Let's look more closely at the tall letters:
cat bio-j-jsa-gut.wds \
| sed \
-e 's/^/_/g' \
-e 's/$/_/g' \
-e 's/cs/z/g' \
-e 's/ij/k/g' \
-e 's/ix/e/g' \
-e 's/is/r/g' \
-e 's/iiu/n/g' \
-e 's/y/i/g' \
-e 's/ci/a/g' \
-e 's/cg/8/g' \
| compare-contexts -lctx 0 -rctx 0 -colw 24 \
'[czlqjg]*lj[czlqjg]*' \
'[czlqjg]*qj[czlqjg]*' \
'[czlqjg]*lg[czlqjg]*' \
'[czlqjg]*qg[czlqjg]*'
571 0.36 lj 230 0.33 qj 10 0.62 lgccc 53 0.34 qgccc
376 0.24 ljcc 161 0.23 qjc 2 0.12 lg 50 0.32 qg
322 0.21 ljc 118 0.17 qjcc 2 0.12 clgccc 8 0.05 qgzcc
38 0.02 cccljc 28 0.04 qjccc 1 0.06 lgcc 8 0.05 qgcc
32 0.02 ljccc 27 0.04 cccqjc 1 0.06 cclg 4 0.03 qgzc
27 0.02 zccljc 18 0.03 ccccqjc ----- ---- ---- 4 0.03 cqgcc
26 0.02 zcccljc 14 0.02 zccqjc 16 1.00 TOT 4 0.03 cqgc
23 0.01 ccccljc 14 0.02 qjzcc 4 0.03 cccqgcc
18 0.01 ljzcc 11 0.02 zcccqjc 3 0.02 cccqg
17 0.01 zcclj 9 0.01 cqjcc 2 0.01 zccqgcc
14 0.01 ccclj 8 0.01 zcccqj 2 0.01 zcccqgc
10 0.01 zccljcc 7 0.01 zccqj 2 0.01 qgcccc
10 0.01 cccljcc 7 0.01 cqjc 2 0.01 cqg
9 0.01 cljcc 5 0.01 zccqjcc 2 0.01 ccqgzccc
9 0.01 cljc 5 0.01 qjzc 1 0.01 zqgcc
9 0.01 cccclj 5 0.01 ccqj 1 0.01 zccqgccc
6 0.00 zcccljcc 4 0.01 cccqj 1 0.01 zccqg
6 0.00 cclj 3 0.00 zcqj 1 0.01 ccqgccc
5 0.00 zccclj 3 0.00 qcqjc 1 0.01 cccqgccc
4 0.00 ljzc 3 0.00 ccccqj 1 0.01 ccccqgcc
4 0.00 ljcccc 2 0.00 zcccqjcc ----- ---- ----
4 0.00 ccccljcc 2 0.00 qjccz 154 1.00 TOT
2 0.00 zclj 2 0.00 qcqj
2 0.00 qcljcc 2 0.00 cccqjcc
2 0.00 ljczcc 2 0.00 ccccqjcc
2 0.00 ljczc 1 0.00 zcqjcc
2 0.00 ccljcc 1 0.00 zccccqjcc
2 0.00 ccljc 1 0.00 qjczc
1 0.00 zzcljcc 1 0.00 qjcz
1 0.00 zzcclj 1 0.00 qjcccc
1 0.00 zzcccljc 1 0.00 qcqjccc
1 0.00 zljcc 1 0.00 qcqjcc
1 0.00 zlj 1 0.00 cqjccz
1 0.00 zcljc 1 0.00 cqj
1 0.00 zccljccc 1 0.00 ccqjc
1 0.00 qlj ----- ---- ----
1 0.00 ljcz 700 1.00 TOT
1 0.00 ljccz
1 0.00 ljccccljc
1 0.00 clj
1 0.00 cccljccc
----- ---- ----
1565 1.00 TOT
These statistics confirm the identification of the \l/ and \q/
in gallows.
The questions to decide now are whether \Hcc/ and \zcc/ are letters
or composites \H/+\cc/ and \z/+\cc/.
cat bio-j-jsa-gut.wds \
| sed \
-e 's/^/_/g' \
-e 's/$/_/g' \
-e 's/[ql][jg]/H/' \
-e 's/cs/z/g' \
-e 's/ij/k/g' \
-e 's/ix/e/g' \
-e 's/is/r/g' \
-e 's/iiu/n/g' \
-e 's/y/i/g' \
-e 's/ci/a/g' \
-e 's/cg/8/g' \
> .bar
foreach f ( z H )
cat .bar \
| enum-contexts -vPAT='[^czH]'"${f}"'[^czH]' \
| sed -e 's/.$//g' \
| wfreq \
> .${f}.L
cat .bar \
| enum-contexts -vPAT='[^czH]'"${f}cc"'[^czH]' \
| sed -e 's/.$//g' \
| wfreq \
> .${f}cc.L
cat .bar \
| enum-contexts -vPAT='[^czH]'"cc"'[^czH]' \
| sed -e 's/^.//g' \
| wfreq \
> .cc.R
cat .bar \
| enum-contexts -vPAT='[^czH]'"${f}cc"'[^czH]' \
| sed -e 's/^.//g' \
| wfreq \
> .${f}cc.R
pr -m -s' ' -t -i' '1 -w 96 .${f}cc.L .${f}.L .${f}cc.R .cc.R \
| expand
end
326 0.74 _zcc 136 0.90 _z 296 0.67 zcc8 54 0.39 cc8
65 0.15 ezcc 6 0.04 ez 108 0.25 zcca 47 0.34 cca
22 0.05 8zcc 6 0.04 az 35 0.08 zcco 23 0.17 cco
14 0.03 azcc 2 0.01 oz ----- ---- ---- 11 0.08 cce
9 0.02 rzcc 1 0.01 qz 439 1.00 TOT 2 0.01 ccr
3 0.01 ozcc ----- ---- ---- 1 0.01 cc_
----- ---- ---- 151 1.00 TOT ----- ---- ----
439 1.00 TOT 138 1.00 TOT
380 0.76 oHcc 668 0.79 oH 319 0.64 Hcc8 54 0.39 cc8
61 0.12 eHcc 87 0.10 _H 169 0.34 Hcca 47 0.34 cca
33 0.07 _Hcc 68 0.08 eH 13 0.03 Hcco 23 0.17 cco
26 0.05 aHcc 18 0.02 aH ----- ---- ---- 11 0.08 cce
1 0.00 8Hcc 1 0.00 qH 501 1.00 TOT 2 0.01 ccr
----- ---- ---- ----- ---- ---- 1 0.01 cc_
501 1.00 TOT 842 1.00 TOT ----- ---- ----
138 1.00 TOT
From these numbers, it seems plausible that `Hcc' and `zcc' are composites.
Collecting again all [czH] patterns, splitting H and P:
cat bio-j-jsa-gut.wds \
| sed \
-e 's/^/_/g' \
-e 's/$/_/g' \
-e 's/[ql]j/H/' \
-e 's/[ql]g/P/' \
-e 's/cs/z/g' \
-e 's/ij/k/g' \
-e 's/ix/e/g' \
-e 's/is/r/g' \
-e 's/iiu/n/g' \
-e 's/y/i/g' \
-e 's/ci/a/g' \
-e 's/cg/8/g' \
| enum-contexts -vPAT='[czHP][czHP]*' -vCTX=0 \
| wfreq
795 0.20 H
493 0.13 Hcc
484 0.12 ccc
482 0.12 Hc
439 0.11 zcc
152 0.04 z
138 0.04 cc
87 0.02 zccc
70 0.02 zc
70 0.02 cccc
65 0.02 cccHc
63 0.02 Pccc
60 0.02 Hccc
52 0.01 P
41 0.01 zccHc
41 0.01 ccccHc
37 0.01 zcccHc
32 0.01 Hzcc
24 0.01 zccH
20 0.01 cHcc
19 0.00 cHc
18 0.00 cccH
15 0.00 zccHcc
13 0.00 zcccH
12 0.00 ccccH
12 0.00 cccHcc
11 0.00 cccz
11 0.00 ccH
9 0.00 c
9 0.00 Pcc
9 0.00 Hzc
8 0.00 zcccHcc
8 0.00 Pzcc
6 0.00 zzcc
6 0.00 ccccHcc
6 0.00 Hcccc
5 0.00 zcH
5 0.00 ccccz
4 0.00 zcccz
4 0.00 cccPcc
4 0.00 cPcc
4 0.00 cPc
4 0.00 cH
4 0.00 Pzc
3 0.00 cccP
3 0.00 ccHc
3 0.00 Hczc
3 0.00 Hccz
2 0.00 zcccPc
2 0.00 zccPcc
2 0.00 ccz
2 0.00 ccPzccc
2 0.00 ccHcc
2 0.00 cPccc
2 0.00 cP
2 0.00 Pcccc
2 0.00 Hczcc
2 0.00 Hcz
1 0.00 zzcccHc
1 0.00 zzccH
1 0.00 zzcHcc
1 0.00 zcz
1 0.00 zccz
1 0.00 zccccHcc
1 0.00 zcccc
1 0.00 zccPccc
1 0.00 zccP
1 0.00 zccHccc
1 0.00 zcHcc
1 0.00 zcHc
1 0.00 zPcc
1 0.00 zHcc
1 0.00 zH
1 0.00 ccccc
1 0.00 ccccPcc
1 0.00 cccPccc
1 0.00 cccHccc
1 0.00 ccPccc
1 0.00 ccP
1 0.00 cHccz
1 0.00 cHccc
----- ---- ----
3906 1.00 TOT
Based on the analysis above, it seems that, after colapsing \ci/,
\cy/, \cg/, \cs/, and the tall characters \[lq]j/ to `H' \[lq]g/ to
`P', we can parse most strings consisting of `c', `z', `H', and `P'
into the following "letters" (the frequencies are for isolated
occurences only):
freq letter code
---- -------- ----
795 H E
52 P P
152 z Z
138 cc M
70 zc R
482 Hc I
484 ccc O
439 zcc A
493 Hcc U
19 cHc X
4 cPc Y
20 cHcc K
4 cPcc G
Here is my best guess for the parsing of those strings:
freq string best parsing alt parsings
---- ---- ---------- ------------ --------------------
795 0.20 H (H)
493 0.13 Hcc (Hcc)
484 0.12 ccc (ccc)
482 0.12 Hc (Hc)
439 0.11 zcc (zcc)
152 0.04 z (z)
138 0.04 cc (cc)
87 0.02 zccc (z)(ccc) (zc)(cc)
70 0.02 zc (zc)
70 0.02 cccc (cc)(cc)
65 0.02 cccHc (ccc)(Hc) (cc)(cHc)
63 0.02 Pccc (P)(ccc)
60 0.02 Hccc (H)(ccc) (Hc)(cc)
52 0.01 P
41 0.01 zccHc (zcc)(Hc) (zc)(cHc)
41 0.01 ccccHc (ccc)(cHc) (cc)(cc)(Hc)
37 0.01 zcccHc (zcc)(cHc) (zc)(cc)(Hc)
32 0.01 Hzcc (H)(zcc) (H)(z)(cc)
24 0.01 zccH (zcc)(H)
20 0.01 cHcc (cHcc)
19 0.00 cHc (cHc)
18 0.00 cccH (ccc)(H)
15 0.00 zccHcc (zcc)(Hcc)
13 0.00 zcccH (z)(ccc)(H) (zc)(cc)(H)
12 0.00 ccccH (cc)(cc)(H)
12 0.00 cccHcc (ccc)(Hcc) (cc)(cHcc)
11 0.00 cccz (ccc)(z)
11 0.00 ccH (cc)(H)
9 0.00 c
9 0.00 Pcc (P)(cc)
9 0.00 Hzc (H)(zc)
8 0.00 zcccHcc (zcc)(cHcc) (zc)(cc)(Hcc), (z)(cc)(cHcc)
8 0.00 Pzcc (P)(zcc)
6 0.00 zzcc (z)(zcc)
6 0.00 ccccHcc (ccc)(cHcc) (cc)(cc)(Hcc)
6 0.00 Hcccc (Hcc)(cc) (Hc)(ccc)
5 0.00 zcH (zc)(H)
5 0.00 ccccz (cc)(cc)(z)
4 0.00 zcccz (zc)(cc)(z) (z)(ccc)(z)
4 0.00 cccPcc (ccc)(P)(cc) (cc)(cPcc)
4 0.00 cPcc (cPcc)
4 0.00 cPc (cPc)
4 0.00 cH
4 0.00 Pzc (P)(zc)
3 0.00 cccP (ccc)(P)
3 0.00 ccHc (cc)(Hc)
3 0.00 Hczc (Hc)(zc)
3 0.00 Hccz (H)(cc)(z)
2 0.00 zcccPc (zcc)(cPc)
2 0.00 zccPcc (zcc)(Pcc) (z)(cc)(P)(cc)
2 0.00 ccz (cc)(z)
2 0.00 ccPzccc (cc)(P)(zc)(cc) (cc)(P)(z)(ccc)
2 0.00 ccHcc (cc)(Hcc)
2 0.00 cPccc (cPc)(cc)
2 0.00 cP
2 0.00 Pcccc (Pc)(ccc) (P)(cc)(cc)
2 0.00 Hczcc (Hc)(zcc)
2 0.00 Hcz (Hc)(z)
1 0.00 zzcccHc (z)(zcc)(cHc) (z)(zc)(cc)(Hc)
1 0.00 zzccH (z)(zcc)(H)
1 0.00 zzcHcc (z)(z)(cHcc) (z)(zc)(H)(cc)
1 0.00 zcz (zc)(z)
1 0.00 zccz (zcc)(z)
1 0.00 zccccHcc (zc)(cc)(cHcc) (z)(cc)(cc)(H)(cc)
1 0.00 zcccc (zc)(ccc) (z)(cc)(cc)
1 0.00 zccPccc (zc)(cPc)(cc) (z)(cc)(P)(ccc), (z)(cc)(Pc)(cc)
1 0.00 zccP (zcc)(P)
1 0.00 zccHccc (zc)(cHc)(cc) (zcc)(H)(ccc), (zcc)(Hc)(cc)
1 0.00 zcHcc (zc)(Hcc) (z)(cHcc)
1 0.00 zcHc (zc)(Hc) (z)(cHc)
1 0.00 zPcc (z)(Pcc)
1 0.00 zHcc (z)(Hcc)
1 0.00 zH (z)(H)
1 0.00 ccccc (cc)(ccc) (ccc)(cc)
1 0.00 ccccPcc (cc)(cc)(P)(cc) (ccc)(cPcc)
1 0.00 cccPccc (cc)(cpc)(cc) (ccc)(P)(ccc)
1 0.00 cccHccc (cc)(cHc)(cc) (ccc)(H)(ccc)
1 0.00 ccPccc (cc)(P)(ccc)
1 0.00 ccP (cc)(P)
1 0.00 cHccz (cHcc)(z)
1 0.00 cHccc (cHc)(cc)
This still doesn't look quite right....
Let's try it anyway.
jsa2hip
-------------------------------------------------
#! /n/gnu/bin/sed -f
# Recoding superanalytic to "hip" encoding:
/^[^#]/s/ij/k/g
/^[^#]/s/ix/e/g
/^[^#]/s/is/r/g
/^[^#]/s/iiu/n/g
/^[^#]/s/y/i/g
/^[^#]/s/ci/a/g
/^[^#]/s/cg/8/g
/^[^#]/s/cs/z/g
/^[^#]/s/iin/m/g
/^[^#]/s/in/m/g
/^[^#]/s/ir/v/g
/^[^#]/s/qj/E/g
/^[^#]/s/qg/P/g
/^[^#]/s/lj/E/g
/^[^#]/s/lg/P/g
# Parsing of [czPE] strings:
/^[^#]/s/[zcEP][zcEP][zcEP][zcEP][zcEP][zcEP][zcEP][zcEP]*/@/g
/^[^#]/s/zccEcc/AU/g
/^[^#]/s/ccccEc/OX/g
/^[^#]/s/zcccEc/AX/g
/^[^#]/s/ccccE/MME/g
/^[^#]/s/zccEc/AI/g
/^[^#]/s/cccEc/OI/g
/^[^#]/s/zccc/RM/g
/^[^#]/s/cccc/MM/g
/^[^#]/s/Pccc/PO/g
/^[^#]/s/Eccc/EO/g
/^[^#]/s/Ezcc/EA/g
/^[^#]/s/zccE/AE/g
/^[^#]/s/cEcc/K/g
/^[^#]/s/cccE/OE/g
/^[^#]/s/cccz/OZ/g
/^[^#]/s/Ecc/U/g
/^[^#]/s/ccc/O/g
/^[^#]/s/zcc/A/g
/^[^#]/s/cEc/X/g
/^[^#]/s/ccE/ME/g
/^[^#]/s/Ec/I/g
/^[^#]/s/cc/M/g
/^[^#]/s/zc/R/g
/^[^#]/s/E/E/g
/^[^#]/s/z/Z/g
/^[^#]/s/P/P/g
-------------------------------------------------
extract-words-from-interlin \
-recode jsa2hip \
-chars "qoa8HPZAEIOUMXRKermnkvc@" \
bio-j-jsa.evt \
bio-j-hip
lines words bytes file
------ ------- --------- ------------
7054 7054 36231 bio-j-hip.wds
1967 1967 13458 bio-j-hip.dic
4658 4658 22234 bio-j-hip-gut.wds
862 862 4575 bio-j-hip-gut.dic
843 843 2464 bio-j-hip-fun.wds
5 5 24 bio-j-hip-fun.dic
1553 1553 11533 bio-j-hip-bad.wds
1100 1100 8859 bio-j-hip-bad.dic
Digraph counts (edited):
q o a 8 R M A O P E Z I U e r m n k v X K c @ TOT
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
. 1239 965 161 363 129 149 440 436 65 91 149 32 29 282 79 . . . . 7 8 16 18 4658
q . . 1227 . . . . . . . . . . . . . . . . . . . . . 1247
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
o 21 . . . 18 . 9 . . 61 714 . 394 380 893 178 9 . . . 7 10 . . 2727
a 2794 . . . 11 . . 14 9 . 23 . 19 26 411 275 333 109 33 19 . . . . 4104
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
R . . 26 22 31 . 107 . . . . . . . . . . . . . . . . . 199
M . . 32 132 142 . 95 . . . 36 11 . . 11 . . . . . . . . . 468
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
A . . 40 125 322 . . . . . 25 . 41 15 . . . . . . 37 . . . 609
O . . 40 167 404 . . . . 7 18 11 77 . . . . . . . 41 . . . 765
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
e 825 . 105 114 53 36 49 71 154 10 76 7 43 61 . . . . . . . . . . 1614
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
P . . 37 17 . . 22 8 66 . . . . . . . . . . . . . . . 165
8 50 . 37 1948 . 9 18 22 16 . . . . . . . . . . . . . . . 2113
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
E 10 . 75 795 . 9 . 32 61 . . . . . . . . . . . . . . . 996
r 401 . 36 64 . . . 9 16 . . . . . . . . . . . . . . . 539
Z 42 . 63 73 . . . 7 . . . . . . . . . . . . . . . . 196
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
I . . 20 173 391 . . . . . . . . . . . . . . . . . 12 . 608
U . . 10 179 321 . . . . . . . . . . . . . . . . . . . 513
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
m 339 . . . . . . . . . . . . . . . . . . . . . . . 342
n 114 . . . . . . . . . . . . . . . . . . . . . . . 115
k 36 . . . . . . . . . . . . . . . . . . . . . . . 37
v 19 . . . . . . . . . . . . . . . . . . . . . . . 20
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
X . . . 86 10 . . . . . . . . . . . . . . . . . . . 101
K . . . 14 10 . . . . . . . . . . . . . . . . . . . 26
c . . . 11 12 . . . . 12 . . . . . . . . . . . . . . 48
@ . . . 11 13 . . . . . . . . . . . . . . . . . . . 24
----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----
TOT 4658 1247 2727 4104 2113 199 468 609 765 165 996 196 608 513 1614 539 342 115 37 20 101 26 48 24 22234
There are some nice things in this table: \ccc/ = `O' and \cscc/ = `A' come out similar, and same for \cc/ = `M' and \csc/ = `R'.
There are some surprises, such as the similarity between \qg,lg/ = `P' and \cg/ = `8'; or between \lj,qj/ = `E', \cs/ = `Z',
and \is/ = `r'.
The slight differences between members of the same class may be telling us something, too.
\cc/ and \csc/ are similar, but
only \cc/ is followed by \lj/, \qj/, \cs/, or \ix/
\ccc/ and \cscc/ are similar, but
only \ccc/ is followed by \lg/, \qg/, or \cs/
only \cscc/ is followed by \ljcc/ or \qjcc/
\ljc/ and \ljcc/ are similar, but
only \ljc/ is followed by unparsed \c/
only \ljcc/ can be preceded by \cscc/
Also, \cgci/ is probably a letter; indeed \cg/ is followed by \ci/ 91% of the time, although \ci/
occurs in other contexts too.
What can we conclude from these bits, really?