The expert (technē) analogy often plays an essential role in the arguments of Xenophon, Plato and Aristotle, and this type of argument can be traced back to Socrates. Yet there has been remarkably little work done on the argument itself. Vlastos, and to a lesser degree Robinson, interprets the majority of expert-analogies as intuitive inductions, where the conclusion is built into the concept of an expert and thus it is not an actual inference. On the other side McPherran, and to a lesser degree Santas, interprets them as probable inductions, i.e. an inference based on an insufficient number of cases or an insufficient number of similar attributes between the analogous cases, yielding a probable inference.
This thesis tries to defend a third alternative, where the expert-analogy is understood as an inference from one species to another species, the inference being valid as there is a common genus to which the attribute inferred belongs per se. Thus the analogy is interpreted to have a valid deductive structure. It is claimed that a similar analogical structure can be found in other types of proofs, e.g. the homological proof found in evolutionary biology. It is further argued that this structure can be found in Aristotle’s discussion of the argument by example (paradeigma), and further that a justification can be found in Aristotle’s four-part division of identity into that of quantity, species, genus and analogy – and it is claimed that the expert-analogy is in fact based on an identity in genus. Indications can also be found in Plato, but these were not developed further by him. And in addition, the Aristotelian principle that a proof should be at its most generic level further justifies the proposed structure of the expert-analogy. Finally this structure is used in the discussion of several controversial cases of the expert-analogy, hopefully showing that the proposed structure is applicable to the various cases and allows for an increased understanding of them.