Probable it is very basic question, I'm in the beginning of understanding Gellish concepts.
As I'd noticed in TOPini, all synonyms of an object have the same Unique id. For example, "is a specialization of", "specialization of class", "is a subtype of" all are "1 146". What puzzles me is that reverse relationships "is a generalization of" and "is a supertype of" have the same id "1 146". I understand that each Gellish expression can be reversed, but I can not imagine any way I can use reverse relationship names correctly.
The only two roles defined for "1 146" are "subtype" for role-1 and "supertype" for role-2. nTherefore I can not use names "is a generalization of" and "is a supertype of" in any Gellish text or message in a sensible way, having in mind that any usage of "1 146" requires "subtype" for role-1 and "supertype" for role-2.
Is it correct interpretation of the table? Why wasn't it decided to give a separate ids to direct and reverse relationships or just don't allow reverse relationships to appear in the table?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
An expression in Gellish, such as 'car <is a subtype of> vehicle', is an expression of a fact. That fact is the case, independent of the way in which it is expressed. The inverse expression requires also an inverse sequence of the concepts car and vehicle. Thus the expression 'vehicle <is a supertype of> car' is another expression of the same fact.
In other words a <subtype-supertype> relation between A and B can also be expressed in a natural language in the inverse direction B - A by using the inverse phrase. Because both expressions express the same type of relation those phrases have the same relation type UID.
In TOPini it is defined that a relation requires two roles: the first role is played by an object at the left hand of the synonym expression, the second role points to the right hand object as role player. In the inverse expression the position of those role players is inverse: the first role then points to the right hand role player.
TOPini defines the meaning of the relation types. The other parts of the dictionary and other Gellish expressions use those relation types by classifying the relations.
In those expressions the roles are usually left implicit, but they are always played.
For example, in the expression of a fact such as: 'the Eiffel tower <is located in> Paris', the Eiffel tower has a role as 'located' and Paris has a role as 'location'.
Those two roles of the relation type <is located in> are defined in TOPini.
Regards,
Andries
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
You de facto have 2 different ontologies for speaking about roles: 1. "Ontological": Role-1 and Role-2. 2. "Representational": LeftHand and RightHand. And no explicit mappings (е.g. subtitles "Role-1" and "Role-2" for Left Hand Object and Right Hand Оbject column titles in Gellish Table).
I suggest to add subtitles to column titles.
And what about synonymy relation: is direction of this relation important (e.i. exist primary name and secondary ones)?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
There is only one ontology that specifies which object shall be the role-1 player and which object shall be the role-2 player in order to have an unambiguous interpretation.
You could also specify the semantics without using left and right hand objects as follows:
A <is a player of> role-1
B <is a player of> role-2
However there is also a grammatical sequence in natural languages. The one used in the normal table is the English sequence.
I understand that there are also languages with a different normal sequence (e.g. Japanese).
But depending on the phrase that is used the player of role-1 can appear in the left as well as in the right hand column. So the term role-1 should not be included in the column header.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
There can not be a single explicit mapping: the roles "change hands" for an inverse relation! If we exchange statements between computers using only id's - we have to identify roles with hands on every use of relations.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The user interface should include an indication of the correct phrase. The software may change the sequence for internal storage in the standard sequence of role-1 relation role-2. This standard sequence can also be used for exchange of UID's only.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Probable it is very basic question, I'm in the beginning of understanding Gellish concepts.
As I'd noticed in TOPini, all synonyms of an object have the same Unique id. For example, "is a specialization of", "specialization of class", "is a subtype of" all are "1 146". What puzzles me is that reverse relationships "is a generalization of" and "is a supertype of" have the same id "1 146". I understand that each Gellish expression can be reversed, but I can not imagine any way I can use reverse relationship names correctly.
The only two roles defined for "1 146" are "subtype" for role-1 and "supertype" for role-2. nTherefore I can not use names "is a generalization of" and "is a supertype of" in any Gellish text or message in a sensible way, having in mind that any usage of "1 146" requires "subtype" for role-1 and "supertype" for role-2.
Is it correct interpretation of the table? Why wasn't it decided to give a separate ids to direct and reverse relationships or just don't allow reverse relationships to appear in the table?
This is a very good basic question!
An expression in Gellish, such as 'car <is a subtype of> vehicle', is an expression of a fact. That fact is the case, independent of the way in which it is expressed. The inverse expression requires also an inverse sequence of the concepts car and vehicle. Thus the expression 'vehicle <is a supertype of> car' is another expression of the same fact.
In other words a <subtype-supertype> relation between A and B can also be expressed in a natural language in the inverse direction B - A by using the inverse phrase. Because both expressions express the same type of relation those phrases have the same relation type UID.
In TOPini it is defined that a relation requires two roles: the first role is played by an object at the left hand of the synonym expression, the second role points to the right hand object as role player. In the inverse expression the position of those role players is inverse: the first role then points to the right hand role player.
TOPini defines the meaning of the relation types. The other parts of the dictionary and other Gellish expressions use those relation types by classifying the relations.
In those expressions the roles are usually left implicit, but they are always played.
For example, in the expression of a fact such as: 'the Eiffel tower <is located in> Paris', the Eiffel tower has a role as 'located' and Paris has a role as 'location'.
Those two roles of the relation type <is located in> are defined in TOPini.
Regards,
Andries
You de facto have 2 different ontologies for speaking about roles: 1. "Ontological": Role-1 and Role-2. 2. "Representational": LeftHand and RightHand. And no explicit mappings (е.g. subtitles "Role-1" and "Role-2" for Left Hand Object and Right Hand Оbject column titles in Gellish Table).
I suggest to add subtitles to column titles.
And what about synonymy relation: is direction of this relation important (e.i. exist primary name and secondary ones)?
There is only one ontology that specifies which object shall be the role-1 player and which object shall be the role-2 player in order to have an unambiguous interpretation.
You could also specify the semantics without using left and right hand objects as follows:
A <is a player of> role-1
B <is a player of> role-2
However there is also a grammatical sequence in natural languages. The one used in the normal table is the English sequence.
I understand that there are also languages with a different normal sequence (e.g. Japanese).
But depending on the phrase that is used the player of role-1 can appear in the left as well as in the right hand column. So the term role-1 should not be included in the column header.
There can not be a single explicit mapping: the roles "change hands" for an inverse relation! If we exchange statements between computers using only id's - we have to identify roles with hands on every use of relations.
The user interface should include an indication of the correct phrase. The software may change the sequence for internal storage in the standard sequence of role-1 relation role-2. This standard sequence can also be used for exchange of UID's only.