Week 1

External world skepticism → Skeptical of whether there is an external world or not.

● Don’t believe there is anything else beyond ‘me’

How can we know that a particular sentence is true?

Analytic sentence → Has truth must be lying within the sentence itself. Without
referring to the world itself. Cannot be false
Synthetic statements → Are true or false based on empirical evidence from the world.

Fallacy → Wrong reasoning

Chat Gtp summary

● What is the analytic-synthetic distinction and why is it important for empiricism? → The
analytic-synthetic distinction is a fundamental concept in empiricist
philosophy, particularly in the logical positivist and logical empiricist
movements. An analytic sentence is one that is true or false based solely on
the meanings of the words within it, independent of how the world actually is.
For example, the statement "All bachelors are unmarried" is considered
analytically true because it is true by definition. On the other hand, a
synthetic sentence is true or false based on both the meanings of the words
and how the world actually is. For instance, the statement "Jupiter has
seventy-nine moons" is synthetic because it is contingent on empirical
observations of the world.
● This distinction is crucial for empiricism as it helps to delineate the nature of knowledge
and the sources of truth. Analytic truths are seen as having no factual content and are
based on linguistic or conceptual relationships, while synthetic truths are contingent on
empirical evidence and observations of the world. Understanding this distinction is
essential for empiricists as it shapes their views on the nature of knowledge, the role of
language, and the relationship between language and the world. It also informs their
approach to scientific inquiry and the validation of knowledge claims based on empirical
evidence.

We add knowledge to the knowledge stock we’re already supplied with.

Analytical reasoning doesn’t bring you extra knowledge

Do analytical claims start out with the world? → Not rooted in behavioral claims.
Derive truth from definitions only (why did we choose these specific definitions)

Empiricism → Believe that all knowledge comes from experiences

● “Analytical claims don’t provide us with knowledge”
The verifiability theory of meaning → A statement is meaningful only if it can be
empirically verified or confirmed through observation or logical analysis. A
statement is meaningful if there exists a way to determine whether it is true or false
based on empirical evidence or logical reasoning.
● E.g. "The temperature of water boils at 100 degrees Celsius at sea level."
● Not: "God exists."
● Bad: Verification conditions for individual sentences unclear

Case study: 2008 financial crisis

● The 2008 financial crisis was a global economic downturn that raised questions about
the effectiveness of economics as an academic discipline. Unlike typical business cycle
crises, the 2008 crisis originated from microeconomic issues in the mortgage market,
where risky loans were irresponsibly sold and bundled into complex financial products.
This led to a domino effect of financial institution failures, global recession, job losses,
and increased poverty.
● What makes this crisis unique is that it required understanding microeconomic details to
explain a macroeconomic phenomenon, contrary to traditional macroeconomic analyses.
Despite the expectation that well-paid economists in key positions could have predicted
the crisis, many argue that economics cannot account for every individual action or error
outside the domain of economic theory. This raises questions about how economists
should test hypotheses against reality when external circumstances are unstable.
● The aftermath prompted a discussion among philosophers of economics: should the
blame lie with the economic models themselves or with the economists who applied
them? This ongoing debate reflects a broader challenge in assessing the role and
limitations of economic theory in understanding and addressing real-world phenomena

Case study: Planets and dark matter

● In the history of astronomy and physics, the discovery of new celestial phenomena and
the development of theories to explain them have often led to significant debates and
controversies. One such example is the case of Neptune and Mercury's orbits, which
challenged established theories of gravity. In the 19th century, the discovery of Neptune
through precise calculations based on deviations in Uranus's orbit validated Newton's
laws of gravity. However, similar deviations in Mercury's orbit prompted the search for a
hypothetical planet, Vulcan, which was never found. Instead, Albert Einstein's theory of
general relativity, proposed in 1916, provided a revolutionary explanation for Mercury's
orbit and reshaped our understanding of gravity.
● In more recent times, astronomers have encountered another puzzling phenomenon: the
anomalous behavior of stars in galaxies' outer regions. This discrepancy, observed by
Vera Rubin in 1976, cannot be explained by known laws of gravity and visible matter.
Instead, astronomers propose the existence of "dark matter," an invisible and exotic form
of matter that affects the motion of stars through its gravitational pull. Despite extensive
searches, dark matter remains undetected, leading some physicists to question whether
modifications to Einstein's theory of gravity might provide alternative explanations for
galactic dynamics.
 ● The debate between proponents of dark matter and modified gravity theories remains
ongoing, reflecting the complexities of understanding the universe's fundamental
properties. Philosophical considerations play a crucial role in guiding scientific inquiry by
addressing questions about the validity of hypotheses, the allocation of research
resources, and the nature of scientific progress. Ultimately, the resolution of such
debates holds the promise of advancing our comprehension of the cosmos and the laws
governing it.

Case study: Evolutionary psychology

● Evolutionary Psychology is a way of studying human behavior that looks at how our
ancestors' environment shaped our minds. It suggests that many traits we have today
helped our ancestors survive and reproduce in ancient times. For example, our sense of
disgust toward things like feces and rotten food likely kept our ancestors from getting
sick. Another example is sexual desire, which helped our ancestors find mates and have
babies.
● Some controversial ideas in evolutionary psychology focus on differences between men
and women in things like sexual behavior. For instance, it suggests that men might be
more interested in casual sex because it offered them reproductive benefits without the
costs of childbearing. Women, on the other hand, might be more attracted to men with
resources like food and shelter because those traits indicated they could provide for a
family.
● Evolutionary psychology has its origins in a field called sociobiology, which studies
animal behavior from an evolutionary perspective. It became more prominent in 1992
with a book by John Tooby and Leda Cosmides. However, evolutionary psychology has
faced criticism for sometimes making speculative claims without enough evidence.
Some argue that it oversimplifies human behavior and ignores how flexible our minds
can be. Despite these controversies, evolutionary psychology continues to be an active
area of research, with ongoing debates about its validity and usefulness

Week 2.1
Hempel (1948) and Good (1967) have given very different responses to the raven problem.
(See p.63-64) Describe these responses and say which you think is more plausible?
● Hempel → Observing a white shoe does confirm the hypothesis that all ravens
are black. He proposed that any observation of an F that is G supports the
generalization "All Fs are G," emphasizing that logically speaking, an "All Fs
are G" statement is not about Fs but about everything in the universe.
According to Hempel, this means that the observation of a white shoe also
confirms the hypothesis that all ravens are green, all aardvarks are blue, and
so on. He was comfortable with this situation, but his view was not widely
accepted.
 ● Good → The relevance of an observation to a hypothesis is not solely
determined by the content of the two statements but also depends on other
assumptions.

Indoor ornithology (knowledge about birds)

The raven problem shows that when two plausible assumptions about confirmation are taken
together, there is a very strange consequence. (Check pages 62-63 for this.) Pondering over
this, Nelson Goodman has noted that there can be a lot of indoor ornithology. What are these
assumptions and what is the strange consequence? How is the latter a consequence of the
former?
● Claim A confirms a universal statement, a logically equivalent of claim A also does this.
● The two assumptions are:
● Observations of black ravens confirm the hypothesis that all ravens are black.
● The hypothesis "All ravens are black" is logically equivalent to "All non-black things are
not ravens."
● The strange consequence is that according to these assumptions, observations of non-
black objects, such as a white shoe, would also confirm the hypothesis that all ravens
are black. This is because if all non-black things are not ravens, then observing a non-
black object supports the idea that all ravens are black. This consequence arises due to
the logical equivalence between the two hypotheses, leading to counterintuitive results
where observations seemingly unrelated to ravens still confirm the hypothesis about
raven coloration

Problem of induction → Why would observations about the past tell us anything
about the future?

Types of arguments:
Logic → Arguments and inferences: What can we conclude, given some set of
premises
● No focus on the specific content of premises
● Focus is on the structural relation between premises and conclusion
● Deductive logic → Premises true → Conclusion guaranteed to be true
(deductively valid arguments)
● Inductive logic → Premises true → Some support for generalization, not widely
accepted as ‘real’ logic, has no accepted standard version.
●

●
 ●

● Conclusion not certain of logic reasoning → Not deductively valid

● Different from projection → Prediction of next case on basis of previous case

Hypothetico-deductivism → Hypotheses are confirmed when their (deductive) logical

consequences are true → test the logical consequences to support hypotheses
● Can observation and evidence be related to each other by using logic?
● Difficulty
○ When using this rule, any observation supports any theory
○ Truth of T entails the truth of T or S. This is because T or S is true if at least one
of its constituents is true
○ The truth of S entails the truth of T or S
○ Observing S supports T by supporting its logical consequence T or S

Raven problem
● Hypothetico deductivism → Observations of black ravens support the
generalization that all ravens are black
● If all ravens are black, then everything that is not black is not a raven
● All observations of non black non ravens also support the generalization

●
● Proposed solutions
○ Whether observations support generalizations depends on context
○ Only those observations count that potentially refute generalization
■ Ravens and non black non ravens satisfy this condition, while black non
ravens do not.
○ These ideas are incompatible with hypothetico deductivism

What is the difference between green/blue things and grue/bleen things? Why are grue and
bleen things problematic for inductive arguments but not for deductive arguments? (Goodman)
● Grue → An object is grue if and only if it was first observed before the year
2050 and is green, or if it was not first observed before 2050 and is blue.
○ E.g. The sky outside on July 1, 2055, will be grue, if it is a clear day.
○ Works perfectly well in deductive arguments.
 ○ A good theory of induction should include a restriction of some kind on the terms
that occur in inductive arguments. “Green” is OK and “grue” is not.
● Bleen → An object is bleen if and only if it was first observed before the year
2050 and is blue, or if it was not first observed before 2050 and is green.
● Green → An object is green if and only if it was first observed before the year
2050 and is grue, or if it was not first observed before 2050 and is bleen.

New riddle of induction

● Nelson Goodman → Formal theory of induction and confirmation is impossible
● Two inductive inferences can have the same structure and both have true premises, but
have conclusions that contradict each other.
● New riddle of induction:
●

● Wrong
○ Word ‘grue’ time relative?
■ Green can also be defined in a time relative way, if we define it in terms of
grue and bleen
■ The property of being ‘grue’ not objective, ‘grue’ not a “natural kind”?
○ The second suggestion is incompatible with a formal “logic” of induction:
“objective” and “natural kind” are loaded, non formal concepts.

Week 2.2
Inductive and confirmation → Based on both inductive and deductive logic

Karl Popper → We never know if a theory is true, but we can know whether a theory
is false. Philosopher of science.
 ● Got PHD in educational science → Knows how children learn
● Falliblist → Certain knowledge about factual matters impossible
● Skeptic about induction and confirmation → There are no good accounts of
induction and confirmation, but science doesn’t need them
● Attentive → We can be sure that a theory is false in case a prediction doesn’t
come out true → Theories can be falsified.

How can holism be used against Propers theory

● Falsification → Discard a theory. Derive from the theory test hypotheses (is)
(hypothesis that is logically derived from the theory. So if theory is true the
test hypothesis is true as well.) In principle it should be possible to falsify the
theory.
● Not per se disproving the theory but mainly disproving the method.
● Holism → Meaning of the test hypothesis is not just the test hypothesis. It is
the whole. Also the method being used.

Asymmetry of confirmation and falsification

● Confirmation
○ If (theory), then (prediction)
○ Prediction correct
○ Theory correct
○ Deductively invalid
● Falsification
○ If (theory), then (prediction)
○ Prediction incorrect
○ Theory incorrect
○ Deductively valid

Falsification
● According to Popper, falsification is central to science
● Popper’s two stage model of the scientific method
○ Conjecture → Creating new theory
○ Attempted refutation → Testing theory in an attempt to falsify it
● Neither induction nor confirmation needed
● For theories: The more falsifiable, the better; otherwise “ad hoc”

Falsifiability criterion
● Problem of demarcation → How to distinguish science from non-science
● Popper’s solution is the criterion of falsifiability → Hypotheses are only
scientific if they are in principle falsifiable, i.e. can be refuted by some
possible observation
● Possible examples of unscientific theories:
○ Conspiracy theories → Evidence against them does not count, also part
of the conspiracy
 ○ Freudian psychoanalysis → Everything can be explained by appealing
to the unconscious.

Problems with falsificationism

● Holism about testing → Apparent evidence against theory can instead be
evidence against the assumptions we make to test the theory → How to
decide?
○ Popper → Scientists should not deflect the blame: We need to decide
whether to accept an observation
● What about theories that predict that certain observations are only unlikely instead of
impossible?
○ Popper → Decide which predictions are so unlikely that they should be
deemed impossible.

Problems with skepticism about confirmation and induction

● No grounds to decide between a theory that has never been tested, and one that has
been tested extensively and never been falsified
○ Why building a bridge according to the latter and not the former theory?
○ Popper → Theory that has been tested extensively but not falsified has
been corroborated
■ Confirmation → Successful predictions are evidence for the truth
of the theory
■ Corroboration → Successful predictions are not evidence for the
truth of the theory success in the past says nothing about the
future
● And if induction is worthless → Why not even use falsified theories? They
might work in the future.

Problems with falsification

● Theories can have scientific and unscientific versions
○ The truth of a scientific theory should make a difference in terms of observations
○ For example, evolutionary theory: difficult to falsify, but certain observations
would show that parts of the theory are wrong or need to be adjusted.

Q: Popper “solves” the problem of confirmation by showing that science does not need the
concept. All that science does and should be doing is to put forward conjectures and to test, i.e.
to try and falsify the conjectures.
● Solves is a big word
● Falsificationism is good, but not sure if it actually ‘solves’
● We mostly agree with Popper's assertion, provided it is understood that science
operates through a process of conjecture and falsification rather than confirmation. This
approach fosters intellectual rigor and prevents complacency within scientific inquiry.

No questions about the following:

Q: Popper's concept of corroboration – as the alternative to confirmation – provides us a
plausible method to choose between rival theories.
● Important: Corroboration is a form of confirmation. Try to reject and don’t
manage to falsify. Meaning there’s still support for the theory. Corroboration
is a special case of confirmation. Can’t falsify a test hypothesis → Support for
the theory → Kind of confirmation. Some test hypotheses are easy to falsify,
and some are not easy to falsify. When they are easy to falsify, they are easy
hypotheses. Confirmed if they are unable to falsify. So corroboration is strong
confirmation.
● Corroboration → Keep track of things that have worked in the past. More
often it has worked the more corroboration.
● Much more complex
● We mostly disagree with the notion that Popper's concept of corroboration provides a
plausible method to choose between rival theories, provided it is understood that
corroboration alone is insufficient for theory selection. While corroboration offers
valuable insights into the resilience of theories, it falls short as a definitive criterion for
theory choice.

Q: Popper uses falsifiability as the criterion of being scientific or not – i.e. of demarcation. This
criterion is useful to show why Darwinian evolutionary theory is scientific, whereas “intelligent
design” is not.
● According to Popper it is true
● Darwin: evolutionary theory
● Intelligent design is not falsifiable (you can't empirically test it) → So not
scientific

Week 3.1
Quizzes

False → Phenomenalism says that perception gives access to an observer-

independent physical reality

Not deductively valid?

If p, then q. Not p. Hence, not q

Hempels view on confirmation

The observation of a non black non raven, confirms the sentence all ravens are black

Inductively valid
All the many emeralds observed, in diverse circumstances, prior to 2050 AD have been green.
Hence, all emeralds are green
Inductively invalid
All the many emeralds….

Universal statements of the form all Fs are Gs are in general falsifiable → True
Universal statements of the …. → False

Popper considers a theory to be scientific → If it can be falsified at least in principle

Poppers perspective on induction

Induction cannot be justified, but that is not a problem because science does not need it

Holism raises problems for popper’s falsificationism?

Because holism entails that, for any theory and any conflicting bit of evidence, it is indeterminate
whether the theory is falsified by the evidence.

Which approach to Uranus and Mercury’s anomalous behavior constitutes normal science?
What about “dark matter” versus “modified gravity?”
● Dark matter would be considered as normal science
● Modified gravity would not be considered as normal science

VIDEO NOG KIJKEN

Week 3.2
True claim according to Kuhn → The fundamentals in normal science are usually not
questioned

Evolution of knowledge occurs in the practice of normal science

Not in revolutionary science → As soon as a new paradigm comes in, it looks as if
you have accumulated knowledge, but Kuhn says this is not the case, you have to
compare two paradigms, but these paradigms can’t be compared.

Incommensurability → Two paradigms explain the same with the same words but
different meanings.

The word ‘species’ had two different meanings for Darwin and Hopkins

Kuhn
● Cycle of scientific change: normal science, crisis, scientific revolution
● Only one active paradigm at a time
● Kuhn interpreted very differently by philosophers influenced by him
 ○ Lakatos and Laudan: Kuhn’s view of science is irrational and personality driven
○ Feyerabend: Kuhn’s view of science is overly constraining and creativity killing
● Pre-paradigm - Paradigm (normal science) - Crisis - Scientific revolution - Paradigm

Lakatos
● Research programs replace paradigms
○ A scientific field can have multiple research programs
○ Competition is essential for progress
○ Two components
■ Hard core of fundamental principles → fixed
■ Protective belt of principles protecting the hard core → changes
only occur here
○ Two kinds:
■ Progressive → Changes lead to wider application (more
predictive power)
■ Degenerative → Changes only cover up existing anomalies
○ Degenerative programmes can recover
● No clear rules for when to abandon a degenerative research program.

Laudan
● Research traditions replace research programmes
○ Looser relation between traditions → Can branch off, be reabsorbed
○ Movement between hard core and protective belt
○ Progress can be measured in terms of problem solving power
○ Two different theoretical contexts
■ Acceptance → Accept the tradition with the most total problem
solving power
■ Pursuit → Pursue the tradition with the most rate of change of
problem solving power
○ Rational to accept one theory while pursuing another.

Feyerabend
● Epistemological anarchism → Anything goes: we should not propose strict
rules for doing science
○ Against Kuhn: Normal science too rigid, narrow minded, dogmatic etc
○ Instead: Science is a creative pursuit → Should be unconstrained and
free
● Important argument → Science often goes against our immediate senses
○ Galileo → If the earth moves around the sun, then why do we not
notice it?
○ To solve this, Galileo needed to why our immediate senses are misleading
○ Serve as a sharp critique of empiricism
● Science is about to overturning worldviews: one should let one’s inclinations go against
reason in any circumstance
Frameworks
● Carnap → A linguistic framework is used to categorize events or organize
things
● Two kinds of moves
○ Within a framework → Change fundamental principles, which are
analytic claims
○ Between frameworks → Assess and test synthetic claims
● Two kinds of scientific change
○ 1 step process: Always within a single framework (Popper, Quine)
○ 2 step process: also between frameworks (Kuhn, Lakatos, Laudan)

Conclusion

Week 4.1
Intelligence → IQ test

Chapter 7 video

Old sociology of science: Robert Merton (1957)

Norms of science:
1. Universalism
2. Communism
3. Disinterestedness
4. Organized skepticism

Reward system
 ● Since, according to the norm of communism, scientific findings belong to everyone,
scientists are not rewarded for their work through the ownership of their ideas.
● Instead, scientist are rewarded by being recognized to have been the first to discover
something or to have published a novel idea or interpretation of something

Strong Programme (Barnes and Bloor, 1970s)

● Symmetry principle → Belief forming in science is no different from belief
forming elsewhere: it is based on local epistemic norms.
● Relativism → It is impossible to prove that the epistemic norms of science are
the best way to learn about the world.

“Manufacture of facts”
● By Latour
● Vague

Two interpretations of the “manufacture of facts”

● Weak “manufacture of facts”
○ What ends up as being considered a scientific fact is the result of a complex
social process and does not only depend on reality
● Strong “manufacture of facts”
○ Facts are created through social processes

Week 4.2
Chapter 8 video

Science is political
● Enlightenment → Science is progressive and anti-authoritarian
● 20th century → Science maintains inequalities
● Today → Science saves the planet to ruins the economy and society
(depending on who you ask)

Feminist empiricism
● Spontaneous feminist empiricism → Wants to address biases in science
without changing traditional scientific methods and norms
 ○ E.g.: Criticize medical studies for using male subjects only.
● Feminist empiricism → Wants to revise traditional ideas about science and
knowledge without giving up empiricism.
○ E.g.: Longino claimed that although objectivity is a goal to strive for, it can only
exist in a community and not in individuals.

Standpoint epistemology
● Standpoint epistemology → Claims that the situatedness of individuals
provides privileged access to particular information
● Science would profit from giving more/special attention to the views of oppressed and
marginalized groups as they can bring something important to the table.
○ E.g.: there are things only women can know (or women can know more easily). If
you want to know how to make Groningen safer, you might want to ask female
students which places they avoid at night.
● Do standpoints matter in every field, even in Logic, Mathematics and Physics?

Feminist postmodernism
● Feminist postmodernism → Claims that objectivity is an illusion and we should
accept that there are multiple equally good views on how the world is like.
○ E.g.: There is no objectively right answer to the outcome of the last US election.
Trump supporters will have a different but equally good view on this than
democrats
● This type of relativism seems to undermine the political goals of feminism. If the
perspective of a macho is as good as that of a feminist, how can we justify fighting for
equal rights?

Science wars
● The relativist idea that science does not have any special authority to knowledge has
been heavily criticized by scientists.
● These debates are called the science wars.
● In 1994, Alan Sokal attempted and succeeded to publish a joke article in the journal
Social Text to show that postmodernists would publish anything with the right
buzzwords.

Week 5.1
Naturalistic philosophy
● Roughly → Philosophy should be continuous with science
○ Use results from science to answer philosophical questions
● Foundationalism → No special assumptions about science should be made
○ Naturalistic philosophy in science is circular
● Is the circle vicious or virtuous?
Kinds of naturalism
● Epistemology naturalized → Quine (1969) suggested psychology would
replace epistemology. What role is left for philosophical questions?
● Normative naturalism → Science gives descriptions, philosophy can make
value judgements of these descriptions
● Example:
○ Psychology → How are beliefs formed in the brain?
○ POS → Are those beliefs reliable, reasonable, etc?

Theory-ladenness of observation
● Hidden theoretical assumptions affect what we observe

●
● Low-level processes in your brain make you think otherwise → Affect
observation
● Fodor → High-level theories don’t affect observation → Awareness of the
illusion doesn’t change the illusion.
● Science tells us how observation operates → understand the illusion

David Hull (1988): Cooperation and competition

● Scientists are rewarded when their work is used → Primary incentive
● Scientists rely heavily on using and trusting the work of others
● To get your work used, it must be reliable → Reinforces trust
● Fraud is especially bad because all users of the bad work suffer.

Division of labor in science

● How to incentivise scientists to divide labor in the most effective way possible
● Kitcher (1990s): Divide reward among all members of successful programme
○ Scientist “hedge bets” by choosing a bigger pie of a smaller field
● Stevens (2003): Divide reward proportional to overall contribution
○ Prevents free-loading
● Zollman (2010): Reduce information flow between scientists
○ Reduces “bandwagon” effects
● Scientific methods are used to shed light on philosophical problems
● Is the circle virtuous?

Naturalistic philosophy → Seeks to understand the world and answer fundamental

questions using empirical evidence and scientific methods, rather than relying on
supernatural or metaphysical explanations. It emphasizes the unity of the natural
world and aims to explain phenomena in terms of natural causes and regularities.
Naturalism → A perspective that sees philosophy as continuous with science. It
rejects the idea that philosophy should be sharply separated from other fields and
argues for a close connection between scientific and philosophical theories.
However, naturalists differ in their views on the nature of this connection.

Quine's version of naturalism → Suggests that epistemological questions should be

absorbed into scientific psychology. He contends that philosophical questions about
evidence and justification are closely tied to questions in psychology and should be
answered by scientific methods. This approach implies that philosophy should
essentially dissolve into science.
● He suggested that epistemological questions—philosophical questions about evidence
and justification—are so closely tied to questions in scientific psychology that
epistemology should not survive as a distinct field at all. Instead, epistemology should be
absorbed into psychology. The only questions asked by epistemologists that have real
importance, in Quine’s view, are questions that can be answered by psychology itself.
Psychology will eventually give us a purely scientific description of how beliefs are
formed and how they change, and we should ask for no more.

Godfrey-Smith version of naturalism → There are philosophical questions distinct

from those addressed by science. This version acknowledges that science can
contribute to answering philosophical questions but does not advocate for replacing
philosophical inquiries with scientific ones. Instead, it views science as a resource
for philosophy, recognizing that philosophy retains its relevance in addressing
normative questions and exploring the relationship between scientific and everyday
views of the world.

While Quine's naturalism suggests that philosophy should be subordinate to science, other
naturalists argue that philosophy should maintain its autonomy and continue to address
questions that transcend the scope of scientific inquiry. They see philosophy as playing a crucial
role in synthesizing scientific knowledge, critiquing scientific theories, and exploring new
possibilities beyond the boundaries of current scientific understanding.

Week 5.2
Common sense realism
● We live in a common reality that exists independently of us.
● We live in a common reality that exists mostly independently of us, with the exceptions
of the parts that we directly affect by our actions.

Scientific realism → Scientific theories describe this common reality

● Our current scientific theories correctly describe reality
● Most of our mature scientific theories approximately describe reality
 ● Our scientific theories try to correctly describe reality.

Arguments for scientific realism

● No miracles argument
○ Why should we believe that science is (to some degree) successful in describing
reality
○ If scientific theories were not (at least approximately) correct about reality (for
example about the existence of a world full of electrons, elements, etc. that exist
independently of us) then the predictive success of our scientific theories would
be a miracle (or a very big coincidence)
● Pessimistic induction from the history of science
○ Why might we be skeptical about scientific theories being successful in
describing reality?
○ Scientific theories in the past have turned out to be false. Therefore, we can
expect our current and future scientific theories to turn out false as well.

Ontology → Discipline. What is this? Ontology in our mind about what we think it is.
● Trust is needed for purchasing power. Something can only be money if it has more than
just the physical appearance.

Social ontology → Research into what does the social consist of

Underdeterminism
● Theories are underdetermined by evidence → No matter how much evidence
we have accumulated, there will always be multiple theories that are
consistent with the evidence
● Why should we believe that our scientific theories are the correct ones?

Empiricists critiques of realism

● Empiricists believe that our sense perceptions are our only source of knowledge
● It is unclear, how we can infer from particular sense perceptions (for example, different
color impressions) that there is a mind independent reality which causes these
perceptions
● So how can it be claimed that a mind independent reality exists?
● Since we have no access to anything beyond our perceptions, it might even be the case
that the only thing we could possibly mean by words like “reality” are patterns in our
perception.

Van Fraassen’s constructive empiricism

● It doesn’t matter, whether scientific theories are true, only whether they are empirically
adequate
● A theory is empirically adequate if it accurately describes the observable parts of reality
● Why should scientific theories only be interested in the observable?
● Where is the boundary between observable and unobservable?
Metaphysical constructivism → Reality depends on us. The world is a human
construction
● If all we mean by “the world” is our phenomenal world (the world as we perceive it), the
world could be thought of as human construction
● Radical opposition to the bad view
○ GS suggests that constructivism might be an expression of the opposition to the
“bad view”, the idea that reality prints itself on the passive mind (compared to the
view that perception is influenced by personal, social and cultural factors.)

Natural kinds
● We use words to distinguish between different types of things (for example between
bananas and apples)
● Do our distinctions reflect real categories with a deeper significance or are they just
arbitrary categories?
● This question has implications for induction (which categories can we use in inductive
inferences?) and for social interactions (are race and gender social constructs or parts of
reality?)

Week 7.1
Chapter 11

Three accounts of explanation

1 covering law
● Covering law explanation, so called deductive nomological (DN) explanation:
○ Explanation is a mere logical relation between explanandum and explanans
○ The explanans need to involve a law of nature (or a well established regularity)
and background conditions
○ Symmetry of explanandum and explanans is a problem. The scheme only tells us
what to expect: can both be an explanation and a prediction.

2 causal
● Explaining by citing causes:
○ Explanation of effects E in terms of the causes C of E
○ Solves (a)symmetry problem: shadow of flagpole caused by length of flagpole
and the sun
○ Difficult questions to crack:
■ What is causation really?
■ What kind of information is important? Surely not all causes of a
phenomenon.
3 by unification
● Unifying explanations
○ Subsume particular facts under a set of general principles and patterns (there
should be as few general principles and patterns as possible.)
○ Theories are more successful the more they explain

Does one account of explanation ‘win’?

● Should philosophy of science propose one single account of explanation?
● There are alternative views:
○ Pluralism:
■ Within all of science, multiple kinds of explanation should be considered
○ Contextualism:
■ Which kind of explanation we should use depends on scientific context
○ Van Fraassen:
■ Explanation is not even an aim of science

What if disturbances mess up the test?

● All explanatory theories - DN, citing causal relationships or by unification - suffer from
the problem that, in a context of justification, we test a particular circumstances.
● Do these test circumstances give an arrangement of the world such that the lawlike,
causal or unificationary relationships actually show up?
○ This is the problem of the ceteris paribus clause
● If the “other circumstances” remain unchanged, the test hypothesis is expected to come
out true in predictions.

What disturbances mess up the test?

● Example:
○ If the theory predicts that a fall in the price of a good will raise demand: … does a
reduction in demand falsify the prediction?
○ The other circumstances may be the preferences of the demanders. Perhaps the
good fell out of fashion explaining both the fall in price and demand? So we need
a clause:
■ Ceteris - the other circumstances
■ Paribus - remain unchanged

The economics case and cp: the financial crisis

● Economic theory of markets assumes (among many other things):
○ That information flows freely, and that agents are instrumentally rational
● That is:
○ The house owners would not accept a mortgage they cannot afford
○ The credit agents would not offer mortgages that incur too high a risk
○ The investment companies would not buy depts at a bad price/risk ratio
○ The expertise rating agencies would not estimate the risks too low
● But they all did precisely this
 ○ Are the assumptions inadequate or can’t we expect the theory to deal with every
detail of behaviors?

Week 7.2

Week 8.1
Confirmation revisited
● Recall from the beginning of the course
● We were looking for a good account of theory confirmation… but we didn’t find any.
● Specific problems
○ Raven paradox
○ Grue/bleen paradox

Bayesian statistics
● About beliefs
● Probility you assign to your own belief
● Prior → Thing you knew/thought beforehand (subjective)

Bayesian epistemology
● Idea: Use probabilities to express gradual confirmation
● This idea has a long history
● Decisive insight (Frank Ramsey, Bruno de Finetti): Probabilities should be construed as
subjective degrees of belief
● Operationalization: Using gambling behavior
● Dutch book argument: Degrees of belief must be probabilities

Hypothesis and evidence

● Assign probabilities to:
○ Hypothesis h: P (h)
○ (New) evidence e: P(e)
○ Conditional probability of e, given h: P(e|h)
● Conditional probability are defined by: P(a|b) = P(a&b) / P(b)
● Note: the expression P(h) is not regarded as well-defined in standard statistics

Confirmation: Bayes style

● Key question: What’s the impact of learning e on one’s degrees of belief in h?
● Answer (“Beyesian conditioning”): Pncw (h) = P(h|e)
● But how do we compute P(h|e)?
 ● Answer: Using “Beyes’ theorem”
● P(h|e) = P(h) * P(e|h) / P(e)

Confirmation solved
● Beyesian account of confirmation:
○ E confirms h if and only if Pncw (h) = P(h|e) > P(h)
● So we seem to have solved the problem of confirmation

At the cost of making it (partly) subjective

● Recall: The original probabilities P(h) and P(e) (“prior probabilities”) are
subjective → There is no right and wrong about them
● That makes also the posterior probabilities Pncw (h) subjective
● Good news: P(e|h) is often (near-) objective
● And it can be shown that probability assignments become similar as more and evidence
comes in if all agents agree on the P(e|h)

Glue/bleen revisited
● Does this solve Goodman’s new riddle of induction?
● Does observing a green/grue emerald confirm that all emeralds are green and/or that
they are all grue?
● Answer: it confirms both, but we can disfavor the hypothesis “All emeralds are grue” in
our priors

The raven revisited

● The Beyesian framework can nicely solve the raven paradox
● Scenario 1: We collect 100 ravens, investigate their color, and find all of them black (this
is our evidence e)
○ Possibly, P(e| All ravens are black) > P(e|Not all ravens are black)
○ So our evidence confirms the hypothesis
● Scenario 2: We collect 100 arbitrary non-black objects, check them for raven-ness and
find that none of them are ravens (evidence f)
○ Ravens are rare, so we are not surprised, hence:
○ P(f|All ravens are black) = P(f|Not all ravens are black)
○ No confirmation in this case