General semantics

#11

Bob Badour wrote:

Quote:

Nilone wrote:

On May 18, 11:19 pm, Erwin <e.sm... (AT) myonline (DOT) be> wrote:

And I fail to see what "general semantics and it's correspondences to
the relational model?" has to do with this delusionary nonsense of
"relations being flat".

I read "multi-dimensional order" as "n-dimensional data structure",
and it seemed to me that Korzybski aptly described the relational
model and its possible application to language and mental models of
empirical data. As I progress through the book, I find that
perspective reinforced.

But in natural language, it is perfectly normal and perfectly
acceptable to employ idiomatic expressions and figurative speech. In
discussions which are supposed to be scientific, that is much less the
case.

I suspect this thread leans too far into the philosophical for the
regulars of c.d.t. I derive my desire to understand the relational
model from its value as a metaphysical model of reality.

Have you read William Kent's /Data and Reality/ ?

I'd say the breakthrough that made Codd famous was that he saw clear,
straightforward ways to implement his structures, as well as his
assimilation of several classical techniques. They had general
usefulness even if there remains much in human experience that they
don't capture, eg., emotion, ethics, sensory feelings. I suppose you
could say that Codd's relations 'fall flat' when it comes to imparting
feelings but that wasn't his purpose. I don't mean that as a joke,
personally it's my main reason for trying to maintain a
'machine-centric' view when talking about db theory - the other
territory is too ephemeral, eg., the 'models' we talk about can't
replicate anything but themselves.

No doubt Alfred Korzybski couldn't have been hip to the machinery that
Codd was but putting that aside the question for me would be what
implementation might his writing imply, something akin to Codd's or
something else?

Eg., I'd be curious as to who first talked about unary relations, which
seem an essential part of Codd's breakthrough. Seems to me that
anything 'new' needs to be compared to what Codd wrote (though
apparently he had such a practical bent that he saw no need for nullary
relations).

(I don't remember seeing them in Bertrand Russell's 'Introduction to
Mathematical Philosophy' which might also be interesting to some of us.
Like Korzybski, he came before Codd and Kent's generation, I think he
might have stopped writing on such topics by the time Godel made his own
breakthrough. I don't know if he gave up because of various logical
difficulties or just found other interests, maybe he saw similar
obstacles to machine 'understanding', eg., how could a machine ever
smell? Even that's a presumptuous question since we know that humans
can never smell the way Rosie the dog can.)

#12

paul c wrote:

Quote:

Bob Badour wrote:

Nilone wrote:

On May 18, 11:19 pm, Erwin <e.sm... (AT) myonline (DOT) be> wrote:

And I fail to see what "general semantics and it's correspondences to
the relational model?" has to do with this delusionary nonsense of
"relations being flat".

I read "multi-dimensional order" as "n-dimensional data structure",
and it seemed to me that Korzybski aptly described the relational
model and its possible application to language and mental models of
empirical data. As I progress through the book, I find that
perspective reinforced.

But in natural language, it is perfectly normal and perfectly
acceptable to employ idiomatic expressions and figurative speech. In
discussions which are supposed to be scientific, that is much less the
case.

I suspect this thread leans too far into the philosophical for the
regulars of c.d.t. I derive my desire to understand the relational
model from its value as a metaphysical model of reality.

Have you read William Kent's /Data and Reality/ ?

I'd say the breakthrough that made Codd famous was that he saw clear,
straightforward ways to implement his structures, as well as his
assimilation of several classical techniques. They had general
usefulness even if there remains much in human experience that they
don't capture, eg., emotion, ethics, sensory feelings. I suppose you
could say that Codd's relations 'fall flat' when it comes to imparting
feelings but that wasn't his purpose. I don't mean that as a joke,
personally it's my main reason for trying to maintain a
'machine-centric' view when talking about db theory - the other
territory is too ephemeral, eg., the 'models' we talk about can't
replicate anything but themselves.

No doubt Alfred Korzybski couldn't have been hip to the machinery that
Codd was but putting that aside the question for me would be what
implementation might his writing imply, something akin to Codd's or
something else?

Eg., I'd be curious as to who first talked about unary relations, which
seem an essential part of Codd's breakthrough. Seems to me that
anything 'new' needs to be compared to what Codd wrote (though
apparently he had such a practical bent that he saw no need for nullary
relations).

(I don't remember seeing them in Bertrand Russell's 'Introduction to
Mathematical Philosophy' which might also be interesting to some of us.
Like Korzybski, he came before Codd and Kent's generation, I think he
might have stopped writing on such topics by the time Godel made his own
breakthrough. I don't know if he gave up because of various logical
difficulties or just found other interests, maybe he saw similar
obstacles to machine 'understanding', eg., how could a machine ever
smell? Even that's a presumptuous question since we know that humans
can never smell the way Rosie the dog can.)

If you mean my Rosie, she still smells a little like skunk, which is
probably achievable for most humans.

#13

Snipped

Quote:

I don't mean that as a joke,
personally it's my main reason for trying to maintain a
'machine-centric' view when talking about db theory - the other
territory is too ephemeral, eg., the 'models' we talk about can't
replicate anything but themselves.
Exactly. The fundamental value of RM is a inherent limitation of

subjectiveness at system analysis and design time.

#14

On May 19, 2:37*pm, Cimode <cim... (AT) hotmail (DOT) com> wrote:

Quote:

*mental models* *empirical data* ?
Quite frankly, I see no advantage into creating obscure terminology
for already established meanings.

My apologies. I chronically abuse terminology, which probably
indicates that my words exceed my content. I will appreciate it if
you can provide the correct phrase so that I can improve.

Quote:

I have hard time understanding how one can better comprehend a theory
starting from a potential application, if such application is
possible. *Do you think one can truly understand the process of
uranium enrichment by exposing oneself to an atomic explosion ?

I only said I derive my motivation from metaphysics, and I should
include programming as well.

Quote:

cdt is not the best place to discuss philosophical and metaphysical
implications of RM, if any.

In that case, I'll try to stick to the usual topics in the future,
although I find it difficult to separate them. If I stray again, I'll
appreciate your correction.

#15

On May 19, 4:12*pm, Bob Badour <bbad... (AT) pei (DOT) sympatico.ca> wrote:

Quote:

Have you read William Kent's /Data and Reality/ ?

No, but the excerpts and information provided by Google look
tantalizing. If you recommend it, I'll place it very high on my list.

#16

Nilone wrote:

Quote:

On May 19, 4:12 pm, Bob Badour <bbad... (AT) pei (DOT) sympatico.ca> wrote:

Have you read William Kent's /Data and Reality/ ?

No, but the excerpts and information provided by Google look
tantalizing. If you recommend it, I'll place it very high on my list.

I still have not managed to get a copy for myself. However, a lot of
people, who I respect greatly, recommend it and recommend it in the
highest terms. It clearly influenced some of the best minds in the
industry. While I cannot make a direct recommendation at this time, you
can consider it recommended by better men than me.

Since you express an interest in the link between data and reality, I
suspect it is prerequisite reading for you. And if you choose not to
read it, I recommend you make sure it is, at least, in the bibliography
of whatever you are reading to show the author has some awareness of the
prior art.

#17

On May 19, 5:12*pm, paul c <toledobythe... (AT) oohay (DOT) ac> wrote:

Quote:

Eg., I'd be curious as to who first talked about unary relations, which
seem an essential part of Codd's breakthrough. *Seems to me that
anything 'new' needs to be compared to what Codd wrote (though
apparently he had such a practical bent that he saw no need for nullary
relations).

I did some checking and found http://fair-use.org/bertrand-russell...athematics/s27,
from which I snip and paste liberally:

"Peirce and Schröder have realized the great importance of the
subject ... their method suffers technically ... from the fact that
they regard a relation essentially as a class of couples, thus
requiring elaborate formulae of summation for dealing with single
relations. ... it was certainly from the opposite philosophical
belief, which I derived from my friend Mr G. E. Moore, that I was led
to a different formal treatment of relations."

Am I correct in thinking that Russell's 'single relations' refer to
unary relations? Although I didn't follow up all the references, some
further checking makes it seem as if Peirce first developed the idea.
According to http://en.wikipedia.org/wiki/Charles...atics_of_logic,
Codd studied under Burks who strongly advocated the ideas of Peirce,
so it seems likely that Codd would build on that foundation.

Searching on Burks netted me this paper:
http://citeseerx.ist.psu.edu/viewdoc...p=rep1&type=ps
(Peirce's Late Theory Of Abduction), which explores some of Peirce's
phenomenology.

Back to relations - from http://fair-use.org/bertrand-russell...athematics/s30,
"If u be any class which is not null, there is a relation which all of
its terms have to it, and which holds for no other pairs of terms."
If a unary relation describes a relation between a class and its
terms, and classes equate to the domains of relations, then can we /
should we allow the direct use of relations as domains? For example:

Carnivore = [x : Animal]
Wolf
Lion

PredatorPrey = [y : Carnivore, z : Animal]
Wolf, Rabbit
Lion, Deer

This goes against the adage "relations aren't domains", and we can
achieve the same via referential constraint expressions, which can
also express more complex relationships between the domains of
relations, but do we need the extra concept?

#18

On May 19, 12:26*pm, Nilone <rea... (AT) gmail (DOT) com> wrote:

Quote:

I did some checking and foundhttp://fair-use.org/bertrand-russell/the-principles-of-mathematics/s27,
from which I snip and paste liberally:

"Peirce and Schröder have realized the great importance of the
subject ... their method suffers technically ... from the fact that
they regard a relation essentially as a class of couples, thus
requiring elaborate formulae of summation for dealing with single
relations. ... it was certainly from the opposite philosophical
belief, which I derived from my friend Mr G. E. Moore, that I was led
to a different formal treatment of relations."

Am I correct in thinking that Russell's 'single relations' refer to
unary relations? *Although I didn't follow up all the references, some
further checking makes it seem as if Peirce first developed the idea.
According tohttp://en.wikipedia.org/wiki/Charles_Sanders_Peirce#Mathematics_of_logic,
Codd studied under Burks who strongly advocated the ideas of Peirce,
so it seems likely that Codd would build on that foundation.

Relation logic by Peirce and Schröder hasn't been mainstream research
area until Tarski. "Applications of Alfred Tarski's Ideas in Database
Theory" by Jan Van Den Bussche suggests several influences including
cylindric algebras and relation algebras. I'm not convinced about the
latter: Codd's relational algebra and Binary relation algebra have
very different look!

Also see "the origins of calculus of binary relations" by Roger
Maddux. Speaking of metaphysics and philosophical reading, much of
Peirce's "Logic of Relatives" is written in that style, so some
paragraphs taken out of context appear like crank musings from
sci.logic usenet forum. (This observation applies to the famous George
Boole manuscript as well.) This is why I find it easier to read it
digested via Maddux paper.

Quote:

Back to relations - fromhttp://fair-use.org/bertrand-russell/the-principles-of-mathematics/s30,
"If u be any class which is not null, there is a relation which all of
its terms have to it, and which holds for no other pairs of terms."
If a unary relation describes a relation between a class and its
terms, and classes equate to the domains of relations, then can we /
should we allow the direct use of relations as domains? *For example:

Carnivore = [x : Animal]
* * * * Wolf
* * * * Lion

PredatorPrey = [y : Carnivore, z : Animal]
* * * * Wolf, Rabbit
* * * * Lion, Deer

This goes against the adage "relations aren't domains", and we can
achieve the same via referential constraint expressions, which can
also express more complex relationships between the domains of
relations, but do we need the extra concept?

"relations aren't domains" sounds like some dogma. In my system
domains are unary relations (or predicates if the term "relation" is
reserved for finite sets of tuples).

#19

Nilone wrote:

Quote:

On May 19, 5:12 pm, paul c <toledobythe... (AT) oohay (DOT) ac> wrote:
Eg., I'd be curious as to who first talked about unary relations, which
seem an essential part of Codd's breakthrough. Seems to me that
anything 'new' needs to be compared to what Codd wrote (though
apparently he had such a practical bent that he saw no need for nullary
relations).

I did some checking and found http://fair-use.org/bertrand-russell...athematics/s27,
from which I snip and paste liberally:

"Peirce and Schröder have realized the great importance of the
subject ... their method suffers technically ... from the fact that
they regard a relation essentially as a class of couples, thus
requiring elaborate formulae of summation for dealing with single
relations. ... it was certainly from the opposite philosophical
belief, which I derived from my friend Mr G. E. Moore, that I was led
to a different formal treatment of relations."

Am I correct in thinking that Russell's 'single relations' refer to
unary relations? Although I didn't follow up all the references, some
further checking makes it seem as if Peirce first developed the idea.
According to http://en.wikipedia.org/wiki/Charles...atics_of_logic,
Codd studied under Burks who strongly advocated the ideas of Peirce,
so it seems likely that Codd would build on that foundation.
...

That's not the first time I had the impression that Russell was
foretelling the future. Maybe there's a clue in a sentence you snipped:

"... This view is derived, I think, probably unconsciously, from a
philosophical error: it has always been customary to suppose relational
propositions less ultimate than class-propositions (or subject-predicate
propositions, with which class-propositions are habitually confounded),
and this has led to a desire to treat relations as a kind of classes."

Looks to me as if the 'error' he's talking about is either the same as
what Date and Darwen call a "Great Blunder" or some kind of
relation-valued attribute! (Can't remember if it would be the first or
second Great Blunder.)

#20

Tegiri Nenashi wrote:

Quote:

...
"relations aren't domains" sounds like some dogma. In my system
domains are unary relations (or predicates if the term "relation" is
reserved for finite sets of tuples).

In that case, unary relations have more operators than other relations,
not a complication I'd like to deal with.