
wowza
Since this guy is willing to throw real money at pointless, poorly-thought-out "experiments", maybe I can get him to just Paypal me the cash instead.
A developer and Simpsons fan has reported his first success in an experiment to test the infinite monkey theorem: that primates could produce Shakespeare by randomly pressing keys. Jesse Anderson set up software to generate random sets of nine ASCII characters and then match them against the complete works of Shakespeare, …
There never was any fun in this. Totally pointless. If I select random 9 character strings, and retain the ones which match a substring of a longer text, will I eventually match the entire text? Well yes, obviously. You can even estimate how long it will take.
If I drop individual grains of sand into a big glass jar, will it eventually fill up? Same answer. Just because something takes a long time doesn't turn it into an interesting philosophical question.
You have to see the rest of the picture.
He didn't care about the experiment, he knew exactly what he was doing. He parametrized it so that it would end in a finite time -- waiting for all needed 9 letter strings to appear has the correct timeframe; 8 letter strings goes 26x faster, 10 letter 26x slower.
As he explained to the Beeb, he did it to learn his way around Hadoop --- it's a marketable skill, and he's been in every non-techical news report around the world: very very cheap career advertising at $20/day for a short while. [He started for free when Amazon was giving away access to attract developers, and he took it offline when there wasn't anything left to learn to save the $20/day.]
The idea of using a fleet of monkeys to produce the works of Shakespeare did not begin with "The Simpsons." The idea stretches back to ancient times. In modern context, it is most often used as an argument against evolutionary common descent by random forces and chance events. Producing the genetic code of even the simplest organisms by random sequences would require several times over more time than the universe is expected to exist.
The problem is that if it hadn't happened then we'd not be here talking about it... Given enough time, enough universes (multiverse believer here) then *everything* is possible.
It's either that or except the much easier concept of their being a divine being, in which case I'm afraid it's turtles all the way down time :)
There can only be one universe, by definition. Either there exists a universe or multiverses but not multiple or parallel universe. There is no honest meaning to a plural of universe inasmuch as a universe is supposed to include everything. It isn't like a unicycle of which you can have several but is more like a multiwheel unicycle which isn't a unicycle at all but is some type of bicycle, tricycle, etc.
As to which it is, either uni- or multi-, I'll leave that to the bodhisattva.
Just want to stress that evolution has absolutely nothing to do with randomness, and that if this idea is used as an argument against it, it is only because the people arguing have no understanding (or intention of understanding) anything about evolution theory.
Yep totally agree with the point about this idea not having anything to do with The Simpsons either - a ludicrous implication by the article writer!
If not random variation, what source of variation are you suggesting allows Natural Selection?
Natural Selection is a process that works on natural variation, not a process that causes natural variation.
Lamarckism and random variation are two of the processes that are commonly suggested as being sources of natural variation for Natural Selection to work on. Radiation was another suggested source of variation. Population restriction enhances variation, but requires a source of original variation.
Evolutionary change based on observable natural variation is not, and never has been, compatible with the calculated age of the Earth. If you depend on observable variation, it doesn't work. This problem has always been obvious, even if 1859. It has been addressed in a number of ways, including by modern theories of speciation, which do not reject a link between randomness and evolution.
This post has been deleted by its author
This post has been deleted by its author
"it is most often used as an argument against evolutionary common descent by random forces and chance events"
If this experiment has any point, it is to demonstrate that such an argument is wrong. Trying to produce a complete works of Shakespeare (~360,000 specific characters) would need more than 10^500,000 attempts, which would be a stretch even using EC2, hence the cheating involved.
It's true that starting from a CHON 'soup' with a few impurities and trying to produce a self-replicating string of RNA by random activity would take longer than the expected lifetime of the universe. But you don't have to do this in one giant leap. There must be (hand-waving involved here) much simpler protein chains that can catalyse the formation of similar chains from the primeval soup. Once you've started to create them, evolutionary forces will start to act to improve their 'reproduction' rate - in some respects the less accurate the reproduction process, the quicker this works. Stanley Miller demonstrated back in the 50s that you can produce proteins from the 'soup' using heat, electricity, stirring and UV light.
Falling back on the 'big universe' argument isn't very satisfying. Anyway, I'm hoping that we will eventually find unequivocal signs of life, currently or historically existing elsewhere in the universe, which will knock this argument on the head (unless you follow Fred Hoyle with his idea of panspermia).
You can do a nice demonstration of the power of natural selection if you have, say, ten dice. Throwing them all and getting 10 sixes will take a long time, as any creationist will happily point out. But if you throw them and just keep the sixes that turn up (let's call them successful mutations) and then throw the rest again and repeat... then you pretty soon have 10 sixes.
I guess this experiment is along the same lines. Analysed logically, its rapid success is obvious. But it's still amazing how many people just can't see the difference between this and throwing all the dice at once.
So I think it's a useful reminder that there is a world of difference.
What's so special about getting nine consecutive characters in a row, if they still have to be selectively cut and pasted to make the actual text? Why shouldn't my monkey, that only generates one-letter sequences from Shakespeare, but generated all 26 of them within the first second of operation, count?
Bunch of self-publicising horseshit. Shame on ElReg for swallowing it.
His "virtual monkeys" generate random sequences of 9 characters, then another process searches Shakespeare's works and, if it finds a hit, says "Yes! This is a genuine piece of Shakespeare!" and marks off that bit as done.
That's hardly the same as the observation that if an infinite number of monkeys randomly hit keys on an infinite number of typewriters, one of them must eventually duplicate the entire works of Shakespeare as a single continuous text. This can not be practically demonstrated, but doesn't need to. It follows intuitively from the concept "random".
Why 9-character sequences? If he'd chosen single characters instead, he could (according to his definition) have duplicated the entire works of Shakespeare in only about 60 iterations (depending how many unique characters you want to count in the works of Shakespeare).
What a waste of money and computing resources.
Eventually?
With an infinite number of monkeys, all typing away at the same time, by the very definition of the word infinite, you will get every book ever written, to be written, or not even thought about yet, produced in the time it takes to type it. That includes all translations in all languages ever created or not yet conceived, human, animal or alien.
Only if you take away the word infinite and replace it with any conceivable finite number can you even begin to think about applying the concepts of time or statistics to the operation.
A good point - if there are truly an infinite number of monkeys, then some of them (also an infinite number) will type out the complete works at their first attempt.
This really only demonstrates that infinity is a concept you need to be careful with in verbal reasoning. Apologies for my sloppiness. For me, infinity usually amounts to any number which will overflow the storage I've allocated it.
Of course, it still works if you say "a large enough number of monkeys" or even "one monkey, given enough time."
The 9 character sequence selection/match process does trivialise the problem, but as a proof of concept of his methodology can be accepted as valid. In order to prove replication of Shakespere, the X number of randomly generated characters must be compared to the original text to determine a match....
What the experimenter now needs to do is increase the number of randomly generated characters to at least match the length of the shortest of Shakespere's sonnets. In that way, the match must be achieved without the jigsaw puzzle approach..
I doubt if the experimenter will be still alive to see his creation actually achieve the more difficult task but, (this being a random universe and all), I am always prepared to be amazed :-)
I reckon monkeys don't type randomly. I mean, the chance of getting SSSSSSSSS is rather more likely than getting ABCDEFGHI or hello there. Not all 9 character sequences are statistically equal, but if this experiment is using a pseudo-random number generator without any additional logic, to me, it's not sufficiently monkey-like in behaviour.
I don't have first experience of monkeys, though my brother once did when an escaped circus money locked himself in my brother's hotel room bathroom. Explaining that to reception was, er, interesting. I digress.
This is an interesting perspective on the American education system.
Firstly, the intellectual dumbing down that is implicit in relying on the Simpsons to be educated about an idea that even Aristotle argued.
And secondly, for having the sophistication to use new technology such as Hadoop.
This quotation from 1938 seems apropos:
"But the interest of the suggestion lies in the revelation of the mental state of a person who can identify the 'works' of Shakespeare with the series of letters printed on the pages of a book..."
-- R. G. Collingwood
This post has been deleted by its author