ἁπλᾶ, διπλᾶ, τριπλᾶ
(Phryn. PS 43.17–9, Moer. τ 22)

D. General commentary

Entries in the Atticist lexica of Phrynichus (A.1) and Moeris (A.2) prescribe the contracted forms of multiplicative adjectives in -πλόος > -πλοῦς, while the uncontracted forms, considered typical of IonicIonic, are proscribed. Similar recommendations are found in the even more numerous entries dealing with adjectives of materialAdjectives of material, which also belong to the contract thematic declension in Attic: a fragment (B.1) ascribed to the Praeparatio sophistica of Phrynichus, but likely resulting from the conflation of multiple sources, treats both adjective classes together (see entry χρύσεος, χάλκεος, φοινίκεος).

The multiplicative adjectives in -πλόος – such as διπλόος, Attic διπλοῦς ‘double’ – have a suffix of unclear origin: it was perhaps remade from -πλος (which is expected, based on the comparison with other Indo-European languages: cf. Latin du-plus, Gothic twei-fl, etc.) under the influence of verbal governing compoundsCompounds with a second member in -CόCos, possibly as a result of confusion with compounds in -πλόϝος ‘sailing’ (see Kretschmer 1923, 218; Schwyzer 1939, 598; Dieu 2022, 386‒7). Forms of the kind ἁπλόςἁπλός (An.Ox. 2.231), διπλόςδιπλός (Opp. C. 2.449, etc.) are late and rare, although they may be older in some non-Attic dialects, as suggested by adverbial forms like Cretan διπλει, τετραπλει (with an ending modelled on locative adverbs in -ει) or Corcyran διπλη, Cretan διπληι (from old instrumentals in *-ē) (see DELG s.v. ἁπλόος and Bile 1988, 212‒3). Meanwhile, the Homeric adverbs τριπλῇ and τετραπλῇ in the Iliad (C.1) may derive either from adjectives in -πλός (like their Cretan and Corcyran equivalents) or in -πλόος, in the latter case with the contraction seen in διπλῆν (Il. 10.34, Od. 19.226), while the uncontracted stem appears in διπλόος (Il. 4.133, 20.415). Homer does not attest ἁπλόος directly but has the nom. fem. plur. ἁπλοΐδες (Il. 24.230, Od. 24.276).

Contract compound adjectives typically received a recessive accent in Attic, so that, for example, gen. sing. εὐνόου ‘friendly’ gave εὔνου in place of the expected *εὐνοῦ after nom. sing. εὔνοος > εὔνους (contract adjectives of material underwent the opposite development, becoming perispomena in Attic; see Dieu 2022, 386‒7). The numeral adjectives in -πλόος, however, evaded this trend, retaining the perispomenon accent that resulted from the contraction, likely because their compound nature had become opaque (see Risch 1974, 177). Moreover, in the direct cases of the neuter plural Attic analogically reintroduced the timbre /a/ for reasons of morphological transparency, while retaining the long quantity and the circumflex accentuation, producing an ending -ᾶ that was not the regular phonological outcome of contraction (which would have been *-ῶ < -όα). In adjectives in -πλόος, this meant that the Attic -πλᾶ was distinct from non-Attic -πλά (from the short forms in -πλός) in quantity and accent type only. This implies that the Atticist prescriptions on the correct form of adjectives such as ἁπλᾶ can be seen as relating to orthoepyPronunciation, rather than orthography (see Vessella 2018, 250 on A.2).

In Attic prose, the multiplicative adjectives, such as those of material, appear only in contracted form, which is also the norm in tragedy (see, e.g., C.2, quoted by Hansen as the possible locus classicus of A.1), with the exception of an Aeschylean fragment (C.3), and in Aristophanes; Attic prose inscriptions also regularly show contraction (see Threatte 1996, 286‒9). Uncontracted forms are found in non-Attic literature, but often in addition to the contracted forms in varying proportions according to genre and author: Homer, as already mentioned above, has a mixture of contracted and uncontracted forms; Archilochus (fr. 196a.36) has a single instance of the uncontracted διπλόη (meanwhile, διπλοῖ occurs in the spuriously attributed fr. 330.2); Pindar employs the uncontracted forms exclusively (ἁπλόος N. 8.36, διπλόος N. 5.52, 10.89, I. 3/4.88, 5.17, τριπλόος O. 9.2); Herodotus primarily uses the contracted forms, with a small number of exceptions (see C.4, C.5, although the possibility of alterations in the course of the textual tradition cannot be excluded: on this point, see F.1); the Hippocratic corpus, whose Ionic-based language admits several Attic traits, presents a confused picture, although the contracted forms outnumber the uncontracted ones (see López Eire 1984, 345, and F.3 for the special case of substantivised διπλόη). Hellenistic literature shows a preponderance of contracted forms, albeit with more instances of uncontracted forms than are found in classical Attic authors. Gammage (2018, 140‒1) observes that, in this period, uncontracted forms are particularly frequent in epic poetry (Apollonius, Nicander, Oppian) and in medical writers (including Galen): this may be due to the respective influence of the Homeric and Hippocratic models rather than the influence of the koine. In fact, contraction is the norm in the New Testament (Blass, Debrunner 1976, 37) and in Roman and Byzantine papyri (Gignac 1981, 210).

Atticising writers align themselves with lexicographers’ prescriptions. Lucian, for example, uses only the contracted forms, including the adverbial τριπλῇ and τετραπλῇ (C.7, C.8), which are not attested in classical Attic but had a Homeric pedigree; for an uncontracted form in his imitation of literary Ionic (C.9), see F.2. Aelian (C.10) has another apparent exception (see F.3). Even Achilles Tatius, whose Atticism may be described as moderate, uses only the contracted forms (14x), with no manuscript variants attesting the uncontracted forms (see Gammage 2018, 148‒9). It should be noted that the multiplicative adjectives in -πλοῦς faced competition from forms in -πλάσιος, such as διπλάσιοςδιπλάσιος ‘double’, attested since Solon (fr. 13.73 West), or τριπλάσιοςτριπλάσιος ‘triple’ (Ar.+). These originally had a slightly different meaning – ‘twice/thrice the size’ rather than ‘double’, ‘triple’ (see van Beek 2022, 462‒3) – but offered the advantage of being morphologically more transparent and became ubiquitous in Attic prose, ultimately surviving into Modern Greek (see E.). Even later are forms such as διπλασίων (Arist.+), τριπλασίων (Archim.+), derived from the previous forms by the addition of the suffix of primary comparatives.

E. Byzantine and Modern Greek commentary

Multiplicative numerals occur very rarely in Byzantine Greek (see CGMEMG vol. 2, 1263) and show the contracted endings more frequently than the uncontracted endings, although some occasional shifts to other inflections (e.g., nom. masc. sing. ἁπλής, ἁπλύς) may attest to writers’ insecurity about the ‘proper’ ending (see CGMEMG vol. 2, 721‒7). Typically, however, these adjectives shifted to the simple o- and ā-stem declensions: this process’ culmination is seen in Modern Greek, in which the forms απλός, διπλός, and τριπλός are in use. However, a trace of the contracted declension survives in the synthetic gradation forms of απλός ‘plain, simple, single’, which still has a comparativeComparatives απλούστερος and superlativeSuperlatives απλούστατος (see LKN s.v.), taken from the classical ἁπλούστερος and ἁπλούστατος: the latter were ostensibly constructed with the extended suffixes -εστερο-, -εστατο-, but they were likely formed on the model of the contracted nominative in -οῦς (see Schwyzer 1939, 535). Multiplicatives of the type διπλάσιος, τριπλάσιος are also used in Modern Greek – not, however, as direct survivals but rather as borrowings from the learned tradition.

F. Commentary on individual texts and occurrences

(1)    Hdt. 3.42.2 (C.4)

The only multiplicative adjective that Herodotus uses – with 5 occurrences – is διπλόος. In the tradition of the Herodotean text, these primarily show the contracted endings (Untersteiner 1949, 86), with a couple of exceptions: one is at 6.104.1 (C.5), where the codex optimus A (= Laur. Plut. 70.3, 10th century CE) has διπλόον; consequently, all editors print the uncontracted form here. At 3.42.2 (C.4), the codd. have διπλέη, a form that is not only contradicted by διπλῆν, transmitted by all codd. at 5.90.1, but is also linguistically suspect, given that the uncontracted form should have been διπλόη. The unexpected ending -έη might perhaps be explained as analogous to the contract adjectives in -εος, or perhaps to πλέος πλέη πλέον ‘full’; otherwise, it is comparable with the Cretan διπλεία (see K.‒B. vol. 1, 378‒9). Smyth’s (1894, 357 n. 1) theory that ‘the adj. in the feminine when α follows is διπλεία = Ion. διπλῆ < διπλέη; when ο follows it is διπλόος’ seems unlikely and ill-founded, given that it is based on this single example. Nevertheless, only Rosén (1987)    and Medaglia (in Asheri, Medaglia, Fraschetta 1990) print the manuscript reading here; other editors normalise it to either διπλῆ or διπλή. Indeed, an additional complication is added by the oxytone accentuation transmitted by some codd. in the various Herodotean occurrences of this adjective; Hoffmann (1898, 299‒300) argued that these should be interpreted as instances of the ‘short’ form διπλός (see D.), and Hude (1927)    and Legrand (1932‒1954) printed them in their editions accordingly, while Stein (1869‒1871) and Wilson (2015)    choose the circumflex accent. However, Hoffmann’s theory is unlikely, given that the forms in -πλός appear to have been foreign to Attic-Ionic during the classical period. Here, as with other linguistic features, it is unlikely that the mix of different forms in Herodotus’ text reflects what the author himself wrote (a point already made by Merzorf 1875, 217: ‘Has omnes formas simul usurpasse Herodotum quis est, qui credat?’, even though his preferred solution of restoring the uncontracted διπλό- in all instances has not been followed in major modern editions), but it is difficult to determine which forms originated with the author and which were inserted at a later time. Leaving aside the problematic διπλέη, it is perhaps not coincidental that in cod. A only the acc. masc. sing. διπλόον appears uncontracted, while the feminine (acc. sing. -ῆν, plur. -ᾶς) and neuter forms (nom.-acc. plur. -ᾶ) are contracted. This distribution is reminiscent of Homer, who has the masc. διπλόος (2x) but the fem. διπλῆν (2x). In light of the widespread ancient notion that Homer’s and Herodotus’ languages shared a special affinity, it is possible that, at some point in the history of the text, a distribution of contracted and uncontracted forms that mirrored the Homeric distribution was introduced. On the problem of Herodotus’ ‘Homericness’, see, recently, Tribulato (2022) .

(2)    Luc. Syr.D. 31.1 (C.9)

The only Lucianic instance of uncontracted ἁπλόος is found in the treatise On the Syrian Goddess, a work of disputed authorship written in an imitation of Herodotean Ionic. In fact, Herodotus himself does not attest ἁπλόος or any other multiplicative adjective except for διπλόος, which appears mostly contracted in the tradition of his text (see F.1). Thus, this use of ἁπλόος belongs to the instances in which the phonology of On the Syrian Goddess ‘tries very hard to be Herodotean, but […] overdoes some features (such as uncontracted forms)’ (Lightfoot 2003, 110). Given that Lucian’s pastiche ‘is written in the light of grammatical prescriptions’ (Lightfoot 2003, 98) regarding the Ionic traits of Herodotus’ language, hyper-Ionisms such as this may reflect the notion – also seen in lexicographical sources such as A.1 – that multiplicative adjectives were categorically uncontracted in Ionic. It is also possible that Lucian was following the idea that in Herodotus (as in Homer), only the feminine forms of -πλόος adjectives were contracted (see F.1); unfortunately, this cannot be verified as no other instances of multiplicatives occur in this treatise.

(3)    Ael. NA 9.4.6 (C.10)

Aelian follows the Atticist prescriptions in using only the contracted forms of ἁπλόος and διπλόος. This isolated instance of uncontracted διπλόηδιπλόη in the On the Nature of Animals is not, pace Schmid (Atticismus vol. 3, 19), to be counted among the occasional uses of ‘open’ forms by Aelian, given that this is not the adjective but its substantivised feminine, meaning ‘fold’, used in a specialised anatomical sense to designate the ‘porous substance between the double plates in the bones in the skull’ (LSJ s.v.; borrowed in modern languages, cf. e.g. English diploe); this noun never appears contracted in Greek literature, likely because it was a technical borrowingTechnical language from Hippocratic Ionic. Additionally, uncontracted διπλόη could claim some Atticist credentials, given that Plato (C.6) had used it in the figurative sense ‘weak spot, flaw’, although like other Attic prose writers, he exclusively used the contracted forms of multiplicative adjectives.