User:Chrysophylax/Golden Afroasiatic

writin' up some stuff

Phonology

Vowels

In stressed syllables the vowel system centers on an opposition between three vowels: /a i u/. They also come in a long variant. These are the only permitted vowels in stressed syllables. In unstressed syllables there is a tendency to de-emphasise the distinctive qualities of the vowel with occasionally extreme coalescence into a schwa /ə/.

Consonants

GA has a rich system of consonants with articulations ranging from labial to glottal and with a family-specific feature opposition: [-voice], [+voice], [+emphatic]. The nature of the [+emphatic] feature is not clear but it presumably involves some sort of glottalised articulation.

A series of class-internal exchanges between phonemes in the [+labial] series and those in the [+voice] / [-voice] series suggests that the features did not carry heavy information load initially and were thus easily exchanged.

Morphology

Bases

There is no particular distinction between nominal and verbal elements or lexical bases as we shall call them. Nomina can be used as predicates and vice versa although the correlation in meaning isn't always instantly clear. E.g. npš ‘to be alive’ or ‘life’ as a transparent example and qt ‘to take’ or ‘hand’ as a less transparent example.

Nominalisation strategies

One basic way of creating nouns (insofar as a noun class separate from predicates can be even postulated) seem to be formed with the base vowel *a as in the templatic vowel patterns -a-, a-a, usw. e.g. Egyp. rmṯ /raːmac/ ‘man’, sn /san/ ‘brother’.

Nominal morphology

Number

PAA seems to have had at least singular vs. plural with hints of a dual. I've chosen to go with a system reflecting the maximal situation and as such Golden Afroasiatic has three numbers. In addition to this singular-dual-plural system a singulative-collective system seems to have existed as well, with the ever-useful -t suffix marking a single item of nouns usually considered mass nouns.

Singular

The singular (sg) is the basic, morphologically unmarked number and represents one instance of a noun, pronoun and so on.

Singulative

The singulative (sgv) is a separate number form used to mark a single item. It is only used with nouns unlike the other numbers and especially those that tend to be interpreted as mass nouns. It is specified by the -t suffix.

*zwr vs. *zwr-t

‘seed(s)’ vs. ‘a single seed’

Dual

The dual (du) is marked with either -(a)n or -(i)y.

gʷinaʕ vs. gʷinaʕan OR gʷinaʕiy

‘a hand’ vs. ‘a pair of hands’

Plural

Plural formation strategies are many in Golden Afroasiatic with it being difficult to predict which form will be the dominant one for a selected noun. Additionally, when context is clear, plural marking may be eschewed completely. Various sociogeographiclects may prefer one pluralisation marker/strategy over the other.

W-affixation

The addition of -w (-ū) is a frequent pluralisation strategy in many Afroasiatic languages for masculine nouns.

ʔab-ū-

‘fathers’

T-derivation

The almost ubiquitous marker -t surfaces again, this time as a derivational suffix forming collective nouns. This elegantly mirrors the role development of Indo-European -h₂.

sin-t

‘a set of teeth’

Reduplication

Golden Afroasiatic offers two competing reduplicative processes for forming plurals, one more common and one more rare.

End-to-front

The most basic formation of the plural is formed by R→L reduplication of the root morpheme. E.g.

*lis- → *lislis

‘tongues’

*maʔ- → *maʔmaʔ

‘waters’

Lipiński provides some examples of this formation in Afroasiatic languages: Hebrew mēmē ‘waters’, Hausa (Chad.) dambe ‘struggle’ → pl. dambedambe, Bedja (Cush.) san ‘brother’ → pl. sanasanā .

Radical reduplication

A less frequent, less widespread reduplicating formation of the plural which only occurs in the South Ethiopian Semitic, Chadic and Cushitic branches of Afro-Asiatic is the one where the last radical is reduplicated L→R. As it seems to be attested in both Semitic, Chadic and Cushitic, we will adopt this formation of the plural as well. The languages differ in how they pluralise a little, with the South Ethiopian languages adding a plural -t on top of this, while both Chadic and Cushitic have zero additional morphemes beyond the partial reduplication. Chadic and Cushitic strategies seem to have primacy in light of the originally derivational meaning of -t.

ḳas-as(V)

‘bones’

Note however that Ratcliffe argues against this being a PAA feature and instead claims it as innovation. For Golden-Afroasiatic however we will stick to this being a retention.

A-apophony

Another possibly ancient and widespread formation of the plural is by infixing -a- which per Lipiński occurs in Semitic and supposedly has clear parallels in Berber, Cushitic, and Egyptian. Chadic is not mentioned at all and for seemingly good reason. Schuch writes on the Chadic language Bade's morphology and mentions that a-infix plurals as postulated by Greenberg and other linguists seem generally to come from a misunderstanding of a Chadic syllabic feature. However, he notes that there does seem to be something similar to an a-infix in a few roots and gives the Bade example ə̀tlkùmən ‘fool’ → ə̀tlkwàm-cən-ən (emphasis mine).

Golden Afroasiatic allows a plural formation with an -a-. This strategy is slightly less common in frequency than the others.

ŝaar

‘roots’

Gender

Golden Afroasiatic has an opposition between the genders masculine (m) and feminine (f), where the feminine gender is overtly marked with -t.

Verbal morphology

Aspect

Radical doubling

• C₁C₂C₃ → (yi-)C₁aC₂C₂aC₃

  swn → (yi)sawwan=i

  ‘npf\know=1sg’

E.g. Mubi (Chadic) gìidì (pf) → gǐttà (npf) ‘to descend’; Egyptian jrj=f (pfv) → jrr=f (pres) ‘to do’.

Yod-prefix

• C₁C₂ → yiC₁aC₂a

  tˤd → yitˤad=šu

  ‘npf\say=3msg’

Syntax

Morphosyntactic alignment

Proto-Afro-Asiatic's daughters seem to consistently hint at an earlier ergative system, whether it be the 'unbound' vs. 'bound' of Berber, the odd accusative marking for predicates and certain genitive constructions in Classical Arabic, Akkadian, and other Semitic languages, structural similarities between agentive markers and instrumentals, usw. I have chosen to fully embrace the ergative theory and construct a following opposition in NPs.

agent-like	*u
patient-like	*a

While several branches, of which, most notably Semitic show postposed forms of these, the Berber and other families in the phylum show preposed forms. No order seems particularly deserving of primacy as far as I can see so I'll allow free variation in Golden Afroasiatic.

Noun phrase

Copulative phrases

Following the example of the oldest attested languages (Egyptian and Semitic), GA accepts the following structures: [NP₁ PRON PRED], [PRED PRON NP₁], [PRON PRED], [PRED.ADJ NP₁]

ʔisim šV[+high] ŝar

(name 3msg root)

‘the name (is) the root’

Adjectival copulative sentences

[PRED.ADJ NP₁]

yimaʔiyan ʔiynaan-ki

‘your.f eyes (are) blue’

Genitival constructions

Possible strategies:

• Semitic-style construct with the head initial. Additionally, the possessum receives the -a marker whilst the possessor -u. Nothing can intervene between them and its treated prosodically as one word.

  ʔadam-a ʔabVr-u

  ‘land (of) the bull’

• Apposition can also be used.
• Adposition agreeing with possessor in gender and number - "genitival adjective".
• Pronominally by suffixing a dependent pronominal element to the noun.

  maʔ-šV[+high]

  ‘his water’

Verb phrase

The basic order of constituents in a verb phrase is VSO with an optional SVO order as a result of a raising strategy when the subject is being emphasised.

Vocabulary

Very little is agreed upon and very little can be reconstructed somewhat “safely” at all; what does seem to be quite retained is the pronominal system and various demonstrative and a bunch of affixes. A small list of words that the only two competing dictionaries of PAA agree upon is also included.

The thirty

Ehret (1995) and Orel & Stolbova (1995) all agree on the following 30 PAA bits so we'll include them. (Source: Afroasiatic languages article on English Wikipedia )

Number	Proto-Afroasiatic Form	Meaning	Berber	Chadic	Cushitic	Egyptian	Omotic	Semitic
1	*ʔab	father	✔	✔	✔			✔
2	(ʔa-)bVr	bull		✔	✔	✔		✔
3	(ʔa-)dVm	red, blood	✔	✔	✔			✔
4	*(ʔa-)dVm	land, field, soil		✔				✔
5	ʔa-pay-	mouth		✔	✔			✔
6	ʔigar/ *ḳʷar-	house, enclosure	✔	✔	✔			✔
7	*ʔil-	eye	✔	✔	✔
8	(ʔi-)sim-	name	✔	✔				✔
9	*ʕayn-	eye				✔		✔
10	*baʔ-	go		✔	✔			✔
11	*bar-	son	✔	✔				✔
12	*gamm-	mane, beard		✔	✔			✔
13	*gVn	cheek, chin		✔				✔
14	*gʷarʕ-	throat		✔	✔			✔
15	*gʷinaʕ-	hand		✔	✔
16	*kVn-	co-wife	✔	✔				✔
17	*kʷaly	kidney		✔	✔		✔	✔
18	*ḳa(wa)l-/ *qʷar-	to say, call		✔				✔
19	*ḳas-	bone	✔	✔		✔
20	*libb	heart		✔	✔			✔
21	*lis-	tongue	✔	✔				✔
22	*maʔ-	water	*aman	*aman		✔		✔
23	*mawVt-	to die	✔	✔		✔		✔
24	*sin-	tooth	✔	✔				✔
25	*siwan-	know	✔	✔		✔
26	*inn-	I, we	✔		✔	✔		✔
27	*-k-	thou	✔	✔	✔			✔
28	*zwr	seed			✔			✔
29	*ŝVr	root		✔				✔
30	*šun	to sleep, dream		✔				✔

Strongly suggested

Proto-Afroasiatic Form	Meaning	Berber	Chadic	Cushitic	Egyptian	Omotic	Semitic
**ɬa-	cow?	?	West, Central ɬà	Agaw ləwi, East lali, loon, West ɬee	?	?	Akk. luʔum, luu; Arab. laʔan; Shehri (MSA) léʔ/lhóti

qūyka nūtan

“ergative” system - distinction between agent case (casus agens )nom-instr.-loc)) and predicate case (casus patiens (predicative, accusative))

possibly even deeper links to Bantu family? (common causative affix -s-, reciprocal verb suffix, m- prefix and some other funky stuff. etc.)

stative/perfective/imperfective?

prefixing conjugation

1sg indep. ʔǎn- ʔǐn

dep. -i | could it be originally enclitic ʔǐn??

Egyptian primarily loss of prefixing conj. mostly suffixing, loans from substrate looks a bit Sudanic.

t- feminine/diminutive/singulative morpheme (cf. semantic development of IE -h₂ from a collective/marker/derivative noun -> feminine)

Cushitic preserves verbal aspects, “ergative” and “non-active”, has causative, passive, reflexive stems; this matches quite well with semitic (s-, -t-, radical doubling)

North Cushitic: ipfv adabbīl, pret adbíl, volitive-jussive-conditional? īdbil

Berber preserves ergative-ish features better: casus agens vs. casus patients. Agens is prefixed with u-, a- is prefixed to patiens, has stative, imperative and jussive (used in subordinate clauses. cf. North Cushitic conditional used in negative clauses).

Vowel lengthening characterizes intensive stem in Tuareg like some(?which?) Semitic and Cushitic languages. Lipiński gives for Tuareg

intensive stem formation: -lkām-; -lākkəm-; 'follow' (√-lkəm-)
causative: -sərtək-, 'fell'
reflexive-reciprocal: -mətrəg- 'be freed'
frequentive: -təffəg̊- 'go often out'
agentless passive: -ttwaddəz-, 'be crushed' which apparently parallels an Egyptian pseudo-passive (?)

Chadic: shares feminine prefix t- (cf. Berber ta-mazig), apparently one way it forms noun plurals is with suffix -n and -a- insertion (Cf. Arabic? and Berber). shares intensive/pluriactional verb with radical doubling (Cf. Sem+Cushitic).

Wobbly-Timey-Wimey

• Akkadian: prolly first wave of Semitic speakers to invade Asia

• Aramaic: attested features in ca 850 BCE. include broken/internal plurals. Two inscriptions from Zincirli (unspec. 8th cent. BCE.) retain case endings in plural and have no emph. state

• Arabic: mentioned in Neo-Assyrian texts per tribal nomina. Ṣafaitic texts show h- and ’al- as articles so no unified dialect. SA script shows drop of nunation and of case but preserving distinction of š and ś, and of the emphatic series. -t of feminine preserved here and there.

• Gafat: ~Blue Nile, W. Ethiopia, died,r eplaced with Amharic. Preserves archaic plural noun kitač¸( <Sem. *kitāti 'children') ~ cf. Egyptian ktt “little one”. Also preserved mossay “child”, cf. Egyptian mś ( < mśi 'give birth')

Phonology

?no distinction between voiced/devoiced stops in Semitic and Egyptian? e.g. ’bd vs. ’bt “perish”, b‘l vs. p‘l “make”, kbd vs. kbt “be heavy”, ndn vs. ntn “give”, nbš vs. npš “breath, life”, nbk vs. npk “well”, šbt vs. špt “full moon”.

Egyptian: k-p-n and k-b-n for Gbl “Byblos”.

Emphatic identity: pharyngealisation in Semitic - Lipiński argues for primacy of this based on ancient phonetic changes and transcriptions, e.g. Ugaritic nṯ̣r > ng̊r “to guard” pointing out that the interdental fricative ṯ̣ had become a velar fricative acos pharyngealization. Also, spread over words, called tafẖim in Arabic, may explain variation of roots and u-vocalism in East Semitic, Lipiński gives ex. qurbum for qarbum “near”, inaṣṣur for inaṣṣar “he guards”.

For PSem. Lipínski (2001) gives vowels: *a, *i, *u with long vowels *ā, *ū. Mentions vocalic *l and *r as very probable as they're attested in oldest phases of daughters.

Consonants here are then labial *p (which did a P-celtic early on in a bunch of langs, e.g. SArab. ḥrf “autumn” < ḥrp) and probably *b (but see note above about a possible lack of distinction between voiced/voiceless), *m. Semitic dialects show partial confusion of *b and *m - not too strange if thinking of a [b] allophone of orig. /p/ causing a bit of havoc.

Berber and Assyro-Babylonian words show change of nominal prefix m- -> n- when before labial. Additionally, Berber and East Semitic point to alternation between m / n, Lipiński gives ESem. wasāmu “to be skilled” and Berber wsn “to be skilled” (modern Tuareg a-mūssen “skilled man”)

possibly phonemic status of t / k and d / r suspect in Afro-Asiatic. Semitic alternates in pronominals, in Cushitic masc. k vs. fem. t. Cushitic and South Ethiopic have traces ( shifting l > d in Gozo), Berber too (Numidic mnkd, Tuareg a-mnukal), additionally, in the macroregion: Bantu has similar (-tund < -tunl) and Amharic, Argobba, Gafat word qänd < qarn (Ge‘ez) (nd < rn).

PSem. : eight monosyllabic root morphemes types:

• (short) Cv (rare, kinship, human body, numerals 1-2)
• (long) Cv̄, CvC, C₁C₂vC₃
• (ultra-long) Cv̄C, C₁C₂v̄C₃, C₁vC₂C₂ and C₁vC₂C₃.

Ca: (*wa- > ) *u-, *-ma “and”, (*ha- > ) *a- interrogative, *ka- “as, like”, *la- “truly”, *pa- “and so”

Ci: *bi-, “in”; *li- “for”

Cu: *lu-

Disyllabic CāCiC 'active participle of triconsonantal'

A comparison of Berber vs. Semitic points to CvC class maybe being originally CCvC or CvCC in PAA. Heb. lēk 'go' possibly *hlik ~ Berber 'llukk' "go on!" as 'll' probably 'hl'

Causal in Semitic -y?

Morphology

? no clear cut distinction between verbal root and nominal root? Lipiński: Somali qufa‘ “cough”/“to cough” / East Semitic qātum “hand” vs. Somal. qād “to take”

Shared Berber-Semitic C₁vC₂C₂vC₃ pattern for (intensive mostly in Sem.) adjectives. (Wonderful tentative example Lipiński : /ṭubbūẖu/ ⟨Ṭù-bù-ẖu-d’À-da⟩ “Very slaughterous is Hadda” in Palaeosyrian). Berber ex. a-məlləl “white”, a-wəssar “old” - Assyro-Babylonian qattanu “very small” (match with Egypt. ktt?? “little one” maybe)

Berber-Semitic show shared root reduplication for noun formation. (meaning?)

Sem. t(a/i/u)- “verbal nouns signifying an action, nouns of place, animal qualifications”

East Semitic infixes -t- for intensive adjectival nominals. Existed also in NSem. also. PAA archaic remnant or innovation?

Gentilitial/adjectival -iy attested in An Eg. and Sem. probably postpositional origin, also forms genitive -i

Feminine -at probably < -a/iyt ~ matching -īy + -t

Arabic placenames ending in -a/ā show a suffix -āwī (Ṣafā’ vs. Ṣafāwī) could be in An. Egypt. ḥmww “craftsman” ~ ḥmt “craft”

-t widely attested in Ancient EGyptian,Berber, Semitic, Cushitic. In Cushitic seems to be originally collective in a bunch of old conserved words. Lipiński gives Oromo abbōtī “elders” ~ Old Bab. ummatum. Probably primacy (cf. IE similar development) ~ abstract/collective/feminine meaning.

Semitic compounds attested: Gafat abälamʷä “shepherd” < abʷä 'father' (cf. *‘ab id. ) and älamʷä “cow”, Phoen. Mlqrt < milk “king” and qart “city” (ASoIaF Quarth anyone?). Assyro-Babyl. ištenšeret “eleven” < ištēn “one” + ešeret (“ten”).

Gender

? Semitic + Berber + Cushitic share t- with the meaning of fem. however, Sem. shows bits of noun classes which are somewhat shared. Can we project to PAA?

• Sem. (only?) *ya- names of animals, plants, proper names.
• m- verbal noun/tool/instrument/agent/participles/nouns of place: tool with which/means by which action is done, but also where action happens. Egypt. mnẖt “clothing” ~ wnẖ “to dress oneself”, mrẖt “fat” ~ wrẖ “anoint”
• *-b ~ wild/dangerous animals (missing partially in Chadic and Cushitic but then again that can be said for any of these affixes it seems, everyone randomly lacks one). Eg. 3-bw “elephant”, d-b “hippo” ~ Cushitic ló-ba (same meaning, for d~l alternation, see above) double check meanings: Faulkner gives hippo as ḫ3b
• *-r~l seems to have been for domestic/tame animals, cf. Egyp. iy-r “deer”, s-rw “sheep”, , Cush. zab-bä “lion” (wild) vs. dáb-el “goat“ (tame)

• *-n body parts, e.g. Cushitic gʷad-n “rib”, Gurage ãfu-na “nose” ( < Sem. ’anf+na), Oromo č̣inā “side” ~ Amharic and co. č̣ə-n “thigh”, Oromo af-ān “mouth” ~ Ge‘ez af “id.”.

Plurality

• a-apophony

There's a partial dual in Semitic which seems to have been enroached upon by the plural suffix. In light of the collective and the sporadic occasional widespread attestation, presumably the -n dual is an archaic vestige of either Berber-Semitic or higher up and projected to PAA.

Chadic, Berber and some Semites have plural in -n beyond vowel lengthening, and -t serves as a plural too in semitic (giving 'feminine' plurals to masc. nouns) and Cushitic.

Root morpheme reduplication seems to be a thing in Egyptian, Chadic, less Cushitic, regular Semitic guise and mystical Eblaite. In Egyptian initial redupl. of hieroglyph marks dual, thrice plural.

Chadic, Cushitic, and South Ethiopian languages and Amharic shows partial reduplication of final radical L->R e.g. Hausa kofa -> kofofi

Spread broken pattern in all?most? of PAA (Cushitic, Chadic, Semitic, Berber, Egyptian). e.g. Eg. ik3 for k3 “soul” in situ usual k3w

Old collective use of such a pattern in CArab. OAss. Assyro-Bab. (damqum “good” -> dumqum “lit. good things”) OAk. C₁vC₂C₂a(C₃) + -ū/-ī or -ūtu /ātu

Cases

North Cushitic shows similar system to Berber where subjects are prefixed with ū-, ā- (sg., pl.) and ō-, -ē (id.) for object, cf. Berber u- vs. a-, i-. (id.) In both, the feminine/collective t-' precedes. Ancient Egyptian has nil and uses word order. Semitic is questionable but as Semitic seems to be a bit closer to Berber it's quite possible. Semitic seems to have suffixed -a and -u.

Lipiński suggests Afrasian languages have verbs which agree with agent in P, G and N but concord with non-active by pronominal suffixes.

Berber-Semitic broadly agree on using the affix /-a/- for construct and nomen rectum with /-u/-

"new" genitive -i in Semitic is probably postposition -īy (ablative-ish in East Cushitic)

East Semitic has -uš which probably -iš which is a mostly postpositional particle, note however use as a prepos. in Paleosyrian. Mix of particle as both post and preposition is Afroasiatic-ish. Cf. Sem. wa- / -ma “and” with alt. w: m

Semitic -ah is orig. postposition directive in meaning, Ug. šmmh 'heavenward'

Predicate state -a projects to PAA as expected, cf. Egyptian old perfective, Classical Arabic perfect, Semitic stative

Adjectival elatives

Sem. has ša/u- (East Semitic, low frequency, archaic?) and ’af‘al

Numerals

• 1

Sem. ‘(i)št could match Berber iǧ~išt 'one'

MSA. Mehri, Soqoṭri have a resembling ṭāṭ~ṭād~ṭd “one” which might be linked to Cushitic *dad “someone”

Egyptian w‘(y-w)[-t] matches Berber yiwən, yiwət “one”.

• 2

Eg. śn-w(y), śn-t(y), Berb. sin, snat, sən, sənt ~ Pr. Sem. ṯin ~ OBab. šinā

• 3

Berber-Semitic match: PSem. *śrat & Berber ḵraḍ/šaṛḍ

• 4

Several words, Berber numeral kkuẓ matches Sem. nouns for "handful, take with the fingertips, close four fingers over hand", third Afroas. word match in Egyp. fdw, Bedja fáḍig and Hausa fu’du (no etymon?).

• 5

Sem ẖamš- is “five” and a hand - seems to be related to EGyptian ẖpš ( m>p) “fist” and Bedja numeral asa used in compound "asa-gwir" “six” (Lip. gives etym. “5+1”) and Berber səmmus “five”

• 6

Psem. šidṯ- ~ Eg. śrś / śiš-w ~ Berber. sḍis and Hausa šidda.

• 7

Egyp. śfẖ-w, Berber sa, Sem. šab‘

• 100

Sem. *mV̄t-/*mv’vt- ~ m-t, Tuareg. təmeḍe. No links outside Berber-Semitic. Cushitic and South Ethiopic have forms of the shape *baqVr- ~ Somali baqol, bäqlä in Gafat.

no further matches.

Chadic does not distinguish gender in numerals, Egyptian and Berber add -t to qualify fem. nouns, Sem. numerals gain -t when used with masculines (collective) but remain bare when qualifying feminines (thus avoiding -t … -t).

Bibliography

• Semitic Languages: Outline of a Comparative Grammar — Lipiński, Edward
• The Genesis of the Semitic Tense System — Rabin, Chaim
• Case in Cushitic, Semitic and Berber — Sasse, Hans-Jürgen
• A typology of Verbal Derivation in Ethiopian Afro-Asiatic Languages — Fufa, Tolemariam
• Akkadian Verb and Its Semitic Background — Kouwenberg, N. J. C.

  Has new data on PSem. verb system

• Broken" Plural Problem In Arabic And Comparative Semitic : Allomorphy And Analogy In Non-Concatenative Morphology — Ratcliffe, Robert R.
• All around good source for Chadic stuff: Current Progress in Chadic Linguistics : Proceedings of the International Symposium on Chadic Linguistics, Boulder, Colorado, 1–2 May 1987

Books to acquire

• Northeast African Semitic: Lexical Comparisons and Analysis — Hudson, Grover
• Semitic and Afroasiatic: Challenges and Opportunities — Herausgegeben von Edzard, Lutz