Friday, January 06, 2017

Evidence Is Ambiguous

A Chinese-American friend who tried to like the Sherlock Holmes books found himself continually frustrated that Holmes kept believing that the clues he discovered had only one possible explanation.  Men of average height sometimes have the stride of a much shorter or taller man, he countered; ex-sailors might carry toys for two children as an errand for a shut-in neighbor. My friend is quite right.

As an example, I have walked in the woods more this past week, and we have had snow. I noted that some person with bigger feet than mine but probably lighter has been walking some of the same trails, coming in from Bog Rd. Large canine prints put down at about the same time suggest it is a taller man taking a big dog out for a walk/run.  The doggie prints sometimes go away from the trail for 5 yards or 30, but always come back. I actually met them at the very end of my hike today, as they were coming in, and they look as imagined.

However, applying this same logic, I would have to conclude that another person of undetrmined sex (not indeterminate sex - don't get me in trouble here) comes in on snowshoes from New Boston, walking a pet deer.  This seems unlikely, even in New Boston. But the deer prints are put down at about the same time, and they run together for over a mile in a loop. The deer tracks do go off to one side or another on occasion, but they come back to the trail soon, just as the dog's do.  I thought I had established that the deer came later because its tracks printed over the human's. (Let's assume it was a human on the snowshoes.  Let's not go completely overboard here.) Yet just a bit further on it was clear the snowshoe prints came after. I suppose that puzzle is solvable if we postulate that there are two deer, but surely, a man with two pet deer does not suddenly leave one on the trail, expecting that the other will appear shortly to complete his walk. That might do well for Tom Stoppard's solution in "Hapgood" to the Seven Bridges of Konigsburg problem, but This Is Reality, Greg.

This comes up in reference to Bible preachers who profess to be reliable because they tell you to bring your Bible along and look up what they are saying as they go along, to "prove" to you that they are not making this up and what they say can be relied on. I have come to conclude that this is a bad sign coming out of any preacher's mouth, and one had best be alert thereafter. I claim this is the real data and I tell you this is what it means.  I insist you confirm for yourself that this is the real data. Therefore, you should believe me that this is what it means. The one does not follow from the other. The same goes for people trying to prove things to you in print or on internet - popular science and social science.  The trick is very common.  I say music is only a commodity in this society.  Here! I have actual data that shows that people purchase sheet music and recorded performances. Some people make their livings on nothing else, and the entertainment industry permeates our entire economy.  There! I've run rings around you logic'ly. Music is only a commodity in our society.

I think entire social sciences are founded on such things. Remember the person with the pet deer. Evidence can point in many directions and have many interpretations.

3 comments:

Texan99 said...

The book I mentioned last week, "How Not to Be Wrong," touches on this, and specifically references the remark about "When you have eliminated the impossible, whatever is left, however improbable, must be true." Far too often, even assuming we correctly eliminate the impossible, we've failed to take account of all kinds of other possibles. Mystery stories often suffer from this weakness, especially ones that have a Sherlock-like character. The Sherlock stand-in Dr. House actually does a good job of avoiding it; they often make mistaken diagnoses before they find one that really stands up to scrutiny--especially, a diagnosed illness that actually responds to the recommended treatment, though even that isn't always a lock-cinch.

Apparently the human brain is aces at intuitive associative leaps but just awful at careful reasoning about causation. I call it "Stands to Reason!" thinking, which leads to belief in Ouija board predictions and the like. It's one reason I so enjoy books about medical mysteries and airplane crash mysteries; if they're rigorous, they tend to avoid post hoc ergo propter hoc errors, those being the errors that began the mystery before some bright soul came along and re-thought things. A good mystery fiction should be at least that careful. "Chinatown" is a good example.

Christopher B said...

I think Texan99 touches on something that this brought to my mind. Sherlock, et al, are dealing with a known conclusion, a dead body, which eliminates a great deal of possible interpretations of data. House might make errors in diagnosis but the fact that he's dealing with a disease similarly limits the universe of solutions. It also reminded me of a meme I've seen a couple of times on Facebook recently showing the old example of an argument over whether a symbol is a 6 or 9, and claiming that the person who, in essence, provides the most context to the argument is going to be correct (so interesting that the Left is now all about getting to the 'correct' answer, not 'you have to consider other perspectives'). It struck me that this doesn't really follow, since unless you fully understand the data surrounding the symbol all you're really doing is heaping interpretation on top of interpretation. Instead of bringing clarity you're just entrenching yourself into a bigger set of conclusions.

james said...

We tried a game with the kids in which we described a simple scene, and the challenge was to devise as many different ways it could be explained as you could. _I_ liked it, but I suppose you had to be in the mood for it, and the kids usually weren't.