Philosophy Discussion Forums

The dictionary is a list of definitions. It has up to 170,000 words and 170,000 definitions for those words. Isn't this similar to a set of equations with X variables and X equations? Whenever the variables are the same number as the equations, the variables can be solved for.

Once the equations are programmed in, you would just have to tell it to solve for "beauty" (or "define beauty") and everything remotely connected to "beauty" would branch out into more and more ideas until all the beautiful things in the world are part of the definition. If this has been done I want to see the results visually if possible. If it has not been done then it needs to be. A computer program could "solve for truth" and could answer philosophical questions, like what is truth, love, or beauty. [I think AI will be a good philosophy tutor in the future.]

If nobody is working on this problem, then PM me and we can work on it together!

Ah, the dictionary. The one place where nuanced understanding is absent. Dictionary "definitions" are specifically geared towards an audience who literally knows nothing of the subject matter and thus only contains the most elementary and basic information about the matter at hand.

What do you mean by "once the equations are programmed in"? What are the "equations"; just the respective definitions of the words?

Scott
Answering your second question: The equations are the mathematical truth-statements as reflected by the definitional statements. These definitions would have to follow mathematically-interoperable logical English-grammar syntax rules so that it is more than just a collection of words. (This could be short-circuited by analyzing a thesaurus instead of a dictionary, although some specificity of meaning might be sacrificed.)
So, yes, the equations are the respective definitions of the words, altered only to capture English syntax.

Answering your first question: "Once the equations are programmed in" means that each word is assigned its mathematical meaning, while each variable remains a variable. Having all the definitions turned into equations lets you "solve" (or define) any one variable (word) apart from the others, so long as the word under question is used elsewhere in the dictionary. (This would not define, nor should it logically be able to define, new words.)

TryingMyBest wrote: ↑December 17th, 2018, 4:03 pmSo, yes, the equations are the respective definitions of the words, altered only to capture English syntax.

I believe this exists already. For example, you can go to Google and type in "define:X" where X is the word you want defined. Then the definition of X will be displayed. That is just one example. There are many online dictionaries which can give you the definition/"equation" of a word. You enter X (the word) and the online dictionary outputs the definition.

The google define function merely regurgitates the equation that was initially programmed in. I'm after much more than that.

By temporarily deleting the equation of "beauty" and asking the computer to generate a new one, the computer would be tasked with encompassing everything that beauty is, that is all references based on other definitions, and derive a new definition for the term. I think this definition would have some similarities and some differences with the original definition, but I sincerely doubt they would be equivalent. This could be done countless times with different words until "new" associations are uncovered. Perhaps truer and truer definitions would emerge with each iteration of this process.

Taking it to the next logical step: The dictionary is supposed to represent the truth about meaning. This could be visualized by assigning a 3-dimensional node to each word and by allowing the nodes to move in 3-D space as they are attracted to (other nodes that correspond to) their meaning. At first, all words would be given random (x,y,z) coordinates but words of similar meanings would quickly coalesce and form physically intriguing structures. Each connection to other words would serve as a level of attraction between the two nodes. I think that after a bit of time, the simulation would balance and there would be structures of meaning floating and swirling and interacting with other structures of meaning. Perhaps an entire structure would represent Reality while another structure would represent that which is false??? Perhaps virtues and values would find distinct places and be able to be categorized more easily. Perhaps emotions could be quantified. Perhaps much more could be learned. To visualize the true connections of all the words in the dictionary is a powerful tool. If I knew how I would build it. I'm sure the folks at Google have the technical expertise, of course, they already have the equations, I just wish someone would build this tool so that I, and others, might better understand the nature of Truth by definition.

We describe our world by using words. I see value in letting those words and definitions work for us to that we may uncover Truth that we weren't even looking for.

TryingMyBest wrote:By temporarily deleting the equation of "beauty" and asking the computer to generate a new one, the computer would be tasked with encompassing everything that beauty is, that is all references based on other definitions, and derive a new definition for the term...

I'm no expert on language interpretation algorithms, but what you're describing in this post sounds a little bit like a neural network system for the interpretation of language. Possibly similar to systems that have already been developed.

But rather than analysing a static, out-of-date-the-moment-it-is-printed dictionary, a system like that can presumably make much more progress by analysing the ways in which words are actually used by people in real communication. In this case, there would be no concept of a dictionary being "solved" because the meanings of words, in use, are continually shifting. As you've described in your second paragraph, the clouds of words would float around in "word-space" as their meanings and associations with other words gradually changed.

It would be interesting to write software to create a visual representation of this - a dynamically changing 3 dimensional word-cloud that constantly trawls the internet consuming word associations.

Actually, I think Google Translate already uses an algorithm along these lines.

https://en.wikipedia.org/wiki/Google_Ne ... ranslation

It's an interesting subject. I bought a book about Deep Learning a while ago with the intention of doing a bit of research on it. It's called "Deep Learning (Adaptive Computation and Machine Learning Series)" and looks promising. I definitely intend to get around to reading it soon.

TryingMyBest wrote: ↑December 18th, 2018, 3:43 pm The google define function merely regurgitates the equation that was initially programmed in.

No, it doesn't.

Scott wrote: ↑December 19th, 2018, 11:25 am

TryingMyBest wrote: ↑December 18th, 2018, 3:43 pm The google define function merely regurgitates the equation that was initially programmed in.

No, it doesn't.

I don't mean to quibble with you about this but Google uses the Oxford English Pocket Dictionary for its results. https://en.oxforddictionaries.com/ is the source. The definitions exactly match.

If there is a dictionary that populates its own definitions from a dictionary that has been "solved", I'd very much like the reference. I think it would be a great tool for philosophy.

Much like how Google search results in general are determined, I believe Oxford also uses soft AI and complex computer data analysis to find and analyze the information on the "150 million words" or so that they track, as it is presumably way too much data for a team of humans to possibly process. Here is some information about it: https://www.oxforddictionaries.com/our-story/corpus

In any case, TryingMyBest, I believe you are right; I was incorrect to state or imply that Google's define function (through Oxford) does not primarily depend on a human-developed input and/or verification of definitions of words, but that's a moot point.

Scott wrote:What do you mean by "once the equations are programmed in"? What are the "equations"; just the respective definitions of the words?

TryingMyBest wrote:So, yes, the equations are the respective definitions of the words, altered only to capture English syntax.

Scott wrote:I believe this exists already. For example, you can go to Google and type in "define:X" where X is the word you want defined. Then the definition of X will be displayed. That is just one example. There are many online dictionaries which can give you the definition/"equation" of a word. You enter X (the word) and the online dictionary outputs the definition.

TryingMyBest wrote:The google define function merely regurgitates the equation that was initially programmed in.

That's what we're talking about: the "programming in" of equations (a.k.a. word definitions) so that you can input the word/"variable" and get the definition/"equation".

That programming in of equations has been done by Oxford and others, and so you can go to countless online dictionaries or Google and type in X and the computer returns the definition of X.

TryingMyBest wrote: ↑December 19th, 2018, 2:13 pmIf there is a dictionary that populates its own definitions from a dictionary that has been "solved", I'd very much like the reference. I think it would be a great tool for philosophy.

I'd suggest that dictionaries are like holy books, read around a bit and you'll find one that you like, but for much the same reason oyu'll never get the others to agree that one is 'right'.

Why not simply say that any word whose meaning is unknown or only vaguely so is defined generically in a dictionary which becomes more specifically defined in context. Only if the meaning is formalized from a remote source is an algorithm required. As a function it only adds complexity on what is simply a word inquiry.

I would assume any well done language has a dictionary fully solvable by mathematics, given you understand at least ONE component of the dictionary

I would consider your idea to be the standardization of the use of the language, based on a numerical value of appropriateness.
(The Truth IS, that which actually is , and verbal truth , IS that which is said ,attempting to abide what actually exists.)
Words are ballpark approximations which depend upon perspective.
So in that vein, yes, it would be possible if everyone cooperated , to arrive at a highly standardized version of English, yet the truth itself is going to forever be imperfectly defined.