Many people have the idea that science progresses in a linear fashion—a straightforward march from ignorance to knowledge. If that were true, then there would be no such thing as scientific revolutions; old ideas would never need to be overturned, only built upon.
However, not only do scientific revolutions happen, they follow a specific pattern. The Structure of Scientific Revolutions lays out this pattern clearly and labels each part of it. The parts are:
There have been many such revolutions in history that completely changed the way people understood and approached the world. There was a mathematical revolution when the Greeks created proofs to show not only that certain mathematical formulae and postulates work, but why they work. Classical science had a revolution when laboratories and experiments became the preferred way of examining the world, a practice that may have begun with Galileo.
In the 19th century, entire fields of science were codified and sorted into paradigms, including heat, light, and electricity. Phenomena that had baffled scientists could now be categorized and understood. This was around the time of the Industrial Revolution, and likely a direct cause of it. There have been more since then, including Einstein’s theory of relativity and Planck’s first steps into quantum theory.
Revolutions happen because science is not a straightforward path toward what’s “true.” It’s more like a path away from what’s wrong.
Karl Popper’s work could be seen as the precursor to this concept of scientific revolutions. Popper taught that scientists come up with broad, testable ideas, and almost inevitably prove them wrong. Then they refine their ideas based on the new information and try again. This cycle of conjectures and refutations is similar to the idea of scientific revolutions, just on a much smaller scale.
The everyday work of scientists could be called normal science. This is when they work within a paradigm, trying to solve puzzles and correct discrepancies in it. While Structure occasionally sounds dismissive of normal science, the author actually had great respect for it.
Normal science doesn’t try to create anything new. Most research journals follow three trends:
However, even though it doesn’t look for them, normal science is exceptionally good at finding anomalies that don’t match up with the current paradigm. This is because, while normal science may restrict the breadth of scientists’ work to what the current paradigm allows, it allows for a depth of study that wouldn’t be possible otherwise. Scientists can devote great amounts of time and energy to studying minute details of the world.
Normal science usually finds what it thinks it’ll find. However, major discoveries tend to come when that doesn’t happen—in other words, when there’s an anomaly. Remember, that was the heart of Popper’s philosophy too.
However, even when anomalies are observed, people tend to ignore them or brush them aside. Even scientists will usually see what they expect to see, and not always what’s actually there. Because of this, it can take a long time for science to realize the significance of an anomaly.
Even when an anomaly is so huge that a paradigm clearly has to be rejected, it’s not a simple decision. The problem is that you can’t just reject a paradigm because there’s a problem with it, you have to substitute another paradigm that both solves the problem and shows promise for solving future problems at least as well as the current paradigm does. Doing that means comparing the two paradigms to each other, and to what you’ve observed in the world.
A crisis is a moment when two or more paradigms are competing to be adopted as the paradigm. It’s a time of scientific upheaval, when scientists are willing to try anything and debate even the most basic understandings of the field. Out of that scientific free-for-all, we get new ideas and, eventually, new theories.
These new theories eventually lead to a new established paradigm and, from some perspectives, a whole new world in which the scientists now have to work.
(Shortform note: The word incommensurable comes up frequently in Structure. It means “unable to be compared to each other.”)
Different paradigms are generally incommensurable. They address different problems, have different standards, and even use different words for things—or use the same words differently.
(Shortform note: This seems to conflict with the idea of “comparing paradigms to each other” in the previous section. That’s because this is a major simplification of what actually happens during a scientific crisis, which the full summary explores in more depth.)
Since it’s impossible to compare one paradigm to another, it can’t really be said that scientists choose a paradigm during a crisis—it’s more accurate to say that they are converted to one. Furthermore, since different paradigms function so differently, it’s often difficult for scientists to communicate and cooperate with each other.
However, this incommensurability can also be helpful. It contributes to specialization, which is how science continues to evolve and progress. The use of the word “evolve” is not accidental—science specializes, branches off, competes, and evolves in a way that’s completely consistent with biological evolution.
Recall that many people think of science as cumulative: New ideas are piled on top of old ones to increase the overall knowledge that we have. This goes hand in hand with the idea that science is moving toward some ultimate “truth” about the universe. And, in fact, this progressive piling-on of new information is exactly what normal science does.
Scientific revolutions, on the other hand, aren’t about moving closer to the truth—they’re about moving away from old ideas that don’t hold up anymore. We may need to reject the idea that there is some objective understanding of reality that science can find, some specific end goal that it’s working toward. Much like evolution, science has to change and adapt with its environment; just like there’s no “perfect” species waiting at the end of evolution, there may be no perfect truth waiting at the end of science.
Our current view of science is pulled mostly from textbooks and similar sources, the formal education we receive in school. However, these sources are like guidebooks or pamphlets for tourists, rather than accurate pictures of scientific history. As a result, some of our ideas about science are wrong.
Many of those books imply that “science” is only the theories and discoveries in the books themselves, rather than a huge assortment of different fields, ideas, and practices. Also, they often focus on experimental science (the so-called Scientific Method), but ignore the importance of connecting the results of those experiments to more general theories.
If our current ideas about science are true, then science is just the complete collection of knowledge we have about the world. In that case, scientific historians should be devoted to figuring out who made which contributions to that collection—and, on the other hand, figuring out how and why so many mistakes and myths were wrongly added to it over the years.
However, recently, people are finding that scientific history isn’t a straight line from ignorance to knowledge like they thought. There are a few anomalies that have led to this conclusion. (Shortform note: This book was first published in 1962.)
For one thing, questions that should have simple answers become harder—not easier—to answer the more you look into them. For example, you might ask, “Who discovered oxygen?” This seems like a simple question until you research it and find that at least three different scientists have reasonable claims to the discovery of oxygen.
People also have trouble explaining how today’s beliefs and understandings are more “scientific” than in the past. The methods of previous eras weren’t any less rigorous, and the conclusions scientists reach today aren’t any freer from human error.
Science has often supported beliefs that are totally incompatible with what we believe today, and yet we haven’t changed our beliefs simply because we got “better” at science. This seems to go directly against the idea that science is simply gathering facts and building on what’s already known.
This is all leading to a shift in how older kinds of science are viewed. Instead of steps along the way to our current ideas, they might have just been products of their time, eventually pushed aside by new ideas. If that’s the case then, rather than precursors to our current scientific beliefs, they simply aren’t relevant to modern science at all.
Therefore, instead of asking how those old views led to today’s science, some historians now ask how they fit into the beliefs of their times. That is to say, they look for internal consistency in the fields rather than some overarching consistency throughout history.
Problems often arise because accepted scientific theories are biased in some way, which leads to inaccuracies. Scientists will inevitably bring their own biases and experiences to whatever they study.
This is especially notable when studying relatively new fields. By examining different aspects of that field, and using different experiments and methods, scientists can come to wildly different—but no less valid—conclusions. For example, consider Jean-Baptiste Lamarck’s theory that physical changes an animal experienced in life could be passed down to its offspring; then compare that to Darwin’s well-known ideas about genetics. The two theories are totally irreconcilable and grounded in the scientists’ personal biases, yet neither could simply be dismissed as wrong.
Bias isn’t necessarily a bad thing—no scientist could work without pre-formed beliefs. Avoiding all established knowledge and methods is impossible. If you somehow succeeded then you’d be practicing philosophy, not science.
Normal science—that is, adding to what we already know—is all about fitting nature into the frameworks that we learn from formal education. Normal science is based on the assumption that we already know what the world is generally like, and we just need to fill in the details. As a result, normal science tends to cover up new ideas and discoveries that go against the established model. (Shortform note: This covering-up doesn’t have to be malicious. Often it’s just because scientists assume something that doesn’t fit the established model must be a mistake or an outlier.)
However, sometimes science will find an anomaly that can’t be hidden or fit into the current model, even after years of trying. This could be a new theory that doesn’t match the paradigm, or simply an observation that doesn’t line up with what was expected (Shortform note: “Theory” in the scientific sense means a well-developed idea that’s supported by data, not the informal meaning which is closer to a guess.)
When that happens, scientists have to break with traditional knowledge and start looking for a new model that fits this new information. That’s the beginning of a scientific revolution—but to truly deserve that name, the new model has to have big implications.
For example, when the movements of planets and moons couldn’t be resolved with the old Earth-centric model of the solar system, scientists of the day had to discard and replace it with the current sun-centered model. Therefore, the new heliocentric model didn’t just mean that scientists had made a mistake before, it meant they had to completely reevaluate how planets move and why. It was as if the scientists had to work in a totally new world with different natural laws.
Because a new model throws a lot of old work into question, scientists will often resist these revolutions at first. This is why it’s so hard for historians to pinpoint when major scientific revolutions happened: They were and are ongoing processes, but historians try to put specific timeframes on them. (Shortform note: Consider that, even today, there are people who don’t believe in evolution and fight to disprove it. While scientific historians can point to major events in the process, such as when Darwin’s On the Origin of Species was published, it would be very hard to say exactly when the revolution “happened.”)
The Structure of Scientific Revolutions expands on this idea of revolutions as ongoing processes, not concrete events. It also goes into more detail about the difference between normal science and revolutionary science.
The rest of the book focuses on three central questions:
Viewing scientific history purely as history isn’t enough to really understand how science progresses and why. Therefore, the book makes a number of generalizations about the psychology and behavior of scientists. It also goes into epistemology—the study of knowledge itself, and how to separate facts from beliefs. Especially in light of those generalizations, and knowledge drawn from fields that are not the author’s specialty, the ideas and theories in The Structure of Scientific Revolutions should be carefully checked for flaws.
The theory of scientific revolutions should be subject to examination and experiments just like any other scientific model; perhaps eventually someone will find an anomaly that completely invalidates it.
Normal science means study based on previous scientific knowledge. That knowledge is usually learned through textbooks and formal education, which outline the main questions of a field and the methods used to answer them. Before textbooks became common, the old classics like Aristotle’s Physica had much the same purpose.
These books became influential for two reasons: first, because they were remarkable enough to attract a loyal following at a time when many schools of thought were competing with each other; second, because they left a lot of questions and unsolved problems—puzzles for their followers to solve.
Scientific discoveries that have those two characteristics become paradigms: They are perfect examples of their fields, and a lot of information can be extrapolated from them. Paradigms are at the heart of normal science. They are the models that future experiments and theories are based on. This is why we use phrases like Newtonian physics: physics based on the paradigm of Newton’s discoveries.
Students prepare for their work in chosen scientific communities by studying paradigms. Physicists, for example, would study Newton’s work and how to apply his principles to modern science.
Because of this, scientists in the same community will follow the same rules and models in their work. This helps lead to scientific consensus, which is needed for normal science to progress.
Without a paradigm, it’s hard to know which facts are the most important. Scientists working without a paradigm will tend to gather whatever facts are easiest for them to find, or most relevant to their other areas of knowledge.
Because of this lack of a singular focus or guide, historians tracing a particular field of study back through time would likely find that, before there was a common paradigm, each scientist had to recreate the field almost from scratch.
For an example of this pattern, let’s look at the history of optics, the study of light. We see a series of scientific revolutions and changing paradigms going back to the 17th century: Modern studies of light are based on the belief that light is made of photons, which have characteristics of both particles and waves. However, before that model was developed in the early 20th century, most physics books taught that light was a wave. If we go back even further, scientists believed light was made of particles and looked for evidence that it puts pressure on solid objects.
However, before the 17th century, there were many competing schools of thought and no one universally-accepted paradigm. There were so many different schools of thought about light because, at the time, there were no common beliefs or understandings that they could take for granted. Due to the biases and past experiences each scientist brought, they came to wildly different—but equally valid—conclusions. Each school of thought supported itself by pointing to certain observations about the world that it could explain better than the others. Those observations were that school’s paradigms. These different schools all made significant discoveries that helped Newton to create the first almost-universally accepted model of light in the 17th century.
Scientific writings from before a paradigm was established are a confusing jumble. They tend to contain incorrect information (or at least information that doesn’t fit with our current paradigm), accurate facts with implications that scientists of the day didn’t realize, and observations that were too abstract or too complex to relate to the rest of the field.
It shouldn’t be surprising that science performed without a paradigm leads to a lot of different schools of thought. In fact, it might be more surprising that those discrepancies usually disappear.
This usually happens because one school of thought becomes prominent over the others. That prominent idea will attract the next generation of scientists, and other schools of thought will die out.
For example, early electricians (in the scientific sense, not the modern sense) came to their first paradigm because a scientist who thought electricity was a liquid tried to capture some in a jar and created the Leyden jar, a primitive battery. What he learned from that experiment formed the basis of the most convincing arguments of the time, which eventually led to his school of thought becoming the paradigm of the time.
However, it’s important to note that even though electricity-as-liquid was the dominant argument of the time, it couldn’t answer every question about electricity. Paradigms must be more convincing than competing ideas, but they don’t need to explain everything. In fact, those unanswered questions are what make normal science possible.
In some ways, establishing a paradigm is what cements a field of study as a science. Once a paradigm is set, scientists don’t need to recreate the field from scratch every time they begin their studies. This allows them to explore problems that are both more concrete and more obscure. For example, once Darwin’s work cemented evolution as a paradigm, future evolutionary biologists could move on to questions like, “What is the mechanism by which evolution happens?”
Most fields of study will follow a similar pattern. However, there are a few exceptions. Fields like Math and Astronomy have paradigms dating back to prehistoric times, so there’s no way of tracing them to their beginnings. Some more modern fields like Biochemistry came about as combinations of already-existing specialties, so there was no early period before Biochemical paradigms existed. Finally, some fields are so new (relatively speaking) that the first universally accepted paradigm hasn’t been established yet; some types of social science are still in this state.
A scientific paradigm is something to be further studied and refined under new or stricter conditions. Paradigms tend to be very limited and imprecise when they’re first established, so normal science is all about expanding the paradigm and filling the gaps in it.
Normal science doesn’t try to find new theories or explanations for things. In fact, new information that doesn’t fit the paradigm is usually ignored or brushed aside completely. However, that’s not to say that normal science should be disparaged. While the paradigm hugely reduces the scope of normal science, it allows the depth of study that’s key to advancing science. In fact, most scientists practice normal science for their entire careers.
Normal science has three common areas of interest. The first is taking what we already know and improving the specificity, accuracy, or breadth of that knowledge. Scientists will try to make observations that line up with the example set by the paradigm, thereby making the paradigm stronger and more convincing.
One example was Foucault demonstrating that the speed of light is slightly reduced by water; his experiment strengthened the then-current wave theory of light, since the previous particle theory said that water would “pull” on the light particles and make them move faster.
The second area of interest is finding problems in the paradigm and solving them. For example, Newton proposed the existence of gravity, but it wasn’t until a century later that scientists were able to measure a universal gravitational constant. Science’s inability to come up with that constant was seen as a major problem with the paradigm, and it may have led to a crisis if it had remained unsolved.
The third area of interest is closely related to the second. It involves applying a paradigm to phenomena other than the ones it was originally based on, proving that it works under more circumstances. Broadening the paradigm is another method of strengthening it.
However, normal science is not purely experimental; it has theoretical aspects too. Theoretical science’s areas of interest are similar to experimental science.
First, theoretical science uses existing theories to predict new facts, preferably ones that can then be confirmed by experiments. Second, it connects theories to observations about the world, thereby making the theory stronger. Third, it finds problems in the accepted theory and solves them—sometimes this happens simply by rewording or clarifying the theory, rather than actually adjusting it.
One of the most notable things about normal science is that it doesn’t try to find new information, but only to confirm and refine current beliefs. Results that don’t fit into the paradigm are often set aside, and sometimes don’t make sense until much later.
For example, early experiments that showed the attractive power of electricity didn’t give consistent or easily explained results, so they were set aside as unexplained phenomena. It wasn’t until the paradigm shifted away from the liquid model of electricity that they started making sense.
Given this fact, it might seem strange that so many people are so interested in normal science. However, normal science is still challenging—it’s often about how to get the expected results, rather than the results themselves. It’s like solving a jigsaw puzzle: Even though you already know what the outcome is supposed to be (the picture on the box), the challenge and the enjoyment come from figuring out how to achieve it.
A scientist can prove his or her skill and creativity by solving a difficult scientific puzzle. This can also involve developing all sorts of new procedures and equipment, which are worthy challenges for any scientist.
A key aspect of puzzles is that they must have solutions. Therefore, scientific fields will tend to look for problems that—if their paradigms hold true—are solvable. Problems outside of that scope will often be dismissed as belonging to a different field of science, or as just too obscure or troublesome to bother with.
Also, both puzzles and scientific problems must have rules by which those solutions can be found. For example, in the eighteenth century, scientists consistently failed to apply Newton’s laws of motion to the movement of the moon. Some scientists suggested replacing one of his mathematical laws with a different one that would work for the problem—however, that would have been changing the paradigm, and thus breaking the rules. Finally, in 1750, a scientist found the solution using Newton’s laws as they were written, and so the paradigm remained in effect.
Scientific “rules” fall into 4 major categories. The first is scientific laws and theories, such as Newton’s Laws of Motion. The second is which scientific instruments may be used to study the paradigm, and how.
The third category is broad, high-level concepts. For example, in the mid 17th century, scientists believed that the entire universe—including all natural forces—was made of particles. They believed that every natural phenomenon could be explained by particle size and shape, how they moved, and how they interacted with each other. Therefore, the “rules” of that time would force them to explain any findings in those terms, such as describing light as tiny, solid bodies that travel in a straight line and have some minuscule amount of momentum.
The fourth category is personal rules for scientists. Unlike the previous categories, this tends to be the same across all fields of study. In short, scientists must be committed to examining and understanding the world, with special focus on some particular part of it. If their studies reveal anomalies, either the methods or the theories must be refined.
However, scientists in a given field have more in common than just rules they follow, which is why we place so much emphasis on the paradigms they use. Rules come from paradigms, but paradigms can be used as guides even without established rules (this is explored in the next chapter).
While historians can generally find the paradigms of a given community fairly easily, finding the specific rules that community followed is often much harder. For one thing, scientists may follow the same paradigm, but disagree on how to interpret it. Also, like professionals in any field, scientists have experience and knowledge that they either don’t communicate clearly or don’t realize others lack. Therefore, historians sometimes don’t recognize or understand the rules of the time because they’re couched in unfamiliar terms and assumptions.
Finally, just because there is a paradigm doesn’t mean that there is necessarily a set of rules to go with it. Sometimes historians are looking for something that simply isn’t there. However, even when paradigms lack specific rules, they can still restrict the scientific field by guiding the work.
To illustrate this point, consider a question asked by the philosopher Ludwig Wittgenstein. He wondered how we know to apply terms like “chair,” or “game,” to something we’ve never seen before.
The brief version of the answer is that we apply those terms to things that resemble other things we know by those names. For example, if something has legs and a seat and a back, then we’ll probably call it a chair because it looks a lot like other chairs. This holds true even if it’s a type of chair we’ve never seen before.
Applying Wittgenstein’s idea to science, scientists following the same paradigm will end up using the same (or similar) terms and methods in their work. Just like we can clearly identify a chair despite there being no concrete laws of “chair-ness,” we can identify a single paradigm by similarities in research methods and the language they use.
There are several reasons to believe that paradigms guide work even when no fixed rules are in place. First, there’s the difficulty of finding concrete rules in some communities, which we’ve already discussed. Second, scientists aren’t given lists of rules and theories to memorize—they pick them up naturally through study and learning how to apply them in practice. In other words, even when such rules exist, they aren’t formalized.
Third, as long as there is a universally-accepted paradigm—or close enough to one—normal science doesn’t need established rules. The scientists all agree on the problems and solutions that their field has, so they’ll naturally work in the same or similar ways, as was demonstrated in the analogy to Wittgenstein's question. However, the other side of this coin is that normal science does need concrete rules when a paradigm is unstable; in other words, just before and during a scientific revolution.
Finally, normal science—even within the same field—isn’t completely uniform. There could be any number of specialties and subspecialties that, to an outsider, seem to fall under the umbrella of one particular field.
Sometimes a scientific revolution can actually be very small, affecting only a particular specialty within a field. Therefore, there can’t be scientific rules that affect the entire field, when paradigms can change for subsets of that field. The next section will go into detail about subsets within fields and how paradigms can affect them differently.
Normal science doesn’t look for anomalies or unexplained phenomena, but it often finds them anyway; in fact, science seems to be uniquely good at finding unexpected things. In order to mesh those apparently contradictory facts, we have to assume that working within a paradigm is an effective way to change that paradigm.
Working by a particular set of rules, then finding facts or theories that don’t fit those rules, requires that the rules then be changed. Far from unusual or contradictory, this may actually be an inevitable result of digging ever-deeper into a particular paradigm, trying to expand and clarify each tiny facet of it.
There are two key ways a paradigm can be changed: discovering new facts and inventing new theories. However, the two processes are closely related, and both follow the same pattern.
New discoveries start with anomalies. When an observation doesn’t match the expectations set by the current paradigm, there is an anomaly; in that sense, a paradigm is actually needed for new discoveries. You must know what to expect before you can recognize that something’s wrong.
However, anomalies don’t always lead to new discoveries—as we’ve said before, sometimes they’re ignored or brushed aside as mistakes, edge cases, or something too inconvenient to deal with. It usually takes some time for scientists to recognize and acknowledge that the anomaly is important, and must be addressed. Sometimes that never happens at all.
After finally being acknowledged, the anomaly is investigated. Assuming that the anomaly is natural, and not the result of human error, a change in paradigm is needed to account for it.
This process ends when the paradigm has changed so that the anomalous event is now expected. This is more difficult than it sounds. At this point scientists aren’t just adding new knowledge, they’re rejecting old knowledge, and many of their colleagues will resist the change. That resistance to change is part of why discovery and paradigm changes are such drawn-out processes.
A psychological experiment helped to demonstrate this effect. In the experiment, subjects were shown a series of playing cards. However, some of the cards were altered, such as a black four of hearts. Subjects identified the familiar cards easily, but hesitated with the ones that were outside their normal experience of playing cards—that is to say, the anomalies.
Some subjects even became upset by the anomalous cards; they started doubting what they were seeing or how they were interpreting it rather than realizing that it was simply an altered card. Others would incorrectly identify the black four of hearts as either a normal four of hearts, or the four of spades. They immediately and without thinking tried to fit it into their current paradigm of playing cards.
However, when shown the cards a second time, most subjects were able to recognize what had happened and correctly identify all of the cards, including the anomalies. This is a parallel to scientists recognizing that they had not made a mistake, and that their paradigm needs to be adjusted.
Tracking this process in science can be difficult because, in order to really claim a new discovery, a scientist has to find, understand, and announce it. In other words, the scientist must present the fact and a corresponding theory, not only the fact itself.
This can take a great deal of time, which makes it hard to pinpoint exactly when a discovery happened, as discussed in Chapter 1. Therefore, the word discovery may actually be misleading. It implies a single event that happened at a certain time. To understand how closely intertwined fact and theory are, and how difficult it can be to pinpoint when a “discovery” happened, consider oxygen gas: At least three different scientists could reasonably take credit for discovering it.
C.W. Scheele may have been the first person ever to make a pure oxygen sample. However, he didn’t publish his work until after it had been discovered and announced in many other places, so he’s not usually credited for this breakthrough.
Joseph Priestley collected a sample of oxygen-enriched air, but not pure oxygen. He tried to claim that he had discovered the element, but if bottling impure air is enough to take the credit, then he was hardly the first to do so. Plus, it was several years before Priestley even understood what he had bottled; at first he thought it was nitrous oxide, a type of gas that scientists already knew about.
Antoine Lavoisier was the first to collect and correctly identify a sample of pure oxygen, and so he is the one usually credited with discovering it. Some scientists even argue against this claim, since Lavoisier didn’t fully understand how oxygen gas was made. However, the paradigms that led Lavoisier to incorrectly identify how oxygen was produced weren’t rejected until almost a century later, and by that point oxygen itself was firmly entrenched in chemistry.
So you see how it can be difficult to pinpoint what should be a straightforward answer. In fact, the deeper you look into the history of oxygen gas, the less clear it becomes who deserves the credit for discovering it.
If recognizing anomalies is a key part of new discoveries, it shouldn’t be surprising that it’s also part of inventing theories—remember that discovering new facts and developing new theories are closely related processes. That whole process is key to scientific revolutions.
This isn’t to say that normal science never solves anomalies—in fact, that’s one of its major functions. This is part of the reason paradigms are so hard to change: Scientists often assure themselves that there is a solution within the paradigm, it just hasn’t been found yet. Of course, this isn’t always the case.
On the very rare occasions that a field answers every question posed, such as the paradigm that has light traveling in the form of rays, that field stops being science and becomes an engineering tool. However, when those anomalies last long enough or reject some key part of the paradigm, it could be said that the fields they pertain to are in crisis.
In other words, an anomaly must seem like something more than a regular puzzle of normal science. Scientists in those fields have a growing awareness that something is fundamentally wrong, that nature is in direct conflict with their understanding of it. That growing sense of crisis is what prompts the move into extraordinary science, and eventually revolution.
The anomaly becomes more generally recognized in the field, and the most prominent members of that field turn their attention to it. In attempting to resolve it, scientists will push the rules of their current paradigm as far as they can go in an effort to see exactly where they break down. Once the breaking point is found, they will start speculating about new theories that can account for the anomaly. Then, if one of those theories holds up under scrutiny, it may form the basis of a revolution and a new paradigm.
A good example of this is Ptolemaic astronomy, which was the paradigm before Copernicus’s time. While it predicted most planetary movements correctly, small details such as when equinoxes would occur never quite lined up with reality.
Scientists kept making small adjustments over the course of centuries, until the system became so patchwork and so convoluted that it couldn’t possibly be right. This is a clear sign of a scientific system in crisis. Finally, Copernicus said that the Ptolemaic system had become a monster, and he rejected it entirely to replace it with his own model.
In the Copernican revolution and almost all others, new theories come about when normal science repeatedly fails to explain anomalies, often over the course of many years. However, while the failure takes a long time to reach the point that it can’t be ignored, the revolution often occurs relatively soon after that.
In many cases, there are only 10-20 years between the crisis point and the establishment of a new paradigm, replacing the old one. While that sounds like a long time, consider that in the previous example it took hundreds of years for scientists to admit the current system was failing. This short timeframe suggests that new theories come about as a direct result of scientific crises.
This may be because many of the key anomalies in the Ptolemaic system were considered to be solved already, or so close that it hardly mattered, needing only some slight adjustments to the paradigm. Therefore, when it turned out that those problems couldn’t be solved, scientists felt the failure very keenly and began looking elsewhere for answers.
One other common element of new theories is that they were frequently developed—at least to some extent—during a time when there was no crisis in that field; but they weren’t seriously considered until much later.
For example, the Greek astronomer Aristarchus first suggested a heliocentric model of the solar system in the 3rd century BC. However, since the then-current geocentric model explained the universe well enough, there was no need for his new idea and it was brushed aside. That would remain the case until many centuries later, when improved technology allowed astronomers to see things that the geocentric model couldn’t explain.
The only time a paradigm is rejected is when another one comes to take its place. This implies that the process of rejecting a paradigm doesn’t just involve comparing it with nature, but comparing it with another paradigm.
All scientific crises seem to resolve in one of three ways.
Paradigms must be based on theories. Scientific history, especially from the time before a given field has an established paradigm, shows that many competing theories can explain the same sets of data. Even so, normal science makes no effort to develop new theories, because there is no need for them as long as the current paradigm holds. That’s exactly why crises are so important to scientific advancement: They are the stimulus for scientists to break the rules and start looking outside of their paradigm for new ideas.
However, even in the face of crises, scientists will never start by rejecting the current paradigm. They don’t treat anomalies as disproving the current paradigm, even if the paradigm has failed to explain them for many years. Instead, scientists will try to explain away the anomalies with arguments and off-the-cuff modifications of the paradigm, like the Ptolemaic scientists did.
Remember that people—and therefore paradigms—resist change. Formal education trains students on the current paradigms, and often fixes them in it. Students are taught how the paradigm is used, and the problems it has solved, and they take that as proof that the paradigm is true. Alternative explanations for those observations are rarely, if ever, part of a scientific education. While science is seen as enlightened and logical, in many ways it is steeped in tradition and habit.
The word “revolution” is usually used in a historical context, not a scientific one. However, there are many parallels that make the term appropriate here. First and perhaps most important of all, revolutions start with a growing feeling that the current systems aren’t working. This is often because they’re failing to solve problems they helped to create.
The revolutions may only be significant to a small part of the population. Like the Balkan revolutions in the 20th century, outsiders might see them as normal and expected parts of progress. Similarly, scientists whose paradigms aren’t affected by a scientific revolution may not see it as a revolution at all, but just as a new piece of information.
Revolutions, whether political or scientific, try to overthrow the current establishment and put a new one in its place. The community is inevitably divided into those defending the status quo and those fighting against it.
Once that division happens, there is no political solution. The people can’t even agree on what type of political (or in this case, scientific) framework should be in place to make that solution happen. It’s impossible to resolve the conflict by the usual methods, because those are set by the current paradigm, and it’s the paradigm itself that’s under attack. The only thing that’s clear to all sides is that the competing paradigms can’t coexist. Finally, the opposing parties use whatever means they can to get popular opinion on their own side, including force if needed.
Revolutions are key parts of both history and science, but they only work because they exist outside of the normal, established structure.
It seems like, in theory, science should be able to progress without revolutions. Many people believe that—if not for human error—science would add perfectly onto itself, with no need to reject and overthrow old paradigms. This is largely based on the idea that what we call “knowledge” is a human construct put onto observed data, which would imply that any errors in our knowledge must stem from our interpretations of the facts.
However, even if humans were perfectly logical and never made mistakes, it’s unlikely that normal science alone would be sufficient. As previously mentioned, many different and incompatible theories can explain the same observations. Until some new information makes one seem more plausible than the others, even a perfect scientist would have no way of knowing which theory is the correct one.
Also, scientists will tend to look for problems that should be solvable with current knowledge and techniques, and novel discoveries only happen when the knowledge proves to be wrong. Therefore, if humans never made mistakes, major discoveries would never happen. The only means of gathering knowledge at that point would be slow accumulation by normal science, without the sudden leaps that revolutionary science allows.
(Shortform note: Logical Positivism is a philosophy which states that the only problems worth considering are those which can be solved through direct observation or logical reasoning. One side effect of this is the belief that any paradigm that has been used effectively must be correct—the paradigm has been observed to work, therefore it’s correct.
This was the dominant philosophy in Kuhn’s day, though not so much in modern times. In fact, Structure itself played a large role in overthrowing it.)
The theory of scientific revolutions is not a popular one. Many people believe that new paradigms must be either offshoots of, or additions to, current paradigms. They would argue that an entire paradigm doesn’t get overthrown, only those parts of it that are provably wrong; and, furthermore, that those parts are based in human error, not a fundamental flaw in the paradigm. The paradigm as a whole must work, because it has been observed to work.
We should reject this logical positivist view of science. Following that logic, paradigms should only work for the specific phenomena they’ve been tested for, and only under those same laboratory conditions. If only things that have been observed can be considered true, then extrapolation—which is crucial to science—would be impossible.
Consider this: If there were not some critical, unresolvable difference between an old paradigm and a new one, there would never be scientific crises. In that case, any anomalies would be explained by the current paradigm and a new one would never be needed. Therefore, the very fact that such crises exist shows that the logical positivist view is flawed.
New paradigms are fundamentally incompatible with preceding ones, and establishing a new paradigm requires that the old one be rejected, right down to its fundamental beliefs and its scientific rules. When that happens, the entire field may need to be redefined. Old problems are shunted to different fields, or brushed aside as unscientific. New problems, or problems that weren’t considered worth the field’s time before, may become the central focus of that field.
In short, the normal sciences practiced under different paradigms are not just incompatible, they’re completely incomparable. The new paradigm bears no relation to the old one, yet both have been observed to work, so logical positivism can’t be correct.
Not just the sciences, but almost every field has certain established beliefs and paradigms. This means that just about anything could be subject to a crisis.
Are there currently competing paradigms in your field of interest? The field doesn’t have to be in crisis for this to be relevant, just consider whether there are other schools of thought fighting to be recognized.
What might that field look like if one of those other paradigms were dominant?
Could you foresee the current paradigm being overthrown by one of the others? If yes, how? If no, why not?
When paradigms change, it could be said that the entire world changes with them. Scientists ask different questions, follow different rules, and even use different terms (or use the same terms differently) after a paradigm change. Perhaps even more importantly, they see new information in studies they’ve done before.
This process resembles a gestalt shift. In psychology, a gestalt shift is when your perception of something suddenly changes—like a picture that you can see as either a rabbit or a duck (but never both at once).
However, there are important differences between a simple gestalt shift and a scientific revolution. Gestalt shifts can go either way—from duck to rabbit or rabbit to duck, and back again. The subject can even learn to make the switch at will. On the other hand, scientific revolutions generally go one way, and they are irreversible. A scientist who now views the world through a Copernican paradigm can never go back to Ptolemy. A period when scientific perceptions can switch back and forth is, by definition, a crisis.
In a gestalt shift, such as with the picture which could be seen as a duck or a rabbit, either interpretation is correct—or at least, one isn’t more correct than the other. However, as we’ve said before, accepting a new paradigm means rejecting the old one, as if seeing a rabbit meant that you could never see the duck again.
Finally, in psychological experiments there are external forces helping the gestalt shift—for example, in the card experiment described earlier, there was someone there to tell the subject that he had been looking at a black five of hearts. This allowed the subject to begin shifting his perspective of what a playing card could be. However, in science, there is no higher source of authority or information. Scientists have only their instruments and observations to rely on.
Scientific crises aren’t resolved by study and interpretation—that would be normal science, which by definition can’t fix a crisis. They’re resolved by a sudden understanding or change in perception resembling the gestalt shift. Scientists have sometimes described that shift like a flash of inspiration, or scales falling away from their eyes, rather than a deliberate understanding of the material.
Many readers will say that what changes with a paradigm is only how scientists interpret their observations, not what they see. In other words, the observations themselves are unchangeably set by nature, and all that changes is the observers’ understanding. However, like normal science, interpretation needs a paradigm in order to exist. Furthermore, interpretation of a paradigm can’t fix that paradigm. It can only explain and expand upon it. It’s impossible for a paradigm to disprove itself; all it can do is find anomalies that eventually lead to crisis.
For example, consider a stone swinging on the end of a string. Aristotle, whose paradigm said that objects were naturally drawn to their places of rest, saw a stone attempting to fall and being held back by the string. Galileo, however, saw a pendulum, with all the complex physics that implied. By this theory of different interpretations, Aristotle and Galileo both saw pendulums, they just interpreted them differently.
This common view isn’t entirely wrong, and it’s part of a long-standing paradigm about the nature of knowledge and understanding. The current paradigm says that there is an objective reality, which we can learn about through unbiased observations and share by using neutral language to describe it. Furthermore, this paradigm states that objective reality is what scientists are studying, and any discrepancies come from different interpretations of it.
However, despite its past effectiveness, modern research is finding anomalies that suggest the current paradigm is flawed. There isn’t a strong alternative yet, so the current paradigm can’t be rejected right now, but it seems clear that different interpretations of data aren’t enough to explain what happens after a change in paradigm.
For one thing, the data scientists collect is not the same. For example, a falling stone is not the same thing as a pendulum. This isn’t just a reinterpretation of the same thing, it’s a fundamental difference with hugely different implications.
An Aristotelian scientist might observe the stone’s weight, the distance it moves from its starting point, and how long it takes to cover that distance. A Galilean scientist, on the other hand, might observe the pendulum’s starting angle and the radius of its swing. The community transitioning from the falling-stone paradigm to the pendulum paradigm is not just reinterpreting data; that would be impossible, since the data itself is different.
Given that the observations, data, and interpretations are all different, it seems fair to say that, after a paradigm shift, science is now happening in a different and incommensurable world.
The problem we now run into is that most of the previous examples of scientific revolutions aren’t usually seen as such—they’re viewed as additions to knowledge, as would be expected from normal science. Any additional examples would run into the same issue.
Part of the problem is that both scientists and laypeople get their ideas of science from authorities that methodically and intentionally hide the existence of revolutions. These authorities are usually textbooks, along with popular works and philosophical treatises based on those same paradigms. All of these are devoted to established paradigms, along with the data, theories, and questions it contains.
These works explore the results of revolutions, rather than the revolutions themselves, which gives the impression that those revolutions never happened. Furthermore, they tend to show revolutions as logical progressions of knowledge, rather than the overthrow of established paradigms.
This is done for practical reasons, not to deliberately cover up the facts. There’s no sense in exploring the entire history of science in every textbook, so for teaching purposes they only cover the parts that are relevant to the current paradigm. Furthermore, going in depth about outdated paradigms would give them more weight and credibility in present-day scientific communities. This would only lead to new scientists rehashing the same issues again and again.
However, as a result, the books end up making all science look like normal science—linear progress from ignorance to knowledge through the accumulation of facts. They can also give the false impression that scientists have always been working to solve the specific questions posed by our modern paradigms.
For example, textbooks often credit the 17th-century chemist Robert Boyle for inventing the concept of a chemical element in his book The Sceptical Chymist. However, the definition of “element” that he gave was little more than a paraphrase of a concept that had existed at least since Aristotle’s time.
Boyle’s writing was significant because he was at the head of a scientific revolution that changed what the word “element” referred to—a key distinction from changing the definition itself. Aristotle favored the four so-called “traditional elements:” earth, water, air, and fire, while Boyle and his contemporaries had the word refer to elements as we understand them today.
The Sceptical Chymist was a major part of a revolution that transformed the field of chemistry. However, by glossing over everything that came before him, modern textbooks make it seem as though Boyle was simply adding to an established pile of knowledge rather than overturning the popular ideas of the day.
Textbook methods of teaching have so shaped our view of science and discovery that we often can’t see the countless revolutions that brought science to its current state.
While Structure is, of course, focused on scientific revolutions, many discoveries and events across all fields have been described as revolutionary.
Think about a field that’s particularly interesting to you, whether scientific or otherwise. What’s something that has revolutionized that field?
What might that field look like today if that revolution hadn’t happened? What would be better, and what would be worse?
Any new paradigm starts with one or, at most, a few people. These people usually have two things in common. First, they have focused their work on the anomalies that led to the current crisis. Second, they are young, or at least new to the field, and so are less set in the current methods and rules than their colleagues.
The challenge then becomes convincing the rest of the scientific community to adopt their new paradigm. Therefore, the two (or more) paradigms must now be compared both against nature and against each other.
It’s impossible to test a paradigm in every way, under every circumstance. Therefore, many scientists would say that convincing the community to adopt a new paradigm isn’t about proving beyond doubt that it’s correct; instead it’s about showing that, given the available evidence, it is more likely to be correct than the other options.
This is called positive verification, and it offers a couple of testing methods. One way is to compare the paradigm or paradigms in question against all conceivable ones that explain the same data. Another is to come up with tests that the paradigm might reasonably be expected to pass.
However, while it has its uses, this method is fundamentally flawed. As previously said, meaningful testing and observation can’t occur without a paradigm already in place. Given that a paradigm is required, there’s no way to test the new ones in all possible ways—only those ways that the paradigm itself allows.
Positive verification could be compared to natural selection: The most fit of the available options at the time of testing will survive and thrive, at least for a while. It’s meaningless to ask whether it’s the best possible option, because there is no practical way to test that.
The philosopher Karl Popper suggested an alternative, which could be called negative verification, or falsification. Under negative verification, tests would be conducted with the express purpose of failing, thereby disproving a theory.
(Shortform note: To give a somewhat outlandish example, if you showed that you could drop a stone and have it fall up instead of down, you would have disproved the theory of gravity. So far, every test conducted has failed to disprove that theory.)
It’s worth noting that negative verification serves much the same purpose as anomalies: Showing fatal flaws in current ideas, thereby paving the way for new ones. However, anomalies and falsifications are not the same thing. In fact, true negative verification might not even exist.
Remember that very, very few paradigms can answer every question they’re presented with. That is the reason why normal science exists. Therefore, if a single failed test were enough to reject a paradigm, they would all be rejected almost instantly.
There would have to be some point at which the failure is great enough that the paradigm can be comfortably rejected. In other words, scientists would have to show that it’s less likely to be correct than other paradigms, and in trying to establish the exact degree of unlikeliness, they’d run into all the same problems as positive verification.
The underlying problem is that both of these popular, yet opposite, ideas try to combine two separate processes. Positive and negative verification are, essentially, attacking the same problem from opposite sides: One seeks to show that a given idea is more likely to be true than others, the other that a given idea is less likely to be true than others. However, both processes are needed to establish which theory the scientific community could be most confident in adopting.
There are significant challenges in converting a scientific community to a new paradigm.
Remember that scientists working under different paradigms are, in many ways, working in different worlds. Because they aren’t always answering the same questions or using the same methods, competing schools of thought often have trouble communicating with each other. Because they have different standards, they can’t agree on which paradigm meets those standards better. Finally, because the paradigms can’t be compared to each other, the switch between them can’t be forced by reasoning or evidence.
Instead, like the gestalt switch, the change happens all at once—if it ever happens. Sometimes current scientists—especially older ones—can’t or won’t be convinced at all, and it takes a full generation for the new paradigm to establish itself.
However, this is not necessarily because of stubbornness or willful ignorance, but because the scientists are confident in the paradigms they’ve been using their whole careers. Oftentimes they believe that the current paradigm, if given the chance, will solve its anomalies and prove that a new paradigm isn’t needed. That same confidence in the paradigm is what makes normal science possible.
However, many scientific communities have switched from one paradigm to another.
One of the most effective ways to convert scientists, naturally, is by showing that the new paradigm solves the problems that caused the current crisis. This is actually quite rare; despite it being the basis of a revolution, a new paradigm often won’t solve the crisis much better than the old one. This is in large part because the new paradigm hasn’t had the time or normal scientific effort to develop. In its early versions the paradigm is still crude and imprecise, and often doesn’t provide the elegant solutions that scientists were hoping for.
Therefore, the major question becomes not which paradigm solves those particular problems better, but which one should guide normal science going forward. In that case, it’s not so much about which paradigm is currently “better,” but which one shows more promise for future achievements.
If resolving the crisis isn’t enough—or if the new paradigm doesn’t solve those problems better than the old one—other areas of the field will often be drawn into the argument. If the new paradigm can predict phenomena that the old one couldn't, even if those are not the problems that caused the crisis, its position will be greatly strengthened.
For example, Copernicus predicted many things about other planets and the universe that improved telescopes showed to be true decades after his death. The fact that his paradigm correctly predicted such things made it much stronger and more popular, especially with non-astronomers.
On top of the purely scientific arguments, people can sometimes be converted to a new paradigm just by aesthetics—if the new paradigm is seen as simpler or more elegant than the old one. Other “unscientific” reasons for adopting a new paradigm may include the reputation, past successes, or even nationality of the one proposing it. In these cases, people adopt the new paradigm for purely human reasons rather than scientific ones.
No single one of these things will convert every scientist in a community. Instead, there is a gradual but ever-increasing shift in belief among the people in it. As the new paradigm gains support, more scientists will devote themselves to exploring it, thereby making it stronger. This trend continues until there are only a few holdouts left.
It’s important to acknowledge that these holdouts aren’t necessarily wrong. They hold to an older paradigm, one that has or had great scientific value. At most one could say that, after an entire profession has moved to a new paradigm, those who still resist are no longer part of that profession. Sooner or later these last opponents die out or retire, and leave the new paradigm largely uncontested.
Most of the preceding arguments have been about progress: the nature of it, why it’s expected, and how it happens. However, “progress” is often a label put on events by their observers. The reason normal science seems to progress is clear: Scientists working within a paradigm have a clear set of problems to solve, and the means by which to solve them.
Normal science benefits from being fairly well insulated from the laypeople. Scientists generally only show their work to other scientists, meaning they can take certain beliefs and rules for granted. Furthermore, the scientists can choose problems they are reasonably confident they’ll be able to solve, without worrying about what needs to be solved, as an engineer or doctor might have to. All of this makes progress both faster and clearer than in other professions.
Some may wonder why unscientific fields such as art and philosophy don’t seem to progress like people assume science does. However, this question may be fundamentally flawed, and the progress of other fields may be closer to science than people believe.
As an interesting example, painting was seen as a progress-based field for centuries. This view was based on the assumption that the goal of painting is to depict the subject as accurately as possible. Breakthroughs in painting technique allowed artists to get closer and closer to realism. However, when some painters began to shun realism and explore other styles, the current divide between art and science began to form.
Still, creative work such as art progresses in other ways. When an artist successfully creates a new piece, it is added to the collection of all available work just as a new piece of scientific data would be. Furthermore, while “art” as a whole may not be seen to progress, individual schools within it certainly do. Philosophy and other non-scientific fields follow the same pattern.
In some ways the question is purely semantic: Does something progress because it’s a science, or is it science because it progresses? There is no firm answer to this, and it’s not even clear how important a question it really is.
Instead, it seems likely that the question stems from some insecurity about one’s own field, and why it doesn’t seem to progress in the same way that the established sciences do. For example, economists don’t seem to spend much time debating whether economics is a science. That’s probably not because they all agree on what science is, but because they agree on what economics is, and are confident in it.
In an apparent contradiction to the previous section, revolutionary science is also seen as progress, even though it doesn’t use accepted methods to solve accepted problems. In large part, this may be because revolutions must end in victory for one side or the other, and it’s unlikely that the winning side will see its victory as anything but progress.
That school’s paradigm is then written into textbooks and other educational materials, and it becomes what the next generation of scientists believes as well. Therefore, a revolution being seen as progress after the fact is practically a given. (Shortform note: As they say, history is written by the victors.)
This makes science sound like some kind of dystopian system wherein all that matters is who’s able to win the struggle. However, while achieving victory through force and rewriting history would make for a revolution, it would not make for a scientific revolution. The key difference here is that the outcome of the revolution is chosen by the members of the scientific community it affects—through professional study, consideration, and argument, not through brute force.
Those scientific communities share a couple of key features, which make them uniquely suited to make such decisions. First of all, scientists study natural phenomena—concrete problems. Paradigms only go into crisis when nature itself seems to disagree with them. Second, scientists’ work must satisfy the scientific community—but not outsiders. Scientists’ work is not (usually) muddied by having to compromise with or appeal to non-scientists, so there is no outside pressure steering the results toward one paradigm or another. That helps to ensure that the outcome of a revolution is progress, rather than just what sounded good to the unscientific majority.
However, to fully resolve the apparent contradiction between revolutions and “progress,” we may need to give up the idea that science has a clear endpoint.
Remember that science shares a lot of characteristics with evolution. Like creatures evolving to fit their environments, science is constantly changing to best fit what we currently know about the world. As in evolution, paradigms fight for limited resources (scientists’ acceptance) until one prevails. The cycle of revolutions and normal science throughout history has resulted in excellently adapted fields of modern science, and further development in those fields comes through narrow specializations, like animals evolving to fit into an ecological niche.
One of the greatest challenges Darwin faced in getting his ideas accepted was not in convincing people that evolution happened, but that it was not working toward some distant goal—that is, evolution isn’t “progress” as we think of it. Science itself may need to undergo a similar shift, with scientists accepting that there isn’t a single, universal truth that some paradigm will eventually explain perfectly.
Naturally, this raises a couple of important questions: Does science need a goal? Is it even helpful to imagine that there’s some absolute, universally true understanding of nature? Or is it better to let scientists do their work, and let science evolve and adapt as it must?
In conclusion, though it seems contradictory, science often progresses by overthrowing past achievements. These revolutions are not only inevitable, but necessary. They happen over and over again throughout history, and they generally follow the same pattern. Finally, there is no paradox between revolution and progress as long as you give up the idea that science has some final, fixed goal.
(Shortform note: This essay is included as a preface to Structure in the 2012 edition, but its author suggests saving it until after you’ve read Kuhn’s book.)
This introductory essay by Ian Hacking discusses how and why The Structure of Scientific Revolutions was written, its impact on the scientific community, and whether it’s still relevant today (Shortform note: The first edition of Structure was published in 1962).
The sciences have changed a lot since Structure was first published. In 1962, physics was in the spotlight. It was the middle of the Cold War, so everyone had nuclear weapons on their minds. In fact, Kuhn himself was trained as a physicist.
Today, however, biotechnology is the hot field of science. On top of that, modern technology like computer simulations and the internet has changed how science is practiced and how ideas are shared.
Another major change since Structure was written has to do with fundamental physics itself. In 1962, physics had two major competing models of the universe: steady state and the Big Bang theory. By 2012, only the Big Bang theory was commonly accepted.
Finally, Kuhn believed that the most important scientific work was theoretical, not experimental. In fact, the modern emphasis on experimental science is as recent as the 1980s. You’ll have to decide for yourself how relevant Kuhn’s ideas about physics are to modern science.
(Shortform note: Steady state physics said that while the universe was constantly expanding—which had been observed—it was constantly recycling old matter to maintain its density. The Big Bang theory says that this is not happening, and the universe is becoming less dense as it expands. To learn more about the Big Bang Theory, read our summary of Stephen Hawking's A Brief History of Time.)
To show how Kuhn is both right and wrong about normal science, consider the Higgs particle. The Higgs particle is a theoretical particle that would fill a gap in the current physics paradigm. Normal science is trying to find this particle.
In looking for the Higgs particle, scientists aren’t trying to discover anything new; they’re trying to find evidence to support the current paradigm. This is what Kuhn got right. However, it’s possible—even expected—that discovering the Higgs particle will open up entirely new branches of physics. This is what Kuhn’s ideas about normal science gloss over. (Shortform note: The Higgs particle—now called the Higgs Boson—was first observed in 2012, the same year this essay was published.)
There are a number of supplemental readings you could do to better understand Structure and the author’s life: