upper ontology

Comment by Stephen Young on February 24, 2011 at 2:57am

Definition of the "broadest" or any element/concept is useful, sure, but I don't believe it's necessary for locating that concept in any "human-focused" ontology or map - or for subsequent use of it.  Most concepts are defined for us by their relationships with others. 

Think about how a one-year-old learns about and defines the concept of "disapproval" for example.  It will be based purely around the aversive behaviours of his or her care-giver.  Only later will he learn enough to be able to define the concept in the way we would.  But that won't stop him from attaching that "disapproval" concept to his own behaviours and (hopefully) changing them.

I think it's important that we allow in any knowledge structure a place for concepts we don't (initially) properly define.  I believe that's how humans build their own internal models of the world.  The map behind wik.me, for example has millions of these.

Comment by Ferenc Kovacs on February 24, 2011 at 9:32am

 

SY:Definition of the "broadest" or any element/concept.

FK: It is necessary, because to quote myself:

Attention is the weakest, but a minimum form of connecting.
It is also the minimum price you pay for getting connected.
Getting connected means exchanging information. "Information is transformation"
(Bela Hamvas) Also, you have not defined (we have not agreed on anything yet, including words as concept or element, and your phrase "broadest". I would like you to tell me what the simplest concept is in your repertory of 1. words you use 2. the repertory of concepts that you have in your mind 3.  The chunk of reality that you refer to by using the items under 1) and 2) as mutually connected.

SY: locating that concept in any "human-focused" ontology or map.

FK: Surely there us a difference between locating something in your mind or in the collection of words in an ontology or a map. Number one, you do not have have direct (manual, physical) access to anything stored in your mind or stored anywhere in your body, number two, an ontology of what exist should be connected to what really exist, which need to be named in a common language, number three, an ontology or its verbal or symbolic representation is not a map without a scale, which is a reference object.

SY: allow in any knowledge structure a place for concepts we don't (initially) properly define.

FK: The fact that we do not define them only indicates that we have a pre-language model of thinking used by animals and kids the elements of which need also be defined or recognised. Moreover, need to be aligned to.

SY: how humans build their own internal models of the world

FK: Even if you do not know how that is done, we can assume that they use building blocks, if we stay with your metaphor of construction and they perform operations as see in reality checks.

Cheers

Comment by Ferenc Kovacs on February 28, 2011 at 3:45pm

As i have sent you in a private mail, you need to start off from definitions, as NLP is a language, it must be brought into a normal form, just as formal languages to make sense in the course of semantic analysis and not syntax parsing. This is one of the misconceptions about language as seen by logicists and mathematicians. They do not realize that number is an abstraction, yet it has a form) and as an abstraction it does not exist without a tangible form of objects. There are two worlds to be handled together - one as experienced based on perception and the sensory organs, and another one creating mental objects with language (not just words) their knowledge representations. But the outcome of mental operations , as the case is with words, etc. are again tangible and thus may be sorted as objects in real life can be. But they are neither sorted that way, nor are they stored and represented that way in the mind (brains). It is false to assume that they are mapped, when mapping calls for scale, which is reality and not abstract, hence wrongly represented by graphs that to do not represent reality of any sort, but an old paradigm , or even just a metaphor as the tree, as in the Bible. But reading the Bible or the proper authors in Philosophy will show you other symbolic representations that will give you a better understanding of what is going on here in terms of semantic primitives.

Comment by Ferenc Kovacs on February 28, 2011 at 3:47pm

What is knowledge? The big picture

 

We are said to be living in a knowledge society while we even do not what knowle is. After Watson some think that knowledge is knowing a word to be given as an answer to a question which is drived from a number of facts recorded in a computer. So is knowledge the right word said at the right place at the right time?

 

Last year the number of English words passed the one million mark, wherever form they they are stored and tallied. But an average person uses 1-2 thousand words, the smart ones over 3 thousand and some authors make in the 20-30 thousadn region. Obviously it cannot be the words, so ned to check in the dictionaries, they may have the answer in the from of a defintion. Here is a sample of what you can get:

 

“All the facts that someone knows about a particular subject” Why all?

“All the facts that are known about different things or about life generally” Why so vague object?

“Epistemically praiseworthy, non-propositional procedural elements of a cognitive system” Why so “technical”?

“Thought to underlie abilities where performance of a task is consistently better than chance” Why related to gambling?

 

Knowledge:

S: (n) cognition, knowledge, noesis (the psychological result of perception and learning and reasoning)

Try cognition:

S: (n) co

Comment by Ferenc Kovacs on February 28, 2011 at 3:49pm

 to be added to the previous comment

 

S: (n) cognition, knowledge, noesis (the psychological result of perception and learning and reasoning)

Better than nothing! At lest we have recursion! But I am not happy, so I am submitting my version of the definition of knowledge

 

It would be nice to have a chance for correcting input after checking too late

 

Comment by Stephen Young on March 1, 2011 at 9:54am

Since we can't see what knowledge looks like inside the mind of a human I guess we can only define it in terms of what we can do with it.  You can tell I'm from the Behaviorist school of thought ;-)

Fortunately (for me at least) I don't believe that we need a comprehensive definition of a concept before we can make use of it.  I know we differ on this point - but I really haven't seen an argument from you that supports your many assertions on this matter.  I think such an argument would be very useful here.

Comment by Ferenc Kovacs on March 1, 2011 at 11:22am

I also hold the view you share below. But. We have the usual triplets of meaning, the referent, the sign and the concept, the last one being the one in the mind, and the whole as meaning. Looking at this triplet that way leaves only concept to be the difficult one to represent - all we know that it is related to the former two as two entities of "material nature". But if you introduce the concept of mental operation, you will get a different triplet, one the members of which can be defined recursively. By being able to define them recursively, I make them the building blocks of a tesselation that allows all the lower level triplets to be nicely interfaced and integrated in an open edged surface what and how knowledge is stored as information on the surface of materials. Within this surface however domain ontologies are not represented as graphs or tree any more, but cycles,  as that is a "more realistic" and practical form of structuring than graphs. First of all they have the advantage of scale, second, they can be sorted better, and with respect to the actual semantic primitives, they are possible to interpret across all domain ontologies, hence a useful tool for learning in any direction by traversing freely and reaching the outer edges where the current state of knowledge finishes. Besides, such an ontology allows you to combine and make a distinction between physical objects and mental objects making reality checks easier. And finally, you have the advantage of exposing relations that are normally hidden and underrepresented by Boolean algebra, (AI or formal logic) but which come in the fewer number than properties which in turn come in fewer number than objects, and which are sorted in nature according to the law of Physics, which are totally ignored in the cognitive models of mind and knowledge representations, because they are influenced by thinkers who believe that abstract things exist, yet without space and time properties and are not accessible to per

Comment