INTRODUCTION: TWO STORIES

This chapter links two stories, each with sub-plots. The first concerns one of the most famous cases in English legal history, the story of Edith Thompson and Frederick Bywaters who were hanged in 1923 for the murder of Edith’s husband, Percy. The second is the story of two scholars in different departments in the same institution who over several years had worked on the case and had each completed a substantial study before they learned of the other’s interest. The first story is both a human tragedy and a historical mystery. The second exemplifies the fragmentation of learning and illustrates contrasting approaches to history by two scholars from different academic cultures. This chapter is a case-study of the methods of a Shakespearean and a jurist in approaching the question: Was Edith Thompson guilty of the murder of Percy Thompson? We shall start with the more modest tale.

THE SHAKESPEAREAN AND THE JURIST

The Jurist’s Tale

Early in the 1970s I started to use original trial records as a vehicle for teaching evidence to undergraduate law students first at the University of Warwick and later at the University of Miami. I was concerned both with the theory of proof in legal contexts and certain practical techniques of inferential reasoning, as part of what is known as the logic of proof. In particular, I was interested in how to construct and criticise complex arguments based on evidence in legal contexts.

After a time, I settled on the case of R v. Bywaters and Thompson as the best vehicle for my purposes just because the evidence was both extremely complex and ambiguous. The case had the additional advantages of having real human interest and a secondary literature which was radically divided on the question of Edith Thompson’s guilt. But the key point was the complexity and the intractability of the material. As one student put it: “If you can analyse Edith Thompson’s prose you analyse anything”. In 1982 I gave a public lecture about the case. The main purpose was to illustrate a particular method of analysing evidence, modified Wigmorean analysis, which I shall introduce later.

I moved to University College London in 1983, and continued to use the case in teaching, but it was not until 1987-88 that I revised and expanded my lecture into a long essay. I was making final revisions to my essay when I first heard that René Weis, a colleague in the UCL English Department who is a Shakespearan scholar, had finished a book on the same case. When I read the book, which I greatly admired, I decided not to revise my original essay. Instead I wrote a second essay comparing our general approaches, our treatment of detail, and our conclusions. Since then we have discussed the case both publicly and in private on a number of occasions. This chapter continues the conversation.

The Shakespearean’s Tale

In the early 1980s judicial issues featured prominently on the parliamentary agenda. In the wake of the debate in the House of Commons on the death penalty in July 1983 I read Arthur Koestler’s polemical study Reflections on Hanging (1956), which paved the way for the abolition of the death penalty in Britain. It was Koestler’s remarks about Edith Thompson which first aroused my interest in the case of ‘Bywaters and Thompson’.

After reading the Trial of Frederick Bywaters and Edith Thompson in the Notable British Trials volume (edited by Filson Young, 1923), which also reproduces Edith Thompson’s letters, I decided to write a biography of her. I was convinced that an account of Edith Thompson’s life and death would make an effective contribution to the case against the death penalty. Although there seemed to be almost universal agreement about her innocence, she remained a convicted felon in the eyes of the law. Furthermore, the manner of her death, and the rumours about it, had caused great controversy at the time, and its repercussions were felt as late as the 1950s when Koestler’s book appeared.

I was impressed by Edith’s minute detailing in her correspondence of life in London in the early 1920s, and by her vivid discussions of books, plays, and music-hall shows. As a professional reader of literature, I was intrigued by the extent to which the boundaries between fantasy and reality were blurred in Edith Thompson’s mind. As well as being a successful career-woman with a good head for business, she was an Emma Bovary figure capable of losing herself in countless romantic novels. Her imaginary escapes from an unfulfilling marriage and the boredom of everyday life might never have translated into reality, if Frederick Bywaters had not appeared on the scene. He seemed to offer everything that Edith’s husband had failed to deliver, and Edith grabbed the opportunity.

In Criminal Justice: The True Story of Edith Thompson (1988), I argue that in the course of her relationship with Bywaters Edith Thompson’s tenuous grip on vital distinctions between reality and imagination slipped altogether. Not only did she discuss the fates of heroes and heroines as if they were real people, but she started to fantasise on similar lines about ridding herself of her husband. Rather than seeking a divorce, Edith Thompson preferred to toy with thoughts of her husband’s death.

Her claims to have attempted to kill him with, specifically, poison and glass were demonstrated by the Crown’s pathologists to be false. I interpreted these protestations as rhetorical flourishes, a way of assuring Bywaters that she would go to any length to prove her love for him. Imaginary murder, it appears, seemed preferable to the real social stigma of divorce. To the extent that Edith Thompson failed to walk out on her marriage and set up home with Bywaters, she was a prisoner (probably) of her temperament, and (certainly) of her class and time.

Whether or not her declared intent to kill her husband influenced Bywaters’s murder of Percy Thompson late one night in Ilford has been debated ever since Edith Thompson and Freddy Bywaters were jointly charged with murder. I argued that the two were not linked, and that Bywaters started to drift away from Edith Thompson in the months before the murder. I maintained that Edith Thompson was innocent of the charges brought against her, and that she was the victim of a moral climate that cast her as the Sirenic temptress of a young, infatuated hothead.

Criminal Justice is anchored in a contextual reconstruction of Edith Thompson’s life, from her primary school-records to a study of the books and plays she read and saw. I analysed the newspapers of the period, and I gained access to closed Home Office and Prison Commission files on the case. At the time of my research in the mid-1980s there were still people living who had known the protagonists of the story (one of them was Bywaters’s best childhood-friend). They shared their memories and their photographs of Edith, Freddy, and Percy with me. This enabled me to flesh out their story in ways that are acceptable, and indeed expected, in a biography. By exploring the locations in which Edith and Freddy grew up I could, for example, establish that on the night of the murder Edith and Percy Thompson returned to their home on the usual and most direct route, an important point in view of the statements in court to the contrary. My narrative strategy was to offer a prospective, day-by-day account, which took the shape of a narrative ‘diary’ for the eighteen months of Edith Thompson’s involvement with Bywaters.

 

PART I. ANATOMY OF A CAUSE CÉLÈBRE

The Facts

About midnight of Tuesday/Wednesday, 3/4 October 1922, Percy Thompson, and his wife, Edith, were returning from a visit to a West End theatre to see a Ben Travers farce. They were within 100 yards of their home in Kensington Gardens, Ilford when a man in a raincoat and hat came from behind them, pushed Mrs Thompson aside and approached Percy Thompson. After a struggle Percy Thompson collapsed on the pavement 50 feet further on and the man ran off. Edith’s voice was heard by a neighbour crying, “Oh, don’t; oh, don’t.“ Shortly afterwards she ran up to a group of people, also returning from Ilford Station, and cried out, “Oh, my God! Will you help me; my husband is ill, he is bleeding.” When a Dr Maudsley arrived from a nearby house some five to eight minutes later, he found that Percy Thompson was dead. He had been propped up against a wall by his wife. Mrs Thompson was in a confused and hysterical condition. When told that Percy was dead, she said: “Why did you not come sooner and save him?” Blood was gushing from the mouth of the deceased, but it was only after he had been taken to the mortuary that several knife wounds were found in his body. There were several slight cuts in the front of the body and two deep stab wounds in the back of the neck.

The police called at the Thompsons’ house at 3 a.m. Edith Thompson was in a very distressed state. She said that all she knew was that her husband dropped down and screamed out, Oh. She said that she had thought that it was one of his attacks. She denied that either she or her husband had a knife. The following morning the police learned of Edith’s friendship with a young Merchant seaman called Frederick Bywaters and, having found out that he had spent the evening with Edith’s parents the Graydons, some two miles from the scene of the crime, the police detained him for questioning at 6 p.m. on 4 October. In Bywaters’s room in his mother’s house they found two notes, five letters and a telegram all from Edith Thompson.

Bywaters made his first statement, after caution, denying both knowledge of the crime and owning a knife. Mrs. Thompson was arrested on the evening of the same day. On the following day she was shown some letters from her to Bywaters. She then made a statement in which she denied seeing anyone at the time her husband fell; she acknowledged her relationship with Freddy Bywaters, admitting that they had corresponded on affectionate terms. Immediately after making a statement she was taken past a room in which she saw Bywaters. She said: “Oh God, Oh God, what can I do? I did not want him to do it.” After caution, she made another statement in which she said a man rushed out and pushed her away. She saw them scuffling. She recognized him as Bywaters by his coat and hat. Bywaters made a further statement on 5 October, after being told that he and Edith Thompson were to be charged with murder. He admitted to killing Percy after a struggle. “Mrs Thompson must have been spellbound for I saw nothing of her during the fight”.

The police found the murder weapon, a knife, on 9 October, as a result of statements made by Bywaters. On 12 October, they found a box on Bywaters’s ship, the R.M.S. Morea, containing sixty-two letters and telegrams from Edith to Frederick, together with fifty cuttings from newspapers enclosed in the letters and a photograph of Edith. A number of the cuttings dealt with murder, mainly by poison. Three letters from Bywaters to Edith were found at her place of work. Rather over half of the letters and a selection of the cuttings were put in evidence by the prosecution. Much of the interest of the case centres on Edith’s letters and the cuttings, which constituted the main, but not the only, evidence against Edith.

It is universally accepted that it was Bywaters who killed Percy Thompson. It is also not disputed that Edith and Freddy were lovers and that they both had a prima facie motive for killing Percy. Freddy pleaded not guilty on the ground that he had killed Percy in self-defence. Since Percy was unarmed at the time, it is hardly surprising that this defence failed. The interesting question was whether Edith was also guilty either on the ground that she incited Freddy to kill Percy or that they jointly planned to kill him. It is clear from the letters that Edith and Freddy had been lovers for nearly two years; that Edith was very unhappy; that she wished to end her marriage and that she had expressed the view that she would not be sorry if Percy were dead.

Edith Thompson and Frederick Bywaters were tried at the Central Criminal Court on 6-11 December 1922. They were both charged with the murder of Percy Thompson, Bywaters as principal, Edith as a principal in the second degree, because she was present at the killing. A second indictment containing a number of counts, including charges against Edith of administering poison and broken glass with intent to murder were not proceeded with. The jury found both accused guilty, after barely two hours’ deliberation. They were sentenced to death; ten days later the Court of Appeal rejected their appeals; and on 9 January — three months and six days after the death of Percy, Bywaters and Thompson were hanged.

The full transcript of the trial has been published in the Notable British Trials series, this includes all the evidence presented in court, together with a complete transcript of letters from Edith Thompson that were not put in evidence at the trial and the judgments of the Court of Criminal Appeal in the case (two appeals). We thus have available to us what Wigmore called “a mixed mass of evidence”. This includes statements made by the two accused, the testimony of witnesses who were nearby at the time or immediately after the attack, a limited amount of circumstantial evidence, including the report of Sir Bernard Spilsbury that there were no traces of either glass or poison discovered in the post mortem examination. The exhibits included the murder weapon, three letters from Freddy to Edith and Edith’s letters to Freddy. There is the evidence given by both accused at the trial. Some additional information, not given in evidence at the trial, will be ignored for present purposes.

I shall concentrate on those of Edith’s letters that were put in evidence, but it is worth mentioning briefly that the other evidence establishes beyond reasonable doubt that it was Freddy who killed Percy, that there is almost nothing to support, and a good deal to confute, Freddy’s claim that he acted in self-defence, that under cross-examination Edith admitted that she wished to give Freddy the impression that she had tried to make her husband ill and that, while the testimony of both accused is an important aid to interpretation of some passages in the letters, there is little outside the letters to support the prosecution case and nothing that might be taken as determinative or dispositive in regard to the other main theories. Accordingly let us treat the letters as the main evidence against Edith and consider the other evidence presented at the trial as supplementary.

A striking feature of the extensive secondary literature about the case is that the commentators are almost equally split on the question of Edith’s guilt. By and large they belong to the literature of popular entertainment but some are very competent examples of this particular genre — carefully researched, often perceptive, usually readable. Since the main objective is to arouse interest or to make a case, detail tends to be sacrificed to readability and more emphasis is placed on amateurish psychological speculation than on the more mundane aspects of meticulous analysis of evidence.

What can a more scholarly or scientific approach add to our understanding of a complex case such as this?

Modified Wigmorean Analysis

In this case-study I shall apply a modified version of Wigmore’s “Chart Method” of analysing evidence (to be supplied) . The original method has been extensively described and discussed elsewhere. Here it is modified in three ways. First, no use will be made of symbols and charts. Second, clarification of standpoint will be treated as an essential ingredient in macroscopic analysis of the case as a whole. Third, there will be much more emphasis on the idea of “theories” of the case.

The analysis will involve the following steps:

1. clarification of standpoint;
2. definition of the ultimate probanda;
3. outlining four main theories of the case, with some variant sub-theories, stories and themes;
   testing one of these theories, by way of illustration, through detailed analysis of just two items of evidence: the facts about the knife and one passage in Edith’s last letter (exhibit 60);
4. finally, suggesting some provisional conclusions on the basis of this partial analysis. This will not resolve the mystery; but I hope that it will show that even this selective and relatively straightforward exercise in analysis significantly narrows the possibilities and throws some light on the particular case and on some of the general themes that it illustrates.

Clarification of standpoint

The first step is to clarify the standpoint from which the analysis is to be undertaken. For present purposes, I shall adopt the standpoint of a historian in 1988, confronted with a finite body of data — the record of the trial — and concerned to answer the narrow question: does the available evidence support beyond reasonable doubt the allegation that Edith Thompson was legally responsible for planning or inciting the murder of her husband?

Adopting the standpoint of the historian enables us to pursue “the truth about Edith Thompson” on the basis of analysis of the available evidence (the data) without regard to the procedural complexities, multiple objectives and “noise” factors that concern participants in actual trials. However, in order to keep matters simple and to maintain a close connection with the actual trial, I shall impose two artificial constraints on our enquiry. First, I shall confine my analysis to the evidence in the trial record. A real historian would normally not accept any such limit. Secondly, I shall concentrate on the single question whether Edith was guilty as charged. This question is, of course, by no means the only question about the case that might interest historians — but it is a standard one which is very close to the question posed to the jury in the case.

Framing the ultimate probanda

The next step is to identify the ultimate probanda — the material allegations on which our judgement of the case as a whole must rest. Historians have no concept of materiality. They are free to frame their questions as they wish without formal constraints. In arguing in court about disputed questions of fact lawyers have no such freedom. The questions arising from the indictment (or other pleading) are defined in advance by law. Where the law is clear this can greatly simplify the tasks of analysis and argument, for it lays down clear and fixed touchstones of relevance. Evidence is relevant if it tends to support or tends to negate the facts that have to be proved beyond reasonable doubt. Where however, the law is unclear, the task becomes much

In respect of Bywaters the law was clear enough. The prosecution had to prove beyond reasonable doubt that he had intentionally with malice aforethought and without lawful justification caused the death of Percy Thompson. The only points in dispute were whether Freddy acted in self-defence or under immediate provocation. The jury apparently had no doubts on this matter. Nor should it unduly trouble a historian. Bywaters’s defence was that he accosted Percy to ask him to agree to a separation or divorce, the argument developed spontaneously into a fight and he stabbed him in self-defence, as he thought Percy was about to draw a gun. Two of the stab wounds were in the back and there was no evidence to suggest that Percy had a gun on this, or any other occasion, nor that Freddy had any reason to think that he did. The evidence supporting provocation is almost as thin. It is not surprising that Freddy’s story was not believed.

The case against Edith was less simple. She was charged as a principal in the second degree. What this meant was that the prosecution had to prove (a) that Bywaters murdered Percy Thompson; and (b) that Edith was physically present at the murder; and (c) that she aided and abetted the murder. The first two requirements are relatively unproblematic, but the third contains the seeds of some genuine difficulties. Aiding and abetting, conspiracy and incitement are three areas of criminal law that involve notorious problems of interpretation and application. In the trial and on appeal these difficulties barely surfaced, although a careful reading suggests that the trial judge took a much narrower view of the law than had been assumed by the prosecution or which the authorities at the time probably warranted. He directed the jury that they must be satisfied that Bywaters and Thompson had planned this attack, an interpretation which was surprisingly favourable to Edith.

In order to succeed on the charge of murder against both accused the prosecution had to satisfy the jury that the truth of the following propositions had been established beyond reasonable doubt:

1. that Frederick Bywaters deliberately and with malice aforethought killed Percy Thompson; and
2. and that:
   1. Edith Thompson conspired with Frederick Bywaters to kill Percy Thompson either on this occasion or whenever opportunity arose; or
   2. that Edith Thompson intentionally incited Frederick Bywaters to kill Percy Thompson and the killing was within the scope of that incitement.

These three propositions were “the facts in issue” — that is the material allegations that the prosecution had to prove beyond reasonable doubt in order to succeed. The test of relevance of any item of evidence in this case is whether it tended to support or to negate one of these propositions.

In order to convict Freddy it was sufficient for the prosecution to prove the first proposition only. As we have seen, this was not difficult. Although it was relevant to prove intent and motive to show that this attack had been planned or that there had been earlier attempts by either or both of them to kill Percy, these were not strictly necessary for the prosecution’s case. At the trial there was no dispute about the cause of death or the identity of the killer; it was not even necessary for the prosecution to show that the attack was premeditated, as they claimed, so long as they established that it was intentional.

The case against Edith was more complex. In order to succeed the prosecution had to show that Bywaters was guilty of murder (proposition 1) and that Edith had either planned Percy’s death with Freddy or had been responsible for inciting Freddy to kill him. In argument the prosecution, while presenting a story of continuous incitement leading to an attack that had been planned in advance, rather blurred the distinction between incitement and conspiracy —not surprisingly, for either was sufficient. But Mr Justice Shearman in his charge to the jury implied that they should convict Edith only if they were satisfied that this particular attack had been planned in advance. There is an ironic contrast between the unequivocal sententiousness of tone of the judge’s charge and his insistence that Edith should be convicted only on the basis of what I hope to show is by far the weakest of the possible theories of the case that are consistent with her guilt.
The judge exuded generalized disapproval indicating guilt, but then analytically posed the issue in such a way as to make conviction of Edith much harder, at least in theory. If, as seems likely, both the jury and nearly all subsequent commentators reacted more to the general attitude of the judge than to his precise words, that is doubly ironic.

Theories of the case as a whole

The next step is to consider, from the point of view of an historian, the range of colourably plausible hypotheses about the case as a whole, and the general lines of argument for and against each hypothesis — what we may conveniently refer to as the main theories of the case. The aim here is to provide a broad framework of competing general hypotheses and strategic arguments within which all the relevant evidence can be organized and weighed.
In adversarial criminal proceedings typically the prosecution presents its case against the accused in the form of a coherent story of what happened. In order to succeed they have to prove each of the material allegations beyond reasonable doubt. Typically, but not necessarily, this involves persuading the trier of fact to accept their account of the story as a whole. The theory of a case is an internally consistent collection of hypotheses that form a coherent argument, which supports the story.

In order to counter the prosecution’s case three main options are available to the defence, either independently or in combination: to deny the prosecution’s story, either in toto or in some material particular(s), without offering an alternative, for example by submitting that there is no case to answer; to explain away the prosecution’s story, for example, by admitting most of the hypothetical facts, but interpreting them in a manner consistent with the innocence of the accused (Freddy admitted killing Percy, but claimed that it was in self-defence); or to present a rival account of what happened in respect of one or more material facts (e.g., by presenting an alibi). In practice many defences involve a combination of all three strategies, but sometimes these are not compatible with each other.

A theory in this sense relates to the case as a whole. It can serve a number of functions: it helps to organize the material; it provides a basis for selecting some of a mass of potentially relevant evidence for inclusion or emphasis — a subsidiary test of relevance; it may serve to fill in gaps in the available information with more or less plausible hypotheses; and it provides a general basis for testing internal coherence and consistency. From the point of view of counsel each side’s theories provide a strategic framework which guides, and, often determines, many specific tactical choices.

In trials, when there are disputed questions of fact, typically the trier of fact is only being asked either to select between competing theories of the case or to assess whether, on the basis of the evidence presented, the proponent has proved his or her version to the applicable standard. Historians do not work with the law’s elaborate body of devices for guiding decision in situations of uncertainty in adversary proceedings: they do not have concepts such as materiality, conclusive or rebuttable presumptions, standards of proof, or burdens of proof or production . However, the notion of competing theories or accounts fits a dialectical view of historical enquiries into truth. One way of proceeding is to test a series of competing hypotheses against the available evidence and to select the most plausible one. In the present context, a series of alternative theories of the events in Bywaters and Thompson serves much the same function as they would in a real trial. Part of the interest of the present case is that even at this general or strategic level there is room for considerable doubt and disagreement.

Four theories of the case

In order to simplify matters, out of the many different theories about the case which can be advanced, each with possible variants, let us outline four main ones, two favouring the prosecution (or the thesis that Edith was guilty as charged), two favouring the defence.

1. The conspiracy theory: that this attack was planned by Edith and Freddy on the day before. Freddy set out carrying a knife with the intention of killing Percy as the Thompsons returned home from the theatre. Edith expected this attack because she had planned it. In most versions of this theory Edith was the “master-mind”, Freddy the instrument. The actual attack was the culmination of a long series of attempts to kill Percy. This was one of the theories advanced by the prosecution and the one that was emphasized by the judge in his charge to the jury. An important variant of this theory is that Edith and Freddy had agreed to try to kill Percy whenever opportunity arose, but that Edith had not necessarily planned or expected this attack.

2. The incitement theory: Edith may not have known of or expected this attack. But she had over a long period deliberately tried to persuade Freddy to kill her husband — working on him by direct incitement, by innuendo, by suggesting different ways of getting rid of him and by claiming (whether truthfully or not) that she was prepared to risk trying to kill him herself and that she had indeed made several unsuccessful attempts to do so. There was, in short, a protracted and continuous incitement of Freddy by Edith to get rid of her husband. The particular timing and method of the attack were immaterial; Edith deliberately influenced Freddy to kill Percy. One variant is that there was some specific act of incitement that influenced Freddy on this occasion.

3. The fantasy theory: The attack on Percy was probably unpremeditated and was certainly totally unexpected so far as Edith was concerned. At no stage had Edith tried to kill her husband; nor had she deliberately, recklessly or even inadvertently incited Freddy to do so. In this view many of the key passages had an innocent explanation, and even the potentially most damaging statements merely represent the outpourings of a vivid imagination and were so interpreted by Freddy. When Edith wrote to Freddy she entered a world of daydreams and make-believe and nothing that she said constitutes evidence of intent to kill or to incite Freddy to kill in the real world. They were, as Freddy unromantically put it, mere “melodrama”. A variant of this is that some of the passages are to be interpreted as references to attempts to kill Percy, but they were merely part of a game, possibly a sexual game, designed to give spice and excitement to their relationship – and again this is how they were interpreted by Freddy.

4. The broken chain theory: This theory denies any connection between Edith’s behaviour and Freddy’s act. At one or more points the alleged chain of connection is broken. The theory takes two main forms. One version consists of a straight denial of one or more key points in the prosecution’s theory or theories without advancing an alternative account of what happened. An obvious example is a denial that the attack was premeditated. An alternative version is to present a rival account of the attack (and of preceding events) which serves the same function. For example, Edith lived in an imaginary world of passion, daring actions and desperate measures, a world in which fact and fiction and fantasy were inextricably mixed. She gave full rein to her imagination in her letters, pretending, perhaps even imagining, that she had tried to kill Percy. All the possibly damaging passages are in the earlier letters which represent a particular phase in her relationship with Freddy.

In this view it is not necessary to reach any firm conclusion as to what these passages might mean, whether for example she was claiming, whether truthfully or not, to have put broken glass in Percy’s food; even if the most damaging interpretation is put on the key passages relating to incitement, conspiracy and attempted murder there is no sufficient connection between the actual killing and the letters because: (a) Freddy did not take them seriously; and/or (b) the relationship had entered a quite different phase in the period before the killing; all the passages which might bear a damaging interpretation were written six months or more before Percy’s death; and/or (c) the actual attack was spontaneous rather than premeditated. Any one of these is sufficient to break the connection between Edith’s letters and Freddy’s act.

To recapitulate briefly: the conspiracy theory suggests that this attack was either planned shortly before by both accused or was the culmination of a more general plan; the incitement theory suggests that this attack, whether or not it was planned or premeditated, was the direct result of a continuous and protracted campaign of incitement by Edith; the fantasy theory suggests that all the potentially damaging passages in the letters are either open to alternative, innocent interpretations or were merely figments of Edith’s heated imagination and were interpreted by Freddy as “melodrama”. The broken chain theory suggests that whatever interpretation is put on the letters there was no direct connection between the killing of Percy and the letters (or Edith’s other actions).

It is worth making some points about these general theories about the case. First, they are not all of the same kind. The first three each involved advancing a coherent account of what happened — constructing an integrated story that hangs together as a whole. So does the second version of the broken chain theory. There is scope for a few loose ends or ambiguities or gaps, but the overall plausibility or credibility of each theory depends in part on its internal coherence. Does the story as a whole form a unity and fit the available evidence?

The first version of the broken chain theory lacks this positive quality. Essentially it consists of a series of negative assertions about key elements in the other theories. It does not necessarily involve constructing an internally consistent story about what happened. One of the notable features of the trial is that, contrary to the strong advice of her counsel, Edith Thompson insisted on going into the witness box. This had several catastrophic results. It gave the prosecution the opportunity to extract some damaging admissions from her in cross-examination; and she had to advance a single coherent account of what happened and to give her interpretation of key passages in the letters and to have these positive assertions challenged by the prosecution. She thus gave the prosecution a specific target to attack and made it difficult for the defence to cumulate doubt about different elements of the prosecution’s case. For the broken chain theory allows what amounts to a series of arguments in the alternative along such lines as these:

1. None of the relevant passages in the letters supports the view that Edith either tried or pretended to try to kill her husband.
2. Even if she had tried, the chain is broken because Freddy did not believe her.
3. Even if Freddy believed her, the chain is broken because he did not act on that belief at the time and there is no evidence of a connection between the letters and his behaviour on the night.
4. The evidence does not support the proposition that the attack was premeditated — that Freddy set out with the intention of attacking Percy; but even if the attack was premeditated the evidence does not support the contention that Edith had inspired Freddy’s plan.

By giving evidence, Edith Thompson gave the jury the chance to choose which between two competing stories they found to be plausible, rather than to judge whether each of the key elements in the prosecution’s case was established beyond reasonable doubt on the basis of the evidence. Or again, instead of arguing that a particular passage might mean A or B or C or D rather than E, as the prosecution suggested, the defence was committed to giving a single interpretation to each key passage and to having that interpretation challenged. Almost all commentators agree that it would have been much more difficult to convict Edith if she had not given evidence.

Of course, historians do not have the option of deciding not to call a witness. That is a matter of legal procedure. However, the broken chain theory could well provide a sufficient answer to our historian’s question. For if one or other version of that theory is accepted as correct, according to whatever standard is thought appropriate, then it establishes that Edith was not guilty as charged.

Each of these “theories” is a sketch of a complex argument. I have only noted some of the more obvious variants of each of these four main lines of argument. The range of possibilities is immense. There is considerable potential overlap between several of them. Enough has been said to illustrate how, even at this stage of general strategy, in constructing and criticizing this kind of argument, judgement and choice are unavoidable. There is nothing “mechanical” about the art of analysing evidence. This is one reason for doubting Wigmore’s claim that his Chart Method “is the only . . . scientific method” .

Experience suggests some useful working rules of thumb in exercising the art at this strategic stage:

1. “Strategy first, tactics later”. Despite the kinds of complexities that have been illustrated, strategic ‘theories’ are powerful simplifying devices: they are a means of structuring and managing and selecting material, however extensive and complex it seems.
2.  “Proceed dialectically.” The historian and the lawyer need to recognize the weak as well as the strong points in their own and rival theories.
3. “Construct the strongest version that you can of each theory and identify its weakest points.” What constitutes strength may, of course, be rather different for a trial lawyer and a historian.
4.  “Go for the jugular.” That is to say select one or two or a few key points in an argument and concentrate on those.

Such precepts are commonplaces of advocacy and, perhaps less explicitly, of analysis of evidence by historians and others. They are essentially heuristic techniques for making complex problems more manageable. All presuppose a reflexive view of the process: one cannot construct theories, judge their strength and identify key points without regard to detail. On the face of it the advice may seem to suggest moving from the general to the particular; in fact familiarity with the particular is almost always a pre-condition for clarifying the general.

Sub-theories, sub-plots and characters. Each theory is dependent to a significant degree on one or more sub-theories. First, an assessment of Edith’s character, or at least some important aspects, is relevant to each of them. For example, the fantasy and incitement theories depend very largely on the view one takes of Edith and a view of her character is relevant, though less important, to the conspiracy theory. It is not so important to the first version of the broken chain theory, which depends much more on an interpretation of Freddy’s likely reactions to the letters and whether the attack was spontaneous or premeditated. Here the focus is more on Freddy, but an assessment of Edith’s character is important for all the theories both as an aid to interpreting the letters and as part of interpreting the nature and course of their relationship. For example, both Edith and Freddy claimed that the phrase “only three and three quarter years left” in Edith’s very last letter (exhibit 60) referred to a suicide pact. That phrase is relevant to all four theories. Whether there was in fact such a pact, whether Edith had suggested it, but Freddy had not taken it seriously or whether this is an example of Edith’s alleged fantasizing depends in part on one’s views of Edith’s and Freddy’s characters, in part on one’s view of their relationship, and in part on other evidence (for example, other passages ostensibly referring to the same topic).

The main facts about the three principal protagonists and of their relations with each other are not seriously disputed. But there is room for different interpretations of their characters, especially that of Edith, and of Bywaters’s attitude to Edith at the time of the killing.

Percy Thompson had married Edith Graydon, who was four years his junior, in January 1916. There were no children of the marriage. Percy appears to have been a steady, respectable, dull, not particularly successful shipping clerk. There is evidence that he was suspicious and jealous of Edith’s relations with Freddy. She expressed fears that he would be violent and there is some evidence that he assaulted her on at least one occasion. She speaks of “submitting” to him sexually on a number of occasions but also refusing him without repercussions.

Edith Thompson had a striking personality. The undisputed facts about her include the following: She had worked as a book-keeper and manageress for a wholesale milliner’s in Aldersgate for several years and was well-regarded by her employers. She earned slightly more than her husband. She had several male admirers in addition to Freddy, but there is nothing to suggest that she had had extramarital sexual relations with anyone but him. At the time of the murder Percy Thompson was thirty-two, Edith was twenty-eight and Freddy twenty. It is clear that Edith’s marriage to Percy was not happy; it is less clear what price she was prepared to pay to escape from the security, comfort and respectability of her domestic situation. We have plenty of material on which to base judgements about Edith Thompson’s character, but it is open to a variety of interpretations.

Frederick Bywaters has been described as “a clean-cut, self-possessed, attractive-looking youth of twenty with a good character and record”. He had gone to sea before he was sixteen, and at the time of the murder was variously described as ship’s writer, a clerk and a laundry steward. During the period of his relationship with the Thompsons he was at sea for well over half of the time and this was the occasion both for the protracted and intense correspondence between them and for the possible fluctuations in their relationship. During the trial he instructed his counsel to conduct the case so as not to prejudice Edith’s chances of acquittal. This won him a lot of sympathy from public opinion, as reflected in the popular press, but it may possibly have damaged Edith’s case, as he appeared as the dignified, loyal lover, whereas she was seen by some to be solely concerned with saving herself.

The variety of interpretations of Edith’s character to be found in the literature is both fascinating and bewildering. The prosecution — backed by the judge, and seemingly by the Home Secretary and most public opinion — saw Edith as an immoral, scheming, manipulative older woman — a sorceress, who insidiously and persistently worked on an impressionable, inexperienced younger man to kill her husband or to help her to kill him. Other writers have compared Edith to Madame Bovary, and even to Marilyn Monroe. What these various interpretations have in common is that they are highly speculative, confidently asserted, impressionistic judgements by amateur psychologists. There are passages in the letters taken singly which can be used to support each of these and no doubt many other interpretations. Nowhere in the literature is there any attempt to build up a careful picture of Edith’s personality and behaviour on the basis of considering the evidence as a whole using any method other than impression. It may be possible for psychologists (or psycho-historians) to construct a less speculative profile of Edith on the basis of the available evidence.

Another sub-theory concerns Freddy’s character. Apart from points already mentioned this is relevant to assessing his credibility as a witness and to the question whether the attack was premeditated or spontaneous (crucial to all but the fantasy theory, important for the broken chain theory and relevant to the other two).

Some sort of picture of the nature and course of the relationships between the main actors is also relevant at least as background to all the theories of the case; it is crucial to parts of the broken chain theory, in that this is greatly strengthened if one can show a cooling off on the part of Freddy after the most damaging letters were written (there is quite cogent evidence in favour of this) and a calmer, less frenetic reaction on the part of Edith. However, even if it were accepted that at one stage Freddy had tried either to break off or to cool down the relationship (“Can we be pals, only”, he had suggested in September) there is a counter-theory, to the effect that there was a revival of passion later, backed not least by the fact that Freddy killed Percy. Without going further into detail here I wish to suggest that in order to assess the credibility of the broken chain theory in toto and responses to it, it is important to consider in meticulous detail the precise nature of the relationship at a number of stages in the story and to test the hypothesis that the early letters belong to a distinct phase which was significantly different from the state of the relationship in the days immediately before Percy’s death.

Macroscopic analysis: a preliminary stock-taking

One of the main claims made for Wigmore’s method is that it provides a powerful tool for organizing and mapping complex arguments based on masses of evidence. The guiding rule of thumb is to start with the conclusions (the ultimate probanda, facts-in-issue) and to work down or back to the evidence. The ideal type of an argument of this kind is pyramid-shaped with the apex being simpler than the base. In adjudication, there is typically only one apex (guilty/not guilty; liable/not liable), but usually multiple facts in issue are involved and we talk of these as the “ultimate probanda”. It is almost always essential to define each fact in issue before proceeding to analyse the evidence relevant to it.

At first sight the preliminary stages of the analysis of the case of Edith Thompson may not seem to be a good advertisement for Wigmorean analysis as a simplifying device. The first stages have not been simple. But this is due to the nature of the case rather than the method. The case is usually considered to be complex because of the volume of the evidence and the difficulties involved in interpreting Edith’s letters. The first stage in our analysis has revealed a number of further complications.

First, the ultimate probanda are more than usually complex. Many famous trials involve ultimate probanda in which the prosecution had to prove A + B + C + D, when only one element (e.g., identity or criminal intent) has been the subject of serious doubt or dispute. In the case of Edith Thompson the facts in issue exhibit a more complex form: the prosecution case involved a mixture of necessary and sufficient conditions which can be represented as follows:

A (Freddy murdered P) + B [(i) or (ii) or ((i) or (ii) or, possibly, (iii))].

Secondly, some doubts surrounded the law governing the facts in issue: was it sufficient to prove a general conspiracy to murder Percy (B (i)) or did the prosecution have to prove that this attack was planned? What precisely is the scope of the proposition “E incited F”? Is the scope of “incitement” a question of law, or fact, or mixed fact and law? And, although it was not raised at the trial, would Edith be responsible, if the prosecution established the “exploding bomb” theory? This merely illustrates the general point that to the extent that the interpretation or application of the governing law is unclear, the task of constructing or evaluating arguments about disputed questions of fact becomes correspondingly more difficult.

Thirdly, the complexity of the ultimate probanda is reflected in the multiplicity of potential theories of the case, each of which has variants. What this means is that there are several potential possible lines of argument both for and against Edith. A “comprehensive” analysis of the case would need to map and evaluate each of these arguments.

Fourthly, some of the lines of argument overlap: for example, as we shall see, the “fantasy theory” may represent both a sufficient argument for exonerating Edith (in respect of both conspiracy and incitement), and a substantial prop for the broken chain theory. This is not uncommon in disputed cases involving multiple lines of argument.

Fifthly, in many causes célèbres the sole or main disputed issue relates to some relatively “hard fact”, such as identity or alibi or the death of the victim. In this case we are dealing with the mental state of not one, but two people: did Edith intend . . . etc? Was Freddy’s act caused/inspired/done in furtherance of Edith’s incitement or plan? Proof of mental states is notoriously more elusive than proof of “simple facts”.

Finally, as we have seen, some of the intermediate probanda that are potentially important in most of the theories involve interpreting the characters of and the relationship between the two main characters.

The case clearly illustrates how complications can arise at the top of the pyramid as well as at other points. Nevertheless the first steps in the analysis have imposed some order on to the material. Our standpoint has been clarified; the range of potential ultimate probanda, albeit unusually complex, has been laid out; four main theories of the case have been articulated, with scope left for dealing with variants of these where appropriate. We are now ready to proceed to the next stage.

Microscopic analysis: premeditation

All Wigmorean analysis involves selection; “completeness” and “comprehensiveness” are only relative matters in this context. However, the purpose of this essay is to illustrate the application of this method to a particularly complex case. Accordingly, I propose to proceed highly selectively by concentrating on one small sector of potential arguments about the case as a whole.
First, let us look at just one element in the broken chain theory — the proposition (intermediate probandum) that “Freddy’s attack on Percy was spontaneous/unpremeditated” and the main evidence used by the prosecution to support premeditation.

I shall use this to illustrate a number of points about Wigmorean analysis (e.g., generalizations, prejudice and the use of emotive language); but there is another reason for selecting this as a focal point. It is a maxim of advocacy that one should “go for the jugular”. If a reasonable doubt is established about the proposition that “the attack was premeditated”, then this breaks the connection between Edith’s thoughts and actions and all the prosecution theories, except possibly the legally dubious and highly speculative “exploding bomb” theory. This is one potential “jugular” which could be sufficient to destroy all versions of the case against Edith. Whether or not the analysis achieves this result here — I leave that to the reader to judge — it may illustrate the potential value of careful microscopic analysis of one or more key points in a highly complex argument the main outlines of which have been firmly set by careful preliminary macroscopic structuring.

Let us start then by constructing the strongest argument that we can, based on the available materials, that Freddy’s assault on Percy was unpremeditated. This sub-theory might be constructed in outline as follows:

When Freddy set out from the Graydons’ he did not plan to meet Edith and Percy.

1. A planned murder would probably have: (a) been carried out in the absence of Edith; (b) been executed without a struggle; (c) been carried out in some other place.
2. Edith was genuinely surprised by the attack.
3. Freddy’s purpose in meeting Percy was to ask him to grant Edith a divorce. The stabbing occurred during an unexpected struggle when the confrontation misfired.
4. Freddy’s carrying of the knife was innocent.

All five props in this argument are consistent with Bywaters’s conviction for murder and the rejection of his self-defence theory. These five theses are separable, but can be combined in various ways. I shall concentrate on (5), which was treated as crucial in the case. Suffice it to say, in my judgement, there is quite strong support for (3) and that there is little or nothing in the record to negate (1), (2) and (4), but there is little beyond Freddy’s story to give them direct support, except some speculative generalizations which give weak support to (2) and (4).
In order to convict Edith the prosecution had to establish not only that Freddy murdered Percy but also that the attack was premeditated. The two main props for the latter proposition are: (a) that the carrying of the knife was not innocent; (b) that there is clear evidence that this attack was planned. Establishing (a) or (b) or (a + b) beyond reasonable doubt would be sufficient for this purpose. Let us look at (a) and (b) in turn.

(a) The knife. Reconstruction of the actual arguments used at the trial in respect of the knife is particularly revealing. The prosecution argument was in effect: Bywaters had purchased the knife with the intention of killing Percy and he had put it in his overcoat pocket on the day of the murder for the same reason. The defence claimed that Freddy had bought the knife a long time previously, that he had taken it abroad with him on his last voyage, that he regularly carried it around with him and that it was natural that he should do so. The arguments actually used by counsel for each side and by the trial judge in his direction to the jury can be reconstructed as in table A. In the process of the arguments about premeditation, the weapon was variously described as an ordinary sheath knife, an English hunting knife, a deadly weapon, a dagger, a stabbing instrument, and a dreadful weapon. Interestingly, some of the more emotive terms were used by the judge.

There were also some striking differences in the generalizations invoked in the context of considering this issue. Counsel for the Prosecution suggested that possession of a knife of this kind is in itself suspicious; in order to cast doubt on the suggestion that Bywaters had possessed the knife for some time, the prosecution suggested that if this were true “it would have been a subject of jocular remarks”; the judge said: “It is suggested that no reasonable man living in London carries a knife like that about in his pocket.” Defence counsel, on the other hand, in the course of a single paragraph said: “It is not strange for a sea-faring man, visiting foreign countries, to purchase a knife”, and, “There are few sailors who do not possess a knife.” The editor of the record in the Notable British Trials series commented on the last remark in a footnote: “Bywaters was not a ‘sailor’ in the technical sense. He was a clerk on board a ship, and had more use for a fountain pen than for a knife.”

In the preceding examples not only is Bywaters variously categorized as a sailor, a sea-faring man visiting foreign ports, a clerk, and a man living in London, but quite different generalizations are invoked about the purchase, possession and carrying of knives. At the trial the jury had the opportunity to inspect the knife itself and thus could form their own judgement on its description and on the specific issue of whether it would fit conveniently in the pocket of Bywaters’s coat. They could also check the accuracy of the various categorizations of Bywaters: for example, it is arguable that he was both someone living in London and a seafaring man, who visited foreign countries, but the jurors were probably only slightly better placed than the modern reader for making confident judgments about the contemporary habits of merchant seamen or ship’s clerks in respect of knives or of the likelihood that if someone possessed a knife of this kind for a considerable period it would be the subject of jocular remarks.

The example of the knife in Bywaters and Thompson is a simple example of the leeways for choice in selecting and in formulating generalizations as part of an argument about a particular issue of fact. It shows how there is room for the use of emotive terms, for distortion, and for selection by emphasizing different aspects of the same situation. Moreover, it is an example of invoking commonplace generalizations, for example about the normality of carrying knives about London and of sailors possessing and carrying knives, about which the jury is expected to rely on its own version(s) of general experience in order to come to a conclusion. Clearly there is scope in this kind of context for the intrusion of bias, prejudice, and sheer speculation. Moreover, it is doubtful whether all the relevant background knowledge can ever be fully articulated. But, as Wigmore might have argued, what better basis is there for making such judgements?

The date of purchase of the knife and the issue whether Freddy was carrying it with the purpose of attacking Percy were perceived by the main participants to be significant. Almost all the available information about it can be gleaned from the explicit arguments advanced in the case. To this can be added a few further propositions: Mr. H.W. Forster of Osborne and Co., tool merchants of Aldersgate, testified that his business sold identical knives for six shillings and had stocked them for about seventeen years. Neither the defence nor the prosecution attempted to adduce evidence corroborating or negating Bywaters’s statement that he had purchased the knife in November 1921, i.e., almost a year before the attack. If the defence had managed to find one or two of Bywaters’ shipmates who would testify that he had owned a knife for a long time and that he regularly carried it about with him, this could have significantly strengthened Edith’s case.
A key-list of propositions relating to the propositions that (a) the purchase and (b) the carrying of the knife were both innocent, based on the available evidence, and constructed from the point of view of a relatively detached historian, might read as follows:

1. The purchase of the knife was innocent. The carrying of the knife was innocent.
2.  The knife was bought in November 1921.
3.  Freddy Bywaters (testimony, 53).
4.  Freddy Bywaters carried the knife everywhere in England.
5.  Freddy Bywaters (cross-examined, 71).
6.  It was handy at sea and handy at home.
7.  Bywaters (70).
8.  Bywaters bought the knife from Osborne and Co.
9.  Knives of this kind were sold by Osborne and Co.
10.  Such knives had been stocked for about 17 years.
11.  H.W. Forster (Director of Osborne’s) testified to 8 and 9 (42).
12.  Defence called no witness to corroborate Freddy’s possession of knife since November 1921.
13.  Shearman, J. (137).  “It was difficult to put the knife into any kind of pocket, except the side pocket.
14.  Generalizations about the normality of carrying such a knife about — see Table

Prima facie this key-list adds little to the arguments presented in the case. However, it can be used as the basis for some other propositions which further weaken the premeditation thesis:

1. There is no evidence that Freddy Bywaters bought the knife on the day of the killing.
2. If Bywaters already owned the knife (1 + 16), he had it with him when he left home.
3. (17) Bywaters carried the knife about with him from the time he left home until after the killing.
4. Bywaters had the knife on him during at least three social engagements (morning coffee and afternoon tea with Edith; visit to the Graydons).
5. The knife fitted easily in his coat.

All of this helps to bolster Freddy’s claim that the purchase and carrying of the knife were both innocent. Furthermore, (3, above) suggests that the premeditation thesis requires that Freddy had formed the plan to attack Percy before he left home. Yet none of the several witnesses who saw him during the course of the day observed anything strange or unusual about his demeanour. The net effect of all this, I would suggest, is to give some, albeit weak, support to this part of Freddy’s story; it suggests that there is nothing in the evidence about the knife (other than the fact of its possession and use in the killing) which supports the premeditation thesis.

In respect of Edith, it was for the prosecution to prove that the attack was premeditated. Even if one totally discounts Freddy’s evidence about the events of the evening (and his story of the period up to 11 p.m. was generally consistent and was largely corroborated by the Graydons), there is almost nothing to support the proposition that the attack was premeditated. There was no evidence to support the proposition that the knife was purchased recently in order to attack Percy; there was no evidence in support of the proposition that Bywaters put the knife in his pocket that morning because he planned to attack Percy — the best that the prosecution could do was point out that there was no corroboration for his claim that he was in the habit of carrying it.

The alternative hypotheses about his state of mind (planned attack, loss of control in the heat of the moment) are essentially speculative. The evidence relating to the wounds may support an inference of intent, but it is about evenly balanced (and weak) in respect of the question of premeditation versus spontaneity. More information about the date of purchase of the knife, whether Bywaters did regularly carry it about, and its size in relation to the pockets of the overcoat might have helped one side or the other. However, on the basis of careful scrutiny of the available evidence, my conclusion is that it gives only weak support to either side and, on balance, it marginally favours Edith. Accordingly there is considerable doubt about this crucial aspect of both prosecution theories.

(b) Edith’s last letter: the ‘tea-room’ passage (exhibit 60). The other specific prop of the premeditation thesis was the closing passage in Edith’s last letter to Freddy (exhibit 60). In presentation, the prosecution did not subject this to close scrutiny, preferring to concentrate on building up a general impression of a continuous process of encouragement and stimulation (“by precept and example, actual or simulated”). Edith incited Freddy and plotted Percy’s death with him and this led directly to his death. In the case Edith’s letters were admitted as evidence of motive and intent, but were used for other purposes, including as direct evidence of conspiracy. Since the judge ruled that it was necessary to establish that Edith incited or planned this attack, the last letter in fact formed a crucial part of the prosecution case.

I shall argue that the letter at the most gives only very weak support to the prosecution and that there is almost no other evidence to support the proposition that this attack was planned. It is worth subjecting one key passage to detailed scrutiny as an illustration of the potential methods and value of microscopic Wigmorean analysis.

There is some uncertainty about when this letter was written and when it was received; let us, for the sake of argument, take the timing most favourable to the prosecution: that it was written on the day before the killing (i.e., on 2 October) and received before Freddy left his mother’s house on the 3rd. It is reproduced in full below. The whole letter was the main evidence used by the prosecution in favour of the conspiracy theory. Let us focus on the following words which were relied on by both sides to support their case: “Don’t forget what we talked in the Tea Room, I’ll still risk and try if you will — we only have 3 1/2 years left darlingest. Try and help.” The prosecution theory was that ‘what’ referred to killing Percy; the defence claimed that ‘what’ referred to Freddy trying to find Edith a post abroad so that they could elope.

If, from the standpoint of a historian, we take a fresh but careful look at this passage, it is clear that it is susceptible to quite elaborate textual analysis. Here I shall only sketch a possible approach. One might, for example, start by listing a series of propositions about possible referents of ‘what’ (also of ‘risk and try’ , ‘try and help’: risk what? try what? and so on) as follows:

1.  ‘What’ could refer to almost anything, i.e., there is no clear referent.
2.  ‘What’ could refer to killing Percy.
3.  ‘What’ could refer to getting a post abroad.
4.  ‘What’ could refer to eloping.
5.  ‘What’ could refer to a secret assignation.
6.  ‘What’ could refer to asking Percy for a divorce.
7.  ‘What’ could refer to a general conversation covering several or all of the above and possibly much else besides.
8.  ‘What’ could refer to whether to use poison or a dagger (Shearman, 151).

This probably covers the main possibilities, though the list is not exhaustive. There is a number of potential aids to interpretation which might give support to one or more of these hypotheses and eliminate others. Such aids include: the immediate context of these twenty-seven words; the context of the letter as a whole; the context of the whole corpus of letters and of particular passages in other letters (for example, Edith often uses the words ‘risk’ and ‘try’); one or other sub-theory about the state of the relationship between Edith and Freddy; extraneous evidence might be sought from people who saw (or possibly overheard them) in the tea-room; the conduct of Edith and Freddy before and after the meeting; the occasion for writing the letter might also be investigated; and, of course, Edith and Freddy could be asked about the meaning of the passage.
Both the prosecution and the defence, in a very crude fashion, used something akin to textual analysis of the ‘tea-room’ passage in their arguments; but, in my view, they both cheated.

This passage is the main evidence in favour of the proposition that this particular attack was planned, i.e., the conspiracy theory. The prosecution emphasized the words ‘risk’, ‘try’, ‘help’ and ‘only’ in the passage itself to bolster the idea that this referred to a sinister conspiracy. Then by taking a number of other words and phrases out of context (‘great big things’, ‘he was suspicious’, ‘do something tomorrow night’) and linking them to the fact of Percy’s death and other letters, some of which were ambiguous and written over six months before, they suggested that the conversation in the tea-room was the culmination of a continuing conspiracy.

The defence emphasized ‘only 3 1/2 years’, (strange if one was planning something immediate) and the fact that Edith and Freddy both testified that ‘what’ referred to eloping and risking her future with Bywaters, but then counsel used the fact of the prosecution’s misuse of this passage to cast doubt on their interpretation of all other passages, while carefully skipping over any explicit reference to any of the most damaging ones.

Even more remarkable is Mr Justice Shearman’s treatment of the passage; for he asserts, without any basis whatsoever and going well beyond what the prosecution had argued, that ‘what’ refers to whether it was better to kill Percy by means of poison or a dagger.
A more careful analysis using the aids mentioned above, can take us some way beyond this. First, we can quite confidently eliminate some of the possible hypotheses. For example, the judge’s suggestion can be attacked on the following grounds:

1. it is sheer speculation, with no evidence to support it;
2. it involves a petitio principii in that it assumes what is seeks to prove;
3. it does not make sense of the passage: ‘Don’t forget what we talked in the Tea Room [about whether to use poison or a dagger], I’ll still risk and try if you will we have only 3 1/2 years left darlingest.’

Again, we can fairly confidently discard the first hypothesis that ‘what’ could refer to almost anything. The context of the letter as a whole, the context of the sentence, the general background of their relationship and the testimony of both Freddy and Edith all support the proposition that ‘what’ concerned some aspect of their relationship, rather than something else. It seems unlikely that it referred to buying a motor car or that morning’s world news, for example.
Of the remaining six hypotheses (2-7), only no. 2 favours the prosecution thesis that they were conspiring in the tea-room to kill Percy. No. 7 (general conversation about their situation) hardly constitutes evidence of conspiracy. The other four hypotheses all positively support the defence contention that the conversation had nothing to do with murder.

Even if we totally discount testimony of both accused that it referred to eloping (the ‘risks’ being financial and/or of social stigma), the context of the letter as a whole and the words ‘3 1/2 years left’ both tend to support the judgement that an innocent explanation is a good deal more likely than the prosecution’s interpretation. At the, very least, such factors seem to me to cast a reasonable doubt on that interpretation, yet this passage was the main item of evidence in support of the conspiracy theory. The passage supports the proposition that they met and talked seriously about their relationship, but it does not give any significant support to the proposition that they were conspiring to kill — it tends to negate it.

Undertaking such analysis and reading about it can appear wearisome to most people. It is much more entertaining to play with stories — and let Edith go hang. But this kind of approach can and does yield results. In this particular instance I do not think that it can do more than eliminate some of the possibilities and cast doubt on the wisdom of making any confident judgement about what was said in the tea-room; but in respect of other passages, including some that played a key role in the case, meticulous analysis can lead to the conclusion that one particular interpretation is very probably or almost certainly or even beyond peradventure, the correct one. However, in this particular case one’s conclusions about different individual passages in Edith’s letters tend to point in different directions. For example, the more carefully one studies the ‘Marconigram’ passage (Appendix B.1), the less easy is it to believe that Edith was merely fantasizing or playing games or that Freddy was not actively involved in whatever was being transacted between them at that time. On the other hand, some of the other apparently damaging passages either become hopelessly ambiguous in the light of careful scrutiny or else seem to help Edith in one way or another. In short, it is unlikely that this kind of analysis will ever conclusively resolve all the doubts about Edith Thompson’s guilt. In my view, it strongly substantiates the judgement that there is at least a reasonable doubt about it.

Conclusion

I have suggested that careful analysis of even such unpromising material as Edith Thompson’s allusive, ambiguous, inconsequential love letters can, at least reduce the range of plausible interpretations even if it cannot resolve all doubts. I would further suggest that such analysis is the best available way of getting as near as possible to the truth, but it was not the method used at the trial — nor is it a method which is well suited to our system of adversarial proceedings and oral presentation of evidence. In my view, the prosecution used selected passages unfairly by taking them out of context, by attributing particular meanings to them on the basis of innuendo, impression and bare assertion rather than analysis and argument. The reading aloud of the letters in court could create only a general impression, which is no substitute for careful study of the texts. Even if the defence had been handled more skillfully and if the emotional atmosphere and the pressure of time had been less, the basic flaw would have been the same: the oral presentation of such material, whether to twelve lay persons or to an experienced professional, cannot provide a satisfactory basis for the careful analysis and evaluation of the evidence. Thus my conclusion is that dispassionate, logical analysis of even such unpromising material as Edith Thompson’s letters is a possible, indeed even a necessary, method of trying to work towards the truth, but such a method is incompatible with oral presentation of the material — as much in a public lecture as in a courtroom.

 

PART II .   EDITH THOMPSON: FRESH EVIDENCE AND NEW PERSPECTIVES

One hazard of academic life is to learn of an important work on one’s subject shortly before or after completing a project. I was making final revisions to “Anatomy of a Cause Célèbre” (hereafter Anatomy) in June 1988 when I learned of the publication of Criminal Justice: The True Story of Edith Thompson by René Weis. This is by far the most substantial study of the life and death of Edith. Remarkably, Dr Weis and I are colleagues at University College London, yet over a period of five years neither of us knew that the other was working on the case. Dr Weis’s is a scholarly contribution to the general literature on causes célèbres and is the first full-scale biography of Edith. I had been using the case in teaching for rather longer, mainly to illustrate a method of analysis; a subsidiary aim was to demonstrate the potential and the feasibility of using trial records as standard materials of study in legal education.

Criminal Justice contains much new information and many insights about Edith Thompson. I share many of Dr Weis’s values and I agree with some, but not all, of his interpretations and judgements. We both conclude that Edith was probably wrongly convicted, but for somewhat different reasons. However, this section deals only indirectly with the “truth” about Edith Thompson. Rather its concern is methodological. It is a case study of the differences in approach of a jurist using modified Wigmorean analysis and a specialist in English literature using a sophisticated version of the “narrative method” to reconstruct and present the “True Story of Edith Thompson”. My objective here is to explore quite briefly and report as honestly as I can what I think a first reading of Dr Weis’s account has added to my understanding of the case in respect of the question: Was Edith guilty as charged? and, secondly, what modified Wigmorean analysis might have added to his account. This is a sort of “thought experiment”. In order to reach for relative detachment I shall refer to both authors in the third person.

Criminal Justice is a substantial book of over 300 pages. It is published by a general commercial publisher, Hamish Hamilton, and is presented in a form designed to attract the general reader. It is based on extensive and meticulous original research and contains some conventional scholarly “furniture”: a table of contents, a Preface, two pages of acknowledgements, a list of illustrations, a “Cast of Main Characters”, an appendix on sources and an index based almost exclusively on proper names. Significantly it has virtually no footnotes. Even more significantly for present purposes it is presented as a story: the arrange¬ment is chronological, starting with a description of Edith’s parents and ending with a harrowing reconstruction of the execution of Edith and Freddy and a brief account of the subsequent lives of Edith’s closest relatives. The whole forms a readable, coherent, and moving narrative biography. Weis states: “My readers will be the final arbiters of whether or not … I have successfully allowed Edith Thompson to argue her innocence.”

Criminal Justice differs from Anatomy in respect of, for example, purpose, length, style, source material, comprehensiveness, literary form and intended audiences. They are contrasting products of significantly different enterprises. There is, however, a solid basis for comparison, for both focus on the question of Edith’s guilt, using her letters as the main, but not the only, evidence. One can thus compare and contrast two different ways of treating selected matters bearing on this question, with particular reference to the relationship between narrative, analysis and argument.

Standpoint

The first step in any such analysis is to clarify standpoint by asking: Who am I? At what stage in what process am I? What am I trying to do? Partial answers to these questions in respect of Criminal Justice and Anatomy have already been indicated. However, posing the questions more sharply suggests some further affinities and contrasts. Both authors purport to adopt “the standpoint of the historian” exploring the question of Edith’s guilt from almost the same vantage-point: University College London, in the 1980s. Neither is a professional historian; both are liberal academics with a humanistic bent. They have produced works for publication at the end of rather different processes of research. Personal factors apart, some of their differences in perspective might be attributable to the kinds of differences one might expect between a legal and a literary mind: Twining more drily analytical; Weis more practised in the interpretation of literary texts and character. But, one may ask: Is it sensible to think in such terms or to draw sharp distinctions between legal and literary historical truth? Are not the similarities in standpoint much greater than the differences, given that they have both been addressing essentially the same central question?

Weis’s story is based on extensive original research. He has located an impressive range of materials in newspapers, official records (some made publicly available for the first time), “evidence” to two Royal Commissions, Hansard, and, of course, the record of the trial (edited by Filson Young) and numerous secondary works, few of which, he says, add much to Young’s volume. Weis can claim some notable coups: he has unearthed some new letters from Edith, Freddy, members of their families and others. Apart from their intrinsic interest some of these are directly relevant to the question of Edith’s guilt. He has also interviewed a number of people who knew some of the actors or who were able to throw light on the environment of the main characters and the central action of the story. Almost as significantly, Weis reports that some potentially crucial documents are missing (including the originals of Edith’s letters) and some official files are still closed. Weis uses his material to give the fullest account to date of the background, biographies and personalities of the main actors, but both authors use the published version of Edith’s letters as their primary source for the events leading to Percy’s death.

The facts in issue: elucidating the question

Weis interprets the charge of murder against Edith almost exclusively in terms of conspiracy , as did Mr Justice Shearman in his charge to the jury. However, he interprets the relevance of conspiracy rather differently from both Shearman and Twining. The former directed that the jury should only convict if they were satisfied beyond reasonable doubt that the lovers had planned this attack. Twining suggests that it would have been sufficient to convict Edith if the jury was satisfied that Freddy acted in pursuance of a continuing conspiracy to kill Percy whenever opportunity arose. Weis tends to gloss over this distinction. He also treats the question whether Edith and Freddy had conspired to poison Percy in the period up to May as central to the question of Edith’s guilt. Twining on the other hand argues that under the broken chain theory, even if Edith were held to have attempted to poison Percy or to have conspired with Freddy to do so, these acts were too remote from the fatal attack to be sufficient to implicate her in the murder. As the prosecution tried to argue, some of the early letters could be interpreted to support the theory of a protracted and continuing conspiracy, but the defence countered that the evidence failed to support the hypothesis that, if there ever was a conspiracy, it was continuing or that Freddy had acted because of it. Twining differs from Weis in being much less confident that all the early letters are susceptible to an innocent interpretation, but believes that far less turns on them than Weis thinks.

Weis pays relatively little attention to the significance of incitement, both in respect of the murder and of interpreting some of the potentially most damning passages in the letters. Twining, on the other hand, argues that Edith was more vulnerable in respect of incitement than conspiracy, even without resort to the legally dubious, but psychologically plausible “exploding bomb” theory. However, on his analysis the chain was also broken in respect of incitement.

It might be argued that since the trier of fact was a jury of lay persons charged with rendering a general verdict, they would not in practice be concerned with such legal niceties. They made an undifferentiated (holistic?) judgement about Edith’s legal (and moral) responsibility for Freddy’s act and this is what society expects of juries. Accordingly this should also be adequate for an historian. On this view, Twining’s differentiations are unduly legalistic and academic. Twining dissents. Such an argument is not qualitatively different from the judgement, which may well be historically correct, that Edith was “hanged for adultery”. In so far as our enterprise is rationally reconstructing the strongest possible arguments concerning Edith’s legal responsibility for Percy’s death, precision is very important indeed. The crucial point is that under Twining’s interpretation of the law the prosecution had several alternative routes open to them and Edith’s defenders need to deal with all of the main alternatives. Wigmorean analysis requires as precise a clarification of the legal issues as legal interpretation will admit; and where, as in this case, there are multiple facts in issue, some of which are open to more than one interpretation, arguments have to be constructed about each of them.

Theories of the case

In Anatomy Twining outlines four main “theories of the case”, each with several possible variants and sub-variants: the conspiracy theory, the incitement theory the fantasy theory and the broken chain theory. A statement of a theory in this context means a summary statement of a general argument about the facts in issue. A theory in this sense is an analytic device of macroscopic analysis: for advocates, such theories typically form the basis of a strategy of argumentation; for historians and the like they are similarly a useful instrument for organizing complex material. In analysing the present case, outlining four main theories and their variants helps to clarify how each side had several different routes for reaching the conclusion for which it was arguing. Twining suggests that the broken chain theory is the most plausible of the four. It has the special advantage that it can be used to cumulate doubt about opposing theories without necessarily claiming to advance a fully coherent account of “the truth” about Edith Thompson.

This theory was the main basis for the defence strategy at Edith’s trial, but was undermined by her insistence on testifying and therefore being compelled to advance one account of key events and one interpretation of some passages in her letters. Most secondary writers who are sympathetic to Edith, including Weis, rely heavily on the fantasy theory. This has the most human interest and fits in with the function of general literature on causes célèbres which tends to be judged by its entertainment value as well as, or more than, its contribution to knowledge. Twining, however, believes that the fantasy theory is very difficult to sustain by evidence and argument. It is both highly speculative and difficult to reconcile with some of the data. It is in his view the least cogent of the four main theories, although it is not as weak as the tea-room conspiracy theory.
Weis does not use the concept of “theory” to differentiate between lines of argument. It is reasonable to interpret Criminal Justice as being based primarily on a version of fantasy theory with a few “broken chain” points used to supplement his case. As noted above, he does not differentiate very clearly between the various competing theories and their variants. This blunts the force of his negative attack on rival hypotheses and leads him to underestimate the importance of incitement in building up the case against Edith, especially in providing alternative interpretations of some of the letters.

On the positive side it is probably fair to say that Weis, by sustained and careful interpretation of Edith’s letters, constructs the most plausible and best supported version of the fantasy theory yet published. Most secondary writers assert rather than try to prove the fantasy theory and, in so doing, commit, or come close to committing, the fallacy of petitio principii. Weis, through skilful use of Edith’s letters to document the theme of the constant intermingling of fact and fiction, at least provides an impressive number of texts to ground the thesis. His subtle reconstruction of Edith’s character, not least through a perceptive use of her reactions and commentaries on the novels she has been reading, lends support to the fantasy theory. At the very least he provides a coherent and plausible interpretation of the letters read as a whole that is consistent with this line of argument. However, it will be argued below, his interpretations too often involve assertion rather than argument. Moreover, the fantasy theory is by its nature heavily dependent on “soft” psychological speculation which is in danger of foundering at crucial points when compared with rather “harder” data.

A first reading of Criminal Justice suggests a further, ironic, twist. Weis’s version of the fantasy theory, even if true, does not necessarily absolve Edith from legal responsibility. Even if those passages which suggest that Edith had tried to poison Percy or that she is “dreaming” of Percy’s death are accepted as mere fantasy, she is still responsible for communicating those thoughts in a form which might reasonably be interpreted as being factual reports on her acts or desires, for example, that she had tried to poison Percy or that she wished him dead. Absent a positive defence of insanity such acts could legally ground a finding of incitement or conspiracy. There are some hard facts that need to be explained away: Edith did make these statements; she admits to having tried to manipulate Freddy on a number of occasions, as lovers do; she did send at least nine cuttings about cases of murder by poison; she did ask Freddy to send her “things” which might be interpreted as poisons; Freddy did supply her with some medicines; Edith more than once expressed a wish that Percy was dead; Freddy did kill Percy. The fantasy theory has a lot of explaining to do; it is inevitably speculative and fantasy can spill over into fact as easily as fact can spill over into fantasy – even if such a distinction can be maintained in legal or other discourse. This is not to say that the fantasy theory is untrue nor to suggest that it has no explanatory power. But as an argument it is not very cogent.

Sub-theories, sub-plots and characters

In Anatomy Twining argues that an assessment of the characters of the main protagonists and of their relationships to each other is relevant to each of the main theories of the case. He reports a number of different interpretations of these matters that were implicit in the arguments at trial or have been advanced in the secondary literature, but he does not seek to develop them in detail.

Criminal Justice, on the other hand, adduces a wealth of new biographical and background facts, and builds up in fascinating detail a picture of Edith, Freddy and Percy, as well as many other members of the cast of characters. It contains a rich, perceptive and plausible account of all these matters which cannot but add to our understanding of the story, even if one does not accept all of the interpretations. At the “macro” level, Weis’s most striking achievement is to piece together a coherent and detailed chronological account of the course of a number of key relationships – not only between Edith and Freddy, but also between the lovers and members of their families and their circle of friends and contacts. Of particular interest is the detailed account of the course of the relationship between Edith and Freddy, including some significant new information about their attempts to communicate with each other after the trial. Some of the detail may be irrelevant or tangential to the question of Edith’s guilt, but the story is well worth telling at length for its own sake and Weis tells it very well. The final chapters have been justly praised as a devastating depiction of the horrors of capital punishment.

Not surprisingly, this meticulous reconstruction has implications for all of the main theories of the case. Although it is presented as an argument in favour of Edith’s innocence, it also provides material that could be used for and against each of the main theories. To take but one example: Weis’s account of the relationship between Freddy and Edith (which had even more ups and downs than had previously been noted) supports the general thesis that the early letters are of very limited value in interpreting the events immediately preceding Percy’s death. It strongly negates any suggestion of either continuing conspiracy or protracted incitement. It gives some support to the contention that the prejudicial effect of the early letters outweighed their probative value and that they ought not to have been admitted at all. On the other hand, given time one could argue that some parts of Weis’s account could be used to argue that Edith was probably guilty of some of the lesser offences charged in the second indictment (which was not proceeded with) in the early phase of the relationship. However, that is beyond the scope of this essay.

Weis’s reconstruction of characters, relationships and sub-plots (including some new ones) is very helpful in interpreting and making sense of many passages in Edith’s letters. One does not need to accept all of his particular interpretations to recognize the value of these intermediate generalizations and of a coherent narrative as aids to developing well-grounded, mutually support¬ive interpretations of the meaning and significance of these elusive texts. The process is, of course, reflexive: one constructs a picture of a character or incident or sequence of events from the letters or other sources and uses these constructs to interpret the letters. This is quite compatible with a Wigmorean approach and Weis is arguably at his best at this level of analysis.

Microscopic analysis

Two main values are claimed for Wigmorean analysis: at the “macro” level it helps to clarify the central issues and impose order on a mass of mixed evidence or data through the development of carefully formulated hypotheses, theories and sub-theories; at the “micro” level it provides tools for sharply focused and detailed analysis of selected phases of an argument that have been identified as being potentially crucial or otherwise deserving special attention. At each level the main claim is that it helps to clarify what precisely is being argued as a preliminary to evaluating its cogency. ln Anatomy some particular examples were chosen to illustrate the application of microscopic analysis. Let us look at the treatment of just one of these in Criminal Justice, the last letter (exhibit 60) with particular reference to the “tea-room passage”.

Twining uses the last twenty-seven words of this letter as a vehicle for illustrating microscopic Wigmorean analysis. The passage was selected because at trial great emphasis was placed on it by the prosecution and the defence and Mr Justice Shearman, particularly in relation to the thesis that this attack was planned in the tea-room. Twining presents a fairly simple example of careful textual analysis, focusing particularly on the possible referents of “what” in the phrase “Don’t forget what we talked in the Tea Room” and, in less detail, on the questions: “Risk what?” “Try what?” in the phrases “I’ll risk and try if you will” and “Try and help”. He argues that by careful analysis of these twenty-seven words, backed by a range of contextual aids to interpretation (not least Edith’s and Freddy’s testimony at trial) the range of plausible interpretations can be narrowed and that, while the passage could refer to a plan to kill Percy, there are several more likely innocent interpretations. He concludes that the passage does not give any significant support to the conspiracy theory; if anything, it tends to negate it. Furthermore, the analysis suggests that the prosecution, the defence and the judge all used the passage in ways that do not stand up to close scrutiny.

Weis reaches a similar conclusion by a different route. He reconstructs the context of the letter and the events to which the key passage refers. He argues plausibly, partly on the basis of new evidence, that the letter was written before 5 p.m. on Monday 3 October (i.e. a few hours before the killing) and that the meeting in the tea-room took place on Friday 29 September and not on 3 October, as the prosecution had suggested. He states, “The Crown will mistakenly assume that the tea room conversation occurred on Monday afternoon, the day before the murder and that it ought to be causally linked to the tragedy.” This is a partial non-sequitur. If it were accepted that Edith and Freddy had been conspiring to kill Percy a few days before his death, that would have been sufficient to implicate Edith. It is only marginally more damaging to claim that the meeting took place a few hours before the murder, rather than a few days. The dating of the letter and of the meeting involves only a minor and immaterial adjustment in the prosecution’s version of the events. In this instance textual analysis is more helpful than contextual analysis, because the point at issue is what the passage means.

The other specific points analysed by Twining can be dealt with briefly. Weis’s account of Freddy’s movements in the twelve hours before the murder is much more detailed but, like Twining’s, casts doubt on the proposition that the attack was premeditated. Weis produces some new information about the knife:

Getting a knife under the circumstances was hardly prima facie evidence to be used to convict, although his carrying it on this night was a different matter. The dagger itself consisted of a double-edged blade, which measured five and a half inches and protruded from a four-inch long, chequered pattern handle. The weapon caused a stir when it was produced in court, not least among the jury. The police might never have found it, had Bywaters not told them precisely where to look, unaware of the damage its visual impact would inflict on his case.

Weis’s account of Freddy’s movements also supports the proposition that the knife fitted easily and inconspicuously in Freddy’s coat pocket: Freddy left home about noon (a crucial moment in the premeditation thesis), met Edith twice, completed a few errands in the city, dined with the Graydons, went out for a drink at a nearby hotel with Avis Graydon, Edith’s sister, and made several journeys by public transport. There is no evidence to suggest that anyone noticed that he was carrying a knife. All this in turn is at least consistent with Freddy’s claim that he always carried it with him.

Weis invokes the defence’s generalizations about merchant seamen carrying knives in foreign parts, but implicitly acknowledges that this does not in itself dispose of the proposition that possessing and carrying such a knife in London is in itself suspicious. It was a quite formidable instrument. Weis confirms that there is no evidence to support the suggestion that Freddy had purchased the knife recently for the specific purpose of threatening or attacking Percy. All in all this corroborates Twining’s contention that the evidence about the knife gives little support to the premeditation thesis; however, its nature and size (Weis follows Shearman in calling it a “dagger”) is damaging to Edith as well as to Freddy in the context of the premeditation argument as that bears on the conspiracy and broken chain theories.

Perhaps because he underestimates the significance of incitement, Weis never squarely addresses the crucial but complex question: Was Edith the dominant partner in respect of significant phases of the story? If she was, it is suggested, this could ground a different interpretation of some key passages in her letters from those advanced by Weis. To put the matter crudely: If one sees Edith as a manipulative, scheming or forceful woman who exerted or tried to exert influence over her younger, less experienced lover, this supports the thesis that several passages constitute incitement or solicitation to murder either by direct exhortation or by more or less subtle innuendo. This could significantly change the texture and emphasis of Weis’s story, although she might still be rescued from responsibility for murder by the broken chain argument. Before applying this to a specific example, it is necessary to consider Weis’s method of interpreting the letters.

Interpreting the letters

René Weis’s enterprise is to reconstruct and present “the true story of Edith Thompson”. He is openly sympathetic to Edith and explicitly claims to argue her case in her own terms. Criminal Justice is in an important sense Edith’s story and it is presented in a form which serves literary, humanistic and publishing values as well as the pursuit of a mundane kind of truth. It is a good read. In considering his “method” it is accordingly important to distinguish between analysis and presentation and between narrative and argument. This is especially significant in the present context in that the main values claimed for the Wigmorean approach relate to analysis rather than presentation (for example it is more useful in pre-trial preparation than in presentation at trial). Furthermore, while narrative may legitimately form part of a rational argument – stories may constitute arguments as well as be vehicles for presentation – narrative may also be used to “sneak in” irrelevant or invalid factors or gloss over weaknesses in an argument or obscure what is being argued or serve other functions such as attracting attention, sustaining interest or winning sympathy. It might be the case that while Weis’s treatment of the case is more readable and eloquent, Twining’s argument is more cogent.

In Criminal Justice Weis uses the letters as the main vehicle for presenting a coherent account of the story of Edith. Only exceptionally does he explicitly discuss alternative interpretations of these texts. If he had done so, it would have broken the flow in ways which would not only have reduced readability, but also impaired the coherence of his presentation. This is not to suggest that as a meticulous and fair-minded scholar he did not analyse and consider some competing interpretations and arguments for and against them. But it is fair to say that in dealing with the letters he generally reports his conclusions and only occasionally considers competing interpretations in detail. This opens up a whole range of issues that will not be pursued here. Let us rather illustrate some quite simple points with reference to a single passage.

Appended to a long letter, dated 1 April 1922, but possibly written over several days, is a separate note which begins: “Don’t keep this piece.” The text of the note is reprinted in full in Appendix B. It is worth reading carefully at this point. It has been widely regarded as the most damaging part of all Edith’s correspondence by the lawyers in the case, by commentators, and by several dozen intelligent law students, most of whom were sympathetic to Edith.
After summarizing the passage rather less scrupulously than usual, Weis comments as follows:

That this piece of fantasy could ever be construed as part of a premeditated murder plot defies belief. Bywaters knew it was fiction and that she had herself tasted the quinine in the tea to be able to give an accurate account of Percy’s complaint. In the court the jury was told that “the passage is full of crime”. Yes, as long as it is understood that crime means “imaginary crime”. It is never easy to separate fact from fiction in Edith Thompson’s extensive and intense correspondence, and though outside evidence is available to help distinguish one from the other, the more intimately acquainted the reader becomes with the correspondence, the more complex its rash interweaving of fact and fiction is bound to appear. In most of our lives such a blurring is not uncommon. It is not always harmless. But it is seldom the matter of life and death into which it is developed here.

This contrasts sharply with Twining’s report of his reactions to the passage. He states: “the more carefully one studies the “Marconigram” passage, the less easy is it to believe that Edith was merely fantasizing or playing games or that Freddy was not actively involved in what was being transacted at the time.” This merely reports a considered judgement reached after quite lengthy discussion in several classes. Rather than attempt an elaborate analysis of the whole passage, which prima facie contains several damaging statements, let us focus on one sentence and outline a strategy for construing it: “I wish we had not got electric light … it would be so easy.” To what might “it” refer in this context? Nearly everyone who reads this for the first time concludes, in the context of the rest of the note, that “it” means killing Percy (using gas) (interpretation 1). A strong case for this can be built up solely from the internal evidence of the passage.

Other possibilities include making Percy ill (but not killing him) (interpretation 2), but the talk of death in the previous sentence casts doubt on that. Some commentators have suggested that this – and some other passages – refer to Edith trying to abort herself (Interpretation 3). Even if these other passages support the sub-theory that Edith was pregnant by Freddy (as Weis argues) or even that she experimented with unconventional methods of contraception, these interpretations do not make sense in this context. For example, how can one reconcile them with the timing of these actions (“I’m not going to try any more until you come back” … “I’m going to try the glass again occasionally”)? And what of “if we are successful in our action” (what action?) or the need to be careful in what Freddy says to Dan? And so on.

If “it” does refer to killing Percy can this be explained away? One possibility is that Edith was pretending to try to kill Percy. If so, why would she do this except to urge Freddy to try too? Or could this all be a joke … a form of black humour or an April Fool? … or a more elaborate game of make-believe? The last is the most common version of the fantasy theory; its application seems pretty dubious in this context.

First, it seems likely from the context that Dan existed, that the “Sunday morning escapade” actually did occur, and that Freddy did supply quinine and other drugs to Edith. It stretches one’s credulity to suggest that the harmless elements are true but the damaging ones are fantasy.

Secondly, whose fantasy? If only Edith’s, how explain the passages that imply Freddy’s participation (“About the Marconigram? – do you mean … … What you said about Dan”)? If this is a joint fantasy, how does this square with our general picture of Freddy as the down-to-earth young fellow who dismisses Edith’s “vapourings” as “melodrama”? The reader is invited to construct an argument that explains away this note as evidence which supports the proposition (a) that at that time Edith and Freddy were in fact conspiring to kill Percy; or (b) Edith was pretending that she has been trying to kill Percy in order to incite Freddy to risk and try too; or (c) that Edith on this, and other occasions, was telling Freddy that she wished Percy’s death.

To return to Weis’s method. In this instance, he asserts rather than argues that the whole note is a “piece of fantasy” and that Freddy knew that. This is tantamount to a petitio principii.

Furthermore, it is not clear what precisely he is arguing when he says that “in our lives such a blurring of fact and fiction is not uncommon, but is seldom the matter of life and death that developed here”. Blurring of fact and fiction about the death of one’s husband is not so very common and the coincidence of harmless fantasizing about such a death with his actual murder by the recipient of such fantasies is unique in my experience. On reconsideration, in my judgement the “Marconigram” note does lend support to several elements in the incitement and conspiracy theories; it also raises some doubts about the fantasy theory itself, for the reasons stated. However, the timing of the note (six months before the attack) still leaves ample scope for the broken chain theory.

Many of Weis’s interpretations of Edith’s letters help to make sense of them and are much more convincing. However, he generally advances an interpretation on the basis of assertion or implicit argument without considering alternatives. To have done otherwise with the great majority of letters would probably have been unrewarding as well as tedious. However, closer textual analysis of competing interpretations of selected key passages would almost certainly have refined his argument and might also have changed it.

Fresh evidence and new perspectives

Weis has collected a mass of new data and revealing insights and has skilfully woven them together into a coherent whole which for most readers is greater than its parts. Much of the new material fleshes out the background; some has a direct bearing on the central issue. Most of the new information helps to build a quite sympathetic picture of Edith; some of it could be used against her. Some of the facts are “hard” – such as the dimensions of the knife; other data are more difficult to evaluate. What, for example, is an historian to make of the following judgement of Margery Fry, who visited Edith in prison and after¬wards became a committed abolitionist? According to Fry’s biographer, Edith struck her “as a rather foolish girl who had romanticised her sordid little love affair and genuinely thought herself innocent, discounting her own influence on her lover”. For an historian, is this evidence that bears on the question of Edith’s guilt?

The central themes of this book – the interplay of law, fact and value, of reason and imagination, of narrative coherence and atomistic analysis – have all resurfaced in this essay. What, if anything, does it suggest about the uses and limitations of Wigmorean analysis? My personal conclusion is as follows: Criminal Justice has added immeasurably to my knowledge and understanding of a case with which I already had an intimate acquaintance. It is an outstanding example of what Wigmore called the “Narrative Method”. The argument of the book in respect of the issue of Edith Thompson’s guilt would almost certainly have been sharpened and refined by the application of modified Wigmorean analysis in respect of clarification of the precise issues of law and fact, differentiating between several distinct theories of the case and the broad lines of argument that bear on them; and of selecting and subjecting to detailed microscopic analysis key phases of these arguments. Wigmorean analysis would have supplied a firmer foundation and clearer lines of argument around which to build Weis’s detailed edifice. Finally, René Weis has reconstructed a version of events which contains the most careful, cogent and coherent case for Edith’s complete innocence that has yet been made. His narrative also contains a good deal of material that could be used for several other lines of argument. For me, the effect of reading, reflecting and reacting to Weis’s account has been to strengthen my belief that some of Edith’s letters could reasonably be construed as involving acts of incitement, but that Weis’s account strengthens the case for the broken chain theory.

 

PART III. THE BIOGRAPHER’S RESPONSE TO A WIGMOREAN ANALYSIS OF ‘THOMPSON AND BYWATERS’

1. Narrative biography and methodology

Although William Twining (hereafter Twining) and René Weis (hereafter Weis) agree about Edith Thompson’s innocence to the extent that both writers believe her to have been wrongfully convicted, Twining’s analysis of much of the same evidence at first seems to differ strikingly from Weis’s, even though both writers take their cue from the same material, Edith Thompson’s correspondence.

It will be the purpose of the ensuing pages to offer some reflections on these two approaches, particularly on the methodology of a narrative account of the case and how it bears on details of the evidence. The strategies of a narrative history are bound to differ from those deployed in an analysis such as Twining’s which is embedded in a structured approach, whose intermediate conclusions are, at every turn, set out.

Different audiences respond differently to different expositions of the same material, and it will be argued here that the audience the writer has in mind may determine the differences in methodology. Weis’s study was conceived as a biographical work from the cradle to the grave. It aimed to evolve chronologically from Edith Thompson’s birth and childhood in the 1890s to her death at Holloway in 1923. Although this seems in many ways the obvious course to follow, it was in fact a decision by the author which conflicted with the wishes of the publishers, who preferred a ‘stronger’ opening in the death cell at Holloway. Such a narrative strategy would have been legitimate in view of the obiter dicta about what it is above all that renders Edith Thompson’s life a legitimate, albeit painful, subject of study. But the long-term effect of a ‘dramatic’ opening would have been the opposite of the one intended by Weis, which was to take the reader into the recesses of Edith Thompson’s inner and outer life, without at first framing her entire existence by the end in the manner of a Sophoclean tragedy. It seemed preferable to chart the mundane details of the life of a reasonably affluent London housewife, who was also an urban professional, and to pretend that nothing sinister lay in store for her rather than to distort our perspective on the case by setting the whole life in the context of violent death.

The ‘plot’ of the narrative of Criminal Justice consisted in minutely detailing the course of Edith Thompson’s life in order to enlist the reader’s empathy, to get him or her to identify with, or at least understand (while not necessarily condoning), Edith Thompson’s every move. Weis was greatly assisted in the task of charting Edith’s movements by Edith Thompson’s own practice in her letters of recreating her days in a kind of sub-documentary writing, as a way of tying Bywaters to her (or attempting to do so). Twining quotes a law student saying that ‘If you can analyse Edith Thompson’s prose you analyse anything’ , to describe the problems posed by what Twining perceives to be the ‘stream of consciousness quality’ of her letters. Coming to them from a slightly different tradition, Weis does not quite take this view, although it is certainly the case that the letters often ramble, freely associate, and seem at times to be the product of an undisciplined imagination. On the other hand, the same prose which can be described as ‘deathly’ (Twining) often sounds colourful notes that more refined and self-conscious writers in the period do not always achieve while striving to do so.

The main object of the first part of CJ was therefore to make Edith Thompson’s life happen as a life, and it is this which decided Weis to write the bulk of the book in the present tense. There were a number of reasons for this, but the clinching one was the extent to which the narrative use of the past tense forecloses reality, whereas the use of the present, in theory at least, leaves everything open: the future is never less anticipated than in a present tense narrative, whereas in a past tense story the future has already happened at the time of its telling. To that extent the biographical narrative proposes a ‘fiction’ of the case which, in its own terms, agrees with Twining’s strategic emphasis on the need to assess the evidence in the case independently from its interpretations in court.

It is of course true that the story of Edith Thompson had emphatically happened by the time CJ was written in the 1980s (and to that extent the use of the present tense might be seen to be a sleight-of-hand), but it would not have happened for an audience of the book. The intention behind the book was twofold: (i) to clear Edith Thompson’s name through putting her case to the jury of a contemporary British audience, and (ii) to convince the Home Secretary of the day that a posthumous pardon would be appropriate. The use of tense could therefore be interpreted as a legitimate ‘forensic’ strategy.

Weis’s training as a textual scholar provided some of the pasting-up and collating expertise required for documenting Edith Thompson’s life from sources. This involved the sorting of Edith’s published (and some unpublished) correspondence into a correct chronological order, reading the novels she read, and finally collating all this with further information gleaned from several newspapers (including her own preferred Daily Sketch) to form a comprehensive diary. Eventually Weis was in a position to tell almost to the hour what Edith was doing at particular times on any given day. This allowed Weis to decode much of Edith’s rhetoric (particularly as it relates to her sexuality) and to formulate the kind of ‘contextual’ hypotheses (Twining) which he deemed to be crucial for refuting a number of charges relating to ‘incitement’ that counted heavily against Edith at her trial.

2. The Letters

If the case in law against Edith Thompson was based on the evidence provided by the letters then it should have stood to reason that all the letters ought to have been submitted, that counsel for the defence ought to have insisted either (a) on their inclusion from the start, or (b) demanded that the trial judge set the record straight on the Solicitor-General’s prevarication when he falsely claimed that crucial evidence about the time and place of the actual murder were to be found in the letters.

The jury, who had access to only half the evidence appealed to, could not know that this was not the case. If anything, they may well have been suspicious of the contents of the withheld letters, since it was the defence who elected not to submit them. Whatever the finer legal points involved were in the end, the jury’s opinion was greatly influenced by the correspondence; and thirty years after the events the foreman of the jury described their contents as ‘nauseous’, and noted that ‘Mrs Thompson’s letters were her own condemnation’.

It was Weis’s firm belief that the microcosmic account of Edith Thompson’s life would assist in the search for the ‘truth’ about the death of Percy Thompson, that Edith Thompson’s innocence could be proven best through the telling of the story of her life and milieu; and through a scrutiny of her fantasies, particularly where those can be measured against texts which inspired them, in this case romantic novels from the period.

3. The fantasy theory

Twining notes that Weis’s study is largely predicated on the fantasy theory. This is a fair assessment. It need not, however, open the study to the charge of evolving in an area where therefore the writer could be thought to be selecting at random what he likes (if it assists Edith’s case) and discard what he dislikes (material prejudicial to Edith’s innocence). It is of course the case that Weis never appears to entertain the idea of Edith Thompson’s guilt, and that this certainly results in a number of short-cuts through some of the evidence that is open to hostile interpretation. Weis’s defence against this would be to maintain that the burden of proof did (and does) rest with the prosecution, and that the Crown never produced the conclusive evidence that would have convicted Edith Thompson in a fairer trial.

Twining also stresses the widely held view (by Marshall Hall among others) that if Edith had not taken the stand she would almost certainly not have been convicted of the charges brought against her, notwithstanding strong suspicions on the other counts which were not proceeded with. In view of this it does not seem entirely illogical of Weis ‘to suggest that the harmless elements [in Edith’s letters] are true but the damaging ones are fantasy’ (Twining 318). If Edith Thompson was innocent, then that would logically be the case: the potentially damaging passages could be innocently explained, and the innocent ones would be either true or untrue, but it would not greatly matter.

Furthermore, Weis believes that the narrative method can contextually prove that ‘the most damaging part of all Edith’s correspondence’, the Marconigram (Twining 317) are open to an innocent reading as long as the principle of the presumption of innocence is respected.

4. The Tea-room and the Marconigram

Before proceeding to the ‘Marconigram’ (NBT 183), it is worth turning to the tea-room passage (Exhibit 60: Twining 297-99; NBT 214-15) which Weis analyses at some length (CJ 168-72), as does Twining.

The two different approaches here coincide in their conclusion, but arrive there from very different directions. Weis was concerned to establish that the (potentially) most incriminating sentence in the letter, ‘Don’t forget what we talked in the Tea Room, I’ll still risk and try if you will’, referred to a conversation five days before the murder. The Crown suggested that it took place a mere twenty-four hours before it, with a view presumably to linking the sentence temporally, and therefore in the minds of the jury, causally, to the murder of 3 October.

Twining rightly queries the usefulness of Weis’s attempt to restore the time-span to its true chronology because the two lengths of time are not ultimately different enough to matter. Instead Twining demolishes, by a closely argued textual rather than contextual analysis, the probative value to the prosecution of Edith’s statement. The judge did not, however, have such scruples and bluntly told the jury that the ‘what’ in the sentence referred to ‘whether it was better to kill Percy by means of poison or a dagger’ (Twining 289). Here again it seems that the judge’s anxieties and Weis’s almost coincide, albeit from opposite sides of the spectrum.

It is a measure of the difference between the approaches of Weis and Twining that Edith’s question about the ‘Marconigram’ (Exhibit 17, NBT 183), starting with ‘Don’t keep this piece’, should be found so damaging by some (the Crown, the judge, Twining and also, Twining reports, students in law classes), but dismissed as fantasy by Weis (CJ 105).

Twining’s particular objection to Weis’s treatment of the passage is Weis’s refusal to entertain the idea that this passage ought to be treated differently from any other of the various passages in Edith’s correspondence which recklessly touch on the idea of murder. Weis fully accepts Twining’s stricture that a blurring of fact and fiction, when it relates specifically to the death of a spouse, ‘is not so very common and the coincidence of harmless fantasizing about such a death with his actual murder by the recipient of such fantasies is unique in my experience’ (Twining). The fact that is it unique in our experience (though probably not in life) does not by itself invalidate Weis’s point, but it stresses the challenge that the defence faced in court.

The reason why Weis passed over this passage was not so much because of ‘soft’-fact speculation, but because the hardest of all facts unambiguously seemed to him to argue against taking the passage as conclusive evidence of an attempt on Percy Thompson’s life by Edith. The ‘hardest’ fact here refers to Sir Bernard Spilsbury’s autopsy report (independently corroborated by a second Home Office pathologist), that there were no traces of poison or glass in Percy’s body. To argue, as the Crown did, that the passage is ‘full of crime’, flies in the face of the pathologists’ reports (there was no poison, no glass, and no attempt even to use gas, which is mentioned here as a possibility).

Within days of sending the ‘Marconigram’ letter, Edith wrote again and, presumably to protest her good faith, announced that she had now ‘used the light bulb three times but the third time – he found a piece – so I’ve given it up – until you come’. The ‘light bulb’ in question was the ‘electric light globe’ that she referred to in the last sentence of the ‘Marconigram’ letter.

It seems to Weis that although one may well be legitimately troubled by the business of ‘Dan’, wonder about the ‘Sunday morning escapade’, and be dismayed by Edith’s dallying with fanciful ideas of murder, these are presumably fanciful in law, unless the concept of incitement makes no allowance for a period of ‘penitence’ in the five-month gap between the letter and the murder; and ‘incitement’ would legally (presumably) be the only way the ‘Marconigram’ would come into play, since the methods of ‘killing’ themselves discussed here were scientifically proven not to have been carried into effect. It seems to Weis, as narrative biographer, that the ‘Marconigram’ is an instance where hard facts bear out ‘soft’ facts such as psychologically speculative interpretations.

That Edith Thompson was economical with the truth is undeniable. That she was capable of embellishing her untruths with shrewd authenticating touches is also true, and Weis at least suspects that the rather colourful claim by Percy in the ‘Marconigram’, to be ‘like a cat with nine lives’, may well be attributed to him by his wife.

From time to time in this essay Weis’s response to the case and the judge’s have been seen to share common ground. Although this is a state of affairs that originally rather discomfited Weis, Twining’s pointing this out helped to shape the approach of the preceding pages.

The judge in the case has received a poor press, not least for his notorious anti-feminism, which suggests that Edith suffered twice at the hands of male ideas about gender, the judge’s misogynism and the Home Secretary’s alleged feminism (Twining). More to the point perhaps may be the fact that the judge’s intellectual and legal limitations led him to simplify the facts for the lay jury too crudely, although a careful perusal of his summing up does not quite bear this out. But it is almost certainly the case that his blunt pronouncements on sordid affairs, marriage, and his clear moral disapproval carried more weight than his ultra-fair legal point about the time and place of the murder as explained above.

The jury, Weis would submit, listened to the facts, their interpretations, and the law. A combination of all three could be expected to sway them, but it would appear that in the end it was the immorality of the letters which above all convicted Edith Thompson. The defence’s repeated attempts to refocus the case away from damaging moral considerations failed. It probably had as much chance of succeeding as judges’ directions have when they instruct juries to ‘un-hear’ a particular piece of strategically planted counsel’s indiscretion. Once the damage is done it cannot be undone. It was Weis’s conviction that if the truth about Edith Thompson were to be told within the boundaries of the possible, it would inevitably become invasive and intrusive, and would have to move far beyond the legal hard evidence that was used at her trial. A deeper knowledge of the law would have been helpful. It would have resulted in a better, and a more dispassionate book.

CONCLUSION

In this essay Twining and Weis address the same question of Edith Thompson’s innocence and wrongful conviction. Their approaches differ widely. Twining rigorously, and exclusively, focusses his analysis on the evidence provided by the trial record, while Weis, using the same material, contextualises it in the interest of a biographical history.

The logical criteria by which evidence is evaluated ought to be the same in both cases. The fact that through a study of the same evidence Twining and Weis arrived independently at the same conclusion on the question of guilt, and by such different routes, would appear to bear this out. That is, it could be claimed that the Wigmorean jurist and the literary biographer demonstrate that Edith Thompson was wrongly convicted, just as two independent experts, be they pathologists or statisticians, might. Although Twining and Weis differ in some of their microscopic analyses, as for example in the readings of the ‘Marconigram’, the overall picture is one of broadly-based agreement. They draw the same conclusions from a substantial body of disparate raw material.

Even at the level of strategy where the gulf between the two authors seems to be widest, they share some common ground of history and narrative. Whereas Weis avails himself of a number of rhetorical devices that are commonly associated with story-telling such as the manipulation of tense, Twining also tells a ‘story’, as indeed did the Crown and the Defence at the original trial. The filleting of the available material by both parties at the trial, with a view to constructing coherent cases, is echoed by Twining’s focus on selected key issues, such as the knife, the ‘Marconigram’, and the meaning of the word ‘what’. The jurist’s ‘story’ is differently conceived, but its moral is the same as the Shakespearean’s.