A System of Interaction and Structure IV: The Exponentials and Decomposition

We study a system, called NEL, which is the mixed commutative/non-commutative linear logic BV augmented with linear logic’s exponentials. Equivalently, NEL is MELL augmented with the non-commutative self-dual connective seq. In this paper, we show a basic compositionality property of NEL, which we call decomposition. This result leads to a cut-elimination theorem, which is proved in the next paper of this series. To control the induction measure for the theorem, we rely on a novel technique that extracts from NEL proofs the structure of exponentials, into what we call !-?-Flow-Graphs.

💡 Research Summary

This paper introduces NEL (Non‑commutative Exponential Logic), a logical system that combines the self‑dual non‑commutative connective “seq” from the BV fragment of linear logic with the exponential modalities (!, ?) of MELL (Multiplicative‑Exponential Linear Logic). In effect, NEL can be seen as MELL enriched with a non‑commutative, self‑dual connective, or equivalently BV augmented with the usual linear‑logic exponentials. The authors’ primary contribution is the proof of a compositionality property they call decomposition.

The decomposition theorem states that any NEL proof can be split into two independent sub‑proofs: one that uses only the non‑commutative fragment (seq together with the multiplicative connectives) and another that uses only the commutative fragment together with the exponentials. This separation is non‑trivial because the seq connective does not satisfy the exchange rule, so the usual techniques for moving exponentials across the proof structure do not apply. By isolating the exponential part, the theorem paves the way for a cut‑elimination proof (presented in the subsequent paper of the series) that can control the growth of proof size.

To achieve this separation the authors introduce a novel graphical tool called the !-?-Flow‑Graph. In this directed acyclic graph each node corresponds to an occurrence of an exponential rule (!‑weakening, !‑contraction, !‑dereliction, or their dual ?‑rules) and each edge records a dependency created by duplication or erasure of a sub‑proof. The crucial property of the flow graph is that it contains no cycles; consequently one can define a well‑founded measure (the maximal length of a path) that serves as an induction metric for the decomposition proof and later for cut‑elimination. The acyclicity also guarantees that the exponential structure can be “peeled off” from the non‑commutative part without interfering with the sequential ordering imposed by seq.

The paper proceeds as follows. After a concise motivation (the need to reason about resources that are both ordered and duplicable), the syntax of NEL is presented: formulas are built from the usual multiplicatives (⊗, ⅋, 1, ⊥), the self‑dual seq connective, and the exponentials. Proofs are described in a sequent‑style calculus where sequents have a two‑layered structure (Γ ; Δ), with Γ handling the non‑commutative layer and Δ the commutative one. The inference rules are listed, emphasizing that the seq rules forbid exchange, while the exponential rules are the standard MELL rules.

The core technical section defines the !‑?‑Flow‑Graph, proves its acyclicity, and shows how to use it to systematically reorder the proof so that all exponential rules appear either at the top or the bottom of the derivation. This reordering yields the decomposition: the “seq‑only” fragment is obtained by erasing the exponential nodes, and the “exponential‑only” fragment by collapsing the seq‑layer. The authors give a detailed inductive argument, using the depth of the flow graph as the induction parameter, to demonstrate that every NEL proof can be transformed into this normal form without altering its conclusion.

Finally, the authors discuss the implications of decomposition. Because the two fragments are now independent, existing cut‑elimination procedures for BV and for MELL can be applied separately, and their results can be combined to obtain a full cut‑elimination theorem for NEL. The paper also sketches possible applications: modeling concurrent processes where ordering matters, designing type systems for languages that need both linear usage and controlled duplication, and exploring the computational complexity of NEL (the flow‑graph measure provides a natural bound).

In summary, the paper establishes a solid theoretical foundation for NEL by defining its syntax, introducing the !‑?‑Flow‑Graph, proving the decomposition theorem, and outlining how this theorem enables a subsequent cut‑elimination result. The work represents a significant step toward integrating non‑commutative and exponential features in a single logical framework, opening avenues for both proof‑theoretic investigations and practical applications in programming language semantics and concurrency theory.