Should You Believe Wikipedia? Social and Virtue Epistemology for HCI

# Should You Believe Wikipedia? Social and Virtue Epistemology for HCI **Video Category:** Technology & Philosophy (Human-Computer Interaction) ## 📋 0. Video Metadata **Video Title:** Should you believe Wikipedia? An introduction to social and virtue epistemology, with questions about the utility of philosophy for HCI researchers **YouTube Channel:** Stanford Center for Professional Development **Publication Date:** May 7, 2021 **Video Duration:** ~1 hour 20 minutes ## 📝 1. Core Summary (TL;DR) This presentation explores how philosophical frameworks regarding how we acquire knowledge—specifically social and virtue epistemology—can be applied to evaluate and design online information systems. It argues that truth on the internet is a socially constructed, collaborative achievement, where the reliability of information is directly proportional to the degree of rigorous peer review it receives. By understanding these epistemic mechanics, Human-Computer Interaction (HCI) researchers and platform architects can design better affordances that scaffold the evaluation of testimony, foster epistemic virtues, and manage online communities effectively. ## 2. Core Concepts & Frameworks * **Concept:** Epistemology (Justified True Belief) -> **Meaning:** The philosophical study of knowledge, classically defined as a belief that is not only true but also justified by a reliable reason or evidence, rather than arrived at by mere luck. -> **Application:** Used to evaluate whether users are justified in accepting a piece of information encountered online based on the reliability of its source. * **Concept:** Evidentialism vs. Reliabilism -> **Meaning:** Evidentialism requires every belief to trace back through a chain of evidence to basic sensory truths, which is practically impossible for everyday internet use. Reliabilism asserts that a belief is justified if it is produced by a reliable cognitive or institutional process. -> **Application:** Instead of verifying every raw fact, users apply reliabilism by trusting historically accurate institutional processes, such as accepting an article from *The New York Times* as generally factual barring specific contradictory evidence. * **Concept:** Internal Realism (Hilary Putnam) -> **Meaning:** A compromise between strict objectivism (a perfect reality is perfectly knowable) and subjectivism (we are trapped in our own perceptions). It posits that while reality exists objectively, it is only knowable through our fallible, subjective senses. -> **Application:** Explains why 135 seminar attendees can all agree a chair exists in a photo, despite subjective perception; truth is what we socially agree upon as the best representation of reality at a given moment. * **Concept:** Social Construction of Knowledge (Latour and Woolgar) -> **Meaning:** The process by which a claim evolves into an accepted fact through social interaction, review, and citation until the original authors no longer need to be cited because the fact is universally accepted. -> **Application:** Understanding how Wikipedia operates; a claim becomes "true" when a large, diverse group of editors reviews, debates, and eventually accepts it. * **Concept:** Virtue Epistemology -> **Meaning:** A synthesis of virtue ethics (Aristotle's focus on becoming a good person) and epistemology. It frames knowledge acquisition as a lifelong practice requiring the cultivation of specific "epistemic virtues" like intellectual humility, curiosity, and open-mindedness. -> **Application:** Designing online communities and platforms that actively encourage and reward these virtuous behaviors rather than optimizing solely for engagement or outrage. ## 3. Evidence & Examples (Hyper-Specific Details) * **The Game Show Door:** To illustrate "justified true belief," Bruckman explains that if a contestant guesses a car is behind Door #3 without evidence, and the car *is* there, it is not knowledge—it is luck. It only becomes knowledge if a stagehand secretly whispers the correct answer to them. * **The Birdwatcher (Richard Feldman, 2003):** An expert and a novice spot a bird and both spontaneously claim it is a rare pink-spotted flycatcher. The expert has a well-founded belief based on a reliable cognitive process (training), whereas the novice jumped to a conclusion out of excitement. Only the expert possesses actual knowledge. * **The Unitary Charge of the Electron (Latour & Woolgar):** Demonstrates the social construction of facts. Step 1: "Millikan and Fletcher claim that the charge... is unitary." Step 2 (later): "Since the charge... is unitary (Millikan and Fletcher, 1909)..." Step 3 (final): "Since the charge on the electron is unitary..." The citation disappears as the claim becomes an unquestioned fact. * **The Soccer Pass (John Greco, 2021):** A midfielder makes an impossible pass through a broken defense, and the striker kicks it in for an easy goal. Just as the goal is a collaborative achievement, knowledge on the internet is a collaborative achievement involving multiple actors passing and validating information. * **John Roberts' Wikipedia Page (Popular Page):** Created in May 2005 at 182 words. On July 19, 2005, when nominated for the US Supreme Court, it received 31 edits that day and 1,200 edits by the end of the month. It currently sits at over 6,800 words and perfectly matches all his judicial opinions. This proves that high-visibility pages with thousands of reviewers represent highly accurate, socially constructed knowledge. * **Massachusetts State House Wikipedia Page (Unpopular Page):** A page with a warning tag for missing citations. It mentions murals by Edward Brodney (a WPA initiative). Bruckman herself added an uncited claim that Edward used his sister Norma Brodney (Bruckman's grandmother) and his brother as models because he couldn't afford to pay professionals. This unverified claim remained on the page simply because no one else cared enough about the mural to fact-check it, proving that low-visibility pages are highly unreliable. * **Presidential Inauguration Edit War:** During the inauguration of Kamala Harris and Joe Biden, editors warred over Biden's status. One editor updated his page to "President" immediately after he took the oath (before noon). Another reverted it, citing the Constitutional rule that the term begins exactly at noon. At precisely noon, it was updated again. This showcases Wikipedia's real-time, highly pedantic consensus-formation process in action. * **Vulnerable Testimonial Networks (Emily Sullivan et al., 2020):** A network graph visualization demonstrating how a user (the red dot) might believe a fact is highly corroborated because they receive it from three different sources. However, tracing the network back reveals all three sources originate from a single, potentially flawed origin node, creating a dangerous illusion of consensus. ## 4. Actionable Takeaways (Implementation Rules) * **Rule 1: Expose the "Degree of Review"** - UI designers must create affordances that make the review history of a piece of information visible. A user should easily see if a page has been edited by 5,000 diverse experts or 2 anonymous users. * **Rule 2: Scaffold the evaluation of testimony** - Build systems that automatically provide metadata on the provenance of information and the historical credibility of the source, allowing users to efficiently apply "reliabilism" without doing deep research. * **Rule 3: Match affordances to knowledge construction needs** - Emulate Wikipedia's successful structures: provide a Revision History (transparency), Talk Pages (spaces for consensus formation), and WikiProjects (organized sub-communities focused on specific quality control). * **Rule 4: Design distinct spaces for distinct norms** - Do not build universal "free speech" or "heavily moderated" monoliths. Create platforms that support smaller, fragmented communities (like subreddits) where local groups can establish and enforce their own specific epistemic rules and values. * **Rule 5: Treat knowledge design as a socio-technical problem, not just a technical one** - Do not rely solely on automated fact-checking or AI. Design systems that leverage human collaboration, where the technology serves only to facilitate the social process of peer review. ## 5. Pitfalls & Limitations (Anti-Patterns) * **Pitfall:** Applying pure Evidentialism to internet research. -> **Why it fails:** It requires infinite regress. Users cannot practically trace every digital claim down to bedrock, basic sensory experiences. -> **Warning sign:** Information paralysis, where a user cannot make a basic decision because they feel they must personally verify the foundational data of every claim. * **Pitfall:** Idealizing scientific peer review as flawless. -> **Why it fails:** In highly interdisciplinary papers, a single reviewer may only understand one-third of the methodology. Reviewers assume other experts will check the rest, often resulting in complex sections going entirely unreviewed. -> **Warning sign:** Accepting a paper's conclusions blindly just because it passed peer review, without critically assessing if the specific claim falls within the reviewers' core competencies. * **Pitfall:** Assuming consensus based on volume in Vulnerable Testimonial Networks. -> **Why it fails:** Information consumers mistake the repetition of a claim across multiple secondary sources as independent verification, when all sources actually trace back to a single original node. -> **Warning sign:** An echo chamber where multiple news outlets or influencers use the exact same phrasing or cite the exact same primary document. * **Pitfall:** Assuming all Wikipedia pages share the same reliability. -> **Why it fails:** Reliability is not inherent to the platform; it is a byproduct of the *degree of social review*. Obscure pages lack the critical mass of editors required to challenge and correct false information. -> **Warning sign:** Treating a niche, rarely-edited page with the same epistemic weight as a highly contested, widely-viewed page. ## 6. Key Quote / Core Insight "A popular Wikipedia page is arguably the most accurate form of information ever created... A less popular page is much less reliable. The degree of review is the measure of the reliability of the information." *(Insight: Reliability is not a property of the platform itself, but a direct byproduct of the intensity and diversity of the social scrutiny applied to the specific information artifact.)* ## 7. Additional Resources & References * **Resource:** *Should You Believe Wikipedia? Online Communities and the Construction of Knowledge* by Amy Bruckman - **Type:** Book (Cambridge University Press, 2021) - **Relevance:** The core text from which this presentation is derived, detailing the design of online communities and knowledge construction. * **Resource:** "Vulnerability in Social Epistemic Networks" by Emily Sullivan, Max Sondag, Ignaz Rutter, Wouter Meulemans, Scott Cunningham, Bettina Speckmann & Mark Alfano (2020) - **Type:** Academic Paper - **Relevance:** Provides the network topology model showing how misinformation spreads when multiple sources unknowingly rely on a single origin. * **Resource:** *Laboratory Life: The Construction of Scientific Facts* by Bruno Latour and Steve Woolgar - **Type:** Book - **Relevance:** The foundational text on the social construction of knowledge used to explain how claims become facts via citation removal. * **Resource:** Richard Feldman (2003) - **Type:** Academic citation - **Relevance:** Source of the "birdwatcher" example explaining Reliabilism. * **Resource:** John Greco (2021) - **Type:** Academic citation - **Relevance:** Source of the "soccer pass" analogy for knowledge as a collaborative achievement. * **Resource:** J. Heersmink (2017) - **Type:** Academic citation - **Relevance:** Source for the compiled list of Epistemic Virtues (Curiosity, Intellectual autonomy, Intellectual humility, etc.).