Prediction and theory evaluation: the case of light bending

Area of Science:

• Philosophy of Science
• History of Science
• Scientific Methodology

Background:

• The evaluation of scientific theories often hinges on their predictive power.
• Novelty of predictions is frequently cited as a hallmark of robust scientific theories.
• Historical examples, such as Einstein's general relativity, are used to illustrate this principle.

Purpose of the Study:

• To investigate whether successful predictions of new facts provide stronger evidence for a theory compared to predictions of previously known facts.
• To analyze the role of novelty in the evidential weight assigned to scientific predictions.
• To examine historical scientific cases to address these questions.

Main Methods:

• Analysis of historical scientific cases.
• Examination of how scientists evaluate theories based on predictions.
• Comparison of the evidential weight given to novel versus previously known predictions.

Main Results:

• Despite gaining publicity, novel predictions (e.g., gravitational light bending) were often not given more evidential weight than predictions of known phenomena (e.g., Mercury's perihelion advance).
• Scientists frequently use the term 'prediction' for deductions of previously known facts.
• Novelty may hold little importance in theory evaluation and can even detract from a fact's evidential value initially.

Conclusions:

• The evidential value of a scientific fact may not be enhanced by its novelty.
• The importance of novel predictions in scientific theory evaluation is less significant than commonly assumed.
• Further assessment is needed to understand the role of novelty and competing theories in scientific validation.