Academiuri-bolo pmd



Yüklə 59,26 Kb.

tarix22.07.2018
ölçüsü59,26 Kb.


© 2007 Bull. Georg. Natl. Acad. Sci.

saqarTvelos   mecnierebaTa   erovnuli   akademiis    moambe

,,,,,   


   

   


   

     17


 17

 17


 17

 175


55

55, 




,     

   


   

   


    ½

½

½



½

½ 1


11

11, 




,     

   


   

   


    200

2 0 0


2 0 0

2 0 0


2 0 07

77

77



BULLETIN  OF   THE  GEORGIAN  NATIONAL   ACADEMY  OF  SCIENCES,  175,  

½

½



½

½

½ 1



11

11, 200


, 200

, 200


, 200

, 2007


77

77

 



All this is true for the palimpsest codex Vind. 2

housed in the Austrian National Library in Vienna, which

is an outstanding monument of Old Georgian indeed.

Most of what has been known so far about the Vienna

palimpsest

1

 was worked out in the early 1970’s by John



Neville Birdsall who published the first editions of three

of the underlying manuscripts, one containing fragments

of the Gospels, one, fragments of the (apocryphal) Old

Testament book of Esdras Zorobabel, and one, a major

part (eight chapters) of the Protevangelium of James, all

of these texts pertaining to the “Khanmeti” period (Vth–

VIIth cc.)

2

. As Birdsall observed, the palimpsest com-



prises several other original manuscripts, representing

at least 12 different hands of different age; among them

there are at least three further Khanmeti fragments and

seven other manuscripts written in Old Georgian majus-

cules, the so-called Asomtavruli  script

3

. Only two of



these fragments have since been published; they con-

tain parts of the treatise on Measures and Weights by

Epiphanius of Cyprus

4

 and a Khanmeti version of the



legend of St. Christina

5

.



The plan to prepare a complete edition of the oldest

layer of the Vienna palimpsest, i.e., the Khanmeti and

Post-Khanmeti texts written in Asomtavruli script, was

developed in 1997 when the late Zurab Sarjveladze so-



History & Philology

The Application of Multispectral Imaging in the Study of

Caucasian Palimpsests

Jost Gippert

Frankfurt University, Germany (Chair, Institute of Comparative Linguistics)

ABSTRACT.

 For the three ancient written languages of the Caucasus, Armenian, Georgian, and “Albanian”, the

first centuries of their literacy are still quite obscure as only very few manuscript remains from this period have

come down to us. As a matter of fact, most of the manuscripts of the Vth–VIIIth cc. were not preserved in their

primary state but as palimpsests, i.e., reused and overwritten after a more or less total erasure, and special tech-

niques are required to uncover and restore their original content.

 © 2007 Bull. Georg. Natl. Acad. Sci.

Key words: palimpsest, manuscript, Caucasus.

1

 The palimpsest originates from the Monastery of the



Holy Cross in Jerusalem; for details cf. the forthcoming edi-

tion (cf. note 12).

2

 Cf. J.N. Birdsall, A second Georgian recension of the



Protoevangelium Jacobi, Le Muséon 83, 1970, pp. 49–72;

Khanmeti Fragments of the Synoptic Gospels from Ms. Vind.

Georg. 2, Oriens Christianus 55, 1971, pp. 62–89; Palimpsest

fragments of a Khanmeti Georgian version of I. Esdras, Le

Muséon 85, 1972, pp. 97–105.

3

 Cf. J.N. Birdsall, A Georgian Palimpsest in Vienna, Oriens



Christianus 53, 1969, 108-112; MS Vindob. Georg. 2: a

progress report, Oriens Christianus 58, 1974, pp. 39–44; Jost

Gippert, Die georgische Palimpsesthandschrift Codex

Vindobonensis georgicus 2, Biblos. Beiträge zu Buch,

Bibliothek und Schrift 1–2, 2003, 31–46.

4

 Cf. Michel Jean van Esbroeck (ed.), Les versions



géorgiens d’Épiphane de Chypre, Traité des poids et des

mesures, Lovanii 1984 [Édition / Traduction] (Corpus scrip-

torum christianorum orientalium, 460–461 / Scriptores iberici,

19–20), pp. 5–9 / 23–25.

5

 Cf. Lamara KaJaia, C.m. Kris.tinas bamebis xanmeti teksti;



Gelatis mecnierebata a>ademiis žurnali / Journal of Gelati Acad-

emy of Sciences 3/2006, pp. 34–59.




The Application of Multispectral Imaging in the Study of Caucasian Palimpsests

169

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

journed as a visiting professor at the University of Frank-

furt and proposed to compile, as an extension of the

TITUS text data base

6

, an electronic corpus of Khanmeti



texts including the fragments contained in the Vienna

palimpsest. To achieve this, a set of ultraviolet nega-

tives of the corresponding pages was procured from the

Austrian National Library

7

, on the basis of a digitization



of these images, the three underlying manuscripts that

had been dealt with by Birdsall were re-read in total, and

one further manuscript (containing the legends of

Cyprianus and Christina), about halfway through. As a

result of the common work undertaken in 1997–8, a first

digital edition comprising the fragments named above

was published on the TITUS server in February, 1998

8

.



Unfortunately, Zurab Sarjveladze did not live to par-

ticipate in the continuation of this work, which was initi-

ated shortly after his untimely death in December, 2002.

In the course of a project devoted to the manuscript re-

mains of the Caucasian “Albanians”

9

, the idea was born



to apply the newly developed technology of multispec-

tral imaging in the analysis of palimpsests of Caucasian

provenance. This idea has been realised since 2004 when,

thanks to a generous grant by the Volkswagen Founda-

tion, a “MuSIS” camera system

10

 could first be used to



cope with the bad readability of the Vienna codex and the

“Albanian” palimpsests of Mt. Sinai

11

. As the application



of this system meant a big step forward in the study of

these manuscripts, it may be appropriate to shortly de-

scribe its function here. The results of the investigations

facilitated by the new technique will be published in the

editions of the palimpsests which are to appear in 2007

12

.



6

 The TITUS data base of electronic texts (University of

Frankfurt) comprises a large collection of Old and Middle

Georgian texts; cf. http://titus.uni-frankfurt.de/texte/

texte2.htm#georgant.

7

 The same specimens of “excellent ultra-violet photogra-



phy” were also used by John Neville Birdsall in his work on the

Vienna palimpsest; cf. his articles Ms. Vind. Georg. 2... (cf. n. 3),

p. 39 and, before, A Georgian Palimpsest... (cf. n. 3), p. 108.

8

 Cf. http://titus.fkidg1.uni-frankfurt.de/texte/etcc/cauc/



ageo/xanmeti/vienna/vienn.htm; the edition has steadily been

improved and extended since.

9

 Cf. http://armazi.uni-frankfurt.de/armaz3.htm (part of



the project “ARMAZI: Fundamentals of an Electronic Docu-

mentation of Caucasian Languages and Cultures”).

10

 The “MultiSpectral Imaging System” is a product of



Forth Photonics, Greece; cf. http://musis.forth-photonics.gr.

11

 The project “Neue Wege zur wissenschaftlichen



Bearbeitung von Palimpsesthandschriften kaukasischer

Provenienz” has been funded by the Volkswagen Foundation since

12.12.2003; it is jointly run by Zaza Aleksi3e, Jost Gippert,

Jean-Pierre Mahé, Wolfgang Schulze, and Manana Tandaschwili.

12

 The editions will be published in the series Monumenta



Palaeographica Medii Aevi, Series Iberica et Caucasica, as vol-

umes I and II; vol. I containing the edition of the Vienna codex

is currently in the press.

The main principle of multispectral imaging consists

in the fact that the resonance of any object differs with

respect to different wavelengths of light, depending on

the consistence of its colour; this is true both for the

visible and the invisible (ultraviolet and infrared) parts

of the spectrum. By applying a photographing method

that is restricted to a certain range of the spectrum, a

specific resonance may be retained or suppressed. This

principle can easily be demonstrated using a bicoloured

manuscript such as the Church Slavonic menaion manu-

script displayed in Fig. 1.

13

 Of the two colours repre-



sented in it, red and dark-brown, the first one will ap-

pear extremely bright when photographed within the red

range of the spectrum, i.e., 620–750 nm, and it will be

more and more dark the more we go down the spectral

range; cp. Figs. 2–5 showing this effect.

In the case of palimpsest manuscripts, the effect

that can be gained from this predisposition depends on

three factors: The colour resonance of the upper script,

that of the lower script, and that of the background, i.e.,

the parchment itself. One might expect that the the first

two are the most decisive factors in this constellation,

Fig. 1. Church Slavonic menaion (bicoloured)

13

 The manuscript (allegedly of the early XVIIth cen-



tury) was donated to the University of Frankfurt in 2003 so

that it could be used for testing the application of the multi-

spectral imaging system.



170

Jost Gippert

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

Fig. 5. Same, at 680 nm (red)

Fig. 4. Same, at 620 nm (orange)

Fig. 3. Same, at 540 nm (green)

Fig. 2. Same, at 440 nm (violet)




The Application of Multispectral Imaging in the Study of Caucasian Palimpsests

171

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

as in many cases, it might be desirable to “enhance” the

lower script in contrast with the upper script covering it.

This, however, is not always possible in parchment pal-

impsests of Caucasian (or, rather, Near Eastern) prov-

enance as both the lower and the upper scripts were

usually written with the same type of inks, which results

in similar resonances. Thus the application of multispec-

tral imaging must concentrate upon two aims:

a) increasing the contrast between the (erased) lower

script and the background. This aim is quite the same as

that of applying ultraviolet photographing to the pal-

impsests as it is mostly in the violet and ultraviolet range

(370–440 nm) that the contrast between a washed or

wiped out ink and a parchment skin bearing it can be

reinforced. The reason is that the inks used in the Near

Eastern Middle Ages had but very few “blueish” por-

tions and appear darker in that range while the parch-

ment appears rather bright, at least if it has not been

discoloured or stained in the course of history. In many

cases, the results to be achieved applying a diversified

spectral imaging will be better than those to be gained

by mere ultraviolet photography as the range of optimal

contrast may vary from leaf to leaf, often extending into

the blue range (450–495 nm); cp. Figs. 6 and 7 which

contrast an ultraviolet photograph and an image taken

at 460 nm of fol. 10r of the Vienna palimpsest, and Fig. 8

showing a well readable extract of the latter. The method

of preparing a large set of digital images covering the

spectrum from the ultraviolet range up to the red range

to be analysed on the computer screen has been ap-

plied extensively in the present project, and in the case

of the Vienna palimpsest, it has led to an increase of the

reading rates upto about 95% (while ca. 80% are achiev-

able using ultraviolet photographs). It goes without say-

ing that this method always remains experimental and

its results depend to a high degree on the light source

used and other conditions;

b) exploiting the difference of several images show-

ing the same object to reduce the preponderance of the

upper script. This aim is mostly connected with the task

Fig. 6. Fol. 10r (ultraviolet photograph)



172

Jost Gippert

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

14

 The recalculation procedures here discribed are inte-



grated in the software provided with the MuSIS system.

Fig. 7. Fol. 10r (spectral image at 460 nm)

of reproducing the contents of the palimpsest in printed

form, and it is the method of recalculating the bright-

ness information of two different spectral images that

can be applied here. To give an example, Figs. 9 to 16

below show another page of the Vienna palimpsest (fol.

28v, containing Mk. 10,29–30) in eight different appear-

ances: first in a plain colour image and an ultraviolet

photograph, then in two spectral images (at 440 and 580

nm), and lastly in four different “recalculations” of the

latter pair of images.

14

 It is clear from these specimens



that neither the addition nor the multiplication of the

data of two spectral images yield noteworthy results

with respect to a better distinction of the upper and

lower scripts. In contrast to this, a subtraction of the

brightness data of the two images brings about a clear

improvement as it effects the inversion of the lower

script. The best results by far are achieved applying a

division of the data as in Fig. 16. Here, the upper script

is nearly levelled to the grey value of the (parchment)

background, and the lower script alone stands out. The

reason for this can be seen in the fact that the ink used

in the upper script is near to black, which means that it

has quite the same (low) level of brightness throughout

the visible spectrum; the same holds true for the parch-

ment background if it is whitish and not stained or

discoloured in a peculiar way, as in the given example.

Different from this, the lower script, which appears yel-

low or brownish to the human eye, will have extremely

divergent brightness values in the blue and yellow range,

resulting in a high quotient when divided. It must be

admitted that this method, too, has narrow limits. Its

efficiency depends not only on the grade of visibility of

the lower script but also on the state and tinct of the

parchment, and many of the pages of the Vienna palimp-

sest illustrate the problems resulting from this. Fol. 28

(and some other pages of the underlying Gospel ms.)

are rather exceptional in the quality of the reproductions

that can be achieved using the “recalculation” method.




The Application of Multispectral Imaging in the Study of Caucasian Palimpsests

173

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

It should be added here that for those parts of the

Vienna palimpsest the lower text of which is written in

small Nuskhuri characters

15

, none of the methods de-



scribed above can be applied successfully. Whenever

the lower script is covered by the upper text in its en-

tirety, optical means will not be apt to make it reappear.

For the lower texts in Asomtavruli script, the application

of multispectral imaging has generally yielded good re-

sults, however.

In the case of the “Albanian” palimpsests from

Mt. Sinai

16

, the successful application of the multispec-



tral imaging method was further limited by two special

conditions. Firstly, very many leaves of the manuscripts

Fig. 8. Same, extract (enlarged)

were badly defaced by fire

17

, which resulted in a dark



stain especially of the margins; cf. Figs. 18–20 show-

ing a specimen of this type. Secondly, one of the origi-

nal manuscripts (containing parts of St. John’s Gos-

pels) had been erased in such a drastic way that on

many pages, hardly any remains of it have survived.

This is why the reading rate is much lower than with

the Vienna palimpsest, ending up with less than 50%

for the latter original.

To be sure, the results of the present project have

not been achieved by applying multispectral imaging

alone. The “normal” ultraviolet photographs that had

been prepared both for the Vienna codex

18

 and for the




174

Jost Gippert

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

Fig. 9. Fol. 28v (colour image)

Fig. 10. Same (ultraviolet photograph)




The Application of Multispectral Imaging in the Study of Caucasian Palimpsests

175

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

Fig. 12. Same (spectral image, 580 nm)

Fig. 11. Fol. 28v (spectral image, 440 nm)




176

Jost Gippert

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

Fig. 13. Same (values of Figs. 11 and 12 “added”)

Fig. 14. Same (values of Figs. 11 and 12 “multiplied”)




The Application of Multispectral Imaging in the Study of Caucasian Palimpsests

177

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

Fig. 15. Same (values of Figs. 11 and 12 “subtracted”)

Fig. 16. Same (values of Figs. 36 and 37 “divided”)




178

Jost Gippert

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007

colour images was prepared for both the Vienna codex

and the Albanian palimpsests in the course of the

present project. For the study of palimpsests, it is there-

fore highly advisable to use all three kinds of images

side by side.

Fig. 20. Same, extract (at 440 nm)

“Albanian” palimpsests before

19

 have been used



throughout the editorial work to verify suggestions and

support the readings. In some cases, even high resolu-

tion colour photographs

20

 have turned out to be indis-



pensable for the decipherment. A complete set of digital

Fig. 17. 8r of the Albanian palimpsest M55 (colour)

Fig. 18. Same (ultraviolet photograph)

Fig. 19. Same, extract (at 440 nm)

19

 The ultraviolet photographs were taken in situ in the



course of the ARMAZI project (cf. n. 9) and have been pub-

lished in digital form on the project server in http://armazi.uni-

frankfurt.de/armaz3.htm since 2002.

20

 A complete set of digital colour images was prepared



for both the Vienna codex and the Albanian palimpsests in the

course of the present project.




The Application of Multispectral Imaging in the Study of Caucasian Palimpsests

179

Bull. Georg. Natl. Acad. Sci. Vol. 175, No1, 2007



Document Outline

  • 168-172.pdf
    • 168-170.pdf
    • 171.pdf
    • 172.pdf
  • 175-dan.pdf
    • 175-178.pdf
      • 175.pdf
      • 176-179.pdf
    • 179.pdf



Dostları ilə paylaş:


Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©genderi.org 2019
rəhbərliyinə müraciət

    Ana səhifə