Skip to content

References

Academic sources and theoretical foundations for ProEthica. The component literature below mirrors the definition-source records (iao:0000119 and dcterms:source) recorded on the corresponding ontology classes and verified against the cited texts. The authoritative copy of each record is browsable on OntServe alongside the class it grounds.

Framework Foundations

The formal specification D = (R, P, O, S, Rs, A, E, Ca, Cs) draws on three foundational works:

  • McLaren, B. M. (2003). Extensionally defining principles and cases in ethics: An AI model. Artificial Intelligence, 150(1-2), 145-181. DOI
  • Berreby, F., Bourgne, G., and Ganascia, J.-G. (2017). A declarative modular framework for representing and applying ethical principles. AAMAS 2017, 96-104. ACM | PDF
  • Tolmeijer, S., Kneer, M., Sarasua, C., Christen, M., and Bernstein, A. (2021). Implementations in machine ethics: A survey. ACM Computing Surveys, 53(6), Article 132. DOI

The dissertation grounding the framework, and the peer-reviewed papers derived from it, are listed on the Publications page.

Component Literature

Each of the nine components anchors its definition in verified sources. The entries below are the definition-source records of the component classes and their principal subclasses.

R. Roles

The role account is individuated by position or standing in professional practice, with obligation generation attributed to professional roles.

  • Oakley, J. and Cocking, D. (2001). Virtue Ethics and Professional Roles. Cambridge University Press. DOI Regulative ideals, role-generated obligations, and the founding-value limit: a professional role cannot license violating the good the profession serves.
  • Kong, K., Lam, P., and Cheng, W. (2020). Corpus-based empirical approach to professionalism: Identifying interactional roles and dispositions in professional codes of ethics. Journal of Applied Linguistics and Professional Practice, 14(1), 3-28. DOI Source of the four relational archetypes (provider-client, professional peer, employer relationship, public responsibility) that form the relational role axis, complementary to the occupational axis.
  • Doernberg, S. and Truog, R. (2023). Spheres of morality: The ethical codes of the medical profession. The American Journal of Bioethics, 23(12), 8-22. DOI The spheres-of-morality reading: the same person may bear a professional role in one engagement and a participant role in another.
  • Davis, M. (1991). Thinking like an engineer: The place of a code of ethics in the practice of a profession. Philosophy & Public Affairs, 20(2), 150-167. Code governance as a professional-side relation.
  • Wendel, W. B. (2024). Philosophical perspective: The challenge of role morality. In Canceling Lawyers: Case Studies of Accountability, Toleration, and Regret. Oxford University Press. DOI Role morality on the employer-relationship archetype.

P. Principles

  • McLaren, B. M. (2003) (above). Extensional definition of abstract, open-textured principles through decided cases, the extensional grounding of the principle subtree.
  • Frankel, M. S. (1989). Professional codes: Why, how, and with what impact? Journal of Business Ethics, 8(2-3), 109-115. DOI Aspirational codes state ideals along a continuum of code types.
  • Morley, J. et al. (2021). Ethics as a service: A pragmatic operationalisation of AI ethics. Minds and Machines, 31(2), 239-256. DOI Origin of the constitutional-principles analogy: principles provide a foundation and require interpretation in context.
  • Taddeo, M. et al. (2024). From AI ethics principles to practices: A teleological methodology to apply AI ethics principles in the defence domain. Philosophy & Technology, 37(1), 42. DOI Teleological balancing of principles in application.
  • Prem, E. (2023). From ethical AI frameworks to tools: A review of approaches. AI and Ethics, 3(3), 699-716. DOI The high-level, context-free statement of principles, and the grounding of domain-specific principles.
  • Hallamaa, J. and Kalliokoski, T. (2022). AI ethics as applied ethics. Frontiers in Computer Science, 4, 776837. DOI Grounding of the fundamental ethical principle kind.
  • Anderson, M. and Anderson, S. L. (2018). GenEth: A general ethical dilemma analyzer. Paladyn, Journal of Behavioral Robotics, 9(1), 337-357. DOI Principles learned from expert examples, cited on the professional-virtue principle kind.

O. Obligations

  • Donohue, B. (2017). Toward a BFO-based deontic ontology. SoLe-BD 2017 at ICBO, CEUR Workshop Proceedings 2137. Obligations as directive information entities issued with normative authority.
  • Dennis, L. A., Fisher, M., Slavkovik, M., and Webster, M. (2016). Formal verification of ethical choices in autonomous systems. Robotics and Autonomous Systems, 77, 1-14. DOI Specification of abstract principles into precise context-dependent requirements.
  • Ross, W. D. and Stratton-Lake, P. (2007). The Right and the Good. Clarendon Press. The prima facie duty account grounding the defeasibility structure: several duties bear on one situation, and the more stringent becomes the duty proper in the circumstances.
  • Anderson, M. and Anderson, S. L. (2011). A prima facie duty approach to machine ethics. In Machine Ethics, 476-492. Cambridge University Press. DOI Computational formalization of prima facie duties overridden on occasion by other duties.
  • Stenseke, J. (2024). On the computational complexity of ethics: Moral tractability for minds and machines. Artificial Intelligence Review, 57(4), 105. DOI Computational analogue of the ought-implies-can principle: an obligation presupposes the capacity to discharge it.

The domain obligation classes for safety, competence, disclosure, faithful agency, reporting, confidentiality, and attribution anchor to specific NSPE Code provisions. See Professional Code Grounding.

S. States

  • Jones, T. M. (1991). Ethical decision making by individuals in organizations: An issue-contingent model. Academy of Management Review, 16(2), 366-395. DOI Moral intensity: the situational characteristics that affect ethical salience.
  • Berreby, F., Bourgne, G., and Ganascia, J.-G. (2017) (above). Inertial and non-inertial fluents: time-varying properties that affect ethical assessment.
  • Anderson, M. and Anderson, S. L. (2018) (above). Ethically relevant features of a situation, the GenEth vocabulary.
  • Almpani, S. and Stefaneas, P. (2022). Argumentation-based logic for ethical decision making. Studia Humana, 11(3-4), 46-52. DOI Context-determined priority ordering, cited on the risk and resource state kinds.
  • Stenseke, J. (2024) (above). Cited on the competence state kind.

Rs. Resources

  • McLaren, B. M. (2003) (above). Professional ethical knowledge exists primarily in accumulated cases, the grounding of the case-precedent kind.
  • Davis, M. (1991) (above). Codes of ethics as conventions that coordinate professional practice.
  • Frankel, M. S. (1989) (above). The profession as a normative reference group.
  • Bench-Capon, T. and Sartor, G. (2003). A model of legal reasoning with cases incorporating theories and values. Artificial Intelligence, 150(1-2), 97-143. DOI Grounding of the legal-resource kind.

A. Actions

  • Sarmiento, C., Bourgne, G., Inoue, K., and Ganascia, J.-G. (2022). Action languages based actual causality in decision making contexts. PRIMA 2022, LNCS 13753, 243-259. DOI | arXiv Volitional actions (an action follows from a volition) and the NESS-based attribution of causal responsibility.
  • Wright, R. W. (1985). Causation in tort law. California Law Review, 73, 1735-1828. DOI The NESS test (necessary element of a sufficient set) that Sarmiento et al. formalize.
  • Kroll, J. A. (2020). Accountability in computer systems. The Oxford Handbook of Ethics of AI, 180-196. DOI Causality, fault, and duty as the three bases of accountability.
  • Floridi, L. and Sanders, J. W. (2004). On the morality of artificial agents. Minds and Machines, 14(3), 349-379. DOI Accountability separated from intentionality-requiring responsibility.
  • Govindarajulu, N. S. and Bringsjord, S. (2017). On automating the doctrine of double effect. IJCAI 2017, 4722-4730. DOI Intention: the ethical evaluation of the same action differs with the knowledge and intention of the agent.
  • Berreby, F., Bourgne, G., and Ganascia, J.-G. (2017) (above). The volitional/non-volitional boundary between actions and events.

E. Events

  • Kowalski, R. and Sergot, M. (1986). A logic-based calculus of events. New Generation Computing, 4(1), 67-95. DOI The Event Calculus: happenings initiate and terminate fluents.
  • Berreby, F., Bourgne, G., and Ganascia, J.-G. (2017) (above). The distinction between agent-caused, exogenous, and automatic occurrences that grounds the event origin subclasses.
  • Sarmiento, C. et al. (2022) (above). Causal chains: actions produce events that initiate new states.
  • Almpani, S., Stefaneas, P., and Frangos, P. (2023). Formalization of ethical decision making: Implementation in the data privacy of wearable robots. International Journal of Extreme Automation and Connectivity in Healthcare, 5(1), 1-13. DOI Emergency contingency protocols on the event side.

Ca. Capabilities

  • Tolmeijer, S. et al. (2021) (above). The four ethical-competence requirements structuring the recognition-and-reasoning competencies: norm competence, situational awareness, learning and adaptation, and explanation and justification.
  • Epstein, R. M. and Hundert, E. M. (2002). Defining and assessing professional competence. JAMA, 287(2), 226-235. DOI Professional competence as the judicious use of knowledge, skill, reasoning, and reflection in practice.

Cs. Constraints

  • Arkin, R. C. (2008). Governing lethal behavior: Embedding ethics in a hybrid deliberative/reactive robot architecture. HRI 2008, 121-128. DOI Constraints as negative behavioral limits that determine the permissible action space, with prohibition constraints evaluated first.
  • Ganascia, J.-G. (2007). Modelling ethical rules of lying with Answer Set Programming. Ethics and Information Technology, 9(1), 39-47. DOI Default rules that retain a general prohibition while tolerating justified exceptions in extraordinary circumstances.
  • Dennis, L. A. et al. (2016) (above). Specification of abstract principles into precise context-dependent constraints.
  • Benzmüller, C., Parent, X., and van der Torre, L. (2020). Designing normative theories for ethical and legal reasoning: LogiKEy framework, methodology, and tool support. Artificial Intelligence, 287, 103348. DOI Deontic-logic encodings that enable formal verification of normative theories.
  • Stenseke, J. (2024) (above). The legal-versus-moral enforcement point, cited on the legal-constraint kind.

Professional Code Grounding

The domain principle and obligation classes anchor to specific provisions of the NSPE Code of Ethics, and two derive from Board of Ethical Review case-law doctrine rather than provision text:

  • National Society of Professional Engineers (2019). Code of Ethics for Engineers. Publication 1102. Fundamental Canons, Rules of Practice, and Professional Obligations. Each anchored class records its provision designation, such as Rule of Practice II.2 on the competence obligation and Fundamental Canon I.1 on the safety obligation.
  • NSPE Board of Ethical Review (2004). Case 04-8, Public Welfare: Client Action Following Engineer's Services. The monitor-to-completion reading behind the corrective-action-monitoring principle.
  • NSPE Board of Ethical Review (1999). Case 99-8, Incomplete Plans and Specifications. The personal-conscience doctrine (originating in Case 82-5) behind the whistleblower principle.
  • National Society of Professional Engineers v. United States, 435 U.S. 679 (1978). Held the ban on competitive bidding an unreasonable restraint of trade. The rescinded provision is the documented grounding of the free-services prohibition principle.

Analysis Methods

ProEthica composes decision points, compares cases to precedents, and generates entity-grounded arguments using structured argument mining and practical professional ethics methodologies.

Toulmin Model of Argumentation

Toulmin, S. E. (1958/2003). The Uses of Argument. Cambridge University Press. DOI

The model decomposes arguments into six components, which map onto extracted entities:

Toulmin Component ProEthica Entity Role in Ethical Argument
Claim Board conclusions The ethical determination
Data/Grounds Actions, events, states What happened in the case
Warrant Principles, obligations Professional duties justifying the claim
Backing Code provisions Authority supporting the warrant
Qualifier Capabilities, constraints Limits on the claim
Rebuttal Conflicting obligations Counter-arguments

Question emergence uses the same model: data (triggering events and actions) invoke warrants, and competing warrants arise when the same data supports different conclusions through different normative pathways.

Argument Validation

Generated arguments are validated with a three-tier test derived from the role ethics of Oakley and Cocking (2001): entity grounding (argument components must reference extracted entities), founding value (no argument may endorse violating the good the profession serves), and professional virtue (virtues invoked by an argument require capability support). All computations use entities extracted during the pipeline steps, not free text.

Ethical Tensions

Extracted tensions document conflicts between obligations and constraints. Each tension is characterized by conflict type (obligation versus obligation, after the defeasible default rules of Ganascia 2007; obligation versus constraint, after the hierarchical constraint management of Dennis et al. 2016) and scored for moral intensity on the six dimensions of Jones (1991):

Component Definition Professional Application
Magnitude Impact on the founding professional good Severity of events
Social consensus Professional consensus via codified standards Explicit code provision citations
Probability Likelihood of violating professional obligations Professional foresight requirement
Temporal immediacy When consequences affect key goods Immediate versus long-term duty timing
Proximity Role-relationship binding strength Direct professional relationships
Concentration Scope of professional impact Number of affected parties

Line-Drawing Method

Harris, C. E., Pritchard, M. S., and Rabins, M. J. (2018). Engineering Ethics: Concepts and Cases (6th ed.). Cengage Learning.

Decision point composition applies the line-drawing approach: identify paradigm cases at the extremes, identify the features that make them paradigmatic, place the test case on the continuum for each feature, and determine where the ethical boundary lies.

Causal-normative links connect professional actions to the obligations they engage: an action fulfills an obligation, violates it, or raises it (puts it in force, to be resolved by a later action). The fulfills and violates assignments are grounded in NESS causal-responsibility analysis (Wright 1985; Sarmiento et al. 2022); the raises assignment is grounded in the Event Calculus treatment of an obligation as a fluent (Berreby et al. 2017; Dennis et al. 2016; Dennis and del Olmo 2021, A Defeasible Logic Implementation of Ethical Reasoning). Each link is materialized as an object-property edge from the action to the extracted obligation individual, so the full chain is queryable, not restated as prose.

Fluents and Temporal Ordering

A state is a fluent: a time-varying property holding over an interval. A happening (an action or event) initiates or terminates fluents, and inertial fluents persist until terminated (Kowalski and Sergot 1986; Berreby et al. 2017). Where the Event Calculus says what holds and why it changed, OWL-Time says when:

  • Allen, J. F. (1983). Maintaining knowledge about temporal intervals. Communications of the ACM, 26(11), 832-843. The thirteen interval relations ordering case happenings.
  • Cox, S. and Little, C. (2017). Time Ontology in OWL. W3C Recommendation. W3C The OWL-Time properties that express the Allen relations in committed case graphs.

Transformation Classification

Marchais-Roubelat, A. and Roubelat, F. (2015). Designing a moving strategic foresight approach: Ontological and methodological issues of scenario design. Foresight, 17(6), 545-555. DOI

Board resolutions are classified by four transformation types from the action-based scenario framework: transfer (the ethical burden moves to another party), stalemate (competing obligations remain in tension), oscillation (duties shift between parties over time), and phase lag (delayed consequences reveal obligations not initially apparent).

Precedent Discovery

Precedent discovery combines sentence embeddings with case-based reasoning to find relevant prior Board of Ethical Review decisions.

  • Richter, M. M. and Weber, R. O. (2013). Case-Based Reasoning: A Textbook. Springer. Publisher
  • Sun, Z., Zhang, K., Yu, W., Wang, H., and Xu, J. (2024). Logic rules as explanations for legal case retrieval. LREC-COLING 2024, 10747-10759. ACL Anthology
  • Wiratunga, N. et al. (2024). CBR-RAG: Case-based reasoning for retrieval-augmented generation in LLMs for legal question answering. ICCBR 2024, LNCS 14775, 445-461. DOI | arXiv

Ethical Framework Labels

The framework categories used in theoretical question generation (deontological, virtue ethics, utilitarian, and similar labels) are exploratory research features and are not formally defined in the ProEthica ontology. Questions labeled Theoretical are generated by prompting the language model to consider the case from different ethical perspectives. The labels indicate the intended analytical lens, not a formally modeled framework.