Most of what happens in an internal system — the particular parts a person has, the specific burdens they carry, the exact configuration of their Manager-Firefighter-Exile ecology — is completely individual. No two people's inner worlds look the same. IFS makes no claims to a universal map of which specific parts everyone has or what everyone's parts want.
But Schwartz names two claims that have held without exception across decades of clinical practice — structural rules that appear to apply to every inner system, every time, regardless of the complexity or severity of the client's history.1 He calls these the Laws of Inner Physics.
They are not axioms derived from theory. They are empirical observations that clinical practice has not yet managed to contradict. That matters for how they should be used: not as metaphysical certainties but as extremely reliable working assumptions that have never been falsified in the clinical record and that practitioners can stake treatment decisions on.
The two laws are:
Law One: The Agreement Principle — Parts can agree not to overwhelm the system, and when they make that agreement, they will honor it.
Law Two: Self Neutralizes — When the Self is genuinely present and unafraid, nothing in the internal system has power over it.
Each of these is simple to state and radical in its implications.
When a practitioner asks a protective part to step back or agree not to flood the system during a session, and the part makes that agreement — it will keep it.1
The clinical version: before approaching an exile, the practitioner negotiates with the Managers and Firefighters protecting that exile. They are asked not to interfere with the session, not to overwhelm the client, not to shut down access to the exile through their usual protective maneuvers. If they agree, they will hold to that agreement for the duration of the session.
This is not always easily secured — protective parts often need significant reassurance, curiosity, and direct engagement before they are willing to agree. But once they agree, they keep the agreement. The agreement is not a polite fiction that collapses under pressure. It is a structural feature of how parts operate.
The Agreement Principle is the clinical anchor for IFS work with highly traumatized or dissociated clients — the populations that most therapeutic approaches struggle with precisely because protective systems are so extreme that standard approaches cannot penetrate them.
The standard clinical problem with severe trauma presentations: attempting to approach the exile directly activates protective flooding that prevents any access. The client dissociates, shuts down, is overwhelmed. The therapist backs off. The session achieves nothing for the exile and reinforces the protective parts' belief that their strategy is necessary.
The Agreement Principle interrupts this loop. Instead of attempting to bypass the protectors, the practitioner engages them directly — explains what IFS proposes to do, addresses the protector's specific concerns about what will happen if the exile is approached, and secures the protector's actual buy-in. The protector knows the exile better than the practitioner does; it knows exactly what has been too dangerous to approach. If the practitioner can earn the protector's trust that this approach is different from the ones that have previously failed or caused harm, the protector will agree to step back — and will hold that agreement.
Schwartz reports that in decades of clinical practice, this law has not been violated. No part that has genuinely agreed — meaning the agreement was made from some degree of trust rather than compliance under pressure — has subsequently overwhelmed the system during the session.1
The key phrase is "genuinely agreed." The Agreement Principle is not magic — it does not mean that parts will agree to whatever you ask them, or that any agreement automatically holds. Parts that have been coerced or manipulated into an agreement, parts that said yes from a place of fear rather than genuine trust, parts that are performing cooperation while secretly running a different agenda — these will not hold the agreement.
The practitioner's job is to earn genuine agreement, not to extract nominal compliance. This requires actually caring about the part's concerns, actually modifying the approach based on what the part says it needs, and giving the part real assurance that its exile will not be hurt worse by the work. The agreement holds because the protector actually trusted the process. This is why the relationship with protective parts must come first — not as a strategy to reach the exile faster but as a prerequisite for the agreement to mean anything.
When the Self is genuinely present in the system — differentiated from the parts, not blended with any of them, not afraid of what it is encountering — nothing in the internal system has power over it.1
The clinical version: parts that had previously overwhelmed the client, Exiles whose charge had been so intense that any approach destabilized the entire system, Firefighters whose interventions had seemed impossible to interrupt — all of these lose their destabilizing power in the presence of a sufficiently differentiated and unafraid Self.
This is the most counter-intuitive of the two laws. The clinical common sense is: more severe and entrenched the parts, more powerful the protective responses, more difficult the work. The Law of Self says something different: the protective responses have power in the absence of Self-presence. In the presence of Self, that power dissipates.
Schwartz uses the film A Beautiful Mind to illustrate the principle.1 John Nash, the mathematician whose schizophrenic hallucinations included vivid, fully articulate characters that dominated his experience, eventually finds that he can operate in the world not by fighting the hallucinations or medicating them into silence but by developing a different relationship with them — acknowledging their presence, declining to engage, learning to move alongside them without being controlled by them. The hallucinations do not disappear. They lose their authority over his behavior.
This is not a clinical endorsement of Nash's approach to schizophrenia — Schwartz is using the film's representation as an illustration of the structural principle. What Nash's cinematic resolution demonstrates is the Law of Self: the "parts" (in the film's representation) that had organized his entire life and work lose their organizing power when he stops treating them as authoritative and starts relating to them from a position of stable, unafraid presence.
In IFS terms: the hallucinations are extreme parts. Their power over Nash came from blending — from the absence of a sufficiently differentiated center that could observe them without being recruited by them. When that center was stable enough, the power differential shifted. The parts still existed, but their power to flood and direct the system depended on the absence of Self, not on their own inherent force.
Self must be present and unafraid. These are two conditions, not one.
A Self that is present but afraid of what it encounters will retreat from the exile rather than remaining. The protectors monitoring the Self's approach will register the fear response as confirmation that the exile is indeed too dangerous to approach — and will respond by escalating their protective intervention. The Self's retreat was not a failure of will; it was the structural response of a Self that was not yet sufficiently differentiated from the exile's charge to remain present without being pulled into blending.
The clinical language for this: the Self needs enough of its own ground to be curious about the exile rather than frightened of it. The shift from fear to curiosity about the exile is the shift from a Self that will be destabilized by approach to a Self with enough presence to do the work. The 8 C's of Self (calm, curious, connected, compassionate, courageous, creative, clear, confident) are not personality traits to be cultivated — they are the experiential signature of a Self that has enough ground to be unafraid.
Law Two has a structural implication for why therapeutic approaches that focus on strengthening ego function or building coping skills often have limited efficacy with severe trauma: they are addressing the wrong variable. The problem is not that the person lacks coping resources (the "building up" model). The problem is that the coping resources they do have are parts running protective functions, and those parts are not Self.
Strengthening a Manager part's capacity to suppress an exile produces more effective suppression, not less exile charge. The exile's charge, suppressed better, finds another route. The Law of Self suggests the alternative: instead of building up the parts that manage the exile, develop the Self's capacity to be present and unafraid in the exile's presence. The exile doesn't need to be better managed. It needs a relationship with a center that is not afraid of it. That relationship is what makes the exile's charge available to be released rather than merely contained.
The two laws together produce a specific clinical protocol for presentations that standard therapeutic approaches cannot penetrate.
For a client whose protective system shuts down every attempt to approach trauma material:
Step 1: Meet the protectors directly. Not as obstacles to bypass but as the first subjects of the work. Use the same IFS approach — curiosity, no blending, genuine interest in what the protector is afraid of — that will eventually be used with the exile.
Step 2: Develop the relationship with each protector until it is solid enough that the protector's concerns are genuinely understood. What exactly is the protector afraid will happen if the exile is approached? What has happened before when the exile was approached in other therapeutic contexts (or in daily life)?
Step 3: Secure the agreement. Not extract compliance — earn agreement. This may require multiple sessions with the protector before any agreement is available. When it is available, secure it explicitly: "Are you willing to step back during this session and allow us to approach [the exile]? Can you agree not to shut the system down?"
Step 4: Trust the agreement. The Law says it will hold. If it does not hold, something in Step 3 was incomplete — the agreement was compliance rather than genuine. Return to Step 2.
Step 5: Develop the Self's capacity to be present and unafraid in the exile's presence. This is where Law Two activates. As the Self maintains genuine curiosity in the face of the exile's charge — without being destabilized, without blending, without retreating — the exile's power to flood the system diminishes. The exile is not in less pain; but the pain is not being amplified by the absence of a relational center capable of witnessing it.
When a session isn't moving — when parts keep flooding despite stated agreements, when the client keeps dissociating, when every approach to the exile produces shutdown — the two laws provide diagnostic leverage.
If the agreement isn't holding: the agreement was not genuine. Return to the protector relationship. Something the protector needs has not been addressed.
If the Self is being overwhelmed despite presence: the Self is still blended with something — either the exile (pulled into its charge) or a protector (running interference). The Self's "presence" is actually a Self-like Manager's performance. The diagnostic question: does the Self's engagement feel curious and open, or tense and goal-directed? Goal-directedness is a Manager signal.
The Laws of Inner Physics are exclusively a No Bad Parts contribution — they are not named or formalized in the 1995 textbook.1 The 1995 text describes clinical practice that implicitly relies on both principles (the practitioner earns trust from protectors; the developed Self works with exiles without flooding), but the abstraction of these into named, explicit clinical laws is a 2021 formalization.
The naming matters clinically: once practitioners have the Laws as explicit working assumptions, they can use them diagnostically. When Law One appears to fail (an agreement doesn't hold), the named law tells them where to look (the agreement was compliance, not trust — return to the protector relationship). When Law Two appears to fail (Self is being overwhelmed), the named law tells them what to check (blending — the Self is not as differentiated as the session assumed).
The 2021 book also introduces the qualification that is absent from simpler summaries of IFS: Self must be unafraid. This precision is the difference between a clinical principle and a clinical rule. Without the qualifier, practitioners might assume that any presence of Self qualities should neutralize the parts. With it, practitioners understand that partial Self-presence — a Self that is present but carrying anxiety about the exile — will not activate Law Two. The anxiety is itself a blending signal.
Psychology — IFS: Self and Self-Leadership: The Laws of Inner Physics operationalize the Self-leadership concept. Self-leadership is described in terms of its qualities (the 8 C's) and its structural position (undamaged center differentiated from all parts). The Laws describe what those qualities and that position actually produce in the system's behavior. Law Two is essentially a statement about what Self-leadership is for: not just a preferred way of being but a specific intervention into how the system organizes itself. The Self's presence doesn't just feel better — it structurally neutralizes the parts' power to dominate. Together the two concepts produce the full picture: what Self is (the qualities + structural position) and what Self does (the Laws).
Psychology — Epistemology of Survival: Leo Gura's framework describes how defense mechanisms operate as cognitive gatekeepers, preventing the person from knowing they have defenses. The Agreement Principle runs counter to this picture: if parts can genuinely agree to step back, they must have some access to the meta-level of what they are doing. They must know, at some level, that they are running a defensive strategy and be capable of choosing to pause it. This is a small but significant gap between the two frameworks. IFS implies that parts have more self-awareness about their protective function than Gura's gatekeeping model would predict. The gap is generative: does the degree of a part's access to its own function predict how available it is for the Agreement Principle? Can parts that are deeply unconscious of their function agree, or does agreement require some minimal meta-awareness?
Behavioral Mechanics — Operator Internal Mindset: The Composure Pendulum in the behavioral mechanics framework describes the same structural phenomenon as Law Two from the outside: the operator whose internal state is grounded and unafraid occupies a different position in any social system than one who is reactive or performing composure. The behavioral mechanics framework observes this from the outside (state as the primary broadcast signal; others register the operator's actual internal state before any content is transmitted). IFS's Law Two describes the same dynamic from the inside. Together: the unafraid Self doesn't just feel better and work better therapeutically — it broadcasts a different signal into every system it enters. The structural reason some people's presence changes a room is not charisma; it is the effect of a genuinely differentiated and unafraid center operating in a field of parts.
The Sharpest Implication
The Agreement Principle's corollary is uncomfortable: if parts that genuinely agree always keep the agreement, and your protective parts are not keeping agreements — then the agreements were not genuine. Every "I'll try not to do this anymore" that didn't hold, every resolution that evaporated, every commitment to change that lasted three weeks before the familiar pattern returned — was not a failure of willpower. It was a failure of genuine agreement. The part that needed to agree didn't agree. It complied under pressure from the decision-making center, which is a very different thing. Genuine agreement requires that the part understands why the change is being proposed, that its own concerns have been heard, and that it actually prefers the proposed alternative to the behavior it is currently running. That is a much higher bar than "I've decided to stop." And it explains why decisions without part-engagement don't hold: you secured compliance from the part's surface, not agreement from its actual operative concern.
Generative Questions