<aside> ๐ก
ANIT is a structural model that enables the complete reorganization of literary works without altering their textual materiality at any level (lexical, syntactic, or semantic).
The model operates under the principle that narrative meaning is not inherent to textual units themselves, but emerges from their structural arrangement.
It demonstrates that interpretative outcomes can be transformed exclusively through reordering, while preserving every original textual unit in its exact form.
</aside>
The model is empirically demonstrated through the literary work:
In this work, the same textual corpus is reorganized into two inverse narrative structures, producing distinct interpretative trajectories without any modification at the lexical, syntactic, or semantic level.
ANIT is not limited to a single language.
The model operates across linguistic systems, preserving textual integrity while reorganizing how narrative is experienced.
A Portuguese structural application is embedded within the literary project: Lรขmina: A Lรขmina entre o amor e as sombras โ poesias e contos.
The validity of ANIT is established through two conditions:
Textual Invariance No insertion, deletion, or modification of any textual element is permitted.
Interpretative Divergence The reordered structure must produce a demonstrably distinct narrative experience.
Let:
T = {tโ, tโ, tโ, ..., tโ} be a finite set of invariant textual units.
Let ฯ be a permutation defined over T.
A narrative structure S is defined as an ordered sequence of T:
S = (tโ, tโ, tโ, ..., tโ)
ANIT allows the transformation:
Sโ = ฯ(Sโ)
such that:
โ t โ T โ t remains textually unchanged
and
I(Sโ) โ I(Sโ)
where I(S) represents the interpretative outcome of structure S.
To ensure strict invariance, ANIT allows the use of cryptographic hashing to verify that all structural variations maintain identical textual materiality.
This enables:
ANIT does not generate new text and does not rely on semantic rewriting.
Its domain is strictly structural.
Any transformation outside reordering invalidates the model.
Working paper (2026) โ under independent distribution and open-access repositories.