Evidence Study Guide

📖 Core Concepts Evidence – Anything that supports a proposition or claim. Strength of evidential relation – Ranges from weak correlation (suggests) to indisputable proof (guarantees). Functions – Justifies beliefs, confirms hypotheses, or establishes/refutes legal claims. Philosophical contexts | Field | What counts as evidence? | Key idea | |------|--------------------------|----------| | Epistemology | Private mental states (percepts, beliefs) – sometimes even unconscious stored beliefs. | Evidence makes a belief rational or justifies it. | | Phenomenology | Self‑given intuitive knowledge (Selbstgegebenheit). | Provides indubitable access to truth, but can be fallible. | | Philosophy of Science | Public, observable information obtained via the scientific method. | Serves as a neutral arbiter; however, underdetermination and theory‑ladeness can blur neutrality. | | Law | Testimony, documents, physical/digital objects that prove or disprove factual assertions. | Governed by procedural rules (e.g., Federal Rules of Evidence). | Evidential relations Probabilistic (Bayesian): Evidence E confirms hypothesis H if $$P(H\mid E) > P(H)$$ Hypothetico‑Deductive (HD): Evidence = true observational consequences of H plus necessary auxiliary assumptions. Positive‑Instance: An observation that is a positive instance of a universal hypothesis and shares the same vocabulary. Types of evidence Scientific (empirical) – Observations & controlled experiments; burden of proof lies on the claimant. Legal – Physical, digital, trace, testimonial, relationship evidence; standards of proof vary (reasonable doubt, preponderance, clear & convincing, etc.). --- 📌 Must Remember Evidence definition – “what supports a proposition.” Bayesian confirmation – $P(H|E) > P(H)$. Underdetermination – Same evidence can equally support multiple theories. Theory‑ladenness – Background theories shape what counts as evidence. Legal standards of proof: Criminal: beyond a reasonable doubt. Civil: preponderance of the evidence. Others: probable cause, prima facie, clear & convincing, substantial evidence. Chain of custody – Must be documented from collection to courtroom to keep evidence admissible. Propositionalism (Epistemology) – Only propositional mental states count as evidence, yet sensory impressions are often treated as evidence anyway. --- 🔄 Key Processes Bayesian updating Start with prior $P(H)$. Acquire evidence E. Compute likelihood $P(E|H)$. Apply Bayes: $P(H|E)=\frac{P(E|H)P(H)}{P(E)}$. Compare posterior to prior; if higher → confirmation. HD verification Form hypothesis H. Derive observable predictions O using H + auxiliary assumptions A. Conduct experiment/observation; obtain data D. If $D$ matches O (and A are plausible), D is evidence for H. Legal evidence handling Collection: Identify relevant items; preserve scene integrity. Documentation: Record chain of custody (who, when, how). Presentation: Apply relevance & hearsay rules; meet the applicable burden of proof. --- 🔍 Key Comparisons Epistemology vs. Phenomenology Epistemology: evidence = private mental states, can be unconscious. Phenomenology: evidence = self‑evident intuition, claimed indubitable. Bayesian vs. HD vs. Positive‑Instance Bayesian: quantitative, probabilistic boost. HD: logical consequence + auxiliary assumptions. Positive‑Instance: single observation that directly instantiates the universal claim. Scientific vs. Legal evidence Scientific: public, repeatable, supports hypothesis testing. Legal: admissibility rules, various standards of proof, includes testimonial & physical items. Burden of proof vs. Standard of proof Burden: who must provide evidence (prosecution vs. plaintiff). Standard: level of certainty required (reasonable doubt vs. preponderance). --- ⚠️ Common Misunderstandings Correlation ≠ evidence – Accidental statistical links can raise $P(H|E)$ without a causal connection. Evidence always decides – Underdetermination shows the same data can support competing theories. All sensory data are evidence – Propositionalism restricts evidence to propositional states; raw sensations may need interpretation. Intuition is infallible – Phenomenological self‑evidence can still be mistaken. Higher quantity = stronger proof – Quantity does not overcome theory‑laden biases or relevance issues. --- 🧠 Mental Models / Intuition Evidence as a “weight” – Imagine belief as a scale; evidence adds weight to one side, shifting the balance toward the hypothesis. Bayesian “thermometer” – Each piece of evidence raises the temperature (probability) of a hypothesis; the hotter, the more confident. Underdetermination as “multiple maps” – Same terrain (data) can be drawn with different maps (theories); you need extra landmarks to choose the correct map. Theory‑laden glasses – Scientists wear conceptual lenses that tint observations; recognizing the tint helps evaluate objectivity. --- 🚩 Exceptions & Edge Cases Unconscious mental states can count as evidence (some epistemologists). Positive‑Instance approach fails when hypothesis and evidence use different vocabularies or involve unobservable theoretical entities. Bayesian liberalism – Very weak correlations may still raise $P(H|E)$, leading to “false confirmation.” Chain of custody breaks – Even perfect physical evidence can be excluded if documentation is incomplete. --- 📍 When to Use Which Choose Bayesian when you have prior probabilities and can quantify likelihoods (e.g., medical diagnosis, statistical inference). Choose HD for classical scientific testing where you can derive clear predictions and control auxiliary assumptions. Choose Positive‑Instance for simple universal claims with directly observable instances (e.g., “All swans are white” → sighting a white swan). Select scientific evidence in research contexts; select legal evidence when dealing with courtroom disputes. Apply higher proof standards (beyond reasonable doubt) in criminal cases; lower standards (preponderance) in civil matters. --- 👀 Patterns to Recognize Auxiliary assumptions lurking – Any “observational consequence” is likely backed by hidden premises. Chain‑of‑custody gaps – Missing documentation often signals inadmissible evidence. Same data, multiple theories – Spot underdetermination when a result is cited by competing hypotheses. Correlation without causation – High $P(H|E)$ but low theoretical justification. Hearsay flags – Statements not within the speaker’s personal knowledge usually trigger exclusion. --- 🗂️ Exam Traps “Evidence always proves a hypothesis” – Distractor; evidence may only increase probability, not guarantee truth. Confusing burden & standard – Remember: burden = who must prove; standard = how well they must prove. Assuming theory‑neutrality – Test items may ask you to identify theory‑laden influences on observation. Misreading “positive instance” – A single observation is evidence only if it logically entails the restricted hypothesis and shares vocabulary. Over‑relying on correlation – Choices that cite a strong statistical link without causal justification are often wrong. ---