1. Introduction

1In 1880 Enrico Ferri, one of the most famous personalities of the Positive School of Criminology, wrote a critical essay on the trial by jury in the Italian system.¹ When Ferri published his study, the institute of the jury in Italy had acquired thirty years of experience and was gathering a growing consensus among the jurists, despite some undeniable drawbacks. Ferri did not share the general enthusiasm towards lay participation in the administration of criminal justice: in his opinion, indeed, the jury had unequivocally revealed more deficiencies than merits, thus the only reasonable solution was to abolish it once and for all.² Abolition, however, was out of the question because, at that time, the jury in Italy was considered a palladium of liberty against the strictness of the law and the severity of the judges.³ The author, therefore, fell back on a more modest program and proposed specific amendments, in the hope of mitigating the worst weaknesses of the system. One of the proposals was the overcoming of the binary model acquittal/conviction and the introduction of a third form of verdict, to be used when the accused was burdened by suspicion, but the prosecution had failed to prove guilt beyond reasonable doubt. According to Ferri, a defendant seriously suspected of being guilty was not entitled to a positive declaration of innocence: in such a case the accused deserved a verdict of not proven, following the example of the Scottish juries.⁴

2The Scottish criminal jury evoked by Enrico Ferri represents «a very peculiar institution» in the jury world.⁵ The Scottish system, indeed, is characterized by some distinctive features, which move the interest of scholars nowadays just as they did in the past.First of all, juries in Scotland are made up of fifteen members, while traditionally criminal juries consist of twelve persons. Secondly, in the Scottish experience juries issue their verdicts by a simple majority of votes, thus marking a significant difference with the English courts of justice, which adopted the rule of unanimity until a few decades ago.⁶ Lastly – and this is the central point of our investigation – Scottish juries can choose among three different verdicts, that is guilty, not guilty, and not proven, while usually the jurors are bound to choose between the guilt or innocence of the defendant.

3The three-verdict system is the most peculiar aspect of the trial by jury in Scotland.⁷ In Scottish criminal law, not proven and not guilty are both acquittals and have the same legal consequences. Ordinarily, not proven comes into play when the guilt of the accused has not been adequately demonstrated by the prosecution.⁸ In recent years, this phenomenon has attracted a growing attention and became the subject of an intense academic debate.⁹ The goal of this research is to examine such a debate between nineteenth and early twentieth centuries, by considering four famous murder trials resulting in a verdict of not proven. The trials of Christina Gilmour, Madeline Smith, Alfred John Monson, and John Donald Merrett became causes célèbres in their time and gave rise to a passionate discussion, whose echo is felt to the present day.¹⁰

4In the wake of the reflection proposed by Enrico Ferri, the Scottish experience will be compared with the Italian jury system over the same period of time. The decision-making procedure in the Italian courts of assize was profoundly different from what took place in Scotland: Italian jurors struggling with circumstantial evidence were required to choose between guilt or innocence, without the possibility of finding the charge not proven. As a consequence, whenever the prosecution failed to prove the case, the defendant was to be acquitted according to the principle in dubio pre reo. In this regard, we will examine the trial of Vito Modugno, which took place at the court of assize of Perugia in the early twentieth century.

5The Italian system radically changed in March 1931, when the trial by jury was abolished by the royal decree n. 249 and replaced by mixed courts, in which professional magistrates and lay judges had to work side by side and decide together on both verdict and punishment.¹¹ A famous trial held at the court of assize of Sassari a few years later reveals how much the procedural rules had changed under the Fascist Regime. Despite the abolition of the trial by jury in the Italian jurisdiction, a comparative perspective is still possible: in the second half of the twentieth century, indeed, the possibility to acquit the defendant for insufficient evidence was the subject of discussion in both Scotland and Italy. As we will see in the conclusion, however, such a debate finally led to different outcomes.

2. The origin and development of the not proven verdict in Scotland

6In 1949 Sir Alfred Denning wrote that a person accused of a crime must either be found guilty or not guilty, because «the freedom of no one of us is to be taken away on a compromise». From his point of view, the verdict of not proven was an unacceptable middle ground.¹² Originally, however, the not proven verdict did not represent a compromise between guilt or innocence: as written by Ian Douglas Willock more than fifty years ago, the appearance of the «bastard verdict» in the Scottish jury experience «was a pure historical accident».¹³

7In the early modern period, in fact, Scottish juries were bound to choose between ‘guilty’ or ‘not guilty’ just like English juries.¹⁴ In the late seventeenth century, however, a new practice arose and gradually replaced the previous custom. At the time of the Restoration, indeed, jurors were increasingly dealing with indictments of considerable complexity in which the facts were set out with circumstantial minuteness. In such cases the jury was required to examine the facts and find them ‘proven’ or ‘not proven’, leaving it to the judge to determine if those facts did or did not amount to the offence charged. This new practice usurped the prerogative of the jury by introducing a distinction between probation and relevancy.¹⁵ Such a development was welcomed by Sir George Mackenzie, who was not a supporter of jury trials. In The Laws and Customs of Scotland in Matters Criminal, the famous Scottish lawyer wrote that «the Justices are only Judges to the relevancy, and Assizers to the probation; yet to distinguish the limits of their different cognitions becomes very oft difficult»: in order to prevent this complication, the Lord Advocate wished that «the Justices were Judges both to relevancy and probation».¹⁶ It is interesting to note that the possibility, or the lack of it, to separate the factual question from its legal qualification was one of the most debated issues in nineteenth century Italy: like Sir George Mackenzie, many Italian authors believed that the distinction between the jurors and the judges’ duties could be quite complicated because fact and law represent «two sides of the same coin».¹⁷

8As a result of this new practice, the form of the verdicts issued by Scottish juries was substantially altered: the choice between ‘guilty’ or ‘not guilty’ was replaced by the alternative ‘proven’ or ‘not proven’, and the role of the jurors in criminal matters was reduced by a significant extent. In the first half of the eighteenth century, however, the juries claimed their right to return a verdict of not guilty and ancient customs were restored.¹⁸ Nevertheless, contrary to what might be expected, the verdict of not proven did not disappear, but took on a different connotation as pointed out by David Hume in his Commentaries on the Law of Scotland respecting Crimes first published in 1797.

9Hume explained the difference between not proven and not guilty as follows: the phrase not proven is used to mark a deficiency of lawful evidence to convict the panel, while the phrase not guilty is used to convey the opinion that the accused is «innocent of the charge».¹⁹ Not proven and not guilty, therefore, are both acquittals but they have a profoundly different meaning. They also have different consequences on the reputation of the acquitted person: if the jury returns a verdict of not guilty, the accused is considered innocent in all respects and his (or her) name is fully rehabilitated; by contrast, whenever a jury returns a verdict of not proven, the defendants receive a sort of black mark and their report is irreparably compromised. When the charge has been found not proven, indeed, the accused, «goes away from the bar of the court with an indelible stigma upon his fame, when the stands recorded against him the opinion of a jury that the evidence respecting his guilt was so strong that they did not dare to pronounce a verdict of acquittal».²⁰

10This indelible stigma is a consequence of the idea mentioned above: the accused has been acquitted only because the prosecution has failed to prove his or her guilt beyond any reasonable doubt.²¹ The idea that the guilt must be proved beyond reasonable doubt did not prevent Scottish juries from adopting their verdicts by a simple majority of votes: as pointed out by Gerry Maher, the principle of reasonable doubt and the rule of majority are not mutually exclusive.²² From this point of view, the Scottish jury differs from the English model. English juries indeed were required to find a unanimous consensus upon the guilt or innocence of the defendant: unanimity was considered a protection for the accused, who could only be convicted if no one among the jurors entertained any doubt about his guilt.²³ In Scotland a comparable protection was offered by the verdict of not proven. Nonetheless, in a three-verdict system, the rule of majority could bring a measure of uncertainty as to the final sentence, since the foreman did not have to report the precise outcome of the voting process. For instance, in a trial held at a sheriff court in 1958, the jury issued a verdict of guilty by a majority. With the permission of the sheriff, the defence counsel inquired the state of the voting: it turned out that six jurors voted for guilty, five for not guilty, and four for not proven. Actually, therefore, the majority of the jurors was in favour of the acquittal. As a consequence, the sheriff ordered that a verdict of not guilty should be reported.²⁴

11As specified by Ian Douglas Willock, the verdict of not proven should not be confused with the verdict of non liquet: the not proven is a form of acquittal used by juries when the guilt of the accused has not been satisfactorily demonstrated, while non liquet means that the jurors declined to vote due to a lack of information on the facts. In the Scottish system abstention seems to have been a tolerated practice, although «no examples of individual assizers delivering non liquet have been found».²⁵ As we will see hereafter, also Italian jurors could decline to vote by leaving their voting-paper blank, albeit this faculty was widely contested by some authors, among whom Enrico Ferri.²⁶

12As we have just seen, the verdict of not proven represents a ‘second-class’ acquittal leaving a black mark on the defendant. This is the reason why in popular perception the not proven is considered a compromise between guilty and not guilty and it is colloquially called «the bastard verdict». This term was coined by Sir Walter Scott after learning the verdict issued by the jurors in the trial of Mrs Mary Elder or Smith, which took place at the High Court of Justiciary in February 1827.²⁷

13Mrs Mary Smith was a 42 years old woman married to a farmer from West Denside in the county of Forfar. Mr and Mrs Smith had two sons and two daughters. They shared their house with their two servants, Jean Norrie and Margaret Warden. Margaret was a widow of 24 years who lived in poor conditions with her three young children. In July 1826, Mrs Smith found out with great disappointment that her youngest son George and Margaret felt an affection for each other. One night in early September, Jean and Margaret were sitting in the kitchen when Mrs Smith came into the room and offered to the girls a glass full of a thick white mixture. Two days later Margaret fell seriously ill, while Jean suffered no ill effects because she had barely tasted the liquid. On Friday afternoon a doctor came to visit Margaret who was very sick: his diagnosis was that she was dying due to cholera. Margaret actually passed that night.

14The body was buried a few days later, but soon after a rumour spread throughout the countryside that «the dead girl had been poisoned by her mistress to avert the consequences of a liaison with her son».²⁸ Thus the body was exhumed and submitted to an autopsy: the post mortem examination revealed traces of acute inflammation. Questioned by the sheriff, Mrs Smith admitted she had given Margaret a dose of castor oil, but claimed that she kept no poisonous substances in the house. The next morning, however, Mary confessed that she had bought «something to put away rats» the Friday before Margaret’s death. Thereupon, the woman was officially indicted for murder.²⁹

15After a few postponements, the trial finally started in February 1827. The prosecution had no doubt: Margaret Warden had died by poison. The main question was: did she poison herself or had she been poisoned by the defendant? Both evidence and witnesses – said the Lord Advocate – excluded any idea of suicide. According to the prosecution, the poison had been administered by the defendant because she believed that the liaison between her son and the victim could bring disgrace and damage the good name of the family. On the other side, the defence claimed that Margaret had died by cholera. Even assuming the rat poison was the cause of death, there was no proof against the prisoner. In a case «so involved in mystery», the defence counsel did not ask jurors for «a triumphal acquittal», but demanded a verdict of not proven.³⁰ The not proven verdict was thus invoked «to mark a deficiency of lawful evidence to convict the panel», exactly as noted 30 years before by David Hume.³¹

16In his summing up, the Lord Justice Clerk specified that the case was one of «circumstantial evidence».³² Two questions were submitted to the consideration of the jury: the corpus delicti and the guilt of the prisoner. If the jurors found themselves convinced that poison was the cause of death, they had to turn upon the following question: by whom had it been administered? The jurors were bound to issue a verdict of guilty if they were fully satisfied that the poison had been administered by the prisoner at the bar; by contrast, if they had any doubts, they should give the accused the benefit of them.

17The jurors unanimously returned a verdict of not proven. The deliberation was unpopular: as anticipated, it is on that occasion that Sir Walter Scott coined the term «bastard verdict». The author found such a Caledonian medium quid a detestable outcome: if the jurors found the charge not proven, the only possible solution should have been a verdict of not guilty, because «one who is not proven guilty is innocent in the eye of law».³³ The author, personally, considered Mrs Mary Elder guilty beyond any reasonable doubt. Some years later, in fact, Lord Cockburn remembered Scott’s remark upon this acquittal: «if that woman was my wife, I should take good care to be my own cook».³⁴

18The trial of Mary Elder exemplifies the traditional use made by Scottish juries of the not proven verdict. Such a deliberation, however, may be issued also when «the jury knows perfectly well that the accused is guilty but thinks that the law needs to be tempered with mercy».³⁵ Trials against abused women charged with the murder of their violent husbands are a prime example of this attitude: in such situations the jury may return a verdict of not proven «presumably because it sees no purpose in further punishing the accused».³⁶ In these particular cases, not proven could be employed in order to assert leniency against the strictness of the law. The same need was felt by Italian jurors in a number of different situations. The classic example are infanticide trials when the defendants were poor, young women who had killed their newborn children to safe their honour and reputation.³⁷ As noted by Raffaello Balestrini in 1888, in infanticide cases the rigour of the legislation often found correction in an overindulgence on the part of the juries, with the result that the law’s threats resulted in a «vain bugbear».³⁸ Italian jurors, however, did not have a third verdict option: in such situations, therefore, they fulfilled their feeling of benevolence by returning a verdict of not guilty, despite any evidence given in the trial against the defendant.

3. Lady poisoners: Christina Gilmour and Madeline Smith

19The trials of Christina Gilmour (1844) and Madeline Smith (1857) were two causes célèbres in the mid-nineteenth century Scotland. Both the women were tried on charge of murder by poisoning: in both cases the charge was found not proven. Christina and Madeline were accused of killing their victims by arsenic poison. The nineteenth century has been «the golden age of arsenic murders» in Victorian Britain.³⁹ Arsenic was put everywhere: in food, drinks, clothing, candles, even in cake decorations. As we have seen in the trial of Mrs Mary Elder, people who bought arsenic usually said they would use it to kill off rats: evidently, there was a multitude of rats in Victorian England. Sometimes the purchase was made in great secrecy or through a third party, as happened in the case of Christina Gilmour.

20Christina Gilmour was the daughter of a wealthy family with a good social position. The woman had a secret love relation with an older man, John Anderson, son of a farmer at Broadley. In 1842 a new suitor came forward, John Gilmour, considered by Christina’s parents an excellent prospect. The woman confessed to Anderson that she was engaged with another man: contrary to her expectations, Anderson gave up their love. Hence, all Christina could do was marry John Gilmour. The wedding was celebrated on 29 November 1842.⁴⁰

21The facts under investigation took place in December, less than a month later the wedding. On 26 December Christina asked a servant, Mary Paterson, to buy arsenic to deal with the rats: Mary was supposed to send somebody else to make the purchase; by contrast, she went personally to the shop and said she was buying arsenic for Mrs. Gilmour. Mary then delivered the package to her mistress and told her how she had bought it. The next day Christina, in Mary’s presence, threw a package – apparently the same package she had received the previous day – into the boiler because «it would be of no use to her».⁴¹ In the following days the facts unfolded rapidly. On 29 December John Gilmour fell ill. His wife took care of him and did not move away from his bedside until Friday 6 January, when she left home for a few hours. That evening a servant found in a black bag a vial with the word ‘poison’ written upon it: Christina said she had purchased some turpentine to treat her husband. On Sunday a doctor visited the patient and found him in a bad state of health. On Wednesday the man was in critical conditions. Before death came, Gilmour seemed to have said ‘Oh that woman! If you have given me anything’, but it was not clear who he was referring to.⁴² Shortly thereafter, the man died.

22The corpse was buried without an autopsy and Christina returned to live with her parents. Soon after, however, a rumour spread in the community: John may have been poisoned by his wife. In April, therefore, the corpse was exhumed and submitted to a post-mortem examination from which it emerged that the man had in fact been poisoned. In the meantime, Christina had boarded a ship to America. Superintendent George McKay sailed to America himself and tracked down the woman: they returned to England in early September, following extradition. As soon as she returned, the suspect was imprisoned and questioned by the sheriff: the woman admitted she bought arsenic before her husband’s death, however her intent was not to kill him but to kill herself. Christina remained in prison until the beginning of the trial, which took place at the High Court of Justiciary in January 1844.

23In court, the Crown prosecutor focused the attention of the jurors on two questions. The first concerned the cause of death. On this point there could be no doubt: the man had been poisoned by arsenic. The second concerned the mode by which the poison had been administered. In this regard, the jury had to carefully examine the circumstances of the case and consider three possibilities: suicide, accident, or murder. According to the prosecution, the poison had been administered by the accused at the bar: first of all, Christina had admitted to having bought arsenic; secondly, she had the opportunity to poison the victim because she stayed at his bedside all the time; thirdly, the woman wanted to put an end to her marriage, thus she had a motive to commit the crime. On the other side, the defence counsel claimed that there was no evidence against the accused, but only suspicions and conjectures.⁴³ The proof given in the trial, indeed, was not so strong as to exclude a reasonable possibility of suicide or accident. As pointed out by the lawyer, the case was of purely circumstantial evidence: the defence counsel, therefore, asked the jurors to return a verdict of not proven.⁴⁴

24After the speeches of the parties, the Lord Justice Clerk proceeded to charge the jury. The judge encouraged the jurors to fulfil their duty regardless of the consequences and without hesitation. If they were satisfied that John Gilmour died from the effects of poison, they had to consider if the poison got into his body either accidentally or intentionally. Intentional administration embraced two distinct hypothesis – suicide or homicide – both of which had to be carefully examined by the jury. The judge concluded his summing up as follows: «if you think… that there are still mysteries unexplained, I will not tell you that you must give, for I know that you will give, the full benefit of that doubt or obscurity to the individual who is charged with the perpetration of this dreadful crime».⁴⁵

25When the summing up was ended, the jurors retired to deliberate. An hour afterwards, they returned the following verdict: «the jury, after careful and mature deliberation of the evidence brought before them in this case, are unanimously of opinion that John Gilmour died from the effects of arsenic, but they find that the charge is not proven against the pannel at the bar as libelled».⁴⁶ Therefore, as had happened 17 years earlier in the proceeding against Mrs. Mary Elder, the jurors believed that the evidence given in the trial was defective and the guilt of the woman had not been satisfactorily demonstrated.

26The trial of Christina Gilmour presents a great number of similarities with the case of Madeline Smith, which is one of the most famous Scottish trials in the long nineteenth century. As noted by John Gray Wilson, the story of Madeline Smith cannot be omitted in any research study focused on the verdict of not proven.⁴⁷ The story begins in April 1855 when Madeline, a girl of 19 years, met the man who would become her lover, Pierre L’Angelier. Pierre and Madeline developed an intense correspondence and in June 1856 they also had a sexual intercourse. At that time, however, Madeline was involved with another man, William Harper Minnoch. In the following months the correspondence between the young woman and her lover continued, but Madeline was losing interest in the relation. In January 1857 she accepted Minnoch’s wedding proposal. On 21 February the woman bought an ounce of arsenic, purportedly to kill rats. On 23 February L’Angelier suffered a severe illness. In March Madeline bought twice again a quantity of arsenic. On 23 March Pierre L’Angelier died. The preliminary report on the post-mortem examination revealed that arsenic poison could have been the cause of death. Suspicions immediately fell on the woman who was arrested a few days later.

27The trial took place at the High Court of Justiciary from 30 June to 9 July 1857. The woman was put on trial on three charges: the first and the second was the administration of arsenic with intent to poison the victim in two separate occasions, on 19/20 and on 22/23 February; the third charge was murder of the victim by the same means on 22/23 March 1857.

28In order to prove the guilt of the woman, the prosecution was required to demonstrate that the illness proceeded from the administration of arsenic and that Madeline met L’Angelier on those days, because, if a meeting had not taken place, the man could not have been poisoned by the defendant. About the first charge the judge had no hesitation in telling the jurors that the prosecution had failed: the Lord Advocate had not proved that on that date Madeline had poison in her possession, neither there was any medical evidence that the illness had been caused by arsenic.⁴⁸ The judge had doubts also about the second charge, but proof of a meeting on that date was much stronger, thus he did not give explicit directions on this point. In the third charge, the prosecution was required to prove beyond reasonable doubt that Madeline met Pierre on 22/23 March 1857 and poisoned him. According to the defence counsel, the evidence on this matter was absolutely defective: the solicitor general, indeed, could not rule out that L’Angelier had ingested arsenic either accidentally or intentionally for the purpose of killing himself. Against the defendant there were only suspicions and conjectures. As pointed out by the judge, the great use of a jury is to separate firmly and clearly suspicions from evidence. Addressing the jurors, the Lord Justice Clerk stated:

29[i]f you can’t say, we satisfactorily find here evidence of this meeting, and that the poison must have been administered by her at that meeting, whatever may be your suspicion, however perplexing may be the probability against her, and however you may have to struggle to get rid of it, you perform your best and bounden duty as a jury to separate suspicion from truth, and to proceed upon nothing that you don’t find established in evidence against her.⁴⁹

30When the summing up was ended, the jury retired to deliberate. Half an hour later, the verdict was issued as follows: on the first charge the jurors found the prisoner not guilty; on the second and third charge, they returned a verdict of not proven. All deliberations were issued by majority of votes. Based on this verdict, Madeline Smith was dismissed from the bar.

31Over 30 years the trials of Mary Elder (1827), Christina Gilmour (1844) and Madeline Smith (1857) led to the same conclusion. These famous cases show that there was an evident reluctance on the part of the juries «to return a verdict adverse to the prisoner upon purely circumstantial evidence».⁵⁰ That said, the juries were required to choose between not guilty or not proven verdict: therefore the choice was between a «triumphal acquittal», on one side, or a «second class acquittal», on the other side. When the case was shrouded in mystery despite the prosecution’s efforts, the jurors usually issued a verdict of not proven. There was indeed a very strong connection between standard of proofs and jurors’ attitude towards the defendant: not proven appeared to be the most suitable verdict when the evidence given in the trial was inferential, but nevertheless able to cast a shadow over the innocence of the prisoner.

4. The Ardlamont mystery

32Firmly and clearly separating suspicions from evidence: this was the burdensome task of the jurors in criminal matters. If the jury was not fully satisfied by the evidence given in the trial, they were bound to give the prisoner the benefit of the doubt, without consideration to the consequences of their deliberation. The verdict, indeed, is «the best approximation to truth at which we could arrive».⁵¹ Struggling with different points of evidence or faced with a real conflict of skilled testimonies, the jurors were asked to issue a verdict according to their own deliberative judgment as pointed out by the Lord Justice Clerk in the trial of Alfred John Monson which took place in December 1893.

33The Monson case – also known as the Ardlamont mystery – is another famous Scottish trial resulting in a verdict of not proven.⁵² In May 1893 Alfred John Monson, his wife and their three children moved to Ardlamont House in Argyllshire. Monson worked as tutor: he was engaged to educate Cecil Hambrough, a boy of 17 years destined for military career. At the time of the facts, Cecil lived with Monson and his family at Ardlamont House. On 8 August 1893 a man, who called himself Scott, came to the estate and spent there a couple of days. The following morning Alfred and Scott planned to do some splash-net fishing and borrowed a boat for two nights. That evening Cecil had a misadventure during a boat excursion: while he was in the boat with his tutor, the boat capsized, the boy climbed on the rocks, Monson swam to another boat, and rescued him. The next morning Alfred, Scott and Cecil went out shooting. At 09.00 o’ clock Cecil Hambrough died due to a gunshot in his head. The death was qualified as an accident until it was discovered that Cecil’s life had been ensured by his tutor for a huge sum of money. As a consequence, Alfred Monson was arrested and put on trial on the double charge of attempted murder and murder.

34The trial lasted for ten days from 12 to 22 December 1893. To the utmost of his ability, the solicitor general gathered evidence against the prisoner.⁵³ According to the prosecution, in order to kill Cecil, the accused had bored a hole in the side of the boat, induced the young boy to embark along with him, and when the boat was in deep water, removed the plug from the hole, so that the boat sank and Cecil was thrown into the sea. The defence counsel, by contrast, classified the fact as nothing but a misadventure. Immediately after the mishap, indeed, Cecil had spent some time at home with Monson in the smoking-room talking in a friendly way: contrary to what may be expected by a victim who has barely escaped death by foul play, Cecil seemed to have no suspicions at all of an attempted murder against himself.⁵⁴

35But the jurors’ attention was focused mainly on the second charge: murder. According to the solicitor general, Monson had shot Cecil to collect the insurance policy. The scene of the death, the direction of the gunshot, and the character of the wound were consistent with murder. The medical witnesses agreed that Cecil was shot from behind: accidents with guns may occur in an infinite variety of ways but «which of any of these would produce the discharge of the gun through the man’s head from behind almost horizontally, at a height from the ground corresponding with the line of shot through the adjacent trees?»⁵⁵

36By contrast, the defence counsel qualified Cecil’s death as accidental. After all, accidents were commonplace during the shooting season, especially if a person was careless with guns as young Cecil was. Accidents could occur in the most unusual ways as proved by the testimony of the colonel Tillard: the witness explained that, while he was shooting in India, he fell backward and shot himself in the head from the front with the extraordinary result that the shot struck the skull and passed through the ear. Finally, the defence noticed that the prisoner had no reason to kill Cecil: the insurance policy indeed was not valid since the boy was minor. The defendant certainly knew that a minor could not assign, hence the motive suggested by the prosecution had no sound foundation.

37After the speeches of the parties, the judge Macdonald proceeded to charge the jury. Before entering into the merits of the case, the Lord Justice Clerk explained which were the duties of a judge in such situations. In criminal matters the magistrate should offer a judicial and impartial point of view; sometimes, however, the judge may suggest to the jury «things which occur to his own mind upon the evidence». It was a widespread practice and, in the course of his summing up, Macdonald himself proposed this kind of observations but he qualified them with an important remark: the jurors could accept them on the whole, accept them in a modified form, or reject them according to their discernment.⁵⁶ The task of the magistrate indeed was simply to assist the jurors in coming to a true conclusion without imposing his idea upon the case. Sometimes, however, the judge’s opinion was so explicit and convincing in its expression that it risked compromising the judgment of the jurors. This could happen not only in Scotland, but also in Italy: also in the Italian system, in fact, the magistrate was asked to summarize the evidence given in the trial after the controversial and passionate speeches from both sides.⁵⁷

38Getting back to the trial, after this preliminary clarification, the Lord Justice Clerk went into the merits of the question. He specified that the case was purely of circumstantial evidence. He also explained that, when «the guilt of a prisoner is only inferential and is not proved as a matter of fact […] the question of motive becomes of vital importance». In such a case the alleged motive was the assignation of the insurance policy: if Monson believed that this assignation was valid, he actually had a motive for murder, otherwise «a great deal of the superstructure built upon the foundation of motive would necessarily fall».⁵⁸

39Then the judge focused on the first charge and observed: «[w]e never heard before of an attempted murder or drowning in which the person against whom the attempt was made never had the slightest idea of anything of the kind, but came home with his companions, had his whisky and water, and laughed and joked, and was on perfectly friendly terms».⁵⁹ About the second charge, the jurors were required to examine with the most scrupulous attention the scene of death, the skilled testimonies upon the direction and the distance of the shot, and the evidence of doctors upon the character of the wound. The Lord Justice Clerk reminded the jurors that, at first, Cecil’s death was classified as an accident. This prevented the prosecution from making his examinations in the immediacy of the fact, but «it is not, certainly, the fault of the prisoner».⁶⁰ The jury also had to keep in mind what happened to Colonel Tillard while he was shooting in India: this story proved that «accidents do happen in very extraordinary ways».⁶¹ The judge concluded his summing up as usual: if the jury was fully satisfied on the evidence given in the trial, they were bound to give a verdict for the prosecution, otherwise they must give a verdict for the prisoner.⁶² After 73 minutes in the jury-room, the jurors returned a verdict of not proven on both charges. Therefore, the jurors had not emphatically said that Monson did not commit the crimes, but only that the prosecution had not brought the charges home. The following day the Glasgow Herald asked its readers: «Shall we ever know the truth of what occurred in that secluded wood on that tempestuous morning?».⁶³

40The day after the verdict, Monson released an interview to the Westminster Gazzette where he said: «I am glad that the verdict was not a conviction … but I think, on the other hand, that after the very strong charge of the Judge, it should have been a complete acquittal, and not merely a verdict of not proven». Monson considered himself as «a sufferer by the different procedure adopted in Scotland»: an English jury, indeed, would have certainly found him not guilty.⁶⁴ Actually, it is difficult to speculate on the outcome of the proceeding if the jurors were forced to choose between ‘guilty’ or ‘not guilty’. According to some authors, indeed, the verdict of not proven was favourable to culprits because, without this Caledonianan medium quid, the juries probably would have found them guilty.⁶⁵

5. The deliberation of the verdict in the Italian system

41In Scottish law not guilty and not proven are both verdicts of acquittal and have the same impact; in popular perception, however, between not guilty and not proven there is a remarkable difference: a verdict of not guilty corresponds to a full declaration of innocence, while a verdict of not proven simply means that the evidence given in the trial is defective and the jurors are not fully satisfied that the accused has committed the crime. This is exactly the reason why in late nineteenth century Enrico Ferri proposed the introduction of the not proven verdict in the Italian system. In his opinion not proven was necessary «in order to distinguish true acquittals from those which are doubtful»⁶⁶. Such a distinction would have had a practical consequence: the prisoners who had been acquitted for insufficient evidence could be tried again if new proofs were afterwards obtained against them. Actually, therefore, the system devised by Ferri diverged from the Scottish model: in Scotland not proven was (and still is) a verdict of acquittal, while the Italian jurist proposed the introduction of an intermediate verdict which ontologically and legally differed from an acquittal⁶⁷. Ferri’s proposal, however, was destined to remain unheeded: as we are going to see, the doubt upon the guilt of the accused did not allow a form of expression in the Italian jury system.

42The first Italian code of criminal procedure (1865) established the trial by jury for the most serious crimes under the jurisdiction of the courts of assize. A few years later, in June 1874, the legislator approved an important reform in order to solve some inconveniences in the selection and functioning of the jury.⁶⁸ Italian juries were made up only of men selected for their education, occupation, or wealth. Those men were required to examine all the circumstances of the case and issue their verdict according to their own intimate conviction. After the deliberation of the verdict, the court, composed by three professional judges, identified the law applicable to the facts and found the appropriate punishment within the minimum and maximum ranges provided by law. The system, therefore, rested on the distinction between factual questions, demanded to the jurors, and legal questions, entrusted to the magistrates. This separation was strongly criticised by almost all the jurists of the time, because factual and legal questions represented two sides of the same coin, thus the desired separation would have been impossible to wholly achieve in everyday practice.⁶⁹

43The decision-making procedure in the Italian system was patterned on the French model which, under this respect, deviated significantly from the English experience.⁷⁰ In Italy as in France jurors were asked a number of questions, to which they could only answer ‘yes’ or ‘not’.⁷¹ These questions were formulated without resorting to any legal terms, in an attempt at implementing the coveted, but controversial, separation of competences between jurors and judges. The questions were proposed by the president of the court after the speeches of the parties but before the presidential résumé which, therefore, assumed a double connotation: in his summing up, indeed, the magistrate not only summarized the evidence given in the trial, but also explained the meaning of the questions and the connections existing among them.⁷² When the presidential résumé was ended, the jurors retired to their room to deliberate. Therein, they could not discuss the matter among themselves, because the verdict was to be found exclusively on the basis of each juror’s intimate conviction. As anticipated, the vote was given by answering positively or negatively to the questions.⁷³ The voting-paper could be left blank: as provided by law, a blank voting-paper was considered a vote for the accused. No other option was given: to Enrico Ferri’s great regret, an intermediate verdict was not contemplated in the Italian jury system.⁷⁴

44When the voting process was over, the jury returned to the courtroom and the foreman read their decision aloud in the presence of the bench, the public prosecutor, and the defence counsel. The verdict was issued by a simple majority of votes: from this perspective, Italian and Scottish juries were on the same page. The choice between unanimous verdicts and verdicts by a majority was subject of discussion in the Italian debate. The matter was examined by Giuseppe Pisanelli in his book Dell’istituzione de’ giurati first published in 1856. As pointed out by the author, the unanimity rule adopted in England represented a protection for the defendant who could only be sentenced if all the jurors found him (or her) guilty. In order to achieve this result, English jurors were kept together without refreshment until they found a unanimous consensus upon the guilt or innocence of the accused. Hence, the much-vaunted unanimity was likely to be the result of mild coercion and a desire by the jurors to be relieved of their unwelcome task. The majority rule was itself rife with problems, a single vote marking the difference between conviction or acquittal. The best option, therefore, was in the middle: the jurors should issue their verdicts by a qualified majority, that is 9 jurors out of 12.⁷⁵ Italian legislator, however, opted for the principle of simple majority, this means that 7 votes were enough to convict the defendant. In the event of a tie, the opinion in favour of the accused prevailed in accordance to the principle in dubio pro reo.

45However, a moderating element was also introduced: under certain conditions, in fact, the bench could suspend the judgment and postpone the case to a new session before another jury. This possibility was allowed when the jurors had issued a verdict of guilty by a simple majority of votes, while the three professional judges were unanimously persuaded that the verdict was wrong.⁷⁶ In the intention of the legislator, such postponement of judgment would prevent an innocent person from being unfairly convicted. This rule, therefore, was a judicial protection for the accused, just like in Scotland the verdict of not proven was supposed to prevent unfair convictions based on defective evidence. In Italy, however, this measure was almost never applied, while in Scotland the «bastard verdict» was widely expected.⁷⁷

6. In dubio pro reo: the trial of Vito Modugno

46In the early twentieth century some famous criminal affairs attracted a growing interest in the Italian public opinion⁷⁸: among these famous cases, the trial against the lieutenant Vito Modugno, charged with the murder of his wife, Vincenzina di Cagno, is worth of special attention in order to understand the attitude of Italian jurors struggling with circumstantial evidence.⁷⁹

47Vito Modugno belonged to a modest family from the city of Bitonto in the south of Italy.⁸⁰ After completing his studies in Turin, Vito entered in the Military Engineering Corps in 1897. In March 1900 Vito married a 20 years old girl, Vincenzina: the spouses lived in Bitonto, in Modugno’s paternal house, for a short period of time, before moving to Pavia. Four months after the wedding, Vito left for an expedition to China and Vincenzina moved back to Bitonto, where she gave birth to a female infant. On his return to Italy in October 1901, the lieutenant was assigned to the military department located in Bari, where he moved with his family. On Christmas Eve 1902, Vito, Vincenzina, and their daughter Maria went to Bitonto to spend there the Christmas holiday. In the morning of 29 December, Vincenzina was found dying in her bed with a gunshot wound in the head. In a suicidal note found by Vito under her pillow, the woman confessed to being tormented by serious regrets which forced her to commit suicide. Vincenzina died a few hours later without regaining consciousness. After two weeks, the lieutenant was arrested on charge of murder.

48The trial started in March 1905, over two years later from the tragedy. In the Italian criminal procedure, a trial should take place before the judges of the territory where the alleged crime had been committed; under certain conditions, however, the venue of the trial could change. A change of venue, for instance, could be arranged when the case aroused considerable local interest to the point of compromising the serenity and impartiality of the judgment. This is the reason why the trial of Vito Modugno moved from south to central Italy and was held at the court of assize of Perugia.⁸¹

49Before the judges Vito proclaimed his innocence and supported the theory of suicide. He said that, at the time of the shot, he was in the bathroom, while Vincenzina was awake but silent in their bed. The testimony of the housekeeper, Domenica Bonasta, seemed to confirm such a theory. Domenica stated that she run into the bedroom as soon as she heard the shot: no one else but Vincenzina was there. In a previous declaration, however, the housekeeper said that she entered in the room five minutes after the shot: questioned about the discrepancy in her two statements, Domenica claimed to have had «a wrong perception of time». Vito’s alibi seemed to be confirmed also by the farmer, Francesco Fusaro, who, at the time of the fact, was working in the vegetable garden: when he heard the shot, he climbed the window and saw the lieutenant running out of the bathroom. According to the public prosecutor, however, both testimonies were consistent with murder theory: after shooting his wife, Vito would have run into the bathroom and then returned to the bedroom, where he found the housekeeper. Thereby, the man provided himself an alibi for the time of the shooting.

50About the weapon of the alleged crime, Vito and his father claimed to have seen a gun on the bed near Vincenzina. According to the prosecution, however, the lieutenant, in the rush to get out, brought the pistol with him: for this reason, the other witnesses did not see any guns in the room. Regarding the character of the deceased, the prosecutor depicted the woman as a sane, well-adapted person who had no reason to commit suicide; the defense counsel, by contrast, claimed that Vincenzina had killed herself, because she was overwhelmed by regret and affliction.⁸²

51Looking at all the evidence, it appears that there were no direct proofs against the defendant: the case was purely of circumstantial evidence. Therefore, as often happened in criminal matters, it was up to the jurors to bring out the truth through conjectures and inferences. After 6 months, more than 100 hearings, and 200 witnesses’ testimonies, the trial reached its conclusion.⁸³ As usual, before the jurors retired to deliberate, the presiding magistrate summarized the evidence on both sides. The summing up lasted for many hours: the judge focused the jurors’ attention on the pistol and the gunshot wound, the calligraphic expertise on the suicidal note, the character of the spouses, Vito’s behavior after the fact, and the possible motive of the crime. The jury was required to answer the following question: is the accused guilty of having fired a gunshot against Vincenzina in the morning of 29 December 1902 with intent to kill her, causing her death after a few hours? The jurors took only 19 minutes to answer: Vito was found not guilty for the murder of his wife⁸⁴.

52Shortly thereafter, it became known that the verdict had been issued by an equal number of votes, thus leading to an acquittal. According to the reporter Augusto Guido Bianchi⁸⁵, the jurors in the secret of their room arranged to return a verdict which, morally speaking, was a compromise between innocence and guilt. In Bianchi’s opinion this verdict concealed «an unproven guilt»: the jurors opted for this solution, since they could not formally acquit the defendant for insufficient evidence.⁸⁶ As we have seen, this kind of deliberation was precluded to Italian juries even when they were struggling with circumstantial evidence. The verdict, therefore, reverberated all the uncertainty felt by the jurors who, under the weight of doubt, decided to acquit the lieutenant in accordance to the principle in dubio pro reo.⁸⁷

53The trial of Vito Modugno can be considered as an emblematic case: Italian juries, as well as Scottish juries, were reluctant to return a verdict against the defendant whenever the evidence was merely circumstantial. In the Italian system, however, doubt was not allowed any legitimate form of expression: under this perspective, an equal division of votes might reveal more than what one might expect.

7. From Italy back to Scotland: an alleged matricide at the High Court of Justiciary

54The story of Vito Modugno presents some similarities with a famous Scottish case which took place more than twenty years later at the High Court of Justiciary: the trial of John Donald Merrett.⁸⁸ In both cases there was a family relation between the accused and the alleged victim: Vito was charged with the murder of his wife, while Donald was accused of killing his mother, Mrs. Bertha Merrett. Bertha, like Vincenzina, died of a gunshot wound in her head. As we are going to see, also in the Scottish case there were three persons at the flat on the morning of the tragedy: the defendant, the alleged victim, and the housekeeper, who provided different versions of the story. Finally, in both trials, the theory of suicide was entertained as a possible alternative to murder.

55Donald, 17 years old, and his mother Bertha lived in Edinburgh in Buckingham Terrace. In the morning of 17 March 1926, the housekeeper, Mrs. Henrietta Sutherland, arrived in the flat about 9 o’ clock, as usual. After having cleaned the fireplace, Henrietta left the sitting room, leaving mother and son alone there. Shortly thereafter, the daily maid heard a gunshot and a scream from Mrs. Merrett: the woman had received a pistol wound to her head. Henrietta called for the police, who arrived in a few minutes. Mrs. Sutherland told the constables that she did not witness the scene because, at the time of the shot, she had been working in the kitchen. To the Inspector Fleming, however, Henrietta later said that, on hearing the shot, she had run into the lobby and saw «Mrs. Merrett falling off the chair on the floor and a pistol falling out of her hand». Asked about the discrepancy in her testimonies, the housekeeper went back to the first one.⁸⁹

56Mrs. Merrett was sent to the Royal infirmary, where the doctors tried unsuccessfully to save her: Bertha indeed died on 1 April 1926. Before dying, however, Mrs. Merrett regained consciousness and reported different versions of the story. In particular, she said to her sister, Mrs. Penn, that at the time of the fact she had been sitting at the table «when a sudden explosion went off in my head as if Donald had shot me».⁹⁰ Mrs Penn believed it impossible and did not give importance to these words. At that time, in fact, no one suspected Donald and the case was classified as suicide.

57The investigation took a turn when a handwriting expert, Gerald Francis Gurrin, made a report about a number of cheques which had been recovered from the banks where Bertha had her accounts. As a consequence, at the end of November, Donald was arrested upon two charges: first, he was accused of killing his mother; second, he was accused of uttering as genuine 29 cheques. The trial took place in February 1927 at the High Court of Justiciary before a jury made up of 6 women and 9 men.⁹¹ The presiding judge was Lord Alness; the prosecution was entrusted to the Right Honourable William Watson; the defence was conducted by Mr. Craigie Aitchison.

58For five days the prosecution presented the evidence against the accused: the Lord Advocate examined the testimony of the housekeeper, the statements made by Bertha before dying, the experiments made with the gun, the direction of the shot, and the absence of blackening near the wound. The absence of blackening was considered an element of fundamental importance, because it means that the shot was fired from a certain distance: for the Crown, therefore, the case was homicide.

59By contrast, for the defence the death of the woman could be classified as suicide or accident. According to Mr. Aitchison, the jurors should not rely on the statements made by Bertha while she was at the infirmary, because «the injuries she received were such as to produce very serious mental changes, including altered consciousness or disassociation».⁹² Furthermore, the forensic experiments were not conclusive: on one side the blackening might have been removed when the blood was washed away, on the other side the direction of the wound was consistent with both accident and suicide. In his closing statement, the defence counsel reminded the jury that «it is for the Crown to prove their case»: this elementary and axiomatic rule of criminal law means that, if the Crown failed to prove the case, the accused is entitled to the benefit of the doubt. Mr. Aitchison thought it appropriate to specify that «suspicion is not proof, and conjecture is not proof, and probability is not proof».⁹³ According to the defence counsel, the guilt of the defendant had not been proved beyond reasonable doubt. Under these circumstances, two kinds of verdict could be issued, that is not proven or not guilty. Mr. Aitchison explicitly asked the jurors not to issue a verdict of not proven, because such a deliberation would have implied a stigma on the accused. The «bastard verdict» brings shame and dishonour on the acquitted person whose reputation would be irreparably compromised: with this verdict, therefore, the jurors take «a tremendous responsibility» upon their shoulders. Thus, the lawyer asked the jury for a verdict of not guilty upon both the charges.⁹⁴

60Thereupon, the Lord Justice Clerk proceeded to charge the jury. Firstly, the judge pointed out that the allegation of murder was the result of an «aforethought» on the part of the Crown. At the time of the occurrence, indeed, the police thought they were dealing with an attempted suicide, not a case of murder. Only at the end of November the sheriff granted a warrant for the arrest of Donald. Important consequences follow from this delay: absolute precision regarding the circumstances of the case and the statements of the alleged victim «is now unattainable». The «aforethought» might be correct, but the delay had inevitably compromised the possibility of achieving definite results.

61The judge agreed with Mr. Aitchison: the onus to prove guilt is upon the Crown. To discharge that onus, the prosecution brought up evidence of what happened at the flat on 17 March 1926 and at the infirmary during the following weeks. The Crown had invited the jurors to draw an inference of guilt from this evidence, but the truth is that the statements considered as unequivocal by the Lord Advocate were neither noted, nor taken under oath, for the simple reason that, at the time of the occurrence, the death of the woman had been classified as suicide.⁹⁵ Not even the direction of the wound was decisive, because it was compatible with both murder and suicide. The jurors were asked to focus their attention on the presence of blackening near the wound, because, as noted by the prosecution, the absence of blackening means a wound inflicted by an outside agent. According to the defence counsel, however, the blackening might have been rubbed off, because the wound had been washed to remove the clotted blood. On this matter the experiments made by each side arrived at totally different results: the jury therefore had to deal with «a direct conflict of skilled testimony».⁹⁶

62Regarding the second charge, the jury was asked to examine three aspects: first of all, if the Crown had proved that the cheques were forged; secondly, if the Crown had proved that the accused presented them for payment; lastly, if the Crown had proved that the accused knew the cheques were forged when he presented them for payment. If the prosecution had successfully proved all these elements, the jurors were required to find a verdict of guilty upon the second charge.⁹⁷

63The judge concluded his summing up in the traditional form: if any reasonable doubt existed with regard to both charges, then the accused, in accordance with the Scottish law tradition, was entitled to the benefit of that doubt; by contrast, if the jurors were fully satisfied by the evidence given in the trial, they were bound to convict the prisoner, however unpleasant their task might be.

64When the résumé was finished, the jury retired to deliberate. About an hour later, the jurors returned to the courtroom and communicated their decision: under the first charge the verdict was not proven by a majority, under the second charge guilty unanimously. The Scottish lawyer William Roughead noted that the vote upon the first charge was 5 for guilty and 10 for not proven⁹⁸: therefore, none of the jury considered John Donald Merrett innocent, nevertheless he was acquitted because the Crown had failed to prove the case beyond reasonable doubt. As a consequence, the accused was convicted only upon the second charge: the sentence imposed by the court was 12 months imprisonment.⁹⁹

65Once released from prison, Donald married Vera Bonnar and assumed the name of Ronald John Chesney. In February 1954 Vera and her mother were found dead by the police.¹⁰⁰ During the investigations, the police learned that 27 years earlier Vera’s husband had been charged, as John Donald Merrett, with the murder of his mother. A few days later, Interpol warned Scotland Yard than Chesney had committed suicide. Nonetheless, the investigations continued until it was determined that John Donald Merrett, alias Ronald John Chesney, had killed both women.¹⁰¹

8. The reform of the trial by jury in Italy under the Fascist Regime

66In the late 1920s, while in Scotland the trial against John Donald Merrett was once again turning the spotlight on the so-called «bastard verdict», in Italy the most eminent legal experts were engaged in discussing the role of lay judges before the courts of assize. The trial by jury in the Italian experience had pointed out some critical issues, resulting sometimes in unexpected verdicts which caused a sensation in both the press and the public opinion. Furthermore, it must be considered that, with the coming of the Fascist regime, the political environment had profoundly changed: the new government, however, did not intend to abolish lay participation in the administration of criminal justice, but wanted to submit the deliberations of the jurors to the supervision of the professional judges.¹⁰²

67Until then, the system had been based on the separation between jurors’ and judges’ duties: the former were required to examine the fact and all the circumstances of the case, while the latter were entitled to find out the law and determine the penalty in the event of a conviction. The new Minister of Justice, by contrast, proposed a decisive change of perspective: lay judges and professional magistrates should collaborate and decide together all the relevant issues of the case. On these premises, the royal decree n. 249 enacted in March 1931 abolished jury trials and introduced mixed courts composed by two judges and five laymen assessors, who were bound to work side by side in mutual respect and cooperation.¹⁰³

68As a consequence of the reform, the decision-making procedure adopted by the courts of assize significantly changed: the monosyllabic verdict of the jurors – so highly criticised by Enrico Ferri – was replaced by a properly motivated judgment usually written by one of the two magistrates. The reporting judge was asked to explain with the most scrupulous diligence the reasons behind the deliberation in favour or against the accused. A point which is most relevant to this study is the introduction of a judgment that jurors had not been allowed to make: the mixed courts could now explicitly acquit the defendant for insufficient evidence. In such a case, the reporting magistrate was bound to illustrate the doubts arising in the panel during the discussion. In this matter there were no specific rules: according to the principle of free evaluation of evidence, indeed, the court was entitled to determine as best they could whether the doubts were such as to prevent a declaration of guilt or innocence. It was clearly a slippery slope, as demonstrated by the case of Francesco Mulas, a Sardinian shepherd charged with murder and robbery in 1939.¹⁰⁴

69On 9 March 1937 Francesco Mulas and his sister Maria Pasqua were going to Sassari to visit their sick brother. During their journey by train, they met Pietro Deschini, entrepreneur by profession, who was moving to Sassari in order to buy some donkeys. Francesco and Pietro slept in the same hotel and spent together the following morning. Since then, no one heard nothing about Pietro, whose body was found hidden in a wall nearby the Mulas’ sheepfold two days later. The suspicions immediately fell on the Sardinian shepherd who was arrested on 27 March. After a few months, also Maria was arrested on charges of participating in the crime by luring the victim to his death.

70The trial took place at the court of assize of Sassari in 1939. All the witnesses depicted Maria as a very honest woman unable to commit any crime: as a consequence, Maria was unanimously acquitted. Francesco’s situation, by contrast, was much more serious: according to the prosecution, the accused had killed the victim to rob him of his money. The crime of aggravated murder under the Fascist Regime was punishable with death, thus the life of the defendant was at stake.¹⁰⁵ The evidence, however, was merely circumstantial and the defence managed to instil a reasonable doubt in the minds of the lay judges: the murder might have been committed by someone else and the corpse hidden nearby the Mulas’ sheepfold to put the blame on the accused. In light of this doubt and in view of some gaps in the investigation, the majority of the panel decided to acquit Francesco for insufficient evidence.¹⁰⁶

71The public prosecutor appealed the decision by noting that the reporting magistrate attached much importance to the evidence given in the trial against the accused and devoted only a few lines to explain the doubts felt by his colleagues upon the case.¹⁰⁷ Therefore, the Cassation should have set aside the judgment and transfer the trial to another court of assize.¹⁰⁸

72The Cassation, however, rejected the appeal as unfounded. As pointed out by the supreme judges, in such a case Francesco was dismissed not because he was found innocent, but because of lack of evidence. It is true that in the first part of the reasoning the reporting magistrate affirmed the guilt of the prisoner, but it is also true that in the last part he explained the doubts which arose within the panel: «faced with the proofs against the accused, the doubts may appear modest and tenuous, but nevertheless be enough to invalidate in the mind of the judges the certainty needed to convict».¹⁰⁹ Whatever the opinion of the reporting magistrate might have been, the judgment explicitly pointed out that the majority of the panel was not fully satisfied by the evidence given in the trial. A perfect explanation of this reluctance could not be demanded, because the state of doubt by its nature escapes a precise definition. As a consequence, the appeal was dismissed.

73We can note that the core of the reasoning proposed by the Cassation was precisely the distinction between full acquittal and lack of proof. The case was one of circumstantial evidence and some doubts had crept into the majority of the panel: they acquitted the defendant because the evidence given in the trial was defective, not because they were convinced that Mulas had not committed the crime. Herein lies the difference between ‘a second-class’ acquittal and a full declaration of innocence.

9. The debate until the present day

74When the World War II was over, Italy became a Republic and adopted a new Constitution which entered into force on 1 January 1948. The Italian Constitution states a number of important principles regarding the organisation of the justice system.¹¹⁰ Art. 102, in particular, establishes that «the law shall regulate the cases and forms of direct participation of the people in the administration of justice». Trial by jury, however, is no more feasible because, ex art. 111, all judicial decisions shall include a statement of reasons. In light of these principles, mixed courts seem to be the only possible solution: nowadays, in fact, Italian courts of assize are made up of two magistrates and six lay judges.¹¹¹

75With special attention to criminal matters, the Constitution affirms that «a defendant shall be considered not guilty until a final sentence has been passed». Such a presumption exonerates the accused from the onus to prove his or her innocence: as we have already said on several occasions, the burden of proving the guilt of the defendant lies on the prosecution. Therefore, if the prosecution fails to prove guilt beyond reasonable doubt, the accused is entitled with the benefit of that doubt.¹¹²

76In the decades following the entry into force of the Constitution, an intense debate developed between those who defended the possibility to acquit the accused for insufficient evidence and those who demanded the abolition of this controversial formula.¹¹³ The authors in favour of the abolition noted that the acquittal for insufficient evidence contradicts the presumption of innocence provided by the Constitution. This kind of deliberation, indeed, put the responsibility for the failure of the trial on the defendant, seriously affecting his or her social reputation.¹¹⁴ By contrast, the authors against the abolition noted that between the certainty of guilt and the certainty of innocence there might be situations of doubt which needed a form of expression. The abolition, moreover, could lead to an increase in the number of convictions, thus frustrating the principle in dubio pro reo. Even those authors, however, realized that this type of deliberation leaves a stigma on the acquitted person: as a general consensus, these detrimental effects were to be abolished or, at very least, reduced.¹¹⁵

77The debate continued until a new code of criminal procedure came into force in October 1989: the great reform abolished the acquittal for insufficient evidence and prescribed that «the judge shall deliver a judgment of acquittal also in case of insufficient, contradictory or lacking proof that the criminal act occurred».¹¹⁶ Nevertheless, in the reasoning of the judgment, the magistrate is required to explain if the defendant has been acquitted due to contradictory or insufficient evidence: by this way, the legislator managed to find a compromise between opposing needs and tendencies.

78While in Italy the legislator opted for the abolition, in Scotland the verdict of not proven is still the subject of an intense debate in both the public opinion and the legal profession. Over 40 years after the trial against John Donald Merrett, the not proven verdict was taken into serious consideration by the Thomson Committee on Criminal Procedure in Scotland (1975). The Thomson Committee set out the arguments for the abolition of the not proven verdict, but finally preferred to retain it, because there was «no evidence that the public regard the present system as working unsatisfactorily».¹¹⁷

79The three-verdict system returned under scrutiny in the aftermath of several high-profile cases during the early 1990s. The most significant among them was the trial against a young man charged with the murder of a female student: this case caused a sensation because most observers believed that the accused had committed the crime, but the jury found the charge not proven. After the trial, the victim’s family organized a campaign for the abolition of the not proven verdict, thus attracting a growing attention by the media. Political pressure increased until the Government decided to include a review of the not proven verdict in the political agenda. The Scottish Office considered the theme Juries and Verdicts: herein, arguments pro and contra the not proven verdict closely resemble those put forward by Italian jurists a few decades earlier in the debate for the abolition of the acquittal for insufficient evidence. In conclusion, however, the Scottish Office proposed the retention of the «bastard verdict», because the consultation did not reveal «a consensus for change».¹¹⁸

80Since then, a number of scholars has analysed the impact of multiple options in jury decision making procedure.¹¹⁹ The researches carried out in the last years show a difficulty in understanding the meaning of the not proven verdict and its legal consequences. If originally not proven implied a failure to prove the facts under investigation, nowadays it is simply one of the two acquittal verdicts provided by the Scottish criminal procedure.

81A large-scale mock jury study commissioned by the Scottish Government a few years ago sheds a new light on this matter.¹²⁰ One of the main questions addressed in this study was formulated as follows: «what are jurors’ understandings of the not proven verdict and why might they choose this verdict over another verdict?».¹²¹ The findings of the study reveal that, most of the times, jurors return a verdict of not proven when they suspect the accused is guilty «but do not think the evidence proves it beyond reasonable doubt».¹²² Another important aspect emerged during deliberations was the perception that «there is an element of stigma attached to a not proven verdict».¹²³ This «black mark» is a logical and natural consequence of the idea mentioned above: the jurors find the charge not proven when they think the accused is guilty, but the necessary standard of proof has not been met. From this point of view, therefore, jurors’ understanding nowadays does not differ so much from jurors’ perception in the long nineteenth century.