The Connected City of Ideas

What is the purpose of a moral metaphor? Think of the boss who says, “our company is a family.” Or someone lobbying for corporate tax cuts, who says a strong business sector is “a rising tide that lifts all boats.” These metaphors seem to be issuing moral appeals, of a sort. But how are they meant to work exactly?

Here’s a rough proposal. The point of a moral metaphor is to highlight an aspect of a thing, and tell us that this aspect matters in how we deal with the thing, or that it matters more than we usually suppose. Moral metaphors provide perspectives, in philosopher Elisabeth Camp’s sense of the word. They organize our thinking “by imposing a complex structure of relative prominence . . . so that some features stick out in our minds,” and by imposing “evaluative attitudes and emotional valences on [a thing’s] constituent features.”1 The rising-tide metaphor tells us that the aggregate benefits of a buoyant economy matter more than how evenly they trickle down. The family metaphor tells us that commerce isn’t the only aspect of corporate life to be valued. Relationships matter too. The way we imagine the world, as philosopher Mary Midgeley says, determines “what we select for our attention among the welter of facts that constantly flood in upon us.”2 Moral metaphors are devices for imaginative reflection that highlight morally underappreciated aspects of things.3

If we want to judge the aptness of a moral metaphor, we have to ask, “does it make sense to shine a moral spotlight on that part of the stage?” Consider the corporate-family metaphor. It highlights the way that companies give us relationships, not just commerce. Its aptness depends on whether this part of corporate life is in fact underappreciated. Maybe we judge that it is. Or maybe we think it isn’t, and that it is mainly about guilting workers into doing unpaid overtime.

I am talking about how moral metaphors work because I think we need to update the metaphors we use around free speech. Everyone can see that our communication tools and practices are evolving fast, with a mix of welcome and unwelcome results. But there is an aspect of this evolution that is seriously under­appreciated. Our communication tools and practices are increasingly subject to standardizing and homogenizing pressures. We are being corralled into a narrower range of devices and methods for talking to each other. We need to actively strategize about how to deal with the threat that this homogenization poses to our abilities as creative, reflective, thinking beings. But first, we need to recognize it as a threat.

The dominant moral metaphor in free speech discourse—namely, the marketplace of ideas—inadvertently desensitizes us to this threat. This metaphor invites us to worry, primarily, about authorities controlling the ideological content of public communication. At the same time, it analogically portrays homogenization in our methods of communication as something benign or even good. We need another metaphor that frames this homogenization as something to worry about.

Cities are more liveable when they are connected, when they have an integrated mix of trains, cars, buses, cycle paths, and walking paths, which provide a diverse array of locomotive affordances. Similarly, societies are more liveable if they enable us to use a variety of idea-transmission media with diverse communicative affordances with respect to expressive formats (text, voice), stylistic options, breadths of audience, and tempos of exchange. We should be able to freely exchange ideas and information, subject to reasonable caveats. But we should not be content with this measure of freedom. We should also be free to exchange ideas using a heterogeneous repertoire of media and methods, suited to various communicative purposes. We should have a connected city of ideas.

John Stuart Mill’s writing inspired the marketplace of ideas metaphor. But that metaphor has become a dead dogma of the kind that Mill saw as inhibiting our mental vitality.4 If we want to carry the free speech tradition’s underlying ideals into the future, and refashion liberal society, we need interpretive lenses that have a deeper focal point than the marketplace metaphor gives us. We need lenses that orient our gaze toward problems that Mill, in the nineteenth century, and the lawmakers who implemented his ideals in the twentieth, couldn’t yet envision.

The marketplace metaphor has established rivals. Alexander Meiklejohn used the image of a town hall meeting to illustrate the normative appeal and pragmatic implications of a democratic conception of free speech.5 Robert Goodin and Robert Sparrow have riffed on marketplace lingo, inviting us to think of free speech culture as a garden of ideas.6 Seana Shiffrin’s “thinker-­based” theory of free speech has, at its heart, a striking simile, likening censorship to solitary confinement.7

By pitting my connected-city metaphor against the marketplace of ideas I am not insisting that the latter is the best of the currently available options. I am targeting the marketplace metaphor mainly because it is so influential. At the same time, I disagree with those critics who regard it as a totally hollow or disingenuous piece of rhetoric.8 I believe it has some enduring merit as a highlighting device.

To appreciate this, we have to decode the metaphor by asking, first, why markets per se are presumed valuable and, second, how the benefits of not having censorship resemble the benefits of using free markets to organize certain activities.

The key convictions behind a promarket ethos, for present purposes, are 1) that preference-satisfaction is good, or a reasonable proxy for the good, 2) that people are decent at knowing their own preferences, and 3) that people do better in acting to satisfy their preferences than third parties. Except in special circumstances, then, we should avoid things like centrally planned economies or protectionist limits on trade. These are bad because they interfere—ineptly, or based on insufficient information—with the satisfaction of our preferences, which are better satisfiable if we are left to conduct voluntary, mutually beneficial exchange. Or so the theory says.9 In essence, markets are good because they distribute stuff in a way that efficiently satisfies people’s preferences and, crucially for our purposes, they take the work of stuff-distribution out of authorities’ incapable hands.10

Analogously, free speech principles take the work of information- and opinion-­distribution out of the authorities’ incapable hands. Authorities are liable to think they know better than the folk themselves which ideas and opinions (as with, which products) are good for the folk to receive. When we say that free speech principles give us a marketplace of ideas, we are highlighting how these principles limit the ability of authorities to use censorship to impose paternalistic controls upon public discourse, much like markets stand in the way of authorities’ centralized, bureaucratic, and ultimately counterproductive controls on product allocation. The critical resemblance is in how free speech and consumer markets both spare us the troubles of having incompetent authorities deciding, on our behalf, what things, produced by others, we may access or consume.11

One objection to the metaphor is that the marketplace of ideas is rigged. The market doesn’t necessarily give people ideas and information they want. Often, instead, it gives people the ideas that ideologues and media corporations want them to have. What exists in most liberal societies is more like an oligopoly of ideas.12 The notion that we have a free market in ideas is libertarian mythology, distracting us from the oppressive power structures that are manifest in, and reified by, liberal society’s communication systems.13

But even if we grant the key premises here, this doesn’t problematize the marketplace metaphor’s prescriptive use. Suppose we are in a society where free speech rules are limiting content-based censorship—just as our metaphor recommends—but where media oligopolies wield great influence over public discourse. In this context, it would be spurious to suggest that public discourse is giving people the ideas they really want. If the marketplace metaphor is used as a way of conveying that suggestion, that’s bad. But it doesn’t nullify our worries about government control over ideas distribution, or make it illicit to highlight these worries using marketplace imagery. We might object to a fixation upon these worries that simultaneously overlooks nongovernment threats to the integrity of public discourse.14 But the problem there, again, is with the metaphor’s context-specific misuse, not with the validity of the moral concerns that it encapsulates. 

Another objection points to a mismatch between what friends of the marketplace of ideas want it to deliver and what it actually delivers, even if it hasn’t been transformed into an oligopoly. The English forefathers of free speech theory, Milton and Mill, seemed to believe that truths will outcompete falsehoods in an open contest.15 Our metaphor is often deployed in defense of this notion.16 However, so this second objection says, in a marketplace of ideas, people don’t reliably “buy” truths. People buy the ideas they like. And people don’t reliably like truths better than falsehoods. What the invisible hand does, all going well, is efficiently allocate goods to people based on what they want. Market-based systems of interaction will not magically popularize truths, then, any more than they will magically guarantee the popularity of higher-quality consumer products.17

All that this objection shows, though, is that some champions of the marketplace of ideas misconstrue their metaphor’s main lesson. Truth-based justifications for free speech are out of favor nowadays, largely replaced with claims about free speech’s role in realizing democracy.18 We have little reason to think free speech reliably furthers our epistemic aims (like truth, understanding), given the fragility of what we know about human rationality and credulity. It is still a mistake to believe that authorities know better than the folk which ideas are good for the folk to receive. But this isn’t because people are in fact great at judging what is plausible or who is credible.19 Rather, it’s a mistake because authorities have the same weaknesses on this front, as well as additional weaknesses that come with trying to advance the folk’s informational interests using centralized bureaucratic processes, which all-too-easily end up pre- or mis-judging complex issues.20

Free speech isn’t a royal road to truth. If it can be justified, it is with reference to other (for instance, democratic) ideals, and claims about how free speech principles help to realize them. The marketplace metaphor’s utility is in supplementing these justifications by highlighting the perennial risk of government overreach. Complex distribution networks, through which people with varied needs interact to try to fulfill their preferences, cannot be micromanaged by authorities, even decent authorities, much less inept or corrupt ones. Many factors bear on how we resist this overreach, in practice. But in principle, in complex networks, a decent strategy for satisfying preferences is to let people themselves decide what they want from whom, while limiting government’s power to dictate how things go. The marketplace metaphor has been one of our ways of culturally encoding this lesson over the last century and reminding ourselves of its relevance to free speech policy. 

So, what’s the problem? What other aspect of free speech policy should we emphasize, and how does the marketplace metaphor get in the way?

Here is a thought experiment. Suppose you’re in a world that has a well-functioning, universally accessible communication network—call it the ­System—that is used by nearly everyone and that has largely displaced the use of other communication tools, including other digital tools, as well as older options like telephone and mail. Use of the System isn’t legally mandated. But it is ubiquitously used all the same because it is a low-cost option that many find useful, and its widespread usage creates network effects that discourage opting out. Moreover, the System is a free speech zone, with few or no ideological constraints on content. Some criminal and tortious expression, which falls outside the coverage of free speech, is restricted. But otherwise, the System’s users can say whatever they please and engage with whomever they please. 

But suppose, also, crucially, that the System has a limited expressive palate, which to a nontrivial extent standardizes the style and format of people’s speech.

We can toy with the setup here depending on how realistic we want to make it. In a less fantastical version, we might imagine the System being roughly similar to Facebook. It is a text-based tool through which you can write posts of varying lengths and decide whether to let others comment. But the System still dictates a number of parameters. Very long messages must be broken up into shorter ones. You can embed links but not footnotes. Fonts and other visual features are uniform. And readers can react to posts using a menu of preset emojis. These parameters may only have a mildly homogenizing effect on how the System’s users communicate there. Nevertheless, the medium partly shapes people’s messages. 

If we wanted to make things more contrived, we could imagine the System being far more restrictive, for example by limiting messages to fifty characters, or not giving users any say over who sees their posts. Granted, the stricter and less user-­friendly we imagine the System, the less realistic it is to envision it as a widely preferred platform. But within the range of ways that the System could be set up, while plausibly retaining its global popularity, we can imagine it building in a more or less stringently homogenizing suite of expressive capabilities. The medium may only shape people’s messages subtly, but it may be more obtrusive. 

The System is a free speech zone, by stipulation. So if you are primarily worried about ideological control over communication networks, you should be happy, in theory, being in a world where the System is the dominant discursive hub. Indeed, we could further stipulate that the System isn’t just a free speech zone but that its governance makes its libertarian character counterfactually stable. The more modally robust the System is in protecting speech, the happier you should be having it locked-in as a dominant discursive hub.21 And if it gains dominance because everyone freely opts into it, then what is there to worry about? 

There is something blinkered in that perspective. If the System’s limited expressive affordances—combined with its ubiquity—homogenize the world’s methods and styles of communication, then something valuable is imperiled, if not already lost. We should worry about the openness and variety of the communicative formats available to people for the same kinds of reasons that we worry about the openness and variety of the viewpoints people are allowed to convey. In both cases, variety and openness support people’s ability to think deeply, and think for themselves. Much like the homogenization of ideas that people can express, homogenization in the style and format in which people can express ideas is liable to inhibit people’s ability to critically reflect upon what they share and hear.22 Whether you care about free speech for democratic reasons, or truth-seeking reasons, or because you place a high moral value on individual autonomy, this inhibition of people’s critical faculties is bad news.

The marketplace metaphor tells us to worry about communication policies that are akin to price-fixing or five-year agricultural plans—to worry about authorities deciding what ideas we receive. Simultaneously, it invites a neutral or positive view of policies that expedite the exchange of ideas. Just like free trade pacts make it easier to exchange goods, ideologically open communication hubs lubricate the flow of ideas. Any big institution can be corrupted, of course. But as long as our central hubs are not commandeered by bad actors, we should be pleased to have them. When operating within the marketplace metaphor’s normative horizons, we have no more reason to worry about the ubiquity and uniformity of the System than to worry about free trade pacts or stable currency exchanges.23

This brings the metaphor’s principal drawbacks into focus. Even a gung ho free market fanatic should recognize that the trade in ideas is critically unlike a trade in goods and services. The platforms mediating our idea transactions more deeply condition the character and texture of—and thus, potentially, affect the quality of—what is being exchanged. Communicative life under the System would make it harder for us to transact in certain kinds of ideas, while also homogenizing and thus depleting the richness and vitality of the cognitive activities involved in those transactions.24 The marketplace metaphor’s spotlight keeps all of this in the shadows, and hence dampens the anxieties we should be feeling about the homogenizing forces that are bearing down on our communication networks. 

These homogenizing forces, to which the marketplace metaphor desensitizes us, are precisely what the connected-city metaphor encourages us to worry about. 

Cars are useful. But it is hard living in a big city like Los Angeles where cars are the only decent way to get around. The car’s privacy and manoeuvrability upsides have corresponding downsides, like space inefficiency. Vast tracts of land must be turned into roads and parking lots in order for the car’s benefits to be realized. And excessively asphalted places are tough to inhabit, to say nothing of how issues of dynamic demand and static supply make traffic jams near-inevitable. This doesn’t mean that roads and cars have no place in a locomotively optimized city. People with mobility challenges can’t always catch trains. No one wants to ride a bike to the hospital to give birth. And roads accommodate buses, delivery vans, ambulances, and fire trucks, as well as cars. The problem with places like Los Angeles isn’t just that they have loads of cars, it’s that they lack (quality versions of) other transport options. 

What is it that makes connected cities—cities with good trains, buses, roads, bicycle lanes, and walking paths that are all linked up so we can move from one to another—more liveable? First, people have diverse locomotive needs, depending on their age, fitness, and sensory/mobility capacities. Second, people have diverse locomotive desires. Some people like walking and cycling, others don’t. Third, locomotive needs and desires vary circumstantially, depending on the weather, what we drank last night, whether it’s a busy day, or whether we are moving tricky cargo, like a cake, a bassoon, or a toddler. Fourth, our locomotive needs can change if we are traveling solo or in smaller or larger groups.25

In light of this diversity, the connected city’s mixture of locomotive affordances makes the incomprehensibly intricate collective choreography of urban transport more fluent at a group level and less frustrating for individuals. It also partly mitigates the drabness and dreariness of a landscape smothered in asphalt.26

When we communicate, we share information and beliefs, while also trading in a range of subtler sociolinguistic currencies.27 The conveyance of these things in speech isn’t perfectly analogous to the conveyance of myself or my family around a city. But it is analogous enough for our purposes. The superficial layer is easy to grasp. Humans have diverse communicative needs and desires much like our diverse locomotive needs and desires. Just as some people can’t easily catch trains and therefore need taxis, some people can’t easily write long emails and therefore need to be able to leave voice messages. Just as some people like bikes but not buses, some people like texting but hate going back-and-forth on X (formerly Twitter).

The analogy’s deeper layers need more unpacking. Apart from pressing concerns about the need to limit our carbon dioxide emissions, we might think of transport options simply as a means to our ends. Other things being equal (time, cost, ease, accessibility), you may be fairly indifferent about whether you travel to work on a bike or a train. It seems like a mistake, though, to approach modes of speech in this blasé frame of mind. Even if two expressive tools are on par, vis-à-vis time, cost, ease, and so on, if the expression that’s involved matters to you, then the choice of medium should matter too. An intimate conversation about a delicate issue might go very differently in person versus by phone. Or think about trying to make a persuasive argument. Your odds of nailing it can vary enormously depending on whether you produce an essay, a podcast, or a tweetstorm, and on how your argument’s specifics and nuances—combined with your own communicative abilities—lend themselves to your chosen medium’s communicative affordances.

The benefits of a connected city of ideas—a system in which we can readily utilize various communicative tools, with varied affordances—seem to run deeper than the benefits of a locomotively connected city. In a liberal society, we want locations and ideas to be accessible to everyone interested in them. Diverse locomotive and communicative options support both kinds of access. Often, though, accessing locations is purely about logistics. With ideas, the means of access are less fungible. Some ideas might not be communicable—not as easily and fluently, or in all their specificity and subtlety—except through a particular medium: a documentary film, a satirical essay, a piece of long-form investigative journalism, a talk radio discussion, a meme on a WhatsApp group, or a slowly unfolding face-to-face conversation. Part of how a connected city of ideas works is by offering assorted communicative options to groups with diverse expressive predilections. But it also makes it easier for anyone’s communicative aims to be pursued in media that are better-suited to their realization, style- and format-wise.

So, here is the argument boiled down. Communicative homogenization is something we should worry about from a free speech perspective. The connected-city metaphor highlights this worry. The marketplace metaphor, which dominates free speech discourse, obscures it. So we should replace the latter with the former.

We can buttress the argument by working through three objections. First, why think that the homogenization of communication methods has the negative effects I am claiming? Second, in what way is homogenization a genuine danger, either now or under emerging technological conditions? Third, why situate these antihomogenization worries within the ambit of a free speech politics?

Objection 1 is simply a flat-footed skeptical rejoinder. Why think that homogenization is such a bad thing? That our communicative abilities are inhibited by the homogenization of our communication tools and practices? Yes, arguments or intimate chats are liable to go differently in different formats. But this is merely due to life’s complexity and unpredict­ability. There is nothing about an essay, a podcast, a phone call, or an intimate face-to-face conversation that dictates what kinds of ideas or other sociolinguistic stuff can be exchanged within it.

What can we say to this? I have mentioned affordances at a few points. I am using this term in the sense pioneered by psychologist James Gibson in the 1960s, and taken up in various research programs in the sciences, such as perceptual psychology, and the humanities, such as philosophy of action.28 In its simplest form, the idea is that locations and objects make some opportunities for action more available to agents than others—certain things that are easier for agents to do or perceive. The deeper conceptual thesis in the background is that thought and action aren’t products of free-floating minds, but of material beings interacting with environments. Agency emerges out of organism-environment interactions.29

Insofar as this is a sound portrayal of agency’s underpinnings, it seems like an important starting point for critical thinking about the moral implications of changing technologies. Technological innovations re-landscape the agential environment from which thoughts and actions interactively arise. As philosopher Shannon Vallor has noted, technologies “afford specific patterns of thought, behavior, and valuing,” while opening up “new possibilities for human action, and foreclos[ing] or obscur[ing] others.”30 Acts of creating or adopting communicative devices are, then, we might say, meta acts that alter the choice architectures contained in an environment. We can differentiate technologies from mere tools in terms of how much they change our sense of which thoughts/actions are available to think/perform.31

This view of affordances and technology undermines the idea that technologies are merely utensils that we use as we please. It helps us see why this idea is misleading, in the same way that “guns don’t kill people, people kill people” is misleading. Tools and technologies elicit certain usages, not inexorably, but probabilistically. They are integral to the processes by which preferences form. 

Trying to fully vindicate this picture would take us far afield. How plausible you find it, in general, or with respect to communicative media specifically, will probably depend on how it chimes with your own experience. Some people may have a livelier sense of how communicative media shape their thinking. Others, including people whose expressive abilities are well-suited to a variety of media, may not feel this much at all. In any case, to say that media provide affordances for thought that affect our critical and interpretative abilities isn’t to say that we are all affected to the same extent or in the same way. Moreover, we can ultimately concede the skeptic’s point that the communicative medium does not dictate its contents. After all, dictation is an overloaded way of characterizing the type of interactive, probabilistic influence that an affordance exerts. At one point in his most famous work, cultural critic Neil Postman—an author who is deeply invested in the affordances framework that I am endorsing—writes:

How we are obliged to conduct . . . conversations will have the strongest possible influence on what ideas we can conveniently express. And what ideas are convenient to express inevitably become the important content of a culture.32

Expressive media have a formidable influence by Postman’s lights—indeed, “the strongest possible.” But the upshot isn’t to dictate exactly which ideas we can access. Rather, our media affect which ideas become communicatively convenient and, in turn, become ready reference points in our culture. Theorizing expressive affordances in a plausible way seems to require some caveat along these lines. A homogenized communicative milieu, style- and format-wise, probably won’t make any communicative purposes totally unachievable. But whichever styles and formats predominate, they will make certain communicative aims easier to realize, and others harder, in ways that influence everyone’s reflective capabilities.33

Patricia Lockwood’s No One Is Talking about This is one of the better (and funnier) English-language novels to date about being on the internet. Its protagonist is interested in how online platforms—she calls them, collectively, the portal—are formatting her language and configuring her thinking. She wonders:

Why were we all writing like this now? Because a new kind of connection had to be made, and blink, synapse, little space between was the only way to make it. Or because, and this was more frightening, it was the way the portal wrote.34

I don’t think Lockwood is misguided in these apprehensions. The advent of social media platforms has given us unfamiliar ways of writing and speaking, and thus, via some alchemy of form and content, novel thoughts. New communication tools have shepherded us toward new ways of accessing and traversing ideas. Were these mental routes totally unreachable before? Maybe not. But accessible tracks into them have been trodden and cleared by many pairs of feet. When things go the other way, though—when communicative options are subtracted or standardized—the opposite occurs. We lose some of our cognitive affordances.

But why believe this homogenization is a genuine danger? Don’t the points I have just made belie the anxieties I am trying to provoke? New communication technologies provide new affordances for thinking and speaking. No one is taking existing affordances away! Some devices, like fax machines, fall into disuse organically, taking their redundant expressive affordances (the redundancy of which is manifest in their organic demise) with them. Overall, though, our repertoire of communicative options, and the richness and diversity of the communicative affordances that they provide us with, needn’t become depleted. Right?

When it comes to futurological claims, we are all in the same speculative boat. This essay appears in a volume on the future of free speech, and I wanted to write it because I have hunches about the trajectory we are tracing vis-à-vis free speech’s future that seem to be out-of-step with many other people’s hunches. If you believe the possibilities I am fretting about are far-fetched, then you won’t see much reason to embrace my proposed shake-up of metaphors. But it is hard to turn futurological claims into anything more than guesstimates. The best I can do is to state the factors that underlie my key hunch: that on our current trajectory, a significant homogenization in our communicative media is likely in our lifetimes. What factors point to such a trajectory?

1. Monopolies. Plausibly, we are seeing a historically unprecedented level of centralization and monopolization in the ownership and management structures of widely used communication technologies, including devices and platforms whose operations are radically global in scope.35

2. Privatization. Many countries have ailing public communication infrastructure, such as phone lines, broadcasting facilities, cable internet, and postal services. The pressures on maintaining public communication utilities may either lead to their collapse, or otherwise enable cashed-up global tech corporations to acquire legacy communication infrastructures and incorporate them into cross-platform networks.36

3. Compulsion. In many countries, it is hard to participate in public life without a smartphone. In some sub-enclaves, social media is similarly de facto mandatory. Participation in public life may always force people to use particular communicative media. But the demandingness of these requirements vis-à-vis the captivating potential of the technologies that they mandate seems historically unprecedented.37

4. Biointegration. The advent of commercially available biointegrated communication technologies, like Neuralink, is just around the corner. It seems possible that by virtue of their biointegrated nature, these technologies will create more resilient network effects compared with existing technologies. This would amplify the costs for those preferring to opt out, thus strengthening the de facto mandates noted above.38

5. Stylistic standardization. Widely used assistive technologies like Grammarly stylistically standardize written expression in unprecedented ways: standardization is faster and more integrated with otherwise familiar expressive tools and affordances than ever before. The power, ubiquity, and integration of these technologies are steadily increasing.39

6. Linguistic standardization. Many languages are dying, English is increasingly entrenched as a global lingua franca, and autotranslation tools are becoming more powerful. Plausibly, the combination of these factors will mean that a larger portion of global communication is conducted in standardized and expressively flattened languages.40

What does this all add up to, once all the relevant counterforces are factored into our conjectures? Could we soon be living under a ubiquitous, homogenizing communication network, like the System, if only to a lesser degree? Some homogenizing forces, potentiated by emerging technologies, are likely to have an impact on near-future communication systems. How these merge with other economic and political forces, and whether everything gets derailed by cataclysmic events, is a giant unknown. But I don’t think the homogenization anxieties are baseless. 

Of course, so much depends on how the tools and practices we end up with interact with each other, and with other political and economic forces. As the supporters of new communication technologies like to point out, there were once great panics over novels and radio. The fact that a transformative suite of communicative media is becoming dominant isn’t yet a reason to think that our communicative interests are in peril. Social systems are super complex, and the devil is in the details. But it is complacent to use these observations as an excuse to ignore worries about homogenization, as we have largely been doing. People who are more optimistic about our current technological trajectory need to explain either why communicative homogenization isn’t going to occur, or why it isn’t such a bad thing. It isn’t enough to just circle back to banal reminders about how society survived the printing press and wireless radio. 

Even if that’s right, why think that these antihomogenization issues fall within the ambit of free speech theory? Free speech principles are principles of restraint, which limit the means governments may use—via legislation or direct coercive and administrative acts—to interfere with the exercise of people’s speech rights. They aren’t principles that oblige governments (or other actors) to support speech or to otherwise try to realize the interests that speech rights purportedly serve. Even if communicative homogenization poses a real threat to human well-being, remedying this problem isn’t on the agenda for free speech politics, except where antihomogenization measures happen to coincide with constraining governments from impinging upon people’s speech rights.

This isn’t an idiosyncratic view.41 But I favor a more capacious conception of free speech, encompassing both duties of restraint and “positive” duties to support our speech-related interests. This is actually an old-fashioned view, from Mill. In the nineteenth-century liberal mind, free speech isn’t just about constraining state power. It is about everyone working to achieve a culture of open discussion, free from conformist pressures of all kinds.42 Even in American constitutional law, in which the narrower, negative conception of free speech principles is widely accepted, there are good reasons to think that the efficacy of these principles depends upon them operating in a culture supportive of free speech.43 And insofar as that’s true, it’s somewhat arbitrary to situate positive speech-related duties outside of free speech’s domain, in some bundle of adjacent supplementary norms.

Recent U.S. legal scholarship on free speech and tech policy lends support to the more capacious conception. In his work on platform regulation, legal scholar Jack Balkin defends a triangulated model of free speech. Free speech isn’t just about states not interfering with citizens. It is a three-way relation between states, citizens, and expressive platforms, in which one of the state’s duties is to create a regulatory environment that incentivizes platforms to support citizens’ expressive interests. Balkin’s overall argument, roughly, is that speech platforms are essential in realizing healthy public discourse, and that a dyadic state-versus-citizen notion of free speech obscures this, while encouraging us to disempower states from fulfilling their regulatory responsibilities in constructively incentivizing platforms. If we follow Balkin in replacing the dyadic framework with a triangulated framework, this ipso facto means including certain positive state duties—that is, duties to regulate speech platforms—within the “official” scope of free speech principles and policies.44

In a similar vein, legal scholar Evelyn Douek argues that the regulation of speech platforms should take a systems approach. It should embed upstream norms—promoting good speech and algorithmically suppressing harmful speech—rather than norms that, in a downstream fashion, identify and remove harmful instances of speech based on case-by-case appraisals of their harmfulness. Her argument, in essence, is that the latter approach is unfeasible with large platforms, given the scale of the regulatory task and the need for relatively quick action, since speech’s harmful potential often depends on how long it remains visible.45 We could still situate the norms for regulating platforms outside the scope of free speech. But this seems odd, given that, again, such norms are essential to realizing the discursive ideals that free speech theory has long revolved around. It seems natural, instead, to include these norms—norms that don’t impose constraints on states, but rather, positive duties of constructive discourse organization—within the scope of free speech. And this means embracing the broader conception. 

It isn’t stretching our normative categories beyond their proper bounds, then, to see a demand for antihomogenization regulations, in government or the private sector, as part of a free speech politics. What might these regulations consist of? They may include 1) special antimonopoly laws for tech companies, 2) regulations to make it harder for public communication utilities to be privatized, and 3) laws protecting workers from being unnecessarily forced to use specific communication platforms. Mapping out the goals of an antihomogenization regulatory agenda, let alone its details, is a big task.46 I’m setting the stage. I want us to see why this agenda warrants our attention as free speech theorists, and, crucially, to see how free speech metaphors can sensitize or desensitize us to the concerns driving them. We need different imagery to enliven our perceptions. We have to shake off the interpretative languor that the marketplace metaphor, despite its legitimate uses, has instilled in us. 

Even if you accept all of my arguments up to this point, you might nonetheless think I haven’t shown that communicative homogenization is a more serious worry than ideological control over the content of public discourse. You might argue that we should only redirect our attention toward the issues highlighted by the connected-city metaphor, and away from the marketplace metaphor’s implicit anti-­authoritarian agenda, if we find that this comparative moral assessment holds.

But there are other good reasons to adjust our focus. Moral metaphors are imaginative devices for highlighting underappreciated aspects of things. Free speech theory’s normative spotlight has long been illuminating anti-authoritarian concerns. I am not questioning those concerns themselves, so much as the amount of attention we have lavished upon them. In any case, whatever priority ordering ought to obtain among these concerns, my earlier point remains. The marketplace metaphor isn’t only failing to highlight antihomogenization worries. It is camouflaging them. The point of the connected-city metaphor is to attune us to what ought to be an urgent concern in free speech theory, but one that our leading metaphor has encouraged us to tune out.

author’s note

For advice and feedback on this essay, many thanks to Lee Bollinger, Noah Chafets, Jonathan Gingerich, Rich Healey, Polly Mitchell, Ethan Nowak, Francesca Perry, Daniel Rothschild, Geof Stone, and Rony Yuria.