Philosophy of Science 80-220


Current News: 

The final reading assignment for Monday is posted.

Note that the Achinstein reading questions have been eliminated.


I. Case Study: The Copernican Revolution.
II. Confirmation and Induction:  What justifies conclusions that go beyond the data? Does anything?
III. Theory Structure:  What do theoretical terms mean?
IV. Explanation:  What is an explanation and is explanation an indicator of truth?
Midterm paper assignment.
Final paper assignment.

Basic Information:

Time: Tuesday and Thursday at 1:30-2:50 PM

Room: Scaife Hall 125

Instructor: Kevin T. Kelly

T.A.: Samantha Smee



The stunning success of modern science occasions a number of very basic questions.

What is an explanation?  What is a cause?  What is confirmation?  What is probability? What justifies drawing conclusions beyond the data?   Does science aim at truth? Explanation?  Prediction?  Do scientific theories describe reality or do they merely predict observations?  Is there a scientific method?  Is there a role for subjectivity?  Is there such a thing as objective evidence or justification?

Reasonable (and famous) people have disagreed sharply on each of these questions.  We will survey some standard answers and the arguments for them.  This is not a closed subject, so it lends itself to discussion and exploration.  In order to get you thinking, I will put issues on the table in a loose lecture format.  Also, I will give some lectures on some of the logical and probabilistic background as the need arises.

Level of the Course:

The class carries a 200  number. As such, it must be offered at a level accessible to Sophomores.  Advanced students can pursue their interests in the rather open midterm and final essay assignments. I will also be happy to discuss further issues and details in office hours.

Although the course material is not so difficult, it demands a certain intellectual maturity to maintain a clear grasp on the point at hand.  At the end of the day, you will not be told what the "right" answer is.  You will have to master several conflicting positions and not confuse them.  You will have to understand and construct arguments instead of calculating the answers to exercises.  All of this is typical of any subject, including science, at the advanced level, but most students encounter it first in undergraduate philosophy classes.

Aims of the Course:

    To provide an introduction to some issues and well-known literature in the philosophy of science.
    To provide practice in argumentation.
    To provide practice in structuring vague issues.
    To provide practice in succinct, technical writing.

It is not an aim of the course to march through a fixed range of material.  The articles we will read stand alone and the schedule gives us little time to explore each topic in depth.  If there is support for the idea, we can extend a topic (e.g., confirmation).

General Requirements

Class participation:

This is a discussion course.  I expect enthusiastic and well-informed class discussion.

Careful reading:

Whenever you read a philosophical article, have a notebook at hand and follow this procedure:

  1. Find the author's conclusion.  Write it as simply and crisply as you can.
  2. Reconstruct the author's argument for the conclusion, noting all the basic assumptions.  Be very concrete about this.  List the assumptions, and then derive the conclusion from them by logical reasoning.
  3. If the conclusion doesn't seem to follow, think of assumptions that would make it follow.
  4. Some assumptions may be left unstated because they are obvious or the author doesn't want to call attention to them
  5. Think about whether you agree with the assumptions.  Try to think of cases in which they would be false.
  6. In the end, you should have

                Missing assumptions

I may call on you in class to state an author's argument from your notes.

Written Assignments and Grading

Reading Exercises (33% of the grade)

Simple reading questions are published with each reading assignment on the web.  These exercises give you some official credit for attendance and for your preparation for class discussion.  They also provide excellent practice for the concise writing style I will expect in the required essays (see below).  Finally, the exercises focus attention on the more important portions of the text.

Since they account both for attendance and preparation for the discussion, we will adhere to a strict policy for submitting them.  For full credit you must either

Turning in an exercise in any other way results in an automatic 20% deduction.  Since critical discussion is a major component of the class, the 20% reduction applies also in cases of illness or other legitimate excuses.   The reading exercises are just a way of recording your presence and preparation for the discussion; they are not a substitute.

Having a friend submit your exercise to cut class is a form of cheating.  Don't do that.

Answers must be typed or written legibly.

Keep the answers as short and crisp as possible.  I expect no more than a couple of sentences per question.  Try to say something that convinces me you read the text instead of guessing.

Midterm Paper Assignment (33% of the grade) 

Length:  5 pages plus references.

Essay must be typed, double spaced, 12 pt. Times Roman, with proper references.  See paper writing guide.

Option 1:  Philosophical theories of scientific justification are often judged by their ability to make sense of the history of science.  Analyze the Copernican Revolution (or some other scientific episode you know about)  from the point of view of one or more of the approaches to confirmation and induction discussed in the course.  Argue either that the philosophical theory does a good job of accounting for the case or that it fails to make sense of important aspects of the case.

Option 2:  Bayesianism is a very flexible approach.  Try to show that some of the other confirmation theories we have looked at can be understood to be applications of Bayesian principles.

Option 3:  Should scientists favor simple theories?

Option 4:  Write an expository essay on any other topic that pertains to more than one of the confirmation readings.  Give me a written proposal with references first.

I. The Copernican Revolution

This introductory section of the course has two purposes.  First of all, we will be visiting a time when no distinction was made between philosophy and science.  Philosophy always becomes more important to scientists during times of revolutionary change, when the certitudes of the past come up for re-evaluation.  Second, the Copernican Revolution is a celebrated example of scientific change that will serve to illustrate many ideas in the philosophy of science.  It is widely agreed that philosophical theorizing must be grounded in knowledge of scientific practice.  A third reason for the choice is Kuhn's admirable writing style, which you may wish to emulate.

Some related links:

1. The Two-Sphere Universe.  Kuhn Chap 1 and Technical Appendix section 1:

Reading Questions:

  1. Define north, the solstices, and the equinoxes in terms of gnomon observations.  Draw a picture illustrating the definitions.
  2. How are north, the solstices and the equinoxes defined theoretically in terms of the two-sphere model?
  3. What were the ancient plausibility arguments for the two-sphere model?
  4. What phenomena are logically systematized by the two-sphere model?  How does Kunn distinguish explanation from usefulness for prediction?   (Wake up when words like "explanation" are discussed!)
  5. Why were the more "modern" views of Leucippus, Pontus, and Aristarchus ridiculed and rejected?  (Give both reasons and remember them.  They will be important when we get to confirmation theory).

2. The Problem of the Planets and Aristotle. Kuhn Chap 2, Chap 3:

Reading Questions:

  1. By what observable property were the planetary orbits ordered?  Which planets were left unordered by this principle?
  2. What problem posed by Plato did Eudoxus attempt to solve?
  3. What obvious phenomenon did epicycles account for that Eudoxus' theory did not?
  4. Define "equant", "eccentric", and "major" and "minor epicycle".  Why didn't Copernicus like equants?
  5. Was Ptolemy's theory ever refuted by the data?  What was wrong with it?  (This is a very important question).
  1. How did later astronomers try to fit epicycles into Aristotelian physics?  How did this lead to spurious measurements of the diameter of the universe?
  2. Distinguish Aristotle's sublunary physics from his celestial physics.
  3. Why would admission of infinite space or a vacuum undermine Aristotle's physics?
  4. How did Aristotelian physics foster astrology?
  5. What two advantages did Aristotle's theory have over those of his contemporaries?

Some links:

3. Medieval Developments.  Kuhn, Chap 4.

Reading Questions:

  1. When did Western Europe lose the Greek texts and when did it recover them?
  2. Contrast Hellenic and Hellenistic astronomy.
  3. What was Aquinas' achievement?
  4. What was the moral significance of the Aristotelian cosmos in the 14th c.?
  5. What did Oresme do?
  6. How did Buridan's impetus theory differ from Aristotle's account of projectile motion?
  7. Contrast the assimilation of Ptolemy with that of Aristotle.
  8. What practical concerns heightened the importance of astronomy in the Renaissance?
  9. Contrast the texts recovered in the 15th c. with those recovered in the 12th c.
  10. How did new translations of Proclus affect Renaissance science?

4. Copernicanism.  Kuhn, Chap 5.

This is an important topic.  We want to figure out what was better about Copernicus' system.  The philosophical question will be: does any of that mean that the theory is true?

Reading Questions:

  1. Why did Copernicus find the current Ptolemaic system to be monstrous?
  2. What does Copernicus himself cite as the main advantage of heliocentrism?
  3. Why is it so important to Copernicus that the ocean and atmosphere make Earth into a perfect sphere?
  4. How does Copernicus reply to Aristotle's argument from natural motion to the fixity of the Earth?
  5. How does Copernicus' account of gravity differ from Aristotle's? (Where did he get it from?)
  6. Describe Copernicus' explanation of retrograde motion. Provide a picture.
  7. Describe Copernicus' explanation of the seasons. Provide a picture.
  8. What is the problem of stellar parallax?
  9. How did Copernicus objectively order the orbits of the planets?
  10. Was Copernicus' system more accurate or simpler than Ptolemy's?

Some links:

  1. My notes on the Copernican Revolution.
  2. Copernicus' Preface to De Revolutionibus .

5. Reception.  Kuhn, Chap. 6.

This is another important topic.  We want to see how scientists received the theory after it became known.  Try to figure out whether the reception says anything about the unstated methods of the scientists involved.  Think about simplistic proposals one hears all the time about scientific method.  Was one of the theories refuted by data?  Was there a crucial experiment?  Did one theory have more positive instances than the other?  Did astronomers critically suspend judgment until one side was proved right?

Reading questions

  1. What was the significance of the Prutenic tables for the reception of Copernicus' work by astronomers?
  2. Contrast Copernicus' early reception by astronomers and non-astronomers.
  3. Did Copernicanism initially fare better among Catholics or Protestants, and why?
  4. What accounts for the changing Catholic attitude toward attacks on the Aristotelian cosmos in the early 17th c.
  5. During which period did the tide turn in favor of Copernicanism?
  6. What was the Tychonic system and what were its advantages?
  7. Why were Tycho's parallax measurements on comets and novas so important?
  8. State Kepler's three laws of planetary motion.  Where did the first law come from?
  9. Why was it so important that stars seem to shrink when viewed through a telescope?
  10. What other telescopic observations supported Copernicanism?

6. The New Universe

Reading questions:

  1. What is Kuhn's analogy between Bohr's atom and the moving Earth?
  2. What three cosmological ideas were revived by the success of Copernicanism?
  3. Why did Copernicanism open the door to Descartes' law of rectilinear inertial motion?
  4. Why did Kepler's universe call out for physical explainations of celestial motion in a way that Copernicus' universe did not?
  5. Describe Kepler's "anima motrix" theory.
  6. Describe Borelli's theory.
  7. What was Newton's achievement?
  8. Why was gravity viewed with skepticism in the 17th century?
  9. How much time elapsed before Newton's approach was victorious over Descartes'?
  10. How did Newtonianism affect the political philosophy of the 18th c?


II. Induction and Confirmation

This section of the course will be quite different from what came before.  Instead of looking at how science actually works, we will consider a few ideas about how it should work.  Whatever else one might say about them, the papers we will be looking at are very widely known in the philosophy of science.

7:  Carl Hempel: Studies in the Logic of Confirmation

This paper shaped a whole approach to the philosophy of science.  Don't be put off by all the technical-sounding stuff.  The motivation and principles he discusses along the way have been very influential in the philosophy of science, so it is worth looking at.  Stay awake and look for presuppositions you might challenge.  The reading questions will help you to focus on some relevant points.  They will be more critical in character than before.

I took a course with Hempel in 1981 and drove him to the airport when he left Pittsburgh for Princeton after his retirement.  He was a very congenial gentleman and an animated lecturer.

Some reading  notes:

  1. Carl Hempel was a close associate of the logical positivist's " Vienna Circle ".  The basic idea of the logical positivist movement was to use logic as a tool to model scientific method the way it had already been used to analyze the foundations of mathematics by Gottlob Frege and Bertrand Russell.  Thus, Hempel expects applause when he repeatedly draws an analogy between scientific confirmation on the on the one hand and logical proof on the other.  That is also his excuse for gratuitous technicality throughout the paper.
  2. When Hempel says "A entails B" he means "if A then B".
  3. "(x)(Red(x))" abbreviates "for all x, x is red".
  4. "(Ex)(Red(x))" abbreviates "there exists an x such that x is red".
  5. "All ravens are black" is translated into logic as "(x)(Raven(x) ---> Black(x))", or "for all x, if x is a raven then x is black".
  6. Don't worry too much about sections 5.11 and 5.12 on pp. 266-267.  Pick up again at 5.2.
  7. Read footnote 24.
  8. Hempel's proposed theory is a mess.  Try to derive some absurd consequences from it.  For example, check that a billion black ravens count for nothing if one raven crosses the sun so you can't take down its color, whereas one black raven counts as confirmation.
  9. Footnote 27 pertains to question 7 below.

Reading Questions

  1. What does Hempel want to provide a theory of?
  2. What is "inductivism" and how does Hempel criticize it?  Do you agree?
  3. Why does Hempel think problem (A) is prior to problem (B)?  Can you think of an objection to that conclusion?  Try to think of an analogous concept where the qualitative version is slippery but the comparative concept is not.
  4. Hempel considers whether his project rests on which false assumption?  How does he respond?
  5. What is Nicod's criterion of confirmation and what are Hempel's two objections?
  6. What are the "paradoxes of confirmation"? (These are famous).
  7. What is Hempel's solution to the paradoxes?  Do you agree?
  8. State the "special consequence condition".  Does Ptolemy's theory provide an intuitive counterexample?
  9. Why can't you adopt the special consequence condition, the converse consequence condition, and the entailment condition all at once?  Give the argument.
  10. What is "material adequacy"?

8:  Two articles by Rudolf Carnap:

1. Statistical and Inductive Probability

2. On Inductive Logic.

Here is a more formidable idea.  Carnap still thinks of confirmation as a logical relation between hypothesis and evidence.  Carnap's idea is that a valid deductive argument completely supports its conclusion, whereas in science, the evidence incompletely supports the conclusion to some degree.  This degree of incomplete support is called the logical probability of the hypothesis given the data.  Like many earlier writers, including the physicists Bernoulli and Laplace and the economist John Meynard Keynes, Carnap thinks of degrees of support as probabilities.  The guiding analogy is:

deductive logic is to complete support as logical probability is to incomplete support.

Perhaps the first recorded exponent of the idea was the ancient skeptic Carneides, who held that complete certainty is not justified but that degrees of certainty are.  All of these proposals have the property that one can talk of the probability of h given e rather than somebody's personal degrees of belief in h given e.  On such a theory, the degree of support of h given e is objective; subjective differences being a matter of different evidential histories and of taste in choosing a particular method of assigning degrees of support or "confirmation"..

Carnap was a leader in the positivist's Vienna Circle , developing many of its main ideas.  He felt that the new mathematical logic of Frege and Russell gave him an advantage over earlier advocates of degrees of empirical support like Laplace and Keynes. He actually picked up the idea of logical probability from a pre-mystical Wittgenstein, during the latter's tumultuous visit to the Vienna Circle .

Reading hints:

You shouldn't have much trouble with the first article.  The article "On Inductive Logic", is needlessly difficult to understand.  The concepts are actually quite simple, so I'll explain them myself. My hints should more or less save you the work of  reading the article up to section 9.  If you aren't used to mathematical definitions, don't panic.  The terms introduced below mean nothing more or less than what the definitions say, so no background is required.

Reading hints.

Reading Questions on first article:

  1. How do inductive probabilities differ from statistical probabilities (e.g., what are they attached to)?
  2. What is the principle of indifference and how did Reichenbach and Von Mises object?  How does Carnap object?
  3. What are prior probabilities and what are methods I and II for assigning them?
  4. How does Carnap propose to avoid the paradoxes of indifference?
  5. Why does Carnap choose method II over method I?

Reading Questions on second article:

  1. To make sure you get the idea:  list all the isomorphism classes of state descriptions for the language with two predicates P(_), Q(_), and two individuals a, b (easiest on a word-processor:  cut and paste a lot and add - signs where required).  Next to each state description, write the value of m* for that state description.  Use the table to calculate c*(P(a), Q(b)).
  2. What is a direct inference?
  3. What is an inverse inference?
  4. What "astonishing" consequence does Carnap's theory have for the confirmation of universal laws?  How does he try to address this consequence?  What do you think about that?
  5. What kind of "justification" of induction does Carnap approve of?


Goodman: The New Riddle of  Induction and

Savage: Implications of Personal Probability

Nelson Goodman's short article is among the most famous in the philosophy of science.  It represents a direct attack on the very idea of Carnap's "logical" account of induction.

After Goodman's article, philosophers of science lowered their ambitions from finding an objective logic of empirical support to merely identifying some "rational" constraints on changes in admittedly subjective degrees of belief.

The view that probabilities on hypotheses are just somebody or other's willingness to bet on the proposition is now called Bayesian methodology, after the Reverend Thomas Bayes.  The idea that there is no justification for belief but, nonetheless, we are psychologically wired to become more confident in light of increasing evidence goes back at least to David Hume in the 18th century.  The view was revived in this century by the philosopher/mathematician Frank Ramsey.  J. M. Keynes, the famous economist who first proposed the logical interpretation of probability pursued by Carnap, was completely converted to the Bayesian or "personalist" position by Ramsey.

L. J. Savage, a Bayesian statistician, was a major figure in laying the foundations of Bayesian statistics, the official view of the Carnegie Mellon statistics department.

I have included some notes on Bayesian methodology that complement Savage's more foundational discussion.  Please study them. They contain ideas that may be useful for your midterm paper.

Reading Questions:

  1. Why is it important to distinguish lawlike from accidental statements?  How does this point relate to Carnap's theory?
  2. What is the point of the blue/green vs. grue/bleen example?
  3. What is the starting point for the theory of personal probability?
  4. What is an example of an objective, normative constraint on personal degrees of belief?
  5. What is Savage's model of the objectivity of scientific knowledge?
  6. What is the "paradox" of the objectivity of classical statistics?
  7. What is Savage's view of logical or "necessary" theories of probability like Carnap's?
  8. How does Savage respond to the objection that a theory of gambling preferences is too mundane to serve as a foundation for science?
  9. How does Savage respond to Carnap's view that there are two kinds of probability?
  10. Why does personalism leave one with no foundation for one's current beliefs?




Gilbert Harman:  The Inference to the Best Explanation.

Clark Glymour:  Relevant Evidence

Harman defends the "hypothetico-deductive" method, which is to to select the theory that best accounts for the data.

Clark Glymour is a professor in our philosophy department.  So if you have any complaints about his theory, go tell him! This article was very influential.  Unlike the preceding papers, Glymour's emphasizes the importance for confirmation of "unification" or "harmony" of the sort we saw in Copernicus' theory.  For this reason, Glymour's theory has lots of applications in real science and my be of use to you in composing your midterm paper.  Glymour claimed that his theory is not Bayesian.  But a Bayesian, Roger Rosencranz, claimed that Glymour's ideas follow from Bayesianism.  While you ponder this, go back and look at my notes on Bayesianism, under the heading of "unification".

The idea is simpler than Glymour makes it sound.  His idea is just that the hypotheses that are confirmed in a theory are the ones that can be "cross-checked" by using the data and other hypotheses in the theory to compute values for their theoretical quantities in different ways that might possibly agree.  So for example, the observation of two planets tests Kepler's third law relative to the assumptions of Copernican astronomy, but the observation of one planet does not, since Kepler's law relates the radii and velocities of different planets.  If the data could not have turned out in such a way that the hypothesis is refuted given the rest of the theory (i.e., if the hypothesis is not "at risk" from the data, given the rest of the theory), then even if the theory is consistent with the data, the hypothesis is not confirmed by the data.  Even though two theories make the same predictions with equal accuracy, the hypotheses in one theory may be cross-checked against one another better than the hypotheses of its opponent.

That sounds good, doesn't it?  Here's a question for you to ponder.  If both theories make exactly the same predictions, how could relying on "internal cross checking" possibly allow you to determine which is true unless God told you in advance that he would produce phenomena in a unified or internally cross-checkable way?

If you find the article rough sledding, make sure to look at the quote of Weyl on p. 334, which presents the idea very simply.  Don't worry about the details of the generalization to logical theories under heading III.

There is typo on p. 333:  the equation should be

X(f1(E1 ... Ek), ...,  fi(E1 ... Ek)) = 0.

Reading Questions:

  1. What is "enumerative" induction and how does Harman argue that hypothetico-deductivism is more fundamental?
  2. What feature of knowledge does the hypothetico-deductive approach explain that enumerative induction cannot explain?  It could have been said more clearly.
  3. Why does Glymour think the hypothetico-deductive account failed?
  4. What is Glymour's objection to subjective Bayesianism?
  5. How do scientists compute values of theoretical quantities?  Use your knowledge of Copernicus' determination of the diameters of the superior planets to illustrate this point.
  6. Why is hypothesis (1) not confirmed by an observed value of A1?
  7. Why is variety of evidence good, and how does Glymour define "variety" with respect to a theory?
  8. How does Thirring's theory illustrate the point that accurate prediction does not suffice for confirmation?
  9. How does Glymour characterize theoretical simplicity and what does it have to do with confirmation?
  10. What is holism?  What part of holism does Glymour agree with and what part does he disagree with?



Imre Lakatos:  The Role of Crucial Experiments in Science

Thomas Kuhn:  Objectivity, Value Judgment, and Theory Choice

Lakatos was concerned that philosophical pronouncements and ideals are unduly harsh on the historical practice of the best science, giving rise to needless sceptical doubts about the justification of science in general.  He was particularly critical of the view that genuine scientists must seek to refute theories with experiments and are unscientific unless they specify the conditions under which the theory would be rejected and then carry through when these conditions are met.  Lakatos was unhappy, however, with the idea that there are no general standards governing scientific rationality.  His approach was to replace what he viewed as naive proposals with a more sophisticated one that does not require scientists do drop theories as soon as they get into trouble.  I will leave it to you to figure out what his proposal is.  Lakatos is very critical of Popper and of vaguely described "inductivists".  But what might a sophisticated personal probabilist say about his proposal?  This could be the basis of a nice midterm paper.

Kuhn agrees that rationality never forces us to drop a theory when it is refuted.  But unlike Lakatos, he does not think any general rational standards dictate when the theory should be rejected.  Instead, he accounts for scientific change in terms of the particular constellation of values of particular scientific communities.  Since methodological standards are community-relative, choices between scientific approaches are more like political revolutions than like logic.  One side survives and rewrites the textbooks and histories of science to make it look like logic forced the result.  You can see how Kuhn came up with this view after writing our text on the Copernican Revolution.  Recall that some astronomers emphasized the unified explanations of Copernicanism and others emphasized the incoherence of Copernicanism with any known approach to physics. 

Reading questions:

  1. What is a "demarcation criterion"?
  2. What question left unanswered by Popper does Lakatos claim to answer?
  3. According to Lakatos, how is Popper's proposal refuted?  Be succinct.
  4. Are theories rejected when they yield contradictions?
  5. What is Lakatos' amended conception of falsificationism?
  6. When should a research program be rejected?
  7. Why does Kuhn think that rational disagreements in theory choice are unavoidable?
  8. What is the "discovery/justification" distinction and how does Kuhn respond to it?
  9. Why is subjectivity essential to scientific progress?
  10. Why is the assumption of shared data and communication problematic for Kuhn?


C. Howson and P. Urbach:  Scientific Reasoning, The Bayesian Approach, Ch. 4

In this chapter, Howson and Urbach argue that the basic historical points urged by Lakatos agree with Bayesian recommendations, after all.  One merely has to see the full implications of the concept of conditional probability.  See if you agree.


  1. How do the authors account for the confirmation of a theory by observing one of its consequences?
  2. How do the authors solve the ravens paradox?
  3. How do the authors account for the survival of a research program when an auxiliary hypothesis is refuted?
  4. How do the authors account for “ad hocness”?

C. Howson and P. Urbach:  Scientific Reasoning, The Bayesian Approach, Ch. 11

In this chapter, Howson and Urbach argue respond to critics. 


  1. Do Bayesians favor weak hypotheses?
  2. How can old evidence confirm a theory now?
  3. What is the “Principal Principle and what is Miller’s objection it?”
  4. How do Howson and Urbach explain the role of simplicity?


For fairness, here are Glymour’s criticisms of Bayesianism in “Why I am Not a Bayesian”, a chapter of his book Theory and Evidence.


Forster and Sober:  “How to Tell when Simpler, More Unified, or Less Ad Hoc Theories will Provide More Accurate Predictions”

Here is a non-Bayesian explanation of simplicity in terms of minimization of risk.  Risk is expected distance from the truth of an estimate from a random sample.  The idea is that risk can be analyzed into two components, the distance of the average value of the estimate from the truth (bias) and the spread or probable error (variance) of the estimate around its average value.  Increasing the number of free parameters in the model used for estimation purposes decreases bias but increases variance even more, so that overall risk can be reduced by estimating by means of a simpler theory.


  1. What is the universal reaction of philosophers to the curve fitting problem?  Would Bayesians agree?
  2. How do the authors define distance of a family of curves from the true curve?
  3. What is “overfitting”?
  4. What is Akaike’s theorem?
  5. How do the authors respond to Kuhn’s account of the Copernican revolution?
  6. How do they respond to the Bayesian account?

K. Kelly:  “How Simplicity Helps You Find the Truth Without Pointing at it”

This paper attempts to explain the role of simplicity in scientific inference in terms of efficiency of convergence to the truth, where efficiency is understood in terms of minimizing worst-case reversals of opinion prior to convergence. 


  1. What is Kelly’s objection to the explanation by Forster and Sober?
  2. What is Kelly’s objection to the Bayesian explanation?
  3. What is the freeway metaphor?
  4.  What is Kelly’s argument for Ockham’s razor?
  5. Why is the debate over scientific realism not a genuine debate?

IV.  Realism

The correctness of a theory depends, to be sure, on how the world is.  But it also depends on how the theory is intended and understood.  For example, James Clerk Maxwell constructed a system of differential equations that predicted all known macroscopic electromagnetic phenomena.  The equations were discovered in terms of mechanical models of vortices in the ether, using such mechanically literal constructions as "idler wheels" to keep the ether vortices from scraping each other when they rotate in the same sense.  One could choose to take all of this quite literally, as Maxwell seems to have at the outset.  Heinrich Hertz, who confirmed the theory's astounding prediction of exlectromagnetic waves,
remonstrated  that Maxwell's theory is Maxwell's equations.  For him, the mechanical metaphors might be suggestive but are speculative dead-weight better dropped from the completed theory.  The same dispute occurred in the early atomic theory.  Some chemists spoke of "atomic weights".  Others, like Humphrey Davy, insisted on the more guarded terminology of "combining weights".  There has always been a tension between those who wish to interpret theoretical structures literally and those who view them as crutches or shorthand notations for more basic evidential relations.

Semantic realists think that theories should be understood literally.  Scientific success is then literal truth.  If the theory refers to quarks, then it is not correct unless quarks really exist and have the properties the theory says they have.

Semantic anti-realists think that theories involving unobservable entities should not be taken literally.  There are several ways not to take a theory literally.

empiricism:  all meaningful discourse must somehow be definable or otherwise reducible to discourse about "observables".

phenomenalism:  obsevables = direct sensory impressions (i.e., colors and tastes).
operationism:  observables = concrete, macroscopic laboratory experiments.

instrumentalism:  a theory is to be evaluated only as a calculator for computing predictions in various specific circumstances.

Another distinction concerns justification.  One may be pessimistic or optimistic about our ability to justify scientific claims.  Usually, higher demands on interpretation (semantic realism) give rise to concerns about justification (skepticism) and weaker demands (semantic anti-realism) lead to optimism.  The possible positions are as follows:

Scientific realism = optimism + semantic realism.
Scientific anti-realism  = optimism + semantic anti-realism.
Theoretical skepticism = pessimism + semantic realism.
Inductive skepticism = pessimism + semantic anti-realism.

Philosophers of science don't like skepticism very much, since they usually assume at the outset that science is great.  Therefore, the realism debate is usually understood to be a debate between scientific realists and scientific anti-realists.  The realism debate is perhaps the main debate in the philosophy of science.  This is no doubt because it is also a perennial debate in physics.

As usual, most of the papers in our textbook are familiar classics that every philosopher of science will have read (and disagreed with).


Rudolf Carnap: Theories as Partially Interpreted Formal Systems

Carl Hempel:  A Logical Appraisal of Operationism

As you may recall, Rudolf Carnap had a theory of "inductive logic" to justify empirical generalizations.  Therefore, Carnap was not an inductive skeptic.  His inductive logic cannot justify conclusions involving theoretical terms like "atom" that do not occur in the data.  He avoided theoretical skepticism by adopting a kind of empiricism, according to which a theory is an uniterpreted formal system for deriving conclusions, together with a set of conventionally true semantical rules that interpret the formal system in terms of concrete observations.  Without such an interpretation, a theory is meaningless.  And with such an interpretation, each concept of the theory is logically tied to some complicated combination of observable properties.

Hempel provides an empiricist critique of physicist Percy Bridgman's operational.  Hempel's version of empiricism, which follows Carnap's, is less restrictive than Bridgman's.  Nonetheless, Bridgman's views are very close in the philosophical spectrum to those of Hempel and Carnap.  Hempel's views are less rigidly reductionistic than Carnap's.  Hempel's views represent the last, most liberal phase of the logical positivist movement.

Reading tips:
By "calclulus" Carnap means simply "formal axiomatic system".
The Y Carnap talks about is the "wave function" in quantum mechanics.  The waves are not waves "in" anything.  They are distributed through space and determine a probability distribution on measured values of physical variables whenever we happen to make measurements.  Until we measure, there is no fact of the matter about the values of all these measurements.  Since the wave function produces physical measurements by this curious procedure, it is not itself the familiar classical state characterized by these measurements.  Aside from producing the measurements, its nature is fundamentally unknowable.

Reading Questions:

  1. According to Carnap, how does mathematics differ from physics?
  2. How do "abstract" terms receive their interpretation?
  3. What are the two different approaches to constructing a calculus?
  4. What is the advantage of formulating a theory as a formal system?
  5. Is it a scandal that quantum physicists still cannot say, intuitively, what Y is?
  6. What is the principal difference between logical empiricism and operationism?
  7. What is the problem posed by "disposition terms"?  Be sure to look at the relevant footnotes.
  8. What is the problem posed by real-valued quantitative concepts?
  9. What is "implicit definition"?
  10. How does Hempel's "interpretative system" idea weaken Carnap's proposal?


Grover Maxwell:  The Ontological Status of Theoretical Entities

Ian Hacking:  Do We See Through a Microscope?

Carnap's system is founded on a distinction between observable and abstract vocabulary.  Maxwell argues that there is no such distinction, because observability is actually a spectrum of notions.  This is called a "slippery slope" argument.  Hacking counters that some microscopes (those working on diffraction principles) are not really similar to ordinary vision.  Nonetheless, we see with them when we can manipulate nature under them.  So Carnap's distinction is still overturned.

Reading tip:
"Ontology" is the philosophical study of what "really exists".  Maxwell attacks the view that obsevable things "really exist" and "theoretical entities" do not.
"Diffraction" occurs when the troughs and crests of light rays add or cancel to make fringes or colors (like when you look at the grooves on the back of a CD).  .

Reading Questions:

  1. What is the "instrumentalist" view?  How does "Pelter's" view differ?
  2. How does Maxwell argue that "unovservabililty in principle" is a self-defeating concept for defenders of anti-realist views?
  3. What position does Maxwell attack with the "men over 14 ft. tall" example?
  4. What kinds of entities does Maxwell take "sense data" to be?
  5. Why does Van Fraassen conclude that we can see through a telescope but not through a microscope?
  6. How do we learn to see under a microcope?
  7. According to Hacking, do we need theory to see through an interference microscope?
  8. How does Abbe's theory of microscopic vision relate to Maxwell's position?
  9. How does one determine that "dark bodies" are not an artifact of electron microscopy?  What is the form of the argument?
  10. What is the grid argument against anti-realism?



Bas Van Fraassen:  To Save the Phenomena

N.b.,  The "Craigian reduction" of a theory is a theory that entails only the observational consequences of the given theory, eliminating all reference to hidden entities and processes. 

Reading Questions:

  1. What is scientific realism, according to Van F.?
  2. What is empirical adequacy and how does it compare with truth?
  3. What is acceptance and how does it differ from belief?
  4. What distinguishes the observable from the non-observable?
  5. How does Van Fraassen respond to the argument that some theories are easier to extend to new phenomena?


Paul Churchland:  The Anti-Realist Epistemology of Van Fraassen's The Scientific Image

Reading Questions:

  1. How does Hume's problem differ from the problem of underdetermination?
  2. Why is the anti-realist gullible?
  3. How does Churchland's view of the role of simplicity differ from VanFraassen's?
  4. What is the point of the thought experiment involving the implanted transducers?
  5. What is Churchland's response to Stich?  What do you think of it?

V.  Explanation


Covering Law and Statistical Relevance Accounts

Carl Hempel and Paul Oppeneim:  Studies in the Logic of Explanation (skip section 4).

Wesley Salmon: Statisical Explanation and its Models

Hempel tried to work out the obvious "covering law" model of explanation, which is based on the idea of exercises in a physics book.  You explain a phenomenon when you "derive" it from "initial conditions" and the laws presented in the book.  The trouble is, Hempel couldn't define what a "law" is supposed to be, except by pointing.  Hempel also extended his view to statistical explanations.  A statistical explanation is supposed to confer a high probability on the explained fact.  This is analogous to the covering law model.

Wesley Salmon proposes, instead, that a statistical explanation consists of a listing of all the statistically relevant factors, even if they lower the probability of the explained fact.  Relevance, rather than high probability is the criterion.  Thus, for Hempel, birth control pills explain why men taking them don't get pregnant; for Salmon, they don't.  Salmon seems to have the problem that barometers are relevant to thunderstorms.  He gets out of this, however, by saying that the atmospheric pressure renders barometer readings irrelevant after all.

Think about what this says, from a Bayesian perspective, about quality of explanation being an indicator of truth.

Typo:  condition 2 on p. 172 should read: P(B|A.C) is not equal to P(B| A.-C).

Reading Questions:

  1. What question does an explanation answer?
  2. What are the conditions for  adequate explanation?
  3. How does Hempel claim that explanation relates to prediction?  Do you agree?  (Consider this example.  The light is off. What is the predicted state of the switch?  Off. Why is the switch off?  Because the light is off.  Does this fit with Hempel's first paragraph?)
  4. What problem does Hempel encounter when he tries to characterize "law-like" statements?  What does that say about his account of explanation?
  5. Why do statistical explanations require maximum specificity?  Give an example.
  6. What is the point of the Linus Pauling and dissolving force examples?
  7. What is the point of the paresis example?
  8. Given an example of a bad explanation that violates condition 5.
  9. Given an example of a bad explanation that violates condition 7.
  10. What is "screening off"?  Illustrate it with the barometer example.

21:  Causal Accounts

Even after all the laws and statistically relevant factors are isolated, it seems that causes explain and effects do not.

Nancy Cartwright:  The Truth Doesn't Explain Much

B. A. Brody:  Towards an Aristotelian Theory of Scientific Explanation

  1. What is Cartwright's complaint about covering law models of explanation?
  2. How does a covering law theorist respond?
  3. What are Cartwright's two replies?
  4. What is the point of the camelia example?
  5. What was Aristotle's account of explanation?
  6. What is the logical empiricist acount of causation and what is wrong with it?
  7. What is Bromberger's example?
  8. How does Brody deal with the sodium example?
  9. How does Brody respond to Duhem's skeptical argument?
  10. How does Brody respond to Popper?

B. Van Fraassen: The Pragmatics of Explanation?

Due in class on Dec 6.

The VanFraassen piece represents an anti-realist attempt to counter realist claims that explanation justifies belief in theoretical entitites.  VanFraassen wishes to argue that explanation is a purely pragmatic virtue that is independent of grounds for accepting the theory.  As such, it cannot justify underdetermined theories over their observable consequences.

Reading Questions:

  1. What are the false ideals of explanation theory according to Van Fraassen?
  2. What is Van Fraassen's objection to Salmon's theory?
  3. What are the two prejudices that stand in the way of progress in understanding explanation?
  4. What is the logical problem raised by explanatory asymmetries?
  5. How can contextual relevance reverse Bromberger's example?

Related reading:

P. Achinstein: Can there Be a Model of Explanation?

Final Paper Assignment (33% of the grade) 

Length:  5 pages plus references.

Essay must be typed, double spaced, 10 pt. Times Roman, with proper references.  See paper writing guide.

Topic suggestions: 

Relate confirmation and/or explanation to the realism/anti-realism debate.

Relate any topic in the class (realism, confirmation, explanation) to a real episode in scientific history.