Standards of Proof

by Christophe Champod

Introduction

Papillary surfaces are covered in fine lines, or ‘ridges’ arranged generally in patterns such as loops, whorls and arches. The ridges form characteristics called minutiae such as bifurcations, ridge endings (or combined minutiae) and also microscopic features (pores, ridge edges and structures). These detailed features are the consequence of a morphogenesis so sensitive to outside influence that they are unpredictable, unlike general patterns, in their position and their form. The variability is such that even monozygotic twins have different fingerprints in this respect.
Following a comparison between a mark (recovered for example in association with a crime) and a control print (associated with a suspect), suppose that the examiner observes a complete agreement between the mark and the print without any significant discrepancy. The question is then often expressed as follows: ‘how many similarities are required to confirm the identification?’ or ‘what is the evidential value of a partial print with such similarities?’
The aim of this article is to try to answer these questions, to address the question of the ‘standard of proof’, by reviewing international views and practices.

Historical Milestone

The first rules establishing the minimum number of minutiae necessary for fingerprint identification can be attributed to Edmond Locard, in 1911-1912. He suggested a tripartite rule, which followed from the discovery of poroscopy. This can be summarized as follows.
1. If more than 12 concurring points are present and the fingerprint is sharp, then the certainty of identity is beyond debate. (The imperative requirement for the absence of significant differences is implicit.)
2. If there are 8-12 concurring points, then the case is borderline and the certainty of identity will depend on:
(a) the sharpness of the fingerprint;
(b) the rarity of its type;
(c) the presence of the center of the figure (core) and the triangle (delta) in the exploitable part of the print;
(d) the presence of pores;
(e) the perfect and obvious identity regarding the width of the papillary ridges and valleys, the direction of the lines, and the angular value of the bifurcations.
In these instances, a certainty can only be established after discussion of the case by at least two competent and experienced specialists. 3. If a limited number of characteristic points are present, the fingerprint cannot provide certainty for an identification, but only a presumption proportional to the number of points available and their clarity.
Locard based his tripartite rule on various sources of information: the discovery of poroscopy, the practice gathered by the identification bureaux around the world and statistical evaluation.
This approach persists throughout the extensive writings of Locard. His considerations (principally the first two) were largely taken up by the most eminent dactyloscopists or criminalists of the first half of this century.

Current Views and Practice

On the fringe of the theoretical view expressed above, the practice has led to various positions. The different positions that currently exist were not established without controversy after the views expressed by Locard.

Predetermined minimum number of minutiae

The majority of European dactyloscopists have favored a purely quantitative approach, leaving qualitative aspects in the background, by fixing a numerical standard – a minimum number of minutiae necessary to establish identification. The numerical standard represents a lower limit; above this value, the identification is beyond doubt regardless of the type of minutiae that are found. The interpretation of the concept of numerical standard may vary from agency to agency; some positions are summarized in Table 1.
Despite the systematic use of a numerical standard, various countries (e.g. Finland, Greece, Holland, Israel and Portugal) have developed methods to bypass the rigid number when certain characters (visibility of pores, ridge structures or rare combinations of minutiae) are observed in the comparison. The adoption of a range from 8 (or 10) to 12 points is a way to relax a rigid threshold. In 1983, an addendum was made to the ’16 point standard’ in the UK, which stated that in extremely rare cases, an expert with long experience and high standing in the profession can make an identification that does not meet the nationally accepted standard.

No predetermined numerical standard

Before the 1970s, American fingerprint examiners followed Locard’s view. As a result, the 12-point rule was generally respected and below this threshold qualitative factors in the comparison were taken into consideration. In 1970 a commission of experts from the International Association for Identification (IAI) was established to study the question of the pertinence of a fixed numerical standard for dactyloscopy. The work of this committee (for ‘standardization’) resulted in a number of articles and reports tracing the state of empirical and scientific knowledge.

Table 1 Principal positions regarding numerical standards adopted in some countries

Apart from the fact that the justification of any number could not be based on statistical researches pertaining to fingerprints, the most powerful argument against any numerical standard derives from knowledge of the morphogenesis of the papillary lines. The development of papillary lines on corresponding surfaces is initiated (for the fingers) after 7 weeks of fetal life with the growth of volar pads on a papillary area. After 10 weeks, as the volar pads regress, primary ridges begin to grow in the dermis followed by secondary ridges. The various stresses involved in this process (regression of the volar pads, development of size, meeting of multiple development fronts) induce a stochastic formation of minutiae (in terms of form and positioning). Around week 25, the development of papillary lines is completed on the dermis and is projected onto the epidermis of the skin. The fact that the complete pattern is secured on the dermis explains the durability of the fingerprints. From that moment, a final differentiations occurs on the papillae which dictates the form of the pores and ridge edges. This final stage also produces a strictly individual pattern.
Hence, it was clearly not legitimate to reduce the fingerprint individuality to the minutiae as is suggested with any numerical standards. The number of specific features is much broader than minutiae alone. The nature of the papillary individuality prevents the adoption of any predefined number of ridge characteristics necessary (without significant differences) for identification.
Following such arguments, the following resolution was adopted by the International Association for Identification in 1973: ‘The International Association for Identification, based upon a 3-year study by its Standardisation Committee, hereby state that no valid basis exists for requiring a predetermined minimum number of friction ridge characteristics that must be present in two impressions in order to establish positive identification.
This resolution is also endorsed by practical observations.
• Some minutiae are rarer than others (for example, a double bifurcation in the periphery of the pattern is six times less frequent than two separate bifurcations). Hence, counting minutiae up to a predefined number without taking into account their relative and distinct values is not adequate.
• The absence of any minutiae; from a papillary surface is as relevant as their presence. It would be very exceptional to observe a core area or a delta area where the ridges do not display any minutiae.
• The pore structure and ridge edge structures are individual and, when visible, contribute to the identification decision. Numerous cases of identifications based on poroscopy and edgeoscopy (i.e. ridgeology) are documented in the scientific literature.
It has been accepted that the concept of identification can not be reduced to counting fingerprint minutiae; each identification represents a unique set of circumstances and the identification value of concurring points between two fingerprints depends on a variety of conditions that automatically excludes any minimum standard. When a fingerprint expert concludes an identification, he reaches a decision threshold. This threshold may be a number of concurring points or, in addition to this quantitative element, consideration of qualitative factors such as the rarity of the general pattern, the type of points observed, and the relative frequencies of these points. In this way, the identification process is a global assessment which balances both quantitative and qualitative aspects. Following this resolution, different minima were instituted at the level of the training of American dactyloscopists. A technical working group driven by the FBI on friction ridge analysis (SWGFAST) has reviewed these guidelines.
In 1995, a conference meeting on fingerprint detection techniques and identification at Ne’Urim (Israel), unanimously approved a resolution, a slight variation of the IAI 1973 resolution. The Ne’Urim declaration states that: ‘No scientific basis exists for requiring that a predetermined minimum number of friction ridge features must be present in two impressions in order to establish a positive identification.’ Various countries are or will follow the IAI as indicated in Table 2.
It must be noted that the work done by the English committee has been a determinant in the various moves towards the adoption of the IAI resolution. In 1989, a collaborative study among 130 fingerprint examiners in England and Wales was conducted. Each participant was sent 10 mark/print comparisons (9 associations and one exclusion) and was asked to mark the number of concordant minutiae and to express an opinion with respect to the identification. No misidentification was reported. But, when examining genuine pairs of prints, examiners varied widely in the number of points of comparison they found. In the case of one print it varied from 11 to 40. In other words, some examiners would not have gone to court with this pair of impressions although most would have done so. An identical survey in Switzerland in 1997 led to analogous results. These studies led to the following conclusions.

Table 2 Main positions regarding the abandonment of a numerical standard

• The dogma adopted by some fingerprint examiners that fingerprint identification is an ‘exact’ science is a misconception. The essence of science is inductive inference. Inference is a mental process that cannot be exact (or deductive). Fingerprint identification is scientific in that sense.
• The precision implied by any number (12, 16, etc.) is also a lure. The determination of individual minutiae or features is highly subjective.
• As a means of achieving quality, a numerical standard is poor. The way forward is to concentrate on professional standards rather than on rules about numbers of minutiae. A scheme of quality management should be instituted including training, certification testing, performance testing, file audits and blind trials.

Range of possible conclusions in the fingerprint field

Most contemporary fingerprint experts refuse to give qualified opinions on fingerprint comparisons that do not meet the minimum requirements for an identification. From this standpoint, it is not authorized to speak of ‘possible’ or ‘probable’ identifications, for example.
The idea that dactyloscopy could provide a presumptive link preceded Locard, notably in the casework dating back to the beginning of the twentieth century by Reiss at the Institut de Police Scientifique et de Criminologie (IPSC) in Lausanne, Switzerland. At present, very few identification bureaux in Switzerland leave this possibility open. However according to a recent European survey, few other countries are using fingerprint evidence as corroborative evidence. For example in Belgium where there are between 8 and 12 points of agreement between a questioned fingermark and a fingerprint (without discrepancies), it is stated that ‘the fingerprint could be the same as the fingermark being examined’.
Therefore, apart from a few exceptions, Locard’s third directive was totally forgotten as the use of dactyloscopy became widespread, thus giving dactyloscopy its very distinctive character compared to other forms of transfer evidence. In practice nowadays, a fingerprint is proof of identification or exclusion or it has no evidential value. In most countries, the probabilistic aspect has been excluded from dactyloscopy, removing the possibility for a fingermark to constitute corroborative evidence for an identification.
Steinwender, a German fingerprint expert, was the first to publish his refusal to consider dactyloscopic evidence as anything but absolute proof. This point of view has largely been followed by the whole profession. This is sometimes described as the ‘positivity’ of the fingerprint field.
In 1979, during the annual IAI conference, a resolution was approved prohibiting members from giving testimony on qualified identifications (likely, possible, probable, etc.). An identification, according to the IAI, can only be negative, impossible or certain. This resolution was widely debated before being revised and accepted in 1980. The IAI experts then rejected the idea of using dactyloscopic proof as corroborative evidence in the following terms (Resolution VII Amended or Resolution V): ‘[The delegates of the IAI] state unanimously that friction ridge identifications are positive, and officially oppose any testimony or reporting of possible, probable or likely friction ridge identifications found on hands and feet, [...]‘
This approach means that in countries using numerical standards, for instance Great Britain, a fingerprint comparison with 16 concordant minutiae is an identification, whereas a comparison presenting only 15 points is not probative evidence and should not be presented in court. Such a rule can easily result in difficult situations when, for example, a suspect arrested and identified in one country on the basis of an identification with 12 points is to be judged in a neighbouring country where 16 points are explicitly required.
The present author considers that the resolution voted by the IAI in 1980 prohibiting this form of testimony is opposed to the scientific principles governing the interpretation of transfer traces in criminalistics. It is erroneous to consider that dactyloscopic evidence is only a dichotomic form of proof, uniquely applicable to identifications or nonidentifi-cations. The interpretation cannot be so drastic, so clear-cut; there is an increasing scale from exclusion to identification. In fact, there is no logical reason to suppress gray levels between white (exclusion) and black (identification). Each piece of evidence is relevant if it tends to make the matter at issue more or less probable than otherwise. Dactyloscopic evidence can be combined with other evidence (forensic or not) and fit into the particular context of each case. The refusal of qualified opinions in dactyloscopy is a policy decision (even if the distinction of the arguments (policy or scientific argumentation) is not so clear in the literature).
Often, probabilities have been excluded from dactyloscopy arguing that each portion of a fingerprint, no matter how small, can only have come from one finger. The individuality of a particular papillary arrangement, and even the individuality of the process and the contours of the ridges, is without question. However, the transfer of this information necessarily implies a loss of detail – a loss of information – that reduces this individuality with respect to the trace evidence itself. It is the individuality of the papillary impression that is at question and not the individuality of the skin surface that produced it. The transfer of material in criminalistics, as defined by Locard, logically implies, by definition, a loss of information. In dactyloscopy, this loss may be of two types: quantitative (due to the limited size of the trace) and qualitative (blurring, bad definition, loss of pore details, etc.). The concept encapsulates a continuum of values in that a feature may have an identifying value ranging from very low to very high. Locard’s third directive is fundamental and permits dactyloscopic proof to be on the same level as other types of transfer traces. Some will no doubt argue that judges expect dactyloscopy to be univocal and without compromise. This idea is false. It is obvious that any judge will prefer indisputable forms of proof, but would also wish to be informed of any evidence that can lead closer to the truth. Fingermarks have already been considered by jurors from a probabilistic point of view. By their admission, these traces may be classed with all other forms of evidential material in the judicial system. This issue has been addressed by New Zealand when the Court of Appeal not only ruled that there are no numerical standards for an identification but has also stated that:
• much useful evidence was being kept out of the courts on decision of technicians;
• fingerprint evidence should (of course) be combined with other evidence to determine whether or not the accused was present.
The absence of extensive statistical data on fingerprint individuality can be viewed as the main reason to prevent giving qualified opinions. Statistical data could make dactyloscopists feel more comfortable in this area and make statements based on sound data obtained from significant samples.
Research results only describe one factor contributing to fingerprint individuality, which is the probability of the configuration of minutiae (basic forms and combinations) on a surface P(C). A computer program has been developed to enable the acquisition of data on an extensive sample of nearly a thousand fingerprints selected from more than 100 000 dactyloscopic forms. Nine types of minutiae (ridge ending, bifurcation, island, double bifurcation, hook, lake, opposed bifurcation, etc.) were considered, taking into account their orientation and length when defined.
A model giving an upper bound estimate P(C*) of the probability P(C) has been proposed and successfully tested. To illustrate the impact of these results, let us consider two different configurations of the same size taken from the periphery region of two ulnar loops (Fig. 1).
The probabilities P(C2*) and P(C2*) have been calculated and are as follows:

These probability figures strongly suggest the possibility of evaluating dactyloscopic evidence in a probabilistic way. When compound minutiae are observed on a surface, the calculation possibilities are extended and lead to probabilities which are of impressive value as evidence. It is possible to obtain, with some specific minutiae on a surface (or in the absence of minutiae on a large papillary surface), probability figures which exceed the value proposed in other forensic fields (e.g. DNA), even with less than 12 points.

Figure 1 The configuration C1 shows a skeletonized fingerprint pattern with two bifurcations and five ridge endings. The configuration C2 shows a pattern with one lake, two bifurcations facing each other, a hook and three ridge endings.

The Move Towards Quality

Proficiency tests results

An indicator of quality may be searched in the results of the proficiency testing in the profession. The Collaborative Testing Service, Inc. (Herndon, VA), now in association with the IAI proposes declared annual proficiency tests to agencies around the world. A summary of the results for the latent prints examination is given in Table 3.
These results clearly indicate that error is possible and that some examiners do not reach adequate quality standards. But the evaluation of the standard of proof (or the quality of the evidence) cannot be limited to the examination of proficiency testing programs. Indeed, the corrective actions taken in an agency, following an erroneous identification, are far better indicators of quality than the report of errors (or successes). A global assessment of quality is then needed.
Table 3 Results of the proficiency tests made through the CTS program from 1995 to 1998

Total quality management

A total quality management system, based for example on the European Foundation for Quality Management (EFQM) model, can be recommended and is the basis of the most recent changes in the fingerprint practice. The various aspects of such a system are illustrated in Fig. 2, the program is divided here into three parts: fingerprint examiners, processes and products.
For each category the quality management scheme must specify the needs and objectives, along with the quality indicators used to monitor the system and the corrective actions available. Audits (internal and independent) permit the claims to be checked regularly. In such a general context, proficiency testing must be regarded as only one block of the building. The value of a fingerprint identification will be improved by a strict quality assurance program.

Figure 2 Outline of a total quality management system for a fingerprint service.

Conclusion

It is obvious that Locard proposed the basic concepts determining the value of fingerprint evidence. Locard’s opinions were enlightening and, above all, the most topical.
With regard to numerical standards, there is no scientific argument in favor of any predefined number of minutiae (in the absence of discordance) in order to pronounce an identification. Every comparison bears its own specificity which refrains from defining a priori the amount or volume of information needed for the identification. It forces us to rely on experience, professionalism and integrity of fingerprint examiners. Even if some practitioners rely on minimum numbers, they do not escape, thanks to the arbitrary limit, from the quality issues because, depending on the examiners, the interpretation of the number of minutiae in a comparison may vary to a large extent.
Finally, following this description of dactyloscopic practices, what is the standard of proof?
Definitely, the standard is not a predefined number of minutiae. Indeed, insisting dogmatically on any particular minimum number of points is not an effective way of insuring quality. The way forward is the adoption of a scheme of total quality management dealing with the various aspects of education and training, laboratory procedures and audits.
Concerning the possibility of giving qualified opinions based on dactyloscopic evidence (likely, possible, probable, etc.), it is clear to the author that from a scientific point of view, there is no argument to justify the so-called ‘positivity’ of the field. Nothing opposes the use of fingerprint evidence as corroborative evidence in the same way as other transfer traces.