Overcoming Quine's objection to the analytic / synthetic distinction

Known unknowns and unknown unknowns!

Moderators: AMod, iMod

Post Reply
Skepdick
Posts: 14362
Joined: Fri Jun 14, 2019 11:16 am

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by Skepdick »

PeteOlcott wrote: Mon Apr 06, 2020 12:18 am For the whole body of knowledge that can be fully expressed using language: **
The only way that people know the meaning of words is that these words are defined using other words.
Pete, that's how dictionaries work. The thing is - dictionaries can handle the fact that one word can have many meanings.

Your system can't.

PeteOlcott wrote: Mon Apr 06, 2020 12:18 am The only aspects of knowledge left out are first-hand physical sensations from the sense organs.
So that's everything that matters then. Know-how. Intuition. All the things that people ACTUALLY use.
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 12:22 am
PeteOlcott wrote: Mon Apr 06, 2020 12:18 am For the whole body of knowledge that can be fully expressed using language: **
The only way that people know the meaning of words is that these words are defined using other words.
Pete, that's how dictionaries work. The thing is - dictionaries can handle the fact that one word can have many meanings.

Your system can't.
Do you know what globally unique identifiers are?
Skepdick
Posts: 14362
Joined: Fri Jun 14, 2019 11:16 am

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by Skepdick »

PeteOlcott wrote: Mon Apr 06, 2020 12:26 am Do you know what globally unique identifiers are?
Yes. We have had that discussion. You are being forgetful.

I am trying to explain to you that absolutely everything you are talking about exists today.
Not even theoretically - practically. We. have those systems - they are implemented exactly the way you describe.

There is nothing remotely human about a MySQL database.

You want unique identifiers? MySQL has that.

https://dev.mysql.com/doc/refman/8.0/en ... ue-id.html

You want uniqueness constraints? MySQL has that.

https://dev.mysql.com/doc/refman/8.0/en ... y-key.html
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 12:35 am
PeteOlcott wrote: Mon Apr 06, 2020 12:26 am Do you know what globally unique identifiers are?
Yes. We have had that discussion. You are being forgetful.

I am trying to explain to you that absolutely everything you are talking about exists today.
The sum total of all human knowledge that can be expressed using language
has been carefully structured in single structure such that algorithmic processing
equivalent to the human degree of comprehension of the meaning of every word
is hardly more than a tree walk?
Skepdick
Posts: 14362
Joined: Fri Jun 14, 2019 11:16 am

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by Skepdick »

PeteOlcott wrote: Mon Apr 06, 2020 12:46 am The sum total of all human knowledge that can be expressed using language
has been carefully structured in single structure such that algorithmic processing
equivalent to the human degree of comprehension of the meaning of every word
is hardly more than a tree walk?
An algorithmic structure has no property called "comprehension" or "processing". It's just a structure - it's static. It does nothing. It interacts with nothing. It doesn't change.

Given any two words X and Y, all you are going to get out of your system is all the relationships between X and Y.

Mathematicians call this a hom-set.

In fact, this is true for all mathematics. This is all that Mathematical proof will ever give you - structural guarantees. It will NOT give you any behavioural/functional guarantees, and certainly no fitness-for-purpose guarantees.

Maths is structuralism. https://en.wikipedia.org/wiki/Structura ... thematics)
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 12:48 am
PeteOlcott wrote: Mon Apr 06, 2020 12:46 am The sum total of all human knowledge that can be expressed using language
has been carefully structured in single structure such that algorithmic processing
equivalent to the human degree of comprehension of the meaning of every word
is hardly more than a tree walk?
An algorithmic structure has no property called "comprehension" or "processing". It's just a structure - it's static. It does nothing. It interacts with nothing. It doesn't change.
Yes that is correct and the reason why I added this part:
algorithmic processing equivalent to the human degree of comprehension of the meaning of every word
is hardly more than a tree walk [of this static data structure].

The static structure is organized in such a way the a mere tree walk of this structure provides
[a human degree of comprehension of the meaning of every word].
Skepdick
Posts: 14362
Joined: Fri Jun 14, 2019 11:16 am

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by Skepdick »

PeteOlcott wrote: Mon Apr 06, 2020 12:58 am Yes that is correct and the reason why I added this part:
algorithmic processing equivalent to the human degree of comprehension of the meaning of every word
is hardly more than a tree walk [of this static data structure].
It is exactly what it is. Graph traversal.
PeteOlcott wrote: Mon Apr 06, 2020 12:58 am The static structure is organized in such a way the a mere tree walk of this structure provides
[a human degree of comprehension of the meaning of every word].
Graph-walkers have no "comprehension" - they are usually greedy.

If you want to learn about them - go look at the query optimizer in MySQL or Postgres. They are both open source.
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 12:59 am
PeteOlcott wrote: Mon Apr 06, 2020 12:58 am Yes that is correct and the reason why I added this part:
algorithmic processing equivalent to the human degree of comprehension of the meaning of every word
is hardly more than a tree walk [of this static data structure].
It is exactly what it is. Graph traversal.
PeteOlcott wrote: Mon Apr 06, 2020 12:58 am The static structure is organized in such a way the a mere tree walk of this structure provides
[a human degree of comprehension of the meaning of every word].
Graph-walkers have no "comprehension" - they are usually greedy.
None-the-less the walk of a very special graph that is a knowledge ontology of the set
of all knowledge that can be represented using language and carefully optimized for
machine processing of natural language semantics can demonstrate the comprehension
of the meaning of words.
Skepdick
Posts: 14362
Joined: Fri Jun 14, 2019 11:16 am

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by Skepdick »

PeteOlcott wrote: Mon Apr 06, 2020 1:11 am None-the-less the walk of a very special graph that is a knowledge ontology of the set
of all knowledge that can be represented using language and carefully optimized for
machine processing of natural language semantics can demonstrate the comprehension
of the meaning of words.
Your graph can't cope with ambiguity. Or adjectives. Or adverbs or any of the other nice things you get in English.

Formal languages give you only verbs (operators) and nouns (objects).
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 1:12 am
PeteOlcott wrote: Mon Apr 06, 2020 1:11 am None-the-less the walk of a very special graph that is a knowledge ontology of the set
of all knowledge that can be represented using language and carefully optimized for
machine processing of natural language semantics can demonstrate the comprehension
of the meaning of words.
Your graph can't cope with ambiguity. Or adjectives.

Formal languages give you only verbs (operators) and nouns (objects).
It is not that it can't cope with ambiguity, it eliminates ambiguity by design.
Ambiguous natural language expressions are translated into multiple mete-language
expressions.

Why would you believe that it can't handle the semantic meaning associated with
every part-of-speech?
Skepdick
Posts: 14362
Joined: Fri Jun 14, 2019 11:16 am

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by Skepdick »

PeteOlcott wrote: Mon Apr 06, 2020 1:18 am It is not that it can't cope with ambiguity, it eliminates ambiguity by design.
It doesn't. How do you resolve the ambiguity in Bachelor - unmarried man/bachelor of science?
PeteOlcott wrote: Mon Apr 06, 2020 1:18 am Ambiguous natural language expressions are translated into multiple mete-language
expressions.
Yes. Branching. How would the computer know which one you meant when you ask it "Bachelor(john)"

How would the computer decide which branch to take?

PeteOlcott wrote: Mon Apr 06, 2020 1:18 am Why would you believe that it can't handle the semantic meaning associated with
every part-of-speech?
Because you can't program a computer in English.
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 1:21 am
PeteOlcott wrote: Mon Apr 06, 2020 1:18 am It is not that it can't cope with ambiguity, it eliminates ambiguity by design.
It doesn't. How do you resolve the ambiguity in Bachelor - unmarried man/bachelor of science?
Like I already said GUID's.
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 1:21 am
PeteOlcott wrote: Mon Apr 06, 2020 1:18 am Ambiguous natural language expressions are translated into multiple mete-language
expressions.
Yes. Branching. How would the computer know which one you meant when you ask it "Bachelor(john)"

How would the computer decide which branch to take?
It would process the one that it has information on, or both or neither.
If neither then the answer is: "No" meaning that neither can be proved.
PeteOlcott
Posts: 1514
Joined: Mon Jul 25, 2016 6:55 pm

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by PeteOlcott »

Skepdick wrote: Mon Apr 06, 2020 1:21 am
PeteOlcott wrote: Mon Apr 06, 2020 1:18 am Why would you believe that it can't handle the semantic meaning associated with
every part-of-speech?
Because you can't program a computer in English.
That is a very lame answer considering you already acknowledged English "verbs" and "nouns".
Skepdick
Posts: 14362
Joined: Fri Jun 14, 2019 11:16 am

Re: Overcoming Quine's objection to the analytic / synthetic distinction

Post by Skepdick »

PeteOlcott wrote: Mon Apr 06, 2020 3:52 am Like I already said GUID's.
That solves absolutely nothing. The computer may have unique identifiers, but humans don't.

We infer meaning from context, we don't (need to) use GUIDs.

If I knew the GUID for the meaning of Bachelor (unmarried man), distinct from the GUID for Bachelor (person who wholds a bachelors degree).
If I could keep the GUID of every distinct meaning in my head, I could just speak in GUIDs, no?

But I don't speak in GUIDs, because it's easier to construct meaning on as-need basis than to remember GUIDs.
Last edited by Skepdick on Mon Apr 06, 2020 11:20 am, edited 3 times in total.
Post Reply