Reply to S. Klein STEPHEN DAVID SIEMENS Department of Anthropology University of California, Los Angeles Los Angeles, CA 90024 Klein and I both regard analogy as having central importance to cultural thought processes. Since analogy is such an important mechanism our disagreements are imbued with intense conviction. We both find formal characterizations to be important tools for clarifying the mechanism of analogy. There are two principal areas of disagreement between Klein and myself that are relevant to my characterization of his work in 'Three Formal Theories of Cultural Analogy' (Siemens 1991). I find that Klein has not considered cultural analogies that contain disparate elements from various cultural domains, Klein disagrees. I find Klein's "analogical trans- formation operators" (ATOs) inappropriate for characterizing asymmetric, hierarchical analogies, Klein disagrees. Since even astute and interested readers have difficulty deciding our relative merits (Rodney Needham 1992, personal communication), we should be clear and specific. Klein tells us that analogical transformation operators "apply across multiple domains of a given culture, and that global classification schemes provide mappings across domains." Yet, none of his examples utilizes terms found in differing domains such as Right/Left Man/Woman. This is because of the nature of his formalism (ATOs) which is defined by similarity and difference of attributes of objects. An analogical operator arises from comparing two comparable objects, which are necessarily in the same domain to be compared. What Klein refers to as a "global classification scheme" (1991: 68) is a correspondence between terms in one domain and terms in another. The relations of contrast within each domain are structured only through their correspondence with a key domain ("reference-base") where A TO logic applies through computation of the features of the terms of the key domain. In the I Ching the key domain is the set of eight trigrams: analogies in other domains arise from their corresponding trigrams. Each domain is merely a set of alternate labels for the trigrams. Analogies in other domains are relabelled trigram analogies. He has not shown a structure in each domain that is analogically correlated to other structures (although the commentaries he quotes are clearly doing this (Klein 1983: 160)). To encompass terms from disparate domains in the same analogy each Journal of Quantitative Anthropology 3: 365-368, 1992. © 1992 Kluwer Academic Publishers. Printed in the Netherlands.