Sam G > 05-01-2017, 12:39 PM
Torsten > 05-01-2017, 06:26 PM
Quote:I still don't see how this really explains it. It seems like a good theory should account not only for which words exist, but also for how common they are.
Sam G > 07-01-2017, 11:14 AM
ThomasCoon > 07-01-2017, 02:42 PM
(07-01-2017, 11:14 AM)Sam G Wrote: You are not allowed to view links. Register or Login to view.[font=Arial]For instance, it seems clear that okaiin[/font] is related to otaiin, okol to otol, etc.
Sam G Wrote:Like I said above I think it's also important to consider the fact that q is almost non-existent in the labels. This also seems straightforwardly grammatical: q is something used in writing complete sentences but that is not needed in labeling individual items.
VViews > 07-01-2017, 03:16 PM
-JKP- > 07-01-2017, 03:41 PM
(07-01-2017, 03:16 PM)VViews Wrote: You are not allowed to view links. Register or Login to view.Sam G & Thomas Coon,
I agree with your observations regarding q.
The rarity of q in the labels and its absence as paragraph initial (but not as page initial) is the reason why I wonder whether q might stand for an ampersand, attached to the word which follows it.
Torsten > 07-01-2017, 05:38 PM
Quote:Sam G Wrote:
To me, many of these patterns look grammatical. [font=Arial]For instance, it seems clear that okaiin[/font] is related to otaiin, okol to otol, etc., and then for some reason relating to the purpose of these words, the k-form calls for the q-prefix more often than the t-form does.
Quote:Sam G Wrote:
I understand that there are lots of similar words, but that does not mean that the words are meaningless. I also don't see how we get from the rules you have proposed to a complete text generation method or how such a hypothetical method would explain the frequencies with which the different words occur.
Quote:Sam G Wrote:
Like I said above I think it's also important to consider the fact that q is almost non-existent in the labels.
Koen G > 07-01-2017, 05:50 PM
Sam G > 07-01-2017, 06:42 PM
(07-01-2017, 05:38 PM)Torsten Wrote: You are not allowed to view links. Register or Login to view.There is a whole network of relations between words for the entire VMS. For instance it is even possible to focus on a part of this network connecting okaiin with okol:
aiin (469) daiin (863) okaiin (212) qokaiin (262) kaiin (65)
ain ( 89) dain (211) okain (144) qokain (279) kain (48)
air ( 74) dair (106) okair ( 22) qokair ( 17) kair (14)
ar (350) dar (318) okar (129) qokar (152) kar (52)
or (363) dor ( 73) okor ( 34) qokor ( 36) kor (26)
ol (537) dol (117) okol ( 82) qokol (104) kol (37)
Therefore it is not possible to describe it as grammatical forms like we know from human language.
Quote:Quote:Sam G Wrote:
I understand that there are lots of similar words, but that does not mean that the words are meaningless. I also don't see how we get from the rules you have proposed to a complete text generation method or how such a hypothetical method would explain the frequencies with which the different words occur.
[font=Trebuchet MS][font=Trebuchet MS][font=Trebuchet MS][font=Trebuchet MS][font=Trebuchet MS]It seems that you are arguing against something I didn't said in this thread. For a text full of similar[/font][/font][/font][/font][font=Trebuchet MS] words multiple explanations are possible. Again, [font=Trebuchet MS]the three simple rules are not a [font=Trebuchet MS]text generation method. They are only a way to describe the relations for the words in the VMS.[/font][/font][/font][/font]
Quote:Quote:Sam G Wrote:
Like I said above I think it's also important to consider the fact that q is almost non-existent in the labels.
This is typical for the VMS. chedy as the third most frequent word for the whole VMS did never occur for pages in Currier A. Words ending in -m are typical for the end of a line. Words at the beginning of a paragraph start typically with a gallow glyph. It is possible to describe numerous patterns which words can occur on a certain part of the manuscript or on a certain place within a page or within a line. Since this is typical for the VMS there is no need to assume a special rule that q was not allowed for labels.
Quote:There are two pages with q-labels. This pages are f66r and f75v:There's a list of all the labels, in alphabetical order, here:
<f66r.L.9;H> qotesy (1 time)
<f66r.L.13;H> qokal (191 times)
<f66r.L.14;H> qolsa (1)
<f75v.L1.8;H> qokal (191)
<f75v.L1.16;H> qoted (4)
<f75v.L3.29;H> qotedy (91)
[font=Trebuchet MS]It is interesting that in both cases multiple instances of [/font]q[font=Trebuchet MS]-labels occur. Moreover on page You are not allowed to view links. Register or Login to view. two similar labels [/font]qoted[font=Trebuchet MS] and [/font]qotedy occur on the same page. Both observations could speak for a local element in the text generation mechanism.
Torsten > 07-01-2017, 08:55 PM
Quote:Sam G Wrote:
Well, you're just completely wrong on this point: patterns like these are actually extremely common in natural languages
Quote:Sam G Wrote:
Surely the relations between words are interesting and relevant but I don't think they show that the text is meaningless or that it was generated via a simple procedure. Quite the opposite, in fact, as argued above.
Quote:Sam G Wrote:
Yeah, but all these other patterns need to be accounted for as well!
Quote:Sam G Wrote:
...that's obviously much rarer than in the main body of the text.