by Patrick C. Ryan
(6/18/97)
A. Adpositions (Case Inflection)
1. We make a distinction today between "cases", on the one hand, and "postpositions" and
"prepositions" ("adpositions") on the other, which is totally inappropriate in view of the origins
of case inflection.
2. All case inflection arises out of originally unbound morphemes which existed as discrete
semantic entities before they were tonically attached to nouns.
3. During the preceding stage (class-type), the agents of transitive phrases remained unmarked except by position. The canonical order of elements in a class-type Proto-Language transitive construction was (S)OV, with the transitive agent ((S)ubject) optionally expressed, while OV is reducible to Topic-Comment.
a. An excellent description of what such a language was like is found in Lehmann (1978:223-266) in a chapter by Charles N. Li and Sandra A. Thompson, entitled: "5. An Exploration of Mandarin Chinese".
1) According to them, Mandarin Chinese "is a language in which 'subject' is not a clearly definable notion". And they believe it is best described as a language in which the topic-comment relation rather than the subject-predicate relation plays a major role" (226-227).
4. Before the introduction of demonstratives, which lexically signified the definiteness of a nominal, the Proto-Language method of indicating definiteness for nominals (during the class-type stage) is unknown but Chinese may offer us the pattern for which other evidence may eventually be found in other language-families.
a. In Mandarin Chinese, the Verb is the center of the sentence. Subjects and objects that are "definite" appear in "preverbal position. . . Since topics . . . are definite by definition, they are always preverbal (227)".
1) In addition, "Preverbal time phrases tend to signal punctual time, while postverbal time phrases tend to signal durative time (229)".
b. A similar marking of punctual (definite) and durative (indefinite) time may have an analog in the Afrasian languages where pre-verbal positioning of the subject (?ak-tabu, "I am/was writing", imperfect [indefinite]) may be fulfilling a similar function as opposed to postverbal placement of the subject (katab-tu, "I wrote", perfect [definite]) is doing if we re-interpret the order as signalling (in)definiteness of the object: "I was writing (on) a letter"; "I wrote the letter".(1)
c. In spite of this characteristic of Mandarin Chinese, Adjectives (and Relative Clauses) always precede their Nouns; these are established characteristics of OV languages.
1) In addition, spatial relationships are frequently implied rather than specified by explicit markers (adpositions): e.g.
a) nèikuài tián wômen zhòng daòzi, "that field we grow rice" = "we grow rice on that field".
2) These are all characteristics which derive from the class-type stage of language evolution while
tone as a lexical device originated in the active-type stage.
3) Chinese also illustrates the circumstances out of which adpositions developed. Consider the sentence: ta: pâo huí jia: qù, "he run back home go" = "he ran home". Here, although pâo by itself means "run", it is combined with qù, "go". The reason for this is that "run" could be understood in terms of motion towards or motion away from its object. By adding qù, the natural spatial relationship of which with its object is "to" (cf. Romam eo, "I go to Rome", where the simple accusative designates the goal of "go"), the phrase can be interpreted as "run (to) home".
a) We could achieve much the same effect in two sentences: *ta: pâo huí ("he ran back") and *ta: qù jia: ("he went home").
4. This is especially clear in many African languages, such as Yoruba: ó fi mí rérIn, "he laughed at
me", where the literal translation is: "he put me laughed".
5. What has hardly been realized except by some of the pioneers of IE studies, is that IE is the object of exactly similar processes.
6. Most IEists, in a proud effort to distinguish their language from other "primitive" (particularly non-"nominative-type") languages would seriously resist the idea that the IE dative is -ei is simply the case employment of the verb listed in Pokorny as 1. ei-, "go", in an exactly analogous employment to Mandarin Chinese qù or Yoruba fi, but, unless IE is totally unconnected with the other languages of the world, this origin for the IE dative is certain.
a. It was the achievement of the ergative-type phase that these phrases were standardized and, through interval adjustment, more closely bound to their nouns. While there are a few common case and local ("where? whither? whence?) inflections based on purely nominal compounds, at which we will look below, the most important and frequent case inflections derive from verbs.
1) A particularly illuminating case of this process is found in Sumerian. Sumerologists designate
the Sumerian nominal -e inflection as "locative-terminative"; it is used in sentences like: e-2-e
lugal-bi gu-3 ba-de-2, literally: "house go king its voice [cut-off (finish) = perfect tense]
express" = "its king had spoken to the house".
2) Here, because the phrase gu-3...de-2, "voice express", means "speak", we actually can identify
the "subject" of the verb *e, "go" (e-3 [Jaritz #684+#410]), "go", written as e [Jaritz #574]) for
convenience); what we really have here is two phrases: "its king had expressed" and "'the voice
had gone to the house", collapsed into one.
3) In other words, the object, expressed or implied, is the subject of a verbal phrase with e-3, "go
(to)".
4) This becomes even more interesting when we see that the Sumerian ergative ending is -e also.
b. In ergative languages, the transitive subject has an inflection of some kind, but the direct object usually has no marking, showing its early primacy.
1) Consider the sentence: e-2.ninnu ki Gar-ra, which would mean: "the Eninnu (a kind of temple), which was founded". If no ergative subject is expressed, the English translation is as a passive.
a) But if we add an ergative subject: e-2.ninnu An-ne-2 ki Gar-ra, the translation becomes: "the Eninnu, which An (a god) founded".
b) Now, some readers will think: well, what goes to An here? The Eninnu? Or, if they know
Sumerian, the "place" (ki) since "found" is a composite of ki ("place") and Gar ("make").
c) The noteworthiness of the ergative formation is that it represents a higher level of abstraction. What is "going to" An is not the Eninnu, or the place, or even yet the making, but the entire phrase: "the Eninnu (which) was founded", i.e. the result is being ascribed to An, (the credit or blame) goes to him.
2. In order to become an ergative-type language, no other inflection was necessary since the direct object normally appears without ending (as it does in Sumerian); however, for the possibility of a higher level of abstraction to take place, the lower level of abstraction represented by "gu-3 e-3 e-2" must already have preceded.
a. However, as might be expected, this lower level abstraction of -e was not the only example of this process.
3.We do not find reflexes of Proto-Language HHE ("go out of, leave"), seen in Sumerian e-3, in
Egyptian as a preposition but the Sumerian -ra, dative ("to / for"), is present in the Egyptian
preposition (i)r, "to" (?E-R[H]O). Although PL RO results in Sumerian la, PL R[H]O would be
Sumerian ra; both PL RO and R[H]O would result in Egyptian r.
4. Readers who have been following closely will remember the IE dative -eimentioned above. This
is also 1. ei-, the phonetic result of a dissimilation of HHE-HHE to HHE-$E. Since we have an
attested Egyptian ii, "go", it seems clears that IE -ei is, at origin, the same as Sumerian -e, with
the exception that it represents a reduplicated form. It is only the lamentable state of our
knowledge of Sumerian that prevents a possible full identification since Jaritz #410 + #684 can
also be read as i-10, and Sumerian *i is the possible result of an earlier *ei.
5. Vasconists like R. L. Trask also have, incredibly, refused to recognize the possible relationship
between the Basque dative -i, to which Trask assigns the greatest "antiquity" (1997:201), and IE
dative -ei of the singular, which we can see is clearly related to the Sumerian
locative-terminative -e, and the ergative in -e. It will surprise no one to learn that similar
obstructions prevent Trask from recognizing the relationship of the Basque formative -i, which
forms Basque adjectives, e.g. gorri, "red", and handi, "large", to IE -i, which "is found in
substantives and adjectives (Brugmann 1972:II,278) "; and is also seen in Old Egyptian -i(i),
which "is employed to form adjectives from nouns and prepositions (Gardiner 1973:61)".
6. In fact, the earliest "genitive" (except the direct genitive reprsented by a simple juxtaposition of
Noun+Noun) I can reconstruct utilizes this formant (from PL $E, "-like"), which can be seen in
some western Indo-European languages as the IE alternative genitive -i: and in the Afrasian
genitive in -i(2); this is, of course, not a true genitive but only an adjective being employed as a
genitive. The long vowel of -i:, I ascribe to the effects of the stress-accent.
7.One other formant will complete the picture of the earliest inflections: F[H]A, "go around", "*concern".
a. This is seen most easily in the Japanese topicalizer -wa ("concerning X") but it can be
discovered with a just little more effort in many other languages.
b.Hurrian, a language generally regarded as having Caucasian affinities, has a dative in -wa, which we regard as originally and primarily a dative of interest. This formant for the dative is also easily isolatable in the Urartian dative plural termination -a-wa.
1) In Sumerian, it is also easily recognizable. In view of Sumerian *Ga-2, "I" and *za, "you", the possessive forms -Gu-10, "my", and -zu, "your", are obviously the basal form *Ga-2 and *za with the addition of a -u formant. A phrase like e-2-Gu-10, "my house", is therefore to be more exactly interpreted as "the house that concerns me".
a) The remaining possessive forms of the singular, -a-ni, "his/her", and -bi, "its", are "genitives/adjectives" in -i built on the bases *an, "this one", and ba, "piece, *thing".
2) The only "case" singular inflection known in Egyptian is -w, the ending which occurs in the
"oblique" forms only of the personal pronouns attached to what Egyptologists usually consider
possessive bases: (-)i (with metathesis: wi), "my"; (-)k, "your"; *(-)z, "his/hers".
3) In view of the direct genitive construction in Egyptian, it is to be wondered why Egyptian i/k/*z need to be considered as possessive forms when Noun+Noun is regularly interpreted as X of X so that pr.f, "his house", would be properly "house of he".
a) This is particularly odd for two reasons:
1)) although it is not uncommon to see local case endings added to genitives, it is unusual to see a
"direct object" ending added to a possessive form; and
2)) In addition to other consideration, the pronominal forms in -w are employed as the subjects of stative verbs, a peculiar employment for an accusative.
b) I propose that the Egyptian pronominal forms in -w are, in reality, topicalized nouns formed with reflexes of the primordial topicalizer derived from PL F[H]Aso that a phrase like nfr zw, "he is good", is better understood as "goodness is around the individual" or "concerning the individual, there is goodness".
1)) In transitive sentences like sDm.n zw Hm.f, "His Majesty heard him", the original construction meant something like: "Concerning him, it was whom His Majesty heard".
c) As in well-known, pronominal inflecions are tenaciously conservative. This topical inflection
exists in IE pronouns also in forms like tewe-, "thee", newe-, "us", and sewe, "self", where its
topicalizing function has been obscured.
d) A final faint trace of PL F[H]A in its earliest meaning is seen in the IE locative plural ending
-su, now a transparent combination of the usual -s for plural + *-u, "around".
e) Thus, now we can see that the Afrasian canonical triad of u-nominative, i-genitive, and a-accusative, is simply the end product of an earlier topical -*wa, adjectival -*ya, and an absolutive in -0 (an ending *-a is a figment of the imagination).
8. Tense
a. During the "ergative-type" period, an increasing distinction was being made functionally between nouns and verbs although in the majority of cases, there was no difference in form between them.
1) In the preceding "class-type" period, "verbs" prefixed by reflexes of PL ?E, "that", were probably basically understood as nominal phrases of "that"+ a noun ("hearing/([what is]heard") but in this period the functional difference between nouns used as nouns and nouns used as verbs implies that this formant was interpreted as a tense formant.
8. Summary
a. During this period, at least within the group of speakers who would become Nostratic
(Indo-European and Afrasian among others), phonological changes were taking place that would
prepare those speakers for the next evolutionary plateau of language, the nominative-type.
b. We really have no way to know when Proto-Language consonants began to be accompanied by
significant glides: a palatal glide (y) before PL E; and a velar glide (w) before PL O. Before PL
A, no glide developed.
c. This is, of course, another application of the maxim of Anttila mentioned above: phonological
systems tend toward "maximal differentiation" (Anttila 1972: 186).
d. But this modification of the original phonological system made possible the final separation of nouns and verbs that was the necessary condition for the development of the nominative-type constructions, which we will examine in Part Five.
END OF PART FOUR
continue to (Part
Five)
or
return to The Proto-Language------------------------------------------(Part
Three)
--------------------------------------------(Part
Two)
-------------------------------------------
(Part One)
or
return to INDEX
the latest revision of this document can be found at
HTTP://WWW.GEOCITIES.COM/Athens/Forum/2803/ProtoLanguage-4.htm
Patrick C. Ryan * 9115 West 34th Street - Little Rock, AR 72204-4441 * (501)227-9947
PROTO-LANGUAGE@WorldNet.att.net