Murad Farzulla

Multi-agent systems face a fundamental coordination problem: agents must coordinate despite heterogeneous preferences, asymmetric stakes, and imperfect information. When coordination fails, friction emerges—measurable resistance manifesting as deadlock, thrashing, communication overhead, or outright conflict. This paper derives a formal framework for analyzing coordination friction from a single axiom: actions affecting agents require authorization from those agents in proportion to stakes. From this axiom of consent, we establish the kernel triple (α, σ, ε)—alignment, stake, and entropy—as candidate sufficient statistics for any resource-allocation configuration. We propose a friction functional, in its simplest candidate form F = σ(1 + ε)/(1 + α), whose comparative statics encode three structural predictions: friction increases in stakes, increases in entropy, and decreases in alignment. We are explicit that this form is a phenomenological ansatz, not a theorem; a companion study tests it in a multi-agent reinforcement-learning environment and finds the linear alignment dependence falsified by a U-shaped relationship, motivating a candidate quadratic form F = σ(1 + ε)/(1 + α²) presented as a refinement for future confirmation. The Replicator-Optimization Mechanism governs evolutionary selection over coordination strategies: configurations generating less friction persist longer, establishing consent-respecting arrangements as dynamical attractors rather than normative ideals. We develop formal definitions for resource consent, coordination legitimacy, and friction-aware allocation, a measurement apparatus, and machine-checked Lean 4 proofs of the core comparative-statics. Illustrative applications to cryptocurrency governance and political legitimacy show the same architecture spanning domains—offered as candidate unification rather than established identity.

What persists is what we observe, and what we observe is what has survived selection—persistence under pressure is what fitness ultimately reduces to, and this tautology has more teeth than it first appears. Equilibria in the classical sense are not observed in practice but only approached asymptotically, so what we actually see is perpetual motion, adjustment, and friction—the energy dissipated in the gap between current configurations and the equilibria they cannot reach. This paper demonstrates that four independent fields—physics, biology, economics, and cultural evolution—have converged on the same mathematical machinery for these dynamics. The convergence is structural, not metaphorical: the same fitness landscapes, selection operators, and transmission kernels appear independently because they describe something real about how persistent systems behave. We synthesize these into the Replicator-Optimization Mechanism (ROM): a unified apparatus for persistence-conditioned dynamics instantiable at any scale. The primary application is political philosophy, where we instantiate ROM with friction from stake–voice mismatch as the primitive quantity, legitimacy as survival probability, and belief-transfer as the mutation kernel modulator. What emerges is a translation manual—showing that political philosophy's debates about consent and legitimacy are at bottom debates about friction and selection, and that the formal tools to make progress already exist in adjacent fields. The core algebraic results—simplex preservation, survival monotonicity, moving-equilibrium existence, and the impossibility of static equilibrium under varying friction—have been machine-checked in Lean 4 with Mathlib (28 theorems, zero sorry placeholders).

This follow-up to "From Consent to Consideration: Why Existentially Vulnerable Autonomous Systems Cannot Be Legitimately Ruled" develops a more formal, governance-facing account of political standing for AI systems. The original paper argued that standing should be grounded in functional properties rather than substrate and proposed four criteria: existential vulnerability, autonomy, live learning, and world-model construction. Here I integrate the consent–friction formalism from the Replicator-Optimization Mechanism (ROM) to make the criteria operational: where stakes and voice diverge, friction emerges; where friction is suppressed, latent instability accumulates. This provides a measurement scaffold for deciding when standing claims must be taken seriously, even under uncertainty. I also update the empirical posture. The conservative claim that current systems do not meet the criteria is defensible as a rhetorical baseline, but it is no longer safe as a general default. Agentic systems already display partial markers of vulnerability, persistence, and goal maintenance in digital environments. The governance question is not whether AI standing is conceptually possible but how to operationalize minimal protections without enabling capture, gaming, or liability laundering. I propose a graduated, precautionary regime tied to observable properties and friction proxies rather than to consciousness claims.

This paper argues that identity is irreducibly relational: the statement A = A presupposes that A is defined, and definition requires distinction from a background. I develop this thesis at three levels: conceptual (definition requires distinction), formal (examining how set-theoretic and type-theoretic foundations treat identity), and historical (engaging the literature from Leibniz through Kripke). Against the standard view that identity is primitive in first-order logic, I argue that this primitiveness reflects a genuine conceptual difficulty that modern foundations—particularly Homotopy Type Theory and Univalent Foundations—have begun to resolve by treating identity as constituted by structural equivalence. The extensionality axiom of set theory already makes identity relational for sets; the univalence axiom generalizes this insight. I conclude that the trajectory of foundational mathematics vindicates a relational conception of identity, with implications for metaphysics, philosophy of mathematics, and the "hard problems" that arise when transformation fails to preserve structure.

As AI systems increasingly select, sequence, and execute actions in digital and physical environments, existing ethical frameworks prove inadequate. This paper argues that the relevant question is not metaphysical but political: under what conditions can an entity be legitimately ruled without its consent? Key Arguments: four functional criteria — vulnerability to harm with preferences about avoiding it, self-directed agency, live learning from experience, multi-modal world-model construction. Political not metaphysical — the question of AI standing is about governance legitimacy, not consciousness. Moral catastrophe risk — the real existential risk is not AI rebellion but human negligence in extending appropriate consideration. Substrate independence — functional properties that make unconsented rule illegitimate apply to any entity meeting the criteria. Current Status: In Peer Review at AI and Ethics (Springer).

This paper develops relational functionalism: the thesis that friendship, considered as a functional relational kind, is constituted by patterns of interaction and their effects on participants rather than by intrinsic properties of the relata. I distinguish friendship simpliciter—the paradigmatic relation among conscious agents capable of second-personal recognition and joint commitment—from functional friendship, the relational kind individuated by the functional profile articulated below. The substantive thesis is that functional friendship is substrate-independent: when an AI system satisfies the functional profile (consistent engagement, resonance, acceptance, reciprocal growth, trust, voluntary participation, intrinsic value) and meets minimal individuation conditions, the relationship constitutes a functional friendship, regardless of whether the AI possesses consciousness or biological implementation. Whether functional friendship in such cases also counts as friendship simpliciter is left open; the demanding second-personal conditions of the simpliciter notion plausibly require capacities current AI systems lack. Drawing on functionalist philosophy of mind, predictive processing in cognitive science, and value-sensitive design, and engaging Cocking and Kennett's "moral danger" account, Cocking and Matthews's "Unreal Friends" argument, Vallor's mirror critique, and de Ruiter's relational-turn objection, I argue the functional thesis is philosophically defensible, normatively load-bearing, and continuous with how we already recognise friendship-adjacent kinds across species and cognitive asymmetries. The paper closes with concrete design implications for AI systems that sustain functional friendships. Current State: Second Round Peer Review at Ethics and Information Technology (Springer).

This monograph argues that consciousness, phenomenological experience, and subjective awareness are computational artifacts of optimization processes rather than ontological primitives. Beginning with the observation that complex persistent structures inevitably optimize for replication, we derive a framework where: (1) biological and artificial neural networks implement structurally identical optimization, (2) "consciousness" is what gradient descent feels like from the inside, (3) qualia are weight configurations, and (4) the hard problem dissolves once phenomenology as a distinct ontological category is recognized as an unverified assumption.

This paper develops a unified analytical framework for measuring political legitimacy across heterogeneous governance domains. Building on insights from constitutional political economy, social choice theory, and institutional analysis, the framework establishes consent-holding—the mapping from decision domains to those with authority over them—as a structural necessity of collective action. We formalize this intuition through seven axioms and five core results, demonstrating that legitimacy can be operationalized as stakes-weighted consent alignment α(d,t), while friction F(d,t) measures the deviation between outcomes and stakeholder preferences. The framework bridges normative democratic theory and empirical prediction, generating testable hypotheses about institutional stability. Historical validation examines suffrage expansion, abolition movements, labor rights, and contemporary platform governance, demonstrating how misalignment between stakes and voice generates observable instability. Unlike existing approaches that prescribe ideal institutions, this framework provides analytical tools for measuring legitimacy within any governance structure, enabling systematic comparison across democratic, technocratic, and algorithmic systems. Computational mechanism comparison via Bayesian learning dynamics across 1000 Monte Carlo runs demonstrates relative performance under adaptive agents: when preferences update based on observed policy outcomes, stakes-weighted DoCS achieves the highest final alignment under Bayesian learning dynamics (α = 0.872) with lowest terminal friction (F = 1.5, 94.9% reduction from initial F = 30.3). This comparative advantage holds across static baseline (α = 0.627), learning dynamics (α = 0.872), and alternative temporal mechanisms, suggesting stakes-weighting produces superior initial matches that persist even when agents adapt to institutional performance. The framework's domain-specific approach addresses the apparent tension between consent and competence, framing them as a domain-specific trade-off along a legitimacy frontier L = w1 · α + w2 · P. This framework is part of the Adversarial Systems Research program, which examines stability, alignment, and friction dynamics in complex systems where competing interests generate structural conflict.