Philosophy Of Science

Flashcards covering key vocabulary from the Philosophy of Science lecture notes.

What is the sense of a word/sentence

Word: the concept it expresses.

Sentence: the proposition it expresses.

What is the topic of a word? When does it refer?

What it is about. It refers when the topic of a word exists.

What are the five things a praxis consists of?

People, Aim, Means, Criteria, Norms and Rules

Axiology

The social discourse on values.

Skopology (Aristotelian and Baconian)

Discourse on aims.

Aristotelian aim: internal, epistemic, for science.

Baconian aim: external, practical, for the good of humanity.

What is CUDOS?

Communalism: no secrecy or private ownership.

Universalism: objectivity; claims to knowledge can be made by anyone regardless of race class gender etc.

Disinterestedness: act selflessly, no conflicting interests.

Originality: scientists should contribute something new.

Skepticism: being critical and thoroughly testing.

What is DECAY?

Differentialism: knowledge claims are not merely epistemic but also subject to political, moral, economic and cultural influence.

Egoism: boasting and bragging of scientists, marketing enterprise around research.

Capitalism: the profiting from ideas, innovations and discoveries in science through patents.

Advocacy: commitment to a social or political agenda.

Scientific Realism in two principles

Skopological principle: the epistemological aim of science is to find everything there is concerning the adequate description/representation of reality.

Epistemological-Semantic Principle: the properties stored in models and theories are made possible by reality, specifically the topic propositions.

Two anti realist views

(Neo) Kantian Idealism: reality (noumena) is unknowable, only phenomena are knowable (made possible by cognitive capacities). Science concerns itself with phenomena.

Nominalism: state that reality is unknowable (like the Kantians), but go one step further: human frameworks are put onto the world by means of inference to the simplest explanation (linguistic idealism).

Explain Popper’s thesis concerning the partial truth in Semantic Realism, and the critique on this view in the Tichy-Miller Theorem.

Semantic Realism: reality gives propositions and concepts their referents.

Popper’s idea was that scientific knowledge should be closer and closer to the truth, meaning asympotically approximating it. However, as summarized in the Tichy-Miller theorem, this implies that one false theory is not any more false than another.

Two arguments for realism

Feyernabend: progress requires criticism, criticism requires proliferation (diversity of ideas/theories) proliferation requires realism (one theory true, another false). Therefore scientific progress is hindered by anti-realism.

No miracle argument: scientific realism is the only theory that accounts for the succes of scientific enterprise without relying on miracles. Although Feyernabend critiques this idea: succesfull theories are self selected because of their success.

Two arguments against scientific realism, and the possible lines of defence.

Duhem-Quine underdetermination thesis: observations cannot determine the reality of models or theories. Observational knowledge also does not show which axiom/belief should be discared when hypothesis O is false (see slides).

Three lines of defense:

1. Other hypotheses possible that are independent of H.

2. Underdetermination is relatively rare in science (it happens mostly in physics), and can be broken by inference to the best explanation.

3. See slides.

The Pessimistic Meta-Induction over the History of Science: many theories and models have been discarded (1) and replaced (2) by better ones in the history of science. Thus, reject current theories, they will be replaced by better ones.

Two lines of defense:

Against 1: not all theories are replaced/discarded, a lot remain.

Against 2: the theories may be replaced, but the referring terms remain.

Constructive Empiricism in two principles

Skopological principle: the aim of scientific models and theories is that they relate with the observable nature of reality.

Semantic-Epistemic Principle: models/theories only relate to the observable part of reality, and those that do not are neither believed nor disbelieved.

Van Fraassen: it is not an epistemological principle that one might as well hang for a sheep as for a lamb. Explain this.

Realism seems the richer view (also includes unobservables) compared to constructive empiricism, but is not, because it is by definition not under attack by a test. Thus: constructive empiricism is the lamb, fat sheep is realism.

Structural Realism in three principles

Skopological Principle: the aim of science is to find the whole truth concerning the structure of reality.

Semantic-Epistemic Principle: propositions, theories and models about the structure of reality are made true by the structure of reality.

Metaphysical Principle: there exist structures in reality that give rise to everything there is independent of us.

Distinctions of context of scientific praxes and their development according to Reichenbach and Laudan.

Context of discovery: field for the history of science.

Context of justification: field for philosophy of science.

Laudan adds a third context:

Context of pursuit and development: context for both philosophy and history of science.

Two kinds of induction

Enumerative induction: sweeping the data together (if N swans are white, all swans are white).

Colligative induction: contains a concept in the conclusion which is absent from the premises.

Two criticisms of induction

Hume’s Problem of Induction: induction is not an epistemic justification, but one of habit.

Induction is too restrictive, not all knowledge can be derived from particulars.

Two criticism of the Hypothetico-Deductive Method

Hume’s Problem of Induction Redux: no matter how strong a hypothesis, it is always infirmable.

The method is not universal (there exist others such as simulation, parameter fitting, data mining, purely theoretical research).

Deductive falsificationalism

Popper: science finds falsifiable propositions that are not yet falsified.

Three criticisms of Poppers Falsificationalist Theorem

It is too weak, as all kinds of weird statements are now scientific.

Too strong: all kinds of justified propositions are unscientific.

Failing to falsify seems the other side of the logical coin of succeeding to confirm.

Four criticisms on Kuhn’s paradigms

Too big of an exaggeration of social/psychological effects.

After a paradigm shift, not all of the old paradigm is lost.

Is there such a thing as normal science? Scientist critique/question the paradigm all the time.

Incommensurability of paradigms is a myth, interdiciplinary discussions happen all the time.

MetaHistorical Criterion by Lakatos and criticism by Feyernabend

A methodology of science should explain to the largest degree (internally) what happens within the history of science rather than externally.

Feyernabend’s response: everything that happens, according to Hegel, is rational, thus his view would be best. But this is nonsense.

Explication of registering, seeing and observing

Registering: S registers x iff an image of x forms on the retina of S.

Seeing: S sees x iff S registers x and S is aware of x.

Observing: S observes x iff S registers x, S is aware of x and S pays close attention to x.

Explication of observable

Concrete entity x is observable iff S under normal circumstances can see x.

Differences between evidence and proof:

Proof consists of a deductive argument of which the premises are known. Evidence is less strong, and science needs to verify and falsify, infer and confirm.

Explication of logically simple sentence, actual sentence, analytical/conceptual sentence, synthetic, observational sentence.

Also: when a sentence is (potential) empirical evidence.

Logically simple sentence: contains no variables, junctors, quantors or modal operators (atomic sentence, always consistent).

Actual sentence: if it is only about concrete entities that exist.

Analytical sentence: if the sentence is true by means of the meaning of the words in the sentence.

Synthetic sentence: if the sentence is true by other means that the meaning of words in the sentence.

Observational sentence: a sentence is observational iff it is synthetic, logically simple, contains only observable predicates and is only about observable entities.

Empirical evidence: a sentence is empirical evidence iff the sentence is observational, actual, and true.

Possible empirical evidence: a sentence is possible empirical evidence iff the sentence is observational and actual.

Five threats posed by theory-ladenness

Anti-Positivist Threat: logical positivists aimed to reduce science to sensory experience, which collapses when observational sentences cannot be delineated because they are infected by theory.

Solution: theory-ladenness is a necessary condition for science.

Anti-Test Threat: when scientific evidence is not described by observation sentences, and is therefore infested by theory, will the observations not only confirm the theory?

Solution: not an issue, observations pile up.

Blinders Threat: when everything is seen through the lense of the theory, is it possible to observe something outside of the theory?

Remark: human condition, but for science it is relevant: saliance/relevance of observations is determined by theories.

Expectation Bias Threat: when we use concepts of a theory to see, we expect to see what the theory tells us.

Remark: confirmation bias, which is part of the human condition.

Circularity Threat: in order to justify theories, one needs observations, which are justified using empirical biconditionals, which are justified using theories, ad infinitum.

Remark: this is not a problem if the theory for biconditional/hypothesis do not overlap. Or if the empirical biconditionals are analytic truths. Or by Popper’s empirical basis based on consensus between experts.

What are empirical biconditionals?

Empirical biconditionals mark the essence of science: evidence published in observational reports must be logically connected to empirical evidence by means of empirical biconditionals.

Three pillars of Bayesianism

Credence principle: quantitative conception of belief.

Principle of chronic coherence: rational believer at moment in time.

Principle of diachronic coherence: beliefs in different moments are related and their change/development is rational.

What is the Measure Principle, Principle of Chronic Coherence and Principle of Diachronic Coherence in Bayes?

See slides.

Two Problems and solutions of Bayesian Confirmation

Hempel’s Raven Paradox: because of the laws of probability, irrelevant evidence works in favor of the hypothesis.

Solved by adding background knowledge.

Goodman’s New Riddle of Induction: evidence in favor of green or grue and blue or bleen are equivalent.

Solved by adding background knowledge in the prior, but the influence of priors vanishes over time!

What forms an explanation?

Explanans (what is used to do the explaining) and explanandum (what is explained).

What is Deductive-Nomological Explanation?

Finite set of sentences G DN-explains explanandum E iff

1. G deductively implies E.

2. The premises of G are only laws of nature (L), mathematical theorems (M) or particular facts (F).

3. G is consistent (otherwise both E and not E would follow from G.

4. E cannot be explained by M and F alone, we also need L.

5. No premise in G can be deduced from E (no circularity).

6. The deduction that G implies E is slender: it contains no redundancies, you cannot leave one premise out and still have G implying E.

Three problems with Deductive-Nomological Explanation

Redundancy Problem: adding a random statement to G, still implies E (given that G already implies E). This is solved by condition 6 of DN-explanation; that G should be slender.

Tacking Problem: if G DN-explains E, it also DN-explains E and the union of an arbitrary statement. This is solved by requiring that all of E should be deducible from G.

Symmetry Problem: weird things happen when F or E can either be implied by G (flagpole example, where the shadow of the flagpole determines its height, but this is determined by the maker!). We can require F in G to imply E to get rid of the problem, but when this is not possible, the problem remains.

What is Inductive-Nomological Explanation?

Finite set of sentences G IN-explains explanandum E iff

1. The premises of G are only laws of nature (L) of which at least one is probabilistic, mathematical theorems (M) and particular facts (F).

2. G implies that the probability of E given F is close to one.

3. The above conclusion (2) cannot only be deduced from F, because L is needed.

4. G is consistent.

5. All of the premisses in G cannot be deduced from E (no circularity).

6. The deduction in 2 is slender (no redundancies).

One problem of Inductive Nomological Explanation

The explanans must contain all relevant evidence, otherwise adding new information (that is relevant but not yet included) can invert the conclusion. Therefore, Carnap added the requirement that the explanans must contain all relevant evidence. But how to know whether all the relevant evidence is included? Therefore Hempel proposed maximum specificity (not required to know the specifics).

Causal explanation

Finite set of sentences G is a causal explanation of event E iff

1. G explains how the occurrence of n events are jointly sufficient for the occurrence of event E.

2. Not one event can be omitted, otherwise the n events are not jointly sufficient anymore for E.

3. G is consistent.

What is the explanation of: causation, event, causally connected, cause, effect

Causation: a relation between events.

Event: physical objects in a finite space-time region.

Causally connected: Event C is causally connected to event C’ iff the objects in event C physically interact with the objects in event C’.

Cause: event C is a cause of event C’ iff event C is causally connected to event C’ and event C happens earlier than event C’.

Effect: event C’ is an effect of event C iff event C is a cause of event C’.

What is Functional Explanation?

Finite set of sentences G functionally explains event E iff

1. Gamma describes a function of the prime object(s) involved in E and shows how the functioning leads to E.

2. G is consistent.

What is Unification Explanation?

To explain the phenomenon is to unify them.

What is science?

Science is the plurality of all and only scientific praxes (five criteria + CUDOS).

Give the sextuplet of theory T (same for classical (Aristotelian or modern)

Domain: Dom(T)

Concepts: Enn(T) & Fund(T)

Propositions: Def(T), Princ(T) & Thrm(T).

What are the three things a modern theory requires?

Ontology: concerning the entities that Th is about.

Ennology: fundamental and defined concepts of Th, maybe including propositions concerning which concepts correspond to which features of domain members.

Nomology: postulates about how domain members behave and relate.

How to compare theories?

Difference measure based on the difference between union and intersection of two theories (straightforward, use a number to quantifiy the difference based on the cardinal number).

What is a representation model and what are the four types of representation models?

Representation model: M is a model of X iff M represents (features of) X, where X is the object or target of M.

Four types:

Data models: using a data structure to describe specification story S.

Mathematical models: use mathematical entities.

Concrete models: uses stuff in real world to represent.

Simulation models: run a simulation based on mathematical model or theory.

Give three cases of idealizations of models

Aristotelian: Truncation (discard irrelevant information).

Galilean: Distortion (make necessary simplifications of complex abstractions into more simple explications).

Newtonian: Approximation (cut of Taylor series, round off decimals, etc).

Give two interpretations of the relation between models and theories.

Models are mediators between phenomena and theory. Phenomenons are modeled by models which obey the theorem.

Models are building blocks of theories.