Varieties of First-Person Testimony

"an article published in the Philosophical Transactions for October 1699 offered an attempt to formulate a mathematical model for assessing the value of eyewitness accounts. . . . After some impressive calculations, he arrives at a ratio of 5/6ths believability per person"

One sixth of what I thought I saw

Could be confirmed by other methods

Devised by other seers. Five sixths

Was either false or disreputable 

Or impossible to confirm. I thought

I saw a strange little man with a beard

And a limp pass through my world

Almost unnoticed by those who should

Have thought to ask themselves, what is that

Out-of-place shadow shape doing here?