Forensic handwriting analysis in the UK

This paper was originally delivered at the Criminalistic Institute in Prague, in 1993, by invitation.

It is a great honour to come to this famous and beautiful city and to talk to an audience of forensic scientists. The more so because there is no doubt in my mind that all of you know more about science than I do. The Department I teach in at the University of Birmingham is the department of English Literature, and what we do there most of the time is study novels and poetry and plays. My training is in that study, and only after I got the job in Birmingham in 1974 did I switch to the study of handwriting from (I hope) a scientific point of view.

I will explain to you how I came to be studying forensic handwriting analysis in the course of this paper; but I thought that the most helpful thing I could do for you was to put my small part in handwriting analysis in the context of a general description of how forensic handwriting analysis is done in the UK. In order to do this, because my first study is poetry, I will organise this talk along lines suggested by a poem by an English writer, Rudyard Kipling. He wrote:
I keep six honest serving-men
(They taught me all I knew);
Their names are What and Why and When,
And How and Where and Who.
So first I will deal with What forensic handwriting analysis is; then I will talk about Where it is done in the UK, and by Whom; next I will talk about the history of it (ie When it was done); and then How an actual case is analysed. Finally, I will talk about Why: why we do handwriting analysis, and why we do research into it.


What is the subject that we will discuss here? Well, it is handwriting as it relates to identification. In England this is the domain of Forensic Document Analysis, itself a subsection of Forensic Science.

Forensic scientists are historians: they study the traces the past has left in the present in order to find out what happened, and, more particularly, who made it happen. Handwriting is unusual because the trace of the crime, the ink line, is also the crime itself.

There are two quite distinct areas within document analysis. The first involves the use of hardware, for analysing ink, paper and so on; the second involves the visual examination of handwriting in order to establish the identity of the writer. My own expertise is not in the use of hardware, but in the other aspect of document analysis: the study of handwriting itself. What I study is the significance for identification of the handwritten lines; their shapes, and what we call in English the 'line quality': the fine detail of the line that is some indication of the speed at which it has been written. This discipline we call 'forensic handwriting analysis'.

The object of study of this discipline is handwriting; or, rather, hand-written documents. And these documents can be defined as those which are suspected of telling lies, or at least concealing the truth, about themselves. Essentially forensic handwriting analysis is based on a belief: that given enough evidence, if you examine two pieces of handwriting, you can tell whether it was written by the same person or not. In other words, that is is possible to use handwriting for purposes of identification, just as we can identify faces, or fingerprints, or teeth.

There is, however, a difference. Handwriting of a given individual can vary, or it can be made to vary, much more than normally happens to faces or fingerprints or teeth. be varied. It can vary because illness, for instance arthritis, or psychological causes, for instance stress; and it can be made to vary because the writer is disguising it, thus making it not resemble his or her normal hand, or because of forgery, where the writer not only writes abnormally, but does so in such a way as to attempt to copy some features of someone else's writing. Forensic handwriting experts would claim that, given enough material, they can distinguish between forged and genuine writings, and in many cases penetrate disguise to identify the concealed writer.

Where and Who?

So, that is what we do, and I am sure that it is obvious to you, because it is what you do too. The question now is, where is it done in the UK and who does it?

First you must understand the English legal system. I will assume that you know as little about it as I do about the Czech system; I am sure that this is not true, and if it is not, I apologise. I should also add that there is another reason why what follows is a very simplified account: ignorance. I am not a lawyer, and what I know of the system is what I have picked up here and there.

The English legal system is contestatory: it is an argument, just like the Parliamentary system of government, between different sides. When a case is contested each side takes up a position and defends it, and attacks the other side; the result is always overseen by a judge or a magistrate, who acts as a very powerful referee. The result is either decided by a jury of 12 people, taken at random, or by the judge, or judges, or magistrates. The lawyers who prepare the case are known as solicitors; the lawyers who actually argue it in court are called barristers. There are two kinds of case: criminal, and civil.

Criminal cases deal with crimes (obviously). All criminal investigation is done by the Police: they examine the crime scene, interview witnesses and suspects, gather evidence, and make an arrest. If they feel there is a need for forensic investigation of whatever kind they call in a forensic scientist, or send data (including documents) to a laboratory. The scientists make a report (I will describe these reports in detail later), which is returned to the Police. The Police add this report to the evidence, which they present to the Prosecuting Solicitors, who are paid by the state. The Prosecuting Solicitors decide whether to prosecute or not; if they decide to go ahead then they will appoint a barrister to argue the case for them. Small crimes are tried before magistrates; large crimes are tried before a judge and jury, though anyone has the right to 'go to jury' if they don't want to be tried in a magistrate's court.

Civil cases, of course, are those where no crime is committed and the police are not involved, but there is a disagreement that two sides have taken to law to have resolved. In that case two sets of private solicitors are hired; they present arguments to each other, and if they can't agree then they hire barristers to argue the case in court and a judge decides the outcome. If either side feels the need for forensic evidence then they can hire a private expert to do the work and present a report.

Because of the contestatory nature of the system, each side will hire an expert, whether the case is criminal or civil. The expert's reports will be exchanged; if they agree, then one side will either give in or, if it feels it can continue, will present its case and attempt to cast doubt on the expert's opinion by cross-examination. This case will of course be much weakened if the knowledge that their expert actually agrees with the opposition expert is known to the opposition, so quite often it is the practice for one side to hide the fact that it is employing an expert until they find out whether the expert's report is favourable or not. If the experts disagree, they are not allowed to discuss the case to find out why they disagree; this is actually illegal. They must fight it out in Court. Many people regard this as a waste of time and money, but the situation has not been changed yet, though it may be. So each expert stands up and presents his or her evidence, and is cross-examined by the opposition barrister, who usually has the help of the other expert sitting behind him and feeding him with nasty questions to ask the expert standing up. I'll go into more detail about the way in which evidence is actually presented in Court at a later point.

So you see the contestatory spirit is actually very deeply engrained in the English judicial system; even though experts are normally telling the truth as best they can, they are, if they disagree, made to fight, like gladiators. And, until about five years ago, there were two kinds of experts: those who normally acted for the prosecution, and those who acted for the defence. In the last five years this has changed, because of privatisation, so a better classification of the two kinds of experts would now be: those experts who are paid a salary by the State, irrespective of their case-work, and those who earn money on a case by case basis. I will call them salaried experts and independent experts.

These salaried experts work in three institutions.

Firstly, a laboratory in London known as the Laboratory of the Government Chemist. This is specialised, since it mainly deals with two classes of item. One is Customs and Excise, which means broadly speaking cases to do with smuggling, for instance the handwriting on a parcel that has been intercepted and found to have drugs in it, or correspondence relating to smuggling drugs. The second is Social Security matters. Social Security is our word for the money paid by the state to support those who are out of work. This money is paid by cheque through the post, and these cheques are commonly intercepted and cashed with forged signatures.

The second laboratory is the Document Section of the Metropolitan Police Forensic Science Laboratory. 'Metropolitan' means London: this laboratory, which is actually an arm of the Police force, deals with any document-related crime committed in London, and works only for the Police.

Finally, there is the Document Section of the Home Office Forensic Science Laboratory. This laboratory, which is situated in Birmingham, Britain's second largest city, which happens to be where I work too, is funded by the government, but it is not part of the Police Force. In principle its expertise is available to anyone, prosecution or defence, but in practice it is only used by the Police, and deals with all document problems in the UK that happen outside London.

Or, at least, this was true until five years ago. Five years ago there was a revolution. I'm not very clear about the details, but I understand that there was increasing discontent in the Birmingham Laboratory because of the pressures towards privatisation. They were extremely over-worked, with a three-month backlog of cases waiting to be worked on, and working conditions kept getting worse and worse. At the same time they were conscious that the independent experts were leading rather good lives: they were free of any kind of pressure from above, and they could choose to spend their time as they wanted. So there was always a steady leakage of experts from the public sector to the private sector, until finally the entire Birmingham Laboratory became so discontented that they emigrated en masse to become independent experts.

Now, previous to this, the situation was that all prosecution work was done by salaried experts, and all defence work, together with all work in civil cases, was done by independent experts. There was some point to this, because there is no quality control system for independent experts: the only test is the trial in Court. Whereas salaried experts are subject to strict quality controls. When the members of the Birmingham Lab went private the Police had no choice but to go on taking their document problems to them, and so for the first time employed private experts to do police work on a large scale.

So, let me repeat: there are no quality controls on independent experts. Anyone can announce themselves as a handwriting expert. It's a free market. They have to give their qualifications to a lawyer, and if that lawyer accepts their qualifications, then they are hired. The qualifications might be disputed in court, or they might not be: there are no agreed qualifications for handwriting experts: no degrees or vocational university courses. The control is the contestatory system: every expert opinion is checked by at least one other expert, and will moreover have to convince a judge and jury, and submit to the attempt of a clever barrister to destroy it.

Salaried document examiners, however, do have a formal training. Firstly, they must have a B.Sc. They are thus trained in--normally--the 'hard' sciences, usually physics or chemistry; several of them have Ph.D's. As a result of this they are very much at home with the hardware side of document analysis, and also have a strong basic training in scientific method, which shows in a very tough-minded and sceptical attitude. But, until enter the Document service, they know nothing about the examination of handwriting as such.

Once a candidate has been appointed to the Forensic Document service, training proceeds by a process of apprenticeship. He or she may well begin with sorting and classifying handwriting for records, with practice at using the hardware, and will be allowed to do--independently--cases that have already been examined by senior experts. The results are compared and the discrepancies discussed. He or she will accompany senior practitioners to Court and watch them in action. Eventually sufficient competence will be acquired and the trainee will be allowed to handle his or her own cases, and to give evidence in Court; to begin with, perhaps, on uncontroversial matters such as the decipherment of erasures, graduating finally to the status of 'Reporting Officer', which assumes the ability to survive a full cross-examination. The emphasis is on gradualism, constant checking, and caution, and one learns by doing, one case after another, rather than from teaching. What will happen to this apprenticeship system now that most experts who work for the Police outside London are in the Private sector, I do not know.

The independent experts, on the other hand, of whom I am one, are a mixed bunch. As I said, there is nothing to prevent anyone from advertising his or her services as a 'handwriting expert', and there are no formal courses of instruction to provide qualifications. In other forensic sciences Universities provide most of the independent expertise, but I am the only academic working in this subject in the UK (and I think in America too, though I am not sure of this). Apart from me, and from the refugees from government employment, there are two classes of independent document experts: there is a mixed bag of retired schoolteachers and bank managers who occasionally offer their services, though this is very rare nowadays, and there are the graphologists: those who believe that they can determine 'personality' from handwriting.

I should say something about the position of graphology in the UK. I can't speak from first hand knowledge about it because I know very little. The plain fact is that it is not recognized as a scientific discipline in England. In Europe I know that there are University courses in the subject, but this is not the case in the UK. Graphology is certainly not part of the training of salaried experts, and in fact they have considerable disdain for graphologists. Nonetheless, graphology is (I believe) quite widely practiced in England: it is used for job selection on quite a large scale. Many people have therefore set themselves up in private practice as graphologists, and what they mainly do is private analyses of character; but occasionally they also are employed by lawyers to offer evidence in Court. Now, because these people are often self-taught, since there are no University courses in the subject, a consequence is that their handwriting reports tend to be, well, dreadful. This has nothing to do with the graphological content: it is simply because they have no training in logical thought. I was given one to examine a few months ago: it was illiterate, for one thing, but worse than that it was impossible to determine what the woman was saying about the problem at hand. So, irrespective of the validity or otherwise of graphology, this has the effect of reinforcing the contempt in which it is held by the government trained experts.

My own view of the validity of graphology is open. I know of some very good studies that have investigated the subject sympathetically and thoroughly, and have found, to put it bluntly, that it doesn't work. I also have serious problems with the whole concept of 'personality', which I believe to be a cultural construct. And I know of no convincing studies that have shown that graphology works. So, though I do claim to have an open mind, perhaps it would be better to say that my mind is not entirely closed on the subject. I do know of at least one graphologist whose judgements quite reliably please and astonish those whose writings she analyses. Why this is, I cannot imagine.

Finally, there is me; or rather us. I was originally trained in the study of English Literature Later, I specialised in what is known as textual criticism: that is, the comparison of different versions of literary texts in order to find out what the author actually wrote in the first place. This may seem a little trivial, so it is worth pointing out that in the case of Shakespeare's Hamlet there are two main printed editions, of 1602 and 1623, and there are over 1500 differences between these two editions. No-one knows, in the case of each of these differences, whether Shakespeare wrote one, or the other, or both, or neither. Textual critics have a lot of work to do.

My training at Oxford, and later at Yale, included deciphering and identifying the handwriting of various authors. I was appointed in 1973 to teach a three-year undergraduate option in Birmingham English Department on textual criticism and related topics, which included handwriting: identifying the handwriting of authors. In 1974 I discovered that my training in literary handwriting identification qualified me for forensic work too, and since then I have practiced as an independent forensic handwriting expert. It used to be that I would average about three completed cases a month, and appeared in court much less frequently than that. All of my work is for the Defence, or in civil cases. I have found that my education in the study of literature did have one advantage: most of literary studies is about argument about value and meaning of poems and novels, and so I had a long training in unfair and ruthless techniques of argument. This has been a great help in the contestatory English judicial system.

I say that I used to do about three cases a month; this has now changed to five. The reason for this is quite a long story, but, briefly, what happened was that I was very involved in 1987 onwards with the investigation of Police corruption. There was an elite detective force in Birmingham known as the West Midlands Serious Crime Squad. I was given a case late in 1987 in which I was able to give evidence that suggested that they had forged a confession, through the use of the ESDA machine. This started a very big ball rolling; it turned out that it could be shown that a number of confessions elicited by this Squad were dubious, including, most famously, the confessions of the Birmingham 6: Irishmen who served many years in prison for allegedly murdering a large number of young people with two bombs. My involvement in all of this meant that I was on TV quite a lot, and my work load increased accordingly.

Each student in the course I teach does a substantial dissertation as part of it; a number of these have made significant contributions to research in handwriting and some students have stayed to do further research, as graduate students and Research Associates. We have, in particular, now undertaken four substantial research projects, funded by the Home Office, and in close collaboration with handwriting experts at the Birmingham Forensic Science Laboratory.


How is document analysis actually done in the UK?

First, I will say what kind of cases are part of the case load of a forensic handwriting expert. Or, rather, I will say what my cases are like. I am unusual, in that I only do this work on a part time basis, and only work in civil cases or for the defence; and I only undertake the examination of handwriting or ESDA analysis. If any physical examination of documents is involved, then I pass the case on. In the last twelve months I find I have completed and sent in reports on 72 cases.

my case load

Of these, almost half are disputed signatures on various kinds of documents: wills, sale transactions, house purchase documents, a passport application, and various legal documents: 34 cases in all.

The next most numerous class (12 cases) concern forged cheques. Here what happens is that a cheque book will have been stolen together with a cheque guarantee card. The forger will buy goods up to the limit of the card (between 50 and 250), write out the text of the cheque, and forge a signature. If he or (more usually) she is caught then often the police in the cheque squad can tie the actual cheque uttered on the occasion when the suspect is caught with in with other cheques that have been sent to them as forgeries; they make a preliminary examination of the handwriting of the texts of the cheques, and if they look similar then they send the whole batch to a document examiner for a report. If his or her report is positive, then eventually the defence will call on another examiner (in this case, me) to check it.

I have done 8 cases in the last 12 months involving disputed confessions. This arises from my work with the West Midlands Serious Crime Squad, and it concerns people who have been imprisoned on the basis of confession evidence who are saying that their confessions were fabricated. It involves applying the ESDA test to the confessions, and making some very complicated deductions from the results. Handwriting identification may also be involved, if the defendants are saying that the police forged their signature on the confession or on various documents concerning the arrest.

The next most common case (6 in all) is peculiar. In England if you are suspected of being drunk in charge of a car you can be stopped, and your breath is tested on a machine. The machine prints out a piece of paper with the results, and if it is positive, you sign the sheet of paper, and are taken to the police station and tested again. Another piece of paper is signed. If the results are still positive you are charged, and sign what is known as the custody record in various places. What happens rather commonly is that someone will deny that he is the person who was arrested, and allege that someone else pretended to be him and forged his signature. Of the six cases, incidentally, five were not forged and one (in my opinion) probably was.

I did four cases that involved denial of authorship of an incriminating document: a letter, or an entry in a diary. These are text writing, which is harder to forge and of course easier to identify, in general.

I did three cases involving anonymous letters, 2 concerning alleged insertion of numerals or letters, one concerning writing on a blackboard, one concerning the doctor's notes in a rape case (the defence thought they may have been fabricated), and one kerbcrawler's questionnaire. What is a kerbcrawler's questionnaire? This is my case study, so I will explain this in great detail later.

So: that is the kind of cases I get; now, what actually happens in the course of a handwriting analysis? I will describe first the progress of a typical criminal case, let us say an allegedly forged signature on a legal document.

The police, acting upon a complaint or as part of a larger investigation, will become aware of the allegation that the signature in question is a forgery. If they think it might be, they will gather the necessary evidence, in the form of comparison samples of signatures, and send them to a forensic laboratory. Part of police training is the awareness of what kind of samples document experts require, and sometimes, particularly in the case of cheques, they have special pre-printed cheque forms to take samples on. This batch of documents arrives in the lab, together with a short description made by the officer in charge of the case of the facts of the case. When the expert comes to examine the case, this is what he does: I am sure that something of the kind is what you do too. Please forgive me if this is all completely obvious.

The first thing that one needs is the original questioned document. Photographs and photocopies are sometimes sufficient, but one can't really know that unless one has seen the original: some of the fine evidence for forgery, for instance, is not reproduced by photocopying. Next, one needs adequate samples. These should be of the same kind as the questioned writing: one can normally only compare capitals with capitals, and moreover /A/s with /A/s; capitals can't be compared with lower case writing, or /A/s with /B/s. The other important characteristic of the samples is that they should be treated with great suspicion. The possibility of disguise is always there. To counteract this, it is necessary whenever possible to obtain two kinds of samples: 'request samples' which are produced for the examination, and, in the case of text-writing, conveniently duplicate the questioned material (from dictation); and naturally occurring samples, which were made without the knowledge that they would be examined. The samples should be as nearly contemporary with the questioned document as possible, and sufficient in number to demonstrate the normal range of the individual's handwriting; the amount of variability varies, of course, from one individual to another.

The process of comparison, briefly, involves the close examination of each letter in the questioned and the sample writings, and then a detailed comparison of each corresponding letter. One is looking for significant matches or mis-matches, on the level first of the form of the letter, and then of the fine detail of letter-formation. Relative proportions of letters, the links or spaces between each letter, the relative height of each letter above the line, and the quality of the pen line itself--is it smooth, tremulous, jagged, confident, does it pause, are there odd pen-lifts, and so on. This sounds rather subjective, and it must be admitted that the judgements of handwriting experts are largely based on experience, and that experience has not, on the whole, been quantified and calculated; nor can it be wholly described in words. Nonetheless no-one will take an expert's opinion for granted. The people they have to convince are, firstly, fellow-scientists: in a criminal case the handwriting evidence is almost always examined by three different experts; secondly, they have to convince juries and judges, who will not necessarily take their word for it.

Before the handwriting examiner gives a positive statement that two handwritings were written by the same person, he or she would hope to find what is taken to be a definite match between every significant element in the handwriting. Anything less than this and the report will be qualified. (I will talk about report writing a little later).

So what the examiner actually does is this. He or she first examines the sample signatures, and in a lab notebook makes careful drawings, with notes, of the features of the genuine signature, until he has what amounts to a complete description, together with illustrations, as far as the evidence permits, of the range of variation of all of the letters in the sample signature. This lab notebook is a rather private document, though it is always checked by another examiner; however sometimes the defence demands to see it, and they then have to give up a photocopy.

The examiner then comes to a conclusion. The conclusion goes into a report, which goes back to the police together with the documents. All that this report contains is the bare minimum: a list of documents, a statement of what the examiner has done (for instance, I have compared the questioned signature on document x with the sample signatures on documents a, b, and c') and a conclusion.

The form of this conclusion varies depending on the laboratory giving the opinion. In the London Police laboratory they have a standard form, which was described by David Ellen, the then director, in Ellen, D M, "Expression Of Conclusions In Handwriting Examinations," Canadian Society Of Forensic Science Journal 12 (1979): 117-20. This is a very thorough and interesting article, and I will only summarise it briefly here.

Before I describe this, I should first describe the standard of proof required in English Courts. The way in which it is put is this: in a criminal case guilt must be established beyond reasonable doubt. Whereas in civil cases, the requirement is that a case should be proved on balance of probability. In the latter case, it means it is more likely than not that (say) two handwritings are in the opinion of the expert written by the same person. Even if it is only marginally more likely, even 50.001% more probable, the opinion is still valid. Whereas in the criminal case, what the expert for the prosecution is looking for is a very considerable weight of proof, whereas what the expert for the defence seeks is any reasonable possibility for doubt. One result of this is that the statements of prosecution experts are very cautious and conservative, and so they should be, in my opinion. It is actually rather hard to get a usable report from a Home Office document expert, and the police frequently make a joke out of the caution of handwriting experts. On the other hand the police know, once they have a positive handwriting report, that it is what they call 'good evidence': it will stand up in court, and they are as happy to have it as to have any other kind of solid forensic evidence. Handwriting evidence in the UK is respected, on the whole.

Ellen's five categories are as follows:

1. 'In my opinion this was written by the writer of that'. 'This indicates a positive finding, enough to indicate to a court that beyond all reasonable doubt the writer of the questioned document is identified.'

2. 'In my opinion there is a high probability that this was written by the writer of that'. If there is no other material evidence there is insufficient reason to convict on the basis of this opinion alone, but it would only require a small amount of additional evidence to obtain this conviction'.

3. The next category is where there is some evidence to connect two handwritings, but the evidence is not strong. No particular formula is given for this, but it is suggested that the words 'consistent with' or 'could have been' may be used, presumably as in 'The evidence is consistent with the two writings being in the same hand'.

4. 'The evidence is totally inconclusive', either because of lack of evidence or because of disguise or other disruptive factors.

5. There is no evidence to suggest that the two writings were made by the same person, and there is some evidence to suggest that they were made by different people.

What is interesting from my point of view, as an expert who works for the defence, is that there are four categories of guilt, as it were, but only one of innocence: the categories were very clearly designed not from a neutral point of view, but from the point of view of someone who is looking for evidence to help the police. I am not disagreeing with this; my point is that the scientist is not neutral. He or she has an aim, to find evidence for guilt or evidence for innocence. It is a contestatory system, and even the scientific expert is affected by this. So if I operated a set of categories, which I don't, then they would be the mirror image of this set, with four degrees of innocence and one of guilt.

Document Examiners trained in other laboratories don't operate precisely this set of fixed categories, although the end result is much the same. So for instance in a recent report from the Birmingham lab, it says 'The questioned handwriting corresponds closely to the specimens of COBBE'--the defendant--'and in my opinion she wrote it, the evidence being such that the possibility of some other person being the author can be excluded'. A less positive report might say something like 'I cannot match all of the letter forms, and therefore cannot exclude the possibility that the writing in question was written by someone else, but I consider that possibility to be unlikely' or 'I consider that possibility to be so remote that it can be excluded'.

My own reports are much less formulaic, since I am looking for grounds for doubt, and I am also also accustomed to work not only with criminal but also civil matters. So the last report I sent out, which was on a questioned signature on a will, went as follows:

It is not easy to come to a definite opinion in this matter because although there is a considerable amount of comparison writing, this only includes two sample admitted signatures, of which one is a carbon copy.

Nonetheless I can say that I can see no evidence of forgery in the questioned signature, and a number of characteristics that it has in common with the sample signatures. I would say that, on the basis of the evidence that I have, in my opinion the signature in question was on balance of probability written by Mr X.

The evidence is then checked by another scientist; double checking is standard practice in all prosecution reports. In the Police laboratories, and in what used to be the state-funded Birmingham laboratory, another level of checking operates: the Home Office Research Centre at Aldermaston would prepare dummy cases and send them in for analysis via the Police. None of the experts involved would know that this was happening, and the resulting opinions would be evaluated in Aldermaston. This is a very efficient method of quality control, which is unfortunately not possible for private document analysts.

Finally, if the opinion is positive, photographs are taken, with a view to preparation of a chart for presentation in court. The positive report is then disclosed to the defence, and another expert will be called in by the Defence to give an opinion on all of the evidence seen by the prosecution expert. He or she will follow essentially the same procedure: make notes on the writing in question in a notebook, issue a bare minimum report, and take photographs if it is felt that the evidence will go to court. The job of the defence expert is in some ways easier, since he or she has access to the report of the prosecution expert, and only has to agree or find adequate grounds for disagreement.

If the case actually comes to Court, the procedure is as follows. You are probably familiar with the system: in criminal trials, briefly, the prosecuting barrister introduces the prosecution case, and then the defence introduces his or her own case. Witnesses to support the prosecution, who are cross-examined by the defence; the defence then produces witness to support the other side, who are in turn cross-examined by the prosecution. Both barristers summarise their cases, the judge summarises the whole thing, and the jury retires to consider then returns a verdict.

The forensic expert is treated like any other expert, with two exceptions. One is that witnesses must wait outside the court to be called to give evidence. This rule is normally, but not always, waived in the case of experts, who are allowed to sit in on the trial. This is very helpful, I find: one gets used to the proceedings and a sense of what the barristers are like; also sometimes one can assist the barrister in cross-examination. The other difference is that the expert witness is normally treated with a certain respect: it is accepted that he or she is telling the truth to the best of their ability, and a normal practice in cross-examination, which is to behave as if the witness is a liar, is not done. It is in fact very difficult to cross-examine experts, particularly if they have any experience, since the golden rule of cross-examination is not to ask a question to which you do not know the answer, and an expert will naturally know far more about the subject than a barrister, making it quite easy on occasion to trip the barrister up.

The handwriting evidence is usually (though not always) presented with the aid of a chart. This is a mounted set of blown-up photographs illustrating the aspects of the handwriting under discussion. It is the convention not to take individual letters in isolation, but to show them in the context of the word to which they belong, so that relative heights, relative spacing, and height above the base-line can be considered. Sometimes also the expert will illustrate the evidence with drawings, done on the spot on a clipboard. Here are some charts that I have made. Normally charts are done photographically, and pasted and mounted by hand. These were done with a computer. There is a loss of detail, but there are compensations for that.

Illustration 8, 9: Charts

Confrontation by handwriting experts is very rare. Normally, their reports coincide, so only one of them ends up giving evidence. When they do differ, it may be that when they see each other's reports one will back down. Or it is clear that one side has access to more evidence than the other, and the two opinions can be reconciled. Or else the difference between them is a difference of degree, rather than a diametrically opposed opinion. This last has never happened to me in 20 years, though I have had experience of all the other categories.

Finally, in this section, I will gave a case history. I had a choice when writing this, either to present a normal and representative case, which would be illustrative of how we go about things in the UK, but unfortunately boring, or an unusual case, which would be interesting, but unfortunately unrepresentative. I chose to be interesting (I hope) rather than boring.

The case of Mr Smith

As usual, there are two stories: the Police story and the Defendant's story. The Police story is as follows. Smith was seen, they say, by two plain clothes policemen driving through a known red light district. They decided to follow him in their car because they thought he was looking for a prostitute. He parked, and one of the officers got out to follow him. This officer says he saw him approach a woman who was known as a prostitute and talk to her, after which the two walked off together. At this point the Police Officer interrupted. Mr Smith was asked to come to the Police car and fill in what is known as a 'Kerb Crawler's Questionnaire'.

A kerb crawler is someone who drives very slowly looking for a prostitute, incidentally. I think the British custom of asking him to fill in a questionnaire is designed to embarrass and therefore discourage kerb-crawlers. While the form was being filled in the officers said they noticed alcohol on Mr Smith's breath, and so they arrested him on suspicion of being unfit to drive because of drink, took him to the station, and breathalysed him: the results were positive. Eventually he was tried before magistrates and found guilty of driving with excess alcohol.

Mr Smith tells a completely different story. He admits that earlier in the evening he had drunk two cans of lager beer. He says that he was driving home, and felt unwell. He is a diabetic, and he recognized the symptoms of an insulin reaction. He needed to take in some carbohydrate, and, though he normally carries sweets for this purpose, he didn't have any with him on this occasion. However, he did have a large bottle of cider. So, he stopped the car and drank a lot of cider, which carried him through the insulin reaction but made him drunk. He decided not to drive home: he got out of the car and looked for a phone box to phone his mother and tell her he was going to sleep it off in the car. Unfortunately, since he had parked quite by chance in a notorious red light district, he was approached by a prostitute. She propositioned him, and he refused, but at this point a policeman arrived, and after this his story and the police story are more or less identical.

Apart from one thing. Smith says that when they were in the police car the police officer did have a piece of paper, presumably the Kerb Crawler's Questionnaire, but he was so irritated by Mr Smith's vehement denials that he had been looking for a prostitute that he screwed it up in disgust and threw it on the floor.

As I said, Mr Smith was convicted of driving under the influence of alcohol, in 1992. He appealed against the conviction, and his appeal was listed for hearing in early 1993. Immediately before the hearing, the police produced a document, which had not been used in evidence at any point before: the Kerb-Crawler's Questionnaire. Mr Smith immediately denied the authenticity of this document, saying, presumably, that the police had concocted the document after the event in order to support their story; and the case was adjourned so that the document could be examined to test its authenticity.

The whole case depended on whether I thought the signature on the document was authentic or not.

Here is a chart I made of the signatures in this case.

Illustration 11: Smith chart

The questioned initials from the Kerb-Crawler's Questionnaire are at the top. Underneath are three examples of Mr Smith's initials obtained at my request by his solicitor, and underneath that two samples of his signature, similarly obtained. They certainly differ from the questioned signature: there is no doubt about that. But perhaps the difference is the result of deliberate disguise? In any case, when the questioned signature was made (if he made it) Mr Smith would have been in a peculiar state: he was presumably stressed, certainly drunk, and possibly suffering from insulin shock. Fortunately some excellent sample signatures were at hand, in the form of the intoximeter printout, which has to be signed twice, and the Custody Record, which has to be signed several times. I asked for these, and have put a sample of these signatures on the chart (signatures 7, 8, and 9).

Now, what can I say about this material? Well, I can offer an opinion. The opinion would be that the questioned and the sample writing are not in the same hand. I would offer as support for this view the following: that there are no letter-shapes in common between the questioned and sample writing, and that the line quality differs.

Here an oddity is that the line quality of the questioned signature is actually superior to the sample signatures; superior to the request samples, 2 to 6, which were presumably produced under relaxed circumstances, but also superior to the signatures done at the time, when Mr Smith may have been suffering from all sorts of influences that might have affected his handwriting, including fear, anxiety, anger, alcohol, and possibly insulin shock. Now, in my experience these things have all sorts of different effect on handwriting, but one thing they are unlikely to do is to improve the line quality. I would also say that when letter-forms degenerate due to the effect of stress or disability they still retain some of their formation; they do not disappear completely to be replaced by something entirely different. As we can see, in fact, on the chart: there are resemblances between the request and the contemporaneous signatures, however illegible, as in the case of signature 9, the latter may be.

However, there is not in the case of the questioned writing a lot of material to go on: just perhaps a couple of initials. So what I would say is that this evidence is such as to give rise to reasonable doubt as to the authenticity of the writing in question.

I took this evidence to Court and gave it. I sustained the cross-examination with my testimony intact. There was no opposing expert opinion. Mr Smith lost his appeal.


Why do I study handwriting? Because it is interesting, and there is a great deal of work to do, and because it is important. People regularly go to prison on the basis of handwriting analysis, so it is necessary to know as much about it as possible. This seems obvious. It is not universally accepted, however. In 1982 we were given a research grant of about £10,000 to attempt to produce a statistically based description of the characteristics of left handed handwriting. Which we duly did. However, I incautiously issued, under the promptings of the University Press Office, a modest press-release describing the project. This caused a remarkable uproar. Some of the less scrupulous newspapers (the worst offender was The Daily Telegraph) took up our project as a clear example of totally useless, utterly academic, academic research. The Home Office was besieged with letters. An MP was narrowly prevented from asking questions in the House. I was amazed.

In answer I can offer another anecdote. Some years ago, a man whom the newspapers called the Black Panther kidnapped a young girl. He sent out ransom notes. There was an enormous man-hunt, costing over a million hours of police time. He murdered her, eventually: she was found hanged, on a steel rope, in a sewer. When the murderer was finally caught, he was found to be left-handed. Had this been deduced from the handwriting of the notes, the number of possible suspects would have been cut by 90%. But this could not happen, because the Home Office had not yet given Birmingham University a grant to do the necessary research. The need for handwriting experts, and the need for research in handwriting, is obvious.

Tom Davis
University of Birmingham