My dissertation’s structure has been on a recalibrating offsite (or something) for the past few weeks. When I first drew up a proposal for my advisors at the end of 2024, I had a clear work plan: start with a robust understanding of Marx’s Capital, work through an argument about his concurrently scribbled mathematical manuscripts, link Marx to the mathematical modernism at the turn of the 20th century, and conclude with some psychoanalytic flourish if I still had more to give. I still find this trajectory compelling, and may return to aspects of it in a future publication or book project. But I have put such substantial work and thought into a range of different software projects over the past year that it has begun to seem strange to not include any part of them in my dissertation proper, especially as its self-elected subject is rethinking computer science from the vantage of Marx’s philosophy, a body of work that famously confounds the distinction between theory and practice.
More importantly, I have found that an intricate, theoretically precise argument about the mathematical consistency in Marx’s thought generally fails to appeal to not only computer scientists, but also many critical theorists and literary scholars, as serviceable thinking. I am in two minds about this, as I do believe strongly in the political nature and even necessity of certain kinds of theory that are often dismissed as overly ornate, or of the ivory tower. Writing need not go down the gullet like honey and lemon to have a real effect, and there can be great value in asking more of the reader rather than less. But there is also always a question about what kind of reader will go fishing with you, a question that depends on what you suggest might come up out of the water at the outset.
Previous versions of my dissertation introduction had suggested that, in order to set the scene for a critical theory of the computer, it was necessary to return to both Marx’s critique of political economy in depth and to the intellectual history of mathematics at the turn of the 20th century. Here’s an excerpt from an early draft:
Marx has on occasion been accused of constructing a theory of capitalism that, while socially and politically persuasive, fails to take into account the real complexities of the human subject. Psychoanalysis, on the other hand, while widely credited for its thoroughgoing treatment of the intricacies of any given individual in the clinic, is sometimes suspected of having limited applicability when it comes to the question of material social change….
This dissertation attempts to clear the air of these twin false dichotomies: between the individual subject and the politics of society, and between the critique of political economy and psychoanalysis. The subject that emerges in Marx’s critique of political economy is a position that conceptually prefigures the notion of the psychoanalytic subject as it is later invented by Freud and then more precisely elaborated by Lacan. The subject cannot be walled off from its work in the world—a conundrum that appears in Marx’s work through the notion of labor, and in Freud’s through the unconscious. It is tempting to conceptually excise ourselves from society because it is an unsettling thought to acknowledge that our ideas are not factually on firm footing. Though the separation of the subject from society is a fundamentally false foundation for thought, it still structures much of the thinking about both terms in today’s critical climate. This dissertation will show both why we should make out a theory of the subject in Marx that prefigures psychoanalysis, and also why this theory critiques the rationalist fantasies of computer scientific reason. Drawing on insights of keen readers of both Kant and Marx such as Sohn-Rethel and Kojin Karatani, amongst others, I argue that the subject in Marx is minted by the social condition into which it is naturally and unavoidably thrown, and that we cannot cordon it off from its point of genesis in the world. Making out the figure of a subject in Marx is the key to seeing his critique of political economy as the groundwork for a critique of philosophical epistemology; to connecting his project in Capital to psychoanalysis, his unfinished critique of calculus, and the imminent mathematical revolution in number.
This research interests me deeply and remains essential groundwork for a more thorough understanding of how capital compromises our claim to freedom in modernity, and whether psychoanalysis offers something more than Marx does in the project of critical philosophy, i.e. thinking ourselves as freely as we can. But it is also a dissertation’s worth—if not much more—of close reading of Marx, Kant, Lacan, Freud, and many more thinkers yet, a serious task that would leave little time for dealing in the history of computer science, let alone presenting real software as contributions within it. This kind of dissertation in philosophy which I had planned out for myself also represents me partially and probably poorly on the job market, where my capability as a software engineer will arguably be what differentiates my application from the mountain of overqualified others.
But I don’t want to misrepresent this pivot as being about my commercial viability as a career academic. It is more properly about the fact that, for the past twelve months, I have spent 80% of my time thinking critically about what kind of software I should build, how I should build it, and actually building it, and about 20% of my time reading deeper on the topics that I don’t yet know enough about to write a dissertation worthy of the promise in the blockquote above. For the first time in a while, I really believe that the software I am writing could make a material difference: not in the sense of convincing a certain number of swing voters; nor of displacing an existing commercial or open source solution; but in the sense that it represents ideas that don’t yet exist in the form in which I mean to create them. Of course this may prove to be misguided or just outright wrong, but the belief gets me up in the morning and glues me to my screen: it is something to live for.
The new version of my dissertation intro is very much a first draft (which took two days to put together), so I won’t regale the reader with an inexact excerpt. The basic structure is to present and historicize one or more softwares in each chapter: Rheo and rookery in chapter 1, developer tooling for working with LLMs in chapter 2, a distributed system I’ve been dreaming up in chapter 3, and interpretability work at the Digital Theory Lab in chapter 4. The fifth and final chapter will repurpose what was previously my first chapter, a deep dive on Marx’s theory of the commodity and capital, into a reflection on the nature of formal systems and their study in theoretical computer science.
I expect that this new structure will be more approachable for both computer scientists, who may take interest in the systems, if not the historicization, and critical theorists, who may find it interesting to see the methods of close reading and historical scrutiny applied to such an irregular object of study as a software system. At least, this is the hope!