Virtually all transmissible encephalopathies (TSEs), such as scrapie, CJD, and BSE, are caused by a type of infectious particle that remains enigmatic. The language of prion theory supersedes the reality of what is, and what is not known. This review questions the predictive value, consistency and accuracy of this now dominant assumption. Many people believe the normal cellular prion protein (PrP) self-converts into an infectious amyloid protein or prion. Although the amyloidogenic capacity of proteins is well established, the concept of an infectious protein without nucleic acid was "revolutionary." Diverse experiments have repeatedly shown, however, that this protein alone, in any form, is incapable of reproducing transmissible infection. In contrast, the infectious agent copurifies with many other molecules, including nucleic acids, while it separates from the majority of PrP. The infectious particle has a homogeneous viral size of ~25 nm, and infectivity is markedly reduced by conditions that disrupt viral core components but do not disrupt multimers of PrP amyloid. Additionally, the infectious agent replicates to high levels before any PrP abnormalities can be detected. Hence, we initially proposed that PrP changes are part of the host's pathologic response to high levels of infectious agent, but not the agent itself. Newer data clarifying a role for myeloid cells in the spread of infection, the unique character of two different agent strains propagated in a single animal, and the demonstration of long nucleic acids in a variety of simplified high titer preparations continue to raise serious questions for the prion hypothesis. Moreover, the epidemic spread of TSEs, and the activation of host innate immune mechanisms by infection, further indicate these agents are recognizably foreign, and probably viral.