SumerianGrammar.htm

This grammar is based on the hypothesis that IE and Sumerian are both descended from a common ancestor, which, I term the Proto-Language — from the form into which it developed between 55-60K BPE.

The date is based on the estimates of Cavalli-Sforza for the separation of the peoples of Asia and Europe (The Great Human Diasporas, p. 123) from the "main" branch of the people speaking the Proto-Language.

During this phase of development, the Proto-Language was passing out of a class-type morphology into an ergative-type morphology (G. A. Klimov).

The word order of Sumerian is — like Basque ( Trask 1997:109) — consistently SOV , what we would expect from any language that preserves early syntax. Although "modifiers overwhelmingly precede their heads" in Basque (genitives and relative clauses; Trask 1997:122) — a further correlate of SOV typology, in Sumerian, genitives and relative clauses follow their referents; however, in Basque, "lexical adjectives follow the nouns they modify (Trask 1997:122)" just as in Sumerian, in which an adjective "stands directly after the noun which it qualifies ( Thomsen 1984:64)".

This discrepancy can be resolved when we realize that a number of Basque adjectives like ilun, "dark", also function as nouns: "darkness", which has led "a number of vasconists to suspect that, at some early stage of the language, there was no distinction between adjectives and nouns (Trask 1997:210)", which Trask admits as a possibility if "at a very remote period". This is certainly the case in Sumerian in which "Adjectives do not differ morphologically from nominal or verbal stems and there are no morphological means to derive adjectives from other stems. An adjectival stem is primarily characterized by its syntactic use . . . (Thomsen 1984:64)".

In certain words, an earlier word order of Adjective/Genitive+Noun seems to be preserved in Sumerian, at least, graphically. The name of the moon-god, Zu-en, 'lord of wisdom' or 'wise lord', is written zu-en. This is the expected word order in a conforming SOV language. I interpret this to mean that when the Sumerians invaded southern Mesopotamia, Sumerian was Adjective/Genitive+Noun, and that a loss of this original syntactic feature is due to contact with the people who were autochthonous in the region, presumably Semitic speakers who preferred the word order of Noun+Adjective/Genitive. Further support for the idea of an earlier Sumerian word order of Adjective/Genitive/Relative-Clause+Noun comes from Abû Salâbîkh (circa 2600 BP), where a passage from a work known as 'Instructions of Šuruppak to his son Ziusudra', reads:

What is enormously exciting about Sumerian is that (unlike Basque and Japanese) it separated from the main branch of the Proto-Language after the stage of development (Pontic) in which the oldest semantic contrasts of CE / CA / CO were replaced by C^yV, C^(-)V, and C^wV, the superscripts indicating semi-consonantal glides or no glide — in keeping with the pattern observed in other Caucasian languages, Sumerian lost the superfluous V before contact with Semitic. The contact with Semitic caused Sumerian C + ^glide to be replaced with vowels (C^y became Ci; C^w became Cu; while C^- became Ca .

Therefore, in open syllables ( in the absence of a following /j/ or /w/), Sumerian preserves a record of and Basque preserves the original vowel quality of the Proto-Language antecedents intact.

In the Table of Correspondence found below after the grammar, the column entitled PROTO-LANGUAGE shows the earliest syllables before vocalic contrasts were replaced by a contrast of glides and no glide (during the Pontic stage: 60-40K BPE).

The interpretation of the Sumerian evidence has unique problems^a. Among which, most of the signs have multiple phonological values; and I have prepared a small series of essays which will explain the significance of these variations.

Current Sumerological practice is to indicate the commonest occurring word of any given form with no diacritical (e.g. mu [**m^yu], 'name'), the next commonest homophonous sign with an acute accent (e.g. mú [**mü₂], 'kindle'), and the third commonest with a grave accent (e.g. mù [**mU₃], 'grind'); and the fourth commonest, with a subscript (e.g. mu₄, 'clothe' [which may actually be, correctly, mur₁₀]).

I will indicate the acute accent by the subscript ₂, and the grave accent by the subscript ₃.

The Proto-Language^b used tone (rising and falling) to indicate the beginning and end of a statement. If the extraordinarily numerous (twelve mu's, e.g.) supposedly homophonous words of Sumerian were not distinguished by tone, for which there is not the slightest evidence, or in some other way, it would have been a most untypical and burdensomely ambiguous language. What is likeliest, is that Sumerian distinguished "homophones" by differences in vowel quality that the Akkadians were unable to maintain after Sumerian ceased to be spoken. The Akkadians were enabled to maintain Sumerian as a literary language because the "homophones" were not written identically, and the different cuneiform signs employed usually predicted the meaning as, in my opinion, they formerly indicated the vowel quality much more precisely.

In addition, modern Sumerian scholars have adopted the highly questionable idea that final consonants — as in mu₄(r₁₀), 'clothe', above — were dropped in certain positions or employments³, compounding the problem of homophonous ambiguity for spoken (but not necessarily written⁴) Sumerian.

Sumerian was not a language that needed greater homophony. In spite of that, however, most Sumerologists believe that "Final consonants are often omitted in the writing"⁵, and, presumably in the pronunciation as well. We shall see, in the appropriate places, that this belief is completely unwarranted; and, it is to Thomsen's credit, that she seems to be acknowledging this by writing: "Because of the uncertainties concerning the actual pronunciation I have, as a principle, rendered the stems with the final consonant: dug₄ (not du₁₁), zid (not zi), níg [nig₂] (not nì [ni₃])"⁶.

1. Sumerologists currently acknowledge the existence of another "case", which they infelicitously call the "locative-terminative", which is, of itself, an oxymoron. Interestingly, the "locative-terminative" is supposedly restricted to inanimates while the "ergative", which has exactly the same form, i.e. -e, is conveniently restricted to animates or personified inanimates as can be seen from the following example:

e₂-e ^(DIG[~]IR)Asar-re šu-si ba-sa₂

Asar put the house in order.



The literal translation is: "With regard to [-e] the house [e₂], (as) Asar [Asar] (was) nearby [-(r)e], (it) had been [ba-] equalized [- sa] completely [šu-] (while) being filled [-si]."

2. Sumerologists currently acknowledge the existence of a demonstrative of the form -e, 'this (**the one nearby, better 'that')', which is rare but seen in: alam-e u₃ kug nu za.gin₃ nu-ga-am₃, 'that [-e] statue [alam] is [am₃] also [-ga- for better -**(n)g³u/û_x⁷] not ever [nu-] lapis lazuli [za.gin] or (again) [u₃] (it is) not ever [nu] silver [kug]."

3. Most Sumerologists currently postulate the existence of a "pronoun" with second person reference, which is used in the chain which precedes the verb stem, admittedly quite rarely, in the form of -e-. However, many older Sumerologists doubted this interpretation of - e-. Deimel 1939 on page 217 writes: "Das Infix -e- findet sich bei allen drei Personen des Singulars." And even the circumspect Thomsen (1984) writes on page 149: "The conclusion must be that there are three pronominal prefixes only for three different 'classes': a) /-e-/ for 2. person (and perhaps also for 1. person), b) /-n-/ for animate; and c) /-b-/ for inanimate. (emphasis added by author)" but then, on page 150, cites a sentence from Enmerkar and Ensuhkešdana 13, which cannot be either 1st or 2nd person reference, verifying Deimel's earlier conclusions:

me.lam₂-bi (...) Aratta^(KI)-a tug₂- gin₇ ba-e-dul, gada-gin₇ ba-e-bur₂

Its radiance covered Aratta like a garment, enveloped it like linen.



This translation neglects to provide any meaning for -e-. And Thomsen adds after it, on page 150: "ba-e-VERB is especially frequent and probably mostly intransitive."

I propose that -e- cannot be reasonably interpreted as an indication of a 1st or 2nd person ergative or absolute agent in view of these facts. We have already seen above that -e has been reasonably interpreted above as 'nearby'. In the sentence just mentioned, a translation of 'nearby' for -e- fits perfectly with the context:

Its radiance had rimmed (around) Aratta nearby (closely) like a garment,
had flowed loosely (around it) nearby like linen.


Why then have so many researchers been led astray? Since 'you' is the second participant of the speech situation, 'you' must always be nearby, at least, at the time of address; and actions that 'you' performs or performed very recently, will often have been performed nearby also. Therefore, the inclusion of an adverb meaning 'nearby' is to be expected in a certain number of cases involved in actions of the speech partner.

Another group of examples illustrates the same point indirectly. According the Thomsen (p. 149), who is passing along the traditional explanation, the pronominal prefix -e- "is written -ù (u₃-) after /mu-/" in the Gudea texts, but then on page 191 quotes the interpretation of this sentence (Gudea, cyl. A VIII 22):

ša₃(better -g₄)-bi nu-mu-u₃-da-zu

"I do not know its meaning".


This is a perfect specimen of the inept translations that Sumerian linguists have foisted off on generations of uncritically acceptive students.


a. The only identifiable "pronominal element" in this sentence is -u₃- (-e- preceded by a glide from the previous -u-), which, however, contrary to what is taught as cited above, is assumed to represent "I" rather than "you"; the better translation is: "Verily (-mu-), there has never (nu-) been a familiarity (-ê- + - zu - 'know nearby') with (-da) the heart (šag₄; "principal motivation or meaning") of it (-bi)."

b. In another phrase, quoted on the same page, nu-u₃-zu is rendered "that he does not know it"; here, it represents "he" (sic!).

c. If u₃ after (m/n)u- can be interpreted as "I", "you", or "he", depending on the context, what value does the statement on p. 149, "The main function of /-e-/ is to denote the 2. person", really have?

d. The only possible reasonable answer is: none!.


It is virtually certain that u₃ was, at some point, read as **wê⁽⁸⁾, which would make the employment of it to render -ê preceded by a **w-glide deriving from the preceding -u- quite understandable.

The conclusion to be reached is that -e(-) should as a grammatical element always should be interpreted as an adverbial or adjectival 'nearby/that', and that the assignment of a second person significance to -e- should be forthwith abandoned as totally unjustified.

4. This element is seen in yet another application: as the verbal prefix e- in Old Sumerian which frequently but not always takes the place of the later verbal prefix i₃-, written occasionally also i-.

Much ink has been spilled on the significance of this prefix (e-/i₃-). Thomsen admits: "It is difficult to attribute any characteristic function or meaning to the prefix /i[~]-/." Notice that one of the idlest speculations concerning the prefix has been that, although nasal vowels are not attested elsewhere in Sumerian, this initial i₃- was supposedly nasalized (/*i[~]/).

This professorial fantasy was concocted to attempt to "explain" the circumstance that the sequence i₃-/i-/e-ba- does not occur in Sumerian when some thought it should. However, i₃-/i(m)-/e-ma- does occur; and to "explain" this, some one came up with the idea that i₃-/i(m)-/e-ma- developed out of i[~]₃-/i[~]-/e[~]-ba- caused by the unique nasalized vowel.

The correct explanation is that i₃-/i-/[e]- is a prefix that indicates non-concomitant actions, i.e. [closely] sequential actions, either past or future, and that it is not combined with -ba- because this prefix indicates the perfect, completed action before the action of subsequent verbs, and the addition of i₃-/i-/e- would be redundant and unnecessary. Translating e- by 'then' or 'thereupon' would do justice to the nuance intended⁹.

The better interpretation of the element -(m)-m(V-) in the sequence i₃- /i(m)-/e-ma- is, in view of Jaritz Sign #639, ma, 'land', rather straightforward: -ma- (better -**mâ-) simply means 'there ('at that place yonder')', the place of the non- participant in the speech situation. This analysis has been supported by the conclusions of Black (1991:8): "2) The texts reveal the unmistakeable existence of a verbal element -m- with by-forms (-)ma- and (-)mu-, and this (-)mu- is clearly distinct from the conjugation prefix mu-/mi-."

The category of 'person' does not exist in the Sumerian verb, which is not finite in the narrow sense of the word. All Sumerian verbs are formally participles, which are indifferent to person (1st/2nd/3rd), number (singular/plural), voice (active/passive), time (immediate [a- ]/ proximate [e-]/ distal [i-]), mode (volitive [ga-]/ necessitative [ha-{better ku₆}]), aspect (perfect [ba-]/ perfective [al-]/ perfective punctual [ -n-]/ imperfective [-**î]/ imperfective partitive [-b-]/ imperfective iterative [-*d(u)]), and others. To compensate for this, many adverbial elements were employed to specify the relationship of the action to the participants in the speech situation.


a. ma, 'there'

It is to Thomsen's credit that she recognized the essential validity of an assignment of meaning to -m- as a spatial adverb; she writes on page 173: "The theory which has the most arguments in its favour, I think, is that /-m-/ is a ventive element (this was put forward by M. Yoshikawa)". Ventive is a technical term usually implying 'movement to a place' but I believe that a simple translation of 'there' is more accurate.


e₂-a ^(G[~]IŠ)hur-bi im- g[~]a₂(r_x)- g[~]a₂(r_x)

Then [i-] (he) laid out [g[~]a₂(r_x)- g[~]a₂(r_x)] this [-a] house [e₂] there [-m-] according to [-bi] the plan [^(G[~]IŠ)hur].





ud (better ul₆) ^(DIG[~]IR)Nin.g[~]ir₂.su-ke₄ uru-ni- še₃(better -ši₄)
igi zid im-ši-bar-ra

Then [i-] with regard to [-(k)ê₍₄₎] the Lady (Nin), the possessor (of) [-k-] Girsu [g[~]ir₂.su; 'sewing needle'], at [-(r)a] the time [ud] (of the) opening [bar; 'split(ting)'] (of) a steady [zid] eye [igi] toward [-ši-] there [-m-], toward [- še₃] her [-ni-] city [uru].



One might more fluently translate: "Then as Ningirsu directed a steady eye thither, toward her city . . ."

This element, ma, 'there', can appear in the verb chain in three forms: as -m, due to the stress-accent of an initial element, such as i- (-im-ši-, 'toward there'; -im-ta-, 'through there, thence'); as -mâ- (written -ma-), the result of a contraction of ma + (?)a, the "locative"; or as (-)mî- (written (-)mi- or (-)mi-₃), the result of a contraction of ma + î (written -i in combination, derived from i₁₀, 'go(ing) (to)', which is usually transcribed e₃), the allative", as in im-mî-sar, 'then he wrote thereto/thither'.





ma, 'there'	+-Ø, locative	+-a, "locative"	+-ta, "ablative"	+-î, allative
	m(a)A, 'at there'	mâB, 'on there'	m(a)-taC, 'from there'	mîD, 'to there, thither'



The employment of these morphemes is not normally superfluous or redundant. They are intended to modify the verb in ways the following examples will make clear:


A. lu₂ na(better ne₆).me inim nu-um-g[~]a₂(r_x)- g[~]a₂(r_x)
man not-say word never-(at)there-place-place

Never shall any man lodge a word (claim)

B. im-mâ-g[~]en(better g[~]in)
then-(on)there-go

Then (he) reached there.

C. lu₂ E₂.ninnu-ta im-ta-ab-e₃-e₃-a
man E-ninnu-from then-there-from-some-go-out-go-out-this

The man that then shall remove something from there, from the E-ninnu . . .

D. na gal-gal-bi lagab-a mî-ni-dê₆
stone-large-large-its boulder-at thither-into-bring

(He) deposited (by rolling) thither all the large stones for it as boulders.


We can also see ma, 'there', in combination as mê (written me), 'where', the result of a contraction of ma + -**î, '-like'.


We reconstruct Proto-Language ?A- ¿E, 'forehead-like = near(by)', for this formant, which will henceforward be indicated as -ê, the anticipated result of early Sumerian **(?)aj: e₂-ê ^(DIG[~]IR)Asar-rê . . ..

Faint traces of this morpheme, which would appear in Indo-European as **H₂ei- or *ai-, can be seen in IE 3. *ai-, 'give, allot, (**place near)'; and, in combinations like Old Indian i-há, 'here', from **ai-, '**near(by)', rather than from *ei-.

1. A verbal prefix in the form of a- is discussed in Thomsen (pp. 155-69). Based on the same faulty reasoning that emended i₃- to a non-existent /iⁿ-/, a- has been re-interpreted as a non-existent /aⁿ-/. For the reasons discussed above, I reject this as preposterous.

In the examples cited for a-, I do not think it will difficult for most readers to see that an interpretation of 'immediate present' for a- fits the context well:

kug ša₃(better -g₄)-g[~]â₂ a-šag₅-ga

(The) silver that [-**(g)a] now [a-] is sweet [šag₅] on [-(-g[~])u₁₀ + -a = (- g[~]â₂] my [-**g[~]u₁₀] heart [ša₃(better -g₄)] {'the silver that pleases me now' . . .}




tukumbi [ŠU.NIG₂.TUR.LA₂.BI] dur(better: tûr₂) ab-tu.lu

If [tukumbi] the rope [dur- (better: tûr₂)] now [a-] is somewhat [-b-] loose [-tu.lu; probably better 'swollen, frayed'] . . .


And certainly, the imperative use of suffixed verbal "prefix" -a is persuasive:

e₂-a-ni gul-a

Destroy his house immediately!


2. as an adverbial element of the verbal chain setting the stage of the action 'in the immediate presence': in the same misguided effort that identified -ê as an indication of the second person, -a-, 'here', has been misidentified as a first person element (Thomsen 1984:149):

šu zid ma-ra-a-g[~]ar

(With) a steady [zid] hand [šu], there [ma-] has been a setting [-g[~]ar] up[-ra-]on [-a-]..


This may be compared to the proffered translation: "I have faithfully performed it for you", in which -a- is supposed to represent "I", an idea that ultimately goes back to the flawed 1923 work of Arno Poebel.

But, once the door to a pedantic playpen was opened by the theory of a loose kind of vowel harmony in Sumerian, even sillier learned expositions have been vainly enunciated¹¹.

a₂ še mu-e-da-a-a-ag[~]₂

(A) single [še] arm [a₂] has verily [mu-] been bent [ag[~]₂] here [-a-(ag[~]₂)] on [-a-(-a-)] (the) side [-da-] (of) that (nearby) [-e-].

= (loosely) Verily, a single instruction ('bend the arm = point out') has been given here in connection with that.


Here, Thomsen quoting Gragg, analyzes that e-da means "me" (!) whereas it had been maintained that -e- meant "you"; and, at least, one of the -a-'s, which is supposed to mean nothing in this position (but previously suggested as "I"), is now suggested to mean "you" though it is normally -e-. I personally do not believe that any rational student of Sumerology can feel comfortable with this tortured rhetoric to justify a fundamentally false theory.

3. as a noun ('immediate presence') modified by the preceding noun, construed as a suffix to indicate immediate presence; this is the "locative" identified by Sumerologists but it is properly a superessive, 'on . . .'.


a. "Genitive" In the same manner as -ê has a generalized use as a proximative, and a specialized use, mostly with animate nouns as the "ergative", the superessive (-a) has a generalized use with mostly inanimate nouns ("locative") and a specialized use with animate nouns as a "genitive", as in e₂ lugal-la, 'the house on which the king is' = 'the king's house'.

Sumerologists usually assert that the Sumerian genitive is -ak but falsely maintain that the final -k "is deleted in final position and is only written when followed by a vowel (Thomsen 1984: 90)". That is not to say that all Sumerologists have held this belief. To Thomsen's credit, she quotes, on the same page, Edmond Sollberger: "Je pense à mon tour que la forme du suffixe est bien -a", an opinion with which we agree.

In order to accomodate this incorrect theory, the possible relationship between the supposed suffix -ak and the verb ak(a) (better **aku to judge by the Akkadian designation of the sign¹²), 'possess', has been suppressed because if the form of the suffix were **-aka/u rather than - ak, there would be no reason for us not to be able to see the final -k, which would be followed by a vowel.

A circumlocution for the genitive is found which, I believe, is explained much more satisfactorily if we assume an adjectival meaning of 'this' for -a.

e₂-a ig-bi

this [-a] house [e₂] its [-bi] door [ig- ] versus of [-a] (the) house [e₂], its [-bi] door [ig-].

= (loosely) the door of this house.


However, there are, of course, certain genitive constructions in which -ak(a/u) exists. We explain these by suggesting that since the normal "genitive" was -a, additional added elements beginning with a vowel would have obscured the semantic integrity of the genitive. For example, if we added "locative" -a to e₂ lugal-la, 'the house of the king', the resulting form (-**â) to e₂ lugal-la/**â, 'at the house of the king', would have only minimally different from e₂ lugal-la, the plain "genitive"; and that is only if we presume that -la + -a would have resulted in an actually lengthened -**lâ.

Therefore, I believe the Sumerians elaborated the "genitive" construction with -ak(a/u), 'possession', when, for the sake of clarity, the normal genitive (-a) might be ambiguous: **e₂ lugal-(l)ak(a/u)- therefore means 'the house, the kingly possession'.


4. as an adjectival suffix meaning 'this': on page 64, Thomsen almost reaches the conclusion that I have: that -a is an adjectival suffix meaning 'this' ('here'): "inim dug₄-ga 'the word which has been spoken', i.e. not any word but this particular word spoken (underline added)" (better: 'the word, this spoken thing') but above that, she rather innocuously states: "Some adjectives occur always, others occasionally with the suffix /-a/". Those adjectives that have -a should be translated as 'this (adjective) (noun)'.

A comparatively frequent use of -a, 'this', is found after participles which function as relative clauses: inim An-ne₂ dug₄-ga, 'this word which An spoke' ('with regard to An, this word spoken'). Thomsen seeks to make a distinction between this, which she terms a "subordinate construction"; and inim An-ne₂ in-dug₄-ga, which she terms a "finite relative clause"; but the only real difference between the two virtual relative clauses is that the second includes additional verbal qualifications: i₃-, 'then', and -n-, which makes the verbal action punctual: 'this word (of) which An then made an utterance'.

In the same way that adjectives can vary as to whether they are qualified by 'this' (-a) or not, virtual relative clauses can also be seen without -a: dig[~]ir sag[~] zig, 'a god who lifted the head'. Similarly, with verbal forms that are imperfective: ^(DIG[~]IR)En.ki . . . di (better: dê) pad₃(better: pat_x)-de₃(better **tî_x [**{pa}ti + î]), '^(GOD)En-ki . . . who reveals/shall reveal the portions (make a legal determination of the division of interests).' For no reason that can presently be conclusively demonstrated, an equivalently determined form with -a does not seem to be attested after an imperfective verb form in -î; this is perhaps because one of the functions of -a is to create an adjectival phrase, and with -î, an adjective has already been formed from the verb.



	a 'immediate presence'	verbal	adverbial	nominal	adjectival
		a-, immediate present 'now'	-a-, 'here'	-a, superessive, 'on' "locative" "genitive"	-a, 'this'

continue to

go to first 35+ root cognates (1-35) ?

PL MORPHOLOGICAL ELEMENTS IN SUMERIAN

press here to see

to investigate these phonological correspondences in detail, see the

NOTATIONAL CONVENTIONS

Combinatory Modifications

for modifications of the vowels and consonants in combination, see the

Table of Modifications

b. I am aware that some readers, who may entertain the possibility of a Pontic-Nostratic connection between Indo-European, Afrasian, and Sumerian, will not be able to accept the possibility of a reconstruction of a language as early as the Proto-Language. To those readers, may I suggest that the Proto-Language reconstructions and interpretations be merely regarded as an expression of an arbitrary system of notation that allows for the regular relationships of correspondence between Indo-European, Afrasian, and Sumerian?

The comments relating Sumerian grammatical forms to prototypes in the Proto- Language will be highlighted in green so that they may passed over by those whose interest is solely in Sumerian grammar.

1. Thomsen 1984, pp. 21.

2. Thomsen 1984, pp. 39-40.

3. Thomsen 1984, p. 42.

4. Although two phonological forms for "enter" are postulated, ku₄ and kur₉, both are written with the same sign: J. 99.

A better explanation of the phenomenon that final consonants are variationally resumed as an initial onset in a following syllable that begins with a vowel is that the notation records a shift in stress-accent.

5. Thomsen 1984, p. 42.

6. Thomsen 1984, p. 42.

7. Jaritz 1967, Sign #594, the sign with which ga is written, has the recorded Akkadian reading gu(- u), which would be the anticipated Sumerian reflex of PL QO, 'attach(ed), also', which we believe is the basis for this element as evidenced in other languages.

8. Jaritz Sign #805, with which u₃ is written, is graphically composed of the signs for 'eye' and 'cord', suggesting a basal meaning like 'the eye is drawn to'. One of its meanings is 'cry of pain', **wê, which I believe represents the Sumerian reflex of PL F[H]A-¿E, 'wail-ing', which is also seen in IE *wai, 'interjection "woe"'; and AA **w-y as exemplified by Arabic wai(H/ha), 'interjection of pity or threat'. In addition, Jaritz Combination #3058, u₃.di, is glossed as Akkadian kûru, 'woe, trouble, misery'; I believe this correlates with IE **waid- in Middle Irish fa:ed, fo:id, 'cry', and Latvian vai[~]di, '(pl.) cry of distress, crying, trouble'.

Interestingly, another word, derived from PL FA-¿E, 'leaf-like=fragile', seen in IE *wa/a:i-, 'weak, miserable', is also present among the meanings listed for Sumerian u₃ but assigned to the reading si₅ (cf. IE *sa:i-, 'pain, sickness, damage'; here we have an alternate reading based on semantic similarity).

The conclusions I draw from these considerations are that Sumerian u₃ originally represented **wê, and that the basal meaning of the sign originated with this meaning.

9. If my derivation for ê from **(?)aj is correct, i.e. from PL ?A-¿E, it is legitimate to ask what would have been the earliest and most appropriate sign to represent it. I believe that question can be answered. Jaritz Sign #599 depicts a wall-mat for a house, and means 'house'; it has the reading ?e₃ however is usually rendered as e₂, showing that the glottal stop (/?/) was lost over time. This may represent PL ?A-¿E, 'grass-like (thing)' or PL ?A-¿E, 'family-like thing' — either of which could lead to 'house'.

Presumably, at the time that writing is first recorded, the glottal stop had been lost except after a vowel, so that initial ?ê-, the non-concomitant verbal prefix ('then, thereupon'), and -?ê, the demonstrative ('this nearby'), at least after a verb ending in a consonant, were pronounced ê, which carried adverbial -?ê- ('there nearby') with them. When Sign #599 acquired a commoner reading of **bêd for 'tent' (PL P[?]A-¿E-T[?]O), the way was clear for a sign which consistently read ê to represent earlier ?ê

This sign was Jaritz Sign #574, which depicts a canal, and means 'canal' or 'water'. This derives from PL HHA-¿E, 'water-like', which probably went through the stage of /**he:/ before becoming /e:/, i.e. ê, presently transcribed e.

It is possible that i/î might represent a later purely phonological development from *(?)aj, perhaps in a dialect other than that which had (?)ê but what I consider to be more likely is that /*?i/ is from PL ?E, 'that yonder, then, over there', the base for which is seen in Jaritz Sign #15, i₅, 'tooth'. When Sign #15 began to be read more frequently as ka, another sign that preferred the reading of /i/i:/ was substituted for it.

Now i₃, Jaritz Sign #456, the primary meaning of which appears to be an 'oilpress', I connect with PL ¿O, 'squeeze'. Although *iu is not an acknowledged reading of Sign #456, Akkadian records a reading of ia?u, which I believe should be corrected to *iu, the anticipated Sumerian result of ¿O. In its later manifestation, i₃ will almost certainly have represented î₍₃₎.

On the other hand, i is Jaritz Sign #270; it has the primary meaning of 'five', which I connect with PL ¿A, 'many', which would have had the anticipated Sumerian reflex of î/ia; I believe that the employment of this sign is purely phonological.

It is highly probable that the earlier reading of Jaritz Sign #456 was *iu and the earlier reading of Jaritz Sign #270 was *ia; it was only after both had been simplified phonologically to /i:/ that they become substitutes for earlier /?i/ , which, apparently carried a stress-accent.

10. Since the formant for the allative is, according to my view, -**î from i₁₀, 'go (to)', which forms a phrase with its target, if 'heaven' were simply an, we would expect **an-î for '(going) to heaven'; however, we find an-ne₂, suggesting that an should be read as ana₂, an attested alternative reading for the sign, since -e/ê should be the result of a + i/î.

A further correlate of this is that the time of the first use of this formant, at least many Sumerian nouns retained a final vowel so that it was regularized as -e rather than -i.

This appears to be the case with the imperfective aspect of verbs as well. The ending - e(n), which is termed the "present-future" by Sumerologists, is, we shall see, a result of a + î, of which the î is the phonological reflex to PL -¿E.

11. According again to G. B. Gragg, as summarized by Thomsen on p. 149-50: "The subject mark -e- in preradical position has changed to -a- because of the preceding -da-; the second -a- probably denotes some kind of transition between the pronominal prefix and the initial vowel of the verb, cf. Gragg SDI p. 85."

12. In my opinion, one of the earliest terms for 'possession' is based on PL HHA-KX[H]O, 'water-mollusc=sea-shell', which would have the Sumerian reflex aku; with a slight shift of meaning, this can also be seen in Egyptian jx.t, 'property, possessions'; and in IE *e:ik- (based on Anglo-Saxon a:gan, I believe it is possible that a related reconstruction for IE in the form of **a:k- may be justifiable), 'own, dispose over'. Obviously, this interpretation is based on the idea that some of the earliest (valuable) non-utilitarian possessions were sea-shells. Hittite aku, 'sea-shell(?)', is also of interest in this context.