Empiricism and Hypothesis Testing
Empiricism and Logic
Empiricism
Definition: Empiricism asserts that the only source of real knowledge about the world is experience (PGS, 8).
Note: This claim is not exclusive to science, per PGS.
Empiricism and Science
Relationship: Scientific thinking and investigation follow a similar basic pattern to everyday thinking and investigation.
Knowledge Source: Both rely on experience for knowledge.
Science's Advantage: Science is particularly successful as it is organized, systematic, and highly responsive to experience (PGS, 8).
Contemplation: Despite its apparent validity, one questions if empiricism is sufficient on its own.
Empiricism in Practice
Common Narrative: Stories are shared illustrating how observation has overturned established ideas.
Notable Examples:
Semmelweis and Disinfection: Refers to hands sterilization to prevent childbed fever.
Snow and Cholera: Focuses on the cholera outbreak and how the understanding changed through observation.
Critical Perspective: These instances are often framed as observations defeating dogma; however, the assertion that the cholera cause was miasmas from swamps also stemmed from observations, thus complicating the narrative.
Need for Clarity: A more intricate discussion of how observation and hypothesis interact is essential for understanding these cases.
Observation and Theory
Surface Level Validity: Empiricism seems evidently correct when focusing on straightforward observable facts (e.g., apples are red, a meter reads .5 V).
Complex Claims: Challenges arise for claims less directly linked to perception, illustrated through the example of color perception involving cone activity, which is not directly observable and necessitates a theoretical framework.
Logical Positivism / Empiricism
Overview: Logical positivism is a strict version of empiricism, founded primarily on linguistic theory.
Aim of Science: Science, and by extension everyday problem-solving, aims to observe and predict patterns in experience.
Schlick’s View (1932-33, 44): Scientists primarily seek rules that govern the connection among experiences to make rational predictions about future occurrences.
Knowledge Path: Experience remains the only true route to knowledge; any alternative proposed by traditional philosophy leads to non-sensical outcomes (PGS, pp. 29-30).
Inference Logic: Purist epistemology posits that understanding the logic of inference (how observations support generalizations) suffices to comprehend science.
Nature of Science: All scientific inquiry is asserted to be knowledge founded exclusively on empirical observation and logical reasoning.
Validity and Invalidity
Definition of Validity: An argument is logically valid if it is impossible for premises to be true while the conclusion is false, considered invalid otherwise.
Example of a Valid Argument:
If the moon is made of green cheese, then Wittgenstein is a genius.
The moon is made of green cheese.
Therefore, Wittgenstein is a genius.
Conclusion Validity: This argument is valid because if the premises were true, the conclusion must also be true.
Common Argument Forms
Form 1:
If P then Q
P
Therefore, Q
Form 2:
If P then Q
Not Q
Therefore, Not P
Importance: These two argument forms are foundational for logical discourse.
Example of Invalid Argument
Invalid Argument Form:
If Tyler is human, then Tyler is mortal.
Tyler is mortal.
Therefore, Tyler is human.
Reasoning: The premises could be true, but the conclusion could still be false (Tyler could be a squirrel, which is also mortal).
Universal Invalidity Rule: If the form is
If P then Q
Q
Therefore, P,
this structure is always invalid.
Validity and Soundness
Valid Argument Function: Valid arguments enable deriving true conclusions from true premises.
Truth Guarantee: True premises in a valid argument ensure a true conclusion.
Limitation of Non-true Premises: Arguments with false premises yield no insight about the conclusion's truth.
Hempel’s View of Hypothesis Testing
Hypothesis Definition: A hypothesis claims to explain certain phenomena of interest.
First Example: Women in the second ward of the Vienna general hospital die from childbed fever at a statistically higher rate than in other wards.
Possible Extensions: It could also refer to unobservable phenomena, like protons carrying charge +1.
Deductive Testing of Hypotheses
Observational Consequence: Deducing observable outcomes from a hypothesis:
If childbed fever is caused by transferred cadaveric matter, then careful handwashing and disinfection will lower childbed fever rates.
If protons are made of quarks, high-energy electrons will collide with protons in a specific manner.
Empirical Testability: The aforementioned conclusions derived logically from hypotheses will be referred to as test implications.
Test Implications
Test Implication Structure: Standard format is conditional (if-then):
If conditions of kind C occur, then kind E event will happen.
Example: “If medical staff wash hands in chlorinated lime, then childbed fever cases will drop.”
Verifying Test Implications: To confirm a conditional’s truth, conditions C must be validated against expected outcome E, providing a binary result on the implication.
Example: Hypothesis Formulation and Testing
Hypothesis Creation:
Hypothesis: Childbed Fever is instigated by Cadaveric Material.
Test Implication Derivation:
If individuals in the First Division wash their hands with chlorinated lime, mortality from childbed fever will decrease.
Implication Check: Implement the handwashing protocol.
Result Evaluation: If false, the hypothesis is invalidated; if true, it supports the hypothesis but does not guarantee its truth.
Addressing Falsehoods in Hypotheses
Logical Fallacy in Valid Argument: If H, then I; Not I; therefore Not H reflects a valid logical structure, verifying H's falsehood based on I's failure.
Experimental Design Objective: Well-structured observation or experimentation can disprove a hypothesis.
Example: If chlorinated lime washing does not reduce childbed fever, cadaveric matter is not the cause.
Addressing Truth in Hypotheses
Argument Limitation: If H, then I; I; therefore H lacks validity, as true premises do not assure H's truth.
Conclusion from Successful Experimentation: A successful design like Semmelweis’ does not equate to core hypothesis truth.
Confirmation of Hypotheses
Process of Testing: Testing can refute hypotheses but cannot solely confirm them as true.
Fallacy Warning: Valid implications do not sufficiently validate a hypothesis’s truth.
Support Degree: Successful tests increase the support for a hypothesis, termed confirmation, which varies in degree rather than binary status. Observations can provide varying levels of confirmation, but cannot achieve absolute proof.
Confirmation vs. Proof
Proof Limitation: Confirmation should not be confused with proof; demonstrating a successful implication does not affirm a hypothesis's truth based on formal logic flaws.
Support Through Testing: Hempel suggests that confirming test implications increases tendency to trust hypotheses but lacks a straightforward logical explanation for this phenomenon.
Refutations Possible: A hypothesis can be disproved even when it cannot be confirmed as true.
Science and Induction
Ideal Scientific Inquiry Stages (A.B.Wolfe, 1924):
Total observation and documentation of facts are paramount.
Facts must be analyzed and classified with no preconceived notions.
Generalizations are compiled from this analysis on relationships and causation between observations.
Deductive and inductive reasoning follows established generalizations.
Deduction in Scientific Reasoning
Truth Preservation: Deductive reasoning preserves truth meaning; starting with true premises yields true conclusions unless errors are made.
Valid Inferences: Exemplified through the claim that all humans are mortal; Hilbert, being human, must also be mortal.
Limitation of Deduction: Deduction can only derive specific observations from general claims, not the reverse.
Induction in Scientific Reasoning
Inductive Argument Definition: Moves from specific observations to form general principles.
Example: Observing traffic patterns, one concludes the traffic condition will persist daily.
Hempel's Perspective on Scientific Inquiry
Narrow Inductivist Conception: Hempel criticizes the ideal of unassisted observation analysis for being untenable and self-defeating.
Principles of Inquiry:
Data collecting should not be conducted void of initial hypotheses.
Analytical frameworks necessitate hypotheses that guide classification.
Hypothesis construction presupposes knowledge beyond mere mechanical processes.
Creative imagination within scientific inquiry must be tempered by empirical scrutiny for theories and hypotheses.
Hypothesis Testing Methodology
Testing Sequence:
Assert a hypothesis.
Outline test implications.
Validate these implications.
Implication Validation: If negated, the hypothesis fails; if confirmed, support exists but does not guarantee truth.
Challenges with Auxiliary Hypotheses
Auxiliary Assumption Integration: Testing hypotheses often entails assumptions (auxiliary hypotheses) which further complicate validation.
Example Case: Testing implies chlorinated lime negates cadaveric matter, adding layers to confirmation or refutation.
Set Members: When testing, it becomes difficult to determine which hypothesis within a set is confirmed or disproven during trials due to auxiliary elements.
Continued Complexity: Auxiliary assumptions encompass assumptions about the absence of unknown variables influencing results (i.e., controls must strive toward this truth but may not achieve it).
Conclusion on Testability:
Empirical Import Claim: A statement or hypothesis lacking testable implications holds no significant relevance to scientific discourse (Hempel, p. 32).
Final Note: A scientifically sound hypothesis must present some empirical implications that are verifiable in principle; otherwise, they hold no merit in the scientific community.