Introduction: Society 5.0

We describe the latest manifestation of the digital revolution as a translation from print capitalism (Anderson 2006/Reference Anderson2006/1983) to data capitalism by taking Japan’s information and communication technology (ICT) development plans as a case study. Particularly, by employing Donzelli’s (Reference Donzelli2025a, Reference Donzelli2025b) semiotic criticism of capitalism as a translation machine, with inputs from science and technology studies (Jasanoff Reference Jasanoff, Sheila and Kim2015; Zuboff Reference Zuboff2019) and biosemiotics (Marais Reference Marais2019), we analyze the discourse of Society 5.0 (Cabinet Office n.d.), Japan’s technology development plan, as contextualized in the country’s infrastructure of dataveillance. We observe a techno-nationalist (Humphrey Reference Humphrey2022, Reference Humphrey2024) justification of dataveillance, reducing practices of reading and writing from hermeneutically open and complex to code inscription.

Society 5.0 (Cabinet Office n.d.) is Japan’s futuristic policy for developing a “human-centric” (ningenchūshin no shakai 人間中心の社会) artificial intelligence (AI) infrastructure. It was launched in 2016 as the latest iteration of Japan’s ICT policy efforts. It builds on the e-Japan project, inaugurated in 2001 to develop broadband infrastructure, and which transitioned to u-Japan by 2006 (MIC 2005-2007), to keep up with the emergence of deep learning. Society 5.0 shifts from ubiquitous digital access and Internet connection (u-Japan) to AI, big data, and the Internet of Things (IoT).

The Government of Japan (Cabinet Office n.d.) describes Society 5.0 as a “vision of the future,” building on “the hunting society (Society 1.0), agricultural society (Society 2.0), industrial society (Society 3.0), and information society (Society 4.0).” In this imagined future, intelligent machines powered by AI and big data create new markets, reinvigorate the economy, reform health care, strengthen social relations, bolster national security, and promote resource sustainability.

The Council for Science, Technology and Innovation (2015, 13) of the Government of Japan thinks of Society 5.0 as a “super smart society” where

While some objectives of this project are commendable, reflecting a concern for (democratic) social development, much of its language is vague and carries ideological presumptions. The project sets out to “integrate” or “merge” infrastructures, but this language of connectivity ushers a discourse that strongly endorses surveillance practices (see Mantello and Olteanu Reference Mantello and Olteanu2026), which shape or, rather, reinforce imaginaries of the Japanese nation. Nationalism stems explicitly in key moments of the Report on the 5th Science and Technology Basic Plan, which is the starting point for Society 5.0, such as for example, that

Technology development is explicitly claimed to be a “national” matter and the desirable future that development should lead to, which this document names “Target National Image,” is to be communicated top-down, from government to citizens.

Reflecting on the vagueness of the document, the Council for Science, Technology and Innovation (2015, 14) claims that “to realize a super smart society, it is necessary to connect various ‘things’ via a network, create highly advanced systems out of these things, and integrate diverse systems so that they can coordinate and collaborate with each other.” Such connectivity, actually, spells out as data collection, and the document explains this clearly: “This integration allows for a wide variety of data to be collected, analyzed, and applied across all the coordinating and collaborating systems in order to continuously produce new values and services” (2015, 14). Collecting data from the “things” connected implies a translation of things into monomodal “data,” which becomes a universal epistemological system. Our present paper discusses the Society 5.0 project with respect to literacy, particularly how it shapes reading and writing.

Like many nations, Japan is transforming brick-and-mortar retail into data-intensive environments. These establishments form a datafication infrastructure aligned with state ICT discourse. Unlike older CCTV that merely recorded, these AI-based retail media initiatives identify and evaluate customer behavior through behavioral and biometric parameters (i.e. facial recognition, age, gender, gait, and eye movements). Retailers now retrieve histories of visits, spending, and preferences, assigning risk scores and alerting staff to offenders. The newest iterations integrate emotion–recognition algorithms that infer affective and psychophysical states (Mantello et al. Reference Mantello, Manh-Tung, Ming-Hoang and Vuong2023; Mantello and Olteanu Reference Mantello and Olteanu2025), classifying individuals by quantified opportunity or risk, thus reducing sensing to counting (Del Percio and Vigouroux Reference Del Percio and Vigouroux2025). This shift from passive recording to a simulation of sensing is an effort to bestow on corporate legal persons what they always lacked in comparison to human legal persons, namely a body (Lat. corporare) that is situated and performative (Hull Reference Hull2025). The (human) subject is written into society through a surveillant infrastructure that, to begin with, determines how the subject is read. Society 5.0 is, thus, the hermeneutic framework for new forms of machinic reading–writing.

Translating the past

Given the (modern) history of media technologies (Hayles Reference Hayles2002, 31–2), reading and writing are the semiotic processes that produce legal persons (Hull Reference Hull2025): an entity has the right to sue and be sued because it is written into society by a state authority. Traditionally equated with skills of representing society through the alphanumeric symbols of the printing press, reading and writing constitute literacy. Literacy has been the aim of modern mass education, as supposed to empower individuals into citizenship (Campbell and Olteanu Reference Campbell and Olteanu2024). In this media history, citizenship became equated with semiotic competences of operating print, thus inculcating a language-centric industry, arguably manifest as what Anderson (2006/Reference Anderson2006/1983) termed print capitalism. What could have been and, due to the recent pluralization of media technologies, is becoming a broad and more inclusive “semiosy” (Stables 2012), has historically been a narrow “literacy.” In brief, logical and critical thinking have been historically confused with skills to manipulate the symbolic code of print (Barton Reference Barton1994; Hartley Reference Hartley2012, 18). Limiting the capacity to exercise civil rights and liberties, as well as to thrive economically, to operating within the affordances of print induces a view of language as separate from other sign systems, which “pre-empts the possibly quite various assumptions about communication around the globe” (Gal, Irvine Reference Gal and Irvine2019, 11). Such sociotechnical imaginaries of print fuel a view of sign systems as arbitrary and closed, championed in linguistics by Saussure (1959/1916). Here we see an important advantage of Donzelli’s (Reference Donzelli2025b) approach to structuralist and Foucauldian discourse analysis, namely construing technological infrastructure as part-and-parcel of discourse. With computing becoming ubiquitous, we find this theory urgent. Two decades ago, Leeuwen (Reference van Leeuwen2008, 132) observed the rise of “new writing,” detached from the linguistic mode, “at once more visual than the old ‘page’ media, and less pictorial than the old ‘screen’ media.” More than shifting from a “grammar of sentences” to a “grammar of the diagram, grid, or network” (Leeuwen Reference van Leeuwen2008, 132), AI codifies writing as self-inscription, rendering it inaccessible to its author (Lassègue Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025). As such, aligned with Donzelli, we think through a concept of “writing” unconfined to language and print, as sociocultural representation modes that constitute hermeneutic sets of competences for thriving in society—in one word, literacy. In this sense, literacy is not singular or fixed. Governments often make abstraction of how ICT development policies reshape literacies and, hence, overlook developing the required educational policies (e.g. Hartley Reference Hartley2012; boyd Reference boyd2014; Campbell and Olteanu Reference Campbell and Olteanu2024). Society 5.0, we explain, is a prime example of how an old imaginary is renovated through a sociotechnical transformation.

The modern equation of literacy with citizenship views linguistic difference as the mark of systemic boundaries that “naturalize hierarchy and domination” (Gal and Irvine Reference Gal and Irvine2019, 1), making imagined communities appear empirically objective states of fact. As Hartley (Reference Hartley2012, 39) explains, “complex networks can be tracked and modeled where once they were what Benedict Anderson famously called ‘imagined communities,’” but now “need be imagined no longer; they are data.” In Japan, we observe a historical legacy of a specific print imagination (Shockey Reference Shockey2020) exploited to entrench the boundaries and hierarchies imagined through print while transitioning to a mediascape of ubiquitous computing. We explain how the business sector and government collaborate toward a dataveillance infrastructure that maintains print imaginaries.

We analyze dataveillance in Japan through a broad concept of translation applicable as “a metasemiotic infrastructure of capitalist valorization” (Donzelli Reference Donzelli2025a, 321), dismissing a traditional, much criticized (Gal Reference Gal2015) but still common view of translation as “carrying over of meaning across self-contained-codes” (Donzelli Reference Donzelli2025a, 321). We emphasize that, following Donzelli’s framework, we do not use the term translation merely as a metaphor for transformation, with a semiotic flavor. Particularly, we ground our analysis in a tradition of thinking on translation that takes inspiration from the semiotics of Charles S. Peirce, first explored by Jakobson (Reference Jakobson and Venuti2004/1959), and which is recently receiving increased attention (Silverstein Reference Silverstein, Basso and Selby1976, Reference Silverstein, Rubel and Rosman2003; Petrilli and Ponzio Reference Petrilli and Ponzio2005, 87; Petrilli Reference Petrilli2016; O’Halloran et al. Reference O’Halloran, Tan and Wignell2016; Elleström Reference Elleström2018; Olteanu Reference Olteanu2021a; Xu, Olteanu, and Campbell Reference Xu, Olteanu and Campbell2025), possibly due to its capacity to encompass intermediality and multimodality, as necessary in digitalizing contexts. This view on translation explains that work with cultural artifacts is translational, not in a metaphorical sense, as extrapolated from interlinguistic translation. Rather, as Marais (Reference Marais2019, 39) explains, because culture is semiotic, “culture emerges from translational actions.”

This Peircean line of thinking, which we explain further below, to ground Donzelli’s criticism more firmly, is at odds with discourse analysis and structuralism. Discursive approaches as analysis of forms (Barthes 1972/Reference Barthes1972/1957; Derrida 1978/Reference Derrida1978/1967; Foucault Reference Foucault and Young1981) miss the transmedial aspect of translation processes because, by distinguishing “between materiality and ideation,” they place “language in an ideational-mental realm” (Gal and Irvine Reference Gal and Irvine2019, 15). These approaches assume the semiological dichotomy of signifier and signified (Saussure 1959/1916), which ushered in a methodology based on observing differences. This has been criticized for its exaggerated relativism, as it tends to invent boundaries by conceiving cultures and languages as closed systems (Gal, Irvine Reference Gal and Irvine1995, Reference Gal and Irvine2019; Olteanu Reference Olteanu2021a, Reference Olteanu2021b). Such criticism (Gal Reference Gal2015, Reference Gal, Duranti, George and Riner2023) inspired Donzelli’s (Reference Donzelli2025a, Reference Donzelli2025b) critical approach to capitalism, illustrated by a special issue in this journal. Donzelli’s framework pays attention to infrastructure, traditionally overlooked in discourse analysis, by building on Berardi’s (Reference Berardi2009, 18) conceptualization of semiocapitalism as “the new regime characterized by the fusion of media and capital,” where “poetry meets advertising and scientific thought meets the enterprise.” This broad notion has the merit of conceiving capitalism as a transmedial process (which may or may not include Jenkins’ (Reference Jenkins2006) convergence), comprising both technological infrastructure and discursive content. We observe convergence of media technologies and economic infrastructure as a goal of Society 5.0, clearly displaying Berardi’s theory.

Translating Japan from print to digital media

As digital datafication enables the flattening of otherwise complex hermeneutics of translation processes, it serves capitalism “to expand without changing the business model” (Donzelli Reference Donzelli2025a, 318; see Tsing Reference Tsing2015). We argue that dataveillance in Japan serves the expansion of capitalism while preserving past social imaginaries. While this requires a work of translation from old to new media, it is also the convenient acceptance of social inertia. Salient for digitalization, Gitelman (Reference Gitelman2006, 4) recapitulates the basics of media theory:

Society 5.0 employs surveillance technologies to reinforce nationalist imaginaries, as the former become the new sites for organizing representation. As societies become constructed through representations that make use of increasingly more varieties of sense perception channels and sign systems, an imaginary that relies chiefly on printed language becomes unsustainable. As such, in Japan, we are witnessing an effort to justify nationhood through new media technologies (Humphrey Reference Humphrey2022). As Inoue (Reference Inoue2016, 152) observes about Japanese society, “new forms of mediated public culture, and new creative imaginaries of the social world of the unprecedented present have emerged, both reflecting and troubling neoliberal common sense.” While Inoue remarks on this to critically reflect on the changes to Japanese gendered language in the context of growing surveillance through digitalization, here we observe similar dynamics in regard to nationalism.

Our analysis also informs Donzelli’s framework with insights from the interrogation of emergence in the biosemiotic approach to translation (Marais Reference Marais2019). Like the anthropological semiotic theory on which Donzelli builds (Gal and Irvine Reference Gal and Irvine2019), biosemiotics favors Peirce’s pragmatic semiotics to Saussurean-inspired discourse analysis. A centerpiece of Gal and Irvine’s (Reference Gu2019, 15) semiotic anthropology is that, following Peirce, “the continuous chains of abduction that constitute both communication and knowledge do not distinguish linguistic signs from any other types of signs.” While he did not address translation much, Peirce mentioned that his semiotic theory leads to a broad notion of “transuasion (suggesting translation, transaction, transfusion, transcendental, etc.)” as “mediation or […] being in creating Obsistence” (CP 2.89). This enabled Marais (Reference Marais2019) to conceptualize translation as energy transfer (labor) that involves semiosis. This theory aligns with the scope of philosophy of technology to see systemic changes not as leaps to new ontological levels, but as the emergence of complexity (Fuller Reference Fuller2022; Xu et al. Reference Xu, Olteanu and Campbell2025). We find this concept not just compatible but a useful underpinning for Donzelli’s critical approach to capitalism in the interrogation of translation, especially in the current context of technological transformation.

Our hypothesis is that if surveillance infrastructures transform capitalism from mass to individual consumption preferences via networked digital spaces (Zuboff Reference Zuboff2019), then emergent sociotechnical imaginaries must be scaffolded on imaginaries fabricated along the historical trajectory of print capitalism. Data capitalism builds on the imagined communities enabled by print, inculcating the perception of correspondence between language and a political community “both inherently limited and sovereign” (Anderson Reference Anderson2006/1983, 6). In Society 5.0 and its pro-business regulatory softlaw (METI 2024), we trace a transmediation of hegemony, a media future (Ernst and Schröter Reference Ernst and Schröter2021) that draws nationalism from the past.

Society 5.0 developed from previous ICT plans, which have transformed along with the fast-paced technological changes in the past three decades. It shifts from ubiquitous digital access (u-Japan) to AI, big data, and the IoT. What Society 5.0 shares with its predecessors, e-Japan and u-Japan, especially visible in the translation from one to another, is an overtly optimistic techno-nationalism that positions Japan as a preeminent (post-)digital model nation that reimagines itself with every new technological development.

Language-centrist myths

This technological trajectory of capitalism rests on three myths that are fueled by language-centrism. The first myth is that technologies augment human thinking without altering the mind of Homo sapiens (Fuller Reference Fuller2022). Second is that behavioral data is a freely available and value-free natural resource (Couldry and Mejias Reference Couldry and Mejias2019). Third is that anonymization ensures data neutrality, which supposes either crude positivism or digital illiteracy. These myths are underpinned by the classic concept of translation as maintaining content (signified) while changing form (signifier), but in a positivist reductionism that is at odds with Saussurean semiology. Equating translation with hermeneutically neutral transcription of code enables capitalism to function as a “translation machine” (Tsing Reference Tsing2015; Donzelli Reference Donzelli2025a, 313). Rather, we think of translation as a complex form of recontextualization (Gal Reference Gal, Duranti, George and Riner2023, 179–80; Donzelli 2025, 317), intrinsic to mind work (Petrilli Reference Petrilli2014, Reference Petrilli2016; Elleström Reference Elleström2018; Marais Reference Marais2019).

These three myths justify ICT development through past media futures, namely that past imaginaries can be entertained as translated into new mediascapes, a conceptual conservatism that is characteristic of Japan (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022; Humphrey Reference Humphrey2024). They assume the dichotomy of form and content that underpins the physical symbol systems hypothesis (Gevarter Reference Gevarter1985), namely that intelligent behavior arises from disembodied symbol manipulation. This hypothesis lies at the foundation of robotics, having inspired the development of computing machines and AI technologies in general. It supposes that sign systems are clearly bounded from one another. Developed in light of the symbol systems hypothesis, digital machines polarize further: writing becomes software coding (Emmeche Reference Emmeche2001, 667), and reading becomes human interaction with spaces designed to conceal code (see Gitelman Reference Gitelman2006, 96, 143; Manovich Reference Manovich2013; Kitchin and Dodge Reference Kitchin and Dodge2014; Paulsen Reference Paulsen2025). In the case of technologies with legal intricacies, such as surveillance, this abstraction removes legislating from those to whom laws apply (Hildebrandt Reference Hildebrandt2018; Lassègue Reference Lassègue, Cappiello and Carullo2021, Reference Lassègue2022, Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025; Hull Reference Hull2025), which is our contention regarding dataveillance in Japan.

If print capitalism was entangled with Illuminist ideals of emancipating citizens who understand civil rights (Hartley Reference Hartley2012; Stjernfelt and Lauritzen Reference Stjernfelt and Lauritzen2020; Hull Reference Hull2025), data capitalism relies on blind trust in machines to legislate. As there is no legal language to express such trust to automatized legislation (Lassègue Reference Lassègue, Cappiello and Carullo2021, 61), it can only be justified through ideological presumptions and the naïve belief that automatization can maintain content while changing form. To illustrate, in early modernity, it was supposed that a legal document that establishes a corporation as a legal person should be public (“letters patens”; Hull Reference Hull2025, 359–60). Moreover, the legal status of the corporation was not understood as deriving from the legal text, but the utterances, purposefully in the present tense, constituted the legal status of the corporation. In this logic, if writing is reduced to code, not only should code be public but also cogently readable. In Japan, datafication is justified through nationalism intrinsic to technological optimism (Humphrey Reference Humphrey2022, Reference Humphrey2024; Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022; Wright Reference Wright2024). Our point is that unregulated AI-powered surveillance erodes democracy by rendering citizens illiterate. This occurs in a historically new condition of literacy, in which “control over the writing process was henceforth lost by individuals and became a collective enterprise nobody alone could have a full grip on” (Lassègue Reference Lassègue, Cappiello and Carullo2021, 63). One aspect for which in modernity critical thinking became confused with skills of using print is that these skills entailed the deciphering not only of printed symbols but also of authorship and legitimacy of vernacularly circulating prints (Gitelman Reference Gitelman2006, 27–8).

Our argument also hinges on distinguishing “data” from “information.” Data is a basic, indexical dimension of meaning-making, and information emerges from reading data. Dataveillance collapses this distinction, masking data’s political transformation into information, where power becomes operational. Through this translation, knowledge and governance coalesce. In the Japanese techno-nationalist context, the conflation of data and information enables the state and corporations to claim neutrality while enacting discreet forms of governance and persuasion (Richards and King Reference Richards and King2013). Treating data as objective truth, interpretative possibilities disappear and, with them, accountability. What is presented as objective quantification is in fact a semiotic inscription that transduces bodies, gestures, and affect into actionable and governable signs. The political force of dataveillance lies not in data collection alone, but in its conversion into information that seemingly speaks without mediation. Still perceived as “new media,” AI surveillance technologies have to become a technology of inscription, to determine what and how writing itself is, as “the success of all media depends at some level on inattention or ‘blindness’ to the media technologies themselves (and all of their supporting protocols) in favor of attention to the phenomena, the ‘content,’ that they represent for users’ edification or enjoyment” (Gitelman Reference Gitelman2006, 6). This illusion of immediacy, stemming from the dichotomy of form and content, legitimizes algorithmic decision-making as rational and ethical, situating surveillance within the domain of Japan’s safe society legacy, masquerading control as care, which blurs the distinction between freedom and control (see Chun Reference Chun2006; Inoue Reference Inoue2016). Thus, Japan’s project of harmony through datafication conceals the monopolization of interpretation by those who wield power and own the means of datafication, in turn silencing many whose lives are rendered into code.

Under print capitalism, reading and writing technologies functioned disciplinarily, shaping subjects through institutional norms of literacy. A Foucauldian perspective reveals this but remaining anchored in the dichotomy of form and content, it is unfit for digital modalities (see Inoue Reference Inoue2016, 154). To account for the complexities of the translation machine, we find it necessary to construe literacy as multimodal and unfixed (van Leeuwen Reference van Leeuwen2008; Machin Reference Machin2016; SReference Stablestables 2012; Campbell and Olteanu Reference Campbell and Olteanu2024). Also, as Society 5.0 evokes imaginaries that are irreducible to linguistic content, we take as guiding concept the sociotechnical imaginary, “collectively held, institutionally stabilized, and publicly performed visions of desirable futures, animated by shared understandings of forms of social life and social order attainable through advances in science and technology” (Jasanoff Reference Jasanoff, Sheila and Kim2015, 4). Digital nationalism is a sociotechnical imaginary enabled by dissociating data from imagination; it forgets that data, too, is imagined (through media technologies), and that what counts as data is subject to social convention. Society 5.0 postulates that the nation consists of data that a surveillance infrastructures senses, as a body would (Humphrey Reference Humphrey2022). As such, it shapes writing and determines what is being recorded, that is, what is the subject of historiography (see Gitelman Reference Gitelman2006). Like print capitalism inculcated the imaginary that proficiency in a (native) language factually corresponds to a culture, which makes a worldview, and which means belonging to a nation (Adami Reference Adami, Schröder, Adami and Dailey-O’Cain2023), Society 5.0 supposes that an entity belongs to a nation or not according to data that objectively corresponds to that entity, making it a legal person. We illustrate this in technological infrastructures such as koseki and MY NUMBER (below).

From print to digital imaginaries

In modernity, writing operated as discursive inscription (Gitelman Reference Gitelman2006; Hartley Reference Hartley2012; Petrilli Reference Petrilli2014), feeding the regimes of truth that defined what was sayable and knowable. In Japan, Shockey (Reference Shockey2020) describes the compulsory engagement with texts regulated through the printing press as a distinct typographic imagination. These disciplinary practices yielded compliant subjects who internalized moral codes, norms, and national imaginaries. Authors were interpellated through print affordances, as storytelling was bound by censorship, societal expectations, and institutional protocols. With the late-19th-century boom of printed outlets, especially during wartime, print capitalism became deeply entangled with Japanese identity.

With data capitalism, regimes of truth yield to regimes of freedom-control (Chun Reference Chun2006). Individuals are interpellated not through ideological enclosures but algorithmic feedback loops. Navigating data flows became essential for reading in an environment where the reader becomes part of data (Abdul-Jabbar and Bhatt Reference Abdul-Jabbar and Bhatt2025). As writing becomes an involuntary inscription of the self, critical literacy needs to be reimagined. Andrejevic’s concept of infoglut (Andrejevic Reference Andrejevic2013) captures this shift: the proliferation of information overwhelms interpretive capacities, turning apparent empowerment into surveillance. The subject is shaped through the very data they generate.

We consider that prevailing sociotechnical imaginaries, where literacy is equated with encoding and decoding, remain anchored in print-based epistemologies. These imaginaries sustain the perception of symbolic expression as a neutral medium, an ideal conduit for unbiased representation (Petrilli Reference Petrilli2014), privileging data as objective truth. The monopoly of print, especially in education and research (Kress Reference Kress2010; Lacković and Olteanu Reference Lacković and Olteanu2024), historically inculcated the assumptions that (1) concepts are solely linguistic; (2) language is the medium of logic and criticality; and (3) print can represent ideation with perfect objectivity. This enduring fallacy to equate interpretation with a rigid notion of translation as changing form while maintaining content (Petrilli and Ponzio Reference Petrilli and Ponzio2005; Marais Reference Marais2019), belongs to the trajectory of datafication that Innis (Reference Innis1964 [1951]) described as a “desert” of classification and abstraction (McLuhan Reference McLuhan and Innis1964 [1951], xi, Reference McLuhan1994, 85).

Data capitalism extends this trajectory. While digital technologies promise diversity and hermeneutic richness, when designed through print imaginaries, they constrain (see Gitelman Reference Gitelman2006, 13). Unlike the reflective communities cultivated by print capitalism, dataveillance deterritorializes subjects (Fawzy Reference Fawzy2023; see Inoue Reference Inoue2016, 154), reconstituting them as data-points. Print capitalism, though bound to nationalist segregation, enabled democracy through literacy, as an Illuminist ideal (Campbell and Olteanu Reference Campbell and Olteanu2024; Lacković and Olteanu Reference Lacković and Olteanu2024). Critical literacy not only challenged feudal orders but also forged transnational publics, such as the Republic of Letters (Stjernfelt and Lauritzen Reference Stjernfelt and Lauritzen2020, 10). Printing infrastructures and steam engines nevertheless constrained criticality within nation imaginaries, which illustrates why technology must be understood as integral to discourse.

If print once summoned a coherent public sphere, data capitalism fragments it through microtargeted feeds, collapsing the distance between expression and meaning. Andrejevic (Reference Andrejevic2013) argues this deluge engenders a new barrenness, a desert not of classification but of literacy, where knowing eclipses understanding, overwhelming the semiotic capacities through which subjects once oriented themselves toward meaning. Consequently, a kind of “epistemic automation” unfolds in which cognition is scaffolded not through reflection or embodied interpretation but predictive algorithms and feedback loops, collapsing the distinction between data as something neutral or apolitical and information as inherently ideological. The goal of this post-interpretative epistemology, Andrejevic asserts, is to automate the work of meaning-making (Andrejevic Reference Andrejevic2020), effectively turning society into a recursive system that interprets itself. Here, correlation supplants causation, anticipation replaces understanding, and ideology becomes algorithmic inference. Users are isolated within digital enclosures rather than gathered into literate publics that negotiate (Pariser Reference Pariser2011; Hildebrandt Reference Hildebrandt2018; Lassègue Reference Lassègue, Cappiello and Carullo2021, Reference Lassègue2022, Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025). As state–corporate partnerships standardize online practices (Stjernfelt and Lauritzen Reference Stjernfelt and Lauritzen2020, 155), the transition from print to data capitalism demands reconceptualizing nationhood. What was imagined through presses, broadsheets, and novels becomes a scaffold for computational constructs of sovereignty.

In Japan, past imaginaries persist through conceptual conservatism masked as progress (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022). Datafication, whether for convenience, profit, or control, operationalizes digital forms of reading and writing. If writing is a practice that enables literacy (Lassègue Reference Lassègue, Cappiello and Carullo2021, Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025), reducing it to coded transcription flattens critical to merely functional literacy. Criticality supposes hermeneutic competences of working with the affordances through which representation and possible interpretations are entangled (Petrilli and Ponzio 2005; Machin Reference Machin2016; Marais Reference Marais2019). Hegemonies, on the other hand, replicate old protocols through new media (Bruce Reference Bruce2018): today’s collected data are yesterday’s imaginaries, “converting individuals and social practices into population data” (Goode Reference Goode2019: 365). As dataveillance operationalizes new literacies intrinsic to the processes of digital capitalism, digital literacy should comprise corresponding semiotic competences. The issue is that the automation of reading and writing scripts self-perception into fixed imaginaries.

Our argument proceeds in three parts: first, a theoretical interrogation of sociotechnical imaginaries surrounding data capitalism; second, an analysis of Society 5.0, echoing Zuboff’s (Reference Zuboff2019) behavioral futures marketplaces; third, an examination of how Japan’s data-protection soft laws (METI 2024), media discourse, and business corroborate.

Theoretical framework

The dichotomy of “form” and “content,” criticized by Gal and Irvine (Reference Gal and Irvine1995, Reference Gal and Irvine2019), results in a relation between infrastructure and discourse. As such, the object of discourse analysis has been traditionally identified with (linguistic) content. Machin (Reference Machin2016; cf. Fairclough Reference Fairclough2002) has pioneeringly indicated the importance of considering technological infrastructure as intrinsic to discourse, in the interrogation of affordance. That is, the infrastructure of circulation is semiotic, intrinsic to discourse (Donzelli Reference Donzelli2025a, 313; Handman Reference Handman2025, 8–10). Bridging this divide can align semiotic analysis with science and technology studies (Jasanoff and Kim Reference Jasanoff and Kim2015) and post-interpretative epistemologies.

The notion that a concept (signified) can take any form (signifier) stems from Saussure’s (Reference Saussure1959 [1916]) theory of arbitrariness, where opposition is the main analytical concept (Jakobson Reference Jakobson1965; Fairclough Reference Fairclough1992, 89; Gal and Irvine Reference Gal and Irvine1995, Reference Gal and Irvine2019; Marais Reference Marais2019, 96). In this view, for a system to be socially useful, it must be conventional and fixed, not subject to individual accident, which leads to policies that impose strict monolingualism at the cost of freedom of expression (Olteanu Reference Olteanu2021a; Adami Reference Adami, Schröder, Adami and Dailey-O’Cain2023; Handman Reference Handman2025, 13–4). Saussure’s concept constitutes the backbone of (post)structuralist deconstruction (e.g. Barthes 1972/Reference Barthes1972/1957; Derrida 1978/Reference Derrida1978/1967) and discourse analysis (Foucault Reference Foucault and Young1981; Fairclough Reference Fairclough1992, 74–5, 89; Lévi-Strauss Reference Lévi-Strauss1963), where methodologies rely on more nuanced concepts than opposition, such as semantic differential and différence. These frameworks analyze society through the identification of differences, which indicate boundaries. Supposing boundaries within a system, within (signifier/signified) and between signs (Saussurean value), assures the system’s conventionality (arbitrariness). If a system is conventional, it functions in isolation from other systems, according to its own internal logic. As such, strong boundaries are implied between systems that may interface (e.g. two languages spoken on the same geographical territory—alphabet and emojis). Lotman’s (2005/Reference Lotman1984) theory of the semiosphere, which Donzelli (Reference Donzelli2025a, 314–135) correctly finds useful for ushering the broadened notion of translation on which she relies, is nevertheless another instance of semiotic analysis that relies on differentiation (e.g. Lotman 2005/Reference Lotman1984, 208). Biosemiotic theory, on which Marais (Reference Marais2019) develops a concept of translation as intrinsic to meaning, advocates a more encompassing semiobiosphere (Petrilli and Ponzio 2005). Without diving into details, we mention that far from being merely a matter of labeling, the difference between (Lotmanian) semiosphere and (biosemiotic) semiobiosphere marks a criticism of both language-centrism and anthropocentrism. The semiobiosphere is a better conceptualization of the continuity of semiotic systems globally, accounting for translation not as “semiotization of incoming materials […] into information” (Lotman 2005/Reference Lotman1984, 210), but always as resemiotization of information into differently situated and interpreted information (Marais Reference Marais2019, 28–9), or circulation (O’Halloran et al. Reference O’Halloran, Tan and Wignell2016, 201, 203).

The structuralist epistemology of differentiation fueled exaggerated relativism in anthropology, linguistics, and semiotics (Gal and Irvine Reference Gal and Irvine1995, Reference Gal and Irvine2019; Olteanu Reference Olteanu2021a). Also, it views human–computer interaction in light of the physical symbol systems hypothesis, polarizing discourse and technological infrastructure as mutually independent, instead of entangled through affordances. This seeded the divide between media theory and semiotics, contradicting current trends to construe technology as mindwork (Hayles Reference Hayles2017; Fuller Reference Fuller2022; Mantello and Olteanu Reference Mantello and Olteanu2025). Also, it propagates language-centrism, the assumption that all sign systems derive from verbal language (Sebeok Reference Sebeok1986; Petrilli Reference Petrilli2014; Olteanu Reference Olteanu2021a). Digitization has inspired the rejection of this dichotomy within semiotics (Olteanu Reference Olteanu2021b), ushering the multimodality framework (O’Halloran and Smith Reference O’Halloran and Smith2012; Machin Reference Machin2016; O’Halloran et al. Reference O’Halloran, Tan and Wignell2016), which calls for conceiving discursive content and technology as intertwined through affordances, not bounded by opposition.

Concerned with relations within sign systems (value), Saussure’s semiology avoids questions on reference. As such, it avoids representation theory (Marais Reference Marais2019, 148), justifying the physical symbol systems hypothesis in robotics and the analysis of discursive content as unaffected by how information circulates through infrastructures (form). Our analysis of dataveillance in Japan contributes to collapsing this dichotomy.

Surveillance is a way of perceiving that becomes social representations of power and control (Lyon Reference Lyon2023; Boczkowski Reference Boczkowski2010, 64). Communities understand themselves through how they perceive themselves. Print capitalism has evidenced how writing is formative for shaping social imaginaries (see Handman Reference Handman2025, 19–20). The spread of print-standardized languages enabled mass literacy, creating literacy-based cultures. In those conditions, communities began to imagine themselves as nations. Data capitalism (Murakami Wood and Monahan Reference Murakami Wood and Monahan2019) shapes imaginaries by reducing writing to automatized algorithms, where data is construed as a natural resource extracted for value (Couldry and Mejias Reference Couldry and Mejias2019). Technocratic and positivist epistemologies of dataism that erase the data–information distinction demand a renewed semiotic attention to technology. In this direction, the multimodality framework (van Leeuwen Reference van Leeuwen2008; Kress Reference Kress2010; O’Halloran and Smith Reference O’Halloran and Smith2012; Machin Reference Machin2016) enables a construal of discourse and technology as intertwined, but which was seldomly pursued. Apart from a few exceptions (Fawzy Reference Fawzy2023), the proposal to understand (new) technologies as inherent to discourse (Boczkowski Reference Boczkowski2010; Paulussen and Harder Reference Paulussen and Harder2014) has so far only been noted.

The language-centric reduction of discourse to “text and talk” arises from mistaking print infrastructures for rationality (Hartley Reference Hartley2012). Typing thus appears as a neutral transcription of thought (Petrilli Reference Petrilli2014). While surveillance is a classic discursive concern (Foucault 1995 [Reference Foucault1995 [1975]]), only a few semiotic studies address digital dataveillance and they do so by treating discourse as mediated representation, distinct from technology (Fernback Reference Fernback2007; Simone Reference Simone2009; Tiainen Reference Tiainen2017; Silva and de Oliveira Maia Reference Silva and de Oliveira Maia2022). In the case of computing machines producing discourse, affordances are shaped by algorithmic computation that feeds into cognition (Hayles Reference Hayles2017; Mantello and Olteanu Reference Mantello and Olteanu2025). Semiotic and discourse analyses often focus on the outputs that machines display for human reading (e.g. Putland, Chikodzore-Paterson, and Brookes Reference Putland, Chikodzore-Paterson and Brookes2023), overlooking computation itself, where writing occurs and collective imaginaries form.

Thus, we see in the Donzelli’s (Reference Donzelli2025a) framework a useful pathway to eschew the Foucauldian approach in analyzing writing with AI. Discourse analysis overlooks how technologies such as AI, cloud computing, and machine learning extend the state’s capacity to collect vast data (Mantello et al. Reference Mantello, Manh-Tung, Ming-Hoang and Vuong2023). The functioning of technology should be understood as part-and-parcel of discourse, not merely a form that carries discursive content.

We reflect on dataveillance through a semiotic concept of writing (Petrilli Reference Petrilli2014; Lassègue Reference Lassègue, Cappiello and Carullo2021, Reference Lassègue2022, Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025), aligned with media and technology studies (Kittler 1999 [Reference Kittler1986]; Hayles Reference Hayles2002). Bridging discourse analysis with science and technology studies enables the joint consideration of content and infrastructure. Following Machin (Reference Machin2016, 327), we interrogate discourse through how semiotic resources and affordances shape each other. To unearth conservative and techno-optimist imaginaries in Japanese nationalist discourse, we examine spaces (Lassègue Reference Lassègue2022, Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025) of dataveillance. To illustrate how dataveillance in Japan marks a transition from print capitalism (Anderson 2006/Reference Anderson2006/1983) to data capitalism (Zuboff Reference Zuboff2019), we bring in conjunction to semiotic analysis, concepts from technology (Jasanoff Reference Jasanoff, Sheila and Kim2015), surveillance (Murakami Wood and Monahan Reference Murakami Wood and Monahan2019; Zuboff Reference Zuboff2019), and data studies (Couldry and Mejias Reference Couldry and Mejias2019; Andrejevic Reference Andrejevic2013). In this view, the capitalist trajectory from print to dataveillance is a translation of print-based imaginaries into fixed digital databases, where citizens become numerical signifiers, or dividuals (Deleuze Reference Deleuze1992; see Inoue Reference Inoue2016, 167).

Japanese techno-nationalism: new writing, new imaginaries

Society 5.0 is a political project that structures imaginary uses of technology (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022; Wright Reference Wright2024), exhibiting “an ideology and doctrine that advocates, endorses, and sanctions a techno-nationalist agenda of ‘future-as-idealised-past’” (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022, 451). The rationale is to overcome current crises through the implementation of AI infrastructures, while maintaining the modern imaginary of the Japanese nation.

Portraying Japan as a model of digitalizing nationhood, this discourse “goes forward by going back” (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022, 451), justifying a past media future: it resemiotizes the nation imaginary in digital infrastructures. Exploiting the particular techno-optimism of Japanese nationalism, conservative ideology obstructs imagining new possible futures. This “vision of the future” of the Japanese nation-state remains bound to outdated dichotomies pertaining to classical humanism. For example, it supposes clear distinctions between human and technology (or machine), by preaching “human-centered” technological development and a differentiation between “cyberspace,” where computational technologies are deemed to operate, and “physical space,” supposedly populated by humans (Cabinet Office, n.d.). The tarrying of these modern dichotomies confuses: what is the human, as distinct from its technological extensions, so that technology is “centered” on it? Where is human agency located, if cyberspace is distinct from physical space? Ultimately, who is the beneficiary of policy? The 5th Science and Technology Basic Plan states that in the super smart society sought “humans and robots and/or artificial intelligence (AI) coexist and work to improve quality of life by offering finely differentiated customized services that meet diverse user needs” (Council for Science, Technology and Innovation 2015, 13). The term “coexist” is confusing in a discourse of “integration”: which of these two is the end goal? In this same equation, it is not clear the quality of whose life should be improved. How is life defined in an unfolding media future where cyber- and physical space merge?

A key motif in Japan’s techno-nationalist discourse is what Robertson (Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022) calls “kinship technologies.” While policy documents explicitly mention diversity, nostalgia for a past that never was, typical of nationalism, is reinforced (Massumi Reference Massumi2015). This concept posits that technologies ideologically shape imagined social ties. Hull (Reference Hull2025, 357) observes that what for human persons are kinship relations, for corporations are political relations (which does not exclude that human kinships can be political). Thus, a corporate person charges the technology it champions with its political relations. Print, for instance, is a kinship technology: it historically inculcated communities imagined through language, producing modern nations via standardization. Nations, then, are corporate persons that rely on the imaginative affordances of print. By imagining themselves as monolingual communities, literate publics required self-governance. Print capitalism marks an early step in datafication (Koopman Reference Koopman2019; Zuboff Reference Zuboff2019), the translation of human life into machine-readable data. Surveillance technologies, we argue, translate old imaginaries into digital databases to justify nation-states in contemporary mediascapes. According to Society 5.0 (see Fig. 1), social media, IoT, and AI should enhance communication among geographically separated kin. Smart homes equipped with IoT devices are meant to support family members living independently. Drones are envisioned to deliver food and medicine to the elderly living in remote areas, and as such, are deemed vital to Japan’s aging population.

Figure 1.

Society 5.0, according to the Government of Japan.

Figure 1 Long description

The illustration shows a family consisting of six people standing on a platform representing Society 5.0. Various technologies are depicted around them, including IoT, AI, big data and robotics. On the left, a smart car and a person using a tablet are shown, indicating connectivity and smart transportation. Above, a person is interacting with a digital screen, suggesting remote communication or telemedicine. On the right, drones are flying over a field, symbolizing agricultural advancements. Below, a smartphone displays a globe, representing global connectivity. The scene emphasizes the integration of technology in everyday life to improve communication, transportation, healthcare and agriculture.

Source: Cabinet Office (n.d.); see also Gu (2019).

Japanese techno-nationalism shapes spatial representations, its acceptance of surveillance rooted in a remix of postwar automation, late-20th-century ubiquitous computing, biometric “sensing,” and conservative “safe society” discourse. In the Report on the 5th Science and Technology Basic Plan, too, security is a central concern, and it is evoked in ways that establish a sense of identity:

This techno-optimist nationalism builds on Japan’s specific “typographic imagination” (Shockey Reference Shockey2020), which activates past media futures as nostalgic imaginaries (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022). Clinging to print imaginaries replicates fixed, monomodal grammar and obstructs emerging critical literacies. Within such imaginaries, the print medium is deemed neutral, giving ideation visible form without altering content. Dataveillance, a complex process of writing, is then misconstrued as neutral transcription. That is, the translation process is made invisible.

Japan’s affinity to neutral transcription functions not simply as an aesthetic nostalgia but as an ideology for automating meaning-making. Society 5.0 transduces cultural interpretation into algorithmic inference (Kinoshita and Mizuno Reference Kinoshita, Mizuno, García Márquez and Lev2017). It allows the techno-nationalist state to reimagine itself as a noble steward of data, protecting citizens with precognitive governance rather than soliciting their voluntary consent. For example, “society must also be capable of anticipating potential needs and […] enabling anyone to be a service provider” (Council for Science, Technology and Innovation 2015, 13–4). This preempting of needs is also a prescription of what are needs. Enabling “anyone” to be a service provider can be unrequested by many citizens. From this paradoxical fusion of safety, security, and control stems the outsourcing of semiotic labor to machines, making them de facto surrogates of sovereign authority (Mantello Reference Mantello2016). The status quo hierarchies of print capitalism are reinscribed into numerical code. Consequently, the automation of meaning becomes an epistemic and moral project with promises of stability through control, readability through digital transduction.

The diversification of writing through new media could foster inclusivity and extend kinship toward what Rifkin (Reference Rifkin2009; see Lacković and Olteanu Reference Lacković, Olteanu and Campbell2024, 54) calls biosphere consciousness, or, in the pragmatist tradition, world consensus (Wirth Reference Wirth1948). New technologies might render the nation imaginary obsolete, yet they often fuel ethnic and political isolationism (Noble Reference Noble2018; Mihelj and Jiménez-Martínez Reference Mihelj and Jiménez-Martínez2020; Wodak Reference Wodak2020, 239). Instrumentalized for radicalization, social media echo chambers intensify isolationism (Stjernfelt and Lauritzen Reference Stjernfelt and Lauritzen2020, 263). While digital technologies could cultivate global empathy, techno-nationalism perpetuates the modern educational program equating learning with language acquisition and reasoning with print literacy. Monolingual policies obstruct critical digital literacies, like feudalism once opposed mass literacy. Reducing literacy to fixed grammatical convention serves ideological conservatism (Kress Reference Kress2010, 7).

For the preservation of nationalism, the database and the dividual must retain the ethos of the print-imagined community, a future-past of kinships. Japan’s MY NUMBER and digital hanko exemplify this translation of print and premodern administrative practices into digital form. MY NUMBER is a digital identifier replicating the typographic imagination: identity is an alphanumeric code. Initially introduced as a mechanism to streamline tax revenue, its function creep now expands to banking, health care, driver licenses, and possibly future smartphone registration. Beyond being plagued with public scandals involving government mismanagement and identity theft, MY NUMBER offers citizens little practical benefits while granting the state traceable, simplified identities. The hanko, on the other hand, is a personal stamp predating mass literacy. As a holdover of print capitalism, it remains central to linking a citizen’s identity to institutional bureaucracy. The turn toward remote work during the global pandemic, however, fast-forwarded wider acceptance of its digital incarnation. Such transcriptions preserve the imagined borders of the nation.

This digitization of bureaucracy exemplifies priming (Massumi Reference Massumi2015), a longing for a memorial past reimagined through the futuristic prism of the database. In Japan, digital governance is a palimpsest of the past, not a rupture from it: the form changes, but the epistemology endures. By transposing the print-era logic of inscription into code, the nation-state maintains its symbolic authority as the custodian of social harmony, security, and prosperity. The database thus inherits the moral function once held by the family register (koseki, 戸籍) to delineate belonging, ancestry, and legitimacy (Chen (Lara) Reference Chen (Lara), Chapman and Krogness2014). The koseki preserves nationalism by admitting only Japanese citizens while excluding immigrants, even those married into Japanese families or holding permanent residency (Liu-Farrer Reference Liu-Farrer2024). In doing so, Japan bureaucratizes ethnicity and race, safeguarding a mythic homogeneity that binds national identity to biological lineage. Where the koseki was static, the database is dynamic and constantly updating, cross-referencing, and predicting, thus creating the illusion of progress while deepening the reach of state-driven surveillance and its underpinning national imaginary. A digital archive of identity results as a clearly defined sign system, which feels both familiar and futuristic, suturing national continuity to algorithmic computation.

A community as a collection of data is now ubiquitous, the database both constituting and representing Japanese society. Kinship technologies (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022, 449–50) reinforce Japanese nation imaginaries, obstructing empathy beyond already imagined boundaries. The database, misconceived as neutral and objective, sets the grammar of imagination. Thus, dataveillance in Japan functions as a conservative force to maintain the nation-state. It transforms the literary marketplaces of print into AI behavioral futures markets. Yet while once literary markets enabled publics to exercise corresponding literacies, behavioral futures marketplaces do not train publics into reading them. In the latter, unintelligible code determines possible futures, and citizens’ needs are preempted. As Donzelli (Reference Donzelli2025a, 314) observes, to maximize the circulation of discursive protocols, contemporary capitalism requires “the assumption of seamless translatability, complete interoperability of coding systems, and all-encompassing transduction.”

Ubiquitous surveillance is intrinsic to digital platforms, transforming social relations into “surveillant exchanges whose coordination must be technologically mediated and therefore made exploitable as data” (Murakami Wood and Monahan Reference Murakami Wood and Monahan2019, 1). The nation was a proto-platform in the history of surveillance, enabling the Taylorist quantification of labor. With data replacing oil as resource (Zuboff Reference Zuboff2019) and media platforms replacing broadcasting (Rifkin Reference Rifkin2011), digital infrastructures do not merely record activity. They prescribe it, shaping affective and behavioral norms under the guise of optimization. This transition from state registry to predictive platform marks a deepening of submission to mandatory opt-in strategies and data extraction. True to Deleuze (Reference Deleuze1992), the social contract is rewritten in algorithmic terms, as signatures are replaced with passwords and citizenship succumbs to a process of continuous data verification through the metrics of compliance, convenience, and trust in automated systems.

Data capitalism fulfills print capitalism’s requirement to imagine closed communities, obstructing the globalization of kinship into a biosphere consciousness. In Japan, nationalism fuels governance and control through technologies predicted on data collection, resisting social change that technology could otherwise enable. It seeks to preserve power by translating an imagined idea of the nation as a homogenous construct into data, defining kinship (see Fig. 1). This requires not just a nostalgia of an imagined past lack of empathy for diversity but a deficit of critical digital literacies, the capacity to understand and act within digital platform societies (Beck et al. Reference Beck, Goin, Ho, Parks and Rowe2021; Reilly Reference Reilly2021).

Justice and literacy in Society 5.0

Society 5.0 exposes the widening gap between legal discourse and practice. It displays a techno-solutionist whitening of social complexity (Wright Reference Wright2024), transnational hurdles to legal and technological interoperability (Miyashita Reference Miyashita2020), and ethical dangers of machines surrogating sovereign authority (Wright Reference Wright2019). It envisions a future where AI is a preordained instrument of governance, production, consumption, and national security, tasked with restoring economic growth and stabilizing a shrinking workforce. Japan’s AI Guidelines (METI 2024), a legally nonbinding document, offer voluntary rules for developers, providers, and other AI users. The rationale of such soft law is to stimulate business by avoiding regulation. A similar logic shapes the Guidebook for Utilization of Camera Images (MIC 2022), which expresses the hope “that businesses can effectively utilize the guidebook to make concrete advances in the utilization of camera images and data.” This endorses techno-nationalism and voluntary self-regulation, favoring data capitalism, analogous to how the printing press’s Enlightenment potential was folded into print capitalism. In data capitalism, citizen and consumer converge.

A prime example is how Japanese retail businesses are lending themselves to an ever-widening surveillant infrastructure that treats perception as a metric of risk and attention as a valuable resource (Humphrey Reference Humphrey2022; Mantello and Olteanu Reference Mantello and Olteanu2026). This infrastructural visibility sets the grammar for a new genus reading and writing driven by the post-optical image. By post-optical image, we refer to an automated visual representation designed not to be seen but to calculate, sort, and infer. Whereas previous CCTV security systems captured images intended for human eyes, either as deterrent or post hoc documentation, AI perception machines reorganize the terms through which interest, engagement, and identification have operational value (Zylinska Reference Zylinska2023).

Security vendors in Japan market their AI security products, by positioning them as conduits of Society 5.0’s mandate to automate efficiency and security for a digitally prosperous nation. Companies such as EarthEyes, ELSYS, and VAAK promulgate their wares as behavioral predictors capable of identifying and monetizing the return customer while simultaneously inferring and preventing those harboring criminal intentions. As customers enter a store, the AI system will read their basic biometric data to determine if they are a repeat customer. If so, it will call up on a manager or store attendant’s screen that person’s store history data, such as average visit time, favorite past purchases, and spending amounts. The AI will track the customer throughout their store visit, reporting behavioral anomalies such as unusual dwell time in a store blind spot and excessive eye and head movements away from a product they are standing in front of or holding in their hand. Based on a constellation of data points (many of which are not disclosed to the public), the customer is assigned a “suspicion” risk score. If the customer hides a product in their clothing or personal bag, the AI will automatically flash red, sending a quiet audio alert to the store attendant. The same systems have also been put in place in self-checkout counters and a growing number of fully automated stores.

Mainstream media echo this techno-optimism, mediating between government and business. Nikkei reports with optimism on the installation of Sony AI cameras in convenience stores (Sato Reference Sato2024), the use of AI to accelerate production to reduce staff (Hara Reference Hara2023), with automated checkout systems and robots remotely controlled by Filipino tele-operators (Beltran Reference Beltran2025). Thus, AI functions as a kinship technology. A purveyor of Japanese nationalism as technology is promoted as a silver bullet to societal ills, such as labor shortages and errant immigrants, perpetuating the long-standing myth of the “bad foreigner” (Yamamoto Reference Yamamoto2010). For instance, mainstream outlets routinely translate anxieties about economic decline and demographic imbalance into moral panics about foreign deviance. Nikkei (Kanauchi Reference Kanauchi2024) warns about “foreigner” criminals allegedly targeting retail businesses, while the Asahi Shimbun (Suzuki and Ito Reference Suzuki and Ito2024) frames a “Vietnamese shoplifting ring” as emblematic of rising lawlessness linked to immigration. Such reporting does not merely indicate isolated incidents; it reproduces an interpretive grid in which the “foreign body” functions as the glitch in an otherwise well-calibrated social machine. AI is promoted as a cure and custodian—an automated corrective that restores social order through algorithmic sapience.

This drive toward technological purification aligns seamlessly with Japan’s evolving doctrine of data sovereignty. Society 5.0 rebrands surveillance as sensibility—a national project to cultivate “trusted” data ecosystems that allegedly reflect Japan’s unique moral and social order. Such rhetoric reframes techno-nationalism as a form of affective governance, where sensing infrastructures are said to embody the nation’s collective empathy. Beneath this language of harmony lies a cartography of exclusion: the codification of belonging through data. In rendering “sensing” synonymous with security, Japan’s digital nationalism converts emotion into infrastructure and subjectivity into a resource to be mined, modeled, and managed. The sovereignty of sense thus becomes the sovereignty of sorting, where to be legible is to be loyal.

Long known as a society reluctant to expand its immigrant population, Japan’s nationalistic rhetoric has recently intensified. In repeated statements, the new prime minister has made her priorities unmistakable, vowing to crack down on illegal overstays with plans to award whistleblowers and to revoke resident permits for foreigners who default on social security or healthcare premiums (Suzuki Reference Suzuki2025), raising fees for visa applications and instituting language requirements for foreigners seeking employment. To implement these measures, a new cabinet position was established, the title of which reveals the Government’s attitude to non-nationals: Minister of Promoting an Orderly and Harmonious Co-existent Society with Foreigners (NHK 2025). This wording does not even try to hide what Bauman (Reference Bauman1995) identified as the main characteristic of discriminatory authoritarianism, namely “making” strangers.

Beyond the obvious disciplinary undertones of the term “orderly” lies the conspicuous absence of adjectives such as inclusive or diverse, which are nevertheless repeated throughout the 5th Basic Plan (Council for Science, Technology and Innovation 2015). Instead, the ethnocentric rhetoric of the new administration flattens tourist/international, student/temporary worker/permanent resident into the single frame of “foreigner.” This population is a priori identified as suspicious, and imagination is permitted to entertain not more than mere “coexistence.” It reveals a political will and agenda to preserve an entrenched us versus them dichotomy and Japan’s defensive sense of homogeneity. By doing so, it creates identities by supposing boundaries.

Society 5.0 models the nation as a digital database. “Data” is the deterritorialized residue of surveillance, and the database fixates the grammar. A technological democracy, as Society 5.0 is praised, assumes that every citizen is digitally literate, competent in the new digital grammar. Yet literacy here is reduced to the replication of old technologies, such as hanko and koseki. Individuals inhabit the kinships of prior infrastructures, now confined within a database, a rigid grammar they do not comprehend. Many are literate functionally but not critically—able to use the hanko without understanding its mechanics. Thus, the nationalist imaginary endures in a new medium; print protocols are transcribed into digital form, as evidenced in the post-COVID trend to digitize the hanko. For the nation-imaginary to persist through digitalization, citizens must remain literate enough to serve the economy but uncritical enough to forestall reimagining community through new media. Whereas print regimes built networks conducive to national imaginaries, data ontologies weave post-optical networks indifferent to historic borders and corporeal boundaries, capable of unsettling conservative and identitarian ideologies, unless tightly regulated.

Japan’s assimilation into computational mega-infrastructures follows risk-based imaginaries of prediction through automated observation (Humphrey Reference Humphrey2022). These techno-solutionist deployments of AI transform the very conception of justice through “a transfer of legality from a spatial and linguistic order to a non-spatial, non-linguistic one based on lines of written code nobody has full access to” (Lassègue Reference Lassègue2022, 145). Predictive algorithms alter the act of writing, separating the legal agent from the patient. The authorship of law is increasingly delegated to machines that do not share the lived experience of those to whom the law applies (Hildebrandt Reference Hildebrandt2018). Unlike writing in print capitalism, software coding does not claim to represent the (oral) languages of human societies (Lassègue Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025, 91).

Lassègue (Reference Lassègue, Cappiello and Carullo2021, 61) identifies “a conflict between two forms of legality: one based on legal texts written in technical but natural languages expressing political sovereignty; the other on unreadable software whose authority derives from collective trust in machines, a trust not yet legally articulated.” Tech companies and the retail sector invest such trust in machines without considering surveillance’s legal implications (Mantello and Olteanu Reference Mantello and Olteanu2026), a stance amplified by major media outlets. Nikkei (Imai Reference Imai2018; Yoshihiro Reference Hara2023; Muto Reference Muto2024; Nikkei Staff Writers 2024) repeatedly celebrates AI surveillance in retail stores as a signpost of modern efficiency, security, and economic prosperity. This enthusiasm rests on the dichotomy of reading and writing, as surveillance technologies are presumed to read objectively without writing individuals into categories of risk. Actually, they shape perception, categorization, and affordances.

This blind faith in machinic reading extends Japan’s postwar fantasy of bureaucratic infallibility into the digital realm. Where once the written statute conferred legitimacy through public deliberation, today code performs governance through opacity. The substitution of legal hermeneutics with computational inference reconfigures citizenship: subjects become data profiles, prejudged by predictive metrics rather than recognized through juridical speech. Within this techno-legal imaginary, the algorithm becomes a silent legislator, authoring decisions that no one claims but everyone obeys. The result is a subtle depoliticization, an aesthetic of order without accountability. Japanese media, celebrating such precision as cultural virtue, frame automation as an extension of harmony rather than a displacement of agency. This erasure of authorship masks the violence of inscription: a system that rewrites legality as logistics, ethics as efficiency. In doing so, it perpetuates the automation of meaning-making, outsourcing interpretation to data systems presumed to know better than the public itself. Justice becomes concerned with perception and time more than fairness (Massumi Reference Massumi2015), and the dream of perfect governance resolves into a feedback loop where code writes the citizen before the citizen can speak.

This machinic legality paradoxically invokes wa (和), the long-idealized principle of social harmony, as its moral alibi. In Japanese political and media discourse, wa has long functioned as both ethical compass and disciplinary logic, a promise of cohesion that demands the effacement of dissent. Within the digital state, wa is reinterpreted as algorithmic equilibrium, a harmony to be achieved through predictive governance rather than mutual understanding. The algorithm, cast as the ultimate neutral arbiter, enacts a depersonalized version of consensus that preempts conflict by preempting interpretation itself. Yet, in transforming harmony into computation, wa is hollowed out: no longer an emergent property of social negotiation but a statistical artifact produced by optimization. This transformation domesticates the political under the guise of balance, rendering inequality and exclusion as technical errors to be corrected rather than injustices to be addressed. It reframes wa as automated conformity, a recalibration of social affect designed to maintain the illusion of unity in an increasingly datafied polity.

The divide between readable and digital law is intensified by behavioral-capture technologies in law enforcement. Beyond the predictive shift from post hoc to a priori approaches to crime (Mantello Reference Mantello2016), AI reshapes the software of observation itself, delegating perception from human social space to machine symbol systems. Because state-of-the-art AI is unintelligible to humans, behavioral harvesting technologies are rendering humans increasingly illiterate (Lassègue Reference Lassègue, Cappiello and Carullo2021, Reference Lassègue2022, Reference Lassègue, Pantsar, Stjenrfelt, Gramelsberger and Olteanu2025). If in print capitalism the fellow citizen was imagined, in data capitalism, they are prescribed in data. Like all media, dataveillance is not semiotically neutral. It frames social and legal action through coded algorithms. To understand these processes is to understand their medium, formed of a code and an interface, the latter designed to obscure rather than reveal the former (Manovich Reference Manovich2013).

A necessary condition for democratic governance is that technology users become “digital citizens” (Lyon Reference Lyon2023, 364), capable of critical engagement. Far from signaling democracy’s surrender, (post)digital literacies mark a critical frontier for research and education (Campbell and Olteanu Reference Campbell and Olteanu2024; Lacković, Olteanu, and Campbell Reference Lacković, Olteanu and Campbell2024). While literacy alone cannot dismantle the discriminatory architectures of data capitalism, it remains indispensable (Beck et al. Reference Beck, Goin, Ho, Parks and Rowe2021; Reilly Reference Reilly2021).

Conclusion: democracy versus unintelligibility

The passage from print capitalism to data capitalism requires transcribing the nation into its digital double, a future reimagined as a memory by the translation machine (Donzelli Reference Donzelli2025a, Reference Donzelli2025b). In Japan, where government and industry jointly endorse datafication, the Rule of Law dissolves into code. As such, translation itself undergoes a process of reduction to transcription. Legislation, once “a process which opens up a future that remains to be written collectively” (Lassègue Reference Lassègue, Cappiello and Carullo2021, 71), now risks being written by technological infrastructures rather than by the communities bound by them. For democracy to hold, law must remain legible to its citizens; writing it must remain a shared, ethical undertaking (Hildebrandt Reference Hildebrandt2018).

Society 5.0 employs emerging technologies to restore romanticized pasts (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022, 451), projecting a homogenous nationhood into the circuitry of the digital age. During technological acceleration, “Japan” becomes a project of preservation: the digital nation as a spectral continuation of the imagined one. What once had to remain unknown to be imagined is now supposedly observed, as fixed in data, which feeds the capitalist translation machine. In this techno-nationalist imaginary, “the imagining of a robotized society of the future is informed by the reimagining of past social structures and relationships” (Robertson Reference Robertson, Bruun, Wahlberg, Douglas-Jones, Hasse, Hoeyer, Kristensen and Winthreik2022, 451).

Against this closure of imagination, we propose a turning outward: to seize the creative affordances of technology as a means of unbinding these inherited imaginaries. This demands a digital criticality, an education in perception and interpretation, that can ask what it means to write a future rather than merely to read one inscribed by machines. Particularly, this implies a criticism of the capitalist translation machine, which liberates imaginaries from perceiving boundaries.

We recognize the democratic peril in AI’s displacement of intelligible language with the opacity of code. Yet ours is not a call for techno-skepticism or a retreat into humanist nostalgia. We even salute the progressive and inclusive language of Society 5.0 (e.g. Council for Science, Technology and Innovation 2015), starting with the very important accessibility to documentation, but recognize in its assumptions the inertia of capitalism as a translation machine. We glimpse the persistence of familiar power struggles, civil freedoms, and responsibilities rewritten in languages that many can no longer read. The task, then, is not to resist technology but to reimagine literacy: to make the codes of our collective future legible, so as to subject them to translating through social imaginaries.