Re: VMs: Re: T-maps later than thought?

[elvogt@xxxxxxxxxxx]

Zitat von Barbara Barrett <barbarabarrett@xxxxxxxxxxxxxxx>:

> ...
> Barbara Babbles;
> Perhaps if you've got the English name of the map I can look it up on the
> net; most of the significant medieval maps are one line somewhere.

It was just a page in a book (oh which I don't know the title anymore); anyway 
I don't think it's spectacular enough to warrant an exhaustive search.

> ...
> > Uhm... No.
> > I recommend Barlow, Roger J.: "Statistics" (ISBN: 0471922951) as a good
> > beginner's book on the subject.
> 
> I had to give up my JP training due to ill health, but in fingerprint
> evidence only 14 points of exact correspondence ( and a vague resemblance
> fore the rest of the prints) were necessary to declare a Match out of
> possible hundreds points of correspondence/features.
> 

I wasn't attacking the fingerprint technique, but your assumption of "6 matches 
make a proof".

Two points:

*) First of all, strictly speaking statistics never prove anything; they only 
refute. (So to "prove" your point you have to disprove all alternatives. In the 
matematical sense you only prove the suspect is guilty if his fingerprints 
match, and you've checked everybody else's to make sure they don't.)

*) The significance of a match depends on the feature and on the sample. If 
you've got ten suspects, one of which is female, and you know that the 
perpretator was a woman, then the significance is high. If you know it was a 
man, you only rule out 10% of the suspects -- low significance.

What you have in the alleged T-O-map is... three strokes of the quill.

> > In a nutshell, two eyes, a nose, a mouth and two ears are characteristical
> for
> > a human being. But they might as well belong to a monkey, a dog, or a
> > pig. "100%" means little to nothing.
> 
> Not if you are specific. IE *Human* nose, eyes, nose, mouth, ears can not be
> found on any other beast. And the correspondences are specific, not general.
> 

Yes, but f68v3 is _not_ specific. All you've got is a circle with two lines 
running through it. I could show you the exact same shape from my last 
PowerPoint presentation. It's a pie chart with percentages -- does this prove I 
was actually drawing a T-O map?

No, what it proves is, that the shape of a T-O can occur in various 
circumstances, and to precisely identify it we have to have context 
information -- like from the labels.

> ...
> Ah, but you're making the assumption that the labels on a T-O map always
> identify its constituent parts, which is not true. 

No, I don't. I'd be happy as well if the African label read "Here Montgomery 
beat Rommel on the head in '43", or if the Asian label read "Home of the 
Vodka". Because then we'd have a clear identification of the shapes as land 
masses. Currently, we don't have that.

> ...
> Furthermore the f68v3 isn't just a circle with two lines; ...

What you have is a circle, divided in a 180° and two 90° segments. Everything 
else is your interpretation. Orientation is arbitrary, and there is only one 
way to arrange these segments, so there's not really a whole lot of information 
in there.

You corroborate your interpretation by the second circle running around it, 
saying "it's the great ocean, and that fits in the T-O interpretation." At the 
same time you ignore the four spirals running outward -- arbitrarily dividing 
the image in parts which belong to the suspected T-O, and parts which do not.

What you're doing is: You show that the interpretation as T-O is consistent. 
(You found no contradictions.) This is not the same as proof.

> 
> Barbara
> 

I'm not quite sure if there's really anybody interested in the argument except 
for the two of us... sorry if we're wasting bandwidth.

Cheers,

   Elmar



-------------------------------------------------
debitel.net Webmail
______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list