AIM Framework glossary
Data-Architecture Distinction
Canonical definition
The Data-Architecture Distinction is the methodological separation of empirical findings (observations that constrain any adequate theory) from theoretical architectures (frameworks that organise findings into explanatory structures), such that the same findings can support multiple architectures and rejection of an architecture does not require rejection of the findings the architecture was originally built around. The Distinction is the enabling condition for symmetric consilience and for the Mutual Architectural Correction Principle, both of which depend on the capacity of an integrating theory to accept the empirical record produced by a neighbouring discipline while treating that discipline's architecture as a separable, revisable object.
Explanation
Data and theory are not the same thing, even though they often arrive bundled together in published work. Data are what has been observed. Theory is the story the discipline tells about what the data mean. The Data-Architecture Distinction is the commitment to keep these two separate in one's own reasoning, so that accepting a finding from a neighbouring discipline does not automatically commit one to accepting the theoretical architecture the discipline built around the finding. A researcher may accept the entire replicated empirical record of behavioural economics while rejecting the assumption that preferences are exogenous. The Distinction makes this possible without breach of consilience, because the empirical record and the theoretical architecture are separable objects. Collapsing the Distinction, treating data and architecture as a single bundle, is the characteristic move of subordinate consilience, and is what prevents disciplines from exporting architectural corrections to their neighbours even when their own data would support such corrections.
Epistemic status
Framework-Specific Principle.