Some people have a really hard time separating “truth” from “fact,” and they also have difficulty with how these relate to science.
A novel may contain truth, in that it is not a factual account of anything, but you might learn a life lesson from it. Indeed, many childrens books (and certainly many other genres) are specifically intended to teach valuable lessons. Religious practitioners often conflate the two. If your scriptures are (as they are taught) “true,” does that mean they are factual? You might learn something from the Bible, but there are many things in it provably non-historical, consistent with the Hebrew penchant for taking other people’s oral traditions and adding a “moral” component. Anything historical in it doesn’t necessarily convey useful truth, and anything non-historical is not necessarily devoid of truth.
Conversely, it’s common to mislead by the use of facts. Propagandists often present accurate factual information, followed by specious reasoning that leads the listener to an incorrect conclusion. It’s all in how you present things, what you emphasize, what you downplay, and what teleological conjectures you want to draw to explain those facts. Politicians are brilliant at making the statistics say whatever they want.
Then there’s science. It is indeed fact-based. And we hope that it is true. But in fact, it is not a truth generating engine. It is a MODEL generating engine. A true model is, of course, better than one that is merely numerically accurate, but there’s only so far you can be sure (or maybe even care) just how true it is. Sometimes, you just need something predictlve. A recent Ars article about zebra stripes mentioned how scientists developed and tested several different explanatory models before they found one or two that were fully consistent with all of the facts. Every single one of those models, even the wrong ones, was scientific, because they were falsifiable (a term that few people really understand). Another example is the prevailing theory of the moon’s origin; we have a model that is consistent with what we can measure today, but there’s so little physican evidence that we only accept the model because we lack any better explanation. It if turned out to be wrong, nobody would be the least bit surpirsed. Even a blind, non-explanatory model, like using a neural net for numerical regression, is of scientific value, because it can be used to do engineering, and it may aid in further analysis that leads to a falsifiable explanatory model. Once a model has gone from postulate to hypothesis to theory, consistent with the evidence, we can say that it is consistent with the FACTS, but as for truth, we can only say that it is PROBABLY MOSTLY TRUE. Each time we discover some more evidence that we haven’t explained or which contradicts the model, we have to adjust it, making it incrementally more probably mostly true.
This brings me to pseudoscientific ideas like intelligent design. Even if it were, hypothetically, true, it isn’t and can’t be science. Why? Because it isn’t falsifiable. Anything you can’t explain, you can dismiss as being the result of some outsider tweak, so it’s impossible to prove it wrong. It’s also not predictive. It makes no interesting testable claims that evolutionary theory doesn’t, so it doesn’t yield any new knowledge. Finally, it’s useless to engineering. Not all scientific theories are necessarily going to be used by engineers, but in the case of intelligent design, it CAN’T be. A potentially useful scientific theory must be based entirely on predictable naturalistic mechanisms. This way, engineers can develop new systems that rely on or leverage those natural phenomena. Intelligent design, on the other hand, requires miracles or alien interference that we’re (by definition) too primitive to understand. And unfortunately, engineers can’t perform magic and don’t have access to alien hyperspace nano-wormhole entangement bioengineering technology.