Derivable Architectures — Volume II
Paper 017 — Branching of Derivations
Research Program: Derivable Architectures
Volume: II — Structural Theory
Document: papers/017-branching-of-derivations.md
Status: Draft 0.1
Authors
Eduardo N. Hering OpenAI ChatGPT
Abstract
Previous papers established derivation spaces, composition, partial derivations, constraints, equivalence, and reversibility.
This paper studies how derivations diverge during implementation.
Rather than viewing branching as merely a practical consequence of engineering decisions, we show that branching is an intrinsic structural property of derivation spaces.
Each architectural decision partitions the remaining space of admissible realizations.
Implementation therefore becomes a progressive reduction of architectural possibility.
1. Observation
At the beginning of derivation, many implementations remain possible.
As derivation proceeds,
choices are made.
Each choice eliminates some future realizations while preserving others.
Implementation therefore progressively specializes the architecture.
2. Observation
The ICA demonstrates this immediately.
Initially,
multiple jurisdictions are admissible.
Once Brazil is selected,
all remaining derivations must satisfy Brazilian law.
The derivation space immediately becomes smaller.
The architecture itself has not changed.
Only the remaining admissible derivations have changed.
3. Definition
Definition 1 — Branch Point
A branch point is a derivational state from which two or more architecturally valid continuations exist.
Each continuation preserves architectural identity.
Each continuation leads toward distinct regions of the derivation space.
4. Proposition
Branch points are structural properties of derivation spaces.
Proof Sketch
Branching depends upon the existence of multiple admissible continuations.
These continuations are determined by architectural constraints,
not by engineering preference.
Therefore,
branch points belong to the derivation space induced by the architecture.
∎
5. Observation
Not every decision constitutes a branch.
Many implementation decisions are completely determined once previous architectural commitments have been made.
Such decisions introduce no structural choice.
6. Definition
Definition 2 — Forced Derivation
A forced derivation is a derivational step for which only one architecturally admissible continuation exists.
Forced derivations contain no branching.
7. Observation
Derivations therefore alternate between
periods of architectural freedom
and
periods of architectural determinism.
Implementation is not uniformly constrained.
8. Proposition
The amount of architectural freedom decreases monotonically along every complete derivation.
Proof Sketch
Each valid decision either
preserves
or reduces
the remaining admissible derivations.
Previously eliminated continuations cannot reappear without violating derivational consistency.
Therefore,
the remaining derivation space cannot increase.
∎
9. Observation
This monotonic reduction of possibility explains an engineering intuition.
Architectural decisions become increasingly expensive to reverse.
The reason is not implementation cost alone.
The derivation has already eliminated large regions of admissible realization.
10. Definition
Definition 3 — Remaining Derivation Space
The remaining derivation space at a given derivational state consists of every architecturally admissible continuation still available.
This quantity decreases throughout implementation.
11. Corollary
Implementation may be viewed as successive navigation through shrinking derivation spaces.
The architecture remains constant.
The reachable future changes.
12. Observation
This perspective naturally separates
architectural commitment
from
implementation execution.
Commitment reduces future possibilities.
Execution realizes already committed decisions.
13. Conjecture
Branching may provide the fundamental measure of architectural flexibility.
Architectures possessing numerous well-structured branch points may support broad implementation diversity without sacrificing architectural identity.
14. Open Questions
Can branching complexity be measured?
Can branch points be automatically identified?
Can equivalent branch structures exist across different architectures?
Can branching predict implementation effort?
Can optimal decision sequences maximize future flexibility?
Conclusion
Branching transforms derivation from a simple sequence into a decision structure.
Implementation becomes the progressive resolution of architectural alternatives.
Each derivational decision simultaneously realizes one future while excluding others.
The architecture itself remains unchanged.
Only the remaining admissible realization space evolves.