{"id":"ae3f507e-21b8-4b69-b5c5-3cd6c5cfe6b4","arxiv_id":"2411.16609","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper introduces Event-Model-F, a formal OWL/DL ontology of events built on DOLCE+DnS Ultralite, with six patterns covering participation, mereology, causality, correlation, documentation, and interpretation.","lead":"This paper presents Event-Model-F, a formal ontology for describing events and how they relate to people, objects, time, and space. It is designed on top of the DOLCE+DnS Ultralite foundational ontology to improve interoperability between event-based systems.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The causality and correlation patterns provide only unconstrained role labels; absent DL axioms tying F:Correlate to a common cause, the claimed comprehensive structural support is not semantically grounded.","rationale":"The reader's weakest assumption identifies precisely the most load-bearing vulnerability: the semantic adequacy of reifying causal and correlative statements as DnS situations. The paper's own functional requirement 4b says the model should let systems automatically check the validity of exchanged knowledge, and requirement 4c defines correlation as common cause. Yet the described patterns contain no axioms that would support such checks, and the comparison with existing models in Section 7 depends on the assumption that marking 'Yes' for Causality and Correlation in Event-Model-F is meaningful. Without DL axioms or reasoning results, the formal nature of the model cannot be independently assessed. I agree with the reader's conditional verdict: the model description is plausible and a useful ontology-engineering contribution, but the full-support claim requires the actual axioms and a demonstration that the causality and correlation patterns capture the intended semantics. No reason to strengthen or weaken the verdict; the paper should remain CONDITIONAL on these clarifications.","tokens_in":11392,"tokens_out":6061,"duration_ms":60623,"concrete_test":"Fetch the OWL ontologies from https://github.com/ascherp/ontologies and load them into a DL reasoner such as HermiT. Then check whether the ontology entails, for every F:EventCorrelationSituation, the existence of some F:EventCausalitySituation involving a common-cause event that is the cause of all events classified as F:Correlate in that situation. If the entailment does not hold, the correlation pattern does not implement the 'common cause' definition from Section 5.4, confirming that the structural support is not semantically grounded.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that Event-Model-F provides formal, comprehensive support for mereological, causal, and correlative relationships rests on the causality pattern (Section 5.3) and correlation pattern (Section 5.4). In the causality pattern, F:Cause and F:Effect are merely EventTypes that classify events included in the same F:EventCausalitySituation; F:Justification classifies some DUL:Description, but no axiom is given that requires that Description to contain a causal law, that the cause precede the effect, or that the situation actually constitutes a causal explanation. Consequently, any arbitrary pair of events can be labeled Cause/Effect under some Description, so the pattern does not distinguish causality from mere co-classification. The correlation pattern is even weaker: F:Correlate is defined as a Role, but there is no axiom linking F:EventCorrelationSituation to any F:EventCausalitySituation or to a common cause, despite Section 5.4 defining correlation as having a common cause. Thus an arbitrary set of events can be asserted as correlated without satisfying the paper's own definition. Because the paper provides no DL axioms and reports no reasoning results, the reader cannot tell whether the actual OWL restricts these patterns; on the face of the paper, the structural support for causality and correlation is purely syntactic. If the reification does not preserve the intended causal semantics, the central claim of comprehensive structural relationship support fails, even though the ontology constructs exist.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents Event-Model-F, a pattern-based event ontology aligned with DOLCE+DnS Ultralite (DUL). It introduces six patterns, namely participation, mereology, causality, correlation, documentation, and interpretation, and claims that they together provide comprehensive formal support for representing time and space, objects and persons, and mereological, causal, and correlative relationships between events. The model is modular and designed to be extended by domain ontologies. The paper evaluates the model against six functional and five non-functional requirements and compares it with eight existing event models, concluding that Event-Model-F is the only one offering full support for structural relationships and multiple event interpretations.","tokens_in":11653,"tokens_out":5033,"duration_ms":44583,"significance":"If the formal claims are substantiated, the model would be a valuable contribution to event-based systems interoperability: it is modular, grounded in a foundational ontology, and explicitly addresses the underrepresented aspects of causality, correlation, and multiple interpretations. The comparative analysis of eight event models is also useful. However, the current manuscript does not deliver the promised formal axiomatization; the patterns are described informally through diagrams and text, and the requirement coverage is self-assessed rather than formally demonstrated. The paper's significance therefore remains conditional until the DL axioms and formal evaluations are actually provided.","major_comments":[{"comment":"The causality and correlation patterns are semantically unconstrained. In Section 5.3, F:Cause and F:Effect are introduced as EventTypes that classify events included in an F:EventCausalitySituation, and F:Justification classifies some DUL:Description, but no axiom is presented that requires the Description to contain a causal theory, that the cause precedes the effect, or that the situation constitutes a causal explanation. Consequently, any arbitrary pair of events can be labeled Cause/Effect under some Description, and the pattern does not distinguish causality from co-classification. In Section 5.4, F:Correlate is defined as a Role but no axiom links an F:EventCorrelationSituation to an F:EventCausalitySituation or to a common cause, despite the section's own definition of correlation as sharing a common cause. The structural support for causality and correlation is therefore purely syntactic in the manuscript, which contradicts the claim of comprehensive formal support in Section 8.","section":"Sections 5.3 and 5.4"},{"comment":"The paper repeatedly asserts that the model is axiomatized in Description Logics, but no DL axioms, class definitions, role/property definitions, or reasoning results appear in the manuscript. The sentence 'The axiomatization of the Event-Model-F has been conducted in Description Logics and is available online at: http://isweb.uni-koblenz.de/eventmodel' is an external pointer, not a specification. Because non-functional requirement (b) explicitly demands sufficient formality and axiomatization for automatic semantic checks, the absence of the axioms in the paper makes the central formal claims unverifiable and prevents a reader from checking whether the online artifact matches the patterns described here.","section":"Section 5, opening; Section 4.2(b)"},{"comment":"The functional requirements are synthesized by the authors from a scenario and existing models, and the model is then evaluated against these same requirements in Sections 5 and 7. The claim that Event-Model-F 'supports all of them' (Section 7) is an assertion based on the design, not a formal demonstration: the paper provides no consistency check, no entailment tests, and no independent benchmark. The comparative table in Figure 3 is also not based on an explicit scoring rubric, so the degrees of support ('Yes', 'No', 'Lim.') are not independently reproducible. This self-assessment does not by itself invalidate the model, but it does mean the full-support conclusion is partly self-referential and should be supported by formal evaluation or at least by a clearly defined criterion.","section":"Sections 4.1 and 7"},{"comment":"The claim that events may be 'arbitrarily temporally related to each other' using 'the provided means of DOLCE such as the formalization of Allen's Time Calculus' is not substantiated. The F:EventCompositionConstraint is described only informally as an instance that 'formalizes' constraints, but no constraints are defined for temporal, spatial, or spatio-temporal cases, and no mapping from Allen's relations to the ontology is given. Without these details, the mereology pattern does not demonstrate the required formal support for relative temporal and spatial relations between events.","section":"Section 5.2 (Mereology Pattern)"}],"minor_comments":[{"comment":"The sentence 'Such enduring entities unfold over space, i.e., they are in time' appears self-contradictory; the authors likely mean that objects endure over time or that they exist in space, and the text should be reworded for clarity.","section":"Section 2"},{"comment":"The comparison table is difficult to parse: columns such as 'Abs.Rel.' are not explained, some cells contain multiple check marks, and the abbreviation 'Lim.' is not defined in the caption or the text.","section":"Figure 3"},{"comment":"The notation 'DUL:hasRegion of a SpaceRegion' should be made precise, for example as a property that relates a Quality to a SpaceRegion, to avoid confusion between the class and the property.","section":"Section 5.1"},{"comment":"The URLs for the online ontology are given as bare links (http://isweb.uni-koblenz.de/eventmodel and the GitHub repository in the header); the authors should provide stable identifiers, such as DOIs, or include the axioms as supplementary material.","section":"Sections 5 and 9"}],"recommendation":"major_revision","confidential_remarks":"The manuscript appears to be a lightly updated version of a 2009 K-CAP paper, and the editor may wish to consider whether the contribution is sufficiently novel for a journal submission, and whether the absence of the actual DL axioms and reasoning results in the text is acceptable for the journal's standards. The external references to online artifacts may also need to be archived or made persistent."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look if you work on event ontologies or interoperability. The paper is a 2009 K-CAP reprint, and it reads like a good conference paper: clear motivation, a concrete emergency-response scenario, and a systematic comparison with eight existing event models. What is genuinely new is the six-pattern decomposition, especially the correlation and interpretation patterns. No cited model offers either, and the comparison table makes the gap vivid. The alignment with DUL is careful, and the pattern-oriented design is a real contribution to ontology engineering. The soft spots are real, but I don't think they sink the paper. First, the actual OWL/DL axioms are not in the paper. The authors say the ontology is axiomatized and point to a repository, but in the submitted text you cannot check whether F:Cause and F:Effect have any constraints beyond being EventTypes. The stress-test note is fair: on the face of the paper, the causality pattern is just a reification with role labels, and the correlation pattern never links F:Correlate back to a common cause, even though Section 5.4 defines correlation that way. That means the claimed 'comprehensive support' for structural relationships is not demonstrated inside the paper. This is a presentation gap as much as a semantic one—the repository may contain the missing axioms—but the paper should have included or at least summarized them, and it should have reported any reasoning or consistency checks. Second, the functional requirements are synthesized by the authors and then the model is scored against those same requirements. The external comparison mitigates the circularity, but it does not eliminate it. The citation pattern is fine. They cite the obvious event models and foundational ontology literature, and the self-citations are to related core-ontology work that is genuinely relevant. Who is this for? Ontology engineers and anyone building semantic event-based systems who wants a modular, well-grounded vocabulary. The paper is not a formal semantics paper, so a logician will be unsatisfied. But as a design proposal and comparative analysis, it holds up. I would send it to peer review. A serious referee should ask for the DL axioms or formal characterization, a tightening of the correlation pattern to make the common-cause link explicit, and an external evaluation or at least a set of competency questions. Those are major revisions, not grounds for rejection. My take: the work is serious and the core idea is sound. It just overstates the formalism visible in the paper.","headline":"Solid ontology-engineering paper whose real novelty is the six-pattern architecture, but the formal axioms are missing from the paper and the causality/correlation patterns are too thin as presented.","tokens_in":673,"tokens_out":1059,"would_cite":false,"duration_ms":24767,"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":"Event-Model-F represents events with six modular ontology patterns, covering time, space, participants, structural relations, documentation, and interpretations.","keywords":["events","core ontology","DOLCE+DnS Ultralite","Descriptions and Situations","event participation","event composition","event causality","event interpretation"],"falsifier":"Open the published OWL axiomatization and ask whether a consistent model can assert the same event as both F:Cause and F:Effect of itself under a single F:Justification; because no constraints on these roles are given, the pattern would accept such a self-cause, showing that it does not yet carry the causal semantics it claims to represent.","tokens_in":11162,"feed_emoji":"🧩","tokens_out":6534,"duration_ms":58142,"temperature":0.7,"pith_summary":"Distributed systems that process events need a shared, machine-readable way to describe what an event is, when and where it happened, who and what took part, and how events relate. This paper argues that existing event models cover participation, time, space, and documentation but fall short on structure and interpretation, and it presents Event-Model-F as a formal remedy. Event-Model-F is built on the DOLCE+DnS Ultralite foundational ontology and packs six functional requirements into six modular ontology patterns: participation, mereological composition, causality, correlation, documentation, and interpretation. If the model works as claimed, components in an emergency-response or similar distributed system can exchange event descriptions with unambiguous semantics, and the same occurrence can carry different, even conflicting, interpretations.","feed_headline":"Six patterns give events a shared formal language","feed_subtitle":"It covers time, place, people, parts, causes, correlations, and rival interpretations for distributed event systems.","key_machinery":"The load-bearing machinery is DUL's Descriptions and Situations (DnS) pattern. In each of the six Event-Model-F patterns, a Situation includes the events and objects being described, and a Description it satisfies classifies them with EventTypes and Roles—for instance F:Cause and F:Effect for causality, F:Participant for participation, F:Composite and F:Component for mereology. Reifying these relations as situations makes them first-class individuals, which is what lets one event be described from several contextual points of view at once. The non-functional requirements—extensibility, axiomatization, modularity, reusability, and separation of concerns—are handled by a pattern-oriented design aligned with DUL.","core_discovery":"The central claim is that Event-Model-F is a formal model of events that satisfies all six functional requirements it identifies: participation of objects in events, temporal duration and spatial extension, structural relationships (mereological, causal, correlative), documentary support, and event interpretations. The paper shows this by instantiating the Descriptions and Situations pattern for each requirement, so each aspect of an event becomes a situation that satisfies a description defining typed roles. It also claims that this combination is what separates Event-Model-F from existing event models: none of the surveyed models offers full structural support together with support for multiple interpretations of the same event. The conclusion states this distinction explicitly.","pith_inferences":["The paper leaves the causal and correlative roles unconstrained by axioms; a natural follow-up is to define consistency conditions, such as Cause and Effect being distinct events with a shared Justification, and then test whether real causal narratives remain representable.","Because every relation is reified as a situation, event descriptions become large graphs of interconnected situations; this may make querying and reasoning more expensive than in flat event models, so an empirical comparison of reasoning performance on realistic event logs is a direct way to probe the model's practical interoperability claim.","The interpretation pattern suggests an integration strategy the paper does not itself develop: map each proprietary event model's relations onto one of the six F patterns, and use interpretations to keep divergent views separate while sharing the common event core."],"forward_implications":["Event descriptions built with Event-Model-F can be exchanged between heterogeneous systems and checked automatically for consistency with the model's axioms, not just their syntax.","Domain ontologies can be plugged into the pattern roles, such as a Citizen or AffectedBuilding role in an emergency-response ontology, so existing domain knowledge is reused without being remodeled.","One real-world occurrence can be represented under multiple event interpretations simultaneously by binding separate participation, composition, causality, correlation, and documentation situations under an interpretation pattern.","Composition constraints let events be related by arbitrary temporal orderings, including disjointness and overlap, using Allen-style relations; relative spatial constraints are expressed through participating objects' locations.","The model explicitly represents correlation as a common-cause relation, so correlated effects can be recorded even when the common cause itself is unknown."],"supporting_citations":[{"why":"Supplies DOLCE, the foundational ontology whose top-level distinction between enduring objects and perduring events the model adopts.","marker":"[10]"},{"why":"Supplies DOLCE+DnS Ultralite (DUL), the ontology that Event-Model-F is aligned with.","marker":"[14]"},{"why":"Introduces the Descriptions and Situations pattern used to reify every event aspect as a situation satisfying a description.","marker":"[9]"},{"why":"Defines the six aspects of event model E that the functional requirements are blended from.","marker":"[29]"},{"why":"Defines the journalism interrogatives in Eventory that inform the requirement analysis.","marker":"[28]"},{"why":"Provides CIDOC CRM as a comparison event model; its limited 'resulted in' causality is contrasted with F's causality pattern.","marker":"[4]"},{"why":"Provides the Event Ontology as a comparison baseline with a simple sub-event part-of relation.","marker":"[23]"},{"why":"Provides the event calculus as a comparison baseline for knowledge-representation events.","marker":"[16]"},{"why":"Gives the common-cause definition of correlation that the correlation pattern formalizes explicitly.","marker":"[25]"},{"why":"Supplies the view of causality as explanation with a justification that the causality pattern's F:Justification role is built on.","marker":"[12]"}],"fun_headline_variants":["Formal event model covers time, space, parts, causes, and views","Event-Model-F: six patterns for shared event semantics","DOLCE-based event model unifies interpretations and relations","New event model supports composition and rival interpretations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The causal and correlative parts of the model depend on the assumption that reifying a causal or correlative statement as a situation with a justification role preserves the statement's intended meaning—yet the ontology supplies no axioms that actually constrain what Cause, Effect, or Correlate may do, and if that reification is semantically inadequate the claim of comprehensive structural support breaks down.","fun_headline_variants_meta":{"raw":{"variants":["Formal event model covers time, space, parts, causes, and views","Event-Model-F: six patterns for shared event semantics","DOLCE-based event model unifies interpretations and relations","New event model supports composition and rival interpretations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00019,"raw_usage":{"total_tokens":1267,"prompt_tokens":803,"completion_tokens":464,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":419,"completion_tokens_details":{"reasoning_tokens":396}},"tokens_in":419,"tokens_out":464,"duration_ms":4536,"temperature":1.0,"reasoning_tokens":396,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:56:02.719175+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Open the published OWL axiomatization and ask whether a consistent model can assert the same event as both F:Cause and F:Effect of itself under a single F:Justification; because no constraints on these roles are given, the pattern would accept such a self-cause, showing that it does not yet carry the causal semantics it claims to represent.","supporting_citations":[{"cited_title":"Oberle and et al","cited_arxiv_id":null,"evidence_quote":"Defines the journalism interrogatives in Eventory that inform the requirement analysis."},{"cited_title":"Casati and A","cited_arxiv_id":null,"evidence_quote":"Supplies DOLCE, the foundational ontology whose top-level distinction between enduring objects and perduring events the model adopts."},{"cited_title":"Ericsson and M","cited_arxiv_id":null,"evidence_quote":"Supplies DOLCE+DnS Ultralite (DUL), the ontology that Event-Model-F is aligned with."},{"cited_title":"Arndt, R","cited_arxiv_id":null,"evidence_quote":"Introduces the Descriptions and Situations pattern used to reify every event aspect as a situation satisfying a description."},{"cited_title":"Oberle and et al","cited_arxiv_id":null,"evidence_quote":"Defines the six aspects of event model E that the functional requirements are blended from."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides CIDOC CRM as a comparison event model; its limited 'resulted in' causality is contrasted with F's causality pattern."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Event Ontology as a comparison baseline with a simple sub-event part-of relation."},{"cited_title":"Franz, S","cited_arxiv_id":null,"evidence_quote":"Provides the event calculus as a comparison baseline for knowledge-representation events."},{"cited_title":"M¨ uhl, L","cited_arxiv_id":null,"evidence_quote":"Gives the common-cause definition of correlation that the correlation pattern formalizes explicitly."},{"cited_title":"Doerr, C.-E","cited_arxiv_id":null,"evidence_quote":"Supplies the view of causality as explanation with a justification that the causality pattern's F:Justification role is built on."}],"review_version":1}