[Paper Review] General Logic-Systems that Determine Significant Collections of Consequence Operators

Robert A. Herrmann|arXiv (Cornell University)|Mar 24, 2006

Control and Stability of Dynamical Systems1 references4 citations

TL;DR

This paper introduces general logic-systems to analyze finite consequence operators in a lattice-theoretic framework, demonstrating that the lattice of finite consequence operators on a denumerable language is not meet-complete and has the cardinality of the continuum. It shows that each finite consequence operator is generated by denumerably many general logic-systems, with distinct permutations of inference rules yielding different systems for the same operator.

ABSTRACT

In this paper, general logic-systems and a necessary and sufficient algorithm are used to substantiate significant consequence operator properties. It is shown, among other results, that, in certain cases, (1) if the number of steps in a deduction is restricted, then such deduction does not yield a consequence operator. (2) In general, for any non-organized infinite language L, there is a special class of finite consequence operators that is not meet-complete. (3) For classical deduction, three different examples of modified propositional deduction yield collections of finite consequence operators that are not meet-complete. Other general logic-system examples are given. In a final section, the notion of potentially finite is investigated.

Motivation & Objective

• To investigate finite consequence operators using general logic-systems as a foundational framework.
• To demonstrate that the lattice of finite consequence operators on a denumerable language is not meet-complete.
• To establish that the set of all finite consequence operators on a denumerable language has the cardinality of the continuum.
• To show that each finite consequence operator is generated by a denumerable collection of general logic-systems.
• To explore how language-specific properties influence the behavior of finite consequence operators.

Proposed method

• Utilizes general logic-systems defined by fixed finitary or infinite sets of n-ary inference rules on a language L.
• Applies a deduction algorithm based on pre-axioms and inference rules to generate consequence sets.
• Constructs distinct general logic-systems by permuting the order of premises in inference rules, preserving the generated consequence operator.
• Demonstrates that for each n ≥ 2, there are n! distinct general logic-systems generating the same finite consequence operator.
• Uses non-organized languages—where only word-formation rules are assumed—to avoid embedding unintended algebraic or logical structure.
• Applies the framework to the GGU-model and intelligence measurement via hyperfinite logic-systems and superagents.

Experimental results

Research questions

• RQ1 Can general logic-systems provide a characterization of the lattice-theoretic supremum of a nonempty collection of finite consequence operators?
• RQ2 Is the lattice of finite consequence operators on a denumerable language meet-complete?
• RQ3 What is the cardinality of the set of all finite consequence operators defined on a denumerable language?
• RQ4 How many distinct general logic-systems can generate the same finite consequence operator?
• RQ5 To what extent do properties of finite consequence operators depend on the underlying language structure?

Key findings

• The lattice of finite consequence operators on a denumerable language is not meet-complete, as demonstrated by constructing a nonempty collection whose meet does not exist.
• The set of all finite consequence operators on a denumerable language has the cardinality of the continuum.
• For any finite consequence operator C on a language L, there exist denumerably many distinct general logic-systems that generate C.
• For each n ≥ 2, there are n! distinct general logic-systems that generate the same finite consequence operator C, obtained by permuting the order of premises in inference rules.
• Different permutations of inference rule coordinates yield non-isomorphic general logic-systems, even when generating the same consequence operator.
• The framework supports a formal measurement of intelligence via hyperfinite logic-systems and superagents, where hyper-deduction enables conclusions unreachable by standard agents.