One of the mathematical facts about 100% certainty is that it means that no _possible_ evidence, no matter how overwhelming, could ever make you less than 100% certain again. Even if the original painting turned up down the back of someone's sofa, along with the notebook kept by Gaston Lévy during the period that goes into great detail about how he copied it, you'd have to stick to 100% certainty at risk of mathematical incoherence. P(x) = 1 -> ∀y∈F P(x | y) = 1.
I take it that you didn't actually mean to make a strict mathematical claim, but were just ironically rebutting the fake maths of the AI 'analysis'. But I think it's interesting to ask under what circumstances you might actually reasonably come to be 100% certain, in the strict mathematical sense, in an attribution of authorship.
The example I think about the most is the Pauline epistles. It is relatively well-known that most scholars agree the 'Pastoral' letters to Timothy and Titus were not written by St Paul. This is usually explained by pointing out that the writing style in the Pastorals is very obviously different from how Paul wrote, and the theology in those letters is also significantly different from Paul's own theology (so significantly different that it's almost impossible to have been the result of a simple change of mind). But some people have a follow-up question: how do we know what Paul's writing style and theology were _really_ like? The answer is 'based on the authentic letters', but now the answer is circular: we know which letters are authentic based on writing style and theology, but we know about Paul's writing style and theology from the authentic letters. Why can't we say that the Pastorals are authentic, and the ones that differ from them must be spurious?
There are some more substantial answers to this question, but I think the most fun way to respond is: the guy we are interested in is the guy who wrote the masterpiece we call 'The Letter to the Romans', _whoever that was_. If it turns out that his name actually was Paul and he really was a first-century Christian missionary, great. If it turns out that his name wasn't actually Paul and he was a Greek pagan trying to write a parody of these Jewish apocalyptic preachers, we could still call the guy 'Paul' for ease of reference, and we'd still be just as interested in what 'Paul' wrote. And if it turned out it was a pious forgery by some second- or third-century Christian, we'd just lose interest in the historical figure, and keep writing about the author, at most maybe adding a 'Pseudo-' to the start of his name. (This is exactly what happened with Dionysius the Areopagite: nobody cares about the historical figure any more since it turned out that he didn't write the works attributed to him.)
In a certain sense, then, we can be 100% certain that Paul wrote the letter to the Romans, and no evidence will ever make us less certain; because, in a certain sense, what we mean by the name 'Paul' is 'the guy who wrote this text', so it is true by definition. The current state of scholarship is that we can be pretty sure that this guy was actually a real historical figure named Paul, who lived in the period he claimed and had the life experiences he wrote about. But even the most radical new bit of evidence wouldn't cause us to file the text away forever as a forgery, and start trying to reconstruct the life anew; at most we would instead file the life away forever, as irrelevant for understanding the text.
I don't think you'd necessarily go this far, but there are echoes of this idea when you write:
> This authentic Rubens is indeed not only a genuine Rubens, but it is a lodestone of Rubensian perfection such that many other works, even autograph works, pale in comparison to this one painting.
The difference with Rubens is that there are a lot of works one might be interested in with him, and many are unambiguously and uncontroversially associated with the life of a historical seventeenth-century figure, so we cannot so easily redefine the name 'Rubens' to mean 'the guy who painted this Samson and Delilah'. But I wonder if you're gesturing at the idea that Rubens only deserves so much attention to the degree that he actually painted this work; if he did not paint this work, our understanding of what Rubens stands for in art history would have to be radically changed. You're not going all the way towards re-definition, and thus not all the way towards certainty; but it's a lot closer than others might realise.
Thanks Peter! This is very interesting on probability and I absolutely didn't think of it in this detail, but you make a fair point.
Yes, I was being pretty lighthearted with my certainty "figure" and taking the mickey out of pseudo-quantitativeness of it all. If I were to write this again I might have said something like 99.999999999999% certainty or something to that effect.
On the 'consequences of Samson and Delilah' for Rubens question I completely agree. I think this feels along the lines of the point I was finishing with, just more precisely expressed. I do think this work being attributed to Rubens is consequential for Rubens for the better and makes him more interesting. This is partly because of the very good condition of the picture (note the risk of admitting this from my point of view) and because the paintwork is exceptionally brave even for him (again note the risk of admitting this).
One of the mathematical facts about 100% certainty is that it means that no _possible_ evidence, no matter how overwhelming, could ever make you less than 100% certain again. Even if the original painting turned up down the back of someone's sofa, along with the notebook kept by Gaston Lévy during the period that goes into great detail about how he copied it, you'd have to stick to 100% certainty at risk of mathematical incoherence. P(x) = 1 -> ∀y∈F P(x | y) = 1.
I take it that you didn't actually mean to make a strict mathematical claim, but were just ironically rebutting the fake maths of the AI 'analysis'. But I think it's interesting to ask under what circumstances you might actually reasonably come to be 100% certain, in the strict mathematical sense, in an attribution of authorship.
The example I think about the most is the Pauline epistles. It is relatively well-known that most scholars agree the 'Pastoral' letters to Timothy and Titus were not written by St Paul. This is usually explained by pointing out that the writing style in the Pastorals is very obviously different from how Paul wrote, and the theology in those letters is also significantly different from Paul's own theology (so significantly different that it's almost impossible to have been the result of a simple change of mind). But some people have a follow-up question: how do we know what Paul's writing style and theology were _really_ like? The answer is 'based on the authentic letters', but now the answer is circular: we know which letters are authentic based on writing style and theology, but we know about Paul's writing style and theology from the authentic letters. Why can't we say that the Pastorals are authentic, and the ones that differ from them must be spurious?
There are some more substantial answers to this question, but I think the most fun way to respond is: the guy we are interested in is the guy who wrote the masterpiece we call 'The Letter to the Romans', _whoever that was_. If it turns out that his name actually was Paul and he really was a first-century Christian missionary, great. If it turns out that his name wasn't actually Paul and he was a Greek pagan trying to write a parody of these Jewish apocalyptic preachers, we could still call the guy 'Paul' for ease of reference, and we'd still be just as interested in what 'Paul' wrote. And if it turned out it was a pious forgery by some second- or third-century Christian, we'd just lose interest in the historical figure, and keep writing about the author, at most maybe adding a 'Pseudo-' to the start of his name. (This is exactly what happened with Dionysius the Areopagite: nobody cares about the historical figure any more since it turned out that he didn't write the works attributed to him.)
In a certain sense, then, we can be 100% certain that Paul wrote the letter to the Romans, and no evidence will ever make us less certain; because, in a certain sense, what we mean by the name 'Paul' is 'the guy who wrote this text', so it is true by definition. The current state of scholarship is that we can be pretty sure that this guy was actually a real historical figure named Paul, who lived in the period he claimed and had the life experiences he wrote about. But even the most radical new bit of evidence wouldn't cause us to file the text away forever as a forgery, and start trying to reconstruct the life anew; at most we would instead file the life away forever, as irrelevant for understanding the text.
I don't think you'd necessarily go this far, but there are echoes of this idea when you write:
> This authentic Rubens is indeed not only a genuine Rubens, but it is a lodestone of Rubensian perfection such that many other works, even autograph works, pale in comparison to this one painting.
The difference with Rubens is that there are a lot of works one might be interested in with him, and many are unambiguously and uncontroversially associated with the life of a historical seventeenth-century figure, so we cannot so easily redefine the name 'Rubens' to mean 'the guy who painted this Samson and Delilah'. But I wonder if you're gesturing at the idea that Rubens only deserves so much attention to the degree that he actually painted this work; if he did not paint this work, our understanding of what Rubens stands for in art history would have to be radically changed. You're not going all the way towards re-definition, and thus not all the way towards certainty; but it's a lot closer than others might realise.
Thanks Peter! This is very interesting on probability and I absolutely didn't think of it in this detail, but you make a fair point.
Yes, I was being pretty lighthearted with my certainty "figure" and taking the mickey out of pseudo-quantitativeness of it all. If I were to write this again I might have said something like 99.999999999999% certainty or something to that effect.
On the 'consequences of Samson and Delilah' for Rubens question I completely agree. I think this feels along the lines of the point I was finishing with, just more precisely expressed. I do think this work being attributed to Rubens is consequential for Rubens for the better and makes him more interesting. This is partly because of the very good condition of the picture (note the risk of admitting this from my point of view) and because the paintwork is exceptionally brave even for him (again note the risk of admitting this).
You’re very convincing !
Thank you!