Some thoughts on new technologies in archaeology

[E.P. Grondine]

Moving on, one of the big problems in archaeology is bags or trays of shards, and sorting them into ceramics, and then gluing them together.

Obviously, for larger sites a computer database program might help.
Identifying Features:
type of ceramic vessel if known;
fabric;
if painted,inside slip; outside slip; any decorative painting;
if not painted, any outside decoration;
top fragment;
side fragment;
bottom fragment;
thickness;
circumference (for side fragments only?);

Then you have the sorting trays; their construction and labeling.

Any thoughts?

[circumspice]

Really E.P., This is a non-issue. This question has been dealt with for decades at the very least, possibly approaching a century now.

Look up the work of Petrie, Aubrey & Shepherd.

https://greekarchaeology.osu.edu/arch-edu/analysis

https://sha.org/conservation-facts/faq/ ... materials/

[E.P. Grondine]

H spice -

The problem is to turn those piles of shards (like the ones shown in your first link) into complete ceramics for display.

Now perhaps AI could be used with holograms or 3d images, but otherwise its sorting trays followed by human eye attempted joins.

In my quick scan (perhaps too hasty) I didn't notice anything about glues and glue solvents, nor about materials for fills.

[circumspice]

H spice -

The problem is to turn those piles of shards (like the ones shown in your first link) into complete ceramics for display.

Now perhaps AI could be used with holograms or 3d images, but otherwise its sorting trays followed by human eye attempted joins.

In my quick scan (perhaps too hasty) I didn't notice anything about glues and glue solvents, nor about materials for fills.

This info is readily available on the internet EP...

https://sha.org/conservation-facts/faq/ ... reatments/


https://www.google.com/url?sa=t&source= ... Y352KMC_eA

[E.P. Grondine]

H spice -

The problem is to turn those piles of shards (like the ones shown in your first link) into complete ceramics for display.

Now perhaps AI could be used with holograms or 3d images, but otherwise its sorting trays followed by human eye attempted joins.

In my quick scan (perhaps too hasty) I didn't notice anything about glues and glue solvents, nor about materials for fills.

This info is readily available on the internet EP...

https://sha.org/conservation-facts/faq/ ... reatments/


https://www.google.com/url?sa=t&source= ... Y352KMC_eA

Hi spice -

I'll have to differ with your first source:
"and that reconstructed ceramics require more space to store than shards"

Well, visitors would rather see reconstructed ceramics, and they provide more information.
It is attention to details like this that determine who is good, and who not so good, IMO.

perhaps Ai and 3D imaging will some day be applied to the task.

[E.P. Grondine]

HOW TO ATTACK CYPRO-MINOAN -

There are probably more elegant linear algebra approaches to this problem,
but I've had a stroke, and getting enough peace and quiet to even roughly work this through is difficult.
That said, (and this is not a complete solution)...

What is known is that planes roughly approximate the conic sections
formed as the product of two parabolas of the vowels and consonants as distributed by Zipf functions.

So what we are looking for is matching strings of the types [k1V1/K2V2, K1V1/K1V3/, K1V1/K1V4, K1V1/K1V5...
or [K1V1/K2/V1, K1V1/K3VI, K1v1/K4V1...], the lines to define a phonetic plane.

First confine the character count to utterances - 3 characters or more.

Very long variable precision maths will be necessary.
Depending on processors and variable types,
the counts for small samples my need to be multiplied by 100, 1,000 or even 10,000.

Now one could try a brute force attack, randomly step wise assigning a set of consonants and vowels to the character set,
and then seeing how close the results approximate a plane.
Both the number of consonants and vowels are unknown,
with likely numbers of consonants ranging from way 26 to 52 (where aspirates attach to consonants instead of vowels),
and the number of vowels ranging from 3, 4, 5 or 6.
On a parrallel computer with a sufficient number of processors,
this task can be distributed and run fairly quickly

But let's see if analysis can limit the task,
down to something a workstation can handle.

Calculate ratio of the occurrence each syllabic sign against all other individual syllabic signs.
The results will have the form of an NxN matrix, where N = the number of signs in the syllabary,
with elements in the form of [Ratio, Sign W/Sign X].
and having a diagonal of 1's,

Next, subtract each ratio from all the others, and sort the results by magnitude;
this will produce elements in the form [difference, (Ratio, Sign W/Sign X)-( Ratio, Sign Y/Sign Z)]
For identities where SignW=SignX=Sign Y=Sign Z the difference will be = 0.

My assumption here is that the ratios [K1/K2]-[K1/K2] or [V1/V2]-[V1/V2] should show as as a series of values near 0.

So sort the rows of this NxN matrrix with values near zero by magnitude.

Our object is to find chains of the type
K1V1/K2V2 = K3/V1/K4/V2 = K5V1/K6/V2... to the number of consonants in order to identify the number of vowels,
where these are most likely long chains,
and of the type
K1V1/K2V2 = K1V3/K2V4 = K1V5/K2V6.. to the number of vowels to identify the consonants,
where these are most likely short. chains of the 3,4, 5 or 6 elements in length.

{For a proper Gausian fit our initial variables should have initially been of the form
[Ratio, Sign W, (number of occurrences of Sign W) /Sign X (number of occurrence Sign X)].
But note that we are using a plane to roughly approximate the a conic section
formed as the product of two parabolas of the vowels and consonants as distributed by Zipf functions.
so the differences have meaning.]

We now have some idea of the number of vowels and consonants in the syllabary's language.

OR

As my aasumption is that the ratios [K1/K2]-[K1/K2] or [V1/V2]-[V1-V2] should show as as a series of values near 0.

Now take the various ratio values and multiply them against the NxN matrix of the original ratios.
Our initial ratios were of the form K1V1/K2V2.

By multiplication, we should now have some products of the forms
K1V1/K2V2 *K2/K1 =V1/V2.
or K1V1/k2V2 *V2/V1=K1/K2.

To identify those regular ratios, take the differences of this matrix's elements and sort by the value of their differences

Now we should have some ratios of the type
K1V1/K2V2 *K2/K1 =V1/V2.
or K1V1/k2V2 *V2/V1=K1/K2.

Now, using these results, make an estimate of the phonic plane-
that is a plane which represents the language's information loading,
and then sort the characters to that phonic plane

Take the total character sign count and multiply it by these ratios, and compare with the actual character occurrences.

Run your programs against Linear B or Japanese kana to check them.

After this, compare that plane with the phonetic planes for IE, PIE, and Semetic.

Finally, make some guesses -
Try to identify common phrases,
assign likely phonetic values based on the phonetic planes for nearby languages and do lexical comparisons,
or compare likely phonetic values based on similarly shaped characters in other systems.

In the case where similarly shaped signs share either consonants or vowels,
these ratios can be used to estimate the total character count for other instances of the same shared phonemes.
Bur as we know from the experience of using the Greek Cypriot Syllabary to break Linear B, this may be of very limited use.

If you are dealing with a vowel inflexive language [say something like Etruscan],
looking for declensions like Ventris did with Linear B will not work too well.
"G' type bomber navigation maths, like those Ventris used, will not work very well
in identifying characters with shared phoneme pairs.

A tip of the hat here to Dr. Brown, and his breaking of Linear A.

[circumspice]

How to attack Cypro-Minoan -

First confine character count to utterances - 3 characters or more.

Second, calculate ratio of each character against all other characters.

Third, sum up each ratio occurrence and bin them.

Fourth, sort characters and construct a phonic plane - that is a plane which represents information loading.

Fifth, compare that plane with the planes for IE, PIE, and semetic.

Sixth, make some guesses -
compare likely phonetic values based on similarly shaped characters,
or assign likely phonetic values and do lexical comparisons.

A tip of the hat here to Dr. Brown, and his breaking of Linear A.

Linear A has been deciphered? When? Where? By whom?

[E.P. Grondine]

Here you go, spice -

http://classics.unc.edu/people-3/faculty/edwin-brown/

http://www.people.ku.edu/~jyounger/LinearA/biblio.html

see also:
https://books.google.com/books?id=a_B8V ... 0A&f=false

As far as the entry of the Hittites into Anatolia,
this text may have some bearing on the problem:

http://abob.libs.uga.edu/bobk/ccc/cc012099.html

Naturally, everyone would like to get at the inscriptions at Carcamesh,
from T.E. Lawrence on.

The relationship between Hurrian and Luwian is of interest.

"The Lycians are in good truth anciently from Crete; which island, in former days, was wholly peopled with barbarians [non Greek speakers].

A quarrel arising there between the two sons of Europa, Sarpedon and Minos, as to which of them should be king, Minos, whose party prevailed, drove Sarpedon and his followers into banishment. The exiles sailed to Asia, and landed on the Milyan territory. Milyas was the ancient name of the country now inhabited by the Lycians: the Milyae of the present day were, in those times, called Solymi.

So long as Sarpedon reigned, his followers kept the name which they brought with them from Crete, and were called Termilae, as the Lycians still are by those who live in their neighbourhood. But after Lycus, the son of Pandion, banished from Athens by his brother Aegeus had found a refuge with Sarpedon in the country of these Termilae, they came, in course of time, to be called from him Lycians. Their customs are partly Cretan, partly Carian.

They have, however, one singular custom in which they differ from every other nation in the world. They take the mother's and not the father's name. Ask a Lycian who he is, and he answers by giving his own name, that of his mother, and so on in the female line. Moreover, if a free woman marry a man who is a slave, their children are full citizens; but if a free man marry a foreign woman, or live with a concubine, even though he be the first person in the State, the children forfeit all the rights of citizenship." -- Herodotus, Histories, 1.173


"As in the Bronze Age syllabary more certainly evolved for the recording of Luvian, the so-called Hittite hieroglyphs, the sound values of the Linear A signary seem designed to represent exactly the Luvian phonetic system having 3 long vowels, 3 short. And like its Anatolian counterpart, where certain of the syllabograms demonstrably match the sound of the first syllable in Luvian of the object that inspired its sign-shapes, the Linear A signary too would appear to have been built on the acrophonic principle.

So far, some three dozen matches between Ventris' Linear B sound values (slightly altered to reflect the Greek 5-vowel system) and the Luvian name for the object often still identifiable in the Linear A sign--e.g., AB40/L28: WI as in WIdula-, the Hittite (and likely Luvian) term for 'chariot box'."
- From Brown's Minos article, not available online.

[circumspice]

Here you go, spice -

http://classics.unc.edu/people-3/faculty/edwin-brown/

http://www.people.ku.edu/~jyounger/LinearA/biblio.html

see also:
https://books.google.com/books?id=a_B8V ... 0A&f=false

As far as the entry of the Hittites into Anatolia,
this text may have some bearing on the problem:

http://abob.libs.uga.edu/bobk/ccc/cc012099.html

Naturally, everyone would like to get at the inscriptions at Carcamesh,
from T.E. Lawrence on.

The relationship between Hurrian and Luwian is of interest.

"The Lycians are in good truth anciently from Crete; which island, in former days, was wholly peopled with barbarians [non Greek speakers].

A quarrel arising there between the two sons of Europa, Sarpedon and Minos, as to which of them should be king, Minos, whose party prevailed, drove Sarpedon and his followers into banishment. The exiles sailed to Asia, and landed on the Milyan territory. Milyas was the ancient name of the country now inhabited by the Lycians: the Milyae of the present day were, in those times, called Solymi.

So long as Sarpedon reigned, his followers kept the name which they brought with them from Crete, and were called Termilae, as the Lycians still are by those who live in their neighbourhood. But after Lycus, the son of Pandion, banished from Athens by his brother Aegeus had found a refuge with Sarpedon in the country of these Termilae, they came, in course of time, to be called from him Lycians. Their customs are partly Cretan, partly Carian.

They have, however, one singular custom in which they differ from every other nation in the world. They take the mother's and not the father's name. Ask a Lycian who he is, and he answers by giving his own name, that of his mother, and so on in the female line. Moreover, if a free woman marry a man who is a slave, their children are full citizens; but if a free man marry a foreign woman, or live with a concubine, even though he be the first person in the State, the children forfeit all the rights of citizenship." -- Herodotus, Histories, 1.173


"As in the Bronze Age syllabary more certainly evolved for the recording of Luvian, the so-called Hittite hieroglyphs, the sound values of the Linear A signary seem designed to represent exactly the Luvian phonetic system having 3 long vowels, 3 short. And like its Anatolian counterpart, where certain of the syllabograms demonstrably match the sound of the first syllable in Luvian of the object that inspired its sign-shapes, the Linear A signary too would appear to have been built on the acrophonic principle.

So far, some three dozen matches between Ventris' Linear B sound values (slightly altered to reflect the Greek 5-vowel system) and the Luvian name for the object often still identifiable in the Linear A sign--e.g., AB40/L28: WI as in WIdula-, the Hittite (and likely Luvian) term for 'chariot box'."
- From Brown's Minos article, not available online.

None of your links show any info on the purported 'cracking' of Linear A.

You also seem uninformed about who made it possible to decipher Linear B... Ventris stood on the shoulders of the scholar who did the real work of deciphering Linear B. He only completed her work after her untimely death at age 43. Look up Alice Kober. You might learn something.

[E.P. Grondine]

None of your links show any info on the purported 'cracking' of Linear A.

You also seem uninformed about who made it possible to decipher Linear B... Ventris stood on the shoulders of the scholar who did the real work of deciphering Linear B. He only completed her work after her untimely death at age 43. Look up Alice Kober. You might learn something.

I am familiar with Kober's work and her untimely death, min.
Her spotting of declensions made Ventris's analysis possible,
in which he used G type bomber navigation maths on the sign counts.
This analysis led to his identification of consonant and vowel ratios
See Ventris's Worknotes for more detail.

For Chadwick's work on Japanese at Bletchley, see:
https://en.wikipedia.org/wiki/John_Chadwick
[He never really spoke about it much, as near as I know.]

Clearly you are not familiar with either Hittite Hieroglyphic nor Cretan Hieroglyphic.

"As in the Bronze Age syllabary more certainly evolved for the recording of Luvian, the so-called Hittite hieroglyphs, the sound values of the Linear A signary seem designed to represent exactly the Luvian phonetic system having 3 long vowels, 3 short. And like its Anatolian counterpart, where certain of the syllabograms demonstrably match the sound of the first syllable in Luvian of the object that inspired its sign-shapes, the Linear A signary too would appear to have been built on the acrophonic principle.

So far, some three dozen matches between Ventris' Linear B sound values (slightly altered to reflect the Greek 5-vowel system) and the Luvian name for the object often still identifiable in the Linear A sign--e.g., AB40/L28: WI as in WIdula-, the Hittite (and likely Luvian) term for 'chariot box'."
- From Brown's Minos article, not available online.[[/quote]

It's too bad Brown's Minos article is not available online, otherwise you would not make such a stupid comment, min.

Brown's sign identifications are pretty well bullet proof, and extensive.
I believe credit should be given when credit is due, and in my opinion Brown broke Linear A wide open.

The only question which may be open is whether the Lycian aspirates were attached to consonants or to the vowels,
in other words is the sign system a 3 vowel system or a 6 vowel system, with the evidence favoring 6 vowel,
until a detailed computer analysis is undertaken.

One wonders how Brown has progressed from when his Minos' paper was published.
I would hope that Brown and Packard could spend a little time together...

[kbs2244]

"in which he used G type bomber navigation maths on the sign counts."
And, once again, military financed research proves it's real worth.

[circumspice]

None of your links show any info on the purported 'cracking' of Linear A.

You also seem uninformed about who made it possible to decipher Linear B... Ventris stood on the shoulders of the scholar who did the real work of deciphering Linear B. He only completed her work after her untimely death at age 43. Look up Alice Kober. You might learn something.

I am familiar with Kober's work and her untimely death, min.
Her spotting of declensions made Ventris's analysis possible,
in which he used G type bomber navigation maths on the sign counts.
This analysis led to his identification of consonant and vowel ratios
See Ventris's Worknotes for more detail.

For Chadwick's work on Japanese at Bletchley, see:
https://en.wikipedia.org/wiki/John_Chadwick
[He never really spoke about it much, as near as I know.]

Clearly you are not familiar with either Hittite Hieroglyphic nor Cretan Hieroglyphic.

"As in the Bronze Age syllabary more certainly evolved for the recording of Luvian, the so-called Hittite hieroglyphs, the sound values of the Linear A signary seem designed to represent exactly the Luvian phonetic system having 3 long vowels, 3 short. And like its Anatolian counterpart, where certain of the syllabograms demonstrably match the sound of the first syllable in Luvian of the object that inspired its sign-shapes, the Linear A signary too would appear to have been built on the acrophonic principle.

So far, some three dozen matches between Ventris' Linear B sound values (slightly altered to reflect the Greek 5-vowel system) and the Luvian name for the object often still identifiable in the Linear A sign--e.g., AB40/L28: WI as in WIdula-, the Hittite (and likely Luvian) term for 'chariot box'."
- From Brown's Minos article, not available online.[

It's too bad Brown's Minos article is not available online, otherwise you would not make such a stupid comment, min.

Brown's sign identifications are pretty well bullet proof, and extensive.
I believe credit should be given when credit is due, and in my opinion Brown broke Linear A wide open.

The only question which may be open is whether the Lycian aspirates were attached to consonants or to the vowels,
in other words is the sign system a 3 vowel system or a 6 vowel system, with the evidence favoring 6 vowel,
until a detailed computer analysis is undertaken.

One wonders how Brown has progressed from when his Minos' paper was published.
I would hope that Brown and Packard could spend a little time together...[/quote]


I'd like to point out something first...

I am not min... You have addressed me as such twice in your latest post. Circumspice... My forum nick is Circumspice.

Secondly, you still haven't posted any links that state unequivocally that Linear A has been deciphered. Any work that proves that Linear A has been deciphered would have been announced to the world with the greatest of fanfare & would have been posted in all the major journals & reported upon by every major news service, worldwide. It would be a triumph of huge proportions to have finally deciphered Linear A...

You haven't provided any links showing any such news.

[Minimalist]

I am familiar with Kober's work and her untimely death, min.

Who are you talking to?

[E.P. Grondine]

My apologies to both of you.
I've been a little distracted.

Assured as he is that combinatorial methods (adopted by all serious students of Linear A) form the sole reliable means for arriving at a full understanding of the relevant texts, Brown is at work on a means of testing the Luvian affinities of the language that these documents encode.

Brown may be waiting for more texts, probably from Thera, before he can complete a combinatorial analysis reliable to the degree he desires.

As far as to why no acknowledgement of Brown's break has been publicly made by those interested in the problem, ask them.
I have no idea at all.

As you pointed out, spice, Kober made the break [spotting the declensions],
and then Ventris and Chadwick exploited it using "combinatorial methods".

[circumspice]

My apologies to both of you.
I've been a little distracted.

Assured as he is that combinatorial methods (adopted by all serious students of Linear A) form the sole reliable means for arriving at a full understanding of the relevant texts, Brown is at work on a means of testing the Luvian affinities of the language that these documents encode.

Brown may be waiting for more texts, probably from Thera, before he can complete a combinatorial analysis reliable to the degree he desires.

As far as to why no acknowledgement of Brown's break has been publicly made by those interested in the problem, ask them.
I have no idea at all.

As you pointed out, spice, Kober made the break [spotting the declensions],
and then Ventris and Chadwick exploited it using "combinatorial methods".

In other words: Linear A has not been deciphered. It is still a work in progress.