1 Here we have an impudent work that performs a beneficial service: making a diagnosis of recent developments in the justice system under the impact of information technology, while inviting the reader to consider the extent of the resulting changes. The two authors, Antoine Garapon (Institut des hautes études sur la Justice) and Jean Lassègue (Centre national de la recherche scientifique), manage to take up the challenge in some 360 pages by combining their talents as specialists in law and epistemology, while highlighting the specificity of their respective approaches – the anthropology of justice and the semiotic theory of culture. The originality of their method lies in its avoidance of the approach via performance and utility, by undertaking a more radical questioning of the symbolic nature of the digital revolution: revealing its written dimension (the code) means suspecting the mythological value of the belief in "delegating to machines" (automation).

2 The first part of the book tries its hand at a definition of digital justice by way of a long anthropological circumlocution. Relying in particular on work by Clarisse Herrenschmidt [1] and Sylvain Auroux, [2] the authors show that a third revolution is taking place in writing today, in the wake of its invention in Mesopotamia (letters/numbers). The alphabet segments language, opening the door to its mechanization; grammatization considers it as an objectifiable tool (grammars and dictionaries), especially in the age of the printing press; the passage from the analog world to the digital world inaugurates a mute form of writing that deploys its own activity and opens new cultural possibilities. Computer science as a technique also constitutes an authentic symbolic form, [3] alongside that of law, and all of the authors’ efforts are devoted to detailing the modes of this conflict of normativities. From the outset, Garapon and Lassègue stress the limits of this form of writing, which is incapable of eliminating the disparity between the physical and the semiotic. Indeed, calculations and programming evacuate semantics, and therefore interpretation, from signs, and computer code does not have all of the formal properties of a natural language, which are a result of reflexivity. Furthermore, such writing is independent of the institution that governed it and the domain to which it was connected, which has consequences for the legal ritual. Lassègue and Garapon offer the suggestive hypothesis that the desymbolization associated with a purely computational nature explains the search for a substitute mythology, in the form of the machine’s control of itself (chapter 1). The useful and succinct but clear overview of the concepts of "computability" (Turing) and incompleteness (Gödel) helps explain the failure of Hilbert’s dream of a basic set of axioms for mathematical thought, and makes it easier to understand the semantic line of flight of formal thought. The calculable is intrinsically limited and its expansion into all areas of life cannot avoid a symbolic moment of reflection that resituates information technology within the space-time of life in common as currently experienced – except in the mythological dream of total subservience to machines (chapter 2).

3 Next, the authors emphasize the specifically anthropological revolution that has taken place: although the computer is a graphic machine, it has found its way into every corner of our lived experience, especially since the mobile phone came into widespread use in the early 2000s. This "total social fact" is conveyed by a libertarian political ideology, which is consistent with a pirate mindset and a neoliberal, globalized era that is wary of the state. The political aspects, as decisive as they are (the new public information resources such as Wikipedia paradoxically seem to owe their shared meaning to private parties), are however derived from this deeper symbolic mutation. The same is true of the social transformations induced by legal technology (the promise of easier access to justice, to the democratization of law, competition between mediation services), and even its cognitive aspects (substitution of prediction for interpretation, algorithmic fetishism). One of the most striking diagnoses of the two authors, inspired by Cassirer as much as by Lefort, Gauchet and Ricœur, is that of an effacement of the third party and of institutional mediation, so ultimately one of an (illusory) disavowal of the symbolic order (chapter 3). The digital revolution does not affect the aims, but rather the very modes of justice, thus giving the impression of bringing about nothing less than a new natural law. Virtual meetings, oaths taken remotely, and elisions in the ritual all attest to the profound disorganization of the spatiotemporal framework of justice, a disorganization that is a consequence of information technology as the new dimension of the world we live in. But the exhaustive subservience to machines amounts to the magical belief in a world that is already preordained by calculations, whose computers need only collect the data. This myth is anchored in a fallacious line of thought that defines technique as an extension of the body, instead of considering the co-naturality of technique and society within culture, and must be understood in the light of the impossible effort of self-grounding on the part of democratic societies. Although information technology presents itself as a new natural law that is an integral part of the world, this is an illusory effect (the scanning of legal data is a new form of representation) that lets its symbolic nature be forgotten. And yet the calculable is no substitution for the symbolic, just as the operational does not exhaust the signifier (chapter 4). Blockchain [4] technology deserves its own detailed analysis, given the considerable implications this offshoot has for justice and the legal profession. Indeed, it capitalizes on a strict determinism (via cryptology) that revitalizes the notion of trust by its ambition of unfalsifiability, to the extent that all political ties are potentially modified, starting with the contractual obligations that claim to do without institutional mediation and the third-party guarantor (smart contracts, decentralized organizations for settling disputes). By occasionally justifying itself through a redistributive vocation (unlike the online platforms), it fosters the illusion of a democracy of individuals; in reality, a form of law that is entirely determinate becomes inhuman and antipolitical, since justice needs space for play, indeterminacy, and even imperfection (chapter 5).

4 The second part of the book examines what digital justice does to the idea of justice. The structure of the judicial hearing is altered by the decoupling of space and time, by the disintermediation of evidence (which no longer aims to satisfy all doubts by the end of the proceedings, but rather "speaks for itself"), by the impoverishment of the trial experience (which, with the shattering of the three classical unities, loses its symbolic effectiveness), and by deritualization (chapter 6). Judges risk soullessness, due to the systemic nature of digital justice (online dispute resolutions; [5] "made-to-order" courts, with "blockchain jurors" but no jury; the self-interested construction of judicial truth; the elision of the adversarial principle and the instance of the law; the automatic application of decisions). By constituting a fourth term that dominates the two opposing parties and the third party of justice, by extolling "solutions" rather than judgments, digital justice threatens to subvert democratic principles (chapter 7, pages 201-203 in particular). Moreover, the dream of a fusion of prescription and prediction, whose epistemology is questionable, risks twisting legal standards by turning jurists into the agents of economic strategies, and by diverting legal reasoning – argumentative, narrative and interpretive reasoning – toward so-called artificial intelligence, whose statistical correlations rely only on the power of calculation and a supposed determinism. There is then a great risk of reducing the number of decisions, of reinforcing majority tendencies, of supplanting the practical wisdom of jurists, of eliminating dissident opinions, and of petrifying judicial time through the weight of an infinite memory (chapter 8). What’s more, law is in danger of disappearing in favor of a personalized, pliable standard, a micro-directive that may vary in time and space, tending toward a personal injunction, and therefore liable to reintroduce inequalities (as a result of the varying mastery of internet tools). Also, the new archetype of power is not so much the centralized one of the panopticon as one of a new feudal relation between individuals and insurance companies, via machines that "calculate" them. Criminal justice risks losing its retributive vocation in order to punish "criminogenic" risks, thereby shifting the criteria for punishment from the act to the intent or tendency (chapter 9). In addition, judgments may be influenced by horizontal control (recommendations, utilitarian observations), pressure from the multitude (and its risks of exclusion), and the sheeplike phenomena associated with influence or the rigidification of behavior (chapter 10). Finally, the authors insist on the "great [cultural] adjustment" that the digital revolution has brought about, an adjustment that modifies the process of legal classification (something automation has attempted), with legal standards that themselves seem to emanate "naturally" from reality by a kind of digital intensification of experiencing the world (acceleration, permanent transformation), moving imperceptibly from an equitable justice of "fairness" [6] to an adequate justice of "fitness" [7] through self-correcting adjustments (chapter 11).

5 The brief third part (chapter 12) – which takes a more normative, long-term view – serves as a reminder of what the idea of justice demands from digital justice. It retraces the fundamental myth of delegating to machines, which is at the heart of current trends in big data, [8] artificial intelligence, deep learning, [9] etc. Justice is a reflexive idea, an appeal (a protest) against evil, with deep and ancient anthropological sources that information technology could never fully explore. The nature of a symbolic form disposes it to spread indefinitely. But this does not mean that it won’t encounter resistance from others, or that it can do without some form of arbitration with its competitors in the arena of shaping culture. This investigation, grounded in symbolic anthropology, teaches us that justice must remain connected to a particular context, to physical presence in a courtroom, and to the institutional presence of a third party, if it wishes to protect human nature from automation and commodification. This is why the author’s conclusion consists in recalling the difference between the two symbolic forms, and in pleading for a reasonable level of coordination between them, so as to better withstand the risks of fusion.

6 Apart from small inevitable mistakes (the possibility that supposedly inviolable blockchains could be hacked was only discovered after the book was published [10]), we may regret slight inaccuracies, such as the lack of distinction made between information technology, the internet (as a network of cables) and the Web (one of the internet’s applications, resulting in a virtual "world"). Similarly, in some examples, the authors confuse the participative Web 2.0 with the "semantic" Web 3.0 (p. 273 [11]). They also fail to differentiate between collaboration and commodification (p. 277), as if coworking in a digital context necessarily leads to a commercial outcome. A more serious problem is the massive and ultimately somewhat dogmatic assertion of a connaturality of information technology and neoliberal globalization (p. 29), as if globalization were not a policy (undermined, for example, by Donald Trump). This closeness would explain the inevitable powerlessness manifested by state actors – as if some of them hadn’t already demonstrated that it is perfectly possible to reestablish institutional limits within the internet, not to mention China’s gigantic efforts, for example, to limit the internet itself. Similarly, the conclusion that information technology is inherently disruptive (p. 313) seems a bit hasty, as it dispenses with considering the fact that the dynamics of destructive creation are much more characteristic of capitalism itself, especially in its current advanced, post-industrial stage.

7 Furthermore, the assertion that "the processing of big data now proposes to provide individualized information by mapping data to metadata" (p. 94) should be qualified, given that the clinical ambition of formal thought (particularly algorithmic thought) is currently encountering cyclical [12] and structural [13] limits. Finally, even more radically, and as strange as it might seem, we may contest the very idea that the concept of "data" [14] supposedly "coding" external reality (p. 32) could be used as a definition of computer science. Indeed, some computer scientists prefer to use the notion of "documents-for-a-certain-interpretation," [15] inventing technologies of suggestion, interpretation, and comparison, in short assisted action and not automated action, human-machine interaction and not delegation to machines (and the reciprocal fantasy that goes with it of machines taking over). The marginal place the authors occupy in the current landscape of computer science takes nothing away from the relevance of their remarks.