Cybernetica Mesopotamica

Graphemics

Giorgio Buccellati – March 2012

home

aims

history
archaeology

Grammar Urkesh
Terqa
graphemics
Ebla
Terqa
OB letters
MA laws
programs
morphemics

Ebla pers. names
OB royal letters

Back to Ebla personal names

Aims and procedures
Graphemic analysis
Triadic concordance
Graphemic indices
The categorization system
Current corpora

Back to top

Aims and procedures

     Linguistic analysis of a "dead" language is based initially on the "decipherment" of the script. But the two notions "dead" and "decipherment" need to be placed in the proper light.
     "Dead" would be a very improper term if meant to describe the nature of the organism to which it refers. Languages were very much "live" organisms, and in our current context, this means that the script conveys the spoken sounds, even if we can no longer hear them.
     "Decipherment" means, accordingly, recovering the silent sounds. Specifically, decipherment is not establishing a simple one to one correspondence between graphic signs and approximate values. It is rather the identification of very specific correspondences with the minimal distinctive units of the language, the phonemes. This is the basis for my definition of graphemics.
Back to top

Graphemic analysis

     Assyriological practice developed from the beginning a highly accurate approach to graphemics. The possibility of applying digital analysis has now refined greatly our methodology, because correlations can be established much more precisely among an immeasurably larger universe of data. It is with a view towards this goal that I developed new approaches to graphemic categorization, one of the interesting tools being the concept of triadic indices.
Back to top

Triadic concordance

     Using mainframe computers in the 70es, I pioneered the notion of a triadic concordance, which is still heuristically significant, besides being very useful philologically. Even though I published a full account of it in my publications on graphemic, this approach has not been otherwise applied in the field.
     A triadic concordance gives truncated text citations, sorted according to the shape of the sign, rather than its value. Here is an example from Ebla:

sign shape

context (sign values in red)reference

A DA NE

TUG2 SA6 DAR / p_ti-la-*NI / g_a-da-bi2-gu2-=KI / 1 AKTUM(A.SU)-=TUG2 /$A1#1 v.11,14
TUG2 SA6 DAR / p_ti-la-*NI / g_a-da-bi2-gu2-=KI / 1 AKTUM(A.SU)-=TUG2 /$A1#1 v.11,14
al6 {ARET1 p.318} / p_đu-D_ma-lik / g_a-da-bi2-ig-=KI / p_du-ur-*NI / g_'a3$A1#8 v.15,8
X)] {ARET1 p.318} / p_đa-[X-]bu3 / g_a-da-bi2-ig-=KI / AL6-TUĐ / g_ga-la$A1#8 v.15,16
TUG2 SA6 DAR / p_ti-la-*NI / g_a-da-bi2-ig-=KI / in / GI6-SA2 /$A1#10 r.10,8
IB2-:2-=TUG2 DAR / p_na-sa / g_a-da-bi2-gu2-=KI / 1 'a3-da-um-=TUG2$A1#15 v.4,11
TUG2 SA6 DAR / p_ti-la-*NI / g_a-da-bi2-gu2-=KI / 1 AKTUM(A.SU)-=TUG2 1$A1#15 v.6,6
5 *"KIN" SIKI / p_ti-la-*NI / g_a-da-bi2-ig-=KI / 2 *"KIN" SIKI / 2$A1#15 v.9,8
ma-bu14(KA)-ru12 / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 GIR2 MAR-TU ba$A1#17 r.4,5
ma-*KA / ln3 / ĐU-DU8 / g_a-da-bi2-ig-=KI / ĐU-DU8 / p_a-'a3$A1#17 v.1,13
ma-lik / wa / {p_*NI}-*LUM / g_a-da-bi2-ig-=KI / 1 AKTUM(A.SU)-=TUG2 /$A1#17 v.2,5
AKTUM(A.SU)-=TUG2 / p_ma-u9-um / g_a-da-bi2-ig-=KI / E3(UD.DU) / NIG2-KASKAL$A1#17 v.2,8
*DU / NIG2-KAS4 / si-in / g_a-da-bi2-ig-=KI / 1 AKTUM<(A.SU)-=TUG2> 1$A2#14 r.9,7
DU@s) / p_ha-{ra-il} / […] / g_a-da-bi2-ig-=KI / 1 gu2-li-lum A$A3#3 r.6,1'
/ g_SA-ZA0-=KI / DAB5-LU2 / g_a-da-bi2-ig-=KI / 10-2 GU-DUL3-=TUG2$A3#59 r.4,4'
NIG2-LA2-GABA / p_ti-la-*NI / g_a-da-bi2-ig-=KI / EN / 2 'a3-da$A3#61 r.3,8'
GI / 1 DIB p_bur-ha-ađ2 / g_a-da-bi2-ig-=KI / 10 *GIN2-*DILMUN(NI.TUK) BAR6$A3#81 r.1,4'
IB2-:3 DAR TUG2v?.6,2' p_mi-*NEv?.6,3' g_a-da-bi2-ig-=KIv?.6,4' 1 'a3-da-um-=TUG2$A3#134 v.6,3'
BU3-BAR6:KU3!! / p_kir-su-ud / g_a-da-bi2-ig-=KI / p_hi-su-ud / p_tal$A3#144 r.2,7''
mu / […] / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 GU-DUL3-=TUG2 1$A3#193 r.5,2
da-ru12-=KI / p_đa-nu-bu3 / g_a-da-bi2-gu2-=KI / p_*KU-D_ma-lik / g_ag$A3#244 r.6,3'
/ 3 NIN(SAL.TUG2)-NI / UGULA / g_a-da-bi2-ig-=KI / 1 IB2-:3-=TUG2 BAR6$A3#322 r.3,8'
/ [(X)] SIKI […] / TUG2-TUG2 / g_a-da-bi2-ig-=KI / […]$A3#382 / MAĐKIM(PA.DU@s)$A3#381 r.2,2'
[GI]BIL {ARET3 p.407} / [X] X […] / g_a-da-bi2-ig-=KI / 1 TUG2-*NI.*NI 1$A3#465 r.8,1
3 DAR TUG2 / p_ig-du-ru12 / p_a-da-bi2-zu2 / p_PUZUR4(KA×GAN2@t.ĐA)-ra-ha-al6 /$A3#468 r.2,6'
/ […]$A3#482 / p_iđ-i3-lum / g_a-da-bi2-ig-=KI / 1 'a3-da-um-=TUG2$A3#482 r.1,2'
d]a?-ga[-*L]UM / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 *SAL-=TUG2 / p_ib$A3#502 r.2,3'
DAR SA6 TUG2 / p_iđ-i3-lum / g_a-da-bi2-ig-=KI / […]$A3#522 / 2 *SAL$A3#521 r.3,4'
TU ZU2 *LAGAB@x*HAL-A / p_i3-gi / g_a-da-bi2-ig-=KI / 9 *GIN2-*DILMUN(NI.TUK) BAR6$A3#635 v.4,7'
GU-DUL3-=TUG2 […]$A3#744 / ce3 / g_a-[da]-bi2-ig-=KI / 5 SIKI ZI-RI2 /$A3#744 v.2,1'
*NE-li / p_PUZUR4(KA×GAN2@t.ĐA)-ra-BE / g_a-da-bi2-ig-=KI / ĐU-DU8 / g_gu-ra$A3#938 r.5,6'
[X X] {X} / […] / […] / g_a-da-bi2-gu2-=KI / 2 *SAL-=TUG2 2 IB2$A4#2 r.2,7
DAR / p_PUZUR4(KA×GAN2@t.ĐA)-ra-a-gu2 / g_a-da-bi2-gu2-=KI / in / g_da-ra-um$A4#4 v.5,13
5 *"KIN" SIKI / p_ti-la-*NI / g_a-da-bi2-ig-=KI / 1 *SAL-=TUG2 1 IB2$A4#4 v.6,4
I3-NA-SUM / p_ti-la-*NI / g_a-da-bi2-gu2-=KI / ap / 1 gu2-li$A4#7 v.8,19
đa-la-ab / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 gu-zi-dum-=TUG2$A4#10 r.4,12
*PI-ar(IGI.RI) / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 *SAL-=TUG2 1 IB2$A4#10 r.6,3
/ p_*EN-ha / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 *SAL-=TUG2 1 IB2$A4#10 v.2,3
/ p_*NI-la / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 *SAL-=TUG2 1 IB2$A4#10 v.3,6
/ p_ri2-zu2 / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 *SAL-=TUG2 / p_al6$A4#11 r.11,12
ZA0(LAK384)-=KI / ĐU-DU8 / in / g_a-da-bi2-ig-=KI / 1 GU-MUG-=TUG2-:1$A4#11 v.1,5
ra-an / ĐU-DU8 / in / g_a-da-bi2-ig-=KI / 1 GADA-=TUG2 *IGI-NITA$A4#11 v.5,5
en-na-*NI / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 *SAL-=TUG2 1 IB2$A4#12 r.8,3
TUG2 / g_'a3-du-=KI / in / g_a-da-bi2-ig-=KI / ĐU-BA4-TI / in$A4#13 r.8,10
EN / *DU-*DU / si-in / g_a-da-bi2-ig-=KI / 1 AKTUM(A.SU)-=TUG2 1$A4#13 r.8,16
ba-ra-i / ĐU-KEĐDA / UGULA g_a-da-bi2-ig-=KI / 3 GIĐ=-*"KIN" SIKI AMA$A4#13 v.13,13
TUG2 SA6 DAR / p_ti-la-*NI / g_a-da-bi2-gu2-=KI / 1 IB2-:4-=TUG2 DAR$A4#14 v.4,11
TUG2 SA6 DAR / p_ti-la-*NI / g_a-da-bi2-ig-=KI / in U4 / […] /$A4#22 v.2,15
u3}(IGI.DIB)-gi / LU2-KAR(TE.A) / g_a-da-bi2-ig-=KI / 1 gu2-li-l[um A]$A7#107 r.2,2
ĐE-HA0(SAG@xHA)-MUL / IB2-:3-=TUG2 / g_a-da-bi2-gu2-=KI / ITI n_za-'a3-dum /$A7#134 r.2,3
đu-a-=KI / wa / KI / g_a-da-bi2-gu2-=KI / p_a-ma-D_ma-lik /$A7#154 r.6,1
/ 5 mi-at GANA2-KEĐDA-KI / g_a-da-bi2-gu2<-=KI> / in KI-SUR / g_ur$A7#156 r.2,15
KI / 1 li 1 mi KI / g_a-da-bi2-gu2-=KI / g_đe3-ra-du-=KI /$A7#156 r.4,8
TUG2 TUG2-MU / p_ti-la-*NI!! / g_a-da-bi2-g[u2]-=KI / 1 GADA-=TUG2 TUG2-M[U]$A8#522 r.6,8
TUG2 TUG2-MU / p_ti-la-*NI!! / g_a-da-bi2-g[u2]-=KI / 1 GADA-=TUG2 TUG2-M[U]$A8#522 r.6,8
TUG2-SA6-DAR / p_ti-la-*NI / g_a-da-bi2-gu2-=KI / p_*UR-*NE / g_i3-mar$A8#522 r.9,15
da / LU2 p_zi-kir-M_da-mu / g_a-da-bi2-gu2-=KI / 1 'a3-da-um-=TUG2$A8#542 r.5,13
[IB2-:3-=TUG2]-S[A6-DAR] / cb2 / g_a-da-bi2-gu2-=KI / 2 'a3-da-um-=TUG2$A8#542 r.8,1
1-1/2 gu2-bar / ENGAR-KINDA2 / g_a-da-bi2-gu2-=KI / n_E2-DURU5-=KI-:2 /$A9#66 r.4,8
total59
*GIĐ.*NI-=KI / g_a-lu-=KI / g_a-da-*NE-du-=KI / g_a-ba-ha-{an}-{đu}$A7#156 r.5,6
4 UDU / 60 ĐE gu2-bar / g_a-da-*NE-du-=KI / 10 *GIN2-*DILMUN(NI.TUK) BAR6$A3#682 r.1,5

     One can easily evince the function of the concordance: out of 61 examples in which the triad A DA NE occurs (where the writing A DA NE stands for the sign shapes), two alternative readings are possible for the triad: a-da-bi2 in 59 occurrences and a-da-*NE in the remaining two. The graphemic import of this display is that the polyphony of the cuneiform signs is considerably limited when the signs are see in a triadic context.
     In fact, the polyphony is practically non-existent within any given corpus, because the incidence where the possibility of alternative readings occurs is in turn extremely limited.
     As I designed it, the original triadic index program was written by John Settles for mainframe computers, and with it we produced concordances for all the corpora that were being produced at the time.
     In 2012, with the support of a grant from the Council on Research of the Academic Senate, I was able to revisit my old graphemic project. Dr. Massimo Maiocchi graciously agreed to write the program for current platforms

Back to top

Graphemic indices

     
Back to top

The categorization system

     
Back to top

Current corpora

     Ebla personal names
     OB Letters from the Royal Chancery of Babylon
Back to top