"A resemblance between the relations among things (rather than between the things themselves) (Adapted from J.W. SUTHERLAND, 1973, p.122)
K. KRIPPENDORFF states: "In the theory of models, an analogy between two systems is explained by one simpler system that models both, but by different homomorphisms" (1986, p.3) An analogy is thus a limiting conceptual constraint.
According to J.W. SUTHERLAND: "… the analogy is in the first and last instance a conceptual device designed to call attention to isomorphisms, and to engine the attempt to induce causality specific to a phenomenon by imposing a general causality" (p.123)
The first aspect does not seem objectionable. However, the second one could easily lead to dangerously sweeping generalizations by "embracing creeds" (POPPER, quoted by SUTHERLAND – p.117)
Further on, SUTHERLAND adds: "… analogies are properly used… as masks with which an a priory chaotic situation can be at least partially ordered, subsequent to empirical validation of the causality proposed by the mask. Grand operational principles then, in the hands of the serious scientist, simply become systemic, encompassing heuristics whose utility lies not in their reification but in their perspective".(p.127)