Abstract
This article challenges English criminal law's approach to causation. In doing so, it proposes replacing the standard tests of causation with a single test, known as ‘INUS’ causation – where a cause is an
Keywords
Introduction
On 21 October 2021 the actor Alec Baldwin was filming on the set of the upcoming film Rust when a prop gun he was pointing at cinematographer Halyna Hutchins discharged a live bullet that killed her. Baldwin denied pulling the trigger and it is alleged that the prop supplier company distributed a mix of dummy and live ammunition on set. Amongst other factors, the question of whether Baldwin should be criminally liable for Hutchins’ death depends on whether his conduct was a cause of her death, which also speaks to his broader criminal responsibility. That specific – but crucial – causal enquiry forms the backdrop for this article. It directly challenges English criminal law's standard tests for causation – the legal doctrines that govern the connection between the defendant's (D's) conduct and outcome in result crimes. 1
It is uncontroversial to say that the standard tests are problematic. The ‘but-for’ test of factual causation leads to both over and under-inclusive results and the additional criteria, known as legal (or ‘proximate’) causation, have rendered the law of causation ‘unstable and irrational’. 2 However, there is a further, neglected problem with these standard tests in the context of criminal-responsibility ascription. This article argues that those tests represent a normative 3 exercise in finding D responsible for a prohibited outcome, often grounded only in D's moral responsibility 4 for that outcome. Such an approach is problematic because moral responsibility is irrelevant to causal responsibility; and a failure to distinguish causal responsibility from moral responsibility results in inappropriate criminal-responsibility ascription for result crimes.
Criminal responsibility is central to criminal justice because it concerns the basis on which individuals are called to account for criminal conduct. 5 The claim that the standard tests contribute inappropriately to criminal-responsibility ascription is therefore important. It reveals that individuals are inappropriately called to account for criminal conduct in result crimes.
This concern prompts my examination of an alternative approach to causation in English criminal law. The article contends that it is preferable for English criminal law to adopt a single, non-normative test of causation; a metaphysical one that offers a robust causal enquiry that focuses only on causal responsibility. It examines the potential utility of the ‘INUS’ test of causation: Which holds that a cause is an
The analysis of causation in terms of necessary and sufficient conditions is the most favoured line of metaphysical causal enquiry, stemming back to Hume and Mill. 6 However, even though INUS causation is considered by some to be the most sophisticated attempt to analyse causation along these lines, 7 it rarely features in academic scholarship on criminal law. This article, therefore, fills an intellectual gap by proposing INUS causation as an alternative approach to causation in English criminal law and examining its practical utility.
The article starts, in the ‘Causation in English criminal law: Ascribing moral responsibility’ section, by analysing the standard tests in English criminal law. It claims that these tests represent a normative exercise in finding D responsible for a prohibited outcome, often grounded only in D's moral responsibility for that outcome. Such an approach to causation gives rise to incoherence and uncertainty. Then, the ‘Criminal law inappropriately ascribing criminal responsibility’ section demonstrates why that approach is problematic in the context of criminal-responsibility ascription. Moral responsibility is irrelevant to causal responsibility; and not distinguishing causal responsibility from moral responsibility results in inappropriate criminal-responsibility ascription for result crimes. This is because both notions – causal and moral responsibility – are required to appropriately determine overall criminal responsibility for result crimes. However, if causal findings are products of only moral-responsibility ascription, then this neglects a significant role in criminal-responsibility ascription – causal-responsibility ascription. Consequently, ‘the metaphysics of causation: INUS causation’ section examines the INUS account of causation as an alternative approach to the standard tests. It contends that INUS is the most rational metaphysical account of causation that English criminal law could adopt. It would therefore be a suitable replacement for the standard tests in English criminal law. Subsequently, ‘the utility of INUS causation in English criminal law’ section investigates the practical utility of adopting INUS causation in English criminal law. It argues that INUS causation would offer several practical benefits as it would facilitate engagement with causal enquires in a broader range of cases on a more principled, clear and consistent basis.
Causation in English Criminal Law: Ascribing Moral Responsibility
This section analyses the standard tests of causation in English criminal law. It argues that these tests represent a normative exercise in finding D responsible for a prohibited outcome, rooted in D's moral responsibility for that outcome. As a result, causal findings, especially in complex cases, are often grounded only in moral-responsibility ascription which gives rise to incoherence and uncertainty.
The Standard Tests of Causation
Factual Causation
Whilst the criteria for the standard tests are haphazard, the courts and legal commentators accept that the attribution of causation consists of two distinct requirements. The first requirement is a ‘factual’ enquiry into D's conduct and its link to the prohibited outcome which asks a counterfactual question: But for D's conduct, would the outcome have still occurred in the way that it did and at the time that it did? 8 If it is established that the prohibited outcome would not have occurred at the time and in the way it did, then typically there is an established causal link between D and the prohibited outcome. 9
However, the courts sometimes proceed to analyse legal causation even where factual causation is not established. For instance, in McKechnie, 10 D attacked V inflicting serious head injuries. V was hospitalised where it was first discovered he was suffering from an unrelated life-threatening stomach ulcer. However, due to the extent of V's injuries, the doctors decided not to remove the ulcer. The ulcer subsequently burst and V died ‘principally from the duodenal ulcer’. 11 D's conviction for manslaughter was upheld. Although the ‘but-for’ test is not established – because the ulcer bursting would have occurred anyway – the court considered that what was crucial was the reason for the ulcer not being able to be removed. 12 As the doctors’ decision not to operate was made specifically because of the injuries D inflicted, 13 D was therefore found to be a cause of V's death. So, whether the ‘but-for’ test is applied or skipped in such cases is the result of a normative judgment about D's responsibility for the outcome.
The ‘but-for’ test is over-inclusive since it generates a large number of spurious causes that might lead to convictions of the morally innocent. To give an example provided by the Supreme Court: ‘the woman who asked her neighbour to go to the station in his car to collect her husband would be held to have caused her husband's death if he perished in a fatal road accident on the way home.’ 14 The reason is that the husband would not have died but for the wife's request. English criminal law therefore refines the results from the ‘but-for’ test by requiring that D's conduct is also a ‘legal’ cause.
Legal Causation
The starting point for legal causation is that D's conduct must be a ‘substantial’ cause of the prohibited outcome. 15 But such conduct need not be the sole, or even main cause of the outcome, it being enough that D's conduct merely has to be a more than minimal (de minimis) contribution. 16 This is the main reason why the wife would not be considered a legal cause of her husband's death – her causal role would be regarded as too minimal or insignificant.
A particular problem with the de minimis test is uncertainty in how the courts determine ‘more-than-minimal’. This is because ‘more-than-minimal’ does not set a definable threshold. It may be that difficulties will not arise in straightforward cases involving one defendant, but the difficulty lies when unequal contributions are made by two potential defendants. Consider the situation where D stabs V, V is bleeding profusely, and X, a doctor, takes a sample of blood from V, hastening V's death by a few seconds. In such circumstances, the current application of the de minimis threshold holds that X's contribution is too minimal to be considered a cause of V's death. But what if X's role is not so innocent in such circumstances? Taking Simester's example, suppose now that there is motive for X to accelerate death because X stands to inherit from V's will. 17 Here, the courts may not be satisfied with a finding that X's conduct is too minimal to deny any sort of criminal liability in these circumstances. As Simester notes, ‘[a]pplication of the de minimis rule to this kind of case is itself normatively sensitive’. 18 In other words, whether or not the de minimis test is applied in such cases is susceptible to normative decisions about D's responsibility for the prohibited outcome.
The courts have determined that D's conduct must also be an ‘operating’ cause. 19 This means that there must have been no new intervening acts (novus actus interveniens) between D's conduct and the prohibited outcome. Under this rule, either the free, voluntary and informed, 20 or reasonably unforeseeable, 21 intervention of another person or natural event is regarded as a ‘break’ in the causal chain that would have existed between D's conduct and the prohibited outcome, thereby relieving D of any causal relationship with that outcome.
However, the free, voluntary and informed (or ‘voluntary act’) and ‘reasonable foreseeability’ tests are in tension. For example, Kennedy (No. 2) held that despite D (a drug dealer) preparing a syringe for V, who was a drug addict, D did not cause the death of V as V injected himself with drugs. V broke the chain of causation as it was a voluntary act. Yet it was certainly reasonably foreseeable on the facts of the case that V would inject himself with the prepared syringe (given that V was a drug addict), so as to not break the causal chain. Because of such tensions in the novus actus interveniens tests, the courts tend to apply them inconsistently. For instance, courts sometimes deem themselves free to pick and choose which rule they think best suits the facts of the case. 22 Occasionally, judges simply neglect to consider any of the novus actus tests. 23 So, in certain cases, considerations of which novus actus rules are to be applied, or decisions as to whether they are skipped entirely, are normatively sensitive.
Another requirement of legal causation is that D must also be a ‘blameworthy’ cause. That is, D's blameworthy – and not, say, accidental – conduct must cause the outcome. 24 In Hughes, D was involved in a fatal road collision with V who was oncoming in the opposite lane. Notwithstanding D's faultless driving, he was uninsured and unlicensed. It is an offence Contrary to s.3ZB of the Road Traffic Act (RTA) 1988 to cause death merely by driving while uninsured or unlicensed. As Sullivan and Simester remark, ‘[t]his offence is a stark example of strict liability. No culpability of any kind on the part of D is required in the matter of causing V's death’. 25 In this case, D is a factual cause of V's death: But for D driving his vehicle, V's death would not have occurred in the way and time that it did. D's conduct is also a substantial and operating cause of death since V died from injuries that resulted from a road collision in which D was involved. However, the Supreme Court held that the causal component in s.3ZB requires at least some act or omission which involves some element of fault on behalf of D. Since there was nothing in the manner of D's driving that was at fault that contributed in any way to the death of V, the conviction was therefore quashed.
A concern with the blameworthy test is that it creates a tension between causation and strict-liability result crimes. 26 This is because it injects blameworthiness into these types of offences which typically do not require any type of culpability. So, in strict-liability result crimes, the courts are having to juggle conflicting principles. Whether or not the blameworthy test is applied in those cases will inevitably be the result of a normative judgment about D's conduct.
Causation Grounded in Only Moral Responsibility
The standard tests reveal that there are no coherent or consistently applied criteria for determining causation. This is because, rather than being a factual causal enquiry, causation in English criminal law is a normative exercise in finding D responsible for an outcome – it concerns an enquiry, through D's conduct, into whether D should be held responsible for a proscribed outcome.
Such a normative exercise is rooted in determining D's moral responsibility for the prohibited outcome. This is most apparent in the ‘blameworthy’ test which partly grounds legal causation. Here, the courts impute onto causation questions of moral blameworthiness which essentially blurs the legal distinction between actus reus and mens rea. Elsewhere, in relation to novus actus interveniens, the ‘voluntary act’ rule is grounded in the idea that an individual should not be blamed for what is caused by another's intervening voluntary actions;
27
whilst the ‘reasonable foreseeability’ test is rooted in D's blameworthiness. This is because the test looks at whether D – based on what would have been foreseen by the reasonable person – should have foreseen that such an outcome would result from his conduct, and foresight is a mental state associated with moral responsibility. Williams confirms that these rules of legal causation have developed because the courts are reluctant to convict someone of a result crime when they are not blameworthy. He writes: When one has settled the question of but-for causation, the further test to be applied to the but-for cause in order to qualify it for legal recognition is not a test of causation but a moral reaction. The question is whether the result can fairly be said to be imputable to the defendant … If the term cause must be used, it can best be distinguished in this meaning as the ‘imputable’ or ‘responsible’ or ‘blameable’ cause, to indicate the value-judgment involved.
28
Furthermore, since the ‘but-for’, de minimis and novus actus interveniens tests admit of normative decision-making, the decision when and how to apply them is also susceptible to decisions based on the perceived blameworthiness of D. 29
Since the standard tests are rooted in D's moral responsibility, causal findings in English criminal law are therefore prone to determinations of only moral-responsibility ascription. This is highlighted in the more complex result crimes – those where there are unusual or complicated facts, such as where another's conduct or another natural event also plays a potential causal role in the prohibited outcome. A useful illustration is the court's approach to situations where V commits suicide following an attack from D. In Wallace, 30 D threw corrosive acid over V and, due to the extent of V's injuries, V subsequently died in Belgium by lawful euthanasia. The trial judge withdrew the charge of murder on the basis that causation could not be established between D's conduct and V's death. However, in allowing the appeal, the Court of Appeal held that a jury was entitled to find that D's initial attack caused V's death. According to the Court, there was no potential intervening act as V's decision to commit suicide was not voluntary ‘given the truly terrible situation he was in’. 31
However, criminal law is not generally interested in the reasons or motives for someone's conduct.
32
Therefore, regardless of V's reasons in Wallace, V made what appeared to be a voluntary decision to die. In addition, the doctors, in carrying out the euthanasia, also made voluntary decisions to act on V's resolve to die and suicide was certainly unforeseeable on the facts, both voluntariness and unforeseeability constituting novus actus interveniens.
33
But according to the Court, V's suicide was: a response by a victim to (extreme) circumstances created by [D's] unlawful act, which were persisting, and which had put [V] into a position where he made a ‘choice’ that he would never otherwise have had to make or would have made.
34
The court therefore clearly blamed D for V's death. As a result, it controversially tightened the notion of voluntariness in order to ascribe liability to D for the death of V. However, while D might have been morally responsible for V's death, D was not a legal cause of it. 35 Indeed, V died from the lethal injection by the doctors. The initial injuries inflicted by D were not a concurrent operating cause of that death. 36
Criminal Law Inappropriately Ascribing Criminal Responsibility
The previous section demonstrated that causal findings in English criminal law, especially in complex cases, are grounded purely in moral responsibility. Such an approach to causation gives rise to incoherence and uncertainty. However, this section demonstrates why that approach to causation is problematic in the context of criminal-responsibility ascription. It contends that moral responsibility is irrelevant to causal responsibility. Therefore, issues of moral responsibility should be kept separate from causal enquiries. Distinguishing causal responsibility from moral responsibility is important in criminal law because both notions are required to appropriately determine overall criminal responsibility for result crimes: Causal responsibility establishes D's actual wrongdoing, whereas moral responsibility determines D's blameworthiness for his wrongdoing. However, if causal findings are products of only moral-responsibility ascription, then this neglects a significant role in criminal-responsibility ascription – causal-responsibility ascription. As a result, there is inappropriate criminal-responsibility ascription for result crimes.
The Irrelevance of Moral Responsibility to Causal Responsibility
In the philosophical literature, causal responsibility is attributed to an individual for an outcome when that individual's conduct causes the outcome. 37 According to Satorio, ‘A is causally responsible for O … only if one of A's actions or omissions caused O’. 38 Legal theorists also agree that individuals are causally responsible for the outcomes that they cause. 39 To give an example of causal responsibility, consider a situation where R, after pushing open a door, accidentally knocks over P's vase that P placed too close behind the door. The vase breaks. Here, we would not say that R is morally responsible for breaking the vase because his actions do not attract any sort of resentment or indignation – he was not aware of the vase being behind the door. But we would nevertheless say that R – via his conduct – is causally responsible for breaking the vase. This is simply because R's action of opening the door and knocking over the vase is, metaphysically speaking, an actual cause of the broken vase.
Philosophical studies suggest that on common-sense views about causation, moral responsibility can have an impact on causal judgements. 40 According to such studies, morally responsible agents are considered more of a cause compared to those agents who are not morally responsible. English criminal law reflects such common-sense views about causation. When dealing with questions of causation, the courts have often declared that it is common-sense notions of causation that are to be applied to the case. For instance, in Alphacell Ltd v Woodward, Lord Salmon said: ‘what or who has caused a certain event to occur is essentially a practical question of fact which can best be answered by ordinary common sense’. 41 In their seminal work on causation in criminal law, Hart and Honoré's fundamental argument is that courts are correct to claim that it is common-sense notions of causation with which criminal law generally seems to be concerned. 42 According to them, the central view of causation, as used in ordinary speech, is that of a causal relation between events (i.e., of ‘causing’ or ‘bringing about’ an effect). 43 On this ordinary person's view of causation, Hart and Honoré claim that counterfactual considerations are prominent in assessing what counts as ‘causes’. They also make the claim that, in ordinary speech, interventions between an ‘original’ cause and its effect by factors such as voluntary human acts and abnormal events will typically negate the causal link from the original cause. 44 Because such views are also adopted in English criminal law, this leads Hart and Honoré to conclude that English criminal law is concerned with common-sense notions of causation.
However, common-sense views of causation are deficient because there are clear-cut situations where moral responsibility is irrelevant to causation. Take the vase example. R's conduct in opening the door and knocking over the vase is a cause of P's broken vase despite R not being blameworthy. But if causation did require moral responsibility, then that possibility would no longer exist and R's conduct would not be a cause of the broken vase since it involved no element of blame. That conclusion should be rejected. As Simester states, [b]lameworthiness arises not from doing the wrong thing, but from how D engages with the reasons why she ought not to have done it’. 45
Natural events also cause outcomes. As Miller says, ‘[h]uman agency has no special status when causal responsibility is being allocated’. 46 If a building fell down because of an earthquake, for instance, we would generally say that the earthquake is responsible for the falling of the building. Hart explains that in such contexts, we sometimes substitute the expression ‘responsible for’ for the words ‘caused’ or ‘produced’, or some other causal expression in referring to causally related outcomes. 47 The idea that causation involves natural events, such events being irrelevant to morality, reinforces the view that causal responsibility does not have a moral basis.
As Tadros points out, ‘causal responsibility is concerned with the proper explanation of events in the physical world’. 48 It gives us an answer to the question, ‘why did E occur?’. So, when it is asked: ‘why did P's vase break?’, we cite R as being responsible (or ‘the reason’) for it breaking simply because it was R's conduct that caused the vase to break. It is causation that provides an explanation as to why E occurred; nothing else. Since moral responsibility does not matter to causal responsibility, issues of moral responsibility should therefore be kept apart from causal enquiries. As such, the issue of Alec Baldwin's moral responsibility should have no bearing on his causal responsibility for the death of Halyna Hutchins. Indeed, Baldwin's relationship to Hutchins’ death should be characterised as causal despite his lack of moral responsibility for her death – his conduct is not open to criticism as he was rehearsing a scene and was unaware of the live bullet in the prop gun.
Why Distinguishing Causal Responsibility From Moral Responsibility Matters in Criminal Law
Distinguishing causal responsibility from moral responsibility matters in criminal law because each notion plays an important role in ascribing criminal responsibility for result crimes.
Retributive notions of punishment focus on the demands of justice in that offenders get their ‘just deserts’. Essentially, offenders are punished because, and only because, they deserve it. 49 The dominant view of retributive justice is that D deserves punishment for his culpable wrongdoing – wrongdoing for which D is both accountable and blameworthy. 50 This view has been accepted by many scholars and has led to claims that the shape and structure of criminal law is, and indeed should be, informed by a core moral thread – culpable wrongdoing. Criminal responsibility, is thus – paradigmatically – attributed to D for his culpable wrongdoing; even if, for instrumental reasons (e.g., deterrence; efficiency), criminal responsibility sometimes targets non-culpable wrongdoing, such as through strict liability crimes. 51
For culpable wrongdoing, the ‘culpability’ requirement reflects the fairly uncontroversial view amongst criminal-law scholars that criminal punishment should be deserved – wrongdoing for which D is blameworthy. But the ‘wrongdoing’ requirement requires identification of the wrong that D actually committed – it requires D's accountability for the actual wrongdoing itself. 52 It is this wrongdoing requirement where causation (thus causal responsibility) plays a significant role in criminal-responsibility ascription for result crimes.
Causation is important in identifying D – via D's conduct – as an author of the prohibited outcome in result crimes. The function of ascertaining ‘who killed V?’ is a starting point in establishing the appropriate defendants: D is linked to V by the fact that D's conduct actually caused V's death. Here, it is causation that links D, through D's conduct, to V's death. The identification of defendants is not a matter of moral responsibility. To be sure, consider Simester's ‘South-side’ example where D1 attacks V leaving V unconscious and dying from blood loss in a dangerous area of the city. But before that occurs D2 comes upon V, who is a member of a rival gang, and deliberately shoots V dead. 53 Here, D2 actually kills V and D1 does not. From a moral responsibility perspective, both D1 and D2 may be equally culpable characters. But from the perspective of causation, only D2 is the author of V's death, through D2's conduct. In this situation, the status of D2 is very different from that of D1, for causal reasons that have nothing to do with the culpability of their behaviour. When D2 shoots and kills V, establishing causation between these two events gives us an answer to the question, ‘who killed V?’. It connects V's death to D2, the culpability of which comes later. 54
In criminal law, the identity of D as an author of the prohibited outcome is important for many reasons. Identifying ‘who killed V’ recognises that we are separate agents who are able to act in and upon the world. 55 From a practical view, failing to distinguish causers from non-causers would cast the net of criminal liability too wide. Without this distinction, all other things being equal, the existence of a prohibited outcome would have the potential to involve anyone who is to blame for its occurrence as a possible defendant. Rule-of-law principles also become significant, since criminal liability may become too unpredictable if anyone who is to blame for a prohibited outcome becomes exposed to criminal conviction. 56
But the most important reason for identifying D as an author of the prohibited outcome relates to D's accountability for his wrongdoing in result crimes. These crimes require that D's conduct cause a prohibited outcome. Many of the legal provisions for these crimes are framed explicitly in terms of D's ‘causing’ a prohibited outcome. For example, contrary to s. 1 of the Road Traffic Act 1988 it is an offence when ‘[a] person…causes the death of another person by driving a mechanically propelled vehicle dangerously…’. However, sometimes the legal provisions governing result crimes do not use the term ‘cause’. Instead, they use so-called ‘causative verbs’. For instance, murder is defined as killing, not causing death. 57 Nevertheless, criminal law treats such causative verbs as equivalent to causation. For example, killing is treated as equivalent to ‘causing death’. However, unlike conduct crimes, result crimes do not usually stipulate the conduct that is required to cause the proscribed outcome. For instance, shooting; stabbing; poisoning; etc., can all suffice for murder. Instead, result crimes typically appear to be specified solely by reference to causing a particular outcome which can be brought about through various forms of conduct. So, for result crimes, causing a prohibited outcome is the pro tanto wrong that concerns criminal law. Identifying D as an author of the proscribed outcome captures the causal connection between D's conduct and the prohibited outcome. Therefore, identifying D as an author of the prohibited outcome actually establishes D's accountability for that pro tanto wrong.
Thus, both causal responsibility and moral responsibility are required to appropriately determine overall criminal responsibility for result crimes: Causal responsibility establishes D's actual wrongdoing, whereas moral responsibility determines D's blameworthiness for his wrongdoing. However, if causal findings are products of only moral-responsibility ascription, as numerous cases in English criminal law suggest, then this merely tells us why D should (or should not) be held liable for a prohibited outcome – not whether D's conduct was the actual cause of that outcome. As a result, such causal conclusions do not establish D's accountability for his actual wrongdoing. So, by losing sight of what is happening in the real world, English criminal law's approach to causation neglects a significant role in criminal-responsibility ascription – causal-responsibility ascription. It therefore results in inappropriate criminal-responsibility ascription for result crimes.
The Metaphysics of Causation: INUS Causation
The previous section established that English criminal law's approach to causation can result in inappropriate criminal-responsibility ascription for result crimes. The problem stems from normative causal findings being grounded only in moral-responsibility ascription. Therefore, it is preferable that English criminal law adopts a single, non-normative test of causation; a metaphysical one that offers a robust causal enquiry that focuses only on causal responsibility, which contributes appropriately to criminal-responsibility ascription. As such, this section proceeds to analyse the INUS account of causation as an alternative approach to the standard tests. It argues that INUS causation is the most rational metaphysical account of causation on offer and would therefore be a suitable replacement for the standard tests.
The INUS Account of Causation
Outside the law, our understanding of causation is rooted in metaphysics. Metaphysics is the branch of philosophy that wonders about what exists in the universe and what existing things are like, in their most fundamental features and interrelationships.
58
INUS causation is a leading metaphysical account of causation, propounded by Mackie.
59
According to INUS causation, a cause (C) is an INUS condition: An
Most of the criteria for INUS causation are redundant. The test insists that the causal factor under consideration must be an insufficient NUS. However, if an insufficient (but NUS) factor can qualify as a cause, then it stands to reason that so can a sufficient (and NUS) factor. If anything, a sufficient and NUS factor will be a stronger candidate for causation. So perhaps the ‘
On this reading, the INUS (or rather, the ‘NS’) account is similar to Wright's ‘NESS’ (Necessary Element in a Sufficient Set) test of causation. 60 According to the NESS test, an actual condition C is a cause of an actual condition E if and only if C is a necessary part of the set of conditions sufficient for E. 61 Wright claims that not only does NESS define the concept of actual causation, it also captures the underlying common intuitions and judgements about causation. 62 The NESS test was established in response to difficulties in causation in tort law in the United States and has since been developed and received support in tort-law literature in several jurisdictions. 63 Given that INUS causation, reduced to the NS test, is essentially the same as the NESS account of causation, most of the analysis of INUS causation in the proceeding sections draws upon that tort-law literature.
The tests inherent in INUS consist of necessary and sufficient conditions. But what is crucial to the success of INUS is the type of sufficiency on which it relies. If X is lawfully sufficient for Y, it means that X's presence guarantees the existence of Y by some law of nature. 64 For example, fire is lawfully sufficient for oxygen because the presence of fire guarantees the presence of oxygen (since fire requires oxygen). The problem with this, however, is that lawful sufficiency does not necessarily indicate a causal relationship. For instance, a barometer needle falling guarantees the existence of a storm but the falling of barometer needles do not cause storms. Storms are caused mainly by unstable air pressure in the atmosphere. As Strevens states in his revision of Mackie's INUS account, ‘it is not enough that a set of conditions be [lawfully] sufficient for the occurrence of e; it must be causally sufficient for e’. 65
Causal sufficiency is a distinct type of lawful sufficiency, according to which C will be sufficient to cause E. Thus, although a barometer needle falling is lawfully sufficient for the existence of a storm, it is not causally sufficient for a storm. Wright employs the term ‘causal sufficiency’ for an identical concept of sufficiency in the NESS account.
66
He elaborates on its definition: ‘causal sufficiency’ … is the instantiation of all the conditions in the antecedent … of a causal law, the consequent … of which is instantiated by the consequence at issue. A causal law is an empirically derived statement that describes a successional relation between a set of abstract conditions … that constitute the antecedent and one or more specified conditions of a distinct abstract event or state of affairs that constitute the consequent such that … the instantiation of all the conditions in the antecedent entails the immediate instantiation of the consequent, which would not be entailed if less than all of the conditions in the antecedent were instantiated.
67
In other words, causal sufficiency requires the complete instantiation of all the factors in sufficient condition. So, these factors need to consist of facts that actually occur and come into play in the causal process.
Invoking the notion of causal sufficiency has, however, been criticised for rendering causation ‘viciously circular’, since causal terminology is used in its elaboration and relies upon the concept of causal laws. 68 Essentially, causation relies on a notion of sufficiency that is itself predicated on a certain understanding of causation, which would effectively amount to defining ‘causation as causation’. 69 However, the definition of the causal law noted in the previous paragraph – that it be a statement that describes a successional relation between distinct abstract events – does not contain any causal language or related suppositions. It is merely a statement that describes a relation between C and E which can be established by empirical observation and experience. This is because we can empirically observe causation in action by seeing if the non-occurrence of a supposed causal factor actually makes a difference in the occurrence of the outcome. 70 For instance, we can determine by experimentation or observation that eliminating the presence of oxygen eliminates the occurrence of fire. Furthermore, the notion of causal sufficiency is defined non-circularly as the complete instantiation of all the factors in the sufficient condition. Consequently, there is no conceptual circularity in defining causation, vicious or otherwise. 71
INUS is arguably the most rational metaphysical account of causation because it captures common causal intuitions and judgments in key situations where other main metaphysical accounts fail to do so. 72 To demonstrate this, INUS, along with two other leading metaphysical accounts of causation – namely, the ‘counterfactual’ and ‘process’ accounts – will now be analysed, respectively, against two causal situations: Cases of overdetermination and circumstances involving omissions. 73
Overdetermination
In their purest form, overdetermination cases involve situations where more than one event brings about another event, but the events are each sufficient on their own to bring about the result. It is clear that INUS can detect causation in overdetermination cases. Consider the Unites States case of Jones v Commonwealth, 74 where D1 and D2 simultaneously shoot V and their respective bullets are each fatal. In this situation, there are two sufficient conditions for V's death. Condition 1: D1 shooting the gun, V being present in proximity to D2, and the fatal physical bodily processes that occur as a result of the gunshot to V's head; and Condition 2: D2 shooting the gun, and the position of V etc. In condition 1, D1's gunshot is necessary for death: If D1 had not fired his gun, death would not have occurred. Similarly, in condition 2, D2's gunshot is necessary for death in the same way. Hence, D1 is an INUS condition of death, and so is D2.
The best-known counterfactual analysis of causation is developed by Lewis. 75 According to a counterfactual account of causation, C causes E when E counterfactually depends on C. That is, if C had not occurred, E would not have occurred at the time and in the manner in which it did. Evidently, this counterfactual account of causation is synonymous with the ‘but-for’ test of factual causation currently used in English criminal law. Because the ‘but-for’ test adopts a metaphysical notion of causation, it is said to be the only truly causal component of the law's two requirements. 76 However, cases of overdetermination present problems for counterfactual accounts. In Jones v Commonwealth, it seems that V's death does not counterfactually depend on either D1 or D2 firing their guns: If D1 had not fired his gun V would still have died at the same time and in roughly the same way (because of D2 firing his gun), and if D2 had not fired his gun V would still have died at the same time and in roughly the same way (because of D1 firing his gun). Consequently, we get the result that neither D1 nor D2 is a cause of death.
In trying to deal with this difficulty, Lewis asserts that even if D1 is not a cause and D2 is not a cause, the mereological sum of D1 and D2 is a cause of V's death. 77 But taking the mereological sum of D1 and D2 as a cause does not mean that we are taking D1 as a cause, and D2 as a cause. This result is inexplicable. In the context of overdetermination, it is unclear why D1 and D2 jointly cause V's death while neither event alone is a cause. Lewis’ manoeuvre, then, fails to match our natural intuitions about causation in cases of overdetermination. 78 In criminal law, the courts would simply ignore applying the ‘but-for’ test in such cases. So, D1 and D2 will both be held to have legally caused V's death on criteria other than satisfying the ‘but-for’ test. 79
Being unable to properly handle overdetermination cases, along with the issue of over-inclusiveness, are the main reasons counting against a counterfactual analysis as a rational metaphysical account of causation.
Omissions
Omissions are situations that involve an absence or failure of some sort of act or event. Intuitively and legally speaking, omissions cause things. For instance, a parent who fails to feed their child is considered a cause of the child's death if the child died of starvation. 80 Similarly, if a patient (P) needs medical treatment to prevent her being seriously ill, but the doctor (R) actively withholds that treatment and P falls seriously ill, R's omission is a cause of P's illness.
The INUS account accommodates omissions as causes. Take the example of R withholding P's medical treatment for, say, diabetes. In such a case, a condition sufficient to bring about P's illness, say, heart disease, would be the conjunction of her underlying medical condition, i.e., the diabetes, and R's refusal to provide medical treatment for her diabetes. In this condition, D's refusal to provide medical treatment is necessary: If R had not refused to provide medical treatment then the heart disease would not have developed (because the treatment would have prevented it). Hence R's omission is an INUS condition of heart disease. 81
Some may argue that the failure to provide medical treatment by everyone else in the world, such as X, a passer-by, may also constitute a cause. So, in cases involving omissions, a potential difficulty for INUS arises in that there are superfluous causes. If INUS merely required lawful sufficiency it would incorrectly treat the omissions of everyone else as duplicative causes of P's death. This is because the omissions of everyone else, along with all the other relevant factors (i.e., diabetes), guarantees the heart disease and thus is lawfully sufficient for its occurrence (abstractly derived). However, for INUS, we require causal sufficiency rather than mere lawful sufficiency. Negative conditions, such as omissions, often cause positive conditions (i.e., heart disease) through ‘negative causation’, by preventing a positive causal process (i.e., medical treatment) that would have averted the occurrence of the positive condition. 82 The omission by R is a negative cause that is fully instantiated since there is an actual absence of medical treatment for diabetes that could have been provided to P and which would have prevented the heart disease. This omission is therefore part of a causally sufficient condition. However, the omissions by everyone else in the world, as negative causes, are not instantiated. This is because everyone else is unaware of P's medical condition, and lacks the proximity (in terms of space, time, knowledge, expertise, resources, etc.) to provide medical treatment for the patient and thus prevent the heart disease from actually occurring. So, unlike R's omission, the omissions by everyone else do not, and indeed cannot, prevent the interference (i.e., medical treatment), which would have jeopardised the heart disease. These omissions, therefore, do not play a causal role and so do not form part of a causally sufficient condition.
Nevertheless, some superfluous causes remain. For instance, an individual watching an aid appeal on television for starving children and not donating when he could have done so might plausibly constitute an INUS condition of a child's death if that death occurred from starvation. However, in English criminal law, like most jurisdictions, liability for omissions is limited to those who have a legally recognised duty to act. 83 It is this duty that restricts the focus of criminal law from all those who failed to act, when they could have done so, to those few whose omissions are considered criminal. So, whilst it may be true that the person watching the aid appeal would be a negative cause of the child's death, he would not be held criminally liable for a homicide offence since there was no duty to act. Therefore, although INUS remains theoretically over-inclusive for some omissions, it is not a practical concern for English criminal law. More importantly, because INUS can deal with omissions as causes, it can detect causation in those omissions that concern English criminal law – those which require a duty to act.
Process accounts of causation are more mechanical in that causation is seen in terms of the transfer of some sort of continuing physical process from C to E. 84 Consider a simple example where D shoots V dead. Supporters of process accounts would claim that D's conduct is a cause of V's death because a sequence of physical processes can be traced from D pulling the trigger, the gun discharging, to the bullet puncturing V's body, killing V. Various criminal-law scholars also subscribe to this account of causation. 85 However, for process accounts, the problem with omissions is obvious. There is no set of continuing physical processes or interactions linking cause and effect where omissions are involved. For example, there are no physical processes between R's failure to withhold medical treatment and P's illness. So, in these cases, process accounts would arrive at the incorrect conclusion – they would claim that R's omission had nothing to do with P's illness. Philosophers and criminal-law theorists who acknowledge this difficulty end up resorting to counterfactual analyses to accommodate omissions as causes. 86 But this approach highlights that process accounts require another account of causation to do the causal work.
Those criminal-law scholars who endorse process accounts concede they are generally incomplete and, in addition to omissions, cannot account for causes in other situations in which they ought to. 87 To use Simester's example, D throwing V in front of a train, or tying them to the track, does not cause V's death: We cannot trace a continuing sequence of physical processes from D's conduct to V's death; it merely creates the opportunity for the train to do so. For Moore, pure physical causation is all the causation there is. So, Moore would accept such a conclusion. Yet D's relationship to V's death seems rightly characterised as causal. Simester, therefore, invokes the notion of ‘indirect’ causation to account for such situations. 88 Indirect causation is essentially where two continuing physical process chains are bridged together through moral principles, such as foreseeability and voluntariness. Thus, D would be considered an indirect cause of V's death since the latter continuing physical process chain (i.e., the train hitting and killing V) is a reasonably foreseeable occurrence from the earlier continuing physical process chain (i.e., D throwing V in front of the train). However, the problem with indirect causation is that it contradicts the very essence of process accounts: It ascertains causation even though there are no continuing sequence of physical processes from C to E. Again, it is effectively another account of causation that supplements process accounts.
Ultimately, unlike counterfactual and process accounts, the INUS account detects causation in both overdetermination cases and omissions. It is the most rational metaphysical account of causation available and should therefore be adopted in English criminal law.
The Utility of INUS Causation in English Criminal Law
The previous section argued that INUS is the most rational metaphysical account of causation on offer and should therefore replace the standard tests of causation in English criminal law. This section investigates the practical utility of adopting INUS causation in English criminal law. It argues that the ability of INUS to explain causation rationally is crucial to its practical utility in criminal law. INUS would be able to engage with causal enquires in a broader range of cases on a more principled, clear and consistent basis.
Practical Benefits
Compared with the standard tests of causation in English criminal law, what makes INUS particularly beneficial is its ability to achieve causal conclusions in a broader range of cases on a more rational basis. With INUS, the courts would also be required to confine their causal analysis to a single non-normative enquiry: Whether or not C is a necessary part of a condition sufficient for E, instead of its current convoluted bifurcated approach. Unlike the standard tests, determinations of whether or not particular factors were the causes would be relatively unambiguous since INUS is restricted to empirical concepts, namely necessity and sufficiency. There would also be no need to depart from INUS or implement additional vague criteria to supplement and attenuate INUS as it would be able to engage with causal enquires satisfactorily. As a result, the law would be clearer and applied more consistently which would enhance the principle of ‘fair warning’: i.e., that citizens should be advised if their conduct will be censured so that the law embodies clarity and certainty. 89
Indeed, as previously discussed, unlike the ‘but-for’ test, INUS causation would be able to establish causation in cases of overdetermination such as Jones v Commonwealth. In establishing causation in these cases, the courts currently ignore the ‘but-for’ test. Ormerod and Laird state that ‘[e]ven this basic rule may have exceptions, but only in very unlikely circumstances’. 90 However, the problem with this is that there are no formal exceptions to the ‘but-for’ test. As a result, the courts are relying on arbitrary and ill-defined exceptions to a causal criterion. In terms of consistency and clarity, it would be preferable to analyse overdetermination cases (along with all the other types of causal scenarios) according to INUS rather than relying on such exceptions to the ‘but-for’ test. Furthermore, the causal conclusions drawn from cases that fail to meet the ‘but-for’ test are illogical: Causation is found despite not satisfying a legally recognised causal criterion. Because INUS would be able to handle overdetermination cases without deviating from its causal criteria, INUS would therefore also establish causation in these types of cases on a more rational basis.
INUS would be similarly useful in cases involving pre-emptive causation. Pre-emptive intervening causes are causes that pre-empt some other event from doing its causal work. Recall the South-side example above, D1 attacks V leaving V to die in a dangerous part of the city. But before that occurs D2 comes upon V, who is a member of a rival gang, and deliberately shoots V dead. The risk that this would occur was reasonably foreseeable and, indeed, foreseen by D1. 91 In such cases, what we are saying is that the first event did not cause the outcome, the second one did. 92 The second event pre-empted the first. Again, the ‘but-for’ test is unable to properly handle cases of pre-emption. It would establish that neither D1's nor D2's conduct is a factual cause of V's death since neither conduct is necessary for death to occur. In such cases, the courts, like in McKechnie, would simply ignore the ‘but-for’ test. As a result, criminal law typically deals with pre-emptive intervening causes through the novus actus doctrines.
As with cases involving overdetermination, INUS causation can also adequately handle pre-emption cases. Consider, again, South-side. According to INUS causation, D2's conduct is a cause of V's death but D1's conduct was not. In such a case, a condition sufficient to bring about V's death would be the conjunction of D2 shooting the gun, V being present in proximity to D2, and the fatal physical bodily processes that occur as a result of the gunshot to V's head. In this condition, D2 shooting the gun is necessary: If D2 had not fired the gun, death would not have occurred at the time and in the way it did. Hence D2 shooting the gun is the cause of V's death.
Some may argue that the INUS account renders both D1's and D2's conduct as causes. 93 But this criticism erroneously assumes that INUS requires lawful sufficiency, rather than causal sufficiency. 94 If INUS causation merely required lawful sufficiency, then it would incorrectly treat D1's conduct as well as D2's conduct as a cause of V's death. This is because D1's conduct, along with all the other relevant factors (i.e., fatal blood loss; lack of medical intervention; etc.), is part of a sufficient condition that would guarantee V's death (abstractly derived). However, for INUS, causal sufficiency rather than mere lawful sufficiency is required. The causal generalisation and thus the underlying causal law for V's death by D1's attack includes fatal blood loss, a factor that was not instantiated since V died before the loss of blood occurred. On the other hand, the causal law for death by shooting, which includes as a necessary factor the gunshot to V's head (and the fatal physical bodily processes that occur as a result of the gunshot to V's head, etc.), were fully instantiated. Thus, D2's shooting, but not D1's attack, is the cause of V's death. D1 would only have constituted a cause if V died from loss of blood from the assault – but V did not.
Compared with the novus actus doctrines, INUS would also be able to handle pre-emption cases on a more rational basis. To see this, South-side will now be contrasted with a slight variation of itself. Recall that in South-side, INUS causation correctly detects only D2's conduct as a cause of V's death. The novus actus doctrines will hold the same causal conclusion as D2's conduct constitutes a voluntary act which thereby breaks the causal chain between D1's conduct and V's death. 95 However, if the South-side example was altered slightly, so that at the time D2 fired his gun he was insane, then problems start to emerge for English criminal law. Although INUS would still detect D2's conduct as a cause of V's death (the mental state of D2 is irrelevant to the causal enquiry), the current novus actus rules would establish, instead, that D1's conduct was a cause of V's death. This is because D2's conduct is an involuntary (and foreseeable) intervention that would not pre-empt D1's attack as a cause. 96 It is an involuntary intervention because D2 lacks the capacity to make a reasoned and informed decision about his conduct. 97 But this would be a difficult causal finding because the attack exerts no influence whatsoever in the way that V actually dies. So, the only way the courts can find D2's conduct as a cause of V's death is if it deviates, again, from legally established causal criteria which is undesirable. Of course, it may be true that even though D2's conduct causes V's death, D2 is not to blame for it, or otherwise should not be held criminally liable for a homicide offence. But if so, this conclusion should result from considerations of doctrines other than causation. 98 Ultimately, because INUS would be able to handle pre-emption cases rationally, the application of INUS to these types of cases would ensure that the causal analysis is more consistent since there is no need to deviate arbitrarily from established causal criteria.
Not only would INUS be able to engage with causal enquires in a broader range of cases on a more principled basis which ensures consistency and certainty, but it would also promote clarity in the law. This is mainly because INUS does not use ambiguous concepts or incompatible tests. For instance, in relation to the novus actus tests, the courts often focus its causal enquiry on the vague notion of ‘voluntariness’. In Kennedy (No. 2), it was held that despite V being a drug addict, his self-injection of heroin was a voluntary act (thereby breaking the causal chain between D's conduct and V's death). Yet it is difficult to reconcile this with the apparently involuntary act of V committing suicide in Wallace (which did not break the causal chain between D's conduct and V's death). Such an ambiguous notion, therefore leads to controversial outcomes. However, analysing cases like Kennedy (No. 2) and Wallace in terms of INUS would not only arrive at the appropriate liability findings, but it would do so in a clearer and more consistent manner. In contrast to the concept of voluntariness, the meaning of which ‘cannot be easily pinned down’ since it can be approached from either a narrow or broad perspective, 99 it would not be so easy for the courts to differ on what constitutes necessity or sufficiency as those concepts are more objective and therefore relatively clear-cut. If INUS had been the test of causation in English criminal law at the time of Wallace, the Court would simply have combined the doctor's administration of the lethal drug to V with the associated fatal effects of that drug to V's vital bodily organs to construct a sufficient condition for V's death. The doctor's administration of the drug would then have been held an INUS condition of death, instead of D's conduct (which does not form part of any condition sufficient for V's death). As a result, INUS would have generated a different, and more appropriate, liability outcome in Wallace. 100 But the same outcome would have also been established in Kennedy (No. 2) for similar reasons: D's conduct does not form part of any condition sufficient for V's death – the condition sufficient for V's death is the conjunction of V's self-injection of heroin with the associated fatal effects of that drug.
Furthermore, due to the incompatibility of the ‘voluntary act’ and ‘reasonable foreseeability’ novus actus tests, it is unclear how causation will be decided in novel cases when there is an intervening act that is, say, both voluntary and reasonably foreseeable. Currently, in line with Girdler, the courts deem themselves free to pick and choose which rule they think best suits the facts of the case. 101 However, INUS would not have this problem since it is a single non-normative test and would therefore not provide the courts room to arbitrarily shape the causal enquiry.
The de minimis test is also ambiguous since it does not set a definable threshold. As previously discussed, a particular problem with the de minimis test is uncertainty in how the courts determine ‘more-than-minimal’, particularly when unequal contributions are made by two potential defendants. Recall the example of the doctor extracting blood from V, who had been stabbed, hastening V's death. Here, it is relatively unclear whether or not the doctor is a ‘more-than-minimal’ cause. It may depend, in part, on the doctor's motive in taking the blood sample. However, INUS would make clear that causation is established in such cases without recourse to such moral considerations. INUS would make the doctor's blood extraction causally relevant by combining it with the blood loss from V's stab wound to make a sufficient condition for V's death. 102
Although INUS would be able to engage with causal enquires in a wider range of cases on a more consistent and clearer basis, it would have implications for liability outcomes in some cases. As noted above, INUS would affect liability outcomes in cases where D is morally responsible for the prohibited outcome but is not causally responsible for it, such as Wallace. However, it would also have implications for liability outcomes in strict-liability result crimes where D is not morally responsible even though D is causally responsible for the proscribed outcome, such as Hughes. Here, D was acquitted because his blameworthy conduct was not a cause of V's death. However, in this case, D is a cause of death of V on the INUS account. A condition sufficient to bring about V's death is the collision which involves V's driving on the road and D's driving on the road. In this condition, D's driving on the road is necessary: If D had not driven the car and had not been in that position at the time, V's death would not have occurred. Hence D's driving is an INUS condition of V's death. Therefore, INUS would have rendered Hughes liable for the offence contrary to s.3ZB RTA 1988.
Of course, a liability finding in Hughes would have been unwelcome. D was driving faultlessly and was not to blame for V's death. In fact, the victim was to blame for the incident himself. The only concerning element of D's conduct was that he should not have been on the road at the time he was (because he was unlicensed and uninsured). In such circumstances, criminal censure and sanction of D for the death of V is unjust – there is no blameworthiness for the death of V. 103 Perhaps this partly explains the Supreme Court's approach in Hughes. But this reaffirms the view of many criminal-law scholars that strict liability is unjustified in criminal law and should be expunged. 104
INUS would also establish that Baldwin's conduct is a cause of Hutchins’ death. Here, a condition sufficient to bring about her death would be the conjunction of Baldwin pointing the gun at Hutchins, the gun discharging, Hutchins being present in proximity to Baldwin, and the fatal physical bodily processes that occur as a result of the gunshot. In this condition, Baldwin pointing the gun is necessary: If Baldwin had not pointed the gun, death would not have occurred at the time and in the way it did. Hence, Baldwin pointing the gun is an INUS condition of Hutchins's death. Although Baldwin's conduct would constitute a cause of Hutchins’ death, INUS may not impact Baldwin's overall criminal liability since his lack of moral responsibility would be dealt with by legal doctrines other than causation, such as mens rea. The point, here, is that INUS would properly detect Baldwin's causal responsibility and the issue of his moral responsibility would then be handled appropriately by legal considerations other than causation. It provides a clearer legal distinction between actus reus and mens rea.
The Potential Danger
Perhaps the main concern for INUS is that it increases the proliferation of causes, which raises serious practical issues in applying the test. 105 Moore has argued, in particular, that the problem is the ‘sheer number of events and states of affairs’ needed to create a sufficient condition. 106 He claims that to compile a list of factors sufficient for him to write his book, he would need to include factors such as oxygen for him to breathe; the neurological processes in his brain to function; his room being sufficiently warm so that his fingers did not freeze; his mother giving birth to him; the Big Bang and an almost infinite amount of other factors. 107 Even Mackie concedes that to establish a truly sufficient explanation of any actual past event, we need to include ‘the whole prior state of the universe’. 108
This may be true if INUS required lawful sufficiency. However, for INUS, we require causal sufficiency rather than mere lawful sufficiency. In addition, if we supplement the notion of causal sufficiency to be properly understood as minimal causal sufficiency, the number of causes (thus the numbers of factors required to construct a sufficient condition) is substantially limited. Minimal causal sufficiency is where we empirically determine, as best we can, what set of factors is minimally sufficient for a certain outcome to occur. 109 When constructing a minimally sufficient condition, non-included existing factors are simply ‘roped off’ while determining whether the described included existing factors constitute a minimally sufficient condition. 110 To give a simple example, when an assassin (A) shoots a President (P) dead the conjoined factors minimally sufficient for P's death would include the conjunction of A shooting P and the fatal physical bodily processes that occur as a result of P's fatal wound. Because of this established minimally sufficient condition, the plethora of additional existing factors, such as A's mother giving birth to A or C providing A the gun etc., become irrelevant and are disregarded as causal – they do not form part of the condition sufficient for P's death. The idea of a minimally sufficient condition is that it is supposed to create a kind of ‘causal boundary’ within which causation can safely be confined. 111
Nevertheless, even within a minimally sufficient condition, there are various factors that play a causal role. This is because, as Mackie acknowledges, outcomes are typically brought about by a conjunction of various factors and it is seldom, if ever, that they are brought about by a single factor. 112 But this is the essence of the INUS account. It recognises, as causes, not only strongly necessary factors (in the sense that without it the outcome would not have occurred on the particular occasion), but factors that were neither strongly necessary (in the sense that it is only necessary to complete a sufficient condition) nor independently sufficient. 113 Minimally sufficient conditions would also have to include an indefinite number of negative INUS conditions – that is, the non-occurrence of an event that was necessary for the condition's sufficiency. Thus, in the assassin example, the fact that P was not wearing his allocated bullet-proof vest; that the assassin's gun was not on safety; that P's personal bodyguard did not jump in the line of fire immediately before the bullet reached P etc., are all possible INUS conditions of P's death.
However, the proliferation of causes is not a significant problem for the practical utility of INUS in English criminal law. 114 Criminal law is not interested in determining all of the immediate and remote causes of every possible event, it is only concerned with the possible causes of legally recognised prohibited harms, and criminal law is interested in only a very few of those possible causes: The potentially culpable causes. So, in the assassin example, because A's culpable conduct forms part of the minimally sufficient condition, the courts would be able to properly establish A's conduct as a cause of P's death and could disregard all the other additional existing factors.
Conclusion
This article has challenged English criminal law's approach to causation. It has been argued that the standard tests of causation represent a normative exercise in finding D responsible for a prohibited outcome, often grounded only in D's moral responsibility for that outcome. This approach is problematic because moral responsibility is irrelevant to causal responsibility; and not distinguishing causal responsibility from moral responsibility results in inappropriate criminal-responsibility ascription for result crimes. This is because both notions are required to appropriately determine overall criminal responsibility for result crimes: Causal responsibility establishes D's actual wrongdoing, whereas moral responsibility determines D's blameworthiness for his wrongdoing. However, if causal findings are products of only moral-responsibility ascription, then this neglects a significant role in criminal-responsibility ascription – causal-responsibility ascription.
Given the proposed challenge to the standard tests, the article contended that it is preferable that English criminal law adopts a single, non-normative test of causation; a metaphysical one that offers a robust causal enquiry that focuses only on causal responsibility, which contributes appropriately to the criminal-responsibility ascription. It argued that INUS causation is the most rational metaphysical account of causation that could be adopted and should therefore replace the standard tests in English criminal law.
The analysis of INUS is broadly consistent with existing case-law. It would, however, generate different liability outcomes in cases where D is morally responsible for the prohibited outcome but is not causally responsible for it, such as Wallace. INUS will also affect liability outcomes in strict-liability result crimes where D is not morally responsible even though D is causally responsible for the proscribed outcome, such as Hughes. Ultimately, INUS causation would facilitate engagement with causal enquires in a broader range of cases on a more principled, clear and consistent basis.
Footnotes
Acknowledgements
The author would like to thank Matthew Gibson, Marie Fox, John Child, Daniel Hill and an anonymous reviewer for their helpful comments on earlier drafts of this article. A different version of this paper was presented at the Society of Legal Scholars Annual Conference at Durham University, September 2021. I am grateful to the participants at this event for their feedback. This paper was made possible by a Doctoral Scholarship from the University of Liverpool for which I am also very thankful. All errors remain my own.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
