{"id":"03d14b40-f8df-4fbe-b3e9-a6188886e11a","arxiv_id":"2507.05519","paper_version":10,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Norms (obligation, permission, prohibition) are encoded as preemptible global constraints in ASP, and the paper claims this resolves classic deontic paradoxes including Chisholm's contrary-to-duty paradox.","lead":"This paper shows how obligations, prohibitions, and permissions, the operators of deontic logic, can be written directly as constraints in Answer Set Programming (ASP), a logic-based AI programming method. The authors argue this encoding resolves classic paradoxes of deontic logic, such as Chisholm's contrary-to-duty puzzle, and they provide runnable ASP examples including a car-borrowing scenario.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Paradox resolution rests on hand-picked preemption conditions; Section 4.3 gives no derivation rule for c, so the central claim is not yet established.","rationale":"The core encoding of deontic operators as ASP denials is concrete and, for the fixed examples, checkable: the Chisholm answer sets {go, tell} and {-go, -tell} do follow from the stated rules. The reader's weakest assumption correctly identifies the absence of a systematic derivation of the preemption condition c as the most serious gap. I agree with that identification, but I would not treat the OLON semantics itself as a separate weakness: under standard stable-model semantics, a rule of the form `q :- B, not q` is indeed vacuous when q is in the model and otherwise forces B to be false, so the paper's informal description matches the formalism. The conditional verdict is appropriate because the central claim of paradox resolution depends on an unformalized choice of c; a CONDITIONAL acceptance requiring a derivation procedure (or a precise admissibility condition for c) would settle the concern without discarding the otherwise useful encoding.","tokens_in":18088,"tokens_out":8422,"duration_ms":106255,"concrete_test":"Formalize the derivation of c in §4.3 as a mechanical function of the deontic theory (e.g., c is the negation of the obligated literal when that negation is entailed by the narrative, or c is the antecedent of a contrary-to-duty norm). Implement that derivation and rerun the programs from §5.1 and Appendix A; the check passes only if, for every paradox, the c used in the paper is exactly the function's output and the resulting answer sets are unchanged. If the derivation is underdetermined or yields different c's for any example, the paradox resolutions are stipulative rather than implied by the representation.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section 4.3 introduces the OLON rule `c :- not p, not c` as the way to preempt an obligation OBp when a violation condition c occurs, and Section 5 claims this resolves the classical paradoxes. The load-bearing gap is that the paper never defines how c is obtained from the deontic theory or narrative. In each example c is chosen to make the intended answer set exist: c=-go in Chisholm (line 1 of §5.1), c=kill in Forrester (§A.1), c=stay in Sartre (§A.2), c=broke in Kant (§A.2). Nothing in the syntax of §4.2 distinguishes these c's from other literals, and no theorem states that the answer sets of the encoded program are independent of, or even a function of, the set of preemption conditions. The abstract's claim that the paradoxes are 'simply and elegantly resolved' is therefore not a property of the ASP encoding of deontic operators; it is a property of the hand-written exception clauses. A variant of Chisholm in which c is taken to be `tell` (Jones tells) instead of `-go`, keeping lines 2-4 and fact -go, has no answer set because line 3 forces -tell while the OLON rule forces tell, illustrating that different choices of c change the outcome. The paper also does not specify how to handle multiple candidate violation conditions or conflicting preemptions. Until a general derivation (or at least a well-defined family of admissible c's) is provided, the central contribution is incomplete.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a translation of a fragment of deontic modal logic into answer set programs (ASP). Obligations, impermissibilities, and permissions are represented as global constraints (denials): OBp as `:- not p`, IMp as `:- not -p`, PEp as `:- -p`, with conditional norms handled analogously. The authors further propose preempting an obligation by an OLON rule `c :- not p, not c` when a condition c violates the obligation, and they claim that this mechanism resolves Chisholm's contrary-to-duty paradox and several other classical deontic paradoxes. A theorem asserts that SDL axioms D, K, and NEC are satisfied. The paper also presents a car-borrowing example with executable s(CASP) code and justification trees.","tokens_in":18353,"tokens_out":13080,"duration_ms":169415,"significance":"If the central claim were fully established, the paper would offer a very simple and executable implementation of deontic reasoning in ASP, with a clean mapping between the two negations of modal logic and ASP's default and strong negation, plus a uniform treatment of contrary-to-duty obligations. The concrete encodings are a genuine strength: the Chisholm program has exactly the two admissible worlds reported, the s(CASP) examples in Section 6 are runnable, and the justification trees in Appendix C support reproducibility. However, the paradox-resolution claim is not established because the preemption condition c in Section 4.3 is chosen by hand in each example, and some appendix resolutions depart from the stated translation fragment. As written, the paper is better described as a framework for normative modeling with explicitly user-specified exceptions than as a resolution of the classical deontic paradoxes.","major_comments":[{"comment":"The preemption condition c is a free parameter with no derivation rule. The paper states that for an obligation OBp and a condition c that violates it, one should encode the OLON rule `c :- not p, not c`, but it never defines how c is obtained from the deontic theory or narrative. In each paradox, c is chosen to produce the intended answer set: c=-go in Chisholm (§5.1 line 1), c=kill in Forrester (§A.1 line 1), c=stay/join in Sartre (§A.2 lines 1-2), and c=broke in Kant (§A.2 line 1). This is load-bearing: replacing the Chisholm OLON rule with `tell :- not go, not tell` (i.e., c=tell) while keeping lines 2-4 and the fact `-go` yields no answer set, so the set of admissible worlds is not determined by the deontic formulas alone. A formal characterization of admissible preemption conditions, or a derivation procedure for c, is required before the claimed 'simple and elegant' resolution of the paradoxes can be assessed.","section":"Section 4.3 and Section 5"},{"comment":"The Ross paradox example is not translated within the stated fragment. The premise 'mail -> post∨burn' is encoded as the two ASP rules `post :- mail` and `burn :- mail`, which assert both post and burn, not a disjunction. Since the syntax in Section 4.2 does not admit disjunction, this example does not demonstrate a resolution of Ross's paradox under the proposed encoding; it demonstrates a stronger, hand-chosen normative theory. The paper should either extend the fragment and give a translation of disjunction, or explain why the stronger encoding is the intended representation.","section":"Appendix A.3"},{"comment":"The Good Samaritan example does not actually resolve the paradox: the encoding `:- not rob` and `:- not help` forces rob and help in every answer set, which yields the very conclusion OBrob that the paradox calls into question. The text says 'in any world in which Smith is robbed, Jones helps him' and concludes the paradox is no longer a paradox, but this is an acceptance of the paradoxical conclusion rather than a resolution. The authors need to explain why conclusion OBrob is acceptable under the proposed semantics, or provide a different encoding that avoids it.","section":"Appendix A.4"},{"comment":"The theorem that SDL axioms D, K, and NEC are satisfied is stated more strongly than what is proven. The proof of D shows only that the two denials corresponding to OBp and OB¬p are jointly unsatisfiable via the built-in `:- p,-p`; this is a consistency property, not the derivation of the formula ¬OB¬p. The proof of NEC is not conclusive in the supported fragment: 'τ is a tautology' is not representable as a single atom under the Horn syntax of Section 4.2, and `not τ` is negation-as-failure over an atom, not classical negation of a formula. The theorem should be restated as a property of answer sets, with the scope of the fragment made explicit.","section":"Section 4.2, Theorem"}],"minor_comments":[{"comment":"The last paragraph says the obligation can be 'even further simplified' as `:- not p, not c`. This is indeed equivalent to the OLON rule, but stating the equivalence explicitly would avoid confusion, since the OLON form also names the exception as a head predicate.","section":"Section 4.3"},{"comment":"The explanatory text for Narrative 7 is repeated verbatim for Narrative 8, including the sentence about using classical negation `-financially_broke(jones)`, which is not present in Narrative 8. This appears to be a copy-paste error.","section":"Appendix B, Narratives 7 and 8"},{"comment":"Line 24 of Figure 5 reads `same_battery_level(C, Tb, Tr). :-` with an extra period before the neck symbol; this is likely a typo.","section":"Figure 5"},{"comment":"The sentence 'we will find that their are only two admissible worlds' contains a typo: 'their' should be 'there'.","section":"Section 5.1"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle to acceptance is the unconstrained preemption condition c in Section 4.3. If the authors can supply a general derivation rule for c, or explicitly narrow the contribution to a framework with user-supplied exception conditions, the paper could become publishable. The Ross and Good Samaritan examples in Appendix A need to be reworked, as they currently do not follow the stated translation scheme. The paper's strengths are its concrete, executable encodings and the detailed car-borrowing case study, which should be preserved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core idea is worth your time: map OBp, IMp, PEp to ASP denials (:- not p, :- not -p, :- -p), and let conditional norms become denials with the antecedent in the body. That is a genuinely direct encoding, and the OLON trick for preempting an obligation (c :- not p, not c) is a neat way to make obligations violable. The Chisholm program works as advertised: exactly two admissible worlds, and the intended one survives when -go is added. If you work on normative reasoning in ASP, the basic translation is a useful building block.\n\nThe soft spots are real, and the main one is the paradox resolution. The paper never tells you where the preemption condition c comes from. In every example, c is chosen so that the intended answer set exists. There is no rule for deriving c from the narrative, no theorem that the answer set is invariant under different choices, and the stress-test variant (take c = tell in Chisholm) breaks the intended reading. So the abstract's claim that the paradoxes are 'simply and elegantly resolved' is not yet a property of the encoding — it is a property of the hand-written exception clauses. That is the load-bearing gap.\n\nThe SDL theorem is also loosely argued. Axiom D is proven only in the trivial sense that p and -p cannot both hold, which is true of any consistent ASP program. NEC is asserted rather than derived. Those are not fatal, but they signal that the formal claim is weaker than advertised.\n\nTwo smaller things. The appendix has duplicated paragraphs (Narratives 7 and 8 repeat the same note) and refers to line numbers that do not match the displayed code, which makes independent audit harder than it should be. And Ross's and Good Samaritan are reclassified as 'not really paradoxes' rather than resolved — that is a defensible position, but it should be argued, not asserted.\n\nWho is this for? People working on executable deontic logic and on knowledge representation of norms. They will get a concrete, runnable starting point, and a clear open problem (how to select preemption conditions). I would not take the paradox-resolution claim at face value, but the building blocks are solid enough to cite.\n\nThis deserves peer review: the idea is new, the examples are checkable by hand, and the flaws are addressable in revision. A serious referee should push hard for a general account of c, a sharper statement of what the encoding actually settles, and a cleaned-up appendix. Conditional accept, major revision.","headline":"Nice concrete idea for deontic operators as ASP denials, but the paradox resolution rests on hand-picked preemption conditions and the paper needs a real derivation rule before that claim holds.","tokens_in":19041,"tokens_out":1896,"would_cite":true,"duration_ms":24749,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that deontic modal logic can be directly implemented in answer set programming by encoding obligations, prohibitions, and permissions as constraints, and that this dissolves its classic paradoxes.","keywords":["deontic modal logic","answer set programming","global constraints","odd loops over negation","contrary-to-duty paradox","Chisholm's paradox","default negation","strong negation"],"falsifier":"Run the paper's Chisholm encoding in a conventional answer-set solver under the standard stable-model semantics and compare the answer sets with the two worlds claimed in the paper; if the OLON rule is treated as inconsistent or yields additional worlds, the resolution depends on the goal-directed execution behavior rather than on the denial encoding alone.","tokens_in":17794,"feed_emoji":"⚖️","tokens_out":6934,"duration_ms":68506,"temperature":0.7,"pith_summary":"This paper argues that the modal operators of deontic logic, obligation, prohibition, and permission, can be translated directly into answer set programming, the logic-based framework for declarative problem solving. In its encoding, an obligation that p is the denial ':- not p', a prohibition that p is ':- not -p', and a permission that p is ':- -p'; a conditional norm Q -> OBp becomes ':- not p, Q'. The authors claim that this constraint-based reading resolves long-standing deontic paradoxes such as Chisholm's contrary-to-duty paradox, Forrester's paradox, Sartre's dilemma, Ross's paradox, and the Good Samaritan paradox, because obligations are preempted when the condition that violates them is provable. If correct, the result gives a practical, executable route from normative language to reasoning programs without extra logical machinery.","feed_headline":"Duties, bans, permissions become simple logic constraints","feed_subtitle":"Encoding duties as denial constraints dissolves Chisholm's paradox and its cousins.","key_machinery":"The load-bearing mechanism is the ASP global constraint, also called a denial: a rule of the form ':- not p, Q' that rejects every answer set in which Q is true and p is absent. Linked to it is the odd loop over negation (OLON) rule c :- not p, not c, which lets a violating condition c preempt the obligation; under the stable model semantics the rule is taken out of consideration when c is derivable by other rules, and otherwise acts as a denial forcing p. The mapping rests on treating default negation (not) as modal negation and strong negation (-) as propositional negation, so that the deontic square of opposition is reproduced by combinations of the two negations.","core_discovery":"The central discovery is that the deontic operators OB, IM, and PE are not primitive modal machinery but shorthand for global constraints on the set of acceptable worlds. In answer set programming, 'it is obligatory that p' becomes the constraint that any answer set not containing p is rejected, written as the denial ':- not p'; 'impermissible that p' becomes ':- not -p'; and 'permitted that p' becomes ':- -p'. Conditional norms of the forms Q -> OBp and OB(Q -> p) both translate to the denial ':- not p, Q'. The paper further claims that when a violating condition c arises, the denial is re-encoded as an odd-loop-over-negation rule c :- not p, not c so that the obligation is dropped exactly when c is derivable, which is how the paradoxes are said to resolve. It also proves that the standard deontic logic axioms D, K, and NEC hold in this representation.","pith_inferences":["An implication the authors leave implicit is that the same constraint-as-denial translation should work for epistemic and temporal modalities, because the distinction between default and strong negation is not specific to deontic operators.","The paper does not give a general rule for choosing which condition c preempts an obligation; a testable extension would derive c automatically from a proof of violation, which would turn the hand-built paradox resolutions into a system.","Because the claimed OLON behavior is tied to the goal-directed execution system used in the paper, a natural probe is to run the same encodings on a conventional answer-set solver; if the answer sets differ, the paradox resolution depends on that system's semantics rather than on ASP itself."],"forward_implications":["Any Horn deontic formula in the supported fragment can be executed directly by an answer set programming system, making normative reasoning computable.","The paradoxes of contrary-to-duty reasoning disappear without introducing dyadic operators, sanctions, or preference orderings over models.","Conditional obligations and prohibitions become conditional constraints, so secondary obligations ('least one can do') can be represented by chained OLON rules.","Because constraints can be named via OLON rules, separate logic can decide when a norm applies, supporting elaboration-tolerant knowledge representation.","The approach distinguishes factual detachment from deontic detachment by whether the norm's antecedent is evaluated in the current or accessible world."],"supporting_citations":[{"why":"supplies the stable model semantics, denials, and OLON rules on which the whole encoding is built.","marker":"[4]"},{"why":"provides the goal-directed answer set programming system used to execute the encodings and realize OLON preemption.","marker":"[3]"},{"why":"defines the SDL axioms, the deontic square of opposition, and the paradox list the paper targets.","marker":"[25]"},{"why":"is the source for the contrary-to-duty (Chisholm) paradox, the paper's main test case.","marker":"[29]"},{"why":"frames obligations as constraints on goal satisfaction and discusses Jørgensen's dilemma, which motivates the denial reading.","marker":"[21]"},{"why":"gives the ASP overview that fixes the meaning of default negation, strong negation, and global constraints.","marker":"[5]"}],"fun_headline_variants":["Duties as denial constraints: ASP dissolves deontic paradoxes","Obligations, prohibitions, permissions as ASP global constraints","Why deontic paradoxes vanish in ASP constraints","Conditional obligations as denial rules in ASP"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The method presumes that for each obligation one can identify, in advance, the violating condition c that should preempt it, and in the paper these conditions are chosen by hand for each paradox rather than derived by a general procedure.","fun_headline_variants_meta":{"raw":{"variants":["Duties as denial constraints: ASP dissolves deontic paradoxes","Obligations, prohibitions, permissions as ASP global constraints","Why deontic paradoxes vanish in ASP constraints","Conditional obligations as denial rules in ASP"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000549,"raw_usage":{"total_tokens":2558,"prompt_tokens":816,"completion_tokens":1742,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":432,"completion_tokens_details":{"reasoning_tokens":1676}},"tokens_in":432,"tokens_out":1742,"duration_ms":15632,"temperature":1.0,"reasoning_tokens":1676,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:28:26.679332+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the paper's Chisholm encoding in a conventional answer-set solver under the standard stable-model semantics and compare the answer sets with the two worlds claimed in the paper; if the OLON rule is treated as inconsistent or yields additional worlds, the resolution depends on the goal-directed execution behavior rather than on the denial encoding alone.","supporting_citations":[{"cited_title":"Anonymity is not privacy: Technical perspective,","cited_arxiv_id":null,"evidence_quote":"supplies the stable model semantics, denials, and OLON rules on which the whole encoding is built."},{"cited_title":"In: Zalta, E.N., Nodelman, U","cited_arxiv_id":null,"evidence_quote":"defines the SDL axioms, the deontic square of opposition, and the paradox list the paper targets."},{"cited_title":"In:Internat Encyclopedia of Philosophy","cited_arxiv_id":null,"evidence_quote":"is the source for the contrary-to-duty (Chisholm) paradox, the paper's main test case."},{"cited_title":"Kowalski & Ken Satoh (2018):Obligation as Optimal Goal Satisfaction.J","cited_arxiv_id":null,"evidence_quote":"frames obligations as constraints on goal satisfaction and discusses Jørgensen's dilemma, which motivates the denial reading."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the ASP overview that fixes the meaning of default negation, strong negation, and global constraints."}],"review_version":2}