His work is science. It is not engineering.

Do Not Ask Neil deGrasse Tyson What Collection of Scientific Limits Can Be Used to Build a Moon Rover

Why You Should Never Ask Noam Chomsky Why Nature Controls His Research

And Do Not Ask Neil deGrasse Tyson What Collection of Scientific Limits Can Be Used to Build a Moon Rover

Here is a story…

Bob is an engineer. He designs gas turbine hot-section components for a living. He understands creep rupture, thermal barrier coatings, and the stress distribution in a fir-tree root attachment at operating temperature. He can tell you the Larson-Miller parameter for IN718 at 1200 degrees Fahrenheit without looking it up. He holds a Professional Engineer license and carries the legal liability that comes with it. When Bob stamps a drawing, he is telling the world that the physics are right, the margins are sufficient, and he will answer for it personally if they are not.

On Tuesday and Thursday afternoons, Bob sits in a social science elective he needs for his degree. He does not dislike the class. The professor is thoughtful, the discussions are interesting, and Bob has learned things about human behavior that genuinely surprised him. But he has noticed something that bothers him more each week: the professor keeps using the word “science.”

This week, the professor assigned a paper. The prompt: “Observation is the purest form of science. Discuss how the systematic recording of human behavior constitutes scientific methodology and why the social sciences represent the highest expression of the scientific method.”

Bob stared at the prompt for a long time. He thought about the thermal cycle test he had run that morning—sixteen specimens, identical conditions, identical results, the crack growth rate matching the Paris Law prediction to within two percent. He thought about the word “science” on that assignment sheet, and then he thought about the word “science” on his fatigue data, and he realized they were not the same word. Not even close.

Bob wrote a different paper. This is what it said.

The Question Nobody Asks

Ask Noam Chomsky, the noted linguist, why nature controls his research and he will look at you the way a cat looks at a ceiling fan. The question does not compute. Not because it is too complex for one of the most celebrated intellectuals of the past century, but because it does not apply. Nature does not control Chomsky’s research. Nature has no opinion on generative grammar, no stake in the minimalist program, no force that prevents a linguist from redefining the rules of syntax tomorrow morning over coffee. Chomsky operates in a world where humans define the constraints, humans modify the constraints, and humans discard the constraints whenever they see fit. His work is rigorous, it is brilliant, and it is a study. It is not science.

Now ask Neil deGrasse Tyson, the noted astrophysicist, what collection of scientific limits can be used to build a moon rover. Watch the pause. He can tell you every force acting on that rover. He can describe the regolith it will traverse, the thermal cycles it will endure, the radiation environment it must survive. He understands the science with extraordinary depth and communicates it with rare clarity. But he has never organized those scientific limits into a system of materials, mechanisms, and controls that produces a repeatable result on the lunar surface. That is not what he does. He observes nature act, and he explains why it acted that way based on the constraints the natural world imposes on the actions he observes. His work is science. It is not engineering.

The engineer who built that rover did something that neither Chomsky nor Tyson was trained, credentialed, or methodologically equipped to do. That engineer identified the relevant science—thermal expansion coefficients, bearing load capacities under reduced gravity, power density of radioisotope generators, fatigue behavior of titanium alloys at cryogenic temperatures—and organized those natural constraints into an integrated system that works. Repeatedly. Predictably. On the moon. Where there is no help desk, no software patch, no interpretive framework that lets you explain away a seized wheel bearing.

Three different disciplines. Three fundamentally different relationships with reality. And we have lost the language to distinguish between them because we allowed the words to be stolen.

The Corruption of Two Words

The words “science” and “engineering” used to mean something precise. Science meant the systematic observation and testing of natural phenomena under conditions that demanded repeatability. Engineering meant the organized application of scientific knowledge to produce systems that function reliably within nature’s non-negotiable boundaries. Both words carried weight because both words carried constraints that could not be argued away.

Then something happened. People began appending these words to things that had no business wearing them. Computer science. Software engineering. Social science. Data science. Sanitation engineering. The logic seemed harmless: if a discipline is rigorous, systematic, and produces useful results, surely it deserves the prestige of being called a science or an engineering practice.

But prestige is not the issue. Precision is. When you call something a science, you are claiming that the constraints governing that discipline are imposed by nature and that the results are independently repeatable. When you call something engineering, you are claiming that natural laws have been organized into a system that produces a predictable, repeatable outcome. These are not honorifics. They are definitional claims about the relationship between the work and physical reality. And those claims are false for most of the disciplines that now carry these titles.

A Clean Hierarchy

There is a hierarchy of words that, properly defined, eliminates the confusion entirely. Each level has a single test: who defined the constraint? The answer determines what the work is and what it is not.

Study

A study operates within human-defined constraints. The rules are created by people, the boundaries are negotiable, and the outcomes change when someone rewrites the rules. Linguistics is a study. Chomsky’s generative grammar is a framework built on axioms that humans selected. If a new axiom serves better, the old one is replaced. No natural law prevents the substitution. Economics is a study. The “laws” of supply and demand are descriptions of human behavior patterns, not physical forces. Change the culture, change the incentive structure, and the “law” bends or breaks. Software development is a study of logic. The constraints of a programming language exist because a human wrote a compiler that enforces them. Write a different compiler and the constraints dissolve. Computational complexity limits are properties of a made-up rule in a code language. They can be modified at will. The result may be less efficient, but the modification is possible because no force in nature prevents it.

This is not a diminishment. Studies produce extraordinary value. The entire infrastructure of modern communication, finance, law, and governance rests on studies. But they are not science, because their constraints are human artifacts, not natural ones.

Science

Science operates within nature-defined constraints. The rules are discovered, not invented. Gravity does not care about your interpretation. The speed of light does not adjust for your convenience. Entropy increases whether or not a committee votes to approve it. The defining characteristic of science is repeatability imposed by nature itself. Drop a ball on Earth and it accelerates at 9.81 meters per second squared. It did this before Newton described it. It will do this after every physics textbook has turned to dust. The observation does not create the phenomenon. The phenomenon exists independent of the observer, and any competent observer who replicates the conditions will observe the same result.

Tyson lives in this world. When he describes the lifecycle of a star, he observes a process governed by gravitational collapse, nuclear fusion, and radiation pressure, and he explains why the star behaves as it does based on the constraints nature places on those actions. The forces operate identically whether the star is observed from Earth or from a galaxy humans will never reach. His observations are constrained by nature. His explanations must conform to measurable, repeatable reality. He cannot redefine nuclear fusion the way Chomsky can redefine a syntactic rule.

Engineering

Engineering is the organization of science to produce technology. The engineer does not merely observe nature’s constraints. The engineer harnesses them. Every material selection, every dimensional tolerance, every thermal management strategy represents a decision made within boundaries that nature will enforce with absolute indifference to the designer’s intentions. The proof is not a publication. The proof is that the thing works, every time, under the conditions it was designed to meet.

Technology

Technology is a collection of scientific constraints arranged to always produce a desired output. A bearing is a technology. The coefficient of friction between the rolling elements and the raceway, the hardness of the steel, the thermal expansion behavior under operating temperature—all of these are scientific constraints organized so that when energy is applied, the result is always the same: low-resistance rotational motion under load. The bearing does not decide to work. It does not interpret the load. It does not reason about friction. The scientific constraints simply produce the output. Every time. The same way. Because nature does not offer alternatives.

This distinction matters. A technology is not a clever invention. It is not a gadget. It is a specific arrangement of natural constraints that guarantees a repeatable output. Consider the humble thermocouple—two dissimilar metals joined at a point. The Seebeck effect, a scientific constraint discovered in 1821, dictates that a temperature differential across dissimilar metal junctions produces a voltage. The thermocouple is a technology because it organizes that constraint to always produce one desired output: a voltage proportional to temperature. It does not interpret the temperature. It does not decide whether the reading matters. It simply converts a thermal condition into an electrical signal, every time, because the science permits no alternative. If the output produced is constrained by human-defined parameters rather than nature-defined, the product may be useful, it may be sophisticated, but it is not a technology in this sense. It is a product of study.

System

A system is a collection of technologies arranged to produce more than one repeatable outcome. A gearbox is a system. It combines bearing technology, gear tooth technology, lubrication technology, and housing technology. Each technology produces its own singular repeatable output. Arranged together, they produce multiple repeatable outcomes: speed reduction, torque multiplication, directional change, power distribution. The gearbox does not choose which outcome to produce. The configuration determines the outcome. Nature enforces every constraint in every technology simultaneously. The system works because every technology inside it works, and every technology inside it works because the science inside it is non-negotiable.

Mulact

A mulact is a collection of systems arranged to act in an infinite number of repeatable actions. The word is new because the concept has never been cleanly separated from the words that have been corrupted around it. A mulact implies no logic, no autonomy, no decision-making. It is the scissors test at the highest level of organization. A pair of scissors is a technology—a pivot, blade geometry, material hardness, all arranged to produce one repeatable output: a shearing cut. There is no logic in what scissors do when energy is applied. The scientific constraints simply execute.

A mulact operates the same way, at a vastly higher scale of organization. Consider a piano. It is a collection of systems—hammer mechanism system, string system, damper system, pedal system, soundboard system—arranged so that when energy is applied and direction is given, it can produce an infinite number of repeatable acoustic actions. Eighty-eight keys. Three pedals. Every combination is a physical input. The human presses a key, a mechanical linkage throws a felt hammer against a string, the string vibrates at a frequency determined by its mass, tension, and length, and the soundboard amplifies it through acoustic physics. Press a different key, different string, different frequency, different sound. Press three keys simultaneously and the physics of harmonic interference produces a chord. Step on the sustain pedal and the damper system physically reconfigures so that every string is free to resonate. The piano does not decide what note to play. It does not interpret the music. It does not reason about harmony. The human supplies the direction. The piano supplies the capability. Nature supplies the constraints. The result is an infinite action space bounded entirely by physics, with no logic required at any level. Press middle C today, press it tomorrow, press it in a hundred years with the same string at the same tension—same frequency. Every time. Because nature does not negotiate.

And the piano proves the infinite in “mulact” is not an exaggeration. Eighty-eight individual keys produce 88 individual outputs. But press two keys simultaneously and you are not producing two separate sounds. You are producing a third thing—a unique acoustic event governed by the physics of wave superposition. The harmonic interference pattern between those two frequencies is physically distinct from either note played alone. Three keys pressed together produce a different interference pattern than any two of them. Four keys, different again. The combinatorics alone are staggering: choose two keys from eighty-eight and you have 3,828 unique combinations. Choose three and you have 109,736. Choose four and you have over 2.4 million. Choose five from eighty-eight keys and you exceed 43 million physically distinct acoustic outputs—and every single one is governed entirely by the physics of vibrating strings in a shared resonant cavity. Now add the three pedals that physically reconfigure the damper system, the hammer distance, and the string selection. Now add variation in strike force—press a key gently versus hard and the hammer velocity changes, which changes the harmonic content of the sound, not merely the volume. Now add duration. Now add the sequence in which keys are pressed and released, because a string still resonating when a second key is struck produces a different interference pattern than two keys struck into silence. You are not approaching infinite. You are functionally there. And every output is repeatable, governed entirely by nature, and requires zero logic. That is a mulact.

This point demands absolute clarity. A mulact does not require software to change its output, and it does not require reconfiguration between actions. Press a C on the piano and the hammer, string, and soundboard systems produce a specific acoustic output. Now press an E. The systems do not have to be externally reconfigured. No die set changes. No selector is turned. No mechanical adjustment is made. One action has no influence on the next, and the mulact responds instantly with no internal or external modification necessary to produce a completely different output. The physics of the E string are different from the physics of the C string, and nature produces the corresponding result the moment energy is applied. This is what keeps the mulact entirely within the hierarchy. No logic selects between the outputs. No software interprets the input. The human chooses which key to press, and the mulact produces the result that the physics of that key demands. Every output is governed by science. Every result is repeatable. The moment you replace that human choice with software that interprets, decides, and chooses, the software portion is no longer engineering. It is a study in execution. Human-defined constraints, modifiable at will. The mulact remains engineering. The software never was.

And this is where the hierarchy reveals something profound. When someone adds software to a mulact—when they add code that makes decisions, interprets inputs, and selects actions—the software portion is a study of logic. Full stop. Not engineering with a software layer. Not an enhanced mulact. A study of logic. The code is human-defined. The decision-making rules are modifiable at will. Every algorithm is a human-created constraint that a different human can rewrite tomorrow morning. The autonomy is a construct, not a natural phenomenon. That is exactly the definition of study: human-defined constraints, modifiable at will. A robot, then, is not a higher form of engineering. A robot is a mulact—which is engineering—with a study of logic bolted onto it. The mulact portion remains bound by nature, repeatable, non-negotiable. The software portion is a study of logic, bound by human rules, changeable, negotiable. And when the robot structurally or functionally fails, it is the technology or system within the robot which is defective. When the software fails, it fails for study reasons—a bad algorithm, an unforeseen input, a flawed assumption. A human-defined constraint encountered a situation its human creator did not anticipate. When the engineering fails, it fails for science reasons—a material exceeded its yield strength, a thermal cycle caused fatigue cracking, a lubricant degraded beyond its functional range. The failure modes are categorically different because the constraints are categorically different. One is negotiable. The other never was.

Why the Distinction Matters

The dilution of these words is not a semantic quibble. It has consequences that affect public trust, professional accountability, and the safety of the systems people depend on.

When a software developer calls himself a software engineer, he borrows credibility from a discipline that is licensed, regulated, and legally accountable for the safety of its products. A licensed professional engineer who stamps a drawing has wagered his career and potentially his freedom on the correctness of that analysis. If the structure fails and the stamp was negligent, the engineer faces legal consequences. The constraints are not just physical. They are institutional, and they exist because the physical consequences of engineering failure are measured in lives.

When a social researcher calls her work social science, she borrows credibility from a methodology that demands independent replication under controlled conditions. But social phenomena cannot be controlled the way a chemistry experiment can. Variables cannot be isolated. Initial conditions cannot be replicated. The “results” are statistical tendencies, not repeatable facts. Valuable insights, but not science in the way that thermodynamics is science, and calling them science obscures a critical difference in the reliability of the conclusions.

The public, understandably, does not distinguish between these uses. When they hear “science says,” they assume the same rigor whether the claim comes from a physicist describing quantum entanglement or a sociologist describing cultural trends. When they hear “engineer,” they assume the same accountability whether the title belongs to a structural engineer designing a bridge or a prompt engineer configuring a chatbot. The words have been diluted to the point where they no longer communicate the thing that made them valuable: a defined, non-negotiable relationship with physical reality.

We have already seen what happens when the distinction is lost. Social studies dressed as science have driven public policy with the confidence of physics but the reproducibility of astrology. Software studies dressed as engineering have produced systems with the complexity of a turbine but the testing rigor of a first draft. The words mattered because they used to communicate something about the reliability of the underlying work. Now they communicate nothing except a vague sense of technical respectability.

The Complete Chain

The hierarchy is not arbitrary. It is a dependency chain, and it flows in only one direction.

Study — human-defined constraints, modifiable at will.

Science — nature-defined constraints, discovered and repeatable.

Engineering — the organization of science to produce technology.

Technology — scientific constraints arranged to always produce one desired output.

System — technologies arranged to produce more than one repeatable outcome.

Mulact — systems arranged to act in an infinite number of repeatable actions.

Engineering depends on science. Science depends on nature. Technology depends on engineering. Systems depend on technologies. Mulacts depend on systems. At no level in this chain is a human-defined constraint sufficient. At every level, nature has the final word. And at every level, the proof is the same: the thing works. Repeatably. Predictably. Without negotiation.

Notice what is absent from this chain. There is no level where someone’s opinion determines the outcome. There is no level where a theory’s popularity substitutes for its accuracy. There is no level where changing the observer changes the result. The entire hierarchy stands on a single foundation: nature is not negotiable. Every word in the chain inherits that foundation, and any discipline whose constraints can be rewritten by the people who work within them does not belong in it. Not because such disciplines lack value, but because they lack the one quality that defines every link in this chain—absolute, repeatable, non-negotiable accountability to physical reality.

Bob’s Conclusion

Bob finished the paper at two in the morning. He read it once more, thought about the professor’s prompt—“Observation is the purest form of science”—and shook his head. Observation is the beginning of science, not the whole of it. Recording what people do in a room is not the same as measuring what a material does under load. One produces an interpretation that might hold true next Tuesday. The other produces a data point that will hold true until the sun burns out.

And science itself is only the beginning. What comes after—engineering, technology, systems, mulacts—is what transforms understanding into capability. Into the bearing that carries the load. Into the rover that crosses the lunar surface. Into the piano that produces an infinite number of repeatable sounds from eighty-eight keys, three pedals, and the non-negotiable physics of vibrating strings, without once asking why.

He thought about Chomsky, brilliant in his study, unconstrained by nature. He thought about Tyson, brilliant in his science, observing nature and explaining why it acts as it does based on the constraints the world places on every action. And he thought about the engineer who sits between them and below neither of them, doing the thing that neither of them can do: organizing what nature demands into something that works.

Do not ask Noam Chomsky why nature controls his research. Nature has nothing to do with it. Do not ask Neil deGrasse Tyson what collection of scientific limits can be used to build a moon rover. That is not his art. Ask the engineer. She will tell you exactly which limits apply, exactly how they interact, and exactly how the system she designed accounts for every one of them. And then she will show you the rover, running across the lunar surface, doing what it was designed to do. Repeatedly. Predictably. In a place where no amount of interpretation, logic, or eloquence will save you if the engineering is wrong.

That is why the words matter. That is why the distinction is not academic. And that is the big why.

Bob turned in the paper. He got a C-minus. He did not appeal the grade. The professor’s grading rubric was, after all, a human-defined constraint.

---

Herbert Roberts, P.E. spent 32 years in aviation R&D across two companies and has spent the last eight years analyzing accidents for attorneys under his PE license, translating engineering findings into legal language. Inventor’s Mind publishes every Tuesday, Wednesday, and Thursday at inventorsmindblog.com.