Falsificationism, Inductivism and Kuhn Notes

(Originally Written November 3, 2006 in Book 8)

Ok, so we are in trouble because of a time crunch. The test is in 1.5 hours. It's time to sum up.

Falsificationism:

I. Naive Falsificationism and its problems

• A hypothesis must be falsifiable in order to be informative.
• Anything that is unfalsifiable can make all situations fit their theory. This makes the theory look good; but, it neither proves anything nor denies anything. It simply explains events according to a theory.
• Since hypotheses must be falsifiable, the more falsifiable a theory is the better it is.
• Problems - 
• "Since hypotheses must be falsifiable, the more falsifiable a theory is the better it is" is very vague
• Naive falsificationism takes too static or too isolated a view on individual theories.
• "A hypothesis must be falsifiable in order to be informative" is in fact, a non-falsifiable statement. How can you falsify it?

II. Falsifiability as a criterion for scientific theories. 

• Science is a set of hypotheses to explain a certain aspect of the world.
• A hypothesis must be falsifiable to state anything informative
• A non-falsifiable hypothesis can look sound but cannot explain anything or deny anything.
• Falsification makes it informative and if science is to be informative it must be falsifiable

III. Ad Hoc Modifications and Popperian (Sophisticated) Falsificationism

• Ad hoc modification - any modification to a theory that has no independent testability
• Relativity of falsifiability - Naive falsificationism claims that a more falsifiable hypothesis is better. Sophisticated falsifiability claims that a falsifiable hypothesis ought to be replaced by a more falsifiable hypothesis
• Aim of science - The aim of science is to falsify hypotheses and replace them with better ones.
• Confirmation is key in sophisticated falsificationism (Chalmers)

IV. Confirmation in the falsificationist model

• Bold conjectures are better than irrelevant truisms
• Bold conjectures that are not falsified lead to novel predictions
• The bolder the conjecture the better
• Refutations help to eliminate falsehoods which in turn reveals more truth

V. Problems with Falsificationism

• Theory dependence of observation: There is no absolute objectivity in science.
• Falsifications are fallible
• Ad Hoc modifications
• Auxiliary assumptions
• Falsificationism is not recommended by the History of science
• The Copernican Theory, for example, would have been rejected before adequate technology could confirm it.

Inductivism:

I. Inductivism and its problems

• Science is inductively derived from the facts
• Science has a factual basis, not a theoretical one
• Observation precedes theory
• Problems
• Facts and experimentation are theory dependent and fallible
• What is the criteria for a good inductive inference?
• Science is not the mere accumulation of facts. Relevant theories and methods must precede observation

II. The problem of induction

• How can knowledge of the unobservable be derived from observable facts?
• Solutions
• Hume - animal faith
• Kant - induction is grounded in a priori, not observation
• Will - Past futures resemble past pasts. Therefore, future futures will resemble past pasts and future pasts. (This begs the question)
• Reichenbach - pragmatic approach: "if anything works, induction will"

The Problem of induction

I. Hume's Fork (Two types of knowledge)

• Relation of ideas
• Matters of fact
• Where does the concepts of uniformity of nature and causality lie? They are neither a relation of ideas nor a matter of fact.
• Hume's solution - Animal faith or a habit 

II. Kant's response - inductive inferences are grounded in a priori concepts

III. Will states that past futures resemble past pasts so future futures will resemble past futures and past pasts.

IV. Reichenbach - (Pragmatic) "If anything works, induction will"

Kuhn's paradigms 

I. Nature of scientific paradigms/incommensurability of paradigms

• Paradigms are relative to their specific field
• Paradigms possess:
• Specific accepted laws and theories
• Specific accepted approaches to experimentation
• Specific accepted metaphysical generalities
• Specific accepted methodologies
• Paradigms are incommensurable because they cannot fit with each other at all
• They are incommensurable because they ask different questions and hold different value systems
• The incommensurability of paradigms leads to the crises and revolutions

II. Paradigm shifts

• Pre-Science - disorganized research, no paradigm
• Normal science (caused by Pre-Science) - puzzle solving activity governed by an accepted paradigm
• Anomalies - unsolved puzzles 
• Crisis (caused by Anomalies) - serious anomalies will cause crisis
• Revolution (caused by crisis) - If the crisis is not resolved revolution will occur and a new paradigm will emerge

III. Criticism of Kuhn

• Ambiguous
• Relativity
• Criteria changes within his method
• Is there progress or is there paradigm shift?