Extended abstract:

Whether they concern crimes, political or military matters, or scientific discoveries, many facts are learned through the intermediation of individuals with special access to information, such as law enforcement officers, employees with a security clearance, auditors or special inspectors, or experts with specific knowledge.
For the public at large as well as for policymakers and other economic agents, there is often no way of learning about facts other than through these specific intermediaries, and even the "exogenous" signals often featured in economic models, perhaps an "accidental" discovery of evidence, must be reported by someone whose motives and discovery may potentially be questioned.
This paper considers whether societies and organizations can learn about such facts when evidence is i) costly to acquire, ii) cheap to fabricate or manipulate, and iii) produced through a sequential process in which each investigator observes the evidence produced by past investigators.
The answer turns out to depend on an asymptotic scarcity condition pertaining to the amount of evidence available about the fact. The condition distinguishes, for example, between reproducible scientific evidence and the evidence generated by a crime. When evidence is reproducible ad infinitum, as in the case of some natural sciences, facts can be learned with a precision that is only limited by the cost of acquisition relative to maximum level of incentives. When evidence is asymptotically scarce, however, there is no way of eliciting the truth from the intermediaries, no matter how numerous they are and no matter how their incentives are structured.
The talk will discuss several extensions and limitations of the result, as well as its various implications for the role of ethical behavior.