Introduction
This presentation lays out a plan for developing a theme followed implicitly across a four-decade engineering career: philosophy must inform engineering. Drawing on systems engineering at Hughes Aircraft and General Motors (including the EV1 electric vehicle), Digital Earth information systems with NASA, OGC, and the Spatial Web Foundation, and recent work on cyber-physical agents under IEEE 2874, it presents an explicit approach to anchoring standard engineering practices in deeper intellectual foundations.
Two background topics set the stage. First, engineering is defined from practice: systems engineering traces to the early 1900s, was formalized in Hall's A Methodology for Systems Engineering (1962), and is carried internationally by INCOSE into IEEE and ISO standards. An engineering project moves among stakeholder concerns, next-generation technology, design and build, and deployment, operation, and assessment — every phase, fully linked, all at once. Second, the fora where philosophy of engineering is discussed remain split between academic venues (fPET, the POET series, the Handbook of Engineering Systems Design, the Cosmos Institute, St. John's College Graduate Institute) and professional societies (IEEE Ethically Aligned Design and the 7000-series standards, IEEE SSIT, ACM SIGCAS, INCOSE) — with few cross-references between the two communities.
The core of the presentation is four engineering practices, each developed in real projects and each now being given a more robust philosophical standing: addressing stakeholder concerns, design as the creation of order, balancing innovation with liberty, and the agency of AI in cyber-physical ecosystems. Engineers exercise power in shaping infrastructure; making that power visible takes moral-philosophy practice.
Practice 1 · Stakeholders
Stakeholders include the direct users of a system and those affected indirectly by it. Methods now exist to make ethical concerns — responsibility and value — explicit inputs to design and governance, alongside conventional user requirements.
ISO/IEC/IEEE 42010 provides the method for identifying stakeholders and their concerns. In commercial automobile development this was the "Voice of the Customer"; on GM's EV1, vehicle range and recharge time were the requirements on which success rested. Safety and ethical concerns are secondary effects the engineer must equally surface.
Ethics supplies theories and reasons for thinking that this or that is right. IEEE 7000 translates them into Stakeholder Ethical Value Requirements (EVRs), drawing on utilitarian, virtue, and duty ethics — the key step being translation of the language of ethics into technical requirements.
IEEE 7007 and IEEE 2874 define ontologies for the entities relevant to a system's definition and operation — spanning virtuous, deontological, consequentialist, and composite ethical theories. As systems become more abstract and virtual, these ontologies too benefit from philosophical grounding.
Evaluating a partially viable product before deployment — including "red teaming" — surfaces stakeholder concerns that surveys and requirement documents miss. A well-developed set of stakeholder concerns is a valuable input to system design.
From Ethical Value Requirement to Technical Requirement
Practice 2 · Design
Aristotle observed in Metaphysics I.2 that it is the mark of the wise to order; Aquinas reformulated it as sapientis est ordinare. Design is the engineer's act of creating order — and the association of order with wisdom is an onus the engineer must take on.
The key design method is the allocation of functional requirements to physical components — function-to-form synthesis. The functional objective is decomposed, and each function is allocated to a physical element of the architecture, making the design traceable from stakeholder concern to component. This rational method sits among complementary design methodologies catalogued by INCOSE (Rechtin, Maier): participative methods grounded in stakeholders, normative methods grounded in standards, and heuristics drawn from experience. GeoRoundtable has exercised the method directly — synthesizing physical architectures from the same functional requirements for the IEEE 2874 Universal Domain Graph and for an electric vehicle, by human engineer and by AI, then comparing the results.
Philosophy gives this practice its depth. Herbert Simon framed design as an artifact's relation to its environment, seeking to survive and achieve (The Sciences of the Artificial, 1969/1996). Christopher Alexander derived order from functional analysis and fitness to the environment (1964, 2004), and at OOPSLA 1996 charged that "software engineers should take aesthetic-moral responsibility for the wholeness of what they build." The Agile manifesto designs while building, with no plans. And designs produce unanticipated results — affordances (Gibson) and mediation (Verbeek) — that no allocation matrix fully predicts.
A philosophy for design must therefore address design practice, wise ordering, and emergent forms together — a need taken up in the Handbook of Engineering Systems Design (Maier, Oehmen, Vermaas, 2022) and in natural philosophy's three themes: change as a constant stream of events, form as persistent structures that emerge, and agency as adaptive form.
"Software engineers should take aesthetic-moral responsibility for the wholeness of what they build."
— Christopher Alexander, OOPSLA 1996
Alexander's alternative to developers as "guns for hire" was a call to create living structures: when you build a thing, you must also repair the world around it and within it, so that the larger world at that one place becomes more coherent and more whole.
Each of the four practices answers that call in a different phase of engineering — making stakeholder values explicit, ordering wisely, governing the freedom to innovate, and taking responsibility for the agents we now set loose in the world.
Practice 3 · Innovation
Restraints on innovation to reduce risk must be balanced with liberty. The tension between innovation and constraint is one of the defining features of the current AI moment — and the tension is ineliminable but governable.
Creative Destruction & Permissionless Innovation
Innovations progress by displacing current technologies (Schumpeter, 1942; Aghion, 2014). Permissionless innovation holds that experimentation with new technologies should be permitted by default (Thierer, 2014), with problems addressed as they develop — against the precautionary principle, which would curtail innovations until proven harmless.
The Harm Principle as the Scope of Restraint
J.S. Mill's On Liberty (1859) offers one approach to scoping limits: the sole end for which mankind is warranted in interfering with the liberty of action is self-protection. Who should decide the limits on innovation, and when they are imposed, are judgments — not predictions.
AI Governance Needs Institutions, Not More Principles
The failure of current AI governance is less a shortage of ethical principles than a shortage of institutions capable of exercising practical wisdom at the speed and scale the technology demands.
Phronesis & Risk Management
Governing the ineliminable tension is precisely what phronesis — practical wisdom — is for. Risk management, a long-standing engineering practice for identifying harm during development, provides the working basis for those judgments. Now is the time to have the discussions and make them.
Practice 4 · AI & Agency
AI motivates us to consider the effect of technology on what it means to be human in the world. Engineering autonomous agents in cyber-physical ecosystems must include philosophical considerations — as the Spatial Web demonstrates.
Agents sense and respond to their environment and take action toward goals. We have long studied agency in the world; now we are creating it in earnest, and that is changing our understanding of the world. Agency is increasingly seen as a key element of natural philosophy, with levels of agency (Tomasello) extending well beyond humans. AI has redefined cognitive space itself through the spatial embedding of concepts; the epistemology of agents now plays out in networks — social physics — and neurosymbolic AI combines neural intuition with symbolic representations.
The Spatial Web (IEEE 2874) was developed as the standard for physical AI serving the needs of people and society — built on ideas that come from philosophy and that now need to feed back into philosophy. Its philosophical commitments are explicit: an ontology for entities in hyperspace; the activities of agents in epistemic networks; governance including norms and contracts; and collective intelligence emerging from social agents.
Where Next
Engineers design the order of artifacts that should contribute to human flourishing. Explicit philosophical engagement is indispensable for responsible socio-technical engineering.
Are these four practices viable paths? Is there value in studying them from a philosophic perspective? Are they a good set to bridge practice and philosophy? And how can the connection between academic and professional-society communities be increased?
Develop the practices in two complementary venues: IEEE — rigorous, consensus-driven engineering standards — and St. John's College — primary-text, seminar-based humanizing inquiry. Then apply them: which development projects could benefit most?
Related Work
Interested in integrating philosophical foundations into your engineering practice, standards work, or AI governance? Get in touch.
✉️ percivall@ieee.org