Mass or Count Noun: Latin Considerations of the Use of sanguis in the Plural

Abstract

Educated by generations of grammarians who state that the term sanguis (blood) is used only in the singular, Fathers of the Church, exegetes, and commentators were confronted with about twenty scriptural, essentially veterotestimentary tokens where sanguis is used in the plural. Justifications for this particular use appear throughout the commentaries. My study will attempt to answer a series of questions. Which passages interested the commentators the most and why? Which grammarians were involved and in what respect? What kind of justifications were provided? Was their interest purely hermeneutic or did the exegetes aim to preserve a state inherited from scrupulous translations? Were the passages treated in isolation or set in resonance? Does this assessment of the commentaries allow us to identity filiations? It will also be seen that this matter is another piece of the puzzle in the relationship between grammar and faith.

1. Introduction

Educated by generations of grammarians who state that the term sanguis (blood) is used only in the singular, Church Fathers, exegetes, and commentators were confronted with about twenty scriptural, essentially veterotestimentary tokens where sanguis is used in the plural. Justifications for this particular use appear throughout the commentaries, from Augustine, whose influence is, once more, striking.

Inherited from translations of Greek and/or Hebrew, languages in which the plurals דָּמִ֖ים and αἵματα are not grammatically problematic,1 the forms affected in the Vulgate are limited to the two obliques sanguinum and sanguinibus (Goelzer 1884, p. 299); see Appendix A 

Table A1. Table of the Biblical tokens.

Locus	Hebrew	Greek	Latin	Meaning
Ex 4.25-26	דָּמִ֛ים damim	αἷμα	sponsus sanguinum	blood of circumcisions
Ex 22.2-3	דָּמִ֛ים damim	---	sanguinis reus	murder
Dt 19.10	דָּמִ֛ים damim	αἷμα	sanguinis reus	murder
Dt 22.8	דָּמִ֛ים damim	φόνον (bloodshed)	sanguinem	accident
1 Sm 25.26	בְדָמִ֑ים ḇəḏāmîm	αἷμα	in sanguine	murder
1 Sm 25.33	בְדָמִ֑ים ḇəḏāmîm	αἵματα	ad sanguinem	murder
2 Sm 3.28	מִדְּמֵ֖י middəmê	αἱμάτων	sanguine	murder
2 Sm 16.7-8	הַדָּמִ֖ים/דָּמִ֛ים haddamim/damim	αἱμάτων	sanguinem	murder
2 Sm 21.1	הַדָּמִ֖ים had.da.mim	αἱμάτων	sanguinem	murder
2 Rg 9.26	דְּמֵ֣י dəmê	αἱμάτων.αἱμάτα	sanguine	murder
1 Par 22.8	דָּמִ֛ים damim	αἷμα	sanguinem	murder
1 Par 28.3	בְדָמִ֑ים ḇəḏāmîm	αἵματα	sanguinem	war
2 Mcc 8.3	---	αἱμάτων	sanguinis	murder
2 Mcc 14.18	---	αἱμάτων	sanguine	war
2 Mcc 14.45	---	αἱμάτων	sanguis	war
Ps 5.6	דָּמִ֛ים damim	5.7 αἱμάτων	5.7 uirum sanguinum	murderer
Ps 9.12	דָּמִ֛ים damim	9.13 αἵματα	9.13 quaerens sanguinem requirens sanguinem lxx	murder
Ps 15.4	מִדָּ֑ם middām	αἱμάτων	de sanguine de sanguinibus lxx	sacrifices
Ps 26.9	דָּמִ֛ים damim	25.9 αἱμάτων	25.9 cum uiris sanguinum	murderers
Ps 50.16	מִדָּמִ֨ים mid-dā-mîm	αἱμάτων	de sanguinibus	sins
Ps 55.23	דָּמִ֛ים damim	54.24 αἱμάτων	54.24 uiri sanguinum	murderers
Ps 59.2	דָּמִ֛ים damim	58.3 αἱμάτων	58.3 a (de lxx) uiris sanguinum	murderers
Ps 105.38	בַּדָּמִֽים׃ baddāmîm	αἷμασιν	sanguinibus in sanguinibus lxx	murder
Ps 139.19	דָּמִ֛ים damim	138.19 αἱμάτων	138.19 uiri sanguinum	murderers
Pro 29.10	דָּמִ֛ים damim	αἱμάτων	uiri sanguinum	murderers
Eccli 22.24	---	αἱμάτων	22.30 sanguinem	bloodshed
Eccli 34.21	---	αἱμάτων	34.25 homo sanguinis	murder
Is 1.15	דָּמִ֛ים damim	αἵματος	sanguine	murder
Is 9.5	בְדָמִ֑ים ḇəḏāmîm	καταλλαγῆς	sanguine	war
Is 33.15	דָּמִ֛ים damim	αἵματος	sanguinem	murder
Ier 19.4	דַּ֥ם dam	αἱμάτων	sanguine	murder
Ez 7.23	דָּמִ֛ים damim	---	iudicio sanguinum	murder
Ez 9.9	דָּמִ֛ים damim	ἀκαθαρσίας	sanguinibus	murder
Ez 22.2	הַדָּמִ֖ים haddamim	αἱμάτων	ciuitatem sanguinum	murder
Ez 24.6	הַדָּמִ֖ים haddamim	αἱμάτων	ciuitati sanguinum	murder
Ez 24.9	הַדָּמִ֖ים haddamim	---	ciuitati sanguinum	murder
Os 4.2	וְדָמִ֥יםwə-ḏā-mîm bəḏāmîm; בְּדָמִ֑ים	αἵματα ἐφ αἵμασιν	sanguis sanguinem tetigit sanguinem sanguini miscuerunt lxx	murder
Mi 3.10	בְּדָמִ֑ים bəḏāmîm	αἵμασιν	in sanguinibus	murders
Mi 7.2	לְדָמִ֣ים ləḏāmîm	αἵματα	in sanguine	murders
Na 3.1	דָּמִ֛ים damim	αἱμάτων	ciuitas sanguinum	murders
Hab 2.12	בְּדָמִ֑יםbəḏāmîm	αἵμασιν	ciuitatem in sanguinibus	murders
Hab 2.17	מִדְּמֵ֤י middəmê	αἵματα	de sanguinibus hominis propter sanguinem lxx	murder
Ioh 1.13	---	αἱμάτων	sanguinibus	conception of a child

at the end of the article. Used twelve times, sanguinum appears in the recurring expressions uir and especially uiri sanguinum (six occurrences, five of which are psalmic), ciuitas sanguinum (four instances, three of which are in Ezekiel). Sanguinibus appears nine times, one of them only in the iuxta lxx version—it is in Ps 15.4, where the singular מִדָּ֑ם is rendered by mean of plural αἱμάτων (οὐ μὴ συναγάγω τὰς συναγωγὰς αὐτῶν ἐξ αἱμάτων), adopted in the iuxta lxx version (de sanguinibus), but eliminated by the Vulgate (de sanguine). While the passages with sanguis in the plural do not reveal a perfect alignment of the scriptural languages, fifteen occurrences in the plural in Hebrew do have plural correspondents in Greek and Latin (Ps 5.6; Ps 26.9; Ps 50.16; Ps 55.23; Ps 59.2; Ps 105.38; Ps 139.19; Pro 29.10; Ez 9.9; Ez 22.2; Ez 24.6; Ez 24.9; Mi 3.10; Na 3.1; Hab 2.12; see

Table A3. Biblical abbreviations.

Ex	Exodus	Eccli	Ecclesiasticus
Dt	Deuteronomium	Is	Isaias
1 Sm	1 Samuel	Ier	Ieremias
2 Sm	2 Samuel	Ez	Ezechiel
2 Rg	2 Regum	Os	Osee
1 Par	1 Paralipomenon	Mi	Michaeas
2 Mcc	2 Macchabaeorum	Na	Naum
Ps	Psalmi	Hab	Habacuc
Pro	Prouerbia	Ioh	Iohannes

). The Scriptures do not, however, provide any instance of cruor in the plural, and the singular is attested only six times (Deut 21.9, 32.42; Ex 7.19; Num 35.33; III Reg 2.5; IV Reg 24.4). Sanguis is thus much more represented: as Ernout (1951, p. 32) points out, the distinction between the two tends to become blurred, and the weakening of the difference between them works in the favor of sanguis, to which the Latin translations of the Bible bear witness, while Isidore of Seville († 636), in his Differences and Etymologies, recalls, yet again, the ancient values of the two terms, with the problems of interpretation that will be seen. Finally, let us note that Jerome († 420), who claims to be a pupil of Donatus († c. 380),2 generally accepted the plurals of sanguis, and even reintroduced a Hebrew plural, that of Ez 9.9 (see below).

In order to better understand how the Church Fathers and other interpreters analysed those passages from the 4th century on, the study will first focus on the observations made by Roman grammarians about the status of sanguis in the plural. Their remarks indeed formed the scientific basis common to all interpreters; therefore, it is crucial to grasp what their had in mind while analysing biblical instances of plural sanguinum or sanguinibus. The second part of the study will concentrate on the role of grammatical rules in the analyses provided by a large number of biblical commentators during late Antiquity and the Middle Ages. The last part will focus on some of the consequences of these practices, such as grammarians taking into account the plural forms of sanguis found in the Scriptures while commenting on Donatus (see Section 3.1), or translators of Eastern tracts of theology using freely those plural forms (see Section 3.2), or even autonomous uses influenced by a continuous reading of the Bible that changed the Latin ecosystem from within3 (see Section 3.3).

2. The Grammatical Point of View

2.1. The Foundation of Ancient Grammarians: Sanguis Is a Singular Noun

The question of singularia tantum is evoked as early on as Aulus Gellius († c. 180), who tells, in his Attic Nights, of his master Fronto († 166/167) smiling at a friend who was saying that he had been cured of dropsy by a bath of “warm sands” (arenis calentibus).4 The conversation that ensues is, for Fronto, an opportunity to remind him that, according to Caesar’s De analogia, caelum, triticum and arena (sky, wheat, sand) can only be singular,5 and to justify this point of view as follows:

Since arena, in the singular, expresses the infinite multitude of parcels that make up what we call arena, it would be no less than idle to say arenae. This word does not include a plural, since the singular implies in itself the idea of multiplicity.

6

However, Fronto then adds:

For, after all, why do we always say caelum in the singular, when mare and terra, when puluis, uentus and fumus have plurals? … Why do mel, uinum and other similar words take the plural, while lac does not ?

7

In passing, Fronto therein endorses as perfectly Latin the plurals of ‘sea,’ ‘land,’ ‘dust,’ ‘wind,’ ‘smoke,’ words that will soon be dealt with again; our sanguis however is not mentioned. Fronto closes out the debate by inviting to seek out only from genuine authorities occurrences of quadrigae and arenae, which can be interpreted as a sign that the use of mass nouns in the plural would not shock the man in the street:

Go then, and when you have the leisure, look and see if quadrigae and arenae are not found in some poet or orator from Antiquity, those who have a rank, who have some authority, and who are not lost in the crowd of the masses.

8

In conclusion, Aulus Gellius admits to not having looked for occurrences of arenae, given that Caesar was actually the only author to have condemned the plural use of the word.9

Discussing, two centuries later, the issue of number, Donatus calls attention to the category of singularia tantum vs. pluralia tantum. This time he illustrates the category of masculine singularia tantum by way of two examples, puluis (whose plural form was actually accepted by Fronto) and sanguis, dust and blood.10 His list is augmented by Diomedes Grammaticus (late 4th century AD), who compares it to genius, fimus, limus, and muscus, that is genius, manure, mud, and moss.11 Grammarians then diverge in function of whether they follow Donatus’ short list, in the case of Consentius (5th c.)12 or Pompeius (5th c.),13 or Diomedes’s long list, in the case of Phocas (late 4th c./early 5th c.),14 which includes fumus, smoke (yet another plural accepted by Fronto). The case appears thus to be resolved and Pompeius insists on the fact that hi sanguines is not attested, conceding that hi cruores is however found (Verg. Aen. 4, 687), but he highlights that it is abusive (usurpatiue), in agreement with Servius (late 4th c./early 5th c.), who evokes abuse in the same terms15 (usurpauit), as well as with Donatus who singles out the Virgilian “abuses” (see note 30). The text takes place nonetheless in the section entitled De numeris quasi retractando, in which Pompeius looks for counterexamples to Donatus’ rules: the fact that sanguis is exclusively singular has no exception, and the contrary uses are placed under the vague category of abuse. His extremely careful analysis is worth citing in its entirety:

We must be attentive to the fact that every example provided by Donatus is done so in reference to grammar, to avoid resorting to an authority and ruining what he is putting forth. There are indeed many counter-examples, in the masculine singular like puluis (dust): no one says ‘dusts,’ and yet Horace says ‘disperse their dusts the ninth day.’16 This is a reference to an author, but it must be known that ‘dust’ is solely singular. And there is a reason. What is it? The thing is indeed indivisible and cannot be divided, and since it cannot be divided, it cannot have a plural form either. Could dust be part here and part there? The thing is by nature indivisible, and is therefore singular. Therefore, the author has committed an abuse. Likewise, ‘blood’ (sanguis): ‘bloods’ is not found; ‘blood’ (cruor): ‘bloods’ is found, and it is abusive.

17

According to Pompeius, Donatus’ examples come from the grammar of the language (secundum artem); one should therefore not turn to authors to counter what he says. Counterexamples do exist, but Horace’s pulueres does not change the fact that puluis is and will always be a singular noun. Pompeius demonstrates this by the fact that dust, as an indivisible thing that we cannot cut, nor divide into two distinct places, cannot have a signifier employed in anything but the singular. Pompeius therefore provides a very clear justification founded on the opposition between continuous and discrete, the importance of which will be considered later.

One century later, Priscian (6th c.) takes a new approach in distinguishing two types of exclusively singular nouns, those which are so by nature and those which are so solely in use (let us be reminded that Priscian never read Pompeius’ commentary, cf. Holtz 2019, 90 n.8). This second category is illustrated by Donatus’ two examples, sanguis and puluis, which are opposed to nouns that can only be singular since they designate, by nature, unique objects (gods, as well as the sun and the moon, and geographical locations.)18 As is specified in what immediately follows this passage, while proper nouns are invariably singular, appellative nouns may be used in the plural, without implying any change in meaning, as in Virgil using, with the weight of his authority, mella (14 occurrences), frumenta (4), hordea (3), farra (4), (honeys, wheats, barleys, spelts), as well as the well-known cruores already seen, which virtually opens the path to sanguines, and the very common cineres which in turn would give legitimacy to pulueres,19 precisely the term used by Horace in the sense of cineres. For Priscian, these singulars do not result from any kind of rule but exclusively from usage.20 Cui si collibuisset, says Priscian, in imagining the fact that plural uses of sanguis could have existed, although it is not the case; let it be noted that he does not draw upon biblical examples in support of this affirmation, leaving aside as always the uses that do not come from the Latin canon (Grondeux 2011).

2.2. Uncountable (Mass) Nouns: Two Systematic Treatments

The question of the plural sanguines is part of the thorny issue of mass nouns, as opposed to count nouns (Fournier 2007; Nicolas 2002; Lauwers and Vermote 2014; Lemghari 2014; Lemghari 2016). The position of Latin grammarians generally consists, once collective nouns have been set aside, in wanting to make the grammatical number correspond to the counting operations of the linguistic reality (Colombat 1993). Among Latin grammarians, two systematic presentations stand out, both due to their synthetic nature and to their points in common.21

The first is the common source of Charisius and Anonymus Bobiensis, dated respectively to the 4th and 5th centuries (de Nonno 1982). This lost text proposed six categories of semper singularia: (1) natural elements, (2) metals, (3) mesurable solids and (4) liquids (the term is used only by Anonymus Bobiensis), (5) geographical proper nouns, (6) incorporeals. Each category, or practically, is nuanced by the mention of their being nonetheless counter-examples (1b, 2b, 3-4b, 5b), but that these are strictly limited to the first cases of the plural (1c, 2c, 3-4c), thus excluding genitive, dative, ablative plurals, an incompleteness that implies that they are not true plurals. The first notable fact is that the author proposes to gloss the plural terrae by terrae regiones (1d), as if this use of the plural were actually an ellipsis. The second point of interest is the first time that incorporeals appear in the issue (6), a point to which I shall return. Presented in the table below, this synthesis shows the common points and the differences between the two uses of the lost source.

	Char. gramm. 38, 3–39, 22	Exc. Bob. gramm. I 550, 21–551, 7
1a	Elementa semper singularia sunt,
	uelut caelum οὐρανός, aether αἰθήρ, aer ἀήρ, sol ἥλιος, iubar {τὸ σέλας τοῦ ἡλίου}, terra γῆ, mare θάλασσα, autumnus et autumnum {φθινόπωρον μετόπωρον}.	aether caelum aer uer sol iubar autumnus terra mare.
1b	maria tamen quamuis dicantur pluraliter,	quamuis maria dicitur,
1c	attamen nec marium nec maribus dicemus.	nec marium nec maribus.
1d	et terras pro terrae regionibus accipimus.
2a	Item metallica semper singularia sunt,
	ut puta aurum stagnum argentum cassiterum {κασσίτερος} ferrum orichalcum aes aurichalcum plumbum.	aurum argentum ferrum aes plumbum cassiterum orichalcum aurichalcum.
2b	aera quamuis dicantur,	et quamuis aera dicimus,
2c	tamen ceteris casibus non utimur.
3	Item quae mensura constant, arida dumtaxat, quae numerari non possunt, semper singularia sunt,
	uelut triticum, <frumentum>; quamuis et frumenta dicamus: hoc far {ἡ ζειά}, hoc ador adoris frumenti species, haec oryza ὄρυζα, haec faba; Vergilius (georg. 1, 215) ‘uere fabis satio’, inusitata declinatio: haec lens {ἡ φακῆ}, hoc cicer {ἐρέβινθος}, hoc milium {κέγχρος}, hoc minium κιννάβαρις, hoc paniceum {ἔλυμος κέγχρος}, hoc git monoptotum μελάνθιον, hoc ordeum {κριθή}.	ut triticum, frumentum; quamuis et frumenta dicimus: far ζεά, ador ζειά, hordeum; ‘serit hordea’ Vergilius: faba, lenticula, lens, cicer, milium κέγχρος, paniceum μελίνεως κέγχρος, oryza.
4	Item haec semper singularia sunt, mel μέλι; quamuis Vergilius mella dixerit: oleum ἔλαιον, uinum οἶνος; quamuis Vergilius uina dixerit: mulsum οἰνόμελι, defrutum ἕψημα, acetum ὄξος, lac γάλα lactis, garum muria liquamen.	item liquida semper singularia, mel; et mella legimus: oleum, uinum; et uina legimus: mulsum οἰνόμελι, defrutum ἕψημα, acetum, lac, muria γάρον liquamen.
3-4b	et siqua horum nominatiuo pluraliter efferuntur,	et siquid horum nominatiuo plurali efferatur,
3-4c	ceteris tamen casibus cessabunt.	ceteris casibus cessabunt.
3-4d	[fel sic declinatur quo modo mel]. et si pluraliter declinata fuerint, non ad quantitatem sed ad genus referuntur, uelut mella, ut sint multae species, quasi locale, uelut Atticum Rhodium. similiter et lanae, ut sint uariae species lanae, quasi Milesia Hispana.
5a	Item nomina
	fluminum et montium et ciuitatium	montium, fluminum, ciuitatium
	semper singularia sunt,
5b	nisi quae natura pluraliter enuntiantur, ut Athenae Thebae
	Baiae Cumae Puteoli Ostia Brixae Abellae, hi Argi.	Puteoli Ostia.
6	Item singularia semper sunt quae nec uideri nec tangi possunt, uerum ab his in alterutram partem doloris aut gaudii adficimur, ut gaudium
	{χαρά}· quamuis Vergilius (A. 1, 502) et gaudia dixerit, ‘Latonae tacitum pertemptant gaudia pectus’: uigor {σθένος}, metus {δέος}, letum {θάνατος}, {fuga φυγή, pallor ὠχρίασις, luxuria ἀσωτία}, timor {φόβος}, terror {πτόησις}, insania {μανία}, sopor {κάρωσις}, ius {δίκαιον} et pluraliter iura tantum, iustitia {δικαιοσύνη}, maestitia {λύπη}, stultitia {μωρία}, senium {μέριμνα φύσεως} senilis morbus, scrupulum {σκάνδαλον}, sapientia {φρόνησις}, salus {σωτηρία ὑγίεια}, fides πίστις, perfidia ἀπιστία, macies {λεπτότης} maciei, maeror <λύπη> maeroris, misericordia {ἔλεος}, inertia {ἀδράνεια}, segnitia {νώθεια}, pigritia {ὄκνος}, uelocitas {ταχυτής}.	et ‘gaudia’ dixerunt: ‘uigor’, et ‘uigores’ dixerunt; ‘metus’, et ‘metus’ pluraliter; ‘letum, fuga, pallor’, ‘pallores’ dicitur; ‘timor, terror’, ‘terrores’ dicitur; ‘insania, sopor’, ‘sopores’ dicitur; ‘furor’ pluraliter dicitur; ‘ius’, pluraliter ‘iura’ tantum; ‘iustitia, maestitia, senium μέριμνα ψυχική’, maeror λύπη; Cicero pluraliter extulit in Pro Cluentio ‘sed multorum medicamentum maerorum’: ‘scrupulum, sapientia, salus, fides, perfidia, macies maciei, inertia, misericordia, segnitia, pigritia, uelocitas’ pluraliter dicitur.
6b	Sed auctores non usque quaque haec obseruauerunt.	Sed auctores non usque quaque hoc obseruauerunt.

To the common base, Charisius adds a justification: the plural uses are linked to the fact that varieties of these objects can be distinguished (3-4d). The designations of these objects can switch from the status of mass noun to that of count noun. One could thus speak of Attic or Rhodian honeys, or of Italian wines.22 The phenomenon is mentioned by Varro (116–27 BC; ling. 9, 40, 66–68, for uina and aquae23), and described thoroughly in Donatus’ circle using a single Ciceronian example,24 as in Servius,25 Pompeius 26 and Cledonius (5th c.).27 It is interesting to note that these grammarians focus here on a prosaic example, which contradicts the mella and other hordea, which are attributed to metrical constraints. As for the Bobbio Scholiast, he adds a mention based on another Ciceronian example ‘sed multorum medicamentum maerorum’ (“the remedy for many sorrows”: not only is maerorum a plural use of an appellative designating an incorporeal, but it is also a use in the genitive, attributable, what’s more, to a prose writer (Cic. Cluent. 201).

The second classification is that of Priscian (Ars V.46, 174.23 sq.), who begins by giving a general presentation of singularia tantum, analyzing this trait as linked to their nature (for proper nouns) or to their use (for certain common nouns28) (section A of the table below). He then points out four parameters likely to induce variations, the first three of which influence only proper nouns: (1) chance (two characters bearing the same name), (2) the partition of regions in two (one will speak, for example, of Citerior and Ulterior Gaul, (3) the diversity of opinions (some may believe in the existence of several suns and several moons). Priscian then inserts a brief development (B) on terms belonging to a kind of gray area: terra, uirtus, pietas, iustitia, probitas, which can designate goddesses (we would say personnified virtues), in which case they do not take plural forms, or things, corporeal or incorporeal, in which case they are completely admissible in the plural form. It is only after this transition between proper nouns and common nouns, in an absolutely remarkable arrangement, that he develops the fourth parameter, authority (4), which only comes into play for common nouns; on this occasion, he introduces a note (C) on the usage with respect to metals and measurables, and stresses once again that it is merely a usage and not a rule.

(A) Sunt quaedam nomina semper singularia uel natura uel usu: natura, ut propria, quae naturaliter indiuidua sunt: ‘Iuppiter’, ‘Venus’, ‘Ceres’, ‘Achilles’, ‘Hector’, ‘Sol’, ‘Luna’, ‘Italia’, ‘Sicilia’, ‘Cilicia’; usu, quae singulariter proferri tradidit usus, ut ‘sanguis’, ‘puluis’.

In his tamen ipsis est quando uel euentus uel partitio regionum uel diuersae opiniones uel auctoritas ueterum cogit uel quando pro appellatiuis quoque proferuntur, ut etiam plurali numero proferantur:

(1) Euentus, quod euenit saepe eodem proprio nomine duos uel plures nuncupari, ut ‘Aeneas’ Anchisae filius et Siluius ‘Aeneas’, ‘Pyrrhus’ filius Achillis et ‘Pyrrhus’ rex Epirotarum, ‘Aiax’ Telamonius et ‘Aiax’ Oilei filius. Haec enim non naturaliter communia sunt, quippe nullam qualitatem communem significant, sed casu diuersis contigere personis.

(2) Idem facit partitio regionum, ut ‘Gallia’ citerior et ‘Gallia’ ulterior,

(3) Diuersae opiniones, quod quidam multos soles et multas lunas esse putauerunt.

(B) ‘Terra’ quoque, ‘uirtus’, ‘pietas’, ‘iustitia’, ‘probitas’ et similia, quando deas significant, propria sunt, quando uero partes uel in hominibus sitas res, appellatiua sunt. Itaque si sint propria, non habent pluralem numerum, sin appellatiua, habent.

(4) Auctoritas, ut ‘mella’, ‘frumenta’, ‘ordea’, ‘farra’ plurali numero protulit Virgilius, cum in usu frequentiore singularis numeri sunt, sicut alia quoque plurima, ut ‘sanguis’, ‘puluis’, ‘pax’, quae tam singulariter quam pluraliter prolata idem possunt significare. Sed pluraliter non utimur eis, quia auctoritas deficit, cui si collibuisset, quomodo ‘cruores’, dicere ‘sanguines’, uel quomodo ‘cineres’, sic ‘pulueres’, nihil impediret.

(C) Sciendum tamen, quod metallorum uel seminum uel umidorum ad mensuram uel ad pensum pertinentium pleraque semper singularia inueniuntur, ut ‘aurum’, ‘argentum’, ‘aes’, ‘plumbum’, ‘stannum’, ‘uitrum’, ‘orichalcum’, ‘electrum’, ‘triticum’, ‘eruum’ ὄρυζα, ‘faba’, ‘lens’, ‘pisa’, ‘conchis’, ‘acetum’, ‘liquamen’, ‘mulsum’, ‘oleum’. Adeo autem haec usus, non regula prohibet etiam pluralia habere, quod quidam propria confisi auctoritate plurali quoque, ut dictum est, haec protulerunt numero, ut ‘ordea’, ‘frumenta’, ‘fabae’, ‘uina’, ‘mella’.

Among the common points shared by these developments, the most striking is the sudden introduction of incorporeals into the issue, a recurrent theme in the definition of the noun as capable of signifying corporeals or incorporeals. On the other hand, it should be noted that the source of Charisius and Anonymus Bobiensis was, unlike Priscian, silent on the question of the proper names of people. In any case, these two analyses synthesize the elements that are more or less common among Latin grammarians. For example, the opposition between countable and measurable is already present in Donatus, who isolates the latter category,29 but the same is not yet true for metals, which initially appear in series that mix all sorts of mass nouns (see Donatus, note 33: aurum argentum oleum ferrum triticum). The author of the Ars Laureshamensis, a 9th century commentary on Donatus, picked up on the problem and points out that if Donatus writes mensuram pondusue, it is because, by mensuram, he means liquids and solids, which are measured, and by pondus, he means metals, which are weighed.30 However, it has long become a topos to oppose the “prohibitions” of grammarians to Virgilian, and more generally, poetic, usages. As such, Pseudo-Probus points out that Virgilian plurals are poetic and do not correspond to the ratio, to the logic of the language, an idea that also appears in Aemilius Asper.31

However, as Charisius points out, plural uses do not modify the grammatical ecosystem and remain outside the norm.32 Moreover, grammarians attempt to limit the scope of these plurals, stressing that they are only used in a few cases (nominative, accusative, vocative), as if the limited number of forms prevented them from being authentic plurals.33 As such, it would be said that mella is but triptote34 or even aptote,35 and it is again Priscian who would qualify the statement by saying that those types of plurals are only rarely used in the genetive and dative.36 It goes without saying that the Cross Database Searchtool Brepolis or the Corpus corporum provide numerous instances of frumentorum, mellibus, iuribus, uinorum, etc.

These analyses are nonetheless very interesting, in that they suggest how watertight the categories of mass and count nouns are, or conversely the possibilities of switching from one category to another (Kleiber 1999, p. 112 apud Lemghari 2014; Kleiber 2014). Two approaches allow for the recategorization of mass nouns and count nouns, the first being to consider a thing either as a raw material (lead) or as an object proceeding from this material and distinct from another (fishing lead, which is not that of the mason37): a Rhodian honey is not an Attic honey; Italian wines are distinct from other geographical appellations. These uses are never identified as figurative, which suggests that they were not considered to be abnormal. The second approach is the pivotal point adopted by Priscian in his presentation of singularia tantum, namely incorporeals. If proper nouns are in essence singular, there is a borderline type, that of personnifications; ‘Earth,’ ‘Virtue,’ ‘Piety,’ ‘Justice,’ ‘Probity’ will then be proper and singular nouns, but when they designate common things, they can, on the other hand, take the plural.38 Let us note, in contrast, that the source common to Charisius and Anonymus Bobiensis clearly stated the impossibility of incorporeal nouns being used in the plural39, that is before listing, of course, counter-examples. The two approaches are thus diametrically opposed, which allows for the possibility that Priscian has here applied his own analysis of incorporeals, developed in Book II. Let us recall that Donatus had replaced, in the definition of the noun, the traditional statement that it meant a corporeal or incorporeal thing by the fact that it meant “a body or thing” (corpus aut rem), “in a proper or common way” (proprie communiterue), but without providing an exemplification for the four types outlined as such.40 Following him, commentators and grammarians set out to illustrate the category of “incorporeals with proper nouns,” and it is in this context that Priscian would turn to the example of Pudicitia (Modesty) borrowed from Livy, uirtus (virtue) being a contrario an example of an incorporeal with a common noun41 (Grondeux 2007, 2009). The consequence is this remodeling of the sequence present in the common source of Charisius and the Bobbio Scholiast, so that incorporeals come artificially to act as a pivot between proper nouns and common nouns. Despite this contrast in approaches, incorporeals are indeed part of the debate, which is not without evoking one of Beauzée’s reflections, in which he proposed to assimilate the names of metals to proper nouns, on the grounds that they did not have individuals contained in their species, and more closely associate their names with those of virtues and vices:

As for the remark in the gramm. gén. part. II. ch. jv. that there are several appellative nouns that have no plural, I am tempted to believe that this idea comes from the fact that what are actually proper nouns are taken for appellative nouns. The names of each metal, gold, silver, iron, are, if you will, specific; but what distinct individuals are found in this species? The same goes for the names of virtues and vices, justice, prudence, charity, hatred, cowardice, &c. & of several other words that have no plural in any language, unless they are taken in the figurative sense.

42

2.3. A Detour by Way of the Plural Cruores

A few late antique and early medieval scholiasts and grammarians (Servius, the Bobbio Scholiast, Priscian, Bede43) draw a parallel between sanguis and cruor,44 with regard to a Virgilian use (Verg. Aen. 4, 687: atros siccabat ueste cruores). Ancient poetry was, however, not sparing with this plural form, which raises questions about the focus on the Virgilian example. I will therefore now consider what is particular about this use, but I must first spend a bit of time on the word cruor in its association with sanguis. Ernout-Meillet’s dictionary explains the difference between these terms: cruor (s.v.) designates “spilled or coagulated blood, a pool a blood,’ as opposed to sanguis ‘blood in circulation,’ a distinction observed by the great authors, cf. Lucr. 2, 194–195, quod genus e nostro cum missus corpore sanguis/emicat exultans alte spargitque cruorem,” while sanguis (s.v.) designates “blood (that flows, different to cruor ‘coagulated blood’).” The two formulations are not completely identical: cruor sometimes designates “spilled or coagulated blood” (which is not exactly the same thing) as opposed to blood “in circulation,” and sometimes “coagulated blood” as opposed to liquid blood. The first opposition thus plays on the location of the matter in question, either outside or inside the body, while the second one plays on the state, rigid or liquid, of this matter.

In the 7th century, long after our grammarians, Isidore posits, in his Differences, and even more clearly in his Etymologies, that sanguis designates the blood contained in the body, and cruor, the same liquid when it is outside of the body.45 The distinction made in the Etymologies, with which Latinists are quite familiar, is the clearest and the most complete, opposing an enclosed liquid (dum in corpore est) and its diffused state (effusus). Even so, is this opposition really supported? In her edition of the Differentiae, Carmen Codoñer points out a problem of transmission between the source, namely the Pseudo Caper (2nd c.), and Isidore.46 Let us indeed compare the three versions:

Caper orth. 7, 99, 20: dum manat (manet BC), sanguis est, effusus uero cruor erit.

Isid. diff. I 374(529): sanguis est dum in corpore manet, effusus uero cruor fit.

Isid. etym. 11, 1, 122 (Gasti 2010): sanguis autem est dum in corpore est, effusus uero cruor dicitur.

Caper’s opposition between dum manat and effusus is hardly difficult: the blood in movement that flows (manat) while escaping from a body is opposed to the blood spilled on the ground, which has become an inert substance. However, the uncertainties of manuscript transmission have interfered with this, so that manat (‘flows’) has become manet (‘remains’), as testified by the Differentiae and confirmed by the reinterpretation found in the Etymologies. Let us note that it is not even sure that this should be considered, with Carmen Codoñer, as “mis-understanding,” because Isidore perhaps already had at his disposal a manuscript with the manet reading. In any case, let us retain that originally sanguis designates blood that flows, whose source can be identified, a single individual, and cruor blood that has flowed.

In ancient poetry, the plural cruores designates blood spilled in large quantities, in the context of sacrifices,47 wars,48 or multiple murders (in this instance, those which Medea premeditated over committing on her children49). Here, the blood spilled is always that of distinct individuals, in variable, but generally significant, quantities. These plural uses persist in the same way in the Middle Ages, and in all sorts of obliques, actually,50 be they sacrifices,51 martyrdoms52, massacres53 or wars.54 Uses designating visually isolatable and therefore potentially countable <drops of> blood55 or <stains> of blood (namely eye petechiae56) were thus quite different, as were, even more so, <flows of> blood continously escaping from a single being.57 This is what makes the Virgilian use unique: the poet indeed depicts Dido, carried by her sister, who, with her clothes, is soaking up the blood pouring out of the mortal wound that Dido has inflicted upon herself. This use, however extreme it may be, is presented as authoritative by Priscian: all that sanguines was lacking was in fact Virgil’s authority, from which cruores, on the contrary, benefited.

This detour through the plural uses of cruor allows us, first of all, to identify a crucial distinction between continuous blood and discrete blood: “discrete blood” is that of distinct entities, the blood spilled by A not being that of B nor of C etc.; they can be distinguished and even counted, and the plural cruores can, figuratively, designate their respective killings. “Continuous blood” is, on the contrary, that of a single individual. It can therefore not be divided, nor, as Pompeius said about dust, can it take a plural form. The detour also allows us to validate the selection made by Latin grammarians of the Virgilian example, which is indeed out of the ordinary, involving a reference to blood spilled by a single person, Dido, blood spilled in such large quantities that it might appear to come from several beings. Let us note that neither Servius nor Pompeius were concerned with identifying a figure here, speaking only, and vaguely, of abuse. The distinction I suggest making between continuous and discrete blood is thus indirectly confirmed by the indifference of grammarians, who were never moved by these cruores as long as they were indeed those of several people. What alarms them, therefore, is not the plural in itself, but the fact that it designates continuous blood and not discrete ‘bloods’. A priori, the analysis seems difficult to extend to sanguis, which is supposed to be specific to a single being, and therefore indivisible and unique, as Pompeius explains; but we shall see if it is the case.

3. The Exegesis of the Uses of sanguis in the Plural

One preliminary clarification is necessary: in the list in which Donatus mentions sanguis, there are many other common terms, classified according to whether they are masculine singular or plural, feminine singular or plural, or neutral singular or plural. A systematic search conducted in the Latin translations of the Bible demonstrates that sanguis is the only singulare tantum to appear in the plural; conversely, three pluralia tantum are used in the singular, namely quadrigae, scalae, and scopae, the Scripture giving four occurrences of quadriga,58 one of scala,59 and one of scopa.60 None of these irregular uses sparks any reaction from the commentators, especially since quadriga is already mentioned in Pliny, Suetonius, Propertius—let us remember that Fronto had urged his listeners to search for examples of quadriga and harenae, and that Aulus Gellius declared having managed to find quadriga in the singular in Varron’s Satires.61 Bede the Venerable († 735) is the only one, in his De orthographia, to discretely point out the singular uses of quadrigae and scalae in the Latin of the Scriptures, designated by in nostrorum litteris scriptorum and nostri,62 but his remarks would not be taken up. With its plural uses, derived from translations, sanguis is therefore an exception.

3.1. Grammarian Exegetes

I will concentrate here on the passages about which commentators make grammatical difficulty explicit, since many occurrences are not even considered from a linguistic perspective, and it is interesting to note that these analyses focus exclusively on passages containing the form sanguinibus. To the Old Testament passages on which the exegesis focuses are in fact the problematic, Ps 15.4 iuxta lxx (non congregabo conuenticula eorum de sanguinibus) and Ps 50.16 (libera me de sanguinibus), must be added Ioh 1.13 (qui non ex sanguinibus neque ex uoluntate carnis neque ex uoluntate uiri sed ex Deo nati sunt), which conveys the same form of ablative plural. They reflect three distinct situations, respectively: a Latin translation from the Greek of the Septuagint that does not correspond to the Hebrew; a Latin translation from the Greek of the Septuagint but ratified by the Iuxta hebr. version of Jerome; a Latin translation of the original Greek. I will consider here three patristic era approaches, those of Augustine, Eucherius, and Cassiodorus.

3.1.1. Augustine of Hippo (354–430)

Following in Donatus’ footsteps, Augustine reminds us several times that Latin does not admit the plural sanguines, even in the nominative63; here we find, implicitly, the question of the plural cases evoked by the Latin grammarians with respect to the singularia tantum (supra 1.2). In De doctrina christiana, Ps 15.4 iuxta lxx64 is invoked to defend the possibility of using the popular form ossum instead of os (Banniard 1995, p. 301), relying on a sanguinibus which is not Latin either:

For if our translators did not shrink from saying ‘non congregabo conuenticula eorum de sanguinibus’ (I shall not assemble their assemblies of blood), because they felt that it was important for the sense to put a word here in the plural which in Latin is only used in the singular; why should a teacher of godliness who is addressing an unlearned audience shrink from using ‘ossum’ instead of ‘os,’ if he fears that the latter might be taken not as the singular of ‘ossa,’ but as the singular of ‘ora,’ seeing that African ears have no clear perception of the shortness or length of vowels?

65

According to Augustine, the interpreters felt it was important (senserunt ad rem pertinere) to use a plural against Latin grammar—but Augustine does not specify why. It is elsewhere, in his De peccatorum meritis et remissione et de baptismo paruulorum, that he would clarify the motivation for the sanguinibus in Ps 15.4 iuxta lxx:

Then he says, I will not gather their assemblies for bloody sacrifices (Ps 16:4), for, when they formerly gathered in a tent or in the temple, the many bloody sacrifices proved them to be sinners rather than purified them. I will not now, he says, gather their assemblies for bloody sacrifices, for the one bloody sacrifice has been offered for many so that they may be truly purified by it.

66

The justification emanates from an author steeped in ancient literature, who spontaneously links this sanguinibus to the plural uses of cruor, which, as has been seen, almost always referred to the discrete blood-letting events of individuals or animals sacrificed in large numbers. These quantities of blood (multis sacrificiorum sanguinibus) served, Augustine tells us, to force the adherance of sinners, rather than to purify them (conuincebantur potius peccatores quam mundabantur). At the end of the passage, Augustine opposes the blood of these victims, who are in fact countless, to the singular blood of Christ. We may compare this to a comment made by Jerome about a difference between his translation of Ezekiel and that of the Septuagint (λαῶν πολλῶν): he points out that he did not follow them in their translation of Ez. 9.9, but wrote sanguinibus (based on the Hebrew), because the blood in question was shed in great quantity (non modicus sanguis effusus est). Again, it is the quantity of discrete bloods that justifies the plural.67

Let us now turn to Augustine’s analysis of Ps. 50.16. Here, the exegete contrasts the Latin language with the form derived from a translation of the Greek, pointing out that “we all know that in Latin it is neither sanguines nor sanguina,” but that since Greek, because of Hebrew, has a plural form, the translator, because of his piety, has favored content over form, and has chosen the ‘less Latin’ over the ‘less appropriate.’ Here again, Augustine comments on the iuxta lxx version. This time he confronts the problem of this plural and its meaning, and his solution is based on that of Ambrose, who posited the equivalence between blood(s) and sins in Ps. 50.16. Ambrose was indeed the first to gloss sanguinibus in Ps. 50.16 as peccatis mortalibus, in reference to the murder of Uriel, but without making any remark on grammar.68 Augustine thus herein combines the linguistic remark with the Ambrosian interpretation and identifies these sins with their source, the sin of the flesh:

The Latin translator has expressed, though by a word not Latin, yet an accuracy from the Greek. For we all know that in Latin, sanguines (bloods) are not spoken of, nor yet sanguina (bloods in the neuter), nevertheless because the Greek translator has thus used the plural number, not without reason, but because he found this in the original language the Hebrew, a godly translator has preferred to use a word not Latin, rather than one not exact. Wherefore then has he said in the plural number, From bloods? In many bloods, as in the origin of the sinful flesh, many sins he would have to be understood. The Apostle having regard to the very sins which come of the corruption of flesh and blood, says, Flesh and blood shall not possess the kingdom of God (1 Cor 15. 50).

69

The commentary on Ioh 1.13 is for Augustine an opportunity to remind us once again that we must not fear the well-known ferules of grammarians,70 for this is the price of truth. This time, the justification of the plural is the fact that humains are born of two bloods, that of their father and that of their mother.71 We find here again the theme of discrete blood, with respect to cruores: it is appropriate to speak of “bloods,” since the “bloods” of two distinct individuals are needed to produce a child.

The sons of men are born of flesh and blood, and of the will of man, and of the embrace of wedlock. But in what manner are these ones born? Not of bloods, as if of male and female. ‘Bloods’ is not Latin; but because it is in plural in Greek, the interpreter preferred so to express it, and to speak bad Latin according to the grammarians, in order to make the matter plain to the understanding of the weak among his hearers. For if he had said ‘blood’ in the singular number, he would not have explained what he desired; for men are born of the bloods of male and female. Let us say so, then, and not fear the ferule of grammarians, so long as we reach the solid and certain truth.

72

In Augustine, the foundation for further reflection is laid: sanguinibus has received two types of justification, the literal meaning when referring to the blood of discrete beings (Ps 15.4 iuxta lxx and Ioh 1.13), and the figurative meaning, inherited from Ambrose, of sins (Ps 50.16).

3.1.2. Eucherius of Lyon (c. 380–c. 449)

A few decades later, Eucherius of Lyon takes inspiration from Augustine73 to point out in his Instructiones, written at Lérins between 430 and 434 for his son Salonius, the problem posed by the sanguinibus form from Ps 15.4 iuxta lxx, and provides him with the same solution. Why, wonders Eucherius, is it written de sanguinibus against the rule of the Latin language (contra regulam Latinitatis) whereas de sanguine would have sounded more Latin (magis Latinum sonasset). The answer is that we read in Greek the plural αἱμάτων which means blood (singular), instead of αἷμα. Like Augustine, Eucherius does not recognize any additional signifying property belonging to this form:

Question. ‘I will not gather,’ he says ‘their assemblies of bloods,’ he says: though ‘of blood’ would have sounded more Latin, for what reason is it written ‘of bloods’ against the Latin rule? Answer. In Greek, we read the plural αἱμάτων which means ‘blood.’ It is the spirutual nature of this particular expression that that translator has expressed in a plural even in Latin.

74

In his Formulae spiritalis intelligentiae, dedicated to his other son Veran, and still following Augustine, Eucherius gives sanguis the meaning of work of the flesh, based on Ps 50.16 and I Cor 15.50, exploiting the textual file assembled by Augustine:

‘Blood’: work of the flesh, in the psalm ‘deliver me from bloods, O God, God of my salvation,’ and similarly ‘the flesh and blood will not possess the kingdom of God.’

75

Augustine’s extralinguistic justification is here turned on its head to give sanguis the originally figurative meaning of work of flesh, for lexicographical purposes.

3.1.3. Cassiodorus (c. 485–c. 585)

The Augustinian analyses are taken up by Cassiodorus in his Expositio Psalmorum, completed at the Vivarium, where the grammatical and rhetorical influence is omnipresent (Grondeux 2013). Cassidorus focuses on the same two passages, Ps 15.4 iuxta lxx and Ps 50.16, and I give here his analysis of the former:

For ‘of bloods’ designates the blood of animals, which was quite abundantly shed during sacrifices; this rite was modified upon the coming of Christ. Nonetheless, it is clear that the noun ‘of bloods’ is used in spite of grammar, according to which there is no plural form of this word; it is for this reason that it must be counted among the idioms, that is, the particularities of the Holy Scriptures.

76

Cassiodorus emphasizes that the plural uses of sanguis go against Latin grammar deliberately. This position is interesting, because he is supposed to have had access to Priscian, whom we have seen to be more conciliatory than his predecessors in the matter, and the reading of whom could precisely have led Cassiodorus to take an innovative approach to the plural form of sanguines. However, as Louis Holtz (1981, p. 245) has shown, Cassiodorus does not actually know Priscian, whom he initially took for a Greek grammarian. Like Augustine, Cassiorodus justifies the plural of Ps 15.4 iuxta lxx by the enormous quantity of animals sacrificed (tunc copiosus in sacrificiis fundebatur), a rite modified by the coming of Christ. Cassiodorus also introduces here the notion of idioma sanctae scripturae: having noted that sanguinibus is indeed used in spite of the grammar (contra artem positum constat esse grammaticam), according to which sanguis has no plural (apud quam pluralis huius uerbi numerus non habetur), he concludes that it should be counted among the scriptural idioms (it could possibily be objected to him that the underlying Hebrew מִדָּ֑ם is a singular, and that the plural goes back only to the Septuagint form αἱμάτων), without justifying it further.

Commenting on Ps 50.1677, Cassiodorus notes similarly the inadequacy of sanguinibus with respect to Latin grammar. The translator who reproduced what he found in Greek (which coincides this time with the Hebrew, but Cassiodorus does not mention it) should nonetheless be congratulated (omnino laudandus est): the interpreter has indeed chosen (elegit) to “write something up against grammar” (aliquid contra artem saecularium ponere) rather than to deviate from the truth, the peculiarity being again due to a scriptural idiom:

The plural ‘bloods’ is obviously employed against the Latin language, but because it is found in the Greek copies, the translator should absolutely be congratulated; he has indeed chosen to inscribe something against this secular art rather than enter into disagreement with an expressed truth. For in saying ‘of blood,’ he might appear to designate a single sin; but in using the plural, he confesses without a doubt that they were numerous; this is what we can call the idiom of the Holy Scripture.

78

Cassiodorus identifies a rhetorical device, before exposing the equivalence between blood and sin linked to the flesh (carnalibus delictis):

This figure is known as an exallage, that is to say a permutation, each time that a genus or a case is exchanged in spite of use. The prophet therefore asks to be delivered from the carnal offenses in order to finally stop sinning in his fragility. ‘Blood’ is indeed written for the human body, because it seems to be superior to the other spirits. For even in the Gospel, it is said to Peter ‘this was not revealed to you by flesh and blood’… And it is also added ‘and my mouth will exalt your righteousness,’ in other words, ‘if you deliver me from the bloods (read ‘the sins’), my mouth will rightly sing your praise.

79

Cassiodorus’ reasoning involves a certain amount of complexity: (1) ‘blood’ is in the plural in Hebrew because it means sin here, and the number of sins imposes a plural; (2) the translator’s piety makes him prefer fidelity to the literal Hebrew and/or Greek than to Latin grammar, since the Septugints were inspired; (3) the Latin text produced being grammatically faulty, the deviation from the norm invites the seeking of a supplementary justification by way of meaning. The problem being, of course, that there is no ‘mistake’ neither in Hebrew nor in Greek. Cassiodorus sets up a two-tiered justification, the first being fidelity to the literal, the second being the excess of meaning that the believer is invited to detect. We can also see in it a warning to copyists, who are invited to respect scriptural particularities without, in a philologist’s reflex, correcting the text.

By the end of the patristic era, we can see that neither Augustine, nor Eucherius, nor Cassiodorus use Ps 15.4 iuxta Hebr. (non litabo libamina eorum de sanguine), and therefore that the Hieronymian version, which restores a singular consistent with the Hebrew, remains unexploited, and is in fact never cited. The Augustinian exegesis will therefore carry its full weight here.

3.2. The Controversy between Agobard and Fredegisus

Around 830, these questions became the cause of a violent controversy between two Carolingian intellectuals, Agobard, archbishop of Lyon (c. 779–840), successor of Leidrad (and distant successor of Eucherius80), and Fredegisus (end of the 8th c.-c. 834), abbot of Saint-Martin of Tours, successor of Alcuin.81 As this quarrel is known to us only through Agobard’s reply to a (lost) text by Fredegisus, who was himself criticizing a text (also lost) by Agobard, it is certain that we do not know all the ins and outs of it. Nonetheless, the letter is detailed enough for us to at least partially grasp the two arguments at hand. In the first episode, it seems Agobard has declared, in the course of a sentence, in reference to his own weakness in grammar, that he was relieved that the Scriptures and their translators did not always scrupulously follow gramatical rules; upon reading these few words, Fredegisus, it seems, became enflamed and violently inveighed against Agobard in a letter. Some passages are known to us, as they are literally quoted by Agobard in the third writing, the only one remaining; I will reproduce them below in quotation marks in order to distinguish them from the rest of Agobard’s text.

In a less literal way, Agobard’s letter also teaches us about the strategy put in place by Fredegisus, who brandishes against his adversary the plural of “blood,” the emblematic example around which the debate centers. Abogard’s response is that Fredegisus has gone to a lot of trouble for nothing, since he himself has never criticized any biblical author or translator, nor has he ever heard anyone criticize them,82 and that he is perfectly aware of the plural use of “blood” in the Scriptures.83 His critique is based, first of all, on Fredegisus’ arguments: the Latin plural is not due to the fact that the “translator found it in the plural in Greek,” but to the fact that it was the only way to render the idea of the Evangelist. Agobard limits himself here to the Greek of Ioh 1.13, but then launches into a series of more of less biased linguistic remarks, borrowed from Jerome, and aimed at producing a demonstration by way of the absurd: the fact that ר֨וּחַ (spirit) is feminine in Hebrew did not lead the translator to feminize spiritus in Latin.84 Second of all, his criticism centers around the supposed amalgam committed by Fredegisus between all kinds of translators and exegetes:

“It is thus all together that the apostles and the evangelists and the interpreters of the Devine Scripture, with their Christian exegetes, can be rationally defended against ignorant criticism. Indeed, they said, against the grammatical rule, only things required by a reason or because of a mystery.”

At this point, Agobard inflicts on his opponent a long historical clarification on the different authors and translators of the Scriptures, underlining that Jerome himself denounces some of them in his prefaces: it is not because a translation exists that it is automatically sheltered from all criticism. This is followed by an avalanche of patristic and scriptural quotations, which conclude with a silent borrowing from the City of God:

It is good to be reminded those words of the Savior by which he says “If you offer with righteousness, but do not part with righteousness, have you not sinned?” (Gn 4, 7 apud Aug. civ. 15, 7, 1). He who is indeed righteous in the Catholic faith makes a righteous offering, but if he attributes to all the doctors or interpreters a veneration equal to that of the apostles and the evangelists, he offers without righteousness, and for this reason he sins.

85

This borrowing is malicious on more than one level: the analysis coming from an irrefutable Father is applied brilliantly to Fredegisus and manages to ruin his argument, by attesting in passing Abogard’s familiarity with a high-flying exegesis of a notoriously complicated passage (Alexandre 1988, p. 353):

These words ‘nonne si recte offeras, recte autem non diuidas, peccasti?’, because we see neither the reason, nor the circumstances following which they were pronounced, engender in their obscurity many meanings when the commentators of the Holy Scriptures attempt to explain them according to the rules of faith. A sacrifice is just when it is offered to the true God to whom it is due. But offering is not just when we have poor judgement of the places, the times, the things we offer, of the one who offers, of the one who receives, of those to whom we distribue food—offering thus means judging—that we offer when we should not, or what we should not offer here but elsewhere; that we offer when we should not or we offer what we should not offer now but at another time; that we offer what we should never offer anywhere or anytime; that among the offerings, he keeps for himself the best of what he presents to God; that the oblation be distributed to a profane or to someone who does not have the right to it.

86

Finally, Agobard faces one last accusation, quoted literally, that of having suggested that the Holy Spirit may have made authors and translators express themselves in a vulgar way:

“It is indeed shameful to believe that the Holy Spirit, who spread the languages of all the nations into the spirits of the apostles, would have spoken through them in any language in a rustic, rather than a noble, way.”

87

Focusing on the stupidity of the statement, Agobard’s response is once again scathing:

With what intention did you say this, none of those who fear speaking and saying nothing will understand it, except in lending me the additional crime of accusing the Holy Spirit of rusticity. If such is the case, in other words, if this is your intention, what can I answer except “that he who knows hidden things, the witness of hearts, he who sounds our kidneys, the all-powerful god, sees and hears” (cf. Dan 13.42). Beyond the fact that you seem to attribute to me such a sacreligious thing, your words also reveal your opinion about the prophets and the apostles, namely that the Holy Spirit did not only inspire their sense of predication, and the modes or literal arguments, he also formed, in their mouths, the words that came out of their bodies. If this is indeed what you think, the absurdity of it is difficult to evaluate.

88

Agobard thus affirms that the Latin biblical plurals are not mere calques, but that they are always motivated by a latent meaning that corresponds to a spiritual message: respecting them and giving an account of them are part of the tasks of the translator and the exegete. Here, there is a clear echo with the Cassiodorian double justification, based on literality and additional meaning.

3.3. The Justification by Way of the Idioma Sanctae Scripturae or by Way of Meaning

Medieval exegesis hardly renews the approach to the three passages whose commentaries by Augustine, Eucherius and Cassiodorus are presented above. The commentators take up, sometimes literally,89 the Augustinian themes. I will limit myself to a few examples: Haimo of Auxerre in the 9th century, Berno of Reichenau in the 11th, and the Glossa ordinaria and the Sentences by Peter Lombard in the 12th century.

3.3.1. Haimo of Auxerre († c. 865)

Commenting on Ioh 1.13, Haimo of Auxerre points out the grammatical irregularity committed by the translator and justifies it in two ways. Firstly, the plural indicates that a child is born of the two bloods of its parents; secondly, the plural is identical to that of Ps 50.16, and means ‘sins.’ Here we find the bridge that Augustine built between these different passages that are supposed to illuminate each other:

By writing here ‘bloods,’ he has done so against the grammar rule, as for the grammarians ‘blood’ is written only in the singular and never in the plural in accordance with the rule. However, the evangelist was not concerned with observing the grammatical rule, in order to better respect the meaning. Indeed, in writing ‘from blood,’ he seems to mean only that of the man, without expressing fully what he wants; while when writing ‘of bloods,’ he highlights the union of a man and a woman. For this reason, so that no one wonders why he said ‘of bloods,’ he adds the explanation ‘neither of the will of a man, nor of the will of the flesh.’ He writes ‘flesh’ for the wife… Or ‘from bloods,’ that is of vices and sins, as it is said by the Psalmist ‘deliver me from bloods, O God, God my savior.’

90

3.3.2. Berno of Reichenau (c. 978–1048)

In analyzing Ps 54.6 iuxta lxx91, which contains the singular tenebra where one would expect the plural form tenebrae, the Abbott Berno of Reichenau appeals to the opposite phenomenon, illustrated by the plural sanguinibus. We therefore find the usual ingredients: the reader must not be troubled (nec conturbet) by the fact that the Scripture transgresses a grammatical rule (regulam grammaticae artis), which is an occasion to invoke Gregory the Great’s hackneyed expression according to which Scripture is not governed by the rules of Donatus,92 before citing Ps 50.16 and Ioh 1.13 to support him, and to underscore the importance of sticking to the Hebraic truth—of which this is the first mention in this context, without however providing any justification by way of meaning. It is quite particular to find the demonstration of the singular tenebra by way of the iuxta Hebr. version which gives the singular caligo, since the Hebrew פַּלָּצֽוּת would have been better translated as horror for instance.

Do not be troubled by the fact that the translators have written, in spite of the grammatical rule, the singular ‘darkness’ for the plural, since, as Saint Gregory says, the words of the celestial oracle must not bend before the rules of Donatus; just as the prophet wrote without fearing the ferule of the grammarians ‘deliver me from bloods’ and in the Gospel of John ‘which were born not of bloods, he says, nor of the will of the flesh nor of the will of a man but of God’; just as we wrote there the plural instead of the singular, similarly we wrote here the singular instead of the plural, in accordance with the hebraic truth, which says ‘fear and trembling hath came upon me, and darkness covered me.’

93

Augustinian filiation is clear here, without any kind of renewal of the themes, or recourse to Priscian, or even recourse to the Hieronymian translation. It is as if the vocabulary, the oppositions, the commonplaces were frozen.

3.3.3. The Glossa ordinaria and the Sentences

The phenomenon is again verified in the commentaries of Anselm of Laon († 1117), starting with the Glossa ordinaria (see Smith 2009; Giraud 2010). The commentary of Ps 50.16 explicitly cites Cassiodorus and Augustine, with a few adjustments, simplifications and rewritings94; the commentary on Ioh 1.13 simply mentions in an interlineary gloss the fact that sanguinibus plural refers to the contributions of the two parents,95 which we find, identically, in the commentary on the Gospel of John by the same author.96 The Glossa directly inspired the Sentences by Peter Lombard (c. 1096–1160), and it is interesting to observe, in the table below, the game of crossed rewritings. The Augustinian reformulation of libera me de sanguinibus in libera me ab iniquitatibus, munda me ab omni corruptione, which figured at the end of the commentary, had been augmented and reused as an interlinear gloss (1) [the numbering follows the order of the Glossa, which serves here as a kind of pivot], but is found yet again at the end of the sequence in Peter Lombard, but in the extended version of the Glossa. The Cassiodorian argument (2) also ends up extended in a final ut iam ab his desinam, which is found as is, in Peter Lombard. Severely rewritten, the Augustinian sequence (3) is also taken up in the Glossa: however, Peter Lombard apparently also turned to the original as the introductory phrase omnes nouimus, abandoned by the Glossa, reveals. Recourse to I Cor 15.50, inherited from Augustine, is treated identically by the Glossa and the Sentences (4).

Augustine		Glossa ordinaria	Peter Lombard
	(1)	(interl.) ab omni corruptione pene et culpe et da incorruptionem
	(2)	Cassiodrvs. Quia multa sunt carnalia peccata, ut iam ab his desinam.
Expressit latinus interpres uerbo minus latino proprietatem tamen ex graeco. Nam omnes nouimus latine non dici sanguines, nec sanguina; tamen quia ita graecus posuit plurali numero, non sine causa, nisi quia hoc inuenit in prima lingua hebraea, maluit pius interpres minus latine aliquid dicere, quam minus proprie. Quare ergo pluraliter dixit ‘de sanguinibus’? in multis sanguinibus, tamquam in origine carnis peccati, multa peccata intellegi uoluit.	(3)	Avgvstinvs. De sanguinibus. Et si minus latine magis tamen proprie pluraliter dixit innuens multa peccata ex origine carnis peccati trahi.	Omnes nouimus Latine non dici sanguines, sed sanguinem, sed etsi minus Latine dicatur pluraliter quam singulariter, tamen magis proprie pluraliter dixit, innuens multa peccata ex origine carnis peccati trahi, id est ex carne per originale peccatum corrupta.
Ad ipsa peccata respiciens apostolus, quae ueniunt de corruptione carnis et sanguinis, ait ‘caro et sanguis regnum dei non possidebunt’ (I Cor 15.50).	(4)	Vnde ‘caro et sanguis regnum Dei non possidebunt’ (I Cor 15.50), id est peccata que de corruptione carnis et sanguinis ueniunt.	Vnde ‘caro et sanguis regnum Dei non possidebunt’ (I Cor 15.50), id est peccata quae de carnis et sanguinis corruptione ueniunt.
(…)			(2) Et hoc est, libera me de sanguinibus, o Deus, id est de carnalibus peccatis quae multa sunt, ut iam ab his desinam.
(1) libera me, inquit, de sanguinibus; hoc est: libera me ab iniquitatibus, munda me ab omni corruptione.			(1) Vel ita ‘libera me de sanguinibus’ id est ab omni corruptione poenae et culpae, et da incorruptionem, quod implebitur in futuro.

Exegesis has thus trivialized the idea of a plurality of meanings, sometimes singular, sometimes plural, and it is not unusual to see commentators make a list of the of meanings of sanguis, as Honorius Augustodunensis (c. 1080–c. 1140), for example, does when he recalls, before commenting on Ps 50.16, that ‘blood’ in the singular designates a part of the body, or man, or lineage, but in the plural refers to sin or death.97 The polyvocity of this plural is inexhaustible, as Denis the Carthusian (1402–1471) testifies when he lays out two ways of glossing the sanguinibus of Ez 9.9 (which is actually only very rarely commented on), either by ‘sins and works of flesh’ (peccatis et operibus carnis), or by ‘bloods shed by murdered innocents’ (innocentium in ea occisorum cruoribus)98; and it is rather fascinating to see sanguinibus glossed here by cruoribus…

4. Consequences

4.1. Grammairian Exegetes

From the end of the 6th century on, grammar books evolved very quickly, including in the tradition of Donatus. The anonymous author of De dubiis nominibus, in a corrupted passage heavily restored by Glorie, introduces a biblical passage (Ecl 2.8) that presents a plural use of uina.99 In the 7th century, the Anonymus ad Cuimnanum would produce a very interesting analysis, that takes the aforementioned passage from Pompeius as a starting point. After recalling the point of view of the late antique grammarian, and faithfully going over his reasoning which treats puluis as a singular, the author points out that the reader should not be troubled if he comes across the plurals sanguines and pulueres. Two kinds of reasons indeed motivate the existence of these forms, either the authority of the ancients, or the diversity of languages. The authority of the ancients is illustrated by an example from Horace,100 and it is no longer a matter, here, of abuse: in a unique shift in perspective, the authority of the ancients alone guarantees the validity of the expression. The diversity of languages, a novelty in this context, is illustrated by way of Ps 50.16, with a typically Cassiodorian justification: the interpreter, who could not translate any other way and preferred a useful translation to a translation that might follow the logic of Latin, did not hesitate to translate a Greek plural by a Latin plural, against the logic (ratio) of the language:

For what reason do ‘dust’ and ‘blood’ not have plurals? For this reason: it is because they are individisible things, and as they are indivisible, they cannot have plurals; because they are of a unique nature. But a reader must not be troubled if he finds ‘bloods’ or ‘dusts’; indeed, they can be found written both ways, either by the authority of the ancients or due to the difference of langauges. For example, as in Horace’s ‘nundinales diis parem pulueres’ (nouendiales dissipare pulueres); but we know it is due to an authority. And furthermore, ‘deliver me from bloods, O God’; and the Latin interpreter, finding in Greek ‘nazon’ (αἱμάτων), which is equivalent to ‘bloods,’ unable to translate any other way, and prefering the utility of the translation to the logic of Latin, did not hesitate to translate ‘bloods’ in spite of logic.

101

If Pompeius is indeed at the root of this development, the pattern is particularly inflected. An individisible thing does not, indeed, have a plural form, but it can be found for two reasons: authority (precisely that which Pompeius denounced as abusive), and the diversity of languages, a radically new argument in this context. The author salutes, in passing, the boldness of the translator, with a distant echo of the congratulations offered by Cassiodorus to the same interpreter. Let us note, furthermore, that grammatical transgression is now the norm, illustrated by a very nice use of puluis in the sense, not of ‘dust,’ but of ‘grain of dust’ in Isidore of Seville, who compares atoms to grains of dust (pulueres) that float in the sunlight:

Isid. etym. 13, 2, 1: hi per inane totius mundi inrequietis motibus uolitare et huc atque illuc ferri dicuntur, sicut tenuissimi pulueres qui infusi per fenestras radiis solis uidentur.

As the editor notes (Gasparotto 2004, 13 n. 20), the almost literal source here is Lact. ira 10, 9, except that Lactantius (c. 250–c. 325) was not speaking of pulueres but of pulueris minutias (sicut pulueris minutias uidemus in sole, cum per fenestram radios ac lumen immiserit), in a way that conformed more with grammatical recommendations.

Pompeius’ reasoning, according to which an indivisible is necessary singular, is regularly taken up by the 9th century commentators of Donatus, and sometimes without objection as in Ars Ambrosiana. The Ars of Lorsch, however, opposes it immediately with the example of Ps 50.16, that it completes, for the first time, with Ioh 1.13. The author removes the paradox that he has just created, by explaining that the translator preferred to break the grammarians’ rule (frangere regulam grammaticorum) rather than silence the idiom of the Greek translation,102 before reproducing Pompeius’ entire reasoning.103 The approach is thus different than that of Anonymus ad Cuimnanum, since the problem raised by the biblical counter-exammples is treated and evacuated from the outset, in order to better consolidate Donatus’/Pompeius’ argument. We can imagine that grammarians were used to including these biblical excerpts in their reasoning, whether it be to contradict Donatus/Pompeius or to reaffirm its content. Indeed, Muretach and Sedulius Scottus104 use the same presentation in the 9th century, and Aelfric, in the 10th century, also includes a similar development in his Exceptiones, objecting, with respect to sanguis, to the fact that the Scriptures contain the expression uirum sanguinum.105 This commonplace can still be read in the 12th century in the Summa super Priscianum by Petrus Helias, which imagines different types of solutions to the problem raised by the biblical occurrences of sanguinum and sanguinibus. The first builds on the famous Gregorian maxim according to which the language of the Scripture is not subject to grammatical rules; the second proposes to attach the forms sanguinum and sanguinibus not to sanguis but to sanguen, thus saving what had been stated by Donatus; the third possibility, which is innovative and subtle, but underdeveloped, consists in saying that one must be wary of the notion of use. The plurals of ‘blood’ may well be found in the Bible, but they remain foreign to what is in common use by all; in a Christian world, Petrus Helias thus reinstates, in the course of a sentence, the idea that Latin resulting from translations remains forever outside the corpus of standard Latin:

He (Priscian) therefore says that certain nouns never take the plural form, either because their “nature,” that is their meaning, repels it, because they mean something that cannot admit several meanings like Iupiter, or because usage forbids it, just as sanguis has no plural because only usage repels it. If someone were to object that this noun has a plural in usage, for the same reason that we find uiri sanguinum and qui non ex sanguinibus etc., either we would say that the Scripture is not subject to the rules of this art, or we would say that it does not come from sanguis but from sanguen. Or this can be resolved in another way. Indeed, although we find sanguinibus here and there, that does not mean that it is in use. In fact, it is said that something is in use when it is used by all.

106

There is a new turn with Guillelmus Brito (13th c.), who after having cited Isidore of Seville introduces Priscian’s opinion, before specifying that the Scripture, which is not subject to “Priscian’s rules” (sic) uses sanguis in the plural in the figurative sens of “sin.”107

More generally, everyone seizes upon these biblical examples to justify a number of phenomena that are ultimately quite disparate. The Ps 50.16 is thus called upon to justify the plural Capreae (insulae), the ancient name of Capri, in a 10th century hagiographic account of dedicated to Constantine of Capri!108

4.2. The Bloods of the Virgin

Coming from Eastern theology, the question of the “bloods of the Virgin,” from which the virginal conception of Christ is operated, appears in several periods in Western history: it is again a form modelled on a Greek plural. The first occurrence is attributed to Bede the Venerable, who seems to have known Pseudo Dionysius109 directly, perhaps via the school of Theodore and Hadrian110:

For Dionysius, remarkable among the authors of the Church, expresses himself in this way in his booklets on the divine names: indeed we do not know how he was formed from the bloods of the Virgin, according to a law other than that of nature, and how, without having feet with bodily weight and a material mass instilled upon him, he moved about in a wet and unstable substance.

111

This theme is found in the translation of the Pseudo-Dionysius by Hilduin, Abbot of Saint-Denis (c. 785–c. 855),112 in very similar terms, as well as in all those that would follow his, those of John Scotus Eriugena (c. 800–c. 877),113 John the Saracen (12th c.),114 Robert Grosseteste (c. 1168–1253),115 Marsilio Ficino (1433–1499),116 Ambrose Traversari (1386–1439).117 The other occurrences, on the other hand, all stem from translations of the synthesis done by John of Damascus (675–749),118 attributable to Cerbanus and Burgundio of Pisa (12th c.), and recognizable by the characteristic expression purissimi sanguines of the Virgin,119 in particular by way of the dissemination that Peter Lombard’s Sententiae120 would procure them. Peter Lombard here is an essential link for the transmission of the uses of John of Damascus’ ideas, leading to a searing but poorly inspired response from Walter of Saint-Victor († c. 1180),121 and above all to all the great summae of the 12th (Zachary of Besançon) and 13th centuries (Albert the Great,122 Alexander of Hales,123 Bonaventure,124 William of Auxerre,125 Henry of Ghent,126 John Duns Scotus,127 Thomas Aquinas128), all the way up to Lawrence of Brindisi in the 17th century, all of whom include a discussion on this point inherited from Eastern theology.

The most interesting point, however, is what is happening in Reichenau, well before the 13th century. It is indeed there that are collected, in the ms Aug. 80, Latin translations of the 9th century, by the hand, perhaps, of the bishop John of Arezzo († 900), of Greek homilies on the Virgin that take up this theme, one by John of Damascus129, the other by his contemporary Andrew of Crete130 (660–740), which convey the same expression sanguines Virginis.

4.3. Autonomous Uses

Outside the field of exegesis strictly speaking, many authors use plural forms of sanguis, which we will call “autonomous” here. Let us note that as early as the 4th century, we find in Vegetius’ Mulomedicina (late 4th c.-early 5th c.) a plural use of sanguis, in which the word designates, like cruores seen above, ocular petechiae: sanguines therefore means here bloodstains that the eye can discriminate individually, and therefore could, in theory, count, which means that sanguis escapes, here, by way of an identical catachresis, the meaning “continuous liquid belonging to a single individual.”131

Various expressions emerge very early on, such as uiri sanguinum, used to designate, outside of any exegetical context, disreputable people, adversaries,132 etc., in a phenomenon related to the “patois de Canaan” (Cabanel 1995). Along the same lines, Aistof, king of the Lombards (749–756), defeated by Pepin the Short at the request of Pope Stephen II, was described by the latter as a ‘devourer of Christian bloods.’133 The expression effusio sanguinum, shedding of blood, appears very early, and is also the most frequent. At an early date, it is particularly represented in translations from Greek, whether it be the revised version by Bede of the hagiographic dossier relating to the martyr Anastasius,134 or a letter to Pope Zacharias, of Greek origin, to Boniface.135 The expressions sanguinum fluores and perfusiones/profusiones sanguinum that we find in the translations of Anastasius the Librarian (c. 810–c. 878)136 can be compared. Although the expected expression effusio cruoris is more frequent, effusio sanguinum is highly present in historical and hagiographic accounts,137 in contexts that closely evoke the aforementioned plural uses of cruores; let it be noted in particular that, as in the case of cruores, it refers to distinct individuals and in general many individuals. These “effusions of blood” would then lead to a variant of uiri sanguinum, that would become, by the hand of Bruno of Cologne (c. 1030–1101), uiri effusores sanguinum, “shedders of blood.”138 We have chosen to set aside the sputamina sanguinum mentioned in the Passion of Saint Leger of Autun (7th c.), because here we are dealing with <drops of> blood, 139 and we have seen an identical use of cruores (note 58).

Though it is the most represented, the plural genitive is not isolated, and Rupert of Deutz (c. 1075/1080–c. 1129) speaks spontaneously of regnare in sanguinibus,140 while a bishop from Reims uses, during the same period, the accusative sanguines to designate murders.141 Finally, sporadically, sanguis in the plural designates the blood of menstruation, a usage attested from Tertullien (155–220) to Coluccio Salutati (1331–1406),142 just as doctors speak indifferently of urine or of urines.

Let us note, in conclusion, that this evolution is thoroughly described by Ménage in his Observations (Ménage 1675, p. 296), going from the ancient language, to biblical Latin, to medieval Latin:

Sang. On dit tousjours le sang, & jamais les sangs. Les Latins ont dit demesme sanguis au singulier seulement. J’entens les Latins des prémiers siécles: car ceux des derniers ont dit sanguines. Arbor sanguinum, dans Luitprandus. Libera me de sanguinibus dans le Séaume 50. Sur lequel endroit S. Augustin a fait cette note: Expressit Latinus Interpres verbo minus Latino: proprietatem tamen ex Greco: nam omnes novimus Latinè non dici sanguines, nec sanguina sed sanguinem. Tamen quia ita Grecus posuit plurali numero non sine causa nisi quia hoc invenit in prima Lingua Hebræa, maluit pius Interpres minùs Latinè aliquid dicere quàm minùs propriè.

The example of Liutprand of Cremona (c. 922–972) seems a priori poorly supported by the modern editions of the Lombard laws143 to which Ménage is referring here in 1675, but it is relayed by Keyssler (1720) and Eckhart (1729), the latter borrowing it from the former.144 If we, however, follow the lexicographer John of Genoa (13th c.),145 it is more like a shrub, so named for its red color (the blood dogwood, Cornus sanguinea), which Bayet (1935, pp. 39–41) also confirms. However, sanguinum turns out to be the accusative of a noun, sanguinus according to John of Genoa, and no longer the plural genitive of sanguis, which invalidates the romanticized deductions of Keyssler and Eckhart, repeated in learned treatises on Germanic mythology (Tkány 1827, p. 44).

5. Conclusions

Grammar and faith shared an entangled history in the context of Western Christianity146. This article has shown how roles could be merging, shedding light upon exegetes acting as grammarians and grammarians acting as exegetes. Exegetes aim at exempting the Holy Scripture from any reproach of grammatical error, while at the same time no grammarian will use these passages to criticize the language of the translators of the Bible. Both agree indeed in seeing them as clues proper to reveal hidden meanings of a text located out of the Latin canon.

What role does grammar ultimately play in the analysis of these plurals? Augustine is the one to draw attention to these plural forms of sanguis in the Scriptures. It was obviously to Donatus that he was referring to implicitly, and his analysis would be taken up and repeated, without Donatus’ name ever being mentioned, except when the Gregorian adage according to which Scripture is not subject to the rules of Donatus is recalled. While no exegete or commentator cites him, Priscian, the reading of which could have encouraged a more tolerant approach to these deviant forms, is never cited either. It is in fact the opposite that occurs when Petrus Helias evokes, in his commentary on Priscian, the biblical plurals, but rejects the Scriptures outside of Latin canon, ultimately confirming Donatus’ postulate. In any case, the two grammatical traditions have never explicitly confronted each other on this point. One should also note the lack of any reference to the question of continuity and discretion: Pompeius had nonetheless posed the problem very clearly, in terms largely borrowed by Carolingian commentaries on Donatus, the same ones that integrate the biblical aspect. Grammar has therefore played an important role, but a static one since Augustine, in an unchanging topos, reinforced by the Gregorian adage: a translator must not fear overriding Latin grammar, nor should he fear being scolded by grammarians, and the reader should not be troubled by this infraction. Exegetes and commentators have long repeated over and over that sanguis exists only in the singular, but none of them have gone further to gain an understanding of the reason for that prohibition. No one even mentions the Isidorean distinction, which is almost more awkward for the correction of these occurrences, but comes much later than the translations of the Bible.

Rule, use, authority: the plural forms of sanguis question the logic of the language, which wants grammatical number and extralinguistic reality to correspond. However, at what level does this transgression occur, and what is its degree of gravity? While the late-antique grammarians did not take sides, Eucherius is the first to have a clear opinion and to say that the plural use of sanguis constitutes an infraction of a grammar rule (contra regulam latinitatis). As for the grammarians, the first to discuss the subject was Priscian, for whom the exclusively singular use of sanguis reflected use and was in no way any kind of rule. The majority opinion would fall with Eucherius: the Ars of Lorsch, Agobard, Haimo of Auxerre, Berno of Reichenau, unanimously consider that sanguis in the plural violates a grammar rule, which, in the Cassiodorian view consisting in awarding congratulations to scrupulous translators, is obviously more valorizing than a mere infringement of a usage. According to Priscian, this is indeed the level on which the use of sanguis in the plural plays out—but Priscian denies that there are examples—and we have seen that cruores, which is just as abusive, gives legitimacy to sanguines, just as cineres allows for pulueres. The lack of recourse to Priscian can be explained by the comfort found in the repetition of patristic exegesis, but perhaps also because there needs to be transgression in order to trigger the “hermeneutical leap” (Dahan 2009, p. 197 sq.).

This statement again raises the eternal question of authority, of the authorities and of the possibility of imitating them. Let us remember that Fronto incited his students to seek out plural forms of arena, but not in just any author, only in the authors who possessed authority. For Pompeius, however, although an author can use pulueres, puluis remains singular: authority is therefore no match for the weight of the logic of the language, the ratio. However, we it is in fact authority that, in function of how we consider it, opens the path to a second type of justification, introduced by the Anonymus ad Cuimnanum, that of the difference of languages, in connection with Cassiodorus and his consideration of scriptural idioms. Biblical Latin thus interferes with grammatical reasoning, and becomes, like the poets, a potential factor of transgression. However, does the Bible have any authority over Latin? We have seen that the range of answers has been wide, extending to the rejection of the corpus and its exclusion from usage, by Petrus Helias, on the grounds that it is not common to all speakers of Latin.

Hearing the Bible nonetheless accustomed Latin ears to the plural of sanguis, in exactly the way that Biblical Latin made popular the use of cutis instead of pellis (Grondeux 2005). In addition, later translations from Greek (lives of saints, theologies of Pseudo Dionysius and of John of Damascus) have caused these plural uses to multiply, and have fully convinced authors that these uses were, in the end, licit. Following on Bede’s observations, it would be interesting to investigate whether early Romance vernacular could have been more permeable to the use of sanguis in the plural. Although a quick survey carried out in Godefroy’s dictionary and in the corpus http://zeus.atilf.fr/dmf/ (accessed on 5 September 2022) does not suggest that the investigation would be conclusive, it is perhaps worth carrying out, but in another framework than that of this study, which is limited to Latin.