Discussion:
PIE
(too old to reply)
DKleinecke
2017-12-27 21:26:18 UTC
Permalink
Raw Message
Contrary to its usual custom these days the current issue of
Language has an article on historical linguistics (PIE to be
exact). One of the things the article has is a tree of IE
languages (we all have seen at least one of these in the past).

The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.

Is this GAA cluster the current consensus among historical
linguists?

PS: The tree does have a (short-lifted) Celtic/Italic node.
Peter T. Daniels
2017-12-27 21:44:10 UTC
Permalink
Raw Message
Post by DKleinecke
Contrary to its usual custom these days the current issue of
Language has an article on historical linguistics (PIE to be
exact). One of the things the article has is a tree of IE
languages (we all have seen at least one of these in the past).
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
If it doesn't mention the other Anatolian languages, you needn't worry about
the omission of Tocharian A.

(I haven't taken it out of its plastic wrap yet, but since Johanna Nichols is
among the authors, it's probably sound.)
Post by DKleinecke
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
PS: The tree does have a (short-lifted) Celtic/Italic node.
Greek/Armenian has been a popular pair for several decades. I forget the name
of the author who publicized it, but I recall that Eric Hamp wasn't persuaded.
Arnaud Fournet
2017-12-28 08:12:40 UTC
Permalink
Raw Message
Post by DKleinecke
Contrary to its usual custom these days the current issue of
Language has an article on historical linguistics (PIE to be
exact). One of the things the article has is a tree of IE
languages (we all have seen at least one of these in the past).
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise.
Indeed. no surprise.
A.
Post by DKleinecke
The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A.
It seems that Tocharian A was already dead at the time Tocharian B scribes still made some mention of it in their manuscripts.
A.
Post by DKleinecke
The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
Personally, I'm not sure Albanian should be lumped with satem languages.
Anyway, little work is made on this language, and there are still several theories about how phonology evolved in Albanian.
Meillet and several French linguists liked to put Greek and Armenian together, but I'm not sure this is really solid. If you look at the words that Armenian shares with only one other IEan language, quite often it's with Balto-Slavic not with Greek.
I don't know what their reasons are for advocating this GAA cluster. IMO it's not a valid node.
A.
Christian Weisgerber
2017-12-28 15:51:14 UTC
Permalink
Raw Message
Post by DKleinecke
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
What's the question here? The grouping of G/A/A, or its position
as the third split?

In _From Proto-Indo-European to Proto-Germanic_, Don Ringe presents
this tree:

PIE
/ \
Anatolian North IE
/ \
Tocharian West IE
/ \
Italo-Celtic Central IE
/ \ /\
Celtic Italic / \


“The ‘Central’ subgroup includes Germanic, Balto-Slavic, Indo-Iranian,
Armenian, Greek, and probably Albanian; its internal subgrouping
is still very unclear, though it seems likely that Indo-Iranian,
Balto-Slavic, and Germanic were parts of a dialect chain at a
very early date.”

This was in 2006, but a quick peek with Amazon.com's “Look inside”
shows no change in the 2017 second edition.

Ringe's area is historical linguistics (Greek, Tocharian, Germanic)
and he has also worked on linguistic phylogenies. I'm pretty sure
this paper from 2002 has been mentioned here in the past:
“Indo-European and Computational Cladistics”
https://www.cs.rice.edu/~nakhleh/CPHL/RWT02.pdf
Abstract: This paper reports the results of an attempt to recover
the first-order subgrouping of the Indo-European family using a
new computational method devised by the authors and based on a
‘perfect phylogeny’ algorithm. The methodology is also briefly
described, and points of theory and methodology are addressed in
connection with the experiment whose results are here reported.
--
Christian "naddy" Weisgerber ***@mips.inka.de
DKleinecke
2017-12-28 18:41:33 UTC
Permalink
Raw Message
Post by Christian Weisgerber
Post by DKleinecke
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
What's the question here? The grouping of G/A/A, or its position
as the third split?
In _From Proto-Indo-European to Proto-Germanic_, Don Ringe presents
PIE
/ \
Anatolian North IE
/ \
Tocharian West IE
/ \
Italo-Celtic Central IE
/ \ /\
Celtic Italic / \
“The ‘Central’ subgroup includes Germanic, Balto-Slavic, Indo-Iranian,
Armenian, Greek, and probably Albanian; its internal subgrouping
is still very unclear, though it seems likely that Indo-Iranian,
Balto-Slavic, and Germanic were parts of a dialect chain at a
very early date.”
This was in 2006, but a quick peek with Amazon.com's “Look inside”
shows no change in the 2017 second edition.
Ringe's area is historical linguistics (Greek, Tocharian, Germanic)
and he has also worked on linguistic phylogenies. I'm pretty sure
“Indo-European and Computational Cladistics”
https://www.cs.rice.edu/~nakhleh/CPHL/RWT02.pdf
Abstract: This paper reports the results of an attempt to recover
the first-order subgrouping of the Indo-European family using a
new computational method devised by the authors and based on a
‘perfect phylogeny’ algorithm. The methodology is also briefly
described, and points of theory and methodology are addressed in
connection with the experiment whose results are here reported.
You should look at the article in Language to see how their
tree was developed. I am not competent to explain it.

I observe it is rather different than Ringe's tree which I was
not familiar with. I was asking about the validity of the GAA
cluster but now I see that it's position is also controversial.
Christian Weisgerber
2017-12-28 21:40:36 UTC
Permalink
Raw Message
Post by DKleinecke
You should look at the article in Language to see how their
tree was developed. I am not competent to explain it.
I don't have access to _Language_. The article you so coyly allude
to seems to be this one:
“NP recursion over time: Evidence from Indo-European”
M Widmer, S Auderset, J Nichols, P Widmer, B Bickel - Language, 2017

A preprint is available here:
http://www.comparativelinguistics.uzh.ch/dam/jcr:10daee1b-7a0b-4d6c-9f4a-e5f00fdb8176/NPrecursion_finalMS.pdf

The tree you refer to is presumably figure 2. From what I understand,
that paper doesn't develop the tree at all. They use it to check
an hypothesis. The tree is adapted from this paper:

“Ancestry-constrained phylogenetic analysis supports the Indo-European
steppe hypothesis”
W Chang, C Cathcart, D Hall, A Garrett - Language, 2015
Preprint:
https://www.linguisticsociety.org/sites/default/files/news/ChangEtAlPreprint.pdf

Figure 2.
--
Christian "naddy" Weisgerber ***@mips.inka.de
DKleinecke
2017-12-29 04:16:39 UTC
Permalink
Raw Message
Post by Christian Weisgerber
Post by DKleinecke
You should look at the article in Language to see how their
tree was developed. I am not competent to explain it.
I don't have access to _Language_. The article you so coyly allude
“NP recursion over time: Evidence from Indo-European”
M Widmer, S Auderset, J Nichols, P Widmer, B Bickel - Language, 2017
http://www.comparativelinguistics.uzh.ch/dam/jcr:10daee1b-7a0b-4d6c-9f4a-e5f00fdb8176/NPrecursion_finalMS.pdf
The tree you refer to is presumably figure 2. From what I understand,
that paper doesn't develop the tree at all. They use it to check
“Ancestry-constrained phylogenetic analysis supports the Indo-European
steppe hypothesis”
W Chang, C Cathcart, D Hall, A Garrett - Language, 2015
https://www.linguisticsociety.org/sites/default/files/news/ChangEtAlPreprint.pdf
With modifications. I didn't get explicit because my question
did not involve how the tree was obtained - only that it existed.

I am not completely sold on cladistic methodology - but the
fact that it compares well to results obtained by other methods
shows that cladistics should not be dismissed out of hand. I
would say that original PIE (prior to Anatolian and Tocharian)
has not resolved any of its internal problems with the aid of
Anatolian or Tocharian. That is, assuming the laryngeals were
already accepted (which is not historically true)
Arnaud Fournet
2017-12-29 08:20:03 UTC
Permalink
Raw Message
Post by Christian Weisgerber
Post by DKleinecke
You should look at the article in Language to see how their
tree was developed. I am not competent to explain it.
I don't have access to _Language_. The article you so coyly allude
“NP recursion over time: Evidence from Indo-European”
M Widmer, S Auderset, J Nichols, P Widmer, B Bickel - Language, 2017
http://www.comparativelinguistics.uzh.ch/dam/jcr:10daee1b-7a0b-4d6c-9f4a-e5f00fdb8176/NPrecursion_finalMS.pdf
The tree you refer to is presumably figure 2. From what I understand,
that paper doesn't develop the tree at all. They use it to check
“Ancestry-constrained phylogenetic analysis supports the Indo-European
steppe hypothesis”
W Chang, C Cathcart, D Hall, A Garrett - Language, 2015
https://www.linguisticsociety.org/sites/default/files/news/ChangEtAlPreprint.pdf
Figure 2.
--
Thanks for providing the references.
A.
Arnaud Fournet
2017-12-29 08:42:32 UTC
Permalink
Raw Message
Post by Christian Weisgerber
Post by DKleinecke
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
What's the question here? The grouping of G/A/A, or its position
as the third split?
In _From Proto-Indo-European to Proto-Germanic_, Don Ringe presents
PIE
/ \
Anatolian North IE
/ \
Tocharian West IE
/ \
Italo-Celtic Central IE
/ \ /\
Celtic Italic / \
“The ‘Central’ subgroup includes Germanic, Balto-Slavic, Indo-Iranian,
Armenian, Greek, and probably Albanian; its internal subgrouping
is still very unclear, though it seems likely that Indo-Iranian,
Balto-Slavic, and Germanic were parts of a dialect chain at a
very early date.”
Personally, I would agree with Ringe's Central subgroup.
Both Germanic and Armenian are a phonetically evolved form of Indo-Iranian.
These three groups created a fourth series of consonants out of stop+laryngeal clusters: like t+H > *th.
This innovation is unheardof in the other languages, not even in Greek.
Old Indian is conspicuous for having stabilized the four series into a consistent system.
Germanic and Armenian have somehow parallel phonetic paths, but for the centum/satem difference.
I can't see a single phonetic innovation that Germanic would share with Celtic or Italic.
A.
DKleinecke
2017-12-29 18:16:33 UTC
Permalink
Raw Message
Post by Arnaud Fournet
Post by Christian Weisgerber
Post by DKleinecke
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
What's the question here? The grouping of G/A/A, or its position
as the third split?
In _From Proto-Indo-European to Proto-Germanic_, Don Ringe presents
PIE
/ \
Anatolian North IE
/ \
Tocharian West IE
/ \
Italo-Celtic Central IE
/ \ /\
Celtic Italic / \
“The ‘Central’ subgroup includes Germanic, Balto-Slavic, Indo-Iranian,
Armenian, Greek, and probably Albanian; its internal subgrouping
is still very unclear, though it seems likely that Indo-Iranian,
Balto-Slavic, and Germanic were parts of a dialect chain at a
very early date.”
Personally, I would agree with Ringe's Central subgroup.
Both Germanic and Armenian are a phonetically evolved form of Indo-Iranian.
These three groups created a fourth series of consonants out of stop+laryngeal clusters: like t+H > *th.
This innovation is unheardof in the other languages, not even in Greek.
Old Indian is conspicuous for having stabilized the four series into a consistent system.
Germanic and Armenian have somehow parallel phonetic paths, but for the centum/satem difference.
I can't see a single phonetic innovation that Germanic would share with Celtic or Italic.
I must leave this matter to PIE specialists. But if forced to
have an opinion I would say it looks like a dialect continuum
somewhere near southeastern Europe fragmented. I speculate
that there was a political entity - an empire if you will -
that held the continuum together for a long time and then
collapsed. The dialects of this empire became the original
PIE languages. Anatolian had already diverged. I don't know
enough about Tocharian to have a opinion about it.
António Marques
2017-12-29 19:42:52 UTC
Permalink
Raw Message
Post by DKleinecke
Post by Arnaud Fournet
Post by Christian Weisgerber
Post by DKleinecke
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
What's the question here? The grouping of G/A/A, or its position
as the third split?
In _From Proto-Indo-European to Proto-Germanic_, Don Ringe presents
PIE
/ \
Anatolian North IE
/ \
Tocharian West IE
/ \
Italo-Celtic Central IE
/ \ /\
Celtic Italic / \
“The ‘Central’ subgroup includes Germanic, Balto-Slavic, Indo-Iranian,
Armenian, Greek, and probably Albanian; its internal subgrouping
is still very unclear, though it seems likely that Indo-Iranian,
Balto-Slavic, and Germanic were parts of a dialect chain at a
very early date.”
Personally, I would agree with Ringe's Central subgroup.
Both Germanic and Armenian are a phonetically evolved form of Indo-Iranian.
These three groups created a fourth series of consonants out of
stop+laryngeal clusters: like t+H > *th.
This innovation is unheardof in the other languages, not even in Greek.
Old Indian is conspicuous for having stabilized the four series into a consistent system.
Germanic and Armenian have somehow parallel phonetic paths, but for the
centum/satem difference.
I can't see a single phonetic innovation that Germanic would share with Celtic or Italic.
I must leave this matter to PIE specialists. But if forced to
have an opinion I would say it looks like a dialect continuum
somewhere near southeastern Europe fragmented. I speculate
that there was a political entity - an empire if you will -
that held the continuum together for a long time and then
collapsed. The dialects of this empire became the original
PIE languages. Anatolian had already diverged. I don't know
enough about Tocharian to have a opinion about it.
I find this unlikely because such a political entity would require an
amount of urbanisation that would have left traces we'd have already found.

Or maybe not. Maybe a number of cities in the Balkans are much older than
they seem.
DKleinecke
2017-12-30 01:48:03 UTC
Permalink
Raw Message
Post by António Marques
Post by DKleinecke
Post by Arnaud Fournet
Post by Christian Weisgerber
Post by DKleinecke
The tree is not very exciting but I was taken by its oldest
splits. The first split in PIE separates Hittite/Luvian from
everybody else - no surprise. The second split separates
Tocharian B from everyone - there is no mention of Tocharian
A. The third split separates Greek/Albanian/Armenian from
all the rest. This is the one that interests me. All the
later splits are obvious.
Is this GAA cluster the current consensus among historical
linguists?
What's the question here? The grouping of G/A/A, or its position
as the third split?
In _From Proto-Indo-European to Proto-Germanic_, Don Ringe presents
PIE
/ \
Anatolian North IE
/ \
Tocharian West IE
/ \
Italo-Celtic Central IE
/ \ /\
Celtic Italic / \
“The ‘Central’ subgroup includes Germanic, Balto-Slavic, Indo-Iranian,
Armenian, Greek, and probably Albanian; its internal subgrouping
is still very unclear, though it seems likely that Indo-Iranian,
Balto-Slavic, and Germanic were parts of a dialect chain at a
very early date.”
Personally, I would agree with Ringe's Central subgroup.
Both Germanic and Armenian are a phonetically evolved form of Indo-Iranian.
These three groups created a fourth series of consonants out of
stop+laryngeal clusters: like t+H > *th.
This innovation is unheardof in the other languages, not even in Greek.
Old Indian is conspicuous for having stabilized the four series into a consistent system.
Germanic and Armenian have somehow parallel phonetic paths, but for the
centum/satem difference.
I can't see a single phonetic innovation that Germanic would share with Celtic or Italic.
I must leave this matter to PIE specialists. But if forced to
have an opinion I would say it looks like a dialect continuum
somewhere near southeastern Europe fragmented. I speculate
that there was a political entity - an empire if you will -
that held the continuum together for a long time and then
collapsed. The dialects of this empire became the original
PIE languages. Anatolian had already diverged. I don't know
enough about Tocharian to have a opinion about it.
I find this unlikely because such a political entity would require an
amount of urbanisation that would have left traces we'd have already found.
Or maybe not. Maybe a number of cities in the Balkans are much older than
they seem.
I wonder how many lost empires have vanished without leaving
a trace. The Inca ran a fairly large empire without much
surviving infrastructure and the shadowy empires that
preceded the Inca have left even less. The monoliths-builders
of England, France, Malta and doubtless elsewhere proved that
very early cultures could be quite energetic.
Christian Weisgerber
2017-12-29 21:32:08 UTC
Permalink
Raw Message
Post by DKleinecke
I must leave this matter to PIE specialists. But if forced to
have an opinion I would say it looks like a dialect continuum
somewhere near southeastern Europe fragmented. I speculate
that there was a political entity - an empire if you will -
that held the continuum together for a long time and then
collapsed. The dialects of this empire became the original
PIE languages. Anatolian had already diverged. I don't know
enough about Tocharian to have a opinion about it.
From
“The Indo-European Homeland from Linguistic and Archaeological
Perspectives”
DW Anthony, D Ringe - Annu. Rev. Linguist., 2015
http://www.annualreviews.org/doi/full/10.1146/annurev-linguist-030514-124812

...
In contrast, the Pontic-Caspian homeland can be accommodated to the
known relationships between the IE daughter branches because migrations
are archaeologically documented to have occurred out of the
Pontic-Caspian steppes into neighboring regions in the sequence and
direction (Figure 2) that are demanded by the oldest three branchings of
the IE tree (Ringe et al. 2002, p. 90). Pre-Anatolian separated to the
west, into southeastern Europe, about 4200–4000 bce with the Suvorovo
migration, the first archaeologically visible migration out of the
steppes (Bicbaev 2010). Then Pre-Tocharian separated to the east, with
the Afanasievo migration into the western Altai Mountains beginning
about 3300 bce, the second archaeologically visible migration out of the
Pontic-Caspian steppes, matching Ringe et al.’s and Winter’s (1998)
expectation that Tocharian was the second branch to separate. After
that, a cluster of western European branches separated to the west, into
the Danube valley on the south side of the Carpathians with the Yamnaya
migration up the Danube about 3100–2800 bce, and into southern Poland on
the northern side of the Carpathians with the expansions of the Usatovo
and the Tripolye C2 cultures about 3300–3000 bce (Ecsedy 1994, Mallory
1998, Klochko & Kośko 2009, Heyd 2011, Anthony 2013). These last
separations match the proposal that the ancestors of Italic and Celtic
(and perhaps pre-Germanic) could have separated in a rather complex
phase of migrations and language spreads. The later spread of
Indo-Iranian languages into Central Asia, Iran, and South Asia from the
steppes after 2000 bce is the same in our hypothesis and in Plan B by
Renfrew (1987, pp. 197–205), which he prefers (Renfrew 2002b, p. 6). A
steppe homeland satisfies the archaeology-and-language relationship test
by exhibiting archaeological evidence for migrations in the direction
and sequence suggested by linguistic evidence.
...
--
Christian "naddy" Weisgerber ***@mips.inka.de
DKleinecke
2017-12-30 01:42:06 UTC
Permalink
Raw Message
Post by Christian Weisgerber
Post by DKleinecke
I must leave this matter to PIE specialists. But if forced to
have an opinion I would say it looks like a dialect continuum
somewhere near southeastern Europe fragmented. I speculate
that there was a political entity - an empire if you will -
that held the continuum together for a long time and then
collapsed. The dialects of this empire became the original
PIE languages. Anatolian had already diverged. I don't know
enough about Tocharian to have a opinion about it.
From
“The Indo-European Homeland from Linguistic and Archaeological
Perspectives”
DW Anthony, D Ringe - Annu. Rev. Linguist., 2015
http://www.annualreviews.org/doi/full/10.1146/annurev-linguist-030514-124812
...
In contrast, the Pontic-Caspian homeland can be accommodated to the
known relationships between the IE daughter branches because migrations
are archaeologically documented to have occurred out of the
Pontic-Caspian steppes into neighboring regions in the sequence and
direction (Figure 2) that are demanded by the oldest three branchings of
the IE tree (Ringe et al. 2002, p. 90). Pre-Anatolian separated to the
west, into southeastern Europe, about 4200–4000 bce with the Suvorovo
migration, the first archaeologically visible migration out of the
steppes (Bicbaev 2010). Then Pre-Tocharian separated to the east, with
the Afanasievo migration into the western Altai Mountains beginning
about 3300 bce, the second archaeologically visible migration out of the
Pontic-Caspian steppes, matching Ringe et al.’s and Winter’s (1998)
expectation that Tocharian was the second branch to separate. After
that, a cluster of western European branches separated to the west, into
the Danube valley on the south side of the Carpathians with the Yamnaya
migration up the Danube about 3100–2800 bce, and into southern Poland on
the northern side of the Carpathians with the expansions of the Usatovo
and the Tripolye C2 cultures about 3300–3000 bce (Ecsedy 1994, Mallory
1998, Klochko & Kośko 2009, Heyd 2011, Anthony 2013). These last
separations match the proposal that the ancestors of Italic and Celtic
(and perhaps pre-Germanic) could have separated in a rather complex
phase of migrations and language spreads. The later spread of
Indo-Iranian languages into Central Asia, Iran, and South Asia from the
steppes after 2000 bce is the same in our hypothesis and in Plan B by
Renfrew (1987, pp. 197–205), which he prefers (Renfrew 2002b, p. 6). A
steppe homeland satisfies the archaeology-and-language relationship test
by exhibiting archaeological evidence for migrations in the direction
and sequence suggested by linguistic evidence.
All this is quite plausible BUT other people have reached other
conclusion and I can see no way to reconcile everybody. My rule
of thumb in such cases is to seek a way in which everybody can
be more-or-less right. Hence my speculation.
Christian Weisgerber
2017-12-30 17:07:48 UTC
Permalink
Raw Message
Post by DKleinecke
All this is quite plausible BUT other people have reached other
conclusion and I can see no way to reconcile everybody. My rule
of thumb in such cases is to seek a way in which everybody can
be more-or-less right. Hence my speculation.
So basically you're suggesting the shape of the earth is a hemisphere.
--
Christian "naddy" Weisgerber ***@mips.inka.de
DKleinecke
2017-12-30 18:39:39 UTC
Permalink
Raw Message
Post by Christian Weisgerber
Post by DKleinecke
All this is quite plausible BUT other people have reached other
conclusion and I can see no way to reconcile everybody. My rule
of thumb in such cases is to seek a way in which everybody can
be more-or-less right. Hence my speculation.
So basically you're suggesting the shape of the earth is a hemisphere.
--
Cute - but no analogy

Loading...