Could AI Replace Diplomats?

By Shinri Furuzawa

New computer algorithms are developing personal intelligences and are capable of outperforming us at games requiring skills once thought to be specific to humans. For decades, computers have surpassed us at games which are primarily logical, syntactic, or mathematical, such as chess, or Go.

Now, however, a recent article in Science describes the Cicero algorithm which can win against humans at the board game, Diplomacy. Enjoyed by the likes of John F. Kennedy and Henry Kissinger, this is a game which requires intuition, persuasion, and deception. Cicero is able to discuss strategy, forge alliances, and carry out subterfuge and betrayal. It mimics natural human language in text conversations that entail negotiation with other players. The ability to observe and evaluate the trustworthiness of other players while convincing others of one’s own trustworthiness, and dealing with imperfect information, are key skills for actual human diplomats.

All things considered, one wonders how close AI could come to replicating the skills of a real diplomat and whether one day, AI could even replace human diplomats.

What Makes a Good Diplomat?

Former high-level American diplomat Robert Blackwill, suggested fifteen qualities which he thought essential for diplomats. Perhaps a third of these characteristics are inherent, and therefore irrelevant to AI, such as resilience to failure, or honesty. In other areas, such as analytical skills, attention to detail, or knowledge of history, AI already surpasses humans.

AI would, however, struggle in any area involving interactions that occur in person when the personal intelligences are especially vital. Diplomats must accurately collaborate, observe and evaluate others, and understand other people’s motivations while taking into account cultural, political, organizational and other differences. These trained professionals form mental models of their antagonists, and update them even unconsciously.

Diplomats are also skillful in interpreting non-verbal cues such as facial expressions, eye movement, and body posture. For decades, it has been common for diplomats to receive specific instruction on these interpersonal skills. While AI has made advances in interpreting non-verbal cues and information, it’s not quite there.

• Facial and emotional recognition: AI is already being used to recognize faces and monitor people’s facial expressions, for example, in airport security systems. The problem for affect detection algorithms arises, however, with the fact that facial expressions of emotion are not universal; the way in which people communicate their emotions can vary according to culture or the situation. AI also performs better at recognizing Caucasians over people of color, a further problem that may lead to racial profiling.

  If AI can’t yet read us well by looking at our faces, it does better at listening to our voices.

• Voice analysis: AI already has voice recognition and realistic voice generation. It can now also be used to detect patterns and characteristics in the voice that cannot be picked up by the human ear. Algorithms can predict psychiatric illness and other health conditions. By analyzing recordings of Vladimir Putin’s voice in February and March of 2022 during the ongoing war in Ukraine and comparing them to a recording of a talk he gave in September 2020, AI was able to detect stress levels 40% above baseline. While AI can collect such data, it must still be interpreted by humans and cannot—or at least should not—be used to predict human behavior. 

  AI capabilities may still be nascent in some areas, but they will only improve in the future.

Could AI RENDER Human Diplomats OBSOLETE?

Diplomacy may involve skills that we have long considered to be quintessentially human. I talked to Steven Siqueira, a former Canadian diplomat and chief of staff for several UN peace operations, and to Dr. Martin Waehlisch who leads the Innovation Cell in the Policy and Mediation Division of the UN Department of Political and Peacebuilding Affairs. I asked them what a diplomat does that an AI could never do. It seems to me, that it comes down to interpersonal intelligence.

Steven Siqueira - former Canadian diplomat and chief of staff for several UN peace operations

Martin Waehlisch - leads the Innovation Cell in the Policy and Mediation Division of the UN Department of Political and Peacebuilding Affairs

• Developing personal relationships: In Siqueira’s view, “You need personal relationships to get things done.”

  He gave the example of when he was tasked with establishing a UN mission in Sudan. Siqueira negotiated with a high number of separate stakeholders, which meant cultivating a myriad of different relationships. AI may be able to form analyses and identify requirements, but actual implementation is a human task. It would be extremely hard for AI to navigate the interface between personalities, and the intricacies behind each stakeholder’s position: their limitations and accountability whether it be to politicians, the military, civil society, or the media, all while working together towards a mutually satisfactory outcome.

Political scientist, Joseph Nye, would agree on the value of human relationships. Nye describes the importance of “soft power” as opposed to traditional “hard power” which relies on military or economic strength. He suggests that agreements and alliances today are fostered more through amicable relations, using tact and warmth, rather than aggressive tactics. According to Nye, even a smile can be a soft power resource. Diplomatic efforts need to be directed at citizens, not just governments, shifting to influence through likeability, attraction, and relationship rather than power—or at least in addition to—force, or coercion. As Waehlisch says, “The future is about soft skills… I was skeptical of emotional intelligence but I’m more and more convinced.”

• Innovative thinking: AI’s ability to think creatively and adapt to circumstances is also questionable. AI cannot respond in innovative ways if it is only drawing from the past. In the Diplomacy game, the chatbot is not creating anything new, it’s regurgitating based on percentages of success rates in past games.

  In the real world, diplomats think on their feet and rely on their training and experience to deal with new situations. This aligns with the last point on Blackwill’s list; diplomats must be quick to recognize opportune moments and know how to exploit fortuitous and unforeseen circumstances when they arise.

• Experience: In diplomacy, experience is crucial. Diplomats are trained through mentoring and vital skills are learned on the job. Blackwill listed learning from experience as an essential skill for diplomats, and as he puts it,“Would you hire a plumber who was academically well-versed in water distribution, but had never installed a pipe?”

What Role Does AI Have to Play in Diplomacy?

AI may fall short in personal intelligences, but it fares significantly better in linguistic and logical-mathematical intelligences. Siqueira and Waehlisch provided some insights into how AI is being used in diplomacy now, and how it could be used in the future.

• Generating text: Blackwill’s list of essential skills includes the ability to write and speak well, or linguistic intelligence. The latest reports on Open AI’s ChatGPT-3 attest to AI’s ability to converse convincingly with a human. It can engage in philosophical discussions, tell (bad) jokes, and debate political issues; it can also write and debug code, write college-level essays, and take tests successfully. Whether the task entails making an after-dinner speech or giving a presentation, AI can be programmed to tailor language, tone, style, format to match an audience. Many of the more mundane report-writing tasks performed by interns today could be carried out by AI. Diplomats will no doubt increasingly rely on AI for research.

• Mediation: AI could be used to support mediation. At the UN, for example, all mediated agreements are in a database. AI could easily draw upon the same language to mitigate similar situations that have occurred in the past. AI could scan and track different clauses—thereby providing valuable insights and perhaps helping to sustain peace efforts.

• Advisory roles: Computers are able to process and instantly retrieve exponentially more information than humans, enabling them to take over traditional advisory roles. Diplomats on the UN Security Council use their smartphones to find information or receive instructions rather than relying on advisors to whisper in their ears. Computer programs and algorithms are superior at assessing data and anticipating outcomes—important skills in negotiation.

• Targeting resources: At the UN, AI capacities in the form of data aggregators are already being used to analyze the press releases and communiques of all foreign ministries, allowing political officers to “mine the sentiment” on a given topic. Knowing which countries are most concerned about an issue enables targeted approaches—for example, by knowing which countries may be open to providing donor resources.

  Geospatial technology has recently made significant advances in providing “eyes in the sky.” These capacities entail data collection and analysis in fragile states which can improve monitoring and allow targeted humanitarian or peacekeeping efforts. It’s important to remember, however, that early warning doesn’t mean early action–political decision-making must still be done by people. Technology can’t fill this gap.

• Increased productivity: AI undoubtedly improves productivity. It offers internal solutions by tackling intrinsic systemic challenges with products aimed at automation and speed. External solutions enable closer human connections which results in inclusivity.

• Access: AI also enables dialogue. Many groups that once could not have been part of the negotiation process due to geographical remoteness, or that were simply not allowed at the table, can now be party to the conversation. Increased opportunities in terms of language and translation capabilities through TV and radio mining enable access to low-resource languages. Such outreach outflanks cultural and language barriers. 

• Training intrapersonal intelligence: New advances in virtual reality (VR) can be used to develop a diplomat’s intrapersonal intelligence. Such technology allows active “body swapping,” so people can “walk each other’s journeys.” Built-in behavioral science experiments may well detect implicit biases and identify cognitive challenges. VR provides a safe space for diplomats to learn about themselves, discover their biases, and better understand their interactions with others. Put differently, it fosters perspective-taking and helps overcome dehumanization. As Waehlisch suggested, “What if Netanyahu went through an Israeli checkpoint as a Palestinian?” In VR, he would see how people looked at him, the weapons pointed at him, and feel the danger, to perhaps reveal a new perspective.

Dangers of AI in Diplomacy

There are some things that can never be left to AI.

• Decision making: Delegating decision-making to AI would be a mistake, even though in some ways it could be seen as desirable.

  It is conceivable that AI could be programmed to make more rational, fair, and evidence-based decisions than humans. After all, AI is not vulnerable to human emotions or weaknesses. For centuries, the ideal diplomat was like a robot, coldly efficient, rational, and devoid of emotion, as codified in diplomatic protocols. Indeed, diplomats are routinely rotated every few years to prevent emotional attachments. In contrast, AI has no problem remaining detached and calm in stressful situations. Without emotions or physical sensations, AI could not be threatened or made to feel vulnerable in the same way as a human, for example, as when Vladimir Putin used his dog to intimidate Angela Merkel—famously terrified of dogs. AI would not be motivated by personal gain, or be tempted to abuse its authority and would be untroubled by the personal cost of resisting political pressure and standing by diplomatic policy decisions (in a nation’s interest) even if unpopular. AI would not be susceptible to exhaustion or lapses in judgment. In fact, a survey conducted by the Center for the Governance of Change at IE University in Spain, one in four Europeans indicated that they would prefer policy decisions be made by AI rather than politicians. However, in decision-making complete rationality is not always best.

Take as an example, the “Prisoner’s Dilemma” from game theory. Even though mutual cooperation would yield a greater net reward, the only possible outcome for two purely rational prisoners is betrayal. And of course, in real life, this stance could quickly lead to escalated military action, or nuclear war and mutual destruction. Even if we set parameters beforehand, these may be incomplete or fail. Would AI have the ability to pull back? If an algorithm were tasked with bringing about world peace, an efficient move might be to eradicate all humans from the planet.

Yejin Choi, a computer scientist and 2022 recipient of the MacArthur “Genius grant,” makes the same point from an ethical standpoint. In one interview, she said that in the most fundamental ways, “AI struggles with basic common sense.”

While humans understand many things, such as common exceptions to rules, AI must be specifically taught, or be at risk of choosing extreme or damaging solutions that humans would never consider. Choi argues the challenge will be to account for value pluralism, to teach AI that values can be broad and that diverse viewpoints need to be taken into account. Ethical guidelines are necessary but there is no one moral framework that can be imposed. The implications for diplomacy are dangerous. While AI will continue to improve, Choi doubts that humans will ever create sentient artificial intelligence, or AI with true intrapersonal intelligence.

• Malevolence: There is potential for AI technology to be used maliciously. We need to work on ways to forecast and mitigate such threats. The threat of AI is easy to see in what has been described as today’s “post-truth era.” AI is being used for negative messaging which leads to greater polarization, destabilization of existing frameworks, and the influencing of elections.

• Bias: There is also the problem of bias in AI systems. While often seen as a technical problem, most AI bias stems from human biases and systemic, institutional biases. For machine learning models to work well, a very large and diverse, and robust set of data involving all ages, genders, ethnicities, and other demographic criteria must be used. In the history of Western diplomacy, key decisions have been made by mostly men of a certain profile which could certainly skew the dataset. Regulations and safeguards are of course necessary. Excessive concentration in AI space and in a handful of technology companies must also be avoided through regulation—for example, encouraging competition and not allowing monopolization.

“The Greatest Threat and the Greatest Opportunity”

French Ambassador, David Cvach, said in a 2018 Tedx talk that AI is both the greatest threat and the greatest opportunity for diplomacy. There is truth to this assertion.

Sophia robot

In the field of international relations and diplomacy, AI is touted more often as a threat, for example, in terms of autonomous weapons. Though AI may have (often unintentional) negative consequences, organizations such as AI For Peace have a different stance: on their account, dialogue between academia, industry and civil society can help ensure the benefits of AI while minimizing the risks. Waehlisch has suggested machine learning and natural language processing can be used to promote peace. His chief concern is how to use new technologies to help de-escalate violence and increase international stability.

I would argue that while AI will augment the work of human diplomats making them more efficient and effective, it will never be more than a useful tool in diplomacy. Indeed, it could not and should not replace human diplomats. AI might outperform humans at most analytical tasks, but humans will still surpass AI at more subtle, “feeling tasks.” Even as algorithms come closer to replicating human interpersonal intelligence, direct person-to-person interaction is probably the most important method of increasing or maintaining “soft power” in diplomacy. Chatbots may be able to fool humans at the Diplomacy game online, but robots such as Sophia (appointed in 2017 as the UN Development Program’s first Innovation Champion), could not yet be mistaken for human.

On the positive side, the opportunities of AI lie in creating a more level playing field, as long as technology is not limited to wealthy countries. The ability of diverse stakeholders to use algorithms could provide more holistic and comprehensive solutions to today’s challenges, such as forced migration or unanticipated pandemics. Perhaps AI can be a means for engaging and uniting people around the world on issues of mutual interest for a more peaceful and sustainable future. We should use all our multiple intelligences, and the possibilities of artificial intelligence, to achieve this end.

In our next blog post, Howard Gardner will discuss the implications of AI in understanding human personal intelligences.

I would like to thank Howard Gardner for his valuable input into this post. I am also grateful to Steven Siqueira and Martin Waehlisch for very kindly agreeing to interviews and sharing their thoughts.

Are all Intelligences Equal? An Issue Raised by Cormac McCarthy’s Recent Novels

Ⓒ Howard Gardner 2023

Cormac McCarthy Throws Down the Gauntlet

In Stella Maris, the second of Cormac McCarthy’s recently published pair of novels, protagonist Alicia Western baldly asserts:

“The Stanford Binet (IQ test) is racist—no questions about music... For instance, music doesn’t count. So here’s a black guy with a measured IQ of 85 who is by any matrix you might care to choose a musical genius. Simply off the charts. But to the IQ folks he’s little more than a half wit.”

Whether or not he’s aware of it (and we have known each other casually since the 1980s), author McCarthy is giving voice to the insight that led to MI theory— “the theory of multiple intelligences.” Put simply, the human intellect cannot be adequately assessed by a single paper-and-pencil (or computer-administered) instrument. Rather, there are various forms of intelligence—linguistic, logical-mathematical, spatial, interpersonal, etc., each deserves to be assessed separately and in an “intelligence-fair” way. Strength—or weakness!—in one form of intelligence does not predict strength—or weakness—in other intellectual realms.

Alicia Western goes on to ridicule other forms of psychological testing—from the MMPI (Minnesota Multiphasic Personality Inventory) to the Raven’s Progressive Matrices. But by no means does she equate all intelligences. Indeed, as spelled out convincingly by reviewer James Wood, Alicia embraces a definite hierarchy or ranking system for intelligences. In her view, numerical gifts are the most important—the single index of intellect by which one should judge individuals. Like mathematics, music is also a gift from the gods—and therefore has a special status. Definitely of a lower rank is language—on Alicia’s account invented far more recently, subject constantly and unpredictably to historical forces and cultural factors, lacking in precision, and inherently incapable of achieving the universality of mathematics.

Clearly, this literary creation is proposing an intellectual pecking order among humans: Mathematicians stand above all others; physicists are those who are not quite gifted enough to carry out “pure mathematics;” those restricted to ordinary language—like novelists and even poets—are not in the same league. Perhaps, as reviewer James Wood suggests, McCarthy is ranking himself well below the mathematicians and scientists with whom he has long engaged at the interdisciplinary Santa Fe Institute. But perhaps as Graeme Wood, another reviewer, points out, it is language itself which may be better equipped to attack and (unpack) the philosophical issues that undergird and motivate many of Cormac McCarthy’s searching writings.

In what follows, I step back from the sheer ranking of importance, or priority, of various forms of intellect. I seek to identify and illuminate some of the various lenses with which one can consider intellect. And I put forth my own views.

The Battle Among Psychologists

Over a century ago, French psychologist Alfred Binet created the first test for intelligence that became widely used. Binet deliberately assembled a potpourri of questions which predicted, with some accuracy, who would succeed readily in (indeed, breeze through) French public schools— and who could be expected to encounter difficulties. Developed (and successfully so) for this limited purpose, these “IQ tests” soon swept much of the Western world. They were used not only in scholastic settings but as a tool for placing individuals in putatively appropriate niches in the military; or for determining whether a person accused of a crime might be treated differently, depending on whether considered a genius, of average intelligence, or–per the lingo of the day–an imbecile.

Psychometricians (along with other psychologists) being a contentious crowd, Binet’s work (particularly as reformulated by American test-makers Louis Thurstone and Lewis Terman) has often been critiqued. As early as the 1940s, psychometrician David Wechsler articulated the need to assess “social intelligence.” And then forty years later, both Robert Sternberg and I put forth more detailed dissections of intelligence—Sternberg introducing his triarchic theory, I putting forth the theory of multiple intelligences (MI); Sternberg created measures, I largely desisted from doing so. And once a pluralistic view of intellect had been proposed, psychologists as well as psychologically-oriented pundits readily proposed other forms of intelligence—ranging from financial to sexual to emotional (the latter put forth initially by psychologists Peter Salovey and John Mayer, and effectively communicated to the rest of the world by science journalist Daniel Goleman.)

Intelligences Beyond Humans: Animals, Living Entities… and other forms of Matter

While tests of intelligence were initially devised for administration to students (as well as soldiers and research subjects), there is no principled reasons why they should be restricted to Homo sapiens. And indeed, while it’s scarcely feasible to administer standard tests of language or mathematics to other species, it’s certainly possible to assess the maze-running ability of rats, or the navigational capacity of pigeons. Indeed, for some capacities, one or another species may well surpass human beings—if indeed, we knew how to assess the spatial abilities of octopi or the signaling acuity of dolphins.

Further: how convincingly can we execute comparisons within and across species? Perhaps rodents or birds can display as much spatial intelligence as humans…and many species (including our close primate relatives) may be proficient in bodily-kinesthetic intelligence. But how could we ascertain their personal intelligences—or, for that matter, their existential intelligences?

More recently, artists—as well as a few bolder scientists—have even raised the possibility of intellectual achievements beyond the animal kingdom. Perhaps, there is intelligence—or wisdom—among trees and plants. Or, indeed, perhaps the earth itself constitutes an intellectual gyroscope. For elaboration of these points, see my blog post (link here).

And of course, problem-solving—and perhaps even problem-finding—are no longer restricted to the natural world. Seventy years ago, computers were already capable of impressive computation—rapid calculation, logical proofs, game playing (such as chess), prediction of election results. And with every passing year—perhaps even week or day!—computers are able to best the most proficient human beings...and to do so even in games, contests, academic feats once thought the solitary purview of Homo sapiens. As for the retort that these machines (however constituted) have been designed, programmed, operated, read by human beings: we have to concede that we members of Homo sapiens no longer know, we no longer understand how computers are achieving these amazing feats… A trend that promises to continue and presumably accelerate indefinitely.

On to… Ontology

Shortly after I published Frames of Mind, my 1983 introduction to the theory of multiple intelligences, I arrived at an important realization.

Writing as a scholar and scientist—one reasonably well-versed in psychology and neuroscience—I had thought of the intelligences as properties of the human mind, the human brain. (A contemporary, data-grounded, criteria-based variant of phrenology, if you will). We are the species that possesses linguistic, logical-mathematical, musical, spatial, bodily-kinesthetic, interpersonal, and intrapersonal intelligences. (Later I added an eighth or naturalist intelligence; and I have also speculated about the possibilities of an existential intelligence (the intelligence which puts forth and tackles “big questions”) and a pedagogical intelligence (the intelligence that enables us to convey our knowledge in appropriate ways to audiences with varying degrees of knowledge and sophistication). No candidate intelligence is confirmed unless it meets the criteria that I have laid out.

But as I began to work with educators, I soon realized that they did not think primarily in terms of the computational capacities of their students. Rather, (and appropriately) most educators think in terms of curriculum—what’s appropriate for math class, physics class, English class, home economics, gym, the marching band, the art or dance studio, etc.)

Of course, there is a rough alignment between cognition and curriculum. Presumably, math class relies significantly on logical-mathematical intelligence, English class largely on linguistic intelligence, band and chorus on musical intelligence, etc.

But cognition and curricula are NOT equivalent. There are many ways to teach math; to construct and to assess mathematical assignments; to remember and draw on mathematical skills and reasoning. The same can be said with respect to almost any subject, any topic. And this realization led, in terms, to an important distinction between DOMAINS and INTELLIGENCES.

DOMAINS consist of the set of knowledge, procedures, insights that constitute a subject taught in school (or covered in a textbook or an online course); the INTELLIGENCES are the mind/brain toolkit on which learners can draw. Taking advantage of this toolkit, teachers can present materials and students can assimilate, apprehend, and display their knowledge of a specific topic. Accordingly: as an example, art history is a domain; but its mastery may call on linguistic intelligence, spatial intelligence, naturalist intelligences, and no doubt others as well. Or, to flip the example, our linguistic toolkit can be drawn on as we attempt to master domains as diverse as history, economics, or biology.

When communicating with educators, I convey this point with a simple assertion. Anything worth learning can be approached in more than one way. If you can teach a topic, or come to master a topic, in several ways, you have a well-rounded, well-textured understanding. Conversely, if you can only teach or understand a topic in one way, your grasp is limited, fragile.

This distinction also has important, indeed, crucial implications for students who exhibit learning deficits. If you teach such students in only a single conventional way, you are virtually bound to fail. The deft teacher of the dyslexic (or of the physically clumsy, or the tone deaf, or the color blind) is the teacher who is able to convey the DOMAIN’S key principles and ideas in novel but appropriate ways. As the innovative educational organization CAST maintains. It’s not the student that is disabled; it is the curriculum! Hence CAST’s call for Universal Design for Learning (UDL).

Then, there’s the opposite side of the coin: In traditional computing, everything can be expressed in bits—in 0s and 1s. Does that mean that everything that a computer can spew forth—from creating a poem to solving a differential equation to writing a fugue—is identical? Obviously not—in any helpful sense. Only if you argue that since all life contains DNA, all living material is the same; or, to switch entities, since all materials, all matter are composed of atoms, all materials, all matter are the same!

To generalize: Be wary if you have a very thin and one-dimensional view of the world(s) in which we live.

Ontology Continued

Let’s start with the extremes, just introduced:

At one horn of the extreme: there is only one form of computation (as epitomized by 0s and 1s in computation); every capacity can and should be expressed and evaluated in light of this primitive (or primary) metric.

On the opposite horn of the extreme: every computation that is not identical is distinct; accordingly, there are infinite numbers of computations—01, 10, 001, 100, 1010 etc.)

Few would be content to endorse either of these extreme positions.

Now consider a very rough analogue in the study of intelligence: There is only one form of intelligence, only one form of computation; and we can align individuals, tasks, programs, computational systems in term of degree or amount of computational power.

One can actually locate a statement to this effect—indeed, two such utterances—in Cormac’s McCarthy recently published novels. We encounter the claim that, for many individuals, physicist J. Robert Oppenheimer was “the smartest person man they’d ever met”—or, indeed, that the world had ever known. Roughly speaking, here’s the assertion: Oppenheimer could compute better or faster or both than any other human being—known or even imaginable.

However, this remarkable claim leaves questions unanswered. Were the evaluators actually taking into account all of Oppenheimer’s skills and capacities, disabilities, frailties, weaknesses as well as his clear prowess?

Consider the evidence. Let’s concede that Oppenheimer was a brilliant physicist and an excellent leader of the Manhattan Project. But it is also known that Oppenheimer did an inadequate job defending himself at the 1954 hearings of the Atomic Energy Commission. He did not demonstrate his capacity to be entrusted with secrets, and hence lost his security clearance. (It was restored in 2022, 68 years later)! Also, his own family life was troubled; as a young adult, he sought to poison one of his preceptors; and he never recovered psychologically from his role in the Manhattan project and the carnage in its aftermath. Nor, reverting to Alison’s hypothetical musical genius who was black, is there evidence that he was unusually gifted in music. Whatever it encompassed, Oppenheimer’s genius was clearly limited.

Back to Alicia, as portrayed in McCarthy’s books. On the one hand, she is willing to credit intelligence to the black musician—enough to undermine the usual psychometric assessments of intelligence, which tap logical-mathematical, linguistic, and perhaps spatial intelligence, but not the several remaining ones.

But there are less generous readings of Alicia’s position—which, Cormac McCarthy implies, would be found in those who populate centers like the Santa Fe Institute, the Institute of Advanced Study in Princeton, or similar “think tanks” in Europe.

Consider her critique of standard intelligence testing:

1. “I’ve never met anyone in the (IQ test) business who had any grasp at all of mathematics. And intelligence is numbers, it is not words. Words are things we’ve made up. Mathematics is not. The math and logic questions on the IQ tests are a joke.”

   
   Or as she puts it elsewhere:

2. When you are talking about intelligence, you’re talking about numbers. Verbal intelligence will take you only so far. There is a wall there… if you don’t understand numbers, you won’t even see the wall”

On reading 1, there may be other “lesser” intelligences—like musical or linguistic—but numerical, mathematical, or perhaps logical-mathematical stands indisputably at the top.

On reading 2, numerical intelligence—and, just possibly musical intelligence—may reflect the way that the universe is—indeed, the way that the universe has to be.

Let me put it in lay terms—the only ones that I (as a non-mathematician) can understand and articulate: Alicia is arguing: The universe is inherently numerical—it cannot be any other way—and mathematicians are the ones—for better or worse, the only ones—capable of decoding, understanding and explicating that reality.

Three apparent codicils:

1. Physics is secondary to math. More human beings can “do” physics than can “do” math. And mathematicians are the only human beings who can grasp—or can even attempt to grasp—the nature of the universe.

2. Just possibly, music is analogous, akin to math—perhaps there is only one music, and (as the philosopher Arthur Schopenhauer claimed), it is a privileged part of the mysterious nature of the universe.

3. Language is definitely secondary in importance and profundity. There are many languages, they have undergone—and continue to undergo—historical developments and unpredictable changes. They are influenced by culture; they are inherently imperfect; they lack the precision and the finality of mathematics, and, perhaps, of music.

Back to Psychology, Sociology, Brain science

With due respect to Alicia, and the author who created her, there’s a quite different way of conceptualizing this space—an entirely distinct ontology, so to speak.

On this account, over millennia, the brains (and minds) of the hominid species have evolved in numerous ways. Our species has different neurological organization and psychological capacities, developed to ensure our survival on the planet—and, accordingly, different ways of knowing the world. At various times, and under various circumstances, certain “ways of knowing” and “ways of expressing” come to the fore. (Small but timely example: until the 20th century, linguistic capacities were particularly valued in higher education; but that has radically changed in the last century, when logical-mathematical capacities are now valued far more). And it’s possible that in the 21st century, with machine computational capacities far outflanking those of human beings, that the personal intelligences (sometimes dubbed “soft skills”) will be ever more valorized.

And of course, the brain and the mind might have evolved quite differently. That’s an essential element of Darwinian evolutionary theory and the one that most disconcerts religious fundamentalists who adhere to a literal version of the account presented in Genesis.

A PeEk at Epistemological Ontology

Based on my glancing knowledge of intellectual history, philosophers René Descartes and Gottfried Wilhelm Leibniz were engaged in a deep effort to identify the most fundamental forms of thought—which they deemed as logical. Centuries later, Bertrand Russell and Alfred Whitehead concluded that the basics of mathematics, the realm of number, could be expressed in symbolic logic. Mathematician Kurt Gödel demonstrated the limitations—indeed the inherent incompleteness—of this mode of analysis. And pioneering computer scientists like Herbert Simon and Allan Newell used symbolic logic as the basis of their computing system; linguist Noam Chomsky (who aligned himself with the program of Cartesian Linguistics) was searching for universal principles of grammar or syntax that underlay superficial differences across spoken languages. Following in the footsteps of their physicist father, who had worked on the Manhattan Project, both Alicia Western and her brother Bobby Western valorized this strand of thinking; author Cormac McCarthy seems to endorse—or at least give voice to—that valorization.

However: In his penetrating (though positive) evaluation of the two McCarthy books, reviewer Graeme Wood points out that across his oeuvre, Cormac McCarthy actually wrestles with the most fundamental human issues—life, death, love, hatred, war, peace, choice, fate, destiny. These are not susceptible to conclusive mathematical analyses (as both Western children ultimately if regretfully came to acknowledge). However much one might wish it were the case, such vexing human concerns can’t lead to a convincing, slam-dunk conclusion—the vicissitudes of life do not yield a neat quod erat demonstrandum. Rather they call for—indeed require—unending reflection, debate, uncertainty—all serving as a stimulus to yet further discussion reading, writing, philosophizing. And this endless dialog, dialectic, deconstruction—far more than strings of 0s and1a, or Compu-chat—may continue to dominate, indeed may constitute, human destiny.

Coda: Two Ways to Construe this Epistemological Puzzle—A Personal Apologia

By inclination, perhaps by nature, I am a synthesizer. I like to lay out a puzzle; reflect on its various dimensions (including promising pathways but also dead ends); and then assemble the pieces in ways that make sense to me—and, perhaps, to others. Here are two guiding principles:

1. The Philosophy of Symbolic Forms
   As laid out by philosopher Susanne Langer, there are two basic forms of symbolization. One is rooted in math, logic, and ordinary language—these discursive forms have essentially the same meanings across usages and can be combined in various ways (as in this essay). The second is Presentational Symbols. Encountered in myth, religion, rituals—and especially in works of art—these symbols cannot be decomposed meaningfully. Change even one feature, one line, one beat, one note, and the work, the presentation as a whole is necessarily, if inexplicably, altered.
   My teacher, Nelson Goodman, expressed this distinction in a more formal stricture. He contrasted notational with non-notational symbols. The arts are the playground of non-notational symbols. A forgery can never be conflated with the original because it is always possible in principle to find differences between two superficially identical symbolic representations.
   Adopting this perspective, we can say that Cormac McCarthy (and other artists in other media) are mining the realm of non-notational symbolization, while the Western siblings valorize only the terrain of notational (or discursive) symbols.

2. The World of Organic Evolution
   The universe has existed for perhaps 13 billion years; our solar system (and presumably our planet) for 4.5 billion years; the first living entities go back 3.7 billion years; animals perhaps 800 million years. The first hominids appeared perhaps 250 million years ago; Homo sapiens perhaps 50,000 to 100,000 years ago.

   We can attribute certain capacities and potentials to each of these layers of existence: presumably the several intelligences that I have delineated only flowered in the last few hundreds of thousands of years. It does not make sense to apply “IQ” or “MI” ways of analyses to earlier strata of life.

   But when it comes to our understanding of the universe, the tools are different. Perhaps the Western siblings are correct: mathematical thinking can illuminate the nature of the physical universe, with musical thought queued up somewhere close by; in contrast, natural language—whether discursive or presentational—is a recent arrival, so to speak. And just perhaps, music may occupy an intermediate spot—with its Platonic features suspended somewhere between word and number. But for that very reason, if we want to understand all of matter—and especially life, living, growing, declining, dying—we will need to draw on the very verbal features that Cormac McCarthy’s characters may denigrate; —but that constitute sense making in its most capacious —Eastern as well as Western—sense.

For very helpful comments on earlier versions of this essay, I thank Shinri Furuzawa, Mindy Kornhaber, Donald Richard, and Ellen Winner.

References

Gardner, H. (1983). Frames of mind: The theory of multiple intelligences. New York: Basic Books.

Gardner, H. (2006). Multiple intelligences: New horizons. New York: Basic Books: New York.

Goleman, D. (1995). Emotional intelligence: Why it can matter more than IQ. New York: Bantam Books.

Goodman, N. (1968). Languages of art. Indianapolis: Hackett.

Langer, S. K. (1942). Philosophy in a new key. Cambridge: Harvard University Press.

Salovey, P., & Mayer, J. D. (1990). Emotional intelligence. Imagination, cognition and

personality, 9(3), 185-211.

McCarthy, C. (2022). The passenger. New York: Alfred A. Knopf.

McCarthy, C. (2022). Stella Maris. New York: Alfred A. Knopf.

Sternberg, R. J. (1988). The triarchic mind. New York: Viking.

Wood, G. (2022). Cormac McCarthy has never been better. The Atlantic, (January-February), 78-81.

Wood, J. (2022). The numbers game. The New Yorker, (December 19), 60-65.

Photo by Stefan Cosma on Unsplash

MI Theory in Bilingual Education

A bilingual school in Hong Kong recently advertised itself as being inspired by the theory of multiple intelligences (link here).

Howard Gardner has been approached many times about how best to use MI theory in the teaching of foreign languages. He believed originally that MI ideas could only contribute modestly to bilingual education, compared to how they could be used in the teaching of subjects such as history or biology. He now offers the following advice on ways in which MI theory can be useful.

1. Youngsters probably learn languages best in different ways. I, for example, love to learn from the written text; many, perhaps most others, prefer to learn through human interactions, watching videos, etc.

2. We all learn the best when we are motivated by the topic and like the setting in which the learning takes place. Especially given the new technologies, it should be possible to "custom-fit" the educational approach to the student.

3. This is probably the most important point, but it is the most difficult to explain succinctly. Every language has certain distinctive features which are non-intuitive to those who do not speak the language. For example, the various subjunctives in Spanish don’t make intuitive sense to me; Latin cases are difficult for those who don't mark cases in their native language, etc. There are many ways in which to convey these concepts, which are important within a language, and multiple intelligence approaches can be quite helpful. I work out these ideas in Chapters 7-9 of my book, The Disciplined Mind.

4. My book, Intelligence Reframed, contains a large bibliography which has various "tests" for the multiple intelligences. I don’t particularly recommend any of them, but if you need some kind of instrument to measure a profile, that is where you should begin. Be sure to distinguish between tests that simply measure preferences, and ones that actually tap skills in the different intelligences.

More broadly speaking, the key ways that schools can use the theory of multiple intelligences is through individuation and pluralization. Whether in language education, or the teaching of any other subject, teachers can use MI theory to design lesson plans that enhance their students' strengths, improve upon any weaknesses, and keep engagement high.

Photo by Towfiqu barbhuiya on Unsplash

MI Depicted in Stained Glass Window

We have just learned, thanks to Jim Reese of Washington International School, that the theory of multiple intelligences is depicted in a stained glass window at Utah Valley University. MI theory is part of one panel in an installation known as the Roots of Knowledge.

The windows installation is 200 feet long, 10 feet high, and contains 80 panes which pay tribute to the greatest human achievements in science, philosophy, medicine, law, literature, music, art, architecture, and technology. It took 12 years to complete and was unveiled in 2016 as part of Utah Valley University’s 75th anniversary celebrations.

Multiple Intelligences is part of the penultimate Places You’ll Go column of windows (top left)

“The windows in this column take the viewer into the modern era and a little bit beyond with projections of what is to come. The theme of education and the pursuit of knowledge continues to the end of the Roots of Knowledge, as displayed by a cutting-edge classroom, luxurious buildings rising into the ether, high-speed transit systems below the seas, and a limitless galaxy beyond our earth waiting to be explored.”

An online description of the window panel reads as follows:

“MULTIPLE INTELLIGENCES:

Howard Gardner (1943 – ) is a psychologist who published Frames of Mind in 1983. In this ground-breaking book, he proposed the theory that humans have multiple intelligences, a big challenge to the previous century's knowledge of human intelligence.”

More information on the Roots of Knowledge windows can be found here.

Photo credit: Jim Reese and Utah Valley University

Temple Grandin and the Theory of Multiple Intelligences

By Shinri Furuzawa

Temple Grandin wears many hats—she is a leading animal scientist in the humane livestock-handling industry, an academic at Colorado State University, and autism-rights activist. She is also a New York Times bestselling author

In her latest work, Visual Thinking: The Hidden Gifts of People Who Think in Pictures, Patterns, and Abstractions, she references Howard Gardner and his theory of multiple intelligences. Grandin has also spoken of Gardner’s ideas at her recent book events and in this interview with The Washington Post (link here).

How does Grandin explain “visual thinking”?

Grandin describes two types of thinkers—those who are “verbal” and those who are “visual.”

• Verbal thinkers: think linearly, tend to do well in school as information is taught sequentially, are often well organized, sociable, and good talkers. They gravitate towards careers as educators, administrators, lawyers, writers, politicians, etc.

• Visual thinkers: see images in their minds and make rapid associations among them, have an excellent sense of direction, are good problem solvers, and easily understand how things work or fit together. On the other hand, they are often late talkers who struggle with traditional educational settings and teaching methods, and they may also be socially awkward.

Visual thinkers can be further divided into two types, object and spatial.

Object-visual thinkers: see the world in pictures, and gravitate towards careers as designers, artists, architects, mechanical engineers, etc.

Spatial-visual thinkers: see the world in patterns and abstractions, and gravitate towards careers as statisticians, electrical engineers, physicists, etc.

What’s the evidence?

Grandin cites several studies to support her claims. Examples from recent brain research include a 2015 study by Kazuo Nishimura and colleagues in which magnetoencephalography was used to measure brain activity associated with “verbal” and “visual.” They demonstrated that when study participants were asked to recall a temple, signs of the zodiac, or a past conversation, “visual thinkers” created images while “verbal thinkers” used language. She also mentions Nobel laureate Roger Sperry and his notion of left-brain vs right-brain thinking to distinguish “verbal vs. visual” thinking.

Grandin cites most heavily the work of Maria Kozhevnikov of the Visual-Spatial Cognition Lab at Massachusetts General Hospital. In one 2002 study, Kozhevnikov tested high visualizers on spatial reasoning and other visual tests. She found that artists and designers tested as object visualizers while scientists tested as spatial visualizers. And while high-spatial visualizers interpreted graphs as abstract representations of spatial relations, low-spatial visualizers saw graphs as pictures. Verbalizers showed no preference for visual or spatial imagery.

What about Grandin herself?

Grandin categorizes herself as an "object-visual thinker.” She sees the world in photorealistic pictures, or film clips. To process any information or solve a specific problem, she must “do the equivalent of a Google search” in her mind to access images. A scan of Grandin’s brain using Diffusion Tensor Imaging showed her visual circuits to be significantly larger than those in the control group. She describes these circuits as a “huge internet trunk line from my rear visual cortex to my frontal cortex.”

Throughout the book, Grandin supports her ideas with anecdotes from her own life. For example, in critiquing the US education system, she claims that the system in effect screens out “object-visual thinkers.” She points out the irony that though she now teaches veterinarians and has proven mechanical engineering skills, she herself could not have qualified for veterinary school or engineering programs due to her lack of ability in algebra.

Am I convinced?

One problem with Grandin’s book is her overreliance on anecdotal and autobiographical examples as evidence. Even the scientific data she gives are not always robust. For example, the 2002 Kozhevnikov study mentioned above had a sample size of only 17, and Sperry’s left brain/right brain hemisphericity premise is now seen as of historic rather than of scientific interest.

Some of Grandin’s claims are farfetched. In her chapter “Visualizing Risk to Prevent Disasters,” she suggests that many disasters, such as Fukushima, could have been prevented. How? In her argument, in the case of that nuclear disaster, a “visual thinker” would have “envisioned the water coming over the top of the seawall.” She also suggests that “most geniuses” are “visual thinkers” including Edison, Einstein, and Picasso—relying on controversial posthumous (not clinical) diagnoses of their neurodivergence. Different groups in the neurodivergent community often claim past luminaries have their particular learning difference. This may be an effort to enhance the prestige and status of their own marginalized group and motivate others in their community. In fact, Einstein is claimed as an (unwitting) “ancestor” by many groups!

To be fair, Visual Thinking is not aimed at a scientific audience. For the lay reader perhaps it is enough to appreciate, as Grandin argues convincingly, that people who think differently deserve to be recognized, and should be recognized, for the betterment of their own self-image and for their possible contributions to society. Teams can be stronger when they include diverse thinkers. She promotes an approach to neurodiversity not as a disability but as an asset. Grandin herself is an inspirational figure; many enjoy and benefit from reading about her ideas and about her life.

how does “Visual thinking” fit into the theory of multiple intelligences?

In terms of MI theory, Grandin’s “verbal thinking” reflects Gardner’s linguistic intelligence, while her “visual thinking” reflects spatial intelligence.

I would infer that Temple Grandin displays impressive capacities with the following intelligences:

• Naturalist intelligence (she has expert knowledge of animal behavior, in particular, cattle)

• Spatial intelligence (she is able to visualize maps and design complex plans by seeing them in her mind)

• Bodily-kinesthetic intelligence (she is very good with her hands)

• Logical-mathematical intelligence (she is knowledgeable in mechanical engineering)

On the other hand, as is commonly the case in people with autism, Grandin admits that understanding others can be a challenge, so interpersonal intelligence would not be a strength.

Howard Gardner does not agree with the concept of “visual thinking.” He is quoted on the jacket of Visual Thinking as saying that he believes Temple Grandin has written a “fine book” that clearly defines how “a self-described ‘visual thinker’ apprehends, understands, and explains the world.” Much of what Grandin describes is what Gardner would call spatial intelligence. He also asserts that visual is different from spatial and the two should not be confounded. Gardner says, “Reading is visual, and appreciating art and sculpture is visual, but they are not particularly spatial—blind people are capable of developing a strong spatial sense.” People who read at a young age can be spatially challenged, while people who have difficulty reading can have excellent spatial intelligence. Moreover, it is important not to tie an intelligence to a single sensory system like sight or audition. Grandin would counter this argument by saying that “visual thinking” is not about simply seeing, but rather it is how the brain perceives what is being seen.

Is there common ground?

In her book, Grandin concedes her differences with Gardner but finds they have similar views on education. She writes, “Though Gardner doesn’t recognize visual thinkers (let alone the different kinds of visual thinkers) as a separate category of intelligence, we are in agreement that our education system fails to recognize different types of intelligence.” Grandin believes there is a crisis in American education that is leading to a dangerous loss of technical skills and ingenuity. She attributes this situation partly to prejudices against community colleges and vocational schools, but also the fact that hands-on subjects like art and shop class are being phased out of schools. She also laments the barriers presented by biased testing systems that screen out otherwise capable and talented learners.

In several respects, Gardner concurs with Grandin’s broader aims. He has long criticized standardized tests, such as the SAT commonly used for US college admission, that unfairly tap primarily linguistic and logical-mathematical intelligences. His MI theory is a sustained critique of the widespread assumption that intelligence can be adequately measured by IQ tests and their ilk. He has argued for recognizing and nurturing “all of the varied intelligences and combinations of intelligence.”

Ultimate goals

By advocating for the neurodiverse members of our society, Temple Grandin hopes her greatest legacy will be that she helped neurodiverse children find careers that contribute to society. She told The Washington Post,

“I want to help the kids that are neurodiverse, they have different kinds of minds—autistic, dyslexic, ADHD, or whatever—to get into satisfying jobs where they can make a positive difference. That's the thing I feel I need to be doing now, as somebody who has had a long career and is now in their 70s.”

Howard Gardner has a similar hope: to encourage people to use their multiple intelligences to do “good work,” that is, work which is excellent, engaging, and ethical. Gardner has said it is the most important thing he can do.

References

Gardner, H. E. (2011). Frames of mind: The theory of multiple intelligences. Basic books.

Grandin, T., & Lerner, B. (2022).  Visual thinking: The hidden gifts of people who think in pictures, patterns, and abstractions. Riverhead Books.

For their comments on an earlier version of this blog post, I thank Tom Hoerr, and my colleague, Howard Gardner.