01 / A forest facing an uncertain future
A forest has more than one possible future.
Let's return to our forest. Imagine it under stress, with the present organization still in place.
Its future might unfold in several ways. It may remain stressed for a time and still continue as the forest we've specified. It may regenerate. It may adapt into a different condition while preserving the required organization. Or it may collapse, losing that organization.
I find this a useful place to begin because continuation does not demand that everything return to an earlier condition. The question concerns the organization that we have defined and the rules under which it can continue.
The formal forest model makes those possibilities explicit. Its small number of states lets us follow each kind of history carefully. The pathways below come from the declared five-state example; they are alternatives across different contexts and horizons, not predictions that all four pathways are available at once.
A lawful stay history can still satisfy continuation.
A lawful, continuing history under mild, long-horizon conditions.
The declared organization is retained through a changed composition.
The model permits this outcome under severe, long-horizon conditions, but it fails continuation.
Formal illustration only. The model's history constructors determine which transitions exist in each scenario and horizon. The more explicit permissive-candidate/lawfulness-filter encoding is developed in Paper IV. No ecological forecast or empirical probability is depicted here.
Already we can see why a mere inventory of imaginable outcomes is too loose. We need to say which histories the system's formal rules permit, and then identify which of those preserve the organization we are evaluating.
02 / Building the viable family
Now let's make the filtering rule precise.
For a starting state x, the model describes histories leading to possible target states y under scenario d and time horizon t. I call an individual history γ (gamma). The complete collection of viable continuations is:
The viable-continuation family
Start with histories. Keep the continuing ones.
Paper I, formal viable-continuation construction. The collection is indexed by scenario and horizon; its members carry both lawfulness and continuation conditions.
There is a lot of notation in one line, but each symbol is doing familiar work.
- Vd,t(x)
- The complete collection of viable continuations from the starting state.
- x and y
- The initial state and a possible endpoint.
- γ
- A particular history or pathway connecting the states.
- Hist
- The model's family of histories for the declared scenario, horizon, and endpoints.
- Lawful(γ)
- That history satisfies the declared admissibility rules.
- Continues(γ)
- That history satisfies the constitution's requirement for continued organization.
- ∈
- “Belongs to.” The braces
{ }collect the entries meeting the stated conditions.
What matters is the construction. First, we have histories belonging to the model's declared history family. Then we require lawfulness and continuation. The resulting viable family includes exactly the histories that satisfy both conditions.
We now have a precise object upstream of any summary we might later call a system's capacity.
03 / Two questions with different jobs
A lawful future can still be a failure.
One of the most important choices in the framework is to keep lawfulness and continuation distinct. Lawfulness concerns whether the history belongs to the admissible possibilities of the declared model. Continuation concerns whether that history preserves the organization we're evaluating.
In the forest example, severe conditions can permit collapse. That collapse is lawful in its declared scenario. It is also a failure of continuation, because the target state no longer realizes the forest organization. The model must represent that possible loss without pretending that a lawful event is automatically healthy.
| History | Lawful? | Continues? | Viable? |
|---|---|---|---|
| Stays stressed | Yes | Yes | Yes |
| Regenerates under mild, long-horizon conditions | Yes | Yes | Yes |
| Collapses under severe, long-horizon conditions | Yes | No | No |
| Regenerates immediately under a short horizon | No | Not evaluated | No |
The last row uses the explicit candidate-lawfulness filter of Paper IV: an imagined short-horizon regeneration candidate is inadmissible. In Paper I's context-indexed history encoding, the corresponding history constructor is simply unavailable at that horizon. These are two ways of expressing the same boundary in the finite model, not different observed forest processes.
This separation is powerful. It lets the mathematics describe failure honestly. Some histories may be possible under the declared rules and nevertheless end in the loss of the organization. Those histories are kept distinct from viable continuations.
04 / What the formal result establishes
The viable family is complete relative to the declared rules.
Paper I makes this selection exact. The lawful-history family can be written as L, and the viable family has an equivalent description: lawful histories that satisfy continuation.
The lawful-history characterization
Two descriptions of the same viable content
The equivalence symbol ≃ records a formal equivalence of the declared collections. It is a theorem about the specified mathematical history system.
For me, the value of the result is that we can identify exactly which conditions admit a history into the viable family. There are no extra viable histories outside the stated criteria within this specification.
The conclusion is rigorous and deliberately bounded. It does not mean that we have enumerated every possible ecological future of a real forest. In real measurement, even identifying the correct state, dynamics, and continuation conditions is a scientific task in its own right.
05 / The next layer
A viable continuation is not the same thing as capacity.
We have assembled the viable histories. The next question is how much of their information a chosen observation needs to retain.
Suppose our forest model contains five distinct viable histories under a declared scenario and horizon. Now imagine an observer that reports only two kinds of response: stable and exposed. Multiple histories may produce the same response label.
The five-to-two illustration is a deliberately simple hypothetical collector. It is not a claim about the exact number of histories or response atoms in the full forest census, and it does not assign probabilities.
This is why I distinguish the family of viable continuations from the capacity representation obtained by observing and collecting them. In formal notation:
From viable histories to represented capacity
Observe, then collect.
The observer o maps each viable continuation to a response; collect forms the capacity representation. The content of that representation depends on the declared specification.
Five underlying histories can thus contribute only two distinct response categories. That compression can be useful: we may not need to retain every detail of every history to evaluate a particular requirement. But if the observer erases a distinction that matters to the question, the resulting capacity representation will be insufficient.
That question—what information a health judgment must preserve—will become central in Papers III and IV.