CLPEX Chat Board

I've never heard it put this way so I thought I'd share. Has anyone else seen the "nearly unique" description?

"Fingerprints leave nearly-unique marks on a surface that can be copied and compared to a database to identify a suspect, a police technique that rose to prominence in the early 1900s." (highlighting added by me).

Source: "Dutch Forensic Scientists Discover How to Date Fingerprints" link: http://www.sbs.com.au/news/article/2014 ... int-dating

I saw that and thought with the new paradigm that you cannot absolutely "individualize," perhaps "nearly unique" is an apt description in today's environment.

But isn't that like saying "almost infinite" or "a little bit pregnant?"

Cindy, you brought me back to posting. I struggle when people say, "very unique." Now I will struggle when people start saying, "nearly unique." I am happy the author cited nobody on that statement. I am not persuaded to abandon, "Always remember, you are unique, just like everybody else!" JohnV

There are two points that need to be clarified in any expert discussion of this topic.

First, what is the definition of "fingerprint" exactly? Are we talking about the friction ridge skin itself on the tip of the finger? Are we talking about a scene of crime mark, a latent print, or a chance impression, the donor of which is unknown? Are we talking about a clear, fully rolled fingerprint in black ink on a white piece of paper? In SWGFAST discussions on documents over the past couple of decades, confusion sometimes resulted when a party on one side of an argument meant by "fingerprint" the skin itself, while a second party meant a borderline quality latent print. So first, do we mean only one of those things, or all of them?

Second, David Ashbaugh advocates in his works on Ridgeology that the criteria for making an identification are "ridge formations, in sequence, having sufficient uniqueness, to individualize." When I testified to that exact wording in the trial of Shirley McKie in 1999, Lord Johnston, the Judge, stopped the proceedings and admonished me that an item either IS unique or it is NOT unique. It cannot be either sufficiently unique or insufficiently unique. I acknowledged the correctness of his observation while pointing out to the Court that I was quoting David Ashbaugh from a peer reviewed and published article, at which point Lord Johnston annoyedly waved me off and said, "Very well, you may continue." But after realizing that Lord Johnston was quite correct, I modified my teaching to instruct students that the conditions for making an identification are "ridge formations, in sequence, having DETECTABLE uniqueness, to individualize."

Combining these two points, it becomes crucial to recognize that at the foundational level, every cell of the skin, every sweat pore, every ridge, and every entire print is unique. Having acknowledged that, we must reach the logical conclusion that every crime scene mark, every latent print, and every inked print is also unique. Even the most worthless smudge showing no detectable ridge detail at all is still fundamentally unique and could only have been deposited by the exact skin cells that contacted the surface with the residue present on the skin at the time of contact and under the precise conditions of contact. Even the slightest deviation in any of those factors would have produced a different print.

The problem then is never the absence of uniqueness. The problem is the examiner's inability to detect and compare the uniqueness that is present in each of two friction ridge skin impressions. The question is not whether the "fingerprint" is unique, the question is whether we can detect and recognize that uniqueness as representing the same region of friction ridge skin in two different impressions.

"ridge formations, in sequence, having DETECTABLE uniqueness, to individualize."

It's why I am an advocate of "features are 'sufficiently discriminating' to individualize". The new(ish) SWGFAST Articulation of Individualization document focuses on two key components in an individualization: 1) the features between the mark and control print are in correspondence and 2) the features are sufficiently discriminating, such that a large, but reasonable, population of potential donors can be filtered down to a single source of the mark.

g.

So Glenn, how does that roll up into the idea of a statistical measure of uniqueness?

The flow out of the recurve of a loop where it starts to meet the horizontal ridges has a large volume of the same type of ridge events. Arguably, the discriminating power of information in this area is less due to the frequency of the type of data and the potential ambiguity affecting the ridge count between them that can occur when bifurcations fail to reproduce fully, giving the appearance of ridge endings.

It seems that would be difficult to account for explicitly with so much tolerance neccesary.

This is so simple compared to all the factors and facets being associated with individualization, but will share it. This is wording taken in an opinion rendered in the US Supreme Court on another issue:

“…hard to define, but I know it when I see it…”
Potter Stewart
Associate Justice of the US Supreme Court,
Jacobellis v. Ohio (1964)

Hi Ernie, I've heard other people say that as well. My only response would be that the statement represents a lack of a definition of value as opposed to an actual definition of value. The Stewart decision boils down into 'training and experience' or 'trust me I'm an expert', but I think you'd agree that demonstrability is rather important.

Being that we're an evaluative discipline it seems that there would be a published theory of value. For instance it would be important to recognize even simple values such as information that differs in kind (level 2 vs level 3, or dot vs enclosure) as opposed to information that differs in degree (thicker ridge vs thinner ridge, curvature vs angularity) etc.

I'm not talking about fully articulating every single thing, as that's impossible, but more along the lines of asking the kind of question that takes us from simple spatial & sequential agreement to a deeper understanding of how the data in a print is functioning in our arrival at a conclusion.

I too have a little bit of difficulty with the "Detectable Uniqueness" position.
My concern is that it could provide a vehicle for people to "cherry pick" the various classes of features that they see while ingnoring the absence of features they should expect to see in a given friction ridge inmpression/image.

FWIW How about "...ridge formations, in sequence, having sufficient value, to support an individualization"?

Shane Turnidge

I personally use "enough reliable and discriminating features in correspondence to individualize" - not really different from the newer version of SWGFAST. But I think you could go to the next step of asking "how do we decide if some feature or event is sufficiently discriminating?" which (I think) gets to the point that Boyd is making about the broader understanding of how we do what we do and how that information leads us to a conclusion. While there is some interesting research on this, I wish we as a field looked toward anthropology a bit more for some potential answers. Humans are naturally quite good at recognizing faces, so much so that we often fool ourselves (think of that rock on the moon that looks like a carved face, or seeing a person on a piece of burnt toast). Even as very young children we can pick our parents out of a crowd with very little information. Additionally, we are hardwired with the ability to read extremely sublte changes in a person's facial expressions - this being necessary for survival in a close-knit social group. These abilities evolved a long time ago, and I can't help but think that they are what allow us as LPEs to recognize, categorize, and differentiate latent prints. We're using the same mechanisms that help us recognize a familiar person from a distance or from the back. Steven Mithen has written some interesting work on cognitive fluidity and how the modern human mind evolved. Whether you agree with him or not, it's an interesting topic. I feel that understanding how these general abilities evolved and how they are used for recognizing ANYTHING seems like a very useful topic for pattern recognition scientists.

"detectable uniqueness" - Pat
"sufficiently discriminating" - g.
my prefered wording is "sufficient distinctiveness", but they're all shorter ways of saying the same thing.

i.e. "Sufficient unique, distinctive and discriminating features of the skin have to be detectable in both the questioned and exemplar prints to conclude that they originate from the same source."

Hello Cindy,
as a teacher/lecturer you appear to be looking for interesting articles to stimulate your students through reading the responses on this site. As you can see the semantics of terminology and lack of consensus and above all the strength of personal opinion are what dominates these discussions. But looking at the article you posted, the author is obviously a research chemist who has found a way of experimenting with fingerprint residues but as yet not found a way to make it field applicable. The terms "nearly unique" are made by someone without fingerprint science knowledge or professional experience. As a teacher/lecturer I dont know if you are also a certified and practising expert. I will stop before entering into the many reasons why fingerprints are unique, and close with a term your students should explore and that is 'reliable predictability'. Using a process of hypothesis development and testing, combining analysis with observations and leading to evaluation and conclusion. As Pat stated, are the fingerprints under examination high quality control or latent impressions and if so what is the place for "nearly unique", the basket in the corner. Cheers from oz Les

Hello Cindy,
you may be wondering why I chose 'reliable predictability' as a good alternative to those terms from your reference article. During the identification process the expert must reach, a point of confidence in the conclusion, or whether it is exclusion. The amount and quality of the data in the fingerprint pattern, the friction ridge details, and the third level details will collectively build to prove that the exemplar and the questioned fingerprint have a common source. That source is the skin area of the person of interest. The second part of the process is for your students to understand how and why that particular skin area has no duplication either on the same person or on any other person. The fundamental principles of fingerprint science provide the belief that no two fingerprints are identical unless from the same skin source. There are two theories and several dogmas that teach how and why fingerprints are unique. Your students should be competent in researching the correct and scientific basis for the biological development of special sense skin, friction ridge skin, or volar skin, or hairless skin of the palmar and plantar areas, or five layer integumentary skin, or thick skin. Needless to say your students should equally be able to research the interdependent relationship between nerves and the integument in the process of forming both uniform ridges and intermals and the uniform spacing of ostia along ridge summits. Cheers from oz. Les