Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771

Journal information: Cognition

Provided by Santa Fe Institute

Phys.org

On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
>
> Journal information: Cognition
>
> Provided by Santa Fe Institute
>
> Phys.org

It's not really about writing systems, but about the shapes of characters. It
quantifies a variety of obvious observations, and though it says it's based
on 133 scripts, it doesn't list them.

And it doesn't use the actual shapes, but it manipulates the Unicode shapes
which often don't resemble what people actually write.

And there are various factual errors -- it excludes Cherokee from its "invented
scripts" category because it mistakenly things the shapes of the letters were
taken from the Latin alphabet. (The shapes used for printing came along later.)

On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> >
> > Journal information: Cognition
> >
> > Provided by Santa Fe Institute
> >
> > Phys.org
> It's not really about writing systems, but about the shapes of characters. It
> quantifies a variety of obvious observations, and though it says it's based
> on 133 scripts, it doesn't list them.
>
> And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> which often don't resemble what people actually write.
>
> And there are various factual errors -- it excludes Cherokee from its "invented
> scripts" category because it mistakenly things the shapes of the letters were
> taken from the Latin alphabet. (The shapes used for printing came along later.)

Excluding invented scripts, ...
Is there a taxonomic classification of scripts?
If so, are there any isolates?

On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:

> > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > Journal information: Cognition
> > > Provided by Santa Fe Institute
> > > Phys.org
> > It's not really about writing systems, but about the shapes of characters. It
> > quantifies a variety of obvious observations, and though it says it's based
> > on 133 scripts, it doesn't list them.

Neither the pdf nor the web page offers a link to the sort of "Supplemental
Information" that would contain such a list.

> > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > which often don't resemble what people actually write.

A bad habit also used by the unfortunately widely cited articles by Changizi
et al.

> > And there are various factual errors -- it excludes Cherokee from its "invented
> > scripts" category because it mistakenly things the shapes of the letters were
> > taken from the Latin alphabet. (The shapes used for printing came along later.)
>
> Excluding invented scripts, ...
> Is there a taxonomic classification of scripts?
> If so, are there any isolates?

My book lists for only about $35 (USD). That's what it's about. You could
encourage your local libraries to buy it.

Isolates are any that we have archeological evidence for that didn't
survive to "reproduce." Meroitic would be an example.

On Friday, June 18, 2021 at 8:45:14 PM UTC+5:30, Peter T. Daniels wrote:
> On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> > On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
>
> > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > Journal information: Cognition
> > > > Provided by Santa Fe Institute
> > > > Phys.org
> > > It's not really about writing systems, but about the shapes of characters. It
> > > quantifies a variety of obvious observations, and though it says it's based
> > > on 133 scripts, it doesn't list them.
> Neither the pdf nor the web page offers a link to the sort of "Supplemental
> Information" that would contain such a list.
> > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > which often don't resemble what people actually write.
> A bad habit also used by the unfortunately widely cited articles by Changizi
> et al.
> > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > scripts" category because it mistakenly things the shapes of the letters were
> > > taken from the Latin alphabet. (The shapes used for printing came along later.)
> >
> > Excluding invented scripts, ...
> > Is there a taxonomic classification of scripts?
> > If so, are there any isolates?
> My book lists for only about $35 (USD). That's what it's about. You could
> encourage your local libraries to buy it.
>
> Isolates are any that we have archeological evidence for that didn't
> survive to "reproduce." Meroitic would be an example.

Wouldn't an isolate also be a script that didn't come from some other script?

On Friday, June 18, 2021 at 1:31:59 PM UTC-4, Dingbat wrote:
> On Friday, June 18, 2021 at 8:45:14 PM UTC+5:30, Peter T. Daniels wrote:
> > On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> > > On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > Journal information: Cognition
> > > > > Provided by Santa Fe Institute
> > > > > Phys.org
> > > > It's not really about writing systems, but about the shapes of characters. It
> > > > quantifies a variety of obvious observations, and though it says it's based
> > > > on 133 scripts, it doesn't list them.
> > Neither the pdf nor the web page offers a link to the sort of "Supplemental
> > Information" that would contain such a list.
> > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > which often don't resemble what people actually write.
> > A bad habit also used by the unfortunately widely cited articles by Changizi
> > et al.
> > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > taken from the Latin alphabet. (The shapes used for printing came along later.)
> > > Excluding invented scripts, ...
> > > Is there a taxonomic classification of scripts?
> > > If so, are there any isolates?
> > My book lists for only about $35 (USD). That's what it's about. You could
> > encourage your local libraries to buy it.
> > Isolates are any that we have archeological evidence for that didn't
> > survive to "reproduce." Meroitic would be an example.
>
> Wouldn't an isolate also be a script that didn't come from some other script?

Obviously. But you said "excluding invented scripts."

On Friday, June 18, 2021 at 8:15:14 AM UTC-7, Peter T. Daniels wrote:
> On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> > On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
>
> > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > Journal information: Cognition
> > > > Provided by Santa Fe Institute
> > > > Phys.org
> > > It's not really about writing systems, but about the shapes of characters. It
> > > quantifies a variety of obvious observations, and though it says it's based
> > > on 133 scripts, it doesn't list them.
> Neither the pdf nor the web page offers a link to the sort of "Supplemental
> Information" that would contain such a list.
> > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > which often don't resemble what people actually write.
> A bad habit also used by the unfortunately widely cited articles by Changizi
> et al.
> > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > scripts" category because it mistakenly things the shapes of the letters were
> > > taken from the Latin alphabet. (The shapes used for printing came along later.)
> >
> > Excluding invented scripts, ...
> > Is there a taxonomic classification of scripts?
> > If so, are there any isolates?
> My book lists for only about $35 (USD). That's what it's about. You could
> encourage your local libraries to buy it.
>
> Isolates are any that we have archeological evidence for that didn't
> survive to "reproduce." Meroitic would be an example.

I'd use Mayan as my example. Nowadays we can read it which is, I
think. not the case with Meroitic.

On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:
> On Friday, June 18, 2021 at 8:15:14 AM UTC-7, Peter T. Daniels wrote:
> > On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> > > On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:

> > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > Journal information: Cognition
> > > > > Provided by Santa Fe Institute
> > > > > Phys.org
> > > > It's not really about writing systems, but about the shapes of characters. It
> > > > quantifies a variety of obvious observations, and though it says it's based
> > > > on 133 scripts, it doesn't list them.
> > Neither the pdf nor the web page offers a link to the sort of "Supplemental
> > Information" that would contain such a list.
> > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > which often don't resemble what people actually write.
> > A bad habit also used by the unfortunately widely cited articles by Changizi
> > et al.
> > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > taken from the Latin alphabet. (The shapes used for printing came along later.)
> > > Excluding invented scripts, ...
> > > Is there a taxonomic classification of scripts?
> > > If so, are there any isolates?
> > My book lists for only about $35 (USD). That's what it's about. You could
> > encourage your local libraries to buy it.
> > Isolates are any that we have archeological evidence for that didn't
> > survive to "reproduce." Meroitic would be an example.
>
> I'd use Mayan as my example. Nowadays we can read it which is, I
> think. not the case with Meroitic.

We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
values of the characters by comparing names written in Egyptian),
but the _language_ wasn't properly deciphered until 2010, when Claude
Rilly realized that the script is an abugida (not an alphabet) and the
language belongs to the Nilo-Saharan subfamily Nubian.

On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:
> On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:
> > On Friday, June 18, 2021 at 8:15:14 AM UTC-7, Peter T. Daniels wrote:
> > > On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> > > > On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
>
> > > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > > Journal information: Cognition
> > > > > > Provided by Santa Fe Institute
> > > > > > Phys.org
> > > > > It's not really about writing systems, but about the shapes of characters. It
> > > > > quantifies a variety of obvious observations, and though it says it's based
> > > > > on 133 scripts, it doesn't list them.
> > > Neither the pdf nor the web page offers a link to the sort of "Supplemental
> > > Information" that would contain such a list.
> > > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > > which often don't resemble what people actually write.
> > > A bad habit also used by the unfortunately widely cited articles by Changizi
> > > et al.
> > > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > > taken from the Latin alphabet. (The shapes used for printing came along later.)
> > > > Excluding invented scripts, ...
> > > > Is there a taxonomic classification of scripts?
> > > > If so, are there any isolates?
> > > My book lists for only about $35 (USD). That's what it's about. You could
> > > encourage your local libraries to buy it.
> > > Isolates are any that we have archeological evidence for that didn't
> > > survive to "reproduce." Meroitic would be an example.
> >
> > I'd use Mayan as my example. Nowadays we can read it which is, I
> > think. not the case with Meroitic.
> We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> values of the characters by comparing names written in Egyptian),
> but the _language_ wasn't properly deciphered until 2010, when Claude
> Rilly realized that the script is an abugida (not an alphabet) and the
> language belongs to the Nilo-Saharan subfamily Nubian.

Aha. I missed that. I suppose Rilly has written a book which I will
never get around to reading. I read a grammar of Nubian once a long
time ago but I don't remember anything about it. I wasn't trying to learn
Nubian - I was looking for possible Cushitic influences.

Fri, 18 Jun 2021 08:15:12 -0700 (PDT): "Peter T. Daniels"
<grammatim@verizon.net> scribeva:

>On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
>> On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
>> > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
>
>> > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
>> > > Journal information: Cognition
>> > > Provided by Santa Fe Institute
>> > > Phys.org
>> > It's not really about writing systems, but about the shapes of characters. It
>> > quantifies a variety of obvious observations, and though it says it's based
>> > on 133 scripts, it doesn't list them.
>
>Neither the pdf nor the web page offers a link to the sort of "Supplemental
>Information" that would contain such a list.
>
>> > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
>> > which often don't resemble what people actually write.
>
>A bad habit also used by the unfortunately widely cited articles by Changizi
>et al.
>
>> > And there are various factual errors -- it excludes Cherokee from its "invented
>> > scripts" category because it mistakenly things the shapes of the letters were
>> > taken from the Latin alphabet. (The shapes used for printing came along later.)
>>
>> Excluding invented scripts, ...
>> Is there a taxonomic classification of scripts?
>> If so, are there any isolates?
>
>My book lists for only about $35 (USD). That's what it's about. You could
>encourage your local libraries to buy it.

The old or the new? The old one is crazily expensive, I recently
noticed.

If there is een ePub version available for a reasonable price, I might
consider buying it.

>Isolates are any that we have archeological evidence for that didn't
>survive to "reproduce." Meroitic would be an example.

--
Ruud Harmsen, http://rudhar.com

On Saturday, June 19, 2021 at 9:49:42 AM UTC+5:30, DKleinecke wrote:
> On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:
> > On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:
> > > On Friday, June 18, 2021 at 8:15:14 AM UTC-7, Peter T. Daniels wrote:
> > > > On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> > > > > On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > > > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> >
> > > > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > > > Journal information: Cognition
> > > > > > > Provided by Santa Fe Institute
> > > > > > > Phys.org
> > > > > > It's not really about writing systems, but about the shapes of characters. It
> > > > > > quantifies a variety of obvious observations, and though it says it's based
> > > > > > on 133 scripts, it doesn't list them.
> > > > Neither the pdf nor the web page offers a link to the sort of "Supplemental
> > > > Information" that would contain such a list.
> > > > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > > > which often don't resemble what people actually write.
> > > > A bad habit also used by the unfortunately widely cited articles by Changizi
> > > > et al.
> > > > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > > > taken from the Latin alphabet. (The shapes used for printing came along later.)
> > > > > Excluding invented scripts, ...
> > > > > Is there a taxonomic classification of scripts?
> > > > > If so, are there any isolates?
> > > > My book lists for only about $35 (USD). That's what it's about. You could
> > > > encourage your local libraries to buy it.
> > > > Isolates are any that we have archeological evidence for that didn't
> > > > survive to "reproduce." Meroitic would be an example.
> > >
> > > I'd use Mayan as my example. Nowadays we can read it which is, I
> > > think. not the case with Meroitic.
> > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > values of the characters by comparing names written in Egyptian),
> > but the _language_ wasn't properly deciphered until 2010, when Claude
> > Rilly realized that the script is an abugida (not an alphabet) and the
> > language belongs to the Nilo-Saharan subfamily Nubian.
> Aha. I missed that. I suppose Rilly has written a book which I will
> never get around to reading. I read a grammar of Nubian once a long
> time ago but I don't remember anything about it. I wasn't trying to learn
> Nubian - I was looking for possible Cushitic influences.

What do you mean by Cushitic? Meriotic was a language of a kingdom
the Egyptians called Kush. The Greeks called its inhabitants Aethiopes.

On Saturday, June 19, 2021 at 12:19:42 AM UTC-4, DKleinecke wrote:
> On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:
> > On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:
> > > On Friday, June 18, 2021 at 8:15:14 AM UTC-7, Peter T. Daniels wrote:
> > > > On Friday, June 18, 2021 at 9:23:22 AM UTC-4, Dingbat wrote:
> > > > > On Friday, June 18, 2021 at 1:58:09 AM UTC+5:30, Peter T. Daniels wrote:
> > > > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> >
> > > > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > > > Journal information: Cognition
> > > > > > > Provided by Santa Fe Institute
> > > > > > > Phys.org
> > > > > > It's not really about writing systems, but about the shapes of characters. It
> > > > > > quantifies a variety of obvious observations, and though it says it's based
> > > > > > on 133 scripts, it doesn't list them.
> > > > Neither the pdf nor the web page offers a link to the sort of "Supplemental
> > > > Information" that would contain such a list.
> > > > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > > > which often don't resemble what people actually write.
> > > > A bad habit also used by the unfortunately widely cited articles by Changizi
> > > > et al.
> > > > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > > > taken from the Latin alphabet. (The shapes used for printing came along later.)
> > > > > Excluding invented scripts, ...
> > > > > Is there a taxonomic classification of scripts?
> > > > > If so, are there any isolates?
> > > > My book lists for only about $35 (USD). That's what it's about. You could
> > > > encourage your local libraries to buy it.
> > > > Isolates are any that we have archeological evidence for that didn't
> > > > survive to "reproduce." Meroitic would be an example.
> > >
> > > I'd use Mayan as my example. Nowadays we can read it which is, I
> > > think. not the case with Meroitic.
> > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > values of the characters by comparing names written in Egyptian),
> > but the _language_ wasn't properly deciphered until 2010, when Claude
> > Rilly realized that the script is an abugida (not an alphabet) and the
> > language belongs to the Nilo-Saharan subfamily Nubian.
>
> Aha. I missed that. I suppose Rilly has written a book which I will
> never get around to reading. I read a grammar of Nubian once a long
> time ago but I don't remember anything about it. I wasn't trying to learn
> Nubian - I was looking for possible Cushitic influences.

I kinda thought I gave enough info to google him ...

Claude Rilly, a 500-page monograph in French proving it's Nubian;
a more reasonable book by Rilly and Alex de Voogt from Cambridge
(in EWnglish) on the language, with little sketches of the history and
such (for which you'd need to go to the archeologists).

On Saturday, June 19, 2021 at 4:37:30 AM UTC-4, Dingbat wrote:
> On Saturday, June 19, 2021 at 9:49:42 AM UTC+5:30, DKleinecke wrote:
> > On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:
> > > On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:

> > > > I'd use Mayan as my example. Nowadays we can read it which is, I
> > > > think. not the case with Meroitic.
> > > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > > values of the characters by comparing names written in Egyptian),
> > > but the _language_ wasn't properly deciphered until 2010, when Claude
> > > Rilly realized that the script is an abugida (not an alphabet) and the
> > > language belongs to the Nilo-Saharan subfamily Nubian.
> > Aha. I missed that. I suppose Rilly has written a book which I will
> > never get around to reading. I read a grammar of Nubian once a long
> > time ago but I don't remember anything about it. I wasn't trying to learn
> > Nubian - I was looking for possible Cushitic influences.
>
> What do you mean by Cushitic? Meriotic was a language of a kingdom
> the Egyptians called Kush. The Greeks called its inhabitants Aethiopes.

Cushitic, as even Wikipedia will tell you, is one of the six families comprising
Afroasiatic. Cushitic languages are spoken from northeastern Sudan (Beja)
through most of Ethiopia (Oromo is the most prominent) and Somalia (Somali)
with some disconnected populations in northern Kenya (Iraqw).

Meroitic was the language of the kingdom of Meroë, south of the present-day
Egypt-Sudan border, at lest as far as Khartoum. Inscriptions survive from the
4th c. CE, so that's usually given as the end of its use, but the Old Nubian
alphabet is the Coptic alphabet plus a couple of letters taken from Meroitic,
and it's attested (dated documents) from the late 9th c. to the 12th, so some
knowledge of Meroitic must have survived for several centuries after the
latest attestation.

On Saturday, June 19, 2021 at 6:43:24 PM UTC+5:30, Peter T. Daniels wrote:
> On Saturday, June 19, 2021 at 4:37:30 AM UTC-4, Dingbat wrote:
> > On Saturday, June 19, 2021 at 9:49:42 AM UTC+5:30, DKleinecke wrote:
> > > On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:
> > > > On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:
>
> > > > > I'd use Mayan as my example. Nowadays we can read it which is, I
> > > > > think. not the case with Meroitic.
> > > > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > > > values of the characters by comparing names written in Egyptian),
> > > > but the _language_ wasn't properly deciphered until 2010, when Claude
> > > > Rilly realized that the script is an abugida (not an alphabet) and the
> > > > language belongs to the Nilo-Saharan subfamily Nubian.
> > > Aha. I missed that. I suppose Rilly has written a book which I will
> > > never get around to reading. I read a grammar of Nubian once a long
> > > time ago but I don't remember anything about it. I wasn't trying to learn
> > > Nubian - I was looking for possible Cushitic influences.
> >
> > What do you mean by Cushitic? Meriotic was a language of a kingdom
> > the Egyptians called Kush. The Greeks called its inhabitants Aethiopes.

> Cushitic, as even Wikipedia will tell you, is one of the six families comprising
> Afroasiatic.
>
Ah! If Cushitic is named after Kush, it seems odd to look into whether a
language of Kush is possibly influenced by a Cushitic language.
>
> Cushitic languages are spoken from northeastern Sudan (Beja)
> through most of Ethiopia (Oromo is the most prominent) and Somalia (Somali)
> with some disconnected populations in northern Kenya (Iraqw).
>
> Meroitic was the language of the kingdom of Meroë, south of the present-day
> Egypt-Sudan border, at lest as far as Khartoum. Inscriptions survive from the
> 4th c. CE, so that's usually given as the end of its use, but the Old Nubian
> alphabet is the Coptic alphabet plus a couple of letters taken from Meroitic,
> and it's attested (dated documents) from the late 9th c. to the 12th, so some
> knowledge of Meroitic must have survived for several centuries after the
> latest attestation.
>
Hieroglyphic Meriotic was an abugida. Hieroglyphs wouldn't go with a Coptic
alphabet well enough to be used to extend that alphabet. Did these couple
of letters come from Demotic equivalents of hieroglyphs?

On Sunday, June 20, 2021 at 3:46:29 AM UTC-4, Dingbat wrote:
> On Saturday, June 19, 2021 at 6:43:24 PM UTC+5:30, Peter T. Daniels wrote:
> > On Saturday, June 19, 2021 at 4:37:30 AM UTC-4, Dingbat wrote:
> > > On Saturday, June 19, 2021 at 9:49:42 AM UTC+5:30, DKleinecke wrote:
> > > > On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:
> > > > > On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:
> >
> > > > > > I'd use Mayan as my example. Nowadays we can read it which is, I
> > > > > > think. not the case with Meroitic.
> > > > > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > > > > values of the characters by comparing names written in Egyptian),
> > > > > but the _language_ wasn't properly deciphered until 2010, when Claude
> > > > > Rilly realized that the script is an abugida (not an alphabet) and the
> > > > > language belongs to the Nilo-Saharan subfamily Nubian.
> > > > Aha. I missed that. I suppose Rilly has written a book which I will
> > > > never get around to reading. I read a grammar of Nubian once a long
> > > > time ago but I don't remember anything about it. I wasn't trying to learn
> > > > Nubian - I was looking for possible Cushitic influences.
> > >
> > > What do you mean by Cushitic? Meriotic was a language of a kingdom
> > > the Egyptians called Kush. The Greeks called its inhabitants Aethiopes.
>
> > Cushitic, as even Wikipedia will tell you, is one of the six families comprising
> > Afroasiatic.
> >
> Ah! If Cushitic is named after Kush, it seems odd to look into whether a
> language of Kush is possibly influenced by a Cushitic language.

What an astonishingly ahistorical attitude.

When was it fashionable to name language groups using biblical
genealogies? When was "Kush" first investigated? Why was "Kush"
called "Kush"?

> > Cushitic languages are spoken from northeastern Sudan (Beja)
> > through most of Ethiopia (Oromo is the most prominent) and Somalia (Somali)
> > with some disconnected populations in northern Kenya (Iraqw).
> > Meroitic was the language of the kingdom of Meroë, south of the present-day
> > Egypt-Sudan border, at lest as far as Khartoum. Inscriptions survive from the
> > 4th c. CE, so that's usually given as the end of its use, but the Old Nubian
> > alphabet is the Coptic alphabet plus a couple of letters taken from Meroitic,
> > and it's attested (dated documents) from the late 9th c. to the 12th, so some
> > knowledge of Meroitic must have survived for several centuries after the
> > latest attestation.
>
> Hieroglyphic Meriotic was an abugida. Hieroglyphs wouldn't go with a Coptic
> alphabet well enough to be used to extend that alphabet. Did these couple
> of letters come from Demotic equivalents of hieroglyphs?

There are two varieties of Meroitic writing, the pictographic and the "linear."
The linear forms are older. Neither has anything to do with either Egyptian
or Semitic letter shapes.

Let me introduce this thing called "Google" to you.

https://en.wikipedia.org/wiki/Old_Nubian#Writing

"The alphabet included three additional letters ⳡ /ɲ/ and ⳣ /w/, and ⳟ /ŋ/, the
first two deriving from the Meroitic alphabet."

On Saturday, June 19, 2021 at 6:03:51 AM UTC-7, Peter T. Daniels wrote:
> On Saturday, June 19, 2021 at 12:19:42 AM UTC-4, DKleinecke wrote:
> > On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:

> > > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > > values of the characters by comparing names written in Egyptian),
> > > but the _language_ wasn't properly deciphered until 2010, when Claude
> > > Rilly realized that the script is an abugida (not an alphabet) and the
> > > language belongs to the Nilo-Saharan subfamily Nubian.
> >
> > Aha. I missed that. I suppose Rilly has written a book which I will
> > never get around to reading. I read a grammar of Nubian once a long
> > time ago but I don't remember anything about it. I wasn't trying to learn
> > Nubian - I was looking for possible Cushitic influences.

> I kinda thought I gave enough info to google him ...

You did, thank you, but I thought others might be interested.
>
> Claude Rilly, a 500-page monograph in French proving it's Nubian;
> a more reasonable book by Rilly and Alex de Voogt from Cambridge
> (in EWnglish) on the language, with little sketches of the history and
> such (for which you'd need to go to the archeologists).

As I recall it is easy to see Nubian is not Afroasiatic. I was working on
Cushitic and I wondered whether there was any clear example of
Afroasiatic influence on Nubian. I found none - but I know so little
Egyptian that I might have missed influences from that direction.
I see now that I really should have looked at Berber and Chadic, too.

On Sunday, June 20, 2021 at 3:01:57 PM UTC-4, DKleinecke wrote:
> On Saturday, June 19, 2021 at 6:03:51 AM UTC-7, Peter T. Daniels wrote:
> > On Saturday, June 19, 2021 at 12:19:42 AM UTC-4, DKleinecke wrote:
> > > On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:

> > > > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > > > values of the characters by comparing names written in Egyptian),
> > > > but the _language_ wasn't properly deciphered until 2010, when Claude
> > > > Rilly realized that the script is an abugida (not an alphabet) and the
> > > > language belongs to the Nilo-Saharan subfamily Nubian.
> > > Aha. I missed that. I suppose Rilly has written a book which I will
> > > never get around to reading. I read a grammar of Nubian once a long
> > > time ago but I don't remember anything about it. I wasn't trying to learn
> > > Nubian - I was looking for possible Cushitic influences.
> > I kinda thought I gave enough info to google h
>im ...
> You did, thank you, but I thought others might be interested.
>
> > Claude Rilly, a 500-page monograph in French proving it's Nubian;
> > a more reasonable book by Rilly and Alex de Voogt from Cambridge
> > (in EWnglish) on the language, with little sketches of the history and
> > such (for which you'd need to go to the archeologists).
>
> As I recall it is easy to see Nubian is not Afroasiatic. I was working on
> Cushitic and I wondered whether there was any clear example of
> Afroasiatic influence on Nubian. I found none - but I know so little
> Egyptian that I might have missed influences from that direction.
> I see now that I really should have looked at Berber and Chadic, too.

While doing my int4roductory chapter for a book on Literacy in
Africa I found that Nilo-Saharan has received by far the least
coverage of the "four phyla." My late friend Lionel Bender (we
roomed together two or three times at meetings -- he insisted
on watching the NCAA basketball tournament) was really the
only synthesizer. He also did a lot on Omotic.

On Sunday, June 20, 2021 at 6:34:22 PM UTC+5:30, Peter T. Daniels wrote:
> On Sunday, June 20, 2021 at 3:46:29 AM UTC-4, Dingbat wrote:
> > On Saturday, June 19, 2021 at 6:43:24 PM UTC+5:30, Peter T. Daniels wrote:
> > > On Saturday, June 19, 2021 at 4:37:30 AM UTC-4, Dingbat wrote:
> > > > On Saturday, June 19, 2021 at 9:49:42 AM UTC+5:30, DKleinecke wrote:
> > > > > On Friday, June 18, 2021 at 2:25:41 PM UTC-7, Peter T. Daniels wrote:
> > > > > > On Friday, June 18, 2021 at 3:13:22 PM UTC-4, DKleinecke wrote:
> > >
> > > > > > > I'd use Mayan as my example. Nowadays we can read it which is, I
> > > > > > > think. not the case with Meroitic.
> > > > > > We could "read" Meroitic in 1911 (F. Ll. Griffith discovered the sound
> > > > > > values of the characters by comparing names written in Egyptian),
> > > > > > but the _language_ wasn't properly deciphered until 2010, when Claude
> > > > > > Rilly realized that the script is an abugida (not an alphabet) and the
> > > > > > language belongs to the Nilo-Saharan subfamily Nubian.
> > > > > Aha. I missed that. I suppose Rilly has written a book which I will
> > > > > never get around to reading. I read a grammar of Nubian once a long
> > > > > time ago but I don't remember anything about it. I wasn't trying to learn
> > > > > Nubian - I was looking for possible Cushitic influences.
> > > >
> > > > What do you mean by Cushitic? Meriotic was a language of a kingdom
> > > > the Egyptians called Kush. The Greeks called its inhabitants Aethiopes.
> >
> > > Cushitic, as even Wikipedia will tell you, is one of the six families comprising
> > > Afroasiatic.
> > >
> > Ah! If Cushitic is named after Kush, it seems odd to look into whether a
> > language of Kush is possibly influenced by a Cushitic language.
> What an astonishingly ahistorical attitude.
>
> When was it fashionable to name language groups using biblical
> genealogies? When was "Kush" first investigated? Why was "Kush"
> called "Kush"?

NatGeo says "Kush" comes from what Egyptians called the civilization. It doesn't say why they called it that and it doesn't identify the Bible as the source of "Kush". I've seen it suggested that "Nubia" "means land of gold" via Egyptian "nub" meaning gold.

<<The Kerma kingdom controlled the Nile Valley between the first and fourth cataracts, making its territory as extensive as its powerful neighbor to the north, Egypt. Egyptian records are the first to identify this Nubian civilization as “Kush.”>>
https://www.nationalgeographic.org/media/kingdoms-kush/

> > > Cushitic languages are spoken from northeastern Sudan (Beja)
> > > through most of Ethiopia (Oromo is the most prominent) and Somalia (Somali)
> > > with some disconnected populations in northern Kenya (Iraqw).
> > > Meroitic was the language of the kingdom of Meroë, south of the present-day
> > > Egypt-Sudan border, at lest as far as Khartoum. Inscriptions survive from the
> > > 4th c. CE, so that's usually given as the end of its use, but the Old Nubian
> > > alphabet is the Coptic alphabet plus a couple of letters taken from Meroitic,
> > > and it's attested (dated documents) from the late 9th c. to the 12th, so some
> > > knowledge of Meroitic must have survived for several centuries after the
> > > latest attestation.
> >
> > Hieroglyphic Meriotic was an abugida. Hieroglyphs wouldn't go with a Coptic
> > alphabet well enough to be used to extend that alphabet. Did these couple
> > of letters come from Demotic equivalents of hieroglyphs?
> There are two varieties of Meroitic writing, the pictographic and the "linear."
> The linear forms are older. Neither has anything to do with either Egyptian
> or Semitic letter shapes.
>
> Let me introduce this thing called "Google" to you.
>
> https://en.wikipedia.org/wiki/Old_Nubian#Writing
>
> "The alphabet included three additional letters ⳡ /ɲ/ and ⳣ /w/, and ⳟ /ŋ/, the
> first two deriving from the Meroitic alphabet."

This shows two scripts used for writing "Meriotic and possibly Old Nubian".
I see a possible predecessor of ⳣ in the Demo column. The legend "Demo" suggests that its glyph came from Demotic but I've seen that script being called Cursive elsewhere. A postscript claims that the parent writing systems of the two scripts are Demotic and Egyptian hieroglyphic.
https://en.wikipedia.org/wiki/Meroitic_script#/media/File:Meroitic.png

Dear Peter Daniels,

I happened on this google groups discussion which discusses my paper with Helena Miton, "Graphic complexity in writing systems".
I would like to attract your attention to the fact that we published our complete dataset, including (obviously), the list of 133 scripts used in this study. This is written in the paper on p. 4: 'Data and R scripts used to produce the results and figures can be found at https://osf.io/9dnj3/ '
The fact that you missed this seems to suggest you barely skimmed this paper before rushing to dismiss it.
Regards
Olivier Morin

On Thursday, June 17, 2021 at 10:28:09 PM UTC+2, Peter T. Daniels wrote:
> On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> >
> > Journal information: Cognition
> >
> > Provided by Santa Fe Institute
> >
> > Phys.org
> It's not really about writing systems, but about the shapes of characters.. It
> quantifies a variety of obvious observations, and though it says it's based
> on 133 scripts, it doesn't list them.
>
> And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> which often don't resemble what people actually write.
>
> And there are various factual errors -- it excludes Cherokee from its "invented
> scripts" category because it mistakenly things the shapes of the letters were
> taken from the Latin alphabet. (The shapes used for printing came along later.)

On Thursday, June 24, 2021 at 4:48:43 AM UTC-4, Olivier Morin wrote:

> Dear Peter Daniels,
>
> I happened on this google groups discussion which discusses my paper with Helena Miton, "Graphic complexity in writing systems".
> I would like to attract your attention to the fact that we published our complete dataset, including (obviously), the list of 133 scripts used in this study. This is written in the paper on p. 4: 'Data and R scripts used to produce the results and figures can be found at https://osf.io/9dnj3/ '
> The fact that you missed this seems to suggest you barely skimmed this paper before rushing to dismiss it.

Thank you. That is not the article that was published in the journal *Cognition*.

I was indeed put off by your reliance on Changizi et al., for my reaction to which
you can consult my book *An Exploration of Writing* (2018), p. 152.

> Regards
> Olivier Morin
> On Thursday, June 17, 2021 at 10:28:09 PM UTC+2, Peter T. Daniels wrote:
> > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > >
> > > Journal information: Cognition
> > >
> > > Provided by Santa Fe Institute
> > >
> > > Phys.org
> > It's not really about writing systems, but about the shapes of characters. It
> > quantifies a variety of obvious observations, and though it says it's based
> > on 133 scripts, it doesn't list them.
> >
> > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > which often don't resemble what people actually write.
> >
> > And there are various factual errors -- it excludes Cherokee from its "invented
> > scripts" category because it mistakenly things the shapes of the letters were
> > taken from the Latin alphabet. (The shapes used for printing came along later.)

On Thursday, June 24, 2021 at 7:49:25 AM UTC-4, Peter T. Daniels wrote:
> On Thursday, June 24, 2021 at 4:48:43 AM UTC-4, Olivier Morin wrote:
>
> > Dear Peter Daniels,
> >
> > I happened on this google groups discussion which discusses my paper with Helena Miton, "Graphic complexity in writing systems".
> > I would like to attract your attention to the fact that we published our complete dataset, including (obviously), the list of 133 scripts used in this study. This is written in the paper on p. 4: 'Data and R scripts used to produce the results and figures can be found at https://osf.io/9dnj3/ '
> > The fact that you missed this seems to suggest you barely skimmed this paper before rushing to dismiss it.
> Thank you. That is not the article that was published in the journal *Cognition*.
>
> I was indeed put off by your reliance on Changizi et al., for my reaction to which
> you can consult my book *An Exploration of Writing* (2018), p. 152.

I find no list of the 133 scripts you used. I find a list of something over 200
scripts, some of which are phantoms, some of which are toys (such as
Shavian and Deseret).

The claim "over 47,000 characters" is not helpful. Does it include all the
characters in most of the scripts, and several hundred Sinitic and Egyptian
characters? Are variants that the Unicode folk unaccountably list separately
counted separately, and ones they unaccountably combined treated asone?

(Changizi et al. must have selected just a few characters from each script
they name in order to make their numbers come out, yet they excluded at
least one of the most complicated-shaped scripts in use, Malayalam.)

GIGO.

> > Regards
> > Olivier Morin
> > On Thursday, June 17, 2021 at 10:28:09 PM UTC+2, Peter T. Daniels wrote:
> > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > >
> > > > Journal information: Cognition
> > > >
> > > > Provided by Santa Fe Institute
> > > >
> > > > Phys.org
> > > It's not really about writing systems, but about the shapes of characters. It
> > > quantifies a variety of obvious observations, and though it says it's based
> > > on 133 scripts, it doesn't list them.
> > >
> > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > which often don't resemble what people actually write.
> > >
> > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > scripts" category because it mistakenly things the shapes of the letters were
> > > taken from the Latin alphabet. (The shapes used for printing came along later.)

Dear Peter Daniels,

The file InventoryScripts.csv contains a list of all the scripts that were considered for inclusion. Of those, we selected 133. It is easy to figure out which by looking at the "ExclusionCriteria" field and setting it to the value "0". It helps to open this csv file in R incidentally. R is a software that people use to read data files.

We do not in any way "rely" on Changizi, on the contrary, we saw flaws in his paper and re-tested his conclusions from scratch. Our conclusions are quite different from his. This comes out clearly from any minimally honest and careful reading of our paper.

You ask:
> The claim "over 47,000 characters" is not helpful. Does it include all the
> characters in most of the scripts, and several hundred Sinitic and Egyptian
> characters?

This question is already clearly answered in our paper. All eligible characters are included for Chinese characters and Egyptian hieroglyphs (more than several hundred if memory serves). There are criteria for character inclusion which are the same across all scripts, please refer to our paper.

Are variants that the Unicode folk unaccountably list separately
> counted separately, and ones they unaccountably combined treated asone?
>

It depends what you mean by "unaccountably". Like any large-scale comparative study ours is based on methodological choices that can be questioned. Any methodological choice can be questioned, since no one is infallible. The interesting question to ask is whether those choices would have changed our results in any way.

Regards,

OM

On Thursday, June 24, 2021 at 2:47:22 PM UTC+2, Peter T. Daniels wrote:
> On Thursday, June 24, 2021 at 7:49:25 AM UTC-4, Peter T. Daniels wrote:
> > On Thursday, June 24, 2021 at 4:48:43 AM UTC-4, Olivier Morin wrote:
> >
> > > Dear Peter Daniels,
> > >
> > > I happened on this google groups discussion which discusses my paper with Helena Miton, "Graphic complexity in writing systems".
> > > I would like to attract your attention to the fact that we published our complete dataset, including (obviously), the list of 133 scripts used in this study. This is written in the paper on p. 4: 'Data and R scripts used to produce the results and figures can be found at https://osf.io/9dnj3/ '
> > > The fact that you missed this seems to suggest you barely skimmed this paper before rushing to dismiss it.
> > Thank you. That is not the article that was published in the journal *Cognition*.
> >
> > I was indeed put off by your reliance on Changizi et al., for my reaction to which
> > you can consult my book *An Exploration of Writing* (2018), p. 152.
> I find no list of the 133 scripts you used. I find a list of something over 200
> scripts, some of which are phantoms, some of which are toys (such as
> Shavian and Deseret).
>
> The claim "over 47,000 characters" is not helpful. Does it include all the
> characters in most of the scripts, and several hundred Sinitic and Egyptian
> characters? Are variants that the Unicode folk unaccountably list separately
> counted separately, and ones they unaccountably combined treated asone?
>
> (Changizi et al. must have selected just a few characters from each script
> they name in order to make their numbers come out, yet they excluded at
> least one of the most complicated-shaped scripts in use, Malayalam.)
>
> GIGO.
> > > Regards
> > > Olivier Morin
> > > On Thursday, June 17, 2021 at 10:28:09 PM UTC+2, Peter T. Daniels wrote:
> > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > >
> > > > > Journal information: Cognition
> > > > >
> > > > > Provided by Santa Fe Institute
> > > > >
> > > > > Phys.org
> > > > It's not really about writing systems, but about the shapes of characters. It
> > > > quantifies a variety of obvious observations, and though it says it's based
> > > > on 133 scripts, it doesn't list them.
> > > >
> > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > which often don't resemble what people actually write.
> > > >
> > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > taken from the Latin alphabet. (The shapes used for printing came along later.)

On Friday, June 25, 2021 at 2:40:01 AM UTC-4, Olivier Morin wrote:
> Dear Peter Daniels,
>
> The file InventoryScripts.csv contains a list of all the scripts that were considered for inclusion. Of those, we selected 133. It is easy to figure out which by looking at the "ExclusionCriteria" field and setting it to the value "0". It helps to open this csv file in R incidentally. R is a software that people use to read data files.

How would one go from the published version to that absurd procedure?

Can you point to other published articles comparing writing systems that
go to such lengths to conceal the database used in the study?

Look at what you just said. First, the reader must somehow get from the
published article to the page you here provided the link for. Second, the
reader must discover from a fairly lengthy outline of contents that among
the links is a list of scripts. Third, one must discover somewhere else some
sort of list of criteria for culling that list. Fourth, one must infer how to apply
those criteria in each of more than 200 items in order to discover which
approximately half of them were considered, though there is no way of
ascertaining whether one's guesses correspond with the actuality..

I don't know what academic field you are coming from, but in that field,
is it not the practice to include "supplementary information" files linked
from the article, as has been the practice in *Science* and *Nature* for
years, and is now even found in *Language: The Journal of the Linguistic
Society of America*?

It's shocking that a top-level journal like *Cognition* would publish an
article that made it so difficult for the reader to discover what it was
even talking about.

> We do not in any way "rely" on Changizi, on the contrary, we saw flaws in his paper and re-tested his conclusions from scratch. Our conclusions are quite different from his. This comes out clearly from any minimally honest and careful reading of our paper.

One cannot "honestly and carefully read" a paper about writing systems
without knowing what writing systems it is about.

You did, however, adopt their unacceptable reliance on the shapes displayed
in the Unicode documentation and their further simplification them before
performing your analysis.

> You ask:
> > The claim "over 47,000 characters" is not helpful. Does it include all the
> > characters in most of the scripts, and several hundred Sinitic and Egyptian
> > characters?
> This question is already clearly answered in our paper. All eligible characters are included for

"eligible"?

Are the variants that were included in Unicode disregarded, or simply counted
as individuals? The Unicode treatment of Cuneiform is especially heinous.
Unicode does not separate Traditional from Simplified Chinese characters,
nor ones that are Japanese-specific, nor ones that are essentially historic
variants.

> Chinese characters and Egyptian hieroglyphs (more than several hundred if memory serves). There

"The largest Chinese dictionaries" include around 50,000 characters, because
they list every variant known to the ancient scholars. Only around 700 Egyptian
hieroglyphs were used during most of Egypt's history (some for only a single
word), but in the Ptolemaic period, when that sort of learning was a niche subject,
new hieroglyphs were invented supposedly bringing the inventory to several
thousand. Do you deal only with character inventories that were actually used,
or with everything you can find?

are criteria for character inclusion which are the same across all scripts, please refer to our paper.
>
>> Are variants that the Unicode folk unaccountably list separately
> > counted separately, and ones they unaccountably combined treated asone?
>
> It depends what you mean by "unaccountably". Like any large-scale comparative study ours is based on methodological choices that can be questioned. Any methodological choice can be questioned, since no one is infallible. The interesting question to ask is whether those choices would have changed our results in any way.

Noth8ng in the published article talks about "methodological choices."

For instance, combining Greek and Coptic; for instance, separating by now
probably about a dozen ancient Northwest Semitic scripts into separate
blocks, and a perhaps greater number of variants of Indic scripts.

> Regards,
>
> OM
> On Thursday, June 24, 2021 at 2:47:22 PM UTC+2, Peter T. Daniels wrote:
> > On Thursday, June 24, 2021 at 7:49:25 AM UTC-4, Peter T. Daniels wrote:
> > > On Thursday, June 24, 2021 at 4:48:43 AM UTC-4, Olivier Morin wrote:
> > >
> > > > Dear Peter Daniels,
> > > >
> > > > I happened on this google groups discussion which discusses my paper with Helena Miton, "Graphic complexity in writing systems".
> > > > I would like to attract your attention to the fact that we published our complete dataset, including (obviously), the list of 133 scripts used in this study. This is written in the paper on p. 4: 'Data and R scripts used to produce the results and figures can be found at https://osf.io/9dnj3/ '
> > > > The fact that you missed this seems to suggest you barely skimmed this paper before rushing to dismiss it.
> > > Thank you. That is not the article that was published in the journal *Cognition*.
> > >
> > > I was indeed put off by your reliance on Changizi et al., for my reaction to which
> > > you can consult my book *An Exploration of Writing* (2018), p. 152.
> > I find no list of the 133 scripts you used. I find a list of something over 200
> > scripts, some of which are phantoms, some of which are toys (such as
> > Shavian and Deseret).
> >
> > The claim "over 47,000 characters" is not helpful. Does it include all the
> > characters in most of the scripts, and several hundred Sinitic and Egyptian
> > characters? Are variants that the Unicode folk unaccountably list separately
> > counted separately, and ones they unaccountably combined treated asone?
> >
> > (Changizi et al. must have selected just a few characters from each script
> > they name in order to make their numbers come out, yet they excluded at
> > least one of the most complicated-shaped scripts in use, Malayalam.)
> >
> > GIGO.
> > > > Regards
> > > > Olivier Morin
> > > > On Thursday, June 17, 2021 at 10:28:09 PM UTC+2, Peter T. Daniels wrote:
> > > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > >
> > > > > > Journal information: Cognition
> > > > > >
> > > > > > Provided by Santa Fe Institute
> > > > > >
> > > > > > Phys.org
> > > > > It's not really about writing systems, but about the shapes of characters. It
> > > > > quantifies a variety of obvious observations, and though it says it's based
> > > > > on 133 scripts, it doesn't list them.
> > > > >
> > > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > > which often don't resemble what people actually write.
> > > > >
> > > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > > taken from the Latin alphabet. (The shapes used for printing came along later.)

To repeat, the link to the full open data is given on p. 4 of our paper (section 2.1). Yes, the paper we published in Cognition. Since you have denied this basic several times, there is probably nothing more to say. I can only assume you had the wrong copy of our paper, so here it is again:

https://www.sciencedirect.com/science/article/pii/S0010027721001906/pdfft?md5=69ac199bd7c11def6498e667741919b7&pid=1-s2.0-S0010027721001906-main.pdf

On Friday, June 25, 2021 at 3:35:08 PM UTC+2, Peter T. Daniels wrote:
> On Friday, June 25, 2021 at 2:40:01 AM UTC-4, Olivier Morin wrote:
> > Dear Peter Daniels,
> >
> > The file InventoryScripts.csv contains a list of all the scripts that were considered for inclusion. Of those, we selected 133. It is easy to figure out which by looking at the "ExclusionCriteria" field and setting it to the value "0". It helps to open this csv file in R incidentally. R is a software that people use to read data files.
> How would one go from the published version to that absurd procedure?
>
> Can you point to other published articles comparing writing systems that
> go to such lengths to conceal the database used in the study?
>
> Look at what you just said. First, the reader must somehow get from the
> published article to the page you here provided the link for. Second, the
> reader must discover from a fairly lengthy outline of contents that among
> the links is a list of scripts. Third, one must discover somewhere else some
> sort of list of criteria for culling that list. Fourth, one must infer how to apply
> those criteria in each of more than 200 items in order to discover which
> approximately half of them were considered, though there is no way of
> ascertaining whether one's guesses correspond with the actuality..
>
> I don't know what academic field you are coming from, but in that field,
> is it not the practice to include "supplementary information" files linked
> from the article, as has been the practice in *Science* and *Nature* for
> years, and is now even found in *Language: The Journal of the Linguistic
> Society of America*?
>
> It's shocking that a top-level journal like *Cognition* would publish an
> article that made it so difficult for the reader to discover what it was
> even talking about.
> > We do not in any way "rely" on Changizi, on the contrary, we saw flaws in his paper and re-tested his conclusions from scratch. Our conclusions are quite different from his. This comes out clearly from any minimally honest and careful reading of our paper.
> One cannot "honestly and carefully read" a paper about writing systems
> without knowing what writing systems it is about.
>
> You did, however, adopt their unacceptable reliance on the shapes displayed
> in the Unicode documentation and their further simplification them before
> performing your analysis.
> > You ask:
> > > The claim "over 47,000 characters" is not helpful. Does it include all the
> > > characters in most of the scripts, and several hundred Sinitic and Egyptian
> > > characters?
> > This question is already clearly answered in our paper. All eligible characters are included for
> "eligible"?
>
> Are the variants that were included in Unicode disregarded, or simply counted
> as individuals? The Unicode treatment of Cuneiform is especially heinous.
> Unicode does not separate Traditional from Simplified Chinese characters,
> nor ones that are Japanese-specific, nor ones that are essentially historic
> variants.
> > Chinese characters and Egyptian hieroglyphs (more than several hundred if memory serves). There
> "The largest Chinese dictionaries" include around 50,000 characters, because
> they list every variant known to the ancient scholars. Only around 700 Egyptian
> hieroglyphs were used during most of Egypt's history (some for only a single
> word), but in the Ptolemaic period, when that sort of learning was a niche subject,
> new hieroglyphs were invented supposedly bringing the inventory to several
> thousand. Do you deal only with character inventories that were actually used,
> or with everything you can find?
> are criteria for character inclusion which are the same across all scripts, please refer to our paper.
> >
> >> Are variants that the Unicode folk unaccountably list separately
> > > counted separately, and ones they unaccountably combined treated asone?
> >
> > It depends what you mean by "unaccountably". Like any large-scale comparative study ours is based on methodological choices that can be questioned.. Any methodological choice can be questioned, since no one is infallible. The interesting question to ask is whether those choices would have changed our results in any way.
> Noth8ng in the published article talks about "methodological choices."
>
> For instance, combining Greek and Coptic; for instance, separating by now
> probably about a dozen ancient Northwest Semitic scripts into separate
> blocks, and a perhaps greater number of variants of Indic scripts.
> > Regards,
> >
> > OM
> > On Thursday, June 24, 2021 at 2:47:22 PM UTC+2, Peter T. Daniels wrote:
> > > On Thursday, June 24, 2021 at 7:49:25 AM UTC-4, Peter T. Daniels wrote:
> > > > On Thursday, June 24, 2021 at 4:48:43 AM UTC-4, Olivier Morin wrote:
> > > >
> > > > > Dear Peter Daniels,
> > > > >
> > > > > I happened on this google groups discussion which discusses my paper with Helena Miton, "Graphic complexity in writing systems".
> > > > > I would like to attract your attention to the fact that we published our complete dataset, including (obviously), the list of 133 scripts used in this study. This is written in the paper on p. 4: 'Data and R scripts used to produce the results and figures can be found at https://osf.io/9dnj3/ '
> > > > > The fact that you missed this seems to suggest you barely skimmed this paper before rushing to dismiss it.
> > > > Thank you. That is not the article that was published in the journal *Cognition*.
> > > >
> > > > I was indeed put off by your reliance on Changizi et al., for my reaction to which
> > > > you can consult my book *An Exploration of Writing* (2018), p. 152.
> > > I find no list of the 133 scripts you used. I find a list of something over 200
> > > scripts, some of which are phantoms, some of which are toys (such as
> > > Shavian and Deseret).
> > >
> > > The claim "over 47,000 characters" is not helpful. Does it include all the
> > > characters in most of the scripts, and several hundred Sinitic and Egyptian
> > > characters? Are variants that the Unicode folk unaccountably list separately
> > > counted separately, and ones they unaccountably combined treated asone?
> > >
> > > (Changizi et al. must have selected just a few characters from each script
> > > they name in order to make their numbers come out, yet they excluded at
> > > least one of the most complicated-shaped scripts in use, Malayalam.)
> > >
> > > GIGO.
> > > > > Regards
> > > > > Olivier Morin
> > > > > On Thursday, June 17, 2021 at 10:28:09 PM UTC+2, Peter T. Daniels wrote:
> > > > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > > > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > > >
> > > > > > > Journal information: Cognition
> > > > > > >
> > > > > > > Provided by Santa Fe Institute
> > > > > > >
> > > > > > > Phys.org
> > > > > > It's not really about writing systems, but about the shapes of characters. It
> > > > > > quantifies a variety of obvious observations, and though it says it's based
> > > > > > on 133 scripts, it doesn't list them.
> > > > > >
> > > > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > > > which often don't resemble what people actually write.
> > > > > >
> > > > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > > > taken from the Latin alphabet. (The shapes used for printing came along later.)

On Friday, June 25, 2021 at 9:42:06 AM UTC-4, Olivier Morin wrote:
> To repeat, the link to the full open data is given on p. 4 of our paper (section 2.1). Yes, the paper we published in Cognition. Since you have denied this basic several times, there is probably nothing more to say. I can only assume you had the wrong copy of our paper, so here it is again:
>
> https://www.sciencedirect.com/science/article/pii/S0010027721001906/pdfft?md5=69ac199bd7c11def6498e667741919b7&pid=1-s2.0-S0010027721001906-main..pdf
> On Friday, June 25, 2021 at 3:35:08 PM UTC+2, Peter T. Daniels wrote:
> > On Friday, June 25, 2021 at 2:40:01 AM UTC-4, Olivier Morin wrote:
> > > Dear Peter Daniels,
> > >
> > > The file InventoryScripts.csv contains a list of all the scripts that were considered for inclusion. Of those, we selected 133. It is easy to figure out which by looking at the "ExclusionCriteria" field and setting it to the value "0". It helps to open this csv file in R incidentally. R is a software that people use to read data files.

Are you referring to this paragraph:

"2. Methods
"2.1. Pre-registration and data accessibility
"We kept a complete research diary on the Open Science Framework
website (https://osf.io/9dnj3/) where all analyses carried out were preregistered
and described. Pre-registration consists in describing both the
research design and analysis plan as independently as possible from data
collection. Data and R scripts used to produce the results and figures can
be found at https://osf.io/9dnj3/."

How is someone who has been reading the linguistics literature for
about 50 years supposed to know that that is where you have buried
(under five layers of seeking) the most basic information?

"Preregistration" is not part of the publication process in linguistics.

Even if one managed to decode "can be found" -- "Data and R scripts"
doesn't exactly suggest that actual data are to be found there, but
some sort of "Data scripts" -- the actual "finding" when given that link
is hardly perspicacious.

Do you concede that nowhere do you provide a list of the 133 scripts
involved? At least Changizi et al. made it easy to discover how distorted
their claim about their database was.

You still haven't revealed what realm of investigation this study
represents. Not the study of writing systems. Possibly the study
of visual processing?

> > How would one go from the published version to that absurd procedure?
> >
> > Can you point to other published articles comparing writing systems that
> > go to such lengths to conceal the database used in the study?
> >
> > Look at what you just said. First, the reader must somehow get from the
> > published article to the page you here provided the link for. Second, the
> > reader must discover from a fairly lengthy outline of contents that among
> > the links is a list of scripts. Third, one must discover somewhere else some
> > sort of list of criteria for culling that list. Fourth, one must infer how to apply
> > those criteria in each of more than 200 items in order to discover which
> > approximately half of them were considered, though there is no way of
> > ascertaining whether one's guesses correspond with the actuality..
> >
> > I don't know what academic field you are coming from, but in that field,
> > is it not the practice to include "supplementary information" files linked
> > from the article, as has been the practice in *Science* and *Nature* for
> > years, and is now even found in *Language: The Journal of the Linguistic
> > Society of America*?
> >
> > It's shocking that a top-level journal like *Cognition* would publish an
> > article that made it so difficult for the reader to discover what it was
> > even talking about.
> > > We do not in any way "rely" on Changizi, on the contrary, we saw flaws in his paper and re-tested his conclusions from scratch. Our conclusions are quite different from his. This comes out clearly from any minimally honest and careful reading of our paper.
> > One cannot "honestly and carefully read" a paper about writing systems
> > without knowing what writing systems it is about.
> >
> > You did, however, adopt their unacceptable reliance on the shapes displayed
> > in the Unicode documentation and their further simplification them before
> > performing your analysis.
> > > You ask:
> > > > The claim "over 47,000 characters" is not helpful. Does it include all the
> > > > characters in most of the scripts, and several hundred Sinitic and Egyptian
> > > > characters?
> > > This question is already clearly answered in our paper. All eligible characters are included for
> > "eligible"?
> >
> > Are the variants that were included in Unicode disregarded, or simply counted
> > as individuals? The Unicode treatment of Cuneiform is especially heinous.
> > Unicode does not separate Traditional from Simplified Chinese characters,
> > nor ones that are Japanese-specific, nor ones that are essentially historic
> > variants.
> > > Chinese characters and Egyptian hieroglyphs (more than several hundred if memory serves). There
> > "The largest Chinese dictionaries" include around 50,000 characters, because
> > they list every variant known to the ancient scholars. Only around 700 Egyptian
> > hieroglyphs were used during most of Egypt's history (some for only a single
> > word), but in the Ptolemaic period, when that sort of learning was a niche subject,
> > new hieroglyphs were invented supposedly bringing the inventory to several
> > thousand. Do you deal only with character inventories that were actually used,
> > or with everything you can find?
> > are criteria for character inclusion which are the same across all scripts, please refer to our paper.
> > >
> > >> Are variants that the Unicode folk unaccountably list separately
> > > > counted separately, and ones they unaccountably combined treated asone?
> > >
> > > It depends what you mean by "unaccountably". Like any large-scale comparative study ours is based on methodological choices that can be questioned. Any methodological choice can be questioned, since no one is infallible.. The interesting question to ask is whether those choices would have changed our results in any way.
> > Noth8ng in the published article talks about "methodological choices."
> >
> > For instance, combining Greek and Coptic; for instance, separating by now
> > probably about a dozen ancient Northwest Semitic scripts into separate
> > blocks, and a perhaps greater number of variants of Indic scripts.
> > > Regards,
> > >
> > > OM
> > > On Thursday, June 24, 2021 at 2:47:22 PM UTC+2, Peter T. Daniels wrote:
> > > > On Thursday, June 24, 2021 at 7:49:25 AM UTC-4, Peter T. Daniels wrote:
> > > > > On Thursday, June 24, 2021 at 4:48:43 AM UTC-4, Olivier Morin wrote:
> > > > >
> > > > > > Dear Peter Daniels,
> > > > > >
> > > > > > I happened on this google groups discussion which discusses my paper with Helena Miton, "Graphic complexity in writing systems".
> > > > > > I would like to attract your attention to the fact that we published our complete dataset, including (obviously), the list of 133 scripts used in this study. This is written in the paper on p. 4: 'Data and R scripts used to produce the results and figures can be found at https://osf.io/9dnj3/ '
> > > > > > The fact that you missed this seems to suggest you barely skimmed this paper before rushing to dismiss it.
> > > > > Thank you. That is not the article that was published in the journal *Cognition*.
> > > > >
> > > > > I was indeed put off by your reliance on Changizi et al., for my reaction to which
> > > > > you can consult my book *An Exploration of Writing* (2018), p. 152.
> > > > I find no list of the 133 scripts you used. I find a list of something over 200
> > > > scripts, some of which are phantoms, some of which are toys (such as
> > > > Shavian and Deseret).
> > > >
> > > > The claim "over 47,000 characters" is not helpful. Does it include all the
> > > > characters in most of the scripts, and several hundred Sinitic and Egyptian
> > > > characters? Are variants that the Unicode folk unaccountably list separately
> > > > counted separately, and ones they unaccountably combined treated asone?
> > > >
> > > > (Changizi et al. must have selected just a few characters from each script
> > > > they name in order to make their numbers come out, yet they excluded at
> > > > least one of the most complicated-shaped scripts in use, Malayalam.)
> > > >
> > > > GIGO.
> > > > > > Regards
> > > > > > Olivier Morin
> > > > > > On Thursday, June 17, 2021 at 10:28:09 PM UTC+2, Peter T. Daniels wrote:
> > > > > > > On Wednesday, June 16, 2021 at 10:28:48 PM UTC-4, daud....@gmail.com wrote:
> > > > > > > > Helena Miton et al, Graphic complexity in writing systems, Cognition (2021). DOI: 10.1016/j.cognition.2021.104771
> > > > > > > >
> > > > > > > > Journal information: Cognition
> > > > > > > >
> > > > > > > > Provided by Santa Fe Institute
> > > > > > > >
> > > > > > > > Phys.org
> > > > > > > It's not really about writing systems, but about the shapes of characters. It
> > > > > > > quantifies a variety of obvious observations, and though it says it's based
> > > > > > > on 133 scripts, it doesn't list them.
> > > > > > >
> > > > > > > And it doesn't use the actual shapes, but it manipulates the Unicode shapes
> > > > > > > which often don't resemble what people actually write.
> > > > > > >
> > > > > > > And there are various factual errors -- it excludes Cherokee from its "invented
> > > > > > > scripts" category because it mistakenly things the shapes of the letters were
> > > > > > > taken from the Latin alphabet. (The shapes used for printing came along later.)