Grades of Uniqueness

[Texas Pat]

In the current issue of the JFI, a commentary by Joseph B. Kadane titled Knuckles and Fingerprints: A Comparison and Case Study, he sets forth the idea of "Grades of Uniqueness." Dr. Kadane was a defense expert arguing in court against a knuckle identification from the back of a hand in a kiddie porn photo to the defendant's knuckles. See an article on the case here:
https://www.vcstar.com/story/news/local ... 973198001/

The concept that concerns me in Dr. Kadane's letter to the JFI is this idea of "Grades of Uniqueness."

While testifying in the UK in 1999, I was explaining Ridgeology as opposed to the 16 Point Standard. I quoted David Ashbaugh, including his term "sufficient uniqueness." The Judge immediately stopped me and admonished that I was making no sense. His exact words, as I recall them, were, "A thing either is unique, or it is not. It cannot be sufficiently unique or insufficiently unique." In the UK, one uses proper English in the courtroom.

In an American dictionary, you usually find two definitions of Unique: 1) One of a kind; 2) Very rare. In correct English, apparently, there is only the first definition. But now Dr. Kadane wants to grade uniqueness across a whole spectrum of gray. I disagree with his overall analysis on several levels, but that phrase is especially disconcerting, especially in light of the current discussions and controversy surrounding statistical analysis of latent print comparisons and probability modeling instead of identifications.

I would be interested in hearing the thoughts of other LPEs on this idea of "grades of uniqueness."

[Bill Schade]

I think this is just more along the lines of

1) Levels of Complexity

2) Degree of suport

3) statistical measurements

4) Degree of ambiguity

etc.

Everyone is trying to get away from a binary decision on anything related to pattern evidence conclusions.

The "state of the art" in 2020

[John Vanderkolk]

Texas Pat, you wrote: "In an American dictionary, you usually find two definitions of Unique: 1) One of a kind; 2) Very rare." Make sure I don't buy that dictionary. If unique is also very rare, can unique also be less than unique? How does that work? Can there be something that is very unique or more unique? Words still mean something.

[Texas Pat]

I looked it up in my dictionary when I got back from testifying in 1999. I distinctly remember that old book gave two definitions for "unique" as mentioned in my first post. I kept the hard-bound Mirriam-Webster's Collegiate dictionary on my desk in those days and replaced it with a new one every three years. The edition I used in 1999 would have been either the 1996, 1997, 1998, or 1999 edition.

I see this morning in Mirriam-Webster's online dictionary that they give four definitions for "unique" now:
1. being the only one : SOLE
2. being without a like or equal : UNEQUALED
3. distinctively characteristic : PECULIAR
4. UNUSUAL

On the grayscale, if "one of a kind" unique is black and "very common" is white, then at least all of these are or tend to be toward the dark end of the scale with 1. and 2. being pure black. But it seems to me that Dr. Kadane is taking his "Grades of Uniqueness" explanation of fingerprints closer to the mid-gray region of the scale.

In other words, it is the statisticians from academia who are watering down our science, not us ourselves.

[John Vanderkolk]

One thing UNique and UNusual have in common is UN. Maybe I will have to change the way I pronounce you-neek to un-eek so it sounds more like unusual. Words used to mean something.

[Texas Pat]

Words used to mean something.

During that trial in the UK in 1999, the British judge used a word in pinning down the defense attorney during one particular argument. I knew the general meaning of the word, but had to look it up in my dictionary also as soon as I got back from the UK. That word was "Pedantic."

It would seem like we in the US have moved from being near-pedantic to the opposite end of that spectrum, too.