Reconsidering prepositions and Case assignment in the text of Revelation 4 and 5 S

In order to describe the government by prepositions in the book of Revelation in terms of the Government and Binding Theory, it is imperative that the sub-theory of Case assignment be considered. With the latter as point of departure one may describe, i) the shifts from autothematic and structural Case to oblique Case, ii) the use of prepositions with oblique Case instead of the structural genitive Case, and( iii) the peculiarities of the Case and case assignment of the preposition evpi, as found in Revelation 4 and 5.


INTRODUCTION
The use of cases with regard to the government by prepositions forms an integral part of the remarkable and peculiar Greek featuring as the language of Revelation (cf Bousset [1906Bousset [ ] 1966:164-8;:164-8;Charles 1915:86;Mussies 1980:167;Dougherty 1992:10;Riekert 1996;2003).Musser (1992:1) mentions a number of suggestions advanced in the past to describe the peculiar grammatical usage of Revelation.Riekert (1996) departs from the hypothesis that the morphological cases in Revelation 4 and 5 can be adequately described and explained within the framework of the Case sub-theory of the Government-Binding Theory.Therefore he presents an analysis of those two chapters regarding the abstract Case assignment and the morphological realisations in terms of proposals made within the Government-Binding Theory and the application of these proposals to New Testament Greek (cf Riekert 1985;1996;2003).This study of Riekert (1996) shows that in terms of abstract Case assignment no irregularities could be found, at all.When the so-called "foreign" usage of morphological case in Revelation is considered within the framework of the Case sub-theory of the Government-Binding (GB) Theory of Chomsky, one should be in a better position to decide whether the "foreign" usage is quite so "foreign".
Bearing in mind the developments in the realisations of the morphophonological cases from the usual in the rest of the New Testament to those in the book of Revelation (cf Charles [1920] 1971: xxxviii-cxi), one should realise that the deviations reflect only change and development within the Greek language which can easily be explained within the framework of the Case sub-theory of the Government-Binding theory (cf Riekert 1996).
Nevertheless we may comment on the following developments.

SHIFTS FROM AUTOTHEMATIC AND STRUCTURAL CASE TO OBLIQUE CASE
We may take them together as shifts from autothematic and structural Case to oblique Case.The use of prepositions in this regard calls for a more detailed discussion.Riekert (1996:128) states with clarity that instrument and manner may be expressed by means of the so-called dative of instrument and dative of manner, but also by means of a preposition and its governed NP.It is remarkable that ev n imati, oij leukoi/ j (cf Riekert 1996:71), in Revelation 4:4, is used as a prepositional phrase expressing instrument.In (1) we have the following analysis:

Autothematic Case and prepositions with oblique Case
( For the analysis of the ev pi, -phrase, see (4)(a).The omitted verb ei= don assigned to the object ei; kosi ... peribeblhme, nouj objective Case, of course realised as accusative.The participle peribeblhme, noujdescribed by Dougherty (1992:330) as a "circumstantial participle" -governs a PP in which P ev n assigns oblique Case to imati, oij leukoi/ j, realised as dative, according to Dougherty (1992:150) as a description of manner.
Charles ([1920] 1971:cxxxix) follows a reading which has the dative imati, oij leukoi/ j without the ev n and comments that the dative of instrument "is mostly replaced in our author by ev n".
(2) (a)  (cf Brütsch 1970:264).We assume that after hv go, rasaj there is an NP-trace which may be recovered after deletion, according to Rienecker (1966:616) as object from the subsequent passage.The NP tw/ | qew/ | shows in the analysis a theta role, governed by the verb, and the dative case is a realisation of inherent Case (cf Dougherty 1992:143,145-6, 156).The preposition ev n governs the NP tw/ | ai[ mati and assigns to it dative case as realisation of the oblique Case.
That ev n governs an NP here which receives a theta role of instrument or of price (cf respectively Blass, Debrunner & Funk 1961:118;Charles [1920] 1971:147), does not affect Case assignment, however the latter NP governs in its turn again the NP sou and assigns to it structural genitive Case.

Structural genitive Case and prepositions with oblique Case
Another confirmation of the prevalence of the use of prepositions is seen in (3).The circumscription ev k tw/ n presbute, rwn (Rev 5:5, cf Riekert 1996:99-100) is an equivalent for a (partitive) genitive, or ev k could be considered as a kind of marker of the genitive (cf Charles [1920]  The NP ei-j is the head of the NP-phrase which serves as subject of the verb le, gei and which is coindexed with the INFL element of the verb.Therefore ei-j receives nominative case and Case (cf Dougherty 1992:74,266 who draws attention to the fact that ei-j is semantically equivalent to an indefinite pronoun).The preposition ev k governs the NP tw/ n presbute, rwn and assigns therefore genitive case which realises Oblique Case.The verb le, gei governs the pronoun moi which fills a theta role position (cf Dougherty 1992:140, 155) and assigns dative case to it as realisation of the inherent Case.

THE PREPOSITION EVPI, WITH SHIFTS OF CASES IN THE REALISATION OF OBLIQUE CASE
Charles ([1920] 1971:cxlii) quotes Moulton who refers to "uncertain use of cases" in Revelations.One such instance is the government of the preposition ev pi, .This preposition is used with different cases and all of them indicate the place where.In Revelation 4 and 5 it is conspicuous that ev pi, governs all three cases of the oblique Case in the sence of on, at (positionally in a place).Even more conspicuous is the use of ev pi, with forms of qro, noj (=throne) as prepositional phrase following a participle of kaqh/ sqai (=sit) (cf Charles [1920Charles [ ] 1971:112;:112;Riekert 1993:69-71).The latter instances are analysed in (4).
(4) (a) (ii) Rev 4:4, two elements selected in the original order.The expletive element iv dou, should be considered as omitted and the functioning is as in Riekert (1996:61-65).The NP (a; nqrwpoj) kaqh, menoj( has moved for the sake of emphasis with deletion of a; nqrwpoj, or it may be a substantiated participle without the article (cf Dougherty 1992:76).The Case assignment to the NP is with coindexation to the expletive element.The realisation of the case of PP ev pi.to.n qro, non draws the attention: the Case assignment is oblique, but the realisation as accusative deviates from that of classical Greek (cf Swete [1908Swete [ ] 1968:67;:67;Charles [1920]  ev pi, which realised its oblique Case as accusative in sentences ( 6) and ( 10) also features in this sentence.In this instance the assignment of oblique Case is problematic.The presupposition by oblique Case as inherent Case is that the assignment would be dependent on the inherent characteristics of the Case assigner.In all these instances ev pi, has the same meaning.We have therefore no justification for more than one case realisation to be found, cf Riekert 1996:132-146.The prepositions eiv j and ev nw, pion also assign oblique Case, respectively realised as accusative (tou.j aiv w/ naj) and genitive (tou/ kaqhme, nou), according to the inherent characteristics of the Case assigners.
The NP tou.j aiv w/ naj also governs the NP tw/ n aiv w, nwn( and assigns to it structural genitive Case and case, all according to the principles of Greek Case assignment, (cf Dougherty 1992:124-5) even if the expression reflects a Hebrew and Aramaic superlative construction.
Only the variant reading according to (4) (c) (iii) disturbs the pattern: in this case ev pi, governs a NP in the genitive case following a dative participle.
It seems that we have here assimilation, or rather assonance, of the case governed by ev pi, to the case of the participle of kaqh/ sqai.Nevertheless, the assimilation is limited, for the nominative participle preceding ev pi, , fails to bring about that the case following ev pi, be nominative, but indeed accusative.
That should be ascribed to the rules regulating Case assignment.Nominative case cannot realise an oblique Case assignment (cf Riekert 1993:70).Assimilation-assonance however cannot completely explain the use of different case of the word qro, noj.
The Case theory determines that the preposition ev pi, just like all the other prepositions assigns oblique Case.The question is now how would it be realised in terms of cases.All three of those cases which realise oblique and inherent Cases, could be considered in principle.The fact that the nominative participle has a different case following ev pi, and breaks the pattern of correspondence between the case of the participle and the case of the NP following ev pi, , confirms that the Case assignment rules are obeyed here.Riekert (1996:136-141) analyses combinations of kaqh/ sqai as well as other verbs with ev pi, where the formulation of ( 5) is not valid.
The instances in (6) come into consideration.
(6) (a) (i) Revelation 4:4 kai.The preposition ev pi, governs its NP ta.j kefala, j and assigns to it oblique Case realised as accusative; the NP ta.j kefala, j in its turn the NP auv tw/ n and assigns to it structural genitive Case and case.The omitted verb ei= don realises as accusative the assigned objective Case of its object stefa, nouj crusou/ j (cf Riekert 1996:71, 72), with the PP ev pi ..... auv tw/ n as a free adjunct to the omitted verb.
( Because of the length of the NP pa/ n ..... auv toi/ j pa, nta which moved in front of the verb to take the topicalisation position, the construction of the sentence is not so obvious. The verb h; kousa governs the NP-construction above and assigns to the heads of the recursive composed NP, viz. to pa/ n kti, sma and ta, , objective Case realised as accusative case (cf Riekert 1996: 64;Dougherty 1992:127-8).
The DET ta, is accentuated by means of the PP ev n which governs the NP auv toi/ j and assigns to it dative case as realisation of the oblique Case.Lohmeyer (1953:57) is of the opinion that ta.ev n auv toi/ j resumes the last three nouns together, thus a resumption of the detailed description of pa/ n kti, sma now from the viewpoint of a totality, or as Bratcher (1984:53) formulates it "the whole universe", according to Bousset ([1906] 1966:262, cf Charles [1920] 1971:150) in terms of a quatro-partition.The fact that the whole is meant seems to be confirmed by the use of pa, nta as an adjective to strenghten the resumption.Charles ([1920Charles ([ ] 1971:136) :136) shows that pa, nta follows its noun only in two other instances in Revelation.Dougherty (1992:220, 225) considers pa, nta as a substantive adjective.In this case we have a different construction and AP with pa, nta changes to NP.It is also part of the chain with (pa/ n kti, sma) ta, and le, gontaj and it has the Case assignment in common with the rest.
The verb av kou, w is sub-categorised to take the participle together with the object and in this case le, gontaj which then like pa/ n kti, sma and ta, have accusative case as realisation of the objective Case which is also a constructio ad sensum (cf Charles [1920Charles [ ] 1971: cxlii;: cxlii;Dougherty 1992:343).Consequently the agreement with regard to number and case and Case is uncomplicated (cf Dougherty 1992:332, 341-2), however with regard to grammatical gender there is a problem, which becomes obvious from the text critical apparatus reflecting attempts to change le, gontaj to neuter plural in agreement with ta, .
(6) (c) (ii) Rev 5:13 ev pi.th/ j qala, sshj on the sea genitive The preposition governs th/ j gh/ j and assigns to it genitive case which is a realisation of the oblique Case.Blass, Debrunner & Funk (1961:96) correctly assume that the PP ev pi th/ j gh/ j is not a substitute for a direct object -it is in fact a free adjunct.
All the principles of case realisations which are gleaned from (4)(a)(i) to (4)(c)(iii), seem to be valid here.The explanation of the assimilation of the cases of qro, noj (=throne) would rather be sought in the rhetorical situation.
The word qro, noj with the participal construction of kaqh/ sqai ev pi, serves as an indication of God being the One who holds sway over everything.Therefore Lohmeyer (1953:45)  In all the instances supplied in (4) ev pi, has the meaning of on, at in the meaning of place where something rests, be positionally.From the perspective of language economy it is problematic to have three cases to serve one function.Charles ([1920Charles ([ ] 1971:112-3;:112-3;cf Swete [1908cf Swete [ ] 1968:67;:67;Riekert 1996:67) draws attention to the efforts made to find different meanings in the usage of three cases, regrettably with scant success, should we ignore the mere listing of cases.
The definition of Liddell e a (1968:621-2) of the meaning of ev pi, and usage, to express locality, may be expessed as in ( 7).
(7) (a) ev pi, with genitive: locality with verbs of resting, or with verbs of movement where the subject is resting on something (like on a horse or wagon): on, with at.A number of idiomatic expressions are only used with the genitive.
(b) ev pi, with dative: locality, similar in meaning to (a). (c) ev pi, with accusative; locality, with verbs of movement, unto an object or place, or to a higher object, place or in hostile sense: to; unto; against.The conclusion (8) may now be drawn: (8) (a) ev pi, in the sense of on, at might take in classical Greek its case in either genitive or dative to realise the oblique Case, and in this regard there was always the freedom of stylistic choice between genitive and dative.
(b) ev pi, with its NP in accusative case in the sense of on, at shows a shift in the case realisation which should be investigated in more detial.Charles ([1920] 1971:191) would like to read the phrase (10) (a) to (c) in the same way when he says: "We should expect either accusatives throughout or genitives".He atributes the alteration to the fact that the author uses certain characteristic phrases.It is possible, but this instance may serve as tangible proof that the distinction between ev pi, with accusative is still maintained (cf Mounce 1977:165-6;166N3;Riekert 1993:73).The translations supplied, present such an understanding for (10)(c) for which we have a parallel in (11).
(11) Rev 3:3 h[ xw ev pi.se, Å I will come against you accusative Riekert (1996:144-5) discusses the uses of ev pi, + genitive and accusative in Rev 7:1 and Rev 3:3.Which leads to the conclusion: Therefore we may assume that the author offers us a subtle distinction here: the wind will blow on the earth and on the sea but against the trees.In this case: (10)(a) to (c) presents a demonstration that the distinction between ev pi, with accusative and ev pi, with the genitive still exist for the author of Revelation.Please note that the change goes in one direction only: the accusative gain usage and assumes usage from the genitive.

CONCLUSION
The question arises as to how we should account for the facts in (4) to ( 11) with the Case theory.In the classical period there it is settled that Case assignment by prepositions was oblique.This oblique Case was potentially realised in three different cases and these cases were inevitably determined by the meaning which should be expressed.The instances in (4) to (9) in contrast with (10) demonstrates that as far as it concerns ev pi, , the distinction between ev pi, with accusative and ev pi, with genitive or dative was on the wane.Furthermore, this indicates that the accusative acquires greater use and it reflects a development similar to the usage of prepositions in English which prevails upon Chomsky (1981:49, 50, 292-3) to conclude that English has the marked characteristic that a P assigns structural objective Case.
Should one assume that Case in Revelation is still assigned obliquely after prepositions, then it follows that an N can be taken from the lexicon in either accusative, genitive or dative case.The Case control filter is indeed functional, but it passes any one of the three forms as realisation of the oblique Case, because of the developments discussed above.In the phase preceding Revelation, the filter has prevented the accusative with ev pi. in the sense of on, at (positional in a place).When the genitive and dative in a subsequent development, were in the process of becoming obsolete, then only would the Case control filter start to prevent them.ev pi, with accusative was already gaining usage at the expense of ev pi, with genitive or dative.The distinction in meaning between ev pi, with accusative and ev pi, with genitive or dative was already disappearing in Revelation.
Otherwise it can be accepted withThompson (1985:88)that they have been generated under the influence of the Hebrew syntactic rules in that position, without any influence on Case assignment.The recursive composed NP do, xan kai.timh.nkai.euvcaristi,an is generated by the verb dw, sousin (from di, dwmi) and it assigns to it objective Case which is realised as accusative (cf Dougherty 1992:187).The NP tw/ | kaqhme, nw| is also governed by the verb dw, sousin which assigns inherent Case, realised as dative in a theta-grid position(cf Dougherty   1992:140,156,179).The NP tw/ | zw/ nti is in apposition with the NP tw/ | kaqhme, nw| and it is antecedent-governed by it and receives the same Case and case by coindexation.As substantiated participles the NP's tw/ | kaqhme, nw| and tw/ | zw/ nti have the possibility to take free adjuncts in the form of prepositional phrases(cf Dougherty 1992:396).The preposition ev pi, governs in the first instance the NP tw/ | qro, nw and in the second instance the NP tou/ qro, nou; both NP's are affected by oblique Case assignment, but the two assignments are respectively realised as dative and genitive.The preposition The NPs ta.zw/ | a and oi` ei; kosi te, ssarej presbu, teroi are generated in the SPEC-or subject position of INFL.These NPs are coindexed with INFL and therefore receive structural nominative Case and case; they are at a later stage moved to a position to the right of the verb (cf Dougherty 1992:392).We may note here that INFL is marked plural where Attic Greek would prefer singular(cf Riekert 1996:84; Dougherty 1992:102).
It does not affect, however in any way the Case assignment regardless from which trace position it would move.The move however eclipses to a degree the observation of the objective Case of bibli, on gegramme, non..... katesfragisme, non..... when the PP 1 moves to a position between ei= don and its NP with objective Case realised as accusative.The preposition ev pi, governs firstly th.n dexia.nandsecondly tou/ qro, nou and asssigns to them oblique Case, respectively realised as accusative and genitive case, but with the same meaning(cf Brütsch 1970:244).
The PP ev pi...... tou/ qro, nou could move from any one of the two trace positions[PPt]marked by an asterisk.This difference indicates a clash with the oblique Case as an inherent Case assignment(cf Riekert 1996:132-146).The NP th.n dexia.ngoverns the NP tou/ kaqhme, nou and assigns to it genitive case and Case.The PP ev k ..... qro, nou receives a theta role from the verb ei; lhfen.The P ev k assigns genitive case as realisation of the oblique Case to th/ j dexia/ j which in turn governs the NP tou/ kaqhme, nou and assigns to it structural genitive Case.As a participle kaqhme, nou has a theta role which has been filled by the PP ev pi.tou/ qro, nou where P ev pi, with genitive case offers a realisation of the oblique Case assignment.(4)(c) (ii) (α) Rev 4:9; 5:13, two elements selected in the original order.
) it becomes conspicuous that ev pi, with the accusative functions in the classical meaning.It would appear that ev pi, with the genitive never functions in this sense.The only possible instances are in(10)