+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Voynich Talk (https://www.voynich.ninja/forum-6.html)
+--- Thread: Human generated nonsense text (/thread-3155.html)
Human generated nonsense text - Mark Knowles - 11-04-2020
I am sure if I have mentioned this before, but I wonder what kind of nonsense or meaningless text a human would be inclined to produce if that was their intention. I think that producing "random" or "randomish" text quickly is too computationally intensive for the human brain, our mind just don't seem to be designed to do such things easily, though of course for a computer this is trivial.(Just in the way that multiplying 2 large numbers together is hard for most people and trivial for a computer.)
It seems to me that a human trying to produce "random" or "randomish" will in fact produce text with a clear structure of pattern to it. Now this is, I suppose, an empirical statement. I suppose in principle one could asks 20 volunteers to produce 2 pages of random text using a sample of invented symbols that they were given to use and then see what the results look like. It would be interesting to see what commonality there are between the different nonsense texts that the different volunteers produced. I suppose for it to be a proper test the volunteers would have to be given some kind of time constraint to produce the 2 pages. They would also have to be forbidden from using any tools such as dice or coins in order to generate the random text. Maybe also all working would have to be done in their head i.e. with a pad to do working on.
I ask this as I hypothese that the volunteer might produce text broadly speaking structured in the kind of way the more repetitive examples of Voynich text is. Phenomena like "copying" or repeating "randomish" sections might be done in order to increase speed of text product. I wonder also if having words structured in essentially to same way, but with 1 or 2 letter differences might be a natural thing to produce.
This question interests me as I have mentioned that I think the Voynich contains a mixture of real text and nonsense. I doubt the nonsense text was generated by any mechanical means such as a cardan grille or anything else, but was just human generated.
Now I don't have 20 volunteers, so this is more of a thought experiment than a proper experiment.
I know discussion has been made of algorithms to generate Voynichese like text. I am not sure of the scope in this context for algorithms to simulate human generated text, just because simulating human behaviour is difficult and complex.(though some human behaviour is easy to model.)
RE: Human generated nonsense text - davidjackson - 12-04-2020
Humans are terrible at random generation.
The problem seems to be short term memory. When we start a repetitive random generation, we tend to remember the last three or four numbers and try to avoid them. So we fall into a pattern where we jump from one end to another. (An incredibly simplified explanation!)
Interestingly enough, random number generation is a test for different types of mental disorders. Schizophrenic patients show a very different pattern of random generation than those of non-schizophrenic people. Likewise autistic people show a distinct style when asked to produce random generation (usually they are unable to comprehend that they are not producing random numbers).
If you are interested in the subject, there are a vast number of scholarly articles on the web.
RE: Human generated nonsense text - voynichbombe - 12-04-2020
also, there has been an experiment about it on this forum recently. My mid-term memory does not spit up the thread name and also not the initiator, as well as I do not remember to have seen the output of it. I'll let it rest for one night.
RE: Human generated nonsense text - RobGea - 12-04-2020
Yes a thread by MarcoP:
Collecting meaningless handwritten text
You are not allowed to view links. Register or Login to view.
A little post by me:
You are not allowed to view links. Register or Login to view.
Actually computers have trouble with random as welll, they mostly use psuedorandom generators, though they may have improved these days.
Though this next relates to humans producing random numbers rather than text, its still illustrative.
You are not allowed to view links. Register or Login to view.
To quote:
"In conclusion, the pattern-based analysis using the Levenshtein-Damarau distance is both able to predict humanly generated random number sequences and to identify person-specific information within a humanly generated random number sequence."
So their algorithm can guess what 'random' number from 1-9 you are going to generate next, 27% of the time and tell the difference beteween a random number sequence that you generated vs a sequence from someone else 88% of the time.
RE: Human generated nonsense text - davidjackson - 12-04-2020
(12-04-2020, 01:26 AM)RobGea Wrote: You are not allowed to view links. Register or Login to view.Actually computers have trouble with random as welll, they mostly use psuedorandom generators, though they may have improved these days.Yes, but only another computer would ever notice
However, given that the Voynich corpus is not "random" in nature I'm not sure where this thread is going.
RE: Human generated nonsense text - bi3mw - 12-04-2020
That's right, the text in the VMS is, however, structured and far from random. There have already been various approaches to understanding this structure (also here in the forum). In my opinion, this is the key to understanding the text. This applies regardless of whether the VMS is written in any (contemporary) cipher or not. A cipher that actually generates "random text" would be anachronistic in itself anyway.
RE: Human generated nonsense text - Mark Knowles - 12-04-2020
(12-04-2020, 02:32 AM)davidjackson Wrote: You are not allowed to view links. Register or Login to view.(12-04-2020, 01:26 AM)RobGea Wrote: You are not allowed to view links. Register or Login to view.Actually computers have trouble with random as welll, they mostly use psuedorandom generators, though they may have improved these days.Yes, but only another computer would ever notice
However, given that the Voynich corpus is not "random" in nature I'm not sure where this thread is going.
I think my point is that truly "random" is probably a theoretical idea. As has been pointed out in practice we use pseudo-random numbers. The Voynich text is not "random", but all of it or some of it may be the product of an attempt to produce "random" text. My point is that it is very difficult for a human to produce "random" text without access to tools and so a human being endeavoring to produce random text at a reasonable rate will inevitably produce text with a lot of structure.
As an exercise one might want to try to produce 2 pages of random words without the use of tools and at a reasonable rate and then see what one produces.
I think one's inclination would be to create structures within which there is some "randomish" element.
RE: Human generated nonsense text - davidjackson - 12-04-2020
As I remember from when I attempted the test references above, it was quite difficult for me. Not that that means anything, many things are!
However, I think we need to think about the mindset of the times. How would a 15th century intelectual go about producing a hoax text? Would they? Would it even make sense for them to attempt it, to produce gibberish on the scale of the voynich?
RE: Human generated nonsense text - Mark Knowles - 12-04-2020
(12-04-2020, 05:37 PM)davidjackson Wrote: You are not allowed to view links. Register or Login to view.As I remember from when I attempted the test references above, it was quite difficult for me. Not that that means anything, many things are!
However, I think we need to think about the mindset of the times. How would a 15th century intelectual go about producing a hoax text? Would they? Would it even make sense for them to attempt it, to produce gibberish on the scale of the voynich?
Well, of course my own personal perspective, as I have mentioned elsewhere, is that the Voynich contains a mixture of meaningless filler text and meaningful real text.
So I suppose a more exact formulation of the test would be one where someone is given a shortish text, say half a page, and they must produce 2 pages of text containing the shortish text and filled in with a page and a half of nonsense words such that it is difficult for someone to tell which is which. All text would then be written in a made up script with a one to one mapping of imaginary characters to letters of alphabet. Then the question becomes what would be the likely pattern and layout of the meaningless text that the author produces? I think it won't be purely random, but how will it be structured?
RE: Human generated nonsense text - -JKP- - 12-04-2020
It's not easy to come up with a system that quickly and easily generates "legal" Voynichese (other than taking full tokens and just copying them in different arrangements but that presumes already having a model in front of you from which to copy).