I’d make a distinction between: “The author believes (almost everything of) what they are writing.” and “The experiments conducted for the paper provide sufficient evidence for what the author writes.”
The first one is probably true more frequently than the second. But the first one is also more useful than the second: The beliefs of the author are not only formed by the set of experiments leading to the paper but also through conversations with colleagues, explored dead-ends, synthesizing relevant portions of the literature, …
In contrast, the bar for making the second statement true is incredibly high.
So as a result, I don’t take what’s written in a paper for granted. Instead, I add it to my model of “Author X believes that …”. And I invest a lot more work before I translate this into “I believe that …”.
This seems like it trusts the authors to be summarizing their own work accurately. Is that correct? Do you/your friend endorse that assumption?
I’d make a distinction between: “The author believes (almost everything of) what they are writing.” and “The experiments conducted for the paper provide sufficient evidence for what the author writes.”
The first one is probably true more frequently than the second. But the first one is also more useful than the second: The beliefs of the author are not only formed by the set of experiments leading to the paper but also through conversations with colleagues, explored dead-ends, synthesizing relevant portions of the literature, …
In contrast, the bar for making the second statement true is incredibly high.
So as a result, I don’t take what’s written in a paper for granted. Instead, I add it to my model of “Author X believes that …”. And I invest a lot more work before I translate this into “I believe that …”.