Discourse Analysis for Knowledge Organization

In its Francophone context, Discourse Analysis (DA) is a theoretical product of the philosophical, scientific, political, and social movements that characterized France in the late 1960s, with Michel Pêcheux as one of its leading figures. Along with a group of collaborators, Pêcheux developed a theory of the materiality of discourse, building on the philosophical work of Louis Althusser. He conceptualized discourse as a distinct form of materiality-historical and linguistic-directly intertwined with ideological materiality, proposing a "semantics of discourse." The theories of Pêcheux and Althusser marked a shift from

Instead, ideology is seen as a productive force with specific operations, and language is theorized to contain ideological meaning inherently. For Pêcheux, discourse analysis is a foundational method for examining the production of meaning in discourse, recognizing that these texts-even those that seem neutral or purely rational-are profoundly influenced by history and ideology. This shift established the groundwork for a new field within linguistics based on materialist principles. Rather than viewing ideology as just a misrepresentation of the economy, Pêcheux (2014) argued that it operates independently and has a concrete mechanism that requires a sophisticated analytical approach. This materialist semantics treated discourse as both an event (momentary utterances or statements) and a structure (a set of rules or norms governed by ideological forces).

Pêcheux’s discourse analysis approach distinguished itself from structural linguistics, notably rejecting Ferdinand de Saussure’s notion of language as self-contained. Discourse analysis, especially as influenced by Pêcheux's framework, scrutinizes how acts within ideological structures, positioning discourse as something more than a linguistic phenomenon-it is inherently ideological and historical.

DA’s connection to Marxism and linguistics enhances its interdisciplinary nature in the study of meaning formation. From the perspective of linguistic sciences, discourse analysis offers an alternative approach to understanding semantic phenomena, employing Pêcheux’s materialist framework to examine language and its political and ideological impacts. Viewed through this lens, DA evolves, enriched by continuous refinements, adjustments, and reassertions, creating a framework where language, meaning, and ideology are inseparable. Therefore, when considering controlled vocabularies, taxonomies, or ontologies, we perceive perspectives on knowledge. However, since knowledge is interwoven with its political, ideological, social, and semantic dimensions, it establishes a limited view of meaning.

This perspective transformed the earlier theoretical fields, providing what some call a radical departure and academic revitalization of language and ideology studies. Texts were seen not just as isolated linguistic structures but as embedded in vast ideological networks regarding social, political, or historic forces. For Pêcheux’s discourse, the concept of “event” anchors the theory since texts originate in events fraught with power struggles and ideological tensions, reflecting broader societal forces. This distinction allows DA to go beyond mere discourse description to uncover its systemic imbalances and its role in reproducing social and ideological dominance. Pêcheux’s radical transformation of DA goes beyond a critique of structural linguistics; it is an inherently interactive, dynamically positioned theory of social and ideological struggle.

DA does not address a homogeneous linguistic system but one prompted by events-signifiers that circulate within ideological formations, constantly shaped by the ever-evolving discursive environment. Such an understanding clearly establishes the foundation for integrating DA into Knowledge Organization (KO) by recognizing that texts, terms, and concepts associated with knowledge are rooted in their ideological, social, and historical context. Discourse analysis and its materialist approach then deliberately critique knowledge systems by identifying not only the conceptual but also the ideological and social mechanisms underpinning them.

By employing metaphors and imagery, such as a “network of meanings,” DA theorists like Barros (2015) and others illustrate the subtle tensions between a “structured whole” and its contradictions. A discourse operates like a complex, porous network that allows for non-representations and deconstructed meanings, which do not adhere to linear paths. These insights are vital in both KO and domain analysis, enabling researchers to move beyond merely conceptual frameworks and toward an understanding of knowledge organization influenced by social and political forces.

Viewing KO as discourse implies that KO structures themselves are influenced by ideological discursive fields-not politically neutral but instead layered in meanings determined by who is disseminating the knowledge and why. For instance, in the construction of knowledge ontologies in KO, DA challenges the linguistic assumptions by injecting a social, ideological, and historical critique of the terms used for knowledge representation. Traditional approaches to ontology construction that treat concepts and terms as fixed entities run counter to an understanding of them as malleable and historically grounded.

DA's contribution in this context is crucial, particularly in situations where ideologies are likely to form and affect these constructed domains, seen, for example, through the political lens of an autocratic or democratic state, directly influencing semantic relations within a hierarchical ontology. Furthermore, DA’s analysis identifies the ideological positions within texts and the discursive formations that allow knowledge representation.

The application of DA to KO is an essential epistemological and methodological shift that complements traditional domain analysis. Discourse analysis delivers interpretations based on a broader scale that challenges conceptual terms and offers new theoretical concepts, such as ideological and social formations, which direct the overall configuration of the domain under study. It allows for a critical examination of knowledge organization systems, providing a deconstruction of their internal structures that highlights power dynamics, subordination, and ideological imbalances present in knowledge representation.

Lastly, discourse analysis provides a framework for KO research that does not take the text, concept, or ontology at face value. Discourse analysis requires a “critical distance,” as theorized by Pêcheux, where the analyst recognizes the duality of being both a subject of study and a participant in the interpretative process. Through an active role in analyzing discursive formations, the researcher recognizes that discourse systems cannot be objectively separated from their social and ideological contexts. When applied to KO, discourse analysis offers groundbreaking potential not merely in technical theory but in human-centered reflections of power, knowledge, and ideology embedded in the systems that govern our access to information.

Viewed through this lens, discourse moves beyond simply representing the world to actively producing meaning within it, so that a text serves as an object in which discourse and ideology intersect, defined by its own historicity and the recognition that such intersections shape meaning at once collectively and individually. This conceptual stance is often conveyed through the metaphor of discourse as a network or fabric, replete with “holes” through which new or alternative meanings can pass, emphasizing that words do not hold meaning in isolation but rather in the ideological, social, and historical contexts in which they appear.

The notion of a discursive formation thus designates a space larger than the text itself, demarcating the range of what can be said and implying that the significance of words is never intrinsic but is determined by shifting configurations of ideology, power, and social positioning. Such formations are fluid, heterogeneous, and continuously reconfigured, introducing internal contradictions and enabling fresh avenues for interpretation.

When applied to Knowledge Organization (KO), discourse analysis frames KO as a form of discourse, one that is molded by significant imperatives such as technological advancement and the imperative to systematically organize scientific knowledge. Adopting a discourse analysis perspective on KO extends beyond traditional conceptual or terminological challenges; it actively engages with the ideological, social, and political dimensions that underlie the frameworks we create for the classification and retrieval of information. For example, when utilizing concept theory and terminology, the development of a hierarchical ontology in a conventional manner effectively aligns terms with their corresponding concepts. Conversely, an approach informed by discourse analysis would integrate the influences of discursive formations and ideologies, acknowledging that meaning cannot be detached from its social, historical, and political contexts. Constructing an ontology within an authoritarian regime, for instance, would significantly shape the representation of democracy and dictate how particular theories or perspectives are presented. Knowledge, therefore, cannot be considered neutral, as language itself is not neutral.

Accordingly, discourse analysis contributes layers of social and historical context to semantic relationships, thus denaturalizing and fragmenting the processes through which KO systems are constructed, and encouraging richer theoretical, epistemological, and methodological discussions.

Practically speaking, discourse analysis requires defining a specific textual or institutional object for investigation, assembling a corpus that the researcher then scrutinizes through a critical lens-an “observatory”-because no analyst is neutral. ^{Pêcheux and Fuchs, in their 1975} discussion of automatic discourse analysis, distinguish linguistic surface, discursive object, and discursive process, emphasizing that textual meaning emerges from both conscious and unconscious “forgettings” (No. 1 and No. 2). These illusions reflect the dual ways in which ideology and enunciation obscure the fact that meaning is a reactivation of historically constituted, socially embedded sense, rather than the original creation of an autonomous subject. Barros (among others) has demonstrated the utility of this approach for interrogating a wide range of KO-related discourses, from archival ethics to indexing policies and conceptions of traditional knowledge, thus reinforcing how discourse analysis reveals historical, ideological, and social threads in seemingly technical or scientific content.

Ultimately, Pêcheux’s concept of a “non-subjective theory of the subject” highlights the interconnected nature of ideology, institutions, and texts, revealing that meaning is influenced as significantly by historical and material conditions as by linguistic frameworks. Consequently, discourse analysis can serve as a potent instrument in knowledge organization (KO), fostering a profound understanding of how socio-ideological factors shape domain analyses, classification systems, and various knowledge structures. Instead of perceiving language or metadata as merely transparent conduits of meaning, discourse analysis emphasizes the fundamental discursive processes that facilitate interpretation. By prioritizing history, ideology, and social context, it enhances the analytical capacity of KO, stimulating considerations on the optimal design of knowledge organization systems that recognize and navigate the intricate interplay of language, power, and context-specific human comprehension.

Discourse Analysis as an Approach to Knowledge Organization

KO occupies a position within information science, embodying a dual objective: to refine the systematic and replicable organization of knowledge domains while simultaneously recognizing the broader contexts-cultural, ideological, and historical-in which classification practices are situated. As Hjørland (2016) observes, domain analysis has emerged as a fundamental methodology for comprehending the formation and evolution of knowledge structures, providing a systematic framework that identifies essential concepts, stakeholders, and publications pertinent to a specific intellectual or professional domain. However, a tension remains within Knowledge Organization: the discipline’s focus on replicable techniques, including controlled vocabularies, classification schemes, and indexing protocols, at times risks neglecting the ideological and discursive frameworks that fundamentally influence these very techniques.

Discourse analysis (DA), mainly as developed within the French tradition by Michel Pêcheux and his collaborators, powerfully addresses this oversight by revealing how no knowledge system is ever free of social and historical influences, and by showing that language serves not merely as a transparent medium for communicating meaning but also as a material site where ideology is enacted and contested.

From its inception, discourse analysis in France set out to rethink the role of language in shaping consciousness, social relations, and institutional structures (Pêcheux, 1969). In contrast to structural linguistics, which often treated language as a self-contained system governed by stable syntactic and semantic rules, Pêcheux insisted that discourse should be seen in its “double materiality,” combining linguistic form with ideological and historical content (Pêcheux, 1990). By extension, any field that organizes knowledge-such as library or information science-cannot escape the fact that its domain-defining vocabularies and ontologies encode worldviews. Hence, KO systems are never purely functional or technical. Instead, they are imbued with social power, cultural norms, and rhetorical conventions that subtly guide how specific ideas gain traction, while others might be marginalized or excluded.

One of the core contributions of Pêcheux’s DA lies in the concept of “discursive formation.” This concept highlights how discrete texts or utterances belong to broader networks of meaning, shaped by ideological forces and recurrent patterns of enunciation (^{Pêcheux & Fuchs, 1975}). In Knowledge Organization, we might see these networks manifest in classification hierarchies: certain categories, subject headings, or terminologies are sanctioned because they align with the prevailing assumptions of an epistemic community or a social group. Likewise, certain narratives, conceptual frameworks, or user perspectives may be omitted or remain underrepresented because they stand outside the dominant “discursive formation” that underlies a classification scheme (Barros, 2015). In this light, domain analysis ceases to be merely a matter of collating the key references, journals, and concepts in a particular area. Instead, it involves examining the social and political processes by which a group’s knowledge claims become standardized. This observation resonates with critical studies in library and information science emphasizing that indexing and retrieval practices can embed hegemonic viewpoints (^{Frohmann, 1994}).

Consider, for instance, how ^{Orlandi (2007}) describes meaning as the continuous interplay of overlapping or “transferring” expressions. According to this perspective, we should not expect to find stable, context-free definitions for specialized concepts. Rather, each term or concept emerges at the intersection of language, ideology, and history, implying that the definitions themselves are variable depending on the user’s or institution’s vantage point (Orlandi, 2009). This insight is especially crucial in Knowledge Organization, where building a shared vocabulary for indexing is often seen as the foundation for efficient retrieval. Even the move toward standardization-like adopting a single authoritative word or phrase to capture a concept-reflects an ideological choice, one that might alienate voices or viewpoints that do not adhere to the dominant discourse.

When we talk about DA in the context of KO, methodological reflexivity is one of the most important practical consequences. Domain analysis has long been presented as a sequence of steps: defining the domain, identifying the key actors, analyzing the literature, discerning the conceptual frameworks in use, and so forth (Hjørland, 2016). A Pêcheux-inspired discourse analytic approach, however, raises deeper questions at each step. First, who decides what a domain is? This question points us to the ideological impetus behind domain boundaries. Are these boundaries historically contingent, formed by powerful institutions, or do they arise through a consensus-based process involving multiple actors and perspectives? Second, which sources are included or excluded in analyzing the domain? Inclusion and exclusion criteria are never neutral: they reflect assumptions about validity, relevance, and legitimacy, all of which tie back to historically emergent ideological formations. Third, which theoretical frameworks are privileged? Some theories-and, by extension, some conceptual vocabularies-garner acceptance precisely because they align with hegemonic thinking, while competing paradigms may be sidelined.

Such inquiries point to the significance of discursive communities, a concept that merges domain analysis with discourse analysis. In addition to identifying an “epistemic community” through references, research areas, or specialized journals, a discourse analytic lens suggests that communities are forged around rhetorical conventions, shared ideological standpoints, and a cluster of repeated textual moves that collectively define who “belongs” to the community and who does not (^{Barros, 2017}). On this view, membership in a domain cannot be reduced to mastery of technical knowledge; it also entails adopting a particular discursive identity-speaking in a way that resonates with community norms.

Conversely, individuals who challenge or subvert these established norms may find themselves marginalized or excluded from prevailing discourses. In the context of Knowledge Organization, the imperative lies in developing classification systems and indexing vocabularies that not only reflect dominant voices but also recognize, or at minimum, accommodate alternative discursive formations.

Michel Pêcheux’s methodology further illuminates the interpretive facet of Knowledge Organization (KO). As ^{Orlandi (2007}, p. 44) posits, meaning manifests through a dynamic interrelationship: “the signified is always a word, an expression, or a proposition for another word, another expression, or proposition; and it is in this relationship-this overlapping, this transfer (metaphor)-that signifier elements come to confront one another.” When we apply this principle to domain analysis, it becomes evident that the conceptual architecture of each domain is not immutable. Instead, it is perpetually in flux, shaped by various historical and social processes converging within discursive enactments. A classification system that regards concepts as neatly delineated or as stable “mental objects” risks overlooking these dynamic transitions of meaning. Consequently, domain analysis (DA) necessitates the examination of discursive shifts, revealing how specific categories or conceptual labels evolve, respond to external pressures, or become subject to ideological tensions.

An especially telling example arises when Knowledge Organization systems have to represent contested political or ethical concepts. Imagine constructing an ontology of governance terms under an autocratic regime, where official discourse might define “democracy” in highly constrained ways or even limit usage of the term altogether, we cited this example before, but let us work with it more.

A conventional domain analysis might take such definitions and descriptors at face value, particularly if it seeks to remain “neutral.” However, a DA perspective would illuminate that no truly neutral stance is possible: adopting the state-sanctioned definitions effectively endorses the ideological framework of the autocracy (Barros, 2015). Conversely, to represent diverse or dissenting definitions might incur political risk or violate official policies-again illustrating how classification decisions are far from purely technical. This tension underscores how discourse analysis reveals the ethical stakes of knowledge organization, obliging practitioners to decide how to navigate conflicts between institutional norms and broader commitments to intellectual freedom or social justice (Orlandi, 2009).

Such conflicts connect to a theme introduced by Pêcheux regarding “forgetting,” particularly the illusions that cause subjects to imagine themselves as the source of meaning, rather than as reactivating historically sedimented discourses (^{Pêcheux & Fuchs, 1975, pp. 180-182}). In KO, an analogous forgetting occurs whenever professionals fail to recognize the historical and ideological baggage that underwrites their classification frameworks. For instance, a domain-analytic approach might emphasize robust referencing or data reliability but may not question the assumptions that have shaped the domain’s recognized canons. Discourse analysis thus invites a more profound reflexivity: classification becomes not just an act of enumerating categories but also a praxis that can either reify or challenge existing power structures.

Moreover, the methodological implications of DA for KO can be far-reaching. While domain analysis has frequently been characterized by systematic literature reviews, citation analyses, or the ethnographic study of professional communities, a discourse analytic angle encourages additional methods such as close textual reading, rhetorical analysis, and genealogical tracing of how specific terms have shifted across time and usage contexts. These tools can identify discursive rupture or ideological dispute points in ways that standard domain-analytic protocols might overlook. They can also highlight how an organization’s internal documents-cataloging manuals, indexing guidelines, or usage policies-reveal latent ideologies shaping knowledge. For example, indexing guidelines might systematically marginalize specific user populations or cultural expressions by labeling them with terms that carry a derogatory historical charge. A discourse analytic approach would examine how these labeling practices came about, which authorities codified, and how they continue influencing users’ navigation of information resources.

Seen in this light, discourse analysis not only adds nuance to KO but may also instigate more democratic and inclusive practices of knowledge representation. By demonstrating how ideologies become embedded in language, DA highlights the possibility of alternative discursive formations that can disrupt or expand the field’s conceptual terrain. Far from merely diagnosing ideological bias, it provides conceptual tools-such as discursive formation, the tension between event and structure, the illusions of forgetting-that can help practitioners critique existing structures and propose more contextually sensitive or participatory classification approaches. These new practices could involve user communities in designing vocabularies or explore multi-vocal strategies that allow for contradictory, contested, and evolving definitions of concepts within the same classification scheme (^{Barros, 2017}).

The implications of such reflexivity are, in fact, deeply aligned with the fundamental motivations of domain analysis. While Hjørland (2016) initially proposed domain analysis as a systematic approach for discovering how knowledge fields are cognitively and socially organized, discourse analysis amplifies this perspective, reminding us that political, historical, and ideological factors shape each domain’s organization. Recognizing these factors is key to capturing the domain’s full complexity and ensuring that the classification system we build is ethically and epistemically sound. This alignment between DA and domain analysis thus marks a significant step forward in theoretical sophistication, bridging the seemingly pragmatic goals of KO-reliability, efficiency, replicability-with the more profound recognition that classification systems are interpretive frameworks steeped in power relations.

In conclusion, integrating discourse analysis into Knowledge Organization reorients the field, compelling it to confront the ideological, rhetorical, and historical dimensions that underwrite any act of classification. Michel Pêcheux’s vision of discourse as a material practice underscores that KO cannot treat language merely as a vehicle for stable concepts; meaning is always dynamic, context-specific, and entwined with power structures. DA enriches and challenges the domain-analytic tradition by encouraging critical reflection on how domain boundaries are drawn, how terms acquire their significance, and how certain voices get amplified while others are omitted. This approach unveils the often-invisible discursive formations that guide classification and opens pathways for more inclusive, context-responsive, and ethically informed knowledge organization. Adopting a discourse analytic lens may help KO practitioners move beyond technical efficiency alone, spurring them to design and maintain information systems that address users’ diverse needs and conscientiously acknowledge the social contexts in which knowledge is produced, contested, and shared.

Applications of Discourse Analysis in Knowledge Organization Systems: an example “democracy.”

Based on our discussions, we constructed a taxonomy to capture how “democracy” is conceptualized, practiced, and contested in an autocratic state. It adopts a straightforward hierarchical structure, reflecting official narratives and alternative or critical viewpoints that challenge the dominant discourse. The primary aim is to offer a coherent system for classifying materials-texts, documents, and digital resources-that discuss the complexities of democracy under authoritarian governance. By dissecting the notion of democracy into logical categories and subcategories, the taxonomy highlights the ideological tensions, social dynamics, and external influences shaping its meaning and application.

At the apex lies the concept of “Democracy in an Autocratic State,” an umbrella category encompassing everything from official political rhetoric to clandestine activism. Subsumed under this main heading are four principal branches, each illustrating a unique dimension of how democracy is claimed, exercised, or restricted. The first major branch, “Officially Sanctioned Democracy,” focuses on the state’s presentation of democracy through recognized institutions and authorized practices. Within this category, one finds “State-Defined Principles,” which serve as the ideological anchors or slogans legitimizing the regime’s curated form of democracy (e.g., “guided democracy,” “cultural suitability”). Another sub-level is “State Institutions,” referring to the advisory parliaments, unity councils, or ministries tasked with creating a facade of representation, often lacking absolute autonomy or checks on power. “Permitted Civil Participation” covers the controlled forms of citizen engagement-like supervised local elections or authorized civil groups-that simulate popular consent while maintaining autocratic control.

A second major branch, “Restricted or Criminalized Discourses,” details how critical or dissenting perspectives on democracy are suppressed. “Opposition Views” addresses overt demands for multiparty elections or transparent governance, which the regime typically deems " subversive. “Underground Mobilizations describes covert activism, from local clandestine networks to encrypted social-media channels, seeking to evade state surveillance. Critical Media captures the role of independent journalists or bloggers who face censorship for publishing unauthorized accounts or critiques.

The third branch, “External or Global Conceptions,” highlights international or exile-based perspectives that either inspire local reformers or threaten the regime’s legitimacy. This includes “International Organizations,” with standards for free and fair elections; “Foreign Democratic Models,” which serve as reference points for those seeking alternatives; and “Exile-Based Narratives,” representing dissident communities abroad who maintain contact with domestic audiences.

Finally, “Mechanisms of Control & Adaptation” encompasses the apparatus by which the state enforces, revises, and protects its authoritarian approach to democracy. Under “Propaganda Apparatus,” government bodies and officially owned media outlets frame the regime’s narrative, asserting that centralized authority best serves national interests. “Censorship & Surveillance” deals with active monitoring, from internet firewalls to security crackdowns, ensuring that critical voices remain stifled. Meanwhile, “Legal-Institutional Enforcement” pinpoints the concrete legislation and judicial mechanisms-anti-subversion laws, loyalty oaths, security courts-that criminalize alternative discourses.

This straightforward, enumerative classification offers a valuable framework for librarians, researchers, or archivists who aim to index or organize democracy-related content in authoritarian contexts. By incorporating both government-sanctioned categories and marginalized viewpoints, the taxonomy illustrates that "democracy" is not monolithic or ideologically neutral but is continuously shaped and reshaped by competing forces. Moreover, it highlights how official doctrines, domestic opposition, and global influences intersect within a broader sociopolitical tapestry. Ultimately, taxonomy is a comprehensive tool for systematically categorizing and analyzing evolution.

1.1 OFFICIALLY SANCTIONED DEMOCRACY

Definition:

The version of democracy formally endorsed by the ruling government, emphasizing controlled participation, top-down authority, and rhetorical alignment with regime interests.

1.2 RESTRICTED OR CRIMINALIZED DISCOURSES

Definition:

Alternative or oppositional understandings of democracy that conflict with the official narrative and are suppressed, censored, or delegitimized by the autocratic regime.

1.3 EXTERNAL OR GLOBAL CONCEPTIONS

Definition:

Perspectives on democracy that originate outside or beyond the regime’s control, posing potential challenges to the official narrative.

1.4 MECHANISMS OF CONTROL & ADAPTATION

Definition:

The regime uses structures, processes, and strategies to control, shape, or neutralize any discourse on democracy that deviates from the official line.