The theory of pemphigus proposed by Stanley and Amagai is a truly dazzling intellectual accomplishment. The concept of the anti-desmoglein autoimmunity as the necessary and sufficient pathogenic event immediately provides a cause-and-effect scenario in which major adhesion proteins are blocked by the autoantibodies. Can a few minor inconsistencies (detailed in Refs 1, 2) undermine the otherwise perfect desmoglein autoimmunity concept? Before addressing this question directly, we would like to recall the old proof of K. Goedel and I. Lakatos that the probability that any scientific theory is true is very close to zero. The refutation is simply a matter of time and occurs when new observations cannot readily be explained by the current theory and requires ad hoc modifications, and when an alternative explanation that accounts for all older as well as new observations can be proposed. L. Wittgenstein used an elegant analogy comparing the theory to a ladder that enables us to reach a higher level of understanding. When this happens, it will serve us better to discard that ladder and find another one to lead us higher. Originally, the anti-desmoglein antibodies were thought to block keratinocyte cohesion by a simple steric hindrance within the desmosome. The available data, however, do not support this mechanism and suggest a much more complicated scenario, in which the pathogenic autoantibodies initiate various signalling processes ultimately leading to acantholysis and keratinocyte death. For instance, careful electron microscopic analysis of pemphigus lesions revealed that the loss of cell–cell adhesion occurs in the interdesmosomal membrane portions and desmosomal disruption is a rather late event. An important observation is that acantholytic keratinocytes die by apoptosis and oncosis (3, 4). This is rather counterintuitive, as normal keratinocytes do not undergo anoikis (cell death due to loss of adhesion) but rather terminal differentiation (5). Thus, the pathogenic antibodies must confer a death signal, but as yet it has not been demonstrated in any cell type that desmogleins operate as death receptors. There are also significant methodological issues, which have not yet been resolved. Furthermore, a truly satisfactory purification of human anti-desmoglein antibodies has not been achieved (2), which is a major obstacle in the investigations on the functional significance of these antibodies. To date, the best attempt to demonstrate that patients with pemphigus vulgaris produce pathogenic anti-desmoglein 3 antibodies was based on the phage display assay (6), a technique which is plagued by a number of problems, including the possibility of artificial autoantibody synthesis – even in non-immune subjects. Also, the in vitro models of acantholysis are rather disappointing and show a very low ability of autoantibodies to block cell cohesion in cell culture – in contrast to their strikingly high efficacy in vivo. Moreover, acantholysis can be produced in a number of ways, including antagonizing the cholinergic receptors, perturbation of intracellular calcium metabolism (for example, as seen in Hailey–Hailey or Darier disease), as a secondary event to cell death, or even by depletion of membrane cholesterol (R. Gniadecki, unpublished). We actually do not know for sure whether desmoglein blockade actually causes acantholysis, as the desmoglein knockout mice do not have pemphigus. There is no doubt that the desmoglein hypothesis is still perfectly suitable for standard textbooks. However, researchers in the field are well advised to seriously entertain and pursue alternative or complementary pathogenic mechanisms as well. Borrowing from juridical terminology, the anti-desmoglein antibodies are found at the scene of the crime, but whether they have the means or the motive to commit keratinocyte murder remains uncertain.