An investigation into the epistemological foundations of science, the nature of technological progress, and the ethical frameworks that govern engineering practice.
May 2026
Epistemology, the study of the nature and scope of knowledge, provides the framework through which we evaluate scientific claims. In the history of science, two primary competing doctrines have shaped our understanding of how we acquire “certain” knowledge: Rationalism and Empiricism.
Rationalism asserts that true, certain knowledge is derived primarily from logical reasoning and mathematical deduction. Proponents such as René Descartes and Baruch Spinoza argued that sensory experience is inherently misleading and fallible.
In the Cartesian view, the senses can be deceived (as in dreams or optical illusions). Therefore, the only reliable starting point for knowledge is a priori reasoning—concepts that are known independently of experience. This approach often leads to a “top-down” model of science, where universal laws are deduced from self-evident axioms.
In direct opposition, Empiricism maintains that sensory perception is the exclusive origin of knowledge. Philosophers like John Locke and David Hume posited that the human mind begins as a tabula rasa (blank slate), and all ideas are the result of sensory impressions.
Under Empiricism, science is a “bottom-up” process of induction: we observe specific phenomena and generalize them into laws. However, David Hume famously identified a critical logical flaw in this approach: the Problem of Induction. Hume demonstrated that inductive generalization—assuming the future will resemble the past—is a psychological habit rather than a strict logical necessity.
Understanding the difference between these inference methods is vital for scientific methodology:
Immanuel Kant attempted to reconcile Rationalism and Empiricism through Transcendental Idealism. He argued that while all knowledge begins with experience (Empiricism), it is not all derived from experience.
The mind is not a blank slate, but rather possesses innate mental categories—such as space, time, and causality—that actively structure sensory data. We don’t see the “thing-in-itself”; we see the world as our mind constructs it through these a priori lenses. In science, this means that while we need data from the world, that data only becomes “knowledge” when it is structured by the rational categories of the human intellect.
Consider the statement: “The sun will rise tomorrow because it has risen every day for billions of years.”
Positivism is a philosophical system that holds that every rationally justifiable assertion can be scientifically verified or is capable of logical or mathematical proof. Founded by Auguste Comte in the early 19th century, Positivism sought to establish science as the final and most superior stage of human intelligence.
Comte argued that the history of human knowledge and society evolves linearly through three distinct stages of development. Each stage represents a different way of explaining the world and the causal mechanisms behind phenomena.
In this initial stage, the human mind searches for the essential nature of beings and the first and final causes of all effects. Phenomena are explained as the results of the immediate action of supernatural beings (gods, spirits, or a single deity). This stage is characterized by belief rather than observation.
Transitional in nature, the metaphysical stage replaces supernatural agents with abstract entities or “personified abstractions.” Rather than gods, “nature” or “essences” are used to explain phenomena. For instance, an object might be said to fall because of its innate “heaviness” or “essence,” a view frequently associated with Aristotelian physics.
This is the final, scientific stage. The mind gives up the vain search for Absolute notions and the causes of the universe. Instead, it applies itself to the study of laws—their invariable relations of succession and resemblance. Reasoning and observation, combined, are the means of this knowledge.
Comte didn’t just define stages of thought; he also proposed a hierarchy of the sciences based on their complexity and their dependence on the sciences that preceded them. According to Comte, sciences developed in a specific order:
He believed that as we move up the hierarchy, the subjects become more complex and less general, eventually leading to the scientific study of society itself.
While many of Comte’s specific ideas have been superseded, the core tenet of Positivism—that empirical evidence and the scientific method are the only valid paths to truth—remains a dominant (though debated) force in modern thought. It established the “Context of Justification” as the focus of science: the idea that it doesn’t matter how you discovered a theory (discovery); what matters is that you can prove it (justification).
In the early 20th century, a group of philosophers and scientists known as the Vienna Circle advanced a radical form of positivism known as Logical Positivism. Their goal was to make philosophy as rigorous as science by purging it of “pseudo-problems” and metaphysical speculation.
The cornerstone of Logical Positivism is the Verifiability Principle of Meaning. This principle asserts that a statement is meaningful if, and only if, it is either:
Under this strict criterion, many traditional philosophical questions—especially those concerning ethics, aesthetics, and religion—were dismissed as “cognitively meaningless.” If you cannot observe or measure something, the Logical Positivists argued, you are not talking about anything real.
Logical Positivists viewed science as a strictly objective and rational enterprise. They made a sharp distinction between two aspects of scientific work:
Despite its initial popularity, Logical Positivism faced internal contradictions. The most famous “suicide” of the movement was the realization that the Verification Principle itself cannot be verified. You cannot use sensory observation to prove that “only sensory-verifiable statements are meaningful.”
Furthermore, scientific laws presented a problem. A law like “all water boils at 100°C” refers to an infinite number of instances across time and space. Since we cannot observe every instance, such laws are technically unverifiable, yet they are the core of science.
Logical Positivism pushed science toward extreme precision and formalization. It championed the idea of Unified Science—that all branches of science (from physics to psychology) should eventually be reducible to a single, consistent logical framework. While the movement eventually collapsed under its own rigidity, its emphasis on clarity and empirical evidence remains a pillar of scientific practice.
Karl Popper rejected the idea that science works by “verifying” theories through induction. Instead, he proposed Critical Rationalism, arguing that science progresses not by proving theories right, but by proving them wrong. This doctrine is known as Falsificationism.
Popper observed that no matter how many white swans you see, you can never logically conclude that “all swans are white.” However, seeing just one black swan can instantly and decisively disprove that statement.
He argued that scientists should not look for evidence to “confirm” their theories (since you can almost always find confirmation if you look for it). Instead, they should make bold conjectures and then subject them to the most rigorous attempts at refutation.
For Popper, the mark of a scientific theory is its falsifiability. A theory is scientific only if it makes definite predictions that can be tested and potentially proven false.
Popper famously criticized Psychoanalysis (Freud) and Marxist historicism for being too “elastic”—they could explain away any contradictory data, making them unfalsifiable and therefore pseudo-scientific.
Popper described the growth of knowledge as a cyclic, evolutionary process. Science starts with a problem, proposes a tentative theory, attempts to eliminate errors through testing, and inevitably ends with a new, more refined problem.
In this framework, we never “know” the truth with certainty; we only have theories that have “withstood the test of time” and have not yet been falsified. These are called corroborated theories. The goal is to move closer to the truth by systematically eliminating falsehoods.
A central question in the philosophy of science is: What distinguishes science from non-science or pseudo-science? This is known as the Demarcation Problem. Solving it is not just a theoretical exercise; it has real-world implications for what we fund, what we teach in schools, and what we trust as evidence in courts.
Two major attempts to solve this problem came from the Logical Positivists and Karl Popper.
Consider Astrology.
Regardless of whether one prefers verification or falsification, most philosophers of science distinguish between how a theory is found and how it is judged:
The demarcation criteria usually focus on the Justification. It doesn’t matter if you came up with a theory in a dream; it’s only “science” if it can be justified through empirical testing (Verification) or withstand attempts at disproof (Falsification).
In engineering, we must distinguish between “rigorous models” and “intuition” or “rule of thumb.” While an engineer might use a heuristic based on years of experience (Tacit Knowledge), the safety of a bridge must be justified through models that are themselves scientific—meaning they are falsifiable and built on verifiable physical laws.
In his seminal work The Structure of Scientific Revolutions, Thomas Kuhn challenged the idea that science is a slow, steady accumulation of knowledge. Instead, he argued that science operates within conceptual frameworks called Paradigms.
A paradigm is more than just a theory; it is a whole “worldview” shared by a scientific community. It includes:
When scientists work within a paradigm, they are not usually trying to falsify their theories (as Popper suggested). Instead, they are engaged in “Normal Science”.
Kuhn described Normal Science as “puzzle solving.” During these periods, the paradigm is taken for granted. Scientists work to extend the paradigm’s reach, improve the accuracy of measurements, and solve localized problems.
If a result doesn’t fit the paradigm (an anomaly), the scientist usually assumes their calculation was wrong, or their equipment failed, rather than blaming the paradigm itself. Normal science is essentially dogmatic; it requires a deep commitment to the existing framework to function.
The PlantUML generation failed (remote + local).
@startuml
skinparam rectangle {
BackgroundColor<> #FFD700
}
rectangle "The Paradigm" <> {
(Theories)
(Techniques)
(Metaphysical Beliefs)
(Standard Problems)
}
[Researcher] --> [The Paradigm] : "Toolbox"
[The Paradigm] --> [Puzzle Selection]
[Puzzle Selection] --> [Normal Science]
@enduml
While normal science tries to ignore them, anomalies occasionally accumulate. When these anomalies can no longer be explained away or ignored, the scientific community enters a state of Crisis. This crisis is the precursor to a scientific revolution, where the old paradigm is eventually replaced by a new one.
A critical part of Kuhn’s argument is that observation is theory-laden. This means we do not see the world “as it is”; we see it through the lens of our theories. A chemist and a physicist looking at the same bubble chamber might “see” two entirely different things based on their training and the paradigms they inhabit.
When a scientific paradigm faces enough anomalies that it can no longer function, a Scientific Revolution occurs. This isn’t just an update to old ideas; it is a fundamental shift in how scientists see and interact with the world.
A paradigm shift is often compared to a “gestalt switch.” Just as an image can be seen as both a duck and a rabbit, but not both at once, a scientist switches from one way of seeing the world to another. Examples include:
Kuhn’s most controversial claim was the principle of Incommensurability. He argued that competing paradigms lack a “common measure” or objective vocabulary.
Because of this, you cannot simply say Paradigm B is “better” than Paradigm A based on pure data. Why?
The PlantUML generation failed (remote + local).
@startuml
left to right direction
skinparam packageStyle rectangle
package "Newtonian Paradigm" as P1 {
component "Gravity as Force" as Gravity
component "Absolute Time" as AbsTime
component "Constant Mass" as ConstMass
}
package "Einsteinian Paradigm" as P2 {
component "Gravity as Curvature" as GravityC
component "Relative Time" as RelTime
component "Variant Mass" as VarMass
}
interface "Wall of Incommensurability" as Wall
P1 - Wall
Wall - P2
note top of Wall : No common language\nNo neutral data
@enduml
Kuhn’s view implies that science does not move linearly toward “The Truth.” Instead, it is a sequence of different ways of looking at the world that are better at solving the puzzles prioritized by each community. While a new paradigm might solve the anomalies of the old one, it might also lose some of the problem-solving capabilities or insights of its predecessor.
Because paradigms are incommensurable, the transition from one to another is not purely logical. It involves a “conversion” similar to a religious experience. Younger scientists are often more likely to adopt the new paradigm, while older ones may never fully make the switch—a phenomenon Planck described by saying that science “advances one funeral at a time.”
While many philosophers sought to define science as a purely objective, detached logical system, Michael Polanyi argued that scientific progress relies heavily on the “personal knowledge” of the practitioner.
Polanyi, a physical chemist himself, rejected the idea that scientists are neutral observers who simply follow a mechanical “method.” He argued that all knowledge involves an active, personal commitment. If you strip away the observer’s passion, judgment, and expertise, you don’t get “pure data”—you get nothing meaningful at all.
Polanyi’s most famous concept is Tacit Knowledge: the idea that “we know more than we can tell.”
There are skills and understandings that we possess but cannot fully articulate in words. Consider:
How do we acquire this knowledge? Polanyi described a process of interiorization. When we learn to use a tool (like a probe or a microscope), we eventually stop “feeling” the tool in our hand and start “feeling” the object at the end of the tool. The tool becomes part of our body; it is internalized.
In science, we “dwell in” our theories. We use them as tools to perceive the world. This means our scientific conclusions are grounded in a framework we have personally accepted and integrated into our way of being.
Because knowledge is tacit and personal, science requires discretionary judgment. Rules and methods cannot cover every possible situation. The scientist must decide when a law applies, when a measurement is “good enough,” and when a theory is beautiful enough to be true despite conflicting data. This judgment is not subjective whim; it is an exercise of expertise within a responsible community of peers.
While the natural sciences often aim for universal, objective laws, the human sciences frequently adopt Interpretivism. This perspective asserts that truth is contextual and depends heavily on the relationship between the observer and the observed.
Though they sound similar, these two branches of Interpretivism focus on different scales of meaning-making:
In these frameworks, there is no “view from nowhere.” Every observation is filtered through a cultural and linguistic lens.
The PlantUML generation failed (remote + local).
@startuml skinparam componentStyle rectangle component "The Object" as Obj component "Cultural/Linguistic Lens" as Lens actor "Observer" as Obs Obj -right-> Lens Lens -right-> Obs : "Interpreted Reality" note bottom of Obs : Truth is contextual\nand inter-subjective @enduml
A specific subset of interpretivism is the Critical Perspective. This approach assumes that what we call “truth” is often defined by those in power. The goal of critical research is not just to understand the world, but to “emancipate” people by deconstructing the power structures (of class, race, gender, or technology) that define our reality. For a critical theorist, science is never neutral; it either reinforces the status quo or challenges it.
This philosophical shift explains the methodological divide between different disciplines:
In the context of engineering and technology, this means acknowledging that a bridge or an algorithm is not just a physical object, but a social artifact that carries specific meanings and power dynamics.
To understand how science and technology interface with human life, we must look at the structures of experience and interpretation. Two key methodologies for this are Phenomenology and Hermeneutics.
Phenomenology is the rigorous study of conscious experience from the first-person perspective. It seeks to describe the essential elements of everyday experience without making hypotheses about their causal or biological origins.
A key technique in phenomenology is Bracketing (Epoché). This involves stripping away your cultural preconceptions, scientific theories, and biases to describe a phenomenon exactly as it appears to your consciousness. For example, a phenomenologist wouldn’t describe “seeing a table” in terms of light waves hitting a retina; they would describe the experience of the table’s “threeness” or its “readiness-at-hand.”
While phenomenology focuses on description, Hermeneutics focuses on interpretation. Originally used to interpret ancient texts, it is now applied to all human phenomena.
The central concept is the Hermeneutic Circle. This is the paradox that:
Understanding is thus an iterative, circular process of moving back and forth between the details and the broader context.
Martin Heidegger, a leading figure in phenomenology, used the example of a carpenter’s hammer to explain how we relate to the world.
In the philosophy of science, these perspectives remind us that data never speaks for itself. It must be interpreted (Hermeneutics) within a world of human experience (Phenomenology). When engineers design systems, they are not just moving matter; they are creating “tools” that will ideally become “ready-to-hand” for the user, disappearing into the background of the user’s life.
What exactly is “technology”? While we often think of silicon chips and steam engines, the philosophical definition of technology has expanded significantly over the centuries, moving from “applied science” to a fundamental mode of human existence.
The word “technology” as we use it today was essentially coined in the English language by Jacob Bigelow. In his 1831 book, Elements of Technology, he defined it as the “principles, processes, and nomenclatures of the more conspicuous arts, particularly those which involve applications of science.”
For Bigelow, technology was the systematic study of the “useful arts.” It represented the bridge between abstract scientific knowledge and practical human labor.
Lewis Mumford provided a much broader definition. He argued that technology is not just hardware (tools and machines), but also the organizational structures that coordinate human effort.
Mumford conceptualized the “Megamachine”: the first machine was not made of metal, but of people. For example, the construction of the Egyptian pyramids required a rigid, hierarchical organization of thousands of laborers. This social machine treated human beings as interchangeable components, operating with the precision and predictability of a mechanical clock.
The PlantUML generation failed (remote + local).
@startuml
skinparam rectangle {
BackgroundColor<> #FFF7AA
}
rectangle "Hardware View" {
(Tools)
(Engines)
(Computers)
}
rectangle "Social View" <> {
(Organization)
(Hierarchy)
(Labor Standards)
}
() "Technology" as Tech
Tech -- "Hardware View"
Tech -- "Social View"
note right of Tech: Beyond the physical object
@enduml
This expansion of the definition suggests that technology is a “human way of doing.” It involves:
When we study the philosophy of technology, we are not just studying gadgets; we are studying the systems that shape how we work, live, and interact with the physical world.
In his influential essay The Question Concerning Technology, Martin Heidegger argued that the essence of technology is “nothing technological.” By this, he meant that to understand technology, we must look beyond its tools and machines to the way it influences our perspective on reality.
Most people hold an instrumental view of technology: technology is just a neutral tool used by humans to achieve certain ends. In this view, a hammer is “good” or “bad” depending only on how you use it.
Heidegger rejected this. He believed technology is substantive—it is a “way of revealing” the world. It fundamentally changes how we perceive everything around us.
Heidegger used the term Enframing to describe the essence of modern technology. Enframing is a way of “ordering” the world such that everything appears as a Standing Reserve (Bestand)—a resource to be used, measured, and optimized.
Consider a river:
When we look through the lens of technology, trees become timber, mountains become ore, and even human beings become “human resources.”
The “danger” Heidegger warned about was not that machines would take over, but that the technological mindset would become the only way we see the world. If we see everything as merely a resource for optimization, we lose our ability to experience the “truth” or the “poetic being” of things. We forget that we are part of the world and start acting like its masters, treating the earth as a giant warehouse.
Heidegger suggested that by recognizing the essence of technology (instead of just using its products), we can develop a more critical and “free” relationship with it. We can “use” technology while still remaining open to other ways of being and revealing.
While science often focuses on success and verification, the philosophy of engineering identifies failure as the primary driver of knowledge. Henry Petroski famously theorized that all engineering calculations are essentially “failure calculations.”
Petroski argued that we learn very little from a bridge that stands. A successful design only “corroborates” (in Popper’s terms) the current engineering paradigms. However, a bridge that collapses provides definitive new information. It reveals the limits of our theories, the flaws in our materials, or the errors in our assumptions.
In this sense, “Design is the active anticipation of failure.” The goal of an engineer is not to ensure a system will never break—which is impossible—but to understand exactly how it will break and to ensure that failure occurs in a safe, predictable manner.
Petroski noted that engineering history moves in cycles. Success breeds confidence, which leads to more daring and efficient designs. Eventually, this push for efficiency leads to “cutting it too thin,” resulting in a catastrophic failure. This failure then triggers a period of conservative design and the development of new, more robust theories.
The PlantUML generation failed (remote + local).
@startuml
skinparam activity {
BackgroundColor #FFFFFF
BorderColor Black
}
start
:Successful Design;
:Increased Efficiency / Confidence;
:Daring Innovation;
:Margin of Safety Decreases;
if (System Fails?) then (yes)
:Catastrophic Analysis;
:New Knowledge / Standards;
:Conservative Redesign;
else (no)
:Continue Scaling;
endif
stop
@enduml
This highlights a key difference between science and engineering:
For an engineer, a “safety factor” is an acknowledgment of our ignorance. We build things 2x or 3x stronger than “necessary” because we know our philosophical and mathematical models of the world are incomplete.
Christopher Alexander, an architect and design theorist (who heavily influenced software engineering patterns), conceptualized the act of design as a search for a perfect “fit” between two entities: Form and Context.
A design is successful when the Form and the Context coexist without friction. However, Alexander pointed out that it is much easier to define a “misfit” than a “fit.” We don’t notice when a door handle works perfectly; we only notice when it is too high, too small, or slippery.
Alexander argued that engineering design is conceptually the “elimination of misfits.” Instead of trying to create “beauty” or “perfection” directly—which are abstract and subjective—the engineer identifies specific points of friction (misfits) and systematically removes them.
This creates a negative feedback loop:
The PlantUML generation failed (remote + local).
@startuml
skinparam componentStyle rectangle
package "The Design Problem" {
component "Form (Solution)" as Form
component "Context (Environment)" as Context
}
Form <-right-> Context : "Interaction"
note top on link : Search for 'Fit'
[Identify Misfit] -down-> Form : "Modification"
Form -left-> [Test in Context]
[Test in Context] -up-> [Identify Misfit] : "New Frictions"
@enduml
Alexander also addressed the problem of complexity. In modern engineering, the number of potential misfits is so high that the human mind cannot track them all simultaneously. His solution was decomposition: breaking the problem down into smaller “sub-problems” that can be solved somewhat independently. This approach laid the groundwork for Modular Design and Object-Oriented Programming.
By solving for “misfits” in small, manageable modules, the engineer can assemble a complex “Form” that fits a complex “Context.”
The process of engineering is remarkably similar to the process of scientific inquiry, particularly the Popperian model of conjecture and refutation. In engineering, this is formalized as the Engineering Design Cycle.
The design cycle is a continuous, iterative loop that brings a concept from abstract requirements to a physical reality. It generally consists of four main stages:
The PlantUML generation failed (remote + local).
@startuml
skinparam activity {
BackgroundColor #ADD8E6
}
start
:Specifications
(Problem);
repeat
:Synthesis
(Design Conjecture);
:Analysis
(Attempted Refutation);
:Diagnosis
(Identify Failures);
repeat while (Goal Met?) is (no)
:Successful Design
(Corroboration);
stop
@enduml
Each loop through the cycle increases the “corroboration” of the design. We never prove a design is “perfect”; we only prove that it has withstood all the tests we have subjected it to.
This iterative nature reflects the “Evolutionary Epistemology” of science. The most successful designs are those that have survived the most brutal “selection pressures” (tests and failures).
One challenge in this cycle is the “Synthesis” phase. Because synthesis is a creative act, it is influenced by the designer’s training, the current paradigms of the engineering community, and even the available tools. This reminds us that technology, like science, is “theory-laden.” The way we define the problem (Specifications) often pre-determines the types of solutions we are capable of synthesizing.
A common debate in the philosophy of technology is whether humans control technology or whether technology, in a sense, controls us. This leads to the concepts of Technological Determinism and the Law of the Hammer.
Abraham Kaplan (and later popularized by Abraham Maslow) famously stated: “I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail.”
In engineering and science, this means that the availability of a specific technology frequently precedes the necessity for it. Rather than humans identifying a biological or social “need” and then designing a tool to solve it, the implementation of a new technology often “invents” its own application.
Technological Determinism is the theory that a society’s technology determines its social structure and cultural values. In this view:
While “hard” determinists argue that technology is the sole driver of history (e.g., “The steam engine created the industrial middle class”), “soft” determinists argue that technology provides the possibilities and constraints within which culture must then choose its path.
Philosopher Jacques Ellul argued that in a technological society, the “means” (the tools) eventually become the “ends” (the goal). Instead of using technology to achieve a specific human good, the “efficiency” of the technology itself becomes the supreme value. We do things because we can, not necessarily because we should.
Recognizing the Law of the Hammer allows engineers to be more deliberate. It asks: “Am I solving a real problem, or am I just applying this new AI/blockchain/hardware because it’s the tool I have?” By deconstructing technological determinism, we regain a sense of “agency”—the belief that we can steer the development of technology toward humanistic goals rather than simply following the logic of the machine.
While many ethical frameworks focus on external rules or consequences, Virtue Ethics is an agent-centric approach. Developed primarily by Aristotle, it focuses on the cultivation of internal character traits (virtues) rather than just the adherence to laws.
Aristotle argued that the goal of human life is Eudaimonia—often translated as “flourishing” or “well-being.” We achieve this not by pursuing pleasure, but by performing our function well. For a human, our unique function is the use of reason. Therefore, a virtuous person is one who uses reason to live a life of excellence.
Virtue is defined as the “golden middle way” between two extremes of behavior:
In engineering, a virtue might be Accuracy.
The PlantUML generation failed (remote + local).
@startuml rectangle "Deficiency (Too Little)" as D rectangle "The #FFD700en Mean (Virtue)" as V rectangle "Excess (Too Much)" as E D --> V V --> E note bottom of V Practical Wisdom (Phronesis) end note @enduml
In a professional context, virtue ethics asks: “What would a morally worthy engineer do in this situation?”
This approach is highly relevant because external codes of ethics cannot cover every edge case of modern technology. A “virtuous” engineer possesses Practical Wisdom (Phronesis)—the ability to perceive which virtue is required in a complex, ambiguous situation. They do the right thing not because they fear punishment, but because they have internalized the values of their profession.
Virtue is not innate; it is a habit. We become just by doing just acts, and we become temperate by doing temperate acts. Professionalism, therefore, is a practice. By consistently striving for the “#FFD700en Mean” in design, communication, and decision-making, an engineer cultivates the character necessary to navigate the ethical challenges of the field.
Utilitarianism is a consequentialist ethical theory founded by Jeremy Bentham and refined by John Stuart Mill. It proposes a simple but powerful rule: ethical actions are those that maximize overall utility (well-being, happiness, or “the good”) for the greatest number of people.
The core of utilitarianism is the “Greatest Happiness Principle.” When faced with a dilemma, one should calculate the potential outcomes of each choice and select the one that results in the highest net benefit.
In this framework, no individual’s interests are inherently more important than anyone else’s. The goal is to maximize the aggregate “good” across the entire affected population.
The engineering profession has operationalized utilitarianism through Cost-Benefit Analysis. When designing a safety system or a piece of infrastructure, engineers and policymakers weigh the financial costs against the quantified benefits (lives saved, accidents prevented, time reduced).
Example: A city is deciding whether to install a new traffic light.
If the “utility” of the benefits outweighs the “cost,” the project is deemed ethical and rational under a utilitarian framework.
In contrast to utilitarianism’s focus on consequences, Deontology (from the Greek deon, meaning duty) focuses on the inherent rightness or wrongness of actions. The most influential deontologist was Immanuel Kant.
Kant argued that morality is grounded in reason and that certain duties are absolute (categorical). He proposed the Categorical Imperative as a test for moral duty. One famous formulation is:
“Act only according to that maxim whereby you can, at the same time, will that it should become a universal law.”
In other words, would it be okay if everyone did what you are about to do? If the result would be a logical contradiction or a world that no one would want to live in (e.g., a world where everyone breaks promises), the action is immoral.
A second formulation of the Categorical Imperative is critical for engineering and research ethics:
“Act in such a way that you treat humanity, whether in your own person or in the person of any other, never merely as a means to an end, but always at the same time as an end.”
This is known as the Respect for Persons principle. It dictates that individuals possess inherent dignity and must be treated as free and equal moral agents. They should never be used as mere “tools” for the benefit of others, even if that benefit is large.
Kantian ethics supports the establishment of human rights:
Under a deontological framework, certain actions are forbidden regardless of their utility.
Generic ethical theories like utilitarianism and deontology can sometimes feel too abstract for daily life. Common Morality refers to the aggregate of widespread moral beliefs and standards shared by a culture. Philosophers have attempted to systematize this “common sense” into practical frameworks.
W.D. Ross argued that morality is not based on a single “master rule” like utility. Instead, we have several prima facie (at first sight) duties that are universally obligatory unless they conflict with a stronger duty in a specific context.
Key prima facie duties include:
Ross acknowledged that these duties often conflict (e.g., your duty to be truthful might conflict with your duty to avoid harming someone). In such cases, you must use your judgment to decide which duty is “actual” or most pressing in that specific situation.
Bernard Gert further systematized common morality into ten fundamental rules. These are primarily prohibitive (“don’t…”) and focused on preventing harm.
Gert made a distinction between moral rules (which are required and whose violation deserves punishment/censure) and moral ideals (which are aspirational, like volunteering or extraordinary acts of kindness). For an engineer, being “ethical” requires following the rules (e.g., don’t deceive users about data privacy), while being “exemplary” involves pursuing moral ideals (e.g., designing low-cost medical tech for underserved communities).
This “pluralistic” approach—recognizing multiple valid duties—reflects the reality of engineering practice. Most ethical dilemmas are not “right vs. wrong” but “right vs. right” (e.g., the duty to protect the environment vs. the duty to provide jobs via a new factory). Systematizing these duties provides a vocabulary for discussing and resolving these conflicts.
When abstract theories fail to provide a clear answer to a complex real-world problem, engineers use specific methodological tools to find a resolution. Two of the most common are Line-Drawing and the Creative-Middle-Way approach.
This method is used to resolve application issues—situations where it is unclear whether a specific act falls under a certain moral category (e.g., “Is this gift a bribe or just a professional courtesy?”).
This approach is used when two or more moral duties compete (e.g., the duty to tell the truth vs. the duty to protect a client’s confidentiality).
Instead of treatings the situation as a zero-sum trade-off (where one duty must be “sacrificed” for the other), the engineer seeks a synthesis. The goal is an innovative course of action that satisfies the maximum number of conflicting obligations simultaneously.
Example: An engineer discovers a safety flaw in a competitor’s product.
The Creative-Middle-Way assumes that ethical problems are a form of design problem. Just as we use creativity to fit a “form” to a “context” in engineering, we use creativity to fit our “actions” to multiple moral “requirements.”
Philosopher Michael Davis developed a set of eight practical “tests” that engineers can use as heuristics to evaluate a proposed course of action. These tests distill complex ethical theories into simple, intuitive questions.
When faced with a dilemma, pass your proposed solution through these tests:
These tests cover different philosophical dimensions:
These tests are not meant to provide a mathematical “score.” Instead, they are tools for reflection. If a proposed action fails the “Reversibility Test” or the “Publicity Test,” it is a strong signal that the action is morally suspect, regardless of its short-term utility or convenience. By applying multiple tests, the engineer gains a holistic view of the ethical landscape, reducing the risk of being blinded by a single perspective or personal bias.