1 Introduction

An important difference between the Old Church Slavic (henceforth, OCS) Gospels and the other five first-millennium CE versions is that OCS does not possess a third mood outside of indicative and imperative, whereas all five other versions (Greek, Gothic, Latin, Classical Armenian, and Old English) are written in languages that possess, in addition to these two moods, a third, irrealis mood,¹ typically called subjunctive.² (In New Testament Greek the optative is not robust enough to be considered an extra mood.) Modern English itself shows a very reduced usage of the subjunctive in non-modal verbs, but its persistence in modals gives us a way of distinguishing, say, He studies all the time, so that he performs well when called upon in class (result clause, actual situation, realis) from He studies all the time so that he might/will/can perform well when called upon in class (purpose clause, non-actualized situation, irrealis).

2 Corpus and data sources

In order to assess the treatment of irrealis contexts in OCS relative to the other versions just noted, I have canvassed the Gospels of Matthew and Mark in the Marianus text (Jagić 1883, repr. 1960), the most complete early version of the OCS Gospels. Because this text does not include Mt. 1.1–5.23, in his edition, Jagić fills out these sections by employing the text of the 13th-century Dečani tetraevangelium, a Bulgarian recension which is not part of the OCS canon. In those instances where we have cited material from this part of Matthew, we have used the text of the Zographensis (Zogr.) (end of the 10th c., like the Marianus of Macedonian provenience; Jagić 1879, repr. 1954); or, where the passage is either not present or corrupt in this text, we have had recourse to the Savvina Kniga lectionary (Savv.) (11th c., presumably East Bulgarian in origin; Ščepkin 1903, repr. 1959), which includes a large number of quotations from the Gospel text.

In treating the OCS text, I have resolved all forms that are overtly abbreviated in the text, and I have generally capitalized proper nouns and names of well-defined groups (e.g., Farisěi but not kŭnižĭnici), which in general are not capitalized in the mss. For the Greek text, I have used Nestle & Aland (2012), occasionally supplemented by Robinson & Pierpont (2018); for Gothic, Streitberg (1919); for Classical Armenian, Künzle (1984); for the Latin Vulgate, Bryson (1994); and for Old English, Bosworth & Waring (1907).

3 Arrangement of data

A conspectus of the way I have arranged the data is provided in the Appendix. I have identified 60 irrealis categories, as defined by the Greek original and its manifold renditions in OCS. These are distributed, numbered consecutively (1–60), among 27 head categories (I–XXVII) based on Greek. In this article, I will discuss these different categories, paying special attention to the distinction between their treatment in OCS and in the other languages making up the focus of this study.

In employing the term ‘irrealis’, I do not limit myself to instances in which Greek employs the subjunctive (or optative). For example, the infinitive, inasmuch as it neutralizes modality, can in some instances signal an irrealis. This means that the larger groupings of the individual categories represent a mixture of morphosyntactic and semantic considerations. Moreover, in some passages where Greek employs the indicative, one or more of the translation languages may respond with a subjunctive, and for the purposes of this study, I consider such instances as well to be irrealis.

The 27 head categories may be divided into eight superordinate categories or complexes, based on the Greek text: 1) the hína/hópōs complex, to which may be added Greek constructions employing infinitives by themselves or introduced by eis tó (Appendix I–IV), a few instances of Gk. hṓste + infinitive consisting of purpose clauses and complement structures (V), and a single passage in which Greek shows an irrealis preterite indicative (VI); 2) the eán (mḗ) complex, consisting of conditional clauses appearing with the subjunctive in Greek (VII), together with contrary-to-fact conditions where Greek shows an irrealis preterite, as well as a few other cases where both Greek and Slavic show indicatives (VIII); 3) the relative + (e)án (mḗ) complex, which consists of various temporal, relative, and conditional clauses, all of which are followed by the subjunctive in Greek (IX–XV), to which we add an isolated instance where Greek shows the indicative (XVI); 4) the mḗ complex (including ou mḗ, mḗpote, and mēkéti), which almost always occurs with the subjunctive in Greek (XVII–XXI); 5) various subjunctive clauses in Greek which function as indirect questions or complement clauses (XXII–XXIII); 6) first person plural exhortatives and deliberatives (XXIV–XXV); 7) third person imperatives, which are often rendered by a subjunctive in the versions outside of OCS (XXVI); 8) those Greek futures and preterites which show subjunctives in Gothic (XXVII). Because of the general familiarity of the gospel text, in presenting the passages to be discussed below, I will adopt the procedure in each case of placing the English translation before each example.

4 Preliminary remarks

Before we begin our analysis, a caveat is in order. OCS does possess an irrealis category independent of those seen in any of the other languages included in this investigation: a conditional. This is a compound formation involving a unique set of forms occurring paradigmatically only with the verb byti ‘to be’. It is unambiguously characterized by a stem form bi-, which is most likely an optative *bh(w)-ī-. The paradigm appears to be in the process of merging with that of the aorist (stem by-), whose forms substitute for it obligatorily throughout the dual (where no distinct bi- forms exist) and optionally throughout the singular and plural (sg. 1 bimĭ/byxŭ; 2, 3 bi/by; pl. 1 bimŭ/byxomŭ; 2 biste/byste; 3 bišę/byšę). In addition, a peculiar and distinctively conditional form bǫ, which apparently represents an old subjunctive/injunctive *bh(w)-ont, is found in the third person plural.

The conditional is nearly always accompanied by the preterite l-participle, in a few cases by the past passive participle in -nŭ-, or rarely, it may even stand alone. The formation is very much alive in the Gospels, appearing 100 times in the Marianus, and may be divided into four distinct categories: 1) conditional clauses showing ašte ‘if’ + conditional in the protasis and conditional in the apodosis (27×, 59 occurrences); 2) passages of the same sort but showing the reverse ordering of clauses (4×, 9 occurrences); 3) purpose or complement clauses signaled by da (21×); and 4) complement clauses involving an indirect question (7×). A small set of miscellaneous passages completes the total. Because clauses of these types in which the conditional is not employed abound in OCS and will be treated throughout this study, the conditional should be considered to be no more than an option available to the OCS language user.

5 Earlier work

The comparative (morpho)syntax of the early Indo-European versions of the Greek Gospels was first studied in detail by Cuendet, who produced monographs on word order and the employment of the imperative in the Greek, Gothic, Classical Armenian, and Old Church Slavic Gospels (Cuendet 1924, 1929). Cuendet’s work seems to have attracted little attention, however, and led to no follow-up multilateral studies by others in the near term.

In 1988, intrigued by the possibilities presented by the existence of the same text in four different branches of Indo-European, with the added consideration that they were the oldest documents in three of them,³ I conceptualized a project that would build on Cuendet’s work using the vocabulary of contemporary linguistic analysis. By the 21st century such a project was “in the air”, as evidenced by the PROIEL enterprise at the University of Oslo (https://syntacticus.org). My most recent work, as represented by this study, has included the Latin Vulgate and the Old English gospels as part of my data set.

With regard to the more modest study of Old Church Slavic and the Greek Vorlage, which stand thematically at the core of this article, the most comprehensive investigation is Večerka’s massive five-volume compendium of OCS syntax (1989–2003). The large majority of passages that are discussed here are commented upon in Večerka’s “wissenschaftliches Lebenswerk”, although for the most part not in ways that are relevant to the points I am trying to make in this article. Another important monographic treatment dealing with ‘the subjunctive’ in OCS is Bräuer (1957), who presents useful date from the Gospels as well as from other canonical OCS texts.⁴

6 The hína/hópōs complex (Appendix I–VI)

In all six of the languages we shall consider, this type is introduced by a complementizer or subordinating conjunction; but OCS stands alone in not supplementing this element with a special mood distinct from the indicative:

(1) a. ‘But in order that you should know that the son of man has power … to remit sins …’ (Mt. 9.6)

Gk.	hína dè eidē̃te hóti eksousían ékhei ho huiòs toũ anthrṓpou … aphiénai hamartías
OCS	nŭ da uvěste ěko vlastĭ imatŭ synŭ člověčesky … otŭpuštati grěxy
Goth.	Aþþan ei witeiþ þatei waldufni habaiþ sa sunus mans … afleitan frawaurhtins
Arm.	Aył zi gitasǰik‘ et‘e išxanowt‘iwn owni ordi mardoy … t‘ołowl zmełs
Lat.	ut sciatis autem quoniam Filius hominis habet potestatem … dimittendi peccata
OE	Ɖæt ge sóþlíce witon, đæt mannes sunu hæfþ anweald … synna to forgyfenne

b. ‘In order that the thing said by the prophet Isaiah should be fulfilled’ (Mt. 8.17)

Gk.	hópōs plērōthē̃i tò hrēthèn dià Hēsaΐou toũ prophḗtou
OCS	da sŭbǫdetŭ sę rečenoe prorokomŭ. Isaiemĭ
Goth.	ei usfullnodedi þata gamelido þairh Esaïan praufetu
Arm.	zi lc‘c‘i bann or asac‘aw i jer̄n Ēsayay margarēi
Lat.	ut adimpleretur quod dictum est per Esaiam prophetam
OE	Đæt wǽre gefylled, ðæt ðe gecweden is þurh Esáiam, ðone witegan

In these purpose clauses, regardless of whether Greek shows hína or hópōs, we see the normal sextet of purpose subordinators Gk. hópōs (/hína), OCS da, Goth. ei, Arm. zi, Lat. ut, and OE ðæt. It appears, then, that it is in the first instance da which in OCS signals to the hearer/reader that this clause is irrealis. However, the verbs which follow da (uvěste, sŭbǫdetŭ sę) are both perfective, in accordance with the fact that the actions or states which they signal have not yet actually occurred. The combination of da plus perfective verb would therefore appear to jointly signal what in all the other versions is established via modality.⁵ Note also that OE witon in (1a), though formally ambiguous, can be safely taken to be subjunctive.

In other instances, Gk. hína or hópōs can introduce a complement clause. Nevertheless, the rendition in all five translational versions remains the same, with the minor exception that Armenian may show et‘e ‘that, if’ as a complementizer:

(2) a. ‘For it is better for you that one of your members be destroyed …’ (Mt. 5.29)

Gk.	sumphérei gár soi hína apólētai hèn tō̃n melō̃n sou
OCS	uněe bo ti estŭ da pogybletŭ edinŭ udŭ tvoixŭ
Goth.	batizo ist auk þus ei fraqistnai ains liþiwe þeinaize
Arm.	zi law ē k‘ez et‘e mi yandamoc‘ k‘oc‘ koric‘ē
Lat.	expedit enim tibi ut pereat unum membrorum tuorum
OE	sóþlíce đé ys betere, đæt án đínra lima forwurđe

b. ‘Therefore, ask of the lord of the harvest that he send out workers into his harvest’ (Mt. 9.38)

Gk.	deḗthēte oũn toũ kuríou toũ therismoũ hópōs ekbálēi ergátas eis tòn therismòn autoũ
OCS	molite sę ubo gospodinu žętvě. da izvedetŭ dělatelę na žętvo̜ svojo̜
Goth.	bidjiþ nu fraujan asanais ei ussandjai waurstwjans in asan seina
Arm.	ard ałač‘ec‘ēk‘ zTR hnjoc‘ zi hanc‘ē zmšaks i hunjs iwr
Lat.	rogate ergo dominum messis ut eiciat operarios in messem suam
OE	Biddaþ đæs rípes hláford, đæt he sende wyrhtan to hys rípe

c. ‘Let your light so shine before men that they see your good deeds and glorify your Father’ (Mt. 5.16)

Gk.	hoútōs lampsátō tò phō̃s humō̃n émprosthen tō̃n anthrṓpōn, hópōs ídōsin humō̃n tà kalà érga kaì doksásōsin tòn patéra humō̃n
OCS	(Zogr.) tako da prosvěti(ti)tŭ sę světŭ vašĭ prědŭ člověky. da uzĭrętŭ děla vaša dobraě. i proslavętŭ otĭca vašego
Goth.	swa liuhtjai liuhaþ izwar in andwairþja manne, ei gasaiƕaina izwara goda waurstwa jah hauhjaina attan izwarana
Arm.	aynpēs lowsaworesc‘ē loys jer ar̄aǰi mardkan. orpēs zi tesc‘en zgorcs jer baris. ew p‘arawor̄esc‘en zhayr jer
Lat.	sic luceat lux vestra coram hominibus ut videant vestra bona opera et glorificent patrem vestrum
OE	swá onlihte eower leoht befóran mannum, đæt hí geseon eowre gódan weorc, and wuldrian eowerne fæder

In (2c) the structure is correlative in all six languages, and this explains the compound complementizer orpēs zi in Armenian, where orpēs, lit. ‘in which way’, stands in correlation with aynpēs ‘in that way’, whereas zi is the item that has served as the complementizer in most of the passages we have seen so far. But in all three passages OCS once again shows da plus perfective verb (pogybletŭ, izvedetŭ, uzĭrętŭ) in lieu of the overt modality seen in the other languages.

In other instances within the hína/hópōs complex the translations are structurally the same as those of (1) and (2), but OCS alone shows a different construction, either employing an infinitive, as in (3a) or a participle plus infinitive, as in (3b):

(3) a. ‘And he would not allow that anybody carry a vessel through the temple’ (Mk. 11.16)

Gk.	kaì ouk ḗphien hína tis dienégkēi skeũos dià toũ hieroũ
OCS	i ne daděaše nikomuže mimo nesti sŭso̜dŭ skvozě crŭkovi
Goth.	jah ni lailot ei ƕas þairhberi kas þairh þo alh
Arm.	ew oč‘ tołoyr et‘e ok‘ anawt‘ inč‘ anc‘owc‘anic‘ē ǝnd tačarn
Lat.	et non sinebat ut quisquam vas transferret per templum
OE	And he ne geþafode đæt ǽnig man ǽnig fæt þurh đæt templ bǽre

b. ‘For he healed many, so that they fell before him in order that they might touch him’ (Mk. 3.10)

Gk.	polloùs gàr etherápeusen, hṓste epipíptein autō̃i hína autoũ hápsōntai
OCS	mŭnogy bo iscěli. ěko napadaxo̜ emĭ xotęšte prikosno̜ti sę emĭ
Goth.	managans auk gahailida, swaswe drusun ana ina ei imma attaitokeina
Arm.	k‘anzi zbazowms bžškeac‘. minč‘[M: -ew]6 gal xr̄nel znovaw zi merjesc‘in ar̄ na
Lat.	multos enim sanabat ita ut inruerent in eum ut illum tangerent
OE	Sóþlíce manega he gehǽlde, … swá đæt hí æt-hrinon his

In the first of these passages OCS says ‘he would not allow for anybody to carry a vessel through the temple’, whereas in (3b) it says, literally, ‘wanting to touch him’, with a preceding participle. Although in this instance the OCS construction is, to be sure, syntactically different from the Greek, it nevertheless retains the voluntative nature of the Greek purpose clause.

Just as OCS shows an infinitive as a complementizer in (3a), so Greek may show an infinitive preceded either by the article alone or by eis tó in purpose clauses, in lieu of hína or hópōs. In these instances, OCS maintains its normal da + perfective indicative, while the other languages may either hold firmly to their particle + subjunctive or respond with an infinitive or gerund construction:

(4) a. ‘For Herod intends to search for the child in order to kill him’ (Mt. 2.13)

Gk.	méllei gàr Hērṓidēs zēteĩn tò paidíon toũ apolésai autó
OCS	(Savv.) xoštetŭ bo Irodŭ iskati otročęte. da pogubitĭ e
Goth.	–
Arm.	k‘anzi i xndir ē Hērovdēs korowsanel zmanowkd
Lat.	futurum est enim ut Herodes quaerat puerum ad perdendum eum
OE	toweard ys, đæt Herodes sécþ đæt cild, to forspillenne

b. ‘They were seeking testimony against Jesus for the purpose of putting him to death’ (Mk. 14.55)

Gk.	ezḗtoun katà toũ Iēsoũ marturían eis tò thanatō̃sai autón
OCS	iskaaxo̜ na Isusa sŭvědětelĭstva da i ubĭjo̜tŭ
Goth.	sokidedun ana Iesu weitwodiþa du afdauþjan ina
Arm.	xndrein hakar̄ak YI vkayowt‘iwn inč‘. zi spananic‘en zna
Lat.	quaerebant adversum Iesum testimonium ut eum morti traderent
OE	sóhton … tále ágén đone Hǽlend, đæt hí hine to deaþe sealdon

Večerka (2002: 339) suggests that the OCS da + perfective present is employed in (4a) in order to avoid successive infinitive clauses. In this instance Old English and Latin both show a gerund (to forspillenne, ad perdendum) and Armenian an infinitive (korowsanel), preceded in each instance by finite clauses. But in (4b) all three of these languages show purpose clauses with the subjunctive, whereas Gothic shows an infinitive (du afdauþjan). Similar to this latter passage is the Matthaian parallel (Mt. 26.59), where Greek shows hópōs and OCS ěko da + perfective indicative, while Armenian, Latin, and OE show the structures seen in (4b).

An instance in which every other language but OCS shows an infinitive or gerund in response to a Greek purpose infinitive is (5):

(5) ‘Behold, the sower went out in order to sow’ (Mt. 13.3)

Gk.	idoù eksē̃lthen ho speírōn toũ speírein
OCS	Se izide sějęi da sěetŭ
Goth.	–
Arm.	Aha el sermanahan sermanel
Lat.	ecce exiit qui seminat seminare
OE	út-eode se sǽdere hys sǽd to sáwenne

Insofar as infinitives and gerunds do not describe an actual state of affairs, they too may be described as belonging to the realm of the irrealis. The OCS verb sěti ‘sow’ would appear to be anaspectual (Kamphuis 2020: 125). It is unclear why OCS employs da + indicative here in lieu of a supine or an infinitive. The simplest ad hoc explanation would be to assume that the translator wanted to recapitulate the phrasal nature of the Greek infinitive. Bräuer (1957: 18) has in any event suggested that OCS shows a particular proclivity for employing da + indicative in purpose clauses.

A nuanced variant of (4a–4b) is (6), where the structure is very close to an indirect question, and all the languages respond to this by using their ‘how’ adverbial. Again, OCS is the only language to show an indicative, as usual perfective in response to the Greek aorist subjunctive:

(6) ‘But the Pharisees, going out, took counsel against him, (as to) how they might destroy him.’ (Mt. 12.14)

Gk.	ekselthóntes dè hoi Pharisaĩoi sumboúlion élabon kat’ autoũ hópōs autòn apolésōsin
OCS	Farisěi že išedŭše sŭvětŭ sŭtvorišę na nĭ. kako i pogubętŭ
Goth.	–
Arm.	Isk P‘arisec‘ik‘n eleal artak‘s. xorhowrd ar̄in znmanē. t‘e orpēs korowsc‘en zna.
Lat.	Exeuntes autem Pharisaei consilium faciebant adversus eum quomodo eum perderent
OE	Đa Sunder-hálgan eodun ðá út sóþlíce, and worhton geþeaht ongén hyne, hú hí hyne forspildon

The employment of t‘e before the ‘how’-word orpēs in Armenian is in line with the tendency in this language to employ this particle before direct quotations as well as in indirect discourse, where the stretch that follows is treated almost as a direct quotation (in this case, roughly equivalent to ‘How might we destroy him?’) (Ouzounian 1992).

Finally, we must note that in 21 instances in all four Gospels OCS shows a conditional in passages of the sort we have been discussing. I initially considered this a marginal phenomenon, and indeed it does have this status in Matthew and Mark (the subcorpus for this study), where it appears only five times in passages showing the Greek hína/hópōs complex. However, in addition to appearing another seven times in this group in Luke and John, it appears nine times in passages where Greek shows an infinitive construction (only once in Matthew and Mark), meaning that it is considerably less marginal within the Gospels as a whole. In all but one instance the preceding main clause is in a past tense, producing a uniform response of ei + preterite subjunctive in Gothic, zi + subjunctive (usually aorist) in Classical Armenian, ut + imperfect subjunctive in Latin, and đæt + preterite subjunctive in OE. Where Greek shows an infinitive construction, the other languages may reproduce it:

(7) a. ‘The man_x who had been possessed by a demon begged him_y that he_x might be with him_y’ (Mk. 5.18)

Gk.	parekálei autòn ho daimonistheìs hína met’ autoũ ē̃i
OCS	molěaše i běsŭnovavy sę. da bi sŭ nimĭ bylŭ
Goth.	baþ ina, saei was wods, ei miþ imma wesi
Arm.	ałač‘ēr zna diwaharn zi ǝnd nma ic‘ē
Lat.	coepit illum deprecari qui daemonio vexatus fuerat ut esset cum illo
OE	hine ongan biddan, se đe ǽr mid deofle gedréht wæs, đæt he mid him wǽre

b. ‘(They sent to him some of the Pharisees and of the Herodians) in order that they might catch him in a word’ (Mk. 12.13)

Gk.	… hína autòn agraúsōsin lógōi
OCS	… da i bo̜ oblĭstili slovomĭ
Goth.	… ei ina ganuteina worda
Arm.	… zi zna orsasc‘en baniw
Lat.	… ut eum caperent in verbo
OE	… đæt hí beféngon hine on his worde

c. ‘(For they disfigure their faces) in order that they may appear to men as fasting’ (Mt. 6.16)

Gk.	… hópōs phanō̃si toĩs anthrṓpois nēsteúontes
OCS	… da b(o̜)7 sę avili člověko(mŭ) postęšte
Goth.	… ei gasaiƕaindau mannam fastandans
Arm.	… orpēs zi erewesc‘in mardkan t‘e parhic‘en [M: -hesc‘en]
Lat.	… ut pareant hominibus ieiunantes
OE	… đæt hig æteowun mannum fæstende

d. ‘(they begged him) that he leave their district’ (Mt. 8.34)

Gk.	… hópōs metabē̃i apò tō̃n horíōn autō̃n
OCS	… da bi prěšelŭ otŭ prědělŭ ixŭ
Goth.	… ei usliþi hindar markos izē
Arm.	… zi gnasc‘ē [M: gnal] i sahmanac‘ noc‘a
Lat.	… ut transiret a finibus eorum
OE	… đæt he férde fram heora gemǽrum

(7a–7b) show Greek hína in complement and purpose clauses, respectively, while (7c–7d) show hópōs in purpose and complement clauses, respectively. (7c) is the only instance in which the preceding main clause is in the present tense, eliciting a present subjunctive in Gothic, Latin, and OE.⁸ It is unclear why OCS should show conditionals in these passages as opposed to the far more common da + perfective indicative seen in (1–6). In (7b–7d) the OCS main verbs (appearing in the l-participle) are certainly perfective,⁹ and I presume that the same can be said at least for the l-participle of byti in (7a). The OCS version of (7c) is particularly important in that it defies the common wisdom (cf. Meillet 1934: 266) that the OCS conditional occurs only following a clause whose verb is in the preterite. The verb of the main clause here is prosmraždajǫtŭ (present) ‘they disfigure’. The only variation seen in the other versions is the Armenian employment of orpēs, lit. ‘in which manner’ in (7c), where the semantics is compatible with a value ‘in such a way that’ (orpēs zi).

A category that neighbors purpose clauses is that of result clauses. These typically differ from the former in that they indicate an actual rather than an envisioned result, so that they cannot be termed ‘irrealis’. In the Greek New Testament, result clauses are normally signaled by hṓste + infinitive; but occasionally, hṓste clauses may breach the result/purpose boundary, as in (8a–8b):

(8) a. ‘(He gave them power over unclean spirits) to cast them out and to heal every sickness’ (Mt. 10.1)

Gk.	… hṓste ekbállein autà kaì therapeúein pā̃san nóson
OCS	… da izgonętŭ ję. i cěliti vĭsěkŭ nedo̜gŭ
Goth	–
Arm.	… hanel znosa. ew bžškel zamenayn c‘aws
Lat.	… ut eicerent eos et curarent omnem languorem
OE	… đæt hig ádryfon hig út, and hǽlden ádle

b. ‘… and it produces large branches, so that the birds of heaven can dwell under its shade’ (Mk. 4.32)

Gk.	kaì poieĩ kládous megálous, hṓste dúnasthai hupò tḕn skiàn autoũ tà peteinà toũ ouranoũ kataskēnoũn
OCS	i tvoritŭ větvi veliję ěko mošti podŭ sěnijo̜ ego pticamŭ nebesĭskyimŭ vitati
Goth.	jah gataujiþ astans mikilans, swaswe magun uf skadau is fuglos himinis gabauan
Arm.	arjakē osts mecamecs. minč‘ew bawakan linel ǝnd hovaneaw nora t‘r̄č‘noc‘ erknic‘ bnakel
Lat.	et facit ramos magnos ita ut possint sub umbra eius aves caeli habitare
OE	and hæfþ swá mycele bogas, đæt heofones fugelas eardian mágon under his sceade

In (8a) only Armenian follows the Greek in showing two infinitive complements. Latin and OE show subjunctives in both instances, and OCS treats the first Greek infinitive no differently from the way it treats purpose clauses with hína and hópōs, while matching the Greek infinitive in its second occurrence. The critical semantic fact is that the infinitive complements here do not signal actual results, so that this passage can be categorized as a purpose clause, and it is the Greek hṓste which is aberrant. Infinitives, it should be remembered, neutralize the realis/irrealis distinction.

In (8b) it is the modal verb dúnasthai ‘to be able’ that marks irrealis in the Greek, with its double infinitive construction; and this is followed by OCS and Armenian (bawakan linel … bnakel ‘to be able … to dwell’), whereas Gothic and Old English show finite verb + infinitive. Latin alone shows its verb of capability in the subjunctive followed by an infinitive. The Gothic and Latin (compound) subordinators swaswe and ita ut can be thought of as signaling, respectively, univerbated and adjacent correlative constructions having the value ‘so that’, whereas this is clearer in OE swá … đæt. Armenian minč‘ew means literally ‘until, to the point that’. A literal translation of the OCS text would be ‘(so) as for the birds of heaven to be able to dwell under its shade’, a construction in which ěko (lit. ‘as’) introduces a dative subject in the subordinate clause. However, as pointed out by an anonymous referee, the absolute identity of word order in the Greek and OCS texts raises the spectre of syntactic calquing, the case difference Gk. accusative + infinitive : OCS dative + infinitive being the result of idiomatic syntax in the two languages.

In other instances, Gk. hṓste + infinitive appears in contexts that clearly demand a resultative interpretation:

(9) a. ‘And he became like a dead man, so that many said that he had died’ (Mk. 9.26)

Gk.	kaì egéneto hōseì nekrós, hṓste polloùs légein hóti apéthanen
OCS	i bystŭ ěko mrŭtvŭ. ěko mŭnozi glagolaaxǫ ěko umĭrětŭ
Goth.	jah warþ swe dauþs, swaswe managai qeþun þatei gaswalt
Arm.	ew ełew pataneakn ibrew zmer̄eal. minč‘ew bazmac‘ isk asel t‘e mer̄aw
Lat.	[25] et factus est sicut mortuus ita ut multi dicerent quia mortuus est
OE	and he wæs swylce he dead wǽre, swá đæt manega cwǽdon, sóþlíce he is dead

b. ‘And picking up (his) bed, he went out before all, so that they all became amazed’ (Mk. 2.12)

Gk.	kaì áras tòn krábatton exē̃lthen enantíon pántōn, hṓste exístasthai pántas
OCS	i vĭzętŭ odrŭ. i izide prědŭ vĭsěmĭ. ěko divlěaxǫ sę vĭsi
Goth.	jah ushafjands badi usiddja faura andawairþja allaize, swaswe usgeisnodedun allai
Arm.	ew vałvałaki ar̄eal zmahičsn el ar̄aǰi amenec‘own. minč‘ew zarmanal amenec‘own
Lat.	et sublato grabatto abiit coram omnibus ita ut admirarentur omnes
OE	and, [underleat bere,] befóran him eallum eode, swá đæt ealle wundredon

Particularly interesting in (9a) is the threefold occurrence in OCS of ěko, each time in a different value: prepositional ‘like’ (ěko₁), ‘subordinating conjunction ‘so that’ (ěko₂), complementizer ‘that’ (ěko₃). Armenian pataneakn ‘the young man’ of (9a) is an exegetical addition, while the adverb vałvałaki ‘immediately’ in (9b) is everywhere else associated with the prior verb ‘he arose’.

The fact that the imperfect in OCS, as well as in Greek (here also the aorist and perfect) and Armenian, the perfect in Latin, and the preterite in OE may be employed to signal an irrealis is illustrated by (10), which provides another instance of the conditional in OCS:

(10) ‘You should have deposited my money with money-changers, and I would have come and taken back what was mine with interest’ (Mt. 25.27)

Gk.	édei se oũn baleĩn tà argúriá mou toĩs trapezítais, kaì elthṑn egṑ ekomisámēn àn tò emòn sùn tókōi
OCS	podobaaše ti ubo vŭdati sŭrebro moe … trŭžŭnikomŭ. i prišĭdŭ azŭ vĭzęlŭ ubo bimĭ svoe sŭ lixvojo̜
Goth.	–
Arm.	part ēr k‘ez arkanel zarc‘atd im i sełanawors ew ekeal es tokaseawk‘ pahanǰei zimn
Lat.	oportuit ergo te mittere pecuniam meam nummulariis et veniens ego recepissem utique quod meum est cum usura
OE	Hyt gebyrede đæt đú befæstest mín feoh myneterum, and ic náme đonne ic come đæt mín ys mid đam gafole

In each of these cases, the verb of the first clause is impersonal ‘it was necessary, fitting’, and the sense of its preterites in all five languages is irrealis, translated here as ‘you should have’. The second clause, however, is also irrealis: in Greek aorist + án, in Armenian imperfect, in Latin and OE subjunctive. The Latin pluperfect subjunctive establishes the main clause as contrary to fact. The OCS conditional here, vĭzęlŭ bimĭ, belongs to a miscellaneous category of contrary-to-fact conditions not involving ašte ‘if’, the only other member of which is the parallel Lk 19.23, which is preceded by a question ‘Why did you not give my money to moneylenders?’ All correspondences to the OCS conditional in that passage are as above.

Based on the evidence of the hína/hópōs complex, we can say that in the important realm of purpose clauses the indicators of the irrealis in OCS are in the large majority of cases not verbal mood but rather the presence of the particle da together with a perfective verb. Users of the language would have reacted to the combination of these two features as a signal that the verbal action did not denote something that has actually occurred. Other means available to signal this condition were the infinitive, which neutralized modality, the conditional, which appears only about 6.7 % of the time in these passages (Breuer 1957: 54) and, under certain circumstances, the imperfect, which normally indicated a past imperfective or habitual. In a few instances, Gk. hṓste + infinitive seems to lie on the boundary of, or even breach, the distinction between purpose and result clauses; and these, too, may show OCS da + indicative or (in line with the normal translation rule Gk. hṓste : OCS ěko), ěko + infinitive or indicative.

7 Conditional clauses (Appendix VII–VIII)

The second group of passages where a Greek subjunctive is matched by an OCS indicative involves conditional clauses that are not contrary to fact. These show a unitary structure, with Gk. eàn (mḗ) being matched by OCS ašte (ne). The other translation languages respond variously, depending on the particular passage:

(11) a. ‘For if you forgive men their transgressions, your heavenly father will also forgive you’ (Mt. 6.14)

Gk.	eàn gàr aphē̃te toĩs anthrṓpois tà paraptṓmata autō̃n, aphḗsei kaì humĩn ho patḕr humō̃n ho ouránios
OCS	Ašte ubo otŭpuštaete člověkomŭ sŭgrěšeniě ixŭ otŭpustitŭ i vamŭ otĭcŭ vašŭ nebesĭsky
Goth.	unte jabai afletiþ mannam missadedins ize, afletiþ jah izwis atta izwar sa ufar himinam
Arm.	Zi et‘e t‘ołowc‘owk‘ mardkan zyanc‘ans noc‘a . t‘ołc‘ē ew jez hayr [M:-n] jer erknawor
Lat.	si enim dimiseritis hominibus peccata eorum dimittet et vobis Pater vester caelestis delicta vestra
OE	Witodlíce gyf ge forgyfaþ mannum hyra synna, đonne forgyfþ eower se heofenlíca fæder eow eowere gyltas

b. ‘If your justice is not greater than that of the scribes and Pharisees, you will not enter the kingdom of heaven’ (Mt. 5.20)

Gk.	eàn mḕ perisseúsēi humō̃n hē dikaiosúnē pleĩon tō̃n grammatéōn kaì Pharisaíōn, ou mḕ eisélthēte eis tḕn basileían tō̃n ouranō̃n
OCS	(Zogr.) ašte ne izbo̜detŭ pravŭda vaša. pače kŭnižĭnikŭ i Farisei. ne imate vĭniti vŭ cěsarĭstvo nebesĭskoe
Goth.	nibai managizo wairþiþ izwaraizos garaihteins þau þize bokarje jah Fareisaie, ni qimiþ in þiudangardjai himine
Arm.	et‘e oč‘ ar̄awelowc‘ow ardarowt‘iwn[M:-d] jer aweli k‘an zdprac‘n ew zP‘arisec‘woc‘. oč‘ mtanic‘ēk‘ yark‘ayowt‘iwn[M:-n] erknic‘
Lat.	nisi abundaverit iustitia vestra plus quam scribarum et Pharisaeorum non intrabitis in regnum caelorum
OE	búton eower rihtwísnys máre sý đonne đæra writera and Sundor-hálgena, ne gá ge on heofonan ríce

c. ‘If you bring your offering upon the altar and there you remember that your brother has something against you, (24) leave your offering there …’ (Mt. 5.23–24)

Gk.	eàn oũn prosphérēis tò dō̃rón sou epì tò thusiastḗrion kakeĩ mnēsthē̃is hóti ho adelphós sou ékhei ti katà soũ, (24) áphes ekeĩ tò dō̃rón sou …
OCS	(23 Zogr.) ašte ubo prineseši darŭ svoi kŭ ol’tar’ju. i tu poměneši ěko br(at)ŭ tvoi imatŭ něčto na tę (24) ostavi tu darŭ tvoi …
Goth.	jabai nu bairais aibr þein du hunslastada jah jainar gamuneis þatei broþar þeins habaiþ ƕa bi þuk, (24) aflet jainar þo giba þeina …
Arm.	Et‘e matowc‘anic‘es zpatarag k‘o i veray sełanoy . ew and yišic‘es [M: e]t‘e ełbayr k‘o ownic‘i inč‘ xēt‘ zk‘ēn. (24) t‘oł zpataragn k‘o ar̄aǰi sełanoyn …
Lat.	si ergo offeres munus tuum ad altare et ibi recordatus fueris quia frater tuus habet aliquid adversum te (24) relinque ibi munus tuum …
OE	Eornostlíce gyf đú bringst đíne lác to weofode, and đú đær geþencgst, đæt đín bróđor hæfþ ǽnig þing ágén đé, (24) Lǽt đær đíne lác …

d. ‘Lord, if you wish, you can cleanse me’ (Mt. 8.2)

Gk.	kúrie, eàn thélēis, dúnasaí me katharísai
OCS	gospodi. ašte xošteši možeši mę ištistiti
Goth.	frauja, jabai wileis magt mik gahrainjan
Arm.	TR et‘e kamis. karoł es zis srbel
Lat.	Domine si vis potes me mundare
OE	Drihten, gyf đú wylt, đú miht me geclǽnsian

The first two of these passages show conditions whose fulfillment is said to lead to a future result (eventualis). Here the only languages to match the Greek subjunctive with subjunctives of their own are Classical Armenian (both [11a] and [11b]) and Old English (11b). In this last instance, the difference between the indicative of (11a) and the subjunctive of (11b) seems to be related to the more factual gyf ge forgyfaþ of the former and the negation, signaled by búton in the latter. No other language in our group makes such a distinction. However, the Gothic structure nibai managizo wairþiþ izwaraizos garaihteins in (11b) is highly likely to represent diachronically an optative nibai ‘it not be’ followed by an unmarked complement clause ‘(that) your justice will be greater’ (lit. ‘there will be more of your justice’), with future-oriented wairþiþ. Gothic afletiþ and OE forgyfþ in the apodosis of (11a) should also be deemed futures. In OCS the futurity of the conditions in both passages is captured by the employment of perfective presents in each. It must be noted in passing that the subjunctive in Classical Armenian is ambiguous between a modal and a future, and its presence in the apodosis of both passages is to be understood in the latter value. The Greek subjunctive in the apodosis of (11b) is regular following ou mḗ. Finally, the Latin future perfect in the protasis of both passages is exceedingly common in eventual conditions in the Vulgate.

(11c) shows a conjoined set of conditional clauses followed by an instruction attendant on its fulfillment. Here Gothic is added to Classical Armenian as a language showing the subjunctive in both instances, whereas Latin shows a future conjoined with a future perfect and OE, like OCS, shows two present indicatives. OCS again shows perfective presents here, the first despite translating a Greek present subjunctive. (11d) shows indicatives in all languages but Greek, with the proviso that in Germanic the indicative of ‘will’ is diachronically a subjunctive (optative).

As a less common variant of (11a), Greek sometimes shows an indicative (without án) rather than a subjunctive, even when the general value of the condition is the same. Thus, (12) is merely the negated version of (11a), and the response of the individual translations is exactly the same as that seen in the latter, with the exception that in the protasis the OCS present is perfective (otŭpustite) rather than imperfective (otŭpuštaete):¹⁰

(12) ‘But if you do not forgive, neither will your Father in heaven forgive your transgressions’ (Mk. 11.26)

Gk.	ei dè humeĩs ouk aphíete, oudè ho patḕr humō̃n ho en toĩs ouranoĩs aphḗsei tà paraptṓmata humō̃n
OCS	ašte li vy ne otŭpustite. ni otecŭ vašŭ iže vŭ nebesexŭ otŭpustitŭ prěgrěšenii vašixŭ
Goth.	iþ jabai jus ni afletiþ, ni þau atta izwar sa in himinam afletiþ izwis missadedins izwaros
Arm.	Zi et‘e dowk‘ oč‘ t‘ołowc‘owk‘. ew oč‘ hayrn jer or yerkins ē. t‘ołc‘ē jez zyanc‘ans jer
Lat.	quod si vos non dimiseritis nec pater vester qui in caelis est dimittet vobis peccata vestra
OE	Gyf ge ne forgyfaþ, ne eow eower synna ne forgyfþ, eower fæder

A passage in which the Greek protasis shows a future verb form is the following. In this instance all the other languages except OCS and Latin respond with a subjunctive:¹¹

(13) ‘Even if all will be scandalized, yet not I’ (Mk.14.29)

Gk.	ei kaì pántes skandalisthḗsontai all’ ouk egṓ
OCS	i ašte vĭsi sŭblaznętŭ sę nŭ ne azŭ
Goth.	–
Arm.	t‘epēt ew amenek‘ean gayt‘agłesc‘in. aył es oč‘
Lat.	et si omnes scandalizati fuerint sed non ego
OE	Đeah đé ealle swicion, ne swicige ic đé ná

Once again, the Armenian subjunctive here is ambiguous between a modal and a future reading, while the OCS denominative sŭblaznętŭ sę is perfective (‘will take offense’).

A more complex structure is seen in (14), where an initial condition expressed as a relative clause is itself followed by a conditional period in which the protasis follows the apodosis. In the lower conditional period every one of the versions except OCS differs from the Greek in showing a modal value, including, in Classical Armenian, the employment of an imperfect in this role, as illustrated in (10) above:

(14) ‘(And whoever [= if anybody] leads astray one of these little ones …) it would be better for him if a millstone were placed around his neck and he were thrown into the sea’ (Mk. 9.42)

Gk.	… kalón estin autō̃i mãllon ei períkeitai líthos mulikòs12 perì tòn trákhēlon autoũ kaì béblētai eis tḕn thálassan
OCS	… dobrěe emu estŭ pače. ašte obložętŭ kamenĭ žrŭnovŭny o vyi ego. i vŭvrŭgo̜tŭ i vŭ more
Goth.	… goþ ist imma mais ei galagjaidau asiluqairnus ana halsaggan is jah frawaurpans wesi in marein
Arm.	… law ēr nma ar̄awel. et‘e arkeal ēr ǝnd paranoc‘ nora erkan išoy. ew ǝnkec‘eal [sc. ēr] i cov
Lat.	(41) … bonum est ei magis si circumdaretar mola asinaria colle eius et in mare mitteretur
OE	… betere him wǽre đæt án cweorn-stán wǽre to his sweoran gecnyt, and wǽre on sǽ beworpen

The Greek text presented here, which is that of Nestle & Aland (2012), sits uneasily on the borderline between realis and irrealis, because it shows present(-oriented) indicatives in both clauses. Indeed, Blass & Debrunner (1979: 292, n. 6) note that the D and W texts show preterite verb forms periékeito and eblḗthē, both of which, as noted above, are frequently used to signal an irrealis; and so it is treated in our translation as well as that of every one of the versions represented here, with the exception of OCS (subjunctive in Gothic, OE, and Latin, irrealis imperfect in Classical Armenian). The Greek indicative of the introductory clause (kalón estin autō̃i) is matched in OCS, Gothic, and Latin, but the Armenian imperfect and OE subjunctive better reflect an irrealis meaning. The OCS impersonal presents are once again perfective.

The situation in OCS is very different, however, in cases where the condition is contrary to fact. Under these circumstances, Greek shows, in the vast majority of instances, an irrealis preterite (aorist or imperfect, occasionally pluperfect) in both clauses, and OCS responds with a conditional in both the protasis and apodosis. There are 31 passages of this sort in the Gospels, only nine of which are in our subcorpus. Of the other languages, Armenian shows imperfects in all clauses, Gothic, where attested, shows preterite subjunctives in both protasis and apodosis, and Latin and OE, with one Latin exception, show preterite/(plu)perfect subjunctives:

(15) a. ‘And if the Lord had not shortened those days, no flesh would have been saved’ (Mk. 13.20)

Gk.	kaì ei mḕ kúrios ekolóbōsen tàs hēméras, ouk àn esṓthē pãsa sárx
OCS	ašte ni bi gospodŭ prěkratilŭ dĭnii. ne bi byla sŭpasena vĭsěka plŭtĭ
Goth.	jah ni frauja gamaurgidedi þans dagans, ni þauh ganesi ainhun leike
Arm.	Ew et‘e oč‘ ēr karčeal AY zawowrsn zaynosik … oč‘ aprēr amenayn marmin
Lat.	Et nisi breviasset Dominus dies non fuisset salva omnis caro
OE	And gif Drihten đas dagas ne gescyrte, nán flǽsc ne wurde hál

b. ‘It would have been better for him if he had not been born’ (Mt. 26.24)

Gk.	kalòn ē̃n autō̃i ei ouk egennḗthē
OCS	dobrěa (Zogr. -ěe) bi bylo emu ašte sę bi ne rodilŭ
Goth.	–
Arm.	law ēr nma t‘e č‘ēr cneal
Lat.	Bonum erat ei si natus non fuisset
OE	betere wǽre đam men, đæt he næfre nǽre ácynned

c. ‘If you knew what is (the meaning of) “I desire mercy and not sacrifice,” you would not have condemned the guiltless.’ (Mt. 12.7)

Gk.	ei dè egnṓkeite tí estin˙ éleos thélō kaì ou thusían, ouk àn katedikásate toùs anaitíous
OCS	ašte li biste věděli. čŭto estŭ milosti xoštǫ a ne žrŭtvě. nikoliže ubo biste osǫdili nepovinŭnyxŭ
Goth.	–
Arm.	Ew [M: ∅] et‘e giteik‘ zinč‘? ē zołormowt‘iwnn[M: -n] kamim ew oč‘ zzoh ‧ apa oč‘ dataparteik‘ dowk‘ zampartsn
Lat.	si autem sciretis quid est misericordiam volo et non sacrificium numquam condemnassetis innocentes
OE	Gyf ge sóþlíce wistun, hwæt ys, Ic wylle mild-heortnesse, and ná onsægdnysse, ne genyđrude ge ǽfre unscyldige

(15b), where the protasis follows the apodosis, is very much like (14); however, the meaning of the second clause (‘if he had not been born’) in this instance clarifies its contrary-to-fact status, as does the Latin imperfect erat. (15c) illustrates a Greek pluperfect in the protasis, while the Latin pluperfect subjunctive in the apodosis is characteristic for sequences whose opening clause shows an imperfect subjunctive.

The final passage belonging to this category which we shall consider is very close to an indirect question and is treated as such in Gothic. All languages but Greek and OCS show subjunctives:

(16) ‘And they were watching him (to see) whether he would heal him on the sabbath’ (Mk. 3.2)

Gk.	kaì paretēroũnto autón, ei toĩs sábbasin therapeúsei (v.l. -eúei) autón
OCS	i naziraaxo̜ i ašte vŭ soboto̜ iscělitŭ i
Goth.	jah witaidedun imma hailidediu sabbato daga
Arm.	Ew spasein nma et‘e bžškic‘ē [M: -kesc‘ē] zna i šabat‘own[M: -t‘ow]
Lat.	et observabant eum si sabbatis curaret
OE	And hí gýmdon, hwæđer he on reste dagam gehǽlde

As was the case with the hína/hópōs complex, in conditional clauses the one constant that we have seen in OCS is the employment of a conjunction, this time ašte, to signal this relationship. In contrary-to-fact conditions, irrealis contexts par excellence, this is accompanied by the conditional. The picture across the other languages is here more varied than that seen with purpose clauses, because conditions may be viewed as more or less factual; and of course in English as well, conditionals do not require a modal/subjunctive. The language most disposed to using the subjunctive is Classical Armenian, but, as noted above, the subjunctive in this language is ambiguous between an irrealis modal and a future. In Armenian the imperfect is widely employed as an irrealis, especially in contrary-to-fact conditions, as is the preterite (aorist, imperfect, and pluperfect) in Greek (15a–15c, respectively) and, less frequently, the imperfect in Latin (cf. the initial ‘apodosis’ of 15b).

8 Relative clauses (Appendix IX–XVI)

The third complex of passages we shall consider consists of relative clauses, whether these are introduced by a simple relative pronoun (expanded to include hósa and hóstis) or a range of subordinating conjunctions (hópou, hótan, epán, héōs). With the exception of one instance involving hópou, these passages all co-occur with (e)án, which is incorporated into hótan and epán; and with only one exception they are accompanied by the subjunctive in Greek. Taken together, these passages comprise 16 of the 60 categories in my data.

We begin with simple relative clauses. These are semantically close to conditionals, especially when the relative pronoun is generalized, ‘Whoever does X’ being roughly equivalent to ‘If somebody/anybody does X’. This contention is supported in real linguistic terms by the fact that in the OCS passages of this type the relative pronoun iže is frequently accompanied by the generalizing particles ašte or koližĭdo, the first of which possesses the basic value ‘if’ which we have observed in conditional clauses.

The most frequent pattern in relative clauses is for the OCS perfective indicative to be matched by the subjunctive in Greek and Armenian, the indicative in OE, and, when the main clause is future, the (ambiguous) future perfect in Latin. Examples are the following:

(17) a. ‘Now then, he who annuls one of the smallest of these commandments and teaches men so will be called least in the kingdom of heaven’ (Mt. 5.19)

Gk.	hòs eàn oũn lúsēi mían tō̃n entolō̃n tóutōn tō̃n elakhístōn kaì didáksēi hoútōs toùs anthrṓpous, elákhistos klēthḗsetai en tē̃i basileíãi tō̃n ouranō̃n
OCS	(Zogr.) iže bo razoritŭ edino̜ zapovědĭ. i sixŭ malyxŭ. i naučitĭ tako člověky. mĭnii narečetŭ sę vŭ cěsarĭstvi nebesĭscěmĭ
Goth.	saei nu gatairiþ aina anabusne þizo minnistono jah laisjai swa mans, minnista haitada in þiudangardjai himine
Arm.	Or ok‘ lowcc‘ē mi inč‘ i patowiranac‘[M: -s] yaysc‘anē i p‘ok‘ownc‘. ew owsowsc‘ē aynpēs zmardik. p‘ok‘r koč‘esc‘i yark‘ayowt‘ean [M: -n] erknic‘
Lat.	Qui ergo solverit unum de mandatis istis minimis et docuerit sic homines, minimus vocabitur in regno caelorum
OE	se þe towyrpþ án of đysum læstum bebodum, and đa men swá lǽrþ, se byþ læst genemned on heofonan ríce

b. ‘He promised to give her whatever she might ask for’ (Mt. 14.7)

Gk.	hōmológēsen autē̃i doũnai hò d’ eàn aitḗsētai
OCS	izdreče ei dati egože ašte vŭsprositŭ
Goth.	–
Arm.	xostac‘aw nma tal zinč‘ ew xndresc‘ē
Lat.	pollicitus est ei dare quodcumque postulasset ab eo
OE	behét he … hyre to syllenne swá hwæt swá heo hyne bǽde

c. ‘Whatever you bind on earth will be bound in heaven’ (Mt. 16.19)

Gk.	hò eàn dḗsēis epì tē̃s gē̃s éstai dedeménon en toĩs ouranoĩs
OCS	eže ašte sŭvęžeši na zemi. bo̜detŭ sŭvęzano na nbesexŭ13
Goth.	–
Arm.	zor miangam kapesc‘es [M: -ic‘es] yerkri. ełic‘i kapeal yerkins
Lat.	quodcumque ligaveris super terra erit ligatum in caelis
OE	swá hwæt swá đú ofer eorþan gebindst, đæt byþ on heofonum gebúnden

d. ‘If two of you agree … concerning anything they ask for, it will come about for them’ (Mt. 18.19)

Gk.	eàn dúo sumphōnḗsōsin eks humō̃n … perì pántos prágmatos hoũ eàn aitḗsōntai, genḗsetai autoĩs
OCS	ašte dŭva otŭ vasŭ sŭvěštaate … o vĭsěkoi vešti ejęže koližŭdo prosite bo̜detŭ ima
Goth.	–
Arm.	et‘e erkow i jēnǰ miabanic‘en … vasn amenayn irac‘ zinč‘ ew xndresc‘en [M: -ic‘en] ełic‘i noc‘a
Lat.	si duo ex vobis consenserint … de omni re quacumque petierint fiet illis
OE	gyf twegen of eow geþwǽriaþ … be ǽlcum þinge đe hig biddaþ, hit gewurþ him

e. ‘Whoever (= If someone) divorces/would divorce his wife, let him give her a writ of divorcement’ (Mt. 5.31)

Gk.	hòs àn apolúsēi tḕn gunaĩka autoũ, dótō autē̃i apostásion
OCS	iže ašte pustitŭ ženo̜ svojo̜. da dastŭ ei kŭnigy raspustŭnyę
Goth.	ƕazuh saei afletai qen, gibai izai afstassais bokos
Arm.	or arjakic‘ē zkin iwr. tac‘ē nma zarjakmann
Lat.	quicumque dimiserit uxorem suam det illi libellum repudii
OE	Swá hwylc swá his wíf forlǽt, he sylle hyre hyra híw-gedáles bóc

f. ‘But whoever swears by the gold of the temple is bound (by that oath)’ (Mt. 23.16)

Gk.	hòs d’ àn omósēi en tō̃i khrusō̃i toũ naoũ opheílei
OCS	a iže klŭnetŭ sę zlatomĭ crŭkŭvŭnymŭ dlŭženĭ estŭ
Goth.	–
Arm.	bayc‘ or erdnow yoskin or i tačarin ē . aržan ē
Lat.	qui autem iuraverit in aurum templi debet
OE	swá hwylc swá swereþ on đæs temples golde, se ys scyldig

These passages show only occasional deviations from the relationship Gk. aorist subjunctive : OCS perfective indicative¹⁴ : Goth. indicative : Class. Arm. subjunctive¹⁵ : Lat. subjunctive¹⁶ : OE indicative. I base my statement concerning Gothic on its general treatment of these passages, including those that I have not included here because they tell us nothing new about the other languages. However, we can say that Gothic is prone to variation between indicative and subjunctive in these construction types. So, for example, in (17a), where we find conjoined clauses under the scope of the relative and therefore expect verbs showing identical structure in each conjunct, Gothic employs as second verb the subjunctive laisjai, as if it were saying ‘he who annuls … and would teach men so’, with a factual opening clause followed by an irrealis. In (17e) the Gothic subjunctive may be related to the generalized subject ƕazuh saei ‘whoever’, which does not pick out a particular individual and hence is removed from actuality. In the first clause of (17a), on the other hand, the subject is simple non-generalized relative saei, and here we find the indicative.

Other generalizing particles widely represented in these passages are OE swá hwæt/hwylc swá (17b–c, 17e–f); Classical Armenian ew (17b, 17d), lit. ‘and’, miangam (17c), lit. ‘one time’ (Germ. Mal, Fr. fois), and the indefinite negative polarity item ok‘ (17a); Lat. -cumque (17b–17e); OCS ašte, lit. ‘if’ (17b–17c, 17e) and koližŭdo (17d); and of course, the Greek particle (e)án, while not morphologically of this type (unlike hóstis), nevertheless serves as a syntactic marker of irrealis. The other anomalies seen in these passages are the employment of the indicative in Armenian in (17f), which lacks the generalizing particle ok‘ and therefore must have been understood by the translator to mean factually ‘the one who swears by the temple is bound’.¹⁷ Finally, the only instance of the subjunctive in Old English in these clauses is seen in (17b), where the translator must have understood ‘He promised to give her whatever she might ask of him’.

Similar to the passages of (17) are those in which Greek shows either hóstis án or hósa (e)án. The first of these is overtly generalizing. In the second, OCS responds with either (j)eliko ašte ‘as much as’ or eže ‘that/which’:

(18) a. ‘Whoever denies me before men, I too will deny him …’ (Mt. 10.33)

Gk.	hóstis d’ àn arnḗsētai me émprosthen tō̃n anthrṓpōn arnḗsomai kagṑ autón …
OCS	iže otŭvrŭžetŭ sę mene prědŭ člověky. otŭvrŭgo̜ sę ego i azŭ …
Goth.	iþ þisƕanoh18 saei afaikiþ mik in andwairþja manne, afaika jah ik ina …
Arm.	Ew or owrasc‘i zis ar̄aǰi mardkan. owrac‘ayc‘ ew es zna …
Lat.	Qui autem negaverit me coram hominibus negabo et ego eum …
OE	Se đe me wiđsæcþ befóran mannum, and ic wiđsace hyne …

b. ‘For whoever does the will of my father in heaven, he is my brother …’ (Mt. 12.50)

Gk.	hóstis gàr àn poiḗsēi tò thélēma toũ patrós mou toũ en ouranoĩs autós mou adelphós … estín
OCS	iže bo ašte tvoritŭ voljo̜ otĭca moego iže estŭ na nebesexŭ. tŭ bratrŭ moi … estŭ
Goth.	–
Arm.	zi or ar̄nic‘ē zkams hawr imoy or yerkinsn ē. na ē im ełbayr …
Lat.	quicumque enim fecerit voluntatem patris mei qui in caelis est ipse meus … frater … est
OE	swá hwylc swá wyrcþ mínes fæder willan đe on heofenam is, he is mín bróđur

c. ‘Therefore, everything that you wish men should do to you, do you likewise to them’ (Mt. 7.12)

Gk.	Pánta oũn hósa eàn thélēte hína poiō̃sin humĩn hoi ánthrōpoi, hoútōs kaì humeĩs poieĩte autoĩs
OCS	Vĭsě ubo eliko ašte xoštete da tvorętŭ vamŭ člověci. tako i vy tvorite imŭ
Goth.	… [tau]jaina izwis mans, swa jah jus taujiþ im
Arm.	Zamenayn or miangam kamiǰik‘ [M: e]t‘e arasc‘en jez mardik. aynpēs ew dowk‘ ararēk‘ noc‘a
Lat.	Omnia ergo quaecumque vultis ut faciant vobis homines et vos facite eis
OE	Eornustlíce ealle đa þing, đe ge wyllen đæt men eow dón, dóþ ge him đæt sylfe

d. ‘And everything that you ask for in prayer, if you have faith, you will get’ (Mt. 21.22)

Gk.	kaì pánta hósa àn aitḗsēte en tē̃i proseukhē̃i pisteúontes lḗmpsesthe
OCS	i vĭsego egože vŭsprosite vŭ molitvě věrujo̜šte priimete
Goth.	–
Arm.	Ew zamenayn inč‘ [M: ∅] zor ew xndric‘ēk‘ yaławt‘s hawatovk‘ [M: -tawk‘]. ar̄nowc‘owk‘
Lat.	Et omnia quaecumque petieritis in oratione credentes accipietis
OE	And ealles đæs đe ge biddaþ ge beoþ tiþa, gyf ge gelýfaþ

In (18b), Kamphuis (2020) treats OCS tvoritŭ as anaspectual. The only modal deviations here from the passages of (17) are seen in (18c), where Latin shows an indicative and OE a subjunctive form of the verb ‘wish’. Unfortunately, the broken context of Gothic does not allow us to say whether its corresponding verb would have been indicative or subjunctive. However, both in (18a) and in response to the following similar Greek construction, differing only in the fact that in Gothic the relative clause follows the main clause, Gothic shows an indicative:

(19) ‘All sins and blasphemies which they utter will be forgiven to the sons of men’ (Mk. 3.28)

Goth.

allata afletada þata frawaurhte sunum manne, jah naiteinos, swa managos swaswe wajamerjand

Finally, in a single passage of this sort Greek shows an indicative. This is the Marcan parallel to Mt. 21.22 (18d), and only Latin adjusts its verbal mood to indicative:¹⁹

(20) ‘Everything that you request while praying believe that you will get, and it will come to you’ (Mk. 11.24)

Gk.	pánta hósa àn proseukhómenoi aiteĩsthe, pisteúete hóti lambánete, kaì éstai humĩn
OCS	vĭsě eliko molęšte sę prosite. věruite ěko priemlete. i bo̜detŭ vamŭ
Goth.	allata þisƕah þei bidjandans sokeiþ, galaubeiþ þatei nimiþ, jah wairþiþ izwis
Arm.	amenayn inč‘ vasn oroy aławt‘s arareal xndrec‘ēk‘. ew hawatayc‘ēk‘ t‘e ar̄nowc‘owk‘ ełic‘i jez
Lat.	Omnia quaecumque orantes petitis credite quia accipietis et veniet vobis
OE	swá hwæt swá ge gyrnende biddaþ, gelýfaþ đæt ge hit onfóþ, and hit eow becymþ

The OCS verb prosite is classified by Kamphuis (2020) as anaspectual.

The remaining passages of this sort involve subordinating conjunctions of time or place (hópou ‘where’, hótan, epán ‘when’). These passages occur only with the subjunctive in Greek in my data, and the first may appear with or without (e)án, which is incorporated into the last two of these forms. The OCS equivalent of the first is usually ideže (1× ěmože) ‘where’, which may be followed by generalizing ašte or koližĭdo. Examples are seen in (21):

(21) a. ‘Wherever the corpse is, there the eagles will gather’ (Mt. 24.28)

Gk.	hópou eàn ē̃i tò ptō̃ma ekeĩ sunakhthḗsontai hoi aetoí
OCS	ideže bo ašte bo̜detŭ trupŭ. tu sŭbero̜tŭ sę orĭli
Goth.	–
Arm.	Zi owr gēšn ic‘ē. andr žołovesc‘in arcowik‘ [M: ew arcowik‘n]
Lat.	ubicumque fuerit corpus illuc congregabuntur aquilae
OE	And swá hwær swá hold byþ, đæder beoþ earnas gegaderode

b. ‘Wherever you enter into a home, abide there’ (Mk. 6.10)

Gk.	hópou eàn eisélthēte eis oikían, ekeĩ ménete
OCS	ideže koližŭdo vŭnidete vŭ domŭ. tu prěbyvaite
Goth.	þisƕaduh þei gaggaiþ in gard, þar saljaiþ
Arm.	yor town mtanic‘ēk‘. andēn awt‘ewans kalǰik‘
Lat.	quocumque introieritis in domum illic manete
OE	Swá hwylc hús swá ge in-gáþ, wunigaþ þar

c. ‘Teacher, I will follow you wherever you may go’ (Mt. 8.19)

Gk.	didáskale, akolouthḗsō soi hópou eàn apérkhēi
OCS	učitelju ido̜ po tebě. ěmože koližŭdo ideši
Goth.	laisari, laistja þuk, þisƕaduh þadei gaggis
Arm.	vardapet. ekic‘ ew es zkni k‘o yor vayr ew ert‘ic‘es [M: dow]
Lat.	magister sequar te quocumque ieris
OE	Láreow, ic fylige þé, swá hwæder swá đú færst

d. ‘And wherever he seizes him, he dashes him to the ground’ (Mk. 9.18)

Gk.	kaì hópou àn [v.l. eàn] autòn katalábēi, hrḗssei autón
OCS	i iže (sic!) ašte koližŭdo imetŭ i. razbivaatŭ i
Goth.	jah þisƕaruh þei ina gafahiþ, gawairpiþ ina
Arm.	[17] ew owr hasanē. zarkowc‘anē zna
Lat.	qui ubicumque eum adprehenderit adlidit eum
OE	Se swá hwær swá he hine gelǽcþ, forgnít hine

e. ‘And wherever he should enter, say to the master of the house that the teacher says, “Where is the apartment where I may eat the paschal lamb with my disciples?” ’ (Mk. 14.14)

Gk.	kaì hópou eàn eisélthēi, eípate tō̃i oikodespótēi hóti ho didáskalos légei· poũ estin tò katáluma, hópou tò páskha metà tō̃n mathētō̃n mou phágō;
OCS	i ideže ašte vŭnidetŭ rĭcěta gospodinu domu. ěko učitelĭ glagoletŭ. kŭde estŭ obitělĭ ideže pasxo̜ sŭ učeniky svoimi sŭněmŭ
Goth.	jah þadei inngaleiþai, qiþaits þamma heiwafraujin þatei laisareis qiþiþ: ƕar sind saliþwos þarei paska miþ siponjam meinaim matjau?
Arm.	ew yor town mtanic‘ē. asasǰik‘ c‘tanowtērn. vardapet asē. owr? en iǰavank‘n owr ašakertawk‘s owtic‘em zzatikn
Lat.	et quocumque introierit dicite domino domus quia magister dicit ubi est refectio mea ubi pascha cum discipulis meis manducem
OE	And swá hwyder swá he in-gǽþ, secgaþ đæs húses hláforde, Úre láreow segþ, Hwar is mín gyst-hús and mín gereord, hwar ete ic eastron mid mínum learning-cnihtum?

The OCS verbs in these passages are perfective, with the exception of ideši in (21c), which is anaspectual (Kamphuis 2020). The employment of the typically future and primarily perfective bǫdetŭ to translate the Greek present subjunctive ē̃i in (21a) is owing to the fact that this Greek subjunctive has no corresponding aorist. It is further notable that all the other translation languages may be interpreted as showing futures in this initial clause. In (21e), which otherwise shows the same configuration of modal usage across all the languages as is seen in (21a), I have included the initial clause, because it provides evidence from Gothic which is lacking in the latter.

In general, Gothic shows the greatest modal variation in these passages, with subjunctive in (21b) and (21e) (2×) and indicative in (21c) and (21d). For Latin, I assume, based on the principle articulated in fn. 11, that only (21d) and the final clause of (21e) are subjunctive, the others future perfect. This is further supported in (21a) by the fact that OE employs byþ and beoþ, respectively, in the two clauses here, both of which are frequently employed in future value. In this language the final verb ete of (21e) is ambiguous as to mood. In (21d) the OCS reading iže ašte koližŭdo ‘whoever’ is most likely an error for ideže ašte koližŭdo ‘wherever’, and Classical Armenian, in contrast to every other instance here, shows a present indicative hasanē ‘he comes upon (sc. him)’.

The passages showing Greek hótan and (rarely) epán + subjunctive are temporal clauses matched in each instance by OCS egda(že) ‘when’ + indicative:

(22) a. ‘Days will come when the bridegroom will be taken away from them’ (Mt. 9.15)

Gk.	eleúsontai dè hēmérai hótan aparthē̃i ap’ autō̃n ho numphíos
OCS	prido̜tŭ že dĭnie. egdaže otŭimetŭ sę otŭ nixŭ ženixŭ
Goth.	iþ atgaggand dagos, þan afnimada af im sa bruþfaþs
Arm.	aył ekec‘en [M: -esc‘en] awowrk‘n yoržam barjc‘i i noc‘anē p‘esayn
Lat.	venient autem dies cum auferetur ab eis sponsus
OE	Sóþlíce đa dagas cumaþ, đæt se brýdguma byþ áfyrred from hym

b. ‘And when you find (sc. him, viz. the child), let me know’ (Mt. 2.8)

Gk.	epàn dè heúrēte, apaggeílaté moi
OCS	(Savv.) egdaže obręštete e povědite mi
Goth.	–
Arm.	ew yoržam gtanic‘ēk‘ [M: zna] azd arasǰik‘ inj
Lat.	et cum inveneritis renuntiate mihi
OE	and đonne ge hyt gemétaþ, cýđaþ eft me

In (22a) the verbs of all the languages except Greek are to be understood as futures.

A temporal relative clause type that deserves a treatment of its own involves the Greek héōs án + subjunctive. These are matched by OCS donĭdeže + indicative. Corresponding subordinating conjunctions in Gothic, Classical Armenian, Latin, and OE are unte, minč‘ew, donec, and ođ đæt, respectively:

(23) a. ‘Reside there, until you go out from there’ (Mk. 6.10)

Gk.	ekeĩ ménete, héōs àn eksélthēte ekeĩthen
OCS	tu prěbyvaite donĭdeže izidete otŭ to̜dě
Goth.	þar saljaiþ, unte usgaggaiþ jainþro
Arm.	andēn awt‘ewans kalǰik‘ minč‘ew elanic‘ēk‘ anti
Lat.	illic manete donec exeatis inde
OE	wuniaþ đar, ođ đæt ge út-gán

b. ‘There are some among those standing here who will not taste death until they see the kingdom of God having come in power’ (Mk. 9.1)

Gk.	eisín tines tō̃n hō̃de hestēkótōn hoítines ou mḕ geúsōntai thanátou, héōs àn ídōsin tḕn basileían toũ theoũ elēluthuĩan en dunámei
OCS	so̜tŭ edini otŭ sŭde stojęštiixŭ. iže ne imo̜tŭ vŭkusiti sŭmrŭti. dondeže vidętŭ cěsarĭstvie božie prišedŭšee vĭ silě
Goth.	sind sumai þize her standandane þai ize ni kausjand dauþaus, unte gasaiƕand þiudinassu gudis qumana in mahtai
Arm.	en omank‘ i soc‘anē or ast kan. ork‘ mi čašakesc‘en zmah. minč‘ew tesc‘en zark‘ayowt‘iwn [M: -n] AY ekeal zawrowt‘eamb
Lat.	(8.39) sunt quidam de hic stantibus qui non gustabunt mortem donec videant regnum Dei veniens in virtute
OE	sume synd hér wuniende, đe deaþ ne onbyrigeaþ, ǽr hí geseon Godes ríce on mægne cuman

Outside of OCS, all languages here show a subjunctive in both passages, with the exception of Gothic, which employs an indicative in (23b). OE alone says, in (23b), ‘ere they see’ rather than ‘until they see’. OCS vidętŭ in (23b) is anaspectual (Kamphuis 2020).

As was the case with conditionals, the absence of an OCS irrealis in relative clauses does not appear to present serious problems of the sort produced (at least from the perspective of English speakers) by its absence in purpose clauses relative to actualized result clauses. From an Indo-European perspective, the detailed opposition between indicative and non-indicative seen in Greek and Latin appears to be a nicety, perhaps spawned by a robust literary tradition replete with sophisticated genre distinctions. The Old Church Slavic translation of the Greek Gospels stands at the very beginning of the Slavic literary tradition, to which must be added that Slavic at an early date lost or did not inherit a subjunctive²⁰ and, unlike Gothic, prehistorically shifted the optative in cases other than that of the verb *bhuh₂ (conditional *bh[w]ī-) to imperative value.²¹ Consequently, unlike Gothic (and Classical Armenian, but in a very different way²²), Old Church Slavic could not oppose an irrealis to an indicative mood. Rather, as was the case in sections 6 and 7, it made do with a set of relative pronouns and subordinating conjunctions (da, iže, ideže, ěmože, egdaže, donĭdeže, eliko) together with a partly ambiguous group of generalizing particles (ašte, koližĭdo) to convey the necessary relations together with their nuances.

9 Negative clauses (Appendix XVII–XXI)

The next complex of passages we shall investigate consists of negative clauses where Greek shows mḗ, ou mḗ, mḗpote, and mēkéti + subjunctive, plus a single instance of mēkéti + optative. The employment of mḗ with the subjunctive is regular in New Testament Greek generally, and ou mḗ has become the basic means of negating the future. Taken together, this complex accounts for 13 categories of Greek-OCS relations, more than 20 % of the total in my data.

The simplest case is where mḗ + aorist subjunctive has imperative (injunctive) value and is correspondingly matched by an imperative in OCS and Classical Armenian, a subjunctive in Gothic, and variably by either a subjunctive or an imperative in Latin and Old English:

(24) a. ‘From the one wishing to borrow from you do not turn away’ (Mt. 5.42)

Gk.	tòn thélonta apò soũ daneísasthai mḕ apostraphē̃is
OCS	xotęštago otŭ tebe zajęti ne otŭvrati
Goth.	þamma wiljandin af þus leiƕan sis ni uswandjais
Arm.	or kami p‘ox ar̄nowl i k‘ēn mi darjowc‘aner zeress
Lat.	volenti mutuari a te ne avertaris
OE	đam đe [wylle] æt đé borgian ne wyrn đú him

b. ‘Now then, when you give alms, do not blow a trumpet before you’ (Mt. 6.2)

Gk.	hótan oũn poiē̃is eleēmosúnēn, mḕ salpísēis emprosthén sou
OCS	egda ubo tvoriši milostynjo̜. ne vŭstrǫbi prědŭ sobojo̜
Goth.	þan nu taujais armaion, ni haurnjais faura þus
Arm.	Aył yoržam ar̄nic‘es ołormowt‘iwn. mi harkaner p‘ox ar̄aǰi k‘o
Lat.	cum ergo facies elemosynam noli tuba canere ante te
OE	Eornustlíce đonne đú đíne ælmessan sylle, ne bláwe man býman befóran đé

The occurrence of the subjunctive in Gothic is owing to the fact that in this language the imperative is limited to employment in instances where a command is to be carried out immediately (Klein 2022). Consequently, it is excluded in apodoses of temporal subordinate clauses and general instructions to be applied outside the context of the immediate here and now. To understand the Latin subjunctive of (24a), it is necessary to look at the larger context. The opening period of this stanza says ‘To the one asking of you, give’, and here Latin does use the imperative followed by a correlative within the diptych qui … ei (qui petit a te da ei). However, in the following period, cited above, the initial dative phrase does not express an overt action but rather a wish (volenti mutuari a te). The internalization of the action here may have provided just enough modalization to deflect the apodosis to a subjunctive (ne avertaris ‘you should not turn away’). The reason for the subjunctive in Old English in (24b) is clearer: the English translator has changed the subject of the main clause to an impersonal, and Old English lacks a third person imperative.

In a second category, OCS shows an infinitive; but this passage is the last of a series of negative instructions on the part of Jesus, who tells his disciples in indirect discourse what not to take on the road with them, ending with the stricture that they should not wear two tunics. OCS continues the indirect discourse, treating this as an infinitive (‘and not to clothe themselves in two tunics’), whereas Latin and OE employ third person subjunctives. But Greek, Gothic, and Armenian conclude this subsection in direct discourse, employing second person subjunctives:

(25) ‘And you should not put on two tunics’ (Mk. 6.9)

Gk.	kaì mḕ endúsēsthe dúo khitō̃nas
OCS	i ne oblačiti sę vĭ dŭvě rizě
Goth.	jah ni wasjaiþ twaim paidom
Arm.	ew mi zgenowc‘owk‘ … erkows paregawts
Lat.	et ne induerentur duabus tunicis
OE	and đæt hí mid twám tunecum gescrýdde nǽron

A third type consists of an admonition in the imperative followed by a complement clause. OCS and OE present an overt complementizer (OCS da, OE đæt) following the imperative, whereas Classical Armenian and Latin follow the Greek in showing no overt complementizer but simply the modal negations mi and ne, respectively. The overall structure here is not far removed from that of a negative purpose clause, and all languages but OCS show subjunctives:

(26) ‘See/take care that nobody/lest anybody lead you astray’ (Mk. 13.5)

Gk.	blépete mḕ tis humãs planḗsēi
OCS	bljuděte sę da ne kto prělĭstitŭ vasŭ
Goth.	–
Arm.	zgoyš kac‘ēk‘. mi ok‘ zjez xabesc‘ē
Lat.	vide ne quis vos seducat
OE	warniaþ, đæt eow nán man ne beswice

A structure very similar to (26) is the following passage, where all languages other than OCS show the same translational strategy as above. The latter goes its own way, replacing the complement clause by an asyndetic imperative structure. In this case, the Gothic passage is preserved, and it shows an overt complementizer (ei):

(27) ‘See that you speak to nobody’ (Mt. 8.4)

Gk.	hóra mēdenì eípēis
OCS	viždĭ nikomuže ne pověždĭ
Goth.	saiƕ ei mann ni qiþais
Arm.	zgoyš ler. mi owmēk‘ asic‘es
Lat.	vide nemini dixeris
OE	warna đé, đæt đú hit nǽnegum men ne secge

Of all the categories of my data, the one that shows the widest range of responses in the OCS text involves the Greek structure mḗpote + subjunctive. These passages indicate that the speaker has some fear that a certain circumstance will eventuate, and they are once again very similar to negated purposes clauses. The English rendition which most exhaustively covers the entire range of these passages is ‘lest’; and indeed the passages of this sort are in Old English marked by đe-læs (đe), the precursor of NE lest. Because Greek pote has under negation literally a temporally indefinite value ‘at any time’, this is occasionally reflected in the translations, as in Gothic ibai ƕan, Arm. zi mi erbek‘, Lat. ne quando, and OCS eda kogda, da ne egda, all of which contain temporal elements (ƕan, erbek‘, quando, and kogda, egda, respectively); but the most frequent rendition in Classical Armenian is gowc‘ē ‘lest’ (lit. ‘it may be’), and in Latin it is ne forte (lit. ‘in order that by chance not’). Examples of these passages are seen in (28):

(28) a. ‘Make friends with your opponent … lest … he give you over to the judge’ (Mt. 5.25)

Gk.	ísthi eunoō̃n tō̃i antidíkōi sou … mḗpoté se paradō̃i … tō̃i kritē̃i
OCS	Bo̜di uvěštaję sę sŭ so̜pĭremĭ svoimĭ … da ne prědastŭ tebe so̜dii
Goth.	sijais waila hugjands andastauin þeinamma … ibai ƕan atgibai þuk … stauin
Arm.	Ler irawaxorh ǝnd awsoxi k‘owm … gowc‘ē matnic‘ē zk‘ez … dataworin
Lat.	Esto consentiens adversario tuo … ne forte tradat te … iudici
OE	Beo đú onbúgende đínum wiđerwinnan … đe-læs đe … đé sylle đam deman

b. ‘Upon their hands they will raise you, in order that you never strike your foot against a stone’ (Mt. 4.6)

Gk.	epì kheirō̃n aroũsín se, mḗpote proskópsēis pròs líthon tòn póda sou
OCS	(Savv.23) na ro̜kaxŭ vŭzŭmo̜tŭ tę. da ne kogda prětŭkneši o kamenĭ nogo̜ svojo̜
Goth.	–
Arm.	i veray jer̄ac‘ barjc‘en zk‘ez. zi mi erbek‘ harc‘es zk‘ari [M: zk‘arē] zotn k‘o
Lat.	in manibus tollent te ne forte offendas ad lapidem pedem tuum
OE	hig đé on hyra handum beron, đe-læs đe đín fót æt stáne ætsporne

c. ‘Lest they see with their eyes and hear with their ears and understand with their heart and return’ (Mt. 13.15)

Gk.	mḗpote ídōsin toĩs ophthalmoĩs kaì toĩs ōsìn akoúsōsin kaì tē̃i kardíai sunō̃sin kaì epistrépsōsin
OCS	eda kogda uzĭrętŭ očima. i ušima uslyšętŭ. i srŭdĭcemĭ razumějo̜tŭ. i obratętŭ sę
Goth.	–
Arm.	zi mi erbek‘ tesc‘en ač‘awk‘. ew akanǰawk‘n lowic‘en. ew sirtiwk‘n imasc‘in. ew darjc‘in
Lat.	ne quando oculis videant et auribus audiant et cordibus intellegant et convertantur
OE	đe-læs hig ǽfre mid eagum geseon, and mid earum gehýron, and mid heortan ongyton, and sín gecyrrede

OCS treats (28a) and (28b) as simple negative purpose clauses (da ne prědastŭ, da ne kogda prětŭkneši). In (28a) the ibai ƕan of Gothic would have originally meant ‘it might be at some time [that]’ (cf. Klein 2011), and OE đe-læs would originally have been đý læs ‘by that less’. In (28c), where the initial main clause has been suppressed, OCS eda is a combination of the old relative pronoun *ye- and the same da whose usage, particularly in purpose clauses, we have traced above. Its negative value is probably the result of ellipsis from *eda ne or the like, whereas kogda contains the loco-temporal suffix -gda added to the indefinite ko-. The Classical Armenian of (28b and c) is completely transparent in its zi mi erbek‘ lit. ‘in order that not at any time’, erbek‘ being a negative polarity item (cf. erb ‘when[?]’).

The remaining categories belonging here show mēkéti + subjunctive or, in a single instance, mēkéti + optative in Greek. These passages convey the notion ‘no longer, never again’, which is expressed in a variety of ways across all six languages, with variant modality in OCS, Old English, and Greek:

(29) a. ‘I command you, come out of him and never enter him again’ (Mk. 9.25)

Gk.	egṑ epitássō soi, ékselthe eks autoũ kaì mēkéti eisélthēis eis autón
OCS	azŭ ti veljo̜ iziti iz nego. i k tomu ne vĭnidi vĭ nĭ
Goth.	ik þus anabiuda: usgagg us þamma jah þanaseiþs ni galeiþais in ina
Arm.	es tam k‘ez hraman. el i dmanē. ew ayl ews mi mtanic‘es i da
Lat.	(24) ego tibi praecipio exi ab eo et amplius ne introeas in eum
OE	ic beode đé, gá of him, and ne gá đú leng on hine

b. ‘May nobody ever eat fruit from you again’ (Mk. 11.14)

Gk.	mēkéti eis tòn aiō̃na ek soũ mēdeìs karpòn phágoi
OCS	kŭ tomu otŭ tebe vŭ věkŭ niktože ploda ne sŭněstŭ
Goth.	ni þanaseiþs us þus aiw manna akran matjai
Arm.	mi ews ok‘ yawitean i k‘ēn ptowł keric‘ē
Lat.	non amplius in aeternum quisquam fructum ex te manducet
OE	Heonon forþ on écnesse ne ete ǽnig man wæstm of đé

In the first of these passages, the clause we are concerned with contains an imperative in OCS and (ambiguously) OE but a subjunctive in the remaining languages. Greek mēkéti is matched here by OCS k(ŭ) tomu ne, Gothic þanaseiþs ni/ni þanaseiþs, Armenian ayl ews mi, lit. ‘another (time) further not’, Latin amplius ne, and OE ne … leng, lit. ‘no longer’. The same combination of adverbial negation and mood is seen in Gothic in (29b), with Latin differing only in showing its non-modal negation non.²⁴ However, OCS shows a perfective present (‘will eat’). The most notable aspect of the Greek passage here is of course its optative, a rarity within the New Testament as a whole.²⁵

The final mēkéti passage is the Matthaian parallel of (29b), which OCS treats in the same way as it does Greek third person imperatives (see below, section 11): with da + indicative. The adverbial value of the Greek negation is matched by OCS nikoliže, lit. ‘not at any time’, OE nǽfre ‘never’ (both with incorporated negation functioning in addition to the primary negation), Armenian aysowhetew ‘henceforth’, and Latin numquam ‘never’ (without additional primary negation):

(30) ‘May fruit no longer come forth from you forever’ (Mt. 21.19)

Gk.	mēkéti ek soũ karpòs génētai eis tòn aiō̃na
OCS	da nikoliže ploda otŭ tebe ne bo̜detŭ vŭ věkŭ
Goth.	–
Arm.	mi aysowhetew i k‘ēn ptowł ełic‘i i [M: ∅] yawitean
Lat.	numquam ex te fructus nascatur in sempiternum
OE	Ne wurđe nǽfre wæstm of đé ácenned

The employment of OCS da here is continuous with and no doubt older than its occurrence in purpose clauses, as noted by Večerka (1993: 81). From a synchronic perspective this usage (‘optatival da’) must be distinguished from the conjunctive employment of this particle that we have seen elsewhere in this study. Yet it is clear that the conjunctive employment ‘in order that’ is a syntactic implementation of the optatival value, inasmuch as the effectuation of the purpose clause is the wish of its subject. This is true irrespective of whether we would wish to pursue the more glottogonic aspect of Večerka’s analysis and trace da back to some primordial exclamatory element.

We complete our coverage of Greek modal negation by mentioning those passages containing Gk. ou mḗ + subjunctive (Appendix XXI). For the purposes of this article, this is a marginal category, because it does not signal an irrealis but rather is the way New Testament Greek negates a future of a particular type. Of the 12 instances of this sort in my data, ten represent Machtsprüche of Jesus amounting to strong predictions (7× amḕn légō humĩn/soi, 2× légō gàr humĩn, 1× éstai gàr tóte), and two represent the prophetic words of Isaiah. This Greek construction is a grammaticalized continuation of Classical usage and is described by Blass-Debrunner (1979: 295, n. 1) as ‘niedere Koine’.

OCS responds in these instances with either a perfective present or a periphrastic future with imati + infinitive, while the Germanic languages employ their presents qua futures, Armenian its subjunctive qua future, and Latin its future. One example of this type with OCS imati + infinitive was passed over without comment in (11b) above, where our focus was on the conditional protasis. Another, showing an OCS perfective present, is the following:

(31) ‘Amen, I say to you, you won’t get out of there until you have paid (your) last penny’ (Mt. 5.26)

Gk.	amḕn légō soi, ou mḕ eksélthēis ekeĩthen, héōs àn apodō̃is tòn éskhaton kodrántēn
OCS	aminĭ glagoljǫ tebě ne izideši otŭ tǫdě. donĭdeže vŭzdasi poslědĭnii kodrantŭ
Goth.	amen qiþa þus: ni usgaggis jainþro, unte usgibis þana minnistan kintu
Arm.	amēn asem k‘ez. oč‘ elanic‘es anti. minč‘ew hatowc‘anic‘es zyetin nak‘arakitn
Lat.	amen dico tibi non exies inde donec reddas novissimum quadrantem
OE	Sóþes ic secge đé, Ne gǽst đú đanone, ǽr đú ágylde đone ýtemestan feorþlinge

The bulk of the clauses considered in this section are very close to negative purpose clauses in which, as we have shown in detail, OCS employs da + indicative. In a smaller group Greek shows mḗ + aorist subjunctive, and in these passages OCS shows, for the most part, imperatives. In a few instances, where the syntax of the Greek text is changed, OCS shows an infinitive.

10 Miscellaneous other clause types involving a Greek subjunctive (Appendix XXII–XXV)²⁶

The remaining occurrences of the subjunctive in Greek are independent employments not involving a subordinating conjunction or particle. These may occur in certain clause types, such as indirect questions and complement clauses, or, particularly in the first person plural, they may be exhortative or signal a deliberative question. Based on the particular type and its OCS responses, eight different categories may be discerned.

We will begin with indirect questions. As might be expected, these do not constitute a widely represented category. There are fewer than 20 such instances in the Gospels, just over a third of which involve the OCS conditional (Trost 1972: 1929). The closer these are to purpose clauses, the more likely they are to show the conditional, as in the following passages:

(32) a. ‘And going out, the Pharisees immediately took counsel with the Herodians concerning him (as to) how they might kill him’ (Mk. 3.6)

Gk.	kaì exelthóntes hoi Pharisaĩoi euthéōs metà tō̃n Hērōdianō̃n sumboúlion epoíoun kat’ autoũ, hópōs autòn apolésōsin
OCS	I abie išedŭše Farisěi sŭ Irodiěny. sŭvětŭ tvorěxǫ na nĭ kako i bǫ pogubili
Goth.	jah gaggandans þan Fareisaieis sunsaiw miþ þaim Herodianum garuni gatawidedun bi ina, ei imma usqemeina
Arm.	Ew eleal artak‘s vałvałaki p‘arisec‘ik‘n Hērovdianosawk‘n handerj. xohowrd [M: xorhowrd] ar̄nein vasn nora ‧ t‘e ziard? korowsc‘en zna
Lat.	Exeuntes autem statim Pharisaei cum Herodianis consilium faciebant adversus eum quomodo eum perderent
OE	Ðá Pharisei mid Herodianiscum útgangende þeahtedon ongén hine, hú hí hine fordón mihton

b. ‘And the scribes and chief priests heard and sought how they might kill him’ (Mk. 11.18)

Gk.	kaì ḗkousan hoi grammateĩs kaì hoi arkhiereĩs kaì ezḗtoun pō̃s autòn apolésōsin
OCS	I slyšašę kŭnižŭnici i arxierei. i iskaaxǫ kako i bǫ pogubili
Goth.	jah gahausidedun þai bokarjos jah gudjane auhumistans jah sokidedun, ƕaiwa ina usqistidedeina
Arm.	Lowan k‘ahanayapetk‘n ew dpirk‘. ew xndrein t‘e orpēs korowsc‘en zna
Lat.	Quo audito principes sacerdotum et scribae quaerebant quomodo eum perderent
OE	Ðá đæra sacerda ealdras and đá bóceras đis gehýrdon, hí þohton hú hí hine forspildon

In (32a) and (32b) the content of the complement clause is identical, the only difference being the precise nature of the complementation, which is dependent upon the variant syntax of ‘took counsel’ and ‘sought’. In the former instance, the complementation is a content clause ‘as to how they might kill him’, whereas in the latter it is more directly adverbial ‘how they might kill him’. Whereas the OCS and Latin renditions of these are exactly the same, and OE shows a minor lexical difference, Armenian employs a complementizer t‘e followed by different subordinating interrogatives in the two. Greek treats the first as a purpose clause (‘they took counsel in order to kill him’), employing hópōs, and the second as an indirect question (pō̃s) and is in this regard followed by Gothic (ei vs. ƕaiwa).

In the following passage of a similar nature, however, OCS employs an indicative:

(33) ‘But the Pharisees, going out, took counsel against him, (as to) how they might destroy him.’ (Mt. 12.14)

Gk.	ekselthóntes dè hoi Pharisaĩoi sumboúlion élabon kat’ autoũ hópōs autòn apolésōsin
OCS	Farisěi že išedŭše sŭvětŭ sŭtvorišę na nĭ. kako i pogubętŭ
Goth.	–
Arm.	Isk P‘arisec‘ik‘n eleal artak‘s. xorhowrd ar̄in znmanē. t‘e orpēs korowsc‘en zna.
Lat.	Exeuntes autem Pharisaei consilium faciebant adversus eum quomodo eum perderent
OE	Đa Sunder-hálgan eodun ðá út sóþlíce, and worhton geþeaht ongén hyne, hú hi hyne forspildon

Here again, Greek shows hópōs + subjunctive, as in purpose clauses. Except for OCS the translations here are like those of either (32a), (32b), or both.

In those instances where the indirect question is less like a final clause, OCS shows variation between a conditional and an indicative:

(34) a. ‘… for their eyes were heavy, and they did not know what answer they should give him’ (Mk. 14.40)

Gk.	ē̃san gàr autō̃n hoi ophthalmoì katabarunómenoi, kaì ouk ḗidēsan tí apokrithō̃sin autō̃i
OCS	běašete bo imŭ oči tęgotĭně. i ne uměxǫ čŭto bǫ otŭvěštali emu
Goth.	–
Arm.	k‘anzi ein ač‘k‘ noc‘a canrac‘eal. ew oč‘ gitein t‘e zinč‘ patasxani tayc‘en nma
Lat.	erant enim oculi illorum ingravati et ignorabant quid responderent ei
OE	hyra eagan wǽron gehefegode. And hí nyston, hwæt hí him andswaredon

b. ‘Do not worry about your soul (as to) what you might eat and what you might drink, nor about your body (as to) what you might wear’ (Mt. 6.25)

Gk.	mḕ merimnãte tē̃i psukhē̃i humō̃n tí phágēte kaì tí píēte, mēdè tō̃i sṓmati humō̃n tí endúsēsthe
OCS	ne pĭcěte sę dušejo̜ vašejo̜. čto ěste li čto piete. ni tělomŭ vašimŭ vŭ čto oblěčete sę
Goth.	ni maurnaiþ saiwalai izwarai ƕa matjaiþ jah ƕa drigkaiþ, nih leika izwaramma ƕe wasjaiþ
Arm.	mi hogayk‘ vasn ogwoc‘ jeroc‘ zinč‘ owtic‘ēk‘ kam zinč‘ ǝmpic‘ēk‘. ew mi vasn marmnoc‘ jeroc‘ et‘e zinč‘ aganic‘ik‘
Lat.	ne solliciti sitis animae vestrae quid manducetis neque corpori vestro quid induamini
OE	(ic secge eow), đæt ge ne sín ymbhýdige eowre sáwle, hwæt ge eton; ne eowrum líchaman, mid hwam ge sýn ymbscrýdde

c. ‘And he sought how he might hand him over at an opportune time’ (Mk. 14.11)

Gk.	kaì ezḗtei pō̃s eukaírō̃s autòn paradō̃i
OCS	i iskaaše kako i vŭ podobŭno vrěmę prědastŭ
Goth.	jah sokida ƕaiwa gatilaba ina galewidedi
Arm.	ew xndrēr t‘e ziard parapov matnesc‘ē zna
Lat.	et quaerebat quomodo illum opportune traderet
OE	And he smeade hú he hine dígellíce sealde

(34a) shows an OCS conditional beside subjunctives in all the other languages. In (34b) the Armenian subjunctives are most likely to be interpreted as futures (‘Do not worry … about what you will eat, etc.’). In (34c) OCS kako functions in a manner similar to da, as may be seen in the parallel Mt. 26.16 iskaše podobŭna vrěmeni. da i prědastŭ ‘He was seeking for a fitting time that he might hand him over’, where Greek shows hína + subjunctive.

In another passage both OCS and Armenian show an infinitive, while the other languages show a structure of the sort seen in (34):

(35) ‘And in those days, the crowd being (v.l. once again) very large and not having anything to eat …’ (Mk 8.1)

Gk.	En ekeínais taĩs hēmérais (v.l. pálin polloũ) ókhlou óntos kaì mḕ ekhóntōn tí phágōsin …
OCS	Vŭ ty dĭni paky mŭnogo so̜štu narodu. i ne imo̜štemŭ česo ěsti …
Goth.	In jainaim þan dagam aftra at filu managai manageim wisandein jah ni habandam ƕa matjaina …
Arm.	Yawowrsn yaynosik darjeal žołovowrd bazown ēr. ew inč‘ oč‘ ownein owtel
Lat.	in illis diebus iterum cum turba multa esset nec haberent quod manducarent …
OE	Eft on đam dagum, him wæs mid micel menigu, and næfdon hwæt hí ǽton

In a second group of passages, Greek shows a subjunctive in a complement clause. An overt complementizer is lacking in these instances in Greek, as it is in Latin. However, it is always present in Classical Armenian (either zi or t‘e) and Gothic (ei). OCS shows variation between the overt complementizer da and its absence, as may be observed in the following similar passages, where all languages but OCS show subjunctives:

(36) a. ‘What do you wish that I should do for you?’ (Mk. 10.51)

Gk.	tí théleis poiḗsō soi;
OCS	česomu xošteši da sŭtvorjo̜ tebě27
Goth.	ƕa wileis ei taujau þus?
Arm.	zinč‘? kamis t‘e araric‘ k‘ez
Lat.	quid vis tibi faciam
OE	Hwæt wylt đú đæt ic đé dó

b. ‘What, then, do you wish that I should do to the one you call king of the Jews?’ (Mk. 15.12)

Gk.	tí oũn (v.l. thélete) poiḗsō (v.l. hòn légete) basiléa tō̃n Ioudaíōn;
OCS	čto ubo xoštete sŭtvorjo̜ egože glagolete cěsarě Ijudeiska
Goth.	ƕa nu wileiþ ei taujau þammei qiþiþ þiudan Iudaie?
Arm.	isk zinč‘? kamik‘ t‘e araric‘ zark‘ayn hrēic‘
Lat.	quid ergo vultis faciam regi Iudaeorum
OE	Hwæt dó ic be Iudea cyninge?

The OE of (36b) is structured as a main clause question.

In another passage of this sort, Latin and OE show indicative verb forms:

(37) ‘Do you wish, then, that we go and gather them?’ (Mt. 13.28)

Gk.	théleis oũn apelthóntes sulléksōmen autá;
OCS	xošteši li ubo. da šedĭše isplěvemŭ ję
Goth.	–
Arm.	kamis? zi ert‘ic‘owk‘ k‘ałesc‘owk‘ zayn i bac‘
Lat.	vis imus et colligimus ea
OE	Wylt đú we gáþ, and gaderiaþ hig?

The logic of this usage is unclear. The only thing separating (37) from the passages of (36) is that in the former the complement clause occurs within a yes/no-question, whereas in the latter we are dealing with a wh-question.

In (38), with a different wh-element, we observe subjunctives in both Latin and OE:

(38) ‘Where do you wish that we should prepare for you to eat the paschal lamb?’ (Mt. 26.17)

Gk.	poũ théleis hetoimásōmen soi phageĩn tò páskha?
OCS	Kŭde xošteši ugotovaemĭ tebě pasxo̜ ěsti
Goth.	–
Arm.	owr? kamis zi patrastesc‘owk‘ k‘ez owtel zatikn [M: zzatiks]
Lat.	ubi vis paremus tibi comedere pascha
OE	Hwær wylt đú đæt we gegearwion đé đíne þénunga to eastron

There remain only independent employments of the Greek subjunctive in the first person plural to signal an exhortative or a deliberative question. In the first of these OCS normally employs its own first person plural imperative. But Greek often employs a preceding exhortative particle (deũte or áphes), and this may be matched in OCS by the imperative of the verb priiti (= deũte) or ostati (= áphes):

(39) a. ‘Let us go into the neighboring towns’ (Mk. 1.38)

Gk.	ágōmen eis tàs ekhoménas kōmopóleis
OCS	iděmŭ vŭ bližĭnęję vĭsi i grady
Goth.	gaggam du þaim bisunjane haimom jah baurgim
Arm.	ekayk‘ ert‘ic‘owk‘ ew [M: ∅] yayl [M: -ł] mawtawor giwłak‘ałak‘sn [M: -k‘s]
Lat.	eamus in proximos vicos et civitates
OE	Fare we on gehende túnas and ceastra

b. ‘Come, let’s kill him and take possession of his inheritance’ (Mt. 21.38)

Gk.	deũte apokteínōmen autòn kaì skhō̃men tḕn klēronomían autoũ
OCS	priděte ubimŭ i. i udrŭžimŭ dostoěnie ego
Goth.	—
Arm.	ekayk‘ spanc‘owk‘ zsa ew kalc‘owk‘ zžar̄angowt‘iwn sora
Lat.	venite occidamus eum et habebimus hereditatem eius
OE	uton gán, and ofslean hyne and habban us hys ǽhta

c. ‘Let us see if Elijah comes to save him’ (Mt. 27.49)

Gk.	áphes ídōmen ei érkhetai Ēlías sṓsōn autón
OCS	ostani da vidimŭ. ašte pridetŭ Iliě sŭpastŭ ego
Goth.	let, ei saiƕam qimaiu Helias nasjan ina
Arm.	t‘oł tesc‘owk‘ et‘e gay Ēłia p‘rkel zda
Lat.	sine videamus an veniat Helias liberans eum
OE	Lǽt; uton geseon hwæder Helias cume, and wylle hyne álýsan

In (39a) the only language that does not show a first person plural subjunctive (or, as in Gothic, a first person plural indicative employed modally) standing alone in translation of Gk. ágōmen is Classical Armenian, which uses an asyndetic double verb construction ekayk‘ ert‘ic‘owk‘, lit. ‘come, let us go’. In this regard it is like (39b), where Gk. deũte is matched everywhere but in Old English by an asyndetic imperative of the verb ‘come’. OE shows its frequent uton + infinitive construction²⁸ extending across two conjoined clauses. Here only Latin treats the second conjoined verb differently from the first, taking the structure to mean ‘Come, let us kill him and we shall take possession of his inheritance’. In (39c) all the languages translate Greek áphes with their own independent verb, but it is only OCS and Gothic that employ a complementizer (da and ei, respectively) between this and the first person plural main verb. However, whereas Gothic ei is regularly a complementizer, OCS da is ambiguous between this value and that of an independent hortatory-like particle appearing with the imperative, which in this case is homophonous with the indicative. OE once again has recourse to its uton + infinitive construction.

In a final passage of this sort, the value of OCS da is ambiguous in a slightly different way:

(40) ‘And they said to him, “Let us go and buy bread for two hundred dinars” ’ (Mk. 6.37)

Gk.	kaì légousin autō̃i, apelthóntes agorásōmen dēnaríōn diakosíōn ártous
OCS	i glagolašę emu. da šedŭše kupimŭ dŭvěma sŭtoma pěnędzŭ xlěby
Goth.	–
Arm.	Asen c‘na. ert‘ic‘owk‘ gnesc‘owk‘ erkeriwr dahekani hac‘
Lat.	et dixerunt ei euntes emamus denariis decentis panes
OE	Ɖa cwǽdon hí, Uton gán and mid twám hundred penegum hláfas bicgan

We have of course presented extensive evidence for the employment of da as a subordinating conjunction in purpose clauses; but it can also be employed as a complementizer following a verb of speaking as well as independently with imperatives, as noted above. In (40) the clause in which da is situated follows a verb of speaking, where it could, uniquely among all the structures presented here, be interpreted as a complementizer (‘And they said to him that …’). On the other hand, it might represent an employment of da in hortatory value with a first person imperative. This would make the OCS structure identical, save for the particle, to that of Greek and Latin, both of which employ a conjunctive participle (apelthóntes and euntes, respectively) together with the main verb. Armenian, on the other hand, once again employs an asyndetic double verb construction (ert‘ic‘owk‘ gnesc‘owk‘, lit. ‘let us go, let us buy’) and OE shows a conjoined uton + infinitive structure (lit. ‘let us go and buy’).

The final category in this section consists of direct deliberative questions in the first person (singular and plural). In all instances OCS shows an indicative:

(41) a. ‘Therefore don’t be worried, saying, “What shall we eat?”, or “What shall we drink?”, or “What shall we wear?” ’ (Mt. 6.31)

Gk.	mḕ oũn merimnḗsēte légontes· tí phágōmen; ḕ· ti píōmen; ḕ· tí peribalṓmetha;
OCS	ne pĭcěte sę ubo glagoljo̜šte. čto ěmŭ li čto piemŭ. li čimŭ odeždemŭ sę
Goth.	ni maurnaiþ nu qiþandans: ƕa matjam aiþþau ƕa drigkam aiþþau ƕe wasjaima?
Arm.	Mi aysowhetew hogayc‘ēk‘· ew asic‘ēk‘ zinč‘ keric‘owk‘. kam zinč‘ arbc‘owk‘. kam zinč‘ zgec‘c‘owk‘
Lat.	nolite ergo solliciti esse dicentes quid manducabimus aut quid bibemus aut quo operiemur
OE	Nellen ge eornustlíce beon ymbhýdige, đus cweđende, Hwæt ete we? ođđe, Hwæt drince we? ođđe, Mid hwam beo we oferwrogene?

b. ‘Is it permitted to pay tribute to Caesar or not? Shall we pay or shall we not pay?’ (Mk. 12.14)

Gk.	éksestin kē̃nson kaísari doũnai ḕ oú; dō̃men ḕ mḕ dō̃men;
OCS	dostoitŭ li dati kinĭsŭ kesarevi ili ni. damĭ li ili ne damĭ
Goth.	skuldu ist kaisaragild giban kaisara, þau niu gibaima?
Arm.	aržan ē hark tal kayser. et‘e oč‘· tac‘owk‘? t‘e oč‘ tac‘owk‘
Lat.	licet dari tributum Caesari an non dabimus
OE	(14–15) Álýfþ gaful to syllanne đam Casere? (15) Hwæder đe we ne syllaþ?

c. ‘What should I ask for?’ (Mk. 6.24)

Gk.	tí aitḗsōmai;
OCS	česo prošo̜
Goth.	ƕis bidjau?
Arm.	zinč‘? xndrec‘ic‘
Lat.	quid petam
OE	Hwæs bidde ic?

In each of these passages, the question may be viewed as having future reference; and this accounts for the exclusive employment of future tense forms of the verb in Latin in (41a) and (41b), leading us to assume the same of the ambiguous petam in (41c). The Armenian subjunctive, too, can be viewed as a future in each instance, and the OCS verbs are all perfective or anaspectual presents in future value. Gothic shows a subjunctive everywhere but in the first two questions of (41a). One wonders whether rhythmic factors may have resulted in the shortening of these from subjunctive to indicative in this sequence of three verbs, with the consequent instantiation of Behaghel’s Law. Certainly, at least (41b) makes it clear that one should expect a subjunctive in the first person plural of deliberative questions; and (41c) suggests that this applies to first person singulars as well. The OE evidence is clear only in (41b), where the verb syllaþ is an indicative used in future value. The first two verbs of (41a) are ambiguous between indicatives employed as futures and subjunctives, because when first (or second) person plural pronouns immediately follow the verb, the ending is shortened to -e (Sievers & Brunner 1965: 276). However, because they translate Latin futures, it is likely that they too are to be understood as indicatives employed in future value. The same is true of OE bidde in (41c). The third verb of (41a), beo, is also formally ambiguous,^{29, 30} but the general future orientation of beon (Campbell 1959: 350), together with the fact that it, too, translates a Latin future, suggests that it is most likely an indicative employed as a future here as well.

In this section the verbs seen in OCS include indicatives, imperatives, and, rarely, conditionals. The determination of the exact value of the verb in those instances where a preceding main clause is in the present indicative is rendered difficult by the vexed question of verbal aspect. According to the traditional understanding, perfective verbs ought to have future value and imperfective verbs should signal present tense. Therefore, verbs such as sŭtvorjǫ in (36ab), isplěvemŭ in (37), damĭ in (41b) (cf. Vaillant 1964: 321), and probably oblěčete sę in (34b) and odeždemŭ sę in (41a) are to be understood as having future tense value, while the rest are unclear.

However, in his recent discussion of the aspectual system of OCS, Kamphuis (2020) demonstrates that it is best described as having three aspectual classes: imperfective, perfective, and anaspectual. Kamphuis assigns to the unaspectual category both the otherwise unmarked verbs ěmŭ and piemŭ in (41a), as well as prošǫ in (41c), imputing to all of these the possibility of being employed as perfective presents under conditions suggesting future value. In English as well, one frequently hears sentences like When he asks me my political affiliation, what do I say?’ (beside what should I say? [somewhat less frequently, in my experience, what will I say?]). In instances where the main verb is in the past tense, such as (33), a future reading would not be felicitous, and therefore we are dealing with the same problem that arises in purpose clauses but without da to disambiguate the mood. It is not surprising, therefore, that in the nearly identical (32b) the conditional is employed as a way of highlighting the irrealis.

11 Passages where Greek does not show a subjunctive (Appendix XXVI–XXVII)

There remain only two categories, neither of which shows a subjunctive in Greek, but which in part show some of the same constructions in OCS that we have seen throughout this article. We therefore include this group in order to fill out the picture of how OCS negotiates the absence of an irrealis. The first of these comprises a very large number of passages where the Greek text has a third person imperative. The overwhelming majority of these show OCS optatival da + indicative, while a few instances show indicatives without da or imperatives:

(42) a. ‘Let your left hand not know what your right is doing’ (Mt. 6.3)

Gk.	mḕ gnṓtō hē aristerá sou tí poieĩ hē deksiá sou
OCS	da ne čjuetŭ šjuica tvoě čĭto tvoritŭ desnica tvoě
Goth.	ni witi hleidumei þeina ƕa taujiþ taihswo þeina
Arm.	mi gitasc‘ē jax k‘o zinč‘ gorcē aǰ k‘o
Lat.	nesciat sinistra tua quid faciat dextera tua
OE	nyte đín wynstre hwæt dó đín swýđre

b. ‘And if the house be worthy, let your peace come upon it’ (Mt. 10.13)

Gk.	kaì eàn mèn ē̃i hē oikía aksía, elthátō hē eirḗnē humō̃n ep’ autḗn
OCS	i ašte ubo bo̜detŭ domŭ dostoinŭ. pridetŭ mirŭ vašŭ na nĭ
Goth.	–
Arm.	et‘e ic‘ē town aržani. ekec‘ē [M: -esc‘ē] ołǰoyn jer i veray nora
Lat.	et siquidem fuerit domus digna veniat pax vestra super eam
OE	And gyf đæt hús witodlíce wyrđe biþ, eower syb cymeþ ofer hyt

c. ‘Just as you have believed, so let it be for you’ (Mt. 8.13)

Gk.	hōs epísteusas genēthḗtō soi
OCS	ěkože věrova bo̜di tebě
Goth.	swaswe galaubides wairþai þus
Arm.	orpēs hawatac‘erd ełic‘i k‘ez
Lat.	sicut credidisti fiat tibi
OE	gewurđe đé, swá swá gelýfdest

Although, to be sure, OCS possesses a living third person singular imperative, as in (42c), it is not of frequent occurrence.³¹ It is easy to understand the OCS indicative of (42b), because the sequence bo̜detŭ … pridetŭ shows a perfective present followed by a so-called determinative (perfective-like) verb of motion producing a future condition (‘If the house will be worthy, your peace will come upon it’) of the sort seen in (11b) and (12) in section 7 above. The same is true in (42b) of OE, where this is made explicit by the employment of biþ in the first clause, and where cymeþ must therefore be reckoned as a future. As for Armenian, Old English, Latin, and Gothic, the former two possess no third person imperative, and in the latter two this category is moribund. Consequently, all four languages must have recourse to subjunctives in all three passages.

Finally, a very small category involves passages in which Greek shows a future verb form. In these cases OCS shows either da + perfective present or an imperative:

(43) a. ‘And whoever of you wishes to become the first will be the slave of all’ (Mk. 10.44)

Gk.	kaì hòs eàn thélēi humō̃n génesthai prō̃tos, éstai pántōn doũlos
OCS	i iže ašte xoštetŭ vŭ vasŭ byti starěi. da bo̜detŭ vĭsěmŭ rabŭ
Goth.	jah saei wili izwara wairþan frumists, sijai allaim skalks
Arm.	ew or kamic‘i i jēnǰ ar̄aǰin linel. ełic‘i amenec‘own car̄ay
Lat.	et quicumque voluerit in vobis primus esse erit omnium servus
OE	And se đe wyle on eow fyrmest beon, se byþ ealra þeow

b. ‘But it will not be so among you’ (Mk. 10.43)

Gk.	oukh hoútōs dè éstai en humĩn
OCS	ne takožde že bo̜di vŭ vasŭ
Goth.	iþ ni swa sijai in izwis
Arm.	Oč‘ noynpēs ic‘ē ew i jerowm miǰi
Lat.	non ita est autem in vobis
OE	sóþlíce on eow hit nis swá

It is unclear why the OCS translator employed da in addition to the perfective present bo̜detŭ in (43a) and the third person singular imperative bo̜di in (43b). However, as we have previously noted, da possesses an optatival value,³² and it is possible that the OCS translator treated the contexts here as having the meanings ‘let him be the slave of all’ and ‘But let it not be so among you’, respectively. The other translations are straightforward: the Armenian subjunctive should in both passages be read as a future, and the Gothic subjunctive is the only option, given the impossibility of employing the present indicative of the imperfective verb wisan as a future. In (43b) the Greek Vorlage for the Vulgate must have contained the reading esti listed as a variant in Nestle & Aland (2012), and this was followed by the OE translator.³³

12 Conclusion

The tenor of this study has been primarily comparative. Its goal has been to demonstrate how the language of the OCS Gospels, which lacks an irrealis mood, handles contexts that in its five first-millennium CE partner versions are rendered in at least some instances in the subjunctive. In the process we have been led to explore clauses indicating purpose, result, conditions, subordination, modal negation, indirect questions, complements, exhortation, deliberation, indirect (i.e., non-second person) orders, and some futures. In certain of these instances, namely conditions and purpose or complement clauses, including those involving indirect questions, OCS utilizes a periphrastic conditional construction; but this usage is found in only a small minority of such clauses overall. From the point of view of English, the context where the absence of an irrealis in OCS causes the greatest degree of inconvenience is in purpose clauses (section 6), where it is the particle da, often together with a perfective verb, that signals, in the absence of any non-indicative modality, the irrealis nature of such clauses. Even more jarring are indirect questions like (34c), where ‘He sought how he will hand him over’ appears to an English speaker to not do justice to the modality inherent in the context.

In conditional clauses (section 7) it is once again a conjunction, in this case ašte, that carries the full load of signaling this category. However, it is important here to distinguish between factual/potential (If you do that, you will be punished/If you should be lucky enough to win, your future would be assured) and counterfactual conditions (If I had gotten the job, I’d be better off financially than I am now). In the latter, OCS employs a conditional. Conjunctions, this time together with a set of generalizing particles, serve as the basic markers of relative and subordinate clauses (section 8).

Another context where OCS is less explicit in signaling modality than some of the other versions we have been tracking is in negative injunctions (section 9), where, to be sure, it employs the imperative but unlike Greek, Armenian, and Latin lacks a distinctive modal negation. In this regard, OCS is like Germanic. Negative purpose clauses, like their positive counterparts, are expressed by da + perfective indicative.

In some indirect questions possessing future reference, such as (34c), where every other language shows subjunctives, OCS must resort, with less than satisfactory results, to its aspect system, which allows both perfective and anaspectual presents to assume future value, the former obligatorily, the latter optionally.

Finally, in all the languages we have been tracking here, with the exception of Gothic, indicative preterites are capable of taking on irrealis (modal) value. This is most clearly seen in (10), and it is true to this day in English (If I had not/Had I not been there, he would have harmed himself.). In OCS and Classical Armenian it is the imperfect which may signal this value, whereas in Greek it may be true of both the imperfect and the aorist, in Latin the imperfect.

In this study we have placed at the fore the idea of an irrealis, whose status as a useful linguistic category has sometimes been questioned (e.g., by Bybee 1998). It is therefore encouraging to see that the cross-linguistic utility of this concept has recently been robustly supported by von Prince, Krajinović, & Krifka (2022). These authors find the irrealis to extend to all the categories we have found relevant here; but as they note, no language morphologically encodes all irrealis categories. Viewed in this light, Old Church Slavic is a language that simply does not go as far as the other languages to which we have compared it here in grammaticalizing this category.