Save

The LXX and Historical Greek Phonology: Orthography, Phonology, and Transcriptions

In: Journal for the Study of Judaism
Author:
Benjamin Kantor University of Cambridge Cambridge United Kingdom

Search for other papers by Benjamin Kantor in
Current site
Google Scholar
PubMed Close
https://orcid.org/0000-0002-7691-7099
Online Publication Date:
28 Sep 2022
Open Access

Abstract

The growing trend to see the language of the LXX as an authentic example of post-Classical Greek may be extended to phonology and orthography. We can situate the phonology of the LXX within its historical Greek phonological context by implementing a restrictive methodology that focuses on transcribed names, the clusters of certain spelling conventions in relation to “early” and “late” books in the LXX, and manuscript-specific phenomena. We find that its language exhibits the same sort of phonological and orthographic features attested in contemporary documentary and epigraphic material. Codex Vaticanus provides the earliest explicit evidence for one of the notable phonological developments in the history of Greek, the fricativization of χ. It is demonstrated that the phonology of the LXX is right at home in its contemporary historical Greek phonological setting, and that it has unique contributions to make to the wider field of historical Greek phonology at large.

Keywords: Koine Greek; Septuagint; phonology; orthography; pronunciation; fricativization of χ

1 Introduction

In recent years, increasing attention has been given to seeing the language of the LXX as just one more example of authentic post-Classical Greek, albeit with some peculiarities—but far less than used to be supposed.1 It turns out that, when compared with contemporary documentary and epigraphic data, some apparent peculiarities of grammar (e.g., lexicon, syntax) are actually quite at home in their post-Classical Greek context. If, then, certain features of lexicon and syntax in the language of the LXX may be appropriately situated (diachronically) between Classical Greek and Byzantine Greek, on one hand, and situated (synchronically) among contemporary documentary and epigraphic evidence, on the other, can we not expect the same principles to apply when dealing with phonology? In short, the answer is yes, but the question itself requires a substantial degree of qualification and defining of terms.

What exactly do we mean by the phonology of LXX Greek? A Greek text by itself cannot have phonology. Although syntax, morphology, and lexicon are ipso facto encoded in a text, phonology is not. Only an author, a reader, or a language community—but not a text—can be regarded as having a phonology. Such phonologies can vary over space and time. A Greek speaker from Ephesus reading LXX Genesis pronounced it differently from the Jews of Alexandria. Philo’s (d. 50 CE) accent when reading ἐν ἀρχῇ ἐποίησεν ὁ θεός “in the beginning God created” would have sounded different from that of John Chrysostom (d. 407 CE).2 People and communities have pronunciation (i.e., phonology); texts have only spelling (i.e., orthography). Typically, spelling is fairly standardized and thus conservative. It does not normally reflect pronunciation. Nevertheless, sometimes developments in pronunciation can affect the spelling conventions used by scribes. While such spelling variation based on developments in pronunciation might show up immediately in lower-level texts written by inexperienced scribes,3 it can take generations for developments in pronunciation to affect more widespread (or “standardized”) spelling conventions.4 Moreover, while certain spelling conventions might have applied across a large swath of time or space—e.g., the Roman Mediterranean—the spelling conventions of any given scribe are determined in part by the school in which he was trained. All we can mean by the phonology of the LXX, then, is the pronunciations of the translators and their communities as they have come to be reflected in the spelling of the text of the Septuagint—but see below as to why we cannot or should not limit our study to such a definition.

The final sentence of the preceding paragraph, namely the spelling of the text of the Septuagint, introduces a whole new host of (text-critical) problems. Recovering the Urtext of any given piece of ancient literature is a difficult enough—some would say impossible—process as it is. This is the case when dealing with identifying which words were original to the text. πολλῷ οὖν μᾶλλον is it going to be an issue when it comes to reconstructing original spellings, given the propensity for scribes to update spelling on an almost subconscious level. The date of composition for the earliest portions of the LXX (i.e., the Pentateuch) goes back to the third century BCE, but our earliest substantial copies of the LXX as a corpus (i.e., witnesses containing multiple books) are not attested until five centuries later.5 There is little doubt that certain spelling conventions may have been updated between the composition of the Pentateuch and the time of manuscripts like Codex Vaticanus and Codex Sinaiticus. If any scribe in the chain of transmission copied by dictation, orthography was likely updated and conformed almost entirely to the conventions of the new scribe. Indeed, the whole endeavour of analyzing the phonology of the LXX on the basis of spelling can be fraught with insurmountable difficulties, especially when we consider what stage and/or version of the text(s) we want to consider the LXX.6

Despite these issues, we do not have to abandon the possibility of situating the phonology of LXX Greek in its historical context. Rather, we only have to restrict—albeit severely—how much we can say about its original context, on one hand, and modify what exactly we mean by its original context, on the other. While studies of syntax and lexicon might readily make reference to Egyptian Greek—or Judeo-Palestinian Greek, for that matter—of the Hellenistic (or Roman) period to explain a phenomenon in the LXX, such appeals can be problematic when dealing with phonology, since such claims about phonology would require identifying “original” spellings. In my view, there are three methodological principles that allow us to bypass such issues:

  1. Reconstructing “original” spellings is generally only permitted when dealing with transcriptions of foreign names, since these are less prone to being updated.7

  2. In rare cases, when certain spelling conventions cluster in “early” books over against “late” books, and vice versa, these may be regarded as “original” and significant for phonology.8

  3. In all other cases, the only meaningful discussions of the phonology of the LXX must depend on dated physical texts: i.e., papyrus fragments and early manuscript witnesses.

Principle (1) allows us to bypass the issue of orthographic conservatism, since the transcription of foreign names is likely to reflect the phonology and orthographic conventions of the period in which it was first transcribed. Moreover, after initial transcription, transcribed names are less likely to be updated orthographically by a scribe unless that scribe has knowledge of the source language (i.e., Hebrew or Aramaic). Principle (2) applies in cases where one specific manuscript (e.g., Codex Vaticanus, Codex Sinaiticus) exhibits distinct spelling conventions in early books over against late books. Although a complex transmission history could bring about such differences, if these differences also align with contemporary documentary and epigraphic evidence, they may relatively safely be regarded as phonologically significant. Principle (3) implies that, at least in some cases, a discussion about the phonology of the Greek of the LXX is really a discussion about the phonology of the scribes (and their scribal culture/communities) who transmitted the text of the LXX. For the purposes of the present study, we will focus on early papyrus fragments and manuscript witnesses up to (and including) Codex Vaticanus (early 4th c. CE) and Codex Sinaiticus (early 4th c. CE). Finally, it should also be noted that principle (2) is not mutually exclusive of the other two, since chronological clustering in “early” and “late” books is significant in any form of analysis.

In the remainder of this article, we will attempt to situate the phonology of the LXX—albeit with our somewhat (but necessarily) malleable definition of that concept—against the backdrop of post-Classical Greek documentary and epigraphic evidence. After a very brief summary of previous research on the phonology of the LXX (section 2), we will begin by outlining (briefly) the principal developments that occurred in Greek phonology during the “Koine” period (section 3). Following this, we will look at a series of features that demonstrate how the phonology of the language of the LXX coheres with contemporary documentary and epigraphic evidence. First, we will look at “original” spellings of transcribed names to provide a snapshot into the phonology of a translator and his community at the time of composition (section 4). Second, we will trace certain phonological and orthographic conventions in early witnesses from the second century BCE to the fourth century CE to demonstrate how contemporary developments in phonology and orthography can be traced over time in the early papyrus fragments and manuscript witnesses of the LXX (section 5). In this way, we provide a brief overview of what might be considered the phonology of the LXX: i.e., the phonology of the translators’ communities and the phonology of the early scribes and transmitters of the text.

2 History of Scholarship

One of the prevailing trends that emerges when reviewing previous scholarship on the phonology of the LXX is just how varied different approaches to the question can be:

Some scholars have approached the issue by noting instances where the phonological shape of the Greek word exhibits assonance with the sounds of the Hebrew word being rendered. Examples claimed to exhibit such a correspondence include κώλυε for תכלאי /tiklaːʔiː/ (Isa 43:6), σής for עש /ʕaːʃ/ or סס /seːs/ (Isa 51:8), καὶ ἀπορίᾳ ἀπορηθήεσται ἡ γῆ for פור התפוררה ארץ /poːr hitpoːrəraː ʔar(ə)ṣ/ and τρανή for תרן /tor(ə)n/ (Isa 35:6).9 While there does seem to be something to this approach, one must wonder to what degree a Greek-speaking audience would have appreciated such cleverness in translation and to what extent the translators needed to depend on this method of rendering.10 This approach might also overlook the fact that there may be more to the Greek usage of a particular lexeme than initially meets the eye. Moreover, from the perspective of the history of Greek, such an approach does not actually analyze the phonology of the LXX in a particularly precise manner, since such apparent correspondences in “phonological translation” can be quite loose.

Other scholars have looked at what might be considered more lexico-phonological or morpho-phonological features of the LXX against the backdrop of contemporary Greek evidence. Already at the beginning of the twentieth century, Thackeray drew attention to the use of words like οὐθείς vs. οὐδείς, μηθέν vs. μηδέν, τεσσαράκοντα vs. τεσσεράκοντα, and ταμιεῖον vs. ταμεῖον. The argument was essentially that, because one of each of these variants is attested later than the other, those LXX books that prefer the late variant might be dated later.11 More recently, Aitken has re-evaluated some of these examples, noting some problems with methodology and calling for a more nuanced approach that takes into account the much greater wealth of data now at our disposal. While phonological data may be relevant to some degree for dating LXX texts, too much weight has been afforded to the distribution of certain forms across the LXX corpus, which is relatively limited in its chronological span. The preservation (or alteration) of original spellings in the transmission of the text can also make phonological analysis problematic, since only the preservation of conservative features is likely to be given to a more straightforward interpretation. In short, though important, phonological evidence is inconclusive by itself (for dating) and must be marshalled cumulatively alongside other (primarily syntactic) evidence.12

Aitken’s caution against affording too much weight to the phonological data for the dating of LXX texts is appropriate. For that reason, the present article will attempt to circumvent most of these issues by focusing on transcribed Hebrew names, which are best equipped to provide a snapshot of (at least part of) the phonology of the translator at the time of composition. The penultimate section (5), on the other hand, will demonstrate just how tenuous spellings in native Greek words can be, since these are regularly updated according to the scribal conventions of the time, even in biblical texts.

3 Phonological Developments in Post-Classical Greek

Before proceeding to analyze the data from the LXX, it will be helpful to outline some of the primary phonological developments in post-Classical Greek. Naturally, the developments outlined here constitute an “idealized” generalization of the situation on the ground. In reality, there were many more nuanced differences from region to region. Nevertheless, we may treat the phonological developments in post-Classical Greek in three separate categories: consonants, diphthongs, and vowels.13

In the realm of the consonants, the following changes took place:

  • C.1. The original combination ζ = /zd/ shifted to /zz/ by the early Hellenistic period. Later, probably around the same time that gemination was generally simplified, /zz/ was further simplified to /z/. This likely took place by the Roman period. This is indicated in part by spelling interchanges like νομιζματος (for νομίσματος) and ζμυρνης (for σμύρνης).

  • C.2. The original voiced stops γ = /g/, β = /b/, δ = /d/ came to be pronounced as the fricatives γ = /ɣ, ʝ/, β = /β/, δ = /ð/. The fricativization of γ was first, occurring early in the Hellenistic period. The fricativization of β likely occurred by the first century CE. The fricativization of δ was last, occurring by the late Roman period. These changes are indicated in part by spelling interchanges like γατρος (for ἰατρός), σιλβανος (for Latin Silvanus), and ζοθηναι (for δοθῆναι). After nasal consonants, however, the stop pronunciation of each of these consonants was retained: e.g., ἐγγύς = [ɛŋˈgys̱]/[ɛ(ɲ)ˈɟys̠] (cf. ἐγώ = [ˈɛɣo̞]), ἐμβῆναι = [ɛmˈbenɛ] (cf. ἐνέβη = [ɛˈnɛβe]), ἀνδρός = [anˈdro̞s] (cf. ἁδρός = [(h)aˈðro̞s̱]). This retention of the stop pronunciation after nasals is indicated in part by spelling interchanges like ενκυς (for ἐγγύς), εμπηναι (for ἐμβῆναι), αντρος (for ἀνδρός).

  • C.3. The original voiceless aspirated stops χ = /kʰ/, φ = /pʰ/, θ = /tʰ/ also underwent fricativization to χ = /x, ç/, φ = /f/, θ = /θ/, though this occurred at a significantly later date. In some regions this change may already have occurred in the Roman period, while in other regions this change did not occur until the Byzantine period. These changes are indicated in part by spelling interchanges like Latin Dafne (for Greek Δάφνη) and φιλοτειμηθαμενος (for φιλοτῑμησάμενος).

  • C.4. Initial aspiration /h/, which is signified in modern print texts by the spiritus asper (e.g., ἁ), was lost. This change had begun by the Roman period, but there is evidence that speakers who pronounced /h/ existed alongside those who had already elided the sound during this period. This change is indicated in part by spelling interchanges like μετ ορκου (for μεθ᾽ ὅρκου).

  • C.5. When immediately following nasal consonants, the originally voiceless unaspirated stops κ = /k/, π = /p/, τ = /t/ came to be pronounced with voicing, so γκ = [ŋg], μπ = [mb], ντ = [nd]. This likely occurred during the Roman period. These changes are indicated in part by spelling interchanges like ενκυς (for ἐγγύς), εμπηναι (for ἐμβῆναι), αντρος (for ἀνδρός).

  • C.6. A general weakening of the nasals μ and ν occurred, especially in word-final and pre-stop position. The nasal could either elide entirely, assimilate to the following stop, and/or leave behind a nasalized vowel. This change may be dated to the Roman period, if not before. These changes are indicated in part by spelling interchanges like την … συνηθεα (for συνήθειαν) and προσηνεκκεν (for προσήνεγκεν).

  • C.7. Moveable ν, which historically was used when the following word began with a vowel and omitted when the following word began with a consonant, was generalized so that it occurred even before a following consonant. This change is indicated in part by spelling interchanges like απεθανεν δε (for ἀπέθανε δὲ).

  • C.8. Geminated consonants were generally simplified. While this likely began in some regions during the Hellenistic period, other regional varieties of Koine Greek apparently retained geminated consonants into the Roman period. This change is indicated in part by spelling interchanges like αλα (for ἄλλα) and γραματα (for γράμματα).

In the realm of the diphthongs, the following developments took place:

  • D.1. The final iota—today this is called iota subscriptum—in the long diphthongs ᾱι (ᾳ), ηι (ῃ), ωι (ῳ) elided so that these diphthongs were simplified to ᾱ, η, ω. This occurred during the Hellenistic period, though some of these might have barely survived into the Roman period. These changes are indicated in part by spelling interchanges like ετοιμω (for ἑτοίμωι).

  • D.2. The historically long diphthong ᾱυ (e.g., ἐμᾱυτοῦ), which first had simplified to ᾱ, later merged with the historically short diphthong ᾰυ. This later merger, perhaps brought about by analogy, likely occurred by the Roman period. This change is indicated in part by spelling interchanges like εματον (for ἐμᾱυτόν).

  • D.3. The diphthong represented by the digraph αι = /ai̯/ monophthongized to [æː] and then eventually lowered to [ɛː] during the Hellenistic period. After the neutralization of phonemic length in the Roman period, it merged with ε = /ɛ/. This change is indicated in part by spelling interchanges like δεομε σου (for δέομαί σου).

  • D.4. The diphthong represented by the digraph οι = /oi̯/ monophthongized to [øː] and then eventually merged with ῡ = /yː/. After the neutralization of phonemic length in the Roman period, it came to be realized as οι = /y/. This change is indicated in part by spelling interchanges like ανυγω (for ἀνοίγω) and προσβοιτεροι (for πρεσβύτεροι).

  • D.5. The second element of the diphthongs αυ = /au̯/, ευ = /ɛu̯/ (and possibly ηυ = /eːu̯/) became consonantal so that the historical diphthongs came to be pronounced as αυ = /aβ, aɸ/ and ευ = /ɛβ, ɛɸ/. This change began in the Hellenistic period and was complete in the Roman period. This change is indicated in part by spelling interchanges like προσαγορεβσε (for προσαγορεῦσαι).

In the realm of the vowels, the following developments took place:

  • V.1. In the early Hellenistic period, if not already in Great Attic (which would become the “Koine”), the vowel η = /ɛː/ raised to /e̞ː/ and the vowel ω = /ɔː/ raised to /o̞ː/. These changes are indicated in part by spelling interchanges like ωδη (for ὧδε) and εχο (for ἔχω).

  • V.2. Early on in the Hellenistic period, the vowel represented by ε likely lowered from /e̞/ to /ɛ/. In other regional varieties, however, it is probable that ε continued to be pronounced as /e̞/. This change is indicated in part by spelling interchanges like τραπαιζιτ[ου] (for τραπεζίτου).

  • V.3. At a relatively late stage, the vowel represented by η, which had already raised from /ɛː/ to /e̞ː/ during the Hellenistic period, continued to raise through /e/ until it eventually merged with ι = /i/. The timing of this change depends on the region, with some varieties exhibiting it as early as the first century CE and others maintaining η = /e/ all the way through the fifth century CE. This change is indicated in part by spelling interchanges like εγραφι (for ἐγράφη) and μηκρων (for μῑκρῶν).

  • V.4. Although the digraph ει had come to represent [iː] generally and [eː] pre-vocalically in Great Attic, it was not until the late Hellenistic period or early Roman period that ει merged with etymologically long ῑ = [iː] in pre-vocalic contexts as well. Eventually, after vowel length stopped being phonemically distinctive (see V.5), ει would come to be pronounced simply as /i/. The early situation (non-pre-vocalic ει = [iː] and pre-vocalic ει = [eː]) is indicated by spelling interchanges like μεικρος (for μῑκρός) alongside interchanges like ηδηα (for ἡδεῖα). Eventually, however, spelling interchanges like ιην (for εἴην) demonstrate that pre-vocalic ει had also come to be pronounced as [iː]. Interestingly, in the Roman period, the use of ει to represent etymologically long ῑ (e.g., τειμη (for τῑμή))—but not etymologically short ῐ—comes to be one of the most popular “Roman” scribal conventions of the period.

  • V.5. At some point during the Koine period, phonemic vowel length distinctions were neutralized, so that there was no distinction in duration between long and short vowels. This phenomenon may be referred to as the onset of isochrony. Different regional varieties attest to the onset of isochrony at different times—e.g., it happened earlier in Egypt than it did in Attica—but it was generally on the rise by the first century CE and the norm by the second century CE. This change is indicated in part by spelling interchanges like ανοθεν (for ἄνωθεν), εχωμεν (for ἔχομεν), πατρει (for πατρῐ́), and δοτη (for δότε).

  • V.6. The shift from a pitch accent, which indicates syllable prominence by means of high and low pitches, to a stress accent, which indicates syllable prominence by features like longer duration, occurred during the Hellenistic period or early Roman period depending on the region. There is indirect evidence for this phenomenon, but it deserves a fuller discussion than can be presented given the scope of the present work.14

4 Phonological and Orthographic Features in Transcribed Names in the LXX

As noted above, it is typically only in the realm of transcribed names that “original” spellings can be fairly reliably reconstructed. While focusing on transcriptions introduces greater complexity, it also allows for a type of diachronic analysis that would otherwise be difficult or impossible. In some cases, careful attention to transcription conventions attested in “original” spellings may confirm the proposed dating of conventionally early (e.g., Pentateuch) and late (e.g., 2 Esdras) compositions (see 4.1 below). In other cases, a particular orthographic convention attested in an original spelling of a transcribed name may convey something about the phonology of the dialect of the author (see 4.2 below). Finally, certain transcription conventions that likely arose secondarily in transmission history may convey something about the phonology of the scribe and the dating of the particular manuscript witness, as in the case of Codex Vaticanus (see 4.3 below).

4.1 The Nature of Pre-Vocalic [eː] → ει or η in Transcribed Names in the LXX

In transcribed names in the LXX, Hebrew long /eː/ normally corresponds with Greek η: e.g., ישראל /jisraːʔeːl/ → ισραηλ, קדר /qeːdaːr/ → κηδαρ, קדש /qaːdeːʃ/ → καδης. There is one particular name, however, which does not follow this standard transcription convention. The name לֵאָה “Leah,” which would have been pronounced in contemporary (i.e., Second Temple period) Hebrew as /leːʔaː/ [leːˈʔɑː],15 is transcribed in the LXX as λεια rather than the expected form **ληα. Although it appears throughout the Bible (Gen 29, 30, 31, 33, 34, 35, 46, 49; Ruth 4; 4 Mac 8), its first appearance is in Genesis, which was one of the earliest books of the LXX to be translated, likely taking place in Alexandria—but not necessarily in Alexandria—at some point in the third century BCE.16 Even though both Ruth and 4 Maccabees were likely translated and composed, respectively, at a significantly later date,17 it is likely that the transcriptions of Genesis established a standardized spelling convention for this name that was imitated by subsequent scribes.

During the period in which the Pentateuch was translated in Egypt, the digraph ει represented a long front close vowel [iː] generally but a long front close-mid vowel [eː] in pre-vocalic position (see V.4 above). This is demonstrated by interchanges found in the papyri like την φυτηαν (for τὴν φυτεῖαν) (P.Petrie 2.32.2a, 238 BCE), σκαφηα (for σκαφεῖα) (P.Tebt. 3.815, 223/222 BCE), αρχηου (for ἀρχείου) (P.Tebt. 3.821.5, 209 BCE), and πληω (for πλείω) (UPZ 1.77.2.17, 161 BCE).18 At the same time, an abundance of interchanges found in the papyri like πολι (for πόλει) (PSI 4.402, 263–229 BCE) and υποκιται (for ὑπόκειται) (UPZ 1.110, 164 BCE) demonstrates that ει represented [iː] elsewhere.19 It was probably not until the Roman period that pre-vocalic ει also came to represent /iː/. After the onset of isochrony (see V.5), this would have been further simplified to /i/.20 It is also worth mentioning that the standard convention for transcribing a foreign [eː] sound in the LXX, namely the vocalic grapheme η, was also undergoing change during these periods. In Egypt, η was raising from [ɛː] to [eː] in the early Hellenistic period. In fact, during the earliest stages of this period, it is possible that in pre-vocalic environments ει was realized as [eː] and η was realized slightly lower as [e̞ː], though homophony is also plausible. Eventually, as η continued to raise, it would merge with ι in the late Roman period.21 In Palestine, however, its merger with ι did not take place until the middle of the Byzantine period.22

At the time of the translation of Genesis, then, Egyptian Greek still exhibited two distinct sounds represented by ει: [iː] generally and [eː] pre-vocalically. The form λεια would thus have been pronounced as [leːɑ(ː)] according to normal phonological and orthographic conventions of early Hellenistic Egyptian Greek.23 Indeed, the name λεια is the only case of ει corresponding to Hebrew long /eː/ in transcribed names in all of the Pentateuch. Everywhere else in the Pentateuch,24 ει in a transcribed name—never in pre-vocalic position—corresponds to long [iː] in Hebrew phonology:

FIG000001

Citation: Journal for the Study of Judaism 54, 4-5 (2023) ; 10.1163/15700631-bja10060

Although we have only one example of a pre-vocalic ει in the LXX Pentateuch transcriptions of Hebrew names (i.e., λεια), it is notable that the distribution of the phonological and orthographic conventions in the LXX Pentateuch matches precisely the phonological and orthographic situation of early Hellenistic Egyptian Greek evidenced in the papyri. The fact that לֵאָה “Leah” was not transcribed as **ληα during this early period may be due to the fact that η was still in the process of raising from [ɛː] to [eː] and had not yet fully raised to [eː], perhaps being realized in the area of [e̞ː]. In that case, ει would have been a better approximation of Hebrew long /eː/, but only in pre-vocalic environments. Because ει would have represented [iː] everywhere else, η would have normally been the best approximation of Hebrew /eː/ in all non-pre-vocalic environments.

There is one final piece of evidence to be added to our discussion of לֵאָה “Leah.” In a funerary inscription from Caesarea Maritima in Palestine dated to the late Roman or Byzantine period (3rd–6th c. CE), the Hebrew name לֵאָה “Leah” is transcribed into Greek as ληα.25 Since funerary inscriptions are not always written by the most professional scribes and thus tend to exhibit more innovative scribal tendencies—they are not as conservative or standardized as monumental inscriptions, for example—this spelling may be regarded as more “up-to-date” with contemporary phonology. In Palestinian Greek of this period, ει would have represented /i/ and η would have represented /e/.26 The form ληα, then, is the best contemporary rendering of Hebrew /leːʔaː/.

Although there are no exact parallels to this sort of pattern in later books of the LXX, incongruence between the Tiberian vocalisation tradition and the LXX transcriptions may actually offer a couple parallels in 2 Esdras, which is thought to have been translated in the latter part of the second century CE, likely in Palestine (though Egypt is also a possibility).27 The form צחא, which is vocalized as צִחָא (Neh 7:46) in the Tiberian tradition, is transcribed as σηα in the LXX. If the LXX transcription actually points to an original Hebrew pattern of /ṣeːħaː/, then this would provide a late example of pre-vocalic (from the perspective of Greek transcription) Hebrew long /eː/ being transcribed into Greek with η rather than ει as in the case of λεια. Similarly, the form מַעֲשֵׂיָה “Maaseiah,” which was likely pronounced as /maʕseːjaː/ [mɑʕ(ɑ)seː(j̞)ɑː] in contemporary Hebrew, is transcribed in 2 Esdras as μαασηα (2 Esdr 10:30). Although there are some significant text-critical issues concerning the ending of this word28 —not to mention the omission in transcription of the Hebrew glide /j/—it may still be significant that the Greek grapheme used to represent the long Hebrew vowel /eː/ in pre-vocalic position (from a Greek perspective) is η rather than ει. As we noted above, at the time of the composition of 2 Esdras, ει represented /i(ː)/ in all environments and η represented /e(ː)/, both in Egyptian and Palestinian Greek.29 The grapheme η would have been the only option for representing the sound [eː].

It is thus significant that, when we trace the representation of pre-vocalic [eː] in transcribed names in the LXX and contemporary extra-biblical evidence, we find that the distribution of ει for long [eː] and η for long [eː] coincides with what we know about the historical development of Greek phonology in Egypt and Judea-Palestine. The early books of the LXX cohere with early Hellenistic Egyptian Greek, whereas the late books (i.e., 2 Esdras) cohere with Roman Judeo-Palestinian Greek.

4.2 Nasal Weakening and Subsequent Gemination in Transcribed Names in the LXX

In transcribed names in the LXX, the sequence μβ normally only occurs due to excrescence: e.g., ממרא /mamreː/ → μαμβρη and זמרי /zamriː/ → ζαμβρι.30 There is one instance, however, in which the sequence μβ appears in intervocalic position: חבקוק /ħabbaːquːq?/ (see below) → αμβακουμ. The original Hebrew pattern of this word is a bit difficult to reconstruct. While both Codex Vaticanus and Codex Sinaiticus have αμβακουμ, some other manuscripts have αββακουμ. In early patristic writings, there is significant attestation of both αμβακουμ and αββακουμ.31 There is even the odd citation of αββακουκ in the sixth-century CE writings of Dorotheus of Gaza.32 All that to say, even though the Tiberian vocalisation tradition exhibits a qătalluːl pattern, it is plausible that a qattaːluːl pattern existed alongside a qătalluːl pattern for this name in Late Antiquity. Assuming that the LXX form reflects the qattaːluːl pattern—and does not preserve the name of a plant from Akkadian ḫambaḳûḳu as some scholars suggest33 —then the sequence μβ in the LXX would correspond to a geminated [bb] in Hebrew.34

Dating of LXX Habakkuk is not precise, but it may plausibly be assigned to second-century BCE Egypt.35 It is likely that the weakening of pre-stop nasals (see C.6 above) had already begun in Hellenistic Egyptian Greek by this time. As nasals weakened in sequences like μβ, there were generally three possible outcomes: (i) the nasal elided entirely so that a word like λαμβάνω would be pronounced as [laˈbano̞], (ii) the nasal elided leaving behind a nasalized vowel so that λαμβάνω would be pronounced as [lãˈbano̞], or (iii) the nasal assimilated to the following stop causing gemination so that λαμβάνω would be pronounced as [labˈbano̞].36 These various developments can be seen in some of the third- and second-century BCE documentary evidence from Egypt: e.g., λαβανον (for λαμβάνον) (P.Cairo Zen. 4.59646, 248/247 BCE), αποτυπανιειν (for ἀποτυμπανιεῖν) (P.Enteux. 86, 221 BCE), συβολης (for συμβολῆς) (SB 22.15238, 225–175 BCE), συβεβλημενωι (for συμβεβλημένωι) (P.Erasm. 2.57, 150/149 BCE).37 Although it is not until the Roman period that we find clear examples of μβ → ββ—note μεταλαββανων (for μεταλαμβάνων) (P.Oslo 3.153, 100–125 CE)—it is probable that the weakened nasals would have brought about gemination at an earlier stage as well.38 After all, the neutralization of phonemic gemination of consonants seems to have already begun in Egyptian Greek during the Hellenistic period.39 As such, interchanges like μβ → ββ, even if pronounced as [bb], would be less likely to appear in the orthography.

In light of the contemporary documentary evidence from Hellenistic Egypt, it is thus not improbable that the sequence μβ in the transcription αμβακουμ represented a geminated [bb], as was prone to happen in certain Greek words due to the weakening of the nasals. The fact that it was not represented by ββ—note that other Hebrew words with geminated /bb/ like רַבָּה /rabbaː/ → ραββα are represented this way in the LXX—may be due to the fact that the dialect of the translator of LXX Habakkuk might have differed from that of other translators, whether chronologically or regionally. What exactly does that mean? In his variety of Greek, two developments might have already occurred that led to his rendering [bb] with μβ: (i) the neutralization of phonemic consonant gemination and (ii) the fricativization of β. If both (i) and (ii) had occurred, then he would have been faced with the following transcription choices for rendering Hebrew [bb]: Greek β, which would have been pronounced as [β], Greek ββ, which also would have been pronounced as [β], or Greek μβ, which would have been pronounced as [b], [_̃b], or [bb]. Clearly, if both the neutralization of phonemic gemination and fricativization of β had already occurred in the dialect of the translator of LXX Habakkuk, μβ would have been the best convention for (phonetically) representing [bb] (see C.2), if we assume that the translator did not want to maintain ββ for [bb] as a conservative transcription convention.40

We actually see a similar transcription convention in use in Judeo-Palestinian Greek epigraphy at a later date. In the Byzantine period, the Hebrew/Aramaic word אַבָּא /ʔabbaː/, which has been loaned into Greek as a title for monks, is often written in inscriptions as αμβα.41 At that point in time, because phonemic gemination had already been neutralized in Judeo-Palestinian Greek,42 the sequence αββα would have been pronounced merely as [ɑβɑ]. However, due to the historical weakening of nasals and subsequent assimilation to a following stop, a sequence like αμβα could have been read as [ɑbbɑ], if not [ɑ̃bɑ] or [ɑbɑ]. Therefore, because the gemination was likely preserved in this title for Aramaic-Greek bilinguals, the best way to represent it in Greek was with μβ rather than ββ. Because Egypt was likely one of the first regional varieties of Koine Greek to neutralize phonemic gemination43 —it probably did so earlier than Judea-Palestine—this same phenomenon might have occurred there already at an earlier stage. For the translator of LXX Habakkuk, a transcription like αββακουμ might have been read as [ɑβɑkum], whereas a spelling like αμβακουμ would have been read as [ɑ(b)bɑkum] or [ɑ̃bɑkum].

4.3 Fricativization of χ and Its Use in Transcription of Names in the LXX

Dating the fricativization of the historically aspirated voiceless plosive χ = /kʰ/ in post-Classical Greek is notoriously difficult. Although it is widely acknowledged that χ, φ, θ must have become fricatives by the Byzantine period in most quarters of the Greek-speaking world, most scholars typically just cite evidence for the fricativization of φ and θ and assume that χ must have become a fricative at roughly the same time as well. The only unambiguous transcriptional data cited in scholarship for the fricativization of χ appears to be dated to the medieval period.44 In fact, until this paper, there does not appear to be any explicit data cited anywhere for a fricative χ dated to the Roman period.

It just so happens, however, that a transcription convention implemented only in the LXX of Codex Vaticanus actually fills just such a desideratum, namely how the Hebrew sound /ʃ/ is transcribed in certain letter names in translations of Hebrew acrostic poems. For a little background, there are certain poetic sections in the Hebrew Bible that are written acrostically. This means that every line in a particular section begins with a particular Hebrew letter, beginning with א (ʔalef), continuing with ב (beth), ג (gimel), and so on. In some LXX manuscripts, each of these sections will be marked by a header comprised of the Hebrew letter name transcribed into Greek: αλεφ, βηθ, γιμελ, etc. This is most common in Psalm 119 (118 in the LXX) and in the book of Lamentations.

In LXX Lamentations in Codex Vaticanus, some of these transcribed letter names actually present unique evidence for the historical development of the Greek consonant χ. In particular, the Hebrew letter names resh (ר) /reːʃ/ and shen (ש) /ʃen/ are transcribed as ῥήχς and χσέν, respectively (fol. 1134–1139). The use of χσ as a convention for transcribing /ʃ/ is unlikely unless χ had already shifted to a fricative. As a fricative, χ would have been realized as [x] generally but probably (at least in some dialects) with a palatal realization as [ç] before front high vowels.45 As far as cross-language perception goes, the sound [ç] can sometimes be regarded as loosely equivalent or perceptually similar to [ʃ].46 In fact, during the Byzantine period, there is evidence that both the Old Athenian group of Greek dialects and a number of insular Greek dialects even progressed the fricative χ all the way through [ç] to [ʃ] or [s].47 Further support for the perceptual equivalence of χ = /x, ç/ and Hebrew שׁ = /ʃ/ is found in the fact that χ by itself can even be used in certain contexts to transcribe Hebrew /ʃ/ during the medieval period: e.g., βερεχὶθ for /bereːʃiːθ/; νεχικὸθ for /neʃiːqoːθ/.48 It seems, therefore, that the scribe of Codex Vaticanus (or its ancestor), who had advanced knowledge of Hebrew phonology,49 thus regarded χσ = [xs̱]/[çs̱] as a better perceptual equivalent of Hebrew /ʃ/ than Greek σ = [s̱]. If χ was still realized as [kʰ], such a perceptual equivalence between χσ = [kʰs̱] and Hebrew /ʃ/ would be extremely unlikely. If, on the other hand, Greek χ had already become a fricative, the perceptual equivalence is entirely plausible. Given the date of Codex Vaticanus, the transcription convention of Greek χσ for Hebrew /ʃ/ may be taken as explicit evidence for the fricativization of χ by the end of the Roman period in the dialect of the scribe.

To date, this appears to be the earliest explicit evidence for a fricative pronunciation of χ cited in scholarship on historical Greek phonology. On this point, in the spirit of the present volume, I cannot help but call attention to the fact that this is not merely a case of data from the LXX proving consistent with contemporary Greek evidence attested in extra-biblical sources. Rather, this is a case where the LXX provides a unique datum that actually advances the wider field of historical Greek phonology. Indeed, it is worth stressing that as a legitimate example of post-classical Greek, the language of the LXX does not only cohere with extra-biblical Greek sources but is actually able to proffer unique and significant data that make meaningful contributions to the field of Greek linguistics and historical phonology at large. Historical treatments of Greek phonology may now cite χσεν and ρηχς in the LXX of Codex Vaticanus as evidence for the fricativization of Greek χ by the end of the Roman period.

5 Phonological and Orthographic Features in Native Greek Words in LXX Transmission

The first three examples each dealt with developments in Greek phonology and orthography as they came to be represented in transcribed Hebrew names in the LXX. Due to the nature of transcribed names, this analysis could proceed on the assumption that reconstructing “original” spellings was not an impossible endeavour. The present section, on the other hand, constitutes an analysis of Greek spelling conventions attested in native Greek words in the LXX. However, any analysis that depends on establishing “original” spellings of native Greek words in the LXX is at least somewhat tenuous. As such, the present section will address “LXX phonology” by tracing certain spelling conventions over time as they are attested in dated LXX papyrus fragments and early manuscripts. These evolving spelling conventions will be set against the backdrop of contemporary Greek documentary and epigraphic evidence. In other words, while “LXX phonology” in the previous section meant the phonology of the author at the time of composition, the term “LXX phonology” in the present section means the phonology of the scribes of the LXX during transmission. We will look at two illustrative examples:

5.1 Iota subscriptum in Dated Fragments and Manuscripts of the LXX

Beginning Greek students are taught that the historically long diphthongs ᾱι, ηι, ωι are to be written as ᾳ, ῃ, ῳ with a diacritic mark known as iota subscriptum ( ͅ)—e.g., ἐν ἀρχῇ, τῇ ἀγορᾷ, τῷ λόγῳ—but iota subscriptum is not to be pronounced. It should be noted, however, that iota subscriptum was not actually invented until medieval times.50 For our purposes, it is essentially a “modern” convention.

In the ancient period, iota was written on the same line or not written at all—i.e., εν αρχηι, τηι αγοραι, τωι λογωι vs. εν αρχη, τη αγορα, τω λογω—the former (also called iota adscriptum) being characteristic of the Hellenistic period and prior and the latter being characteristic of the Roman period and after. This shift in orthographic conventions reflects a development in phonology: i.e., ᾱι → α reflects [aːi̯] → [aː]; ηι → η reflects [ɛːi̯]/[eːi̯] → [eː]; ωι → ω reflects [ɔːi̯]/[o̞ːi̯] → [o̞ː] (see D.1 above). We noted earlier, however, that phonological developments often precede shifts in orthographic conventions by several generations. So, while the form without iota does not become the orthographic norm until the Roman period, the data suggest that these monophthongizations had already taken place in Egypt by the third century BCE.51 This is demonstrated by occasional interchanges already in early Hellenistic documentary evidence: e.g., δυνη (for δύνηι) (PSI 6.584, 259–229 BCE), πανοικια (for πανοικίαι) (SB 5 7524.8, ca. 249 BCE), ετοιμω (for ἑτοίμωι) (P.Hib. 2.253.8, ca. 250 BCE).52 In Attica, the evidence suggests that ηι → η occurred by the fourth or third century BCE, then ᾱι → α and ωι → ω around the second century BCE.53 While various regional varieties of post-Classical Greek exhibit slightly different phonological developments and orthographic conventions for these historically long diphthongs ending in iota, the general pattern is that these forms were phonologically monophthongized during the Hellenistic period so that by the late Roman period the orthographic norm is to write them without iota. In the Greek of Judea-Palestine, for example, the forms like -ηι (-ῃ) and -ωι (-ῳ) with iota tend to be attested at about a 75% rate in the Hellenistic period, whereas the forms like -η and -ω without iota tend to be attested at about a 90% rate in the Roman period.54

As we might expect, this pattern is mirrored in the earliest witnesses of the LXX. This may be demonstrated by a survey of the papyrus fragments and early manuscript witnesses of the LXX up to Codex Vaticanus and Codex Sinaiticus. In the survey below, the situation regarding ᾱι, ηι, ωι is briefly described for each LXX witness alongside several examples. Examples that show iota being written are marked in red and examples that show iota being absent are marked in green:

2nd c. BCE—P.Ryl. 458 (Rahlfs 957)55

Examples: [αριμθω]ι τεσσαρακοντα

The fragment exhibits one case of a historically long diphthong ending in iota, which is written with final iota.

2nd /1st c. BCE—P. Fouad 266 (Rahlfs 847, 848, 942)56

Examples: [ε]ν ανομιαι, δωσις αυτωι, τωι ονοματι, παρα·δωι [σοι], εν φονωι μαχαιρας, υπαρχηι, τηι πολει, [εαυ]τωι, [κλη]ρωι, εν τηι οδωι, σεαυτωι, etc.

The historically long diphthongs ηι, ωι, and αι are attested, all of which are written with iota.

125–1 BCE—4Q119 (Rahlfs 801)57

Examples: [τ]ηι γηι υμων εν καιρωι αυ[του], [μαχαιρα]ι

The historically long diphthongs ηι, ωι, αι are all attested and written with iota.

1st c. BCE—4Q120 (Rahlfs 802)58

Examples: [εν κλιβαν]ωι, [εν ελαι]ωι, [κυ]κλωι, [τωι] δακτυλωι, αυτωι, αποδωι, etc.

The historically long diphthongs ηι and ωι are attested and written with iota.

40 BCE–10 CE—4Q121 (Rahlfs 803)59

Examples: επ αυτηι, [καλυμμα]τι δερματινωι υ[α]κινθιν[ωι]

The historically long diphthongs ηι and ωι are attested and written with iota.

50 BCE–50 CE—8HevXIIgr (Rahlfs 943a, 943b)60

Examples: επι τη [κ]ακια, αδης, ελθη, εν μεσω, etc.

The historically long diphthongs ηι, ωι, αι are attested and never written with iota. What is more, there is even a case of a hypercorrection, where word-final η is instead written with ηι: [παρ]εκληθηι (Jonah 3:10). This likely indicates that the scribe may have even wanted to implement a more conservative usage of final iota but had not been trained to do so.

150–200 CE—P.CtYBR inv. 419 (Rahlfs 814)61

Examples: τη πολει, εν τη ερημω, τω περατη, τη δρυι τη μαμβρη

The historically long diphthongs ηι and ωι are attested, both of which are written without iota.

2nd c. CE—Chester Beatty Library, Pap. VI. (Rahlfs 963)62

Examples: ζηλωση (Num 5:14), τη γυναικ[ι] (Num 5:19), τω λαω (Deut. 2:4), etc.

The historically long diphthongs ending in iota are written without iota.

early 3rd c. CE—Princeton, University Library, Scheide Pap. 1 (Rahlfs 967)63

Examples: εν τω διασκορπισαι, εν μεσω, εν ημερα, εν ημερα μια, etc.

The historically long diphthongs ending in iota are written without iota.

250–300 CE—Chester Beatty Library, Pap. V. (Rahlfs 962)64

Examples: υιω, τη ο[ψει], και εσται η παρθενος η αν εγω ειπω, etc.

The historically long diphthongs ending in iota are written without iota.

4th c. CE—Codex Vaticanus (Vat. gr. 1209) (B)65

Examples: εν τη ερημω (Deut. 1:1), εν τω τεσσερακοστω ετει εν τω ενδεκατω μηνι μια του μηνος (Deut. 1.3), etc.

The historically long diphthongs ending in iota are written without iota.

4th c. CE—Codex Sinaiticus (BL Add. 43725) (א)66

Examples: συν τη κοπρω (Num 19:5), τη συναγωγη (Num 19.9), εν αυτω (Num 19:13.), etc.

The historically long diphthongs ending in iota are written without iota.

It is immediately apparent from this brief survey of papyrus fragments and manuscript witnesses of the LXX that the orthograhpic norm for ηι (ῃ), ωι (ῳ), ᾱι (ᾳ) in the Hellenistic period is ηι, ωι, αι, whereas the orthographic norm for ηι (ῃ), ωι (ῳ), ᾱι (ᾳ) in the Roman period is η, ω, α. The papyrus fragment of the Twelve Prophets (8HevXIIgr), which is dated to the transition from Hellenistic to Roman times, is the earliest LXX witness that exhibits the η, ω, α convention. From that point on, the absence of iota is regular. Though it is not surprising, it is worth underscoring that the phonological and orthographic conventions exhibited in early LXX papyrus fragments and witnesses match up exactly with the data attested in non-biblical contemporary documentary and epigraphic evidence.

5.2 ει for Etymologically Long ῑ in Dated Fragments and Manuscripts of the LXX

Some of the most common words in post-Classical Greek, such as γῑνώσκω, γῑ́νομαι, and ἀναγῑνώσκω, actually exhibit considerable variation in spelling in the ancient period. Some scribes tended to write γινωσκω, γινομαι, and αναγινωσκω, whereas other scribes would write γεινωσκω, γεινομαι, and αναγεινωσκω. At first glance, one might be quick to chalk up such variation merely to the general equivalence of ει and ι in the post-Classical period, but the data do not actually support this. In documentary and epigraphic texts throughout the Roman Mediterranean, we do not find ι interchanging with ει generally. Rather, interchanges of ι → ει are normally found in cases where ι goes back to etymologically long ῑ. While it is true that etymologically short ῐ also interchanges with ει, the interchange ῑ → ει is about ten to fifteen times more frequent than the interchange ῐ → ει in the Roman period.67 Indeed, throughout the Greek-speaking Mediterranean, the usage of ει to represent etymologically long ῑ appears to have been a popular scribal convention, which was first introduced in the Hellenistic period but flourished especially during the Roman period.

In Egyptian Greek, the ῑ → ει interchange is attested in the Roman period in words like τειμην (for τῑμήν) and μεικρος (for μῑκρός).68 In Italy, in Greek transcriptions of Latin in the Jewish catacombs of Rome, Latin ī → Greek ει is far more common than Latin ĭ → Greek ει.69 In Anatolian Greek, etymologically long ῑ is almost always written as ει during the Roman period, as in spellings like ετειμησαν (for ἐτῑμησαν) and νεικη (for νῑ́κη). Brixhe notes that this was a convention introduced at the beginning of the Hellenistic period, but he does not specify its relative frequency over time.70 In Attic, the popularity of the scribal convention ῑ (= [iˑ]) → ει increases significantly in the early Roman period; it was not used very extensively in the Hellenistic period. Already by the mid-first century CE, ephebic texts attest to ῑ → ει more frequently than ῑ → ι. In the second century CE, ει becomes the normal orthography for ῑ, being attested roughly twice as much as ι for ῑ. After 200 CE, ῑ → ει is roughly ten times more common than ῑ → ι. This convention is frequently attested in μειλήσιος, τειμ-, νεικ-, ἐλευσείνιος, ὁπλείτης, ἀφροδείσιος, and names derived from ῑ῎σῐς (e.g., εἰσίδωρος).71 In Judeo-Palestinian Greek, ῐ → ει is attested 0.8% of the time in the Hellenistic period but 2.6% of the time in the Roman period. ῑ → ει, on the other hand, is attested 3.8% of the time in the Hellenistic period but 36.5% of the time in the Roman period. By the Byzantine period, however, the convention ῑ → ει becomes far less common, being attested at only a 2.1% rate and ῐ → ει at a 1% rate. The ῑ → ει spelling convention, then, is a feature overwhelmingly characteristic of the Roman period.72

In sum, all around the Mediterranean we find a steep increase in the frequency of ῑ → ει in the transition from the Hellenistic period to the Roman period. In fact, it becomes more common than just about any other scribal interchange during the Roman period. While it lies far beyond the scope of the present article, I have argued elsewhere (with Clackson) that this chronological distribution is no accident. Rather, it reflects the imitation, adoption, and expansion of a Roman scribal practice. Its specific use in orthography all the way into the late Roman period reflects—embedded in the scribal curriculum—educated pronunciation prescriptions that still carefully maintained a distinction between long ῑ and short ῐ. This is in contrast to the common pronunciation system, in which isochrony was already becoming prevalent in the early Roman period.73 The prevalence of this convention in the Roman period, then, reflects a scribal tradition influenced both by Roman scribal practices and (in the late Roman period) by “Atticist”-like attempts to maintain a more conservative pronunciation.

Although not as straightforward as the previous example, we see a similar chronological distribution in the early witnesses of the LXX. Similar to before, a selection of early papyrus fragments and manuscripts is surveyed below to illustrate the distribution of this feature in early LXX witnesses. Examples that have ῑ → ι are marked in red and examples that have ῑ → ει are marked in green:

2nd c. BCE—P.Ryl. 458 (Rahlfs 957)74

Examples: μαστιγωσιν (Deut 25), μαστιγω[σαι] (Deut 25)

Note that μαστῑ́γωσῐν and μαστῑγῶσαι exhibit etymologically long ῑ → ι.

2nd/1st c. BCE—P. Fouad 266 (Rahlfs 847, 848, 942)75

Examples: υμιν (Deut 31), λευιτης (Deut 18), etc.

Etymologically long ῑ is normally represented as simple ι. Note that ι in λευῑ́της is long since it both comes from a Hebrew long /iː/ and is part of the Greek suffix -ῑτης.

125–1 BCE—4Q119 (Rahlfs 801)76

Examples: υμιν (Lev 26), συνετριψα (Lev 26)

Both cases of etymologically long ῑ in 4Q119 are represented with simple ι.

1st c. BCE—4Q120 (Rahlfs 802)77

Examples: λειαν? (Lev 5)

It is not actually clear if there are any cases of etymologically long ῑ attested in 4Q120. The one potential example is λειαν (for λιαν), but the etymological length of iota in this word is unclear. The fact that Codex Vaticanus and Judeo-Palestinian epigraphy of the Roman period tend to spell this word as λειαν suggests that it might be historically long ῑ, but we cannot be sure.78

40 BCE–10 CE—4Q121 (Rahlfs 803)79

Examples: λευιτα[ς] (Num 3), ιματι[ον] (Num 4)

The two examples of etymologically long ῑ are both written with simple ι.

50 BCE–50 CE—8HevXIIgr (Rahlfs 943a, 943b)80

Examples: γεινε[σθε] (Zech 1), ειματια (Zech 3), υμειν (Mic 1), [α]περρειψας (Jonah 2), αναγεινωσκων (Hab 2), etc.

cf. [κ]αταπιειν (for καταπῐεῖν) (Jonah 2)

cf. [ποτιζοντ]ει (for ποτῐ́ζοντῐ) (Hab 2)

Etymologically long ῑ is routinely represented with ει in this fragment. What is especially significant, however, is that etymologically short ῐ very rarely exhibits an interchange with ει—only in the fragmentary reading [ποτιζοντ]ει. What is more, even words that have etymologically long ῑ in other parts of the paradigm—note present (κατα)πῑ́νειν but aorist (κατα)πῐεῖν—do not exhibit ει when the particular form has etymologically short ῐ: e.g., [κ]αταπιειν. This suggests that the scribe generally had good control over ῑ → ει and ῐ → ι. The apparently hyper-corrective example [ποτιζοντ]ει, however, suggests that this might have been a learned convention rather than one native in the speech of the scribe.81

150–200 CE—P.CtYBR inv. 419 (Rahlfs 814)82

Examples: σηειρ (Gen 14)

cf. ηρειθ[μησεν] (Gen 14)

This fragment provides mixed evidence. On one hand, the transcribed Hebrew name /seːʕiːr/ is rendered into Greek with ει for long /iː/. On the other hand, ἠρῐ́θμησεν with etymologically short ῐ is also represented with ει. But this stands as the exception, since other cases of etymologically short ῐ in the fragment are represented with ι. This is, in fact, consistent with the epigraphic record. Because it was a learned convention, the more likely a scribe was to implement the ῑ → ει representation, the more likely he was also to hypercorrect a small number of cases of ῐ → ει.83

early 3rd c. CE—Princeton, University Library, Scheide Pap. 1 (Rahlfs 967)84

Examples: εις μεισος (Esth 14), εμεισησα (Esth 14), etc.

cf. την βασιλεισσαν (Esth 7)

As was the pattern in the previous example, cases of etymologically long ῑ are frequently written with ει and cases of etymologically short ῐ are normally written with ι; however, probably due to the scribe attempting to implement a learned convention, there are rare hyper-corrective examples of etymologically short ῐ → ει.

250–300 CE—Chester Beatty Library, Pap. V. (Rahlfs 962)85

Examples: ιματισμον (Gen 24), χειλιαδας (Gen 24), ημειν (Gen 34), υμειν (Gen 34), λευει (Gen. 34), etc.

Etymologically long ῑ is frequently written with ει. Note that the term λευει comes from Hebrew /le(ː)wiː/ with long /iː/. There is an exception, however, in the form ιματισμον, which begins with etymologically long ῑ. It may be that the learned convention ῑ → ει did not apply in this word since it is not as common as some other forms and/or because the initial vowel was not regarded as long due to its distance from the stress.

4th c. CE—Codex Vaticanus (Vat. gr. 1209) (B)86

Examples in Scribe A (fol. 41–334): γινου (Exod 34:2, fol. 90c), τιμην (Exod 34:20, fol. 91b), γινεται (Num 6:4, fol. 147b), γεινεσθε (Josh 8:4, fol. 246c), τιμιον (1 Kgs 3:1, fol. 312b), γεινεται (1 Kgs 5:9, fol. 315b), etc.

Examples in Scribe B (fol. 335–675; 945–1234): γεινομενο(ν) (1 Kgs 13:32, fol. 425b), γεινου (3 Kgs 22:13, fol. 440a), υμιν (2 Esdr 1:3, fol. 594b), χειλιοι (2 Esdr 1:9, fol. 594c), τειμιος (2 Esdr. 4:10, fol. 597c), γεινονται (2 Esdr 4:19, fol. 598b), γεινονται (2 Esdr 4:20, fol. 598b), γεινεται (2 Esdr 5:8, fol. 599a), κριμα (2 Esdr 7:10, fol. 601b), εντειμοις (2 Esdr 12:16, fol. 608c), γεινομενον (2 Esdr 15:18, fol. 612c), γινομενη (Jer 15:18, fol. 1081c), etc.

Examples in Scribe C (fol. 675–944): γινεται (Prov 13:11, fol. 727b), τιμων (Prov 14:31, fol. 729b) αποκρεινεται (Prov 15:28, fol. 730b), γινονται (Prov 15:28, fol. 730b), γινωσκων (Eccl 8:7, fol. 758a), τιμιον (Eccl 10:1, fol. 760b), γινομενα (Job 38:33, fol. 804a), τιμην (Wis 8:10, fol. 817b), etc.

The data in Codex Vaticanus are a bit more complex. There were probably three different scribes that worked on the LXX portion of Codex Vaticanus: Scribe A copied Genesis–1 Kingdoms 19:11 (fol. 41–334). Scribe B copied 1 Kingdoms 19:11b–Psalms 77:71 (fol. 335–674) and Hosea–Daniel (fol. 945–1234). Scribe C copied Psalm 77:71–Tobit (fol. 675–944).87 It appears that the representation of etymologically long ῑ with ει is more common in scribe B than in scribe A or C, though both scribes exhibit mixed data. Naturally, internal variation is likely also possibly due to differences already present in the exemplars.88

4th c. CE—Codex Sinaiticus (BL Add. 43725) (א)89

Scribe A (History and Poetry): γεινεσθε (Num 16:16), χειλιαδες (Judg 7:3), χειλιαδες (1 Chr 12:25), χιλιοι (1 Chr 12:35)

Scribe B (Prophets): τιμη (Isa 11:10), γεινωσκω (Isa 48:4), παραγεινομενος (Isa 63:1), αναγινωσκων (Jer 28:63), υμειν (Jer 34:15), αναγινωσκο(ν)τος (Jer 43:13), αναγινωσκων (Hab 2:2), θλειψεως (Zech 8:10), etc.

Scribe D (Tobit, Judith, first half of 4 Maccabees, first two thirds of Psalms): γεινεται (Jud 11:4), θλιψεσι (Ps 9:10), κρεινον (Ps 25:1), δαυειδ (Ps 42:1), κρεινον (Ps 42:1), etc.

Similar to Codex Vaticanus, multiple scribes also worked on Codex Sinaiticus. The present state of research suggests that there were at least three scribes who worked on Codex Sinaiticus: Scribe A copied the History and Poetry sections. Scribe B copied the Prophets. Scribe D copied Tobit, Judith, the first half of 4 Maccabees, and the first two-thirds of Psalms.90 There are examples of ῑ → ει in each of the scribes, but examples of ῑ → ι are significantly more common across the board. There do not appear to be as many cases of ῑ → ει as in Codex Vaticanus. Nevertheless, it is significant that both orthographies occur in Codex Sinaiticus.

All in all, once again we see a pattern in the early papyrus fragments and manuscript witnesses of the LXX that mirrors that reflected in contemporary documentary and epigraphic evidence. Those witnesses from the Hellenistic period regularly represent etymologically long ῑ with ι. When we transition into the Roman period, however, there is a clear preference for representing etymologically long ῑ (at least in some lexemes) with ει. The transition point that exemplifies this change is the Twelve Prophets scroll from Naḥal Ḥever. Once we get to the end of the Roman period and/or beginning of the Byzantine period—we are now talking about Codex Vaticanus and Codex Sinaiticus—we see more of a mix. While Codex Vaticanus—in particular Scribe B—tends to write ει for etymologically long ῑ, there are also plenty of cases of ι for ῑ. In Codex Sinaiticus, even though ῑ → ι is common, though there are still plenty of cases of ει for ῑ. If we were to continue this survey into later Byzantine manuscripts, we might hypothesise that cases of ῑ → ει would become even less frequent, though this lies beyond the scope of this article.

As a brief aside, it is also worth mentioning how the ῑ → ει convention can even be relevant for exegesis and history of interpretation. In the case of a verb like κρῑ́νειν “to judge,” for example, the written form κρινω is ambiguous. It can be read and interpreted either as κρῑ́νω “I judge” or as κρῐνῶ “I will judge.” Those scribes of the Roman period who used the ῑ → ει convention, however, would have made a distinction, writing the former as κρεινω and the latter as κρινω. Therefore, when dealing with a scribe who regularly implements the ῑ → ει convention, spellings like κρεινω, κρεινει, etc. may be taken as evidence of a present interpretation and spellings like κρινω, κρινει, etc. may be taken as evidence of a future interpretation.91 Depending on the scribe, however, this does not always hold.92

6 Conclusions

The concept of the phonology of the LXX is and always will be a bit elusive. There are, in fact, different ways that the phonology of the LXX can be defined and approached. We could have just as well surveyed the phonologies of Egyptian and Judeo-Palestinian Greek during the periods in which the books of the LXX were composed. In this article, however, it seemed more appropriate to address the question of the phonology of the LXX from the perspective of what data from the text of the Old Greek, papyri witnesses of the LXX, and early uncial witnesses of the LXX explicitly convey.

This approach only permitted two avenues of analysis: (i) looking to transcribed Hebrew names, which can be reliably taken to reflect “original” spellings (at least in some cases), for “snapshots” into the phonology of the translator at the time of composition and (ii) tracing certain phonological developments in the history of Greek over a roughly five-century long period of attestation of the LXX text in papyrus fragments and early manuscripts. With respect to the former, we found that the phonology of the LXX reflected in transcribed names matches the phonology attested in contemporary documentary and epigraphic evidence, even when this varies by time and place (e.g., Pentateuch in third-century BCE Alexandria, 2 Esdras in second-century CE Palestine). Although we only looked at three features, further research in this area may lead to more secure evidence for the dating and provenancing of the various books of the LXX. With respect to the latter, we found that certain (phono-)orthographic phenomena (attested in native Greek words) that change over time in the documentary and epigraphic record also tend to change over time in the earliest witnesses of the LXX. While this finding shows that the earliest witnesses of the LXX cohere well with contemporary documentary and epigraphic evidence, it also demonstrates that, in many (but not all) cases, spellings of native Greek words in the LXX are probably only as “original” as the scribe of the copy in which they are attested. Little should be supposed about “original” spellings of Greek words in the LXX. The evidence of the earliest witnesses suggests that scribes regularly updated the spelling of native Greek words according to the conventions of their time.

Finally, it is also worth mentioning that this analysis has actually unearthed a new piece of evidence relevant for the wider field of historical Greek phonology. The transcriptions ρηχς (for Hebrew /reːʃ/) and χσεν (for Hebrew /ʃen/) attested in Codex Vaticanus now provide the earliest explicit evidence for the fricativization of χ in post-Classical Greek. We have thus demonstrated not only that the phonology of the LXX coheres well with that attested in contemporary papyri and epigraphy, but also that the LXX has unique contributions to offer the field of historical Greek phonology at large.

Bibliography

  • Aitken, James K. “Phonological Phenomena in Greek Papyri and Inscriptions and their Significance for the Septuagint.” In Studies in the Greek Bible: Essays in Honor of Francis T. Gignac, S.J, ed. Jeremy Corley and Vincent Skemp (Washington, D.C.: The Catholic Biblical Association of America, 2008), 256–277.

    • Search Google Scholar
    • Export Citation

  • Aitken, James K. The T&T Clark Companion to the Septuagint (London: Bloomsbury T&T Clark, 2015).

  • Aitken, James K. “The Ptolemaic Setting for the Translation of the Greek Pentateuch.” Hebrew Bible and Ancient Israel 9 (2020), 398–414.

    • Search Google Scholar
    • Export Citation

  • Allen, W. Sidney. Vox Graeca: A Guide to the Pronunciation of Classical Greek (Cambridge: Cambridge University Press, 1968).

  • Aly, Zaki, and Ludwig Koenen. Three Rolls of the Early Septuagint: Genesis and Deuteronomy (Bonn: Habelt, 1980).

  • Bons, Eberhard. “Ruth.” In The T&T Clark Companion to the Septuagint, ed. James K. Aitken (London: Bloomsbury T&T Clark, 2015), 118–126.

    • Search Google Scholar
    • Export Citation

  • Boyd-Taylor, Cameron. “Faithful Scribes and Phantom Texts: Jewish Transmission of the Septuagint Prior to the Amoraic Period.” In Die Septuaginta: Geschichte, Wirkung, Relevanz, ed. Martin Meiser, Michaela Geiger, Siegfried Kreuzer, and Marcus Sigismund (Tübingen: Mohr Siebeck, 2018), 359–382.

    • Search Google Scholar
    • Export Citation

  • Brixhe, Claude. “Linguistic Diversity in Asia Minor during the Empire: Koine and Non-Greek Languages.” In A Companion to the Ancient Greek Language, ed. Egbert J. Bakker (Chichester: Wiley-Blackwell, 2010), 228–252.

    • Search Google Scholar
    • Export Citation

  • Brown, Francis, S.R. Driver, and Charles A. Briggs. The Brown-Driver-Briggs Hebrew and English Lexicon (Peabody: Hendrickson, 2008).

  • Bubenik, Vit. Hellenistic and Roman Greece as a Sociolinguistic Area (Amsterdam: Benjamins, 1989).

  • Caird, G.B. “Homophony in the Septuagint.” In Jews, Greeks and Christians, Essays in Honor of W.D. Davies, ed. R. Mamerton-Kelly and R. Scroggs (Leiden: Brill, 1976), 74–88.

    • Search Google Scholar
    • Export Citation

  • De Troyer, Kristin. “The Septuagint.” In The New Cambridge History of the Bible. Volume 1: From the Beginnings to 600, ed. James Carleton Paget and Joachim Schaper (Cambridge: Cambridge University Press, 2013), 267–288.

    • Search Google Scholar
    • Export Citation

  • de Waard, Jan. ““Homophony” in the Septuagint.” Biblica 62 (1981), 551–561.

  • Depauw, Mark and Joanne Stolk. “Linguistic Variation in Greek Papyri: Towards a New Tool for Quantitative Study.” Greek, Roman, and Byzantine Studies 55 (2015), 196–220.

    • Search Google Scholar
    • Export Citation

  • Dines, Jennifer M. “The Minor Prophets.” In The T&T Clark Companion to the Septuagint, ed. James K. Aitken (London: Bloomsbury T&T Clark, 2015), 438–455.

    • Search Google Scholar
    • Export Citation

  • Fritsch, Ch. “Homophony in the Septuagint.” In Proceedings of the Sixth World Congress of Jewish Studies 1973 (Jerusalem: World Union of Jewish Studies, 1977), 1:115–120.

    • Search Google Scholar
    • Export Citation

  • Gignac, Francis Thomas. A Grammar of the Greek Papyri of the Roman and Byzantine Periods: Volume I: Phonology (Milan: Cisalpino–La Goliardica, 1976).

    • Search Google Scholar
    • Export Citation

  • Greenspoon, Leonard. “The Use and Abuse of the Term ‘LXX’ and Related Terminology in Recent Scholarship.” Bulletin of the International Organization for Septuagint and Cognate Studies 20 (1987), 20–29.

    • Search Google Scholar
    • Export Citation

  • Hiebert, Robert J.V. “4 Maccabees.” In The T&T Clark Companion to the Septuagint, ed. James K. Aitken (London: Bloomsbury T&T Clark, 2015), 306–319.

    • Search Google Scholar
    • Export Citation

  • Horrocks, Geoffrey. Greek: A History of the Language and Its Speakers (Oxford: Wiley Blackwell, 2014).

  • Howard, George. “The Oldest Greek Text of Deuteronomy.” Hebrew Union College Annual 42 (1971), 125–131.

  • Jannedy, Stefanie, Melanie Weirich, and Luisa Helmeke. “Acoustic analyses of differences in [ç] and [ʃ] productions in Hood German.” In Proceedings of the 18th International Congress of Phonetic Sciences, ed. The Scottish Consortium for ICPhS 2015 (Glasgow: University of Glasgow, 2015), paper number 328.

    • Search Google Scholar
    • Export Citation

  • Jongkind, Dirk. Scribal Habits of Codex Sinaiticus (Piscataway: Gorgias, 2013).

  • Kantor, Benjamin. “The Thirteenth-Century Hebrew Tradition of the Jews of Southern Italy: Greek Transcriptions of Hebrew Words in Nikolaos of Otranto’s ‘Disputation against the Jews’.” Leshonenu 82 (2020), 148–184.

    • Search Google Scholar
    • Export Citation

  • Kantor, Benjamin. The Pronunciation of New Testament Greek: Judeo-Palestinian Greek Phonology and Orthography from Alexander to Islam (Grand Rapids: Eerdmans, 2022).

    • Search Google Scholar
    • Export Citation

  • Kantor, Benjamin. “The Hebrew Knowledge of the Scribe of Codex Vaticanus.” (forthcoming)

  • Kantor, Benjamin and James Clackson. “The Use of ει to Represent Etymologically Long ῑ: A Greek Imitation and/or Adoption of a Roman Scribal Practice.” (forthcoming)

    • Search Google Scholar
    • Export Citation

  • Knobloch, Frederick W. Hebrew Sounds in Greek Script: Transcriptions and Related Phenomena in the Septuagint, with Special Focus on Genesis, Dissertation (University of Pennsylvania, 1995).

  • Lee, John A.L. The Greek of the Pentateuch: Grinfield Lectures on the Septuagint 2011–2012 (Oxford: Oxford University Press, 2018).

  • Leon, Harry Joshua. “The Language of the Greek Inscriptions from the Jewish Catacombs of Rome.” Transactions and Proceedings of the American Philological Association 58 (1927), 210–233.

    • Search Google Scholar
    • Export Citation

  • Mayser, Edwin and Hans Schmoll. Grammatik der Griechischen Papyri aus der Ptolemäerzeit: Band 1: Laut- und Wortlehre (Berlin: De Gruyter, 1970).

    • Search Google Scholar
    • Export Citation

  • Metzger, Bruce M. Manuscripts of the Greek Bible: An Introduction to Greek Palaeography (Oxford University Press: New York, 1981).

  • Myers, Peter. A Grammar of Transcriptions in 2 Esdras: Text, Philology, Phonology, Dissertation (University of Cambridge, 2019).

  • Petrounias, E.B. “Development in Pronunciation during the Hellenistic Period.” In A History of Ancient Greek: From the Beginnings to Late Antiquity, ed. A.-F. Christidis (Cambridge: Cambridge University Press, 2007), 599–609.

    • Search Google Scholar
    • Export Citation

  • Regnault, L. and J. de Préville. Dorothée de Gaza. Oeuvres spirituelles (Paris: Cerf, 1963).

  • Scarlata, Mark W. “Genesis.” In The T&T Clark Companion to the Septuagint, ed. James K. Aitken (London: Bloomsbury T&T Clark, 2015), 13–28.

    • Search Google Scholar
    • Export Citation

  • Schweizer, Eduard. Grammatik der pergamenischen Inschriften: Beiträge zur Laut- und Flexionslehre der gemeingriechischen Sprache (Berlin: Weidmann, 1898).

    • Search Google Scholar
    • Export Citation

  • Shipp, G.P. Modern Greek Evidence for the Ancient Greek Vocabulary (Sydney: Sydney University Press, 1979).

  • Skehan, Patrick, Eugene Ulrich, and Judith E. Sanderson. Qumran Cave 4: IV: Palaeo-Hebrew and Greek Biblical Manuscripts (Oxford: Clarendon, 1992).

    • Search Google Scholar
    • Export Citation

  • Teodorsson, Sven-Tage. The Phonology of Ptolemaic Koine (Göteborg: Acta Universitatis Gothoburgensis, 1977).

  • Thackeray, Henry St.J. A Grammar of the Old Testament in Greek according to the Septuagint, vol. I, Introduction, Orthography and Accidence (Cambridge: Cambridge University Press, 1909).

    • Search Google Scholar
    • Export Citation

  • Threatte, Leslie. The Grammar of Attic Inscriptions I: Phonology (Berlin: De Gruyter, 1980).

  • Tov, Emanuel. The Greek Minor Prophets Scroll from Nahal Hever (8HevXIIgr) (Oxford: Clarendon, 1990).

  • Tov, Emanuel. Textual Criticism of the Hebrew Bible (Minneapolis: Fortress, 2012).

  • Tov, Emanuel. “Loan-Words, Homophony and Transliterations in the Septuagint.” Biblica 60 (1979), 216–236.

  • Williams, Peter J. “Semitic Long /i/ Vowels in the Greek of Codex Vaticanus of the New Testament.” In Studies in Semitic Linguistics and Manuscripts: A Liber Discipulorum in Honour of Professor Geoffrey Khan, ed. Nadia Vidro, Ronny Vollandt, Esther-Miriam Wagner, and Judith Olszowy-Schlanger (Uppsala: Uppsala University Library, 2018), 15–26.

    • Search Google Scholar
    • Export Citation

  • Williams, Peter J. “The Bible, the Septuagint, and the Apocrypha: A Consideration of Their Singularity.” In Studies on the Text and Versions of the Hebrew Bible in Honour of Robert Gordon, ed. Geoffrey Khan and Diana Lipton (Leiden: Brill, 2012), 169–180.

    • Search Google Scholar
    • Export Citation

  • Wooden, R. Glenn. “2 Esdras.” In The T&T Clark Companion to the Septuagint, ed. James K. Aitken (London: Bloomsbury T&T Clark, 2015), 195–202.

    • Search Google Scholar
    • Export Citation
1

For the purposes of this paper, a broad and “elastic”—as Leonard Greenspoon might call it—definition of “LXX” is intended, spanning anything from the Old Greek to the Greek text exhibited in Codex Vaticanus. This is necessary because of the relationship of phonology to texts (see discussion in the “Introduction”). In some cases, such as when dealing with “original” readings of Greek transcriptions of Hebrew names, we are attempting to address the Old Greek. In other cases, when dealing with orthographic changes in scribal transmission attested in various papyri and early uncials, the relevant papyri and manuscript witnesses necessarily restrict the definition to their own context. For more on the problematization of the term “LXX,” see Greenspoon, “Use of LXX”; Williams, “Bible”; De Troyer, “LXX.”

2

Philo likely would have pronounced it as [ɛn arcʰe(ː) ɛpø(ː)e(ː)s̱ɛn (h)o̞ tʰɛo̞s̱], whereas Chrysostom likely would have pronounced it as [e̞n arçi e̞piːs̱e̞n o̞ θe̞o̞s̱/θʝo̞s̱] (note that IPA conventions are used).

3

Take, for example, the spelling of the word κεῖται “lies” (pronounced [ki(ː)tɛ(ː)]) in various ancient epigraphic sources of the Roman period. While inexperienced scribes might spell this word as κιτε, κιται, or κειτε in funerary inscriptions, it will normally be spelled as κειται in formal or monumental inscriptions.

4

Note that even though the digraph ει had represented [iː] (in non-pre-vocalic environments) throughout the Hellenistic period, a word like γῑ́νεται “becomes” (pronounced [ʝi(ː)nɛtɛ(ː)]) comes to be spelled frequently as γεινεται (even in higher-level texts) during the Roman period due to the popularizing of a particular “Roman” scribal convention that sought to represent etymologically long ῑ with the digraph ει (see V.4 in section 3 below).

5

Even the most significant earlier source, Ra 848 (1st c. BCE), contains only a portion of a single book. See Boyd-Taylor, “Faithful Scribes,” 367.

6

Note the discussion in Aitken, Companion to the LXX, 5–6.

7

But cf. Williams, “Long /i/ Vowels,” on etymological long [iː] in transcriptions of Hebrew names. Nevertheless, when updates in transcribed names occur, they are often clearly discernible and significant (note below on χ in transcribed letter names).

8

Because it is possible that “early” and “late” books attested in a single manuscript (e.g., Codex B) reflect different stages in their respective textual histories, this principle can actually be a bit problematic when applied to Greek spelling on a wide scale. Therefore, principle (2) is applied only in the case of Greek transcriptions of Hebrew or Aramaic names, where reconstructing an original Greek spelling dated to the time of composition is more tractable.

9

De Waard, “Homophony,” 554–58. For other examples of this approach, see Caird, “Homophony”; Fritsch, “Homophony”; Tov, “Loan-Words.”

10

Cf. Lee, Greek of the Pentateuch, 190, n. 40.

11

Thackeray, Grammar, 58–65.

12

Aitken, “Phonological Phenomena.”

13

For an overview of these changes, see Kantor, Pronunciation, §7.3. There are, of course, more nuanced and detailed changes, but those cited here cover the principal phonological developments of the period. See also, for example, Gignac, Grammar of the Greek Papyri I; Mayser and Schmoll, Grammatik (1970 version); Petrounias, “Development in Pronunciation.”

14

For more on this phenomenon, see Kantor, Pronunciation, §9.5.2.

15

While one could explain the transcription λεια by positing an original pattern */liːʔaː/, this is unlikely since it would imply a II-y root rather than a III-y root. While the meaning of the latter comports with that of לֵאָה “Leah” in the literary context, the former is unattested in Biblical Hebrew.

16

See Tov, Textual Criticism, 131; Lee, Greek of the Pentateuch, 272–73. For an origin in the Fayum, rather than Alexandria, see Aitken, “Ptolemaic Setting.” For a slightly later date, see Scarlata, “Genesis,” 15.

17

See Bons, “Ruth,” and Hiebert, “4 Maccabees.”

18

Teodorsson, Ptolemaic Koine, 114; Depauw and Stolk, “Linguistic Variation.”

19

Teodorsson, Ptolemaic Koine, 62–81; Depauw and Stolk, “Linguistic Variation.”

20

Kantor, Pronunciation, §9.3.4.13.

21

Gignac, Grammar of the Greek Papyri I, 235–49; Teodorsson, Ptolemaic Koine, 216–22.

22

Kantor, Pronunciation, §9.3.6.

23

Cf. the explanation of Knobloch, “Hebrew Sounds,” §750, who suggests that the spelling λεια indicates that the Greek scribe replaced the glottal stop, a foreign sound to a Greek speaker, with the glide sound [j]: i.e., [lej(j)ɑː]). This also resulted in the preceding vowel being shortened. This is a valid argument, but not the only possible explanation.

24

These names are tabulated according to the Göttingen editions of the LXX.

25

CIIP 1497: μνημα ληα [τη]ς θυγατρο[ς βεν]ιαμιν “the tomb of Leah, the daughter of Binyamin.”

26

See Kantor, Pronunciation, §9.3.4; §9.3.6.

27

Wooden, “2 Esdras,” 196–97.

28

There are some text-critical issues here. See Myers, “Transcriptions,” 251.

29

See Kantor, Pronunciation, §9.3.4.

30

Excrescence simply refers to the insertion of an unoriginal consonant to a word, usually in between two other homorganic consonants. In other words, it is “epenthesis” of a consonant into a particular phonological sequence, often facilitating the pronunciation of two consecutive consonants in that sequence.

31

αμβακουμ is found in Clement of Alexandria (e.g., Strom. 1.21.121), Origen (e.g., Sel. Ps. MPG 12.1124), Eusebius (e.g., Dem. Ev. 4.16), etc. αββακουμ is found in Theophilus (Autol. 2.35), Origen (Or. 14.4), etc. A text-critical analysis of all the relevant witnesses may, of course, reveal some of these readings to be unoriginal.

32

Regnault and de Préville, Dorothée de Gaza.

33

Brown, Driver, and Briggs, Lexicon.

34

Cf. also forms like σαμβατον, σαμβαταιος. See Shipp, Modern Greek Evidence, 490–91.

35

See Dines, “Minor Prophets.”

36

Note that γ, β, δ maintain their plosive pronunciation after nasals (see C.2 above).

37

Teodorsson, Ptolemaic Koine, 247–48; Bubenik, Hellenistic and Roman Greece, 239–40; Depauw and Stolk, “Linguistic Variation.”

38

Kantor, Pronunciation, §8.5.

39

Kantor, §8.9.2.13.

40

It should be noted—especially when we consider the transcription conventions of other translators, who may have had similar phonologies to that of the translator of LXX Habakkuk—that ββ for [bb] can still be maintained at a late date as a conservative transcription convention. Notably the second column of Origen’s Hexapla uses such a convention, even though phonemic gemination had been neutralized in Greek by the date of the composition of that text.

41

Kantor, Pronunciation, §8.1.2.6.

42

Kantor, §8.9.2.

43

Kantor, §8.9.2.13.

44

Allen (Vox Graeca) notes that after the tenth century CE, Armenian starts to use x and š to render Greek χ. Horrocks (Greek, 170–71) does cite Schweizer, Grammatik der pergamenischen Inschriften, 114–15, who argues that interchanges of χ with κχ, χκ, χχ (e.g., μετηλλακχοτα (for μετηλλαχότα)) in Anatolian epigraphy point to a fricative χ in second-century CE Asia Minor, but this is by no means the only interpretation of such interchanges. Brixhe (“Asia Minor,” 235), when commenting on Roman Anatolian Koine, explicitly states that “even if there is no clear clue for Χ, it is probable that the phoneme represented by this sign has undergone the same development, i.e., /ç/ before /e,i/ and /x/ before /a, o, u/.” Brixhe thus does not appear to regard Schweizer’s data as clearly indicating a fricative χ.

45

Horrocks, Greek, 172.

46

Jannedy et al., “[ç] and [ʃ].” Note that in Hood German, for example, sometimes there is no actual contrast perceived between /ç/ and /ʃ/.

47

Horrocks, Greek, 283.

48

Kantor, “Nikolaos of Otranto,” 9–10.

49

Note, for example, the apparent productive use of ει for etymologically long Hebrew /iː/ in transcribed names Williams, “Long /i/ Vowels.” For more examples and further discussion, see Kantor, “Vaticanus.”

50

Metzger, Greek Palaeography, 28.

51

Teodorsson, Ptolemaic Koine, 220–21.

52

Teodorsson, 117–22, 123–29, 161–68; Depauw and Stolk, “Linguistic Variation.”

53

Threatte, Grammar of Attic Inscriptions I, 353–67.

54

Kantor, Pronunciation, §9.1.1–§9.1.3.

55

Howard, “Greek Deuteronomy”; Metzger, Greek Palaeography, 60.

56

Aly and Koenen, Three Rolls; Howard, “Greek Deuteronomy”; Metzger, 60.

57

Skehan, Ulrich, and Sanderson, Qumran Cave 4.

58

Skehan, Ulrich, and Sanderson, Qumran Cave 4.

59

Skehan, Ulrich, and Sanderson, Qumran Cave 4.

60

Tov, Minor Prophets Scroll.

61

Metzger, Greek Palaeography, 62–63. Text also based on consultation with the Yale Library website: https://findit.library.yale.edu/bookreader/BookReaderDemo/index.html?oid=15761328#page/1/mode/1up.

62

Text based on consultation with CSNTM images: http://www.csntm.org/manuscript/View/Rahlfs_962.

63

Metzger, 70–71. Text based on consultation with CSNTM images: http://www.csntm.org/Manuscript/View/Rahlfs_967.

64

Metzger, 72. Text based on consultation with CSNTM images: http://www.csntm.org/manuscript/View/Rahlfs_962.

65

Metzger, 74. Text based on consultation with Vaticanus Library images: https://digi.vatlib.it/view/MSS_Vat.gr.1209.

66

Metzger, 76. Text based on consultation with Vaticanus Library images: https://digi.vatlib.it/view/MSS_Vat.gr.1209.

67

Kantor, Pronunciation, §9.3.1; §9.3.2.

68

Note, however, that Gignac, Grammar of the Greek Papyri I, 190–91 does not note any special distribution of ῑ → ει in comparison with ῐ → ει. It may be that a careful statistical analysis of the Egyptian material would reveal a higher relative frequency of ῑ → ει for the Roman period.

69

Leon, “Jewish Catacombs,” 221–23.

70

Brixhe, “Asia Minor,” 232.

71

Threatte, Grammar of Attic Inscriptions I, 195–99.

72

Kantor, Pronunciation, §9.3.1; §9.3.2.

73

Kantor and Clackson, “Etymologically Long ῑ.” (forthcoming).

74

Metzger, Greek Palaeography, 60.

75

Howard, “Greek Deuteronomy”; Metzger, Greek Palaeography, 60.

76

Skehan, Ulrich, and Sanderson, Qumran Cave 4.

77

Skehan, Ulrich, and Sanderson, Qumran Cave 4.

78

Personal communication with Peter Williams.

79

Skehan, Ulrich, and Sanderson, Qumran Cave 4.

80

Tov, Minor Prophets Scroll.

81

Kantor, Pronunciation. §9.3.2.

82

Metzger, Greek Palaeography, 62–63. Text also based on consultation with the Yale Library website: https://findit.library.yale.edu/bookreader/BookReaderDemo/index.html?oid=15761328#page/1/mode/1up.

83

Kantor, Pronunciation. §9.3.2.

84

Metzger, Greek Palaeography, 70–71. Text based on consultation with CSNTM images: http://www.csntm.org/Manuscript/View/Rahlfs_967.

85

Metzger, 72. Text based on consultation with CSNTM images: http://www.csntm.org/manuscript/View/Rahlfs_962.

86

Metzger, 74. Text based on consultation with Vaticanus Library images: https://digi.vatlib.it/view/MSS_Vat.gr.1209.

87

Personal communication with Jesse Grenz.

88

Personal communication with Jesse Grenz.

89

Metzger, Greek Palaeography, 76. Text based on consultation with Vaticanus Library images: https://digi.vatlib.it/view/MSS_Vat.gr.1209.

90

For a discussion of the issues, see Jongkind, Sinaiticus.

91

Note, for example, the conflicting orthography and (subsequently added) diacritic marks on the form κρεινει in Gen 49:16 in Codex Vaticanus.

92

Williams, “Long /i/ Vowels.”

Title:
The LXX and Historical Greek Phonology: Orthography, Phonology, and Transcriptions
Article Type:
Research Article
DOI:
https://doi.org/10.1163/15700631-bja10060
Language:
English
Pages:
494–526
Keywords:
Koine Greek; Septuagint; phonology; orthography; pronunciation; fricativization of χ
In:
Journal for the Study of Judaism
In:
Volume 54: Issue 4-5
Received:
27 Aug 2021
Accepted:
25 Jul 2022
Publisher:
Brill
E-ISSN:
1570-0631
Print ISSN:
0047-2212
Subjects:
General, Jewish Studies

Content Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 819 724 91
PDF Views & Downloads 1146 985 161