The algorithmic turn promises to (ope-)rationalize the social at its roots, inaugurating a model of (un-)governance/leveraging based on pre-semiotic, infra-political modulation within the digital anarchive —an ever-expanding repository of digital data (signals) which, unlike traditional archives, operates without curatorial principles or hermeneutic distance, collapsing past and future into an eternal computational present. This perspective revives the techno-utopian dream of a post-representational, post-hermeneutical, immanent governmentality. Such "algorithmic realism" collides with the semiotic and heterochronic nature of law—its reliance on signs, delay, contradiction, and the collective labor of interpretation. In this intervention, I contend that, beyond issues of fairness, transparency, accountability, explicability of algorithmic black-boxes, or beyond issues of performance, reliability, and alignment of AI systems, what is most profoundly threatened by the algorithmization of law is the functioning of legal language as an anarchitecture of the common. This threat reactivates the historical tension between legal formalism's quest for determinate outcomes and the interpretive flexibility that has persistently resisted such closure since the legal realist critique of the early 20th century. When legal language is algorithmically (de-)termined, it loses precisely the productive in(de)terminacy that enables new articulations of justice to emerge in response to previously unheard claims and revolving social conditions. If human language, and legal language in particular, can be said to participate in this anarchitecture of the common, it is not because of some 'natural' or 'ontological' property of language, but because of its capacity to open space for what has not yet taken place, to dis-architecture instituted speech, to poke holes in the fabric of dominant discourses. Anarchitecture, whether spatial (Wigley, 2018) or symbolic, does not designate a 'place' but a ‘process’: that by which a closed structure is made open to its possible opening. Whereas, for example, LLMs model language as a closed flow of probabilistic signifiers, the anarchitecture of legal language, or its open texture (Hart, 1961), maintains the scene of dissensus or negative dialectics —the very condition of politics and collective invention. In this sense, human language, including legal language, has never been the expression of a totality, but always a public use of the unfinished. Hence, the spirit of law as anarchitecture lies not in mastery but in the mastery of mastery, not in the power to dominate, but in the power to dominate domination, and thereby to abolish it (Endnotes, 2011) . Legal anarchitecture, expressing law's negative dialectics, abolishes all totality, including its own. References: Endnotes (2011) "What Are We to Do?", in Noy, B. (ed.) Communization and its Discontents. New York:Minor Compositions, p. 29 Hart, H.L.A (1961). The Concept of Law. Oxford: Clarendon Press. Wigley, M. (2018) Cutting Matta-Clark. The Anarchitecture Investigation. Lars Müller Publishers
Antoinette; id_orcid 0000-0003-0554-9492 Rouvroy (Wed,) studied this question.