Catherine Malabou. Morphing Intelligence: From IQ Measurement to Artificial Brains. Trans. Carolyn Shread. New York: Columbia University Press, 2019. 198 pp.
Review by N. Katherine Hayles
17 May 2019
When was the last time you heard a prominent male theorist admit he was dead wrong? My answer: about the same as the last time I heard those words from Donald Trump. Yet to her credit, that is exactly what Catherine Malabou does in Morphing Intelligence: From IQ Measurement to Artificial Brains (2019). In this insightful analysis of the long-standing resistance of philosophy to quantifications of intelligence, Malabou calls for fresh perspectives on what intelligence means and how it functions and for new rapprochements between the humanities and sciences. Her prose is precise, her research carefully articulated, and her conclusions realistic yet hopeful.
Before arriving at the revelatory moment indicated above, she traces three major paradigms in how intelligence can be measured and understood. The first begins at the dawn of the twentieth century, when Francis Galton and Alfred Binet established a quantitative intelligence measure that would later become standardized as the Stanford-Binet test. The second launches in the 1970s as epigenetics begins to modify ideas about genetic inheritance, giving exchanges with internal and external environments new importance in the regulation, expression, and inhibition of genes. The third is the development of artificial intelligence, especially the neuromorphic architectures of recurrent neural nets and neurosynaptic chips such as IBM’s SyNAPSE.
In analyzing philosophy’s resistance, she playfully evokes the “tustudo” (tortoise) Roman military formation, in which the front row soldiers and those on the sides held their shields upright and those in the middle over their heads, thus creating a turtle-like shell that protected the mass from incoming arrows. The first row defense in philosophy objects that psychological quantification is “nothing but policing” (p. 42); the second row argues that intelligence testing is a method of biopolitical control; and the third row attacks testable intelligence as necessarily machinic, understood as inferior to human creativity and unpredictability. Although not necessarily disagreeing with these critiques in their times, Malabou argues they are now obsolete. “I could stop right here,” she writes, having concluding from the epigenetic revolution that the long-standing conflict between intellect and intelligence, philosophy and psychology, has been resolved (p. 81).
She doesn’t, of course, going on to discuss the implications of ANI (narrow artificial intelligence), AGI (artificial general intelligence), and ASI (artificial super intelligence). Using primarily John Dewey and Piaget, she recuperates a sense of intelligence as an effective means of adjusting to the environment (which thus implies plasticity), and collective intelligence as the wisdom of the community in adapting to larger-scale changes. She riffs on the interplay between habit, as the solidification of experience, and intelligence as the plastic ability to adapt to change and novelty. “Without habit, intelligence has no past. Without intelligence, habit has no future” (p. 101).
In a “Postscript to the English Translation,” she argues that the dangers of strong AI “lies not with the machines but with humans” (p. 151), acknowledging that “the burning question today is humanity’s possible loss of control to machines” (p. 153). She sees this loss as inevitable but argues it can be interpreted in worse or better ways. The best, she suggests, is a view that sees it as an opportunity for “the democratic construction of collective intelligence” (p. 153), that is, what I call a cognitive assemblage that includes both human and computational actors. To ensure this outcome, it will be necessary to have strong, informed regulatory agencies such as the proposed European agency on robotics and artificial intelligence. “Regulate to leave us free” (p. 161).
In a US context, there is less cause for optimism that this form of resistance can be effective, given the influence of corporations over regulation during the Trump administration. Regulations are also notorious for leading to unanticipated consequences, especially in complex human-computational systems. She also dismisses too hastily transhumanism as “hypernarcissism” (p. 163), although enhancement of humans through implants and other cognitive devices is sure to be part of the future, in which the boundary between artificial intelligence and intelligence augmentation (of humans) becomes increasingly porous. Despite these missteps, Morphing Intelligence is a provocative analysis of where we have been and where we are headed as we both suffer and benefit from the realization that humans are no longer the only symbolic cognizers on the planet, and indeed not necessarily the most powerful.