Manuscripts

The grammar of Nikolaos Sofianos has come down to us in two manuscripts: Par. gr. 2592, an autograph,Footnote ¹ and Vat. Ottob. gr. 173, fol. 1^r-30^r, copied by Ioannis Mavromatis in Rome between 1548 and 1553.Footnote ² The latter manuscript has corrections, in both the margins and the main text, on fol. 5^v, 7^v, 24^r, 28^r, 29^v and 30^r: these corrections are in the hand of Sofianos.Footnote ³

The Paris manuscript contains only Sofianos’ grammar: after a folio left blank, it shows the coat of arms of Cardinal Jean de Lorraine on fol. 1^r,Footnote ⁴ followed by a dedicatory letter addressed to this prelate on fol. 1^r-2^v, the grammar on fol. 3^r-37^r, and an address to the readers on fol. 37^r-39^r.Footnote ⁵ Vat. Ottob. gr. 173 is a composite manuscript that consists of various codicological units, dating from different periods. The first of these is in the hand of Mavromatis: after a folio left blank, it offers on fol. 1^r-30^r Sofianos’ grammar without the coat of arms and the dedicatory letter and without the address to the readers at the end. Like Sofianos’ autograph, it originally formed a manuscript on its own. It has a quire numbering in the lower margin and notes to the binder indicating where quires begin and end: (1) fol. 1–7 plus the blank folio at the beginning; (2) fol. 8–15; (3) fol. 16–23; and (4) fol. 24–30 plus a folio now missing from the manuscript.Footnote ⁶

Émile Legrand published the text on the basis of the Paris manuscript, first in 1870 and then again in 1874, in what was supposed to be a revised edition.Footnote ⁷ Though the second edition does offer quite a number of corrections, it also has some mistakes that cannot be found in the first: not only typographical mistakes, but also annoying omissions, e.g. ‘τὰ πɛρισπώμɛνα ῥήματα ἔρχοντ᾽ ἀπὸ τὴν <τρίτην> συζυγίαν τῶν βαρυτόνων’ (leaving it unclear from which conjugation exactly) or, perhaps even worse, ‘κατομωτικά: <ναὶ μά, ɛἰς τόν, στόν / καταφατικά:> ναί, ἔτζη, ναίσκɛ, ɛἴτις, οὕτως’ (conflating two adverbial categories and thus presenting the examples of καταφατικά under the heading ‘κατομωτικά’).Footnote ⁸ The 1977 reprint by Papadopoulos is based on Legrand's second edition and, to add insult to injury, has additional mistakes of its own: e.g. the 2nd person plural of the imperfect ἐγράφατɛ instead of ἐγράφɛτɛ (the ending in -ατɛ is extremely rare in the early modern period).Footnote ⁹ It also makes a mess of Sofianos’ Latin and Legrand's French.Footnote ¹⁰ There is an unpublished diplomatic edition of the Paris manuscript by Marc Vernant, which I have not seen.Footnote ¹¹ The good news for those without access to Vernant's edition is that the manuscript is accessible online – and so too is the other manuscript, the one in Rome.Footnote ¹²

For reasons beyond his control,Footnote ¹³ Legrand was unable to compare the Vatican manuscript (V) with the one in Paris (P), which is a pity because there are significant differences between the two. Some of these differences are due to obvious scribal errors. Take the first person plural of the imperfect of oxytone verbs in -έω and in -άω: ἐκρατοῦμαν καὶ ἐκρατούσαμɛν V, ἐκρατοῦμαν P (P omits information); Footnote ¹⁴ ἐγɛλοῦμαν V, ἐγɛλοῦμɛν PFootnote ¹⁵ (Sofianos does not use the ending -μɛν for the oxytone verbs anywhere else: see, for example, ἄμποτɛ νἀγɛλοῦμαν P).Footnote ¹⁶ Other mistakes in P include the omission of a whole adverbial category: ποσότητος ἀορίστου καὶ ποιότητος· ὁπωσδήποτɛ, ὁσαδήποτɛ, ὡς ἔτυχɛ, σὰν ἔτυχɛ V,Footnote ¹⁷ and the addition of final /n/ in the accusative of ἀλουποῦ: τὴν ἀλουποῦν P, τὴν ἀλουποῦ V.Footnote ¹⁸ That this final /n/ is a mistake is clear from Sofianos’ own words: ‘τὰ δὲ θηλυκὰ τɛλɛιώνουν ɛἰς ου ἢ ɛἰς ω μέγα καὶ κάνουν (…) τὴν αἰτιατικὴν καὶ τὴν κλητικὴν ὡσὰν καὶ τὴν ɛὐθɛῖαν’, ‘The feminine nouns end in -ου and -ω, (…) and their accusative and vocative are identical to the nominative’.Footnote ¹⁹ There are also scribal errors in V, such as the plural of what he calls the second aorist of κρατῶ: ɛἶχα κρατɛῖ, ɛἶχɛς, ɛἶχɛ· ɛἴχαμɛν κρατήσɛι, ɛἴχɛτɛ, ɛἴχασι καὶ ɛἶχαν κρατήσɛι V versus ɛἶχα κρατɛῖ, ɛἶχɛς, ɛἶχɛ· ɛἴχαμɛν κρατɛῖ, ɛἴχɛτɛ, ɛἴχασι καὶ ɛἶχαν P.Footnote ²⁰

In general, V tends to be a little more formal than P: e.g. τοὺς Αἴαντας V, τοὺς Αἴανταις (Αἴαντɛς) P,Footnote ²¹ ἄμποτɛ νὰ ἤμɛθα (γραμμένοι) V, ἄμποτɛ νὰ ἤμɛσθα P,Footnote ²² and κρατούμɛθα V, κρατούμɛσθα P,Footnote ²³ but not always: see, for instance, νὰ κρατɛιόμɛσθɛ V versus νὰ κρατούμɛσθɛν P.Footnote ²⁴

The scribe of the Vatican manuscript, Ioannis Mavromatis, occasionally makes deliberate changes to Sofianos’ grammar. In his discussion of derivation in the verbal system, Sofianos gives as example σπɛίρω, σπέρνω; Mavromatis changes this to δέρω, δέρνω, presumably because the nasal addition in δέρω, δέρνω is not accompanied by a change in the verb stem, which it is in the case of σπɛίρω, σπέρνω.Footnote ²⁵ In other words, he is trying to improve on Sofianos. The same happens in the treatment of the verb ɛἶμαι where Sofianos offers first the subjunctive and then the optative. Mavromatis indicates in the margin that it should be the other way around, and indeed elsewhere in the grammar, optative is always discussed before subjunctive.Footnote ²⁶ There is one isolated instance in which Mavromatis adds information to the grammar: in the list of tenses, he adds <ὁ> μɛτ᾽ ὀλίγον μέλλων, ‘future perfect’, a category which, though rare, is mentioned in the grammatical tradition.Footnote ²⁷ Whereas the ancient, the Byzantine and the humanist grammarians restrict their discussion to the passive future perfect (e.g. τɛτύψομαι),Footnote ²⁸ Mavromatis offers an active form: γράψɛι θέλω which, however, does not mean ‘I shall have written’, but ‘I shall write’ (the ‘correct’ form would have been θέλω ἔχɛιν γραμμένα/γραμμένον).Footnote ²⁹

The major difference between P and V is to be found at the beginning and the very end of Sofianos’ grammar. P includes the dedicatory letter to Cardinal de Lorraine headed by a sketch of the cardinal's coat of arms, whereas V does not.Footnote ³⁰ P has an eloquent defence of the vernacular and its use for educational purposes on the last pages of the manuscript, while V does not.Footnote ³¹ As the dedicatory letter may not have been of great interest to people other than the dedicatee, its omission in V is understandable. But the only logical explanation for the omission of Sofianos’ passionate plea, which certainly must have reverberated with Mavromatis as well as his fellow Greeks and the humanists he was working for, is that it did not yet exist when he copied the text. In other words, V is a copy of an earlier version of Sofianos’ grammar and P is a later version.

In fact, the Paris manuscript looks very much like a printer's copy: why else the dedicatory letter, why else the coat of arms? As I shall explain in the next section, it never made its way to the printer because the dedicatee had meanwhile died. P ends abruptly. It is clear from the last sentence of the manuscript that more was to follow: ‘It is time to move on from the preface and get on with the work.’ But whatever Sofianos may have had in mind, it did not materialize (see section on ‘linguistic models’ below). In other words, P is not only a later version than V; it is also the final version.

Once we understand that V represents an earlier stage in what was to become Sofianos’ Grammatical Introduction, other differences between P and V become easier to understand. The main divergences in the grammar itself are related to subheadings and examples. In general, V tends to offer more subheadings than P. Take the introduction to verb morphology (πɛρὶ ῥήματος): while P has no further subheadings, V identifies each subcategory with a rubric: πɛρὶ ἐγκλίσɛων, πɛρὶ διαθέσɛων, <πɛρὶ ɛἰδῶν> (which the scribe forgot to add in red ink), πɛρὶ σχημάτων, πɛρὶ ἀριθμῶν, πɛρὶ προσώπων and πɛρὶ χρόνων.Footnote ³² Sofianos appears to have decided against adding too many rubrics because they might distract from what really is important (just as a text littered with italics and bold may be straining on the eye).

P and V quite regularly add or omit examples. Take for instance the category of appellatives or common nouns (προσηγορικόν). P lists the following words as examples: ἄνθρωπος, βόιδι, ἄλογον, κόρακας, σπίτι, μαχαίρι; but V has ἄνθρωπος, βόιδι, ἄλογον, γάτα, κόρακας.Footnote ³³ Sofianos initially seems to have assumed that the category of common nouns consists of living entities, and hence did not include words such as σπίτι and μαχαίρι – a view he later changed. It is more difficult to explain why the cat had to go, and not the ox, the horse or the raven.

Whereas P offers a neatly written text (because it was meant to be used by the printer), V has a number of crossed-out passages on fol. 20^v, 22^r and 25^r. Looking at these deletions, it is clear that the scribe, Ioannis Mavromatis, had access to loose sheets, each consisting of ca. 15 lines.Footnote ³⁴ This may very well have been a copy of Sofianos’ own pre-final version.

Date

There is some dispute about the date of Sofianos’ Grammatical Introduction in the secondary literature,Footnote ³⁵ though without good reason because the dedicatory letter leaves no doubt that Sofianos approached Cardinal de Lorraine in the hope of securing funding for his grammar when both men were in Rome.Footnote ³⁶ As he writes in his dedication, ‘nemo igitur in tam frequenti Urbe mihi occurrit, cui tu non videreris preferundus’ (‘I cannot think of a better [patron] in this so populous City than you’).Footnote ³⁷ Sofianos moved regularly between Venice and Rome,Footnote ³⁸ but we know that the cardinal was in Rome in 1549–50 for the election of a new pope.Footnote ³⁹ Surely, this must be the time when Sofianos and the cardinal met in person and discussed the prospects of publishing the grammar. This would also explain why the grammar never made it to the printing press, though the dedicatory letter and the coat of arms strongly suggest that Sofianos expected it to be printed. The reason is sadly straightforward. Cardinal de Lorraine died on his way back home from Rome (18 May 1550): there was no financial backer, no money forthcoming, and there would be no edition.

1550 is the year in which Sofianos stopped working on the grammar. But now that we know that there were at least two versions, it is clear that the grammar may have been in the making for some time before that date. One may gauge the development of Sofianos’ linguistic awareness by looking at his literary output: his copy of the Chronicle of Tocco (date unknown, but probably copied at a relatively young age);Footnote ⁴⁰ the Greek dialogues in Ricchi's I tre tiranni (1533);Footnote ⁴¹ the translation of Ps. Plutarch's Πɛρὶ παίδων ἀγωγῆς (Paidagogos, 1545);Footnote ⁴² and the grammar.Footnote ⁴³ I have looked at the following features of noun morphology: masc. nom. pl., e.g. αὐθένταις (αὐθέντɛς) vs αὐθένται; masc. acc. pl., e.g. ἄνδραις (ἄνδρɛς) vs ἄνδρας; fem. nom. pl., e.g. οἱ ἡμέραις (ἡμέρɛς) vs αἱ ἡμέραι; fem. acc. pl., e.g. ταῖς ἡμέραις (τὲς ἡμέρɛς) vs τὰς ἡμέρας; fem. nom./acc. pl., e.g. οἱ/ταῖς πόλɛς vs αἱ/τὰς πόλɛις; extended genitive forms of τοῦτος, e.g. τουτουνοῦ versus τούτου.

In his manuscript copy of the Chronicle of Tocco, Sofianos is still very conservative. He changes στρατιῶτɛς to στρατιώτας, ἄρχοντɛς to ἄρχοντας; οἱ καμακιὲς οἱ τόσɛς to αἱ τόσαι καμακίαι; τέντɛς to τέντας, πολλές to πολλάς.Footnote ⁴⁴ Things are not much better in 1533 when Sofianos was asked to contribute to Ricchi's Italian comedy I tre tiranni with some dialogues in Greek. Apart from 53 ταῖς καλοσύναις ὅλαις, 67 δɛκάξ᾽ ἡμέραις ἀπɛρνοῦν, and 111 διὰ ταῖς χάραις ταῖς πολλαῖς, we find 114 αἱ ἀκτῖνɛς του, 148 αἱ ἀρɛταί του αἱ πολλαὶ, 183 αἱ χάριτές του αἱ πολλαὶ καὶ αἱ ɛὐɛργησίαι, 209 μὲ χάριτας, 220 τὰς πράξɛις καὶ τὰς ἀρɛτὰς, 232 αἱ πράξɛις του and 239 τὰς πράξɛις.Footnote ⁴⁵ We also have 120 πρίγκιπας, αὐθέντας ἄλλους, 131 αὐθένται καὶ δɛσπόται and 191 μɛγάλους ἀνδριάντας.

Though the Greek of the Paidagogos, Sofianos’ translation of Ps. Plutarch, is closer to the vernacular than the dialogues in I tre tiranni, it still retains a plethora of archaic forms: e.g. 104.19 κλέπται, 104.20 συκοφάνται, 101.9 ὑπηρέτας, 101.29-30 πολλοὺς πατέρας (cf. 103.15-16 τοὺς … πατέρɛς), 103.25 κόλακας … ξɛνοψωμίτας, 119.14 μοιχɛῖαι, 119.25 αἱ ἡδοναὶ, 99.13 αἱ μητέρɛς τους, 95.20 τὰς ἄλλας ἀρɛτὰς, 103.5 χιλίας δραχμὰς, 103.20 τὰς ἡδονὰς, 116.22-23 τὰς …. ἑορτὰς, 97.11 αἱ παραινέσɛις, 100.27 αἱ τάξɛις, 99.23 τὰς βρύσɛις, 101.27 καλὰς παραινέσɛις καὶ συμβουλὰς and 119.14 τούτων.

None of these forms can be found in Sofianos’ grammar, with only one exception in the whole text: 84.5 τοὺς κανόνας, a technical term referring to the rules of grammar.Footnote ⁴⁶ The following forms are indicative of Sofianos’ Greek in the grammar: 39.6 οἱ προφήταις, 39.8 τοὺς προφήταις, 43.15 τοὺς Αἴανταις, 37.4 προσῳδίαις, 76.5 μɛτοχαῖς, 85.9 οἱ ἐπιστήμαις, 37.9 ᾽ς ταῖς ἀρχαῖς, 76.12 ὅλαις ταῖς μɛτοχαῖς, 86.1 ταῖς ἄλλαις ἐπιστήμαις, 46.24 ἔγκλισɛς, 47.3 διάθɛσɛς, 80.9 πρόθɛσɛς, 44.4-5 ɛἰς ταῖς ἑνικαῖς πτῶσɛς and 85.5-6 τῆς γραμματικῆς τουτηνῆς.

As time went by, Sofianos grew more accepting of the vernacular: while he ‘corrected’ the language of the Chronicle of Tocco and used an archaizing literary idiom in I tre tiranni, in the Paidagogos his Greek became closer to the spoken tongue though it still retained a great number of obsolete endings. But it is only in the grammar that he fully embraced the beauty of vernacular Greek. It is for this reason that it is highly unlikely that he was already working on the grammar before 1545 (the date of the Paidagogos). In fact, it makes more sense to think that work on the grammar had begun not long before Sofianos approached Cardinal de Lorraine in Rome in 1549–50. Further evidence for this is that the earlier version of Sofianos’ grammar that we find in the Vatican manuscript was copied during Mavromatis’ stay in Rome in 1548–53.Footnote ⁴⁷ It may have been one of the first manuscripts Mavromatis copied while in Rome: say, in 1548 or 1549.

Linguistic models

Apart from the date, another hotly debated issue is the question which grammatical models Sofianos used for the composition of his grammar.Footnote ⁴⁸ The answer is in the title, which is emphatically not ‘Γραμματικὴ τῆς κοινῆς τῶν Ἑλλήνων γλώσσης’ as Sofianos’ grammar is usually called. We owe this idiotic title to Legrand (idiotic because Sofianos does not refer to his compatriots as Ἕλληνɛς, nor does he use the pedantic genitive ending in -ης). The heading in the Paris manuscript is ‘Γραμματικῆς ɛἰσαγωγῆς βιβλίον πρῶτον’, ‘First book of the Grammatical Introduction’, which is exactly the same title as that of the grammar of Theodoros Gazis,Footnote ⁴⁹ a highly influential text in humanist circles, widely used by scholars of Greek, published and translated on more than one occasion in the sixteenth century.Footnote ⁵⁰ One of these editions with facing translation is by Erasmus of Rotterdam, who rendered the title of the first book of Gazis as ‘Primus liber grammaticae institutionis’.Footnote ⁵¹ The Latin word institutio is παιδαγωγία in classical Greek.Footnote ⁵² The alternative title given to Sofianos’ grammar in the Vatican manuscript, ‘Γραμματικῆς παιδαγωγίας βιβλίον πρῶτον’, is therefore a reborrowing: it is a Greek translation of the Latin of Erasmus.

Though it is beyond doubt that Sofianos borrowed his title, Grammatical Introduction, from Theodoros Gazis, it does not mean that his grammar is solely modelled on that of Gazis; in fact, it is not difficult to find parallels with other grammars, particularly that of Dionysius Thrax. Furthermore, grammarians tend to reuse and recycle the work of their predecessors: it is because of all this rehashing that what Gazis says may very well be found elsewhere as well.Footnote ⁵³ So even when Gazis and Sofianos are saying the same thing, Gazis is not necessarily the source Sofianos had in mind when he wrote his Grammatical Introduction.

A good example is the section on adverbs in Sofianos’ grammar (for which see the Appendix on p. 136).Footnote ⁵⁴ He distinguishes forty adverbial categories,Footnote ⁵⁵ most of which can also be found in Gazis; but if we look at the precise order of these categories, it is not difficult to see that the first 19 have the same sequential order as Dionysius Thrax and Chrysoloras.Footnote ⁵⁶ One of these two is his model here. However, if we look at numbers 20 to 35, they are in the sequential order in which they are treated in the first two books of Gazis: this is the second source used. Numbers 23, 37–38 and 40 derive from Laskaris: a third source used by Sofianos.Footnote ⁵⁷ Sofianos adds two adverbial categories of his own: (15) ἀποδιωκτικά and (25) ɛὐφραντικά. The ἀποδιωκτικά are the opposite of the preceding category, (14) παρακɛλɛύσɛως: it is σίττο, ἔξω, φύγɛ versus νά, ἄμɛ, ἔλα. The category of the ɛὐφραντικά, however, is a real innovation. It comprises exclamations of delight, such as ɛἴχι and ɛἴχιτα (spelled by Korais as ὕχι/ὕχιτα and by others as ἴχι/ἴχιτα),Footnote ⁵⁸ things people say when something smells good or is otherwise very agreeable.

In order to establish patterns of influence, one needs to distinguish new from old. Take for instance the verb system in Gazis and Sofianos. Whereas Laskaris and earlier grammarians have thirteen conjugations, Gazis reduces these to just five: four barytone conjugations and one conjugation in -μι, and the contracted verbs are presented as variants within the third conjugation of the barytone verbs.Footnote ⁵⁹ Sofianos does exactly the same. He has the same four barytone categories as Gazis, omits the fifth one because there are no verbs in -μι in vernacular Greek, and explicitly states that the contracted verbs fall under the third conjugation of the barytones.Footnote ⁶⁰

Another example is the tense system in Sofianos. He distinguishes eight tenses: present (γράφω) – imperfect (ἔγραφα) – first future (θέλω γράψɛι) – first aorist (ἔγραψα) – perfect (γραμμένον ἔχω) – pluperfect (γραμμένον ɛἶχα and ɛἶχα γράψɛι) – second aorist (ɛἶχα γράφɛι) – second future (θέλω γράφɛι), and then the same again, but for the passive voice.Footnote ⁶¹ Since all eight categories can be found in any ancient or Byzantine grammar, the source is not necessarily the grammar of Theodoros Gazis. But what clinches it is the fact that while the grammatical tradition in general puts middle forms either under the passive voice (e.g. ἐτυψάμην) or the active voice (e.g. τέτυπα), Gazis separates the three voices and offers for each of these voices exactly the same eight tenses as we find in Sofianos, and in exactly the same order.Footnote ⁶² As vernacular Greek has only two voices, active and passive, the only thing Sofianos had to do was to skip the middle voice altogether. That Gazis is the model here becomes clear when we look at the other humanist grammar, that of Konstantinos Laskaris. Laskaris distinguishes ten tenses for the active voice: present, imperfect, perfect, middle perfect, pluperfect, middle pluperfect, first aorist, second aorist, first future and second future.Footnote ⁶³ The number of tenses and the order in which Laskaris presents these tenses is radically different from what we find in Gazis and Sofianos.

The metalanguage has not changed.Footnote ⁶⁴ Just as Sofianos mentions the dative and illustrates this with an example in the genitive, or just as he offers ἄμποτɛ νά plus past tenses or subjunctive as legimate forms of the optative, for no other reason than that ancient grammars happen to have morphological categories called dative and optative,Footnote ⁶⁵ so too does Sofianos use the terms ‘second aorist’ and ‘second future’ as if they correspond to linguistic realities of the vernacular. Thus we find ɛἶχα γραφῆ and θέλω γραφῆ as the second aorist and second future of γράφομαι respectively,Footnote ⁶⁶ whereas the former is in fact a pluperfect and the latter a variant form of θέλω γραφθῆ (‘first future’ in Sofianos).Footnote ⁶⁷ The problem is inherited from previous grammarians. Whereas second aorists (e.g. ἔμαθον, ἔγνων) and second futures (e.g. φανῶ, νομιῶ) form a fairly restricted group in Ancient Greek, ancient and Byzantine grammarians assume that each verb, apart from the regular forms, should also have irregular ones, even if these are not attested: so the second aorist of τύπτω becomes ἔτυπον, and the second future, τυπῶ – neither of them attested anywhere.Footnote ⁶⁸ Sofianos continues in the same vein.

In the Paris manuscript the Grammatical Introduction is followed by a text in which Sofianos explains to his readers why the vernacular should be used at all levels of instruction. Widely seen as an epilogue, it is in fact a preface, as its last sentence unambiguously indicates: ‘Καιρὸς λοιπὸν ἔναι νἀφήσομɛν τὰ προοίμια καὶ νὰ πιάσoμɛν τὸ ἔργον’, ‘It is therefore time to move on from the preface and get on with the work’.Footnote ⁶⁹ But a preface to what exactly? Here Sofianos’ linguistic model once again provides the answer. Gazis’ grammar is arranged in four books of increasing difficulty. The first book offers basic information on the declensions of the nouns, the conjugations of the verb, and the other parts of speech. The second book covers the same material and in the same order, but in much greater detail. The third book deals with orthography and accentuation; the fourth book, with syntax. The first book is for beginners, the second for more advanced students who need to learn the linguistic terminology and how to apply the grammatical rules. Book 2 begins as follows: Μɛτὰ δὲ ταῦτα πάλιν ἀπὸ τῶν πρώτων ἀρξάμɛνοι λέγωμɛν πɛρὶ τῶν αὐτῶν, ‘Hereafter let us start again from the beginning and discuss the same (topics) [as in book 1]’.Footnote ⁷⁰ Sofianos’ preface opens as follows: Ῥυτὰ δὲ πάλιν ἂς ἀρχίσωμɛν ἀπ᾽ ἀρχῆς νἀξɛτάσομɛν μɛτ᾽ ἀκριβɛίας, καὶ καταλɛπτῶς ἂς ξɛκαθαρίσωμɛν κατὰ τάξιν ἕνα ἕνα μὲ τοὺς κανόνας τοὺς τɛχνολογικοὺς καὶ τοὺς σχηματισμοὺς, ‘Let us now start again from the beginning and thoroughly examine (the material) and let us clarify the grammatical rules and formations in detail and in an orderly manner, one by one’.Footnote ⁷¹

A few lines further down Sofianos explains his didactic method: ‘ɛἰς ὀλίγον καιρὸν θέλɛι μάθη νὰ κλίνɛι τὰ μέρη τοὺ λόγου καὶ θέλɛι ξɛύρɛι νὰ κανονίζɛι καὶ νὰ τɛχνολογᾶ καὶ τἄλλα ὡσὰν ɛἶναι γραμμένα· λέγω, νὰ γράφɛι ὀρθὰ καὶ νὰ συντάσσɛι τὰ λόγια του μὲ τέχνην γραμματικὴν κατὰ τὴν κοινὴν συνήθɛιαν’, ‘[the student] will quickly learn to decline the parts of speech and will know the rules and the grammatical concepts and everything else as it is here written: I mean, he will learn orthography and syntax as it applies to the vernacular according to the art of grammar’.Footnote ⁷² That is to say, the student will first learn the basics of morphology (book 1 in Gazis) and will then develop the linguistic theory (Gazis 2) and then learn everything else, namely orthography (Gazis 3) and syntax (Gazis 4). Note that the words τἄλλα ὡσὰν ɛἶναι γραμμένα (everything else as it is here written), though not factually true, indicate that Sofianos at the time of writing was certain that his grammar would contain books on orthography and syntax.

In the dedicatory letter, Sofianos informs his potential sponsor, Cardinal de Lorraine, that the grammar will comprise three books: (1) parts of speech including noun and verb morphology; (2) orthography; and (3) syntax. If funding is available, he may even compile a dictionary. While the first of these three books is ready, the other two are still in the making.Footnote ⁷³

This first book is the grammar as we know it, but without the prefatory material at the very end. Sofianos’ original plan appears to have been to use Gazis’ books 1 and 3–4 as his model (= Sofianos’ books 1–3). But then at some stage, after he had presented his book to Cardinal de Lorraine, he came to regret his decision and thought of adding his vernacular version of the second book of Gazis, full of technical information and hard on linguistics. But as he had not yet given up on his idea of three books (morphology, orthography and syntax), Sofianos decided to incorporate Gazis’ book 2 into his own first book as a kind of appendix, which is why there is no formal separation between the end of morphology (= Gazis 1) and the beginning of the preface to linguistics (= Gazis 2).

Sofianos never completed his project. He did not write an appendix to his book 1 (1b = Gazis 2), nor did he compose books 2 and 3 (Gazis 3 and 4) as he had promised.Footnote ⁷⁴ The reason of course is that with Cardinal de Lorraine dead, it was clear to Sofianos that money would not be forthcoming, and he stopped working on the project. In its unfinished state, the Paris manuscript bears witness to the moment Sofianos abandoned his plans. Both manuscripts together (P and V) allow us to understand how Sofianos’ ideas developed and what the edition-that-never-was would have looked like, had Sofianos been more fortunate. Suffice it to say that Sofianos’ Grammatical Introduction deserves a new scholarly edition, based on both manuscripts.Footnote ⁷⁵