Debates among scientists sometimes look almost indistinguishable from public debates on science. The difference is in the premise: in the former case the merit

Debates among scientists sometimes look almost indistinguishable from public debates on science. The difference is in the premise: in the former case the merit of the methods is questioned, and in the latter case the merit of science itself is questioned. My feeling was this critique was of the former kind, particularly with regards to scientists' communication. His view is fairly similar to that of most academics I know, though less nuanced. In that light, I think you may have misinterpreted this particular comment. The word "truth" is a diabolical invention that has a fairly subjective interpretation that changes when you start looking into what it means. However, once you dig deep enough, you find the inevitable: There is no spoon. Empirical science is not about truths. It's about models. It's about modelling our world 'as well as we can'. The theory of relativity does not explain how the universe is architected, it's just a model that makes predictions. Any conclusions about the nature of the universe drawn from the structure of the model or any valuations of the model are entirely your own. What does "as well as we can" mean? Here it comes down to merit. Accuracy is a virtue, but accuracy often implies complexity and limited applicability (e.g., simulations are relatively accurate, but only concern a single set of situations). Practicality is a virtue, because few people use relativity to calculate gravity. Correctness is a virtue, but probability can make things much simpler. Simplicity is a virtue, but not for any fundamental reasons. Generality is a virtue, but often comes at the cost of accuracy. Completeness is a virtue, but we do not know how to achieve it. Communicability is a virtue, because many good models are forgotten for lack of sex appeal. Usefulness is perhaps the greatest virtue, but it's always too early to judge that. As scientists, we are constrained by the same reality that all systems face: There is no right solution. Why make models if they don't reveal any truths? Two reasons. One is that they (sometimes) allow us to calculate things, which is the most powerful concept in human civilization. It allows us to build bridges, launch satellites into orbit, produce the materials that we need, etc. without relying solely on trial-and-error. The second reason is that while the model does not reveal the truth, it does reveal a piece of the mathematics or processes that play a role in shaping the system being studied. Usually the model only reveals one aspect of the system, while in reality all things are related and influence each other and no calculable formula could describe it. However, it gives us a sense of direction, like a compass, rather than a map. Once you grasp these subtleties of science, you become much more careful about your claims. Not so much because you think they might be flawed, that option should already be off the table, but because you know your claims do not (intend to) fairly represent the complexity of the truth. Part of the author's frustration, which many in the sciences department would share, seems to be that some scientists don't see or willfully ignore the subtlety when making claims in public. This leads people to misrepresent science as a magical truth producing machine, and make preposterous claims that distract attention from what science actually does: create models, for whatever purpose they have.