Evidence - Scientific Perspectives

Scientific Perspectives on Evidence Introduction How do scientists use evidence to test theories? The answer is more complex than it might first appear. This chapter examines three interconnected ideas that philosophers and scientists use to understand how evidence relates to scientific knowledge. These concepts help explain why reasonable scientists sometimes disagree about what evidence means, and how we can still maintain confidence in scientific conclusions despite these challenges. Underdetermination: Multiple Theories, Same Evidence Underdetermination is the idea that the same empirical evidence can logically support multiple competing scientific theories. This is one of the most important concepts in philosophy of science, and it directly challenges a common assumption about how science works. Many people imagine science like this: you gather evidence, and the evidence clearly points to one correct theory. In reality, the relationship between evidence and theory is more flexible. Consider a simple example: suppose you observe that the night sky appears to rotate around the Earth over the course of hours and days. This evidence is fully consistent with an old theory (the geocentric model, where Earth is the center) and also with the modern theory (the heliocentric model, where Earth orbits the Sun). The same observations support both theories. Why is this possible? Because evidence alone never fully determines which theory is correct. Multiple different theories can always be constructed to explain the same set of observations. This doesn't mean all theories are equally good—other factors matter, like simplicity, consistency with other knowledge, and predictive power—but it does mean evidence alone cannot eliminate all competing theories. This has an important implication: when scientists choose one theory over another, they're not simply reading the answer off the evidence. They're making a reasoned judgment that involves evidence plus other considerations. Understanding underdetermination helps explain why scientific progress sometimes involves disputes about which theory is better, even when all sides agree on the basic facts. Theory-Ladenness: How Theory Shapes Observation Theory-ladenness refers to the fact that a scientist's existing theoretical commitments and background knowledge shape how they interpret and even perceive observations. This is a crucial insight: observation is not a purely passive, objective process. What we observe depends partly on what we're looking for and what we already believe. Here's a concrete illustration of how this works. Consider two astronomers observing the same astronomical phenomenon. Astronomer A has theoretical commitments based on a certain model of stellar evolution. Astronomer B has different theoretical commitments. When they both look at the same star data, they may literally "see" different things because they interpret the data through different theoretical frameworks. The raw measurements might be identical, but what the data means differs. This applies even to relatively simple observations. If you know that a particular pattern in medical imaging indicates cancer, you'll see the pattern differently than someone without that theoretical knowledge. You might recognize it immediately as dangerous, while someone unfamiliar with the theory might see only an ambiguous mark. Neither observer is being dishonest—their theoretical background genuinely shapes their interpretation. Theory-ladenness also means that scientists don't observe the world with "fresh eyes." We all come to observation with background assumptions. A geologist studying rock formations interprets them through theories of tectonic activity, erosion, and mineral composition. A historian examining an old document interprets it through theories about the period's language, social context, and likely authorship. The theories guide which details seem important and how those details should be understood. An important implication: this doesn't mean observation is useless or that "anything goes" in science. Rather, it means we must recognize that observation involves interpretation, and therefore multiple observers with different theoretical backgrounds might interpret the same evidence differently. This is why reproducibility and peer review matter—they provide checks against any single observer's theoretical biases. Scientific Objectivity: Procedural Safeguards, Not Perfect Neutrality Many people misunderstand what scientific objectivity means. A common misconception is that objectivity means observers have no personal biases or theoretical commitments. That's not realistic—all scientists have biases and theoretical viewpoints. Instead, scientific objectivity is about implementing procedural safeguards designed to minimize the influence of personal and theoretical bias on conclusions. Think of scientific objectivity like the legal system's approach to justice. A single judge's personal preferences might bias their judgment, so the legal system doesn't ask judges to somehow transcend their perspectives. Instead, it uses procedures: adversarial testing of claims, rules of evidence, appeals processes, and multiple judges in higher courts. These procedures are designed to catch and correct for individual biases, even though no procedure eliminates them entirely. Science uses similar strategies: Replication and verification require that important findings be confirmed by other researchers. If one scientist's theoretical bias led them to misinterpret data, other scientists repeating the work may catch the error. This is why "reproducibility" is so fundamental to science. Peer review has specialists in a field critically examine research before it's published. These peers are looking specifically for errors and unwarranted leaps of interpretation. They have their own theoretical commitments, but these typically differ enough from the original researcher's that they catch problems the original researcher might miss. Blinding and controls in experimental design prevent researchers from unconsciously influencing results. In a blinded study, the researcher doesn't know which subjects received which treatment, preventing expectations from subtly influencing measurements or interpretations. Documentation and transparency about methods allow other scientists to scrutinize exactly what was done and where biases might have entered. If procedures are transparent, errors can be identified and corrected. Theoretical diversity within the scientific community means that different groups often hold competing theories. This natural competition creates incentives to carefully evaluate competing interpretations of evidence. If a scientist with Theory A's biases dismisses evidence too readily, scientists with Theory B's biases will investigate it more carefully. The key insight is this: scientific objectivity doesn't require scientists to be unbiased individuals. It requires that we use systematic procedures to detect and correct for bias. This is actually more realistic and more reliable than hoping individual scientists will somehow transcend their theoretical perspective.