γίγνομαι, γιγνώσκω
(Moer. γ 3)

D. General commentary

Atticist lexicography correctly identifies γίγνομαι and γιγνώσκω as the old Greek and Attic spellings. A strict Atticist such as Moeris (A.1) thus recommends γιγν-. The Atticist lexicon preserved by P.Oxy. 1803 (A.2) takes a more tolerant position: γίγνομαι and γιγνώσκω are attributed to the παλαιοίοἱ παλαιοί, but γιν- is acceptable (this is in line with the moderate Atticism professed by the compiler of P.Oxy. 1803, see F.1). In later times, Eustathius too acknowledges that γιγν- is Attic and more accurate; while he does not object to γίγνομαι (B.4), however, in the case of γιγνώσκω, he does not easily renounce γιν-, which he considers more euphonic (B.3; on these passages see F.3).

The phonetic development behind the changing spelling γιγν- > γιν- is best explained as due to dissimilation with initial [g], followed by nasalisation and assimilation [gn] > [ŋn] > [nn] > [n], ultimately causing the compensatory lengthening [i] > [i:] in the first syllable (see Schwyzer 1939, 214–5, Lejeune 1987, 78–9, and Threatte 1980, 562). The chronology of this development is unevenly distributed across the Greek dialects where it took place (see Schwyzer 1939, 215 and Buck 1955, 74). In Ionic it is rather early: γιν- is the standard spelling in Herodotus and the corpus Hippocraticum, and the inscriptions confirm its early diffusion in Eastern Ionic (in LSAM, no. 44 = Greek Ritual Norms, no. 39 [Miletos, 400 BCE] γιν- occurs at lines 2, 14, and 15 [certain restoration]).

The Attic evidence is rich but problematic. With regard to literary texts, γιγν- is the only spelling present in 5th- and early 4th-century writers (e.g. Plato, Xenophon, and the orators). This spelling is generalised in modern editions, although no systematic investigation has analysed the extent to which the competing spelling γιν- is attested in the manuscripts. The situation becomes far more complicated in later 4th-century authors, as the two spellings alternate in the manuscript traditions. In the corpus Aristotelicum, γιγν- and γιν- alternate at random (quite often within the same line, as in EE 1220a.23–7 and 1239a.36–8 or Po. 1452a.18–20 and 1453b.27–8). Most editors therefore print what appears at each occurrence in the manuscripts. A remarkable exception is the Constitution of the Athenians, where editors normally restore γιγν-, even though γιν- is consistently adopted on the papyrus (C.1; see F.4). Theophrastus’ Characters is a similar case: although γιν- is the spelling consistently adopted in the three main manuscripts, editors tend to generalise γιγν- (thus Diels 1909, Navarre 1964, and Diggle 2004; Immisch 1923 and Steinmetz 1960 retain γιν-). New ComedyNew Comedy also poses major challenges (C.3, C.4, C.5, C.6, C.7; see F.5, F.6). The fact that γιν- first crept into later 4th-century texts may be corroborated with evidence from the Derveni papyrusDerveni papyrus (dating to around 340–320 BCE), which regularly adopts γιν- for both γίγνομαι and γιγνώσκω (that the language of the Derveni papyrus is Attic is convincingly argued by Willi 2014, 54–60, who also comments on γιν-). Another 4th-century work which similarly attests to the spread of the spelling γιν- is Aeneas Tacticus’ Poliorcetica, even though γιγν- remains the more common option in this text (γιγν- occurs 39x, γιν- 7x; see Vela Tejada 2018, 100–1). These two texts indicate that γιν- had become the spelling of international Attic. As a further complication, the chronology of the spelling γιν- is different in literary and documentary texts (see Threatte 1980, 562–5 for the epigraphic data). In fact, the first documented occurrence of γιν- on an Athenian inscription dates only to 306/5 BCE (IG 2².2499.38). This is because of the tendency of inscriptions to retain conservative orthography. In the early decades of the Hellenistic period, γιγν- and γιν- still alternate, but from around 250 BCE, the spelling γιγν- disappears and γιν- becomes the standard.

The distribution of γιγν- and γιν- in post-classical authors is also a thorny issue (see Dhont 2018, 122 and n. 90 on the book of Job). We can, however, map some general tendencies. In Hellenistic and early imperial prose, γιν- is clearly the standard spelling (γιγν- is unattested in Polybius, and Strabo only uses γιγν- three times in passages where he is not quoting an earlier source verbatim). However, γιγν- returns to use from the 2nd century CE onward. Plutarch offers a test case: although he mostly uses γιν-, γιγν- is also well documented, with around half the number of occurrences as γιν- (see Vela Tejada 2019, 297). Galen alternates γιν- and γιγν- (sample data in Vela Tejada 2009, 46).

In the case of the Atticising writers of the Second Sophistic, the picture is highly variable. Some writers clearly make the choice to use just one of the two spellings. Thus, a strictly Atticising writer such as Aristides exclusively uses γιγν- (see Schmid, Atticismus vol. 2, 29), whereas a less markedly Atticising writer such as Aelian chooses to use γιν- on all occasions (see Schmid, Atticismus vol. 2, 18 and 39). In several cases, however, the situation is much less clear, and issues of transmission play a major role in shaping our understanding of the evidence. Consider, for example, the conflicting editorial practices of Bost-Pouderon 2011, XV and Panzeri 2011, 146 with regard to Dio Chrysostom. As for Lucian, Deferrari (1916, 36–9) (who does not consider verbal compounds) shows that γιν- was the normal spelling in the archetype, but that γιγν- was restored in some of Lucian’s writings, according to the habits of the scribe. Scholars working on Lucian’s individual works have sought to identify some general tendencies. With regard to De dea Syria, for instance, Lightfoot (2003, 109) argues that γιγν- is more common in the present participle and γιν- in the finite verb. According to Tomassi (2011, 130–1), meanwhile, γιν- is the only attested spelling in Timon or the misanthrope. The situation in Philostratus is even more polarised. In Lives of the sophists, γιγν- is the only attested spelling, and in the Life of Apollonius of Tyana too, γιγν- has the upper hand. In the Heroicus, however, γιγν- only occurs twice in the manuscripts, and more recent editors choose to generalise γιν- (see De Lannoy 1977, XXIV and Follet 2017, CLXXVIII). A likely conclusion is that one or the other spelling was generalised at a very early stage in the transmission of these works, and that the spelling was then retained via the archetypes in the medieval transmission.

Papyri written in high-style, Atticising Greek also offer some interesting material. A useful discussion is provided by Luiselli (1999, 157, 160, 161, and 162). I present his main findings below, together with some additional comments.

(i) The spelling γιγν- may be adopted alongside other purist features. For example, in P.Herm. 2 (= TM 21121)P.Herm. 2 (= TM 21121) [Hermoupolis (?), 317–23 CE], γίγνονται occurs at line 22; other purist elements are οἶσθα at line 3 and φυλάττω and ἁρμόττω at lines 11 and 13.

(ii) The spelling γιν- also co-exists with purist features. In P.David 14 (= TM 27544)P.David 14 (= TM 27544) [provenance unknown, 2nd century (?)], for example, γινώ[σκειν] occurs at line 19, but purist elements are represented by ἀμφοῖν at line 6, μέχρι at line 18, and οἶμαι at line 26 (although, as Luiselli notices, οἴομαι might have been more proper).

(iii) In some cases, γιγν- and γιν- alternate. In P.Neph. 8 (= TM 21449)P.Neph. 8 (= TM 21449) [Alexandria, post 352 CE], there is a blending of purist and non-purist elements: the former are γίγνωσκε at line 4 and τέως at line 20 (see also χάριν ἔσχον at line 3, which is, however, a classical rather than a purist usage); the latter are γίνεσθαι, γίνωσκε, and γινωσκέτω at lines 11, 17, and 30, καθώς at line 16, and the verbs in -σσω at lines 6, 8, and 12. This distribution of purist, classicising, and non-purist elements in P.Neph. 8 invites comparison with P.Oxy. 1.113 (= TM 28405)P.Oxy. 1.113 (= TM 28405) [Oxyrhynchus, 2nd century CE], where the spelling γιγνώσκων at line 13 is paralleled by the classicising formula χάριν ἔχω at line 13, although the obviously non-purist forms τάχιον and θέλεις occur at lines 24 and 30.

This evidence provides a comparandum to the doctrine of less strict Atticists lexicographers, such as the compiler of the lexicon preserved by P.Oxy. 1803 (A.2), who allows for γιν- alongside γιγν-. It seems that systematically adopting γιγν- could be perceived as an excessively purist practice even in stylistically marked texts. For instance, the spelling γιγν- is absent from the 6th-century Petra papyri, even though most of these texts adopt Atticising orthography and morphology (see further Vierros 2018).

There is an array of factors to explain why even Atticising writers allowed for γιν-. For a start, the use of γιν- in the texts of Aristotle and Menander make it appear to be an ‘ancient’ spelling. Further, even though grammarians normally indicate γίγνομαι and γιγνώσκω as the proper forms from the point of view of etymology (see B.3, B.4), at some point the spelling γιν- in γινώσκω was justified through analogy. In the Epimerismi Homerici (B.2), mention is made of the doctrine of those who leave out the second gamma in γιγνώσκω, and their views are set in parallel to μισγητή > μισητή and γέφυγρα > γέφυρα. A fragment of the Herodianic treatise Περὶ παθῶν (B.1), meanwhile, offers a comparison between πινύσκω and γινώσκω (see F.2). These passages show that γινώσκω is considered the proper outcome, or at least one of the proper outcomes, of the derivation process. This may well contribute to the idea that the spelling γινώσκω, and by extension γίνομαι as well, should be deemed acceptable.

The consensus view is that the use of γιγν- in Post-classical Greek was solely a matter of spelling and did not imply any change in pronunciation (see Threatte 1980, 562). However, one wonders whether on occasions such as that of a sophist giving a public speech, efforts were made to distinguish the pronunciation of γιγν- from that of γιν-. This cannot be proved, but we may consider some indirect evidence. We may ask, for example, whether Eustathius’ claim (B.3) that while γιγν- is Attic and more precise, γιν- is more euphonic might recall ancient sources who commented on the general dislike of attempts at reviving the pronunciation of γιγν- (see F.3). Further, if trying to revive the pronunciation of γιγν- was regarded as cacophonic, this would also explain why this spelling is avoided or adopted only to a very limited extent in several Atticising writers and texts.

As regards post-classical documentary texts, γιν- remains the standard spelling throughout Hellenistic and Roman times (see Mayser, Gramm vol. 1,1, 156–7 on the Ptolemaic evidence). In Attic inscriptions, on the other hand, γιγν- comes back into use from the second half of the 2nd century CE, a form of ‘learned revival’, according to Threatte (1980, 562). The older spelling γιγν- is adopted not as an element of purist language: no other obvious attempt at using a more Attic language is discernible in the inscriptions that adopt γιγν-; further, it γιγν- and γιν- sometimes alternate within the same text (see Oliver 1941, 125–141 no. 31 lines 14 and 48 [Athens, ca. 230 CE]). The case of the twin inscriptions IG 2².13209IG 2².13209 and IG 2².13210IG 2².13210 (mid-2nd century CE) may be a rare example where the adoption of γιγν- is socio-linguistically significant, considering that these texts are plausibly connected with the world of Athenian rhetoricians and philosophers (see Robert 1978, 244–5). The spelling γιγν- enjoys continued popularity in early Byzantine papyri (see Gignac 1976, 176). A sample search I have made for γιγνώσκω in Greek papyri confirms these general conclusions, since 37 out of a total of 44 instances of the spelling γιγν- (also written γειγν-) date from the 4th to the 8th century CE. Still, the increasing adoption of γιγν- is equally documented in official and informal texts, and so again the adoption of γιγν- counts more as a renewed orthographic convention than an attempt at adopting a purist language.

E. Byzantine and Modern Greek commentary

The spelling γιν- is dominant in Byzantine texts and remains standard in Modern Greek (only for γίνομαι, since γιγνώσκω does not survive). Judging from the available critical editions, the evidence for γιγν- is meagre. In the chronographers, the spelling γιγν- is attested only once in George Syncellus (on the absence of γιγν- in chronicles, see Psaltis 1913, 99). Similarly, γιγν- is unattested in the writings of Michael Choniates, Nicetas Choniates, and Anna Comnene. As for Arethas, γιγν- is attested only five times, whereas γιν- 300x. In Eustathius’ corpus, leaving aside the Homeric commentaries, γιγν- only occurs five times. This situation is true also for other genres (see, for example, Mitsakis 1967, 19 on Romanus the Melodist), although in other authors the situation is slightly less unbalanced. The writings of Procopius of Caesarea contain 31 occurrences of γιγν-, though only one with γιγνώσκω. In each case, γιγν- is used with the simple forms, while γιν- is attested several hundreds of times both with the simple verbs and with its compounds. In Photius, γιγν- is limited to around 25 instances, over 20 times with γίγνομαι and only four times with the compounds of γιγνώκσω. As for Michael Psellus, γιν- is massively predominant (several hundred occurrences), but there also are 150 examples of γιγν-. In Psellus too the disproportion is particularly clear with γιγνώσκω and its compounds: with this verb, γιγν- is attested only three times and γιν- over three hundred times. It remains difficult to account for the distribution of the two spellings, but it seems that γιγν- is more common with the simple forms of γίγνομαι than with the compounds, and that γιγν- is particularly rare with γιγνώσκω.

Palaeologan writers are seemingly a turning point. In Maximus Planudes, Nicephorus Chumnus, and Philotheus Coccinus, γιγν- occurs hundreds of times. This increasing use of γιγν- may perhaps relate to the Atticist revival of Palaeologan Byzantium (see Gaul 2011, 272–310). A more systematic investigation would be required to confirm this, although we may touch on other pieces of evidence here. In the manuscript transmission of classical authors, γιν- is a common trivialisation, but Planudes’ manuscripts of Hesiod, Theognides, and the Greek Anthology offer examples where γιγν- was restored deliberately (if unnecessarily, since these are Ionic texts; see Condello 2018–2019, 36–7, who at 36 n. 135 also collects bibliographical references and comparable examples of Planudean restoration of γιγν- in Hesiod and the epigrams of the Greek Anthology).

F. Commentary on individual texts and occurrences

(1)    P.Oxy. 1803.col. i verso.11–5 (A.2)

All surviving glosses of this lexicon begin with sigma, and it is puzzling that the orthography of γίγνομαι is discussed in a gloss about συγγίγνομαι. One explanation may be that P.Oxy. 1803 only included words beginning with sigma, and thus the general treatment of the orthography of γίγνομαι appears under the compound. Alternatively, συγγίγνομαι may derive from a locus classicus, though we have no evidence of it. The formulation that the correct forms γίγνομαι and γιγκνώσκω are κατ’ ἐπένθεσιν τοῦ γ̅) is odd, given the ancient etymological analyses of γίνομαι and γινώσκω, which explained them as created, respectively, from γίγνομαι and γιγνώσκω (see F.3). It is likely that the compiler of P.Oxy. 1803 simply described γίνομαι and γινώσκω in a rather imprecise way.

(2)    Hdn. Περὶ παθῶν GG 3,2.179,1–5 (= EM 672.45–50) (B.1)

According to the Herodianic source, πινύσκω is created either by addition of iota (πνύσκω > πινύσκω) or as the effect of reduplication, elimination of pi, and lengthening of [i] (i.e. πνύσκω > πιπνύσκω > πῑνύσκω). In the second explanation, the comparison with γιγνώσκω > γῑνώσκω is apparent. This latter option, however, is regarded as less likely: the cogency of postulating reduplication, loss of pi, and lengthening is hindered by the fact that πῑνύσκω does not exist (the correct prosody is πῐνύσκω). This explains why the grammarian inclines more towards the first interpretation.

(3)    Eust. in Il. 3.862.9–22 (B.3), Eust. in Od. 2.25.6–7 (B.4)

Eustathius does not indicate from which treatise of Heraclides of MiletusHeraclides of Miletus he is quoting, but these passages are plausibly attributed to the Περὶ δυσκλίτων ῥημάτων, a work devoted to the study of verbs with an irregular conjugation (see Cohn 1884, 6–7). Heraclides’ doctrines are similar to, though they do not always overlap completely with, a larger body of erudite treatises concerning the formation of γιγνώσκω and γίγνομαι (see Hdn. Περὶ ὀρθογραφίας GG 3,2.486.25–30 [= Orth. 188.26–32 [AO vol. 2], EM 232.24–34 ~ Et.Sym. γ 97 Baldi], Choerob. Orth. 187.6–16 [AO vol. 2], Epim.Hom. γ 28 [part of this gloss appears as B.2], Et.Gud. 311.20–312.3 [cf. Et.Gud. 311.1–6]).

These two passages derive from the same source and present roughly the same doctrine. Eustathius makes the connection between γίγνομαι and γιγνώσκω explicit, but we have no evidence that Heraclides did so also. The derivation processes of γιγνώσκω and γίγνομαι are different from one another, and so they may originally have been discussed independently. We can also detect some inconsistencies in Eustathius’ treatment of γιγνώσκω and γίγνομαι. In the first passage (B.3), γινώσκω is presented as Homeric and thus more archaic than the Attic γιγνώσκω, which in turn is called a later usage. In the second passage (B.4), γίγνομαι is attributed to οἱ μὲν παλαιοίοἱ παλαιοί, while γίνομαι to οἱ δὲ ὕστερονοἱ ὕστερον, and no attempt is made to defend γίνομαι; nor is there any remark about γίγνομαι as less pleasant to the ear than γίνομαι. It is unclear whether this inconsistency is accidental. Yet, the fact that in Byzantine texts γίγνομαι is far more common than γιγνώσκω (see E.) might help to explain Eustathius’ inconsistency. In other words, Eustathius’ remark that γιγνώσκω is cacophonic may indicate the Byzantine dislike for γιγνώσκω, while, in the case of γίνομαι, he acknowledges that γίγνομαι is the correct spelling adopted by the ancients and regards γίνομαι as only a means of ‘sparing ink’.

(4)    Arist. Ath. 2.2 (C.1)

Even though γιν- is consistently adopted in the papyrus, modern editors typically restore γιγν- qua the 4th-century spelling as reconstructed from the inscriptions (see Chambers 1994, IX). This is problematic. For a start, we do not know whether 4th-century literary texts followed the same conventions as inscriptions. By way of comparison, before the reform of the Athenian alphabet took place in 403/2, elements of the Ionic alphabet were already in use for private and possibly also literary texts (see Threatte 1980, 33–45, Immerwahr 1990, 177, and D’Angour 1999). Secondly, γιν- is extremely common in the Aristotelian writings transmitted via Byzantine manuscripts: should one generalise the spelling γιγν- that has been adopted by the editors of the Constitution of the Athenians? This would be highly hazardous. Considering how widespread γιν- is in the manuscripts, it is an easy guess that γιγν- and γιν- already alternated in the ancient transmission. These observations should warn us against generalising γιγν- in the Constitution of the Athenians.

(5)    Men. Sic. 27 (C.2), Men. Sam. 65 (C.3), Men. Sam. 212 (C.4), Men. Sam. 223 (C.5), Men. Sam. 490 (C.6)

The Sorbonne papyrus of Sikyonioi is the earliest available witness of Menander’s comedies (P.Sorb. inv. 2273b fr. IV A [last third of the 3rd century BCE]). Besides C.2, no other example of γίγνομαι or γιγνώκσω occurs to allow a comparison. In the case of Samia, the Bodmer papyrus (P.Bodm. 25 [late-3rd or early-4th century CE]) and the Cairo codex (P.Cair. 43227 [5th century CE]) also provide evidence for γιγν- (C.3, C.4, C.5, C.6). These occurrences are at variance with other cases, where γιν- is the only attested spelling in either or both papyri (see Sam. 43, 65, 426, 594, 619, 686).

According to the data collected by Arnott (2002, 195), in the papyri of later Greek comedy γ(ε)ίνομαι occurs at least 32 times and γ(ε)ινώσκω 12 times. Likewise, in Menandrean fragments known via the indirect tradition, the best attested spelling is γιν-, which editors generalise (see Men. frr. 683, *871.7, 770.1, 844.5; see also Kassel and Austin in PCG vol. 6,2, 353 ad Men. fr. 717.2). This is similar to the situation in Diphilus, where the spelling is γιν- for both occurrences. Arnott (2002, 195–6) comes to the negative conclusion that we simply cannot say which spelling was employed by Menander, nor whether we could expect him to be consistent in his orthography. Thus, modern editions are often in disagreement with one another. Some observations concerning Sikyonioi and Samia may be offered nonetheless.

As regards Sikyonioi, there is only one occurrence, and the editors thus retain γίγνεται. However, γιγν- is not necessarily to be considered genuinely Menandrean qua the minority spelling, despite the early chronology of the Sorbonne papyrus. The use of γιγν- is consistent with the adoption of another literary spelling: ηὐ- in augmented verbs beginning with εὐ- (ηὐχόμην at Sic. 48, on which see Arnott 2000, 198). Still, it is noteworthy that the copyist of the Sorbonne papyrus is an otherwise rather careless scribe.

In the case of Samia, editors make different choices. Arnott (2000, 62 and 122) and Sommerstein (2013, 70 and 82) restore γιν- against the manuscript evidence. Sommerstein 2013, 46, writes that by Menander’s time ‘γίνομαι and γινώσκω have largely supplanted γιγν-’, but this claim must face the objection that γιγν- is still used in inscriptional texts until around 250 BCE. Thus, it is not easy to accept the claim that during Menander’s lifetime, γιν- had already become the standard spelling. Unlike Arnott and Sommerstein, other editors retain γιγν- at lines 212 and 490, where it is the only reading of the Bodmer papyrus (see Austin 1969, 39 and 49, Jacques 1971, 15 and 33, Sisti 1974, 32 and 48, and Sandbach 1990, 240 and 253). However, the same editors also print γιν- at Sam. 223: that is, in the only case where γιγν- and γιν- are competing variant readings in the papyri, they choose γιν-. The spelling γιν- is clearly regarded as less exceptional than γιγν-.

It is difficult to account for this conflicting evidence, but we can propose some considerations. At Sam. 65, the Bodmer papyrus’ first reading γίγνομαι is corrected by the same hand into γείνομαι (i.e. γίνομαι). We might therefore suspect that the scribe aimed to restore the spelling of the antigraphon, while γιγν- was a slip caused by scribal habit. The Bodmer papyrus dates to the 3rd or 4th century CE, and the papyrological evidence shows that in later imperial times the spelling γιγν- came back into general use (see D.). In such a case, the spelling of γιγν- that occurs on the Bodmer papyrus at Sam. 212 and 490 might equally be suspected to be a slip, which in this case passed unnoticed. This scenario would also explain the occasional instance of γιγν- in the Cairo codex vis-à-vis γιν- in the Bodmer papyrus at Sam. 223.

(6)    Philem. fr. 64.1 (C.7)

In Philemon’s fragments, γιγν- and γιν- are both attested, although γιν- is minoritarian and quite rarely the only reading (see C.7 and further frr. 9.4, 59.2, 72.1, 101.6, 106.2, 125.2, 126.3, 174). While previous editors retained γιν-, Kassel and Austin systematically restore γιγν-. This is sensible, but it is not obvious why Philemon presents such a different case to Menander and Diphilus (see F.5). A possible explanation might be chronological: Philemon was born in 368 BCE, Menander and Diphilus in 342 BCE; further, Philemon was sometimes regarded as a poet of Middle Comedy (see Philem. test. 1 and 7). We might infer, then, that the spelling γιγν- was still the norm in (the editions of) comic texts from a generation, if that, before Menander. This relative chronology is also consistent with the adoption of γιν- in the Derveni papyrus, which dates to the beginning of ca. 340–320 BCE (see D.). Still, the substantial evidence of γιγν- in the transmission of Philemon’s fragments is remarkable. In the indirect transmission of all comic fragments, there is a strong tendency in the quoting sources to render γιγν- as γιν-. In the large corpus of Alexis’ fragments, for example, Kassel and Austin systematically restore γιγν-, although in virtually all cases the manuscripts have γιν- (see Alex. frr. 37.7, 63.2, 76.3, 103.2, 153.17, 160.4, 167.9, 178.15, 215.5, 257.2, 280.2; the only occurrence of γιγν- is in fr. 47.7).