Source-linked AI summary
Evidential-Based Higher-Order Set Argumentation Framework
Shuai Tang
TL;DR
Existing evidential argumentation formalisms do not jointly represent evidential support, higher-order relations, and collective interactions. The paper introduces EHSAF with two complete semantics and propositional and fuzzy encodings, showing that the semantics diverge with support cycles but coincide under support-acyclicity.
Problem
Existing evidence-based formalisms lack a unified treatment of evidential support, higher-order interactions, and collective set-based interactions.
Method
The paper develops EHSAF, two complete semantic approaches, and normal propositional and continuous fuzzy encodings for its structures.
Results
The two EHSAF semantics diverge with support cycles but coincide under support-acyclicity, while encoded models correspond to adjacent complete labellings and fuzzy solutions satisfy core properties.
Takeaways & Limitations
EHSAF provides a single expressive setting for studying complex evidential, higher-order, and collective argumentative structures with qualitative and quantitative semantics.
Takeaways & Limitations
The Łukasiewicz fuzzy semantics is presented as suitable for competitively additive attack and support strengths, but its expanded derivation is omitted.
Abstract
from arXiv · showhide
Evidential argumentation extends Dung's abstract argumentation by requiring arguments and interactions to be backed by chains of evidence rooted in prima-facie elements. However, existing formalisms lack a unified treatment of evidential support, higher-order relations (attacks and supports targeting arbitrary elements), and collective interactions (sources as sets). In this paper, we introduce the Evidential-Based Higher-Order Set Argumentation Framework (EHSAF), which conservatively generalises several existing frameworks within a single expressive setting. We develop two complete semantics for EHSAFs: an \emph{adjacent complete labelling semantics} that admits multiple truth values (true, false, undecided) for arguments in support cycles, reflecting an open epistemic attitude toward future evidence; and an \emph{extension-based complete semantics} that follows a strict evidentialist stance, accepting only arguments with well-founded support chains. We show that these two semantics diverge in the presence of support cycles, and prove their equivalence under support-acyclicity. To enable computational reasoning, we provide a normal propositional encoding of EHSAFs and prove that, in three-valued Łukasiewicz logic, its models correspond precisely to the adjacent complete labellings. We further extend this encoding to continuous fuzzy logics (G{ö}del, Product, and Łukasiewicz), defining a continuous fuzzy normal encoded semantics. We establish that this fuzzy semantics satisfies key properties---continuity, monotonicity, boundary conditions, and solution existence---and that its ternarisation recovers the adjacent complete labellings under natural t-norm conditions. Our framework thus unifies expressive argumentation with principled three-valued and fuzzy semantics, bridging the gap between qualitative and quantitative reasoning about evidence.
1 Introduction
EHSAF unifies evidential support, higher-order interactions, and collective set-based interactions in one framework. It introduces complementary semantics and logical encodings for reasoning about these structures, including support cycles and fuzzy values.
- Motivation: Existing argumentation frameworks separately address evidential support, higher-order relations, or collective interactions, leaving their unified treatment open.Prior formalisms are described as largely isolated or limited to first-order binary interactions.
- Framework: EHSAF combines evidential support rooted in prima-facie elements with higher-order and collective interactions over arbitrary elements and sets.Attacks and supports may target arguments, attacks, or supports, while interaction sources may be sets.
- Framework: The framework conservatively generalizes EAS, REBAF, and HSAF while extending their treatment of interaction sources and support cycles.The resulting setting supports systematic comparison among these existing formalisms.
- Semantics: Adjacent complete labelling semantics allows multiple truth-value assignments for unsupported arguments in evidential support cycles, whereas extension-based semantics adopts a strict evidentialist stance.The adjacent semantics includes true, false, and undecided possibilities when future evidence may emerge.
- Semantics: The two semantic approaches are equivalent under acyclicity conditions, and EHSAF adjacent complete semantics is equivalent to three-valued equational semantics.Their divergence is associated with cyclic support dependencies.
- Logical and fuzzy encodings: Normal propositional and fuzzy encodings provide computational foundations, with fuzzy semantics defined for G{ö}del, Product, and Łukasiewicz t-norms and established core properties.The fuzzy development includes continuity, monotonicity, boundary conditions, solution existence, and correspondence with three-valued adjacent semantics.
2 Preliminaries
The preliminaries review evidential, higher-order, and collective argumentation frameworks, then introduce propositional and fuzzy logic foundations used later.
- Evidential Argumentation Systems: EAS generalizes attacks and supports to nonempty sets of arguments while restricting η from being attacked or supported.The attack and support relations target arguments, with η excluded as an attacker and support target.
- Evidential Argumentation Systems: EAS arguments are valid only when traceable to environmental evidence through evidential support chains.The special prima-facie argument η represents indisputable environmental support, and η is trivially e-supported to avoid support-chain deadlock.
- Recursive Evidence-Based Argumentation Frameworks: REBAFs name attacks and supports, allow set-valued sources, and permit targets to be arguments, attacks, or supports.Their structures record accepted arguments, valid attacks, and valid supports; prima-facie elements require no further support.
- Recursive Evidence-Based Argumentation Frameworks: REBAF acceptability requires evidential support and unactivability of every attack targeting the element.Complete, admissible, preferred, and stable structures retain the usual inclusion relations, and REBAF conservatively generalizes Dung-style frameworks under stated restrictions.
- Propositional Logic Systems: The propositional-logic preliminaries define formulas, connective precedence, three-valued Łukasiewicz truth functions, and continuous fuzzy truth domains.The fuzzy setting uses numerical truth domains containing 0 and 1, with connectives interpreted through negation, t-norms, and residuated implications.
3 Syntax and Basic Semantics of EHSAFs
EHSAFs unify set-valued, higher-order attacks and evidential supports, with labelling and extension-based semantics linked through support and defeat operators.
- EHSAF Syntax: An EHSAF is a finite universal set of arguments, named attacks, and named supports, with prima-facie elements and automatically added auxiliary elements.Sources are nonempty sets of arguments, while targets may be arguments, attacks, or supports.
- Basic Semantics: EHSAF labellings assign values to every universal-set element while fixing bottom at 0, top at 1, and certain auxiliary elements at 1.Equational systems associate each target element with an equation based on source elements and interaction names.
- Extension-Based Semantics: The support operator is monotone and continuous on the complete lattice of subsets, and its least fixed point is obtained by iterating from the empty set.For finite U, the iteration stabilizes after at most |U| steps.
- Semantic Relationship: Extension-based complete labellings always induce adjacent complete labellings, while the converse holds when the EHSAF is support-acyclic.Under support-acyclicity, the two semantics are equivalent.
- Semantic Relationship: Support cycles separate the semantics: adjacent labellings allow true, false, or undecided outcomes, whereas extension-based semantics rejects unsupported cyclic arguments.The distinction reflects tolerant open epistemic and strict evidentialist attitudes, respectively.
4 Encoded Semantics of EHSAFs
The encoded EHSAF construction embeds universal-set elements into propositional constants and variables while fixing auxiliary truth values.
- Encoded EHSAFs: The propositional encoding uses constants and variables so that the universal set U is exactly the set of elements represented in the encoded formula.Bottom is interpreted as 0, top as 1, and interaction names absent from the original relations as 1.
4.1 General Encoded Semantics
General encoded semantics maps EHSAFs to propositional formulas whose models are labellings, with skeptical and credulous variants aggregating all models.
- General Encoded Semantics: An EHSAF encoding is a propositional formula whose variables and distinguished constants coincide exactly with the universal set U.The induced encoded semantics treats models of the formula as valid EHSAF labellings.
- General Encoded Semantics: Skeptical encoded semantics assigns each element its minimum truth value across all basic encoded models, while credulous semantics assigns the maximum.These values correspond to the infimum and supremum of possible interpretations and produce unique labellings.
- Normal Encoded Semantics: Normal encoded semantics is induced by a normal encoding function and the selected propositional logic system.The normal encoding and its semantics are introduced as distinct constructions within the general encoded framework.
4.2 Discrete Encoded Semantics
This section presents theorem-level results connecting discrete encoded semantics with EHSAF semantics, including a corollary derived from preceding theorems.
- Theorem 4 states a relationship for every EHSAF between the relevant discrete semantic constructions.
- A displayed result identifies ecn(EHSAF) with LSac(EHSAF).
- Corollary 3 is presented for every EHSAF and follows immediately from Theorem 1 and Theorem 4.
- Theorem 5 provides another universal result for EHSAFs concerning the discrete encoded semantics.
4.3 General Fuzzy Encoded Semantics and Its Core Properties
The general fuzzy encoded framework defines continuous operator-based equations and encoded semantics for EHSAFs, then establishes their equivalence and core structural properties.
- The CFOE system assigns values in [0, 1] using a continuous negation and continuous t-norm over attack and support conditions.Assignments satisfying all equations are solutions, and CFOE semantics consists of those solutions.
- CFNE semantics is induced by the EHSAF normal encoding and a fuzzy propositional logic system with continuous operators.
- Theorem 6 proves that CFOE and CFNE semantics coincide for every EHSAF.
- The aggregation function fβ is continuous and permutation-invariant with respect to variables within source sets.
- fβ is non-increasing in attack-related elements and non-decreasing in support-related elements.This reflects the stated roles of attacks and evidential supports in determining acceptability.
- Every EHSAF has at least one CFOE solution and at least one corresponding CFNE model.Existence follows by applying Brouwer’s Fixed-Point Theorem to the continuous self-map on [0, 1]^n.
- Under continuous negation and a zero-divisor-free t-norm, ternarising CFNE models yields adjacent complete labellings, and the converse set correspondence holds under the stated conditions.
4.4 Key Instances of Fuzzy Encoded Semantics
The framework instantiates fuzzy encoded semantics with Gödel, Product, and Łukasiewicz t-norms, yielding distinct aggregation interpretations while retaining general semantic properties.
- Gödel Fuzzy Encoded Semantics: Gödel semantics uses minimum aggregation, so collective attacks follow a weakest-link principle and targets are constrained by the strongest attack.
- Gödel Fuzzy Encoded Semantics: Gödel fuzzy encoded semantics is equivalent to its CFOE counterpart and inherits continuity, boundary conditions, monotonicity, and solution existence.
- Product Fuzzy Encoded Semantics: Product semantics multiplies values, modeling probabilistic independent aggregation across multiple attacks and supports.
- Product Fuzzy Encoded Semantics: Product CFOE and CFNE semantics coincide, while continuity, boundary conditions, monotonicity, and solution existence continue to hold.
- Product Fuzzy Encoded Semantics: Unlike Gödel aggregation, Product aggregation is non-idempotent, allowing multiple independent attacks and supports to accumulate gradually.
- Łukasiewicz Fuzzy Encoded Semantics: Łukasiewicz semantics provides threshold-based additive aggregation for competitively additive attack and support strengths.
- Łukasiewicz Fuzzy Encoded Semantics: The detailed Łukasiewicz derivation is omitted because its expanded form is cumbersome and its core properties follow from the general framework.
5 Related Work
EHSAF unifies evidential support, higher-order targets, and collective sources while conservatively extending established argumentation frameworks. Its propositional and fuzzy encodings connect these expressive structures to logical and gradual semantics.
- Framework relationships: EHSAF generalizes EAS, REBAF, and HSAF within a unified syntax for evidential, higher-order, and collective interactions.Its sources may be arbitrary sets of arguments, attacks, or supports, and its targets may be arbitrary elements.
- Framework relationships: REBAF is a proper subframework of EHSAF because EHSAF permits sources containing attacks and supports, not only arguments.Under argument-only source restrictions, EHSAF’s extension-based semantics coincide with REBAF’s structure semantics.
- Labelling semantics: EHSAF differs from incremental evidence-based semantics by prioritizing expressive uniformity over dynamic-update algorithms.Chen et al.’s approach is restricted to first-order argument-to-argument interactions, whereas EHSAF supports arbitrary sources and targets and adds logical and fuzzy encodings.
- Labelling semantics: EHSAF’s adjacent complete labellings align with evidential-support labellings from the taxonomy of complete bipolar semantics under first-order singleton restrictions.Both approaches use complete labellings and include evidential interpretations of support.
- Logical encodings: The normal propositional encoding preserves semantic correspondences for REBAF and HSAF restrictions while extending them to arbitrary-element sources and targets.It provides a propositional counterpart to first-order encodings and enables direct use of SAT solvers and related propositional reasoning tools.
- Fuzzy semantics: EHSAF’s fuzzy semantics gives attacks decreasing and supports increasing interpretations through continuous, monotone aggregation, unlike threshold-based fuzzy approaches.Under suitable t-norm restrictions, it recovers three-valued complete semantics on the ternary domain.
6 Conclusion
The paper introduces EHSAF as a unified framework for evidential, higher-order, and collective argumentation, together with two complete semantics. It establishes their relationship under support acyclicity and supplies propositional and fuzzy encodings with corresponding semantic guarantees.
- Conclusion: EHSAF integrates evidential support, higher-order interactions, and collective sources while conservatively generalizing EAS, REBAF, and HSAF.The framework provides one expressive setting for complex argumentative structures.
- Conclusion: Two complete semantics differ on support cycles: adjacent labellings allow true, false, or undecided values, whereas extension semantics require well-founded support chains.The first adopts an open epistemic attitude; the second follows a strict evidentialist stance.
- Conclusion: Under support-acyclicity, the adjacent labelling and extension-based semantics coincide, so their divergence is attributable to cyclic dependencies.The equivalence is established as a property of support-acyclic EHSAFs.
- Conclusion: In three-valued Łukasiewicz logic, models of the normal propositional encoding correspond exactly to adjacent complete labellings.The encoding supplies a logical basis for computational reasoning about EHSAFs.
- Conclusion: The continuous fuzzy encoding satisfies continuity, monotonicity, boundary conditions, and solution existence, while its ternarisation recovers adjacent complete labellings under specified t-norm conditions.Future work includes incremental computation, preference or weighted extensions, and applications to legal and scientific evidence reasoning.