Testimony is Evidence

[Amarok]

I will spell it out more fully:

Take your claim "Witness testimony is demonstrably unreliable as a form of evidence". That is simply not true. We rely on it to some degree millions of times a minute all over the world: In court cases of all types (criminal, civil, family), the running of governments of all levels, the running of corporations, the reporting of news, writing of articles/books, etc. These are all defeaters to your premise A. 

Perhaps you will backpedal and say "some witness testimony is demonstrably unreliable as a form of evidence". I would agree with this premise. But there are ramifications of this backpedaling: The converse is also true: some witness testimony is reliable as a form of evidence. If that is true, your conclusion is no longer a conclusion that follows from the premises--but a statement of opinion. Now you have:

    1. Some witness testimony is demonstrably unreliable as a form of evidence
    2. Some witness testimony is demonstrably reliable as a form of evidence
    3. Therefore the evidence is reliable on a case by case basis. 

Wait! that looks familiar.

For the third time, tell me why this is not more accurate:

     1 A witness's recollection could be wrong
     2 The witness's character, cognitive ability, subject knowledge, experiences, and track record serve can minimize the possibility of error
     3 The context of the event can minimize the possibility of error
     4 Therefore the reliability of testimony varies depending on the witness and the context

Since you've omitted a critical piece of my position twice now in order to bolster your own, I will phrase my response such that you cannot ignore it again.  

Just because we are forced to trust (or 'rely on', if it pleases you) for practicality purposes, the most mundane and inconsequential testimony in order to be functional living beings, the fact that we do so does not change the inherently unreliable nature of witness testimony as a form evidence.

It just means that in order to function from day to day, we must make choices about about which types of claims are worth accepting solely on an extremely fallible form of evidence, and which ones aren't.  We do this by evaluating the risk; the consequences of the testimony being wrong.  More serious claims beyond the mundane, (religious ones, for example), carry far-reaching consequences that effect our world views, and the ways in which we value our lives, and the lives of others.

Example 1.   My husband testifies to me that he fed our son Macaroni and cheese for lunch.  Maybe he did.  Or maybe, due to sleep deprivation, he forgot that he actually fed him French toast, and was remembering mac and cheese from the day before.  What do I do?  Dig further?  Pull the garbage can apart in search of the empty mac and cheese container?  Interrogate my three year old?  Smell his breath?  Assess the color and consistency of the food stains on his shirt?  Of course not.  I'm going to go ahead and accept his testimony alone, despite the fact that I know it could be wrong. Why?  Because, there are literally ZERO negative consequences to my son having French toast instead of mac and cheese.  Any further investigation for corroborating evidence is simply not worth the trouble.

Example 2.   On the other hand, if my husband tells me god came to him in a dream and said our son has been chosen to be an angel, so we must toss him into the Grand Canyon, well...lol.  Needless to say, I'm not going to accept on his testimony alone, that he's having conversations with God about our son's divine future.

In both examples, my husband's testimony could be erroneous.  This is why testimony, as a form of evidence, sucks.  It's been demonstrated to suck over and over.  It's is just a matter of which claims we're willing to risk being wrong about, and what the implications of being wrong about them are.

Agreed testimony is awful evidence. When it really matter .