So, I was sleepy and going off to bed when I made that comment and now I've realized that how massively I botched it. Major failure of communication on my part,

So, I was sleepy and going off to bed when I made that comment and now I've realized that how massively I botched it. Major failure of communication on my part, sorry. But it's good that you mentioned the Hemingway article, because the difference between that article and this interview highlights pretty well what's my issue with trying to "reconstruct" Camus from quotes. A living person is able to perceive as well as react to context. What questions have been asked, who's asking them, what is the level of understanding of the world, where we are ... these all might be grouped into "context", and the point is that they're all relevant in conversation. Now, when a person takes up the mannerisms of someone else, it doesn't really detract from their ability of adhering to context. But when they're trying to "reconstruct" a person from things they've said at some point for the purposes of an interview, while trying to preserve those quotes to some degree, there can never be such a close fit between question and answer because each was made in different contexts. Another way to look a it: what the author did is essentially search the Camus corpus for continuous segments of text which most closely fit his questions. But the answer which Camus would have most likely given could depend on discontinuous segments of text. This is actually pretty common in academic literature: whenever people are trying to reconstruct the meaning of certain concept, like Aristotle's "entelechy" or like Kant's "noumenon", sometimes we have to refer to different passages of the relevant corpus in order to get a clearer picture of what it might have meant. And there's an interaction between these passages, because they don't just heap over one another as in a concatenation of disjointed facts, but instead build upon another. The interaction between different passages might actually change the perceived meaning of individual passages. If there were a living person to answer these questions, whether it be Camus or just someone who really knows Camus, he already has these higher degree interactions in his head. He doesn't have the excerpts of text, he has the model which generates them. So if there's any question which might warrant a more subtle or complex answer, he'll able to do it. Maybe even a question which wasn't previously considered by Camus and for which we can find no immediately relevant passages in the corpus. My issue isn't that this is some bit of necromancy, but rather that it is a Procrustean bit of necromancy. I wasn't skeptical because there's some degree of uncertainty either (though I think that in between chopping text and answering based on some model there would be less uncertainty regarding the latter), I was skeptical because I think this is an awkward and useless approach. It doesn't "reconstruct" anything, it merely deconstructs the Camus corpus.