I’ve always wondered about this. Is it even scientifically provable? Checking with google I found the page below. That statement from the US Department of Justice, that fingerprint identification has a ‘zero error rate’ is somewhat alarming. How can any human process have a zero error rate?
From ABC Australia :
They are unique just like snowflakes are: The chances are ever finding two that are alike is statistically improbable. Now we only have billions or so people on this planet so odds are there are no two finger prints alike. i.e. There are more combinations then there are people. The funny thing is about snow flakes is that there are so much of them there has to be duplicates. The tick is to find them…
[Hijack]If an Amish quilt-maker forgot to do this on one quilt, would the lack of mistakes count as their mistake?[/hijack]
I’ve found two identical snowflakes…on a SPOON!
I believe The Master has something relevent to say about this.
My racist Eskimo friend says that all snow flakes look alike to him. Claims that Eskimos have over 20 words for snow is the same as saying that the many terms used for any hated group makes someone perceptive. In fact, it is just the opposite.
Huh?
I think he’s making a joke about how all black people look alike? Or all Asians? Or something?
No. He’s saying that having a lot of words for something doesn’t necessarily show a keen observational skill, and deriding bigots at the same time. Multitasking response.
I have no idea why I submitted that but that is exactly what it means. I take comfort in the fact that someone, somewhere can make sense of what I write.
Apparently Simon Cole has found no less than 22 cases where two people had fingerprints that were functionally identical. Yeah, there may be differences on the molecular scale, but since fingerprint reading technology isn’t that sophisticated those differences really don’t matter.
I remember Cecil Adams writing about a case where a British police officer was convicted of a crime because they found his fingerprints at a crime scene. Eventually, someone else confessed to the crime, and apparently the other guy’s fingerprints were so similar that even the experts couldn’t tell them apart. The really scary part is that we only know about the mistake because the real criminal confessed. How many other mistakes have there been where the bad guy didn’t have the guts to own up to his actions?
As for snowflakes, I suspect that you could probably find hundreds of matches in any decent snowfall, but who’s going to to take the time to look for them?
Well, here’s the point- they almost never get a perfect print from a crime site. Criminals are not well known for smearing their fingers with ink and “rolling” them carefully (and even when this is done, the results often have to be re-done).
So, there’s “points”, and it’s not hard to have two people with the same points.
How many words does English have for water?
Water
Steam
Mist
Rain
Ice
Snow
Frost
Fog
Sleet
Hail
Snow
I mean, we come up with a new word every time water changes form; you might even include such words as river, lake, ocean, creek, stream, pond, etc.; they simply denote different shapes of water.
Well, in using them for about 125 years now, nobody has ever come up with identical fingerprints from different people. Even identical twins have different fingerprints.
Trying to actually prove it, though, is pretty near impossible – you would have to obtain fingerprints from every person in the world, and compare every one against every other one. And add in to your tests the thousands of new people being born every hour. (To say nothing about all the dead people, most of whom never had their fingerprints taken at all.)
But the fact that nobody has ever found matching fingerprints on different people is good enough for legal evidence. It’s probably more accurate than a lot of witness testimony in courtrooms.
And, like DrDeth said, the fingerprint experts are usually matching points from imperfect prints at a crime scene to prints on file. Then the matching one(s) are checked with other evidence, to see if that person(s) is a likely suspect. It’s pretty rare for a single, imperfect fingerprint to be enough evidence to convince a jury that a person is guilty – there’s usually lots more evidence.
And it’s quite possible for a fingerprint expert to make an error while comparing fingerprints. That’s why you sometimes see the defense in a trial bringing in their own fingerprint expert.
Doesn’t Diceman’s post contradict this?
Cites?
Simon Cole’s webpage, with a partial list of his articles is here: Simon A. Cole – UCI School of Social Ecology
I zeroed in on his 2007 Skeptical Inquirer article, available here: http://findarticles.com/p/articles/mi_m2843/is_4_31/ai_n27296432/pg_1?tag=artBody;col1
His Nov 2005 NAS talk is available on this page: See Session V, “Presentation”:
http://www.nasonline.org/site/PageServer?pagename=sackler_forensic_presentations
Interesting stuff.
From the 2005 presentation, I gleen that the problem isn’t so much that any two people in the world are likely to have identical fingerprints.
Rather, the problem is that any fingerprint sample could very well match multiple suspects. Indeed, he has dug up 22 faulty matches connected with criminal cases, according to this Science Daily article.
Cole believes that a scientific approach would essentially try to report a false positive rate during testimony, rather than just assert that it’s de minimus. That’s the DNA framework; fingerprint matching in contrast appears to be based more on an apprentice system.
That’s exactly my point. There might be billions of possible fingerprints, but we don’t have the ability to differentiate billions of different fingerprints. If, let’s say, there are 10 billion possible fingerprints, but forensic science can only reliably differentiate between 100 million, then for all practical purposes there are only 100 million different fingerprints.
Ok, Diceman, but I understand that Cole goes further than that.
At a crime scene, fingerprint samples are taken. They tend to be imperfect. The smaller or less resolved they are, the more humans they can be plausibly matched to.
That doesn’t disqualify fingerprint evidence. But it suggests that a methodology that does not rigorously address the possibility of such errors will be faulty. An image of one such false positive is here.
It’s possible that science can reliably distinguish between the fingerprint cards of all 6.6 billion people on earth. But even in that case, we’d still want to upgrade forensic fingerprint methodology, since criminals typically don’t leave perfect fingerprint cards at crime scenes.
