Fully Evaluated Left-Sequential Logics
Pith reviewed 2026-05-24 03:13 UTC · model grok-4.3
The pith
Four fully evaluated left-sequential logics form a hierarchy distinguished by memorization and conditionals, each with complete axiomatizations over evaluation trees.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The paper establishes four two-valued FELs ordered by strength: Free FEL (immune to atomic side effects), Memorising FEL (stores prior evaluations), Conditional FEL (incorporates conditional evaluation), and Static FEL (equivalent to sequential propositional logic). Evaluation trees provide the distinguishing semantics, and each logic has a complete axiomatization. The three-valued versions (with constant U) inherit the two-valued axioms plus two for undefinedness, and the strongest three-valued FEL coincides with Bochvar's strict logic.
What carries the argument
Evaluation trees, which encode the left-to-right evaluation steps of propositional expressions and separate the logics according to whether results are discarded, memorized, or conditioned.
If this is right
- Any equivalence between closed terms that holds in a given FEL can be derived from its axioms.
- The three-valued axiomatizations are sound and complete for reasoning with undefinedness under the respective sequential regimes.
- Static FEL supplies a sequential presentation of Bochvar's strict logic.
- The hierarchy isolates the minimal additions (memorization, conditionals) needed to recover classical behavior from a side-effect-free base.
Where Pith is reading between the lines
- These systems could serve as precise models for short-circuit operators in programming languages where evaluation order and side effects matter.
- The evaluation-tree semantics might extend to first-order or quantified sequential logics while preserving the completeness results.
- Embedding the weaker FELs into the stronger ones could yield conservative extension theorems useful for stepwise verification.
- The link to Bochvar's logic suggests possible translations between sequential undefinedness and other three-valued systems.
Load-bearing premise
That evaluation trees correctly and exhaustively capture the intended semantics of left-sequential evaluation without missing relevant distinctions or side-effect behaviors.
What would settle it
A closed term whose truth value or equivalence class under the axiomatization differs from the value obtained by direct traversal of its evaluation tree.
read the original abstract
We consider a family of two-valued "fully evaluated left-sequential logics" (FELs), of which Free FEL (defined by Staudt in 2012) is most distinguishing (weakest) and immune to atomic side effects. Next is Memorising FEL, in which evaluations of subexpressions are memorised. The following stronger logic is Conditional FEL (inspired by Guzm\'an and Squier's Conditional logic, 1990). The strongest FEL is static FEL, a sequential version of propositional logic. We use evaluation trees as a simple, intuitive semantics and provide complete axiomatisations for closed terms (left-sequential propositional expressions). For each FEL except Static FEL, we also define its three-valued version, with a constant U for "undefinedness" and again provide complete, independent axiomatisations, each one containing two additional axioms for U on top of the axiomatisations of the two-valued case. In this setting, the strongest FEL is equivalent to Bochvar's strict logic.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript defines a family of two-valued fully evaluated left-sequential logics (FELs): Free FEL (weakest, side-effect immune), Memorising FEL, Conditional FEL, and Static FEL (strongest, sequential propositional logic). It employs evaluation trees as semantics for closed left-sequential propositional expressions and asserts complete axiomatizations for each two-valued variant; for the three-valued versions (with undefinedness constant U) of the first three variants it supplies two additional axioms each and claims the strongest three-valued FEL coincides with Bochvar's strict logic.
Significance. If the completeness claims hold, the work supplies a uniform semantic and axiomatic treatment of left-sequential evaluation regimes that differ in memorization and side-effect handling, extending Staudt (2012) and Guzmán-Squier (1990). The evaluation-tree semantics is presented as intuitive and the independence of the added U axioms is a potential strength, but neither machine-checked proofs nor reproducible derivations are supplied.
major comments (2)
- [Abstract] Abstract and main text: the central claim that complete axiomatizations are provided for each FEL variant (two-valued and three-valued) is unsupported because neither the explicit axiom sets nor any proof sketch (inductive argument that every tree-valid formula is derivable) appears in the manuscript; soundness and completeness therefore cannot be verified.
- The load-bearing assumption that evaluation trees correctly distinguish the four evaluation regimes (Free, Memorising, Conditional, Static) and induce the intended consequence relations is stated but not accompanied by explicit verification or counter-example checks that would confirm the distinctions are non-trivial and that the strongest three-valued case is indeed Bochvar-strict.
Simulated Author's Rebuttal
We thank the referee for the careful reading and constructive comments. Below we respond point by point to the major comments and indicate the revisions we will make.
read point-by-point responses
-
Referee: [Abstract] Abstract and main text: the central claim that complete axiomatizations are provided for each FEL variant (two-valued and three-valued) is unsupported because neither the explicit axiom sets nor any proof sketch (inductive argument that every tree-valid formula is derivable) appears in the manuscript; soundness and completeness therefore cannot be verified.
Authors: We agree that the manuscript would benefit from greater explicitness. Although the abstract asserts that complete axiomatizations are supplied, the full axiom sets for each variant and the corresponding completeness proofs (or even sketches) are not presented in the main text. In the revised version we will list the complete axiom sets for Free FEL, Memorising FEL, Conditional FEL, Static FEL and the three-valued extensions, together with inductive proof sketches establishing that every evaluation-tree valid formula is derivable. revision: yes
-
Referee: [—] The load-bearing assumption that evaluation trees correctly distinguish the four evaluation regimes (Free, Memorising, Conditional, Static) and induce the intended consequence relations is stated but not accompanied by explicit verification or counter-example checks that would confirm the distinctions are non-trivial and that the strongest three-valued case is indeed Bochvar-strict.
Authors: The evaluation-tree semantics is defined so that the four regimes differ by construction in their handling of memorization and side effects. We nevertheless accept that explicit verification would strengthen the claim. In the revision we will add concrete counter-examples and small-model checks showing non-trivial distinctions among the four two-valued regimes and confirming that the strongest three-valued FEL coincides with Bochvar's strict logic. revision: yes
Circularity Check
Minor self-citation on base Free FEL definition; new axiomatizations and semantics are independent
full rationale
The manuscript defines four FEL variants and their three-valued extensions, supplies evaluation-tree semantics, and lists axiom sets claimed to be complete. The sole self-citation is the 2012 definition of Free FEL by co-author Staudt; this is used only as the starting point for the family and is not invoked to justify completeness or the claimed equivalence of static FEL to Bochvar's logic. No equation, prediction, or central claim reduces by construction to a quantity fitted or defined inside the present work, nor does any uniqueness theorem or ansatz rest on an unverified self-citation chain. The derivation therefore remains self-contained against external benchmarks.
Axiom & Free-Parameter Ledger
axioms (1)
- domain assumption Evaluation trees provide the semantics for left-sequential propositional expressions
invented entities (2)
-
Memorising FEL
no independent evidence
-
Free FEL
no independent evidence
Reference graph
Works this paper leans on
-
[1]
Bergstra, J.A., Bethke, I., and Rodenburg, P.H. (1995). A propositional logic with 4 values: true, false, divergent and meaningless, Journal of Applied Non-Classical Logics , 5(2):199-217. https://doi.org/10.1080/11663081.1995.10510855
-
[2]
Evaluation trees for proposition algebra
Bergstra, J.A. and Ponse, A. (2015). Evaluation trees fo r proposition algebra. https://doi.org/10.48550/arXiv.1504.08321 [cs.LO]
work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1504.08321 2015
-
[3]
Bergstra, J.A. and Ponse, A. (2023). Conditional logic a s a short-circuit logic. https://doi.org/10.48550/arXiv.2304.14821 [cs.LO]
-
[4]
Bergstra, J.A., Ponse, A., and Staudt, D.J.C. (2010 / 201 3). Short-circuit logic. (First version appeared in 2010.) https://doi.org/10.48550/arXiv.1010.3674 [cs.LO,math.LO]
work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1010.3674 2010
-
[5]
Bergstra, J.A., Ponse, A., and Staudt, D.J.C. (2018). Pr opositional logic with short-circuit evaluation: a non-commutative and a commutative variant. https://doi.org/10.48550/arXiv.1810.02142 [cs.LO,math.LO]
work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1810.02142 2018
-
[6]
Bergstra, J.A., Ponse, A., and Staudt, D.J.C. (2021). No n-commutative propositional logic with short- circuit evaluation. Journal of Applied Non-Classical Logics , 31(3-4):234-278 (online available). https://doi.org/10.1080/11663081.2021.2010954
-
[7]
Blok, A. (2011). Side-Effecting Logic, [Unpublished manuscript], University of Amsterdam
work page 2011
-
[8]
Bochvar, D.A. (1938). On a 3-valued logical calculus and its application to the analysis of contradictions (in Russian). Matematicheskii Sbornik, 4(2):287-308
work page 1938
-
[9]
Burris, S.N. and Sankappanavar, H.P. (2012). A Course in Universal Algebra - The Millennium Edition . Available at http://www.math.uwaterloo.ca/~snburris/htdocs/ualg.html. 6Suppose A = {a}. Then there are 5 ·44 ·33 terms composed of four elements of {a,¬ a, a ∨q T, a ∧q F} and the three connectives in { ∧q , ∨q , ↔← } that have evaluation trees of depth 4...
work page 2012
-
[10]
Church, A. (1956). Introduction to Mathematical Logic . Princeton University Press
work page 1956
-
[11]
Cornets de Groot, S.H. (2012). Logical systems with lef t-sequential versions of NAND and XOR. MSc. thesis Logic, University of Amsterdam (June 2020). https://eprints.illc.uva.nl/id/eprint/1743/1/MoL-2020-02.text.pdf
work page 2012
-
[12]
Guzm´ an, F. and Squier, C.C. (1990). The algebra of cond itional logic. Algebra Universalis , 27:88-110. https://doi.org/10.1007/BF01190256
-
[13]
Hoare, C.A.R. (1985). A couple of novelties in the propo sitional calculus. Zeitschrift f¨ ur Mathematische Logik und Grundlagen der Mathematik , 31(2):173-178. https://doi.org/10.1002/malq.19850310905
-
[14]
Kleene, S.C. (1938). On a notation for ordinal numbers. Journal of Symbolic Logic , 3:150-155
work page 1938
-
[15]
McCarthy, J. (1963). A basis for a mathematical theory o f computation. In P. Braffort and D. Hirschberg (eds.), Computer Programming and Formal Systems , pages 33–70, North-Holland, Amsterdam
work page 1963
-
[16]
McCune W. (2008). The GUI: Prover9 and Mace4 with a graph ical user interface. Prover9-Mace4 Version 0.5B. (Prover9-Mace4-v05B.zip, March 14, 2008). https://www.cs.unm.edu/~mccune/prover9/gui/v05.html
work page 2008
-
[17]
Papuc, D. and Ponse, A. (2022). Non-commutative propos itional logic with short-circuited biconditional and NAND. https://doi.org/10.48550/arXiv.2203.09321 [cs.LO]
-
[18]
Ponse, A. and Staudt, D.J.C. (2018). An independent axi omatisation for free short-circuit logic. Journal of Applied Non-Classical Logics , 28(1), 35-71 (online available). https://doi.org/10.1080/11663081.2018.1448637
-
[19]
Staudt, D.J.C. (2012). Completeness for two left-sequ ential logics. MSc. thesis Logic, University of Amsterdam (May 2012). https://doi.org/10.48550/arXiv.1206.1936 [cs.LO]. A Proofs - Section 2 Correctness of f . To prove that f :SPA→ FNF is indeed a normalization function we need to prove that for all FEL-terms P , f (P ) terminates, f (P )∈ FNF and EqF...
work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1206.1936 2012
-
[20]
x∧q (y∧q (z∧q F)) = ( x∨q y)∧q (z∧q F)
-
[21]
¬ x∧q (y∨q T) =¬ (x∧q (y∨q T)) Proof. x∧q (y∧q (z∧q F)) = x∧q ((¬ y∧q z)∧q F) by (FFEL4) and Lemma 2.6.1 = (¬ x∧q ¬ y)∧q (z∧q F) by (FFEL4) and Lemma 2.6.1 =¬ (¬ x∧q ¬ y)∧q (z∧q F) by Lemma 2.6.1 = (x∨q y)∧q (z∧q F), by (FFEL2) 32 ¬ x∧q (y∨q T) =¬ x∨q (y∧q F) by (FFEL10) =¬ (x∧q ¬ (y∧q F)) by (FFEL2) and (FFEL3) =¬ (x∧q ¬ (¬ y∧q ¬ T)) by (FFEL8) and (FFEL...
-
[22]
m(fa(X[T↦→ F, F↦→ T])) = m(fa(X))[T↦→ F, F↦→ T],
-
[23]
and by induction La(X1[T ↦→ Y ]) = La(X1)[T ↦→ La(Y )] = La(X)[T ↦→ La(Y )]
f a(X[T↦→ Y, F↦→ Z]) = fa(X)[T↦→ fa(Y ), F↦→ fa(Z)]. Proof. Clause 1 is Lemma 3.3.4 in [6]. We prove clause 2 by induction o n the structure of X. The base cases are trivial. If X = X1 ⊴ b ⊵ X2, distinguish the cases b = a and b̸= a: Case b = a, subcase f = L: La(X[T↦→ Y, F↦→ Z]) = La(X1[T↦→ Y, F↦→ Z]) and by induction, La(X1[T↦→ Y, F↦→ Z]) = La(X1)[T↦→ L...
-
[24]
˜∧q La(Y )) ⊴ a ⊵ m(Ra(X ′
-
[25]
˜∧q Ra(Y )) by IH and (Aux3) = ... = m(X ′ ˜∧q Y ). (B) m(X ˜∧q Y ) = m(X ˜∧q Y ′). This follows by induction on the depth of X. The base cases X∈{ T, F} are trivial. If X = X1 ⊴ a ⊵ X2 derive m(X ˜∧q Y ) = m((X1 ⊴ a ⊵ X2) ˜∧q Y ) = m(X1[fulland Y ] ⊴ a ⊵ X2[fulland Y ]) = m(La(X1[fulland Y ])) ⊴ a ⊵ m(Ra(X2[fulland Y ])) = m(La(X1)[fulland La(Y )]) ⊴ a ⊵...
-
[26]
Let P, Q ∈SP A. We have to prove that mfe(˜Fβ ∨q (P ∧q Q)) = mfe(˜Fβ ∨q (Q∧q P )), where β∈ As o satisfies{β} = α (P ∧q Q). By soundness of EqMFEL (Lemma 4.6), it suffices to prove that for any β∈ As o that satisfies α (P )⊂{ β}, EqMFEL⊢ ˜Fβ ∨q (P ∧q Q) = ˜Fβ ∨q (Q∧q P ). Derive EqMFEL⊢ ˜Fβ ∨q (P ∧q Q) = ( ˜Fβ ∨q (P ∧q F))∨q (P ∧q Q) by Lemma C.2 = ˜Fβ ∨q ((P...
-
[27]
The soundness of EqC ℓFELU follows as in 1, the additional case is that for any P ∈SP A U\SP A and Q∈SP A U, clfe(P ∧q Q) = clfe(Q∧q P ) = U. 40
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.