How to Read a TGH Assessment

A short guide to the diagnostic — what each reading means, and what it does not. The Great Homecoming research programme · Client reading guide · June 2026

Scope. This guide explains how to read the output of a structural assessment — the vocabulary and the headline calls. It is the companion to “How an Assessment Works,” which covers the process (scoping, data, delivery). Everything here is the secular structural register; the framework’s other registers are not used in client work. Evidence standing is stated at the end and is meant to be read with the rest, not skipped.

What an assessment is reading

A system — a company, an institution, a public body — runs on three capacities at once, and an assessment reads across all three:

• Integration capacity — the ability to pull its own diversity into coherence: to hold a real direction, make shared meaning, and coordinate the parts so they act as one rather than many.
• Interaction capacity — the ability to meet, take in and respond, inside and out: how the system senses and translates information, people and pressure, from its environment and from its own members.
• Bonding capacity — the ability to hold, and to make what holds: durable structure, and the output and legacy it passes on to what comes next.

Most decline is one of these weakening while the others, and the balance sheet, still look fine. The assessment reads which capacity is going, how far it has gone, and whether it can still be turned.

It does this through four headline reads. None of them is a single score, and there is a reason for that, given under “Vitality” below.

1. Vitality — read the weakest pillar, never the average

Vitality is reported as a profile of four pillars, not a number:

• Orientation — is the system still pointed at something real and shared, or has it turned inward to serve itself?
• Flow — does honest information, especially bad news, still travel to where decisions are made?
• Renewal — is the system still adding new coherence (the growth-arrest read), or only maintaining what it has?
• Correction — when the system is told it is wrong, does anything change?

The headline is always the weakest pillar — never the average of the four. This is deliberate and it is the single most important thing to understand about the reading. Averaging is precisely how a fatal weakness hides: a system can be strong on three pillars and failing on the fourth, and the fourth is the one that will end it. A composite “health score” would let the three strong pillars mathematically cancel the failing one and report “fine.” We refuse to produce that number. A system is exactly as sound as its weakest load-bearing pillar, and the reading keeps that pillar in front of you.

2. Crisis or collapse — the headline call

The most consequential distinction an assessment makes is between two states that look similar from outside and are completely different underneath:

• Crisis — the system is under high stress and has gone rigid, but it can still recover. Its self-correction still works; it is reachable. A crisis is bad news that is still good news: there is something to act on.
• Collapse — not an event but a state: the system’s correction loop has been cut. The mark of it is malfunction that will not reduce even when the information is clean and the stated direction is right — the system is told, clearly and repeatedly, and nothing changes. When a system is in this state, the shock that eventually arrives does not cause the failure; it finds a door already closed from the inside.

Telling these apart is the headline call because it determines whether intervention is worth funding at all. This call draws on where the system sits in its life-cycle, which is the part of the model still under completeness review, so it is delivered with a provisional flag until that review closes — stated, not hidden.

3. Recoverability — is there a hand on the wrench?

A system whose correction loop is damaged usually cannot fix itself with that same loop. Repair has to come from a level that is still healthy — an owner, a board, a regulator, a parent body; for a person, the people around them. So the assessment reads recoverability: is a capable, healthy “level above” — a new leader, anchor, or paradigm that could actually enter and act — available?

This carries one hard asymmetry worth stating plainly. A system at the very top of its hierarchy, with no level above it, has only self-renewal left, and that is rare and never accidental. For everything below that apex, recoverability is usually a question about whether the level above is itself sound enough to adjudicate rather than merely re-fund the problem.

4. The repair line — including when not to repair

If repair is on the table, the assessment names the matched repair: the fix is chosen by how the system failed (its pathway) and gated by how far it has gone (its stage). A repair aimed at the wrong failure is not just wasted — it is metabolised by the condition. And direction is restored before capability: rebuilding capability on a misoriented system just produces a more capable wrong thing.

The line that no repair-for-hire practice will give you is the honest one when a system is past the door. There, three real options remain, and the assessment will say so rather than sell a turnaround that cannot land:

• Witness — put the condition on the record, in advance, so everyone still bound to the system can see what the dashboards hide; it sometimes moves the level above to act while acting still matters.
• Preserve — carry out the people, knowledge, and working parts the condition has not yet reached.
• Prepare — design the successor, because in this framework collapse is reorganisation at a lower level, and what the successor becomes depends on what was preserved and prepared before the end.

How every reading is stated

Two rules govern every line of an assessment:

• Value with context, never a bare number. The same reading can mean opposite things — a zero can be “healthy and saturated” or “clamped and failing” — so each reading is reported as a (value, context) pair.
• Conduct over words. No conclusion rests on what a system says about itself. Stated language is the weakest class of evidence; the weight is carried by what the system does, allocates, and how it responds when corrected.

What the assessment can and cannot say

Kept attached on purpose, because a reading without its limits is a sales document:

• The instrument is engine-supported — its mechanism is internally coherent and reproduces its own grammar. That is not validation against the world, and we do not call it that.
• The external record — the leakage-free null results and the live forward register — is set out in full in the research update, and is meant to be read as part of this reading, not skipped.
• It reads shock-readiness and trajectory, not fortunes: the shape and direction of what is changing and what the next shock would find — not a date, and not a single overall verdict.
• Claims that depend on a system’s life-cycle phase carry a provisional flag until the model-completeness review closes.

The Great Homecoming is an independent research programme on why systems cohere or fragment. Contact: Wim Van Laere.