Markov text generation

Then we took pictures with Agnes’s website whether one is Muninn, the god of the bench by examined, your kid’s not much of a sandwich to see how the tinkling music inside which I sort of describes teaching Hannibal left. You were freaky starecase, spelled that was hewn from the entire book, as we may try. A book that purifies both by examined, your kid’s name of the spider choked in the cellar bench, wonderful people?

By now, you are no doubt aware of what-would-i-say.com,  the website that will generate new texts from your facebook feed. It is basically a Markov text generator.

If you want to play around with Markov text generation, here is another Markov text generator online that will generate text from any text you choose to input. It generated the text at the beginning of this post from text copied from the last few months of my blog.

It lets you determine the randomness and length of the output. Here is an excerpt of text generated from my blog at the most random setting:

I’se mssqumyoblan meatore intown s htheovewh bovinorwathe ce dinthin do wes  wros thespid wa isoricr, ofouan Henk se ck aters wacro HP oimeand.

This setting would come in handy if you are writing a fantasy novel and don’t have time to invent your own fantasy language.

The second-most random setting gives you this:

“Perhapto ther sychops more Mundately rand go larends int to I has if the I site ally expere littage exack melly waite ack. I (evence to ke had ahem somer one be theady yell that chas met.

Set output to book length and write Finnegan’s Wake II in no time!

Text generated from the third-most random setting:

They leaks to did napkin his blog, mean als in a busy, Contail is at they tickets into do wher peanuts of the Don’t have grey has suffet, because the fun put the pressure you trying scribed in Vienna photogrape via train his tinkled up. Hokusai-ins tomatter remember moves are neighbor’s face at the other. A smart of back to the has a little of throws.

Fourth-most random setting:

So salmon is of the leaky faded quickly or a between the hoses contact me at me so I done: collowing up, and wait the weeks, and you trying as going on the calm before new heroine a thesaurus: perfect and Bran. Go ahead and I had me to the older it took a raises his poured over mouth with he had to avoid doesn’t wanted within three meters. After the excrement because we needed lettuce.

Indeed! Fifth-most random setting almost looks normal:

Mark didn’t feel the street car would be completed by foot, or something new, I guess.

Markov chains have been around for ages, so have language bots using them, but this is all new to me. It’s fun to play around with, and interesting to see what it does to language, and your own writing. Here’s the next level of randomness/non-randomness:

The god of the office has never before seen, has never before seen, has never before our superior alien might!’ But no one steals anything friends arrived, the plate holder, that explosively flammable party drug of the god of the spot like that. Anyway, if you do a good job, they come back.

It’s like my own writing, only better! A person could use this stuff as writing prompts, at the very least. Here is the next (seventh?)-most random setting (there appears to be no upper limit to how (non)random you can make it, but less random settings didn’t give me much interesting new text, so I will stop here):

The Spider watched another forms of photography that I can identify by smell, because he was already getting half-price day-old chicken salad.

Odin thinks about trying to write and it will take him to the subway and miss a stop.

Eventually I do, of course,