David Deutsch (DD) wrote in Quantum theory, the Church-Turing principle and the universal quantum computer (1985), p. 3:
Church (1936) and Turing (1936) conjectured ... This is called the ‘Church-Turing hypothesis’; according to Turing,
Every ‘function which would naturally be regarded as computable’ can be computed by the universal Turing machine. (1.1)
And from Deutsch's references (p. 19):
Turing, A. M. 1936 Proc. Lond. math. Soc. Ser. 2, 442, 230.
Now we'll compare with Turing's paper: On Computable Numbers, With An Application To The Entscheidungsproblem (1936), p. 230:
the computable numbers include all numbers which could naturally be regarded as computable.
Turing wrote "numbers", but DD misquoted that as "function". Turing also wrote "could" which DD misquoted as "would".
There's also a problem because Deutsch uses what appears to be an italicized block quote. You'd expect the whole block quote to be a quote of Turing, but instead it's a paraphrase. Inside the paraphrase are quotation marks surrounding the misquote of Turing that I criticized.
DD's citation is also incorrect. DD cites Turing's paper to volume 442 of the Proceedings of the London Mathematical Society, but it was actually in volume 42 not 442.
To determine what's correct, we can check how Turing himself cites it. In a correction to his paper, Turing cited himself:
Proc. London Math. Soc. (2), 42 (1936-7), 230-265.
You can also see that the latest volume of the journal, published in 2021, is volume 122. Volume 442 is unlikely to exist for over 100 more years. And the journal's website has archives showing that the Turing article was in volume 42.
Tangentially, I hope this lowers your opinion of academic peer review. DD's paper was published in Proceedings of the Royal Society of London, a prestigious and peer-reviewed journal that started in around 1830. It has published work from many famous scientists.
Thanks to Dec for finding this misquote.
Note that DD has published a lot of misquotes.
Update 2021-07-15: Dec pointed out that a similar Turing misquote is in DD's book The Fabric of Reality:
He [Turing] conjectured that this repertoire consisted precisely of ‘every function that would naturally be regarded as computable’.
No, Turing wrote "all numbers which could" not "every function that would".
It appears that DD got this misquote from his own paper, and also modified it. There's a recurring pattern where every time DD touches a quote, there's a significant chance that he changes something. Here, he took the word "every" which was outside of quote marks in his paper and moved it inside quote marks for his book.
Update 2021-09-14: I contacted the academic publisher (proceedings of the royal society). They looked into the matter and said:
Apologies for the delay in getting back to you on this. A board member has had a look at the paper and does not think the misquote affects the outcome of the research presented in the paper. Although the error in the refences is unfortunate, we do not believe it will prevent readers from finding the correct article. Given the age of the paper we therefore do not think any further action is necessary.
I have several criticisms of this response.
They agree with me that DD misquoted and miscited.
Why won't they put up errata on their website? Is that too hard for them (they are bad at websites?) or do they actually not want to?
Errata serves several purposes. Academics working in the field could find out about the issue. People debating the issue could also refer to it – it would e.g. let a student whose professor repeated the error borrow the journal's authority to correct the professor. It's risky to correct your professor in general, but much easier with an official errata to point him to.
Is correcting professors a real issue? I think so because professors have been teaching Deutsch's error (there are some examples posted in the comments below). And they've been doing it out of context. In other words, even if the error did not affect the conclusion of Deutsch's paper, it still can affect other conclusions about other issues. So spreading the error matters, and it has in fact been taught in schools. Also, any reader of the paper may remember the Turing quote and use it for something else, and it may negatively affect the conclusion of their usage, even if it didn't affect the conclusion of Deutsch's paper. (Admittedly, some of the professors don't cite a source and might have been getting the error from Deutsch's book The Fabric of Reality where he repeated a similar error. But the fact that Deutsch put roughly the same error in his book is, IMO, an additional reason to errata it and at least do a little bit to stop the spread of the error.)
If they published an errata or other note about the error, they could also state their reasons for why they believe the paper's conclusion is unaffected. Other people could consider that reasoning and potentially disagree. This could be an area for critical thinking and truth seeking rather than an unaccountable authority pronouncing judgment for secret reasons. Even if it's no big deal in this case, their general attitude is concerning. How many other judgments do they make with no transparency? What is the nature of those judgments? Are any of those judgments mistaken? Do they gloss over many errors in papers they published? Could they be doing that partly out of bias and not wanting to draw attention to their own involvement in errors?
People expect academic science journals with peer review to have high standards and to be really picky about errors. They are not living up to this reputation. So much for their unlimited interest in truth for the sake of truth or whatever they were supposed to be doing.
They are still sharing the paper electronically and could update it there. Deutsch is still alive and available and could actually write or approve a small update, or they could do an update which is labelled as written by a journal editor not Deutsch.
How did this error happen? How did every step of the publishing process miss it? Did anyone intentionally cause or allow the error? Were any biases involved? They did no post mortem, no root cause analysis, no investigation into their peer review and editorial process, etc.
There are major causes for concern here. This errors calls into question how effective their reviewers and editors are. It also calls into question Deutsch's integrity. Maybe it was an accident but they have given no account of how it could have happened accidentally nor asked him to give one.
Do peer reviewers or editors not check quotes or cites? Should they? How widespread a problem is misquoting? How many other misquote reports do they receive, validate as correct criticism, and then bury? Might they be hiding a pattern revealing that many papers contain misquotes? Instead of hiding misquotes should they be doing something different like e.g. paying people enough money for misquote reports to make finding the misquotes worth the time and effort? If they actually wanted to find out about misquotes, and find out how big a problem it is, wouldn't they do something more like that? They could have responded to me by offering me money to find more misquotes since I've proven I can do it. That seems reasonable if they were better and more interesting in correcting errors.
Deutsch had an argument with a referree which was related to the text Deutsch misquoted:
But I soon found out that not everyone saw it that way. I also had referee problems. The referee of the paper in which I presented that proof insisted that Turing’s phrase “would naturally be regarded as computable” referred to mathematical naturalness – mathematical intuition – not nature.
(BTW, as a first impression, without reading Turing's paper or investigating the issue, I agree with the referree. When talking about naturally regarding something, that sounds like it's talking about what is natural or intuitive to people and their opinions, not about nature, due to what the key word "regard" means.)
Could Deutsch have intentionally misquoted in order to help win a specific logical point he was arguing about with the reviewer? Could the horrible, misleading presentation of the quote (as a block quote with an internal quote – which btw has tricked some people into thinking the whole thing is a quote) have been some kinda compromise worked out between Deutsch and the peer reviewer? Was the misquote in earlier drafts of the paper? Do they have records of what changes were made to the paper during peer review? In any case, there is some possible motive here for Deutsch falsifying the quote on purpose or just being biased and more careless in his own favor. Deutsch has a history of repeated misquotes throughout his career and most of them favor him in some way and I don't recall any that were bad for him, so it seems like whatever's going on involves bias if not actual deliberate, fully-conscious misquoting.
Seriously, how do wording errors in quotes happen accidentally? I understand typoing a letter or two when typing a quote in from a paper book or journal. But how do you just change the word? That seems more like Deutsch quotes stuff from memory – and his memory is biased in his favor (or there's selection bias – if he likes the version he remembers then he uses it, but if it's not ideal then he looks up the exact wording). Quoting from memory in your books and papers (and scripted speeches) is a serious scholarship violation that should lead to repercussions and major reputational damage. That's totally unacceptable. Another possibility, which there have also been potential indicators for, is that Deutsch changes quotes during his editing process without double checking the original. I suspect Deutsch thinks certain minor changes to quotes are OK, and maybe this somehow escalates to more major wording changes after multiple editing passes. Deutsch's editing could be like the game "telephone" where you whisper something to the guy next to you, who whispers it to the next guy, and so on. The goal is to repeat exactly what you heard. After something has been whispered a dozen times, often all the words are different and the meaning is totally changed.
In my experience, people are often willing to view things as "an accident" or "a mistake" without thinking about how exactly it happened. Some mistakes are simple like a one letter typo happening because you pressed the wrong keyboard key by accident because your finger dexterity is good but imperfect so occasionally you hit the wrong key (and then you usually notice and fix the typo, but not always). But many errors don't have such simple explanations and merit actual analysis. Changing the word "numbers" to "function" is not a typo due to flawed finger dexterity. That's bias, misremembering (while incorrectly believing quoting from memory is OK), intentionally falsifying the quote, or perhaps a horribly unreasonable editing processes that edits words within quotes similarly to how it edits words that are not within quotes. Or there are other possibilities like maybe a peer reviewer or editor caused the error and Deutsch didn't have full control over the final wording of his paper.
And how did the journal miss the error? Was it anyone's job to catch the error? Would the journal like to catch such errors in the future? And how did the error remain unnoticed in the archives for decades? Do they have a tiny readership? Do their readers not care about errors? Do their readers fail to report errors? Do their readers report errors but nothing is done? Would it make sense to hire people to review the archives for errors or should they focus on catching more errors before publication or should they just continue to not even post errata about errors and pretend nothing happened?
For more info, see my reply email to the journal: