wildcard fallacy - pannous/hieros GitHub Wiki

wildcard fallacy

Only five consonants in hard / weak form:

one sign generating all words 'PTRM=VCLN'

𒌓 Values: a₁₂, ah₃, al₅, amna, ara₇, ayaₓ, bab₂, babar, babbar, bar₆, baraₓ, ber₂, beš₁₂, bir₂, da₇, dadagₓ, dag₂, dam₂, e₈, gal₈, ha₅, had₂, hada₂, haš₃, heš, his, hiš, hud₂, kum₄, la₇, lah, lih, luhₓ, par, para, per, peš₁₂, pir, pirig₂, piriŋ₂, ra₃, sa₁₆, salah, siki₂, šamaš, šara₉, šarag, ta₅, tam, tama, tu₂, ṭam₂, u₄, ud, ug₄, ul₆, um₂, un₅, ut, uta, utu, uṭ, waṣitₓ, zalaₓ, zalag, zalaga, zalah, zalaqaₓ

TODO: rewrite

There are universal sound changes and seemingly odd changes specific to languages.

It is often reasonable to assume a very blurred out distribution for a particular Reading:

Look at the cluster Pod-Pfote-bein-boot-foot-ferse ... covering everything from p to f

after 3000 years in the Egyptian language we can suspect many such broad traces combined :

Unfortunately sound-change and even alphabetic-change are very real, we just need to be very careful.
Linguists today are very careful, which is good. Unfortunately this carefulness might go too far and result in Bad-Linguistics of the second kind: Highly plausible connections get rejected due to some dubious sound change law, as if language strictly abides to laws at all times.

D-DT-T
T-TH-S-C
D-DS-S (demi - semi)
C-CH-CK-K
D-DZ-DJ-DCH-CH

combining two of these can lead to a Spektrum leading from K to D, which is not a usual sound change.

asides from the ordinary sound changes and R-L in Egypt we find an affinity of
NV
NM
MB
MW mit/with mo/wood

Traces between (semi)vowels don't help
W-U-OO-O-OA-A-AR W-R etc

quickly all letters of the alphabet can be reached with just three or four consonants: M T P R

Thus giving a sign the infamous sequence µPTR can lead to almost all words in the dictionary if we allowed substrings. µPTR=map.tr µPTR⋍pthr⋍prth⋍price … pick any word.

Thus one needs to be very careful avoiding this sequence or extending known words with 'just an extra m or r'.

That's a dilemma because some words actually might have had such a sequence / cluster
P-T-R father ...
𓆑 PTHN? probably PTH ⇔ 'PSI' though
CEPHAL (head) ⇔ KOPPER ⇔ CAPUT.ma

For 𓆑 at least the pain can be alleviated:
Axiom 8: Most letters had a full (original) reading and an alphabetic one.

𓏶 MAS/PRO/CEN

fortunately these appear in quite specific sequences & contexts.

In those cases one must be careful abusing the blurriness for words which probably have such sequence.

TODO: rewrite

atomization

Usually splitting words into components is the epitome of the wildcard-fallacy, yielding almost any desired reading.
Thus whole word comparison must be preffered under all circumstances even if compartmentization seems very plausible as in the case of chariot 𓌝.

One must not overemphasize or misinterpret the direct translations either.
Even thouth 𓌝 𐎎𐎗𐎋𐎁𐎚 mrkbt /markabtu/ is a perfect match, it could still be a loan with secondary connections.
𒁍 buʳ bear in 𒁕𒁍 da-buʳ the Bear דבא debā 大 !

In some cases a split reading is strongly supported by Coptic, but even this might be some kind of PSM :
𓂋 𓈖 𓊪 𓆶 𓆰 renew ⲓⲣⲓ ⲛⲃⲉⲣⲓ

Examples in Sumerian

𒌓
𒌓 UTU UT UD húd white ⇔ 𒎕 𒇺
𒌓 DAG2 Tag 𒍴 DAG day dayn ⇔ dry
𒌓 TAM time
𒌓 PER PAR BABBAR bright
𒌓 LAH la7 liḫ liš ḫiš Licht 日 light solara 𓇳 RA
𒌓 Šamaš 𒀭𒌓 sun'ra 𒌓 •rat radiate

𒋻 esirra :
𒋻 SILA calla street sluice 𒀀𒋻
𒋻 TAR track 𒋻𒊒 tar-ru 𓅂 tyw® طير ṭīr
𒋻 KU course
𒋻 qud, qut cut ḫaš
𒋻 •šud shade, shadow, side-wall
𒋻 šel šil sil

Ambi Partial wildcard:
𒀀 A : eau aqua
𒀀 BA AB Ob hebb
𒀀 C aqua?
𒀀 D/T/S forms NOT attested standalone, only in 𒀀𒇉 ÍD water
𒀀 mû ≈ मेंह mẽh 𓈗 mar maim 💧 💦

𒀀𒉌 a-zal pelvis
𒀀𒉌 mu.ni/zal მენჯი menǯi
𒀀𒉌 pal/va.zal pelvis ≈ Fluss
𒀀𒉌 aqua.ni אַגָּן agán
𒀀𒉌 ID/a.zal ᵗazal таз taz --

𒀀𒍪;𒀀𒍫;𒉙 a-zu;a-zu5;azu doctor
𒀀𒍪 azu Arzt #asû 博士 はくし hakushi はかせ hakase
𒀀𒍪 medicus 𒀀 mû 𒋢 KUŠ 𒀀 mûter < mwater
𒀀𒍪 buzu #ᵇasû 博士 bóshì "boss" 박사(博士) baksa √abundant; broad ;plentiful;

ambivalence

𒃾 ĝeštin /ɱeštin/ Most ĝ ≈ ɱ ! 𒃾 GEŠTIN = 𒄑𒁷 GIŠ+DIN Gersten (Wein) 𒃿 √gär gud @ 𒆳