Answers are grounded in this paper and cite their sources.
This paper examines what kinds of evidence and explanation are needed to attribute intelligence and understanding to machines. It offers a historical-conceptual reconstruction of Turing’s operational approach to machine intelligence, computationalism, functionalism, Searle’s syntax–semantics challenge, and connectionism. Rather than presenting these positions as a linear sequence in which each replaces the last, the paper treats them as addressing distinct, related questions. It proposes a five-dimensional framework that distinguishes performance, computation, functional organisation, representation, and understanding. Applied to contemporary artificial intelligence, the framework clarifies why successful task performance, computational processing, or access to a model’s internal states does not by itself settle whether the system understands. It also avoids assuming that understanding must be non-computational: the question is what additional account connects a system’s capacities and organisation to semantic content.
No takes yet. Share an insight, caveat, or question.
Touraj Banirostam (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: