Mark Dsouza
* This set of posts is written with a view to being accessible to a wide range of readership. I’ve therefore kept footnotes and references to a minimum. The ideas I discuss here are set out in much greater detail, with references, bibliography, index, and other bells and whistles in my 2017 monograph, ‘Rationale-Based Defences in Criminal Law’, which you can find in larger public and university libraries, in addition to being available to purchase online and in bookstores. Some of the argument also comes from my paper ‘Criminal Culpability after the Act’ (2015) 26(3) King’s Law Journal 440. If you have no way of accessing a copy but would like to check my sources on any of these ideas, please reach out to me at m.dsouza@ucl.ac.uk, and I’ll try to help.
Part 1 of this series of posts outlined my ‘quality of reasoning’ hypothesis about how rationale-based defences work, and identified various things that I needed to show to make it good. I started on the first of these in my previous post, in which I argued that criminal law conduct rules that restrict liberty should be drafted by reference to the facts as they really exist. The next task is to consider whether the same is true for decision rules; in particular, those that guide a criminal court in making personal blaming judgments in respect of a particular case and a particular agent. But this is a big task, best broken into small pieces. The first step, which I take in this post, is to distinguish between two ‘threads’ of our reasoning process, viz. norm-reasoning and functional-reasoning, and consider how blaming evaluations relating to these two threads ought to be made. I will argue that the criminal law ought to blame only for poor norm-reasoning as reflected in our choices, and show that this proposition does not unduly narrow the reach of the criminal law.
The threads of reasoning
According to Raz, we use reasons not just to guide conduct, but also to evaluate it. When conduct is evaluated with a view to judging the agent whose conduct it was, we look at the quality of the agent’s reasoning. A rational agent is blameworthy or praiseworthy depending on whether her conduct, in her circumstances, reflects poorly or well on the quality of her reasoning; that’s why one prerequisite of any blaming evaluation of an agent is that she must have been capable of exercising rational facilities.
Now let’s separate out two threads of a person’s reasoning; the first, to do with her exercised capacity to observe facts and reach reasoned factual conclusions and judgments on their basis (call this functional-reasoning), and the second to do with her understanding of and responsiveness to normative guidance (call this norm-reasoning). As an example of the distinction I’m proposing, consider this (based on R v Scarlett [1993] 4 All ER 629):
T, a burly drunk at a pub, is asked by the pub landlord, D, to leave for being aggressive. This just provokes T, who starts swinging his arms around threateningly. So, D pins T’s arms to his side, and bundles him onto the landing outside the pub door. Because D uses a bit too much force in doing so, T loses his balance, and falls down a short flight of steps into the street, striking his head, and eventually dying from the head injury.
When questioned, D might explain that he did not intend to use as much force as he had ended up using. Perhaps he made a mistake about how strong he was, or how heavy T was, or how drunk and off-balance T was, or how slippery the floor was, and so underestimated the effect that the force would have. D employs functional-reasoning in order to make all of these assessments, and so in offering these explanations, D admits that his functional-reasoning was deficient. Alternatively, D might say that he intended to use exactly as much force as he used, but believed that he was entitled to use that much force, or didn’t care whether or not he was entitled to use that much force. In acting on this basis, D employs norm-reasoning, and so by these explanations, D admits that his norm-reasoning was deficient.
Similarly, consider this example (based on Beckford v R [1988] AC 130):
D, a policeman, is called to a house to investigate a report of a dangerous gunman terrorising someone there. On arrival, he sees T running away with what appears to be a gun. He chases T and coming upon him as he rounds a corner, shoots T dead. At the time, T is unarmed, on his knees, with his hands in the air, begging not to be shot.
When questioned, D might explain that at the time, he believed (albeit wrongly) that T was about to shoot at him. Or perhaps he might say that he had aimed to fire a warning shot above T’s head (but missed). In making judgments like whether T was about to shoot at him, or whether he was aiming correctly, T exercises his functional-reasoning. Mistakes he makes in doing so, potentially reveal his functional-reasoning in poor light. Alternatively, D might explain that he intended to shoot T dead despite knowing that T was not trying to shoot at him, because he believed that he was entitled to shoot with deadly intent at suspects who flee, or because he didn’t care whether or not he was so entitled. In acting on this basis, D employs norm-reasoning, and so mistakes he makes potentially reveal his norm-reasoning in poor light.
We can characterise the blame attaching to someone who displays poor functional-reasoning as ‘functional-blame’. Functional-blaming judgments include judgements of carelessness and negligence. Functional-blameworthy persons exhibit deficient functional-reasoning in failing to observe certain situational facts, or in reaching incorrect conclusions on the basis of those facts, or in underestimating or overestimating the potential risks of a course of action, or the extent of the consequences it might generate. The common thread in all these errors of reasoning is that (in the absence of any normative guidance as to how good the agent’s functional-reasoning should be) they do not definitively reveal her the agent in a bad light in respect of her commitment to, and appreciation of, normative rules of conduct. All these errors definitively show is that the agent made an error of judgment in her assessment of aspects of the physical world.
Similarly, let’s call the blame attaching to someone who displays poor norm-reasoning, ‘norm-blame’. Norm-blaming judgements attach to persons who exhibit norm-reasoning that is deficient in respect of their commitment to, and appreciation of, normative rules guiding conduct. Since norm-blameworthiness depends on conduct rather than outcomes, a person can be norm-blameworthy even if the harm that the norm was aimed at preventing does not occur, and norm-blameless even if that harm does occur. Moreover, if we accept that the norms relating to core substantive criminal law offences are morally derived, a person with a norm-blameworthy attitude to these core norms also shows herself to be prima facie morally blameworthy or evil.
The applicable evaluative standards
If blameworthiness depends on the quality of a person’s reasoning, then we must ask, “To what standard should the defendant’s reasoning be compared?” By dividing the defendant’s reasoning into functional-reasoning and norm-reasoning, we can treat this as two separate questions vis-a-vis the different types of reasoning.
It is immediately clear that for the purposes of attributing blame, there is no point in comparing the defendant’s functional-reasoning to the perfect functional-reasoning of someone with a ‘god’s eye’ awareness of all circumstantial facts. No human can attain a ‘god’s eye’ awareness of everything, so that is an unrealistic yardstick for judging a human. Any person who makes an epistemic mistake falls short of the ‘god’s eye’ standard of functional-reasoning, but even outside legal contexts, when we call someone careless or negligent, we are not just saying of her that she happened to make an epistemic mistake. We are asserting that she fell short even of the less demanding societal standards governing how much care she should have taken when exercising her functional-reasoning skills. The context within which the agent acts often determines how demanding the applicable standard of functional-reasoning is – a trained professional is societally expected to exercise more care in her functional-reasoning when acting within the scope of her training than a layperson. But even outside the realms of special training, an agent would be careless if she did not take as much care as we’d expect from an ordinary member of society in the same situation. This means that one can meet societal expectations pertaining to the exercise of functional-reasoning skills and still get it wrong. What matters for the attribution of functional-blame is not whether one ultimately gets it right or wrong, but whether one took as much care as was societally expected (even if not enshrined in what Raz would call an ‘institutionalized system’) when exercising one’s functional-reasoning skills.
Frequently, the law institutionalises such societally determined standards of functional-reasoning, by referring to them. It does so when adjudicating both criminal and civil liability. But note that these benchmarks are societal facts that exist independently of the institutional systems into which they are incorporated. Independently of the law, these non-institutionalised societal norms tell us how good our functional-reasoning should be. We should judge the quality of an agent’s functional-reasoning by reference to them, and if her functional-reasoning is found deficient, the agent is liable to functional-blame.
We turn now to identifying the appropriate benchmark for assessing norm-reasoning. The first thing to consider here is whether the quality of an agent’s norm-reasoning should be assessed by reference to the norms applicable to the facts as they objectively exist, or by reference to the norms applicable to the facts as the agent perceives them. As a practical matter, an agent can only apply her norm-reasoning capacities to the norms that she thinks apply in her circumstances. So, assuming that the agent is not mistaken as to the content of any norm,[1] the quality of her norm-reasoning can only be assessed by reference to the norms that apply to the facts as she perceives them. Therefore, even if our conduct norms are framed by reference to facts as they objectively exist, decision rules governing an agent’s desert of norm-blame should require us to evaluate her norm-reasoning by reference to the normative guidance applicable to the facts as she perceived them. If she was not appropriately guided by those conduct norms, she is norm-blameworthy. And when is an agent not appropriately guided by conduct norms? When she chooses not to exercise her capacity to be guided by the norm in her behaviour (which may be because she inculcates contrary normative values, or because she chooses to ignore the norms).
Note also that since norm-reasoning and functional-reasoning are conceptually distinct, in principle, the evaluation of the quality of a person’s norm-reasoning is distinct from, and should not be influenced by the quality of her functional-reasoning. To unpack this proposition, consider the following permutations:
- Because of her poor functional-reasoning, D misidentifies the applicable normative guidance. However, she exercises her capacity to follow the (misidentified) norm guidance. She does not display poor norm-reasoning, even though on the whole, she does the wrong thing in the circumstances. Her error here is purely one of functional-reasoning.
- D correctly identifies the applicable normative guidance, but chooses not to exercise her capacity to follow it. Again, on the whole she does the wrong thing in the circumstances, but this time she displays poor norm-reasoning, and not poor functional-reasoning.
- D exhibits poor functional-reasoning in misidentifying the applicable norm guidance, and chooses not to be guided even by those norms. Here too she does the wrong thing in the circumstances, but this time she displays both poor functional-reasoning, and poor norm-reasoning.
- There is normative guidance directing D to meet a certain standard in terms of her functional-reasoning – say the standard of a reasonable person – which also happens to be the expected societal standard. D’s functional-reasoning does not meet the societal standard. In this case, the same exercise of D’s rational capacities attracts both functional-blame and norm-blame. But although D’s desert of norm-blame and functional-blame coincides, her functional-reasoning and norm-reasoning remain distinct, and are benchmarked against standards set by different systems. The benchmark for D’s functional-reasoning is the non-institutionalised system of societal expectations, whereas the benchmark for her norm-reasoning is the institutionalised normative guidance directing D to meet the standard of the reasonable person in her functional-reasoning.
The potential payoff
So we can separate out functional-reasoning and norm-reasoning, and consequently functional-blame and norm-blame, in this way, but why should we? Because, I argue, keeping the two threads of reasoning distinct improves our understanding of criminal culpability (both inculpation, and exculpation). My overall project in this set of posts is concerned with exculpation, and the payoff relating to exculpation will become evident by the end of this project. For now, let me outline the payoff for our understanding of inculpation.
Traditionally, theories of culpability try to trace our negative evaluation of some prohibited outcome ‘X’ back to D, the person who brings X about. Choice theorists do this by arguing that D is criminally blameworthy for causing X only if she chooses not to exercise her capacity to avoid causing X. This choice reveals something important about her qua an agent: that she is not motivated by good values, or that she is motivated by bad values. But this approach means that choice theorists struggle to link X to a negative evaluation of D, in the absence of anything as unequivocal as advertent choice to do, or cause, or risk doing or causing X. They therefore struggle to explain, or deem illegitimate, negligence-based criminal liability. Character theorists argue that the moral (and therefore criminal) appraisal of D is properly focused on D’s character, rather than merely her choices, and having ‘insufficient concern for the interests of others’ is a blameworthy character trait. But these theorists attract criticism for defining the ‘in-principle’ domain of criminal culpability too broadly; on this view, since all negligence shows ‘insufficient concern for the interests of others’, all negligence is, in principle, criminally blameworthy. Both approaches to culpability then generate pictures of the criminal justice system that are intuitively unfamiliar in terms of their reach.
But let’s re-examine the foundational premise that both types of theories share, viz. that a culpability theory (understood as a theory of when a person deserves criminal blame) must trace our negative evaluation of a prohibited outcome X back to D, who brings X about. We have no reason to assume, as traditional theories of culpability do, the connection traced must be direct. In fact, keeping the two threads the reasoning distinct shows that it is much more plausible to think that this connection is indirect.
Here’s how. I have argued that norm-blame flows from a failure to respond appropriately to the guidance in the conduct norms applicable to the situation in which one believes oneself to be. The conduct norms relevant to the criminal law are contained in the system of normative guidance underlying the criminal law system, and a person who fails to respond appropriately to these norms is criminally blameworthy. A large chunk of the guidance in the criminal law’s system of norms is contained in offence stipulations which imperatively restrict a person’s original unfettered capacity to behave as she chooses (though as I suggested in Part 1, some of the criminal law’s conduct guidance is also contained in justifications). Now consider an offence stipulation designed to prevent X. The branch of criminal law theory relating to the identification of things that we ought criminally proscribe – the X’s of this world – and the formulation of normative conduct guidance designed to prevent them from happening, is not culpability theory; it is criminalisation theory. The conduct guidance concerned requires us to choose our behaviour – including, where appropriate, how much care we take in exercising our functional-reasoning skills – so as to prevent X from happening. Once criminalisation theory has done its bit, and generated an offence stipulation designed to prevent X, all culpability theory needs to do is connect this normative conduct guidance to a proper blaming judgement for the agent who brought about X. In other words, the connection between the occurrence of some X, and a criminal blaming judgement for D for bringing about X, is indirect; the intervening junction is the normative conduct guidance designed to prevent X from occurring, i.e. the relevant offence stipulation. Criminalisation theory connects the occurrence of X to this normative conduct guidance, and culpability theory connects the normative conduct guidance to a judgement about the blameworthiness of the agent D.
This means that with a slight modification, the choice theory can explain criminal blameworthiness including blameworthiness in the vast majority of negligence liability cases. Since criminal blameworthiness is norm-blameworthiness traceable to conduct-guiding norms in the criminal law’s institutionalised system of conduct guidance, we need only ask whether there was advance criminal law guidance requiring D not to X, and if so, whether D chose not to exercise her capacity to be guided by this advance guidance. This view also explains culpability in the vast majority of negligence liability cases – these will be cases in which, at the criminalisation stage, it was decided to issue guidance as to how much care we should take when exercising one’s faculties of functional-reasoning. D’s failure to exercise her capacity to be guided by this advance guidance straightforwardly attracts criminal law norm-blame.
* This post summarises the argument in Chapter 2.2, 2.2.1, 2.2.2, and 2.2.3 of my 2017 monograph, ‘Rationale-Based Defences in Criminal Law’. It also contains a condensed form of the argument in my paper ‘Criminal Culpability after the Act’ (2015) 26(3) King’s Law Journal 440. For much more detail, please consult these resources. As background, I recommend reading the Introduction to my monograph as well – this sets out the overall plan of action for the monograph, and explains how I hope to convince readers about my thesis despite it addressing a puzzle in criminal law theory to which there is no uniquely correct answer.
[1] These may differ from the norms actually applicable to those circumstances if the agent has mischaracterised the situation, or if she is mistaken as to the content of the norms. The latter possibility is worthy of much more sustained attention than I can give it here, but fortunately, it can be bracketed without affecting the rest of the discussion. For the present purposes, I assume that the agent correctly understands the content of the norms.
Criminal Justice Theory Blog 