σμῆγμα, σμήχω, σμῆμα, σμάω
(Phryn. Ecl. 224, Moer. σ 35)

D. General commentary

Phrynichus (A.1, on which see also F.1) prescribes the deverbal noun σμῆμα and the aorist σμῆσαι and proscribes the deverbal noun σμῆγμα and the aorist σμήξαι. Moeris’ gloss (A.2) is concerned with the present participle σμώμενος, which Moeris recommends over σμηχόμενος. Thomas Magister (A.3) blends together Phrynichus’ and Moeris’ glosses. Phrynichus and Moeris condemn σμήχω as non-Attic in opposition to good Attic σμάω (see Lobeck 1820, 253–4 and Rutherford 1881, 321–3).

σμάω (GE s.v.: ‘to cleanse or clean by scrubbing, wipe, scrub’) and σμήχω belong together with a small group of verbs that have a dental or velar expansion to the stemVerbal expansions. The closest comparanda are τρύω > τρύχω and ψάω > ψήχω (see further K–B, § 69.4, Schwyzer 1939, 701–2, and Chantraine 1961, 227–8; according to Chantraine, the velar expansion indicates the completion of the process described by the verb). These extended forms were never a productive category in historical times, but are rather residual elements of an older phase of the language.

In the case of τρύω > τρύχω and ψάω > ψήχω, the form with the velar expansion spread rapidly and thoroughly, and it is well-attested in 5th- and 4th-century Attic writers at the expense of forms without the velar expansion. Thus, while τρύω is attested only once in Prometheus bound (27) and once in Plato (Lg. 761d.2), τρύχω is somewhat common in 5th- and 4th-century Attic, and in all literary genres. Even more strikingly, ψάω is attested only once in Hipponax (fr. 19.2 West) and once in Sophocles (Tr. 678). By contrast, ψήχω is used by Hipponax (fr. 84.17 West), Euripides (Hel. 1567), Aristophanes (fr. 43.2), and quite often by Xenophon, Plato, and Aristotle (yet another form, ψαίω, is adopted by Theophrastus; see Schwyzer 1939, 676 on -άω > -αίω). The case with σμάω is different from τρύω and ψάω, though, in that, despite an early Homeric occurrence (Od. 6.266 ἔσμηχενHom. Od. 6.266), and even though σμήχω, its compound forms, and derivatives are standard in the corpus Hippocraticum (where σμάω is unattested), the form with the velar expansion does not seem to have outplayed σμάω in classical Attic. In fact, in 5th- and 4th-century sources, the distribution of σμάω and σμήχω is as follows: σμάω and its compound forms occur four times in Herodotus (2.37.1 διασμῶντες, 3.148.1 ἐξέσμων, 4.73.2 σμησάμενοι, 9.110.2 σμᾶται), once in Cratinus (fr. 97 ἐπισμῇ), twice in Aristophanes (Th. 389 ἐπισμῇ, fr. 360.1 σμωμένην), once in Antiphanes (fr. 146.4 σμῆται), and once in Alexis (fr. 192.5 σμήσας), while σμήχω and its compound forms only occur twice in Aristophanes (C.1, C.2). Such a distribution gives the strong impression that σμήχω was the minority form in classical Attic. The available documentary evidence also confirms this, as proven by the compound form σμηματοθήκησμηματοθήκη (first attested in IG 2².1469.97–8 [ca. 320 BCE]) and σμηματοφορεῖονσμηματοφορεῖον (which, after Ar. fr. 16, is first attested in IG 2².1485.49 [end of the 4th century BCE]). The persistence of σμάω is further proven by the four occurrences in Herodotus, which provide a counterweight to the generalised use of σμήχω in the corpus Hippocraticum.

Even though τρύχω and ψήχω are evidence that the formation of σμήχω raises no objection per se (indeed, these forms do not attract opposition from the Atticists, see e.g. Phryn. PS 27.17 on ψήκτρα Phryn. PS 27.17), Phrynichus and Moeris concluded against the admissibility of σμήχω on account of its rarity in classical sources. How, then, do we explain the two occurrences of the -σμήχω compounds in Aristophanes (C.1, C.2)? Why is this insufficient proof that σμήχω too may be good Attic like τρύχω and ψήχω, which have never provoked complaint from Atticist lexicographers? Note that, besides the Aristophanic occurrences, σμήχω is also presupposed by the adjective ἄσμηκτος in Pherecrates (C.3) and by σμηκτρίς in Eupolis (C.4) and Cephisodorus (C.5). Should we regard Phrynichus’ and Moeris’ proscription of σμήχω as ill-founded? A more nuanced answer can be proposed. In Aristophanes, the preverb in the -σμήχω compounds has an intensifying function, and it might provide an element of Fachsprache that is used for comic purposes (see F.2, F.3). Concerning Pherecrates, as nominal derivatives with a suffix -τος are created only from σμήχω and never from σμάω, a form like ἄσμηκτος probably does not need any further justification. Additionally, this fragment is likely to be an example of philosophical parody, and so there might thus be a humorous note to ἄσμηκτος (see F.4). A similar explanation may also apply to σμηκτρίς in Eupolis and Cephisodorus (see F.5). One conclusion regarding the evidence for σμήχω in classical Attic could then be that while σμάω remained in use throughout the classical period, σμήχω was the productive form used for creating derivates such as verbal adjectives and adjectives in -τρις (C.3, C.4, C.5). This conclusion is confirmed if we compare the (later) forms σμήκτης and σμηκτικός, whereas there is no evidence of nominal or adjectival derivatives created from σμάω. Further, the compounds of σμήχω were perhaps perceived as having a technical ring to them (C.1 and C.2).

In post-classical times, too, σμάω enjoyed a different fate in comparison with τρύω and ψάω. While these forms are marginal at best or essentially disappear, to some extent σμάω stood its ground, despite the growing diffusion of σμήχω. Compounds of -σμάω are much better attested than the simple form, and their diffusion may also have contributed to σμάω standing firm. In the case of τρύω and ψάω, the former survives only in the perfect (which is also the best-attested form in classical authors), while the latter disappears, save for ψαίω in Porphyrius (on whose formation see above). The same is true for the deverbal nouns, that is, *ψῆμα and *ψῆσις from ψάω are unattested, and we only encounter ψῆγμα and ψῆξις from ψήχω (the form ψῆσις, which, according to the TLG, occurs in the Byzantine Euchologia, is a ghost word, see F.6). In contrast to these verbs, σμάω is still a relatively common form in Imperial Greek, even though σμήχω takes centre stage. For instance, σμήχω is the only form attested in post-classical poetry, possibly because of the Homeric antecedent (for which see above).

As far as prose texts are concerned, the growing popularity of σμήχω is probably connected with the typically technical use of this form. An influence of medical and technical texts such as the Hippocratic writings should probably be taken as a major factor in this. Indeed, the main areas in which we have evidence for the use of σμήχω and σμάω in post-classical prose is medical literature, which is unsurprising, given the semantics of these verbsMedical literature. Up until late antiquity, even though σμήχω is clearly more common, σμάω is still a relatively high-frequency form in Dioscorides, Galen, Oribasius, Paulus, and Aetius. Yet, the distribution of σμάω and σμήχω is not random, and while σμάω is mostly limited to the participle (in Dioscorides always, in Galen and Aetius mostly), σμήχω appears in a larger variety of forms. Outside medical literature, σμάω occurs in Diodorus Siculus (5.28.2, who does not use σμήχω) and in Lucian (C.6). The Lexiphanes passage is an evident parody of pretentious language, and σμάω may be part of this humorous strategy: not only is σμῆν a jussive infinitive and thus a syntactically marked construction (see jussive infinitive), but Lobeck (1820, 61) also noticed that the use of σμῆν rather than σμᾶν may be another element of Atticising language (see Phryn. Ecl. 39Phryn. Ecl. 39). σμήχω occurs in Strabo (3.4.16, 17.3.7), Josephus (BJ 2.123), and in a proverbial expression in Lucian (C.7), which may be a sign of a lower, or at least unmarked, linguistic register as opposed to the passage of Lexiphanes. These passages are evidence that after starting out in the language of medicine, σμήχω may then have become standard in Greek overall, while σμάω remained a minoritarian, and possibly more learned, choice of word.

The only deverbal noun in Post-classical Greek to be found in literary sources is the use of σμῆμα in Aristides (C.8). Keil (unlike previous editors) correctly promoted this reading to the text. Koechly’s conjectural σμητήρ (‘cleaner’) in pseudo-Manethonis’ Apotelesmatica (4.423)[Man.] 4.423, though reasonable in the context and well-formed, remains uncertain (see Massimilla 2020, 266–7). Apropos of this form, if one considers the parallel with σμηκτρίς, then σμηκτήρ could be given consideration in place of Koechly’s σμητήρ (see C.4 and F.5). Our picture of σμάω deverbal nominals is significantly enriched by evidence from documentary texts, spanning from the 3rd century BCE to the 4th century BCE. Besides the verb σμάω (C.9), there is good evidence for σμῆμα (C.10; add P.Palaurib. 28.4 [= TM 26157, provenance unknown, 1st century CE], SB 16.12375.73 [= TM 4111, Arsinoites (?), ca. 180 BCE], P.Abinn. 8.23–4 [= TM 10065, Philadelphia/Arsinoites (?), ca. 346 CE]) and for σμῆσις (C.11, a hapax unrecorded in modern lexica). The testimony from documentary sources is important and shows that in imperial times, σμάω and its derivatives lived on also in linguistic contexts where Atticising concerns did not play a role. Hesychius also attests the forms σμηματοδοκίςσμηματοδοκίς (‘box of unguents’) and σμηματοδόχος, -ονσμηματοδόχος (‘for holding unguents’), which he uses in the interpretamenta to (respectively) ρ 492Hsch. ρ 492 ῥύμμα (alongside σμηματοθήκη, for which see above) and λ 1153 λιτρίςHsch. λ 1153. According to GEW s.v. σμάω, the perfect participle προεζμησμένος also occurs in a 2nd-century CE papyrus, but no further indication is provided, and I have been unable to find this occurrence.

E. Byzantine and Modern Greek commentary

In Byzantine times, σμάω fell out of use. Apart from medical writers in early Byzantine times, such as Aetius and Paulus (see D.)Medical literature, all occurrences of σμάω and its derivatives appear in grammatical writings. For instance, late-antique Christian writers use only σμήχω. In contrast to the situation in antiquity, compound forms like ἀποσμάω and διασμάω enjoyed limited diffusion too, and were taken over by the compounds of σμήχω. Differently from σμάω, σμήχω remained in common use, and this explains why the corruption of σμάω into σμήχω and of σμῆμα into σμῆγμα is relatively common in Byzantine manuscripts of classical authors (examples in Lobeck 1820, 254 and Rutherford 1881, 321–2). Notice that σμήγμα (‘sebum’) also exists in Modern Greek scientific vocabulary, but it is an 18th-century semantic loan from French (see LKN s.v.).

In light of this, forms like σμάω and its derivatives piqued the interest of Eustathius, who comments that σμῆμα is a noteworthy formation (B.2) (Eustathius is making a passing remark on σμῆμα in a passage of Antiphanes, which he derives from Ath. 409c–d). On the contrary, when Eustathius mentions σμῆγμα, he does not feel any need to discuss its derivation (see in Il. 3.602.26 and 4.865.9), and σμῆγμα is the form Eustathius would use in his writings (see Exegesis in canonem iambicum Pentecostalem 1.152.1–2 Cesaretti–Ronchey). It should not be excluded, but can never be proven, that Eustathius may have had an Atticist source in mind when commenting on σμῆμα and its formation, although he does not actually say that this form is an Atticism. It remains a suggestive, though speculative, hypothesis that such a source might be Phrynichus’ gloss (A.1), which is the only ancient source (among those known to us) to comment on σμῆμα. But Eustathius may surely have had more erudite sources at his disposal than we do today.

F. Commentary on individual texts and occurrences

(1)    Phryn. Ecl. 224 (A.1)

Fischer (ad loc.) compares Phrynichus’ gloss (A.1) with Ecl. 162Phryn. Ecl. 162 (σαλπικτής is correct Attic, σαλπιστής is not) and Ecl. 211Phryn. Ecl. 162 (παῖσαι is correct Attic, παῖξαι is Doric). The aim of these two glosses is, however, quite different from the one under discussion: namely, the object of Ecl. 162 and 211 Phrynichus is the interplay and confusion between dental and velar stems in the verbal forms ending in -ίζω. Phrynichus recommends retaining the original velar stem in the forms σαλπικτής and σαλπίξαι against the later innovations σαλπιστής and σαλπίσαι (which presuppose a dental stem), while he also advises using παῖσαι, παίσατε, and συμπαίστης with the original dental stem over the Doric forms παῖξαι, παίξατε, and συμπαίκτης with an innovative velar stem.

(2)    Ar. Nu. 1237 (C.1)

Strepsiades pats the creditor’s belly and says that one could work it to produce good wineskins (see Dover 1968, 241). Rutherford (1881, 322) suggested restoring διασμηθείς in place of what he calls the ‘debased’ διασμηχθείς. Although it is true that copyists often trivialise σμάω into the more usual σμήχω (see E.), in this case there is no compelling reason for emending the received text (and indeed, no editor of Aristophanes ever changed the paradosis). Retaining the paradosis is even more advisable considering the Aristophanic parallel of ἐκσμήχω (C.2). The variant reading διαμιχθείς in Choeroboscus (in Theodos. GG 4,1.259.20Choerob. in Theodos. GG 4,1.259.20, where Hilgard unnecessarily restored διασμηχθείς; the passage is also given as Hdn. Περὶ κλίσεως ὀνομάτων GG 3,2.716.34Hdn. Περὶ κλίσεως ὀνομάτων GG 3,2.716.34, where Lentz correctly retains the transmitted διαμιχθείς) is best seen as an ancient corruption, possibly deriving from a faulty understanding of the process alluded to (one may compare schol. Ar. Nu. 1237 a, e, f).

Previous scholarship considers διασμηχθείς to be quite simply a tolerable exception to standard Attic usage (see Lobeck 1820, 254: ‘sed διασμηχθείς Nubb. 1237 excusari potest’). We might explore the possibility that διασμηχθείς is not a lexically neutral choice, but that it instead contributes an element of verbal expressivity used for comic purposes. For a start, διασμήχω has the intensifying preverb δια-, to which one may compare διασμῶντες in Hdt. 2.37.1Hdt. 2.37.1 ἐκ χαλκέων ποτηρίων πίνουσι, διασμῶντες ἀνὰ πᾶσαν ἡμέρην, οὐκ ὁ μέν, ὁ δ’ οὔ, ἀλλὰ πάντες (‘They drink from bronze cups, thoroughly polishing them every single day, and it is not like one does polish and one does not, but they all do’). In Herodotus, the compound verb indicates the priests’ painstaking care in polishing the bronze cups, and this is probably the same nuance that Aristophanes aims to evoke. Further, the fact that σμήχω and its compounds are standard in the corpus Hippocraticum might convey that these forms are an element of Fachsprache, which Aristophanes’ διασμήχω imitates in alluding to the process by which leather is worked to produce wineskins. Thus, διασμηχθείς could contribute a degree of subtle verbal irony. This suggested interpretation of διασμηχθείς may be paralleled in the case of Ar. fr. 656 (C.2, on which see F.3).

It should be pointed out, though, that I could not find any instance of the aorist passive of σμάω nor of its compounds. This form of the paradigm is also rare in the case of σμήχω and its compounds. In fact, the second instance, after the passage of Aristophanes, is in Dioscorides (προαποσμηχθείσας in 1.106.3). Hence, although the evidence is late, it is also possible that, not unlike Pherecrates’ ἄσμηκτος and γῆ σμηκτρίς in Eupolis and Cephisodorus, -σμήχω was the only suitable form for the aorist passive. Still, one could not extend this explanation to account for ἐκσμήχω in Ar. fr. 656 (C.2), and so the problem would remain partly unsolved anyway.

(3)    Ar. fr. 656 (C.2)

Aristophanes is making fun of Euripides’ excessive verbal subtlety, and ἐκσμήχω has the same function as ἐκσμάω does in a passage of Herodotus (3.148.1), where it describes polishing gold and silver cups (see further Taillardat 1965, 296 and Bagordo 2016, 215–8). As in the case of διασμηχθείς (see F.2), Aristophanes’ use of ἐκσμήχω gives a more expressive and slightly technical ring to the verb, which, together with the intensifying preverb ἐκ-, emphasises Euripides’ careful polishing of his characters’ words as a professional, technical procedure.

(4)    Pherecr. fr. 210 (C.3)

The only attested -τος derivatives are from σμήχω, never from σμάω. The only pre-Hellenistic parallel is νεόσμηκτοςνεόσμηκτος in Hom. Il. 13.342, which is then re-used in Hellenistic poetry, while ἁλίσμηκτοςἁλίσμηκτος in Lyc. Alex. 994 is new (on these forms see Pellettieri 2021, 35–6). Pherecrates’ ἄσμηκτοςἄσμηκτος therefore looks like the only available option rather than a deliberate choice, that is, a form like *ἄσμητος would be completely unparalleled. Given that νεόσμηκτος and ἁλίσμηκτος are all poetic, one may also consider the possibility that Pherecrates’ ἄσμηκτος may be an element of parodic language. The person whom Pherecrates is describing is presented as dirty and with disordered hair, both standard features in the comic presentation of philosophers (one may compare Ar. Nu. 836–8). Thus, it is not inconceivable that ἄσμηκτος, if a comic neologism, might be an Ionic-sounding form that mockingly depicts the philosopher using high-sounding forms to describe his unkemptness. The fact that ἄσμηκτος is juxtaposed with ἀπαράλεκτος, which also looks like a comic neologism, further corroborates this idea.

(5)    Eup. fr. 412 (C.4), Cephisod. fr. 6 (C.5)

The γῆ σμηκτρίςσμηκτρίς (‘cleansing earth’, calcium montmorillonite) is a mineral used for cleaning people and clothes (see Orth 2014, 340–2 and Olson 2014, 178). It is also known as γῆ Κιμωλία (see Ar. Ra. 710–3), after the island of Kimolos where it was found. The suffix -τρίς is common in adjectives indicating the function of an object (good exemplification in Orth 2014, 341 n. 561). Besides Eupolis and Cephisodorus, σμηκτρίς is only attested in the Hippocratic corpus. Since there is no evidence for adjectives in -τρίς based on σμάω or its compounds, we might infer that, as in the case of ἄσμηκτος, σμήχω was the only productive verbal basis for creating an adjective in -τρίς.

(6)    Euchologia 27.7.41 (Goar 1730, 509) (see D.)

According to the TLG, the form ψῆσις occurs in this passage concerning the preparation of the holy anointing oil. However, there can be no doubt that ἡ ψῆσις δευτέρα must be a mistake for ἥψησιςἥψησις δευτέρα: the previous sentence is ἡψεῖσθαι δι’ ὅλης τῆς ἡμέρας (‘let it cook for the whole day’), and the correct reading ἥψησις occurs in an alternative version of the same text, that is, Euchologia (cod. Barb. gr. 336) 140 Parenti–Velkovska. The TLG derives the mistaken ψῆσις from Goar’s editio princeps, but since Goar’s translation (‘cum secundo decoquitur’) does imply ἡ ψῆσις δευτέρα, we must conclude that ἡ ψῆσις was originally only a typo.