Re: VMs: No stats no fun

["Jeff" <jeff@xxxxxxxxxxxxxxxxxxxxxx>]

Keep going Nick. Every dog has his day :-)


----- Original Message -----
From: "Nick Pelling" <incoming@xxxxxxxxxxxxxxxxx>
To: <vms-list@xxxxxxxxxxx>
Sent: 09 August 2003 12:06
Subject: Re: VMs: No stats no fun


> Hi Gabriel,
>
> At 11:14 09/08/2003 +0100, Gabriel Landini wrote:
> >What makes the statement of "making stats useless" sound objectable is
that
> >it seems to imply unawareness of the relevance of the statistical
properties
> >of the vms.
>
> Statistics does not equal causality. Statistical equals correlations.
> AFAICT, 91 years (or perhaps several hundred years, if you include
Kircher,
> etc) of statistical analysis of the VMs have given us *not one* definitive
> correlation with an external statistical source - such as natural
language,
> syllabic language, artificial language, cryptography, etc - that has had
> sufficient persuasive power to bring us to even a partial consensus.
>
> TTBOMK, the strongest inferences drawn to date have been "suggestive of"
> one or other of these various categories, without being at all definitive.
> I have seen it argued (in one of my stats textbooks) that many of the
> central successes of human endeavour in the 20th Century came about
> directly as a result of statistical analysis & the emerging statistical
> mindset - how does the VMS continue to resist assaults from this otherwise
> fearsome analytical machinery?
>
> Whether you like it or not, the manifest failure of our general
statistical
> exercise to resolve even the most basic questions points to one of three
> conclusions:-
> (a) the VMS is a truly strange beast which we are unable to categorise;
> (b) the statistical process by which we collect information is somehow
> flawed; or
> (c) the transcription we are using to apply stats analysis to the VMS is
> somehow flawed.
>
> Broadly speaking, (a) points to a conceptual failure, (b) to a processual
> failure, and (c) to a representational failure. You seem to think I'm
> describing (b), when I'm actually describing (c). I think stats are
great -
> but you have to be analysing the right thing for them to move you forward.
>
> >Sounds like if you are intending to dismiss all currently gathered
statstical
> >information with an analysis that has not been done yet.
>
> I'm developing a transcription methodology quite different from other ones
> that have been used, which is obvious "a road less travelled" - perhaps
> you'd prefer I didn't post anything further on this subject until I reach
> the destination?
>
> Cheers, ....Nick Pelling....
>
>
> ______________________________________________________________________
> To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
> unsubscribe vms-list
>

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