI and VII of CLS, he adopts wholesale Von Wright's notation for distinguishing different kinds of norms for distinguishing different kinds of laws (O-laws, P-laws, etc., just as Von Wright distinguishes between O-norms and P-norms). Raz even develops the notation further to develop a more sophisticated taxonomy of laws than Von Wright's of norms. So in CLS we get such things as "DS-laws" - sanction-stipulating laws which make the application of the sanction a duty - and "MS-laws", sanction-stipulating laws which merely permit the application of the sanction.

The influence of deontic logic is even more pronounced in PRN, in which Raz of course first articulates the idea of an exclusionary reason. He quite explicitly draws on Von Wright's Norm and Action, Alchouron and Bulygin's Normative Systems, and Ross's Directives and Norms, among others, in his presentation of mandatory and non-mandatory norms and the discussion of normative systems in Chapter 4.

I have never seen this connection between Raz and the deontic logic literature acknowledged in print, or ever mentioned in conversation. (Well, I myself mentioned it to Les Green a few years ago, and he agreed the connection is often overlooked.) It is true that Raz moved away from thinking about things this way, from the mid-80s onwards, perhaps for the reasons John Gardner mentions. But I think we would be reckless not to try to appreciate what value Raz (and many others) saw in the program of deontic logic. My impression is that Raz learned a lot from Von Wright and others - and we should all be glad he did. In my view, his work would be far less interesting and valuable otherwise.

I'll speculate on the value he did see in their work. An idea that Raz takes as axiomatic in the early work is, as he says in the very first sentence of CLS, "every law necessarily belongs to a legal system". This is emphasized again in the 1980 postscript to that work, and I think is presupposed in the discussion of legal systems in PRN, Chapter 5. And it seems to me Raz has never doubted its truth - and moreover, I think the idea is widely accepted today.

The promise of deontic logic for the legal philosopher then is that it might be able to explain the systematic character of the law. In short, many legal relations are deontic relations. And a better understanding of one would lend understanding of the other. It is unclear to me why (pace Shapiro) this much at least can't be granted.

However, the extent to which "traditional deontic logic has the power to represent adequately normative relations in law" is another matter. But it seems that if one really were a "Razian", one would be a bit more sympathetic to the complaint expressed in the original post.

Posted by: Michael Sevel | Oct 19, 2007 2:10:06 PM

I agree with Michael Sevel that deontic logic of the Scandinavian type both helped to advance our understanding of the logical typology of norms and helped us to understand the systematicity of legal systems. I also agree that on these topics Raz learnt much from von Wright (as well as finding many errors in von Wright's work that were generated by the excessive urge to formalize, an excessive urge which Raz himself manifested in CLS and later had cause to regret). These points are consistent with what I said above about the limited role for formal deontic logic in moral, political and legal philosophy.

Posted by: John Gardner | Oct 19, 2007 2:22:28 PM


Keep in mind that I have conceded, for the time being, that deontic logic cannot represent exclusionary reasons. The limited question right now is whether this is a problem for deontic logic’s ability to represent a legal system. I don’t see how it is.

Deontic logic would represent a legal system by showing how a legal recommendation for action is generated by the legal norms of the system. Exclusionary reasons for action would come into play only if legal norms are exclusionary in the sense that, in the light of the exclusionary legal norm, the legal recommendation of the system can be determined only be excluding other *legal* norms. It seems false that legal norms are exclusionary in this sense. And, from what I remember of Raz, he does not suggest anything like this (if I’m wrong, I’m willing to be convinced).

Consider a simple legal norm – don’t drive above 55 m.p.h. Is that norm exclusionary in the sense that one can determine what the law recommends only by not considering other *legal* norms (for example a norm that allows one to drive faster in the event of an emergency)? I don’t see how. When coming to a recommendation, the law takes all legal norms into account.

Now, of course, Raz thinks that this legal recommendation might be an exclusionary reason in the sense of excluding other *non-legal* reasons for action. But that’s not a problem for representing a legal system. That is a problem for representing the normative system that includes all reasons for action.

Now the situation with prima facie reasons may be different. I’ll concede, once again, that deontic logic cannot represent prima facie reasons. And here we might have a real problem in representing a legal system. For it may be that the way that legal norms come together to recommend a legal course of action is one in which legal norms act as prima facie reasons. So it may be that the problem of prima facie reasons would thwart the application of deontic logic to legal systems.

[As a matter of Raz exegesis, however, I'm not at all sure it is true that *Raz* thought that legal norms act as prima facie reasons when determining the course of conduct recommended by the legal system. Indeed, this is arguably incompatible with his exclusive legal positivism.]

Now I have only conceded for the sake of argument that deontic logic cannot model prima facie or exclusionary reasons. No one has yet provided an argument that such representation would be impossible. Deontic logic might be able to model prima facie reasons by assigning weights to these reasons to capture how they yield an all-things-considered recommendation. And if deontic logic could represent how reasons for action come together to yield an all-things-considered recommendation, I would think that it could represent an exclusionary reason by simply adding the exclusionary reason and removing the reasons the exclusionary reason excludes.

One more point that might make my claims about the usefulness of deontic logic more palatable. Let’s assume that deontic logic can represent all our moral reasoning. This does not mean that we could simply use deontic logic alone to tell us what to do.

An analogy with the logical analysis of a natural language is appropriate here. Let’s say that we have a complete Tarski-style truth functional account of a natural language. Such an account would pass over what is most difficult about learning a language, namely learning the meanings of words. All the account would do is show how the meanings of these words (or more generally the meaning of semantic primitives) contribute to the meanings of sentences of which they were a part.

So a perfect representation of a moral system in deontic logic would leave out the most important part of learning what to do, namely, learning the content of our various duties (and if they are prima facie, their weights). What would it do then? One thing it might do is show how these contents come together to recommend a course of action. This is not much but it is something.

Posted by: Michael Steven Green | Oct 19, 2007 2:41:32 PM

Kelsen´s late theory of norms (Allgemeine Theorie der Normen) is grounded on psychologism and scepticism about normative implication. His conception is all but a contribution to recent deontic logic.

Posted by: Tomáš Sobek | Oct 21, 2007 2:38:56 AM

"Deontic explosion, you might argue, suggests that we should not represent deontic relations through a system under which obligations are closed under entailment. But it certainly looks as if deontic logic is illuminating something."

Michael, we should represent deontic relations through a system under which obligations are closed under entailment, but paraconsistent entailment!

Posted by: Tomáš Sobek | Oct 21, 2007 5:38:38 AM

Tomas -

Yes, paraconsistent logic is one possible solution to the problem.

Posted by: Michael Steven Green | Oct 21, 2007 7:49:50 AM

I understand, why anglo-american legal theory ignorates outdated Kelsen´s theory, but why (my goodness) ignorates recent sofisticated normative expressivism (A. Gibbard, S. Blackburn, R. Brandom)? One exception is Kevin Toh´s beautiful paper: "Hart's Expressivism and his Benthamite Progress".

Posted by: Tomáš Sobek | Oct 21, 2007 2:40:39 PM

Interestingly, Green doesn't argue that Anglo-American philosophers ought to study deontic logic because it is really fruitful, interesting, and would advance legal scholarship in the Anglo-American world. Instead, he argues that they should study it because lots of other philosophers are studying it, including some very good philosophers. I don't find this reasoning compelling at all. Shouldn't the content of a field of study--whether it really is fruitful, for example--determine whether it should be studied, and not facts about its popularity?

Green seems to be saying that, since a sufficient number of philosophers elsewhere in the world find a particular topic worth studying, Anglo-American philosophers ought to study it, too. They are criticizable, or at least "odd," in not studying it. I'm unconvinced.

Posted by: EM | Oct 22, 2007 8:22:16 PM

I'm not competent to comment on the importance of deontic logic for legal philosophy (in relation to legal philosophy, I'm just an interested bystander, not a participant).

But some of the contributors to this blog who have belittled the significance of formal deontic logic for legal philosophy have also gone on to claim that deontic logic has "a limited role in moral .... philosophy" as well.

I confess to finding this claim quite shocking. Isn't it perfectly obvious that the study of the logic of 'ought' is part of metaethics, which is a crucial part of moral philosophy?

In general, it strikes me as deplorable that there is so little interest in deontic logic (and in the formal semantics of normative concepts) on the part of metaethicists.

Why are so many metaethicists content simply to ignore deontic logic? I suspect that the explanation in some cases is just plain laziness. In other cases, it is a sort of diffidence about formal technicalities that is typical of those moral philosophers who prefer to retreat from engaging with philosophical logic into the relative safety of the Ethics Ghetto. (Or perhaps, to put it really bluntly, it is just that a higher proportion of the philosophy graduate students who went into ethics were the ones who weren't very good at logic?)

[Because this has so much more to do with metaethics than with legal philosophy, I'll cross-post this on PEA Soup.]

Posted by: Ralph Wedgwood | Oct 23, 2007 4:52:12 AM

John Gardner says: "so far as I know there is no version of formal deontic logic in which an obligation to phi does not entail the absence of a permission not to phi (such that a conflict between an obligation to phi and a permission not to phi cannot be consistently represented)."

In the jargon of deontic logicians, Gardner is alluding to the principle that the 'O'-operator 'It ought to be that ...' and the 'P'-operator 'It may permissibly be that ...' are duals of each other (i.e. 'It may permissibly be that ...' is logically equivalent to 'It is not the case that it ought not to be that ...', etc.). Although I can't right now remember which deontic logicians explored logics in which the 'O'-operator and the 'P'-operator are not logically related in this way, I feel absolutely sure that some deontic logicians have devised systems of this sort.

So readers of this blog should not be misled by John Gardner's claim: there are indeed "versions of formal deontic logic in which an obligation to phi does not entail the absence of a permission not to phi".

Posted by: Ralph Wedgwood | Oct 23, 2007 5:20:15 AM

In standard deontic logics, permission to p is defined as the absence of an obligation to not-p. But there are some systems of deontic logics that do not define permissions in terms of obligations in this way. There can be deontic gaps: Simply because not-p is not obligatory it does not follow that p is permitted. A positive permission is still necessary.

That said, I think Gardner's point still stands. Under these alterantive systems it still follows from an obligation to p that there is an absence of a permission to not-p. It just doesn't follow from an absence of a permission to not-p that there is an obligation to p. Only a system that can accommodate prima facie duties can take on Garder's problem (but that doesn't mean this isn't possible).

Posted by: Michael Steven Green | Oct 23, 2007 7:40:47 AM

Deontic logic is not a logic of norms but the logic of deontic propositions. Michael, what is specifically normative about truth-conditional semantics of deontic logic?

Posted by: Tomáš Sobek | Oct 23, 2007 10:23:58 AM

You’re right, of course, that standard deontic logic represents normative propositions (which have truth values) not norms. But it represents norms indirectly. Consider a simple theorem of deontic logic (the one mentioned by Gardner): Op then –P-p. The theorem is about the relationship between the truth values of the two propositions. But it also tells us that in a normative system where there is an obligation to p there will not be a permission to not-p. The obligations and permissions spoken of here are norms, without truth values.

Posted by: Michael Steven Green | Oct 23, 2007 11:05:33 AM

Ralph's comments have generated some interesting replies here: http://peasoup.typepad.com/peasoup/2007/10/why-so-little-d.html#comments

Posted by: Brian | Oct 23, 2007 3:43:33 PM

Michael, needs legal theory any formal deductive calculus for normative reasoning at all? For what? If legal theory take advatage of something formal then of adequate logico-semantical analysis of normative expressions. Ought claims have a normative import because they express what to do and every semantical machinery of true normative logic must embody this to-be-doneness. Deontic logic is (relative to this purpose) the non-starter but I think, that Gibbard´s normative logic is promising project.

Posted by: Tomáš Sobek | Oct 23, 2007 10:51:03 PM

OK, a friend pointed me at this thread and I thought I would chime in. (But it is my first blog post anywhere, ever, so please don't jump all over me.)

I'm not going to address the question of whether deontic logic is a ghetto. Sometimes I think all of logic is just a bunch of ghettos. Proof theory is one of the grand old traditions in logic, but these days proof theorists only talk to each other, and there aren't many of them. Who pays attention to relevance logic except other relevance logicians?

I also can't say whether American legal theorists should should not pay more attention to deontic logic, since I know so little about legal theory (but I'll come back to this). And I have a lot of sympathy for the general feeling in many of these posts that deontic logic is empty, sterile, a mass of notation, not worth the effort, etc, etc. I understand where that feeling comes from. I think that, with only a few exceptions, most of the work in deontic logic for several decades until about the early 90's had been going sideways, rather than forward, and the field was not in good shape.

But then an odd thing happened. From a purely numerical standpoint, at least, the field was taken over by computer scientists who were interested in using the formalism to specify the rules, permissions, and so on governing multiple interacting computer systems (as in: does this electronic agent have permission to access the medical records in this database?), and also, and more relevant to this thread, by researchers in artificial intelligence and law. This has given the field a much more practical orientation, introduced new problems, and also, I think, increased the quality of the work. Not all of this work is good, but some is very good - and I think some might be relevant to philosophy, and maybe to legal theory as well.

So, going through this thread, a couple of things jumped out at me.

First, there was Scott's claim that prima facie oughts are important, but that they can't be represented in traditional deontic logic (very true), that maybe nonmonotonic logic could help, but that this is new and unproven technology and we'd better wait to see how it shakes out before adopting it.

Again, I understand the feeling, but I here I disagree. Nonmonotonic logics have been around since 1980, which is 27 years now. If we suppose logic started in 1879 (just to pick a random date :)), that means there have been formal nonmonotonic logics for at least 20% of the time there has been logic at all - that's hardly bleeding edge technology! More seriously, these systems make sense, they have been studied carefully by smart people and are pretty well understood. What's true, though, and I think what's responsible for the "this is all new stuff" feeling, is that there are very few accessible presentations of the work - much of it is unforgivably rebarbative, for which all of us in the field share the blame. (I have an expository book under contract now, but it's going to take some time before I get to it.)

Still, the idea of basing deontic logic on nonmonotonic logic is now a standard option (I would be happy to provide reference to anyone who cares), and it is used, especially, by people in AI and Law. So, I don't think we should be so bleak about this idea. It's out there, and I think it would be great if someone in legal theory would try to do something with it.

Second, there was another of Scott's claims (and Michael and others agreed),that deontic logic can't model exclusionary reasons, and so can't represent Raz's ideas. OK, I know this is really rude, and I hate to plug my own work, but I think the following paper from Philosopher's Imprint shows exactly how we can understand exclusionary reasons, using the tools of deontic logic:


I won't go into details, but here's the general idea. There are reasons for action (the paper provides a recipe explaining how all the various reasons come together to support whatever actions they do). The reasons have priority ranking, and more important reasons defeat less important reasons in case of conflict. One of the things we can do is adjust the priority assigned to our reasons, making them higher or lower, and there may be reasons for doing this as well (Jonathan Dancy calls these intensifiers and attenuators). Sometimes, on the basis of some reason, we attenuate the priority assigned to a reason to such an extent that it falls below some threshold, so that we then pay no more attention to it at all. This reason is then excluded from consideration. A reason for excluding a reason from consideration - placing its priority below threshold - is an exclusionary reason. (In the paper I go through Raz's own example of Colin's promise to his wife not to consider, in some decision, those reasons not bearing directly on the interest of their son.)

OK, anyway, I'm pretty sure this all works out technically, so that deontic logic can in fact help to model exclusionary reasons, but the main point I wanted to make here was that the formal tools used in this analysis were mostly not developed by me (I wish they had been), but by people working in AI and Law, notably Tom Gordon, Henry Prakken, and Giovanni Sartor, who were concerned to model the way in which a court might apply certain reasons in working out the result of some case while at the same time, as part of the same process of reasoning, applying other reason to work out the priorities among those reasons from the first set, and so on, where circularities aren't ruled out. (Interestingly, these three individuals, two of whom (Henry and Giovanni) are law professors in Europe as well as computer scientists, either didn't notice or didn't feel like working out the connection of their work to exclusionary reasons - which may show you something right there about the differences between European and Anglo legal theorists. Oh, and I should also say that Mark Schroeder, in his Princeton PhD thesis, had independently worked out a similar approach to reasons, and I learned a lot about doing this from talking to him, as well as the European AI and Law people.)

Anyway, back on topic, the point is that these tools arose as part of a detailed study of legal reasoning, and that's where I think deontic logic, particular in its default or nonmonotonic versions, may have some interest for legal theory - not in mapping out broad positions in jurisprudence, but in helping us to understand legal reasoning in some detail.

I was thinking about this a while back when I read Schauer's review of Lacey's Hart biography, particularly its discussion of Goodhart, Hart's predecessor, and his work on precedent. Schauer writes (page 867, footnote 50): "As Lacey makes clear, Hart thought that this kind of work on precedent and related topics of legal argument and legal reasoning was of little theoretical interest and beneath the attention of a serious legal philosopher." I don't know enough about legal philosophy to know if this claim is true, either about Hart, or whether in fact it persists today. But if it does, that would help to explain the lack of interest in logic. Logic helps us understand detailed patterns of legal argument and legal reasoning, but if that's not on the agenda for legal theorists, for whatever reason (and I don't mean to judge this - legal theory may have bigger fish to fry!), then no, I don't think legal theorists need to study logic either.

(Oh and finally, finally, back to Brian's original comment that Legal Theory hasn't published anything in deontic logic. Well, technically, that's right, but they did publish a short paper I wrote in 2003 or 2004 on precedent, which used some tools from AI and Law, and honestly deontic logic was right there, really it was, I just didn't bother to write out all the O's and P's.)

Posted by: Jeff Horty | Oct 25, 2007 6:55:47 PM

Thanks Jeff- that was very interesting and helpful.

Posted by: Matt | Oct 26, 2007 7:21:50 AM

Thanks Jeff - I was waiting for some real deontic logicians to start defending the topic.

Posted by: Michael Steven Green | Oct 28, 2007 7:40:52 AM

Thanks, Jeff. Very nice post. Let me remind everyone that Michael's post accused anglophone legal philosophy of prejudice for ignoring deontic logic. My remarks were intended to rebut that claim, not to show that deontic logic is of no use for legal theory. (At some point in the thread, Michael says that he is not criticizing any one particular legal philosopher, but the profession in general. I am not sure, then, that I understand Michael's original point.)

As for your claim that non-monotonic logic is old enough to be incorporated into legal theory, let me clarify my remarks in this regard. When I was working on the topic in the mid-90's, I remember it being controversial whether deontic modus ponens (i.e., A>B, A |- T>B) should be satisfied in an adequate theory of nonmonotonic deontic logic. I thought it should but many of the systems, such as Lewis' V system, didn't validate it.

Perhaps the issue of DMP has been resolved by now -- I stopped working in this area a decade ago so I don't know. But it was issues like that (which are pretty central) which I had in mind when I said that nonmonotonic logic is still new technology. I thought it was unreasonable to blame legal philosophers from getting involved in such issues until the real logicians have sorted them out.

Posted by: Scott Shapiro | Oct 28, 2007 11:31:14 AM

The last sentence of the previous post should read: "I thought it was unreasonable to blame legal philosophers *for not* getting involved in such issues until the real logicians have sorted them out." Sorry about that.

Posted by: Scott Shapiro | Oct 28, 2007 11:36:40 AM

Hi Scott – let me see if I can make the individual/profession point a bit clearer and, as this thread is apparently winding down, remind everyone of the topic of my original post. I brought up deontic logic in connection with a general pattern among American philosophers of law of ignoring the structural character of legal systems – the fundamental relations between laws that organize them together in the same legal system. Lack of interest in Kelsen was, I argued, part of the same pattern.

This claim is about the profession in general. As an analogy, imagine that no one did tort theory in this country (but it was done a lot by non-Anglophones). I think that such a situation would be bad, but I could not say to any particular American philosopher of law – you must do tort theory! He can choose to do other things. My claim would be about the profession as a whole. It would be better off if someone did tort theory.

In fact, the only person who has any real obligation is the person making the criticism. And I have started trying to discuss these ignored structural issues (although not deontic logic) in my own work, e.g. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=455620 and http://papers.ssrn.com/sol3/papers.cfm?abstract_id=881073 .

The model for me of the type of work that I think is missing in this country was in fact Raz’s The Concept of a Legal System. It is a nice confirmation of my original thesis that, as Michael Sevel has helpfully noted, Raz was influenced not merely by Kelsen but also by deontic logic. My original point was that interest in deontic logic seems to come hand in hand with interest in these structural questions. They are all ignored in this country.

As the example of Raz shows, the missing interest in deontic logic is not that closely identified with the *formal* quality of deontic logic, but instead with its goal of understanding the structure of normative relations. And in fact much of the deontic logic literature that I cited in my original post does not have that much formalism.

There is, incidentally, an interesting discussion in the Pea Soup thread, http://peasoup.typepad.com/peasoup/2007/10/why-so-little-d.html#comments about the scope of deontic logic when one gives up the demand of formalization (e.g. whether Bernard Williams could be considered to have done deontic logic in “Ethical Consistency”). I think the best criticism of my post would be that once informal deontic logic is included the topic *is* discussed by philosophers of law in this country. But I think there still is a real difference between the Americans and the non-Anglophones (with the British as a transitional case).

Now you challenged me to identify where formal deontic logic helps philosophers of law. It is a good question and, in a kind of mission creep, I tried to take it up (because I am in fact sympathetic to the benefits of formalization). But it was awkward for me because, as you no doubt know, I’m not a deontic logician. I’m interested and sympathetic, but my exposure to formal deontic logic is spotty. That’s why I’m happy that people like Jeff have joined the debate.

It is also worth noting concerning the formalization debate that your attitude is different from John Gardner’s. Gardner apparently has resistance to the very possibility of formalization. You apparently don’t have the same worries – you’re just skeptical about the benefits of formalization for the philosophers of law.

Posted by: Michael Steven Green | Oct 31, 2007 2:32:44 AM

Thanks, everyone, for the kind comments.

I want to make it clear that I didn't mean to take any stand on the original debate - wasn't it something like whether English speaking legal theorists should be faulted for ignoring deontic logic. Honestly, I don't know enough about legal theory to say, and even if I did, it's hard to fault someone for being interested in what they're interested in and not something else.

I mostly just wanted to point out a line of research, in AI and Law, that I think might interest legal theorists, at least if legal theory includes the detailed study of legal reasoning, and that does make frequent use of deontic logic. And also to correct some misconceptions (though I think completely justified, mind you) about what's happening now in deontic logic.

Which brings be now to John Gardner's post (I think it's his first one - I still haven't read this whole thread). On John's point (1), I just want to say that reasoning in the presence of conflicting norms like these is what a lot of people now study. It's something I've looked at (see above post), but others too. We are not trying to strip out the conflicts before studying the reasoning in some formal way. We want to understand how normative reasoning in the presence of conflict.

(Quick techie note to Scott: I didn't quite get the example, because I couldn't tell what your symbol > meant - was that a conditional ought? But anyway, I think I get what you're saying, and yes, modus ponens is tricky, because you want it to hold in general but be defeated under certain conditions. The kind of conditional logics you're talking about have a lot of trouble with this (and the AGM lots you mentioned earlier, and model preference logics in general). But default logics don't have these problems. So that's what people tend to work with now. Maybe a few exceptions - possibly David Makinson is still an exception.)

Posted by: Jeff Horty | Oct 31, 2007 10:11:53 PM

Post a comment