Lesson 18 · What the measurement cannot see

When a perfectly accurate measurement hides something essential

Paper V has shown how missing observations and imperfect predictive laws affect prospective Health. Now I want to examine a different problem: a representation that accurately preserves everything it was designed to store, yet cannot answer a particular juvenile-support question.

By Zed JamesPaper V · Sections 5.10, 6.8 and 7.3Proposition 5.1 · Juvenile-query non-factorization

← Revisit Lesson 17 All lessons

01 / Our original cohort measurement

A perfectly correct average can omit the distinction that matters to the question.

In Lesson 14 we used 64 cohort cells to represent a large collection of forest stems. In each cell the model stores the number of stems, n, and their root-mean-square diameter, d_RMS.

The root-mean-square diameter

Squaring diameters before averaging preserves their squared sum.

d_{RMS} = \sqrt{\frac{d_{1}^{2} + d_{2}^{2} + \dots + d_{n}^{2}}{n}}

This is exactly suited to a basal-area calculation, because basal area is proportional to the sum of diameter squares. The stored count and RMS diameter allow us to recover that aggregate.

But the continuation criteria also ask for a count of individual stems with diameter strictly below 10 centimeters. Knowing a cell's RMS diameter does not generally tell us how many stems lie below that threshold.

02 / The exact construction in Paper V

Let's build two collections that look identical to the coarse observer.

Both collections contain exactly 100 stems, but their individual diameters are arranged differently. These are constructed mathematical examples from Proposition 5.1, not directly observed Harvard Forest cohorts.

Collection A100 stems

100 stems × 10 cm each

All diameters exactly at the 10-centimeter boundary.

Collection B100 stems

50 × 9 cm + 50 × √119 cm

√119 ≈ 10.91 cm. Half the stems are below the threshold.

Below 10 cm10 cm or above

We can see the different stem-size distributions immediately at the individual-stem level. Now we compress both collections into the count/RMS representation.

03 / Calculate, rather than assume, the equality

The two RMS diameters are exactly the same.

Paper V · Proposition 5.1

Both collections have RMS diameter 10 cm.

\begin{array}{l} d_{RMS,A} = \sqrt{\frac{100 \times 10^{2}}{100}} = 10 cm \\ d_{RMS,B} = \sqrt{\frac{50 \times 9^{2} + 50 \times 119}{100}} = 10 cm \end{array}

Collection A contributes 100 × 100 = 10,000 to the squared-diameter sum. Collection B contributes 50 × 81 + 50 × 119 = 4,050 + 5,950 = 10,000. Divide by 100 and take the square root: both give 10 cm.

Coarse measurements retained exactly by both representations
Quantity	Collection A	Collection B
Stem count	100	100
RMS diameter	10 cm	10 cm
Sum of squared diameters	10,000 cm²	10,000 cm²
Total basal area	π/4 ≈ 0.7854 m²	π/4 ≈ 0.7854 m²

No error has occurred. The count is correct, the RMS calculation is correct, and the basal area is preserved exactly. The two fine collections map to the same coarse information.

04 / The information that disappeared

One group has no juvenile stems. The other has fifty.

In this specification, juvenile-support membership means an individual diameter strictly less than 10 cm. A stem measuring exactly 10 cm is outside the juvenile class.

The individual-stem question

Count stems below the threshold.

\begin{array}{l} J (X) = \sum_{i} 𝟙_{d_{i} < 10} \\ J_{A} = 0, J_{B} = 50 \end{array}

The symbol 𝟙 is an indicator: a stem contributes one if its diameter is below 10 cm, and zero otherwise. Thus J_A = 0 and J_B = 50.

Collection A0 / 100

Actual juvenile stems

Collection B50 / 100

Actual juvenile stems

But the original cohort rule treats the full cohort as juvenile only when its RMS diameter is below 10 cm:

The coarse cohort calculation

An entire cohort receives one threshold membership.

J_{RMS} = \sum_{c} n_{c} 𝟙_{d_{RMS,c} < 10}

Both collections have RMS diameter exactly 10 cm. Thus the coarse representation assigns J_RMS,A = J_RMS,B = 0—even though Collection B actually contains fifty stems below 10 cm.

The coarse representation preserves the basal area accurately while erasing a distinction required by the juvenile-support query.

Look through two different observers

What can the representation actually tell us?

Change the observer to see what the two collections reveal. The original count, RMS diameter, and basal area never change. The juvenile query needs a distinction that the coarse representation lacks.

Choose which information to examine

Collection A100 · 10 cm

0.7854 m² basal area

Collection B100 · 10 cm

0.7854 m² basal area

The two coarse records are identical. A calculation using only those records cannot distinguish the collections.

The table above contains both the preserved coarse quantities and the distinct actual juvenile counts even when interactive controls are unavailable.

05 / A mathematical impossibility, not a calibration failure

No function of the coarse data can answer every juvenile-threshold query correctly.

Define a simple query: does the collection contain at least one juvenile stem? Let q_J(X) equal 1 when J(X) ≥ 1, and 0 otherwise.

Proposition 5.1 · Juvenile-query non-factorization

Equal coarse representations, unequal correct answers.

\begin{array}{l} π_{RMS} (X_{A}) = π_{RMS} (X_{B}) \\ q_{J} (X_{A}) = 0, q_{J} (X_{B}) = 1 \end{array}

The two fine states X_A and X_B have the same coarse projection π_RMS. But the individual-stem query produces q_J(X_A) = 0 and q_J(X_B) = 1.

Suppose a deterministic function of coarse data could always recover q_J. It would receive identical inputs for X_A and X_B, so it must return the same answer for both. The correct answers differ. This is a contradiction, so the function cannot exist over the declared collection of fine states.

This result has a precise scope: it rules out exact recovery of the particular juvenile-existence query from the coarse representation for all fine states. It does not assert that every possible juvenile-sensitive query differs for these two collections, or that their full prospective Health judgments must differ under every scenario.

We have recovered the essential insight of Paper III: representation sufficiency is relative to the question being asked. Exactness for basal area does not imply sufficiency for juvenile support.

06 / What the observed E1 census reveals

This kind of disagreement also occurs when measured diameters are audited.

The paper compares individual-stem juvenile membership against the membership assigned from cohort RMS diameters for 53,611 securely observed living stems in the E1 census.

Audited living stems53,611

Incorrectly called juvenile1,187

Actual juveniles missed18

Total disagreements1,205Approximately 2.25% of the tested support

These are disagreements in juvenile-threshold membership on the particular observed support. They demonstrate that the RMS-based classification can differ from individual diameters in real records.

The 2.25% figure is not a known error rate for the fully reconstructed January 2020 forest state, nor does it directly quantify error in a prospective probability of viable continuation.

07 / Restore the information needed by the query

Split the cohort on the two sides of 10 cm.

The proposed enriched representation carries separate counts and RMS diameters for stems below 10 cm and stems at least 10 cm. The initial-state projection reconstructs the old cohort count and squared-diameter total exactly.

Fine componentd < 10 cm

Juvenile-side count and RMS

Fine componentd ≥ 10 cm

Nonjuvenile-side count and RMS

Coarse projectionn, d_RMS

Original count and basal-area summary

Paper V · Equation (112)

Projection preserves count and squared-diameter sum.

\begin{array}{l} n_{c} = n_{c, <10} + n_{c, ≥10} \\ n_{c} d_{c}^{2} = n_{<10} d_{<10}^{2} + n_{≥10} d_{≥10}^{2} \end{array}

In the second equation the terms n and d with subscripts <10 and ≥10 refer to the two size components within cohort c. The squared-diameter identity is exactly what retains the original basal area.

The enriched state can distinguish juvenile counts across the threshold while still returning the coarse counts and basal-area measures. For a query about the number below 10 cm, that additional information is relevant.

08 / What changes in prospective simulations?

An initial-state projection does not establish a future-capacity preservation theorem.

The paper compares the enriched size-split simulation with the original coarse implementation. In a selected finite-design comparison, only three of 1,536 Health classification cells change. But a challenged state exhibits a viable-mass change of 67/256 ≈ 0.262, or approximately 26.2 percentage points.

Classification changes3 / 1,536

Finite-design Health cells

Selected local viable-mass change26.2%

67 / 256, percentage-point difference

The experiment demonstrates heterogeneous numerical sensitivity. Most tested classifications remained unchanged, while a selected state produced a much larger local capacity difference.

Yet an important qualification follows from Lesson 13: these simulations use different effective future representations and operations. Preserving initial count and basal area does not prove that the enriched and coarse stochastic future kernels commute under projection. The paper does not establish that exact compatibility.

The enriched computation may answer a more detailed question, but its empirical correctness and compatibility with earlier future Health judgments still require further demonstration.

09 / What Paper IV and Paper V each establish

The enrichment theorem has conditions; the ecological example shows why each condition matters.

Precisely what the two papers guarantee
Paper IV · Exact formal enrichment	Paper V · Forest size-split experiment
Richer response atoms distinguish cases merged by the original observer.	Smaller and larger individual stems can be distinguished inside a cohort.
Projection exactly recovers the established prospective capacity.	Initial projection exactly preserves stem count and basal area.
Dynamics and constitution are fixed under the preservation theorem.	Future simulation operations are altered by the enriched state description.
Established Health judgments are preserved under adequacy pullback.	General equality of future Health judgments has not been proved.

Nothing here contradicts conservative enrichment. Rather, the empirical implementation has not established all of the conditions under which the formal preservation result applies.

10 / Check your understanding

Two stem collections share count, RMS diameter and basal area, but have different juvenile counts. What follows?

Choose the strongest mathematically justified conclusion. The RMS calculation must have been performed incorrectly. Equal basal areas guarantee identical counts below 10 cm. Count, RMS diameter and basal area do not suffice for all juvenile-threshold questions.

Select an answer to check the information-sufficiency result.

Paper V · Where the research has brought us

The formal calculation is defined. Its measurement requirements are now visible.

Across Lessons 14–18 we've investigated incomplete-state reconstruction, the probability mass of viable and failed futures, decision reliability near adequacy thresholds, independent mortality calibration, and exact information lost by a cohort representation. Those are different issues, and each needs its own evidence and remedy.

Paper V brings formal prospective Health into contact with an auditable empirical pipeline. The calculations can be exact relative to declared specifications while the ecological models, observation processes, and query-sufficient representations remain subjects for validation and improvement.

What I want you to carry forward

Correct measurement of one quantity cannot recover information that was never retained.

Both 100-stem collections are identical under count, RMS diameter, and basal area, yet different under the juvenile-existence query. The non-factorization theorem proves that no calculation using only those coarse values can recover the correct answer for every fine state.

Lesson 19 will bring together the entire five-paper series. We'll distinguish exact mathematical results from empirical findings, assumptions, and open scientific questions, and examine how prospective Health relates to neighboring work in viability theory, resilience, and capability-based approaches.

← Previous: Lesson 17 All lessons ↗Read Paper V (PDF)

Source: Zed James, Prospective Health under Declared Specifications, Paper V in Health, Formally Defined (2026), Sections 5.10, 6.8, and 7.3, especially Proposition 5.1 and Equation (112). Full publication record · Zenodo DOI. The 100-stem collections are exact mathematical witnesses, distinct from the observed-stem audit and the finite prospective enrichment comparison.