#idk if its denoted w a space or not

[senzacaponecoda]

pAA?

(Decided to reduce the names to E, O and S because this isn’t something that should be treated seriously or like there’s anything in particular respectable or responsible about it, but might be helpful to someone like me trying to find something out for personal reasons.) So since I finally got my hands on a copy of O and S’s reconstruction of PAA, I thought I’d go through and pick out what I thought were their best etymons against what I think are E’s best etymons. As I’ve said before both reconstructions seem to be basically mass comparison reconstructions.

The wiki page on Afroasiatic claims they agree on basically nothing, and out of 1000ish and 2600ish I still got less than 100 shared roots, so that’s probably accurate. The page and what I pulled out didn’t exactly line up though, but also there were some roots especially on E’s end (due to a need to reconstruct verbs to nominalize for his PS theories) that I outright changed the meaning of because they actually made his individual arguments stronger; like, say, if you had 4 roots that mean knife in each sub-branch, why assume the etymon means cut?

I compared the roots in case there was something that was an obvious oversight, and while I don’t think their methods were all that good (although O and S’s are a lot better), I didn’t want to discount everything in case there was something like “well their vowel systems are different so I’m not counting roots where even V[os] = V[e]” or like “O/S posit (g,x) < *q while E has (g,x) < *ÿ” being overlooked and miraculously they agreed on like 300 more roots than the page gives them credit. There’s also issues like, well, “Is this particular author doing such a bad job but the critics aren’t seeing that, that he/she/they are dragging the other author/s down by undermining their ok etymons?” etc.

So, like, I’ve messed with the data in a way that’s irrepairable and shit. Mind that. This isn’t scholarly work and isn’t intended to be. And I messed with some of the data - even if I think it’s more representative of their works at a level, from any kind of responsible position I need to make it absolutely clear that it’s not really reflective of their work, but a reflection of my amateur, shitty thoughts on their concordances.

So, for the curious:

So when putting this list together I accidentally kept mixing which side I wrote whose roots together. E tends to have labiovelars where OS have o, E has long vowels, OS don’t, E has tones, OS don’t, E has the voiced velar fricative I transcribed as ÿ and ŋ, OS have q and q’ vs k’ and k. E has p’, the rest I guess you’ll have to guess.

I also marked a few words with * for what might be imitative or baby speak words. *? like on horn is “I’m on the fence of it being imitative”

Vowels are pretty much ignored, but otherwise definite correlates:

kol-f bark (n) dam blood k'os~k'as bone naf breath ben~bin build di3, da3, du3 call dumn~deman cloud k'at' cut* 2ab father* pir fly*? ba2 go sim hear lib heart k'ar horn*? inkwal~2ankol kidney 2er~2âr know lVk'~lak lick* tir liver sum,sim~süm,sim name wan open bu~baw place dak,duk~dik pound k'u2 rise tuf spin ra3-raa3 sun dab,dib-dub tail les,lis tongue ma2 water

Close matches, differing by not a lot of phonological space, but the correlations seem unsystematic or quasi-systematic (E *k = OS *x? or the other way). A lot of this might be due to obfuscation of the subgroups; it’s known beyond these reconstructions that Semitic b, p, f correspond to Egyptian b, p, f but there doesn’t seem to be a pattern to the correspondences beyond [+labial][-nasal]=[+labial][-nasal]... except some roots where b = m, mostly Egyptian to Semitic (iirc it’s believed that snb = slm).

! maaw~mawut die ! har~heraw day ! t(l)'ok' beat ! bak~bax burn ! k'al~k'ar burn ! (t)san brother ! kor~kw'al angry #related to kidney? ! pak~pax break apart* ! yar~3er burn #n.b. OS *e = E *ya pretty consistently ! ka2(up) cover ! t'ub~duf drip* ! g'arub~ÿar(b) dusk ! 2et~iit eat*? ! gur~guud enclose ! 2ir~2il eye #2il is also given for both by the wiki author ! 2aakw~2ax fire ! ŋiiwr~gir flames ! pur~fir flower 1 k'ur~kâr go around ! ĉa3ar~ła2r hair ! qafV3~gâf hold ! qam~kam hold ! fil~bul hole ! ĉa(2i)d~gwi/ad land ! ne2ul~ñaw moist ! bakr~bar morning ! âf~2ap mouth* ! ĉer~sar,sir root ! cab~sVp sew ! sur~tsur sing ! tsoon~soon smell ! bak'~p'ak' split* ! da2~daw walk

Really kind of stretched correspondences, usually requiring twice the amount of special pleading as above:

!! rip~2erib sew !! dabn~zab hair !! yawr~yabil bull !! büł~fil skin !! 3ir~raw sky

A root that’s probably wrong but I didn’t delete it:

!!! c'eyg~c'a3ek shout

Roots on the wiki article it says are part of “the fragile consensus” but either escaped my standards for good etymologies from E and OS or were lost when I changed E or OS’s proposed definitions to fit the data they presented for etymologization:

>> (2a)bVr bull >> (2a)dVm land >> 2igar~kw'ar enclosure >> 3ayn eye #I think for both this was only found in Sem and Egn, which is proto-Sem-Egn, not PAA >> bar son >> gamm mane, beard >> gVn cheek, chin >> gwar3~(gora3?) throat #These words weren’t in E’s but look like it and I guessed on what the OS reconstruction would look like >> gwina3~(gona3?) hand >> kVn cowife >> kwaly??? kidney #that this isn’t VnkwVl damns something >> k'awal, qwar say, call >> sin tooth >> siwan know >> zwr seed >> łVr root >> šun sleep, dream

These I sort of gathered based on a decent scholarly Egyptologist’s work

** (n)i,ku,nak(u) 1sc ** nVn 1pc ** kumV, kV 2sm ** kimV, ki 2sf ** suwa, sV 3sm ** si(y/t)a, si 3sf **-ú~-aw nom (earlier an ergative) **-á abs **Egyptian had an outright ergative participle so like yeah Some AA assumptions $$ m- participle, possible 3 or 4 denoting instruments, resultatives, passive and active meanings, and more. $$ s- causative (likely su-) $$ -í nisba = adjectivizer $$ -át/-út ending for abstract, mass, diminutive, augmentative, female nouns $$ top two combine frequently $$ maybe k became a masculine? maybe u/w too? $$ perhaps an -n- passive? -t- passive? reciprocal voice? (kick each other) $$ emphatic coronal series $$ prefixing conjugation {2 t t y t n t t y t} on imperfect-type stems $$ core verb stems seem to be perfect vs imperfect $$ a lot of sources report *-r has spreading out semantics at the end of a Semitic verb and I believe it. $$ suffixing possessive pronouns $$ probably all of pAA derives its verb system from those same possessive pronouns suffixed to participles, which probably helped the ergative to nominative alignment change. Again, Egn had an ergative participle at one point. It actually used a slightly different form of the possessed agreement system for that though, maybe reflective of a dead case system. $$ personally I think egn broke off early, sem and amz were geographically between it and ethiopia, the cushitic languages broke early but stayed in contact, chadic is cladistically cushitic that broke off in the middle of the ethiopianization of cushitic in the post-islamic era, since it seems to be in cultural memory. beja wasn’t as ethiopianized and is slightly closer to semitic, amz, and in many ways egn surprisingly, omotic was probably adopted into the family, ongota is too creolized to know anything, tones, vowel rebalancing, probably influences of Central/Eastern Sudanic. I have no idea why languages with ATR (i.e. pharyngealization) distinctions would make pharyngealized languages switch to ejectives, so if sem-amz is a thing you just have to postulate why they switched, since egyptian is no help. for all i know egyptian always had an aspiration distinction (which would explain some insane correlations) and then they fell together and split again for coptic.  Don’t think this can be projected to the proto-lang: $$ ablauting everywhere, in nouns and verbs. masdars and singular nouns probably reflect original vowel patterns. !! E and OS reconstruct consonant systems similar to PS and idk about that. Both give 5 vowel core systems but for different reasons and different results. E has length and the central short vowels fall together in all but Omo. Some shenanigans happen to get Cu uu, ii, aa, ə, and ä. Ch ends up with a i u ə, and the other three a i u. OS has e yield ya and o yield wa, ü for u~i results, and does overall less movement.

These are numbers from my number project:

## tsin~tsar two ## kwrad three ## fVd'w four ## magw ten

It’s interesting that many of these shared roots are in fact Swadesh roots and that that mostly happened on accident. E reconsructed 1000 roots, OS 2600, and just using judgements like “a root needs to mean the same thing in two (standard I held E to) or three (standard I used for OS) subbranches” to sift through them brought me down to these. But like most of these roots are fundamentally two segments long. There were even ~false matches in some of them with the phonaesthetic shape of their English glosses - horn would go back to something like *karn in PIE, which looks like the root here. Maybe horns make a kar sound, but with 2iit~2et, idk, I don’t think of eating as making an “eat-eat” sound; I actually threw away a number of roots that had a “kwa-l-kwa-l” (whence Ar. 2akulu) sound which at least kind of sounds like swallowing.