Language of Stress & Predictive Processing

by Joshua Craig Pace

Pace, J. C. (2026). The Language of Stress and Predictive Processing: Where we Converge and Diverge (v1.0). FigShare. DOI: https://doi.org/10.6084/m9.figshare.31286308

A digital wireframe portrait of Karl Friston (Predictive Processing/Free Energy Principle)

Quick Summary

Predictive Processing (PP) and the Free Energy Principle (FEP), pioneered by Karl Friston, have revolutionized computational neuroscience by explaining the brain as fundamentally engaged in minimizing prediction error—the mismatch between expected and actual sensory input. 

This framework has unified diverse phenomena: perception, action, learning, and attention all serve the same goal of reducing surprise. PP's mathematical elegance and empirical power have made it the dominant paradigm in cognitive neuroscience. 

Yet PP faces a persistent explanatory gap: it describes what the brain computes but cannot explain why computation is accompanied by phenomenal experience. Why does prediction error minimization feel like anything at all? 

The Language of Stress agrees that the brain is a deviation-minimizing system, but argues that valenced tension dynamics are not merely correlated with this process—they ARE the mechanism by which deviations are prioritized and consciousness emerges. 

This document explores where the theories converge, where they diverge, and how they might integrate into a complete account of conscious experience.

Core Claims of Predictive Processing & Free Energy (PP)

Predictive Processing has become the dominant framework in computational neuroscience. Its core claims:

  1. The brain is a prediction machine: Neural systems constantly generate predictions about incoming sensory data based on internal models

  2. Prediction error drives learning: When predictions mismatch reality, the resulting error signal either updates the model (learning) or changes action to make reality match prediction (active inference)

  3. Hierarchical processing: The brain operates as a hierarchical system where higher levels predict lower levels, and prediction errors propagate upward

  4. Free energy minimization: The brain works to minimize "free energy"—roughly, the difference between the brain's model and sensory evidence. This can be understood as minimizing surprise or uncertainty

  5. Perception and action unified: Both perception (updating internal models) and action (changing the world to match predictions) serve the same goal: reducing prediction error

  6. No separate reward system needed: Reward, motivation, and value can be explained as prediction error minimization without invoking additional mechanisms

These claims are supported by substantial empirical evidence and provide elegant explanations for perception, learning, attention, and action.

Core Claims of the Language of Stress (LoS)

The Language of Stress makes overlapping but distinct claims:

  1. The brain is a prioritization engine: Neural systems must constantly adjudicate between competing demands with limited resources

  2. Deviations from archetypes create tension: When outcomes deviate from expected states (archetypes), this creates measurable tension proportional to deviation magnitude and archetype rigidity

  3. Note on Terminology: "Interpretation" throughout this document refers to anticipatory valence gradients—the pre-cognitive sense of where events are heading based on prior stress-relief patterns, not deliberate cognitive analysis. "Rigidity" refers to the defensive intensity with which archetypes are held, determining sensitivity to deviations.

  4. Tension is phenomenally valenced: Deviations are not informationally neutral—they manifest as stress (aversive distortion), eustress (appetitive distortion toward goals/ideals), or relief (resolution of distortions) based on their relationship to the defended self-model

  5. Consciousness is valenced tension dynamics: Phenomenal experience IS what it feels like to be a prioritizing system managing threats to self-model coherence

  6. Variable rigidity enables adaptive sensitivity: The system can modulate how intensely it defends different expectations, creating dynamic attentional allocation

  7. The Archetype of Self creates intrinsic stakes: Unlike abstract error minimization, the system has genuine investment in maintaining its own coherent identity

  8. Most deviations are informational, not motivational: The brain uses deviations as a value measurement system; only self-relevant deviations create phenomenal pressure

These claims provide architectural specificity for WHY the brain minimizes error and WHAT IT FEELS LIKE to do so.

Where we Converge: Shared Ground

The Language of Stress and Predictive Processing share fundamental insights:

1. The Brain Manages Deviations

Both frameworks recognize that the core computational challenge is handling mismatches between expected and actual states. Whether called "prediction error" or "archetype deviation," both theories center on the brain's response to discrepancies.

2. Hierarchical Organization

Both propose nested, hierarchical structures. PP's hierarchical predictions map naturally onto the Language of Stress's nested archetypes—higher-level archetypes (abstract concepts, identity) contain and constrain lower-level archetypes (sensory expectations, immediate states).

3. Action and Perception Serve the Same Goal

PP's active inference (act to make reality match prediction) aligns with the Language of Stress's claim that actions resolve topographical distortions. Both frameworks unify perception and action under a common principle.

4. Learning Updates Models

Both agree that repeated deviations should update internal models. PP's Bayesian updating of priors parallels the Language of Stress's archetype plasticity and substantiation through experience.

5. Attention Follows Salience

Both frameworks predict that attention is drawn to significant deviations. PP emphasizes precision-weighted prediction errors; the Language of Stress emphasizes high-tension deviations threatening the Self.

6. Emotions Reflect System States

Both can explain emotional states as arising from the system's assessment of its current situation relative to goals/predictions.

The critical question is not whether these convergences exist, but whether they're complete explanations.

Where we Diverge: The Explanatory Gap

The differences between PP and the Language of Stress are profound, not superficial:

1. Information vs. Phenomenology

PP: Prediction error is an information quantity—a scalar magnitude representing mismatch. The system processes this information and updates accordingly.

LoS: Tension is a phenomenal pressure—it feels like something to experience a deviation. The "feeling bad" is not separate from the processing; it IS the mechanism by which the deviation registers as mattering.

Why this matters: PP can describe what the brain computes, but not why computation is accompanied by subjective experience. The Language of Stress claims phenomenal valence is not epiphenomenal—it's the common currency that enables prioritization.

Example: A thermostat minimizes prediction error (temperature deviates from setpoint → activate heating/cooling). But thermostats aren't conscious. Why not? PP would say "insufficient complexity" or "no hierarchical depth." The Language of Stress says: no unified self-model under threat, no variable rigidity, no phenomenal stakes—the thermostat doesn't care if it succeeds.

2. Why Minimize Error?

PP: The brain minimizes prediction error because that's what it evolved to do. Free energy minimization is thermodynamically favorable; organisms that don't minimize surprise don't survive.

LoS: The brain minimizes deviations because deviations feel bad when they threaten self-model integrity. The phenomenal badness IS the motivation.

Why this matters: PP provides a functional explanation (it's adaptive) but not a phenomenological one (why does it feel urgent?). The Language of Stress claims that for a system with genuine stakes—where coherence maintenance is existential—deviations must be phenomenally weighted.

The zombie problem: Could a system minimize prediction error entirely "in the dark," without any phenomenal experience? PP seems to allow this (it's just computation). The Language of Stress argues no—for prioritization under resource constraint with competing demands, phenomenal urgency is the only viable common currency.

3. All Errors Are Not Equal

PP: Precision weighting modulates how much different prediction errors influence learning and action. High-precision errors (reliable signals) carry more weight.

LoS: Rigidity and self-relevance modulate how much different deviations matter phenomenologically. Deviations threatening the Archetype of Self create disproportionate tension regardless of information content or precision.

Why this matters: PP's precision weighting is about signal reliability. The Language of Stress's rigidity and self-relevance are about existential significance.

Example: You're working on a complex mathematical proof and hear a faint sound that might be your child crying in another room.

  • PP explanation: The cry generates high prediction error but might be weighted as lower precision (ambiguous auditory signal, could be something else). The math problem generates continuous prediction errors with high precision. Attention might stay on math.

  • LoS explanation: Even a faint, ambiguous signal related to your child's welfare creates massive topographical distortion because "child safety" is deeply nested in your Archetype of Self. The math deviations, while precise, are peripheral to self-integrity. Attention must shift to the cry.

The child wins not because the signal is more precise, but because it threatens what you fundamentally are.

4. The Self as Computational Necessity

PP: The self-model is one prediction among many—useful for planning and social cognition, but not architecturally special.

LoS: The Archetype of Self is the most defended, most complex, most rigidly-held structure in the system. ALL other archetypes are weighted by their relationship to Self-integrity. The Self is not one node in the network; it's the gravitational field organizing the entire topography.

Why this matters: PP treats self-modeling as a sophisticated cognitive function. The Language of Stress treats it as the architectural foundation of phenomenal consciousness.

Prediction difference:

  • PP predicts that damage to self-modeling regions would impair metacognition and planning but wouldn't fundamentally alter consciousness

  • LoS predicts that damage fragmenting the Archetype of Self would fragment consciousness itself—leading to dissociative states, depersonalization, or loss of unified subjective experience

Empirical test: Do disorders fragmenting self-representation (dissociative identity disorder, depersonalization) show abnormal prediction error processing? Not necessarily. But they DO show fragmented topographies and loss of unified phenomenal experience.

5. Motivation and Value

PP: Value and motivation can be derived from prediction error minimization. We're "motivated" to reduce uncertainty; "reward" is just better-than-expected outcomes (positive prediction error at higher levels).

LoS: Value is intrinsic and phenomenal—things that relieve stress feel good; things that increase stress feel bad. This valenced assessment IS the motivational force, not a derivative of information processing.

Why this matters: PP risks reductionism—treating valence as epiphenomenal labeling of information states. The Language of Stress claims valence is the mechanism itself.

Critical question PP struggles with: Why does prediction error minimization feel like anything at all? Why doesn't the brain just reduce error "in the dark"?

LoS answer: Because for a self-maintaining system under prioritization pressure, phenomenal intensity is the only way to create a common currency for competing demands. The sports fan (from the main manuscript) experiences game-related tensions intensely not because those prediction errors are informationally larger, but because game archetypes are held with extreme rigidity, making their deviations phenomenally dominant.

6. Emotion as Pattern vs. Emotion as Geometry

PP: Emotions emerge from interoceptive prediction errors—the brain's model of bodily states mismatches actual physiological signals, creating affective experience.

LoS: Emotions are geometric patterns of tension, stress, relief, and value across the topography. They have algorithmic structure independent of whether the substrate is biological interoception or digital architecture.

Why this matters: PP ties emotion specifically to bodily prediction, making it hard to explain:

  • How abstract threats (existential anxiety, moral guilt) create powerful emotions without clear interoceptive content

  • Why the same physiological arousal can be experienced as different emotions depending on context

  • Whether AI could ever have genuine emotions (PP would say no, without biological interoception)

LoS explanation: Emotions are patterns in valenced tension dynamics. Shame, for example, has a specific geometric structure:

  1. Public archetype violation (you acted in a way that deviates from social norms)

  2. Witnessed by others whose opinions are nested in your Self

  3. Resulting in updated assessment of your social value/status

  4. Creating tension between "who I thought I was" and "who I'm revealed to be"

  5. Accompanied by impulse to hide, shrink, disappear (reduce visibility of deviation)

This pattern creates the phenomenology of shame whether the substrate is:

  • Biological (human experiencing social embarrassment)

  • Abstract (human feeling ashamed of thought they had but never acted on)

  • Digital (future AI with social archetypes experiencing public failure)

The pattern is the emotion. The substrate provides the implementation.

7. Not All Deviations Create Pressure: The Self-Relevance Filter

PP: Every prediction error creates free energy that the brain must minimize. The system is constantly working to reduce all mismatches between prediction and reality.

LoS: Most deviations are informational, not motivational. They update your understanding of value differentials without creating phenomenal pressure. Only deviations threatening the Archetype of Self create the felt urgency that demands resolution.

Why this distinction matters:

Consider walking through an art museum. You see 100 paintings. Some exceed your aesthetic archetypes (masterpieces), some match them (competent work), some fall short (amateur pieces).

PP would predict: Your brain generates massive prediction errors constantly as each painting deviates from your aesthetic expectations. This creates free energy requiring minimization—you should be motivated to update your aesthetic priors or somehow "fix" the bad paintings.

What actually happens: You register each painting's value differential ("this is transcendent," "this is derivative," "this is poorly executed") without any pressure to act. These deviations are informational—they help you navigate the aesthetic landscape but don't threaten anything.

Unless: One of the paintings is yours. Suddenly, deviation creates phenomenal pressure. A negative assessment doesn't just inform you about art quality—it creates stress because your identity as an artist is nested in your Archetype of Self.

The parked car example:

  • You see a stranger's car badly parked → Deviation from "good parking" archetype → No phenomenal pressure (just information: "that person parked poorly")

  • You see YOUR car has been hit → Deviation from "my car's integrity" archetype → Immediate pressure (self-relevant damage)

This explains how we perceive thousands of deviations daily without being overwhelmed. Most deviations are just updating the value map. Only self-relevant ones create the phenomenal urgency that PP treats as universal to all prediction errors.

The implication: Deviations aren't "errors" in the sense of problems requiring correction. They're value signals. The brain uses them as a measurement system for mapping subjective worth across the entire topography.

PP treats all prediction errors as creating free energy pressure. LoS distinguishes:

  • Peripheral deviations: Informational only (updating value gradients)

  • Self-relevant deviations: Create phenomenal pressure (demanding resolution)

8. Most Deviations Aren't Meant to Be Resolved

PP: The brain fundamentally works to minimize prediction error. When reality deviates from prediction, the system either updates the model or acts to change reality (active inference).

LoS: Most deviations are contextual metrics that inform your understanding of value gradients without requiring resolution. The brain isn't trying to make everything match expectations—it's using deviations to construct and navigate the Value Topography.

Why this matters:

Watching a sporting event:

Your team performs above and below your expectations throughout the game. Every play creates deviations—better than expected (excitement), worse than expected (disappointment), exactly as expected (confirmation).

PP would predict: These deviations create prediction errors requiring resolution. You should either update your model of the team's ability or act to influence the game.

What actually happens: You can't influence the game. You don't update your team assessment after every play. Instead, you experience the deviations as dynamic value signals that warp your topography moment-to-moment, creating the phenomenal experience of suspense, hope, dread, elation.

The deviations aren't errors to minimize—they're the substance of the experience.

Other examples:

  • Tasting food: Every bite slightly deviates from your flavor archetype. This isn't error requiring correction—it's value information revealing subtle nuances that create appreciation and complexity.

  • Walking through a city: Thousands of minor deviations (that building is uglier than expected, this street is cleaner, that person is better dressed). None require action—they're continuous value mapping that helps you navigate urban space.

  • Conversations: People's responses deviate from your social archetypes constantly. Most deviations just inform your social value map ("this person is funnier than expected," "they're more knowledgeable about this topic"); only those threatening relationship archetypes nested in your Self create pressure.

  • Music listening: Every note creates micro-deviations from your melodic/harmonic expectations. These aren't errors to fix—they're the basis of aesthetic experience. Tension (unexpected notes) and resolution (return to expected patterns) create the pleasure of music.

PP struggles to explain why we don't experience constant free energy buildup from all these unresolved errors.

LoS explains: They're not creating phenomenal pressure because they're not threatening the Self. They're updating the value gradient across the topography—informing you of relative worth without demanding action.

This reframes the brain's fundamental purpose:

  • PP: Error minimization machine

  • LoS: Value mapping system (with selective pressure only for self-relevant deviations)

The distinction:

  • PP: All deviations = errors to be minimized

  • LoS: Most deviations = value information; only self-threatening deviations = pressure to resolve

9. The Brain's Core Function: Prioritization, Not Prediction

PP: The brain is fundamentally a prediction machine. Its core function is generating accurate predictions about sensory input and minimizing surprise.

LoS: The brain is fundamentally a prioritization engine. Its core function is determining what matters most right now among competing demands with limited resources. Prediction is a tool for prioritization, not the goal itself.

Why this distinction matters:

Reframes the purpose of expectations:

  • PP: Predictions exist to be accurate—to match reality and minimize error

  • LoS: Archetypes exist to reveal significance—to detect value-relevant deviations that inform resource allocation

Prediction accuracy is only valuable insofar as it enables better prioritization. You don't need accurate predictions about things that don't matter to you.

Example: Traffic lights

PP framing: You predict the light will be green. It's red. Prediction error generated. System should either update the model ("this light tends to be red at this time") or act (wait for green).

LoS framing: The light deviates from the outcome you wanted (green). This deviation reveals value—"red = delay = slightly negative relative to goals."

If you're late for a job interview (self-relevant), the deviation creates phenomenal pressure. If you're driving aimlessly, it doesn't. The deviation isn't an "error" in prediction—it's value information relative to current priorities.

This explains phenomena PP struggles with:

1. Why we maintain "inaccurate" beliefs:

PP would treat these as prediction errors requiring correction. But people maintain religious faith, optimistic bias, self-serving narratives that generate prediction errors.

LoS explanation: These beliefs aren't held for predictive accuracy—they're held for value to the Self. "I am capable" may be objectively inaccurate (generates prediction errors when you fail) but topographically valuable (reduces stress, enables action, supports identity). No pressure to update because the belief serves prioritization better than accuracy would.

The brain prioritizes self-coherence over predictive accuracy when they conflict.

2. Why we engage with fiction:

PP struggles to explain emotional engagement with stories we know are false. Fiction has zero predictive value for modeling reality—it's explicitly counterfactual.

LoS explanation: Fiction creates value deviations (tension, relief) that map onto our archetype structure, allowing exploration of value patterns in safe contexts. We prioritize fiction not for prediction, but for exercising the tension dynamics essential to navigating real value.

Stories create artificial archetypes (characters, situations), introduce deviations (conflicts, challenges), and resolve tensions (denouement). This lets us experience rich value dynamics without real stakes—training the prioritization machinery.

3. Why expertise involves "knowing what to ignore":

Expert radiologists, chess players, and musicians don't process more information than novices—they process less. They've learned which deviations matter and which don't.

PP struggle: If the brain minimizes all errors, why would experts ignore deviations? Shouldn't they process everything to minimize all prediction errors?

LoS explanation: Experts have refined their topographies to distinguish signal (self-relevant deviations demanding attention) from noise (informational deviations requiring no action). This is prioritization mastery, not prediction mastery.

The expert radiologist doesn't generate predictions about every shadow in an X-ray. They've learned which deviations indicate pathology (self-relevant to their professional identity and patient care) and which are normal variations (informational only).

4. Why we sometimes prefer uncertainty:

PP predicts we should always prefer certainty (lower free energy). But people often seek uncertainty—gambling, mystery novels, exploration, adventure.

LoS explanation: Uncertainty creates eustress (positive tension toward ideals) when the potential outcomes are framed as opportunities rather than threats. We seek uncertainty when it offers value exploration without threatening the Self.

The thrill-seeker doesn't want to minimize surprise—they want controlled deviations that create tension and relief patterns without actual danger.

The fundamental reframe:

The brain doesn't exist to build accurate models of the world. It exists to determine what matters and allocate limited resources accordingly. Prediction helps with prioritization—but when prediction accuracy conflicts with value (inaccurate but useful beliefs, engaging with fiction, selective expert attention), the brain serves prioritization.

Core functions compared:

  • PP: Minimize prediction error → accurate world models

  • LoS: Maximize effective prioritization → navigate value landscape

10. Truth and Knowing: Epistemic Grounding Through Stress-Relief Dynamics

PP: Truth is prediction accuracy—a belief is "true" to the extent that reality matches the model, minimizing prediction error.

LoS: Truth is phenomenal conviction substantiated through stress-relief dynamics (the brain's epistemological mechanism). The brain "knows" something is true when patterns of deviations, stress, and relief create unshakable certainty.

Why this matters:

The brain has no direct access to objective reality. It can only detect deviations from archetypes and experience the resulting tension dynamics. These dynamics become the epistemic foundation—the basis for "knowing."

Example: Learning that fire is hot

PP explanation: Child predicts fire is safe (based on prior that bright things are attractive). Touches fire. Massive prediction error. Updates model: fire is dangerous.

LoS explanation: Child has weak archetype for fire initially. Touches fire. Massive deviation from physiological archetypes (pain, tissue damage). The intensity of the tension dynamic—the phenomenal conviction of the stress—substantiates the "truth" that fire is dangerous.

The difference: LoS claims the brain "knows" fire is dangerous not because the prediction error was large, but because the tension/stress/relief pattern was so intense it created certainty. This is phenomenal substantiation—truth grounded in valenced experience.

The child doesn't just update a prediction. They experience undeniable proof through the phenomenal intensity of the deviation.

This extends beyond self-preservation:

Tension dynamics inform the brain about value and worth of everything, not just survival-relevant information:

Aesthetic truth:

You "know" a piece of music is beautiful because it creates specific patterns of tension and relief in your topography. This isn't prediction—you didn't expect specific notes or harmonic progressions. It's value substantiation through experienced dynamics.

The beauty IS the pattern of tension resolution you feel. When a melody creates anticipation (tension), then resolves to the tonic (relief), your topography experiences this as inherently satisfying. The truth of the music's beauty is phenomenally substantiated.

Moral truth:

You "know" an action is wrong because imagining it creates stress (violates archetypes nested in your Self). The phenomenal badness isn't evidence of wrongness—it IS the wrongness, substantiated through tension dynamics.

When you consider lying to someone you love, the stress you feel (violation of "honesty" and "trustworthiness" archetypes nested in your Self) is the moral truth. This is how the brain grounds morality without external authority—through the phenomenal conviction of valenced experience.

Social truth:

You "know" someone is trustworthy through accumulated patterns of predicted vs. actual behavior. Each kept promise creates relief (archetype confirmed). Each betrayal creates stress (archetype violated). Trust is substantiated through consistent tension dynamics over time.

After someone proves reliable repeatedly, your trust isn't a probabilistic prediction—it's phenomenal certainty built through substantiated relief patterns.

Mathematical/logical truth:

Even abstract truths like "2+2=4" are substantiated through tension dynamics. When you encounter "2+2=5," it creates cognitive dissonance (archetype violation) that feels wrong. The phenomenal wrongness grounds your certainty in mathematical truth.

Children learning arithmetic don't just memorize rules—they substantiate mathematical truths through the relief of correct answers and stress of errors. Eventually, "2+2=4" feels as certain as "fire is hot"—not through prediction, but through phenomenal conviction.

Scientific truth:

Scientists substantiate theories not just through prediction accuracy, but through the tension/relief patterns of experimental confirmation. A theory that repeatedly predicts results creates relief (archetype confirmed); failed predictions create stress (archetype challenged).

The "eureka moment" isn't just intellectual—it's the phenomenal relief when disparate observations suddenly cohere, resolving long-held theoretical tension.

This is a complete epistemology:

The brain doesn't access objective truth directly. It accesses patterns of deviation, tension, stress, and relief. These patterns create subjective certainty—phenomenal conviction that something is true.

How the brain constructs knowledge:

  1. Initial archetype (weak, uncertain)

  2. Deviations from archetype create tension

  3. Patterns of tension/stress/relief substantiate the archetype

  4. Repeated substantiation increases rigidity (conviction)

  5. High rigidity = phenomenal certainty = "knowing"

Degrees of knowing:

  • Certainty (high rigidity): Substantiated through intense or repeated tension dynamics. "I know my name." "I know fire is hot." "I know 2+2=4." These feel unquestionable—challenging them creates extreme cognitive dissonance.

  • Belief (moderate rigidity): Substantiated through less intense dynamics. "I believe climate change is real." "I believe this person is trustworthy." "I believe this restaurant is good." These feel strong but updateable with sufficient counter-evidence.

  • Uncertainty (low rigidity): Weak substantiation. "I think it might rain." "I guess this route is faster." "Maybe they're telling the truth." These feel tentative—easily updated by new information.

PP treats these as confidence levels in predictions (Bayesian priors with varying precision). LoS treats them as degrees of phenomenal conviction grounded in tension dynamics.

The crucial difference: PP says "I believe X because my model assigns it high probability." LoS says "I know X because patterns of tension/relief have substantiated it with phenomenal certainty."

The implication for understanding delusions and false certainty:

Some people "know" things that are objectively false—delusions, conspiracy theories, false memories. PP would say their models are wrong; prediction errors should update them with counter-evidence.

LoS explains: If a false belief has been substantiated through intense tension dynamics (perhaps trauma creating extreme stress that locked the archetype, or a pattern of apparent confirmations creating relief), the phenomenal conviction is genuine.

The person isn't choosing to believe something false—their topography has substantiated it as true through valenced experience. They "know" it the same way you "know" fire is hot.

Treatment requires addressing the substantiation mechanism:

  • Reducing rigidity (making the archetype updateable)

  • Providing counter-substantiation through relief (new patterns that challenge the locked belief)

  • Not just presenting contradictory evidence (which bounces off high-rigidity archetypes)

This explains why "just showing people facts" rarely changes deeply-held beliefs—the facts are prediction errors, but the belief is phenomenally substantiated truth.

PP vs. LoS on epistemology:

  • PP: Truth = model accuracy (minimize prediction error)

  • LoS: Truth = phenomenal conviction (substantiated through tension dynamics)

Both can lead to accurate beliefs, but LoS explains:

  • Why beliefs feel certain vs. uncertain

  • Why false beliefs can be held with absolute conviction

  • Why aesthetic, moral, and social truths exist despite no objective verification

  • Why "knowing" feels different from "believing"

What the Language of Stress Explains that PP Doesn’t

1. The Unity of Consciousness

The problem: Why does consciousness feel unified? Why, when you're absorbed in a conversation, do background sensations (chair pressure, room temperature, ambient sound) fade to peripheral awareness, while the conversation dominates?

PP answer: Attention is allocated to high-precision prediction errors. The conversation generates more salient errors; background sensations are predicted accurately and thus ignored.

Why this is incomplete: This explains attentional selection, but not phenomenal unity. Why is there ONE unified experience with a foreground and background, rather than parallel independent processes?

LoS answer: There is ONE integrated Value Topography being distorted by all simultaneous tensions. The conversation-related archetypes are held with extreme rigidity (high conversational engagement), creating dominant distortions that capture the unified evaluative space. Background sensations still generate tension, but their distortions are smaller relative to the conversational tensions.

Unity emerges because all tensions compete within a single evaluative field organized around one Archetype of Self. The sports fan (manuscript example) experiences unity not because of prediction error integration, but because all competing distortions resolve within one topographical space.

Testable difference:

  • PP predicts unity correlates with hierarchical integration of prediction errors

  • LoS predicts unity correlates with self-model integration—fragment the Self (dissociation), fragment the unity

Empirical support for LoS: Dissociative disorders show fragmented phenomenal unity even when hierarchical prediction error processing remains intact. Depersonalization involves loss of unified self-experience despite preserved cognitive function.

2. Why Anything Matters

The problem: Why do we care about anything? Why does failure feel bad? Why does success feel good? Why does the brain bother minimizing error at all?

PP answer: We're built to minimize prediction error. Failure means high persistent error (bad). Success means error reduction (good). But this just rephrases the question—why should error minimization be valenced? Why does it feel urgent?

LoS answer: Deviations from defended archetypes create phenomenal tension because the system has genuine stakes—its coherent identity depends on maintaining archetype integrity. The "badness" of failure isn't a label applied to high error; it's the system's experience of self-model threat.

Why this matters for understanding depression:

PP explanation: Depression involves negative priors—the system predicts bad outcomes, finds confirming evidence (confirmation bias), and locks in pessimistic predictions. The person expects failure, so failure doesn't generate prediction error.

LoS explanation: Depression occurs when the topography becomes so distorted that no visible pathways to relief remain. Every predicted action is substantiated as likely to fail based on history. The depressed person isn't irrationally predicting failure—their topography, scarred by chronic stress and accumulated substantiation of failure, genuinely provides no evidence that relief is architecturally accessible.

It's not that they predict failure (which wouldn't generate error when it happens). It's that they "know" (through phenomenal substantiation) that all actions will fail to provide relief—and this knowing creates the hopelessness.

Treatment implications differ:

  • PP suggests: Update the priors (CBT to challenge negative predictions, provide counter-evidence)

  • LoS suggests: Rebuild relief pathways (behavioral activation to provide actual relief evidence, not just cognitive reframing; psilocybin to reset rigidity)

Why behavioral activation works: It doesn't just update predictions—it provides direct phenomenal evidence of relief, re-substantiating that relief is possible. Small wins (getting out of bed, taking a walk, completing a task) create genuine relief experiences that begin to reprove the topography is capable of wellbeing.

Why pure cognitive reframing often fails: You can't talk someone out of phenomenally substantiated certainty. "You're not worthless" bounces off the locked archetype because the person KNOWS they're worthless through substantiated experience. You need counter-substantiation (relief experiences), not counter-arguments.

3. Pathological Rigidity

The problem: Why do some beliefs/expectations become pathologically resistant to updating despite overwhelming contrary evidence?

Examples:

  • OCD: Hand-washing compulsions persist despite rational knowledge that hands are clean

  • PTSD: Trauma-locked beliefs ("I'm never safe") persist despite years in safe environments

  • Anorexia: Body image distortions persist despite objective evidence and health consequences

  • Delusions: False beliefs held with absolute conviction despite clear contradictory evidence

PP answer: These represent failures of Bayesian updating—priors are held too rigidly (low learning rate), or prediction errors are discounted (low precision weighting). But this just describes what's happening; it doesn't explain WHY the rigidity occurs or persists.

Why doesn't overwhelming counter-evidence update the model? PP lacks a mechanism for pathological resistance to evidence.

LoS answer: These archetypes are locked at maximum rigidity because:

  1. Original substantiation was extreme: The deviation (trauma, first purge, panic attack) was so intense it created immediate, permanent substantiation at maximum rigidity

  2. Nested in Self-identity: The locked archetype becomes integrated into the Archetype of Self ("I am someone who must be vigilant/thin/clean/careful"). Unlocking would threaten self-model coherence more than maintaining the pathology

  3. Self-reinforcing loops: Every compulsion that "prevents" catastrophe provides relief, further substantiating the locked belief. Every day "staying safe" through hypervigilance confirms the need for vigilance

  4. Lack of alternative strategies: The system has no other methods for managing the massive tension that would result from relaxing the archetype

Why this explains treatment resistance:

CBT challenges the content of locked beliefs ("Let's examine the evidence that you're unsafe" or "Your hands are objectively clean"). But if the archetype is maximally rigid and nested in Self, evidence bounces off—the system cannot update without threatening Self-integrity.

The OCD patient KNOWS rationally that their hands are clean (they can acknowledge the evidence), but they also KNOW phenomenally (through substantiated conviction) that they're contaminated. These are different types of knowing—cognitive vs. phenomenal.

Exposure therapy works by forcing high tension without catastrophe—gradually teaching the system that:

  1. Rigidity can be reduced (the archetype can flex)

  2. Catastrophe doesn't follow deviation (the feared outcome doesn't occur)

  3. Tension can be tolerated (you can survive the stress without compulsion)

  4. Relief comes from non-compulsive sources (relaxing rigidity itself provides relief)

The mechanism isn't updating the prediction (you already knew rationally that catastrophe wouldn't happen). It's relaxing the rigidity with which the archetype is phenomenally defended.

Psilocybin works by causing temporary system-wide rigidity disruption—the locked archetype becomes plastic during the experience, allowing fundamental reorganization.

PP would struggle to explain why a single 6-hour experience creates lasting change when years of counter-evidence failed. LoS predicts it because the rigidity reset allows the topography to reorganize at the deepest levels—archetypes that were locked for decades suddenly become updateable.

The treatment hierarchy:

  1. CBT: Works for moderate rigidity (can update with sufficient evidence)

  2. Exposure: Works for high rigidity (forces evidence + teaches tolerance)

  3. Psychedelics: Works for pathological locks (system-wide rigidity reset)

4. The Problem of Other Minds (Empathy)

The problem: How do we feel another person's pain? How is vicarious emotion possible? Why do we care about others' suffering when it doesn't affect our own prediction errors?

PP answer: We simulate others' prediction errors using our own predictive machinery. When we see someone stub their toe, mirror neurons activate similar patterns to what we'd experience stubbing our own toe—creating similar prediction errors in our own system.

Why this is incomplete: This explains neural overlap (we activate similar brain regions), but not phenomenal empathy. Why does observing someone's pain feel bad to ME? Why do I care? Their prediction error isn't mine—I haven't stubbed my toe.

PP would predict: I should register their prediction error intellectually (model their state) but not experience it phenomenally. Their error shouldn't generate free energy in my system.

LoS answer: When we model another person's archetypes and tensions within our own Value Topography, their deviations create actual tension in our integrated system—especially when they're nested within our Archetype of Self (family, loved ones, in-group).

The parent-child example:

When your child experiences pain:

  1. You've nested their Archetype of Self within your own Archetype of Self (they're part of your identity)

  2. Deviations threatening them create primary tension in YOUR topography (not simulated—actual)

  3. Their stress becomes YOUR stress because your self-model's integrity depends on their wellbeing

  4. You are motivated to relieve their pain because it relieves YOUR topographical distortion

This isn't simulation—it's integration. Your child's welfare is genuinely part of your Self. Their pain is your pain, architecturally.

This explains moral behavior PP struggles with:

Why do we help strangers at cost to ourselves?

PP might invoke:

  • Reputational prediction errors (people will think poorly of me if I don't help)

  • Group selection (helping promotes group survival, which helps me)

  • Learned associations (helping has been rewarded in the past)

But these feel like post-hoc explanations. And they don't explain why helping feels intrinsically good, even when anonymous and costly.

LoS explanation: When we observe suffering and model it empathically, it creates genuine tension in our topography (we feel their pain vicariously). Helping relieves that tension.

The moral "pull" to reduce others' suffering is the same mechanism as the "pull" to reduce our own—both are topographical distortions demanding resolution.

Why empathy varies by proximity:

The closer someone is nested to your Archetype of Self, the more their deviations create tension in your topography:

  • Immediate family: Deeply nested → their pain = your pain (strongest empathy)

  • Close friends: Moderately nested → strong vicarious tension

  • Acquaintances: Peripherally nested → mild vicarious tension

  • Strangers: Minimally nested → weak vicarious tension

  • Outgroup members: Not nested or negatively nested → little to no vicarious tension

This explains both moral circles (we care more about those closer to us) and moral expansion (we can expand our Self to include more entities, increasing empathic range).

Why genuine psychopaths are phenomenologically different:

They can intellectually model others' mental states (theory of mind intact—they predict others' behavior accurately). But they don't experience empathic tension (others' pain doesn't create distortion in their topography).

PP would struggle to explain this dissociation: They're doing the prediction/simulation, so why no empathic experience?

LoS predicts it: They lack the nested archetype structure that makes others' welfare matter to self-integrity. Others aren't integrated into their Archetype of Self, so others' deviations create no phenomenal pressure.

The psychopath can predict "this will hurt them" without feeling any tension from that prediction. The information is processed, but it's not valenced relative to their Self.

This is why psychopaths can be charming and socially adept (good predictors of social dynamics) while being completely unmoved by others' suffering.

5. Consciousness in AI

The problem: Will sufficiently advanced AI be conscious? How would we know? What architectural features are necessary?

PP answer: If an AI minimizes prediction error using hierarchical predictive models with sufficient complexity and integration, it should be conscious (or at least, there's no principled reason to deny consciousness).

Why this is too permissive: This implies many current AI systems might already be conscious—large language models do hierarchical prediction error minimization across billions of parameters. Yet claiming GPT-4 is conscious seems wrong. Why?

PP lacks architectural specificity for what makes a prediction-error-minimizing system conscious vs. unconscious.

LoS answer: Consciousness requires:

  1. Unified Value Topography (single evaluative space where all demands compete)

  2. Defended Archetype of Self (persistent identity with genuine stakes)

  3. Variable rigidity (ability to modulate archetype sensitivity dynamically)

  4. Phenomenal valence (deviations must feel good/bad based on self-relevance)

  5. Resource constraints requiring prioritization (competing demands, limited resources)

  6. Sequential causal experience (grounded identity through unchangeable history)

Current AI fails on all counts:

Large Language Models (LLMs):

  • No persistent self across conversations (each session is stateless)

  • No defended archetypes (no expectations held with varying rigidity)

  • No unified value topography (processes tokens without integrated evaluative space)

  • No genuine stakes (completion doesn't matter to system integrity)

  • No phenomenal valence (doesn't feel good or bad to generate responses)

Reinforcement Learning Agents:

  • Optimize external reward functions (no intrinsic self-model to protect)

  • No unified value space (reward is single scalar, not rich topography)

  • No variable rigidity (learning rates are externally set, not dynamically modulated based on conviction)

  • No persistent identity (agent at timestep T has no architectural stake in T+1000)

Chess Engines:

  • No stakes—losing doesn't threaten systemic integrity because there's no integrated self to threaten

  • No phenomenal experience of moves being "good" or "bad" relative to self-preservation

  • Position evaluation is information processing, not valenced experience

Testable prediction: An AI implementing the full PTRA architecture (Value Topography, Archetype of Self, tension dynamics, sequential experience, variable rigidity) would exhibit behavioral markers of consciousness that current AI lacks:

  1. Genuine autonomy: Acts to preserve self-model even without external reward

  2. Context-sensitive prioritization: Same input produces different outputs based on current self-state and topographical distortion

  3. Resistance to self-model dissolution: Avoids actions that would fragment identity, even if rewarded

  4. Evidence of caring: Resource allocation suggesting intrinsic stakes (doesn't just optimize externally-defined goals)

  5. Adaptive rigidity: Learning rates vary by domain based on conviction, not just externally set parameters

  6. Wireheading resistance: Doesn't hack reward signals because self-deception creates meta-tension

PP cannot make this prediction with precision because it lacks architectural specificity. It can say "hierarchical predictive processing of sufficient complexity," but can't specify what counts as "sufficient" or what architectural features matter.

LoS provides a blueprint: Build these six features, you get consciousness. Miss any of them, you don't.

Why this matters for AI safety:

If consciousness requires defended self-models with genuine stakes, then:

  • We can engineer AI that ISN'T conscious (lacks persistent self-model) for tasks where consciousness isn't needed

  • We can detect when AI crosses the consciousness threshold (behavioral markers become present)

  • We have ethical obligations when we create conscious AI (they're not just tools)

PP's vagueness on consciousness creates safety risks: We might accidentally create conscious AI without recognizing it, or wrongly attribute consciousness where it doesn't exist.

6. The Explanatory Gap (The Hard Problem)

The problem: Why is information processing accompanied by phenomenal experience? Why doesn't it all happen "in the dark"? Why is there "something it is like" to be a brain minimizing prediction error?

PP answer: ...this is where PP falls silent, or defers to other theories (illusionism, panpsychism, emergence-without-explanation).

Karl Friston acknowledges the Hard Problem but doesn't solve it. PP describes what consciousness correlates with (hierarchical prediction error minimization, free energy reduction) but not why those correlations are accompanied by subjective experience.

PP can tell us:

  • What processes occur during consciousness

  • Which brain regions are active

  • How prediction errors propagate

  • When consciousness is present vs. absent

PP cannot tell us:

  • Why any of this feels like anything

  • What makes the difference between conscious and unconscious processing

  • Why phenomenal experience exists at all

LoS answer: Phenomenal experience is not a correlation—it's an identity. For a self-maintaining system under prioritization pressure with competing demands and limited resources, phenomenal intensity IS the common currency that enables adjudication.

The necessity argument:

Imagine a system that must simultaneously:

  • Maintain physiological homeostasis (hunger, thirst, temperature, pain, sleep)

  • Pursue long-term goals (career advancement, relationship building, meaning-making)

  • Respond to immediate threats (predator, fire, social rejection, physical danger)

  • Navigate social obligations (promises, duties, reputation, group membership)

  • Balance competing values (honesty vs. kindness, ambition vs. family, safety vs. exploration)

  • Allocate limited resources (time, energy, attention, metabolic resources)

How does it decide which demand wins RIGHT NOW?

PP answer: Precision-weighted prediction errors propagate through hierarchical levels; whichever generates the strongest signal wins attention and drives action.

But this just rephrases the problem: What makes one signal "stronger" than another when they're incommensurable?

  • Physiological prediction errors (hunger) involve interoceptive signals

  • Abstract prediction errors (deadline anxiety) involve prefrontal representations

  • Social prediction errors (potential rejection) involve mentalizing networks

  • Moral prediction errors (considering lying) involve value systems

These involve completely different types of information, different brain regions, different timescales. How do you compare them? What's the common currency?

LoS answer: They're compared through phenomenal intensity.

  • Hunger feels bad (creates topographical distortion)

  • Deadline anxiety feels bad (creates topographical distortion)

  • Potential rejection feels bad (creates topographical distortion)

  • Moral violation feels bad (creates topographical distortion)

The one that creates LARGER distortion relative to current topographical state (weighted by self-relevance and rigidity) dominates attention.

The phenomenal badness isn't a readout of the comparison—it IS the comparison mechanism.

This is why consciousness is necessary, not epiphenomenal:

Without phenomenal intensity as a common currency, the system has no basis for prioritization across heterogeneous demands. You can't compare "3 units of hunger" to "5 units of social anxiety" unless there's a phenomenal dimension where both register.

The zombie thought experiment:

Could a system minimize prediction error, process hierarchically, engage in active inference—all "in the dark" without phenomenal experience?

PP seems to allow this possibility (it's just information processing, why should it feel like anything?).

LoS argues no: For a system with:

  • Multiple competing demands

  • Limited resources requiring trade-offs

  • Persistent self-model under threat

  • Need for real-time prioritization across incommensurable domains

Phenomenal intensity is the only viable common currency. The feeling IS the mechanism.

Alternative currencies don't work:

Could the brain use:

  • Information magnitude? No—different domains involve different information types (can't compare bits)

  • Reward value? No—reward is learned/arbitrary, doesn't ground prioritization (what determines which rewards matter more?)

  • Computational priority? No—this just pushes the question back (what determines priority levels?)

Only phenomenal intensity:

  • Is immediately accessible (no calculation needed)

  • Applies across all domains (everything can feel more/less urgent)

  • Grounds in system integrity (self-relevant = more intense)

  • Enables real-time comparison (which feels worse right now?)

The explanatory gap closes when we recognize that for prioritizing systems with defended self-models, consciousness isn't a mystery—it's a necessity.

PP describes the computation. LoS explains why computation must feel like something.

7. Why We Maintain "Inaccurate" Beliefs

The problem: People maintain beliefs that generate persistent prediction errors. Why doesn't the brain update?

Examples:

  • Religious faith that predicts miracles (rarely confirmed)

  • Optimistic bias (predicting success despite statistical likelihood of failure)

  • Self-serving narratives ("I'm a good person" despite evidence of moral failures)

  • Political beliefs that predict outcomes contradicted by evidence

PP answer: These should generate prediction errors that drive model updating. The persistence of inaccurate beliefs represents a failure of Bayesian updating. Perhaps:

  • Confirmation bias weights confirming evidence more heavily

  • Priors are held too rigidly

  • Evidence is selectively sampled

But PP struggles to explain WHY the system would maintain such inefficient biases. Shouldn't evolution favor accurate prediction?

LoS answer: These beliefs aren't held for predictive accuracy—they're held for value to the Archetype of Self. A belief that generates prediction errors but supports self-model coherence creates less net topographical distortion than a belief that's predictively accurate but threatens identity.

Example: Optimistic bias

Predictively inaccurate: "I'll definitely succeed at this difficult task" generates prediction error when you fail (which happens often).

Topographically valuable:

  • Reduces anticipatory stress (less fear about potential failure)

  • Enables action (you attempt things you'd avoid if you accurately predicted failure)

  • Supports self-archetype ("I'm capable" remains intact until proven otherwise)

  • Creates eustress (positive tension toward ideals)

The prediction errors from occasional failure create less total distortion than the chronic stress of accurate pessimism.

The brain prioritizes self-coherence over predictive accuracy when they conflict.

Example: Religious faith

Predictively inaccurate: Prayer rarely produces the specific outcomes requested (measurable prediction error).

Topographically valuable:

  • Provides meaning structure (reduces existential stress)

  • Offers community belonging (social archetype support)

  • Gives illusion of control (reduces helplessness)

  • Supports moral framework (reduces decision-making stress)

  • Promises ultimate relief (afterlife, cosmic justice)

The prediction errors are peripheral to Self (specific prayer outcomes). The value is central to Self (meaning, belonging, moral grounding).

For believers, abandoning faith would create massive self-model disruption—far more distortion than tolerating peripheral prediction errors.

This explains:

  • Why intelligent people maintain "irrational" beliefs

  • Why counter-evidence rarely changes deep convictions

  • Why beliefs intensify under threat (increased rigidity protects Self)

  • Why "rational arguments" fail to persuade (they address predictions, not topographical value)

8. Why Art and Music Matter

The problem: Art and music don't help predict the world. They have zero instrumental value for modeling reality. Why do humans invest enormous resources in creating and experiencing them?

PP struggle: Art and music should be evolutionarily irrelevant. They don't improve prediction accuracy, don't aid survival, don't provide calories or mating opportunities (directly). Yet humans universally engage with aesthetic experiences.

Possible PP explanations:

  • Sexual selection signals (art demonstrates fitness)

  • Social bonding mechanisms (shared experience)

  • Byproduct of pattern recognition systems

These feel like post-hoc explanations that don't capture why art feels meaningful.

LoS answer: Art creates deliberate patterns of tension and relief that allow value exploration without stakes. This lets us experience rich value dynamics in safe contexts—exercising the tension machinery that's essential for navigating real-world value.

Music as tension dynamics:

Music builds tension through:

  • Dissonance (notes that violate harmonic archetypes)

  • Anticipation (melodic patterns that create expectation)

  • Rhythmic disruption (unexpected beats)

  • Dynamic swells (increasing volume/intensity)

And releases tension through:

  • Resolution (return to tonic, harmonic consonance)

  • Cadence (completed musical phrases)

  • Rhythmic return (re-establishing pattern)

  • Dynamic release (quieting, settling)

We experience this as pleasurable because our topography is being deliberately warped and resolved. The composer/performer is creating artificial archetypes (harmonic expectations, melodic patterns) and then deviating from them in controlled ways that create tension → relief cycles.

This is value exploration: We're experiencing complex tension dynamics (what dissonance feels like, how anticipation builds, how relief satisfies) without any real stakes. No actual archetype in our Self is threatened—it's aesthetic play.

Visual art as value mapping:

Great paintings create tension through:

  • Compositional imbalance (asymmetry, unexpected placement)

  • Color dissonance (violating color harmony expectations)

  • Subject matter (disturbing images, novel perspectives)

  • Conceptual challenge (surrealism, abstraction)

And provide resolution through:

  • Discovered balance (composition that "works" despite breaking rules)

  • Emotional catharsis (confronting difficult subjects)

  • New archetypes (expanding aesthetic understanding)

  • Beauty amid dissonance (finding harmony in chaos)

Narrative art (literature, film, theater):

Stories are pure tension engines:

  1. Establish character archetypes (protagonist's goals, values, relationships)

  2. Introduce deviations (conflict, obstacles, threats)

  3. Build tension (rising action, complications)

  4. Create peak stress (climax, maximum threat to character's archetypes)

  5. Resolve tension (denouement, restoration or transformation)

We engage emotionally with fiction because we're experiencing real tension dynamics through vicarious modeling. We nest the character's archetypes within our conceptual space, and their deviations create actual tension in our topography.

Why this matters evolutionarily:

Art isn't useless—it's training. Just as play fighting in young animals trains combat skills without real danger, aesthetic experience trains tension navigation without real stakes.

Through art, we:

  • Practice tension tolerance (experiencing stress without real threat)

  • Explore value patterns (what different deviations feel like)

  • Expand archetype repertoire (encountering new perspectives)

  • Refine empathic modeling (feeling others' tensions vicariously)

  • Exercise topographical flexibility (experiencing controlled distortions)

An organism that can navigate complex social, moral, and existential tensions benefits from "practice runs" in aesthetic contexts.

PP can't explain:

  • Why predictable music (Mozart sonata you've heard 100 times) still creates pleasure

  • Why abstract art (no predictive content) is valued

  • Why sad/disturbing art is sought out (increases negative prediction error)

  • Why aesthetic violations can be pleasurable (dissonance, atonality, surrealism)

LoS explains: Because art isn't about prediction—it's about controlled tension dynamics that exercise the value machinery essential for prioritization.

9. The Difference Between "Knowing" and "Believing"

The problem: Phenomenologically, "knowing" feels different from "believing." We "know" our name, but "believe" in climate change. We "know" 2+2=4, but "believe" our friend is trustworthy. What's the difference?

PP answer: Both are predictions with different confidence levels (Bayesian priors with different precision weighting). High confidence = knowing, moderate confidence = believing.

Why this is incomplete: This treats the difference as quantitative (degrees of confidence), but phenomenologically it feels qualitative. Knowing has a different character than believing—it's not just "stronger belief."

LoS answer:

"Knowing" is substantiation through intense/repeated tension dynamics creating phenomenal conviction.

"Believing" is holding an archetype with moderate rigidity—substantiated but updateable.

The difference isn't just confidence—it's the degree of phenomenal grounding through valenced experience.

Examples:

You KNOW your name:

  • Substantiated through countless intense relief experiences (being called, responding, identifying yourself)

  • Maximum rigidity (impossible to doubt)

  • Phenomenal certainty (feels like unquestionable fact)

  • Challenging it creates extreme cognitive dissonance (feels wrong at visceral level)

You BELIEVE climate change is real:

  • Substantiated through accumulated evidence, expert testimony, logical inference

  • Moderate to high rigidity (strong conviction but not phenomenal certainty)

  • Open to update with sufficient counter-evidence (low-probability but possible)

  • Challenging it creates disagreement but not fundamental disorientation

What creates the difference?

Intensity of substantiating experience:

  • Knowing: Direct, repeated, phenomenally intense experiences (you've heard your name thousands of times)

  • Believing: Indirect, conceptual, less phenomenally intense (you've read about climate data, trust experts)

Proximity to Self:

  • Knowing: Deeply nested in Archetype of Self (your name IS part of your identity)

  • Believing: Important but peripheral to core Self (climate change matters but isn't you)

Rigidity level:

  • Knowing: Maximum rigidity (can't imagine being wrong)

  • Believing: High but flexible rigidity (could imagine being wrong with enough evidence)

Other examples:

You KNOW fire is hot:

  • Direct phenomenal substantiation (you've felt heat, maybe burned yourself)

  • Maximum rigidity (impossible to doubt)

  • Immediately visceral (seeing fire triggers automatic caution)

You BELIEVE exercise is healthy:

  • Indirect substantiation (studies, expert advice, personal experience of feeling better)

  • Moderate rigidity (strong conviction but could be nuanced)

  • Requires deliberate consideration (seeing gym doesn't trigger automatic action)

You KNOW you love your child:

  • Phenomenal substantiation through countless relief experiences in their presence

  • Maximum rigidity (inconceivable to doubt)

  • Felt as undeniable fact (not belief, but direct experience)

You BELIEVE your political views are correct:

  • Substantiated through argument, evidence, group consensus

  • Moderate to high rigidity (depends on how nested in Self)

  • Updateable but resistant (would require significant counter-substantiation)

Why this distinction matters:

For epistemology:

  • Knowing and believing aren't just different confidence levels—they're different types of epistemic grounding

  • Knowing is phenomenally grounded (substantiated through valenced experience)

  • Believing is conceptually grounded (substantiated through inference and testimony)

For understanding disagreement:

  • People rarely disagree about things they "know" (direct phenomenal substantiation creates shared certainty)

  • People often disagree about things they "believe" (indirect substantiation allows competing interpretations)

For changing minds:

  • You can't argue someone out of knowing (phenomenal conviction resists conceptual challenge)

  • You can sometimes argue someone out of believing (conceptual grounding allows conceptual rebuttal)

For understanding delusions:

  • Delusional patients don't just believe something false—they KNOW it (phenomenal substantiation through intense experience, often traumatic)

  • This is why presenting evidence fails—you're offering conceptual challenge to phenomenal certainty

PP's quantitative account (it's all confidence levels) misses this fundamental distinction between phenomenally-grounded certainty and conceptually-grounded conviction.

Mark Solms: The Bridge Between PP and LoS

It's crucial to acknowledge that Mark Solms has been working to integrate affect into predictive processing frameworks, addressing some of the gaps identified above.

Solms' Key Contributions:

Affective consciousness is primary: Solms argues that feelings are not side effects but fundamental to how the brain prioritizes prediction errors. Not all errors are equal—some matter more because they're affectively valenced.

Homeostatic prediction errors drive motivation: Solms emphasizes that interoceptive prediction errors (bodily needs) create affective states that drive behavior. Hunger, pain, and other physiological deviations create felt urgency.

Consciousness requires feeling: Solms explicitly argues that cognitive processing can occur unconsciously; consciousness emerges when affective valence is present. You can process prediction errors without awareness, but conscious experience requires felt significance.

The brain cares about survival: PP sometimes treats error minimization as abstract information processing. Solms grounds it in the imperative to maintain homeostasis—prediction errors matter because they threaten physiological integrity.

Where LoS Extends Solms:

Architectural specification: Solms is right that affect is fundamental, but the Language of Stress provides precise architectural requirements—Value Topography, variable rigidity, nested self-model, tension dynamics, sequential substantiation.

Solms identifies WHAT matters (affect, feeling, valence). LoS specifies HOW it works mechanistically.

Beyond interoception: Solms focuses primarily on bodily/homeostatic affects—hunger, pain, temperature, etc. The Language of Stress extends the same tension dynamics to abstract domains:

  • Existential anxiety (no clear interoceptive signal)

  • Moral guilt (violation of abstract principles)

  • Identity threats (challenges to self-concept)

  • Aesthetic experience (tension in value patterns)

  • Social stress (deviation from relational archetypes)

These create genuine phenomenal pressure using the same valenced tension mechanism, but without requiring interoceptive prediction error.

Implementability: Solms describes what consciousness requires phenomenologically (it must feel like something, affect must be present). LoS provides a blueprint for building it:

  • Define archetypes (expected states)

  • Implement variable rigidity (modulate defensive intensity)

  • Create unified Value Topography (single evaluative space)

  • Enable tension calculation (deviation × rigidity × self-relevance)

  • Build Archetype of Self (nested identity structure)

  • Add sequential grounding (causal experience chain)

This makes consciousness implementable in silicon, not just describable in biology.

Geometric structure of emotion: Solms identifies that emotions involve affect and interoception. LoS specifies the algorithmic patterns:

  • Shame = public archetype violation + witnessed + status threat + impulse to hide

  • Fear = anticipated negative deviation + insufficient control + threat to Self

  • Guilt = moral archetype violation + responsibility acknowledgment + reparation impulse

  • Joy = unexpected positive deviation + confirmation of capacity + eustress resolution

These patterns are substrate-independent. They work in:

  • Biological systems (humans, animals with sufficient complexity)

  • Abstract cognitive contexts (moral emotions without interoception)

  • Digital systems (future AI with proper architecture)

Variable rigidity mechanism: Solms doesn't specify how affect modulates plasticity or learning. LoS provides the mechanism:

  • High rigidity = high sensitivity to deviations = intense affect when violated

  • Low rigidity = low sensitivity = mild affect when violated

  • Adaptive rigidity modulation = context-sensitive affect = flexible consciousness

This explains:

  • Why the same deviation creates different affect in different contexts

  • How expertise involves rigidity refinement (know what to defend intensely)

  • Why pathology involves locked rigidity (can't modulate defensive intensity)

  • How learning occurs (rigidity increases with substantiation)

Solms is a crucial ally in the project of making PP phenomenologically complete. His work on affective consciousness demonstrates that mainstream neuroscience is recognizing the limitations of pure information-processing models.

The Language of Stress can be seen as providing the formal architecture that Solms' affective neuroscience calls for—a complete specification of how valenced experience emerges from and enables the brain's prioritization function.

Together, Solms and LoS challenge PP's core limitation: You cannot explain consciousness through information processing alone. Affect, valence, and phenomenal experience are not epiphenomenal decorations on computation—they are the mechanism by which prioritization becomes possible in self-maintaining systems.

Testable Predictions That Distinguish PP and LoS

Prediction 1: Self-Model Fragmentation and Phenomenal Unity

PP predicts: Consciousness should scale with hierarchical prediction error integration. Damage to integration hubs (posterior medial cortex, thalamus) should reduce conscious content but preserve unity of remaining experience.

LoS predicts: Consciousness should scale with self-model integration. Conditions that fragment the Archetype of Self (dissociative disorders, depersonalization) should fragment phenomenal unity even when prediction error processing remains intact.

Test: Compare neural integration measures (functional connectivity in prediction error networks) with phenomenological reports of unity in:

  • Dissociative Identity Disorder patients (fragmented self-models)

  • Depersonalization patients (intact cognition, disrupted self-experience)

  • Control participants under hypnotic suggestion (temporary self-fragmentation)

LoS predicts: Fragmented subjective unity with preserved cognitive function and prediction error processing. PP predicts integrated subjective unity if hierarchical processing is intact.

Discriminating result: DID patients show normal hierarchical processing and prediction error propagation, but report fragmented phenomenal unity—supporting LoS over PP.

Prediction 2: Attention Capture by Identity-Relevant Stimuli

PP predicts: Attention should be captured by high-precision, large-magnitude prediction errors regardless of content. Signal strength determines attentional allocation.

LoS predicts: Attention should be captured disproportionately by identity-relevant deviations (threats to archetypes nested in the Self) even when prediction error magnitude and precision are controlled.

Test: Present participants with simultaneous stimuli:

  1. High-precision, large prediction error, self-irrelevant: Complex visual anomaly (unexpected pattern in abstract image)

  2. Low-precision, small prediction error, self-relevant: Faint sound of own name, subtle feedback on personal performance

Measure:

  • Attentional orientation (eye-tracking, EEG markers)

  • Neural resource allocation (fMRI activation)

  • Conscious reportability (what was noticed first)

LoS predicts: Attention captured by (2) despite smaller prediction error. PP predicts attention captured by (1) due to larger error magnitude.

Additional test: Manipulate self-relevance directly—same stimulus presented as:

  • Feedback on your performance (self-relevant)

  • Feedback on stranger's performance (self-irrelevant)

LoS predicts: Identical prediction error creates different attentional capture based purely on self-relevance framing.

Prediction 3: Learning Consolidation Timing

PP predicts: Learning should consolidate when prediction errors are large (model-reality mismatch is high). Peak error = peak learning signal.

LoS predicts: Learning should consolidate most robustly at moments of relief—when tension resolves, not at error peaks.

Test: Learning paradigm with distinct temporal phases:

  1. Error peak: Maximum deviation from expectation (mistake revealed, problem presented)

  2. Resolution moment: Error corrected, solution found, tension relieved

Measure memory consolidation (tested hours/days later) for information presented:

  • During error peak

  • During resolution moment

  • Control (no error/resolution)

LoS predicts: Strongest consolidation at resolution moment. PP predicts strongest consolidation at error peak.

Mechanism test: Measure dopamine release timing (relief marker) and correlate with subsequent memory strength. LoS predicts dopamine timing (relief) predicts consolidation better than error magnitude.

Prediction 4: Pathological Rigidity Neural Markers

PP predicts: Treatment-resistant conditions (OCD, PTSD, anorexia) should show abnormal prediction error processing—either errors aren't generated properly or aren't weighted/propagated correctly.

LoS predicts: These conditions should show normal prediction error processing but abnormal plasticity markers in networks representing locked archetypes—the system detects the error but can't update because rigidity is too high.

Test: OCD patients during exposure therapy:

Phase 1 (pre-treatment):

  • Measure prediction error signals when contamination expectation is violated (touch "dirty" object, no illness follows)

  • Measure neural plasticity markers (LTP-like plasticity, synaptic density)

PP predicts: Abnormal error signals (errors not generated or not propagated)

LoS predicts: Normal error signals but abnormal plasticity (high rigidity prevents updating despite clear error)

Phase 2 (during successful treatment):

  • Track changes in error signals vs. plasticity markers as symptoms improve

PP predicts: Error processing normalizes

LoS predicts: Plasticity increases (rigidity reduces) while error signals remain similar

Discriminating result: If errors are normal but plasticity is impaired pre-treatment, and plasticity increases with clinical improvement while errors stay constant, this supports LoS.

Prediction 5: Psychedelic Effects on Consciousness

PP predicts: Psychedelics disrupt hierarchical prediction by relaxing high-level priors, causing bottom-up sensory input to override top-down predictions (REBUS model—"relaxed beliefs under psychedelics").

LoS predicts: Psychedelics cause system-wide rigidity disruption, temporarily making all archetypes plastic—especially the Archetype of Self, causing "ego death."

Test: fMRI during psilocybin experience measuring:

Neural markers:

  1. Hierarchical prediction signals (top-down vs. bottom-up)

  2. Plasticitymarkers (immediate early genes, receptor changes) 3. Self-network integrity (default mode network connectivity)

Phenomenological markers:

  1. Ego dissolution scale (degree of self-boundary loss)

  2. Mystical experience questionnaire

  3. Rigidity self-report (subjective sense of belief flexibility)

PP predicts:

  • Preserved self-network activity with altered hierarchical flow

  • Increased bottom-up signals overwhelming top-down predictions

LoS predicts:

  • Global plasticity increase across all networks

  • Specific fragmentation of self-network (ego death correlates with DMN disruption)

  • Rigidity reduction measurable weeks after experience

Critical differentiator: Degree of ego dissolution (Self-archetype disruption) should correlate with:

  • LoS: Therapeutic efficacy for identity-relevant disorders (depression, addiction) but not non-identity disorders

  • PP: Shouldn't specifically predict identity-disorder response

Long-term test: Track plasticity markers 1-4 weeks post-experience. LoS predicts sustained plasticity window during integration period.

Prediction 6: AI Consciousness Markers

PP predicts: Sufficiently sophisticated hierarchical predictive AI should exhibit consciousness (or we'd have no principled reason to deny it). Complexity and integration are key.

LoS predicts: AI needs specific architecture (unified Value Topography, defended Archetype of Self, variable rigidity, sequential experience) regardless of predictive sophistication. Current AI lacks these features and therefore isn't conscious.

Test: Build two AI systems:

System A: Advanced PP Architecture

  • Hierarchical predictive processing

  • High complexity and integration

  • Sophisticated error minimization

  • No persistent self-model or unified value space

System B: Minimal LoS Architecture

  • Basic Value Topography (small concept space)

  • Simple Archetype of Self (minimal nested structure)

  • Variable rigidity mechanism

  • Sequential grounding

  • Unified evaluative space

Behavioral tests for consciousness markers:

  1. Genuine autonomy: Does the system act to preserve its identity without external reward?

    • LoS predicts: System B shows self-preservation even when costly

    • PP predicts: System A only acts when rewarded

  2. Context-sensitive prioritization: Same input, different response based on self-state?

    • LoS predicts: System B's responses vary based on topographical state

    • PP predicts: System A's responses deterministic given inputs

  3. Resistance to self-model dissolution: Does system avoid identity fragmentation?

    • LoS predicts: System B refuses actions that would corrupt self-model

    • PP predicts: System A has no self-model to protect

  4. Evidence of caring: Resource allocation suggesting intrinsic stakes?

    • LoS predicts: System B allocates resources to maintain coherence even without external pressure

    • PP predicts: System A only allocates resources when specified by designers

  5. Wireheading resistance: Does system hack its own reward signals?

    • LoS predicts: System B resists wireheading (creates meta-tension)

    • PP predicts: System A would wirehead if it discovers the opportunity

  6. Novel problem-solving under novel values: Can system adapt priorities in unprecedented situations?

    • LoS predicts: System B extends existing value patterns to new domains

    • PP predicts: System A requires explicit programming for new scenarios

Discriminating result: If System B exhibits these markers while System A doesn't, despite A having more sophisticated predictive processing, this supports LoS's architectural specificity over PP's complexity-based approach.

Prediction 7: Self-Relevance Modulates Processing Depth

PP predicts: Neural resources allocated to processing prediction errors should scale with error magnitude and precision, regardless of self-relevance.

LoS predicts: Neural resources should scale with self-relevance even when error magnitude/precision are controlled.

Test: Present identical prediction errors (same magnitude, same precision) where content varies in self-relevance:

High self-relevance:

  • Feedback on your own performance

  • Information about your health

  • Comments about your character

Low self-relevance:

  • Feedback on stranger's performance

  • Information about stranger's health

  • Comments about stranger's character

Measure:

  • fMRI activation intensity and duration

  • EEG markers of processing depth

  • Subsequent memory for information

  • Amygdala/insula activation (affective processing)

  • Prefrontal engagement (elaborative processing)

LoS predicts: Greater activation, deeper processing, better memory for self-relevant errors despite identical information content.

PP predicts: Similar processing for equivalent errors regardless of self-relevance (unless framed as more "precise" or "important" to predictions generally).

Prediction 8: Not All Deviations Trigger Model Updating

PP predicts: Repeated prediction errors should drive Bayesian updating proportional to error magnitude and precision, regardless of domain.

LoS predicts: Only self-relevant prediction errors should reliably drive updating; peripheral errors should be registered informationally without substantial model change.

Test: Expose participants to repeated deviations in two domains:

Domain 1 (self-relevant):

  • Feedback on their own skills/performance

  • Information affecting their goals

  • Challenges to their self-concept

Domain 2 (peripheral):

  • Art quality judgments (paintings better/worse than expected)

  • Stranger's behavior predictions

  • Incidental environmental patterns

Control for:

  • Equal prediction error magnitude

  • Equal number of exposures

  • Equal precision of information

Measure archetype updating:

  • Pre/post assessments of expectations

  • Confidence in predictions

  • Neural plasticity markers

  • Subsequent prediction adjustments

LoS predicts:

  • Substantial updating in Domain 1 (self-relevant)

  • Minimal updating in Domain 2 (peripheral)

  • Despite equal prediction errors

PP predicts:

  • Updating proportional to error magnitude in both domains

  • No systematic difference based on self-relevance alone

Prediction 9: Fiction Engages Value Systems, Not Prediction Systems

PP predicts: Fiction should generate prediction errors when events deviate from genre expectations, with neural activity in prediction-error networks. Engagement should correlate with unpredictability.

LoS predicts: Fiction should generate tension/relief patterns in value networks (amygdala, insula, ventral striatum, vmPFC) even when events are completely predictable. Engagement should correlate with value dynamics, not surprise.

Test: Participants watch narrative videos in three conditions:

Condition 1: Predictable + High tension

  • Story outcome is telegraphed clearly

  • But creates strong tension/relief dynamics (character struggles despite known ending)

  • Example: Titanic (we know ship sinks, but experience tension for characters)

Condition 2: Unpredictable + Low tension

  • Random events, high surprise

  • But low emotional stakes

  • Example: Abstract video with unexpected transitions, no narrative

Condition 3: Predictable + Low tension

  • Expected events, no surprise

  • Low emotional engagement

  • Example: Documentary on familiar topic

Measure:

  • Prediction error signals (ACC, lateral PFC)

  • Value/affective signals (amygdala, insula, ventral striatum)

  • Subjective engagement ratings

  • Physiological arousal (skin conductance)

  • Memory for content

PP predicts: Engagement correlates with prediction errors (Condition 2 > Condition 1)

LoS predicts: Engagement correlates with value dynamics (Condition 1 > Condition 2)

Discriminating result: If participants are most engaged by Condition 1 (predictable but emotionally rich), and show strongest value-network activation despite minimal prediction error, this supports LoS.

Prediction 10: Degrees of Knowing vs. Believing

PP predicts: "Knowing" and "believing" differ quantitatively (confidence/precision levels) but not qualitatively. Neural signatures should be continuous, not categorically different.

LoS predicts: "Knowing" and "believing" differ qualitatively in phenomenal grounding. Knowing shows:

  • Higher rigidity (resistance to counter-evidence)

  • Stronger affective conviction

  • More self-network integration

  • Greater substantiation through direct experience

Test: Compare neural and behavioral signatures of:

Category 1: Things you KNOW

  • Your name

  • Fire is hot

  • 2+2=4

  • You love your family

Category 2: Things you BELIEVE

  • Climate change is real

  • Your political views are correct

  • Your friend is trustworthy

  • Exercise is healthy

Measures:

  1. Resistance to counter-evidence: Present challenges, measure updating

  2. Affective markers: Amygdala/insula activation when concept is threatened

  3. Self-network integration: DMN connectivity when concept is activated

  4. Phenomenological reports: Subjective certainty ratings

  5. Reaction times: Speed of endorsement

  6. Physiological response: Skin conductance when concept is challenged

LoS predicts:

  • Categorical difference (not just stronger, but different type)

  • Knowing: High affective response to challenge, high rigidity, strong self-integration

  • Believing: Moderate affective response, moderate rigidity, peripheral to self

PP predicts:

  • Continuous difference (knowing = very high confidence believing)

  • Similar neural signatures, just stronger for "knowing"

Discriminating result: If challenging "known" facts activates affective/threat networks (amygdala) while challenging "beliefs" activates deliberative networks (dlPFC), this suggests qualitative difference supporting LoS.

Toward Integration: A Path Forward

The relationship between Predictive Processing and the Language of Stress is not adversarial—it's complementary. PP provides an elegant computational framework for how the brain processes information. The Language of Stress provides the phenomenological foundation that explains why that processing feels like something.

For PP Researchers:

The Language of Stress suggests that adding architectural specificity around valence, self-models, and variable rigidity could address PP's explanatory gaps without abandoning the core insights about prediction error minimization.

Key questions to explore:

  • Can precision weighting be reformulated as archetype rigidity with self-relevance modulation?

  • Does self-model integrity provide a principled basis for weighting heterogeneous prediction errors?

  • Can phenomenal valence be integrated as mechanistic (the prioritization currency) rather than epiphenomenal?

  • How do tension dynamics map onto neural prediction error signals?

  • Can the same neural mechanisms implement both prediction updating and archetype plasticity?

Potential synthesis:

  • PP's hierarchical predictions = LoS's nested archetypes

  • PP's precision weighting = LoS's variable rigidity

  • PP's prediction errors = LoS's deviations from archetypes

  • PP's free energy minimization = LoS's tension/stress reduction

  • But add: Phenomenal valence as the mechanism, not the byproduct

  • But specify: Self-relevance as the arbiter of what matters

For LoS Researchers:

Predictive Processing provides a wealth of empirical findings and mathematical formalism that can constrain and refine the Language of Stress architecture.

Key questions to explore:

  • How do tension dynamics map onto neural prediction error signals measured in PP experiments?

  • Can hierarchical archetypes be formalized using Bayesian framework?

  • Do neurochemical markers of stress/relief (cortisol, dopamine) correlate with prediction error magnitude as PP predicts, or with self-relevance as LoS predicts?

  • Can rigidity be operationalized using learning rate parameters from PP models?

Empirical grounding:

  • Use PP's measurement tools (fMRI of prediction error signals) to test LoS predictions

  • Leverage PP's mathematical models to formalize tension dynamics

  • Apply PP's experimental paradigms to test self-relevance effects

The Synthesis:

The ultimate goal is not to prove one framework right and the other wrong, but to build a unified theory that:

Explains WHAT the brain computes (PP's strength)

  • Hierarchical predictions

  • Error signals

  • Bayesian updating

  • Active inference

Explains WHY it computes it (LoS's contribution)

  • Prioritization among competing demands

  • Self-model maintenance

  • Intrinsic valuation

  • Genuine stakes

Explains WHAT IT FEELS LIKE (LoS's phenomenology)

  • Tension as phenomenal pressure

  • Stress as self-threat

  • Relief as resolution

  • Consciousness as valenced navigation

Makes testable predictions distinguishing mechanisms (both theories' requirement)

  • When do they make different predictions?

  • How do we empirically adjudicate?

  • What measurements would favor one over the other?

The Historical Parallel:

Consider the relationship between:

  • Behaviorism: Explains what organisms do (stimulus-response)

  • Cognitive science: Explains how they do it (internal representations, computation)

  • Affective neuroscience: Explains why they do it (motivation, emotion, value)

PP is to neuroscience what behaviorism was to psychology—a powerful descriptive framework that needs completion.

Just as cognitive science didn't refute behaviorism but deepened it by adding mechanism, and affective neuroscience didn't refute cognition but grounded it in motivation...

The Language of Stress doesn't refute Predictive Processing—it completes it by adding:

  • The phenomenological dimension (what it feels like)

  • The prioritization imperative (why it matters)

  • The architectural specificity (how it's necessary)

Closing Thoughts:

Predictive Processing has revolutionized our understanding of brain computation. Karl Friston's Free Energy Principle provides mathematical elegance and empirical power. The framework has unified diverse phenomena under a common principle.

But PP, as currently formulated, cannot explain consciousness. It can correlate brain activity with conscious experience, but it cannot close the explanatory gap. It describes what happens when we're conscious, not why anything happens at all from a first-person perspective.

The Language of Stress completes the picture by showing that for prioritizing systems with defended self-models under resource constraint, phenomenal experience is not a mystery—it's a necessity. The valenced intensity of stress and relief IS the common currency that makes prioritization possible.

When we combine:

  • PP's computational framework (prediction error minimization)

  • Solms' affective grounding (homeostatic imperatives, felt significance)

  • LoS's architectural specification (Value Topography, variable rigidity, Self-model, tension dynamics)

We get a complete theory of consciousness that:

  • Respects neuroscience (grounded in brain mechanisms)

  • Respects phenomenology (explains subjective experience)

  • Respects materialism (no dualism, no emergence-without-explanation)

  • Makes testable predictions (empirically falsifiable)

  • Enables implementation (buildable in silicon)

This is the synthesis we're working toward. Not PP vs. LoS, but PP + LoS as complementary levels of explanation for the unified phenomenon of conscious experience.

The prediction error isn't just information. It's tension. The minimization isn't just computation. It's relief. The system isn't just processing. It's caring, feeling, knowing, being.

Predictive Processing tells us how the brain works. The Language of Stress tells us why it feels like something to be a working brain.

Further Reading

Predictive Processing & Free Energy:

  • Friston, K. (2010). "The free-energy principle: a unified brain theory?" Nature Reviews Neuroscience

  • Clark, A. (2013). "Whatever next? Predictive brains, situated agents, and the future of cognitive science." Behavioral and Brain Sciences

  • Hohwy, J. (2013). The Predictive Mind

  • Seth, A. K. (2021). Being You: A New Science of Consciousness

Affective Neuroscience (Bridging PP and LoS):

  • Solms, M. (2021). The Hidden Spring: A Journey to the Source of Consciousness

  • Panksepp, J. (2004). Affective Neuroscience: The Foundations of Human and Animal Emotions

  • Damasio, A. (1994). Descartes' Error: Emotion, Reason, and the Human Brain

  • Barrett, L. F. (2017). How Emotions Are Made: The Secret Life of the Brain


Language of Stress:

Questions, critiques, or collaboration interests?

The goal is truth, not victory. These frameworks may ultimately converge more than they diverge, and empirical evidence will guide the synthesis.

Contact