摘要：By bringing forward and interpreting the results from a collaborative research project that used contemporary computing techniques to investigate Giovanni Morelli’s nineteenth-century method for making stylistic attributions of old master paintings, this article examines how humanists make claims to knowledge and how this process may or may not be modellable or mechanizable within the context of classical, deterministic, digital computation. It begins with an explanation of the rationale behind choosing the Morellian practice of attribution, continues with a survey of another effort at computationally implementing Morelli’s method, and then presents our own computational techniques and results. The article concludes with what we have come to understand about the roles of responsibility, trust, and expertise in the social practice of art attribution, and the dangers in assuming that such human entailments are native to digital computers.
