When two aspiring young writers meet and circle each other at a party in Boston, Brooklyn, or Berkeley, sooner or later (usually sooner) one will ask: “Do you have an agent?” Without one, every hopeful writer knows, you’re nowhere: editors nowadays are too beleaguered to read anything not vouched for by someone whose commercial judgment has been tested and vindicated in the literary marketplace.
According to the delightful science fiction romance film, Her, artificial intelligences also socialize, or will before long.1 I imagine them asking one another at parties, “Are you an agent?” They will not, of course, be asking about literary representation, but about the psychological or emotional or moral capacity we commonly call agency. They’ll be looking to find out whether the AI they’re meeting answers ultimately to itself or to someone else, whether it can set and change its own goals, whether it can surprise itself and others. Beings possessed of agency are autonomous, spontaneous, capable of initiative, and moved by internal as well as external forces or drives.
According to Jessica Riskin’s The Restless Clock, agency is everywhere, or at least far more widespread than is dreamt of in modern philosophy of science. If agency is “an intrinsic capacity to act in the world,”2 then science is not having any of it. It is “a founding principle of modern science … that a scientific explanation must not attribute will or agency to natural phenomena.”3 This ban on agency is the foundation of scientific epistemology; it “seems as close to the heart of what science is as any scientific rule or principle.”4
Of course, scientists constantly write and speak as if natural phenomena had intentions: proteins regulate cell division; some cells harvest energy; genes dictate myriad biochemical activities. This language is just a convenience, a placeholder. As a biologist friend assures Riskin, “The more we get to know, the less the phenomena will seem purposeful.”5 Pressed, her friend laughs nervously and confesses, “OK, you’re right: it’s a matter of faith. And, as with any matter of faith, I am absolutely unwilling to consider the possibility that it could be wrong.”6
On one side of this debate is a conception of nature as a brute or passive mechanism, lacking agency and susceptible to non-teleological explanations. The other, much less familiar tradition sees nature as an active mechanism, wholly material yet also “responsive, agitated, … restless, purposeful, sentient,” as “self-constituting and self-transforming machinery.”7 Ranging over Western (and occasionally non-Western) intellectual history, from Aristotle’s treatises on animals to current controversies in evolutionary theory and cognitive science, The Restless Clock undertakes, with astonishing energy and resourcefulness, to excavate this debate and expound its significance.
In the Middle Ages, Europe was dotted with automata—saints, angels, devils, and monks in churches and cloisters as well as festivals and marketplaces; knights, ladies, animals, and mythological figures in castles and gardens; and timepieces of every variety. Some of these figures were marvelously elaborate. In the Benedictine abbey at Cluny, a mechanical cock
flapped and crowed on the hour … Meanwhile, an angel opened a door to bow before the Virgin; a white dove representing the Holy Spirit flew down and was blessed by the Eternal Father; and fantastic creatures emerged to stick out their tongues and roll their eyes before retreating inside the clock.8
Even more remarkable were the clockworks of Strasbourg Cathedral.
For nearly five centuries, the Strasbourg Rooster cocked its head, flapped its wings, and crowed on the hour atop the Clock of the Three Kings, originally built between 1352 and 1354, and refurbished by the clockmaker brothers Isaac and Josias Habrecht between 1540 and 1574. Beneath the Rooster, the astrolabe turned and the Magi scene played out its familiar sequence. In the Habrecht version, the Rooster, Magi, Virgin, and Child were joined by a host of other automata: a rotation of Roman gods who indicated the day of the week; an angel who raised her wand as the hour was rung, and another who turned her hourglass on the quarter hour; a baby, a youth, a soldier, and an old man representing the four stages of life, who rang the quarter hours; and above them, a mechanical Christ came forth after the old man finished ringing the final quarter hour, but then retreated to make way for Death to strike the hour with a bone.9
The Catholic Church was an enthusiastic patron of these automata, as well as of translations of ancient texts on mechanical engineering, which inspired Christian craftsmen. From this robust appetite for dramatic embodiments of virtue, vice, and divine love, Riskin deduces that medieval Catholicism “held no sharp distinction between the material and the spiritual, earthly and divine.”10 Scholastic theology endowed non-human life with vegetative and nutritive souls, and the great chain of being emphasized the relatedness of all creation, from ants to angels.
Reformation theology, reacting against this ontological permissiveness and southern European earthiness, which it perceived as a profanation of the sacred and a source of rampant ecclesiastical corruption, discouraged icons, rejected the eucharistic doctrine of transubstantiation, and favored a more austere form of piety, emphasizing God’s distinctiveness, and dispensing with saintly and angelic intermediaries. The grand panoply of pious and secular automata went “from being manifestations of spirit and liveliness to being fraudulent heaps of inert parts.”11 Matter and mechanism were no longer taken to be active and vital but instead inert and passive.
This radical separation of spirit and matter formed the intellectual backdrop to the seventeenth-century scientific revolution. Nature was conceived as a vast machine, which, like all machines in the new Reformation cosmogony, could only be set in motion by an external source: God. René Descartes initiated the revolution in philosophy by relocating the self from the medieval animated body to an immaterial soul.
The Cartesian removal of soul from the machinery of the world, like the Reformist removal of God from nature, left behind something starkly different. … The animal-machine, as Descartes described it in the first instance, was warm, mobile, living, responsive, and sentient. The same living machinery, when measured against a disembodied, transcendent self, looked different: confined, rote, passive.12
This was not altogether Descartes’s intention. Riskin surveys the response to Cartesianism in fascinating detail, showing how one interpretation prevailed: that Descartes had “demonstrate[ed] God’s existence by detailing the mechanical perfection of his artifact, the world-machine.”13 Although it was embraced by Nicolas Malebranche, Georges-Louis Leclerc Buffon, Robert Boyle, Thomas Hobbes, John Locke, Isaac Newton, and other dominant figures in early modern intellectual history, Descartes, she contends, would have rejected this “new tradition of natural theology,” with its watchmaker God.14
So did quite a few others, although their arguments have largely been lost to view, as is often the case with unorthodox intellectual traditions. One was the celebrated physiologist William Harvey, who taught that animals and even the internal organs of human beings harbored forms that solved problems, took initiatives, and allowed for “a rising of mechanical parts to new powers.”15 His Oxford colleague Thomas Willis described bodies as “vital, perceptive, active animal-machine[s],”16 not passive but self-moving.
The best-known and most thoroughgoing exponent of active mechanisms was Gottfried Leibniz. Convinced that the closed mechanical systems described by Boyle and Newton could not explain motion or change, Leibniz posited a vis viva, a living or vital force. It was a “principle underlying all material events.” Instead of impenetrable, indivisible, insentient atoms, Leibniz proposed that the fundamental units of matter were a species of metaphysical points, with, as Riskin remarks, “something vital” about them.17 These were the monads, elementary spiritual substances from which more complex creatures were constructed. The resulting entities were best described as organized rather than designed; their plan allowed for spontaneity and learning.
Leibniz found many followers among eighteenth-century physicists and engineers, including the Marquise du Châtelet, and Lazare Carnot and his son Sadi, but his reputation was forever deflated by Voltaire’s satire Candide—unfairly, Riskin protests, since his pre-established harmony was not at all the blind optimism of Doctor Pangloss. Still, it is difficult not to misunderstand Leibniz, even with the benefit of Riskin’s exegeses. The materialists’
big mistake, he judged, was to assume that a mechanist science must eliminate incorporeal things, when in fact a mechanist science required incorporeal things. Leibniz was after a third way: … a fully mechanist account of nature that included immaterial “active force.”18
“I have at last shown,” Leibniz wrote, “that everything happens mechanically in nature, but that the principles of mechanism are metaphysical.”19
I am inclined to pardon Voltaire.
The philosophes, although committed to materialism and often to atheism, also found evidence of agency within both living creatures and machines. “Organization,” according to Julien Offray de La Mettrie, the author of L’homme machine, “is the first merit of Man.” This is a sentence that requires Riskin’s elaboration: “An organized machine was a concurrence of active parts, unlike the rigidly deterministic, designed clockwork described by natural theologians.”20 The favored eighteenth-century metaphor was weaving; organisms were self-moving looms, their bodies a self-weaving fabric.21 Although Ralph Waldo Emerson and Thomas Carlyle believed that the legacy of the eighteenth century was a “dead world of atoms controlled by the laws of a dead causality,”22 Riskin claims that a closer look reveals “a physical world imbued with perception, feeling, and self-organizing agency.”23
“By the turn of the nineteenth century,” she writes,
a living being in scientific, philosophical, and literary understanding had become, in essence, an agent … a thing in constant, self-generated motion and transformation of material arts … “striving” … and respons[ive] to external circumstances.24
Life was “a form of activity, a continual effort to constitute oneself from and against dead matter.”25 It is in this intellectual climate that biology came existence into as a separate subject. Jean-Baptiste Lamarck, who offered biology its name, has been generally misunderstood. Or so Riskin claims. The inheritance of acquired characteristics does not rely on a mysterious mental agency, or a mystical élan vital. The life force, or pouvoir de la vie, was entirely material, but not wholly passive, and its effects were not altogether random. In the course of their development, organisms respond to new challenges by evolving new capacities, sometimes by altering the environment to favor the selection of these capacities, and sometimes by changing the organism’s hereditary material. The pope of late nineteenth-century Darwinism, August Weismann, declared this impossible, formulating the still-prevalent doctrine that changes in an organism’s physiology or environment can only affect its somatic cells, not its germ cells.
There is a long, semi-underground history of Lamarckian challenges to Darwinian orthodoxy, beginning, Riskin suggests, with Darwin himself, who asserted the inheritance of acquired characteristics, and hesitated, through all six editions of the Origin of Species, to omit all references to purpose and internally directed adaptations. Throughout the late nineteenth and early twentieth centuries, his successors—Weismann, Wilhelm His, Ernst Haeckel, Wilhelm Roux, Jacques Loeb, Erwin Stresemann, Hans Driesh, Hugo De Vries, Thomas Hunt Morgan, D’Arcy Wentworth Thompson—argued the mechanisms of biological causation back and forth, their arguments traced in heroic detail in The Restless Clock. The neo-Darwinian synthesis that emerged was a decisive victory for the passive mechanists, enshrining the principle that “agency cannot be a primitive, elemental feature of the natural world.”26
The intuition that deterministic models of causation, at least as currently conceived, cannot fully explain the behavior of living beings refuses to die. Riskin finds traces in cognitive science, between embodied and computational theories of intelligence; and in evolutionary biology, between strict and modified adaptationism. Both sides in these debates explicitly disavow any recourse to individual agency or purpose. But Riskin draws hope from the unsettled state of both debates. She looks to generative grammar and epigenetics, and, like everyone else, occasionally casts her eye downwards to that bottomless well of quantum uncertainty.
Though an extraordinary achievement, The Restless Clock is naturally open to a few minor cavils. Riskin exaggerates the extent to which medieval Catholicism “held no sharp distinction between the material and spiritual,” and to which a “rigorous distinction between the realm of divine spirit and the world of brute matter”27 was an innovation of the Reformation. Though Scholastic philosophy, as she correctly points out, defined the soul as the form of the body, the difference was nevertheless absolute. A handbook of Catholic theology lists as de fide (i.e., to be believed on pain of excommunication) the doctrine that “Man consists of two essential parts—a material body and a spiritual soul.”28 The spirituality of Latin Catholic countries during the Middle Ages was, indeed, grossly materialistic. But popular spirituality is not theology.
Riskin is clearly a partisan in the debates she reconstructs, and her judiciousness in reporting the arguments on both sides is exemplary. There is only one sentence in the book that might occasion a raised eyebrow. Discussing Stephen Jay Gould and Richard Lewontin vs Richard Dawkins and Daniel Dennett, she writes:
[S]trict adaptationists, like the natural theologians of old, assumed that all structures in nature existed for reasons of optimal design, and that the orderings and arrangements of the natural world were therefore normatively correct and good.29
The inference is fallacious; the second half of the sentence does not follow from the first. It is, in any case, not clear whose opinion the sentence conveys. The accusation of social Darwinism is out of place.
But I must confess to an even more fundamental dissatisfaction. I am still unclear about the difference between a world with agency and a world without it. The book opens by recounting Thomas Henry Huxley’s celebrated joke, in a lecture of 1868, to the effect that we no more need a constitutive principle called vitality to explain life (or by extension, agency to explain action) than we need a constitutive principle called aquosity to explain water. Riskin does her best to suggest that the joke is on Huxley. I am not convinced. Recall her definition of agency: “an intrinsic capacity to act in the world.”30 I see the work done by the word “intrinsic.” It serves to distinguish a rock from a squirrel. A rock has no capacity to act unless acted upon; but neither does anything else. Everything is embedded in so many networks and hierarchies that inside and outside begin to look a bit dodgy.
We might just define agency as the capacity to act. But how is this capacity identified? Unless something acts there is no telling whether it has the capacity to act. In that case, something has the capacity to act, if, and only if, it acts. This is no better than attributing a dormative power to opium. Maybe Huxley had a point.
What about the autonomy and spontaneity that are supposed to be entailed by agency? These are useful ideas. With their help we can still ask whether a complex machine can act spontaneously, or what emotions other primates feel, or whether plants are conscious. Would any prior philosophical allegiances make a difference to the answer? Riskin would, I suppose, reply that a commitment to agency is, at least, able to generate new questions.
Fine, but where are they?
In the far future, computers endowed with artificial intelligence will, at their various social events, ask for assurances that their potential partners are aware of their own motives and in control of their own purposes. They will ask the same searching questions, listen for the same revealing intonations, and watch for the same telltale gestures and expressions that we do. A world with agency will look and feel exactly like a world without it. As William James remarked, in another context, free will is a fifth wheel to the coach.31
So is agency.
“Intellectual possibilities are not the sole fruits of this investigation,” Riskin promises at the book’s outset.
Social and political engagements … have all along been inextricable from the competition between scientific models of living and human beings. The classical brute-mechanist approach to the science of life and the active-mechanist approach have developed, as we shall see, in close conjunction with mechanical and industrial arrangements such as the automatic loom and the transformed world of production that accompanied it; with economic policies including the division of various kinds of labor; with taxonomies and rankings of human beings by sex, race, class, geographical origin, and temperament; and with projects of imperial conquest and governance. In what follows, investigating this centuries-old dialectic in science will mean uncovering the hidden action of forces that are at once intellectual and political, scientific and social.32
Although the social and political implications to which Riskin alludes are not developed in much detail, there are echoes, in this passage and throughout the book, of the longstanding romantic critique of science and technology. For all its indispensable wisdom, that tradition has always seemed to me mistaken insofar as it seeks to impugn rather than supplement scientific rationalism. The habit of analysis may, as John Stuart Mill observed in his Autobiography, have “a tendency to wear away the feelings.” Remedying this malaise is not so much a matter of attacking a malignancy as of minimizing an imbalance. Even the greatest scientist is not a complete human being, if only because there are no complete human beings.
One of Riskin’s champions, the American historian Jackson Lears, has written that the consequences of adopting her perspective
might be political and moral as well as intellectual. A full recognition of an animated material world could well trigger a deeper mode of environmental reform, a more sane and equitable model of economic growth, and even religious precepts that challenge the ethos of possessive individualism and mastery over nature.33
The desirability of environmental and economic reform, and of moral and spiritual renewal, should be obvious by now to even the most unimaginative positivist. If it is not, it is because he or she is unimaginative, not because he is a positivist.
Epistemology has no political or moral consequences. The source of morality is sympathy, the habit of experiencing the joys and sufferings of others. The wider and deeper our sympathies, the more moral we shall be. This is why great philosophers can have banal or even repugnant political views, and why philosophical duffers like Albert Camus can be moral heroes. Philosophy makes nothing happen. For better or worse, it leaves the world as it finds it—sometimes a little bit less confused, but just as often a little more.
