Digifesto

Further distillation of Bostrom’s Superintelligence argument

Following up on this outline of the definitions and core argument of Bostrom’s Superintelligence, I will try to narrow in on the key mechanisms the argument depends on.

At the heart of the argument are a number of claims about instrumentally convergent values and self-improvement. It’s important to distill these claims to their logical core because their validity affects the probability of outcomes for humanity and the way we should invest resources in anticipation of superintelligence.

There are a number of ways to tighten Bostrom’s argument:

Focus the definition of superintelligence. Bostrom leads with the provocative but fuzzy definition of superintelligence as “any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest.” But the overall logic of his argument makes it clear that the domain of interest does not necessarily include violin-playing or any number of other activities. Rather, the domains necessary for a Bostrom superintelligence explosion are those that pertain directly to improving ones own intellectual capacity. Bostrom speculates about these capacities in two ways. In one section he discusses the “cognitive superpowers”, domains that would quicken a superintelligence takeoff. In another section he discusses convergent instrumental values, values that agents with a broad variety of goals would converge on instrumentally.

  • Cognitive Superpowers
    • Intelligence amplification
    • Strategizing
    • Social manipulation
    • Hacking
    • Technology research
    • Economic productivity
  • Convergent Instrumental Values
    • Self-preservation
    • Goal-content integrity
    • Cognitive enhancement
    • Technological perfection
    • Resource acquisition

By focusing on these traits, we can start to see that Bostrom is not really worried about what has been termed an “Artificial General Intelligence” (AGI). He is concerned with a very specific kind of intelligence with certain capacities to exert its will on the world and, most importantly, to increase its power over nature and other intelligent systems rapidly enough to attain a decisive strategic advantage. Which leads us to a second way we can refine Bostrom’s argument.

Closely analyze recalcitrance. Recall that Bostrom speculates that the condition for a fast takeoff superintelligence, assuming that the system engages in “intelligence amplification”, is constant or lower recalcitrance. A weakness in his argument is his lack of in-depth analysis of this recalcitrance function. I will argue that for many of the convergent instrumental values and cognitive superpowers at the core of Bostrom’s argument, it is possible to be much more precise about system recalcitrance. This analysis should allow us to determine to a greater extent the likelihood of singleton vs. multipolar superintelligence outcomes.

For example, it’s worth noting that a number of the “superpowers” are explicitly in the domain of the social sciences. “Social manipulation” and “economic productivity” are both vastly complex domains of research in their own right. Each may well have bounds about how effective an intelligent system can be at them, no matter how much “optimization power” is applied to the task. The capacities of those manipulated to understand instructions is one such bound. The fragility or elasticity of markets could be another such bound.

For intelligence amplification, strategizing, technological research/perfection, and cognitive enhancement in particular, there is a wealth of literature in artificial intelligence and cognitive science that addresses the technical limits of these domains. Such technical limitations are a natural source of recalcitrance and an impediment to fast takeoff.

Bostrom’s Superintelligence: Definitions and core argument

I wanted to take the opportunity to spell out what I see as the core definitions and argument of Bostrom’s Superintelligence as a point of departure for future work. First, some definitions:

  • Superintelligence. “We can tentatively define a superintelligence as any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest.” (p.22)
  • Speed superintelligence. “A system that can do all that a human intellect can do, but much faster.” (p.53)
  • Collective superintelligence. “A system composed of a large number of smaller intellects such that the system’s overall performance across many very general domains vastly outstrips that of any current cognitive system.” (p.54)
  • Quality superintelligence. “A system that is at least as fast as a human mind and vastly qualitatively smarter.” (p.56)
  • Takeoff. The event of the emergence of a superintelligence. The takeoff might be slow, moderate, or fast, depending on the conditions under which it occurs.
  • Optimization power and Recalcitrance. Bostrom’s proposed that we model the speed of superintelligence takeoff as: Rate of change in intelligence = Optimization power / Recalcitrance. Optimization power refers to the effort of improving the intelligence of the system. Recalcitrance refers to the resistance of the system to being optimized.(p.65, pp.75-77)
  • Decisive strategic advantage. The level of technological and other advantages sufficient to enable complete world domination. (p.78)
  • Singleton. A world order in which there is at the global level one decision-making agency. (p.78)
  • The wise-singleton sustainability threshold. “A capability set exceeds the wise-singleton threshold if and only if a patient and existential risk-savvy system with that capability set would, if it faced no intelligent opposition or competition, be able to colonize and re-engineer a large part of the accessible universe.” (p.100)
  • The orthogonality thesis. “Intelligence and final goals are orthogonal: more or less any level of intelligence could in principle be combined with more or less any final goal.” (p.107)
  • The instrumental convergence thesis. “Several instrumental values can be identified which are convergent in the sense that their attainment would increase the chances of the agent’s goal being realized for a wide range of final goals and a wide range of situations, implying that these instrumental values are likely to be pursued by a broad spectrum of situated intelligent agents.” (p.109)

Bostrom’s core argument in the first eight chapters of the book, as I read it, is this:

  1. Intelligent systems are already being built and expanded on.
  2. If some constant proportion of a system’s intelligence is turned into optimization power, then if the recalcitrance of the system is constant or lower, then the intelligence of the system will increase at an exponential rate. This will be a fast takeoff.
  3. Recalcitrance is likely to be lower for machine intelligence than human intelligence because of the physical properties of artificial computing systems.
  4. An intelligent system is likely to invest in its own intelligence because of the instrumental convergence thesis. Improving intelligence is an instrumental goal given a broad spectrum of other goals.
  5. In the event of a fast takeoff, it is likely that the superintelligence will get a decisive strategic advantage, because of a first-mover advantage.
  6. Because of the instrumental convergence thesis, we should expect a superintelligence with a decisive strategic advantage to become a singleton.
  7. Machine superintelligences, which are more likely to takeoff fast and become singletons, are not likely to create nice outcomes for humanity by default.
  8. A superintelligent singleton is likely to be above the wise-singleton threshold. Hence the fate of the universe and the potential of humanity is at stake.

Having made this argument, Bostrom goes on to discuss ways we might anticipate and control the superintelligence as it becomes a singleton, thereby securing humanity.

And now for something completely different: Superintelligence and the social sciences

This semester I’ll be co-organizing, with Mahendra Prasad, a seminar on the subject of “Superintelligence and the Social Sciences”.

How I managed to find myself in this role is a bit of a long story. But as I’ve had a longstanding curiosity about this topic, I am glad to be putting energy into the seminar. It’s a great opportunity to get exposure to some of the very interesting work done by MIRI on this subject. It’s also a chance to thoroughly investigate (and critique) Bostrom’s book Superintelligence: Paths, Dangers, and Strategies.

I find the subject matter perplexing because in many ways it forces the very cultural and intellectual clash that I’ve been preoccupied with elsewhere on this blog: the failure of social scientists and engineers to communicate. Or, perhaps, the failure of qualitative researchers and quantitative researchers to communicate. Whatever you want to call it.

Broadly, the question at stake is: what impact will artificial intelligence have on society? This question is already misleading since in the imagination of most people who haven’t been trained in the subject, “artificial intelligence” refers to something of a science fiction scenario, whereas to practitioner, “artificial intelligence” is, basically, just software. Just as the press went wild last year speculating about “algorithms”, by which it meant software, so too is the press excited about artificial intelligence, which is just software.

But the concern that software is responsible for more and more of the activity in the world and that it is in a sense “smarter than us”, and especially the fear that it might become vastly smarter than us (i.e. turning into what Bostrom calls a “superintelligence”), is pervasive enough to drive research funding into topics like “AI Safety”. It also is apparently inspiring legal study into the regulation of autonomous systems. It may also have implications for what is called, vaguely, “social science”, though increasingly it seems like nobody really knows what that is.

There is a serious epistemological problem here. Some researchers are trying to predict or forewarn the societal impact of agents that are by assumption beyond their comprehension on the premise that they may come into existence at any moment.

This is fascinating but one has to get a grip.

intelligibility

The example of Arendt’s dismissal of scientific discourse from political discussion underscores a much deeper political problem: a lack of intelligibility.

Every language is intelligible to some people and not to others. This is obviously true in the case of major languages like English and Chinese. It is less obvious but still a problem with different dialects of a language. It becomes a source of conflict when there is a lack of intelligibility between the specialized languages of expertise or personal experience.

For many, mathematical formalism is unintelligible; it appears to be so for Arendt, and this disturbs her, as she locates politics in speech and wants there to be political controls on scientists. But how many scientists and mathematicians would find Arendt intelligible? She draws deeply on concepts from ancient Greek and Augustinian philosophy. Are these thoughts truly accessible? What about the intelligibility of the law, to non-lawyers? Or the intelligibility of spoken experiences of oppression to those who do not share such an experience?

To put it simply: people don’t always understand each other and this poses a problem for any political theory that locates justice in speech and consensus. Advocates of these speech-based politics are most often extraordinarily articulate and write persuasively about the need to curtail the power of any systems of control that they do not understand. They are unable to agree to a social contract that they cannot read.

But this persuasive speech is necessarily unable to account for the myriad mechanisms that are both conditions for the speech and unintelligible to the speaker. This includes the mechanisms of law and technology. There is a performative contradiction between these persuasive words and their conditions of dissemination, and this is reason to reject them.

Advocates of bureaucratic rule tend to be less eloquent, and those that create technological systems that replace bureaucratic functions even less so. Nevertheless each group is intelligible to itself and may have trouble understanding the other groups.

The temptation for any one segment of society to totalize its own understanding, dismissing other ways of experiencing and articulating reality as inessential or inferior, is so strong that it can be read in even great authors like Arendt. Ideological politics (as opposed to technocratic politics) is the conflict between groups expressing their interests as ideology.

The problem is that in order to function as it does at scale, modern society requires the cooperation of specialists. Its members are heterogeneous; this is the source of its flexibility and power. It is also the cause of ideological conflict between functional groups that should see themselves as part of a whole. Even if these members do see their interdependence in principle, their specialization makes them less intelligible. Articulation often involves different skills from action, and teaching to the uninitiated is another skill altogether. Meanwhile, the complexity of the social system expands as it integrates more diverse communities, reducing further the proportion understood by a single member.

There is still in some political discourse the ideal of deliberative consensus as the ground of normative or political legitimacy. Suppose, as seems likely, that this is impossible for the perfectly mundane and mechanistic reason that society is so complicated due to the demands of specialization that intelligibility among its constituents is never going to happen.

What then?

the state and the household in Chinese antiquity

It’s worthwhile in comparison with Arendt’s discussion of Athenian democracy to consider the ancient Chinese alternative. In Alfred Huang’s commentary on the I Ching, we find this passage:

The ancient sages always applied the principle of managing a household to governing a country. In their view, a country was simply a big household. With the spirit of sincerity and mutual love, one is able to create a harmonious situation anywhere, in any circumstance. In his Analects, Confucius says,

From the loving example of one household,
A whole state becomes loving.
From the courteous manner of one household,
A whole state becomes courteous.

Comparing the history of Europe and the rise of capitalistic bureaucracy with the history of China, where bureaucracy is much older, is interesting. I have comparatively little knowledge of the latter, but it is often said that China does not have the same emphasis on individualism that you find in the West. Security is considered much more important than Freedom.

The reminder that the democratic values proposed by Arendt and Horkheimer are culturally situated is an important one, especially as Horkheimer claims that free burghers are capable of producing art that expresses universal needs.

a refinement

If knowledge is situated, and scientific knowledge is the product of rational consensus among diverse constituents, then a social organization that unifies many different social units functionally will have a ‘scientific’ ideology or rationale that is specific to the situation of that organization.

In other words, the political ideology of a group of people will be part of the glue that constitutes the group. Social beliefs will be a component of the collective identity.

A social science may be the elaboration of one such ideology. Many have been. So social scientific beliefs are about capturing the conditions for the social organization which maintains that belief. (c.f. Nietzsche on tablets of values)

There are good reasons to teach these specialized social sciences as a part of vocational training for certain functions. For example, people who work in finance or business can benefit from learning economics.

Only in an academic context does the professional identity of disciplinary affiliation matter. This academic political context creates great division and confusion that merely reflects the disorganization of the academic system.

This disorganization is fruitful precisely because it allows for individuality (cf. Horkheimer). However, it is also inefficient and easy to corrupt. Hmm.

Against this, not all knowledge is situated. Some is universal. It’s universality is due to its pragmatic usefulness in technical design. Since technical design acts on everyone even when their own situated understanding does not include it, this kind of knowledge has universal ground (in violence, sadly, but maybe also in other ways.)

The question is whether there is room anywhere in the technically correct understanding of social organization (something we might see in Beniger) there is room for the articulation of what it supposed to be great and worthy of man (see Horkheimer).

I have thought for a long time that there is probably something like this describable in terms of complexity theory.

structuralism and/or functionalism

Previous entries detailing the arguments of Arendt, Horkheimer, and Beniger show these theorists have what you might call a structural functionalist bent. Society is conceived as a functional whole. There are units of organization within it. For Arendt, this social organization begins in the private household and expands to all of society. Horkheimer laments this as the triumph of mindless economic organization over genuine, valuable individuality.

Structuralism, let alone structural functionalism, is not in fashion in the social sciences. Purely speculatively, one reason for this might be that to the extent that society was organized to perform certain functions, more of those functions have been delegated to information processing infrastructure, as in Beniger’s analysis. That leaves “culture” more a domain of ephemerality and identity conflict, as activity in the sphere of economic production becomes if not private, opaque.

My empirical work on open source communities is suggestive (though certainly not conclusively so) that these communities are organized more for functional efficiency than other kinds of social groups (including academics). I draw this inference from the degree dissortativity of the open source social networks. Disassortativity suggests the interaction of different kinds of people, which is against homophilic patterns of social formation but which seems essential for economic activity where the interact of specialists is what creates value.

Assuming that society its entirety (!!) is very complex and not easily captured by a single grand theory, we can nevertheless distinguish difference kinds of social organization and see how they theorize themselves. We can also map how they interact and what mechanisms mediated between them.

Land and gold (Arendt, Horkheimer)

I am thirty, still in graduate school, and not thrilled about the prospects of home ownership since all any of the professionals around me talk about is the sky-rocketing price of real estate around the critical American urban centers.

It is with a leisure afforded by graduate school that I am able to take the long view on this predicament. It is very cheap to spend ones idle time reading Arendt, who has this to say about the relationship between wealth and property:

The profound connection between private and public, manifest on its most elementary level in the question of private property, is likely to be misunderstood today because of the modern equation of property and wealth on one side and propertylessness and poverty on the other. This misunderstanding is all the more annoying as both, property as well as wealth, are historically of greater relevance to the public realm than any other private matter or concern and have played, at least formally, more or less the same role as the chief condition for admission to the public realm and full-fledged citizenship. It is therefore easy to forget that wealth and property, far from being the same, are of an entirely different nature. The present emergence everywhere of actually or potentially very wealthy societies which at the same time are essentially propertyless, because the wealth of any single individual consists of his share in the annual income of society as a whole, clearly shows how little these two things are connected.

For Arendt, beginning with her analysis of ancient Greek society, property (landholding) is the condition of ones participation in democracy. It is a place of residence and source of ones material fulfilment, which is a prerequisite to ones free (because it is unnecessitated) participation in public life. This is contrasted with wealth, which is a feature of private life and is unpolitical. In ancient society, slaves could own wealth, but not property.

If we look at the history of Western civilization as a progression away from this rather extreme moment, we see the rise of social classes whose power is based on in landholding but in wealth. Industrialism and the economy based on private ownership of capital is a critical transition in history. That capital is not bound to a particular location but rather is mobile across international boundaries is one of the things that characterizes global capitalism and brings it in tension with a geographically bounded democratic state. It is interesting that a Jeffersonian democracy, designed with the assumption of landholding citizens, should predate industrial capitalism and be consitutionally unprepared for the result, but nevertheless be one of the models for other democratic governance structures throughout the world.

If private ownership of capital, not land, defines political power under capitalism, then wealth, not property, becomes the measure of ones status and security. For a time, when wealth was as a matter of international standard exchangeable for gold, private ownership of gold could replace private ownership of land as the guarantee of ones material security and thereby grounds for ones independent existence. This independent, free rationality has since Aristotle been the purpose (telos) of man.

In the United States, Franklin Roosevelt’s 1933 Executive Order 6102 forbade the private ownership of gold. The purpose of this was to free the Federal Reserve of the gold market’s constraint on increasing the money supply during the Great Depression.

A perhaps unexpected complaint against this political move comes from Horkheimer (Eclipse of Reason, 1947), who sees this as a further affront to individualism by capitalism.

The age of vast industrial power, by eliminating the perspectives of a stable past and future that grew out of ostensibly permanent property relations, is the process of liquidating the individual. The deterioration of his situation is perhaps best measured in terms of his utter insecurity as regards to his personal savings. As long as currencies were rigidly tied to gold, and gold could flow freely over frontiers, its value could shift only within narrow limits. Under present-day conditions the dangers of inflation, of a substantial reduction or complete loss of the purchasing power of his savings, lurks around the next corner. Private possession of gold was the symbol of bourgeois rule. Gold made the burgher somehow the successor of the aristocrat. With it he could establish security for himself and be reasonable sure that even after his death his dependents would not be completely sucked up by the economic system. His more or less independent position, based on his right to exchange goods and money for gold, and therefore on the relatively stable property values, expressed itself in the interest he took in the cultivation of his own personality–not, as today, in order to achieve a better career or for any professional reason, but for the sake of his own individual existence. The effort was meaningful because the material basis of the individual was not wholly unstable. Although the masses could not aspire to the position of the burgher, the presence of a relatively numerous class of individuals who were governed by interest in humanistic values formed the background for a kind of theoretical thought as well as for the type of manifestions in the arts that by virtue of their inherent truth express the needs of society as a whole.

Horkheimer’s historical arc, like many Marxists, appears to ignore its parallels in antiquity. Monetary policy in the Roman Empire, which used something like a gold standard, was not always straightforward. Inflation was sometimes a severe problem when generals would print money to pay the soldiers hat supported their political coups. So it’s not clear that the modern economy is more unstable than gold or land based economies. However, the criticism that economic security is largely a matter of ones continued participation in a larger system, and that there is little in the way of financial security besides this, holds. He continues:

The state’s restriction on the right to possess gold is the symbol of a complete change. Even the members of the middle class must resign themselves to insecurity. The individual consoles himself with the thought that his government, corporation, association, union, or insurance company will take care of him when he becomes ill or reaches the retiring age. The various laws prohibiting private possession of gold symbolize the verdict against the independent economic individual. Under liberalism, the beggar was always an eyesore to the rentier. In the age of big business both beggar and rentier are vanishing. There are no safety zones on society’s thoroughfares. Everyone must keep moving. The entrepreneur has become a functionary, the scholar a professional expert. The philosopher’s maxim, Bene qui latuit, bene vixit, is incompatible with the modern business cycles. Everyone is under the whip of a superior agency. Those who occupy the commanding positions have little more autonomy than their subordinates; they are bound by the power they wield.

In an academic context, it is easy to make a connection between Horkheimer’s concerns about gold ownership and tenure. Academic tenure is or was the refuge of the individual who could in theory develop themselves as individuals in obscurity. The price of this autonomy, which according the philosophical tradition represents the highest possible achievement of man, is that one teaches. So, the developed individual passes on the values developed through contemplation and reflection to the young. The privatization of the university and the emphasis on teaching marketable skills that allow graduates to participate more fully in the economic system is arguably an extension of Horkheimer’s cultural apocalypse.

The counter to this is the claim that the economy as a whole achieves a kind of homeostasis that provides greater security than one whose value is bound to something stable and exogenous like gold and land. Ones savings are secure as long as the system doesn’t fail. Meanwhile, the price of access to cultural materials through which one might expand ones individuality (i.e. videos of academic lectures, the arts, or music) decrease as a consequence of the pervasiveness of the economy. At this point one feels one has reached the limits of Horkheimer’s critique, which perhaps only sees one side of the story despite its sublime passion. We see echoes of it in contemporary feminist critique, which emphasizes how the demands of necessity are disproportionately burdened by women and how this affects their role in the economy. That women have only relatively recently, in historical terms, been released from the private household into the public world (c.f. Arendt again) situates them more precariously within the economic system.

What remains unclear (to me) is how one should conceive of society and values when there is an available continuum of work, opportunity, leisure, individuality, art, and labor under conditions of contemporary technological control. Specifically, the notion of inequality becomes more complicated when one considers that society has never been equal in the sense that is often aspired to in contemporary American society. This is largely because the notion of equality we use today draws from two distinct sources. The first is the equality of self-sufficient landholding men as they encounter each other freely in the polis. Or, equivalently, as self-sufficient goldholding men in something like the Habermasian bourgeois public sphere. The second is equality within society, which is economically organized and therefore requires specialization and managerial stratification. We can try to assure equality to members of society insofar as they are members of society, but not as to their function within society.

Horkheimer on engineers

Horkheimer’s comment on engineers:

It is true that the engineer, perhaps the symbol of this age, is not so exclusively bent on profitmaking as the industrialist or the merchant. Because his function is more directly connected with the requirements of the production job itself, his commands bear the mark of greater objectivity. His subordinates recognize that at least some of his orders are in the nature of things and therefore rational in a universal sense. But at bottom this rationality, too, pertains to domination, not reason. The engineer is not interested in understanding things for their own sake or the sake of insight, but in accordance to their being fitted into a scheme, no matter how alien to their own inner structure; this holds for living beings as well as for inanimate things. The engineer’s mind is that of industrialism in its streamlined form. His purposeful rule would make men an agglomeration of instruments without a purpose of their own.

This paragraph sums up much of what Horkheimer stands for. His criticism of engineers, the catalysts of industrialism, is not that they are incorrect. It is that their instrumental rationality is not humanely purposeful.

This humane purposefulness, for Horkheimer, is born out of individual contemplation. Though he recognizes that this has been a standpoint of the privileged (c.f. Arendt on the Greek polis), he sees industrialism as successful in bringing many people out of a place of necessity but at the cost of marginalizing and trivializing all individual contemplation. The result is an efficient machine with nobody in charge. This bodes ill because such a machine is vulnerable to being co-opted by an irrational despot or charlatan. Individuality, free of material necessity and also free of the machine that liberated it from that necessity, is the origin of moral judgement that prevents fascist rule.

This is very different from the picture of individuality Fred Turner presents in The Democratic Surround. In his account of how United States propaganda created a “national character” that was both individual enough to be anti-fascist and united enough to fight fascism, he emphasizes the role of art installations that encourage the view to stitch themselves synthetically into a large picture of the nation. One is unique within a larger, diverse…well, we might use the word society, borrowing from Arendt, who was also writing in the mid-century.

If this is all true, then this dates a transition in American culture from one of individuality to one of society. This coincides with the tendency of information organization traced assiduously by Beniger.

We can perhaps trace an epicycle of this process in the history of the Internet. In it’s “wild west” early days, when John Perry Barlow could write about the freedom of cyberspace, it was a place primarily occupied by the privileged few. Interestingly, many of these were engineers, and so were (I’ll assume for the sake of argument) but materially independent and not exclusively focused on profit-making. Hence the early Internet was not unlike the ancient polis, a place where free people could attempt words and deeds that would immortalize them.

As the Internet became more widely used and commercialized, it became more and more part of the profiteering machine of capitalism. So today we see it’s wildness curtailed by the demands of society (which includes an appeal to an ethics sensitive both to disparities in wealth and differences in the body, both part of the “private” realm in antiquity but an element of public concern in modern society.)

Arendt on social science

Despite my first (perhaps kneejerk) reaction to Arendt’s The Human Condition, as I read further I am finding it one of the most profoundly insightful books I’ve ever read.

It is difficult to summarize: not because it is written badly, but because it is written well. I feel every paragraph has real substance to it.

Here’s an example: Arendt’s take on the modern social sciences:

To gauge the extent of society’s victory in the modern age, its early substitution of behavior for action and its eventual substitution of bureaucracy, the rule of nobody, for personal rulership, it may be well to recall that its initial science of economics, which substitutes patterns of behavior only in this rather limited field of human activity, was finally followed by the all-comprehensive pretension of the social sciences which, as “behavioral sciences,” aim to reduce man as a whole, in all his activities, to the level of a conditioned and behaving animal. If economics is the science of society in its early stages, when it could impose its rules of behavior only on sections of the population and on parts of their activities, the rise of the “behavioral sciences” indicates clearly the final stage of this development, when mass society has devoured all strata of the nation and “social behavior” has become the standard for all regions of life.

To understand this paragraph, one has to know what Arendt means by society. She introduces the idea of society in contrast to the Ancient Greek polis, which is the sphere of life in Antiquity where the head of a household could meet with other heads of households to discuss public matters. Importantly for Arendt, all concerns relating to the basic maintenance and furthering of life–food, shelter, reproduction, etc.–were part of the private domain, not the polis. Participation in public affairs was for those who were otherwise self-sufficient. In their freedom, they would compete to outdo each other in acts and words that would resonate beyond their lifetime, deeds, through which they could aspire to immortality.

Society, in contrast, is what happens when the mass of people begin to organize themselves as if they were part of one household. The conditions of maintaining life are public. In modern society, people are defined by their job; even being the ruler is just another job. Deviation from ones role in society in an attempt to make a lasting change–deeds–are considered disruptive, and so are rejected by the norms of society.

From here, we get Arendt’s critique of the social sciences, which is essentially this: that is only possible to have a social science that finds regularities of people’s behavior when their behavior has been regularized by society. So the social sciences are not discovering a truth about people en masse that was not known before. The social sciences aren’t discovering things about people. They are rather reflecting the society as it is. The more that the masses are effectively ‘socialized’, the more pervasive a generalizing social science can be, because only under those conditions are there regularities there to be captured as knowledge and taught.