Quaestio facti. Revista Internacional sobre Razonamiento Probatorio
Quaestio facti. International Journal on Evidential Legal Reasoning
Sección: Ensayos
2025 l 9 pp. 101-129
Madrid, 2025
DOI: 10.33115/udg_bib/qf.i9.23105
Marcial Pons Ediciones Jurídicas y Sociales
© Ho Hock Lai
ISSN: 2604-6202
Recibido: 23/12/2024 | Aceptado: 01/04/2025 | Publicado online: 27/06/2025
Editado bajo licencia Reconocimiento 4.0 Internacional de Creative Commons
SHOULD WE BE CONVICTING PEOPLE WE DON’T BELIEVE TO BE GUILTY?
Ho Hock Lai*
National University of Singapore
lawhohl@nus.edu.sg
Abstract: It is doubtful that knowledge of guilt is a necessary condition for a criminal conviction. More plausibly, justified belief is required. But a criminal conviction is not grounded in belief as straightforwardly as is sometimes assumed. While epistemology sheds light on legal proof, a full understanding would require also taking on board considerations of practical reasoning and political morality. What is necessary for a criminal conviction is not first-personal belief in the accused’s guilt. Instead, the judge is required to make a third-personal judgment of whether one would be justified in believing—on the evidence adduced before the court and within legal constraints—that the accused is guilty as charged. If the judge concludes that one would not be justified in believing in any of the facts which must exist to constitute the offence, the judge cannot claim as his or her reason for convicting the accused that the accused has committed the offence. The law should not allow the judge to convict the accused in these circumstances as his or her motivating reason would fall short of the normative reason needed for the conviction. Exceptionally, though, the law permits a lack of congruence between the motivating reason and the normative reason. This is unsettling because it undermines an important aspect of the rule of law.
KEYWORDS: evidence, legal proof, criminal conviction, epistemology, practical reasoning, presumption, burden of proof, rule of law, normative reason, motivating reason.
SUMMARY: 1. INTRODUCTION.— 2. TWO KINDS OF LEGAL RULES: 2.1. Stephen and Dan-Cohen; 2.2. Ofence rules and decision rules.— 3. OFFENCE RULES: 3.1. An example: the rule against drug trafficking; 3.2. Offence rules and normative reasons.— 4. DECISION RULES: 4.1. Disposition rules; 4.2. Simplistic epistemic account of legal proof; 4.3. Objections; 4.4. Responses to the objections; 4.5. Legal constraints of other decision rules.— 5. THE THIRD DISPOSITION RULE: 5.1. Presumption of innocence, 5.2. Presumption against offence elements, 5.3. Conflicting decision rule: conditional presumption of an offence element; 5.4. Conviction without proof of guilt.— 6. OBJECTIONS TO REVERSING THE LEGAL BURDEN OF PROOF: 6.1. Objection based on presumption of innocence; 6.2. Objection based on the rule of law.— 7. CONCLUSION.— BIBLIOGRAPHY.
Whether legal proof should be understood in epistemic terms has been heavily debated by epistemologists and legal scholars. Much of the debate has revolved around the sufficiency of naked statistics to support an affirmative verdict and concentrated on some epistemic property that the necessary evidence must supposedly have 1. One can insist on an epistemic property being necessary for legal proof without insisting that there must be knowledge of or justified belief in the target proposition. Here are just two examples. Gardiner borrows from the relevant alternatives theory of knowledge in developing the theory that legal proof requires evidence that eliminates relevant error possibilities (Gardiner, 2019). But she does not think that legal proof requires knowledge (Gardiner, 2024). Applying his normic theory of epistemic justification to law, Smith (2018) claims that a proposition is legally proved «only if the evidence makes the falsity of that proposition less normal, in the sense of calling for more explanation, than its truth» (pp. 1209-1210). However, he leaves it open whether it is «acceptable to base a verdict of guilt or liability on evidence that would be insufficient to ground justified belief in guilt or liability» (p. 1212, note 170).
Whatever the crucial epistemic property might be, is knowledge or justified belief necessary for legal proof? This question arises generally and not only in relation to statistical evidence. My discussion will, however, be confined to criminal cases. Specifically, is knowledge of or justified belief in the accused’s guilt required for a criminal conviction? The conclusion that I shall reach is that there is no convincing reason why we should go so far as to insist on knowledge. What about justified belief? Test it this way: should we be convicting people we don’t believe to be guilty? While the devil is in the details, a widely shared intuition, I surmise, is that we shouldn’t 2. I have previously offered a belief-centred account of legal proof (Ho, 2008, pp. 89-99). This paper develops it further. To be sure, others have also interpreted proof in terms of justified or rational belief. Nelkin (2021) claims that «legal verdicts…, to be justified, depend on the rational belief that the agent in question is culpable» (p. 25). Günther (2024) proposes that «[a] defendant should be found guilty just in case a fact-finder is justified in believing that the defendant is guilty» (p. 129). For Ross (2022), «guilty verdicts are appropriate only if a full belief in guilt would be rational, given the admissible evidence» (p. 1601) 3. Buchak is more circumspect. She is sympathetic to the idea that rational belief in guilt is necessary for a criminal conviction. However, she holds back from claiming as much because she sensed that the «relationship between it being licensed for a court to conclude that p on the basis of some evidence and it being rational for an epistemic agent to believe that p on the basis of that evidence» is «vexed» (Buchak, 2014, p. 291).
Buchak is right. The relationship is indeed vexed. One of the aims of this paper is to demonstrate that a criminal conviction is not grounded in belief as straightforwardly as is sometimes assumed. The judge or jury is expected to make an epistemic judgment from a third-personal rather than first-personal point of view, a judgment on what one would be justified in believing. Verdicts—including guilty verdicts—are not governed by an identical set of norms as beliefs. Apart from rules that exclude inadmissible evidence, there are rules that regulate reasoning on evidence that has been properly admitted. These rules impose constraints on evidentiary reasoning that do not necessarily apply to epistemic agents outside of the legal context. The law requires guilty verdicts to be reached by engaging in epistemic reasoning from a third-personal perspective and within those legal constraints. Indeed, occasionally, the law may displace epistemic reasoning altogether with practical reasoning.
This paper aims to make a further, and more novel, contribution. While epistemology sheds light on the legal proof of guilt, I will attempt to show that understanding it fully requires taking on board considerations of practical reasoning and political morality. The following line of thought will be pursued. If a court convicts the accused person while finding that one would not be justified in believing that he or she is guilty as charged, the court’s motivating reason for the conviction is inadequate. It is objectionable to convict people for inadequate reasons. I will buttress this claim by examining an exception that proves the rule so to speak. The law exceptionally compels courts under certain circumstances to return a guilty verdict and to convict an accused person even when they find that one would not be justified in believing him or her to be guilty. This is widely seen as legally problematic, and hence in need of special justification—and rightly so because, as I will argue, the rule of law is compromised.
My argument will proceed as follows. Part 2 sets the stage by noting the general distinction and relationship between two kinds of legal rules. I call them offence rules and decision rules. Part 3 takes a closer look at the nature and function of offence rules. The rule against drug trafficking will serve as an example. Offence rules do not merely create normative reasons to guide the conduct of citizens; they also provide normative reasons for the court to convict citizens for the relevant offences. Parts 4 and 5 turn to decision rules. They focus specifically on rules of evidence law that guide the court in fact-finding, in deciding what verdict to return and in choosing between a conviction and an acquittal. The effect of those rules is that, in general, the court may return a guilty verdict and convict a person of an offence only if the court concludes (on the evidence before it and within legal constraints) that one would be justified in believing that the facts are such that they constitute criminal guilt; however, exceptionally, a decision rule may require a conviction even where the court concludes otherwise. The exception will be illustrated by analyzing the presumption of trafficking. Part 6 explains why the exception is unsettling and the price it exacts on our commitment to the rule of law.
My argument rests on a distinction between two kinds of legal rules. In making the distinction, I draw on, but will not adhere strictly to, the works of Stephen (1872) and Dan-Cohen (1984).
Stephen (1872) situates the law of evidence within the larger system of legal rules. He distinguishes and notes the relationship between rules of substantive law and rules of procedural law. His account may be said to represent the conventional thinking among lawyers. According to him, «[e]very judicial proceeding … has for its purpose the ascertaining of some (legal) right or liability», and such «rights and liabilities are dependent upon and arise out of facts» (p. 7). Two things are required of the law under this scheme.
The first is to provide for the «legal effect of particular classes of facts in establishing rights and liabilities» (Stephen, 1872, p. 8). This is the province of substantive law. The second is to set out «a course of procedure…by which persons interested may apply the substantive law to particular cases» (p. 8). This is the job of procedural law. A subset of procedural law is the law of evidence. The law of evidence «determines how the parties are to convince the court of the existence of that state of facts which, according to the provisions of substantive law, would establish the existence of the right or liability which they allege to exist» (p. 8).
Inspired by Bentham, Dan-Cohen (1984) distinguishes between conduct rules and decision rules. These two types of rules differ in that they serve different purposes and the norms they create are directed at different actors and acts. Conduct rules are addressed to ‘the general public’ and this is for the purpose of guiding the behaviour of its members. Rules against theft, robbery and drug trafficking are examples. Decision rules are addressed to ‘officials’, telling them how they are to make decisions in the exercise of powers over the general public. They include rules of evidence law. The judge or jury must reach a decision on the verdict in accordance with the legal rules of evidence.
For Dan-Cohen (1984), a rule can be a conduct rule, a decision rule, or both; on the last possibility, he considers ‘statutes defining offences’ to be ‘decision rules as well as conduct rules’ because they do not just guide the behaviour of the general public, they also «specify for the courts some of the preconditions to the imposition of punishment» (p. 649).
Dan-Cohen’s distinction between conduct rules and decision rules does not map exactly onto Stephen’s distinction between substantive rules and procedural rules. Some rules that Stephen (1872) would view as rules of substantive criminal law, such as rules providing for the defences of duress and necessity, are treated as decision rules by Dan-Cohen. For Dan-Cohen (1984), defences such as duress and necessity are not meant to guide the behaviour of members of the general public (in the sense of guiding their decision whether to commit crimes); they are instead rules (based on some consideration of fairness or compassion) that apply after the event to guide officials in deciding on conviction. For Stephen, on the other hand, these defences are part of substantive law in that they set out the legal effect of duress or necessity on criminal liability. The broad insight I want to draw from their separate works is the distinction and relationship between legal rules that are norms of criminal liability and legal rules that are norms of decision-making.
By norms of criminal liability, I mean roughly what Dan-Cohen calls conduct rules. However, he gives the impression that a difference between conduct rules and decision rules is that the former are addressed to members of the general public whereas the latter are addressed to officials. But, as Dan-Cohen is undoubtedly aware, conduct rules may also apply, and apply only, to officials; an example is the rule of criminal law that forbids public servants from accepting gratifications in respect of an official act 4.
Norms of criminal liability may be found in rules defining offences as well as rules defining defences. To avoid unnecessary complications, this paper will ignore criminal defences. The focus will be on rules that define offences. I will call them ‘offence rules’ and refer to their norm-subjects as ‘citizens’.
By norms of decision-making, I mean specifically the norms that guide the appropriate officials in applying norms of criminal liability to citizens, and even more specifically, in the subset of such norms that consists of rules that guide judges and juries on fact-finding, verdict deliberation and the choice between conviction and acquittal. Henceforth, I will refer to them as ‘decision rules’.
While it is true that we expect judges and juries to apply the relevant offence rule to the accused person in the case before them, for my purposes and unlike Dan-Cohen, I do not treat the offence rule as itself a decision rule. Judges and juries are the norm-subjects of decision rules whereas the accused person is the norm-subject of the offence rule in question. While the judge and jury must follow decision rules in discharging their responsibilities, the offence rule is something that they apply to the accused person in determining his or her guilt. For ease of exposition, I will henceforth use ‘judge’ as a generic term to refer to the relevant official tasked with the relevant decision-making.
For a closer look at offence rules and their functions, and to remove the abstraction of the preceding discussion, let us consider the offence of drug trafficking. One can find in the substantive criminal law of most, if not all, legal systems today a rule against drug trafficking: the rule defines drug trafficking, makes it a crime, and forbids citizens from engaging in it. (This offence rule is invariably accompanied by a punishment rule that spells out the punishment for drug trafficking. Punishment rules are conceptually separate from offence rules.) The offence rule against drug trafficking is addressed to citizens and provides them with an authoritative normative reason not to do the act which the law has defined as drug trafficking.
While most, if not all, legal systems have a rule against drug trafficking, the formulation of that offence rule is jurisdiction-specific. Assume that under the relevant offence rule of a particular jurisdiction, a person commits the offence of drug trafficking if and only if (a) the person possesses a controlled drug, (b) while having knowledge of its nature, and (c) his or her possession of it was for the purpose of trafficking. 5
Given that all three elements are necessary to constitute the offence of drug trafficking, the accused person has not committed the offence of drug trafficking if (a) he or she was not in possession of the substance in question, or (b) was in possession of the substance but did not know what it was, or (c) was in possession of the substance and knew that it was a controlled drug but had no intention of trafficking in it.
The three elements (which constitute the actus reus and mens rea) of the offence definition are types of facts—or, as Stephen (1872) puts it, ‘classes’ of facts—that constitute the offence of drug trafficking. These fact-types are open to different factual instantiations. The instantiations of the types of facts that constitute a crime on a particular occasion—as we might say, tokens of the fact-types—have been given different legal names, including ‘material facts’, ‘ultimate facts’, ‘dispositive facts’ and ‘facts in issue’. I will stick to the term ‘material facts’. The criminal charge must cite the alleged offence and must be drafted with sufficient particularity to give the accused person fair notice of the material facts 6. It must indicate the time of the offence, the location where it was committed, the kind of controlled drug that the accused person is alleged to have been trafficking in and so forth.
What the judge must determine at the trial is not whether the accused person is guilty of drug trafficking but whether the person is guilty of the instance of the offence as particularized in the charge and as alleged in the narrative advanced by the prosecution at the trial (henceforth, the ‘charged offence’). While the discussion that follows will, for ease of exposition, refer to fact-types described in the offence definition, to be precise, the dispute at a trial is over at least one of the material facts.
As already noted, offence rules provide authoritative normative reasons to guide the conduct of citizens. The rule against drug trafficking gives citizens an authoritative normative reason to refrain from drug trafficking. Offence rules have another function: they authoritatively set out the normative reasons only for which judges may convict citizens of the relevant offences.
I shall sidestep the objectivism versus perspectivism debate by adopting the widely held objectivist view. On the objectivist view, normative reasons are facts; they are independent of the agent’s belief. In our example, the normative reason provided in the offence rule for the judge to convict a person of drug trafficking is the following set of necessary facts: (a) the person was in possession of a controlled drug, (b) while knowing what it was, and (c) his or her possession was for the purpose of trafficking. So long as one element—(a), (b) or (c)—does not obtain in the case at hand, there would be insufficient normative reason for the judge to convict the accused person for drug trafficking; this is because the person would then not have committed the offence and not deserve to be found guilty. As I have argued elsewhere, his or her conviction would be incorrect even if the judge believes, and is justified in believing, (wrongly) on the evidence that he or she is guilty as charged (Ho, 2021) 7. This is not to say that the judge deserves blame for the error for that is a separate matter.
Let me turn now to decision rules. Decision rules guide the judge in deciding, among other things, on (i) the disposition of the case (by either convicting or acquitting the accused person); (ii) the type of verdict to return (guilty or not-guilty); and, (iii) issues of proof (by making findings on disputes of fact).
Where a person is put on trial for violating an offence rule, there are typically only two possible outcomes by which I mean two possible ways in which the judge may dispose of the case 8: either convict or acquit. The accused is convicted following a guilty verdict 9. In general (though, as we shall see, not always), the judge must return a guilty verdict only if the judge finds that his or her guilt has been proved. An acquittal follows a verdict of ‘not guilty’. The judge may return a not-guilty verdict only in one of two situations: the first is when the judge finds that the evidence supports the finding that guilt has been disproved and the second is when the judge finds that the evidence support neither the finding that guilt has been proved nor the finding that guilt has been disproved. In the second situation, the accused is declared ‘not guilty’ on the presumption that he or she is innocent and not because his or her guilt has been disproved.
Scotland is unusual in allowing a third type of verdict. This is the verdict of ‘not proven’. It is controversial what a ‘not proven’ verdict signify and when it is called for 10. What is uncontroversial in any legal system is the logical possibility of a fact not being proven. It is entirely possible, even outside of Scotland, for guilt to be not proven. Indeed, legislation in the Commonwealth that is based on the Indian Evidence Act of 1872 explicitly provides for the possibilities of a fact being ‘proved’, ‘disproved’ and ‘not proved’; the last is defined as the situation where the fact is neither proved nor disproved 11. The distinctions that I have drawn may be tabulated as follows (leaving out the Scottish aberration):
|
Disposition of the case |
Verdict |
Finding on proof |
|
Convict |
Guilty |
Guilt is proved |
|
Acquit |
Not-Guilty |
Either: (i) Guilt is disproved, or (ii) Guilt is neither proved nor disproved = innocence is presumed |
The applicable decision rules include a set of three disposition rules. Disposition rules pertain to the outcome of the trial and instruct the judge on how he or she is to dispose of the case. There are only two possible outcomes: either a conviction or an acquittal. Two of the disposition rules may be stated as follows (I will come to the third in Part 5):
(DR1) The judge must return a guilty verdict and convict the accused person of the charged offence if the guilt of that person for that offence is proved, and
(DR2) The judge must return a not-guilty verdict and acquit the accused person of the charged offence if the guilt of that person for that offence is disproved.
Four clarifications are in order. First, it might be said that criminal guilt is a legal construct. No one is guilty unless declared by the court to be guilty. While this may be correct in one sense, as I use the term in this paper, ‘guilt’ is factual and independent of proof in court; I will treat it as possible for a person to be (in fact) guilty and for the prosecution to be unable to prove that he or she is (in fact) guilty.
Secondly, to avoid misunderstanding, I should make it clear that DR2 does not imply that the accused person carries a general burden of proving his or her innocence. What it does is to instruct the judge that if the evidence brought before the judge is such that guilt is disproved, the judge must return a not-guilty verdict and acquit the accused person. This, of course, is not the only situation in which the accused must be acquitted.
Thirdly, a person is guilty of an offence if and only if he or she has committed the offence and does not have any defence. A person who has committed the offence of drug trafficking must nevertheless be acquitted if he or she did it, say, under duress 12. As previously mentioned, criminal defences will set aside to avoid unnecessary distraction; I will assume that the dispute is entirely on the commission of the offence and that the case for the accused rests solely on challenging the presence of one or more elements of the offence.
Fourthly, my discussion is in the context of the adversarial trial system. In such a system, the parties are in the driving seat; they have control over the scope of the dispute and the production of evidence, and either one or the other of the parties is assigned the burden of proving or disproving the disputed facts 13. The disposition rules presuppose that there is a dispute between the parties as to whether the accused person is guilty as charged and a dispute on at least one material fact. There may be no dispute as to whether the accused person is guilty as charged. In an adversarial system, if the accused pleads guilty, proof of guilt is dispensed with. The person is convicted without any guilty verdict. Where the accused person pleads not guilty, the dispute may be limited to only some of the material facts. In a charge of drug trafficking, the accused person may admit formally to having possession of a substance (element (a)) but deny knowledge that the substance was a controlled drug (element (b)); in that event, the prosecution carries the burden of proving the latter but does not need to prove the former 14.
The disposition rules are addressed to the judge. They guide the judge in deciding whether to convict or acquit the accused person by providing the judge with authoritative reasons for convicting or acquitting the accused person. The motivating reasons in applying DR1 and DR2 may be stated as follows:
Where a judge follows DR1 in convicting an accused person of drug trafficking, his or her motivating reason for convicting the accused person is the fact that the person’s guilt for the offence has been proved.
Where a judge follows DR2 in acquitting an accused person of drug trafficking, his or her motivating reason for acquitting the accused person is the fact that the person’s guilt for the offence has been disproved.
Here, ‘motivating reason’ carries the usual meaning. It is the reason for which or in the light of which the agent (in the present context, the judge) acts.
To apply the disposition rules, the judge must come to a conclusion on whether guilt is proved. Some legal scholars and philosophers contend that legal proof turns on knowledge. Duff et al. (2007, p. 89) are of the view that a criminal «conviction is appropriate only if the fact-finder knows that the defendant is guilty». For Pardo (2010, p. 38), «the goal or aim of legal proof is knowledge (or something approximating knowledge)». Other knowledge-based accounts of legal proof are more technical and have been offered by philosophers such as Moss, 15 Blome-Tillmann 16 and Littlejohn 17.
If the concept of legal proof is to be construed in terms of knowledge, and if verdict deliberation requires a search by the judge for epistemic justification in the resources to which he or she has access, it seems the natural account of knowledge for present purposes would have to be an internalist one. As Smith has remarked (Smith, 2018, pp. 1205-6):
If two courts were presented with equivalent bodies of evidence against two individuals charged with equivalent crimes, could it really be acceptable for them to reach different verdicts—for one individual to be found guilty and the other innocent—even if there was some variation in external circumstances? No doubt we can persuade ourselves to say all manner of things about epistemological thought experiments, but there seems to be something almost viscerally bad about such a turn of events.
On the traditional internalist account of knowledge, one knows p only if one believes that p, one is justified in believing that p, and p is true. There is consensus that these three conditions are insufficient but no consensus on what else needs to be satisfied. On the traditional analysis, knowledge is non-primitive and requires justified belief. There is general agreement that belief (that is, outright or full belief), as opposed to partial belief or credence, is needed for knowledge.
So, on both the justified belief account and knowledge account of legal proof, proof of p requires justified belief in p. Since proof is factive, p is proved only if p is true. Shouldn’t justified belief in guilt and the truth of that belief suffice for the legal proof of guilt? Why should we insist, more strongly, on knowledge? 18 Suppose the judge, having considered the evidence, comes justifiably to the true belief that the accused is guilty as charged. Should it matter in law that the belief was true by luck and therefore, according to the received view in epistemology, doesn’t amount to knowledge? I cannot see why it should. Examples given to illustrate the importance of knowledge fail to do so.
Here is an example by Pardo (2010, p. 50):
Framed Defendant: The police arrest a motorist and plant drugs in his car. He is convicted at trial of illegal possession based solely on testimony from the arresting officers and the planted drugs. As it turns out, the defendant did have illegal drugs in his car at the time that never were discovered. The verdict that the defendant possessed drugs is therefore both true and justified (that is, the evidence at the time of the trial is sufficient to establish a conviction beyond a reasonable doubt), but the truth and the justifying evidence are disconnected. The truth of the verdict is purely coincidental or accidental.
As noted in Part 3.1, the issue before the judge is not whether the accused has committed the type of offence with which the person has been charged; what the judge must decide is whether the accused is guilty of the charged offence, or more elaborately, guilty of the particularized instance of the type of the offence cited in the charge and as alleged by the prosecution. Pace Pardo, the verdict is false in Framed Defendant. The defendant was not in illegal possession of the drugs that were allegedly found in his car. He was unaware of their existence until they were revealed by the police officers. In law, one cannot be in possession of something the existence of which one is (non-culpably) ignorant. So, the defendant was never in possession of the alleged drugs that formed the subject matter of the charge. Take an even more far-fetched scenario: I am charged with the murder of A. I didn’t do it. The police set me up. If I am convicted, the verdict is false. We do not consider the verdict true because, ‘by luck’ and unbeknownst to the judge, I happened to have murdered B. My conviction is for the murder of A and not B. Similarly, in Framed Defendant, the defendant was falsely convicted for possessing the drugs (A) that were found in his car. He was neither charged with nor convicted for possession of the drugs (B) that were never discovered by the police.
Consider yet another example by Pardo (2010, p. 52):
Fake Cabs: The plaintiff files a lawsuit against the defendant, who owns and drives the only taxicab in town, claiming she was hit by the defendant’s cab while crossing the street. She saw the cab drive away but did not see the driver. A video camera at the intersection filmed the accident, and it shows what appears to be a cab (but not the driver) hitting the plaintiff, exactly as she claimed. Now, suppose the car in the video really is the defendant’s, but also that—unknown to the jury—along with his real cab there are hundreds of other cars in the town that look identical to his cab. The jury finds for the plaintiff based on the video.
Ex hypothesi, the jury was justified in believing that the defendant is liable, the belief is true, and the verdict is correct. It is doubtful that we should ask for more. Epistemologists committed to externalism might claim that the belief was unsafe due to the number of fake cabs in town and hence cannot amount to knowledge 19. But, so far as the law is concerned, so what? Rather than bemoan the lack of knowledge, it seems to me that the natural reaction is, as Nance suggested, to breathe a sigh of relief (Nance, 2021, p. 105). Luckily, all turned out well. What if, contrary to the example above, the belief and verdict were false? Well, that (the falsity) should be our objection. It seems unnecessarily distracting to frame our criticism in terms of the lack of knowledge.
Suppose we are faced with a different situation where evidence of the existence of fake cabs emerges after the verdict against the defendant was delivered, and it is not a given that the verdict is true, or for that matter false. Arguably, with the introduction of the fresh evidence, belief in the defendant’s liability is no longer justified and, in the light of the new evidence, we think that the verdict may well be false. This may supply sufficient grounds for granting the defendant some legal recourse such as allowing the introduction of fresh evidence in an appeal or perhaps even a retrial. The concepts of belief, justification and truth seem sufficient for analytical purposes. Again, it is unclear what good there is in bringing knowledge into the analysis.
Since knowledge appears to be an unnecessary distraction, an epistemic interpretation of the rules guiding the decision on proof could be based more simply on justified belief. On a simplistic formulation of the rules (and I will clarify later why the formulation is simplistic):
(Psimplistic) The judge must find that guilt for the charged offence is proved if and only if the judge believes justifiably that the accused person is guilty of that offence, and,
(Dsimplistic) The judge must find that guilt for the charged offence is disproved if and only if the judge believes justifiably that the accused person is not guilty of that offence.
These rules guide findings on the issue of proof (henceforth, ‘proof findings’). Let us return to our drug trafficking example. A person is guilty of the offence of drug trafficking only if all of the elements necessary to constitute that offence are factually instantiated in the case. Consequently, according to the decision rules above:
(Psimplistic-DT) The judge must find that guilt for the offence of drug trafficking is proved if and only if the judge believes justifiably that (a) the accused was in possession of a controlled drug, (b) while knowing what it was, and (c) his or her possession was for the purpose of trafficking, and,
(Dsimplistic-DT) The judge must find that guilt for the offence of drug trafficking is disproved if and only if the judge believes justifiably that at least one of the above—(a), (b) or (c)—is not the case.
Critics have voiced objections to the simplistic epistemic construction of the rules guiding proof findings 20. Let me highlight three objections.
First, if proof of guilt supervenes on the judge’s belief, and if doxastic involuntarism is true, as many think it is, the judge cannot be responsible for making the finding which he or she did since the judge has no control over his or her belief. But we do hold the judge responsible for his or her proof findings. So, proof findings cannot simply reflect the judge’s beliefs 21.
The second objection rests on the view that belief is non-contextual 22. At common law, there are usually only two standards of proof 23. Where a legal burden of proof is carried by a party in a civil case or by an accused person in criminal case, the standard of proof that must be satisfied in order to discharge the burden is the balance of probabilities, also called preponderance of evidence or preponderance of probabilities. The standard of proof applicable to the prosecution in criminal cases is the higher one of beyond reasonable doubt. It is possible for the same set of evidence to be sufficient to prove a fact on the lower standard (of balance of probabilities) but not on the higher standard (of beyond reasonable doubt). But we cannot say the same of belief—either the evidence justifies the belief that p or it does not; it matters not whether the question ‘whether p?’ arises in the context of a civil case or in the context of a criminal case, or whether it is the prosecution or the accused person who is alleging that p. Hence, legal proof cannot be based on belief.
The third objection is related to the earlier ones. A person cannot rationally believe at the same time and on the same set of evidence both p and not p. The judge, unless irrational, cannot at the same time and on the same set of evidence believe (in his or her official role) that the accused is guilty and believe (in some other capacity or for some other purpose) that the accused is not guilty. The judge either believes in one or the other or neither. But, so the argument goes, it is possible for a person to believe involuntarily that p while his or her better judgment tells the person that not-p. With this in mind, we can see how the judge might be expected to decide and act in accordance with his or her better judgment and contrary to the judge’s belief regarding the accused’s guilt. Lackey (2021, p. 192) offers this hypothetical:
Racist Raymond: Raymond was raised by racist parents in a very small-minded community and, for most of his life, he shared the majority of beliefs held by his friends and family members. After graduating from high school, he started taking classes at a local community college and soon began recognizing some of the causes, and consequences, of racism. During this time, Raymond was called to serve on the jury of a case involving a young, affluent white man on trial for raping a Black woman. After hearing all of the evidence presented by both the prosecution and the defense, Raymond is able to recognize that the evidence clearly supports the conclusion that the defendant committed the crime of which he is accused. In spite of this, however, he can’t shake the feeling in his gut that the man on trial is innocent of raping the woman in question, repeatedly calling to mind how the defendant just doesn’t look like a rapist. Upon further reflection, Raymond begins to suspect that such a feeling is grounded in the racism that he still harbors, and so he concludes that even if he can’t quite come to outright believe that the defendant is guilty himself, he nonetheless has an obligation to follow the evidence, not his gut. Despite the fact that he does not believe, and hence does not know, that the defendant in question is guilty, Raymond votes to convict, a verdict that is unanimously shared by the other jurors.
Unless he is irrational, and we are assuming that he is not, Raymond cannot at the same time and on the same evidence believe (for the official purpose of determining the verdict) that the accused person is guilty and believe (privately, ‘in his gut’) that the accused is not guilty. It must be, so the reasoning goes, that what is happening here is that Raymond believes that the accused is not guilty, and he is voting for a guilty verdict against his belief—as he rightly should.
The three objections are either unpersuasive or surmountable. The first objection—that we have no direct control over what we believe—is largely beside the point. Judging whether something is true is a phenomenon that surely all of us are familiar with. By judging that p, I mean roughly making up one’s mind that p is true for the reason that one finds sufficient evidence to support the truth of p. Admittedly, the mental state that p (whether it be the belief that p or the doxastic acceptance of p) arises involuntarily from judging that p 24. Even so, there is agency in the judgment that brought forth the mental state and the judge is responsible and open to rational criticisms for the judgment. It does not follow from the involuntariness of the mental state that we cannot hold the judge epistemically responsible for his or her judgment. In Racist Raymond, Raymond is open to criticisms if he had judged the defendant innocent and come to believe in his innocence out of racism.
The second objection is not easy to overcome. One convenient response, which this paper will adopt, is to confine an epistemic account of proof findings to ‘proof of guilt’ or ‘proof beyond reasonable doubt’ while remaining agnostic on ‘proof on the balance of probabilities’. A second response is to insist that proof on both standards requires belief and to locate the difference in the content of the belief: the higher standard requires belief in the existence of the material facts to a higher degree of probabilities than the lower standard. As I have argued elsewhere in relation to the criminal standard, there are problems with this probabilistic approach (Ho, 2021). A third response is to confront the second objection, deny that belief is non-contextual, and take the position that epistemic standard is relative to the interest at stake. This thesis is attractive but its defence falls outside the ambit of this paper 25.
The third objection loses much of its force if we follow Lehrer in distinguishing between belief as a ‘first-order doxastic state’ and acceptance as «a metamental state ordinarily based on positive evaluation of belief» (Lehrer, 2000a, p. 209). For Lehrer, knowledge requires epistemic acceptance and not mere belief.
While ‘acceptance’ has often been offered as an alternative to ‘belief’ in literature on legal fact-finding, it is taken to mean or encompass pragmatic or practical acceptance (Cohen, 1991; Ferrer Beltrán, 2004 and 2006). Cohen offers this definition which he thinks is an apt description of what goes on in the judge’s mind during verdict deliberation (see also Cohen, 1992, p. 4):
… to “accept that p” … is to treat it as given that p. It is to have or adopt a policy of deeming, positing, or postulating that p—that is, of going along with that proposition in one’s mind as a premise or license for inference…. (1991, p. 466).
We can deem, posit or postulate that p for all sorts of purposes without believing that p. If deeming, positing or postulating that p is to accept that p, the acceptance is not necessarily epistemic, and clearly not acceptance in Lehrer’s sense. Lehrer acknowledges that acceptance is an attitude defined in terms of some purpose and the purpose can be of the practical sort. But, according to him, the acceptance that is required for knowledge is acceptance «for the epistemic purpose of attaining truth and avoiding error with respect to the very thing that one accepts»: it is to accept that p if and only if p (Lehrer, 2000b, p. 13). Not everyone agrees with Lehrer’s theory of knowledge; this is hardly surprising since nearly everything in philosophy is controversial. It is a puzzle to me what sort of mental state epistemic acceptance is, and how it is, if it is at all, different from (reflective) belief. Indeed, acceptance aimed at truth is so close to belief that Lehrer uses the terms interchangeably in his book Theory of Knowledge where he thinks precision is not needed 26.
We may agree with the premise of the second objection to this extent: one cannot at the same time rationally believe both p and not-p. In Lackey’s hypothetical, we are told that ‘Raymond is able to recognize that the evidence clearly supports the conclusion that the defendant committed the crime of which he is accused’. Given this and Raymond’s rationality, Lackey’s assumption that he does not believe that the defendant is guilty seems questionable. But even if we concede this point to Lackey, the hypothetical is not fatal to an epistemic account of proof findings. As Lehrer would see it, Raymond accepts (in the sense necessary for knowledge) that the defendant is guilty even though he cannot help believing that the defendant is not guilty. So, even if we concede that proof findings do not rest on belief, it is still possible to maintain an epistemic account of proof findings by basing them on epistemic acceptance.
While the three objections canvassed above are not fatal, the simplistic epistemic formulations of the rules on proof findings are indeed simplistic. The third objection does raise a genuine difficulty. As I shall now proceed to explain, proof findings do not rest on first-personal belief or first-personal epistemic acceptance. Cohen (1991) and Ferrer Beltrán (2004, 2006 and 2021) have separately and rightly emphasized that there are other legal rules that any account of legal proof must contend with. Although they do not put it this way, the effect of those other rules is to require proof findings to be made from a third-personal instead of a first-personal point of view.
First, there are rules on what the judge must not treat as evidence in determining whether guilt has been proved. For various policy reasons, the judge may only take into account admissible evidence that has been properly adduced at the trial 27. This has several implications. One is that the judge may not rely on his or her private knowledge of facts of which no evidence has been presented in court 28. Indeed, if the judge has any personal knowledge about the case, that will likely disqualify him or her from acting as a judge. Another implication is that if the judge has somehow been wrongly exposed to inadmissible evidence in the course of the trial, the judge must ignore those evidence during verdict deliberation.
Secondly, there are rules on corroboration or the sufficiency of evidence. Such rules are nowadays relatively rare since modern systems of legal adjudication pride themselves on free evaluation of evidential weight 29. One instance of such a rule is section 6 of the Sedition Act of Singapore 30. It provides that no person shall be convicted of the offence of sedition on the uncorroborated testimony of one witness. So, in a sedition case, however reliable the testimony of a witness against the accused may be, and even if the judge is justified in believing the witness, the judge is not permitted to return a guilty verdict on the testimony alone. But notice that, so far as a guilty verdict is concerned, section 6 works as a negative constraint rather than a positive compulsion. While section 6 requires that there must be more than the testimony of one witness, it does not compel the court to convict the accused so long as there is some corroborating evidence.
Thirdly, in relation to the evidence that has been properly admitted at the trial and which the judge must consider in determining whether guilt has been proved, there are rules that constrain reasoning on the evidence or the uses to which the evidence may be put 31. These constraints are imposed by law. Two examples will suffice.
Under the common law rule against hearsay, evidence of an out-of-court statement may not be used as a basis for inferring the truth of the fact asserted in the statement. An out-of-court statement by X that the accused person stabbed Y cannot be used as evidence of the fact that the accused person stabbed Y. But evidence of an out-of-court statement may be used as a basis for drawing such other logically relevant inferences as may be justified in the circumstances of the case. Thus, where the accused person is relying on the defence of duress, evidence of the fact that someone had made an out-of-court statement threatening harm to him or her may be used as a basis for inferring that he or she was in fear 32.
The second example is the rule that is represented in Rule 404(b)(1) of the Federal Rules of Evidence in the United States. It states that «[e]vidence of any other crime, wrong, or act is not admissible to prove a person’s character in order to show that on a particular occasion the person acted in accordance with the character». There is an equivalent rule at common law which forbids a similar line of reasoning 33. But evidence showing a past misdeed may, in some circumstances, support a factual inference in a manner that avoids the forbidden line of reasoning, and if so, the evidence is admissible for that purpose 34. Thus, where the accused is charged with the offence of possessing insignia relating to a triad society and claims lack of knowledge that the items found in his possession were related to the particular triad society, the prosecution may adduce evidence of his previous conviction of membership of the same triad society to prove that he had such knowledge 35. The inference of knowledge is not legally forbidden as it is not drawn from the evidence via any assumption about the person’s character or propensity (‘once a thief, always a thief’).
In both examples, some form epistemic paternalism is arguably at work 36. The hearsay inference is unreliable because it rests on the credibility of the maker of the out-of-statement who cannot be tested in cross-examination. Reasoning from past misdeeds is unsound where it involves jumping to the conclusion of guilt from one’s perception of the kind of person the accused is.
Notice, further, that the legal regulation of evidential reasoning is only by way of negative and narrowly defined constraints. Save for specifying some limited uses to which the evidence may not be put, the law in general does not compel the judge to draw any particular inference of fact from the evidence. As Thayer (1898) famously put it, «The law has no mandamus to the logical faculty; it orders nobody to draw inferences» (p. 314, note 1).
If an epistemic construction of the rules guiding proof findings is not to be simplistic, it must acknowledge the legal constraints of decision rules of the sorts that we have just seen. Proof findings must be made within those constraints and from a detached perspective. It is not the first-personal belief of the judge that matters; what the judge must ask himself or herself is essentially, ‘would one be justified in believing that the accused is guilty based on the evidence before the court and within the constraints of the applicable decision rules?’ I will leave it as an open question whether belief for present purposes should include or be replaced by epistemic acceptance of the kind, or similar to the kind, proposed by Lehrer (2000a and 2000b).
The following is a summary of the preceding discussion:
(P) The judge must find that guilt for the charged offence is proved if and only if, based on the evidence before the court and within legal constraints, one would be justified in believing that the accused person is guilty of that offence, and
(D) The judge must find that guilt for the charged offence is disproved if and only if, based on the evidence before the court and within legal constraints, one would be justified in believing that the accused person is not guilty of that offence.
In what follows, I will speak simply of whether one would be justified in believing the accused to be guilty or not guilty; the qualifications to which I have alluded should be taken as tacit.
(P) and (D) are rules guiding proof findings and not definitions of proof and disproof respectively 37. While proof is factive, and a falsehood cannot be proved, 38 it is possible for the judge to follow (P) and find as proven the guilt of an accused person who (unbeknownst to the judge) is in fact innocent, and for the judge to follow (D) and find as disproven the guilt of an accused person who (unbeknownst to the judge) is in fact guilty; in both situations, and as explained in Part 3.2, the judge’s finding is objectively incorrect. Although the judge was objectively wrong in making the finding which he or she did, the judge is excused from blame as he or she reasonably believed that the relevant finding on proof was justified. Or so I have argued elsewhere (Ho, 2021) 39.
It is not enough to have the two disposition rules, DR1 and DR2. Suppose the evidence is such that one would neither be justified in believing that the accused is guilty nor be justified in believing that the accused is not guilty. Proof finding under these circumstances is guided by the following rule:
(¬P&¬D) The judge must find that guilt for the charged offence is neither proved nor disproved if and only if, based on the evidence before the court and within legal constraints, one would neither be justified in believing that the accused person is guilty of that offence nor be justified in believing that the accused person is not guilty of that offence
Where the judge finds that guilt is neither proved nor disproved, how is the judge to dispose of the case? As I have said, there are only two options: either convict or acquit. In the present case, DR1 does not permit conviction (since guilt has not been proved), and DR2 does not permit acquittal (since guilt has not been disproved). The following third disposition rule breaks the impasse:
(DR3) The judge must return a not-guilty verdict and acquit the accused person of the charged offence if the guilt of that person for that offence is neither proved nor disproved.
The reason why the accused person is acquitted under DR(3) is because, where guilt is neither proved nor disproved, the accused’s person is presumed to be innocent. Hence, DR(3) is more fully stated as follows:
(DR3poi) The judge must presume that the accused person is not guilty of the charged offence and, on that basis, acquit that person of that offence if and only if the guilt of that person for that offence is neither proved nor disproved.
DR3poi is commonly known as the presumption of innocence. It is a rule of practical reasoning that serves as a means of extrication; as suggested by Ullmann-Margalit (1983), it offers the judge a way out when the judge is called upon to act in a state of ignorance or doubt. It entitles the judge to make an assumption that the judge is otherwise not justified in making, and to ground the subsequent course of action on that assumption.
The presumption of innocence is against each element of the offence. To elaborate: guilt for an offence is proved only if all elements of the offence are proved. So long as one of the elements is disproved, guilt is disproved. So long as one of the elements is neither proved nor disproved, guilt is neither proved nor disproved; more fully (with the refinements in italics),
(¬P&¬D:element) The judge must find that guilt for the charged offence is neither proved nor disproved where, based on the evidence before the court and within legal constraints, and in relation to any element of the offence, one would neither be justified in believing that the element obtains nor be justified in believing that the element does not obtain.
In the above situation (¬P&¬D:element), DR3poi instructs the judge to deliver an acquittal. A more specific articulation of the presumption of innocence (DR3poi) that descends to the level of offence elements is as follows (again, with the refinement underlined):
(DR3poi:element) The judge must presume that the accused person is not guilty of the charged offence and, on that basis, acquit that person of that offence if any element of the offence is neither proved nor disproved.
Let us return to the example of drug trafficking. Recall that under the offence rule against drug trafficking, a person commits that offence if and only if (a) the person was in possession of a controlled drug, (b) while having knowledge of its nature, and (c) his or her possession was for the purposes of trafficking. Imagine that only the third element is disputed. The judge must, without the accused person having to do anything, presume that his or her possession was not for the purposes of trafficking, and acquit the person accordingly unless the prosecution proves the possession was for the purpose of trafficking.
But the presumption against offence elements may be overturned with respect to an offence element by a conflicting decision rule that provides conditionally for a presumption of that offence element. In other words, the burden of proof in relation to an offence element may be reversed by a conflicting rule of presumption. There are many variants of presumption rules—or various ways of interpreting how they work. On one variant, once the conflicting rule of presumption is invoked, the prosecution is released from the burden of having to prove the instantiation of the presumed element and it is the accused who now carries the burden of disproving it.
One example of such a rule—a rule that provides for a rebuttable presumption of law that shifts the legal burden of proof—is section 17 of the Singapore Misuse of Drugs Act 1973 (hereinafter, ‘section 17’). Many other jurisdictions have a similar provision. Section 17 states that where «[a]ny person is proved to have had in his possession» more than a specified quantity of a controlled drug (for example, more than two grammes of diamorphine), the person is «presumed to have had that drug in possession for the purpose of trafficking unless it is proved that his or her possession of that drug was not for that purpose».
The logical structure of this variant of a conditional rule of presumption may be represented as follows:
(RP) (i) Upon proof of A and (ii) unless B is disproved, the judge must find that B
On the standard terminology, A is the ‘basic fact’ and B is the ‘presumptive fact’. Under section 17, A is the fact of possessing more than two grammes of diamorphine and B is the fact that trafficking was the purpose of the possession. The party seeking to invoke the rule of presumption is the prosecution and the party seeking to resist the application of the rule is the defence.
To trigger the rule of presumption (RP), the prosecution must prove A. Once the rule is triggered by the prosecution, the judge must make a finding of B unless the defence disproves B. If the rule is triggered by the prosecution and the defence fails to disprove B, the judge must find that B in the sense of accepting that B for the practical purposes of determining the verdict. The judge is required to act as if B were true. In drawing the presumption, the propositional attitude is not belief but, as Ullmann-Margalit and Margalit (1992, p. 171) have proposed, ‘holding as true’ 40. It is irrelevant whether the judge believes that B. Indeed, the rule of presumption (RP) is needed precisely because A on its own does not justify the belief that B.
Having outlined the logic of this type of rule, let us see how section 17 works. To recap: our scenario is one in which an accused person is on trial for the offence of trafficking in diamorphine. There is no dispute that (a) the person was in possession of diamorphine and (b) the person knew what it was; the only point of contention at the trial is (c) whether his or her possession of the diamorphine was for the purpose of trafficking. Suppose the prosecution fails to adduce sufficient evidence to prove the third element; the evidence is not strong enough such that one would be justified in believing that the possession of the drug was for the purpose of trafficking.
However, the prosecution manages to prove the basic fact of possessing diamorphine exceeding two grammes. Under section 17, the accused now has the burden of rebutting the presumption by disproving the presumptive fact that his or her possession was for the purpose of trafficking. If the accused succeeds in doing so, he or she would have disproved a necessary element of the offence and thus disproved his or her guilt: in that event the accused must be acquitted under DR2. As we may recall, according to DR2, ‘the judge must acquit the accused person of an offence if the guilt of that person for that offence is disproved.’
Suppose the accused person attempts to rebut the presumption by claiming in his or her defence that the drug was meant entirely for personal consumption. If it is true that the drug in the accused person’s possession was wholly for his or her own consumption, it is false that he or she possessed it for the purpose of trafficking. So, if the accused person succeeds in proving that the drug was meant entirely for personal use, he or she would have succeeded in disproving that he or she had it for the purpose of trafficking.
Assume that the evidence adduced by the defence, while not fanciful, is more likely false than true. We might even suppose that the evidence is only marginally more likely false than true. On the evidence, one would not be justified in believing that the drug was for the accused person’s own consumption. Consequently, the presumption that he or she possessed the drug for the purpose of trafficking stands unrebutted. At the same time, as postulated at the beginning, it is not proved that the accused person had the drug for the purpose of trafficking; on the available evidence, one would not be justified in believing that that was true.
So, the situation at hand is one where a necessary element of the crime is neither proved nor disproved, and as such, one where it is neither proved nor disproved that the accused is guilty of the offence of drug trafficking. Whereas the presumption of innocence (DR3poi:element) would require the accused to be acquitted, the effect of applying section 17 is that the accused must be convicted. The accused must be declared guilty of an offence in circumstances where one is neither justified in believing that his or her possession of the drug was for the purpose of trafficking nor justified in believing that it was not, and hence where it is neither proved nor disproved that the accused has committed the offence. Section 17 permits—and, indeed, may require—the judge to convict the accused person despite being of the view that one would not be justified in believing that he or she is guilty as charged.
A decision rule like the rule of presumption in section 17 reverses the burden of proof. Rules that have this effect, sometimes called ‘reverse onus’ provisions, are controversial. Courts have struck down such provisions on the ground that they violate the constitutional or fundamental right to be presumed innocent. At other times, they have chosen instead to ‘read down’ these provisions, limiting their effect, when triggered, to placing only an evidential burden on the accused person. «An evidential burden is not a burden of proof. It is a burden of raising, on the evidence in the case, an issue as to the matter in question fit for consideration by the tribunal of fact» Sheldrake v DPP [2005] 1 AC at 289 41. But courts have also upheld reverse onus provisions, typically only when there is special justification for their existence. A rule of presumption similar to section 17 would be upheld in Canada only if it is within «reasonable limits prescribed by the law as can be demonstrably justified in a free and democratic society» (R v Oakes (1986) 26 DLR (4th) 200 at 224; Canadian Charter of Rights and Freedoms, section 1). In Hong Kong, the test is whether the presumption has (a) a rational connection with a legitimate societal aim (the rationality test); and (b) it is no more than necessary to achieve that legitimate aim (the proportionality test) (HKSAR v Lam Kwong Wai & Anor (2006) 9 HKCFAR 574; HKSAR v Hung Chan Wa (2006) 9 HKCFAR 614 at [39]). In England (Sheldrake v DPP [2005] 1 AC 264, 297) and Malaysia (Alma Nudo Atenza v PP [2019] 4 MLJ 1), some sort of proportionality approach is also taken. I do not intend to engage in doctrinal analysis here. My interest is in identifying why reversal of the burden of proof is widely seen as legally problematic and calls for special justification. What is the objection that needs to be overcome with weightier countervailing interests or considerations?
As noted, decision rules that reverse the legal burden of proof in relation to an offence element can be seen as exceptions to, or qualifications of, the presumption of innocence. Courts and lawyers have often expressed objections to such decision rules on the ground that they violate the presumption of innocence. This is circular. It merely describes the state of affairs; it does not explain why the state of affairs is objectionable. There is nothing intrinsically objectionable, and it is a commonplace, for one rule to qualify another and for rules to have exceptions. After all, some interests may be outweighed by other interests and a pro tanto wrong may be justifiable all things considered. As I will now argue, such decision rules are indeed objectionable. This is because they derogate from the rule of law.
In our example, given the judge’s finding that elements (a) and (b) are proved, the judge may claim as his or her reason for convicting the accused person the fact that the person (a) was in possession of a controlled drug (b) while having knowledge of its nature. But (a) and (b) are not sufficient to constitute the offence of drug trafficking. The offence is created and defined by the offence rule against drug trafficking and under the offence rule, there is a third condition that must obtain: the possession of the drug must be for the purpose of trafficking.
But the motivating reason for convicting the accused person for drug trafficking in the present case cannot include the fact that his or her possession was for the purpose of trafficking. As Alvarez (2016) has noted, «it seems undeniable that a person cannot act for the reason that p, or on the grounds that p, unless she stands in some epistemic relation to p: she needs to believe, know, accept etc that p» (section 3.2). The relevant offence rule sets out authoritatively the facts that must obtain for the commission of the offence. Guilt is not constituted merely by the probability of those facts obtaining. To convict a person for the reason that the person has committed the charged offence is to convict the person in the belief that the person is guilty as charged. We could also rest this view on the ground that conviction conveys blame, and blame implies belief 42. The necessary epistemic relation is lacking in our case. This is because the judge has concluded that one would not be justified in believing that the person’s possession was for the purpose of trafficking. As such, the judge cannot claim as part of his or her reason for convicting the accused that the accused’s possession of the drugs was for the purpose of trafficking 43. The reversal of the burden of proof for the third element is objectionable because it, in effect, allows—indeed, it may require—a criminal conviction to be motivated by a reason that falls short of the normative reason announced in the offence rule under which the person is convicted.
One might see the present situation as derogating from one of the formal conditions of legality articulated by Lon Fuller (1969) 44. This formal condition is consistency of laws, or the lack of contradictions in the laws. Under the offence rule, a person is guilty of drug trafficking only if the person’s possession of controlled drug was for the purpose of trafficking. However, because of the decision rule in section 17, the judge may find a person guilty of drug trafficking even if it is not proved that he or she had the drug for the purpose of trafficking. There appears to be an inconsistency in the laws.
Another of Fuller’s conditions, one which he considers «the most complex» (p. 81), with some tweaking, helps us to be more precise in identifying the inconsistency. It might be said of our situation that there is, to borrow some phrases from Fuller, a lack of «congruence between official action and declared rule» (p. 81) or «a failure of congruence between the rules as announced and their actual administration» (p. 39). But those phrases were applied by Fuller to situations in which the congruence is destroyed or impaired by mistaken interpretation of the law, corruption and so forth. Our situation is different; the judge is convicting the accused person in accordance with decision rules, and the decision rules (including the rule that provides for the presumption of purpose of trafficking) are also part of the law and as much ‘declared’ or ‘announced’ as the offence rules.
As we are supposing in our example, it is true that officials are following the rules that are addressed to them, rules that guide them in determining the verdict. The problem, however, lies in the fact that the state has laid down a decision rule that permits, indeed sometimes requires, officials to administer the offence rule—find the accused person guilty—for reasons that are insufficient under the offence rule. In short, the lack of congruence is between two sets of rules: the offence rule that defines the offence and a decision rule that guides officials in determining whether that offence has been committed. Fuller was not contemplating the kind of situation that I am talking about. But I think the objection that he has identified would extend to the situation that I have just described.
The decision rules are addressed to officials and not citizens. They are not aimed at guiding the behaviour of citizens. It is true that a decision rule like section 17 might influence the behaviour of citizens; it might prompt the Holmesian ‘bad man’ (Holmes, 1897) to reduce the quantity of diamorphine in his possession below 2 grammes in the hope of avoiding the presumption of trafficking should he be caught. But that is merely a possible incidental effect of section 17. The rule that guides the behaviour of citizens is the offence rule against drug trafficking, and its message is not to traffick in drug, period, in whatever quantity.
One may think of offence rules as authoritative definitions by the state of acts for which citizens may be convicted and punished. These offence rules are legal norms of which citizens have been given notice. They may be seen as laying down the terms of engagement between the state and citizens. Just as citizens are expected to obey the offence rules, the state is also expected to abide by them. The rule of law disciplines power by binding the state to its declared definitions of offences and by insisting that the law be such that judges may convict and punish citizens for an offence only for reasons that are sufficient under the rule addressed to citizens that authoritatively defines the offence 45.
For Raz (1979), a necessary feature of all legal systems is the existence of primary organs such as courts, tribunals and other judicial bodies. Primary organs are «concerned with the authoritative determination of normative situations in accordance with pre-existing norms» (p. 108). «[T]he norms by which the courts are bound to evaluate behaviour … are the very same norms which are legally binding on the individual whose behaviour is evaluated» (p. 112).
The legal system is not operating fully as Raz thinks it should when the legal burden of proof for an offence element is reversed 46. Rules that reverse the burden have the effect of requiring courts to determine the criminal liability of citizens in accordance with norms that are more prejudicial to citizens than the offence rules which the courts purport to be applying to citizens. Citizens may be convicted for reasons that are inadequate under the offence rules. To allow or require judges to convict and punish citizens for an offence for reasons that are inadequate according to the authoritative definition of that offence is to undermine the rule of law in this important respect: there is incongruence between the offence rules as announced to citizens and the decision rules that guide judges in administering the offence rules.
To conclude, let me return to the intuition-testing question that I pose in the introduction: should we be convicting people we don’t believe to be guilty? The answer, it turns out, is complicated. At least in common law systems, judges do not dispose of a criminal trial by simply answering the first-personal question: do I believe that the accused is guilty as charged? 47 They are instead required to make a third-personal judgment of whether one would be justified in believing—on the evidence adduced before the court and within legal constraints—that the accused is guilty as charged. If the judge concludes that one would not be justified in believing any fact which is necessary to constitute the accused’s guilt for the charged offence, the judge cannot claim as his or her reason for convicting the accused that he or she has committed the charged offence. The law should not allow the judge to convict the accused in these circumstances as his or her motivating reason would fall short of the normative reason needed for the conviction. Exceptionally, though, the law permits a lack of congruence between the motivating reason and the normative reason. This is unsettling because it undermines the rule of law. And that is why special justification is needed.
Achinstein, P. (1978). Concepts of Evidence. Mind, 87, 22-45.
Adler, J. E. (2002). Belief’s Own Ethics. MIT Press.
Alvarez, M. (2016). Reasons for Action: Justification, Motivation, Explanation. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy, (Winter 2017 Edition). https://plato.stanford.edu/archives/win2017/entries/reasons-just-vs-expl/
Blome-Tillmann, M. (2017). ‘More Likely Than Not’ Knowledge First and the Role of Bare Statistical Evidence in Courts of Law. In J.A. Carter, E.C. Gordon, and B. Jarvis (Eds.), Knowledge First: Approaches in Epistemology and Mind. Oxford University Press.
Buchak, L. (2014). Belief, Credence, and Norms. Philosophical Studies, 169, 285-311.
Cohen, L.J. (1991). Should a Jury Say What It Believes or What It Accepts? Cardozo Law Review, 13, 465-483.
Cohen, L.J. (1992). An Essay on Belief and Acceptance. Oxford University Press.
Damaška, M.R. (1997). Evidence Law Adrift. Yale University Press.
Dan-Cohen, M. (1984). Decision rules and conduct rules: on acoustic separation in criminal law. Harvard Law Review, 97, 625-677.
Duff, A., Lindsay F., Marshall S. and Tadros V. (2007). The Trial on Trial (Volume 3 – Towards a Normative Theory of the Criminal Trial). Hart Publishing.
Enoch, D., Spectre L. and Fisher, T. (2012). Statistical Evidence, Sensitivity, and the Legal Value of Knowledge. Philosophy and Public Affairs, 40, 197-224.
Ferrer Beltrán, J. (2004). ‘It is Proven that p’: The Notion of Proven Fact in the Law. Associations, 8, 29-54.
Ferrer Beltrán, J. (2006). Legal Proof and Fact Finders’ Beliefs. Legal Theory, 12, 293-314.
Ferrer Beltrán, J. (2021). Proven Facts, Beliefs, and Reasoned Verdicts. In C. Dahlman, A. Stein and G. Tuzet (Eds.), The Philosophical Foundations of Evidence Law. Oxford University Press.
Fuller, L. (1969). The Morality of Law (rev ed). Yale University Press.
Gama, R. (2021). The Search for El Dorado – The Rationalist Conception of Evidence and the Formulation of Precise and Objective Standards of Proof. Revus – Journal for Constitutional Theory and Philosophy of Law, 43. https://doi.org/10.4000/revus.7353
Gardiner, G. (2019). The reasonable and the relevant: legal standards of proof. Philosophy and Public Affairs, 47, 288-318.
Gardiner, G. (2024). Legal Evidence and Knowledge. In C. Littlejohn and M. Lasonen-Aarnio (Eds.), The Routledge Handbook of the Philosophy of Evidence. Routledge.
Günther, M. (2024). Legal Proof Should be Justified Belief of Guilt. Legal Theory, 30, 129-141.
Ho, H. L. (2008). A Philosophy of Evidence Law. Oxford University Press.
Ho, H. L. (2015). The Legal Concept of Evidence. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy, (Winter 2021 Edition). https://plato.stanford.edu/archives/win2021/entries/evidence-legal/
Ho, H. L. (2021). Justification, Excuse, and Proof beyond Reasonable Doubt. Philosophical Issues, 31, 146-166.
Holmes, O.W. (1897). The Path of the Law. Harvard Law Review, 10, 457-478.
Lackey, J. (2021). Norms of Criminal Conviction. Philosophical Issues, 31, 188-209.
Lehrer, K. (2000a). Acceptance and Belief Revisited. In P. Engel (Ed.), Believing and Accepting. Kluwer.
Lehrer, K. (2000b). Theory of knowledge (2nd ed). Westview Press.
Leiter, B. (1997). The Epistemology of Admissibility: Why Even Good Philosophy of Science Would Not Make for Good Philosophy of Evidence. Brigham Young University Law Review, 803-819.
Littlejohn, C. (2020). Truth, Knowledge, and the Standard of Proof in Criminal Law. Synthese, 197, 5253-5286.
Littlejohn, C. (2021). Justified Belief and Just Conviction. In Z. Hoskins and J. Robson (Eds.), The Social Epistemology of Legal Trial. Routledge.
McBride, M. (2011). Reply to Pardo: Unsafe Legal Knowledge. Legal Theory, 17, 67-73.
McHugh, C. (2011). Judging as a Non-voluntary Action. Philosophical Studies, 152, 245-269.
Moss, S. (2018). Probabilistic Knowledge. Oxford University Press.
Moss, S. (2023). Knowledge and Legal Proof. In T.S. Gendler, J. Hawthorne and J. Chung (Eds.), Oxford Studies in Epistemology. Oxford University Press.
Murphy, C. (2005). Lon Fuller and the Moral Value of the Rule of Law. Law and Philosophy, 24, 239-262.
Nance, D. (2021). Truth, Justification, and Knowledge in the Epistemology of Adjudication. In B. Zhang, S. Tong, J. Cao and C. Fan (Eds.), Facts and Evidence – A Dialogue Between Philosophy and Law. Springer.
Nelkin, D. (2021). Rational Belief and Statistical Evidence – Blame, Bias, and the Law. In I. Douven (Ed.), Lotteries, Knowledge, and Rational Belief. Cambridge: Cambridge University Press.
Papineau, D. (2021). The Disvalue of Knowledge. Synthese, 198, 5311-5332.
Pardo, M.S. (2010). The Gettier Problem and Legal Proof. Legal Theory, 16, 37-57.
Pardo, M.S. (2018). Safety vs. Sensitivity: Possible Worlds and the Law of Evidence. Legal Theory, 24, 50-75.
Pritchard, D. (2015). Risk. Metaphilosophy, 46, 436-461.
Pritchard, D. (2018). Legal Risk, Legal Evidence and the Arithmetic of Criminal Justice. Jurisprudence, 9, 108-119.
Pritchard, D. (2022). In Defence of the Modal Account of Legal Risk. Synthese, 200, 290.
Raz, J. (1979). The Authority of Law. Oxford University Press.
Ross, L. (2022). The Foundations of Criminal Law Epistemology. Ergo, 9, 1581-1603.
Ross, L. (2024). The Philosophy of Legal Proof. Cambridge University Press.
Schauer, F. (2022). The Proof – Uses of Evidence in Law, Politics, and Everything Else. Harvard University Press.
Smith, M. (2018). When Does Evidence Suffice for Conviction? Mind, 127, 1193-1218.
Stephen, J.F. (1872). The Indian Evidence Act (I. of 1872): with an Introduction on the Principles of Judicial Evidence. Thacker, Spink & Co.
Thayer, J.B. (1898). A Preliminary Treatise on Evidence at The Common Law. Little, Brown & Co.
Thomson, J.J. (1986). Liability and Individualized Evidence. Law and Contemporary Problem, 49, 199-219.
Ullmann-Margalit, E. (1983). On Presumption. The Journal of Philosophy, 80, 143-163.
Ullmann-Margalit, E. and Margalit, A. (1992). Holding True and Holding as True. Synthese, 92, 167-187
Whiteley, J.H.L. (2024). The Three-Verdict Problem. Legal Theory, 30, 105-127.
Williams, G. (1980). A Short Rejoinder. Criminal Law Review, 103-107.
Wittlin, M. (2023). Theorizing Corroboration. Cornell Law Review, 108, 911-992.
Alma Nudo Atenza v Public Prosecutor [2019] 4 MLJ 1.
Attorney General of Hong Kong v Siu Yuk-shing [1989] 1 WLR 236.
Canadian Charter of Rights and Freedoms (1982). https://laws-lois.justice.gc.ca/eng/const/page-12.html
Center for Competency-Based Learning and Development Pte Ltdv SkillsFuture Singapore Agency [2024] SGHC 121. https://www.elitigation.sg/gd/s/2024_SGHC_121
Center for Competency-Based Learning and Development Pte Ltdv SkillsFuture Singapore Agency [2024] 5 SLR 481.
Federal Rules of Evidence (United States). https://www.uscourts.gov/sites/default/files/federal_rules_of_evidence_-_dec_1_2019_0.pdf
Indian Evidence Act 1872. https://www.indiacode.nic.in/bitstream/123456789/15351/1/iea_1872.pdf
Jayasena v R [1970] 2 WLR 448.
HKSAR v Lam Kwong Wai & Anor (2006) 9 HKCFAR 574.
HKSAR v Hung Chan Wa (2006) 9 HKCFAR 614.
Lim Wei Fong Nicman v PP [2024] 1 SLR 1041.
Lim Wei Fong Nicman v PP [2024] SGCA 33. https://www.elitigation.sg/gd/s/2024_SGCA_33
Loo Chay Sit v Estate of Loo Chay Loo [2009] SGCA 47. https://www.elitigation.sg/gd/s/2009_SGCA_47
Loo Chay Sit v Estate of Loo Chay Loo [2010] 1 SLR 286.
Makin v Attorney-General for New South Wales [1894] AC 57. https://vlex.co.uk/vid/makin-v-attorney-general-803355473
Masoud Rahimi bin Mehrzad v PP [2017] 1 SLR 257.
Mohamed Shalleh bin Abdul Latiff v PP [2022] SGCA 23. https://www.elitigation.sg/gd/s/2022_SGCA_23
Muhammad Ridzuan bin Md Ali v PP [2014] 3 SLR 721.
Perjury Act 1911. https://www.legislation.gov.uk/ukpga/Geo5/1-2/6/contents
PP v Wee Teong Boo [2020] SGCA 56. https://www.elitigation.sg/gd/s/2020_SGCA_56
PP v Wee Teong Boo [2020] 2 SLR 533.
R v Mohamed Humayoon Shah [1874] 21 WR Cr 72.
R v Oakes (1986) 26 DLR (4th) 200.
Ramesh a/l Perumal v Public Prosecutor [2019] 1 SLR 1003.
Road Traffic Regulation Act 1984. https://www.legislation.gov.uk/ukpga/1984/27/contents
Singapore Criminal Procedure Code 2010 (2020 rev. ed.). https://sso.agc.gov.sg/Act/CPC2010
Singapore Evidence Act 1893. https://sso.agc.gov.sg/Act/EA1893
Singapore Misuse of Drugs Act 1973. https://sso.agc.gov.sg/Act/MDA1973
Singapore Penal Code 1871. https://sso.agc.gov.sg/act/pc1871
Singapore Sedition Act.
https://sso.agc.gov.sg/Act-Rev/SA1948/Published/20130831?DocDate=19870330
Sheldrake v Director of Public Prosecutions [2005] 1 AC 264.
Subramaniam v Public Prosecutor [1956] 1 WLR 965.
* This article is based on a paper presented at the workshop on ‘Knowledge and Legal Standards of Proof’ held in November 2024 at the Faculty of Law, University of Girona and co-hosted by the Barcelona Institute of Analytical Philosophy. I am grateful to the organisers, Professors Jordi Ferrer Beltrán and José Juan Moreso, and Dr Diego Dei Vecchi, for their invitation, and to the workshop participants for their valuable feedback and discussion. Professor Vincent Chiao offered helpful comments on an earlier and shorter draft of this paper that forced me to clarify my thoughts. I also thank the anonymous reviewers for their queries and suggestions.
1 Among the contenders for the necessary epistemic property are ‘safety’ (eg, Pritchard, 2015 and 2018 and 2022; Pardo, 2018), ‘sensitivity’ (cfr. Enoch, Spectre and Fisher, 2012—claiming that sensitivity is of epistemological importance but denying that it should matter to the law), and ‘causal connection’ (eg, Thomson, 1986). This is just a very small sampling of a vast and still growing body of literature.
2 It is less clear to me what the usual reaction would be if the question is whether we should convict people we don’t know to be guilty.
3 See also Ross (2024, pp. 13-16, 41-44).
4 See s.16 of the Singapore Penal Code 1871.
5 This is, in broad outline, the analysis of the elements of the offence of drug trafficking adopted in Singapore. See, eg, Mohamed Shalleh bin Abdul Latiff v PP [2022] SGCA 23 at [21] citing Masoud Rahimi bin Mehrzad v PP [2017] 1 SLR 257. See also Muhammad Ridzuan bin Md Ali v PP [2014] 3 SLR 721 at [59]; Ramesh a/l Perumal v PP [2019] 1 SLR 1003 at [63].
6 Section 124(1) of the Singapore Criminal Procedure Code 2010 (2020 rev. ed.) states: «The charge must contain details of the time and place of the alleged offence and the person (if any) against whom or the thing (if any) in respect of which it was committed, as are reasonably sufficient to give the accused notice of what the accused is charged with». See also PP v Wee Teong Boo [2020] 2 SLR 533; [2020] SGCA 56 at [101], citing R v Mohamed Humayoon Shah [1874] 21 WR Cr 72 at 82: «The charge [is] first, a notice to the prisoner of the matter whereof he is accused, and it must convey to him with sufficient clearness and certainty that which the prosecution intends to prove against him and of which he will have to clear himself; second, it is an information to the Court which is to try the accused, of the matters to which evidence is to be directed».
7 It is from this point of view that we can make best sense of Williams’s remark that ‘[t]here is a miscarriage or failure of justice whenever an innocent man is convicted;… to acknowledge this is not necessarily to imply a criticism of the law or of the trial’ (Williams, 1980, p. 104).
8 Other options, such as a stay of criminal proceedings and a discharge not amounting to an acquittal (nolle prosequi), are atypical.
9 But a guilty verdict is not necessary for a conviction; for instance, the accused may be convicted following a guilty plea.
10 See Whiteley (2024).
11 See, eg, s.3 of the Indian Evidence Act 1872 and s.3 of the Singapore Evidence Act 1893, discussed in Loo Chay Sit v Estate of Loo Chay Loo [2009] SGCA 47; [2010] 1 SLR 286 and Center for Competency-Based Learning and Development Pte Ltdv SkillsFuture Singapore Agency [2024] SGHC 121; [2024] 5 SLR 481.
12 Cf. Lim Wei Fong Nicman v PP [2024] 1 SLR 1041; [2024] SGCA 33.
13 See Damaška (1997, pp.74-75).
14 Under s.267(1) of the Singapore Criminal Procedure Code 2010, an admission of a fact under that provision is treated as ‘conclusive evidence’ of the fact admitted. Under s.4(3) of the Singapore Evidence Act 1893, «[w]hen one fact is declared by this Act to be conclusive proof of another, the court is, on proof of the one fact, to regard the other as proved, and is not to allow evidence to be given for the purpose of disproving it.»
15 Moss claims that «[c]onviction requires proving beyond a reasonable doubt that the defendant is guilty, and this conclusion is proved if and only if the judge or jury knows it» (Moss, 2023, p. 177). For her, probabilistic beliefs can constitute knowledge. For the knowledge that is required by standard of proof beyond a reasonable doubt, «[a] compelling conjecture is that the standard requires knowledge of the thoroughly probabilistic content of the full belief that the defendant is guilty» (Moss, 2018, p. 213).
16 Blome-Tillmann (2017) takes the position that knowledge is «the normatively ideal state, but epistemic success in courts of law can be measured and understood in terms of something less than knowledge—namely, the evidential probability that knowledge has been achieved» (p. 284).
17 According to Littlejohn: «[w]hether the jury objectively ought to convict depends upon what they can know. In another sense, the evidence suffices for convicting in ways we prospectively ought to when it is sufficiently probable that we know the defendant to be guilty» (Littlejohn, 2021, pp. 118-119).
18 Skepticisms on the relevance of knowledge to legal proof have been expressed on different grounds by some philosophers, including Enoch, Spectre and Fisher, 2012), Gardiner (2024) and Papineau (2021).
19 On potential difficulties with an account of legal proof that incorporates safety as a necessary condition of knowledge, see McBride (2011).
20 See eg, Cohen (1991), Ferrer Beltrán (2004), (2006) and (2021), and Lackey (2021).
21 See eg, Cohen (1991), Ferrer Beltrán (2004, 2006 and 2021).
22 See eg, Ferrer Beltrán (2004, 2006 and 2021).
23 These are only these two standards in most common law jurisdictions. In the United States, there is a third standard of ‘clear and convincing evidence’.
24 To borrow an example by McHugh (2011), I cannot judge that p—in the doxastic or epistemic sense of judgment—just for the reason that I want to get the million-dollar reward for doing so.
25 For my earlier attempt at supporting the thesis, see Ho (2008, pp. 205-207).
26 See Lehrer (2000b, p. 14).
27 The concept of evidence has received philosophical treatment that gives it a meaning unfamiliar to lawyers: see eg, Schauer (2022, pp 26-27), citing Achinstein (1978), and criticizing the idea that evidence is only potential evidence unless the conclusion it supports is true. ‘Evidence’ in the sense familiar to lawyers and as used in this paper refer simply to that which is adduced by a party in legal proceedings as a means of proving factual claims and it may take the form of oral testimony, documents and other objects (‘real evidence’). Evidence is ‘admissible’ if the law allows it to be received in legal proceedings. See generally, Ho (2015).
28 Unless it is a commonly known fact of which ‘judicial notice’ may be taken.
29 This was not always the case. The romano-canonical system of proof was, on a popular account, artificially regulated by the law, with rules stipulating the number of witnesses needed for proof and assigning specified weights to different types of evidence. See references in Ho (2008, pp. 39-40).
30 Cap. 290, 1985 Rev. Ed. Sing. For other examples in the United Kingdom, see s. 13 Perjury Act 1911 and s. 89(2), Road Traffic Regulation Act 1984, and in the US, see the discussion of corroboration sufficiency rules in Wittlin (2023, pp. 976-979).
31 Arguably, rules of admissibility are better conceived not as rules that exclude evidence but as rules of evidential reasoning.
32 Subramaniam v PP [1956] 1 WLR 965.
33 See Makin v Attorney-General for New South Wales [1894] AC 57, 65, per Lord Herschell:
It is undoubtedly not competent for the prosecution to adduce evidence tending to shew that the accused has been guilty of criminal acts other than those covered by the indictment, for the purpose of leading to the conclusion that the accused is a person likely from his criminal conduct or character to have committed the offence for which he is being tried. On the other hand, the mere fact that the evidence adduced tends to shew the commission of other crimes does not render it inadmissible if it be relevant to an issue before the jury, and it may be so relevant if it bears upon the question whether the acts alleged to constitute the crime charged in the indictment were designed or accidental, or to rebut a defence which would otherwise be open to the accused.
34 See Rule 404(b)(2) of the Federal Rules of Evidence in the United States: «This evidence may be admissible for another purpose, such as proving motive, opportunity, intent, preparation, plan, knowledge, identity, absence of mistake, or lack of accident».
35 A-G of Hong Kong v Siu Yuk-shing [1989] 1 WLR 236.
36 On epistemic paternalism and the law of evidence, see Leiter (1997).
37 They were originally formulated as definitions of proof and disproof respectively. I am grateful to Professor Timothy Williamson for pointing out my error at the workshop mentioned in the acknowledgments.
38 It is of course possible to claim that p is proved when p is false. When the claim is wrongly made, p is not proved even though the claimant may think that it is.
39 For a similar suggestion, see Moss (2023, pp. 204-206).
40 According to Ullmann-Margalit and Margalit (1992, p. 171), where «a person is instructed to hold something as true, ie, to act as if that something were true», the «question of whether the person actually believes that it is true does not arise».
41 Similarly, see Jayasena v R [1970] 2 WLR 448 at 452: an evidential burden ‘can be discharged by the production of evidence that falls short of proof.’
42 See Adler (2002, pp. 216-217), Buchak (2014), Littlejohn (2020, pp. 5268-5269), Nelkin (2021, p. 24-26).
43 In a similar vein, see the account given by Littlejohn (2020, pp. 5275-5276).
44 Fuller describes these formal conditions (for some, inaptly) as desiderata of the internal morality of law.
45 For a persuasive discussion in this connection, see Murphy (2005).
46 Raz’s remarks were not made in relation to the reversal of the burden of proof.
47 On the debate on whether, and the extent to which, this is also true in continental law, see Gama (2021).