The game of telephone involves whispering a phrase in a person’s ear. That person then whispers what they thought they heard into the next person’s ear. This continues until the final person is reached. The final person says the phrase out loud. This is compared with what the first person initially whispered, often to comical effect. A lesson from the game is that successive assertions of testimony can distort the content (truth-value) of that which is asserted. Out of this game comes the following principle, which is often implicit in accounts of testimony:
(RIPD-T) The reliability of a testimonial assertion is inversely proportional to the distance of that assertion from the original testimonial assertion.
The Reliability Inversely Proportional to Distance principle (RIPD-T) accounts for the value of eye-witness testimony. If a person actually witnessed an event, then (ceteris paribus) their testimony is typically more reliable than, say, someone who heard about the event from another person. Setting aside cognitive biases and the reliability of agents according to the reliability of the senses (i.e., whether someone is hard of hearing, has poor eyesight, etc.) a problem arises.
There are cases where the original testimony is less strong than subsequent testimony based on the original assertion. Such is the case when an event does not cognitively register, but subsequent assertions of the testimony better approximate the truth. There are historical examples where native people, upon seeing ships off the coast, either did not see the ships or did not see the ships as ships. A native person who is an eyewitness to a ship, which seeks to conquer the person’s native land, may initially report that they saw a large, sculpted floating piece of wood today. As the testimony is transmitted throughout the tribe it may be refined to a floating boat, a floating boat with people, until it reaches the Chief of the tribe and it is reported as a floating boat full of foreigners. This happens in everyday life when things are new and one does not yet possess the correct concept or language required to approximate the truth in testimony. Subsequent assertions of the testimony, especially in a dialectical context, can result in positive refinement of the assertion. This leads to a principle contrary to (RIPD-T) called the Reliability Proportional to Distance principle:
(RPD-T) The reliability of a testimonial assertion is proportional to the distance of that assertion from the original testimonial assertion.
The (RPD-T) principle holds that testimony becomes reliable in proportion to its distance from the original assertion. I mention these principles to bring out a point about evidence.
Testimony is often regarded as evidence, or as a source of evidence. Legal contexts assume (RIPD-T). In addition, philosophical accounts of testimony, like Robert Audi’s, associate testimony with non-inferential knowledge. One way to support such an account is by using an (RIPD) principle, namely, the closer you get to an experience, unimpeded by inferences that can go astray (probabilistically), the more veridical the testimony. Some accounts of evidence (like an account endorsed by Clayton Littlejohn here and in other places) is that p is part of your evidence-set if you non-inferentially know p. I wonder if these accounts of evidence implicitly endorse something like (RIPD) with regard to the strength of evidence. If so, the following principle is possible:
(RIPD-E) The strength of a piece of evidence is inversely proportional to the distance of that evidence from the experience (fact) that grounds that evidence.
Again, this seems to work in a legal context. If the fact is the evidence, as in the case of physical evidence presented in court, then that evidence is regarded (ceteris paribus) as stronger than a report about the physical evidence, which can leave out details and otherwise distort characterization of the evidence. But, an (RPD) principle is possible to formulate, even in a legal context. A jury may not know what they are looking at or what is relevant about a piece of physical evidence until it is explained. However, that explanation about the evidence may involve an interpretation of the evidence. Some interpretations of the evidence and its relevance to the guilt of the defendant may be more accurate than others. It is possible that given enough subsequent refinements of reports on the evidence, perhaps refinements that fill out the factual details, that the evidence becomes stronger in the eyes of the jury. In such a case, an evidence-specific (RPD) principle is possible:
(RPD-E) The reliability of a piece of evidence is proportional to the distance of that evidence (or propositional report on the evidence) from the original physical evidence, experience, event, or evidential proposition.
The proportional and inversely proportional principles pull in opposite directions. Philosophers whose account of testimony or evidence implicitly assume (RIPD) need to hang up the telephone (couldn’t resist) and sort out how an (RPD)-type principle can be explained away, undercut, or otherwise discharged. Otherwise, it seems reasonable to disconnect veridicality from closeness to the original bit of testimony or evidence.