Philosophy Of Science

Philosophy of Science course at Erasmus University

How to distinguish between the context of discovery and justification?

The Context of Discovery concerns how scientific discoveries are made - the process of having new ideas, “happy thoughts” or eureka moments.

The Context of Justification concerns the validation and testing of this discovery. Essentially justifying why the discovery is true or scientifically sound.

Reichenbach (1938) made this distinction, Popper used it later to argue that only justification matters in science, discovery is purely psychological.

What are some problematic aspects of the distinction between the context of discovery and justification?

The distinction is too sharp and artificial. As Whewell noted, a "Happy Thought" only becomes happy in hindsight after it has been tested. There is always a prehistory to an idea, earlier ideas and trials that led to it.

Additionally, discoveries have an aftermath, subsequent work that completes them. The retrospective nature means what counts as "the discovery" is determined later, not at the moment.

Relegating discovery entirely to psychology also ignores that there are describable patterns and conditions that enable discoveries, making it philosophically relevant after all.

What is Serendipity?

Coined by Horace Walpole (1754), it describes discoveries made "by accidents and sagacity", combining luck/chance and wisdom/skill. It refers to clues for unexpected knowledge and the ability to be alert to such clues. It involves finding something valuable while looking for something else.

What is Bahramdipity?

Bahramdipity describes instances of failing to give credit to individuals who would have played a role in the discovery process. Specifically, it refers to scientists whose potential discoveries are elided or neglected by those higher up in the hierarchy of the scientific community.

What are examples of Serendipity?

• Fleming's penicillin (1928): Returned from holiday to find fungus had inhibited bacterial growth in a petri dish, leading to discovery of antibacterial properties

• Edison's phonograph (1877): While working on telephone improvements, accidentally discovered sound recording

• Descartes' analytical geometry: Searching for certainty in one domain, created coordinate geometry

• Princes of Serendip: Made observations on the road that later proved valuable when they encountered a camel driver's problem

What are examples of Brahmandipity?

Graduate student Ernest Duchesne reported on the therapeutic effects of a Penicillium mold before Fleming. His work remained unknown and he died a few years later. While Fleming - who had better social connections and timing - became famous for the discovery.

What conditions are required for Serendipity to occur?

• Unexpected events or observations that weren't anticipated

• A prepared mind with skills to recognize potential value

• Reflection and follow-up on the unexpected finding

• Community epistemic expectations that make the finding remarkable/unexpected

• Contingency in the whole process (forks in the road, U-turns)

• Open science and contacts that create opportunities

• Time for reflection to recognize significance

• Retrospective recognition of value

What more is there to a scientific discovery than a mere 'happy thought' of a 'genius'.

Retrospecivity: The importance of a discovery can only be determined in hindsight.

Prehistory: There is always a prehistory of other ideas and trials that led to the finding.

Aftermath: Discoveries require subsequent articulation, precision, testing, and development before they become scientifically valuable.

What is the paradoxical nature of a scientific discovery (Copleston)

What is new cannot be expected: one can only find the unexpected by searching on the basis of some other expectation. You cannot search for what you don't know, yet all discovery is of something previously unknown. The resolution is that enhancing the chance of the unexpected can only be done while doing other things (or doing "normal science").(Referenced as Plato’s paradox)

Why is a scientific discovery a process rather than an event? (Kuhn)

Observation doesn't equal recognition - Scientists often observe anomalies or phenomena long before they understand their significance.

Conceptual assimilation takes time - A discovery requires the scientific community to develop new concepts, terminology, and theoretical frameworks to make sense of observations.

Multiple contributors and iterations - Discoveries typically involve many researchers making incremental contributions.

Theory-laden observation - What scientists "see" depends on their theoretical framework (paradigm). During paradigm shifts, the same data gets reinterpreted.

Resistance and acceptance - Scientific communities don't immediately accept discoveries that challenge existing paradigms. The "discovery" only becomes complete when the field accepts and integrates the new knowledge.

Why does serendipity qualify both as a weak and strong emergent phenomenon?

Weak emergence = Results from limits in our understanding, emergent property appears novel and unpredictable but could in theory get deduced from lower-level properties:

Serendipity arises from epistemic limitations—we can't predict what we'll find. With perfect understanding, we might explain how serendipitous events occurred.

Strong emergence = the emergent phenomenon cannot be deduced from the properties of the underlying context or process even in principle:

Serendipity cannot be deduced even in principle from underlying conditions. The complex interaction of factors enabling serendipitous discovery won't repeat, conditions constantly change. Being able to explain a process backward doesn't mean we could predict it forward. Epistemic limitations are part of serendipity's nature, it occurs precisely when there's a failure to expect that a particular process will produce its valuable outcome.

What is Abduction?

Abduction consists in studying facts and devising a theory to explain them. It is the only form of reasoning that introduces new ideas. Philosophers of science often define abduction as "Inference to the Best Explanation" (IBE). It was introduced by C.S. Peirce (1903).

How is abduction different from deduction and induction?

Peirce characterized these three forms of reasoning in terms of uberty (potential to generate new knowledge) and security (certainty of conclusions):

• Deduction: 100% security, 0% uberty. Provides certainty but no new knowledge.

• Induction: Moderate security and uberty. Generates a new rule based on data already known.

• Abduction: High uberty, low security. Introduces genuinely new ideas and content. Argues for a speculated case based on a rule and a result (the observed fact).

What is an example of abduction?

1. The problematic trajectory of Uranus led astronomers to hypothesize that there must be an eighth planet yet unobserved (Neptune), rather than give up on Newton's laws. The surprising orbital behavior would make sense if another planet existed.

2. You come home to find the fridge door open and food missing. You could hypothesize that a thief robbed your fridge or your housemate is fooling you, but you abduce that your housemate ate something in the middle of the night—this is the most reasonable explanation that makes the surprising fact rationally expectable.

What is the relation between abduction and scientific discovery according to Peirce and Grinnell?

Peirce focused on abduction precisely because deduction offers no new knowledge and science needs new knowledge. Abduction is the process by which scientists generate explanatory hypotheses when confronted with surprising facts.

Grinnell argues that abduction in its unconventional interpretation relates to experiencing surprising (serendipitous) observations that trigger a Gestalt-switch, opening up entirely new, previously unthought-of lines of research. The term "abduction" may even refer to being "abducted" by this Gestalt-switch toward a new research question.

What are some differences between common sense and science?

• Common sense relies on familiar causal explanations and mechanisms we can imagine

• Science uses a higher degree of systematicity in order to explain facts

• Descartes: Science doesn’t have to prove facts, they already exist, we need to search for something that explains them in science

Historical progression shows increasing distance:

• Descartes: Mechanistic explanations (still somewhat common-sensical)

• Newton: Novel explanations in terms of macroscopic object behavior guided by "forces"

• Modern physics: Increasingly hard to interpret in terms imaginable mechanisms

What is Fleck’s three different categories of scientific experiments?

• Heuristic experiments: Offer researchers new insights into the problem under investigation. These generate genuine discoveries and new understanding.

• Demonstrative experiments: Clarify heuristic findings into a form suitable for making discovery claims public. This is what philosophers typically call the logic of justification—making results presentable and convincing to others.

• Failed experiments: Perhaps the most common. Yield results that are inconclusive or uninterpretable. May occur for many reasons including technical errors, mistaken assumptions about methods, and poor study design.

What is the relation between abduction and failed experiments?

Grinnell emphasizes that surprising observations, which initially appear as failed experiments, are critical sites for abduction. As Vannevar Bush noted, "many of the most important discoveries have come as a result of experiments undertaken with quite different purposes in mind."

Opportunities for abduction often occur when implicit assumptions turn out incorrect—when assumptions about experimental design taken for granted lead to unexpected results. The key is noticing the surprising observation rather than dismissing it as merely technical failure.

As Claude Bernard wrote: Those with excessive faith in their theories make poor observations and often neglect important facts because they don't further their preconceived aims. Being open to surprise is essential.

What is Delbrück's Principle of Limited Sloppiness?

Our knowledge about any system under investigation is always "sloppy"—muddy, never completely clear. This isn't about carelessness but about the fundamental epistemic condition of doing research at the edge of knowledge.

Because of this, every experiment tests both explicit and implicit assumptions.

This creates the ambiguity problem: when an experiment fails to create expected results is might be due to either a wrong hypothesis or an inadequate experiment design. One can never know initially which is which.

Why is the question of what science and technology is of philosophical concern?

"What is X?" questions about basic concepts are philosophy's duty. Demarcating science from pseudo-science, metaphysics, religion, technology, and politics solves the demarcation problem (Popper). Science belongs to the ontic category of praxes, specifically scientific praxes, and philosophy must determine what distinguishes scientific from non-scientific praxes.

What general category of entity does science fall under, and which subcategory constitutes science?

Science belongs to the ontic category of praxes (collective organized activities), specifically the subcategory of scientific praxes, distinguished by having scientists, aims of knowledge production, scientific methods, confirmation criteria, and CUDOS values.

What is ontology?

The philosophical study of being. It is traditionally understood as the subdiscipline of metaphysics focused on the most general features of reality.

What are CUDOS values?

Communalism, Universalism, Disinterestedness, Originality, Skepticism

What are the 6 required parts of a Praxis?

(1) People: scientists

(2) Aims: production of scientific knowledge and contributing to prosperity

(3) Product: scientific knowledge

(4) Means: scientific methods, conceptual frameworks, domain of discourse

(5) Criterion: large and various body of confirming evidence

(6) Values: CUDOS (communalism, universalism, disinterestedness, originality, skepticism

What is a praxis and scientific praxis?

A praxis is a collective organized activity with 6 elements: people, aims, products, means, criteria, and norms/values. A scientific praxis has these elements specified for knowledge production using scientific methods and CUDOS values.

What is Skopology?

Discourse on aims: e.g. what should we aim for in science?

What is the difference between Aristotlean and Baconian aims of science?

Aristotelian aims (☞): epistemic/internal—knowledge, truth, explanation, understanding as ends in themselves.

Baconian aims (✌): practical/external—prosperity, welfare, wisdom, using knowledge as means to societal ends (knowledge is power).

How to distinguish between an aim of science and personal motives of a scientist?

Aims of science are goals of the praxis itself (e.g., producing knowledge).

Personal motives are individual reasons for participation (fame, overcoming loss, profit)—they may differ from or coincide with praxis aims but don't define them.

What is axiology?

Value theory, also called axiology, studies the nature, sources, and types of values.

What are the different values of science as discussed in discourse on axiology?

CUDOS (traditional): Communalism, Universalism, Disinterestedness, Originality, Skepticism.

Vs.

DECAY (modern alternative): Differentialism, Egoism, Capitalism, AdvocacY—reflecting science's industrialization and intimacy with technology/industry.

What are hominum-objects and what are their classifications?

Objects produced by human beings.

• Artifacts: concrete objects by engineers (technological) and artisans (artisanal)

• Works of Art: concrete objects by artists (artistic)

• Waste: physiological (urine, faeces) and garbage (litter, trash)

What is the dual nature of artifacts?

Artifacts have a dual nature, both as a material object, but it also has a function. Artifacts have proper functions (intended by maker) and accidental functions (unintended uses—butter knife as screwdriver, slipper to kill cockroach).

What is technology and what is the difference between scientific and technological praxes?

Technology aims to produce useful artifacts, it generally follows an engineering design process. Hence, profit is acceptable/the aim in the technological praxis, doesn’t follow CUDOS, and aim is not purely knowledge.

What is Woolgar & Latour's study Laboratory Life (1979) about? (Anthropology)

L&W studied Salk Institute lab anthropologically.

Scale of facticity: 5-stage process propositions pass through to become facts.

Social constructionism: Scientific facts are constructed in collective social processes influenced by contingent factors, not just discovered by following method.

What are the five stages associated with the scale of facticity?

1. Speculation/conjecture: Could be true/false, deemed relevant, considered interesting (value judgment)

2. Assumes modalities: How it relates to trusted regularities

3. Active empirical inquiry: Becomes object of judgment by laboratory personnel and field

4. Confirmed and fitted: Firmly fitted into existing scientific knowledge, especially recent findings

5. Scientific fact: Taken for granted, history stripped away, becomes morsel of knowledge

How do we use the world real?

Reality (R) = everything and every thing that exists, has existed, or will exist. "Real" refers to what exists in R independent of our beliefs/theories about it.

What are the two principles of Scientific Realism? (ScR)

Skopological Principle: Science aims to discover truth about all of R (observable and unobservable).

Semantic-Epistemic Principle: Mature scientific theories are approximately true; their central terms genuinely refer to entities in R; scientific knowledge includes truths about unobservables.

What does Feyerabend argue in terms of Methodology for ScR?

Argument: progress requires criticism → criticism requires proliferation → proliferation requires realism (scientists must believe theories can be true/false about R to motivate developing rival theories).

Criticism: Non-realists could also proliferate ideas for pragmatic reasons, not just truth-seeking.

What is Putnam’s no miracle (NoMir) argument regarding ScR?

Realism is the only philosophy that doesn't make science's success a miracle. If mature theories' terms don't refer and propositions about unobservables are false, yet they yield empirical/technological success—that would be miraculous. Therefore, ScR must be true (inference to the best explanation).

What is the Duhem-Quine UnderDetermination Thesis (UndTh)?

UndTh: The reported observations, performed measurements and gathered data in science underdetermine our models and theories—they do not and cannot guarantee the truth of a single model or theory.

To test a theory experimentally, we need auxiliary assumptions (A), facts (F), and instruments functioning properly according to certain regularities. We can't test an isolated hypothesis, only whole groups. This leads to the problem of never being able to know exactly what theory is true.

Since Scientific Realism (ScR) claims our theories approximate truth, UnDeT says: "How can you know that? Observations can't tell you which theory is true when multiple incompatible theories fit all the data."

When observations disagree with O (so ¬O is true), this implies:

¬ThD ∨ ¬A ∨ ¬F ∨ ¬PanBox, but you cannot know which one of these failed.

What are the defences of ScR against UndTh?

1. Appeal to explanation: When observationally/empirically equivalent theories exist, choose the one that explains phenomena best—that theory approximates truth more closely. (Problem: History shows false explanations once worked satisfactorily. Which is "best"? What if we have a "bad lot" of explanations?)

2. Temporary underdetermination: Perhaps future scientific research will break observational equivalence—new observations, measurements, data will decide. Underdetermination may be temporary, not permanent.

3. Rare in practice: Observationally equivalent theories are rare in history—usually constructed long after a theory's acceptance. In practice, theories usually make different observable predictions eventually.

What is the Pessimistic Meta-Induction over the history of science (PessMI)?

History's "dunghill" is full of discarded entities once thought real (ether, phlogiston, caloric). Current theories will also be discarded someday. Better remain agnostic about present theories' truth/posits' existence.

Laudan (1981): This poses two challenges to ScR: (1) Suggests ScR's principle "If a theory is successful, we can reasonably infer its central terms genuinely refer" is dubious—there have been highly successful theories with non-referring central terms (2) Suggests the realist's claim to explain science's success is false—part of historical success came from theories whose central terms didn't refer

How is ScR defended against PessMI.

• Divide et impera (Divide and conquer): Not all terms in discarded theories failed to refer. Separate referring from non-referring terms. Example: 'ether' didn't refer, but 'electromagnetic field' did. Mature sciences' central terms in successful theories DO refer, even if peripheral terms don't.

• Structural continuity: Mathematical structures and relations often survive theory change even when entity descriptions change. Focus on structures, not specific ontological commitments (leads toward Structural Realism).

• Progress in reference: Earlier theories had partially correct reference—later theories refined/corrected rather than completely replaced. Electrons aren't billiard balls, but they still exist and have properties continuously refined through theory development.

• Selection bias: PessMI focuses on failures. But massive numbers of theoretical posits HAVE survived and been vindicated (atoms, molecules, genes, electromagnetic radiation, etc.). Success rate improves over time—mature sciences are more reliable.

What is Van Fraassen's alternative explanation of the empirical and technological success of science?

Theories compete like organisms—only successful ones survive ("jungle red in tooth and claw"). We select theories on observational adequacy, so shouldn't be surprised our current theories work.

What are some criticisms of Van Frassen’s alternative explanation of the empirical and technological success of science?

Doesn't explain why theories are empirically adequate, just why we accept them; evolutionary analogy may be superficial.

Van Fraassen's response focuses on empirical adequacy, but NoMir also appeals to technological success—our ability to build artifacts based on theories about unobservables. Evolution doesn't obviously explain why theories enable such successful technology.

What are the three pillar distinctions of Constructive Empiricism? (CEmp)

(1) Observable vs. unobservable entities (scientific knowledge is only about entities that can be observed in reality)

(2) Acceptance vs. belief (accept theory = believe empirically adequate; don't have to believe it to be true)

(3) Empirical adequacy vs. truth as aim of science.

What is the skopological thesis of CEmp?

Skopological Principle: Science aims to give us empirically adequate theories—theories that tell the whole truth about observables in R. NOT to discover truth about unobservables. Epistemic modesty is virtue, epistemic hubris is a vice.

What are the implications of the skopoligcal thesis of Constructive Empiricism? (CEmp)

(1) Only empirical + logical adequacy are epistemic virtues (explanation, unification, parsimony are pragmatic).

(2) Propositions about unobservables are accepted but not believed—don't constitute knowledge.

(3) UnDeT and PessMI don't threaten CEmp (only unobservables are underdetermined/discarded).

What are the two vicissitudes (problematic cases) of observability?

(1) Glasses: Don't change observable extensions—can still see things with/without glasses.

(2) Microscopes: Create new observable phenomena (images on screen), but what's viewed isn't observable (blood cells are too small to ever see directly, no matter how close). (Van Frassen)

What does Van Fraassen mean by: “it is not an epistemological principle that one might as well hang for a sheep as for a lamb”?

The saying (among ovine thieves): Hanging was the penalty for stealing sheep. Since the penalty for stealing a valuable sheep (ScR) was the same as for stealing a less valuable lamb (CEmp), a thief might as well "hang for a sheep as for a lamb"—steal the more valuable animal since the punishment is identical.

The risk of being proven wrong is THE SAME in both cases—observations/experiments can only test empirical adequacy. Believing additionally in unobservables' existence doesn't make you "additionally vulnerable" to empirical refutation. So the realist's "richer, fuller picture" is illusory wealth—"empty strutting and posturing, this display of courage when not under fire."

What is the core of structural realism?

Scientific theories tell us the truth about R's structure/mathematical relations, not necessarily about nature of entities themselves. Structures survive theory change even when entity descriptions change. Aims to preserve ScR's benefits while avoiding UnDeT and PessMI.

How does Strctural Realism (StR) aim to evade UndTh and PessMi?

vs. UndTh: Mathematical structure is less underdetermined than specific ontological interpretations, often share the same mathematical structure even while differing in ontological interpretation.

Example: quantum mechanics vs. Bohmian mechanics—same mathematical structure (make identical predictions), different interpretations of what's "really there."

vs. PessMI: Entities get discarded, but mathematical structures survive theory changes (Maxwell's equations survived ether's demise). Progress = discovering structures, not entity-types.

Example: Maxwell's equations (structure of electromagnetism) survived transition from ether theory to field theory—even though ether was discarded

What are some criticisms to Structural Realism?

1. If reality consists ultimately of structures (as ontic structural realism claims), what are these structures structures of?

2. Structural Realism (StrR) has a field day with physics, which is replete with unobservables and mathematical structures. Does StrR work equally well for other branches of science, such as biology, geology, economics, psychology, and history?

3. What exactly is a structure?

4. Still vulnerable to some PessMI - not all mathematical structures survive, and continuity of the structures is also debatable as they get revised.

5. If multiple theories share the same structure, how do we choose between them? StrR may still face underdetermination problems, just at a different level.

How do you distinguish; registering, seeing, observing and seeing conceptually?

Registering: light hits retina, image forms (no awareness).

Seeing: registering + being aware of entity.

Observing: purposeful action of looking for something specific (trained cognitive faculty).

Seeing conceptually: seeing + possessing language/concepts + forming observational belief that F(x).

What are observation concepts?

Any human can determine whether it applies to observable entity by simple observation with naked eye, without possessing scientific knowledge. Children of any culture can learn them. Examples: red, round, larger, longer. Contrast with theoretical predicates (require theory: electron, mass, temperature).

What is monadic and polyadic?

Monadic: "x is Red," "x is Round"

Polyadic (relational): "x is Larger than y," "x is Longer than y"

What is the criterion for observability?

Explication: Entity x is observable iff for every human S under normal circumstances: if S were in front of x in broad daylight with eyes wide open, long enough, then S would register x. Uses subjunctive conditional (were-would). Observability = anthropocentric, extrinsic property.

What is an observation sentence?

Sentence using only observational predicates about observable entities that any competent language speaker could verify/falsify by observation. Examples: "The cat is on the mat," "This cherry is red." Logical Positivists wanted to reduce all science to these.

What is empirical evidence?

Weaker justification than proof (deduction) but strong—the kind science requires. Etymology: Latin ex+videre = "to see," evidens = "obvious." Evidence comes from observation/measurement—provides justification for promoting hypotheses to knowledge when sufficiently large and diverse.

What are theory-infected observation reports?

Theory-infected observations: observational reports using concepts from scientific theories. If concept F comes from theory T, observations are "T-infected via F." Examples: "sun moves across sky" (Ptolemaic), "Earth rotates" (Copernican). Every observational judgment is necessarily conceptually infected—cannot have judgments without concepts.

Why are theory-infected observations necessary for science?

(1) All observations require concepts

(2) Scientific observations require scientific concepts from theories

(3) Without theory-laden concepts, scientists couldn't describe, explain, or test anything scientifically

(4) Observation isn't passive reception but active, conceptualized interpretation.

What are the three main threats of theory-infection?

(I) Anti-Positivist: undermines Logical Positivism's project of reducing science to observations.

(II) Anti-Test: rival theories can each describe phenomena to confirm themselves (Ptolemaic vs. Copernican seeing sun move vs. Earth rotate).

(III) Circularity: using theory-laden observations to test that very theory seems circular. We’re using T in order to test T, can be self-confirming.

How can these threats be allayed?

(I) Anti-Positivist: Positivists can do rational reconstruction anyway (though harder).

(II) Anti-Test: Theories eventually make different predictions; phases of Venus decided Ptolemaic vs. Copernican.

(III) Circularity: Use empirical biconditionals connecting theory concepts to observations; inferences remain semantically unsound but informative; accept Popper's view of "piles driven into swamp."

What are empirical biconditionals, and why are they needed in science?

Statements that connect theoretical concepts to observable conditions through if-and-only-if relationships. Also called "bridge principles" or "correspondence rules. ." Example: "x has mass m ↔ pointer on balance shows m." Needed to:

(1) give empirical meaning to theoretical terms,

(2) enable testing,

(3) alleviate circularity by providing independent observational basis for theoretical concepts.

What is the circle of justification, and how can one break out of it?

Using theory-laden concepts in observations to justify that very theory = circular reasoning.

Breaking out:

(1) Empirical biconditionals provide independent observational conditions;

(2) Accept semantic unsoundness but informativeness of testing;

(3) Popper's "piles in swamp"— we accept pragmatic sufficiency because an absolute foundation doesn’t exist - basic statements are not discovered at the bottom but are imposed from above by the scientific community through decision;

(4) Coherentism—mutual support rather than foundational justification.

What is epistemic growth?

Epistemic growth (or epistemic progress/scientific progress) is the growth of scientific knowledge over time, including both propositional knowledge (knowledge-that: true propositions based on empirical evidence) and practical knowledge (knowledge-how: conducting experiments, building technology). The traditional explanation is that employing the scientific method produces scientific knowledge - the process explains the product.

What is the three-fold context distinction?

(✶) Context of Discovery (Reichenbach): where scientists create hypotheses, models, theories, experiments—imagination dominates, no norms or methods for genuine novelty.

(✹) Context of Pursuit (Laudan): where scientists pursue, communicate, refine, test, and develop their created ideas. (After discovery but before justification)

(●) Context of Justification (Reichenbach): where judgments are made—has the hypothesis been confirmed or infirmed? Should theories be rejected or adjusted? Originally, philosophers focused only on justification (normative), leaving discovery to historians (descriptive), but these contexts actually interlock rather than being neatly separated.

What is the difference between Enumerative induction and Colligative induction?

Enumerative induction (Mill): Simple generalization from experience - observing many instances and inferring a universal conclusion (e.g., every observed elephant has a trunk → all elephants have trunks).

Colligative induction (Whewell): The conclusion contains a concept absent from the premises - "binding" data into a new concept. Examples: Kepler colligating Mars's observed positions into an "ellipse" (concept absent from position data); curve-fitting where a mathematical formula brings in a concept absent from raw data points. Enumerative induction merely sweeps data together; colligative induction binds data into a concept.

What are the main criticisms of inductivism?

Francis Bacon (1561-1626) claimed that Induction was THE distinctive scientific method,The main criticism of Inductivism is that it is one but not the only method employed in scientific research - there are others besides induction. Examples include: gathering data without any specific hypothesis to test and parameter fitting in models.

Hume’s problem of induction: No matter how many white swans we observe, we can never know whether "All swans are white" is true. We cannot infer a universal law from a finite number of observations. In this sense, we cannot acquire scientific knowledge from observations.

What is Hypothetico-Deductivism (HD)?

Hypothetico-Deductivism is the thesis that science proceeds by following the HD-method, that the HD-method is the one and only scientific method. The HD-method works as follows:

From background knowledge (Kn), particular facts (F), and a hypothesis (H), you deductively derive a test-implication O (an observation sentence):

(Kn ∧ F ∧ H) ⊢ O. You then test O empirically - if O is verified, H is confirmed; if O is falsified, H is infirmed. The Logical-Positivists (Carnap, Hempel, Reichenbach) championed this view.

How does HD-confirmation and infirmation work?

When a test-implication O is verified by observation/experiment, we say hypothesis H is deductively confirmed relative to background knowledge Kn.

When O is falsified, H is infirmed (disconfirmed). HD-confirmation provides empirical justification for claims to knowledge - deductively confirmed claims are promoted to scientific knowledge.

However, Popper rejected this epistemic use, arguing that confirmation cannot provide justification due to Hume's Problem of Induction.

What can be used as an illustration of the HD-method?

Joule-Lenz Experiment (1840-1842): James Prescott Joule investigated: What's the relation between electric current (I) and heat (Q) in a conducting wire? He immersed a copper wire in water, ran current through it, measured temperature rise. Initial hypothesis: Q ∝ I. Deduced: T/I = constant.

Result: Falsified! But partially true - higher I means higher T. New hypotheses tested: Q ∝ √I, Q ∝ I², Q ∝ I³. Discovery: Q ∝ I² and Q ∝ R (resistance). Final law: Q = RI² (Joule-Lenz Law), connecting electrodynamics and thermodynamics.

What are criticisms of the HD method?

Criticism is similar to Inductivism's: HD-method is one but not the only method in science. Other methods exist: gathering data without specific hypotheses to test, parameter fitting in models (climate science), purely theoretical activities without empirical testing (superstring theory, quantum gravity), and computer simulations.

However, even activities not following HD-method often yield results that eventually play a role in HD-testing, so the HD-method remains indispensable for understanding scientific practices

What are the two fundamental epistemological problems according to Popper?

I. Demarcation Problem ("Kant's Problem"): (I.a) How to demarcate science from non-science (pseudo-science/metaphysics)? (I.b) What is scientific knowledge?

II. Acquisition Problem ("Hume's Problem"): (II.a) Can we acquire scientific knowledge through sensory experience/observations? (II.b) Are sense organs sources of knowledge? If not, how do we acquire it? Popper answered (II.a) negatively and provided novel solutions to (I.b) and (II.b).

What are wrong answers according to Popper?

Inductivism/HD confirmation and verification. Popper rejected that observations provide justification (confirmation/verification) for scientific knowledge. Example: Freud, Adler, Marxism, astrology - you can always find confirming facts; they're "always fitted, always confirmed" - but this apparent strength is their weakness.

Their theories are compatible with any observation, making it impossible to prove them wrong. Einstein's general relativity was different - it took a risk: if starlight passing the sun wasn't bent, theory would be refuted. Genuine science runs enormous risk of being proved wrong.

What are the correct answers according to Popper?

Falsifiability is the key. For (I.b) What is scientific knowledge: q is scientific knowledge iff q is true, falsifiable, and has undergone several failed falsification attempts.

For (I.a) Demarcation: Science studies falsifiable propositions and tries to falsify them; pseudo-science/metaphysics don't.

For (II.a) Acquisition: NO - observations cannot justify knowledge (Hume's Problem).

For (II.b) How we acquire it: Scientists conjecture and refute. All observations can do is falsify. The only legitimate use of HD-method is trying to refute conjectures: "Conjectures & Refutations," not "Conjectures & Confirmations.”

What is falsifiability according to Popper?

q is falsifiable iff some observation sentences make q false. q is scientific iff q is falsifiable. q is falsified iff some empirical evidence has made q false. Falsifiability demarcates meaningful scientific from meaningful non-scientific sentences (NOT meaningful from nonsensical). For Popper, metaphysics was meaningful and heuristically valuable.

What is the falsification method of science?

The falsificationist method uses the HD-method but ONLY to try to FALSIFY hypotheses, never to confirm them. Science inquires into falsifiable propositions only. The method: scientists conjecture bold hypotheses, then attempt to refute them through severe testing. Failed falsification attempts corroborate (but don't confirm) hypotheses. The game is "Conjectures & Refutations," not confirmation.

What are some criticisms of Popper’s falsifiability?

Criticism 1: Falsifiability is too weak (marks "I'll eat pasta tonight" as scientific) and too strong (marks "There are gravitational waves" as unscientific). It's descriptively inadequate as a demarcation criterion.

Criticism 2 - The Corroboration Dilemma: Corroboration (failed falsification attempts) is identical to confirmation. But confirmation is inductive, which Popper rejects!

Dilemma: (D1) Embrace induction = Falsificationism goes kamikaze, OR (D2) Disown corroboration = problematic view of scientific knowledge and can't explain why we rely on it with our lives.

What are the characteristic features of a paradigm?

A paradigm consists of:

(1) exemplars - successful puzzle solutions that establish the paradigm,

(2) accepted theories that function like dogmas and are never tested,

(3) shared values, norms, and methods for puzzle-solving, and

(4) concepts that determine what can be observed. Paradigms require near-universal consensus among practitioners to exist. Young scientists undergo an initiation rite through textbooks and PhD training to become paradigm members.

What is incommensurability?

Incommensurability means different paradigms cannot be compared using shared standards because:

(1) there are no paradigm-transcendent norms to evaluate them,

(2) their concepts cannot be translated into each other, and

(3) they evoke fundamentally different pictures of the world. What scientists "see" when observing changes completely between paradigms (Ptolemy saw the moon as a planet; Copernicus saw it as a satellite).

What is mostly going on in a paradigm?

During normal science, scientists primarily engage in puzzle-solving - applying the paradigm's exemplars to solve similar puzzles. Theories are never tested (they're treated as dogmas); only proposed solutions are tested. Scientific knowledge accumulates through the growing repository of solved puzzles. Failures to solve puzzles reflect on the scientist's competence, not the paradigm itself.

What is an anomaly in the paradigm?

An anomaly is a puzzle that resists solution within the paradigm - when models based on accepted theories don't work or observations can't be satisfactorily interpreted using the paradigm's concepts. Anomalies are typically archived (to be solved later) or forgotten/dismissed as mistakes. They don't threaten the paradigm during normal science; only when anomalies pile up unpredictably does crisis emerge.

What is a scientific revolution according to Kuhn?

(paradigm shift) occurs when anomalies pile up unpredictably, triggering a crisis where paradigm members question everything. A fresh-minded youngster proposes brand new ideas with alien concepts that solve anomalies. The shift involves a Gestalt-switch (like religious conversion) rather than purely rational choice. History is rewritten, the old paradigm sinks, and survivors must "adapt or drown" in the new paradigm.

What was Kuhn’s views criticized for?

The main criticism: "Kuhn was epically exaggerating and egregiously extrapolating." Popper dismissed Kuhn's "mob psychology" as dangerous - normal science is a threat to good science.

M. Masterman found more meanings of "paradigm" than fingers on two hands. Critics argued Kuhn undermined scientific rationality, reducing theory choice to the "number, faith and vocal energy" of supporters, making "truth lie in power."

What is a scientific research programme according to Lakatos?

Consists of: (1) a hard core of one or more theories adhered to like dogma and never tested,

(2) a protective belt of auxiliary assumptions, ad hoc hypotheses, and models that protect the hard core from falsification, and

(3) positive and negative heuristics prescribing what problems to work on and what to ignore. Rival programmes exist simultaneously, fighting for the same phenomena.

What is empirical and theoretical progression, and degeneration? (Lakatos)

A programme is empirically progressive if it predicts new phenomena or unexpected observations that are then confirmed; otherwise it's empirically degenerative. It's theoretically progressive if it can accommodate new phenomena or unexpected observations (but didn't predict them beforehand); otherwise it's theoretically degenerative. These concepts provide norms for comparing and appraising rival programmes.

What is a scientific revolution according to Lakatos?

For Lakatos, a scientific revolution is a historical period of accelerated degeneration of one programme and accelerated progression of a rival programme. Progressive programmes gradually replace degenerative ones over long historical periods (decades). No instant rationality - rationality operates in the long run. The Copernican Revolution: ArPt (Aristotle-Ptolemy) was degenerative while CopKG (Copernicus-Kepler-Galileo) proved progressive.

What does scientific rationality consist in according to Lakatos?

For Lakatos, scientific rationality consists in following methods that provide internal (methodological) rather than external (psychological/sociological) explanations of epistemic progress. Scientists act rationally by working in progressive rather than degenerative programmes. The hard core and the positive and negative heuristics should be followed. Rationality operates in the long run, not instantly.

What is Lakatos’ MetaHistorical Criterion?

Methodology M is better than M′ if and only if M can explain internally more of what happened in the history of science that led to epistemic progress than M′ (thereby making it more rational), and therefore M needs to explain less externally of what happened that led to epistemic progress than M′ (thereby making it irrational).

Can you distinguish between the different types of scientific methodologies?

Universal Methodology: One scientific method employed in every scientific praxis (Inductivism, Hypothetico-Deductivism, Falsificationism).

Discipline Methodology: Distinct methods per discipline.

Paradigm Methodology: Distinct methods per paradigm (Kuhn).

Contextual Methodology: Methods vary by research context within paradigms (Feyerabend).

No Methodology: No scientific methods at all ("Nothing Goes").

What is methodological anarchism?

It is best understood as Contextual Methodology - not that there are no methods, but that all methodologies have limits and scientists sometimes must violate rules to make progress. Feyerabend's goal: convince readers that "all methodologies, even the most obvious ones, have their limits." He showed that rules were violated by perceptive scientists and had to be violated - insistence on rules would have arrested progress.

Why does inductivism, Hypothetico-Deductivism, Falsificationism, Kuhn's paradigm view, and Lakatos' methodology fail to account for the Copernican Revolution?

Feyerabend's central claim: None of the previous methodologies can explain the Copernican Revolution, and scientific revolutionaries made progress by contravening these rules.

Inductivism fails: observations didn't favor heliocentric view (Copernicus never used a telescope; Kepler assumed CopKG to find his laws).

Hypothetico-Deductivism & Falsificationism fail: CopKG was falsified twice (falling objects, no stellar parallax) but kept anyway.

Kuhn fails: No crisis preceded the revolution; ArPt could solve its problems.

Lakatos fails: Even by his lights, it was never irrational to defend ArPt, so cannot explain why everyone abandoned it.

Why is the dichotomy of belief too coarse to do justice to our doxastic lives?

The simple belief/disbelief dichotomy is inadequate for representing our doxastic lives (lives as believers). We don't just believe or disbelieve - we have gradual beliefs with varying strengths. Example: Thomas can be totally convinced it's raining now, confident it will rain today, less confident about tomorrow, uncertain about next week, and strongly disbelieve Earth is flat. We need degrees of belief (credences) to capture this richness, not just binary yes/no.

What is Bayesianism?

Bayesianism is NOT a descriptive psychological theory of how people actually change beliefs. It's a normative philosophical theory of doxastic rationality - it prescribes how to deal with beliefs in order to be doxastically rational. It models belief quantitatively, by degrees using Probability Theory to assign numerical credences (degrees of belief) to propositions. It tells us how rational believers ought to structure and update their beliefs.

What are the three principles of Bayesianism?

I. Credence Principle: For every subject S, there's a function assigning a real number g ∈ [0,1] to each proposition p - the degree of belief or credence.

II. Principle of Synchronic Coherence: Degrees of belief must satisfy Probability Theory's two Kolmogorov axioms (Unity and Additivity) at any fixed time.

III. Principle of Diachronic Coherence: Beliefs at different times must be related by Bayes' Rule - posterior Pr₁(h) = [Pr₀(e|h)/Pr₀(e)] × Pr₀(h). (prior)

What are the two axioms of Unity and Additivity?

Unity: Every probability function Pr assigns 0 to absurdum (⊥), 1 to veritas (⊤), and a real number between 0 and 1 to every other proposition p.

Additivity: If p and q are incompatible, then: Pr(p ∨ q) = Pr(p) + Pr(q), two propositions are incompatible iff their conjunction implies absurdum: (p ∧ q) ⊢ ⊥, meaning if one is true, the other must be false.

What is the conditional probability Pr(p|q)?

Probability of p given q ≡ Pr(p ∧ q) / Pr(q)

What is the belief-change factor?

The belief-change factor f(e, h) is the factor by which to multiply the prior degree of belief Pr₀(h) to obtain the posterior Pr₁(h).

Formula: f(e, h) = Pr₀(e|h) / [Pr₀(e|h) * Pr₀(h) + Pr₀(e|¬h)(1 - Pr₀(h))]. Pr₀(e|¬h) = likelihood if hypothesis is false.

Then Pr₁(h) = f(e, h) × Pr₀(h). If f(e, h) > 1, evidence confirms h. If f(e, h) < 1, evidence infirms h. If f(e, h) = 1, evidence is irrelevant.

How to use Baye’s rule?

(1) Estimate the prior Pr₀(h) - your initial degree of belief in hypothesis h.

(2) Estimate likelihoods: Pr₀(e|h) = how probable is the evidence if h is true? Pr₀(e|¬h) = how probable is the evidence if h is false?

(3) Calculate belief-change factor: f(e, h) = Pr₀(e|h) / [Pr₀(e|h)Pr₀(h) + Pr₀(e|¬h)(1-Pr₀(h))].

(4) Calculate posterior: Pr₁(h) = f(e, h) × Pr₀(h).