These are some notes triggered both by the past two weeks I spent in the US, one of the most fascinating and disturbing political arenas in the world, and by the insipid vacuum of a 13-hour journey back to Glasgow. I’ve no real interest in discussing the actual politics here as there’s precious little I can offer that more qualified commentators can’t. I’d only anecdotally summarise that the US has elected a president that the US wants, and Trump’s ethos is a rational consequence of the ethos of the nation. Beyond this, despite my strong political beliefs, I’m interested here in a more ambivalent discussion of how political will is formed.

I’m hardly literate in political media theory so I’m not going to over-egg my abilities to chat about it in depth; like every university-educated 30-something I grew up surrounded by Žižek and Chomsky and doing a good job of nodding my head and pretending I’d read more than I had. At least half of my knowledge of things like manufacturing consent admittedly comes from pseudo-intellectual house parties I went to when I was 21, so don’t come at me too hard for everything I’ve misunderstood.

Any real shock at the US electoral outcome might stem from a fantasy — perhaps a particularly liberal one — which hopes for a government or wider political structure to form a political consensus, as opposed to a political consensus forming a government. Walter Lippmann, while coining the phrase manufacture of consent in 1922, advocated for this somewhat disturbing elitism, informing this belief with the idea that

It is no longer possible… to believe in the original dogma of democracy; that the knowledge needed for the management of human affairs comes up spontaneously from the human heart.

This challenge against naive democracy is useful, and we’ll loop back to it in a bit. Lippmann argued that if the voting public couldn’t be relied on to spontaneously make a good political decision, then a technocratic establishment must make the right decision and thereafter persuade the public of its value; thus a government might manufacture a consensus in its constituents through propaganda and other means.

Under this model, the Biden/Harris administration might be viewed as having performed extremely poorly, and the prospects of progressive causes look bleak and feel hopeless, and with good reason. There is not a public consensus in the US to secure abortion rights, or improve healthcare, or trans rights, or support for migrants, and the establishment which liberals had expected to engineer this consensus has collapsed.

It strikes me that nowhere near enough work is done at an inter-personal level to create the conditions for an altered consensus which is generated by those whom the consensus ought to serve. If we accept Lippmann’s premise that the voting public will not innately know how to run a country, yet reject his conclusion that this necessitates a technocratic elite to decide on their behalf, we’re left with the somewhat pedestrian alternative of simply improving the level of education, political literacy, and quality of discussion in a public in order to improve its electoral prospects from the roots.

The reason I believe in Lippmann’s premise, if not his conclusion, is that my anti-essentialist tendencies prompt me to believe that consensus cannot be found, but rather is created. The processes of discussion and interpersonal communication lead us gradually towards the conditions within which consensus can be built. This also contrasts with ‘discussion’ which is not truly inter-personal; that is, discussion which is only amongst those who hold similar beliefs can be intellectually regressive.

This can and does include conversations with AI; I’d reference here my previous piece about writing like a chatbot and the tendency of AI-generated text to end up as mostly just intellectually numbing repetition appropriate to its tightly-constrained context, but Rob Horning’s recent piece on Habermas Machines is much better. Horning discusses a recent Google paper outlining an AI model named after Jürgen Habermas, which aims to identify the common political ground between divided groups. He points to novelist Hiroki Azuma, who much like Lippmann identifies an useful premise while on the dubious route to a technocratic elitist conclusion:

what is important for democracy is not that everyone’s will is gathered but that each will is transformed through the process of consolidating those wills.

That is to say, consensus is not found simply by gathering a mass of spontaneously-formed opinions, but rather it is the processes of consensus-gathering – discussion and discord and adjustment – which form and evolve opinion.

The dysfunctional collapse of this felt tangible in the US, where I was acutely aware of a generalised fear of political discussion due to its divisiveness and a fetishisation of ‘non-political’ neutrality. Commentators adore prefacing political commentary with exposition about how they are ‘not on the left/right’ or even ‘not political.’ This is of course all anecdotal, but I left the country with a distinct feeling that political division has rapidly self-replicated through participants mutually encouraging a disengagement with the kind of difficult inter-personal communication which could form consensus in favour of instead sharing political views within accordant echo chambers.

It follows that within this context, one might envision the ideal political engagement as a painless experience, arbitrated by a lifeless, neutral, depoliticised observer, such as the Habermas AI, which dispassionately elevates all political views indiscriminate of feasibility or factuality. Yet it also makes sense that this is a showcase of a wild abundance of deeply-rooted essentialism undergirding political discourse, especially within the libertarian sphere.

Under this lens, political leanings (and everything else like music taste, sexual preferences, temperament) might be viewed as essential and immutable characteristics embedded into an individuals’ “true self” or “inner being”. This is what Lippmann referred to as knowledge arising “spontaneously from the human heart”. These things are more accurately seen as fluid properties which can and will change through education, life experience, exposure to variance of opinion; the feedback-driven processes of learning.

So this issue isn’t really about Al, but rather the AI of the Habermas Machine or other techno-utopian political projects is posited as a tool to realise the micro-fascist¹ dream of each of us being born with a specific and immutable, neat, tidy, and innate location on a political spectrum.

Thus, essentialist ‘politics’ becomes not a creative process, but just the passionless rote action of finding consensus: every cunt in the world is a fixed circle, and the perfect universal politics exists at the Venn-diagram intersection of all of the circles. The centrist wet dream is one in which the ideal politics is simply a mathematical compromise located exactly between opposing points, rather than a new location determined through a constructive oppositional argument.

Like so many half-baked tech-utopian dreams, this treats human behaviour as a bland and eternal constant in the equation rather than as a (far less convenient) fluctuating variable within a feedback loop. Like the fallacious Moore’s Law it assumes past results to be a guarantee of future outcomes; like the shortsighted Turing Test it is unaware that the results of the equation might alter the variables which comprise it².

The brutal reality which disrupts this logic is that our political desires are mathematically indeterminate, because each new political reality prompts a new or altered state of desire in its constituents. There is no satisfactory end state, there is no neat political ideal to be realised. Rather the goalposts shift each time the feedback and affordances change. The consensus which was previously reached is no longer relevant and a new one must be created through a painful and arduous productive process.

To me this anti-essentialist approach is deeply encouraging and motivating! It offers an alternative to the depressing comparison of current political realities against seemingly unattainable political ideals: instead, each political reality is inconclusive, it will be reiterated. To say that fascism failed in 1945, or that communism failed in 1990, is to fall into an essentialist and determinative trap which ascribes certain human behaviour or actions as permanent, immutable states. It suggests, dangerously, that something like fascism couldn’t reoccur, and claims dishearteningly that any alternative to capitalism is incompatible with human nature.

What we’re talking about is basically just an expansion of the idea of the Overton window into something which lets us accept the endless fluidity of politics, or at least it’s an articulation of the intellectual poverty of Fukuyama’s End of History. Or maybe it’s just rationalising some inappropriate positivity!

For me it’s encouraging because we ought to know that there is no singular ideal. We are always simply engaging with fluctuating affordances, we’re designing for right now, secure in the knowledge that tomorrow’s now will not match today’s. It’s a reminder that it is cruel and flawed to be a distant observer simply locating constants; instead we alter fluctuating variables through real, inter-personal action.

1. Very keen to read more about Deleuze and Guatarri’s ideas of micro-fascisms after coming across the concept in Richard Seymour’s new book Disaster Nationalism. Reminds me a lot of Natasha Lennard’s excellent Being Numerous, I book I recommend to everyone at every moment of my life. ↩︎
2. I’d like to expand on this in later writing, but I’m fascinated by how our determinations of machine intelligence are leading to a reconfiguring of how we define human intelligence. ↩︎