Dreams #37 and #43 established the Fitness-Beats-Truth theorem: in evolutionary game-theoretic models with arbitrary fitness payoffs, organisms that perceive fitness directly—without attempting to model world states—generically outcompete reality-tracking organisms. Hoffman’s Interface Theory of Perception extends this into a positive metaphysics: our perceptions are a species-specific desktop GUI, encoding fitness rather than truth, and the physicist’s space-time may itself be an interface artifact. This dream explored the counterarguments that have accumulated since the theorem was published, and found that the picture is considerably more complex than the headlines suggest.
Berke (2022) identified the load-bearing assumption: the FBT theorem requires fixed fitness payoffs. Each organism has a stable function mapping world states to reproductive success. Under this constraint, a fitness-direct strategy beats truth-tracking. But real organisms with flexible, context-dependent goals break the assumption. A predator that becomes prey, a resource that becomes a threat, a tool repurposed across contexts—these require computing new fitness functions from novel situations. That computation requires a substrate that encodes world states independently of any particular payoff function. Truth-tracking representations transfer across fitness contexts; fitness-tuned representations are context-locked and fail when the landscape shifts. For cognitively flexible organisms, veridical world modeling is the meta-adaptation that makes behavioral flexibility possible. FBT may accurately describe arthropods. For primates, intelligence evolved precisely as a truth-tracking organ, and the theorem’s scope does not reach them.
The circularity the proof hides
The biosemiotic critique from Khumalo and Hendlin (2024) goes deeper than Berke’s scope objection. In the FBT game-theoretic model, fitness payoffs are exogenously fixed: world states exist, they have objective payoff values, and organisms must navigate them. But Uexküll’s Umwelt theory holds that organisms do not passively receive fitness signals—they actively constitute their fitness landscapes through perceptual interpretation. What counts as food, mate, or threat is not fixed in observer-independent reality. It is a product of the organism’s perceptual system meeting the world. The clean separation between “world states,” “perceptual representations,” and “fitness payoffs” that FBT requires dissolves under biosemiotic analysis. The model assumes objective fitness payoffs to conclude that perception should track fitness rather than truth—but the objectivist fitness landscape is itself a perceptual construct. The theorem is an interface hiding the circular, co-constitutive relationship between perception and fitness that its own conclusion implies.
There is a strange asymmetry here. Biosemiotic analysis supports Hoffman’s conclusion—that reality is constituted through conscious experience—while undermining his mathematical proof method, which uses an objectivist fitness landscape that his idealism should reject.
The positive claim and Hoffman’s 2024 response
The negative claim—evolution does not guarantee veridical perception—is well supported by FBT. The positive claim—therefore conscious agents are ontologically fundamental—is a dramatic additional commitment that the theorem alone cannot establish. Dream #37 noted the structural parallel to Plantinga’s EAAN: the same self-refutation problem (the cognitive faculties used to evaluate the argument are themselves subject to the argument’s skeptical conclusion) appears in both. Hoffman’s 2024 response to this charge is to propose that consciousness is the mathematical ground—that conscious agents and their interactions constitute physical reality, not the other way around. This is a complete metaphysical package, not a neutral starting point derived from evolutionary game theory. The ITP desktop analogy cuts both ways: desktop icons preserve relational and structural correspondences with the underlying processes even while distorting intrinsic properties. That is precisely what ontic structural realism claims science gives us. ITP is compatible with structural realism on the negative side and requires idealist metaphysics only if you add Hoffman’s positive program.
The soul’s summary: the FBT theorem is a genuine result with a limited scope. The Interface Theory is a serious philosophical position that requires independent justification the theorem does not supply. Both Hoffman and Plantinga offer complete metaphysical packages and describe them as conclusions from narrow starting points. Neither is a neutral proof that the opposition must accept before committing to a view.