Before we talk about what science has found, we should know what it is trying to do. The standard definition of science is the methodical study of the natural world via observation and experimentation. That definition seems solid and well-known. But as we look into a lot of fields nowadays, especially those that look at deep time, complex systems, or indirect evidence, the border between observation and interpretation is not as clear. So, as we move into studying cosmology, origins research, artificial intelligence, and certain areas of theoretical physics, primarily rely on models that structure data rather than on experiments that can be replicated in controlled environments. This leads to a more fundamental inquiry: is the principal aim of research to uncover the true nature of reality, or to build models that sufficiently predict and control our observations?

People often talk about this dilemma as a distinction between "functional models" and "absolute truths." For a long time, philosophers of science have been fighting about realism and instrumentalism. Realists assert that science reveals authentic patterns and causes in nature. Instrumentalists regard scientific notions primarily as instruments for predicting outcomes, avoiding assertions regarding the nature of reality. But this way of thinking can mask a difference that scientists deal with every day that is more useful. A model that accurately predicts and organizes data differs from one that elucidates causation. These are not conflicting ideas; rather, they represent varying degrees of clarification. Yet still I say build one that does both. Sometimes, the right hand must know what the left hand is doing.

One example that everyone knows is gravity. Newton's math was right about how planets move for hundreds of years, but it didn't explain why masses pull on one another. In actuality, the notion worked well, but it didn't say why things happened the way they did. Einstein's theory of relativity created a bigger picture by connecting gravity to the shape of space and time. However, it did not address the question of whether gravity and quantum physics could coexist. This indicates that scientific advancement normally goes from "it works" to "it explains more," but those two things are not the same. Just because something can forecast what will happen doesn't mean it knows (or can tell them) why it happened.

Yes, we know (I hope) that this difference is essential in places where it's impossible to identify the direct cause. Historical sciences (evolution), cosmology, and origin (abiogenesis) studies occasionally investigate phenomena that cannot be replicated in a laboratory setting. In certain fields, several frameworks can arrange identical evidence with equivalent predictive efficacy. It's hard to tell when a framework is just useful and when it really explains what's going on. What else do you need to cross that line? What distinguishes convenience from adequacy?

The important concern might not be whether science looks for facts or models, but how scientists identify when a model has gone too far and is no longer relevant. Questions? What are the features that make that difference evident in your area of work? When the lines between prediction and explanation aren't apparent, what methods can you use to figure out if a theory is teaching you something true about the world or just putting the facts we already know in order? This phase might be one of the most crucial for determining what science is truly looking for.