Meaning Dissolution: A Structural Model of Coherence Failure Under Distributed Exposure
A formal articulation of how relational context loss prevents the stabilization of coherent meaning despite informational accuracy and accessibility
This paper introduces the Meaning Dissolution Model, a structural account of how coherent meaning fails under conditions of distributed exposure. Existing frameworks in information and epistemic theory largely explain breakdowns in understanding through falsity, bias, or fragmentation. This model addresses a different condition: one in which information remains accurate, accessible, and intact, yet cannot stabilize into coherent meaning because the relational context required for that stability is either stripped in transmission or not carried with the information. Positioned within the broader Psychological Architecture framework, the model defines a distinct mechanism of coherence failure that operates across systems, documents, and extracted components, and establishes a foundation for further inquiry across disciplines concerned with information, interpretation, and knowledge formation.
Meaning Dissolution (March, 2026)
DOI: https://doi.org/10.13140/RG.2.2.34484.10886
Download PDF
Abstract
Existing frameworks for understanding epistemic failure — including misinformation theory, filter bubble theory, and epistemic fragmentation — presuppose conditions of falsity, selective exposure, or competing truth claims. None accounts for the failure of coherent meaning formation in conditions where information remains accurate, accessible, and intact. This paper introduces Meaning Dissolution, a structural model describing the process by which information cannot stabilize into coherent meaning because the relational context required for that stability is either stripped in transmission or not carried with the information under conditions of distributed exposure. The model is demonstrated across two domains. In AI-distributed information environments, accurate content is recombined and redistributed at scale without the relational structure that organized it, producing coherence failure that existing frameworks do not capture. In legal reasoning and case law extraction, the identical mechanism operates independent of technology, where extracted holdings and principles retain factual accuracy but lose the relational context required for valid application. Together these cases establish Meaning Dissolution as a scale-invariant structural process operating wherever relational context fails to persist under distributed exposure. The model occupies a distinct explanatory space within information theory and epistemic theory, with direct implications for how epistemic failure is defined, evaluated, and addressed in distributed information environments.
1. Introduction: The Explanatory Gap
The dominant conversation about artificial intelligence and epistemic life is organized around a specific class of problems: inaccuracy, bias, manipulation, and the displacement of human judgment. These are real problems. But they share a common assumption — that epistemic failure requires something to have gone wrong with the information itself. The frameworks built to explain epistemic failure in the age of AI are, without exception, frameworks about bad information.
What they do not account for is a different and increasingly prevalent condition: the failure of meaning that occurs not because information is wrong, but because the relational structure that made it coherent is no longer present. Information circulates accurately. It is accessible, retrievable, and in many cases precisely correct. And yet it does not cohere. It cannot be reliably assembled into stable meaning because the context that organized it — the sequence, the sourcing, the qualifying conditions, the argumentative relationships — has been stripped in transmission or was never carried with it in the first place. The information persists. The structure that made it meaningful does not.
This is not a problem of misinformation. Misinformation theory requires falsity as its operating condition. Remove the false content and the framework has nothing to explain. The condition described here survives the removal of all false content entirely. Accurate information, fully accessible, still fails to cohere.
It is not a problem of filter bubbles. Filter bubble theory is a theory of selective exposure — certain information reaches certain people and not others, producing epistemic isolation. The condition described here does not require isolation. It occurs in conditions of broad, open, and undifferentiated distribution. Wider exposure does not resolve it. It may accelerate it.
It is not epistemic fragmentation in the established sense. Fragmentation theory addresses the proliferation of competing truth claims across ideological or cultural contexts. It presupposes agents with positions. The condition described here is agentless and structural. No position is being advanced. No claim is competing with another. The failure is not about which version of reality prevails. It is about whether meaning can stabilize at all.
The gap these frameworks leave is precise: there is no existing structural model for coherence failure in conditions where information is intact, accurate, and broadly distributed, but the relational context required for meaning formation is absent.
This paper introduces Meaning Dissolution as a structural model that occupies that space. The model is demonstrated across two domains — AI-distributed information environments and legal reasoning and case law extraction — chosen specifically to establish that the mechanism is structural and scale-invariant rather than technological or context-dependent. The paper concludes with implications for AI governance, epistemic theory, and the structural study of meaning under conditions of fragmented distribution.
2. The Meaning Dissolution Model
2.1 Construct Definition
Meaning Dissolution is the process by which information remains accessible but cannot stabilize into coherent meaning because the relational context required for that stability is either stripped in transmission or not carried with the information under conditions of distributed exposure.
This definition requires unpacking at the level of mechanism. Information, as used here, refers to discrete propositional content — claims, findings, holdings, assertions, descriptions — that can be transmitted, stored, and retrieved independently of the broader structure in which it was originally embedded. Relational context refers to the structural relationships that organize information into coherent meaning: the sequence in which claims appear, the qualifications attached to assertions, the evidentiary basis on which conclusions rest, and the argumentative dependencies between propositions. Coherent meaning refers not to subjective interpretation but to the structural capacity for information to be assembled into stable, reliable, and applicable knowledge.
The mechanism of dissolution operates through two entry points. The first is stripping in transmission, where relational context is present in the source but removed as information moves through distribution channels. The second is absence at origination, where information is generated, summarized, or presented in a form that never carried relational context to begin with. Both entry points produce the same structural outcome: information that is factually intact but relationally unmoored.
The term dissolution is used deliberately and precisely. This is not fragmentation, which implies a formerly whole structure broken into pieces. It is not degradation, which implies a decline in quality. Dissolution names a specific process: the failure of structural bonds that hold components together. The components persist. The bonds do not. The information remains. The structure that made it meaningful is gone.
2.2 Theoretical Positioning
Meaning Dissolution occupies a distinct explanatory space that can be defined by its relationship to adjacent frameworks.
Misinformation theory (Wardle & Derakhshan, 2017; Lewandowsky et al., 2012) addresses the creation and spread of false or misleading content. Its operating condition is falsity or intent to mislead. Meaning Dissolution requires neither. It applies specifically in conditions where content is accurate. This is not a refinement of misinformation theory but a structural complement to it that addresses a category of epistemic failure misinformation theory cannot reach.
Filter bubble theory (Pariser, 2011; Sunstein, 2017) addresses selective exposure produced by algorithmic curation, arguing that individuals encounter information confirming existing beliefs while being shielded from disconfirming content. Its operating condition is differential distribution. Meaning Dissolution is independent of differential distribution. It operates in conditions of open, broad, and undifferentiated exposure. A piece of information encountered by everyone, distributed without restriction, can still dissolve under the model’s conditions.
Epistemic fragmentation (Nguyen, 2020; Sunstein, 2001) addresses the proliferation of competing truth claims across communities with different epistemic standards or values. Its operating condition is the presence of competing agents with divergent positions. Meaning Dissolution is independent of competing agency. No position is being advanced. No claim is competing with another. The failure is structural rather than competitive.
Information overload theory (Eppler & Mengis, 2004; Bawden & Robinson, 2009) addresses the cognitive and organizational effects of excessive information volume. Its operating condition is quantity. Meaning Dissolution is independent of volume. A single sentence, a single extracted holding, a single circulated claim can dissolve under the model’s conditions regardless of how much or how little information surrounds it.
What Meaning Dissolution introduces is a structural account of coherence failure that is independent of falsity, selective distribution, competing claims, and information volume. It proposes that the relational context of information is a load-bearing structural element — not a feature of presentation but a condition of meaning — and that when that context is absent or stripped, coherent meaning cannot form regardless of the accuracy or accessibility of the underlying content. Meaning Dissolution therefore operates as a relational structural model of meaning formation, specifying the conditions under which meaning formation fails when relational context is absent or stripped.
2.3 Scale-Invariance
Meaning Dissolution is a scale-invariant structural process. It operates wherever the underlying mechanism is present, regardless of the unit of analysis.
At the level of information systems, Meaning Dissolution occurs when large-scale distribution channels consistently strip or fail to carry the relational context that organizes the information they transmit. At the level of individual documents, it occurs when a text is excerpted, summarized, or recirculated without the internal relational structure — argument sequence, evidentiary basis, qualifying conditions — that made it coherent. At the level of discrete components, it occurs when a single sentence, finding, or principle is extracted from its relational frame and encountered as a standalone unit.
In each case, the mechanism is identical: relational context is absent or stripped, and meaning cannot stabilize. The scale changes. The mechanism does not.
2.4 Application Conditions
The model applies only where three conditions are simultaneously present:
First, the information is intact. The content has not been falsified, distorted, or selectively omitted in ways that would make it the subject of misinformation or selective exposure analysis.
Second, the relational context required for coherence is absent or stripped. The structural relationships that organize the information — sequence, qualification, evidential basis, argumentative dependency — are not present alongside the information itself.
Third, the exposure is distributed such that the full relational structure is not reliably encountered as a whole. The information circulates in conditions that systematically prevent the relational context from accompanying it.
When any one of these three conditions is absent, Meaning Dissolution is not the appropriate explanatory framework. If the content is false, misinformation theory applies. If distribution is selective, filter bubble analysis applies. If the full relational structure reliably accompanies the information, coherent meaning formation remains possible and dissolution does not occur.
3. Demonstration Case One: AI-Distributed Information Environments
3.1 The Mechanism at System Scale
Artificial intelligence systems engaged in information retrieval, summarization, and generation operate through processes that systematically strip or fail to carry relational context. This is not a design flaw but a structural feature of how these systems process and distribute information.
Large language models are trained on vast corpora of text from which statistical relationships between tokens are extracted. The training process does not retain the relational structures — argumentative dependencies, evidentiary chains, qualifying conditions, contextual anchors — that organize information into coherent meaning within source texts. It extracts patterns of association. When these models generate outputs, they operate through associative reconstruction rather than relational preservation: recombining information according to learned statistical co-occurrence rather than according to the relational logic that structured the original material.
The result is content that is often factually accurate — drawing on genuine information from genuine sources — but relationally unmoored. A summary generated by an AI system may accurately represent the claims made in a source document while stripping the qualifications, the evidentiary basis, the argumentative sequence, and the conditions under which those claims hold. The information persists. The structure that made it reliable does not.
This is distinct from hallucination, which refers to the generation of false content. Meaning Dissolution in AI-distributed environments occurs in cases where the content is accurate. The problem is not that the AI has introduced false information but that it has distributed accurate information without the relational context required for that information to stabilize into coherent meaning.
3.2 Distribution Without Context
The mechanism operates across multiple modes of AI-mediated information distribution.
In retrieval-augmented generation, source documents are broken into chunks, embedded numerically, and retrieved based on semantic similarity to a query. The retrieved chunks carry propositional content but often lack the relational context that gave that content its specific meaning within the source: the argument it was part of, the evidence that preceded it, the qualification that followed it. The information is accurate. The relational structure is absent.
In AI summarization, source texts are compressed into shorter representations. Compression necessarily involves selection, and what is systematically excluded in the compression process is precisely the relational material — the qualifying clauses, the evidential elaborations, the argumentative transitions — that does not survive the move toward propositional density. A summary can accurately represent all the major claims of a text while stripping the structural relationships that made those claims coherent in their original form.
In AI-mediated search and citation, information surfaces as fragments — sentences, findings, conclusions — detached from their original argumentative contexts and recombined in response to queries. A finding from a study may accurately represent the study’s conclusion while omitting the methodological conditions, the sample parameters, and the stated limitations that constituted the relational frame within which that conclusion was valid.
In each case, the information is intact. The relational context is absent or stripped. The mechanism of Meaning Dissolution is present.
3.3 Scale and Magnitude
What makes AI-distributed information environments the primary contemporary site of Meaning Dissolution is not that the mechanism is new — it is not — but that AI systems operate at a scale and velocity that makes the phenomenon structurally systemic rather than episodically incidental.
Prior to AI-mediated distribution, information stripped of relational context circulated at human scale. A sentence quoted out of context, a finding summarized without its qualifications, a principle extracted from its argumentative frame — these were isolated events whose epistemic effects were bounded by the speed and scale of human distribution. AI systems distribute information stripped of relational context at a scale and velocity that human distribution cannot approach. The phenomenon that was previously incidental has become infrastructural.
This does not mean that AI systems are uniquely responsible for Meaning Dissolution or that the model applies only in AI contexts. The following section demonstrates that the mechanism is pre-technological. What the AI context establishes is that Meaning Dissolution has become a structural condition of contemporary information environments rather than an occasional epistemic hazard.
4. Demonstration Case Two: Legal Reasoning and Case Law Extraction
4.1 The Relational Structure of Legal Meaning
Legal meaning is paradigmatically relational. A judicial holding does not possess independent meaning as a propositional unit. It acquires meaning — and validity as a legal rule — through its relationship to a specific set of facts, a particular procedural posture, a defined jurisdictional context, the interpretive standards applied by the deciding court, and the line of prior authority from which it emerges and to which it contributes.
The doctrine of precedent, stare decisis, makes this relational structure explicit as a matter of legal principle. A holding is binding not as an abstract proposition but as a ruling on a particular legal question arising from particular facts in a particular context. When that question, those facts, and that context change, the holding’s applicability changes with them. Legal meaning, in other words, is not carried by the proposition itself. It is constituted by the relational structure in which the proposition is embedded.
This relational structure is what legal education labors to install. The case method does not teach doctrine by presenting holdings as standalone propositions. It teaches doctrine by immersing students in the full relational context — the facts, the procedural history, the court’s reasoning, the counterarguments addressed, the limiting principles acknowledged — from which the holding emerges. The explicit pedagogical rationale for this method is that a holding encountered without its relational context cannot be reliably understood or applied. It can only be repeated.
4.2 Extraction and the Absence of Relational Context
The practice of legal extraction — reducing a case to its holding or rule — is widespread, necessary, and structurally hazardous under the conditions the model defines.
Legal summaries, secondary sources, practice guides, and increasingly AI-generated legal research present holdings as propositions detached from their relational frames. The information is accurate. The holding is a genuine legal rule. But the relational context — the factual conditions under which it applies, the jurisdictional limits on its authority, the interpretive conditions that define its scope, the subsequent cases that have qualified or limited it — is absent or stripped.
The result is structurally predictable under the Meaning Dissolution model. A practitioner who encounters an extracted holding without its relational context possesses accurate information that cannot be reliably applied. The holding accurately states what the court said. It does not carry the structural conditions under which what the court said has legal meaning in the practitioner’s context. Application of the extracted holding to a case with different facts, different jurisdictional context, or different procedural posture produces error not because the information was false but because the relational context required for its valid application was absent.
This is not a new problem in legal practice. The phenomenon of “ripping language out of context” has been a persistent concern in legal reasoning, and courts regularly distinguish between the holding of a case and dicta, between the rule as stated and the rule as limited by facts, between the proposition as extracted and the proposition as it functions within its relational frame. The Meaning Dissolution model does not introduce this observation. It provides a structural account of why the problem is structurally inevitable rather than episodically avoidable.
4.3 Scale-Invariance Demonstrated
The legal case demonstrates Meaning Dissolution operating across the three scales defined by the model.
At the system level, legal information infrastructure — databases, research platforms, secondary literature, and increasingly AI-generated legal research tools — distributes case law in forms that systematically strip or fail to carry the relational context of holdings. The structural conditions of Meaning Dissolution are present at the level of the system that produces and distributes legal information.
At the document level, a judicial opinion is a structured relational argument. Its holding acquires meaning from its relationship to the facts section, the procedural history, the discussion of prior authority, and the reasoning that connects these elements to the conclusion. When the opinion is reduced to a summary, a headnote, or a case brief, the relational structure that constitutes the full document is compressed or eliminated. Accurate content remains. The relational architecture that made it coherent does not.
At the component level, a single sentence extracted from a judicial opinion — the canonical form of legal citation — carries propositional content and zero relational context. The sentence accurately states what it states. Whether that statement has legal meaning in a new context depends entirely on relational conditions the sentence does not carry and cannot convey.
The mechanism is identical across all three scales. The unit of analysis changes. The structural process does not. This confirms the scale-invariance claim and demonstrates that Meaning Dissolution is a structural phenomenon rather than a contextually specific one.
4.4 The Pre-Technological Confirmation
The legal case serves a specific theoretical function in this paper beyond demonstration. It establishes that Meaning Dissolution is a pre-technological structural process that AI systems have made newly visible and newly prevalent, but did not introduce.
Legal practice has encountered the conditions of Meaning Dissolution — accurate information stripped of relational context, distributed in forms that cannot stabilize into coherent meaning — for as long as law has required the transmission of precedent across time and jurisdiction. The structural problem is not a product of the information age. It is a structural feature of any system in which meaning is relational and distribution is partial.
What AI systems introduce is not the mechanism but the scale. The mechanism that produced episodic misapplication in legal practice now operates at the infrastructure level of information distribution generally. Understanding Meaning Dissolution as a structural and pre-technological process is essential to understanding why it will persist regardless of the specific technological forms through which information is distributed.
Meaning Dissolution is not contingent on technological mediation. It is a structural consequence of any distribution condition in which relational context fails to accompany information — a condition that law has navigated for centuries and that AI has now made the default state of information environments at scale.
5. Boundary Conditions and Theoretical Positioning
5.1 What the Model Explains
Meaning Dissolution explains the failure of coherent meaning formation in conditions where information remains accurate, accessible, and intact, but the relational context required for that coherence is absent or stripped as a structural consequence of distributed exposure. It provides a structural account of a specific category of epistemic failure that existing frameworks do not reach.
The model’s explanatory contribution is most precise in three contexts: where information has been accurately summarized or compressed but stripped of its structural relationships; where information has been extracted from its original argumentative frame and circulated as a standalone proposition; and where information is generated, retrieved, or distributed by systems that process propositional content without preserving relational structure.
5.2 What the Model Does Not Explain
Meaning Dissolution makes no claims about cognitive processing. It does not address how individuals receive, interpret, respond to, or attempt to reconstruct meaning from dissolved information. The cognitive and psychological consequences of encountering dissolved information — what happens to comprehension, belief formation, decision-making, or identity when meaning cannot stabilize — are important questions that lie outside the model’s structural scope. They constitute a natural extension domain for future theoretical development but are not internal to the model as defined here.
The model does not account for intentional manipulation. Disinformation, propaganda, and strategic decontextualization involve directed agents whose goal is the distortion of meaning. Meaning Dissolution is an agentless structural process. The relational context fails to persist not because an agent has removed it for effect but because the structural conditions of distribution do not carry it. Where intentional manipulation is present, the appropriate frameworks are those that account for agency, intent, and directed influence.
The model does not address attention, engagement, or the affective dimensions of information processing. It is not a theory of what people do with information or how they feel about it. It is a structural account of whether information can cohere at all under specific conditions.
5.3 Relationship to Psychological Architecture
Meaning Dissolution emerges from and extends the Psychological Architecture framework (Starr, 2024), which organizes psychological life across four interdependent domains: mind, emotion, identity, and meaning. Within that framework, the meaning domain addresses how individuals situate personal experience within broader structures of significance, purpose, and interpretive coherence.
The Meaning Dissolution model operates at the interface between the mind domain and the meaning domain within Psychological Architecture. It describes the structural conditions under which the interpretive processes of mind fail to receive the relational material required to construct stable meaning. When the information environment systematically strips relational context, the interpretive engine operates without the structural inputs that meaning formation requires. The result is not misinterpretation but the structural inability to interpret at all.
This positioning within Psychological Architecture establishes Meaning Dissolution not as a standalone theoretical intervention but as a structural extension of an existing integrative framework. It situates the model within a cumulative program of psychological scholarship while identifying the specific mechanism through which environmental information conditions interact with the psychological structures responsible for meaning formation.
6. Implications
6.1 Implications for AI Governance and Design
If Meaning Dissolution is a structural consequence of how AI systems process and distribute information, then evaluation frameworks for AI outputs that focus exclusively on accuracy are systematically incomplete. A response can be factually accurate and epistemically harmful under the model’s conditions, because the harms produced by meaning dissolution — misapplication of accurate information, inability to form reliable judgments, collapse of coherent understanding — do not require falsity.
AI governance frameworks that treat accuracy as the primary or exclusive criterion of information quality will not capture the category of failure described by this model. Evaluation criteria need to include relational fidelity — the degree to which distributed information carries the structural relationships required for its coherent application. This criterion is operationalizable: it can be assessed by examining whether AI outputs preserve or strip the qualifying conditions, evidentiary basis, argumentative dependencies, and contextual anchors of their source material.
Design implications follow from this. AI systems intended for use in high-stakes epistemic contexts — legal research, medical information, policy analysis, scientific communication — should be evaluated not only on the accuracy of their propositional outputs but on the integrity of the relational context those outputs carry. A summarization system that consistently strips qualifying conditions while accurately representing primary claims is producing Meaning Dissolution at scale, regardless of its accuracy score.
6.2 Implications for Epistemic Theory
Meaning Dissolution proposes a new category of epistemic failure — structural incoherence — distinct from the established categories of falsity, bias, overload, and selective exposure. This has implications for how epistemology accounts for the conditions of knowledge in distributed information environments.
Standard epistemological accounts of justified belief require, among other conditions, that the believer have access to the reasons that support the belief. In conditions of Meaning Dissolution, the structural conditions of distributed information make it systematically difficult to access not just the reasons but the relational architecture within which reasons function as reasons. The conditions of justification are structurally undermined not by the absence of information but by the absence of the relational structure that makes information epistemically usable.
This suggests that epistemic evaluation in distributed information environments requires attention to structural conditions — the presence or absence of relational context — in addition to the propositional conditions that have traditionally defined epistemological analysis. Meaning Dissolution provides a structural framework for this extension.
6.3 Implications for Education and Communication
In educational contexts, the model has direct implications for how information is prepared and delivered for learning. Instructional materials that present accurate content stripped of the relational context that makes it coherent — correct answers without the reasoning that produces them, conclusions without the evidence that supports them, rules without the conditions that qualify them — create conditions for Meaning Dissolution at the pedagogical level. Students may accurately repeat content without being able to apply it, precisely because the relational context required for application was absent from what they received.
In professional communication and writing, the model argues for the preservation of structural relationships as a component of communicative responsibility. The compression of complex material into accessible summaries is legitimate and often necessary. But compression that strips relational context as a cost of accessibility produces, under the model’s conditions, content that cannot cohere reliably in the hands of those who receive it.
7. Limitations and Future Directions
The Meaning Dissolution model is presented here as a formal theoretical model rather than an empirically tested account. The demonstration cases provide conceptual evidence for the mechanism’s operation across domains and scales, but they do not constitute empirical validation of the model’s predictions. Future research should undertake empirical examination of the conditions under which Meaning Dissolution occurs, the degree to which different distribution contexts strip or preserve relational context, and the measurable epistemic effects of encountering information without relational structure.
The model’s boundary conditions require further specification in edge cases. The distinction between compression that strips relational context and compression that preserves relational context in condensed form is conceptually clear but practically complex. Developing operational criteria for this distinction will be necessary for applying the model in empirical and design contexts.
The cognitive and psychological consequences of Meaning Dissolution are explicitly outside the current model’s scope but represent a necessary extension domain. What happens to comprehension, judgment, and meaning-making when the structural conditions of meaning formation are systematically absent is a significant empirical and theoretical question. The interface between the structural model and cognitive and psychological accounts of information processing will require dedicated theoretical development.
The model’s application to non-linguistic information — visual, auditory, and multimodal content — requires examination. The relational context that the model defines is developed here in primarily linguistic and propositional terms. Whether and how the mechanism applies to forms of information that do not operate through propositional structure is an open question with significant implications for how broadly the model can travel.
Finally, the relationship between intentional and structural dissolution requires further theoretical attention. The model defines Meaning Dissolution as an agentless structural process, distinct from intentional decontextualization. In practice, these may interact: systems or agents may intentionally exploit the structural conditions that produce dissolution, using accurate information stripped of relational context as a tool of strategic influence. The boundary between structural and intentional dissolution, and the analytical frameworks appropriate for cases where both are present, will require dedicated examination.
References
Bawden, D., & Robinson, L. (2009). The dark side of information: Overload, anxiety and other paradoxes and pathologies. Journal of Information Science, 35(2), 180–191.
Bratman, M. (1992). Practical reasoning and acceptance in a context. Mind, 101(401), 1–15.
Bruns, A. (2019). Are Filter Bubbles Real? Polity Press.
Dworkin, R. (1986). Law’s Empire. Harvard University Press.
Eppler, M. J., & Mengis, J. (2004). The concept of information overload: A review of literature from organization science, accounting, management information systems, and related disciplines. The Information Society, 20(5), 325–344.
Frege, G. (1892). Über Sinn und Bedeutung. Zeitschrift für Philosophie und philosophische Kritik, 100, 25–50.
Goldman, A. (1999). Knowledge in a Social World. Oxford University Press.
Hart, H. L. A. (1961). The Concept of Law. Oxford University Press.
Lazer, D. M. J., Baum, M. A., Benkler, Y., Berinsky, A. J., Greenhill, K. M., Menczer, F., Metzger, M. J., Nyhan, B., Pennycook, G., Rothschild, D., Schudson, M., Sloman, S. A., Sunstein, C. R., Thorson, E. A., Watts, D. J., & Zittrain, J. L. (2018). The science of fake news. Science, 359(6380), 1094–1096.
Lewandowsky, S., Ecker, U. K. H., Seifert, C. M., Schwarz, N., & Cook, J. (2012). Misinformation and its correction: Continued influence and successful debiasing. Psychological Science in the Public Interest, 13(3), 106–131.
Llewellyn, K. N. (1960). The Common Law Tradition: Deciding Appeals. Little, Brown.
MacCormick, N. (1978). Legal Reasoning and Legal Theory. Oxford University Press.
Nguyen, C. T. (2020). Echo chambers and epistemic bubbles. Episteme, 17(2), 141–161.
Pariser, E. (2011). The Filter Bubble: What the Internet Is Hiding from You. Penguin Press.
Pasquale, F. (2015). The Black Box Society: The Secret Algorithms That Control Money and Information. Harvard University Press.
Raz, J. (1979). The Authority of Law. Oxford University Press.
Shannon, C. E., & Weaver, W. (1949). The Mathematical Theory of Communication. University of Illinois Press.
Starr, R. J. (2024). Psychological Architecture: A Structural Integration of Mind, Emotion, Identity, and Meaning. Depthmark Press.
Sunstein, C. R. (2001). Republic.com. Princeton University Press.
Sunstein, C. R. (2017). #Republic: Divided Democracy in the Age of Social Media. Princeton University Press.
Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146–1151.
Wardle, C., & Derakhshan, H. (2017). Information Disorder: Toward an Interdisciplinary Framework for Research and Policy Making. Council of Europe.
Wittgenstein, L. (1953). Philosophical Investigations. Blackwell.
Zuboff, S. (2019). The Age of Surveillance Capitalism. PublicAffairs.