While this is an interesting idea, if this becomes common practice, the only possible result is a chilling effect on the willingness of informers to inform or whistleblowers to blow said whistles. It's another needle in the coffin of anonymity.
Plus, if this was written by a journalist (or someone else) based on a good-faith account from a legislator, the result will be either that person giving up his source (see paragraph 1, above) or being ethical and refusing to, potentially casting a bad light on the article whether it's appropriate or not.
One might expect that scary agencies already have machine-learning-based approaches as common practice...
Could one defeat it by writing something, then having someone else "translate" it into their own vocabulary sentence-by-sentence, or even paragraph-by-paragraph, and having the original author approve the "translation"?
Should there? There's definitely a lot of writing, but the difficulty of separating all the ghostwritten stuff from the politician's own words might muddy the statistics.
I actually doubt that you can find a large body of writing from enough Congress members to succeed. I don't think members write a word of anything that comes from them, in office or in campaigns. They're written by ad/consulting companies in campaigns, and by staffers in office.
Everything I've received in response to a message to a member in the last two decades sounds like it was written by a corporate spokesman. Which they probably were, since I'm sure these canned responses are reviewed for compatibility with a member's donors.
It doesn't work that way. You can narrow the range of possible authors down to a short list, but you'll never be able to publicly point the finger at someone and say "It was you!". And even your short list will be highly debatable. The analysis could be conducted easily enough and someone could make a strong case for why they feel the conclusions are accurate, but it would be far from hard evidence. Such a thing wouldn't be admissible in court today, but eventually it could be accepted in the same way that handwriting analysis was.
I really don't think this was written by an actual congressman. It seems like more of a political piece using the "confession" motif as a rhetorical device.
The problem is that Members of Congress don't write the stuff that goes out under their name--staffers do it. Just like President Obama doesn't write his own speeches.
