Evidence

From SkepticWiki

1 Definition
2 Discussion
3 Fallacies of Evidence
- 3.1 Argument from Ignorance
- 3.2 Fallacist's Fallacy
4 Related Articles
5 External Links

[edit] Definition

Evidence, from the Latin for "that which is seen," simply refers to a part of the physical (or mental) world which will support an inference -- for example, a mark from which conclusions may be drawn, the testimony of a witness in court in support of a theory or assertion, or grounds for believing such-and-such to be the case.

It is specifically not proof except under very unusual circumstances, usually only within the realm of the purely logical or mathematical. For example, I might offer as "evidence" the fact that I was in Finland last week (and have hotel bills to support this) and therefore couldn't have been committing a murder in Australia at that time. Of course, this isn't "proof" as I might have been able to invent a teleportation device that would let me travel to Australia and back in an instant, but the possibility of this happening is remote and so unlikely that a reasonable jury would accept that I was innocent of the murder.

[edit] Discussion

Interpretation of evidence is often a complex issue, made more so by the fact that evidence can often legitimately be brought to support many different sides of an issue. Three key aspects to consider are the strength/reliability of any given piece of evidence, the ambiguity of the evidence, and the so-called "burden of proof".

[edit] Strength of Evidence

The first aspect to be considered in a discussion of evidence is how reliable it is, which is to say, "how likely is it that what is stated to be evidence is actually true?" When a pathological liar tells us something, we have no reason to believe him or her. Advertisements are often little better in this regard. Perhaps surprisingly, the testimony of eyewitnesses is also often weak; for example, eyewitnesses will often mis-identify participants in an action, omit details, or even get them wrong, often in unbelievable ways. In one recent experiment, people were asked to watch a videotape of a basketball game and count the number of passes thrown. Of the "eyewitnesses" to the video, about half (in a later experiment rising to 90%) failed to notice a woman in a gorilla costume wandering across the scene for nine seconds.

Perhaps needless to see, if the "evidence" is likely to be false, then it shouldn't be given much if any weight. For this reason, many so-called paranormal events, for which the only "evidence" available are contemporary eyewitnesses, are often disbelieved by skeptical observers, on the grounds that there is no "reliable" evidence for them.

[edit] Ambiguity of Evidence

Another aspect is the question of ambiguity, or how many other (alternative) explanations does the evidence in question support? For example, if I claim to have seen a werewolf walking through my back garden, and cite as "evidence" a large canine footprint, is it more likely that the footprint was made by a werewolf, or by a dog?

In many cases, there are several possible explanations for any given phenomenon, and the evidence supports several or even all of them. There is also an aspect of cherry picking to be considered, where the "evidence" is only looked for after the fact, and mere coincidence is taken to be evidence. (How often do I check my back garden for footprints, anyway? Perhaps packs of dogs wander my neighborhood, and I always have their footprints in my garden, but I only looked when I thought I saw a werewolf.) This kind of selective memory is another reason why eyewitness testimony can be unreliable.

[edit] Burden of Proof

The final aspect to consider is the total strength of evidence needed before drawing conclusions. The legal system often considers two different threshholds before delivering a verdict : "preponderance of the evidence" and "beyond a reasonable doubt." In a case that boils down to a he-said/she-said disagreement (for example, perhaps "he" claims that she agreed to mow his lawn, and she says she didn't), whom should the court (or you) believe? Based on an assessment simply of the two people's characters, we may consider it to be more likely that she would lie than that he would, and thus believe "him," but it is not beyond doubt or expectation that he might lie sometimes.

Scientific experimentation and Internet debates do not have this formal notion of a "burden of proof" that must be met, but there is often an underlying belief among all participants of the general nature of the world. Any proposition that is inherently implausible or that violates the (assumed) nature of the world is therefore treated with greater skepticism.

As a simple example, consider the view out of an office window. If I claim to see a grey Volkwagen, parked in the garage across the street, there is nothing in that statement to contradict anything we believe about the universe. If I claim to see a lime-green and plum-purple striped Volkswagen, that's slightly unusual (since VWs don't usually come in that colour), but custom paint shops are nevertheless an accepted part of the universe. On the other hand, if I claim to see a fire-breathing dragon, that contradicts the accepted wisdom that dragons don't exist; unless I can produce better evidence for the dragon, people in general would be justified in concluding that I'm lying or deranged (because lying/deranged people, unfortunately, do exist).

This is often summarized as the maxim "Extraordinary Claims Require Extraordinary Proof".

[edit] Fallacies of Evidence

The technical definition of a "fallacy" is a piece of reasoning where the premises might be correct, while the conclusion is incorrect. In real life, we rarely have the luxury of indulging in entirely unfallacious reasoning, but there are a number of specific fallacies that involve the misinterpretation of evidence: