The Authority of Fingerprint Experts

[Neville]

Thank you Les for your very good article in the JFI.

Here in Auckland NZ over the last 10 years the fingerprint section has employed fingerprint trainees, most with Bsc or higher, and who as you suggest have arrived at the same philosophical belief as most fingerprint experts.

My only nit pick with your article would be that it is not the patterns that are unique it is that what makes up the patterns that is unique.

I for one was always told that fingerprint Identification is based on 3 biological factors, so when was biology removed from the sciences. Also when was it that only something that can be backed up by statistics is a science. If that is the case how come my son got his Masters at the Department of Psychology in the Faculty of Science at the University of Auckland, may be some one should inform the University of Auckland that this is no longer a science. For that matter just about the whole of medical studies is not a science.

Sorry, I got side track, well done Les.

[Gerald Clough]

Attempting to point out to a university that Psychology was not a science would obviously be futile, because schools are structured on practicalities, rather than arguable strict definitions. If the most strict definition of "science" was invoked, schools of science would have few departments. What matters is both the current and prospective nature of the field. Quite high authority in American psychology has bemoaned the field's failure to realize its long standing hopes of developing a more rigorous science. Obviously, such development is difficult, perhaps impossible, if the value of the discipline is to be retained and itself developed. Medical practice, too, would be largely ineffective, if nothing could be done that had not passed rigorous scientific validation. In both fields, clinical experience counts for a great deal. It produces great benefits through the practice of clinical psychology and the obvious beneficial practice of clinical medicine. Both have gained significantly from valid pure science. It wold be unconscionable to deny people the benefits of treatments, simply because there was not an certain knowledge that the diagnosis was correct and that the therapy addressed, in a specifically defined way, the condition. When we consult a physician, we know his determination of the condition may be wrong and that, even if correct, the treatment may not work, may not even be the most certain approach. But we conclude to follow that course in this most important of all endeavors.

I would not quite agree with the conclusion of the last paragraph of the commentary and the implication that it is correct and so therefore satisfies the stated purpose of the piece. That is the statement that were academically trained scientists to make the same examination (presuming they possessed similar experience, which I took as implied), they would conclude that fingerprints were unique and not duplicated. I think that is taking a step too far. I do believe that an academic scientist set to the task of providing assistance to a legal body deciding an issue of fact would conclude exactly what the latent print examiner concludes. That is that the scientist would form an opinion that a particular individual was the source of the impression to an extremely high degree of certainty. I do not believe that such a scientist would abandon fundamental logic to declare there was no possibility of duplication. The requirement of rigorous logic would compel the scientist, as indeed it should compel any person, to recognize that a finite set of objects can only be arranged in a finite number of ways. This is true no matter how finely the objects are defined.

The scientist would, I believe, were he interested in practical usage and not only rigid scientific proof, declare that he cannot say just how many objects are being observed (we can't, either, since we are taking in many aspects of what we observe) and therefore cannot say how many variations are possible. But he would also say that the observed complexity, the many ways in which even a small area of impressed skin can be said to present different features, leads to the conclusion that the number of possible variations is very, very high. There is, I think, an important point here. That is that the examiner's judgment of the ways in which the observed objects might be counted and characterized is the knowledge from which his conclusion springs. It is a knowledge gained by much close observation and cannot now be gained in any other way. That knowledge is essential to any reasonable formation of a conclusion of great rarity, the conclusion that at any given moment in human history, that particular set of features is so unlikely to duplicated in a way that defies discrimination is so extraordinarily unlikely that the possibility can be disregarded for any imaginable practical purpose. Any purpose, that is, except for forming a statement of scientific knowledge. To the scientist with the knowledge of the latent print examiner, his observation and conclusion lead him in the direction of seeking a more exact knowledge. To the working latent print examiner, his observation and conclusion is the end product. The two positions are obviously not mutually exclusive, and there are latent print examiners who are also scientists, producing the appropriate products of examination and also designing research hopeful of more exact knowledge.

Being an academic scientist does not preclude applying a knowledge of the nature of very large numbers to a problem of concluding with extremely high confidence in a practical matter. And being a working latent print examiner does not preclude recognizing that such a conclusion does not grow out of a provable scientific theory of reality. It is not flippant to point out that both scientists and latent print examiners stake their professional reputations on unproven conclusions they espouse. If you think scientists don't risk a great deal when they argue a belief that is not yet subject to rigorous proof, talk with some of them about the consequences within their own community of being wrong. It is a mistake to pretend that the scientists has nothing really at stake when they argue their belief. The differences are superficial. The two, science and fingerprint identification, can only come into conflict when the scientist denies the value of fingerprints or the examiner denies the legitimacy of seeking a scientifically validated more precise conclusion.

And I think it serves the fingerprint discipline poorly to say we must somehow fend off the possibility that valid research might point to statements of probability as the most accurate conclusions. I don't see that as much different from the 1899 proposal to close the US Patent Office on account of there being nothing more to invent or the 1903 declaration that the flying machine might come to be - in one to ten million years. (At that same moment, the Wrights were unpacking their equipment in North Carolina.) Argument to the end of supporting fingerprint identifications conclusions as extremely powerful and practical are appropriate. Argument to the end of pretending that there is nothing more to learn about how conclusions are stated is retreat to dogma.

[Neville]

and to suggest the best way to present fingerprint evidence is by statistics is another way to retreat to dogma. Let's face it, was it not the scientific community that was adamant the world was flat? However my point all along as in other threads is that we already have a statistical value that has been used over the last 20 yrs in courts all around the world.

If DNA is the Golden god to be idolised why is it they base their figures on a population data base that represents the prison population and not the actual population, by that I mean in NZ the vast majority of prisoners are of Pacific Island or Maori back ground (they make up something like 10% of the population and 90% of the prisoners, this is not actual but a wild guess, but you get the idea), where as the actual major population base of the country is of European origins. I am sure this is the case in many other countries also.

After all the hours I have spent in courts over the last 30 plus years, the one thing I do know from experience is that when evidence is being presented by scientist in court relying on mathematical models the jury eyelids start to get very heavy, their heads start to drop, they start to yawn, they are obviously very over it long before there is an end to the evidence being given. I have to say when the fingerprint expert finally gets on the stand the jury looks like they have been through hell and back they look at you, like, PLEASE NO NOT AGAIN, you give your evidence and within 10 to 20 minutes they are feeding out of your hand they are actually smiling and nodding in a very different way. They appreciate the black and white evidence you are giving them there is no doubt, sure you may not actually be proving the offence but you are clearly joining some of the dots for them, they and the Judge are very thankful for that. I have even been thanked after such times in the box by the judge, the defence and the prosecutor on the same case, on more than one occasion. So my advise to any new experts out there is to stay focused on what you are there for and it aint to put them to sleep.

As a foot note for yo awl up there 'On or about 31st March 1903 a reclusive New Zealand farmer Richard Pearse climbed into a self-built monoplane and flew for about 140 metres......' I have not completed the quote as it serves me no purpose to and those that know the rest of this history need no reminding. But he flew way further than I ever have on my own.

[Gerald Clough]

I don't think either of us have to worry much about probabalistic fingerprint conclusions in court. I've worked through what it will take to achieve that, and the work involved, even to begin gathering data, won't begin anytime soon. And how it's presented in court is another whole problem. In fact, I think that if those who complain (correctly) that the standard conclusions are not subject to rigorous scientific validation would set out to attempt to develop conclusions AND validate them, they might well (okay - some of them might) turn their efforts to the less scientifically rigorous but more practically valuable examination of the question of present reliability. If the objection is that reliability has not been validated, propose the best practically possible validation study and carry it out.

On the other hand, just because I can't immediately see a solution to the problem of designing probability studies doesn't mean some bright spark won't see it. I just think that neither long-standing practice nor being more palatable to jurors can be legitimate arguments against trying.

Re: Farmer Pearce

I suppose of the history of early manned flight (and attempts) back to at least 1848 teaches something, it's that, if you want historical credit, you have to:

1 - Survive, and
2 - Build another one.
3 - Figure out how to make money from it.

[Les Bush]

Hi Neville, Gerald,

Its been a while since I posted and with the Inquiry over I can return to a normal life. We must all appreciate that the science of latent fingerprint identification its use and abuse has resulted in this significant case for modern times. Lets hope the Inquiry has to the knowledge and experience to resolve the primary issues of mis-attributions of Y7 and Q12. Given the poor presentations of both latents it would be possible for an untrained person to give weight to arguments about the observations of features which rely heavily on practical experience. Fortunately the global community of experts has the IAI to thank for making its declaration that within this professional organisation those who attribute Y7 to Shirley McKie have committed technical errors.

The article in JFI was meant as an immediate response before another incident repeats the same claims. This type of message delivered through the venue of a local jurisdictional court requires balance in the forensic argument. Hopefully the article has achieved a realistic view of the logic employed by experts. This logic is not mathematical but more biological as it relies upon the variables associated with the formation of the source skin. To answer your comment Neville about what makes the patterns unique my view is that the sophistication of the nervous system is a primary influence since the special sense of touch is the resultant outcome. Research also supports this understanding and it makes logical sense as the network of nerves requires a uniform and dense distribution of receptors.

In answer to your opposition Gerald about what an academic would state with regard to an examination result. Firstly you have applied mathematical reasoning and conditioning to the situation which in my view isnt appropriate. Secondly you have assumed that a finite set of objects is what the science of fingerprint identification is dealing with when solving problems of source attribution. Lastly you have taken a view that the article is relating to latent examinations and excluding the impact that 'duplication' has on ten print examinations. The biology of friction skin formation begins with the genetic history of the parents. The new organism will have unique conditions of ontogeny infinitely different to any other human. Within this generation of humans that postion has been substantially proven. I would be interested to know what finite set of objects you rely on that represent and demonstrate the human fingerprint condition. I would also be interested to know your view about ten print examinations in contrast to our collective experience of 'ground truth' that fingeprints can differentiate human populations to the level of an individual.
Lastly, my next research project is to design a validation study for both ten print and latents on the reliability of fingerprint examinations to achieve 'ground truth' outcomes. Regards to you both from oz. les

[Gerald Clough]

I most of the discussion in this vein, aside from the practical application issues, I try to keep us mindful that we often use language to convey the working principles when the language isn't compatible with a strict scientific viewpoint. This is not to say that we have to limit ourselves to that strict viewpoint, but we have to remember that (1) scientists will object to what they view as improper terminology, and (2) there is an as yet unknown potential for extending knowledge in our field through the kind of inquiry that science critics demand. I'm sorry. This is a bit long.

The notion of a finite set situation I propose is that there is a limit to the density of characterized features we can observe in any portion of skin. As we differentiate smaller and smaller observable differences, there is a point where the differences become so small that they can't be observed. And the ability to observe differences is limited by the nature of what we work with where the physical effects of impression prevent us from differentiating as finely as we theoretically might, since we have to allow considerable variation among the impressed representations of the same physical skin feature. I think this becomes most clear if you imagine trying to develop a coding to define a feature at whatever practical level of observation you might choose. At some point, the number of possible coded representations of a feature is limited by the requirement that impressions of one actual feature produce the same code. Even if you're not seeking a way to code feature data, you reach the same kind of limits. In observational comparison, there's and internal brain coding that tells you when you accept that this impressed feature is like that impressed feature. The limit is the point at which we can't tell the difference. So it doesn't work to say that as we look closer and closer, we can always look a bit closer and see the difference.

Because there is a practical limit to how finely we can differentiate, and because the features of friction ridge skin can differ in ways more finely than we can observe, we must admit that we many not be able to tell the difference between two different physical features. You could theoretically construct a very large number of examples of skin that could not be differentiated by an examiner, because you kept the differences below the observational limit. In the set we just constructed, we could go on adding to the set, so long as we kept the changes below the observable limit, but to the examiner, they would all appear to be identical. This true, no matter how finely you observe, within the practical limits of our ability to observe in examinations. I suppose the finest observation we are capable is at the pore level. We certainly don't observe at the cellular level.

It is logically invalid to hold that what we can observe in examinations is infinitely variable and that therefore the arrangement we observe in any given portion of skin cannot be duplicated. If I take my supposed infinite set of skin portions, I can take every one of them and create a variant by an alteration below the observation limit and create a different skin portion. If I can do that, my set of observably different skins wasn't infinite after all. I had to god below the observational limit to continue on toward infinity. It is perhaps more impressive if I have created my proposed infinite set of variations by taking into account the various ways features may be impressed, allowing for the fact that I accept as alike impressions that could not be exactly overlaid. I could alter the actual source skin by making a change that was within the tolerance I allow in interpreting my observation.

It does not cure the objection to my argument to point out that the sheer quantity of observable detail used to differentiate is extraordinarily large and the number of ways in which the details can be varied and still observably different is therefore truly enormous. Enormous is merely very large, not infinite. And no finite set can produce an infinity of variations. The best way to mathematically represent the number of variations in our problem is difficult to determine and beyond the scope of this discussion (determining that would, I think, be quite useful), but the resulting value is never infinity. And so, inescapably, there is always the possibility that the impressions of two different portions of skin may appear that cannot be differentiated by an examiner. This is the science objection to the uniqueness claim and the absolute conclusion.

But the science objection misses the practical point that one need not hold to absolute uniqueness of any portion of skin, even only at the observable level, in order to make an absolute conclusion. The conclusion states a belief, formed by a qualified expert, that this impression was made by that individual. It is not and need not be a statement of scientific fact guaranteed to hold true in an imagined examination of every portion of skin that has existed, does exist, or ever will exist. Nevertheless, the scientist is compelled to object to language that seem incompatible with plain scientific knowledge and mathematical deduction. It lies in the realm of highly qualified human opinion, derived from careful observation and considerable experience. As such, it is immensely valuable to its purpose and that opinion is fully worthy of the most serious consideration by a judicial finder of fact asked to exercise their best judgment from the best evidence available. As you stated, every examiner's knowledge and experience has led them to the belief that they can perform an examination and can determine, to their most firmly and confidently held belief, the identity of the source.

The difference between concluding identification from a small impression and concluding from even a full impression of the major care set sort is nonexistent for this argument. While the lay person may be even more confident in an opinion drawn from a large impression with hundreds of times more detail than the small impression, the threshold of expert belief expressed in the conclusion from the small impression is precisely the same. (If an examiner can't confidently identify from a least sufficient portion of the ten-print impression, as if that smaller portion was all that was available, there is a very great problem.) The examiner is arguably not more strongly convinced or more powerfully possessed of an opinion of identification by the examination of all ten fingers and palms. In both instances, the scientifically honest examiner must admit that there is incredibly small possibility that either impression was made by someone other than the source, even less likely with the larger quantity is available, but the word "incredibly" is used precisely and literally. The opinion is that the proposition of an alternate source has no practical credibility whatsoever.

So, I think there are two approaches for validation. One is scientific validation, which for the reasons expressed here and in my past series on validation in The Detail, can only produce probabalistic conclusions. The other approach recognizes that fingerprint identification exists in and for practical forensic purposes. While the science approach is worthy of attention, I believe it is going to take a very long time to develop even the methodology and will require many preliminary studies and, like most scientific study, will take many attempts at refinements and much argument over methods to produce anything like a scientific consensus, if indeed any such consensus can be found. It will, however, point up the kind of numbers involved, which will speak strongly in favor of the confident opinions of examiners. We don't know what's possible yet, and we don't know what might come of it. But we cannot hold that anything is out of bounds for science to explore. We often react badly, because critics use language like "leap of faith." There's no leap of faith. It is, indeed a leap of faith to attempt to substitute clinical experience for scientific validation. But we're not making that leap. It's not a leap of faith to form a professional opinion. There's a misapprehended conflict there between the scientists and the fingerprint discipline that needs better understanding of the two realms.

The more practical validation can, I believe, be aimed at demonstrating that the beliefs formed by the examiner are justified, albeit that they cannot now be assigned any particular probability. It would be, I think, a sort of attempt to test to destruction. How far can you push the practice of fingerprint identification before you observe it fail. In other words, how little data can I input before the results become demonstrably unreliable. Although it requires a very large effort, I think it's extremely valuable, because it effectively tests our practice of forming conclusive beliefs. That's a very appealing prospect, given the realm in which we apply the practice and the fact that, no matter how correct scientists may be in their view of what they would require for scientific validation, our contribution is one of the most highly valued and one that is absolutely not going to be discarded or suspended until scientific validation is satisfied. If we had to protest the scientific objections, we could borrow from our GLBT friends and march around the university chanting, "We're here. We're clear. Get used to it."

[Les Bush]

Hello Gerald,

Your reply as usual was expansive and philosophical and a little critical. There were things I agreed with and others I didnt but this posting will focus on just a few. The science of fingerprint individualisation (ten print or latent) is a forensic science rightfully belonging in the scientific community of criminalistics. This has been earned by the continuing practice of examinations, the reliability of results (ten prints and latents) and the continual professional development of experts. This weeks Detail really does lay down the challenge to our ability to produce expert results to the court and it diminishes the authority of fingerprint experts. We have a choice to lay down and let it happen or stand and face the wind of criticism. Your response hasnt answered the questions of what finite objects represent the science of fingerprinting or what impact ten print examinations have on the reliability of fingerprint results. In your response you refer to the impact it has on examiners but I'm interested in the scientists view. Your response also mirrors the message in Minn v Hull and similarly it is filled with hypothetical propositions whereas the facts of the argument in support of the reliability of fingerprint science are established over the past 100 years of testing. There are two ground truths that support the belief in the uniqueness of fingerprint patterns of detail. No duplication has been discovered under testing conditions focussing on distal phalanges. The ontogeny of monozygotic twins as the extreme biological condition has substantially proven that there is infinite arrangements to the combinations of detail. Lastly on the assertion of using language and improper terminology. A fingerprint expert produces a result (tenprint or latent) to the court expressed in evidence as opinion. The expert has formed a professional conclusion about an examination requiring the careful application of forensic science in solving a problem. The language in your response also includes the use of identification whereas individualisation is the correct term. And finally the current use of the term 'pore' is a dogma which should be avoided if scientific credibility is sought. I know you can look at figure 2.25 on page 58 of Quan-Qual FRA and it is a generally used term. However you need to appreciate that scientifically the anatomical name for what is observed on friction ridge surfaces is ostia. The figure 2.25 should be labelled as the anlagen or resultant dermal duct of the eccrine sweat gland. Regards once again from oz. Les

[Gerald Clough]

I think the fundamental difference in how we view this is how we view such things as the motion reproduced in The Detail this week. Such motions are understandably one-sided and intended to lead the court to imagine that there has been some sea change in the scientific community. This is, of course, not the case. The NAS report clearly states that it contains nothing new. All of your points are valid and are obviously what has been in the judges' minds over the years. We are on firm ground and of one mind with almost every court when we argue as you have. I happen to believe that we can only make ourselves vulnerable if we pretend to any statement that is scientifically invalid. This does not mean that every piece of knowledge we cite must be accepted by the scientific community as scientifically validated. The court is not adjudicating competing claims of scientific truth. They are ultimately judging the question of whether the fingerprint conclusion is more prejudicial than probative. Thus, the final defense prayer is "The fingerprint testimony should accordingly be excluded. In the alternative, the fingerprint testimony should be excluded as more prejudicial than probative."

I believe our best defense is to keep the argument in the vein of what I would call "high-value clinical experience that is supported and nowhere contradicted by scientific research." And, of course, we should always point out misstatements when they appear. One of these appearing in the Motion is the use of the quote: "Court acceptance without much scrutiny in adversarial legal systems has led to an almost blind faith in fingerprint evidence." It is merely a quote which itself is a mere opinion and certainly not to be taken as any kind of valid scientific conclusion at all. Obviously, courts have not "blindly" accepted fingerprint evidence. It is true, in fact, that they are not blind and are fully aware of its long-standing value and reliability. And it is well to remind the court (although I don't think they necessarily need reminding) that many kinds of expert analysis have extremely useful and even critical roles in litigation without having been exhaustively explored and stamped as proven by valid scientific research.

One of significant weaknesses in the argument that science doesn't support fingerprint conclusions is that valid scientific research has not shown it to be unreliable. Courts are historically quite insistent that any research presented in opposition to expert evidence be very specific to that evidence and very clear in its results. An example in the US is eyewitness testimony, which although not strictly expert testimony, is offered as the product of a common human expertise. Many attempts were made to offer all sorts of research on the general nature of human perception, primarily attempting to offer general experts to talk about it. Almost universally, those attempts were rejected by the courts, because the research was not specific to eyewitnesses. Now that there is some specific research, it's getting admitted in some circuits. By no means all, because courts are loathe to muddy what they see as a plain process of identification. The probative-prejudicial argument works both ways.

The other thing we have to hammer home is the fact that the NAS report did not declare fingerprint identification invalid and, in fact, did not question its validity at all. They merely and properly call for the long-term effort to see what science can contribute. I think the general attitude is, "We should be able to do something in this are. Why haven't we?" I can't argue against that without prejudging the unknown results of such a project. All of these motions pretend to shock and dismay that fingerprint identification has not been scientifically validated. I don't think courts are often fooled, and they are quite accustomed to these sorts of exclamations in pleadings. (I kind of think this particular motion rather talks down to the court, as if the court is ignorant of the whole issue. An error on their part, I think.)

Bottom line is experts form specific opinions of what reality is revealed through their analysis. The opinion is reasoned, experienced, long found to be reliable, logically could be expected to be reliable, and not contradicted by any valid research. And, as we say, it's grounded in good scientific knowledge, so far as what is known can be applied. Pretenses that this one particular discipline should be held to a higher standard than any other are not going to sway the court. And the misrepresentation of expert fingerprint conclusions as statements of absolute scientific fact do not fool the courts. They understand very well the nature of expert opinion testimony and well understand what an expert means when they state a conclusions. The ultimate question before the court is whether the jury is seriously misled to a degree that prejudices the defendant by testimony that simply should not be given to them because it has little value. It is never a question of having to be absolutely, scientifically proven and reliable. It is strictly a judgment of whether it is fair to present the conclusions. Courts want every piece of valuable evidence and expect the defendant to bring to the jury any argument that it should not be taken seriously. Our meaningful efforts will be from the witness stand, because that's where the courts want the arguments to be heard. And the courts are very aware that this is not strictly a fingerprint issue, and their decisions always consider expert evidence in general.

And I may well use "pore" again. (That was a bit snarky. If I offended in that way, it was not intended.) I'm not seeking scientific credibility. I don't have to. I'm not arguing scientific validation. I'm arguing forensic credibility, what's credible in the legal forum, where for all that they'd like to make it a matter of scientific validity, the test of value is quite different.