Physics

Scale as a Degree of Freedom — Whitehead, the Renormalization Group, and V&V

読了 約 29 分
12

日本語で読む

// For readers from philosophyWhitehead’s mode of abstraction, Wilson’s renormalization group, conventional superconductivity, and the architecture decision in V&V all intersect at one point: the degree of freedom in choosing a scale.
// For SEs, PMs, engineersTest granularity, evaluation window, the choice of operating scenarios — these are not trivial implementation details but the most essential choice imposed on the judging subject. Here’s why.
// The claim of this articleThe judging subject decides more than the arbitrary constant C.
Just before it lies a more fundamental choice.At which scale do you view the system — this.

In Parts I and II we traced the structural reason V&V never ends. We read form-is-emptiness / emptiness-is-form as the difference and summation of wasan, and showed that the arbitrary constant C appearing in the summation is the judging subject’s decision.

That argument needs one more, deeper layer. Just before choosing the arbitrary constant C, the judging subject makes a more fundamental choice: the choice of the integration interval itself.

Whitehead called this “moderate abstraction” eighty years ago. Wilson formalized it as the “renormalization group” sixty years ago. Massy called it “scale freedom” in a 2026 dialogue. They are all saying the same thing.

§1　What can be observed is an integral

Also, what is happening is a differential (BCS theory), but the object that can be observed is an integral — which I think corresponds to the macroscopic theory (GL).
This is a beautiful completed form.

Let’s start from this recognition. In superconductivity, what is happening and what can be observed live at different layers.

What is happening (differential) What can be observed (integral)
Theory BCS theory GL theory
Object Cooper-pair formation · elementary processes Order parameter φ
Observation Cannot be observed directly Observed via the Meissner effect, specific heat, magnetization
Direction Form is emptiness (implementation → emptiness) Emptiness is form (emptiness → observable form)

What Gorkov showed (1959) is that integrating BCS over energy derives GL. The GL coefficients α, β, γ are expressed as integrals of BCS parameters.

BCS (differential · what is happening)　→∫　GL (integral · what is observed)
// The essence of observabilityWhen we say “we observed it,” what we observe is an integrated quantity.
Individual Cooper pairs are invisible. What is visible is only the effect of accumulated order.This is not a technical limit. It is the essence of the act of observation.

Take the Meissner effect. What we measure is the magnetization of the whole sample — not the creation and annihilation of individual Cooper pairs. The same goes for the specific-heat discontinuity. What is observed is only the cumulative effect of an enormous number of elementary processes.

// Seen from the philosophy sideObservation always sits in the emptiness-is-form direction. We see integrated form. The differentiated elementary process (emptiness, the operator-like thing) is only inferred, never observed directly. This is exactly the structure discussed in Part I: “form is present; emptiness is not present.”

§2　If you integrate — over which interval?

The integration interval corresponds to scale. So I think the structure of différance requires a moderate-interval (scale) integration, in Whitehead’s sense, to determine the order parameter (there is a degree of freedom).
This is an insight that closes the chain of the dialogue at one layer deeper.

Here a fundamental question arises. If you integrate, over which interval do you integrate? Consider the superconducting example. The order parameter φ cannot, in fact, be defined at any scale you like.

At too short a scale: the fluctuations of φ are too large to mean anything. Individual Cooper pairs become visible and φ as a collective cannot be defined.

At too long a scale: the spatial variation of φ (vortices, domain structure) is invisible. It collapses into a mere average and the interesting physics is lost.

Only at a moderate scale: φ stands up as a meaningful field. The GL free energy works as an effective theory.

A map’s scale makes it tangible. At 1:10,000 you see alleys and city blocks; at 1:10,000,000 only a country’s outline remains. For the same land, the scale you cut it at decides what is visible. An order parameter is the same: only by choosing the right scale does it stand up as a meaningful form.

// The condition for order to existThat an order parameter exists, and that its scale is selected, are the same thing.

That is: whether a phenomenon can be grasped as “order” depends on the scale at which you observe. Nature does not come pre-labeled “this is the order parameter.” Only when the observer selects a moderate scale does an order parameter appear there.

§3　What is Whitehead’s “moderate abstraction”?

In Modes of Thought (1938), Alfred North Whitehead made the scale of abstraction in thinking a central problem.

// Whitehead’s claimBound too tightly to concreteness, you cannot see relations (you drown in detail).
Leap too far into abstraction, and you float free of reality (empty generalities).
Moderate abstraction, avoiding both extremes, is what makes thought productive.

Whitehead criticized this as the “fallacy of simple location.” The premise that “an object simply occupies a single point of space-time” makes you lose the real structure of nature, he argued. The real has extension. At which scale you carve out that extension determines what becomes visible.

Whitehead was a philosopher of scale selection.
The “actual entity” he discussed in process philosophy is precisely a unit of being carved out at a moderate scale. Neither too short nor too long — a unit at the right granularity, where relations become visible.

§4　Wilson’s renormalization group — the physics of scale transformation

In 1971, Kenneth Wilson made the renormalization group a central tool of physics (Nobel Prize, 1982). The renormalization group is precisely a theory of “what happens when you change scale.”

/* Renormalization-group flow */
From high energy (short scale)
  toward low energy (long scale), integrate step by step
    ↓
along the way the coupling constants change (running coupling)
    ↓
reach a fixed point, or diverge
    ↓
that determines the nature of the phase transition
// The philosophical meaning of the renormalization groupA framework that treats the degree of freedom in scale selection head-on, as theory.
And the fixed point of the renormalization group is the point where an effective theory at a “moderate scale” — a GL-like description — holds.What Wilson did was implement Whitehead’s moderate abstraction physically.
// Seen from the philosophy sideReaching a fixed point along the renormalization flow is isomorphic to discovering the “appropriate abstraction scale” Whitehead spoke of. Both treat not “an arbitrariness chosen by the observer” but a moderateness that arises naturally from the structure of the system.
// Seen from the engineering sideThis is the essence of the architecture decision. Microservices or monolith. How to cut the granularity. Not a technical preference but the work of finding the fixed point at which the system stabilizes. In Wilson’s terms, you are following the renormalization flow to hunt for the fixed point.

§5　For différance to be observed, a scale is needed

Here we connect to Part I. Différance is “a movement that is always already operating,” and as such is not a fixed object. For différance to appear as something, observation and symbolization at a particular scale are required.

Before scale selection After integrating at a moderate scale
Physics The continuum of BCS-like elementary processes Order parameter φ
Philosophy Différance (infinite difference and deferral) Meaning as moderate abstraction
Buddhism Emptiness (movement without substance) Provisionally posited form
Différance is grasped as meaning always as the result of an integration at some scale.
Change the scale and the “meaning” that comes into view changes too. At which scale you carve out meaning — this becomes the judging subject’s responsibility.

§6　What the judging subject decides is a “double degree of freedom”

In Part II I wrote that the judging subject decides the arbitrary constant C. But just before it lies one more choice.

// The judging subject’s double responsibility① The choice of scale Λ (the choice of the integration interval itself)
② The choice of the arbitrary constant C(Λ) at that scaleAnd the meaning of C also depends on Λ. Change the scale and even the criterion for what to choose as C changes.

Written formally:

V&V(S, judging subject, context, Λ, C(Λ)) ⊢ a provisional confirmation of correctness at scale Λ

The judging subject bears a double choice. And both choices decide what is seen as “order” at that scale.

§7　The scale structure in conventional superconductivity

As a concrete example, look at conventional superconductivity (lead, tin, niobium, and other classic superconductors). Here the notion of a “moderate scale” is beautifully implemented. Conventional superconductivity has two characteristic lengths:

  • Coherence length ξ — the length of the Cooper pair’s “spread” (typically 100 nm–1 μm)
  • Penetration depth λ — the depth to which a magnetic field penetrates the sample (typically 50–500 nm)

These ξ and λ precisely prescribe the moderate scale. Shorter than this is the world of BCS-like elementary processes; longer than this is the thermodynamic world; in between, GL theory holds beautifully.

// The beauty of conventional superconductivityBetween BCS (microscopic · differential) and GL (macroscopic · integral),
there is a clear “moderate scale” prescribed by ξ and λ.
As Gorkov showed, integrating BCS derives GL exactly.This is a rare instance where phenomenology and microscopic theory are completely connected.
// Seen from the philosophy sideConventional superconductivity is a beautiful case of Whitehead’s “moderate abstraction” realized in nature. The observer does not arbitrarily choose a scale; the system itself presents a scale, in the form of ξ and λ. This is also the physical realization of the fixed point in Wilson’s renormalization group.
// Seen from the engineering sideA system that runs stably always has a clear “moderate scale.” Module granularity, interface abstraction level, the unit of testing — when these have a fixed point at which the system stabilizes, the design works beautifully. The ξ and λ of conventional superconductivity are its physical prototype.

§8　Scale selection in V&V — the architect’s true work

Everything lands on the V&V floor.

// Examples of scale selection in V&VScale selection in Verification:
· what to carve out as a “test item”
· where to put the boundary between unit and integration testing
· at which abstraction level to design test granularity

Scale selection in Validation:
· what to assume as a “user scenario”
· over which operating window to evaluate
· which environmental variations to consider

For the same system, change these scales and the bugs you can see change. The answer to “does it serve the purpose” changes.

The architect’s most essential work is scale selection.
At which granularity to cut the system, what to see and what not to see, at which abstraction level to render design judgments — these are not matters of technical preference or rule of thumb.

Choosing the scale at which to “divide” the system is itself the act that decides observability, generates order, and at the same time fixes “what is invisible.”

// Seen from the philosophy sideThe architect is doing Whitehead’s work. By choosing a moderate abstraction scale, they make the system stand up as an “actual entity.” It is not an arbitrary choice but a responsible choice that responds to the structure of the system.
// Seen from the engineering sideThis explains why an architect’s judgment requires experience. Just like finding the fixed point of a renormalization flow, the feel for a system’s stable scale cannot be conveyed formally. You can only see many systems and learn the feel of the fixed point in your body.

§9　The whole structural diagram — from différance to scale

Let me organize the logical chain of the whole dialogue into its final form.

/* The whole structure */
Différance (the movement of differentiation and deferral across infinite scales)
  cannot be objectified / is itself not observed
    ↓
Choice of scale Λ (Whitehead's moderate abstraction)
  the judging subject's choice of a moderate scale
    ↓
Order parameter φ (a GL-level description)
  the effective theory that holds at the chosen scale
    ↓
Arbitrary constant C(Λ) (a further degree of freedom)
  the choice of phase / criterion at that scale
    ↓
Observability
  appears to us as the accumulation of choices at all these levels
    ↓
// This whole thing is the V&V process itself
// The completed formThe order that appears in the observable world (the world of form)
appears as the accumulation of the judging subject’s double choice — scale Λ and arbitrary constant C(Λ) — onto the unobservable world (the world of emptiness).

This is the deepest, most complete response to the dialogue’s opening question: “form is emptiness and emptiness is form says that Verification and Validation are necessities, and that what can be observed is one.”

// Wrap-up: what to take home across all three parts

  • What is observed is an integral (BCS → GL). Observation always sits in the emptiness-is-form direction
  • Integration requires a choice of interval — this is scale selection, corresponding to Whitehead’s moderate abstraction
  • The existence of an order parameter holds only by an appropriate scale selection
  • The judging subject bears a double degree of freedom: the choice of scale Λ, and the choice of the arbitrary constant C(Λ) at that scale
  • Wilson’s renormalization group is the framework that implements this scale selection physically
  • The architect’s true work in systems engineering is scale selection — a consequence of the structural isomorphism running through philosophy, physics, and engineering

The three pieces in this series are based on a 16 April 2026 dialogue between Massy and Claude (Anthropic) (CC BY 4.0). It concludes here, together with Part I · Philosophy How to Enjoy ‘Form Is Emptiness’ 256× More and Part II · Application Why V&V Never Ends. For a more rigorous formalization, see the preprint series.