{"id":"02030a89-2ce6-4da2-bc03-842665a2a73a","arxiv_id":"2502.09232","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":4.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper proposes, without presenting, a formalization of smart contracts in the Situation Calculus extended with obligations.","lead":"This paper is a thesis proposal rather than a result-bearing preprint: it sketches a plan to formalize smart contracts using the Situation Calculus with obligations, executed in GOLOG. A generalist might read it to see a plausible research direction for contract verification, but there are no formal definitions or proofs to evaluate.","discovery_kind":"unclear","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed extension of Scherl–Levesque to obligations is asserted, not established; the load-bearing step is an analogy that may break regression.","rationale":"The reader's verdict is UNVERDICTED, and my stress test agrees with that judgment. The paper is a clearly labeled thesis-in-progress extended abstract, and much of it is background and roadmap material. The actual scientific claim, however, is stated as completed work: Section 7 says steps (2)-(4) and parts of step (5) have been completed, including the formalization of obligations and a proof that Reiter's regression operator remains a reasoning mechanism. The strongest claim is that this yields logical foundations for smart contracts and an implementable GOLOG framework. For that claim to hold, the proposed obligation extension must remain inside the class of theories for which regression and GOLOG execution are known to work. That is precisely where the paper is silent. The reader's weakest-assumption analysis identifies the same point: the analogy with Scherl and Levesque is plausible but unproven, and if the analogy fails, the central construction collapses. My concern sharpens this in one direction: knowledge and obligation are not interchangeable modalities, and the regression theorem is sensitive to the syntactic form of successor state axioms. The manuscript does not show that the obligation fluent can be given the required form, nor that the intuitive properties follow from it. This is not a claim of internal inconsistency; it is a claim of insufficient evidence. A constructive response would be to write out the axioms and prove the two properties named in the concrete test. Until then, the appropriate verdict remains UNVERDICTED, so no change to the reader's verdict is needed.","tokens_in":6537,"tokens_out":6132,"duration_ms":64902,"concrete_test":"Reconstruct the proposed theory in full: define the obligation fluent Oblig(phi, ag, s), state its successor state axiom, and define the regression operator for sentences containing Oblig. Then (1) prove from Pirri and Reiter's metatheory that regression is sound and complete for all regressable sentences of the resulting basic contractual theory, and (2) prove from the successor state axiom alone that an obligation-producing action changes no fluent other than the obligation fluents it explicitly enacts, and that Oblig(phi,s) persists across every action a with not Viol(a,phi,s). If the proof of (1) requires an assumption not present in the standard basic action theory framework, or if (2) fails on a simple two-party payment contract, the extended foundation collapses.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Secs. 4 and 7, Step 3) is that a new obligation fluent can be added to a basic action theory by analogy with Scherl and Levesque's solution to the frame problem for knowledge, yielding intuitive obligation properties and leaving Reiter's regression operator sound and complete. The load-bearing assumption is that this analogy preserves the syntactic form that makes regression work. For knowledge, regression works because the possible-worlds operator is eliminable via a successor state axiom with a specific relational structure, relying on the accessibility relation and on relativizing ordinary fluents across accessible situations. Obligations are a distinct normative modality: they are agent-relative, violation-tolerant, and interact with deadlines and contrary-to-duty conditionals. If the successor state axiom for the obligation fluent must quantify over future actions or over fluents in its own content, the resulting theory is no longer a basic action theory and the regression theorem does not apply. If, instead, the axiom is purely definitional, then the claimed intuitive properties (persistence unless violation, no side effects on other fluents) must be proved as theorems from that axiom. The manuscript supplies neither the axioms nor the proofs. Consequently, the foundational mechanism on which the GOLOG implementation and verification framework depend is currently unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes to give logical foundations to smart contracts using the Situation Calculus and the GOLOG programming language. The intended approach is to extend Scherl and Levesque's solution to the frame problem for knowledge-producing actions to a new notion of obligation-producing actions, thereby obtaining 'basic contractual theories' that formalize legal contracts, an implementable GOLOG specification, and an obligation-based programming framework for verifying contract properties. The paper is a brief extended abstract: it surveys related work, states the problem, outlines the methodology, lists expected contributions, and reports on the current status of the research. It contains no formal definitions, axioms, theorems, proofs, or worked examples; the technical content that would support the central claims is entirely absent.","tokens_in":6775,"tokens_out":3898,"duration_ms":34809,"significance":"If the proposed extension succeeds, it would offer a uniform formal semantics for smart contracts, with the potential to support executable specifications and automated verification of contract properties in a well-studied action logic. The combination of deontic notions with Situation Calculus and GOLOG is a plausible and potentially valuable direction, building on a solid existing body of work. However, as it stands, the manuscript is a research proposal rather than a research contribution. No technical claim can be checked, and the significance of the approach cannot be assessed beyond the general interest of the idea. The paper explicitly claims that several steps of the research are complete (Section 7), so the absence of any formal development is a substantive gap, not merely a question of presentation.","major_comments":[{"comment":"The central technical claim is that obligations can be formalized by extending Scherl and Levesque's successor state axiom for knowledge to obligation-producing actions, but the manuscript never defines the obligation fluent, the class of obligation-producing actions, the successor state axiom for that fluent, or the underlying basic action theory. Without these definitions, the claimed 'intuitive properties' (obligation-producing actions affect only the new fluent and those fluents made obligatory, persistence unless a contrary action occurs) cannot be stated precisely, let alone proved. This is the load-bearing step of the entire proposal, and it is entirely missing.","section":"Sections 4 and 7 (Step 3)"},{"comment":"The claim that 'Reiter's regression operator for reasoning about actions back to the initial situation is a reasoning mechanism for this setting as well' is unsupported. Regression is sound and complete only for basic action theories whose successor state axioms have a specific syntactic form. Adding a modality for obligations may break this form, for instance if the successor state axiom for the obligation fluent quantifies over future actions or mentions fluents within its own scope. The author must state the axiom and prove that the resulting theory is a basic action theory (or otherwise justify that regression remains applicable).","section":"Section 7 (Step 3), final sentence"},{"comment":"The proposed translation from Symboleo specifications to Situation Calculus theories is mentioned only in one sentence. Since the paper promises 'implementable specifications' and automatic checking of contract properties, the translation relation, the syntactic categories of Symboleo being covered, and at least one worked example are necessary to evaluate whether the approach is feasible and whether the resulting theories have the claimed properties.","section":"Section 7 (Step 4)"},{"comment":"The 'Current Status of the Research' section states that steps (2)-(4) and parts of step (5) have been completed, but none of the corresponding definitions, theorems, or artifact descriptions appear in the paper. If this is an extended abstract of a thesis, the author should either include the key technical content or explicitly state that the full development is forthcoming; as written, the claim of completion is unverifiable and is not a substitute for presenting the results.","section":"Section 7 (Steps 2-5)"}],"minor_comments":[{"comment":"The sentence 'Smart contracts are going to be implement with Golog (written en Prolog)' contains grammatical errors and a typo; it should read, for example, 'Smart contracts are going to be implemented in GOLOG (written in Prolog).'","section":"Abstract"},{"comment":"The phrase 'le formal languages are equipped...' contains a French article; it should be 'the formal languages...'.","section":"Section 1"},{"comment":"The text says 'the situation calculation' in Section 7; the correct term is 'situation calculus'.","section":"Section 3"},{"comment":"The description of basic action theories as 'mostly first-order' should be made precise, since the foundational axioms typically include a second-order induction axiom; the author should clarify which version of the theory is being used.","section":"Section 5"},{"comment":"Several references are incomplete or contain visible errors (e.g., reference [3] has a garbled title); the reference list should be carefully checked against the original sources.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a work-in-progress extended abstract. The main obstacle is not a demonstrable error but a complete lack of technical substance. For a journal publication, the author would need to include the full formalization and proofs. I recommend major revision rather than outright rejection because the proposed direction is plausible and the missing material can, in principle, be supplied. The editor may also wish to consider whether a short position paper is the appropriate format for this work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nQuick take on arXiv:2502.09232. It's a short extended abstract from ICLP 2024 presenting a thesis in progress, not a research result. The title promises logical foundations, but the manuscript contains no definitions, axioms, theorems, or proofs. Section 7 claims steps (2)-(4) and part of (5) are complete, yet none of that material appears. So there is nothing to verify.\n\nThat said, the plan is coherent and honest. The author correctly identifies Daskalopulu and Sergot [4] as having already suggested Situation Calculus for contracts, and the proposed route—adapting Scherl and Levesque's knowledge regression to obligations—is a plausible thesis topic. Using Symboleo as a source ontology for translation is a reasonable choice. If carried out, the result would be a useful contribution to the formal-methods-for-contracts subfield, though not a revolution.\n\nThe soft spots are proportional: the paper overstates progress by claiming completed work without showing it. The stress-test concern about regression is legitimate. Scherl and Levesque's regression works because the knowledge fluent has a successor state axiom with a specific relational form. Obligations are a different modality—agent-relative, violation-tolerant, interacting with deadlines—and nothing in this abstract shows that a successor state axiom for an obligation fluent preserves the syntactic conditions for regression. If it doesn't, the whole verification story collapses. But absence of proof is not a counterexample; the approach might well work. The paper simply doesn't give the reader enough to judge.\n\nCitation pattern looks fine. The author credits prior work and doesn't claim priority for the basic idea. No red flags there.\n\nWho is this for? Someone supervising or reviewing a thesis proposal might find it a useful status report. A reader looking for results will be disappointed. My recommendation to an editor: this does not merit referee time as a research paper. Desk reject with a note pointing to the missing technical chapters, and invite resubmission once definitions and proofs are actually included. If the venue is a workshop for work-in-progress, it could be acceptable as a short abstract.","headline":"A thesis proposal that claims completed work but shows no technical content; the plan is plausible but unverifiable.","tokens_in":7278,"tokens_out":2289,"would_cite":false,"duration_ms":21371,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03B70","68Q60","68T27"],"pacs":[],"model":"deepseek-v4-flash","headline":"Smart contracts can be given a formal, executable logic by modeling obligations as a Situation Calculus fluent.","keywords":["smart contracts","situation calculus","obligation","deontic logic","frame problem","successor state axioms","formal verification","GOLOG"],"falsifier":"Build the smallest possible contract theory with two actions, one that makes a payment obligatory and one that fulfills it, together with the proposed successor state axioms for the obligation fluent. If a query about the situation after the fulfillment action still yields that the payment is obligatory, or if a regression of that query differs from evaluating it directly in the situation, the analogy with knowledge fails. More generally, any basic contractual theory in which the successor state axioms let an obligation-producing action alter a fluent other than the obligation fluent would falsify the claim that obligations are isolated from the rest of the action theory.","tokens_in":6304,"feed_emoji":"📜","tokens_out":8484,"duration_ms":74055,"temperature":0.7,"pith_summary":"This paper proposes to give smart contracts a formal, executable semantics by modeling them in the Situation Calculus, a logic designed for reasoning about how actions change a world. The central move is to treat contractual obligations the way that logic already treats knowledge: add a dedicated obligation predicate, and write general rules describing how actions create and destroy obligations. With those rules, the author constructs 'basic contractual theories' that axiomatize legal contracts, represents contracts as programs in the GOLOG language, and argues that contract properties can then be proved automatically by regression back to the initial situation. If the proposal is right, a legal contract can be turned into a machine-checkable theory whose clauses are executable and verifiable, addressing an open problem in the area. The author reports the formalization of obligations as completed within the thesis, though it is not reproduced in this short paper.","feed_headline":"Contract obligations become machine-checkable logic","feed_subtitle":"A thesis proposes modeling contracts as action theories with an obligation fluent, enabling automatic verification.","key_machinery":"The central mechanism is an obligation fluent: a predicate that records, in each situation, which conditions an agent is obliged to bring about. It is equipped with successor state axioms written by analogy with the standard knowledge solution to the frame problem, so that the only actions that change the fluent are obligation-producing actions, and the fluent otherwise persists. These axioms are what turn a basic action theory into a basic contractual theory. The other load-bearing piece is the regression operator of the Situation Calculus, which takes a formula about a later situation and rewrites it, using the successor state axioms, into an equivalent formula about the initial situation; the paper relies on this to make contract properties automatically checkable.","core_discovery":"The paper's central claim is that smart contracts can be given logical foundations by reducing the notion of contractual obligation to a single new fluent in the Situation Calculus. Building on the established solution to the frame problem for knowledge, the author proposes that obligation-producing actions are governed by successor state axioms for this fluent: such actions make certain conditions obligatory, obligations persist through unrelated actions, and an action contrary to the obligation is what removes it. These axioms define 'basic contractual theories,' which the paper presents as the formal semantics of legal contracts. In the paper's account, a contract is a process in the Situation Calculus, written in GOLOG and executed by an interpreter, and properties of the contract are proved by the regression operator, which reduces a query about an execution situation to a query about the initial situation. The author states that this extension yields the intuitive properties of obligations and that regression remains a sound and complete reasoning mechanism for the extended language.","pith_inferences":["Beyond the paper, the same extension-by-analogy could be tried for permissions, prohibitions, and powers, yielding a fuller deontic layer on top of the Situation Calculus rather than a single obligation predicate.","A natural testable extension is to implement the obligation successor state axioms on a small payment contract and check whether regression stays complete; a failure there would suggest restricting to a fragment of the language.","If the thesis matures into an implementation, it could supply a formal bridge from natural-language contracts through a legal-contract specification language to auditable blockchain code.","The proposal implicitly predicts that contract vulnerabilities will show up as logical consequences of the basic contractual theory, so a deployed contract that violates a proved property would be an implementation bug rather than a limitation of the logic."],"forward_implications":["A legal contract can be expressed as a basic contractual theory, giving it a single formal meaning instead of an informal code description.","Contract properties such as 'payment becomes obligatory after delivery' can be checked automatically by regression on the theory.","Because the same theory is executed by the GOLOG interpreter, the specification and the implementation of a contract coincide.","Exceptions and subcontracts fit naturally into the same framework, since they are just additional actions and fluents in the logical language.","Contract verification becomes a logic problem: proving that a deployed contract satisfies its legal obligations is a theorem about its basic contractual theory."],"supporting_citations":[{"why":"Supplies the knowledge-based solution to the frame problem that the paper extends to obligations.","marker":"[27]"},{"why":"Provides the basic action theory framework, successor state axioms, and regression on which the contractual theories are built.","marker":"[24]"},{"why":"Gives the original simple solution to the frame problem and the completeness result for goal regression that the paper assumes carries over.","marker":"[25]"},{"why":"Defines the GOLOG programming language and interpreter used to execute and reason about the contract specifications.","marker":"[13]"},{"why":"First suggested Situation Calculus as a possible basis for legal contracts, motivating the paper's program.","marker":"[4]"},{"why":"Introduces the Symboleo specification language that the paper plans to use to express contracts before translation.","marker":"[28]"},{"why":"Supplies the foundational axioms for situations that structure the space of situations in basic contractual theories.","marker":"[21]"}],"fun_headline_variants":["Obligation fluent grounds smart contracts in logic","Contract logic via action theories and regression","Making smart contracts machine-checkable with logic","Situation Calculus formalizes contractual obligations","Obligation-aware action theories for smart contracts"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole approach depends on an unproved analogy: that a new obligation predicate whose changes are specified in the same way as knowledge will behave like a real obligation, with actions creating it, unrelated actions leaving it alone, and only contradictory actions removing it, and that the standard regression method for reasoning about actions will still be correct for the extended language. The paper asserts these properties but does not supply the definitions or proofs.","fun_headline_variants_meta":{"raw":{"variants":["Obligation fluent grounds smart contracts in logic","Contract logic via action theories and regression","Making smart contracts machine-checkable with logic","Situation Calculus formalizes contractual obligations","Obligation-aware action theories for smart contracts"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00022,"raw_usage":{"total_tokens":1435,"prompt_tokens":925,"completion_tokens":510,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":541,"completion_tokens_details":{"reasoning_tokens":444}},"tokens_in":541,"tokens_out":510,"duration_ms":5276,"temperature":1.0,"reasoning_tokens":444,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T22:11:47.526009+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Build the smallest possible contract theory with two actions, one that makes a payment obligatory and one that fulfills it, together with the proposed successor state axioms for the obligation fluent. If a query about the situation after the fulfillment action still yields that the payment is obligatory, or if a regression of that query differs from evaluating it directly in the situation, the analogy with knowledge fails. More generally, any basic contractual theory in which the successor state axioms let an obligation-producing action alter a fluent other than the obligation fluent would falsify the claim that obligations are isolated from the rest of the action theory.","supporting_citations":[{"cited_title":"Scherl & H.J","cited_arxiv_id":null,"evidence_quote":"Supplies the knowledge-based solution to the frame problem that the paper extends to obligations."},{"cited_title":"Reiter (2001): Knowledge in Action: Logical Foundations for Specifying and Implementing Dynamical Systems","cited_arxiv_id":null,"evidence_quote":"Provides the basic action theory framework, successor state axioms, and regression on which the contractual theories are built."},{"cited_title":"Reiter (1991): The Frame Problem in the Situation Calculus: A Simple Solution (Sometimes) and a Completeness Result for Goal Regression","cited_arxiv_id":null,"evidence_quote":"Gives the original simple solution to the frame problem and the completeness result for goal regression that the paper assumes carries over."},{"cited_title":"Levesque, R","cited_arxiv_id":null,"evidence_quote":"Defines the GOLOG programming language and interpreter used to execute and reason about the contract specifications."},{"cited_title":"Daskalopulu & M.J","cited_arxiv_id":null,"evidence_quote":"First suggested Situation Calculus as a possible basis for legal contracts, motivating the paper's program."},{"cited_title":"Sharifi, A","cited_arxiv_id":null,"evidence_quote":"Introduces the Symboleo specification language that the paper plans to use to express contracts before translation."},{"cited_title":"Pirri & R","cited_arxiv_id":null,"evidence_quote":"Supplies the foundational axioms for situations that structure the space of situations in basic contractual theories."}],"review_version":1}