We the Stakeholders (Philadelphia, 1787) The Declaration of the Rights of Man and of the Stakeholder (Paris, 1789)
If you have been following French politics over the past year you may have the impression that democracy is in as much trouble in France as in the United States. And you wouldn’t be wrong1: President Macron was only able to raise the retirement age by resorting to a procedure, built into the constitution of the Fifth Republic, that bypasses parliament; protest demonstrations are banned by police prefects under the authority of an Interior Minister who just two years ago accused far-right presidental candidate Marine Le Pen of being “soft”; and opinion polls indicate that Le Pen rather than Macron would be elected president if a rematch were held this year.
In other respects, however, there has never been a better time in France, where I am spending the summer, to think about democracy. It has been well understood here, at least since 1789, that representative democracy neither begins nor ends with the electoral process. Although the word “stakeholder” has a French translation — partie prenante — the participant in democracy here is called the citizen, just as in 1789.2 The “Festival of Ideas” that was held the first weekend of July, near France’s geographic center in La Charité sur Loire, under the slogan “In the face of perils, let’s grow a desirable world”3, was, among other things, a three-day exercise in participatory democracy. On one side were politicians — mayors of small villages and towns, municipal councilors, parliamentary deputies, think tank researchers, and party officials. On the other side were activists in NGOs and local associations, teachers and agents of social services — and self-described “ordinary citizens.” Not a “stakeholder” in sight!
An original feature of the Festival of Ideas, now in its fourth year, is the reverse debate: instead of sitting through a panel discussion and then speaking up in the Q&A, the debate starts off with a series of questions from the audience, to which the panelists then respond. This took an unexpected turn in the closing panel, when one citizen after another stood up to express their frustration to the top party officials at the room’s focal point.4 They complained, often with surprising vehemence, that the would-be leaders of France’s democratic renewal sound like “elected officials very remote from the people.”5 The headline of the regional paper chose a quotation from a grassroots activist as its headline:
Vous nous écoutez, mais vous ne nous entendez pas.
(You listen to us but you don’t hear us.)
Is mathematics ready for such a vigorous and uncomfortable reverse debate? The role of AI is just one of many potential subjects.
"People of differing views, including some skeptics"
You have probably noticed that practically every day a new article appears in which a journalist or an academic writes something like
or
Because you have these conversations, and you get this somewhat terrifying sense that people understand themselves to be making extraordinarily high stakes decisions, decisions that I think the average person on the street would say that should absolutely be the prerogative of a government with a lot of careful consultation, maybe some kind of international framework. And you have people who are like, we’re going to do it. We’re going to do it ourselves. We thought about it, and we think it’s going to go best if we do it ourselves, and we do it now.…if you want to stop dangerous technologies, that has to be on a structural level with governance and regulation and international agreements and voluntary between lab agreements. You can’t count on people, unfortunately, to just say this is ludicrously dangerous. We’re not going to do it, because I guess, some share of people are just like, this is ludicrously dangerous, and I’m going to do it.
or
or
It’s such an irony seeing a posture about the concern of harms by people who are rapidly releasing into commercial use the system responsible for those very harms. (Sarah Myers West, managing director of AI Now Institute, quoted in NY Times, May 16, 2023.)
or
The message from Altman at al. seems to be "AI is so dangerous, powerful, and mysterious that only people at the top AI companies know enough to regulate it." Regulatory capture is the point. (Melanie Mitchell, Twitter, May 30, 2023)
or
This last is from a NY Times article by Ezra Klein, entitled “The Surprising Thing A.I. Engineers Will Tell You if You Let Them.” The bigger question is embedded in Klein’s answer: what is “society”? It is, he believes, the kind of thing that can “consider,” but what does that mean?
I’ve been reading lately, and probably you have too, about numerous institutes and agencies and deliberative bodies, some of them doubling as “policymakers,” proposing guidelines to regulate AI before it regulates us. The Klein article quoted above lists several of them. It does not mention the industry’s own proposals for AI “governance,” structured by their business model:6
The entire architecture behind the AI “is in the automation mode,” [Daron Acemoglu] says. “But there is nothing inherent about generative AI or AI in general that should push us in this direction. It’s the business models and the vision of the people in OpenAI and Microsoft and the venture capital community.”
If you believe we can steer a technology’s trajectory, then an obvious question is: Who is “we”? And this is where Acemoglu and Johnson are most provocative. They write: “Society and its powerful gatekeepers need to stop being mesmerized by tech billionaires and their agenda … One does not need to be an AI expert to have a say about the direction of progress and the future of our society forged by these technologies.”7
One of these tech billionaires laid out his agenda a few years ago:
Housing, education, and health care are each ferociously complex, but what they have in common is skyrocketing prices in a world where technology is driving down prices of most other products and services. … I think we should build in the next decade new technologies, businesses, and industries that break these price curves -- and in fact reverse them, and make these three primary markers of the American dream easier and easier for regular people to attain.
What, for example, does Andreessen Horowitz have in store for education?
…record numbers of domain experts in education (disillusioned teachers, etc.) are available to team up with record numbers of product builders trained to build modern consumer experiences. Together, these partnerships can deliver the core of any education business: products that either aggregate and deliver content, or products and communities that establish behaviors key to learning such as accountability and consistency.8
Is society best viewed as a collection of “businesses”?
In the article quoted above, Klein welcomed the announcement of Senate majority leader Charles Schumer’s initiative on AI regulation. Details of the procedure have since been made public
Skeptics may legitimately be skeptical that their concerns will be given as much weight as geopolitical and national economic considerations, but this sounds very much like a blueprint for a reverse debate squeezed into one panel, in one room, sitting at one table. I was heartened to learn that Joy Buolamwini, founder of the Algorithmic Justice League, was invited to a roundtable with President Biden. Was she invited as a stakeholder or as a citizen?
How does mathematics measure up to democratic ideals?
Although I am a citizen and not an AI expert, I don’t expect either Schumer’s committee or the European parliamentarians who have begun deliberating on the EU AI Act to invite me to participate in the debate (reverse or not). In mathematics, on the other hand, I do qualify as some kind of gatekeeper. So before I complain about the democratic deficit in decision-making with regard to the development of AI, consistency obliges me to account for how decisions are made in the domain in which I hold a share of power.
If you ever visited the UCLA mathematics department you probably recognize the image in the photo at the top of this essay, but you may not know that it was the product of a collaboration between UCLA professor Raymond Redheffer and the design studio of Charles and Ray Eames. The same collaboration produced the IBM “Men of Modern Mathematics” poster that used to be ubiquitous in mathematics departments. In the second chapter of Mathematics without Apologies, I used the poster and a different picture of the UCLA gallery to illustrate how the designation of a hierarchy of great mathematicians provides structure to the field.9
For $1485 you can buy an original copy of the poster from the Eames Office, and a version of the poster covers one wall of the Mathematica exhibit at Boston’s Museum of Science, but I haven’t seen the poster itself for quite some time. The reason, undoubtedly, is that the poster’s name is no longer acceptable and the fact that there was only one woman among the men even less so. At one point IBM had created an iPad app, entitled “Minds of Modern Mathematics,” with the information on the poster. It’s claimed on the MAA website that the app “rightfully add[ed] coverage of women in mathematics,” and Wikipedia says that a panel including “contributions of both men and women” has been added to the Boston exhibition, but the IBM app no longer works and I have been unable to find details about the Boston panel.
The news from UCLA is that the gallery has been taken down, and Redheffer’s collection dispersed, precisely because nearly all the images were of men.
I’ve spoken to UCLA colleagues who opposed this move and others who thought it was overdue. Upon hearing the news, one female colleague (not at UCLA) said something to the effect of “good riddance.” But it’s not clear to me in what way taking the photos down has made mathematics less hierarchical or more democratic. It is believed, and it may be true, that eliminating symbols of past discrimination is a step toward eliminating discrimination in the future. But where do symbols end? The (almost) uniformly male images are no longer on the wall, but the mathematics taught in UCLA courses hasn’t changed, and the textbooks are filled with the same men’s names. And even if serious plans were made to diversify the contents of the mathematics sequence, along the lines of what has been done in literature departments,10 would this really make mathematics more democratic? Will university courses still be hierarchical, but with different faces on the hierarchy, and eventually on the walls of mathematics department corridors?
"In the Western metropoles, diversifying the curriculum has proven to be easier than diversifying the student population and very much easier than making university faculties and student bodies truly inclusive.” [Guillory's book] suggests that “Western ‘high culture’ has become the low-hanging fruit for the decolonial project”: because we are powerless to change the real injustices and inequalities that have resulted from the long history of colonialism, we take out our aggressions on “figures who …have been roused from their dormancy to serve as … the face of imperial domination.”… in the absence of our collective ability to do anything about capitalism or racism, we offer a course on … its representation in literature … and consider that a sufficient political act.11
“Decolonial project” refers to observations like:
The colonized is elevated above his jungle status in proportion to his adoption of the mother country's cultural standards. (Frantz Fanon, Black Skin, White Masks)
Applied to mathematics, Fanon’s sentence means: any member of what is usually called an “underrepresented group” is welcome in mathematics, but on the terms set by those already represented. When a mathematician refers to decisions that affect the community of mathematics as such, it’s common to use the word “we” to designate that community. For example, in the article that framed last October’s Fields Medal Symposium, Venkatesh wrote
how we value mathematics is an active process inextricably woven in with the actual doing of mathematics. (italicized in the original)
The question is then: who is this “we” who decides on what terms one gets to join the profession’s cruise ship — or its lifeboat? The ones already on board? Is it up to “society” to decide? And then who is “society?” Is “we” speaking for the “you” who listen to “us” but don’t hear “us”?
To regard teachers—in our entire educational system, from the primary grades to the university—as the priests of our democracy is therefore not to indulge in hyperbole.
(Felix Frankfurter 1952)
The need to conform to digital designs has created an ambient expectation of human subservience.
(Jaron Lanier, 2023)
The questions raised in the previous section are not artificial or frivolous. Diversifying the image mathematics presents to itself can make the prospect of mathematical research more inviting to sectors of the population that have historically felt excluded. Attenuating the effects of hierarchy can transform mathematics from an old boys’ network to a democratic community. Opening up the curriculum, it is said, can improve access to STEM degrees that are increasingly necessary for the job market.12
Several colleagues have cited the prospect of opening up the profession as a prime motive for their enthusiasm for mechanization of mathematics.13 On the one hand, while reference to the “mathematical community” masks the profession’s hierarchical nature, the “Lean community” genuinely appears to function as a community of equals, with collective projects the norm. There are also anecdotal reports suggesting that teaching formal methods fosters a more egalitarian sense of mathematical practice.14
On the other hand, access was one of my scorecard’s topics most extensively covered at The Workshop. As reported in my earlier posts, the record thus far is ambiguous. Although Stella Biderman described a rather egalitarian Open Source community, it’s already clear from the first Workshop that, without a reorganization of research on a scale that no one is considering, academic departments and institutes will never be able on their own to match the computing power of the tech giants. Other creative “businesses” are concerned by a similar imbalance of power.
With the encroachment of AI into music, there are important questions about what happens to creativity, imperfection and the livelihoods of music producers, but the bigger questions are about who gets the rewards and who gets to define culture.…The risk is that the technology will concentrate power in ways that devalue and homogenise music.
(Jack Stilgoe, “Give the Drummer Some,” February 28, 2023, aeon.co)
So I have some suggestions for colleagues organizing the follow-up to The Workshop:15
Invite “people of differing views, including some skeptics.”
Consider the model of reverse debate, at least for closing sessions.
This one is a bit longer. One aspect of The Workshop’s respect for democratic process intrigues me. The debate was organized in the traditional top-down direction: panelists spoke, and then the moderator read questions the audience had submitted via Microsoft’s Slido, which bills itself the ultimate Q&A and polling platform for live and virtual meetings and events. With vanishingly few exceptions, the questions expressed sympathy for the excitement shared by panelists and organizers. Did the moderators filter out all skeptical questions, or were there no skeptical questions to filter out?
The organizers can allay any doubts on this score by making the complete list of questions on Slido available for inspection.
Devote an entire panel devoted to externalities of mechanization of mathematics. These include but are not limited to: questions of intellectual property in training models; energy consumption; water consumption.
And easiest of all: scrap that word stakeholders! Try citizen instead — not in the exclusionary sense defined by birth or identity papers but in the expansive sense, as understood by the Paris Commune of 1871, where it referred to membership in a collective defined by common goals.16
The Berggruen Governance Index rates France’s Quality of Democracy at a pretty steady 88 from 2000-2019, while the US ranking slipped from 90 to 83 over the same period, after peaking at 95. More recent figures are not available.
The main difference is that the word is now often written citoyen-ne. The terminology makes a difference. One is a citizen by virtue of being a human being, and not through participation in an economic interest. Google and Amazon can be stakeholders, and in the United States thay can be persons; they will never be citizens. See also note 16 below for the independence of the status of citoyen-ne from identity papers.
Face aux périls, faisons grandir un monde désirable. The “perils” in question include the climate crisis and the danger of social conflict — unexpectedly demonstrated by the wave of urban revolts just the week before the festival — as well as the prospect of right-wing extremism in the next presidential election.
Three of the four constituent parties of NUPES — the left/green coalition that has the largest bloc in the National Assembly — were represented on the closing panel.
A local media collective recorded this astonishing lesson in participatory democracy in its entirety; it can be watched at https://www.webtvdoc.fr/video/749/pleniere-de-cloture-1/#.
For example, Open AI proposes “to experiment” with designing a “mechanism” for “strong public oversight”:
…the governance of the most powerful systems, as well as decisions regarding their deployment, must have strong public oversight. We believe people around the world should democratically decide on the bounds and defaults for AI systems. We don't yet know how to design such a mechanism, but we plan to experiment with its development. We continue to think that, within these wide bounds, individual users should have a lot of control over how the AI they use behaves.
D. Rotman, “ChatGPT is about to revolutionize the economy. We need to decide what that looks like.” MIT Technology Review, March 25, 2023. The article discusses the forthcoming book Power and Progress: Our 1000-Year Struggle Over Technology & Prosperity, by Acemoglu and Simon Johnson.
https://a16z.com/2022/09/16/the-new-learning-economy-its-time-to-build-in-education/, on
To avoid misunderstandings: to say that the hierarchy is “designated” is to say that mathematical greatness is “socially constructed.” See Martin Andler’s talk on the practical effects of hierarchy in the specific case of probability in mid-20th century France. However, the claim is not that it is merely socially constructed, nor that the designation is invalid.
There have been serious proposals to reform the teaching of mathematics to reduce the reward that accrues to stereotypical behavior of dominant groups. Eugenia Cheng’s distinction between ingressive and congressive traits — corresponding, roughly, to competition vs. cooperation — is a notable example. And I am a consultant for an NSF-sponsored project to incorporate the teaching of ethics in the mathematics program. But I am aware of no suggestions to change the subject matter.
Evan Kindley, “Departments on the Defensive,” review of Professing Criticism by John Guillory, NY Review of Books, March 9, 2023.
Often when the DEI industry comes calling on mathematics departments the job market is taken as a fact of nature. Who decided that STEM degrees are a requirement for meaningful integration into the international economy at a level permitting a middle-class lifestyle? For that matter, who decided on the middle-class lifestyle? From one skeptical article: "where did DEI come from? Who runs DEI initiatives? And, most importantly, why does DEI so often inhibit the real change for which many are calling?”
The Math Genome Project promises to formalize “[e]very single” mathematical statement, explicitly for the sake of “Democratizing Higher Mathematics.” I may have more to say about this in the future.
I wonder, still, whether interaction with Lean software is less “top-down” than interaction with a human teacher.
"Considérant que le drapeau de la Commune est celui de la République universelle ; considérant que toute cité a le droit de donner le titre de citoyen aux étrangers qui la servent [...], la commission est d'avis que les étrangers peuvent être admis, et vous propose l'admission du citoyen Frankel. " Marie-Danielle Demélas, Alain Boscus, Militantisme et histoire, Presses Univ. du Mirail, 2000, p. 177, my emphasis.
OK, Mr. Harris. You are no exception to the common leftist woke:
As soon as an opinion is not shared by you or - heaven behold - is polit-inco, it is deleted. Ciao then.
Has mathematics ever been democratic?
Conjuring up Alexis de Tocqueville's spirit one might address this question as follows:
In the business of mathematics as a discipline, demos is sidelined: the truth is determined by the proof being valid, rhetoric (and other democratic devices) notwithstanding.
Whether or not the society a mathematician happens to be born into is democratic or not, seems to play a tangential role. Few would argue that the USSR was democratic 1917-1987. Fewer still would argue that Soviet mathematics did not flourish during that period.