ifc-0086

6.1 A map by creative object

Table 6.1 gives a first orientation. The rows overlap: AI Feynman, for example, is simultaneously a search algorithm, a symbolic regression system, and a scientific-discovery benchmark. The classification marks its primary output and certificate rather than forcing each system into one exclusive school.

Tradition

Creative object

Generative mechanism

Typical certificate

Information and surprise

an unexpected description or repertoire

probabilistic or order-theoretic comparison

information gain, improbability, or surprise

Heuristic discovery

a concept, conjecture, law, or search path

heuristics over a represented problem space

usefulness, rediscovery, proof, or prediction

Computational creativity

an artifact and sometimes the producing process

generation, transformation, appraisal, and interaction

novelty, value, intentionality, and observer judgment

Neuroscience of creativity

a creative cognitive process realized in a brain

interacting neural systems supporting generation and evaluation

behavioral tasks, neuroimaging, stimulation, and convergent evidence

Symbolic regression

an equation explaining numerical data

numerical fitting plus symbolic decomposition and search

exact or predictive recovery with controlled complexity

Automated science

a hypothesis–experiment–revision cycle

active experimentation and machine interpretation

reproducible intervention and held-out prediction

Synthetic creativity

a typed construction or admitted extension relative to a maintained theory

typed composition, local probing, finite theory surgery, and transport

proof, countermodel, simulation, or independent evidence

Table 6.1 Traditions organized by their principal creative object. The final column is decisive: different objects require different admission certificates.

The final row does not subsume the others. Information-theoretic scores can guide search; computational-creativity criteria can assess process and presentation; symbolic regression can propose laws; and automated laboratories can test them. Synthetic creativity asks what additional structure is needed when the intended output is a reusable theory rather than a single artifact or formula.