Expressiveness Preorder in Formal Systems

• Expressiveness preorder is a reflexive and transitive relation that defines one formal system’s ability to encode or simulate another based on specific semantic criteria.
• It provides a uniform framework for comparing diverse domains such as programming languages, process calculi, logics, and graph types through rigorous equivalence measures.
• Its application offers practical insights into establishing hierarchies, guiding automated analyses, and addressing open classification challenges in theoretical computer science.

An expressiveness preorder is a reflexive and transitive relation on a class of formal systems or structures, indicating that one system can encode or simulate another according to specified semantic or syntactic criteria. This concept provides a uniform mathematical framework for formalizing and comparing the expressive power of a wide array of objects: programming languages, process calculi, formal logics, and even graph types. Formalized in seminal works such as van Glabbeek’s framework for process calculi, rainbow forbidden subgraphs in extremal combinatorics, and subexponential logics, the expressiveness preorder clarifies the hierarchy of encodability and the precise boundary between what can and cannot be simulated in a given paradigm (Glabbeek, 2012, Glabbeek, 2018, Ford et al., 2020, Maezawa et al., 2 Feb 2025, Maezawa, 9 Jan 2026, Chaudhuri, 2010, Thomas et al., 2021).

1. Formal Definition and Mathematical Properties

A canonical instantiation of the expressiveness preorder arises in the comparison of system description languages, as in van Glabbeek (Glabbeek, 2012, Glabbeek, 2018). Given two such languages

$$L_1=(\mathbb{T}_1, \llbracket - \rrbracket_1), \qquad L_2=(\mathbb{T}_2, \llbracket - \rrbracket_2)$$

with respective sets of terms and semantic mappings to domains $D_1$, $D_2$, one fixes a semantic equivalence or preorder $\sim\subseteq (D_1\cup D_2)^2$. $L_1\leq_\sim L_2$ iff there exists a translation $\tau:\mathbb{T}_1\to\mathbb{T}_2$ such that:

• (Relatedness) For all $d\in D_1$, there exists $d'\in D_2$ with $d'\sim d$.
• (Correctness) For all $E\in\mathbb{T}_1$, for all valuations $D_1$0, $D_1$1 with $D_1$2, $D_1$3.

This relation is reflexive (by the identity translation) and transitive (by composition of encodings), provided $D_1$4 has the requisite congruence properties. Thus, $D_1$5 is a preorder on the class of languages or systems under study (Glabbeek, 2012, Glabbeek, 2018).

Analogous definitions are used in other domains. For example, for connected graphs $D_1$6 in rainbow subgraph theory, the preorder $D_1$7 holds iff for some threshold $D_1$8, every $D_1$9-coloring of a complete graph that is rainbow $D_2$0-free is also rainbow $D_2$1-free (Maezawa et al., 2 Feb 2025, Maezawa, 9 Jan 2026). Similar structures organize classes of logics ($D_2$2 if every theorem of $D_2$3 is encodable in $D_2$4 under some translation) or graph types (Chaudhuri, 2010, Thomas et al., 2021).

2. Key Examples Across Domains

The notion of expressiveness preorder is instantiated in several research areas.

• Process calculi (CCS, CSP, $D_2$5-calculus): $D_2$6 formalizes the intuition that $D_2$7 can simulate $D_2$8 up to a semantic equivalence, e.g., trace equivalence or weak bisimulation. Van Glabbeek’s framework rigorously formalizes this, allowing for fine-grained hierarchies depending on the choice of $D_2$9 (Glabbeek, 2012, Glabbeek, 2018).
• Rainbow subgraph theory: For graphs, $\sim\subseteq (D_1\cup D_2)^2$0 if rainbow $\sim\subseteq (D_1\cup D_2)^2$1-free edge-colorings enforce rainbow $\sim\subseteq (D_1\cup D_2)^2$2-freeness in complete graphs, yielding a nontrivial preorder whose structure encodes combinatorial containment phenomena (Maezawa et al., 2 Feb 2025, Maezawa, 9 Jan 2026).
• Subexponential logics: Expressiveness preorders order subexponential signatures according to focal-adequate encodings. Equivalence holds (modulo signature-splitting) between classical and intuitionistic logics, refuting the naive view that logic classes are strictly ordered by expressive power (Chaudhuri, 2010).
• Graph type modeling: A graph type $\sim\subseteq (D_1\cup D_2)^2$3 is at most as expressive as $\sim\subseteq (D_1\cup D_2)^2$4 ($\sim\subseteq (D_1\cup D_2)^2$5) if every structure in $\sim\subseteq (D_1\cup D_2)^2$6 can be embedded injectively with preservation of all structural and attribute data. All attributed graph types collapse to a top equivalence class (Thomas et al., 2021).

3. Structural Refinements, Congruence, and Hierarchies

The expressiveness preorder one obtains depends critically on auxiliary parameters:

• Choice of semantic equivalence $\sim\subseteq (D_1\cup D_2)^2$7 or preorder: Finer $\sim\subseteq (D_1\cup D_2)^2$8 (e.g., bisimulation) yield stricter preorders and admit fewer encodings; coarser $\sim\subseteq (D_1\cup D_2)^2$9 (e.g., trace equivalence) lead to more comparabilities (Glabbeek, 2012).
• Congruence requirements: For preorders to be robust, $L_1\leq_\sim L_2$0 must be a congruence on at least the image of translations, sometimes on all process expressions (Glabbeek, 2012, Glabbeek, 2018).
• Equivalence class quotients and partial orders: The induced equivalence $L_1\leq_\sim L_2$1 iff $L_1\leq_\sim L_2$2 and $L_1\leq_\sim L_2$3 enables the formation of a poset structure on equivalence classes (Maezawa et al., 2 Feb 2025, Maezawa, 9 Jan 2026).

Hierarchies can be strict and nontrivial. For example, the rainbow-subgraph preorder organizes tree classes, with only specific (star-plus-edge, etc.) pairs lying in non-singleton equivalence classes; almost all pairs are incomparable (Maezawa, 9 Jan 2026).

4. Relation to Alternative Notions and Criteria

Van Glabbeek’s approach to expressiveness preorders subsumes or contrasts with several alternative frameworks:

• Compositional/operational encodings: Gorla’s five criteria (compositionality, name-invariance, operational correspondence, divergence reflection, success sensitiveness) are shown to be captured by van Glabbeek’s correctness condition for suitable choices of semantic equivalence (Glabbeek, 2018).
• Full abstraction: The preorder definition generalizes Milner’s full-abstraction paradigm by organizing languages with encodings that preserve observations equivalently (Glabbeek, 2012, Glabbeek, 2018).
• Classical/intuitionistic logics: Subexponential logics, parameterized by preorders on exponentials, admit a preorder (modulo focal adequacy and signature translation) that collapses via bijective correspondences between derivations (Chaudhuri, 2010).

Expressiveness preorders thereby provide a language- and context-independent unification of prior, more ad hoc, expressibility comparisons.

5. Case Studies and Explicit Classifications

Significant case studies elucidate the subtlety of expressiveness preorders:

• CCS vs. CSP: There exists a compositional encoding of CSP into CCS up to trace equivalence, but no encoding up to convergent weak bisimilarity; fragment-level equivalence under bisimilarity is possible on acyclic processes (Glabbeek, 2012).
• Rainbow subgraphs (trees): Apart from the family $L_1\leq_\sim L_2$4, no two non-isomorphic trees are equivalent in the preorder; candidate edge-swap pairs almost always fail comparability (Maezawa, 9 Jan 2026).
• Subexponential logics: Classical and intuitionistic fragments (modulo possible preorder and signature-splitting modifications) are strictly equivalent in expressiveness; bijections are constructed between focused proof derivations (Chaudhuri, 2010).
• Graph types: Once arbitrary attributes are permitted, all differences between directed, undirected, hyper, multi-, heterogeneous, and dynamic graphs disappear from the standpoint of modeling power (Thomas et al., 2021).

Domain	Preorder Notation	Distinguished Features
Process Calculi	$L_1\leq_\sim L_2$5	Parametric in $L_1\leq_\sim L_2$6; encodings are syntax/semantics preserving
Rainbow Subgraphs	$L_1\leq_\sim L_2$7	Based on forbidden rainbow subgraphs in edge-colored graphs
Subexponential Logics	$L_1\leq_\sim L_2$8	Focal-adequacy of translations; signature-splitting
Graph Types	$L_1\leq_\sim L_2$9	Structural and attribute-preserving embeddings

6. Limitations, Algorithmic Aspects, and Open Problems

Construction of explicit encodings, identification of equivalence classes, and practical decision procedures are in general nontrivial:

• Congruence and domain assumptions: Expressiveness preorders require that all systems share enough semantic structure for meaningful comparison. In some frameworks, auxiliary side-conditions (freshness, name invariance) must be imposed (Glabbeek, 2012).
• Decidability and algorithms: For rainbow forbidden subgraphs, efficient decision is tractable for graphs up to moderate order using invariants, though a general complexity dichotomy remains open (Maezawa et al., 2 Feb 2025).
• Open classification problems: Determining maximal elements, immediate successors, or full Hasse diagrams for specific preorders (e.g., on trees, graphs, or logical fragments) is largely open (Maezawa et al., 2 Feb 2025, Maezawa, 9 Jan 2026).
• Automation: Future directions include automating congruence checks and structural analyses as part of automated expressiveness proofs (Glabbeek, 2012).

A plausible implication is that expressiveness preorders, while robust and uniform, are sensitive to the precise choice of semantic invariants and often hinge on subtle structural or combinatorial properties. Their study continues to generate nontrivial technical and classification challenges across logic, concurrency, and combinatorics.

7. Impact and Future Directions

The introduction and rigorous development of expressiveness preorders have unified disparate strands of comparative system theory:

• Uniform foundations: They provide precise, parameterized tools for comparing widely different systems on equal footing, clarifying hierarchy, equivalence, and incomparability.
• Paradigm-invariant unification: By abstracting away from specifics of syntax or operational semantics, expressiveness preorders streamline the analysis of inter-system encodability problems.
• Ongoing research: Key directions include the extension to higher-order calculi, systematic comparison with alternative frameworks such as Gorla’s criteria, full classification of rainbow-subgraph classes, and further algorithmization for expressiveness checking (Glabbeek, 2012, Maezawa, 9 Jan 2026).

Expressiveness preorders now stand as central organizing principles in both the meta-theory of programming languages and in combinatorial and logical modeling, supporting robust and extensible comparative analysis across formal systems.