• The Framing Phase: Talking to Viktor Mayer-Schönberger

    When you need to make an important decision, why should you focus on formulating good options? When you start from conventional assumptions, why might even the most skillful strategizing not help much? When I want to ask such questions, I pose them to Viktor Mayer-Schönberger. This present conversation focuses on Mayer-Schönberger’s book (co-written with Kenneth Cukier and Francis de Véricourt) Framers: Human Advantage in an Age of Technology and Turmoil. Mayer-Schönberger is the Professor of Internet Governance and Regulation at Oxford. His research focuses on the role of information in a networked economy. He has published 11 books, including Big Data (co-written with Kenneth Cukier, and translated into more than 20 languages). He also is the author of over one hundred articles and book chapters. In 2014, Mayer-Schönberger received a World Technology Award in the law category for his work.


    ANDY FITCH: Could we start with Framers not providing a catalog of genius revolutionary inventors, so much as probing “the secret ingredient in the stream of continuous, incremental improvements across human history”?

    VIKTOR MAYER-SCHÖNBERGER: When we began thinking about this book several years ago, we first thought to focus on humans’ decision-making processes, and compare these to how machines make decisions, and ask whether (or when) machines might surpass humans in this capacity. But we came to realize two basic points. First, machines might eventually be able to pick more rationally than humans between two specified options. But before you can decide, you need to have some options to decide among. That’s where the real power of human cognition lies — coming up with pretty good options pretty fast. Computers still can’t do that.

    Our second basic point then stems from the fact that, if you start from default conventional options, even very skilled decision-making won’t help much. That further clarifies the foundational importance of eliciting good options in the first place. Over the past 20 years or so, cognitive psychologists and anthropologists increasingly have recognized this cognitive capacity as something that differentiates us from the rest of the animal world, as well as from computers. And as we exercise our cognitive muscles, as we learn to wield this tool of cognitive framing, we tend to get even better at it. So this book focuses on how improved framing can help us come up with better options, leading to better decisions, and to better lives.

    For a quick definitional account, how do frames help us to conceptually organize an otherwise overwhelming stream of stimuli, and how does agile application of frames improve our capacities when it comes to taking practical action?

    Well of course we humans are lazy [Laughter]. We have figured out well how to rely on tools that help us with our laziness. Our thinking, for example, relies on mental models, models representing reality while reducing the level of complexity. This simplification of reality highlights certain elements that our mind deems most important, and disregards much of the rest. This method of focusing allows us to deal with the onslaught of sensory impressions we receive. Mental models provide an extremely useful filter, offloading vast quantities of information-processing that otherwise would consume every instant of our attention. Mental models, in this regard, operate as a crucial feature (not a bug) within our conscious experience.

    Of course, if we focus on something trivial, if we ignore something crucial, that can cause significant problems. But if we can efficiently focus our attention and energy on what matters, while more or less ignoring the constant bombardment of most sensory data, we’ll be much better off.

    To take one quick step back, when I think of actionable frames playing an essential role in human life, I think of accounts from Friedrich Nietzsche, William James, Hans Vaihinger, Ludwig Wittgenstein. Framers acknowledges such predecessors, while suggesting that they relied on “all theory, no empirics.” Here could you describe how a late-20th-century “quiet transformation in understanding how people think” has shaped this book?

    Early research in the cognitive sciences often started out from very behaviorist concepts of human thinking. As experiments failed to confirm the ubiquity of such concepts, an alternative approach was needed. In a way, serious human-cognition research based on mental models began right there. It then gained traction as better experiments, and later better equipment (such as fMRIs), were used to capture more of the essence of human decision processes.

    In terms of these processes, Framers presents causality, counterfactuals, and constraints as elemental components of constructive frame application. Could you describe how humans’ dexterous applications of causal mechanisms allow us to address novel situations, and how these inductive capacities have taken our species far beyond bearing passive witness to lived experience — both at an individual level, and at a broader social level?

    Humans have a kind of implicit understanding that the world operates through cause and effect. Our minds almost can’t help linking effects to causes. This allows us to experience a particular situation, and abstract from it a more generalized causal mechanism — which we then can apply to related situations. Already this means reducing complexity, looking at a lot of trees and seeing the forest (and remembering that more generalized concept of a forest), rather than the details of every single tree trunk. We now can take this template and apply it, transferring our knowledge and understanding to novel contexts.

    By contrast, if a dog learns that giving a paw means getting a treat, it will eagerly repeat this gesture. But the dog will never abstract and generalize that if it does similar tricks, like performing a somersault, that probably also will result in treats. The dog might learn this likely consequence of performing somersaults as well, but never will step back and think: Now how else can I be cute?

    Now in terms of human communication fostering transferability of causal models, it turns out that even very early languages enabled us to create abstractions. You could think of these as some of our foundational frames. Steven Pinker, for example, argues that certain metaphoric capacities provided a key element enabling humans to differentiate ourselves from other species. Metaphors assist us not only in abstracting from immediate experience, but in expressing extracted ideas through language. Visual representations like drawings and written language further transcend space and time, passing on human knowledge to future generations. The world’s great myths, for example, might foreground great stories. But behind these stories themselves you can find humans passing on effective abstract frames.

    Passing down knowledge this way even improves our own understanding. When required to explain a causal mechanism to somebody, we actually gain a better grasp on that mechanism ourselves. This provides our species with a powerful incentive to explain to others what we have learned, and that beautiful evolutionary impulse makes us crave even more understanding. We like to explain because it benefits us to do so — while also of course benefitting others.

    According to Framers, counterfactuals then allow us to anticipate potential changes in our environment, prompting us to envision strategic modifications, broadening our range of practical options. Could you sketch the catalyzing force that counterfactuals bring to a wide range of human projects, from child’s play to the modern scientific method?

    I’ll start by posing a tiny riddle as an example. Suppose you have two unmarked pitchers (one five quarts, one three quarts) that you can fill with water. You have an unlimited water supply, and you want to fill one pitcher with precisely four quarts of water. How do you go about this?

    To answer that question, your mind needs to imagine multiple steps, pouring water from one pitcher to the other, and these kinds of things. Your mind needs to engage in counterfactual thinking, picturing a world that doesn’t exist. We constantly ask and seek to answer these “What if” questions — playing the game of life just a couple moves ahead, speculating on how one development might lead to the next. This counterfactual thinking occurs basically as a form of dreaming, or of imaginative projection.

    It also happens within a set of constraints. When I think about those pitchers, I don’t imagine divine intervention solving this problem for me. My imagination incorporates certain laws of physics and rules of practicality. I create counterfactuals carefully tailored to how these kinds of situations tend to play out. My dreaming has some discipline to it.

    We actually start doing all of this as toddlers. We might think of toddlers as engaged in pretend play, as trying out certain social dynamics. But toddlers’ play accomplishes much more than that. It activates and trains their cognitive muscles. It coaxes them to imagine in a disciplined way, coming up with counterfactuals and constraints that fit whatever fantasies or pretend worlds they’ve devised. It allows them to practice precisely what this book defines as framing. We start framing early on, around one year old, and we never stop. We might have the most time to practice this playful type of framing as toddlers and young children. But even adults of course routinely pose counterfactuals to themselves at work, or in their leisure time.

    As we read a novel or watch a movie or play a video game, we constantly ask “What if” questions: “What if this character is the killer? What if this person falls in love with that other person?” So while I perhaps unfairly described humans as lazy, I don’t think of us as idle couch potatoes. We regularly train our cognitive muscles, much like we might go to the gym. Through such practice, we can further improve on our decision-making — again first by coming up with better options, and then by making consequential decisions.

    So here again, Framers led me to think of well-honed constraints not as forcing us to accept the world before us, but as keeping our counterfactuals actionable, as reducing our cognitive load in ways that channel momentum towards constructive change.

    That’s right. Constraining our counterfactuals further increases the efficiency of our decision-making. It helps us focus on those options most likely to succeed. Several basic elements help to make this happen. First, effective constraints tend to direct us towards relatively small (rather than huge) changes. Our book offers a minimal-change principle, since small changes are significantly more likely to happen, and much easier to control.

    Here effective constraints also adopt a principle of malleability, directing us to changes that we humans can initiate and shape and influence. Again, if we need to rely on divine intervention, or on evolutionary adaptation to catch us up to this immediate moment, the right decision will likely arrive too late. So effective counterfactuals point to where we ourselves can affect real-world change. Actionability, in turn, depends on effective constraints quickly filtering out less useful counterfactuals, helping us foresee good options and make good decisions.

    For one quick example, we can consider the Cuban Missile Crisis. This high-stakes decision-making process has been summarized and analyzed countless times. But one key element often lost in these accounts involves the US military brass around President Kennedy, upon discovering that the Soviets had sent nuclear-tipped mid-range missiles to Cuba, suggesting some retaliatory responses like invasion or bombardment. Kennedy and his immediate staff worried that these sorts of obvious options started from too tight of a constraint.

    When you impose tight constraints, you might only come up with two or three conventional choices, greatly limiting the options space. In such situations, you can significantly improve your decision-making by first saying something like: “Maybe we don’t need to destroy the missiles outright. Maybe, since these missiles still require additional supplies, fuel, and support staff, we can focus on stopping further shipments of materials and personnel to Cuba. Maybe if we impose a blockade on Cuba, we can prevent these Soviet missiles from ever becoming operational. Maybe, at the same time, we can provide the Soviets with a kind of tit-for-tat face-saving win of some sort.”

    Looking at the situation this way, Kennedy’s team realized that US nuclear missiles stationed in Turkey didn’t offer much strategic advantage. So they said: “You know what? We could give these up. We won’t lose out on much. And we’ll change a zero-sum game into something more like a win-win situation.” In this particular case, that loosening up of constraints to introduce a different type of counterfactual worked — and quite possibly saved the world.

    That also takes us closer to reframing, which again involves deliberation, but while propelled a bit more by the occasional stroke of insight. Could you outline various ways in which we might reframe by drawing on a repertoire of frames, or by repurposing frames from seemingly unrelated fields, or by reinventing frames when no known mental model can suffice?

    This book’s conception of framing covers two overlapping but distinct mechanisms. One mechanism involves thinking within a mental model, such as a counterfactual honed by realistic constraints. This particular kind of frame can be very efficient and useful and powerful. It works especially well when radical changes in circumstances seem unlikely. It allows for both familiarity and for some creativity.

    But when more radical changes do take place, an established mental model just might not work anymore. Shifting circumstances might break this mental model, and demand a new model. Now, if you already have many effective models in your repertoire, that can help a lot. That’s something we all should strive for. Here I think, for example, of a school Elon Musk helped to start, whose ideas are now utilized in an online educational service called Synthesis. Synthesis tries to offer a broad repertoire of mental models to young kids, by introducing foundational concepts like evolution, gravity, homeostasis, game theory and the Nash equilibrium, or even just cooperation. It prioritizes these archetypes, basically, of powerful mental models. It operates from the principle that the next generation needs a very solid repertoire of such mental models. That principle fits well with our own research findings.

    Our book also makes the argument, however, that while it’s useful to build up a repertoire of mental models, that’s not sufficient. Because what happens when no available model applies? What about when nothing in your repertoire fits? In preparation for that kind of situation, we need not just to have absorbed prominent models, but to have cultivated a taste for cognitive foraging. We need to already have become willing to rummage around, to look in other fields and other sectors, to let curiosity inspire us to keep exploring and experimenting with alternate models (not just those that have been made to seem all-important).

    And then sometimes, after borrowing models from a couple of different fields, and sensing these alternative applications not working either, you really do just need to think afresh. In the book, for example, we discuss the early-20th-century German chemist Otto Hahn studying the effects of bombarding uranium with neutrons. Hahn saw a sudden surge in energy output. He couldn’t make sense of this. He didn’t understand these results because his chemistry frame had no explanation for them. So he sent a letter to his good friend and former colleague, the physicist Lise Meitner. Meitner quickly recognized: “I know what that is. That’s fission. That’s a physical reaction, not a chemical one.” Meitner wrote back to Hahn and also published her findings. Hahn immediately realized she had comprehended what was happening, while he had not. But nonetheless, Hahn received the Nobel Prize for this discovery, whereas Lise Meitner, nominated 48 times for the Nobel Prize, never did. The scientific community clearly had a gender problem.

    Here could you also make the case (perhaps especially for professional or political mindsets in which nothing significant seems to have been accomplished without a heroic reinvention of the status quo) on why “it is… better to reframe infrequently”?

    Reframing has two big disadvantages. This type of high-risk proposition rarely succeeds, and you don’t get better at it even when you do have occasional success. Again with counterfactuals, even watching a movie can help to sharpen your constrained thinking within internally consistent frames. We can improve significantly when it comes to framing, but we don’t get much better at reframing. So given the relative likelihood of success, if you can solve a problem within an existing frame, this will be your best option, pragmatically speaking.

    Unfortunately, as you said Andy, we live in a world where “radical disruption” has this magic ring. It grabs our attention. We hear about Jennifer Doudna’s breakthrough with CRISPR Cas9 technology, and about Doudna winning the Nobel Prize, and we all want to be Jennifer Doudnas. But much better life strategies exist. We can focus on improved frames for our daily decisions, rather than hope to win the lottery with a radical reframing. If I have one wish for this book, it would be persuading readers to use their cognitive-framing muscles far more frequently and deliberately and also creatively — instead of searching for some radical reframing strategy to remake their lives.

    If you look back to the late-1980s, for a while everybody talked about incremental change outperforming perfectionism, and the Japanese method, and all of that. But in more recent decades, particularly with the rise of the Internet and social media, people who want to portray themselves as innovative and influential present the image of a radical disruptor — even though all economic indicators tell us that over the past 15 years, rates of innovation have dropped off from where they were in the 1990s, itself a decline from where they were in the 1960s and 70s. As the speed of innovation has in many ways slowed, we’ve built up these portraits of heroic disruptors, again even though incremental but meaningful change and choices would make much more sense for most of us.

    You mean both in terms of our individual and our collective well-being?

    Right. Just to come back to Elon Musk quickly, we won’t solve our environmental problems by becoming interstellar and sending a million people to Mars. What does solve human problems tend to be comparatively small but meaningful changes, like switching from incandescent light bulbs to LEDs, or installing timers so that our air conditioners don’t run all day, but instead switch on half an hour before we get home. If we can focus on these seemingly small-scale improvements, we still can innovate — while also strengthening our creative muscles through actionable situations with real consequences. We should never forget these framing capabilities every one of us has, even if we can’t help feeling starstruck by reframers.

    Could we now pivot towards a broader mindset of cognitive pluralism? Could you first make the case, perhaps on pragmatic rather than moral grounds, for cultivating such pluralism not as an end, but as a means?

    Well as you say, we don’t argue here for the moral value of diversity (which certainly may be worthwhile). We argue in favor of diversity as an instrumental value, as the means to a better end (again through better decisions). Mountain ranges provide a useful metaphor here, as Scott Page has suggested. If you climb to a mountain’s top and look around, you may know whether you’ve reached the immediate vicinity’s tallest peak. Yet you can’t tell whether you’re looking down from the overall mountain range’s highest point. And if you descended, then climbed up the next mountain, you’d get a slightly different perspective, but this would consume a lot of time and energy. It works far better to have various people (from very different locations, so to speak — cognitive locations) climb their respective mountains, then compare with each other. If we have different starting points, we’ll also end up with different points of view.

    Now, why do we need a broad spectrum of options? Well, if nothing changes, if our task and its challenges forever stay the same, then we can stick to the options already before us. But if the context changes, if our lived situation changes, if our environment starts changing quite dramatically, then we need not just more options — we need different options than just the familiar ones. At that particular moment, societal diversity pays a big decision dividend.

    No single frame holds a monopoly on truth. The dialogue between Otto Hahn and Lise Meitner shows us that sometimes you need mental models from chemistry as well as from physics. Or as the Cuban Missile Crisis example shows, you might need frames both of the military and of the diplomats. Only by drawing on a multiplicity of frames, and by eliciting a wide spectrum of options, can we optimize our response to fundamentally changed circumstance.

    So in what ways does cultivating such cognitive diversity almost inevitably mean proactively pursuing demographic diversity? And when considering cognitive diversity as a means rather than an end, why must robust institutional diversity welcome the frictions that come with perspective pluralism — rather than rushing to some more tepid social convergence?

    Cognitive diversity does bring a cost by creating societal friction. Societal friction does mean some inefficiency. So while the going remains easy, we might feel tempted to ask: “Why introduce any added friction here? Why overcomplicate something running so smoothly?” But if you fail to promote social diversity, then you leave yourself cognitively naked. When circumstances shift, your society will find itself unprepared. Without even realizing it, your society had bet on one particular future trajectory, but the trajectory has now changed.

    By contrast, when we actively promote cognitive diversity, even during times of relative calm, we prepare ourselves better for the eventual moment when some radical change does arise. And as this book tries to make clear, these moments can arrive much more suddenly than we might imagine. The 1930s offer a striking example. After decades of incredibly stimulating cognitive diversity, complacency had set in, followed by laziness. Society said: “Oh, we know what’s right, and we stick to what’s right. We don’t need all this extra cost from this extra friction.” By reducing its frictions and narrowing its focus in these ways, society lost its resilience, its depth of perspectives, and its ability to deal with powerful shocks.

    Pluralistic mindsets might seek to avoid absolutism in almost every form, but where and why must they respond vigilantly to intolerant frames? How should we, as a collectivity, go about tackling both wrong frames, and problematically “bad” frames?

    To a large extent, bad or even wrong frames don’t exist. At worst, we mostly deal with wrong framing, where somebody picks a mental model unfit for the circumstances and the goal. Typically, whether or not we have a good mental model depends on how well it fits the context. When I measure my garden’s dimensions, for example, I can safely assume that the Earth is flat. For practical purposes, I don’t need more precision than that. Incorporating this slightest of curvatures would only complicate my calculations. So real-world uses certainly do exist even for a flat-Earth frame. We can’t categorize this perspective as being a bad frame, even if it sometimes plays a part in bad framing.

    However, as you rightly point out, one huge exception exists. A frame that categorically denies the existence or the value of other frames, that stresses its exclusive claims to truth, that by its core argument would weed out the diversity and colorfulness of human framing, deprives us both on an individual and a societal level. It leaves us unprepared and vulnerable. It reduces our resilience. That we can’t afford — and therefore, this is the one type of frame we must deny.

    Numerous examples, from across the political spectrum, of course come to mind. And you likewise present certain advocates of a hyper-rationalist tech-centric approach as seeking to remove the biases and compulsions to which human framing seems forever prone. Yet here could you say more about why widespread AI implementation won’t absolve humanity from the need to frame, but will in fact make human framing all the more critical?

    Sure. I’ll first return to the idea that computers have little success when it comes to devising options — which we can characterize as the framing phase. Computing power allows for very complex calculations. But humanity’s real strength, our cognitive muscles, exert themselves by coming up with good options to begin with.

    From my perspective, this particular strength guarantees humans, for the foreseeable future, a central place in the ongoing evolution of cognition. We still provide the best options — the most suitable options for a given situation. Machines have become increasingly powerful tools, but they still require human direction. And for humanity to flourish in that role, we (as individuals, and as a society) need to keep honing and expanding our powers for eliciting the best and most fitting frames.

    By contrast, you present emotionalists as following their gut, rehashing ineffective frames, failing to find the grounded salience that allows for forward-thinking problem-solving propulsion. When confronting emotionalist appeals, why and how might you recommend not taking sides in some reductive binary, not compromising by blending two caricatured extremes, but maintaining our deliberative process of honing actionable frames where we can, and deftly reframing where we must?

    This emotionalist approach of responding to one’s gut definitely resonates today. So we need to figure out how to persuade emotionalists that they’re undervaluing their own problem-solving capacities, and selling themselves short. This isn’t about machines usurping human decision-making powers, or educated intellectuals lording it over everybody else. Effective human framing has never required a college degree. Framing is quotidian. Everybody has the ability to frame. Everyone also at times feels the impulse to just let their gut decide. But we all can succeed and keep improving as astute human framers.

    When we let our emotions override this capacity, I do see a real danger. And the book presents as most dangerous in this regard not the chaotic emotionalist constantly shooting from the hip, but the purported rationalist clinging to very tight frames, reinforcing a false impression that they simply permit no additional options. If you reduce your options space in this way, you might think of yourself as just being rational, structured, logical.

    Engaging somebody reliant on excessive constraints like this can get even trickier than engaging a more volatile type of emotionalist. Here we could compare, for example, a problematically impulsive person like Donald Trump, to those so ensconced in a conspiracy theory like QAnon that they actually consider themselves perfect framers. I’d characterize the QAnon tendency as even more dangerous today than a capricious public figure like Trump.

    Could we close then on your call to foster collective frame pluralism by “emphasizing conditions rather than rights”? While recognizing the need to protect individual liberties, why might you foreground even more “the concrete ability for humans to rise, and stand up against power fearlessly”? Why, only then, can we realize true cognitive diversity?

    Here the book to some extent resurrects the views of Judith Shklar. Today we often hear assertions such as: “This is my right.” I frequently encounter these kinds of statements, for example, when it comes to defining privacy protections within the context of information technologies. But all of these rights add up to naught if we don’t possess certain basic conditions, including a sense of agency, and a chance to shape our lived environment through constructive human framing. I don’t just need some authority to grant me official rights. I need to thrive in conditions allowing me to make effective use of my cognitive powers. It’s not enough to tell somebody: “You have the right to frame.” We need to create social conditions in which humans can frame in the most practical and creative senses imaginable, coming up with new options that address the pressing problems of our time.