LJIL Symposium: Darryl Robinson comments on James Stewart’s “End of Modes of Liability”

[Darryl Robinson is Assistant Professor at Queen’s University Faculty of Law]

James Stewart’s article “The End of ‘Modes of Liability’ for International Crimes” is an impressive piece of scholarship.  It is one of the most sophisticated works to date in bringing the rich scholarship on criminal law theory to bear on problems of international criminal law (ICL).  Stewart brings admirable—and frankly enviable—skill in succinctly explaining major controversies in criminal law theory and weaving that literature into an analysis of ICL issues.  While I will raise some critical questions here, I applaud and share his objective of theorizing about ICL from a liberal perspective.

Of course, the value of any comment lies in the disagreements.  Surprisingly, I agree with the aspects with which I might be presumed to disagree, and I disagree with some aspects with which I might be presumed to agree.  Namely, I agree with his proposition that many problems of ICL flow from domestic law and not from international influences, even though that proposition is presented as oppositional to the views of myself and others.  Conversely, I ask whether Stewart’s critiques of complicity might be unnecessarily stringent.

Domestic v. International Influences
Stewart quite rightly points out that domestic legal systems often contain doctrines that arguably contradict fundamental liberal principles, and that some of ICL’s problematic doctrines were drawn from domestic legal systems.  This is a valuable complement to the point made by scholars such as Danner, Martinez, Fletcher, Ohlin, Damaška and me that some problematic doctrines are fostered by reasoning particular to ICL.

However, when contrasting his position with that of other scholars, Stewart seems to somewhat overstate their position.  For example, he refers to and refutes the “thesis that broad modes of liability are necessarily hatched internationally” (p. 179), and the “assurance that unprincipled international rules necessarily reveal the triumph of international agenda over the restraining force of the criminal law” (p. 203) (emphasis added).  As far as I know, none of the cited scholars have ever suggested any such necessary linkage.  They have not suggested that all, or even most, of the problematic doctrines of ICL flow from international influences.  Similarly, Stewart demonstrates that some departures are not “nefarious creations of an illiberal international system” (p. 198) or a “nefarious utilitarian agenda derived from [ICL’s] international political status” (p. 182).  I certainly agree, but the refutation is misplaced in that nefariousness has not been suggested.  I also don’t think scholars have suggested that domestic systems are free of problematic doctrines (p. 169).

ICL drew heavily on domestic criminal law as a source, and in so doing it inevitably absorbed a great many problems and controversies from criminal law systems.  The point of these scholars was that, in addition to problems typical of criminal law systems, ICL also has particular features that can generate additional difficulties.

The proposition that “A sometimes leads to B” must not be conflated with the very different proposition that “all B are caused by A”.  For example, my piece on “the identity crisis” of ICL demonstrates that “A” (certain habits of reasoning in human rights law) sometimes leads to “B” (internal contradictions with the system’s declared fundamental principles).  I emphasized that these influences are only part of the problem and not the sole cause of departures from principles, that national systems also feature controversial doctrines, and that my interest was in certain popular method of reasoning.  Stewart acknowledges the latter points, for which I am grateful.  My reading of Danner, Martinez, Fletcher, Ohlin, Damaška and Greenawalt is that they also were not putting forward any absolutist hypotheses, but rather investigating specific tendencies of ICL.

Stewart’s point is nonetheless timely and important.  In the wake of articles highlighting the distinct influences in ICL, many thinkers may understandably have formed a halcyon impression of domestic criminal law or an impression that ICL is uniquely and irremediably flawed.  Thus, his argument about domestic influences is a harmonious counterpoint, rather than counterargument, to arguments concerning international influences.  The propositions are complementary, not contradictory.

The Critique of Modes and the Differentiated Model
I share Stewart’s objective of ensuring a liberal system of blame attribution.  Stewart provides a meticulous and well-reasoned argument which aims to show that the liberal critique of particularly problematic doctrines applies equally to complicity in general.  The analysis is skillful.  However, I ask whether Stewart might be applying the criteria too stringently, and throwing out one or more babies with the bathwater.  If one’s theory disapproves of such a large part of the content of the criminal law, it is always possible that the problem lies not with the doctrine but with one’s theory.

Indeed, a differentiated model might actually help save some of the impugned modes of liability.  A differentiated model (such as in current ICL) distinguishes between principals, who cause or control the crime, and accessories, who contribute more indirectly, by influencing or assisting the principals.  Plausible liberal theories indicate that accessories may make a much more indirect causal contribution to the crime.  Further, accessories may not need the paradigmatic mens rea for the offence; other mental states related to their influencing or assisting can be deontologically justified, because they are only being blamed for influencing or assisting.  For example, Stewart discusses JCE-III, a doctrine which holds a person liable for crimes of special intent (such as persecution) in the absence of the requisite intent.  However, part of the problem with JCE-III is that it deems the person to “commit” the crime, i.e. it holds them liable as a principal.  The critique may not apply with equal force to accessorial modes of liability, such as aiding and abetting persecution.  A plausible case for deontological justification could be advanced.  So, some of the sweeping away of doctrines may be overly hasty.

Stewart argues for a unitary system, which would dispose of modes of liability and the distinction between principal and accessory.  In such a system, degrees of responsibility would be reflected at the sentencing stage.  I agree that a unitary system is certainly defensible; some legal systems are unitary and many legal theorists argue for such a system.  But is such a system preferable?  Is it so preferable as to warrant the radical change it would entail in ICL?  In light of the path dependence of law, such a sweeping change (including amendment of the ICC Statute and reversal of existing jurisprudence) would be difficult indeed.  But let us set aside feasibility and focus on the normative question.

A common argument for a differentiated model is the expressive function: to reflect meaningful moral differences between those who cause or control the crime and those who made blameworthy but minor and secondary contributions.  Stewart advances some strong counter-arguments to the expressive argument.  He notes that the Tribunals often do not even specify the mode of liability when entering convictions.  He notes that labels (eg. aiding and abetting, perpetration, commission through another) may not meaningfully communicate culpability to general audiences.  Most devastatingly, they may even miscommunicate culpability if different legal cultures using different tests.

The arguments are excellent, but the solution of dispensing with the differentiated model is arguably a move in the wrong direction.  Stewart argues that on a unitary model, one can add detailed descriptions of the nature of the responsibility.  One could explain that “I was convicted of crimes against humanity because I provided arms with such and such knowledge and such and such intent”.  However, a differentiated system can also provide additional narrative details, just as well as a unitary system can.  Moreover, modes of liability do offer a useful and meaningful shorthand referent for the nature of one’s responsibility.  Labels such as “aiding and abetting a war crime of murder” or “instigating genocide” provide succinct meaningful information about morally significant differences.

A differentiated model also allows for some more calibrated deontological assessments.  For example, I believe that command responsibility is a mode of accessorial liability (I realize this once-obvious point is now intensely controversial; I discuss this in a forthcoming work on how command responsibility got so complicated).  I also believe that a “should have known” standard is deontologically-justifiable in the context of a command relationship.  However, a unitary model would have to either allow the “should have known” standard in all contexts, making it over-inclusive, or else prohibit it across the board, making it over-restrictive, or else allow varying standards in varying contexts, in which case it would break down into a differentiated model in disguise.

Thus, I think there are reasons for pause before moving from a differentiated to unitary model.  Despite these disagreements with his ultimate conclusions, I think Stewart’s piece is a masterful and commendable work.  It is an exemplar of the emerging tradition of sophisticated and thoughtful ICL scholarship.  In particular, Stewart’s sophisticated and adroit handling of issues such as moral choice, fair labeling, conduct and harm, moral luck, causation and culpability is a standard-setter for future thoughtful work in this area.  It is sure to be an essential reference point in discussions on liberal criminal theory in ICL.