“Send your best and brightest,” read an e-mail invitation sent by a MacArthur Award–winning neurobiologist to every department on my former campus, “to an informal, bi-weekly seminar whose purpose this semester will be the restructuring of the current organization of knowledge.” This appeal stood out for its ambition and confidence—“this semester”—but the goal of restructuring knowledge has appeared on many agendas since the late twentieth century. The blame that has fueled the will to truth in the modern disciplines—driving their embedded systems to include more and more things so as not to fall short—appears to be assuming the form of desire: a call from inside and outside the academy for new forms and combinations of knowledge—for what is most often termed interdisciplinarity. From the proliferation of interdisciplinary humanities centers to E. O. Wilson’s dream of “consilience”—a leaping together of the disparate disciplines (Wilson 1999)—to this invitation to take on the “Whole Enchilada,” blaming has given way to hoping for a new resolution. As with the blame, the timing and intensity of this desire are keyed, I will argue, to how system works within our organization of knowledge. The change it signals is taking shape—like the previous shifts this book has described—as a redeployment of system.
My experience with the Whole Enchilada group provided the first clue as to the nature and timing of that change. In culinary terms, an enchilada is about adding in the chili. The secret to making a whole enchilada is to leave nothing out. The invitation to this one was an invitation to totalize—to put this more precisely, to totalize knowledge in a different way. As department chair, I sent myself to reserve a place for my discipline in this new whole. What neither I nor the organizer nor the other participants even entertained back then—at the turn into the twenty-first century—was that “discipline” itself, as the assumed shape of knowledge since the time of Britannica, might itself be a candidate for change. While interdisciplinarity preserves discipline as that which is being mixed or shared, the possibility of a reshaping of the basic unit of knowledge was not then on the table.
As with the recipe for a whole enchilada, we added everything into this group. The first few meetings, variously attended by historians, sociologists, psychologists, biologists, physicists, medical doctors, engineers, and others, were not promising. At best we enacted the interdisciplinary norm: wary encounters at a distance, punctuated by the borrowing of desirable content. But, then, what was only a whisper from across the table at the start of the semester became a chorus. Darwin was somehow the lingua franca of this group, but exactly how and why was not clear. The portability of “evolution” was certainly a significant factor, but since that porting had already been going on for quite a while, there had to be something else in play that more fully explained the connection between Darwin and our totalizing.
The explanation that follows was not the consensus of that group—and even now, since it entails a turn to the present and future, this account is necessarily more speculative than the ones I have offered for earlier historical moments. It benefits, of course, from what I have learned about the past in writing this book and what has happened since the group met. But what has been most crucial to mounting this explanation has been the positing of different kinds of histories. As an alternative to the history of ideas, the history of mediation tells us to look to form as well as to content, and the history of the real, although a necessarily blunt instrument, scales up our perspective on how forms change and how change functions over time.
My primary clue back then for identifying Darwin’s purchase on the group was a book published just a year before we convened. The philosopher Daniel Dennett, in trying to explain why he considered Darwinian thought to be a “universal acid,” altering all that it touches, came up with a catchy—but I think misleading—title: Darwin’s Dangerous Idea (1995). The real catch in his argument is that he defines the idea in two different ways. At first, he claims that natural selection is the idea. But then he complicates matters by calling natural selection an “algorithm.” Darwin, he points out, actually wrote up natural selection in the same “If … then” format that we use today in computer programming. The dangerous idea is then significantly reworded as follows: “All the fruits of evolution can be explained as an algorithmic process” (60). Without exception, in other words, every manifestation of life on earth is a product of the “mindless mechanicity” of the algorithm of natural selection running again and again; no “skyhooks”—Dennett’s term for any form of external help, such as God, beauty, or manifest destiny—allowed (74).
In the first version, the danger lies entirely in Darwin’s particular subject matter—what natural selection means. In the second, it shifts toward a more general and consequential issue: what an algorithm can do. That shift, I would argue, is crucial, for it is what makes the acid universal. The importance of natural selection for knowledge beyond the biological lies not in the content but in the form it assumes, as well as in what that form—as form—is capable of performing. If, in the instance of biological diversity, a simple algorithm can account for the greatest complexity, then it would be important to identify other instances of this formal phenomenon.
The clear and present “danger” in Darwin, I am claiming, lies not in natural selection itself but in that phenomenon. An algorithm, in Dennett’s words, is a “certain sort of formal process” (50)—that is, a process written up in a certain form. Although the idea of evolution has been very powerful in general and certainly was for the Enchilada, Darwin’s purchase on the group’s drive to wholeness—the primary danger it posed and continues to pose to the “current organization of knowledge”—is better understood as a matter of a form and its effects, of mediation by a genre.
The genre at issue here may at first glance appear to be unrelated to our generic focus on system. However, as Dennett’s use suggests, “algorithm” can and should be engaged as more than local to maths and computer science. In fact, even in those fields, algorithm and system are deployed more generally as variations of each other. In a semantic sense, for example, The Free On-line Dictionary of Computing defines system as “any method or algorithm” (foldoc.org, accessed Oct. 26, 2013). Seeing their variation from each other as a matter of degree rather than kind highlights system’s scalability and thus its historical role in negotiating the relationship between complexity and simplicity. To put this in more practical terms, if we reverse-engineer any complex system, we will find—per Dennett’s example—the simple algorithm that generated it.
Algorithm is thus a necessary part of our history of system as a genre, and we can tie it in by connecting Darwin to our earlier analysis of the role of system in Malthus. He claimed that the principle of population had evaded even “the most penetrating mind” because of things that had been left out: “the histories of mankind that we possess are histories only of the higher classes.” With the lower classes included, the regular oscillations of population could be detected. Translated from history into system, their regularity became a “law of nature” (1798, 32, 15).
Sound familiar? Here is the more famous biological version: expand the histories of life on earth to include lower species and thus make visible previously undetected patterns that suggest a natural principle governing all forms of life over time. Like Malthus, Darwin raised the ratio of system by making the principle generically travel—in his case, he also literally traveled on the Beagle—to include more things. He did so by innovating on system, deploying it algorithmically as his new mode of inclusion. Whereas the standard account of Darwin’s debt to Malthus features an idea—that Darwin was struck by the notion that population pressure would induce a contest for survival—the history of mediation calls our attention to form. The systematizing of political economy that depressed Malthus’s contemporaries with the prospect of a starving world enabled Darwin.
We saw earlier how Malthus’s Essay exemplified the power of embedded systems—of how that generic mix became newly masterful by avoiding the liabilities of the all-inclusive Master Systems that preceded it. Darwin’s work, in turn, became an example for us, for it addressed the new set of liabilities that emerged from embeddedness. Unlike Malthus, he did not take those problems of fit and incompletion as an occasion for apology or revision; rather, he took it on as an opportunity to make his case in the manner Dennett describes, betting his system on the still unproven reach of its informing, algorithmic principle. “If it could be demonstrated,” he wrote, “that any complex organ existed, which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down” (Darwin 1909, 46, emphasis mine). By asserting not the truth of his theory but the conditions under which it could be proven wrong, Darwin lured even his opponents into the tasks of accumulating more examples and of trying to fit them to evolution. Together, they raised the ratio of the system by increasing the numerator—the number of things explained.
That placed the pressure, of course, on the denominator: would more principles be required? Newton let God haunt his system from the outside as a final cause. Malthus initially worked differently, making his system travel, but once it reached the crucial juncture—where population was supposed to explain morality—he opted for the same skyhook: God engineered the population problem, he claimed, to turn the earth into a hornbook for the soul, testing and teaching us through suffering (Malthus 1992, 79, 241, 385). Darwin, however, introduced a different mode of amplification: by habitually articulating his informing principle in the conditional “If … then” format, he harnessed the amplifying power of repetition. Every time the condition is met, the principle reasserts itself; thus the more it travels, the more powerfully it resonates.
Algorithm thus played a major role in Darwin—so major that we need to focus in on a twist in how it was deployed. Actually, it was more a doubling than a twist—a doubling that echoes the doubling of system I highlighted at the start of this book. We speak of system, I observed then, as something both conceptual—a way of knowing the world—and as something that is really there, that constitutes part of that world. In pinning down Dennett’s sense of danger, we have enacted the same phenomenon in describing the algorithmic: we engage it as both a formal mode of producing knowledge about the world and—thanks to its simple-to-complex “mechanicity”—as constitutive of that world. In a sense, the world takes shape in the same way that knowledge does: both are products of an algorithmic system.
This is neither error nor ambiguity in Dennett’s argument or my own. Doubling of this kind—think of our earlier example of Adam Smith reacting to Newton’s system “as if” it were “real”—points to how the structural resemblance between knowledge and its object can signal that the object is indeed knowable. In the Enlightenment, it generated the conviction that the world could be known—known because the representations we call knowledge seemed to resemble it so closely. One way to grasp the importance of Darwin’s turn to algorithm is to see it as putting real pressure on that “seem.” In his algorithmic use of system, Smith’s “as if” becomes a superfluous buffer. Instead of just an object to be known, the world (or nature in Darwin’s case) becomes the product of a form of that knowledge: the simple instruction that we call an algorithm.
Strangely enough—or not so strangely, given the focus on “life” they shared with Darwin—that is what Watson and Crick’s discovery of the structure of DNA confirmed a century later: genes carry a set of simple instructions, written in the repetitive syntax of the algorithmic, that are read off for purposes of reproduction. In David Deutsch’s words following Richard Dawkins, genes “embody the knowledge necessary to render their organisms.” “What the phenomenon of life is really about,” he claims, “is knowledge” (Deutsch 1997, 180–181).
Deutsch uses the term virtual in a very specific way to up the ante in considering the alignment of knowledge and the world:
Our external experience is never direct; nor do we even experience the signals in our nerves directly—we would not know what to make of the stream of electrical crackles they carry. What we experience directly is a virtual-reality rendering, conveniently generated for us by our unconscious minds from sensory data plus complex inborn and acquired theories (i.e., programs) about how to interpret them. (1997, 120–121)
Virtual reality refers here not just to computer simulations but to “any situation in which the user is given the experience of being in a specified environment” (122). The fact that it is possible, argues Deutsch, tells us something not only about our mode of knowing but also about the structure of that which is known. If “every last scrap of our knowledge” is encoded in programs for rendering the physical world on the brain’s virtual reality generator, then we experience those scraps as knowledge because they allow us to survive in that world. “The fact that virtual reality is possible,” observes Deutsch, “is an important fact about the fabric of reality” (122).1
Bear with me here, for this argument points to a new and important answer to the debate we have been tracking about the relationship of system to the world: are systems solely made by us in an effort to know the world, or is the world itself a system that we learn to discern? Adam Smith, as we have seen, opted for the former. Systems, he insisted, were “mere inventions of the imagination.” The abbé de Condillac, however, declared just as forcefully that “systems should not be invented. We should discover those which the author of nature has made.”
What bridged these two positions was the emphasis not on the nature of systems but on the nature of their makers: human or divine? What Deutsch’s virtuality argument highlights, however, is the nature of nature itself. And that is where Darwin’s algorithmic workings of system come into play. Should we take the doubling to its logical conclusion? Have we been able to know the world through system because the world is, in fact, itself structured as a system? And is it a system because system itself does the making? If so, then system would not be a product of man or god—as in Smith and Condillac—but would itself be a mode of (re)production.
Pushing the argument in that direction would pry open what we might call Darwin’s door—the door to what the scientist Stephen Wolfram (2001) proclaims “a new kind of science.” With Newton as its model, classical science and its relativistic variants have been using systems cast in the language of mathematics to try to approximate the real, constantly massaging the equations with new constants and particles to make them match increasingly stranger versions of that reality. For Wolfram and others—he has been roundly blamed for taking too much credit for this type of argument2—the running of system is not an approximation; it is the very nature of the real. The same, simple algorithmic systems with which we compute are not simply rules for taking the measure of the real; they actively account, he claims, for reality in all of its complexity.
I am neither advocating nor judging these related but often conflicting claims; Deutsch parts ways with Wolfram, for example, on what is “inherently simple” (Zenil 2013, 12694). I wish only to add them to my account of system’s role in reshaping knowledge—and thus to clarify what may be at stake in making them now. While in the three primary parts of the book I recovered the earlier adventures of that genre, my job here is to report and synthesize its more current exploits. There are many of them, and they are hotly debated. In Wolfram’s case, the complaints by fellow scientists and the media that he has hijacked the work of others to sell as his own have themselves hijacked the initial reception of his arguments. But once we recognize that both his claims and the acts of blame they have generated are episodes in much longer histories, then questions of a different kind—of form as well as of content—arise.
Why, for example, the big book? To rivals angered or intimidated by his 1,263 pages, Wolfram could have pointed out that it worked for Newton—another scientist plagued by debates over originality, as in Leibniz’s claim to the calculus. Newton, as I noted earlier, deliberately opted for huge, totalizing books over journal publication, a practice that was eventually reversed as the grand ambitions of Enlightenment gave way to essays, articles, and the other specialized practices of the disciplines.
Wolfram’s scaling back up, however, should not surprise us, for it highlights how efforts at new kinds of knowledge require shifts in form as well as different ideas. “I gradually came to realize,” recounts Wolfram, “that technical papers scattered across the journals of all sorts of fields could never successfully communicate the kind of major new intellectual structure that I seemed to be beginning to build” (2001, ix). The result is both an echo of earlier efforts to totalize and a departure from them: his big book has been heavily remediated by new technologies. Wolfram wrote and produced it in then new kinds of software—Framemaker using Mathematica—thus allowing for its extraordinary array and display of illustrations. But that was just the first step. The scaling up has been an ongoing project that exceeds the physical bounds of the book itself. A New Kind of Science has its own website, where the full text is free online, accompanied by complementary software as well as message centers, directories, and other forms of explication, communication, and publicity.
Why are algorithmic systems at the center of all of this attention now? System’s presence at a scene of knowledge should certainly not surprise at this point in my argument. Neither should the timing if we focus on the machines that algorithmic systems animate: computers. Wolfram’s entire project is based on the growing capacity of that technology to run algorithms over and over again, producing patterns of complexity from their simple rules. “Every detail of our universe,” he asserts, “everything we see will ultimately emerge just from running this program” (545). He means “see” literally, for what emerges are patterns on the computer screen—patterns of cells that change color and number with each iteration of the algorithm.3
But—and this is the crucial point in understanding Wolfram-like arguments as a historical event—it is not the running of algorithms that is new; it is the speed at which the electronic and digital allow us to do it. Algorithms were, per Darwin, “run” by earlier technologies—and, despite the slower speed, to great effect. It made Darwin the start of something that led much later to both the convening of the Whole Enchilada and—just five years later—the publication of Wolfram’s book.
I know how strange this connection sounds in terms of both its duration and its dramatic personae: 142 years between Darwin and Wolfram. But making that connection can help to illuminate how shifts in technologies entail shifts in form and thus in knowledge. Just as the proliferation of print in the late eighteenth and early nineteenth centuries contributed to the shift from Master Systems to the modern disciplines, so the dramatic increase in computer clock speed (the running of algorithms)4 in the late twentieth century has occasioned claims to new kinds of knowledge.
If this connection is load bearing, then Darwin’s legacy is clearly changing, extending far beyond the religious, philosophical, and scientific controversies of evolution itself. Its new reach can help us to identify the Enchilada’s call for restructuring as yet another effect of the type of formal change I have tracked throughout this book—the redeploying of system. What does that type of change look like now? How would it feel?
One manifestation has been the new episode of clubbing I have described. As with the Fair Intellectuals, there was much pleasure around the table as we gathered together to take on the “character of members” of the Whole Enchilada. It was a pleasure linked both to the use of a relatively new and empowering technology—computation this time instead of writing—and to the sense that this change in character was also a resystematizing; per their nine muses, we were finding our places in a newly constituted whole. The shape of that whole was and still is unclear, and so keeping all options open requires an open-ended term for the desire that brought us together. Instead of interdisciplinary, with its baggage of preserving discipline, we might only need to set a vector for change. The term dedisciplinary points simply to a sense of moving from (“de-”) the structure we have now, leaving open the possibility that even the basic building block of that structure—“discipline”—might itself change.5
The history of blame provides one way to frame that possibility. One reason that blame has accompanied system, I have argued, has been that system was understood to be something made—made to approximate the real and thus always falling short of it. Thus Locke complained, as we saw earlier, that
though the world be full of systems …, yet I cannot say, I know any one which can be taught a young man as a science, wherein he may be sure to find truth and certainty, which is what all sciences give an expectation of. (1693, 229–230)
But what if now, in a way Locke could not have anticipated, system came to be understood as truth—its algorithmic repetition yielding the very certainty that science seeks? If, to push Darwin’s door wide open, system was understood to produce the world rather than approximate it, could the disciplines still function as disciplines without system to blame for falling short? What if the effect of system traveling into those specialized fields was not the attenuated authority I described earlier but the sense of confident consilience that motivated the Enchilada? And what if system’s new reshaping of knowledge is turning the epistemological logic of disciplinarity on its head, reformulating the once startling conviction that the world could be known through specialized, complex systems into the notion that knowledge—in the form of simple, iterative systems—renders the world.
My purpose here, of course, is neither to predict nor root for or against this future or any of the many variations of it.6 Instead, I am extrapolating this particular one to bookend my turn back to Darwin with a look forward to what might be on the other side of the door—to the shape that Dennett’s sense of danger and the Enchilada’s ambition may yet assume. Doing so highlights the issue of duration—the apparent delay, that is, between what Darwin did and how we are realizing it now. For students of technology, such a delay would appear to be yet another example of new tools and technologies taking time to take hold—to mediate, as William Warner and I have put it, in the particular ways that we have come to expect. They do so only after assuming forms and functions that were not obvious, or even possible, when introduced. There is, for instance, Lisa Gitelman’s often hilarious study of how telegraphy morphed from a visual and writerly tool that used electricity to record messages on paper tape to an aural and speechlike medium for electronic communication. The computer’s long metamorphosis from Charles Babbage’s programmed gears of the 1830s to the ENIAC’s missile-targeting tubes of the 1940s to the iPhone is another now familiar example (Siskin and Warner 2010, 10, 120–135).
That delay corresponds in length and dates with the Darwinian one—a surprise at first given that we do not think of Darwin as part of the history of the computer. However, once we recognize what they have in common—the centrality of algorithm and its iterative potential—then Darwin’s legacy might be seen to rival Babbage’s.7 Equally illuminating as both a precedent and a context for Darwin’s fate is the delay with which I began this book. Galileo’s message from the stars, received and sent 250 years before Darwin, broke the seventy-year-old logjam preventing Copernicus’s legacy from being realized. His discovery of a second lunar system, I argued, put abstract debates about system finally and firmly into the physical and the plural—and thus into the history of system(s) this book compiles. This type of delay—let us designate it a Copernican Delay8—was less about finding a purpose for a technology than finding a technology adequate to a purpose. For Galileo, that meant an “improved” spyglass. For our engagement with Darwin, it means an improved processor—one fast enough to generate the message that Darwin, like Copernicus, may belong squarely within the histories I have brought to bear on system.
If Darwin does land there, his turn to the algorithmic may mark the initial unfolding of the next chapter in the history of the real. Processors and processing have not only improved; they have also now taken over center stage in debates over the nature of the real. The year 2013 began with an 856-page volume, from over fifty authors, titled A Computable Universe: Understanding and Exploring Nature as Computation. Most of the contributions argue for and explore what the editor, Hector Zenil, calls “the informational nature of reality”9
The disagreements are sharp and deep, but one common theme is that there has been, in Seth Lloyd’s words, a “paradigm shift in how we think about the world at its most fundamental level.” That shift posits a universe no longer made up of atoms or just energy, but of basic bits of quantum information (qubits). It thus unfolds computationally—it computes itself into existence—and, according to Lloyd, out of it. In his own monograph, he uses an estimate of the number of qubits that make up that universe to calculate exactly when they will all be processed, and thus when everything they constitute—which is everything—will cease (Lloyd 2006).
This is stimulating stuff with much to pursue, but this is not the place to do it. I have turned to it in this coda because—no matter how these explanations play out—they tell us something very important about system. It persists. If the turn to information and computation signposts a change—especially a paradigmatic change—it is not because everything has changed or changed in the same way. Many ideas have been altered or replaced, but one of the primary forms in which we think them—the genre that has played such a central role in shaping our knowledge—is still recognizably present. System saturates this volume as thoroughly as it did the Enlightenment—the word appears over 550 times in those 856 pages. How it is being used, and thus how it mediates knowledge, is changing, but it has been holding its central position in the history of mediation.
I wrote this book to explore that persistence. In doing so I hope I have shed light on how and why it has persisted and on the different ways it has shaped and continues to shape knowledge. Clarifying those continuities and discontinuities can also make system of use to others as a marker of change in any of the areas of knowledge it has informed. The computational universe provides a telling example with which to conclude. “If I have one new message to convey in my book,” Seth Lloyd declared in an interview, “it’s that the universe is a system where the very details and structures in it are created when quantum bits de-cohere.”10 The primary continuity here is that system remains, as it was for Newton, the genre of choice—a genre still adequate to the universe. The discontinuity from Newton to Lloyd and Wolfram lies in the details of structure.
Structure is also where the discontinuity lay between Newton and his predecessors. The author of the Principia, according to Edward Grant in a passage I cited earlier,
regarded his treatise as if it were revealing the mathematical structure of physical nature, rather than as the mere application of mathematics to nature, as virtually all previous natural philosophers would have perceived it. (Grant 2007, 313, emphasis mine)
A structural change of that severity had, not surprisingly, a domino effect on system itself. After Newton, Stephen Wolfram observes, “systems could only meaningfully be described by the mathematical equations they satisfy” (Wolfram 2001, 860). That is why, in a nutshell, Newton rejected his own first draft of Book 3 of the Principia, opting instead for a more mathematical account of the “System of the World”—a system described in terms of an equation for measuring the force of gravity.
To arrive at that equation, Newton set in motion the interplay of the calculus and system—of the dividing of wholes into parts and the assembling of parts into wholes. The result was a System of the World that proliferated systems in the world, giving shape, as I have described, to new kinds of knowledge. By changing the structural details of their System—information overwriting mass and energy per Lloyd and the foregrounding of algorithmic iteration in Wolfram—the architects of the computational universe are not only reshaping their knowledge systems but are also repositioning them. For “fundamental reasons,” Wolfram claims, “systems, rather than traditional mathematics, are needed to model and understand complexity in nature” (860).
System and math are still at play here, but whereas Newton foregrounded mathematical principles—those are the first two words of his title, while system trails in Book 3—this restructuring puts system forward as the primary mode of understanding. The twist is that this is not really a turn from mathematics but a transfer of function; systems, in a sense, now do the math. Engaging system as a genre can clarify what has happened. As a historical grouping of features, the genre of system changes by changing the mix, and, starting with Darwin, the algorithmic has gradually become a featured feature of system. System is now coming to count in a new way.
Repositioning and repurposing system reconfigures, in turn, what I have been calling our knowable spaces. When Newton brought system into his treatise, choosing it as a primary form of guesswork, its subsequent popularity in the form of Newtonianism transformed the knowable spaces of moral philosophy, scaling them up into the Master Systems of the “human sciences.” And when systems changed position again, traveling into other forms such as the essay, knowledge was again reshaped, yielding the modern disciplines. If system is now securing a new position in the computational universe—one in which it generates the world it helps us to know—then we may be entering a new chapter in the shaping of knowledge with system in a newly performative role.