• Non ci sono risultati.

Aspectual and Actional Analysis of Post-Hellenistic Greek 'Εἶναι + Participle' Periphrases in Christian and Hebrew texts

N/A
N/A
Protected

Academic year: 2021

Condividi "Aspectual and Actional Analysis of Post-Hellenistic Greek 'Εἶναι + Participle' Periphrases in Christian and Hebrew texts"

Copied!
175
0
0

Testo completo

(1)

!

DIPARTIMENTO DI

FILOLOGIA, LETTERATURA E LINGUISTICA

CORSO DI LAUREA IN LINGUISTICA

TESI DI LAUREA

Aspectual and Actional Analysis of Post-Hellenistic Greek ‘Εἶναι +

Participle’ Periphrases in Christian and Hebrew Texts

CANDIDATO

RELATORE

Edoardo Nardi

Chiar.ma Prof.ssa

Domenica Romagno

CONTRORELATORE

Chiar.mo Prof.

Franco Fanciullo

(2)
(3)

Ringraziamenti

Sebbene questo lavoro sia scritto in inglese, voglio effettuare i ringraziamenti in

italiano, perché è nella madrelingua di ognuno di noi che le parole assumono un

significato più vero e sincero.

Vorrei dedicare un caloroso ringraziamento al Professor Fanciullo, al Professor

Borbone e, in special modo, alla Professoressa Romagno dell’Università degli Studi

di Pisa, per i loro indispensabili suggerimenti e la loro assidua guida.

E voglio dedicare un sentito e profondo ringraziamento ad ogni singolo membro

della mia famiglia, per il loro incrollabile sostegno ed i loro consigli, talvolta

discutibili, ma pur sempre fondamentali.

(4)

INDEX

6

9

9

9

13

17

19

23

26

27

27

28

30

31

31

35

39

40

43

44

44

46

51

59

59

§0. Introduction

§1. Theoretical Background

§1.1. Periphrasis and

Grammaticalization

§1.1.1. A Brief History of Greek Perfect

§1.1.2. Earlier Studies: Chantraine,

Björck and Aerts

§1.1.3. Earlier Studies: Rosén, Gonda,

Dietrich and Blass & Debrunner

§1.1.4. More Recent Studies: Porter,

Ceglia and Amenta

§1.1.5. Bentein

§1.1.6. Conclusions

§1.2. Semitic Influence

§1.2.1. Earlier Studies: Björck, Aerts,

Mandilaras and Blass & Debrunner

§1.2.2. More Recent Studies: Ceglia,

Amenta, Drinka, and Bentein

§1.2.3. Conclusions

§1.3. Aspect, Actionality and Semantic

Roles

§1.3.1. Introduction to Aspect and

Actionality and Further Observations

§1.3.2. Aspect

§1.3.3. Actionality

§1.3.4. Valency and Syntactic Roles vs.

Semantic Roles

§1.3.5. Conclusions

§2. Analysis of the Occurrences

§2.1. Texts

§2.1.1. Preliminary Indications

§2.2. Adjectival Periphrases

§2.2.1. Analysis and Discussion

§2.2.2. Participles of Stative Verbs

(5)

59

62

64

65

67

67

69

73

75

76

91

91

94

97

104

108

110

113

113

116

121

122

125

125

125

125

138

142

165

170

§2.2.3. Participles of Telic Verbs

§2.2.4. Participles of Activity Verbs

§2.2.5. Conclusions

§2.3. Property-Referring Periphrases

§2.3.1. Analysis and Discussion

§2.3.2. The Issue of Coordination

§2.3.3. Interpretation of

Property-Referring Periphrases

§2.3.4. Actionality and Aspect

§2.3.5. Conclusions

§2.4. Verbal Periphrases

§2.4.1. Analysis and Discussion

§2.4.2. Problematic or Dubious

Occurrences

§2.4.3. Substitutive Periphrastic

Occurrences

§2.4.4. Functional Analysis and

Aspect-Tense Correlation

§2.4.5. Aspect-Actionality Correlation

§2.4.6. Conclusions

§2.5. Future Periphrases

§2.5.1. Analysis and Discussion

§2.5.2. Problematic or Dubious

Occurrences

§2.5.3. Functional, Aspectual and

Actional Analysis

§2.5.4. Conclusions

§3. General Conclusions

§4. Appendix

§4.1. Corpus

§4.1.1. Preliminary Notes

§4.1.2. Adjectival Periphrases

§4.1.3. Property-Referring Periphrases

§4.1.4. Verbal Periphrases

§4.1.5. Future Periphrases

Bibliography


(6)

§0. INTRODUCTION

The purpose of the present work is to provide a comprehensive analysis of the periphrastic constructions occurring in a number of Christian and Hebrew texts, belonging to the Post-Hellenistic Age and dating from I b.C. to VII A.D. (the majority of the examined texts date from I to III A.D.). While the addressed Hebrew texts are all apocryphal, as for 1 2 the Christian texts, we examined both a number of apocryphal books and all the canonical ones (i. e. the New Testament).

It is also important to point out what kind of periphrastic constructions we will take into account: although there is not total agreement on this issue among scholars, several Greek patterns can be considered periphrases, such as ὑπάρχω + present participle, βαίνω / ἔρχοµαι + present / future participle, ἄρχοµαι / παύοµαι + present participle, ἔχειν + present / aorist participle, et cetera. In the present work, we will focus on the construction ‘εἶναι + participle’ (present, aorist, perfect and future). 3

We have selected this corpus for three main reasons: first, as emerged from the previous studies about these periphrases, this kind of pattern mainly occurs in texts from that period (earlier occurrences do exist, but their frequency is extremely low). Second, given the ancient and popular tradition of studies and research on Ancient Greek, nowadays it is not easy to find issues that have not been carefully and deeply studied or even generally addressed: periphrastic constructions do not make exception, although they are relatively less studied compared to other issues; however, it is the perspective that we adopted that may represent an innovative approach (see infra). Lastly, we are extremely interested in religious texts, especially christian, and it was a joyful coincidence that these periphrases appear so frequently in this sort of texts.

As far as the state of the art is concerned, the periphrastic constructions that we are going to deal with have been the topic of only a few studies, such as articles (for example Rosén (1957), Gonda (1959), Dietrich (1973), Ceglia (1998), etc.), or have been treated as part of wider research, which did not specifically address these patterns (for example in Chantraine’s (1927) work on Greek perfect, in Blass & Debrunner’s (1997) New Testament Greek grammar, etc.); however, comprehensive and specific research on periphrases in Ancient Greek have been mainly carried out by four scholars: in chronological order, Björck (1940), Aerts (1965), Amenta (2003) and Bentein (2016); while Björck’s research is rather

In the present work we will refer to the different chronological phases of the Ancient Greek language as 1

follows: Archaic Greek (earlier than VI b.C.; with ‘Homeric Greek’ we will specifically refer to the language used in the homeric poems) - Classical Greek (VI - IV b.C.) - Hellenistic Greek (IV - I b.C.) - Post-Hellenistic Greek (I b.C. - IV A.D.; New Testament Greek is included in this linguistic phase).

A brief definition of the term ‘apocryphal’ may be necessary: in the opening page of his work, Craveri 2

(2014) claims that “l’aggettivo ἀπόκρυφος vuol dire ‘segreto’, ‘nascosto’ (the adjective ἀπόκρυφος means ‘secret, hidden’)”, but if it is “attribuito a uno scritto di contenuto religioso (Vangelo, Atto, Epistola,

Apocalissi ecc.) è considerato sinonimo di ‘non autentico’, ‘erroneo’, ‘eretico’ in contrapposizione a

canonico che significherebbe, invece, ‘autentico’, ‘veritiero’, ‘ispirato’ (‘attributed to a writing of religious content (gospel, act, epistle, apocalypse, etc.) it is considered synonym of ‘not authentic’, ‘erroneous’, ‘heretical’ in contrast with canonical that means ‘authentic’, ‘truthful’, ‘inspired’’)”.

Actually, periphrases selecting the future participle are very rare in the examined corpus. 3

(7)

criticized (particularly by Aerts (1965), but, indirectly, also by others), the other three scholars make a number of relevant points and observations, that turned out to be fundamental for the present work.

In general, all previous research addressing Ancient Greek periphrastic constructions focuses on specific aspects of this topic (of course, not every single work concentrates on every single aspect from the list below; this is a summary of the most relevant and common issues):

-

characteristics that are peculiar to periphrastic patterns (in other words, when it is the case in which we are dealing with a periphrasis) and how a construction becomes periphrastic (i. e. its process of grammaticalization): in a way, this is the same issue approached respectively synchronically and diachronically;

-

distinguishing between verbal periphrases (e. g., ἦν διδάσκων, ‘(s)he was teaching’) and adjectival periphrases (e. g., ἦν κεκλεισµένος, ‘it was shut’);

-

types of participles selected in the constructions and their temporal / aspectual values;

-

supposed influence of Semitic languages (specifically, Hebrew and Aramaic) on these constructions;

-

motivations for preferring periphrastic patterns to synthetic forms;

-

other issues of higher specificity (for example the constituent order of the construction, markedness, etc.).

All these issues will be considered in the present work, although we will more accurately address some (such as the aspectual and temporal values encoded) to the detriment of others (such as describing a precise grammaticalization process that led these patterns to become periphrastic).

At this point, it is fundamental to clarify what is the perspective that we will adopt in the present work: our analysis will specifically focus on the categories of Aspect and Actionality, alongside with reference to Tense; in particular, we will focus on the aspectual values encoded by the examined constructions, on the actional classes selected by the predicates involved in the constructions and on the interrelation between the categories of Aspect, Actionality and (when relevant in relation to one (or both) of the other two) Tense.

The motivation for this choice can be somehow inferred from what we have written above: even though some scholars have addressed Ancient Greek periphrastic constructions from an aspectual perspective, with some reference to Actionality (it should be noted that these categories, especially in the earlier studies, are often confused or even misinterpreted) – as far as we know – a specific study about Aspect and Actionality and their interrelation on this constructions has never been carried out. Briefly, what has emerged from our analysis is that, despite the heterogeneity of the morphosyntactic, functional and aspectual characteristics displayed by the examined constructions, some regular patterns of behavior have been detected in the aspectual values encoded, the actional classes selected and their interrelation; in a number of cases, Tense plays a crucial role too. As we will show, quite expectably, these points are not without exceptions.

Lastly, we will provide a quick overview of the present work, that is divided into four chapters:

1) In the first chapter, an overview of the theoretical background is provided, with special attention to the macro-issues of periphrasticity, grammaticalization and

(8)

aspectual / actional categories: all the theoretical and methodological tools that are necessary to our analysis are described and discussed in details; beside this, the state of the art is illustrated in this section. The issue of Semitic influence is also briefly addressed in this chapter.

2) In the second chapter, all occurrences of periphrastic constructions are collected in analytical tables, divided into four groups and analyzed from the aspectual and actional viewpoints; furthermore, other significant issues are discussed (such as coordination between periphrases and exceptions to the general patterns of behavior).

3) In the third chapter, the conclusions of our work will be presented.

4) In the fourth chapter, all sentences and verses containing the examined periphrastic constructions are collected (together with their translations).

(9)

§1. THEORETICAL BACKGROUND

§1.1. Periphrasis and Grammaticalization

First of all, we believe that it is necessary to provide a brief overview of previous research on periphrastic constructions in Ancient Greek, with special attention to the criteria by which different scholars identified and classified these forms: this is fundamental, because unitary and common criteria for identification and classification of periphrases are far from clear.

As a preliminary observation, we would like to quote Hopper & Traugott (1993, 52): “When VO order arises from OV, the change will often be accompanied by the innovation of new phrasal (‘periphrastic’) ways of coding what at an earlier stage was coded inflectionally [i. e. synthetically]”. Even if Ancient Greek did not display a strict (S)OV order, but rather a generally free word order, Modern Greek shows an (S)VO order, as argued by Taylor (1994, 1-2): “although the order of major sentence constituents at all stages of Ancient Greek is quite free, the distribution of clause types is not random over time, but changes from predominantly verb-final [(S)OV] to predominantly verb-medial [(S)VO], suggesting a change in progress. […] the data reflect an ongoing change from verb-final to verb-medial structure beginning at or before Homer and nearing completion 800 years later with the Koiné of the Hellenistic period”; therefore, the “spread” of periphrastic constructions registered since Hellenistic Age may be an epiphenomenon of the more general principle pointed out by Hopper & Traugott.

On the other hand, Hopper & Traugott (1993, 95) make another observation that seems suitable for our dissertation: “A particular grammaticalization process may be, and often is, arrested before it is fully ‘implemented’, and the ‘outcome’ of grammaticalization is quite often a ragged and incomplete subsystem that is not evidently moving in some identifiable direction”. In other words, this means that often items undergoing grammaticalization do not reach what is supposed to be the final stage of that process, but they stop at an intermediate phase of the process itself. In this regard, Hopper & Traugott (1993, 8) point out the tendency of “periphrastic constructions to coalesce over time and become morphological ones”, as happened, for example, in the formation of the Italian simple future (Lat. cantare habemus (‘we have to sing’) > *cantarabemus > *cantaremus > Ital.

canteremo (‘we will sing’)): the periphrastic constructions in our data did not reach the

grammaticalization phase that we have just described (which is supposed to be the above-mentioned final stage), because their components are always neatly separated, and, despite their quite frequent synchronical usage, it seems that they have disappeared in Modern Greek, which selects a different type of periphrasis. 4

§1.1.1. A Brief History of Greek Perfect

The issue of the continuity between Ancient Greek and Modern Greek periphrases do not concern the 4

present work, so we will not deal with it; however, some of the scholars that we have examined address this topic: specifically, Chantraine (1927), Aerts (1965), Bruno (2012) and Drinka (2017).

(10)

Before discussing various scholars’ research about periphrastic constructions, we think that it is necessary to provide a quick diachronical description of perfect in Ancient Greek.

First of all, the Indo-European function of perfect was to indicate a state of the subject; beside this, it was originally intransitive. Perfect in Archaic / Homeric Greek retains both these characteristics: in this regard, Chantraine (1927, 4) claims that “le rôle propre du

parfait est […] d’exprimer l’état. […] le parfait indique un état atteint la suite d’un procès antérieur (‘the specific function of perfect is […] expressing the state. […] perfect indicates

a state resulting from a previous process’)”; it is also to remark that this verbal form “mettait

en lumière un résultat actuel, l’action passée était laissée au second plan (Chantraine

(1927, 152): ‘[it] highlights a present result, while the past action is left in the background’)”. Beside this, Chantraine (1927, 146) also argues that “de l’étude de la langue

homérique, on peut conclure que le parfait [...] se rapporte généralement au présent. Ce sens c’est conservé en attique. Toute une série de parfaits s’emploient comme des présents exprimant l’état (‘from the study of the Homeric language, one can conclude that perfect

[...] generally refers to the present. This meaning is preserved in Attic. A whole series of perfects are used as presents expressing the state’)”; it must be clarified, though, that “le

sens de présent n’était pas exceptionnel [...]. Il est parfois difficile de distinguer dans un verbe donné le présent et le parfait. La nuance de sens est fugitive (Chantraine (1927, 150):

‘the meaning of present was not exceptional [...]. It is sometimes difficult to distinguish in a given verb the present and the perfect. The nuance of meaning is fleeting’)”: in fact, perfect did not primarily encode Tense (this observation will be fully explained a few lines below), but it generally referred to a present situation with a link with the past.

Romagno (2005, 9-10) specifies that “il prefetto greco viene tradizionalmente suddiviso

in due tipi; e cioè (semplificando una classificazione in realtà più confusa): 1) il cosiddetto perfetto con valore di presente (γέγηθα ‘gioisco’, δέδορκα ‘ho un determinato sguardo’, etc.); 2) il perfetto di stato risultante (πέπηγα ‘sono piantato’, ἔφθορα ‘sono distrutto, rovinato’, etc.). […] il primo tipo rappresenta lo stato dell’autore dell’evento, il secondo lo stato dell’entità che subisce gli effetti di un evento: cioè, dell’oggetto della diatesi attiva (φθείρω τὴν πόλιν : ἡ πόλις ἔφθορε) (‘the Greek perfect is traditionally divided into two

types; that is (simplifying an actually more confused classification): 1) perfect with present value (γέγηθα ‘I rejoice’, δέδορκα ‘I have a certain gaze’, etc.); 2) the resulting-state perfect (πέπηγα ‘I am planted’, ἔφθορα ‘I am destroyed’, etc.). […] while the former type describes the state of the author of the event, the latter indicates the state of the entity that undergoes the effects of the event: i. e., the object of the active Diathesis (φθείρω τὴν πόλιν [‘I destroy the city’] : ἡ πόλις ἔφθορε [‘the city is destroyed’])’)”.

However, Romagno (2005, 10) addresses a problem about this classification: given that “l’interpretazione vulgata attribuisce al perfetto la funzione di significare lo stato risultante

dal compimento di un processo (‘the traditional interpretation assigns to the perfect the

function of encoding the state resulting from the realization of a process’)”, the Italian scholar points out that “questa tesi, però, può dar ragione di (alcuni) perfetti del tipo 2, ma

non di quelli del tipo 1 (‘this thesis can explain (some) perfects of the second type [ἔφθορα],

but not those of the first type [γέγηθα]’)”. Therefore, Romagno (2005, 49) proposes an alternative view: she suggests that (in Homeric Greek, at least) “la nozione di stato non

(11)

necessarily correspond with the one of resulting state’)” and claims that “la funzione del

perfetto indoeuropeo era, piuttosto, quella di significare lo state inteso come proprietà del soggetto, indipendentemente dalla realizzazione effettiva dell’evento stesso (‘the function of

the Indo-European perfect was to encode the state as a property of the subject, independently of the actual fulfilment of the event itself’)”. This function indeed explains both types of perfect (for more detailed explanations, see Romagno (2005)).

So, according to Chantraine (1927), the old perfect had the specific function of indicating a present (at the moment of the utterance) state resulting from a previous event; it should be observed that this is the traditional interpretation of Greek perfect. By contrast, an alternative interpretation is proposed by Romagno (2005, 57), who argues that perfect (at least in Indo-European / Homeric Greek) “rappresenta lo stato in sé, e non come

acquisizione di una condizione nuova (‘represents the state in itself, and not as acquisition

of a new condition’)” (the whole issue is actually more complex; for a detailed explanation see Romagno (2005)).

The above-mentioned original function of perfect endured through the Homeric / Archaic Age. Later, this verbal form developed a new function, i. e. it became what scholars call ‘resultative perfect’: as Romagno (2005, 97) claims, “il perfetto è, nel suo valore primitivo,

tipicamente intransitivo. Tale è la stragrande maggioranza dei perfetti omerici. […] Il perfetto risultativo è transitivo e […] segnala lo stato dell’oggetto ‘coinvolto’ nel processo significato dal verbo (‘perfect is, in its original value, typically intransitive. Such is the

majority of Homeric perfects. […] Resultative perfect is transitive and […] it specifies the state of the object ‘affected’ in the process encoded by the verb’)”; in this regard, Chantraine (1927, 122) observes that “le parfait expire bien encore un état; mais ce n’est plus l’état du

sujet, c’est celui de l’objet (‘perfect still describes a state; but it is not the state of the

subject, it is the one of the object’)”.

Actually, this function of perfect already appears in Homer (in fact, Romagno writes “such [intransitive] is the majority of homeric perfects”), but it was in V-IV b.C. that it fully developed: Chantraine (1927, 122-123) argues that “dans la langue homérique on entrevoit

l’existence d’un parfait résultatif. Au V siecle, les exemples son plus nombreux. […] Pour étudier le développement du parfait résultatif en ionien-attique, il convient de marquer deux étapes. Il apparaît dans les plus anciens textes: Tragiques, Hérodote, Thucydide, Antiphon, mais à l’état de nouveauté: il n’est pas usuel. En moyen et nouvel attique, au contraire il se répand très rapidement, dans la Comédie, chez Xenophon, chez Platon, chez les Orateurs. L’espace de temps qui sépare ces deux groupes n’est pas grand: la différence linguistique pour l’étude du parfait est nette (‘in the Homeric language one can find the existence of a

resultative perfect. In V century b.C., there are more examples. […] In order to study the development of resultative perfect in Ionian-Attic, two steps need to be noted. It appears in the older texts: Tragic, Herodotus, Thucydides, Antiphon, but in the state of novelty: it is not common. In Middle and Early Attic, on the contrary, it spreads very quickly in the Comedy, Xenophon, Plato, the Orators. The space of time separating these two groups is not great: the linguistic difference for the study of the perfect is clear’)”.

So, alongside the original intransitive value of perfect, indicating the state of the subject (subject-oriented), a new transitive object-oriented value gradually developed, which eventually substituted the original function of perfect.

(12)

It was in the Late-Classical and Hellenistic Ages that perfect underwent a crucial change. In Pre-Classical Greek there was a strong opposition between perfect and aorist, which, as Chantraine (1927, 154) claims, expressed “le fait pur et simple” and had an anterior value (it was a praeteritum), i. e. encoded a past Tense: e. g., τέθνηκα (perfect, ‘I am dead’) vs. ἔθανον (aorist, ‘I died’). It is now necessary to point out that, as Chantraine (1927, 157) observes, “le parfait […] exprime avant tout l’aspect, et […] est en soi indifférent au temps (‘perfect […] expresses aspect above all and it is in itself indifferent to Tense’)” (the term “aspect” here risks to be misleading, because it is not meant in the same way as the category of “Aspect” that concerns us in this work; furthermore, what Chantraine means is indeed closer to the domain of Actionality than to that of Aspect (both notions will be explained in details in §1.3.): nonetheless, we have to use this term in order to make our explanation the clearest): so, perfect had an aspectual value. This property gradually disappeared since Late-Classical Greek (IV-III b.C.), when the specific function of the resultative perfect faded and this verbal form “ended up […] being quite equivalent to an emphatic aorist (Chantraine (1952, 199): parfait dit « résultatif », […] qui finira, au cours de l’histoire du grec, à

équivaloir à peu près à un aoriste emphatique)”: in other words, perfect and aorist 5 overlapped and the former underwent a temporalization process, that is it began to encode Tense (specifically, a past Tense), developing an anterior value (the same as aorist) and becoming a praeteritum form. The process of temporalization and its consequences are described by Chantraine (1927, 233-245) in the following terms: “le sens du parfait devient

d’autre part de plus en plus voisin de celui de l’aoriste. Les papyrus de l’époque ptolémaïque et de l’époque impérial laissent bien apercevoir cette évolution. […] Le parfait perd peu à peu sa valeur propre pour rapprocher de l’aoriste [...] d’autre part le parfait tend à disparaître. Mais la décadence du parfait qui sort peu à peu de l’usage est précisément la preuve la plus nette de sa tendance à se confondre avec l’aoriste. Dès que le parfait ne se distinguait plus bien de l’aoriste, il devait disparaître. [...] Dans les papyrus ptolémaiques il est possible de suivre une évolution toute parallèle du parfait. [...] Dans les papyrus chrétiens la confusion du parfait et de l’aoriste est continuelle. [...] Il apparaît dès maintenant que dans la koine grecque, à l’époque impériale, la confusion du parfait et de l’aoriste est près d’être un fait accompli. […] A l'époque byzantine, le parfait n’existe plus comme thème vivant dans la flexion verbale (‘the sense of perfect gets, on the other hand,

closer and closer to that of the aorist. The papyri of the Ptolemaic and Imperial periods clearly show this evolution. […] [in the New testament] perfect gradually loses its own value to get closer to that of the aorist […] on the other hand, perfect tends to disappear. But the decadence of perfect, which gradually comes out of use, is the clearest proof of its tendency to merge with the aorist. As soon as perfect was no longer distinguishable from the aorist, it had to disappear. […] In Ptolemaic papyri it is possible to follow a parallel [to the New Testament Greek] evolution of perfect. […] In the Christian papyri the confusion of perfect and aorist is continual. […] It now appears that in the Greek Κοινὴ, in the Imperial age, the merger of perfect and aorist is ready to be an accomplished fact. […] In Byzantine age, perfect no longer exists as a productive stem in the verbal inflection’)”.

Chantraine does not explain what he means by “emphatique”. 5

(13)

Actually, though, the situation is quite complex, as Chantraine (1927, 233-234) himself admits: “la situation est complex. Pour chaque passage il est souvent malaisé de marquer

pourquoi l’auteur a employé ici un parfait, là un aoriste. […] Très souvent le parfait est employé pour insister sur un état présent (‘the situation is complex. In some passages it is

often difficult to understand why the author used here a perfect and there an aorist. […] Perfect is very often used to insist on a present state [i. e. it still retains its original value]’)”. These facts can be explained by the observation that the process of temporalization was, of course, gradual and it did not change the status quo in one day: however, what concerns us is that the result of this process was the eventual disappearance of the synthetic perfect and that this phenomenon brought with itself crucial consequences for our work, as pointed out by Chantraine (1927, 246): “quand une forme grammaticale perd sa valeur caractéristique

et tend à disparaître, la langue lui substitute des procédés expressifs et en particulier des formule périphrastiques. C’est ce qui est arrivé pour le parfait grecque (‘when a

grammatical form loses its characteristic value and tends to disappear, the language substitutes it with expressive processes and in particular periphrastic formulas. This is what happened to the Greek perfect’)”. In agreement with Chantraine, Bentein (2016, 55 [“Perfect Aspect” chapter]) claims that “the most important development for our present purposes is the disappearance of the synthetic perfect. This development is generally attributed to the functional overlap between the aorist and the perfect when the latter became an anterior perfect [i. e., a praeteritum]”.

§1.1.2. Earlier Studies: Chantraine, Björck and Aerts

Even though Chantraine (1927, 246) admits that “le parfait périphrastique constitué par

le participe parfait accompagné de εἰµί a joué en grec un grand rôle dès les plus anciens textes (‘the periphrastic perfect constituted by the perfect participle accompanied by εἰµί

played in Greek a great role since the most ancient texts’)”, the French scholar also claims that, particularly in the New Testament, we can observe a development of the periphrastic conjugation, especially in constructions that involved εἶναι and a participle in middle Diathesis (often used as adjective), that supposedly had to replace the disappeared synthetic perfect.

As far as εἶναι + perfect participle periphrasis and its formation are concerned, Chantraine (1927) makes two points:

5) “it first developed in the Moods of perfect, that the language had some difficulty to constitute (Chantraine (1927, 246): il s’est d’abord développé dans les modes du

parfait, que la langue avait quelque peine à constituer)”, like optative (pour l’optatif, les exemples sont nombreux) and subjunctive (au subjonctif la situation est la même, le parfait périphrastique est extrêmement fréquent), and it later extended to the

indicative (however, it is not clear what Chantraine is referring to when he writes “quelque peine à constituer”). In agreement with Chantraine, Smyth (1984, 198-199

(14)

& 201) claims that both active and middle perfect optative and subjunctive were “commonly formed periphrastically”. 6

6) In turn, within the indicative Modality, the periphrastic perfect developed from a process of analogical extension starting from verbs whose stem ended with a consonant: in the formation of the third person plural of the synthetic form, an accumulation of three consonants occurred, leading to the regular vocalization of the interconsonantal nasal: e. g. with κρύπτω (‘to hide’), kekrupʰṇtai > κεκρύφαται instead of *κεκρύφνται. As Chantraine (1927, 247) observes, the problem is that “une

forme comme κεκρύφαται ne présente pas au point de vue grec une désinence claire; on a perdu le sentiment que -αται est un représentant phonétique de -νται (‘a form

like κεκρύφαται does not present, in the Greek point of view, a clear ending; the feeling that -αται is a phonetic representative of -νται was lost’)”: in other words, the ending -αται was not distinctive enough. Consequently, in the Classical period, the periphrastic construction κεκρυµµένοι εἰσὶν was increasingly more often used as an alternative form, first in verbs whose stem ended in consonant; later, it was extended to the third person singular and to verbs with stems ending in liquids (e. g., ἠγγελµένοι ἦσαν) and in vowels (e. g., ᾑρηµένοι εἰσίν). 7

Björck (1940) defines a periphrasis as a combination of a copula (specifically εἶναι) and a participle, such that it can substitute a synthetic form without any variation in meaning (i. e., a form like ἦν λέγων – when truly periphrastic – encodes exactly the same meaning as the synthetic form ἔλεγε). Furthermore, we cannot talk about real verbal periphrasis when there is no semantic integration between the auxiliary and the participle, as the latter gains an adjectival value and it expresses a characteristic feature of the subject or a permanent situation, rather than an action: this is what Björck (1940) calls adjectival periphrasis (Björck’s concept of ‘semantic integration’ might be something similar to the three Amenta’s (2003) auxiliarization parameters: see §1.1.4.). In our data, though, we actually found a number of verbal periphrases expressing a permanent or temporal state or property of the subject, i. e. what Aerts (1965; see infra) defines ‘situation-fixing periphrases’: e. g., Mt 19,22: ἦν γὰρ ἔχων κτήµατα πολλά (‘because [he] had a lot of possessions’), where the property of having possessions is expressed, or AcPhi 46: ἡ δὲ διδασκαλία αὐτοῦ ἐστι διαχωρίζουσα ἄνδρας καὶ γυναῖκας, […] (‘but his teaching distinguishes men from women, […]’), where the property of distinguishing is expressed.

Furthermore, Björck (1940) claims that a sort of ‘halfway’ periphrasis between adjectival and verbal can be found as well; these periphrases are the ones where εἶναι conveys an existential meaning, i. e. where εἶναι retains its independent original meaning of ‘to be, to exist’: e. g., Lc 2,8: καὶ ποιµένες ἦσαν ἐν τῇ χώρᾳ τῇ αὐτῇ ἀγραυλοῦντες καὶ φυλάσσοντες φυλακὰς τῆς νυκτὸς ἐπὶ τὴν ποίµνην αὐτῶν (‘and there were shepherds in the same country abiding in the field, and keeping watch by night over their flock’). In disagreement with

“The perfect [active] subjunctive is commonly formed periphrastically by the perfect active participle and 6

ὦ, ᾖς, ᾖ, etc. […] The perfect [active] optative is commonly formed periphrastically by the perfect active participle and εἴην, εἴης, εἴη, etc. […] The perfect middle subjunctive is commonly formed by the periphrasis of the perfect middle participle and ὦ, ᾖς, ᾖ, etc. […] The perfect middle optative is commonly formed by the periphrasis of the perfect middle participle and εἴην, εἴης, εἴη, etc.”.

Both examples are taken from Xenophon. As for point 2, see also Bentein (2016). 7

(15)

Björck (1940), but in agreement with Aerts (1965; see infra), we do not hold these constructions to be periphrastic: for further details see §2.1.1. [point 2].

Anyway, Björck (1940) points out that the only true verbal periphrasis is εἶναι + present participle (progressive periphrasis) and that it can be considered as the Greek counterpart of English progressive tenses: in fact, Gr. ἦν διδάσκων is translated ‘(s)he was teaching’. Even though a very large amount of periphrases in the examined corpus convey the just mentioned meaning, we disagree with this observation, because firstly, our data shows lots of verbal periphrastic constructions with perfect or aorist participles, instead of present participle (e. g., EpAb 3: ἀπεσταλµένος εἰµὶ ὑπὸ Αὐγάρου θεάσασθαι Ἰησοῦν τὸν Ναζωραῖον (‘I have been sent by Abgar to see Jesus from Nazareth’)), and secondly, our data show periphrases selecting present participles that do not encode a progressive meaning (e. g., PasAn 8: ὑπέρ τοῦ ὀνόµατος αὐτοῦ ἔσοµαι διαµένων εἰς τὰς βασάνους σου ὁµολογητής (‘and for his name’s sake I will remain in torments, as his witness’)).

Finally, Björck (1940) refutes the hypothesis that ἦν / ἦσαν + aorist participle is a ‘suppletive periphrasis’ (i. e. a periphrastic construction that replaces a no-longer existing synthetic form: see Aerts (1965) and below in this paragraph) for the disappearing (or disappeared) synthetic pluperfect, for two main reasons:

1) why does εἰµί + aorist participle so seldom substitute a perfect, while ἦν / ἦσαν + aorist participle so often substitute a pluperfect? What seems strange to Björck is the mismatch in frequency between these two forms; in this regard and in open disagreement with Björck (1940), Aerts (1965, 76-77) observes that “Björck makes no allowance for the fact that a language often develops according to very capricious patterns, as is witnessed by the evolution from the ancient Greek to modern Greek future. In the ancient Greek verbal system the future, with its purely temporal function, is an anomaly among the aspect-functions of the other tenses. This anomaly has disappeared from modern Greek, […]. The capriciousness of the way a language evolves, therefore, makes it possible to accept a certain development for the pluperfect which cannot obtain for the perfect” (see below for more details).

2) More importantly, ἐστί / ἦν + aorist participle was long in use before the synthetic pluperfect disappeared.

By contrast, it should be observed here that Mandilaras (1973), in disagreement with Björck (1940), points out that both in the papyri and in the New Testament there occur periphrastic constructions such as ἦν + active aorist / perfect participle and ἦν + passive aorist / perfect participle, encoding respectively an active pluperfect and a passive pluperfect.

Aerts (1965) defines a periphrasis as a combination of a participle and an auxiliary, that most commonly is εἶναι, but it can also be γίγνεσθαι, ἔχειν, etc. (but, as pointed out in the Introduction, we will concentrate on those with εἶναι). On the other hand, he also clarifies what is not a periphrasis: any combination of εἶναι + participle where 8

1) εἶναι has an existential meaning (very roughly, when it can be translated as ‘there is’), e. g. Herodotus, II, 65,3: νόµος δέ ἐστι περὶ τῶν θηρίων ὧδε ἔχων (‘but however that may be, there is a custom with regards to animals which is as follows’);

The following examples are the one proposed by Aerts (1965, 7-11). 8

(16)

2) εἶναι appears with adjuncts of place or time (adjuncts, or complements, are phrases that add any sort of specification to the sentence: time, place, cause, etc. For instance, ‘at dawn, Mark woke up in his bed because of a horrible nightmare’ (adjuncts are underlined)), e. g. Herodotus, I, 206,1: οὐ γὰρ ἂν εἰδείης, εἴ τοι ἐς καιρὸν ἔσται ταῦτα τελεόµενα (‘after all you cannot know, if this will be to your advantage should you carry it out’);

3) εἶναι appears with a Dative of possession (or ‘Dative of interest’, as Aerts calls it), e. g. Herodotus, VII, 209,3: νόµος γάρ σφι οὕτω ἔχων ἐστί (‘they have, namely, the following custom’);

4) the position of εἶναι indicates emphasis (for instance, if εἶναι precedes the participle, this indicates a certain grade of independence of the auxiliary and that “emphasis is given to an ‘essive’ aspect of the verb (Aerts (1965, 11))”), e. g. Herodotus, VIII, 46,1: ἦσαν µέν σφι καὶ ἄλλαι πεπληρωµέναι νέες (‘and they had also other filled ship’).

While we agree with points 1 and 3, we do not with point 2, as our data show a number of periphrases occurring alongside adjuncts of place and / or time (e. g. Lc 19,47: καὶ ἦν διδάσκων τὸ καθ’ ἡµέραν ἐν τῷ ἱερῷ (‘and he was teaching daily in the temple’), where both time and place adjuncts occur, or AcAp 12,6: […], τῇ νυκτὶ ἐκείνῃ ἦν ὁ Πέτρος κοιµώµενος µεταξὺ δύο στρατιωτῶν δεδεµένος ἁλύσεσιν δυσίν (‘[…], the same night Peter was sleeping between two soldiers, bound with two chains’)). As for point 4, it is a bit controverted: if, as Aerts (1965, 11) claims, “emphasis is given to an ‘essive’ aspect of the verb”, i. e. εἶναι retains its existential meaning, the situation is the same as point 1; on the other hand, given the overall free word order of the Greek language (see Taylor (1994) and §1.1.), the position of the auxiliary in relation to the participle − according to our data − is not very significant in the identification of a periphrasis: in fact, we have found periphrases where the auxiliary precedes the participle, but the ‘essive’ aspect of the verb does not seem to be emphasized (e. g. MarMa 12: καὶ νῦν ἰδοὺ ἐκπορεύεται ὁ Ματθαῖος καὶ ὁ Πλάτων σῦν αὐτῷ, καὶ εἰσιν εἰσπορευόµενοι ἐπὶ τὴν πύλην τὴν λεγοµένην βαρείαν· (‘and behold, Matthew is now coming out and Plato with him, and they are making for the door called Heavy;’) or AcAnMa 3A: τότε οὖν ὁ Ματθείας ἐκαθέσθη ἐν τῇ φυλακῇ καὶ ἦν ψάλλων (‘then Matthias sat down in the prison and began to sing psalms’)).

Moreover, Aerts (1965, 3) introduces an innovative classification system for periphrases, by dividing them into three types on paradigmatic / semantic basis:

1) ‘Substitute periphrasis’: when it replaces a synthetic form with no distinguishable change in meaning, e. g. γεγραµµένον ἦν vs. ἐγέγραπτο (‘it had been written’).

2) ‘Suppletive periphrasis’: when it is used to replace a no longer existing synthetic form, e. g. γεγραµµένοι εἰσίν vs. γεγράφαται (‘they have been written’), or when it is used to encode a meaning which a given language does not encode by mean of a synthetic form.

3) ‘Expressive periphrasis’: when it is used with a “special purport”, i. e. when it encodes a marked meaning, e. g. δεινῶς ἀθυµῶ µὴ βλέπων ὁ µάντις ᾖ (‘I have dread fears that the seer can see’). 9

Sophocles, Oedipus Rex 747, translation by Sir R. Jebb. 9

(17)

Aerts (1965) also makes a distinction between the two possible characters that, according to him, periphrases (mostly, εἶναι + present participle) can display, that is either situation-fixing or progressive (it is to remark that these two characters do not complete all the possible values encoded by the periphrases occurring in the examined corpus): while the former describes a situation that can be either permanent or temporal (in disagreement with Björck (1940), who does not reckon these to be ‘true’ periphrases: e. g., Mt 19,22: ἦν γὰρ ἔχων κτήµατα πολλά (‘because [he] had a lot of possessions’)), the latter describes an action that is taking place at the moment of the utterance of the sentence, exactly like the English progressive Tenses (e. g., Mt 7,29: ἦν γὰρ διδάσκων αὐτοὺς […], ‘he was teaching them […]’). Moreover, even though Aerts (1965) argues that situation-fixing periphrases are much more common than progressive ones, he also admits the fact that the boundaries between the two are often very hard to establish. Actually, we have to disagree on both these points: firstly, our data show that progressive periphrases are indeed much more frequent than situation-fixing ones (159 (160?) progressive vs. 44 situation-fixing); secondly, we 10 have not found any specific difficulty in discerning progressive periphrases from the property-referring ones, as they encode very different meanings from each other (see §2.3.). Moreover, we believe that a further Aerts’ (1965) observation should be pointed out as well: εἰµί + aorist participle is very rarely used as a perfect in the Κοινή; by contrast and in open disagreement with Björck (1940), Aerts (1965, 76) claims that ἦν + aorist participle is frequently found as a pluperfect and that “the increase in this use is accompanied by a proportional decrease, and eventual disappearance, of the true [meaning ‘synthetic’] pluperfect”. The same conclusions can be drawn from our analysis: the examined corpus shows 51 occurrences of pluperfect vs. 26 occurrences of perfect (almost exactly the half); by contrast, it should be pointed out that our data also show εἰµί / ἦν + perfect participle (instead of aorist participle) encoding respectively perfects or pluperfects (see §2.4.4.).

As far as adjectival periphrases are concerned, Aerts defines them as combinations of a copula and a completely adjectivized participle; he admits (1965, 12-13), though, that “a state of pure adjectivization […] is usually a very difficult condition to determine, and judgment on that point will depend on the one hand on the individual linguistic feeling and on the other on a given, concrete linguistic situation”. However, as for what has emerged from our data, we believe that adjectivization in itself is not so difficult to determine; what indeed appears to be more problematic is distinguishing between adjectival periphrases and the above-mentioned situation-fixing type (that we prefer calling ‘property-referring’ type; this issue will be discussed in details in §2.3.).

§1.1.3. Earlier Studies: Rosén, Gonda, Dietrich and Blass & Debrunner

Rosén (1957) − it should be noted that his article focuses specifically on Herodotus − proposes the hypothesis that a periphrastic construction is a marked verbal form, in which it is the adjuncts added to the construction that are highlighted, rather than the action denoted by the verb; in other words, it is comparable to an English cleft sentence (for instance:

Our data show a suspect progressive periphrasis that is actually impossible to verify, because the text is 10

heavily corrupted (AcPhi 138B; the fact that a progressive periphrasis occurs here is indeed a suggestion of ours; see §2.4.2.). Moreover, we think that it should be pointed out that four periphrases of the progressive type have their auxiliary omitted (see §2.1.1. [point 5]): therefore, their occurrence was inferred.

(18)

unmarked sentence ‘we got married in June’ vs. cleft sentence ‘it was in June that we got married’). Whether this hypothesis holds to be true in Classical Greek, we cannot tell, as our work focuses on a later type of Greek. However, as for what our data show, we do not think that Rosén’s hypothesis turns out to be true for Hellenistic and Post-Hellenistic Greek, for two reasons.

Firstly, in a number of sentences with periphrases no complements occur, so there are simply no adjuncts to highlight: e. g., AcTho 16: καὶ οὐχ εὗρον αὐτόν· πλεύσας γὰρ ἦν. (‘but they did not find him: for he had set sail’) or ApHen 14,14: καὶ ἤµην σειόµενος καὶ τρέµων, καὶ ἔπεσον. (‘and I was fidgeting and trembling, and I fell’).

Secondly, we have found some utterances which occur identically to each other, except for the verbal form, that is in one case synthetic, while in the other periphrastic; an example will help to clarify: AcAnMa 25: αἱ σάρκες αὐτοῦ ἐκολλῶντο ἐν τῇ γῇ, καὶ τὸ αἷµα αὐτοῦ ἦν ῥέον ὥσπερ ὕδωρ ἐπὶ τὴν γῆν vs. AcAnMa 28: αἱ σάρκες αὐτοῦ ἐκολλῶντο ἐν τῇ γῇ, καὶ τὸ αἷµα αὐτοῦ ἔρρεεν ὥσπερ ὕδωρ ἐπὶ τὴν γῆν (‘and his limbs were stuck together on the ground and his blood was flowing like water on the ground’); the same periphrastic expression occurs at AcAnMa 26. The point is, given that the same expression is repeated identically for three times, we believe that it is unlikely that the sentence is marked in one 11 case, and not in the other. What we reckon more likely – but not beyond doubt – is that, at least in this sort of occurrences, we are dealing with substitute periphrases (in Aerts’ sense) and that, in this specific case, the periphrastic construction corresponds to a synthetic imperfect (ἦν ῥέον = ἔρρεεν). In this regard, Bentein’s (2016, 8 [“Theoretical Background” chapter]) suggestion appears to be significant: “one clear pragmatic motivation for the use of periphrasis is to achieve variation, which is evidenced by its appearing alongside a synthetic form”).

Gonda (1959) − whose article is about Classical Greek − suggests that there is a slight semantic difference between a periphrastic expression and a synthetic form, that is whereas the former is purely descriptive, i. e. describes a state of affairs, the latter encodes a narrative connotation: in order to exemplify his claim, Gonda (1959, 101) writes that “in Herodotus II, 95,1 the words πρὸς δὲ τοὺς κώνωπας […] τάδε σφί ἐστι µεµηχανηµένα are purely ‘descriptive’ stating that the Egyptians have definite contrivances against the gnats, but the sentence τοῖσι δὲ περὶ τὰ ἕλεα οἰκέουσι τάδε […] µεµηχάνηται means, in a ‘narrative’ style of expression, that some people had contrived some defence”. Actually, though, the periphrases in our data do not seem to encode this specific semantic distinction: there are indeed some descriptive periphrases, but many other have a narrative function (we should remind that Gonda’s article is about Classical Greek, that is a slightly different type of Greek in comparison with the one found in the examined corpus).

Moreover, in contrast with Rosén (1957), Gonda (1959, 107) claims that in some contexts a periphrasis does not emphasize its complements, but it highlights “the ‘parenthetical character’ of the sentence which whilst interrupting the course of the narrative informs the reader of a past event explaining a state of affairs”. This could be true for some

We suggest that the repetition of the same expression in this case might aim at reproducing the typical 11

language of christian religious ceremonies, that is full of fixed formulaic expressions. If this held to be true, it would support our hypothesis that both synthetic and periphrastic forms here encode the same meaning, because in formulaic contexts expressions are not expected to vary in any way.

(19)

instances, but, as for what we can tell from the examined corpus, we do not think that Gonda’s observation should be pointed out as a general – or even peculiar – characteristic of periphrastic constructions.

Among the scholars that we have been addressing, Dietrich (1973) is chronologically the first one who gave a quite accurate and precise definition of ‘periphrasis’. He came up with three distinct criteria: 1) a periphrastic construction is a combination of two verbal elements such that it constitutes a syntactic unit (syntactic criterion), 2) whose meaning is not inferable from its parts considered singularly, but from their combination (semantic criterion) and 3) such that it stands in functional opposition with a synthetic form (paradigmatic criterion). As far as point 1 is concerned, despite generally agreeing about it, we would like to make two specifications concerning the type of Greek periphrasis that we are dealing with: first, it is not strictly necessary that the two verbal elements composing the periphrasis occur adjacently to each other (this is most likely due to the freedom in Greek word order); second, these elements have to be an auxiliary and a participle, because there can be combinations of two verbal elements constituting a syntactic unit, but such that they do not form periphrases in the sense that we intend: for example, patterns of ‘modal verbs 12 + infinite’ as in ‘Mark could answer the question, but he did not’.

Dietrich (1973, 203) also points out that aspectual periphrases (i. e. periphrases that primarily encode an aspectual value, most commonly a progressive one (‘to be doing something’); the notion of Aspect will be described in details in §1.3.2.) are typical for a 13 narrative pace, or, as he calls it, a narrative “Erzählungsmanier”: indeed, even though some instances of periphrastic constructions in our data appear in direct speeches or descriptive passages, plenty of them occurs in narrative contexts.

In contrast with Björck (1940), Blass & Debrunner (1997) explicitly point out that εἶναι (specifically, ἦν) + aorist participle conveys the same meaning as a pluperfect. While this is true for some instances (Blass & Debrunner, for example, cite Lc 23,19 - EvPe 23 - EvPe 51 - etc.), our data remarkably show that in some cases εἶναι + aorist participle encodes the same meaning as an aorist (e. g., AcPePa 47: ἦν δὲ ὁ Πέτρος εὐλογήσας […] ἄρτον, (‘and Peter blessed the bread […]’); see §2.4.3.).

As for ἔσεσθαι + present participle, Blass & Debrunner (1997) claim that it offers the chance to encode a durative meaning (i. e. describing a process or an action that last in time (‘to run’), in opposition with an instantaneous event (‘to die’); see §1.3.3.) in the future, that was – according to them – apparently impossible in Classical Greek (we do not feel as certain about this point, but it does not closely concern our purposes).

§1.1.4. More Recent Studies: Porter, Ceglia and Amenta

In an absolute sense, every occurrence of more than one element constituting a syntactic unit is, literally, a 12

‘periphrasis’.

Cf. Björck (1940), who claims that the only truly periphrastic constructions are those involving εἶναι and a 13

(20)

Porter (1989, 452-453) gives the following definition of ‘periphrasis’: “a periphrastic construction must contain an aspectually vague auxiliary verb [i. e. a verb that does not encode a definite type of Aspect (see §1.3.2.)] and a participle in agreement with its referent. The auxiliary verb must be aspectually vague to avoid aspectual conflict with the participle”. (Therefore, Porter implicitly argues that εἶναι is so common as auxiliary in periphrases because of its aspectual vagueness and its generic lexical meaning).

Moreover, Porter (1989) claims that the participle must be adjacent to its auxiliary, either before or after, and that they can be separated only by adjuncts or complements syntactically agreeing with the participle. This is indeed a condition that we have generally met in our data, except for very few instances where the constituents order is pretty confused: e. g., AcBa 15: ἦν δὲ ὁ Βαρνάβας µαθήµατα παρὰ Ματθαίου εἰληφὼς βίβλον τῆς τοῦ θεοῦ φωνῆς καὶ θαυµάτων καὶ διδαγµάτων σύγγραµµα (‘and Barnaba had learnt the teachings from a book of Matthew’s, a treatise about the voice of God and the amazing deeds and the teachings’; the order that we have followed in the translation is: ὁ Βαρνάβας δὲ ἦν εἰληφὼς [τὰ] µαθήµατα παρὰ βίβλον [τοῦ] Ματθαίου τῆς τοῦ θεοῦ φωνῆς καὶ θαυµάτων καὶ διδαγµάτων σύγγραµµα); we should not be surprised by this ‘twist’, though, as it is probably attributable to the freedom of the Greek word order. 14

Beside this, Porter (1989) points out the tendency of the Greek language not to employ different forms for one single function: this implies that a periphrastic construction should 15 encode a different shade of meaning, be it slight or not, in comparison with its synthetic counterpart.

Ceglia (1998) describes a periphrastic construction as the union of two (or more) elements, such that one is [+verbal][−lexical] and the other [−verbal][+lexical], being the former an auxiliary and the latter a participle.

Moreover, Ceglia (1998) distinguishes between verbal and adjectival periphrases: the latter type is not considerable as a periphrastic construction, because the participle involved underwent a process of category shifting (i. e. the participle, from its original condition of verb, turns into an adjective), that is particularly evident for neuter present participles, whose adjectival value is “come irrigidito (Ceglia (1998, 26): ‘kind of stiffened’)”: he refers to instances, really common even in Classical Greek, such as πρέπον ἐστὶ (‘it is fitting’) or ἔξον ἐστὶ (‘it is possible’). Furthermore, Ceglia (1998) argues that another feature that differentiates verbal periphrases from adjectival ones is the fact that, whereas in the former the auxiliary encodes Tense and Modality and the participle conveys both aspectual and lexical meanings, in the latter verbal information is totally encoded by the copula (Tense, Modality and Aspect) and the participle has got a purely lexical meaning (exactly like true adjectives).

Finally, as far as εἶναι + present participle is concerned, Ceglia (1998) observes that in the New Testament Greek this pattern displays a plurality of possible aspectual values (specifically, Progressive, Habitual and Ingressive: for detailed explanations see §1.3.2.): as Ceglia (1998, 32) remarkably claims, “in effetti è proprio la pluralità di valori aspettuali

connessi alla perifrasi con il participio presente a suscitare il legittimo sospetto che nessuno

For the word order of Greek, see Taylor (1994). 14

We may illustrate this concept as follows: “form X : function X = form Y : function Y”. 15

(21)

di questi valori sia specifico di questo costrutto; non è cioè l’espressione di un aspetto determinato il motivo per cui esso viene preferito a una forma sintetica (‘it is the plurality of

aspectual values encoded by the periphrasis with the present participle that raises the legitimate suspect that none of these values is specific for this construction; i. e. it is not the encoding of a determined aspect the reason why it is preferred to a synthetic form’)”. In agreement with Ceglia (1998), Drinka (2011, 46) argues that εἶναι + present participle periphrases occur both in Classical and New Testament Greek, but it is in the latter that we observe an increase in the semantic range: “while the progressive meaning […] can be found in Greek in limited fashion as early as Herodotus [Classical Greek], the habitual and ingressive meanings […] represent innovations […], since classical Greek would have used an imperfect and an aorist, respectively, in these contexts”.

Amenta (2003) claims that the most relevant issue in identifying periphrastic forms is to determine the level of auxiliarization reached by finite verb involved in the construction, that is its degree of grammaticalization as auxiliary. Following Heine (1993), Amenta 16 (2003, 21) claims that “secondo un approccio di tipo scalare, l’ausiliarità va intesa in

termini prototipici […] come categoria al cui interno possono essere comprese varie realizzazioni più o meno grammaticalizzate e cristallizzate nel loro uso grammaticale

(‘according to a scalar approach, auxiliarity should be intended in prototypical terms as a category within which different forms, more or less grammaticalized and crystallized in their grammatical usage, can be included’)”; Amenta (ibidem [note 1]) also specifies that 17 “i prototipi di ciascuna categoria sono quelle entità che presentano tutte le proprietà

prototipiche, cioè le caratteristiche che permettono di definire ciascuna categoria (‘the

prototypes of each category are those entities that display all the prototypical properties, that is the characteristics that allow to define each category’)”. 18

In fact, Amenta (following Heine) suggests three scalar parameters to determine the degree of auxiliarization:

1) the decrease and loss of phonological material (erosion) is found in the most fully grammaticalized forms, such as clitics or affixes: e. g., Lat. *cantare habeo > Ital.

canterò. This parameter, though, is not relevant in our analysis, because the

components of a periphrasis are always neatly divided (in our data, indeed, they are often not even contiguous).

2) Desemantizzazione (‘desemanticization’ or semantic bleaching): it is the loss of lexical meaning and acquisition of grammatical meaning: e. g., the auxiliary in Ital. sono malato (“I feel sick”) is less semantically bleached than in Ital. sono

arrivato (“I have arrived”), as the former sono is to be meant as the copula of a

More specifically, Heine (1993, 30) argues that “many instances of the [grammaticalization] process 16

involve at the same time more than one linguistic item”, rather than a single word or morpheme: for example, the process of grammaticalization of perfect aspects in many European languages involves an auxiliary (typically ‘have’ or ‘be’) and a non-finite form (typically a passive participle).

See also Ramat (1987) and Luraghi (1993). 17

See also Luraghi (1993). 18

(22)

nominal predicate, while the latter sono is a sheer auxiliary. By contrast, it is to 19 remark that Hopper & Traugott (1993) claim that in the grammaticalization process a redistribution or shift of meaning, rather than a loss, takes place, especially at the early stages of the process; consequently, “since the initial phase 20 of grammaticalization involves a shift in meaning, but not loss of meaning, it is unlikely that any instance of grammaticalization will involve a sudden loss of meaning (Hopper & Traugott (1993, 89))”.

3) Decategorizzazione (‘decategorialization’): it is the shifting of the status of the finite verb from ‘true’ verb with an argument structure to auxiliary or modifier. Hopper (1991) points out that this parameter entails the loss of some verbal properties (for example, a verb that has become an auxiliary or a modifier cannot be nominalized or take the passive Voice) and some syntactic restrictions (it does not have a free position in the sentence and it cannot be negated separately from the other element(s) of the periphrasis).

On the whole, though, the linearity of the grammaticalization process is just a tendency and, in fact, a verb becoming an auxiliary may display different behaviors in respect to these parameters: for instance, the auxiliaries of the periphrases that we are addressing show a high degree of desemanticization, but, as far as decategorialization is concerned, they display a quite free position in the sentence.

Amenta (2003) observes that in Ancient Greek there were some constructions, involving a finite verb and a participle, that cannot be considered periphrases: e. g., Plato, Theages, 122C: αἰσθώµεθα γελοῖοι ὄντες (‘we realized to be ridiculous’, in subjunctive Modality). She argues that the property that defines real periphrases and that differentiates them from non-periphrases is the fact that periphrastic constructions encode an Imperfective aspectual meaning (or Imperfective sub-categories; again, for a detailed explanation about Aspect, see §1.3.); in fact, it is not a coincidence that the finite verb that serves as auxiliary is found in Tenses that encoded an Imperfective aspectual meaning: present or imperfect (more precisely, the auxiliary can be found in future Tense too, but this Tense was aspectually neuter so it does not influence in any way the aspectual meaning of the whole periphrasis). However, we do not agree with the just mentioned claims for two reasons: first, the auxiliaries of the periphrastic constructions that we are addressing do not encode the category of Aspect, but they only convey Tense and Modality (as pointed out by Ceglia (1998), Aspect is encoded by the participle); thus, the selection of present or imperfect (or future) Tenses for the auxiliary neither involves nor affects the aspectual encoding of the whole periphrasis, but it only concerns its Tense. Second, we disagree with Amenta’s claim that real periphrases encode the Imperfective aspectual meaning only (or its sub-categories), because our data show periphrases encoding Perfective aspectual meanings: e. g., AcPhi 85: ἐλθόντες δὲ ἐν τῷ τόπῳ ὅπου ἦν ὁ παῖς ἐγηγερµένος, […] (‘and having arrived at the place

Dik (1987) doubts that a copula had a fully lexical meaning (what he calls “independent meaning”) and he 19

would rather describe the process of auxiliarization in terms of ‘expansion’ instead of ‘semantic bleaching’. His article, though, is about a very specific issue (copula auxiliarization) and a discussion about it would be quite irrelevant for our purposes. For further explanation, see Dik (1987).

By “shift or redistribution” Hopper & Traugott (1993, 96) mean “a balance between loss of older, typically 20

more concrete, meanings, and development of newer, more abstract ones that at a minimum cancel out the loss”.

(23)

where the child had woken up, […]’), that is a pluperfect, or EvThoA 7,2: τοῦτο τὸ παιδίον γηγενὴς οὐκ ἔστι, τοῦτο δύναται καὶ πῦρ δαµάσαι· τάχα τοῦτο πρὸ τῆς κοσµοποιίας ἐστὶν γεγεννηµένοι (‘this child is not earthly, he can even tame the fire: perhaps, he was born (literally ‘has been created’) before the creation of the universe’), that is a perfect.

Finally, Amenta (2003) motivates the occurrence of periphrastic constructions in terms of stylistic reasons and markedness (Aerts’ expressive periphrasis), rather than lexical necessity; for instance, periphrastic constructions are often used in the gospels in order to clearly divide between background and foreground information, being the former expressed by periphrases and the latter by synthetic forms: e. g., Lc 15,1-2: ἦσαν δὲ αὐτῷ ἐγγίζοντες πάντες οἱ τελω̃ναι καὶ οἱ ἁµαρτωλοὶ ἀκούειν αὐτου̃. καὶ διεγόγγυζον οἵ τε Φαρισαι̃οι καὶ οἱ γραµµατει̃ς λέγοντες […] (‘now all the publicans and sinners were drawing near unto him to hear him. And both the Pharisees and the scribes murmured, saying […]’).

§1.1.5. Bentein

According with Amenta (2003), Bentein (2012, 35) argues that “the category of verbal periphrasis is prototypically organized” and provides two possible different models:

1) Criterial-Attribute Model: with this model, “members are identified by means of a list of defining features and the category has fixed boundaries”;

2) Prototype Model: this “model recognizes that a category can have both central and more peripheral members, and that there are not always clear-cut boundaries”. In agreement with both Amenta and Bentein, we also believe that the status of the periphrastic constructions we are addressing is better described in prototypical terms (see Amenta (2003) and §1.1.4.). At the grammaticalization stage these periphrases are, it is very hard – not to say virtually impossible – to provide a satisfying grammatical description of them as belonging to a discrete category (as Bentein’s Criterial-Attibute Model would request), because of the large freedom displayed in their syntactic behavior and the 21 remarkable amount of semantic values that can be encoded (also Bentein (2013, 168) agrees on this point: “that the component parts of a construction are (syntactically / functionally) still comparatively «free» is typical for the early stages of grammaticalization”). This should be explained more clearly: what we mean by ‘freedom in syntactic behavior’ is that these periphrases do not always show a regular syntactic behavior: for example, in some cases the auxiliary precedes the participle and in other it is the contrary (AcAnMa 5A vs. AcAp 20,13), or in some cases the periphrasis “allows” more or less heavy lexical material to occur between the auxiliary and the participle while in other it does not (AcBa 15 (heavier) / AcIo 56 (less heavy) vs. AcAp 10,24). As for the semantic values that can be conveyed, these periphrases do not always encode the same grammatical meaning (the lexical meaning obviously depends on the verb used): the same construction may encode different values (e. g., ἦν διδάσκων conveys an Imperfective Progressive meaning in AcAnMa 21, but an Ingressive meaning in AcPhi 108), or, vice versa, the same meaning may be conveyed by different constructions (e. g., the Ingressive meaning is encoded by εἶναι + present participle in AcPhi 108, but by εἶναι + aorist participle in TeRu 1,4). Therefore, given the whole wide range of values and behaviors that these periphrases are capable to encode and display, we

Apart from the general observation that these periphrases are composed of an auxiliary and a participle in 21

Riferimenti

Documenti correlati

[r]

[r]

[r]

[r]

Answer: As a basis of Im A, take the first, fourth and fifth columns, so that A has rank 3. Find a basis of the kernel of A and show that it really is a basis... Does the union of

If S is not a linearly independent set of vectors, remove as many vectors as necessary to find a basis of its linear span and write the remaining vectors in S as a linear combination

Dipartimento di Metodi e Modelli, Matematici per le Scienze Applicate, Universit` a di Roma “La Sapienza”,

Sono i due nuovi coinquilini della vecchia casa di Thomas e Emma, che sentono “the piano playing / Just as a ghost might play”, infatti, il poeta