Evolution has long been an idea in search of a principle. A concept as old as science itself—Aristotle, for example, suggested that nature was ruled by a desire to move from lower to higher forms—“evolution” has been invoked through the millennia to describe change over time. Nowadays, this single word encapsulates Darwin’s work about biological life and the subsequent research that has refined and elaborated his insights. It is also employed much more loosely to describe the development of just about everything. Library shelves sag from the weight of tomes describing the “evolution” of science, nations, written languages, and social values; of religion, war, technology, art, cooking, and even the beautiful game of soccer.
This history is the story of good hunches. Every discernible thing, every design in nature, does evolve. It is dynamic, not static. What has been missing is the single principle of physics that unites these phenomena and allows us to predict how they should evolve in the future. Using the constructal law, we recognize that not only biological species but also technology and language, religion, education, and all the rest are flow systems that configure and reconfigure themselves so that the bodies that possess these designs (we, the cultured) move more easily on the globe. It shows us that evolution is far broader than Darwinians have believed and far more specific and powerful than other thinkers have imagined.
“Evolution” means design modifications over time. How these changes are happening are mechanisms, which should not be confused with the principle, the constructal law. In the evolution of biological design, the mechanism is mutations, biological selection, and survival. In geophysical design, the mechanism is soil erosion, rock dynamics, water-vegetation interaction, and wind drag. In sports evolution, the mechanism is training, selection, rewards, and the changes in the rules of sports competitions. In technology evolution, the mechanism is innovation, technology transfer, copying, theft, and education.
What flows through a design that evolves is not nearly as special in physics as how the flow system acquires and improves its configuration in time. The how is the physics principle—the constructal law. The what are the currents and the mechanisms, and they are as diverse as the flow systems themselves. The what are many and the how is one. Hierarchy more simple than this does not exist.
The constructal law advances our understanding of evolution by proclaiming that design should emerge across nature to facilitate flow. It also holds that these configurations should morph with a clear direction in time: to provide better and better flow access. Evolution, then, is measurable in terms of how much easier and farther things move on Earth.
As it predicts why design should emerge and evolve, the constructal law reveals the broad patterns that abound in nature. Despite their great diversity, flow systems faced with similar challenges and constraints tend to acquire similar designs. Inanimate and animate designs evolve as if they are “intelligent,” because they appear to come up with the same answer to the problem of how to flow more easily. They also generate the same designs that we come up with to facilitate flow; that is why their designs are predictable.
Pattern generation is evident in the predictable design that emerges among the various components of a flow system. The vascular, hierarchical designs we find throughout nature strike a balance between the speeds of their currents (each of which selects the mode of flow, slow and fast, that works best for them) by generating multiscale channels. It is also apparent in the evolving design of even larger flow distribution networks, including the hierarchical distribution of tree sizes in the forests and in the emergence of human settlements—a few large cities, with many small communities on the map.
This overriding natural phenomenon has been noted by researchers. But they have described it empirically, as power-law correlations, hierarchies, allometric scaling, and Zipf distributions of frequency versus rank. Researchers have observed but could not predict. They have known the what but not the how.
This how, the constructal law, sparks especially surprising insights when applied to social dynamics. The prevailing view holds that the institutions built by humanity are subject to the desires of people, not the laws of nature. This view is wrong, not just philosophically and not just as a practical matter. It is wrong as physics, because social designs emerge and evolve as a result of the selfish urges of many individuals who do not consult one another. Each has the tendency to flow more easily; all find it is easier to flow together, with design. This means that social designs, like other flow designs, occur naturally.
We have already seen how the constructal law predicts the evolving designs of engineered entities such as the wheel, roads, airports, and human settlements. In this chapter we will focus on two areas that seem less concrete: academia and human relations. We will see why they, like every other social system, are hierarchical designs that evolve to cover an area with current, and manifest the same patterns that emerge in other natural phenomena.
Each new release of the rankings of America’s best universities by U.S. News & World Report is the talk of the campus. Some administrators discount the importance of rankings, while the rest declare that the university is finally (now) poised to execute “the great leap forward.” This reaction has not changed in years, because the rankings have not changed in any meaningful way in years.
The usual suspects—Yale, Harvard, Princeton, MIT, Stanford, and Duke—are annually reconfirmed as the leading national universities; Williams and Amherst remain among the highest-ranked liberal arts colleges. The graduate school rankings—Yale and Harvard tops in law, Stanford in business—continued this dog-bites-man story.
There is often a little more movement down the lists—a few schools rise a notch or two; others fall slightly. But for all the concern about administrators gaming the system, the absence of change stands out. To quote the Talking Heads song, with academic rankings, it’s the “same as it ever was.”
Why do the rankings seem carved in stone? The answer cannot be found by studying the metrics used by U.S. News editors but by applying the constructal law. As we have seen, a pattern that persists in time by resisting big forces to change speaks of the much bigger forces of nature. It speaks of physics and the design of nature, of the evolution of shape and structure that facilitates our movement on the globe. Yes, you read it correctly—science and education facilitate our movement. The urge to move more easily is what drives the tendency to acquire knowledge, not the other way around. Without science and education we would still move but not much, because we would be hiding in caves. Because knowledge and information are currents that enhance our own movement, they acquire evolving design in accordance with the constructal law.
It should be no surprise that the architecture we find in river basins and forests (few large, many small) is the same one we see in our system of higher education—a tiny number of top-ranked universities, a few more second-tier institutions, and many lower-ranked schools. What may seem remarkable is that this hierarchy is as rigid as the one we find in those other “natural” systems.
We answer this question with pure theory, by predicting that education is an evolving global flow system with design that is governed by the constructal law. The rankings are an expression of this phenomenon. We begin our analysis by putting aside the Darwinian interpretation of the rankings, which casts individual schools as competitors in an epic struggle for survival/supremacy. Instead, all the colleges and universities are components of the single larger flow system that covers the entire globe. Just as the Mississippi River is not competing with its tributaries but working hand in glove with them to move water, the schools of various rank—both high and low—together form the river basin of education that spreads knowledge across the global landscape.
Next, we identify what current is flowing through the design. The answer is ideas, and the pedigree that the density of ideas attaches to those touched by the flow. A scholar and a university become known because of the ideas they generate. Good ideas travel and persist (to “persist” means to keep on morphing and traveling from those who know to those who need to know). The good ideas are the ideas that are adopted by others worldwide. Most ideas are replaced and forgotten; like the vast majority of published research papers, they are not even noticed.
If ideas are the current flowing through the academic system, what is the measurable characteristic that makes one school more highly ranked, or “better,” than another? That is, if we rank cities by the size of their populations (Paris is a bigger channel than Lyon), what do the highly ranked schools possess more of than their lower-ranked brethren? It is certainly not their physical sizes. The top schools do not have the most students. It is, instead, the visibility, the fame, the usefulness of the ideas they generate. In education, fame, or visibility, is synonymous with greater access through the vascular structure of societal flows. Students flock to high-ranking schools because they know these schools can help them enter the main channels of society. Education flows in one direction: from those who have it to those who seek it. When both ends of each such river basin have it and know it, the flow stops. What is not news does not travel.
Seen constructally, a university is not the piece of land in a particular spot. It is the professors, their disciples, and the disciples’ disciples. It is the ideas that flow through these human links and into the books of our evolving science and culture with which we walk on Earth. Because hierarchy occurs at every level of the design, each university is the central node, heart, and aorta, nourishing and sustaining its students and others with the ideas it generates.
It is also a channel on the entire world map, a component of the highly complex global vascular flow network of knowledge. In time, this global vasculature evolves like a river basin during the rainy season: All the streams swell, but their hierarchy remains the same.
The historical institutions—from the Universities of Bologna and Padova to the Sorbonne, Oxford, Cambridge, Coimbra, and Harvard—have earned their rankings in this global system because of the fame of the ideas they have and continue to generate. This entrenched hierarchical design persists because it facilitates the flow of ideas across the world.
With these ideas in mind, the constructal law predicts that all the universities should generate a hierarchical design to facilitate this flow. That is, they should produce the distribution of design features (in this case, of universities) that we find in the design of river basins, forests, and other natural phenomena. The ranking of these schools should be based on the fame, the usefulness, of the ideas they generate.
To test this prediction, I took the only unbiased measure of academic visibility available in my field of engineering—the number of citations of an author’s creative output compiled by the Web of Science. It is an unbiased sample because the researchers who cite an author’s work do so because they read it, valued it, and used it. These numerous voters are not recommended by anybody. A magazine does not handpick them. They do not belong to a club. The best part is that one can see who they are and why they cited the author.
For each U.S. graduate engineering school ranked in the top 50 by U.S. News, I counted the number of names that appear on the most-cited list. I plotted this number on the ordinate in Figure 47. The abscissa indicates the ranking in U.S. News. This figure provides a bird’s-eye view of where university rankings come from. The highly ranked engineering schools are homes to researchers who are highly visible. The lowly ranked schools are not. The left end of the scale is dominated by schools with ordinates in the 5 to 10 range. The right end is dominated by schools with 0 on the ordinate.
Figure 47. Fame versus rank: the number of most-cited researchers in each top U.S. engineering university versus the rank of the engineering university.
This is not a chicken-and-egg argument. The direction is one-way. The university rankings come from the highly cited, not the other way around. Authors are famous because of their creativity, not because of the name of their employer. In my own field, we cite Ludwig Prandtl all the time because of his boundary layer theory, not because of the fame of his employer, the University of Göttingen.
The scatter in Figure 47 does not diminish the firmness of this conclusion. One can argue that “size matters,” which is why some highly ranked schools (those ranked numbers 4 and 19, for example) do not have any high-cited researchers. These examples are the exception, not the rule. To stress this, I replotted the points of Figure 47 by scribing the same values on the ordinate, and using a new abscissa: the rank of the particular school on the list of the most-cited researchers in all engineering. The result is Figure 48. For example, rank 1 on the abscissa of Figure 48 belongs to the school with the most names on the most-cited list (that school was ranked 2 on the abscissa in Figure 47). Because of the logarithmic ordinate in Figure 48, the points with 0 on the ordinate are not shown.
In the new representation of Figure 48, the points descend smoothly from left to right, producing the same pattern we find in other natural flow systems. Practically all the points that were on the left in Figure 47 are still on the left in Figure 48. Immobility also characterizes the points on the right in Figures 47 and 48. The 30-to-32 abscissa range of Figure 47 is essentially the same as the 25-to-40 range of Figure 48.
Figure 48. The number of most-cited researchers in each top U.S. engineering university versus the rank of that university on the most-cited list.
Figures 47 and 48 show that the ranking of universities is hierarchical, like the airways of the lung, the channels of the river basin, and the cities of a country or continent. The more highly ranked, the fewer the candidates for the high positions. The trachea, the Danube River, and Paris are not to be confused with the other airways, river channels, and human settlements. The opposite is true in the other direction: The lower the rank, the more numerous the potential candidates; hence we see more apparent movement the farther down we go on the U.S. News rankings. Why?
The clue lies in the nearly straight line that the data form on the log-log plot in Figure 48. This line has a slope between −1/2 and −1 and is coincidentally the same as the distributions of city sizes throughout the modern history of Europe (Figure 41). The similarity between Figures 48 and 41 suggests that the distribution of sources of knowledge is intimately tied to geography, geology, and history (to the evolving drawings of the flows on the landscape), and to the tissue of information channels on the surface of the globe.
This insight allows us to take another step in our constructal view of education. So far we have shown that the fame (usefulness) of the ideas generated is the current that flows through universities and accounts for their rankings. Now we will see how this also allows us to predict the evolution of the global education system, from simple to more complex constructs (from a few schools, or channels, to many across the landscape). To do this, we will employ the same type of proof that allows us to predict the size and distributions of channels in a river basin, trees in a forest, or cities in a country or on a continent. All hinge on the prediction that as flow systems become larger, covering a bigger area, they should facilitate the access for the currents that move through them. For universities, this means that a hierarchical vasculature should emerge that facilitates the flow of ideas.
Here is how to use flow geography to predict the linear-logarithmic trend visible in Figure 41 (the same trend would appear fuzzier but still linear if Figure 47 were replotted in log-log coordinates). Imagine an area element A1 with a population N1. The inhabitants produce things (students, agricultural products, timber, game, minerals, etc.), the flow rates of which are proportional to A1. These flow rates sustain a human settlement located on A1, where the number of inhabitants is N1 and the production is of a different sort (education, knowledge, services, devices). There is a balance between what flows from the area A1 to the human concentration N1, and what flows from N1 to A1. The key idea is that both classes of flow rates (area-to-point and point-to-area) are proportional to A1, and this means that the size of the human settlement N1 is proportional to A1.
One type of service that flows from the human concentration N1 to the humanity spread over A1 is education, educated individuals, books, knowledge, and science. The human settlement in this case is the university, and the area A1 is the territory that the university serves. The constellation of universities on the landscape is a reflection of the area constructs of land-city counterflows that cover the entire globe.
As we saw in chapter 7 with our discussion of the Atlanta airport and the evolution of urban transport systems, if the objective is access (a shorter travel time), then the distribution of human movement on the Earth’s surface can be viewed as the compounding of area constructs, as shown schematically at the top of Figure 41. Like an area element in a river basin, which feeds the big stream that leaves the area, each area construct sustains the flows that reach a human concentration on the boundary of the construct. It follows that the human concentration on the boundary is proportional to the size of the construct. If the human concentration represents the university, then the university (flow of ideas, impact) is proportional to the size of the area construct that it serves. Over time, the landscape is covered by more and more universities that should have multiple sizes and are assembled hierarchically.
The construction sequence made in Figure 41 is based on area doubling. This construction is how we discover theoretically the pattern hidden in the present-day rankings. Note that the construction of Figure 41 (see this page) is not a “time sequence” (from small to large) of how the landscape might have been covered by the flowing tapestry of knowledge in history; it is simply a mental viewing of how the patches of the quilt are pieced together. The construction is shown in the bottom left of Figure 41, where the size of the black dot is meant to indicate the rank (that is, the flow rate of knowledge) that the human settlement generates. Given an area, the top-ranked university serves not only the area but also the lesser-ranked universities that are spread on that area.
The bottom left of Figure 41 shows the distribution of multirank universities on the landscape after deleting the construction lines used earlier. The hierarchy of ranks is evident: one top university, two universities tied for places 2 and 3, four universities tied for places 4 through 7, and so on. This pattern is discovered here based on pure theory and is represented by the same stepped line as for city sizes in Figure 41. The slope of this line is −1/2, in acceptable agreement with what we saw in Figure 48. The important conclusion is not the predicted slope but the fact that the line should be straight and that it has its origin in the area-to-point access for the flow of information between many inhabitants who live on the same landscape. This approach is validated by the fact that we find the same slope in the actual rankings of universities.
Why is the hierarchy rigid?
The short answer is that ideas, science, and education flow all over the globe like water in all the river basins. When numerous researchers value and use an author’s work, the idea flows from the author to the user. It flows “well” because of the long history and entrenched geography of the flow network, which are due to the evolutionary process that brought the whole world of information sharing to the present level of effectiveness. The success of this evolutionary process goes unnoticed. And yet, it is the reason the user from one end of the globe actually looks for, finds, and trusts the ideas and young professors produced by a famed university or a professor located at the other end of the globe.
There are many intermediary channels along each route: other universities, disciples of known professors, journals, books, libraries, etc. The intermediaries have evolved into a hierarchical flow structure—the right sizes, put in the right places, nourishing and sustaining each other. Each route is a vascular point-to-area flow (from one source to the entire globe) or a vascular area-to-point flow (from the entire globe to the famed source).
These hierarchical flow designs serve all the scholars well. A hierarchical design that concentrates leading scholars in certain schools is a more effective design for facilitating the flow of ideas than a design that spreads these bright lights evenly across all the schools. Intuitively, we understand that shared resources and the ability to bounce ideas off colleagues should help spread the flow of knowledge. Just as a river basin needs a few large channels and many small ones, so, too, does the river basin of education. The design that has evolved is much older and more polished than a new design that someone may promise to put in place today. The highly ranked and the lowly ranked go together. The flow of science improves in time because each university improves while maintaining the place that it has earned in the global structure.
University administrators who promise to change the rank of their schools by simply stealing one top name from a highly ranked school are defeated by nature every time. Sure, the school’s ranking might change a little, moving from thirtieth to twenty-sixth place, but it will still be part of the large group trailing the leaders unless there is a cataclysmic change. The same fate awaits the one who wishes to change rankings by building something artificially big—artificial, because it is not demanded by the natural evolutionary flow and geography that created the tapestry of academic flows that covers our world. An example of artificially big is when a president suddenly decides to spend and build to double the size of his school, because “size matters” in the formula used by U.S. News & World Report. Such wishes are analogous to damming, blocking, or digging river channels. The artificial features of the flow network require constant maintenance (spending), more when the artificial does not resemble the natural. In the end, the water knows how and where to flow, the dams break, the dug channels dry up, and the natural design wins.
Age matters in this evolutionary design as it does in all others because it is good for performance. Over time, the river basin improves the positions of its channels, and the channels stay in roughly the same places. The channels have hierarchy: A few large channels flow in harmony with the many small channels. A sudden downpour is served well by the “memory” built into the old riverbeds.
Similarly, the older universities have dug the first channels, which are now some of the largest channels that irrigate the student landscape. Again, “largest” does not mean the greatest number of bodies moving in and out of the classrooms. It means the streams of the most creative, that is, the channels that attract the individuals who generate new ideas and who develop disciples who produce and carry new ideas farther on the globe and into the future. The swelling student population is served well by the “memory” built into the education flow structure.
From this view follows the prediction that the hierarchy of universities should not change in significant ways. This hierarchy is as permanent as the hierarchy of channels in a river basin. It is natural because it is demanded by the entire flow system (the globe) in which huge numbers of individuals want the same thing (knowledge).
Is there a way to change rankings? There is, but it takes time, and the river basin provides the perfect metaphor for it. Cataclysmic change (for example, plate tectonics) in the landscape of flow access is the answer. Likewise, the flow of higher education can be diverted through major changes in the loci of generation of new ideas and channels for the flow of information.
Freedom is good for design. We have seen this many times in the evolution of the flow of knowledge, from the movement of Leonardo da Vinci from sponsor to sponsor to the abrupt transformation of nobodies into famous research universities in the United States right after World War II and again after Sputnik. Then, the cataclysmic change was the freedom that attracted the brain drain from postwar Europe, and, after Sputnik, the enormous jump in funding for fundamental research (that is, basic science). These changes had the effect of instituting a marketplace where the flow of ideas was freer.
Not a richer one, not a bigger one, and certainly not one that was to be used as a generator of profit for ancillary and politically correct projects on campus. No. The way to create true academia on a plot of dirt was by putting up a table with free food called ideas. And the truly creative came, to create.
Once you know to look for it, you recognize the evolving vasculature in a river basin because it is a single flow system. It can be much harder to see this design in complex entities like universities because they are channels for many different flows, all superimposed on each other. Education, after all, covers all forms of the transfer of knowledge. It is a global flow vasculature composed of a very large number of flow trees that connect the few who know with the many who need to know in various fields that include mathematics, biology, business programs, and various other areas of inquiry, all of whose ideas flow through tree-shaped, hierarchical channels. It will come as no surprise that the same school can occupy very different positions in the various hierarchies of education—MIT, for example, is a main channel for engineering, not English literature.
To examine this phenomenon, my student Perry Haynsworth and I extended the constructal view of higher education by examining one of the prominent aspects of modern university life: athletics. Specifically, we asked a question often raised by fans of college basketball: Why does it seem that the same schools battle it out each year in the NCAA basketball tournament? Why are there a few university basketball programs that are always successful, while many more continually struggle? Is there a hierarchy in this most competitive arena as rigid as that found in other natural phenomena?
We predicted that the ranking of college basketball teams should be just as rigid because college basketball, too, owes its existence and robustness to a geographical tapestry of area-to-point and point-to-area flows of multiple sizes. The movement of basketball players from high school to the professional level is a flow with its own architecture. There are over 23,000 high schools in the United States. Practically all have basketball teams. The talent ranges from those who would never dream of playing basketball in college to those who aspire to the National Basketball Association (NBA). Several years ago the NBA instituted a minimum-age rule, requiring players to be nineteen years old before entering the NBA draft. As a result, basketball players are essentially forced to choose a university path to the NBA. There are 330 Division I basketball programs that channel players to 30 NBA teams. The high schools and universities are tributaries to the big river that leads to the NBA.
Figure 49 shows how the top university basketball programs arrange themselves when ranked according to their total number of appearances in the semifinal round (the “Final Four”) of the NCAA Tournament. When plotted on a log-log field, the data trace a nearly straight line with a descending slope. This feature is important because it unites all natural flow systems that cover the land (see Figures 40, 41, and 48). This distribution is a characteristic of the organization of all flow systems that morph freely and compete for access on the same finite-size territory.
The ranking of the top teams tells a similar story when the measure is the number of players that each team sent to the NBA from 1949 to 2007. These data are plotted log-log in Figure 50, and their alignment is the same as in Figure 49. This conclusion is reinforced by Figure 51, in which we cross-plotted the two rankings (Figures 49 and 50) as one abscissa against the other. There is a correlation between success in the Final Four tournament and success in sending players to the NBA. The cloud of data embraces the rising diagonal, and the scatter diminishes greatly in the lower-left corner, that is, at the top of the rankings. The more successful NCAA teams serve as larger and faster streams to the NBA. In conclusion, the hierarchy is rigid.
Figure 49. The number of appearances in the Final Four of the NCAA Tournament versus the rank of each team on that list.
Figure 50. The number of players selected by the NBA from each team versus the rank of the team on that list.
Figure 51. The robustness of hierarchy: the rank based on the number of players recruited by the NBA (Figure 50 abscissa) versus the rank based on Final Four appearances (Figure 49 abscissa).
The robustness exhibited by university and basketball rankings also contradicts the appealing argument that rankings depend on the formula used to calculate the rank. This is explained by the fact that the hierarchy of natural point-to-area flows has two main features: pattern and diversity. These features are evident in the distribution of tree sizes and numbers in the forest and of cities on a continent. They are also present in Figures 47 through 51. The scatter represents the “diversity,” which is located primarily in the lower ranks, where there are many competitors for the same rank. It is for this large group that the chosen formula matters, but it matters little; that is, both criteria—the Final Four and the NBA draft—produce very similar results.
To leapfrog a few bicycle racers in the peloton (the thick end of the cloud of data in Figures 48 and 51) is to remain in place, inside the peloton. The runaway racers are well in front, and they have names. Their alignment on the diagonal (Figure 51) represents the “pattern.” This is hierarchy, and it transcends all the scheming that goes into ranking formulas and claims that a university (academics or basketball) can be redesigned to score higher in the rankings.
These features (robustness, pattern, and diversity) reinforce the physics view that basketball education is a flow system that sweeps the land, while constantly generating flow structures that are more and more efficient. In this evolving design, the top schools are the big branches. They are the few, not the many. Their identity is permanently carved into the geography of the global flow system.
Basketball is just one kind of education that flows with evolving design on the landscape. Every other discipline in which training is pursued by students living on the same area is a flow system with lasting architecture, in which a few large channels flow in harmony with the many smaller channels. The large channels are the highways on which the faster- and farther-moving students travel.
If we superimpose on the global geography all the flow structures of the various disciplines, we begin to imagine how universities constitute their natural global design. Consider now the comparison between the ranking of universities and the ranking of basketball programs (Figure 52). There is no relation between the two rankings. Had they been related, their data would have fallen near the rising diagonal. Most of the universities appear in only one of the rankings. This is why most universities fall on the sidelines of Figure 52. They separate themselves into two different worlds, two distinct flow systems on the globe.
Figure 52. The ranking of universities according to U.S. News & World Report (x) versus the ranking according to the number of players drafted into the NBA (y). Most of the universities fall outside the 75 × 75 area: They are plotted on the sidelines (those with x > 75 are plotted at x = 75, and those with y > 75 are plotted at y = 75).
When educators and sports announcers refer to college players as “scholar athletes,” they misrepresent both worlds. “Basketball students” is a more accurate name, as is “engineering students” for those who study engineering. This stresses again the notion that the global flow of education is a superposition of evolving vasculatures associated with the various disciplines, just as the grid pattern of city streets is the superposition of various flows to various points of interest.
The channels of basketball excellence are not the same as the channels of excellence in academia. The two flow architectures have different histories, memories, and channels. This dissonance is physics, and it is worth contemplating because it runs against one of the pillars of modern education: mens sana in corpore sano (“a healthy mind in a healthy body”; from Juvenal’s Satires). Modern education has been right to adopt this doctrine, because it works. This doctrine, however, is not happening by itself, as the two-world reality of Figure 52 demonstrates. The university design needs constant improvements to tolerate and maintain this doctrine; it needs reminders and reinforcement, just like the dams that protect the city from the big river that passes through it.
Our look at university and college basketball rankings underscores two insights derived from the constructal law. First, all natural flow structures that are free to evolve—from the rankings of schools or teams to the size and distribution of channels in a river basin or trees in a forest—are characterized by rigid hierarchies. Second, when we plot these multiscale designs on a log-log graph, we should find a rigid distribution line.
While all flow structures are improving, some are hidden from view as they morph. In social dynamics, the hidden constitute a field of study called “dark networks” and “mafias.” Until now we have examined social systems whose flows are relatively easy to recognize and are, on the whole, meritocratic. The selection of athletes discussed in chapter 4 is based on an obvious criterion: their speed on the track or in the pool. The school rankings we’ve described in this chapter also provided the expected results: Those with greater academic impact or success on the basketball court enjoy greater prestige in their domain. But we all know the world doesn’t always work this way. For many flow systems, access to the best channels is based on personal connections, on whom you know and who needs you for the safety and perpetuation of the network. I explored dark networks in my paper “Two Hierarchies in Science: The Flow of Free Ideas and the Academy.”
Briefly, I started with the assumption that the membership of the National Academy of Engineering should align with the list of most highly cited researchers. That is not what I found. The resulting comparison—between 171 highly cited authors and 2243 academicians—had a ratio of 1:13. Furthermore, only one-third (60) of the highly cited individuals are also in the Academy, and they represent only 2.7 percent of the Academy membership.
Thus we see that the pattern of generation of good ideas is in disaccord with the pattern of admission to the Academy. The reason is that knowledge and Academy membership are two very different flow systems in the same landscape. The first concerns the flow of ideas; the second, the flow of people already in the Academy.
This phenomenon is prevalent in human relations. It’s no accident that the phrase “it’s not what you know but who you know” is one of our most enduring clichés. Businesses, for example, are flow systems for goods and services. But they are also vehicles through which owners and managers reward family members and friends with jobs and money. This hiring strategy offers many advantages to the business, especially as it reduces the time spent finding loyal employees. And as long as the company does not become weighed down by mediocrity, it may flourish. This strategy is also a vestige of our feudal past, when the names of the insiders are known to everybody, like the names of the few powerful families in a certain area. Once inside the house, the family invites in the relatives, not the strangers.
Finally, though the constructal law focuses on construction and the coalescence of entities (whether they be raindrops or people) into larger flow systems, the individual remains important.
In my paper “Constructal Self-Organization of Research: Empire Building Versus the Individual Investigator,” I noted that empire building is a phenomenon that dominates today’s research landscape. Large groups, national priorities (for example, nanotechnology, fuel cells), and research centers dwarf the spontaneous individual investigators. Administrators and the thirst for higher rankings encourage this trend. Yet the individuals do not disappear. The paper explained this by linking the emergence of the large group to the pursuit of greater visibility for the institution as a whole. The visibility (V) was modeled as a product of the production (P) of ideas in the institution, and the support (S) that the institution secures for the production of ideas.
I showed that the coalescence of some investigators into a large group tends to increase S and decrease P. On the other hand, an increase in the number of individual investigators has the opposite effect. From this trade-off emerges the main and well-known features of the contemporary research organization: the proportionality between the size of the large group and the size of the entire institution, the strong relationship between the visibility of an institution and its size, and the fact that large groups (empires) occurred first in the largest and most research-intensive institutions. I also showed that as the incentives for large-group research become stronger, smaller and smaller institutions find it beneficial to abandon the individual investigator mode and seek a balance between research empires and individual investigators. Thus, the individual researcher will not disappear, for the same reason that older types of movement and ancient animals are not always replaced by newer designs. The invention of carts and automobiles did not spell the end of walking, because that is still a good way to move in many circumstances. Similarly, insects were not replaced by birds, because global flow is enhanced by components of varying sizes. The tendency toward hierarchical organization is not a push toward large, entrenched structures. It is a balancing act in which the few large and many small work together to enhance flow. It takes all sizes.
I know this firsthand. Before my “empires versus individuals” article, I thought I was alone, an anachronism in the eyes of progressive administrators. I was wrong. After this article, I was stopped on campus and contacted by colleagues from around the world who see the world of ideas the way I do. No, the individual is not disappearing; far from it. The individual is everywhere.