Space of Possible Minds

• Space of possible minds is a multidimensional framework that defines all cognitive systems—natural, artificial, and hybrid—using precise computational and geometric models.
• Research integrates formal methods from Turing machine theory, Hilbert space quantum models, and continuous Euclidean morphospaces to map and measure cognition.
• The study addresses challenges in metric formulation, structural topology, and ethical implications while guiding design strategies for advanced cognitive architectures.

The space of possible minds denotes the mathematical, computational, and organizational domains encompassing all conceivable cognitive systems—natural, artificial, hybrid, or entirely synthetic. This notion underpins analysis in cognitive science, artificial intelligence, philosophy of mind, and evolutionary computation. Modern approaches treat minds as instantiations of informational processing architectures, rigorously formalized through software-theoretic, geometric, algebraic, or multi-scale morphospace representations. Existing physical, biological, and engineered cognitive agents are seen as points or subsets within these vastly multidimensional spaces, whose full extent, topology, and structure remain largely unconstrained by evolved or designed exemplars.

1. Formal Models and Definitions of the Space

Multiple formalizations define what constitutes a "possible mind."

A. Software and Computational Space

Yampolskiy (Yampolskiy, 2014) explicates a programmatic view: the space of possible minds $M$ is the subset of all finite binary strings interpreted as programs on a universal Turing machine $U$ that meet a requisite notion of intelligence,

$M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$

where $I_U$ is the decision predicate for "mindhood."

This space is countably infinite ($|\mathbb{N}|$), and every element is representable as a finite-length code. Equivalent automata-theoretic representations (Moore/Mealy machines) are possible. Structural and behavioral equivalence, complexity class membership, and simulation relations induce further structure and orderings on $M$.

B. Hilbert Space Frameworks

Brody (Brody, 2023) employs the quantum formalism, modeling the space of (pure) mind-states as the complex projective manifold $\mathrm{CP}^{N-1}$, where $N$ may be large or infinite:

• Each mind-state is a normalized vector $|\psi\rangle$ in a complex Hilbert space $\mathcal{H}$; physical irrelevancies are quotiented, so the manifold of distinct minds becomes $U$0.
• Mixed states are density operators (convex combinations of pure states) occupying the closed convex hull of this manifold.

The geometry is non-Euclidean, typically equipped with the Fubini–Study (or induced Riemannian) metric, which encodes informational distances and angles between mind-states.

C. Continuous Mind-Spaces and Morphospaces

Levin (Levin, 2021) generalizes to a continuous $U$1-dimensional Euclidean space $U$2, where each axis $U$3 is a quantitative trait (e.g., behavioral, bioelectric, regulatory, spatial). An agent $U$4 is thus assigned a mind-vector

$U$5

with optional composite metrics:

$U$6

where $U$7 are weights. No categorical boundaries exist; the space is inherently graded and continuous.

Solé et al. (Solé et al., 19 Jan 2026) introduce non-metric, heuristic morphospaces where each dimension reflects organizational, computational, or interactional complexity in basal, neural, or hybrid systems.

2. Taxonomies, Axes, and Metrics of Mind-Space

Rigorous treatment of the space of minds mandates explicit axes and taxonomies.

A. Discrete and Algorithmic Taxonomies

• Architectural: Serial vs. parallel, fixed vs. reprogrammable, embodied vs. disembodied, etc. (Yampolskiy, 2014)
• Functional/Capability: Problem-solving class (P, NP, AI-Complete), social communication, degree of self-improvement, etc.
• Origin and Informational: Evolved, designed, uploaded, nested; blank-slate or preloaded knowledge.

Complexity-theoretic metrics (Kolmogorov complexity $U$8, minimal program length $U$9) quantify representation and behavioral intricacy.

B. Continuous Axes (TAME)

Levin (Levin, 2021) delineates empirically grounded axes such as:

• Morphological Regulative Complexity (MRC)
• Physiological Homeostasis Range (PHR)
• Developmental Bioelectric Competency (DBC)
• Transcriptional Feedback Adaptability (TFA)
• Behavioral Problem-Solving Capacity (BPC)
• Spatial Cognitive Boundary (SCB)
• Temporal Cognitive Horizon (TCH)
• Gap Junction Coupling Index (GJC)
• Persuadability Index (PI)
• Composite Multi-Scale Competency (MSC)

Each is operationalized by empirical or computational measures and admits scalable, aggregative metrics (e.g., weighted sums, geometric means).

C. Morphospaces of Cognition

Solé et al. (Solé et al., 19 Jan 2026) construct three morphospaces:

Morphospace	Axes	Notes
Basal (aneural) cognition	Spatial, computational, developmental complexity	Non-metric, heuristic placement
Neural cognition	Agency $M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$0, computational, agent-agent interaction	$M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$1 given by $M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$2
Human–AI hybrid cognition	Human complexity, machine complexity, exchange depth	Bandwidth and persistence of loop

Clusters and voids (unoccupied, potentially possible regions) are visually, not algorithmically, identified.

3. Structural, Geometric, and Dynamical Properties

A. Topologies and Metrics

• The discrete topology, Cantor-space topology (prefix tree basis), and simulation-induced quasi-orders are all applicable (Yampolskiy, 2014).
• In the quantum-Hilbert space view (Brody, 2023), the Fubini–Study metric governs informational distance.
• In continuous spaces (TAME), standard $M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$3 (Euclidean), weighted $M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$4, or custom aggregators serve as metrics.

B. Convexity and Dynamics

Quantum models embed mind-states in convex, compact sets of density operators; dynamics are described by stochastic differential equations where information acquisition is orthogonally projected (Lüders rule), minimizing uncertainty (analogous to free energy minimization) and recapitulating Bayesian updating in the commutative limit (Brody, 2023).

Dynamical systems and self-improvement maps $M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$5 induce orbits, attractors, and fixed points in mind-space—corresponding to evolutionary or designed optima.

C. Clusters, Voids, and Constraints

Clusters represent regions densely populated by existing biological or AI systems; voids are unpopulated regions, posited to exist due to evolutionary contingency, physical constraints, or design limits (Solé et al., 19 Jan 2026). Physical and chemical constraints (e.g., maximal size), as well as historical path dependency, define soft boundaries; precise inequalities are rare, but plausible constraints include channel and network connectivity and metabolic cost (Levin, 2021, Solé et al., 19 Jan 2026). Quantitative void volumes and cluster densities remain uncalculated.

4. Concrete Instantiation and Navigation of Mind-Spaces

A. Mapping Real and Synthetic Minds

TAME provides direct mappings of agents onto mind-space axes by normalized scoring (see Table below; axes abbreviations as in section 2B) (Levin, 2021):

Agent	MRC	PHR	DBC	TFA	BPC	SCB	TCH	GJC	PI
Planaria	1.0	0.6	0.8	0.7	0.1	0.01	1 d	0.7	0.3
Human (mammal)	0.2	1.0	1.0	1.0	1.0	10 km	10 y	0.2	0.9
Swarm robot	0.0	–	0.0	–	0.7	1 km	hr	0.0	0.6
Synthetic reorg.	0.7	0.5	0.8	0.6	0.3	0.05	3 d	0.6	0.4

Placement in previously unpopulated regions is possible via design or bioengineering—there is no unique "axis of humanity" (Levin, 2021).

B. Hybrid and Chimeric Cognition

Novel minds spanning biological and artificial substrates or blending neural, bioelectric, and software architectures occupy newly accessible volumes in mind-space (Levin, 2021, Solé et al., 19 Jan 2026). Tuning parameters such as gap junctional coupling ($M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$6), software architecture modularity, or interaction loop depth actively navigates toward target mind-space regions.

C. Composite Metrics and Evolutionary Search

Composite mind indices (additive, norms, geometric means) guide rational engineering and evolutionary trajectories. For instance, multi-scale homeostatic competency ($M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$7) scales with system size $M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$8 and elementary subunit IQ $M = \{ p \in \{0,1\}^* \mid I_U(p) = 1 \},$9 (Levin, 2021).

Evolution and rational engineering differ in trajectory: evolution acts as a gradient-free hill-climber potentiated by network connectivity, whereas human design explicitly computes metric gradients ($I_U$0parameter) for targeted optimization in mind-space.

5. Implications, Open Problems, and Research Directions

A. Limits and Decidability

No general algorithm decides $I_U$1 by Rice's theorem; only sound heuristics or resource-bounded approximations exist (Yampolskiy, 2014). Physical computation imposes upper bounds on system size, duration, and energy (cf. Lloyd's limits).

B. Measurement and Ethics

Substrate-independent intelligence metrics remain underdefined; TAME advocates for continuous, graded degrees of mindhood with direct ethical implications: any agent with nonzero competency should be accorded interest proportional to its mind-space coordinates (Levin, 2021).

C. Key Open Problems

• Statistical density of minds and attractor basins under self-modification and evolution (Yampolskiy, 2014).
• Characterization of topologies and proximity metrics that render nearby minds behaviorally or functionally similar.
• Formalization and discovery of voids, quantification of unmapped regions, and algorithmic search for realizable points in these spaces (Solé et al., 19 Jan 2026).
• Definitional, operational measurement of consciousness and free will as continuous (not binary) variables over $I_U$2 (Yampolskiy, 2014).

D. Prospects

Hybridization (biohybrid robots, brain–computer interfaces, chimeric tissue–AI systems) is a primary frontier in exploring unexplored cognitive regimes (Solé et al., 19 Jan 2026, Levin, 2021). The field of intellectology has been proposed for unified, systematic inquiry into $I_U$3 and its structural, algorithmic, and ethical properties (Yampolskiy, 2014).

6. Comparative Summary Table

Approach	Mathematical Foundation	Mind-Space Structure	Key Metrics/Axes
Yampolskiy (Yampolskiy, 2014)	Computability, program theory	$I_U$4; countable, discrete	Kolmogorov complexity, simulation relations
Brody (Brody, 2023)	Hilbert space, quantum theory	$I_U$5, convex hull; geometric	Fubini–Study distance, density operators
Levin (Levin, 2021)	$I_U$6-space, morphospace	Continuous, multidimensional, graded	MRC, DBC, GJC, BPC, composite indices
Solé et al. (Solé et al., 19 Jan 2026)	Morphospaces, heuristic mapping	Non-metric, qualitative, clustered & voids	Complexity, agency, exchange depth

Each formulation illuminates different invariants, dynamical mechanisms, and engineering levers within the universal space of possible minds. The diversity of such models reflects the scope, richness, and as-yet largely unmapped potential of mind-space as an object of scientific inquiry.