Ai Dreams Forum

AI Dreams => General Chat => Topic started by: frankinstien on June 17, 2022, 08:30:07 pm

Title: Is natural language too verbose?
Post by: frankinstien on June 17, 2022, 08:30:07 pm
I've been trying to learn to speed read and the methods are obviously to increase the chunking size of reading multiple words, if not entire sentences rather than a single word at a time. I then thought what if we do the opposite. Literally shrink the amount of words in a sentence? I mean that's already happening with text messaging now, right?

Symbolic representation of data can be enhanced from natural language by applying concepts like adjectives into single expressions such as “an old man” could be represented as “man” and even the individuals ethnicity could be expressed by modifying other characters or again imposing another modification to the letter m, e.g. “man” could represent an older man from China that is married. This type of codification compacts information very elegantly and applying this kind of notion to adverbs could work as well. But imagine stating a young married man from china is running from the police as “man++” or a older married man was walking to the store as: “man+

You could change the plus sign colors for different location types. This kind of symbolization could shrink the amount of verbosity in current literature. I mean technical books could shrink from hundreds of pages to less than hundred.

Obviously we would still need to speak English in the verbose way we do now, but we could represent it in a compact way.

Title: Re: Is natural language too verbose?
Post by: MagnusWootton on June 18, 2022, 02:40:14 am
thats cool.

but we are being silly.   :D
Title: Re: Is natural language too verbose?
Post by: MikeB on June 19, 2022, 07:16:58 am
It can work to compress adjectives/modifiers (removing them from pattern sentences, and adding tags).

EG. "I walked a lovely lovely lovely lady in a red polkadot dress across the street."
to: "I walked a lady across the street."

But to uncompress the sentence into it's original form is another issue. If that needs to be done.
Title: Re: Is natural language too verbose?
Post by: MagnusWootton on June 19, 2022, 08:49:42 am
just make sure its the most obvious thing youve ever read, and then thats clear right?
Title: Re: Is natural language too verbose?
Post by: ivan.moony on June 19, 2022, 02:36:15 pm
Chinese calligraphy?

They have a bunch of base words of which more complex words are combined. I have a book on Chinese calligraphy history right here, but it's on German, and I don't understand it. Maybe it's time for me to learn German? Maybe from Spanish book?
Title: Re: Is natural language too verbose?
Post by: WriterOfMinds on June 19, 2022, 04:50:54 pm
I've seen it claimed that all human languages transmit information at roughly the same rate. So languages that convey less information per syllable and tend to have longer words are spoken faster, but do not convey concepts any faster. Conversely, a very information-dense language would have to be spoken or read slowly. The limiting factor is not the speed of communication but the speed at which the receiving brain can interpret and comprehend the data. Source:,the%20speed%20of%20Morse%20code.

So I propose that increasing the information density of text, by adding formatting modifiers to increase the information content of each symbol, would end up forcing people to read more slowly. Papers would be shorter but take just as long to consume.
Title: Re: Is natural language too verbose?
Post by: frankinstien on June 19, 2022, 06:41:19 pm
each language gravitated toward an average rate of 39.15 bits per second

That is the bit rate of the symbolism which could be very compressed information, where the more compression you have the more computational horse power required  to decrypt the information. So compacting language so that its less verbose means more information is conveyed but requires brain power to decrypt. In any case, even with current speed reading techniques it's a matter of chunking the information as groups of words rather than just single words.  Finding strategies to group multiple words so as to build the habit of consuming more information is the trick.  As my example proved such a technique could convey more information faster than using verbose approach.