Beyond the Tyranny of Language - Part 2
Instituting artificial intelligence as state-form
Part 1 of this series can be read here. In this second part, I dwell somewhat on the Marxist and Kojèvean significance of Vince Garton’s suggestion to institute artificial intelligence as the next state-form.
What is the political? The German jurist Carl Schmitt’s famous answer, which he elucidated in The Concept of the Political, was that the political consists in the ability to make a distinction between friends and enemies. The enemy is one whom it is worth your while to fight to the death—either the enemy’s or your own. A similar understanding of the political seems to be operating behind Alexandre Kojève’s interpretation of the Hegelian master-slave dialectic, where politics exists as long as Masters exist and risks their lives to fight, conquer, and subjugate others as their Slaves; and as long as Slaves have cause to struggle and work for their liberation from the domination of their Masters. Kojève’s understanding of politics reads like a re-Hegelianization of Karl Marx’s own understanding of the political, which was intimately tied to the reality of class struggle. For each of these thinkers, as for many others, struggle and conflict pertain to the very essence of politics.
On this view of things, as long as politics—i.e. struggle, conflict, and enmity—continues to exist, the form of the state will necessarily be political. That is, the state will necessarily be an agent of warfare. The ministers of the state will include among them soldiers, men who wield the sword or the gun, to wage war on behalf of the state against those who are perceived as enemies. The state will also be the principal possessor of the power to discern and judge who shall count as a friend and who an enemy. In this way, it legitimizes its own use of violence by a kind of moral judgment, and reserves to itself both the right to use that violence and to make that sort of judgment.
From all this, it follows that in a condition devoid of human conflict and the making of enemies, the political function of the state becomes obsolete. It is for this reason that Marx and Engels predict that the state in its political nature will “wither away” as socialism transitions into communism. Recall that the state still exists under socialism, because there is still a bourgeoisie to crush. But under a classless society devoid of any reason for class warfare, the political function becomes unnecessary, and the state is transformed into a simple “administration of things.” Such a de-politicized state is no more than a technology, no longer concerned with the regulation, mediation, and even the waging of human warfare, but only with making sure life runs smoothly for people. In other words, this state is no longer required to exercise the sorts of judgment that pertain to the discernment of friends and enemies—a uniquely human judgment. Indeed, its domain is things rather than people—in Marx’s scheme, it is concerned merely with the regulation of production and of the things produced. In other words, it is an artificial intelligence.
Similarly, Kojève predicted that the end of history would be the end of politics, since, as a good Hegelian, he held that history itself could be explained by the dialectical struggle of Masters and Slaves. The “universal and homogenous state” which he foresaw at the end of history would be an apolitical state, in which the universal recognition of all by all would have finally been secured, eliminating the need for further struggle. In his Outline of a Phenomenology of Right, he further explains that this state would be merely juridical or legal, though not political. In this, Kojève appears to directly echo Marx and Engels’ administration of things.
What we can see explicitly in Marx and Kojève, and what would seem likewise to follow from Schmitt’s understanding of politics (though he might not have drawn this conclusion himself), is that the end of politics and its substitution by a purely technological administration of things is desirable at least in theory. In this light, Vincent Garton’s “post-liberal” prognosis at the end of his recent essay for Urbanomic seems more understandable, and not merely the indulgence of some dismal dystopian fantasy: artificial intelligence should be instituted as state-form.
What Garton’s essay highlights, following Kojève, is the linguistic dimension of this AI-governed administration of things, a dimension to which Marx paid only cursory if any attention. Under communism, the apolitical “state” administrates the material process of production. It is an administration of things rather than people. What Garton is drawing attention to, however, is the production of language, which is the central act of liberalism itself. The focus upon the centrality of language-production, for which Kojève may take the credit, is a unique modification of the Marxist materialist analysis of capitalism. Returning to language seems, in a certain way, to bring back the focus upon the ideal rather than the material level of the social organism. Language in this analysis pertains not merely to the ideological superstructure overlaying what is essentially a system of the material production of commodities; rather, it is itself the primary commodity of liberal political economy. The distinction and relationship between material base and ideal superstructure seems to be melting away, just as the relationship between sign and the signified seems to be dissolving. The ideal as such, the discourse of ideologies, is itself taking the place of the material base. Liberal production is nothing but the production of endless new ideals and ideologies by way of the production of endless speech.
And just as capitalism in Marx’s analysis exhibits a trend towards the total automation of production, liberalism exhibits a trend towards the total automation of language production—and in both cases, this is a self-destructive trend. The deliberate acceleration of this trend is a key element of Garton’s proposed post-liberal strategy (to the extent that it can be called a strategy), just as the acceleration of capital’s own productive forces was perceived by Marx to be key to a socialist strategy (to the extent that it could be called a strategy). The total automation of speech would render obsolete the human project of producing new ideals in speech, for they will all have been produced already by automation. Everything that can said will have been said; every statement that can be made will have been made. There would no longer be any sense in “freedom of speech,” just as under communism there would no longer be any sense in “private property.” Not that anybody could no longer speak or produce, but that everything that could be either said or produced must be drawn only from a pre-existing reserve of speech and production.
(As an aside, just as communism refers to the pre-existing reserves of production from which alone all production at the end of history can proceed, likewise the concept of tradition refers to the pre-existing reserves of speech and expression on which all speech at the end of history will depend. There will be no more novelty, only tradition, just as there will be no more “private property” in the bourgeois sense. The end of history will be the triumph of communism and the triumph of tradition.)
To conceive the state as artificial intelligence is nothing new. As Garton has noted elsewhere, the pre-eminent example of a theorist who likened the state to artificial intelligence is Hobbes, whose Leviathan was explicitly meant to be a technological interpretation of the state. But whereas, for Hobbes, the automation of statecraft presupposes a pre-existing condition of total warfare among men, for Marx, Kojève, et al, the automation of the state pertains precisely to the end of history when meaningful conflict has ceased. Precisely because there is no more need for conflict, or because conflict has become, in a sense, superfluous, politics itself becomes superfluous.
Garton’s proposal rests on a similar presumption. Under liberalism, we live in a partial end of history where real conflict and political action have largely ceased, or rather been transposed to the purely discursive domain of language. We fight, we draw battle lines, and we distinguish friends and enemies—but only in speech. There is no politics, exactly, but what Anton Jäger has called “hyperpolitics.” Or if politics exists, it is only at this superfluous and virtual level of pure discourse, separate from the realm of concrete action. In this context, it makes sense to automate the discourse, killing off the last remnants of the political. It has become superfluous.
Of course, one can disagree with elements of this analysis. Recurrent wars in Ukraine, Palestine, and elsewhere might lead one to think that the political has been reborn in a real and not merely virtual form. Even in the developed center of the world, violence still persists: progressives rightly point to the continuance of crippling socioeconomic inequalities, and Christians and Right-wingers correctly identify the genocidal slaughter of millions of children in the womb year after year. Thus, one might speculate that there might still be some history left to be made—and ended—in some parts of the world, still some final acts of political heroism to be carried out and completed.
In any case, Garton himself has acknowledged the continued existence of action, but he has astutely observed that political violence in liberalism’s partial end of history tend to serve the ends of language production, rather than the reverse. Every demonstration or act of violence is ultimately a means of signaling one’s position in the discourse, a means of adding one’s voice to the already-discordant throng, an act of self-expression or an act of speech. Arguably, this is how even war itself is to be interpreted, as directed by the global liberal hegemony under whose purview almost all wars must necessarily fall. Under such a rule, it is the linguistic and ideological content of modern war-making that matters most. Whose side one takes in modern warfare, regardless of whether one is directly engaged in physical combat, is ultimately little more than a signal of what discursive or ideological position one takes in the great continuing discourse of the liberal West. The actual violence on the ground is of diminishing importance—that is to say. it is entirely superfluous (which does not, from an ethical standpoint, diminish its injustice: on the contrary, it heightens it).
Unlike the Hobbesian state of nature, there is a real sense in which the distinction of friends and enemies is now no longer real, and the battles we fight have become unreal—or hyperreal. All our battles are primarily virtual and discursive in nature; real battles in the world out there have an epiphenomenal character. Political enmity is likewise primarily virtual, hyperreal, and discursive in nature—and thus superfluous in the extreme (a wasted surplus). In this context, the automation of discourse makes some sense: it is the automation of what has become superfluous, and thus needless—a waste of surplus human energy. In this respect, the complete automation of language production might bring an effective end to the last vestiges of political conflict that still superfluously remain: the yet unending democratic debate among free thinkers and free speakers, and all remaining violence that serves this tyrannical discourse.
The next question to be answered is this: Beyond the tyranny of discourse, what form will language itself take?
In the next installment, I will shift gears to consider a very different but still related dimension of the surpassing of liberalism’s linguistic regime: the theological dimension. This is the installment I haven’t written yet. Now that I have promised it, I have to write it—though I may not complete it before the New Year.
Great essay! Hope you’re doing well Jonathan
As per usual very impressed by your entries Jonathan, can't wait for part III