Talk:Infinite monkey theorem

Monkey caption

Previously, the caption said "A Chimpanzee probably not writing Hamlet." This was funny, but perhaps too humorous for an article. A user opted to change it to "a chimpanzee sitting at a typewriter," which, although changing the caption is reasonable, strikes me as redundant. Worth changing to something else? Delukiel (talk) 13:09, 21 October 2023 (UTC)

I changed it to "A chimpanzee typing random characters", which seems to capture both the sense of the image and the article. — Loadmaster (talk) 16:39, 22 October 2023 (UTC)

Works for me! Delukiel (talk) 00:35, 24 October 2023 (UTC)

This was discussed (briefly) in talk:Infinite monkey theorem/Archive 4#probably not. I still think it's fine to have it. It's a little bit The Economist-style, but I don't think it outrages the encyclopedic form. --Trovatore (talk) 00:51, 24 October 2023 (UTC)

Either one works for me (admittedly, I prefer the original, but I love goofball captions), I just thought what it was changed to was redundant. Delukiel (talk) 01:12, 24 October 2023 (UTC)

+1 for the original, it really doesn't hurt anyone. Dialmayo (talk) (Contribs) she/her 11:21, 24 October 2023 (UTC)

I've changed it back to the original. Delukiel (talk) 06:34, 26 October 2023 (UTC)

:D Dialmayo (talk) (Contribs) she/her 19:47, 26 October 2023 (UTC)

Now changed to "A chimpanzee writing Hamlet"! Didlidoo (talk) 01:05, 29 May 2024 (UTC)

too far, man _Dial^mayo 01:53, 29 May 2024 (UTC)

Can it stay, or do we need to get rid of it :( Didlidoo (talk) 02:07, 29 May 2024 (UTC)

@Didlidoo the edit was reverted 3 minutes after you made it _Dial^mayo 02:09, 29 May 2024 (UTC)

D: Didlidoo (talk) 03:27, 29 May 2024 (UTC)

an encyclopedia is supposed to not say stuff like that in image captions, this is only for non serious wikis like rationalwiki (wikipedia also is not serious but at least somewhat serious/trustworthy compared to rationalwiki) 2A02:3100:3A3F:7B00:55C0:7B8F:9CAD:F16C (talk) 03:54, 20 July 2024 (UTC)

Because you enjoy a joke doesn't mean it overrides MOS. I am not the first editor you have reverted in order to keep the joke on the page. It is disruptive to readers who will not understand the joke. Please keep the humor to essay pages or other parts of the project. I am happy to propose an alternative caption, but the caption isn't a place for jokes or quips. Dreameditsbrooklyn (talk) 19:28, 4 November 2024 (UTC)

Apologies, above comment directed towards @Trovatore. Dreameditsbrooklyn (talk) 19:29, 4 November 2024 (UTC)

A caption literally describing the image is perfectly useless here; anyone can see what the image is, and the article is not about chimpanzees or typewriters. "Anyone can see" excepting of course our vision-impaired users, but for them there's alt-text, and I'd be happy for your description to go there.

The "no-fun party" does seem to be in almost total control of Wikipedia, but it wouldn't hurt for them to allow this harmless deviation. However, if not, the remaining option, it seems to me, is to have no caption at all. I could live with that. --Trovatore (talk) 19:34, 4 November 2024 (UTC)

We should have a caption that follows MOS. This isn't about 'no-fun,' it's about making the article accessible to people who might not grasp the joke. I find it odd that you propose to throw out the caption all together if you don't get to have your fun. Dreameditsbrooklyn (talk) 19:36, 4 November 2024 (UTC)

OK, please remember that the MOS is a guideline. It should "ordinarily be followed". As I said, a literal caption is entirely useless for this image. --Trovatore (talk) 19:38, 4 November 2024 (UTC)

I agree that MOS is a guideline, but I don't think "because I think it's funny" is a compelling reason to ignore them or the handful of other editors that have rightfully fixed the caption in the past. There were better captions in previous revisions of the article. I'm happy to work on a new one that is helpful to readers. Dreameditsbrooklyn (talk) 19:51, 4 November 2024 (UTC)

I mean, the image itself is really there just to be funny. Or more precisely to provide visual interest that might attract attention and get people to think about the content of the article. It doesn't actually elucidate anything in the article, so it's not clear how a caption would help it do something it's not supposed to do in the first place. --Trovatore (talk) 20:04, 4 November 2024 (UTC)

Please review my attempt at a caption useful to readers. I am perfectly happy to have the caption blanked if you think better, though I do agree with you that there is an argument for visual interest (without a quip or joke) Dreameditsbrooklyn (talk) 00:32, 5 November 2024 (UTC)

Would it be reasonable to use the language "almost surely" in the caption as it is used in the lead? e.g. "A chimpanzee almost surely not writing Hamlet"? _Dial^mayo 18:11, 5 November 2024 (UTC)

Current version looks good :) Didlidoo (talk) 08:02, 6 November 2024 (UTC)

"A monkey, not an ape" is logically impossible. Apes are monkeys. An animal can be "monkey, not an ape" but the opposite doesn't exist. — Preceding unsigned comment added by 2001:4C4D:218C:E500:F1A3:78A2:B90F:685 (talk) 18:06, 24 October 2024 (UTC)

This is completely off-topic but indulge me for a moment. At first I thought the claim "apes are monkeys" was just obviously wrong. But I looked around our articles and apparently it's true for some value of "true". Apes are monkeys in the same sense that birds are dinosaurs and there's no such thing as "trees". That is, if you insist on a taxonomy where you consider only monophyletic groups, then you would say apes are monkeys. In normal English, though, apes are not monkeys. --Trovatore (talk) 20:45, 24 October 2024 (UTC)

New (and old) Articles in The New York Times

Just ran across this article in this morning's New York Times, "Could Monkeys Really Type All of Shakespeare? Not in this universe, a new study concludes." Nothing too startling; I think it's the second time this year I've seen something similar. Might be worthy of mention in this article, however. Not sure of the distinction between this article and the use of this trope in popular culture, which is a separate article. What brings me here, however, is a particular experiment involving this theorem that I read about nearly forty years ago, and whether there's a reason to mention it in this article or elsewhere:

In elementary school I obtained a used copy of Encyclopedia Brown's Second Record Book Of Weird And Wonderful Facts by Donald J. Sobol (Dell Yearling, 1981), which included at page 49:

A Yale University professor, Dr. William Bennett, has calculated that a trillion monkeys, each hitting 10 keys on a typewriter per second, would need a trillion times longer than the universe is old to reproduce Shakespeare's famous line, "To be, or not to be: that is the question." And how long for the answer?

And for many years, I thought that was the original source and version, mentally chiding every account that needlessly substituted "the complete works of Shakespeare", and forgetting all of the other details (I was astonished when I dug out the book a little while ago to see it was "a trillion monkeys"). Of course I found out a while back it was a much older idea!

The Dr. William Bennett referred to seems to be William R. Bennett Jr., and though his article on Wikipedia makes no mention whatever of monkeys, I did find a 1979 New York Times article about his foray into monkey typing probability using computer simulations: "Computer Says Monkeys Couldn't Write 'Hamlet'—At Least Not So Far", which includes samples of text produced using various iterations of his program, beginning with purely random key strikes, and progressing to variations in which letters are produced in frequencies approximating English, the language of Hamlet, and focused on shorter combinations of letters. The best efforts of the virtual monkeys, even with all these refinements, still failed to produce intelligible English (not surprisingly), but there were some intelligible strings of words, of which one sample was:

TO HOIDER THUS NOW GOONS ONES NO ITS WHIS KNOTHIMEN AS TOISE MOSEN TO ALL YOURS YOU HOM TO TO LON ESELICES HALL IT BLED SPEAL YOU

And the "winner" was:

TO DEA NOW NAT TO BE WILL AND THEM BE DOES DOESORNS CAI AWROUTROULD.

Which even as I've just summarized is probably too lengthy to include here or in the "popular culture" article, but maybe Dr. Bennett's experiment should be mentioned—and maybe its inclusion in Encyclopedia Brown would fit in the pop culture article. It's certainly the reason I know about this topic! I don't know—I think editors who are actively engaged in editing this and the other article would be better judges of the significance of the experiment in early computer research on probability, particularly as the researcher, the Times, and the book series citing his experiment, are all individualy notable. P Aculeius (talk) 18:23, 3 January 2025 (UTC)

This is already treated in the "Probabilities" section. As for the NYT article, as far as I can tell there's nothing new there; the fact that they reported it as though it were some novel result is pretty embarrassing for them, or should be.

It is true that we should probably mention this in the lead section. One sentence would probably do. I might take a crack at it later if no one else gets around to it first. --Trovatore (talk) 20:18, 3 January 2025 (UTC)

UPDATE: I couldn't find a natural place to mention it in the existing text, and I was too lazy to rewrite the whole lead, so I made it an explanatory footnote. I didn't go into details about the observable universe and so on because I thought they might distract from the main point. Feel free to tweak if you can do better. --Trovatore (talk) 20:32, 3 January 2025 (UTC)

I don't see how this has anything to do with my question, which was whether it was worth mentioning William Bennett's use of computer simulation in 1979, as discussed in The New York Times and later used as a factoid in children's literature. It wasn't "already treated" in any section; I checked and nowhere does the name "Bennett" occur in the article, before or after the addition of a vaguely-worded footnote alluding to nothing in particular. I don't know which of two articles in the Times you're calling an embarrassment—reporting on one scientist's use of computer simulation to illustrate a point and elaborate on a popular thought experiment in 1979, or today's article, describing a new paper indicating that previous analyses dramatically overestimated the probability, and providing an overview of the theorem's history. But just because it seems passé to someone who's researched the topic and written about it doesn't mean that modern journalists covering science should be castigated for reporting on it. I suppose I will now have to consider carefully how to square my inexpert opinions with an article written from a purely mathematical perspective. P Aculeius (talk) 21:27, 3 January 2025 (UTC)

Oh, I see. To be honest, my feeling is that the calculation is trivial, and a simulation that merely confirms a trivial calculation is probably not worth mentioning. So my answer to your direct question in the first sentence of your latest response would be "probably not".

But it's not a hill I'm going to die on. If other contributors think it's worth mentioning, then we can. Let's avoid making it sound more spectacular than it is, and I certainly don't see any call to mention it in the lead. --Trovatore (talk) 21:49, 3 January 2025 (UTC)

I never suggested mentioning it in the lead. That was your idea. P Aculeius (talk) 22:36, 3 January 2025 (UTC)

True. In any case I looked up the Woodcock paper and am not impressed. Essentially he's done trivial calculations and put pretty graphs with it. It's published in "Franklin Open", which claims to be peer-reviewed but puts its publication fee up front on the landing page (I am looking into its status at WP:RSN). By the way I see no evidence whatsoever for your assertion that "previous analyses dramatically overestimated the probability" — where did you come up with that exactly?

The Bennett simulation could possibly be of interest as a historical note, not as something that has much to teach us about the topic of the article, but simply as an early example of doing that sort of thing. If it really is an "early" example (I wouldn't be surprised if RAND had done that sort of thing considerably earlier). --Trovatore (talk) 22:48, 3 January 2025 (UTC)

Well, the fact that the Woodcock and Falletta paper has been reported and discussed in The New York Times would seem to make it notable enough to mention or describe in brief, even if the paper itself isn't a reliable source; I've had a go at adding both studies, and carefully limited my text to say that's what was reported, without passing on whether their conclusions were correct (on the other hand, they do seem to agree with every other version of the experiment in concluding that random strings of text including any recognizable passages from Shakespeare are highly improbable, even given vast quantities of time).

I expect that you or another editor will probably want to cut down my text, even though I tried to be economical with it. It might be consolidated with some of the other sections, but the fact that the theorem is still the subject of various published—if not always particularly serious—research, and keeps making the news every so often seems worth mentioning with some details and context. I'm not prepared to die on this hill either, but it'd be nice if what I wrote wasn't simply dismissed out of hand because it's not perceived as serious enough. P Aculeius (talk) 23:32, 3 January 2025 (UTC)

My view is we should not mention the Woodcock thing at all. Just because a NYT reporter fell for it doesn't make it of any value. Of course they agree with all the other sources, because their results are not in any meaningful sense new. It's stuff everyone already understood. --Trovatore (talk) 23:59, 3 January 2025 (UTC)

You may be right, but as it's being reported as news, I'm going to err on the side of caution and assume there's actually something new about it, unless I can be certain otherwise. And I can't see that yet. Maybe if someone who has more expertise in the topic than I and is less convinced that the story is rubbish has a look, they'll strike a better balance. P Aculeius (talk) 03:28, 4 January 2025 (UTC)

I don't see how that's "erring on the side of caution". News outfits report stuff all the time that isn't "notable" as the term of art has it. My view is that there is no "balance" to be had here; we should not cover the Woodcock thing at all. --Trovatore (talk) 05:06, 4 January 2025 (UTC)

Normally when a minor, if popular and not very consequential thought experiment makes international news, you can assume that that in itself is significant. You consider it old hat, but you're an expert in the field, and other readers may disagree. The fact that this recent study has been covered by the BBC, The Guardian, Smithsonian Magazine, and The New York Times (among others—ScienceDaily, Physics World, New Scientist) would seem to argue that it's worth mentioning. If I'm right, then I've written something helpful—a couple of sentences that describe who published it and what the study was about. If I'm wrong, nobody is harmed by seeing it mentioned. That seems like erring on the side of caution to me. P Aculeius (talk) 05:18, 4 January 2025 (UTC)

It was significant in that these news outlets got fooled into thinking there was something new here. The harm in mentioning it is that it gives credit to this narrative the authors seem to be pushing that it wasn't understood previously that, for any reasonable number of monkeys, there is no significant chance of getting Shakespeare.

But that was thoroughly understood from the beginning. --Trovatore (talk) 05:32, 4 January 2025 (UTC)

I don't think the risk that people might think this is important is significant enough to suppress any mention of it; readers can be expected to wonder if it's mentioned here, or whether there have been recent studies related to it, and the greater harm would be pretending that there haven't been, just because experts on the topic dismiss them as insignificant. P Aculeius (talk) 05:36, 4 January 2025 (UTC)

Oh, and I see I forgot to respond to your question about the Woodcock and Falletta paper being about overestimation in past analyses. That seems to be what they're getting at when they mention that "our universe" doesn't allow for an infinite number of monkeys or an infinite quantity of time; their study is based on the maximum possible number of monkeys that actually exist, and bounded by the loosest possible time frame (the maximum age attainable by the universe, as science currently estimates). I'm pretty sure they say explicitly that this is a fault in previous analyses, placing no practical limits on key variables employed by the theorem. In fact the novelty of their paper seems to be based primarily on the unique conditions they placed on the equation, just as Bennett explored different iterations using linguistic patterns to see how significantly they could increase the probability (which they did, albeit not enough to produce meaningful language out of random strings—and I think that was the real point, rather than whether monkeys were involved). P Aculeius (talk) 05:36, 4 January 2025 (UTC)

(Just as an aside, they assume facts not in evidence when they speak of "our finite universe". As far as I know it's still an open question whether the universe is finite or infinite. But this is a detail; we can give them the benefit of the doubt and assume they meant "observable universe".)

There is no novelty here. It was almost the point of Borel and Eddington's remarks on the theorem, and it's thoroughly covered in the third paragraph of the "Probabilities" section of the article. Woodcock and Falletta may have assumed constraints that differ in detail, but not in the conclusions to be drawn. --Trovatore (talk) 06:32, 4 January 2025 (UTC)

The universe is assumed to be finite not in the sense that it must have physical boundaries, which I agree is an open question, albeit one in which there is no evidence of such boundaries and no means of learning anything about them if they do exist in the present state of science, or the foreseeable future, so quibbling over the use of "observable universe" seems to introduce a new element of speculation needlessly into the discussion of whether the study produces anything novel.

However, that is not what I meant by "finite universe". The assumption of the paper is not a universe with physical boundaries, but one of finite duration, which science predicts must exist, although there are several possible scenarios that might constitute the end, only one of which (the most probable, according to astrophysicists) is used here. And that is a constraint not necessarily applied by other permutations of the "infinite monkey theorem" (which, after all, also posits infinite monkeys rather than the finite number of monkeys that do or might reasonably exist).

Of course the study can be dismissed because as either a science topic or a thought experiment with vague and variously undefined and practically impossible parameters it's essentially trivial, as its authors plainly admit. But then we would have to ignore the fact that the topic itself is trivial, despite the considerable amount of attention it's received over more than a century, and that would be self-defeating.

Perhaps we should consider this: from July 31 to October 30, 2024, this article received just over 800 daily page views. With the release of the Woodcock and Falletta paper, which first seems to have come to public attention through the news media on October 31, page views more than doubled to over 1700 for the next month, and since December 1 they are just shy of 900, meaning they still have not returned to their previous levels. We might anticipate another bump in traffic with the Times article, and perhaps other media outlets that follow from that.

But either way, it indicates that thousands of readers who would not have come to this article have recently done so because of the Woodcock and Falletta paper. Even if you believe it is poorly written, unoriginal, or that its conclusions are wrong, it has demonstrated significance because it brought readers here, so mentioning and briefly describing it would seem to be necessary; otherwise those readers will be confused about why it's not mentioned at all, and almost inevitably someone will try to add it.

I think my text is modest and of reasonable length and detail, although I have no objection to someone monkeying around with it if it can be improved—though shaving it down to a footnote or its functional equivalent would probably not be reasonable, since it would defeat the purpose of mentioning or describing a study that readers likely came here expecting to see mentioned. I'm happy to leave to other editors whether it or any similar studies could eventually be split off into their own articles, but IMO they would still need to be mentioned here. P Aculeius (talk) 15:00, 4 January 2025 (UTC)

It certainly made a brief news splash. Nevertheless there is nothing in it worth covering here. --Trovatore (talk) 19:12, 4 January 2025 (UTC)