Cattafi ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 ISSN 2159-3159 Article copyright held by the author(s) and made available under the Creative Commons Attribution License CC-BY https://creativecommons.org/licenses/by/4.0/ The Choice between Finite and Non-finite Relative Clauses: Participial Strategy in Documentary Papyri Eleonora Cattafi HIS ARTICLE focuses on the features and uses of at- tributive participles in documentary papyri, in order to provide new evidence about the competition between participial structures and finite relative clauses in Ancient Greek. Participles constitute an extremely varied and productive category1 in Ancient Greek, which has caused Greeks to be known in the scholarship as φιλοµέτοχοι, lovers of participles.2 At the same time, participles are significantly involved in language change especially in the Post-Classical Greek period, due to the increasing presence of periphrastic constructions appearing with the participle.3 1 The category of participle is also per se a non-prototypical one, as it can be situated between the categories of adjectives and verbs. A discussion of the participle as a linguistic category is outside the scope of this article: for some discussions see A. Pompei, “Participles as a Non-prototypical Word Class,” in E. Crespo et al. (eds.), Words Classes and Related Topics in Ancient Greek (Lou- vain-la-Neuve 2006) 361–388, and “La categoria del participio tra grammati- ci antichi e teoria linguistica contemporanea: i tratti verbali,” in A. Bartolotta et al. (eds.), Perspectives on Language and Linguistics. Essays in Honour of Lucio Melazzo (Palermo 2021) 435–455. 2 See A. N. Jannaris, A Historical Greek Grammar, chiefly of the Attic Dialect (Lon- don 1897) 505; R. Kühner and B. Gerth, Ausführliche Grammatik der griechischen Sprache. Satzlehre (Hannover 1904) 98; E. Schwyzer and A. Debrunner, Grie- chische Grammatik II (Munich 1950) 386. 3 K. Bentein, Verbal Periphrasis in Ancient Greek Have- and Be-Constructions T ELEONORA CATTAFI 495 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 In traditional grammars, the three main uses of the Ancient Greek participle have been described as the adjectival (or attrib- utive) participle, the adverbial (or circumstantial) participle, and the verb complement (or supplementary) participle.4 While the forms and usages of participles in the latter two groups have been thoroughly investigated in different periods of the Greek language, the subcategory of attributive participles, which, as outlined in §2, corresponds to—and competes with—relative clauses, has received less scholarly attention. At the same time, studies on Greek relativisation have focused on finite relative clauses, while the participial forms have mostly been excluded. To the best of my knowledge, only two contributions have been dedicated specifically to Greek relative participles5 in their comparison with finite relative clauses. First, Mugler especially concentrates on Homer and Herodotus in order to understand the factors governing the preference for finite or participial rela- tive clauses.6 According to Mugler (160), the choice between the two options can be interpreted in Homer from a stylistic point of view, being based on the degree of “amplitude or concision” needed in the passage, while also taking into account metrical and euphonic reasons. In Herodotus, the participle is found to be especially frequent in the nominative, whereas nominatives are rare as antecedents of finite relative clauses; moreover, par- ticiples are avoided when their use could lead to an effect of heaviness (“lourdeur”), for instance in the repetition of parti- ciples in the oblique cases. (Oxford 2016), and E. Nardi, “The Spread of Participial Clauses in Biblical Greek: Semitic Interference and Multilingualism,” Journal of Historical Linguis- tics 14 (2024) 427–471. 4 See B. Fox, “The Discourse Function of the Participle in Ancient Greek,” in F. Klein-Andreu (ed.), Discourse Perspectives on Syntax (New York 1983) 23– 41. 5 The term Relativpartizip is used by C. Lehmann, Der Relativsatz (Tübingen 1984) 49–58, who attributes its origin to Jacobi. 6 C. Mugler, “Sur la concurrence entre le participe et la subordonnée rela- tive,” RPhil 16 (1942) 146–160. 496 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 More recently, Hayes has presented a comprehensive investi- gation about the choice of relative participles in New Testament Greek.7 The distribution of participle and finite clauses is here interpreted mainly in terms of restrictiveness. With respect to this distinction, “these two constructions could in no way be de- scribed as equivalent” (229): when both options are possible, i.e. in the subject function, attributive participles in the New Testa- ment are used to express a restrictive meaning, while finite relative clauses relate predominantly non-restrictively to their antecedents.8 It is not clear whether similar or additional factors to those presented by Mugler and Hayes are at play in other corpora and periods of the Greek language. Rijksbaron briefly mentions that participles standing in apposition to a noun as their modifier resemble digressive (i.e. non-restrictive) relative clauses,9 and in recent monographs about Ancient Greek relative clauses, par- ticipial clauses are touched upon only by Perna,10 while they are not considered in Probert,11 Faure,12 and López Romero.13 Also in the current most extensive investigation about relativisation 7 M. E. Hayes, An Analysis of the Attributive Participle and the Relative Clause in the Greek New Testament (New York 2018). 8 Exceptions to this general tendency are also documented by Hayes, who explains some of these in ad hoc ways. 9 A. Rijksbaron, The Syntax and Semantics of the Verb in Classical Greek: An Intro- duction (Amsterdam 1984) 132. 10 E. Perna, La frase relativa in greco antico. Analisi sintattica basata sul dialetto attico di Platone (diss. Univ. Padua 2013) 63–72. They are also briefly discussed in S. Fauconnier, “Internal and External Relative Clauses in Ancient Greek,” JGL 14 (2014) 141–162, at 156–160. 11 P. Probert, Early Greek Relative Clauses (Oxford 2015) 4. Probert mentions, however, that participial relative clauses, in early Greek, tend to be “very short and syntactically simple.” 12 R. Faure, The Syntax and Semantics of Wh-clauses in Classical Greek: Relatives, Interrogatives, Exclamatives (Leiden 2021). 13 M. López Romero, Oraciones de relativo en el drama ático clásico (diss. Univ. Sevilla 2023). ELEONORA CATTAFI 497 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 in Greek documentary papyri,14 participial relative clauses are not taken into account.15 As for Greek documentary papyri, the focus of this investigation, some research has been carried out on the syntactic variation between finite and non-finite clauses, in terms of both complementation16 and subordination,17 while the same comparison in terms of relativisation remains a missing piece. In other words, as stated by Probert, “a proper study of the competition between participial and finite relative clauses in Greek is needed, both for early Greek texts and for later peri- ods.”18 The aim of this article is therefore to better understand the relationship between participial and finite clauses in a so-far unexplored corpus of Greek documentary papyri, investigating both the linguistic and the extralinguistic variables involved in this variation. The corpus considered for the analysis includes around 4600 texts from Egypt dating from the first to the eighth century A.D., which are part of the corpus of the ERC project EVWRIT (Everyday Writing in Graeco-Roman and Late An- tique Egypt: A Socio-Semiotic Study of Communicative Varia- tion):19 more specifically, the corpus is composed of 1725 letters, 14 E. Kriki, Morfosintaktiki neoterismi sti glossa ton mi filologikon Papiron I Ana- forikes protasis (diss. Aristotle Univ. Thessaloniki 2013). 15 “η ανάλυση δεν επεκτείνεται για παράδειγµα στις αναφορικές µετοχές,” (Kriki, Morfosintaktiki neoterismi 7). 16 K. Bentein, “Finite vs. Non-Finite Complementation in Post-Classical and Early Byzantine Greek. Towards a Pragmatic Restructuring of the Com- plementation System?” JGL 17 (2017) 3–36. 17 K. Bentein et al., Subordination and Insubordination in Post-Classical Greek: From Syntax to Context (Berlin 2025). 18 Probert, Early Greek Relative Clauses 3. 19 For more information about the EVWRIT project see https:// www.evwrit.ugent.be/, and K. Bentein and Y. Amory, “Everyday Writing in Graeco-Roman and Late Antique Egypt: Outline of a New Research Pro- gramme,” Comparative Oriental Manuscript Studies Bulletin 5 (2019) 17–27. For more information about the digital infrastructure of the project and its future prospects, including the public release of tools and data, see K. Bentein, 498 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 557 petitions, and 2356 contracts, which have been linguistically annotated for the purposes of this investigation.20 I will take as a starting point the current debate around par- ticiples as a relativisation strategy (§1) and apply it to Greek doc- umentary papyri, establishing the degree of equivalence between participial and finite relative clauses (§2) and describing the types of participial relative clauses attested in papyri (§3), also in rela- tion to the presence and the position of the article (§4). I will then move to the comparison between finite and participial relative clauses in papyri, discussing the role played by case-marking (§5) and restrictiveness (§6), examining the contexts in which the two constructions are directly competing (§7) and looking at some extra-linguistic factors (§8), before reaching some conclusions (§9). 1. Typological overview about participles and relativisation Some preliminary aspects about participial relative clauses need to be addressed: first, whether it is possible to interpret the status of participles as equivalent to finite relative clauses and, second, if a comparison is possible, which specific types of par- ticiples correspond to them in Ancient Greek. In general terms, both relative clauses and participles have been regarded as adjectival constructions,21 but the syntactic “Socio-semiotic, Multimodal Annotation of Documentary Sources,” in N. Reggiani (ed.), Digital papyrology III (Berlin 2025) 221–256. 20 Finite relative clauses in the ERC corpus were manually annotated by me, while participial clauses were annotated by Morris Callens during his appointment as intern in the EVWRIT project, and subsequently revised by me and implemented with the annotation of case and further annotations about the definite article. While the identification of participles in the corpus was carried out automatically, the distinction between those involved in relativisation and those involved in complementation or subordination was made through close reading of the texts. Some issues related to their inter- pretation are discussed in §4. 21 See M. Haspelmath, “Passive Participles across Languages,” in B. A. Fox et al. (eds.), Voice: Form and Function (Amsterdam 1994) 151–178, at 152. Cf. also E. Benveniste’s definition of a relative clause as a “syntactic adjective”: Problèmes de linguistique générale (Paris 1966) 222. ELEONORA CATTAFI 499 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 structure of participial relative clauses is considered as “cross- linguistically impoverished in comparison to that of regular (or full) RCs.”22 This is due to various factors such as the absence of complementisers, the absence of overt subjects, and the absence of independent temporal reference, features which, in contrast, are present in finite relative clauses.23 These properties are, however, not always cross-linguistically valid: one notable exception is Ancient Greek, which marks verbal tense also on the participle. The syntactic restrictions on participles also exhibit strong cross-linguistic differences:24 for instance, the possibility of expressing only syntactic subjects holds for English participial relative clauses,25 but cannot be extended to other languages, such as Arabic. On a typological basis, Doron and Reintges therefore distinguish four types of languages in relation to the relativisation of participles:26 (i) the temporally invariable pattern which can relativise only the sub- ject, as in Hebrew, (ii) the temporally variable pattern which can relativise only the subject, as in Ancient Greek, (iii) the tem- porally invariable pattern which can relativise other syntactic positions alongside the subject, as in Arabic, and (iv) the tem- porally variable pattern which can relativise other syntactic positions alongside the subject, as in Older Egyptian, Turkish, and Tamil. The competition of participial structures with finite relative 22 A. Belikova, “Syntactically Challenged rather than Reduced: Participial Relatives Revisited,” in Actes du Congrès de l’ACL 2008 (https://cla-acl.ca/pdfs/ actes-2008/CLA2008_Belikova.pdf) 1–15, at 1. 23 Cf. Belikova, in Actes 1–15. For these reasons, relative participles are also called “reduced RCs.” 24 P. Sleeman, “Participial Relative Clauses,” Oxford Research Encyclopedia of Linguistics (Oxford 2017). 25 R. Quirk et al., A Comprehensive Grammar of the English Language (London 1985) 1263; Sleeman, Oxford Research Encyclopedia 17. 26 E. Doron and C. Reintges, “On the Syntax of Participial Modification,” unpublished paper (Hebrew Univ. Jerusalem 2006), https://pluto.huji.ac.il/ ~edit/papers/DORON%26REINTGES2.pdf. 500 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 clauses is also typologically attested.27 However, in light of the differences between the two structures and especially the re- strictions involving participles, scholars have expressed different opinions whether the use of participles should in fact be con- sidered a relativisation strategy. This is usually connected to the weight given to syntax in their description of relative clauses: for instance, the semantic per- spective chosen by Keenan and Comrie in their definition “justifies considering as RC’s certain constructions that would perhaps not have been so considered in traditional grammar,” which includes “the participial construction,”28 while Hendery considers participial relative clauses as relative clauses but “with a deranked verb form,”29 and de Vries is mostly negative about the status of non-finite relatives, since it would be hard to separate them from other attributive structures.30 However, according to de Vries, who also follows the examples of relative participles presented by Lehmann,31 some “true participial stra- 27 “Manchmal sind sie selbst in stark ausgebauten Konstruktionen attri- butiv verwendbar und machen dann den RSen Konkurrenz” (Lehmann, Der Relativsatz 157); “the complementary distribution of relative complementizers and participial modifiers has been extensively documented both within and across languages” (Doron and Reintges 2006, 6). 28 E. L. Keenan and B. Comrie, “Noun Phrase Accessibility and Universal Grammar,” Linguistic Inquiry 8 (1977) 63–99, at 64. 29 R. Hendery, Relative Clauses in Time and Space: A Case Study in the Methods of Diachronic Typology (Amsterdam 2012) 169–179. 30 M. de Vries, The Syntax of Relativization (diss. Univ. Amsterdam 2002) 58– 60. 31 It is particularly interesting to see what Lehmann, Der Relativsatz 157, writes in relation to Ancient Greek and Latin relative participles: “Partizipien bilden oft auch gar nicht in erster Linie Attribute, sondern nebensatzartige Adverbialien, sog. prädikative Partizipialien oder Participia Conjuncta. Dies ist gerade auch dann der Fall, wenn, wie etwa im Altgriechischen oder klas- sischen Latein das Partizip ziemlich frei erweiterbar ist. Hier machen die Partizipien den RSen keine Konkurrenz da sie syntaktisch nicht eigentlich Verbaladjektive, sondern eher Verbalsubstantive sind und wie Gerundialien fungieren.” ELEONORA CATTAFI 501 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 tegies” can be found, especially when there is the possibility of relativising functions other than the subject. As there is no typological consensus about the category of relative participles, a case-by-case approach to the different lan- guages seems the most viable option to evaluate the features of participial strategies.32 2. Equivalence between Greek participial and finite clauses The equivalence between attributive participles and relative clauses is a debated topic also in Ancient Greek. Although parti- cipial relative clauses present some restrictions in the language (see below), the scholarly consensus supports their assimilation to relative clauses: according to Perna,33 some occurrences of the Ancient Greek participles are “perfectly comparable to the Participle Strategy of Relativization attested in many other lan- guages.” As for Post-Classical Greek, Mollo and Tronci observe that, in the Septuagint, relative clauses are translated from Hebrew into Greek alternatively with an attributive participle and with a relative clause.34 Moreover, as pointed out by Hayes,35 some authors have explicitly asserted that attributive participles belong to the category of relative clauses, at least in New Testa- ment Greek,36 while in other authors the same concept is im- 32 For instance, Sleeman, in Oxford Research Encyclopedia, tackles this issue about participial relative clauses in English. See Doron and Reintges 2006 for further discussion about the equivalence of participles to finite relative clauses from a generativist perspective. 33 E. Perna, “Typology of Relative Clauses in Ancient Greek,” Studies in Greek Linguistics 33 (2013) 320–333, at 325. 34 P. Mollo and L. Tronci, “Lost (and Gained) in Translation: Syntactic Features of Relative Clauses in Septuagint Greek,” in Subordination and Insubor- dination 133–164, at 153. 35 Hayes, An Analysis of the Attributive Participle 4–9. 36 F. Blass, Grammar of New Testament Greek (London 1898) 242, E. Burton, Syntax of the Moods and Tenses in New Testament Greek3 (Edinburgh 1898) 164– 166, A. T. Robertson, A Grammar of New Testament Greek in Light of Historical 502 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 plied.37 Hayes does not however take into account Levinsohn, who states, on the contrary, that “English versions often use a relative pronoun to translate ‘substantive’ or articular participial clauses (…) However, they are not classified as relative clauses, since they are not linked to their main clause by a relative pro- noun.”38 In Greek documentary papyri, as already noted by Mayser,39 the possibility of using participles with the function of relative clauses is widely attested: (1) µαρτυροῦσιν̣ [οἱ θε]ο̣ὶ οἱ τὰ πάντα ὁ[ρ]ῶντες (IV A.D., P.Ammon. I 3.16) the gods who see all things are my witness40 In this article I follow Perna,41 Mayser,42 and most New Testa- ment scholars in considering participles such as οἱ ὁ[ρ]ῶντες in (1) as equivalent to relative clauses, based on their common func- tion. Additionally, the presence of coordinate constructions, as in (2) and (3), where a finite relative and a participle depend on Research (New York 1919) 1105, N. Turner, A Grammar of New Testament Greek II (Edinburgh 1963) 152, J. W. Voelz, Fundamental Greek Grammar2 (St. Louis 1993) 139, and C. C. Caragounis, The Development of Greek and the New Testament: Morphology, Syntax, Phonology, and Textual Transmission (Tübingen 2004) 175. 37 H. W. Smyth, A Greek Grammar for Colleges (New York 1920) §2488, J. L. Boyer, “The Classification of Participles: A Statistical Study,” Grace Theological Journal 5 (1984) 163–179, at 163–164, and D. B. Wallace, Greek Grammar be- yond the Basics: An Exegetical Syntax of the New Testament (Grand Rapids 1996) 617–618. 38 S. H. Levinsohn, Discourse Features of New Testament Greek: A Coursebook on the Information Structure of New Testament Greek (Dallas 2000) 192. I believe, how- ever, that such a position would be problematic also in the consideration of English relative clauses not explicitly linked by a relative pronoun to the main clause, such as “the woman you see.” 39 E. Mayser, Grammatik der griechischen Papyri aus der Ptolemäerzeit II (Berlin 1926) 55–57. 40 Translations of Greek examples are taken from the edition of the papyrus whenever possible. 41 Perna, La frase relativa 325. 42 Mayser, Grammatik der griechischen Papyri II 55. ELEONORA CATTAFI 503 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 the same head noun,43 further points towards the analogy of the two structures: (2) Λεωνίδης Πτολεµαίου Ἀλθαιεύς, ὅς ἐστι νῦν ἐπʼ Ἀλ̣εξανδρείας, ἔχων πατρῴαν ἡµῶν ὑποθή[κ]ην (…) ἀρούρας τεσσαράκοντα [ἓ]ξ̣ (I A.D., P.Berl.Moeller 2.4–9) Leonides, son of Ptolemy, Althaieus, who is now in Alexandria, and who has, as a pledge from our father, (…) forty-six arourae (3) τιµῆς̣ τῆς συνπεφωνηµένης ἀργυρίου δραχµῶν τριακοσίων εἴκοσι ἃς ἀπεσχηκέναι αὐτὴν παρὰ τοῦ Ἀµµωνίου διὰ χειρὸς ἐκ πλήρους οὔσας τοῦ Ἑρµοκλέους (II A.D., PSI XII 1228.13–15) for the agreed price of 320 silver drachmas, which she has re- ceived from Ammonius in full, which are of Hermocles At a morphosyntactic level, two main differences distinguish Greek finite relative clauses from participial relative clauses. First, participles agree with their head noun not only in gender and number, as relative pronouns do, but also in case.44 In (4), for instance, the participle λεγοµένων agrees with the head noun τῶν ἀνύδ̣ρων ἀρουρῶν not only in the masculine plural, but also assuming the same case marking, i.e. the genitive, whereas in (5) the pronoun αἳ agrees in gender and number with the head noun, but maintains its own case marking: (4) µέρ[ος] τρίτον τῶν ἀνύδ̣ρων ἀρουρῶν λεγοµένων Τανα̣ρῆς (VI A.D., P.Berl.Brash 17.14–16) one-third of the unwatered arouras called Tanares (5) τειµῆς τῆς συµπεφωνηµένης πρὸς ἀλλήλους ἀργυρίου δραχµῶν χειλ[ί]ων διακοσίων, αἳ προσεχώρησαν εἰς συνωνὴν πυροῦ (III A.D., P.Oxy. VI 909.18–21) at the price agreed upon between us of 1200 drachmae of silver, which sum was devoted to the purchase of wheat In documentary papyri, agreement does not always happen, as mismatches in case with the participle can be observed: 43 One might also say that the participle is a modification attached to the first relative clause. 44 This agreement can of course apply as well to a plural participle when it is referring to two singular head nouns, e.g. πρ̣[άγµ]ατα καὶ ζη̣µιώµατα πα̣ρὰ [τῶ]ν εἰρηµένω̣ν̣ [Θ]ε[οφίλ]ης τε [καὶ] Διοσκόρ(ου) (VI A.D., P.Cair.Masp. III 67279.23–24) 504 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 however, as also noticed by Manolessou,45 this lack of agreement should be rather attributed to the analogical reformations of endings in the Post-Classical period; in (6), for instance, the par- ticiple τ̣ι̣θέντες only apparently fails to agree with the head noun αὐτοὺς, with τ̣ι̣θέντες representing a new (analogical) accusative ending:46 (6) ἀπὸ δὲ τοῦ νῦν ἐπιτελ(έσειν) αὐτοὺς Πάπ̣ο̣ν̣ κ̣α̣ὶ̣ Π̣τ̣ο̣λ̣(εµαῖον) ̣ ̣ ̣ ̣ ἀµπ(ελ ) ̣ ̣ ̣ ̣ο̣ι̣ν̣ο̣π̣( ) ̣ ̣ ̣ ̣ τὰ τῆς φυ[τεί]ας τῶν δύο ἀρουρῶ(ν) τ̣ι̣θέντες τὸ φυτὸν ἐν̣ ̣ ̣ ̣ (I B.C., BGU IV 1122.14–15) from now on, Papos and Ptolemaios shall complete … the vine- yard (work) … wine (including?) the planting of the two arourae, putting the plant in … The second difference between finite and participial relative clauses lies in the fact that participles cannot express all functions of the Accessibility Hierarchy47 as finite relative clauses do,48 but can only express syntactic subjects. The second constraint can, in my view, be linked to the first: as the case of the participle is not assigned on the basis of its function in the relative clause, the possibility of expressing different syntactic functions could lead to ambiguity, as these would not be recoverable on the basis of the morphological information provided.49 45 I. Manolessou, “From Participles to Gerunds,” in M. Stavrou et al. (eds.), Advances in Greek Generative Syntax: In Honor of Dimitra Theophanopoulou-Kontou (Amsterdam 2005) 241–283, at 243–245. 46 Moreover, the agreement of the participle depends, in papyri, on factors such as the native language of the writers and their educational level. 47 Keenan and Comrie, Linguistic Inquiry 8 (1977) 63–99. 48 Not all the functions of the Accessibility Hierarchy are however equally attested in the papyri, cf. E. Cattafi, “The Accessibility Hierarchy in Post- Classical Greek between Syntax and Sociolinguistics,” in Subordination and Insubordination 209–236, at 219. 49 This is similar to what happens in Hebrew, but, for instance, different from Turkish, where the participle does not agree with the case of the head noun, and a suffix of case refers to its syntactic function in the relative clause. Although languages can tolerate a certain degree of ambiguity, when the par- ticiple can relativise multiple functions, strategies of disambiguation (e.g. con- nected to additional morphology or word order) are generally present in the language. ELEONORA CATTAFI 505 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 Therefore, the prototypical choice between the participial and the finite strategy is located in the subject position: for this reason, as was done by Hayes, the finite relative clauses con- sidered in the comparison will also belong to the same category of subject relative clauses.50 3. Types of participial relative clauses On a formal level, relative clauses can assume different syn- tactic configurations based on the presence (“headed relative clauses”) or the absence (“free relative clauses”) of a head noun.51 In Ancient Greek, the head noun can be external or internal to the relative clause in the finite constructions, while, in the case of non-finite constructions, the head noun can appear before or after the relative participle.52 Let us consider some examples stemming from the same docu- ment, a long petition in which Dionysia addresses the prefect Pomponius Faustianus about a monetary dispute with her father Chaeremon. 50 It is however worth noting that examples of non-subject finite relative clauses might also compete with participial clauses. For example, compare the participial clause in (7) τῆς ἐπιστολῆς τῆς γραφείσης σοι (“the letter written to you”) with the non-subject finite relative clause in ἀντιφώνησιν ἧς ἔγραψα αὐτῷ ἐπιστολῆς (“copy of the letter which I wrote to him,” P.Oxy. III 531.26– 27, II A.D.). A comprehensive comparison of the two structures including also these other functions would pose additional methodological challenges about which specific clauses can be considered actually equivalent. For instance, can only passive participial constructions with an expressed agent as in (7) strictly compete with finite relative clauses with an active verb? Unlike subject finite clauses, should we include in the dataset only non-finite relative clauses when a correspondent participial type is attested in the corpus? 51 See de Vries, The Syntax of Relativization 18–63, for a typological approach; A. Kirk, Word Order and Information Structure in New Testament Greek (Leiden 2012) 184–196; and Perna, Studies in Greek Linguistics 33 (2013) 320–333, for dis- cussion of these categories in Ancient Greek. 52 Also this aspect shows significant cross-linguistic variation: for instance, in Romance languages participial relative clauses can only be postnominal, see G. Cinque, The Syntax of Adjectives. A Comparative Study (Cambridge [Mass.] 2010) 70, quoted in Sleeman, Oxford Research Encyclopedia. 506 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 In postnominal participial relative clauses, the head noun ἐπι- στολῆς is placed before the participle γραφείσης: (7) τῆς ἐπιστολῆς τῆς γραφείσης σοι ὑπὸ τῆς στρατηγίας ἀντίγραφον (II A.D., P.Oxy. II 237.32) the copy of the letter written to you by the strategia In prenominal participial relative clauses,53 the head noun ἐπι- στολήν is placed after the participle γραφεῖσάν: (8) τὴν περὶ τούτων γραφεῖσάν σοι ὑπὸ τοῦ στρατηγοῦ ἐπιστολὴν (237.10) the letter written to you on the subject by the strategus In free participial relative clauses, the participle γραφεῖσι is not referring back to an antecedent but is employed as a substan- tivate participle:54 (9) φροντίσῃς ἀκόλουθα πρᾶξαι τοῖς π[ε]ρὶ το[ύ]του πρότερον γραφεῖσι ὑπὸ Λογγαίου Ῥούφο[υ] (237.34) see that the matter is decided in accordance with the previous in- structions of his excellency Longaeus Rufus The distribution of the different syntactic types is outlined in Table 1: in both participial and finite relative clauses, the headed types are always more frequent than the free counterparts. Externally-headed Internally-headed Free Finite RCs 395 (67%) 99 (17%) 94 (16%) Postnominal Prenominal Free Participial RCs 8014 (42%) 6980 (36%) 4227 (22%) TABLE 1: Types of finite subject relative clauses and participial relative clauses The proportions of the headed types appear quite different: in finite relative clauses, externally-headed relative clauses are here almost four times more numerous than the internally-headed counterparts, while a more balanced distribution can be found between postnominal and prenominal participial clauses. In other words, while the pronoun and the participial strategy can 53 See Fauconnier, JGL 14 (2014) 157–160, for a discussion of these con- structions in Classical Greek. 54 On this see W. A. Stevens, “On the Substantive Use of the Greek Par- ticiple,” TAPA 3 (1872) 45–55. ELEONORA CATTAFI 507 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 relativise all the different types of relative clauses, they show different preferences with respect to the syntactic configurations that they assume. 4. Presence and position of the article: contexts of ambiguity As noted by Mugler,55 the history of the attributive participle in Ancient Greek is strictly linked to that of the article. The range of variation of participial clauses involves not only the position of the participle with respect to its head noun, but also the presence56 and position of the definite article. Moreover, in addi- tion to the participial form, its head noun too can be accom- panied or not by the article.57 This leads to variation among different possible constructions: without the article preceding the head noun or the participle (10), with the article preceding the head noun but not the par- ticiple (11), with the article preceding the participle but not the head noun (12), and with the article preceding both the head noun and the participle (13): (10) ἀντίγραφα ἐπιστολῶν δύο γρ[α]φεισῶν µοι (II A.D., P.Achm. 8.7) copies of two letters written to me (11) ταύτην δέ σοι τὴν ἐντολὴν ὑπ’ ἐµοῦ ἰδίο[ι]ς γράµµασι γρ̣α[φε]ῖσαν ἐπιδίδωµι (IV A.D., P.Ammon. II 27.21–22) I deliver this order to you, written by me in my own hand (12) Ἡρᾶς Παυσιρίωνος ὁ καὶ τὸ σῶµα γράψας (II A.D., PSI IX 1030.27–28) Heras, son of Pausirion, who also wrote the contract (13) τὰ γράµµατα τὰ γραφέντα εἰς κοινὸν ἐµοί ται καὶ προέδρῳ (III A.D., P.Panop.Beatty 1.193) 55 Mugler, RPhil 16 (1942) 147. 56 For instance, in examples (6) and (7) the relative participle was accom- panied by the article, but this was not the case in (4). 57 The behaviour of the article in participial constructions in documentary papyri was investigated from a cartographic perspective by M. Callens, ‘Poly’- Definiteness. Windowing for Definiteness in Poly-Articulated Participial Constructions in Post-Classical Greek (thesis Univ. Gent 2023), who calculates that poly-articu- late constructions in postnominal participial relative clauses, i.e. constructions with a double article such as τῆς ἐπιστολῆς τῆς γραφείσης, amount to only 5% of the occurrences in the corpus. 508 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 the letters addressed jointly to me and to the president Perna58 provides some observations on the behaviour of the article in the participial strategy in relation to the type of relative clause.59 In particular, she states that all free relative clauses are accompanied by the article; in contrast, the article is never present in appositive relative clauses, which exhibit, therefore, the same structure as a conjunct participle. The scenario is more complex in the case of restrictive relatives, where the article is present in the majority of cases, but the participle can appear without the article both in postnominal and prenominal con- structions. The situation described by Perna for relative participles in Classical Greek is less clear-cut when we move to documentary papyri of the Post-Classical period. First, in free relative clauses, although the participle is usually accompanied by the article, there are exceptions: (14) [Βελλίκιος] Π̣ερεγρεῖνος στρατηγῷ καὶ προπολ[ιτευοµ]ένοις Ὀξυρυγχίτ[ου χ]α̣[ί]ρ̣ειν (III A.D., SB XVIII 13932.17)60 Bellicius Peregrinus to the strategus and leading citizens of the Oxyrhynchite, greetings Second, non-restrictive relative clauses can also be introduced by the article: (15) ἵνα ταῦτα καταβάλῃς ἐπὶ τὸν εἰρηµένον Φιλόξενον τὸν καὶ δεδω- κότα [τὰ α]ὐτὰ τρία νοµίσµατα τῷ εἰρηµένῳ Ἀπολλω̣ν̣[ίῳ] (V A.D., P.Oxy. XVI 1932.8–9) So that you may pay them to the aforementioned Philoxenus, who gave the three gold coins to the aforementioned Apollonius The lack of coherence in the use of the article with participles can also be seen in its variation within a text: in two almost identical examples of participial relative clauses in the same text, 58 Perna, La frase relativa 67. 59 The role of the article in Greek participial structures is also analysed by A. Pompei, Il participio come nome e come aggettivo. Il caso del greco antico (Rome 2013). 60 προπολ[ιτευοµ]ένοις can also be intended as a lexicalised participle, see below. ELEONORA CATTAFI 509 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 a marriage contract from the Late Antique period, the first structure (16a) involves a participial relative construction with the structure Noun-Article-Participle and the second (16b) has the structure Noun-Participle:61 (16a) Χριστοδότῃ τῇ σὺν θεῷ ἐσοµένῃ αὐτο[ῦ] συµβίῳ (VI–VII A.D., CPR I 30 fr.2.9–10) Christodote, who will be his spouse with God’s will (16b) Χριστοδότης σὺν θ(εῷ) ἐσωµένης νύµφησς Μεγάλου τοῦ λαµ- προ(τάτου) (fr.2.41–42) Christodote, who will be the spouse of the most illustrious Megas with God’s will The presence of the article is particularly relevant for the inter- pretation of the participle as relative or conjunct: when the par- ticiple is not in an attributive position, the level of ambiguity between the verbal and adjectival uses of the participle increases. An important context for ambiguity is the oscillation between a relative and a causal meaning, which is especially present in the illiteracy formula (17). In this formula, which can also be ex- pressed with variation of the participle such as αὐτοῦ ἀγραµµάτου ὄντος, one person is writing on behalf of another, who is illiterate; however, a causal meaning is present, since the person is also writing because the other cannot write. As this constitutes one of the most widespread formulas with the participle in documen- tary papyri, considering it as a relative participle involves the problem of formulaicity:62 61 The only different element is here represented by the participle being in the dative in (16a) and in the genitive (16b), although this can hardly be inter- preted as a reason of this variation. 62 Formulaic participial clauses in documentary papyri can be reduced to several main areas. In addition to the common illiteracy formula, formulae with relative participles include variants of temporal expressions such as τὸ ἐνεστὸς ἔτος and τοῦ διελθόντος ἔτους, references to letters such as τὰ παρόντα γράµµατα, τὰ γραφέντα, τῶν δοθέντων (µοι) βιβλίων, and formulae used in con- tracts such as χρόνον λογιζόµενον, φόρον νέον καθαρὸν κεκοσκινευµένον, and ὑπατείας τῆς προκειµένης. In turn, there are also finite formulaic relative clauses in a subject position: the ἐὰν ὦσι formula, τῷ δὲ ἀβρόχῳ, τὸ µὴ εἴη, σήµερον ἥτις ἐστὶ, and ἥτις ἐστιν κυρία. 510 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 (17) ἔγραψα ὑπὲρ αὐτῶν γράµµατα µὴ εἰδότων (IV A.D., BGU XII 2152.16) I have written for them who do not know letters Relative and causal readings are also found beyond this for- mula: for instance, in (18) the participle ποιήσοντι, depending on µοι, can be interpreted as corresponding to a relative clause (“who am very happy”) or to a causal subordinate (“because I will be very happy”). The presence of ὁς (i.e. ὡς) favours here the second interpretation: (18) καὶ σὲ δὲ προτρέποµαι ἐπιτρέπειν µοι περὶ ὧν βούλει ὁς ἥδιστα ποιήσοντι (II A.D., P.Brem. 21.9–11) and I also ask you to give me orders about whatever you want, as I will be very happy to carry them out Another context for ambiguity, this time between an adjectival and nominal use of the participle, is the expression of political roles (cf. also (13)),63 a large category of occurrences found mostly in administrative documents: here the participles express- ing a political role can assume the function of an apposition rather than a relative clause: (19) [Αὐρήλιος Ἑρµόφι]λος Ὡρίωνος κοσµητεύσας Ἑρµουπόλεως τῆς µεγάλης ἀρχαίας καὶ λαµπρᾶς καὶ σεµνοτάτης (III A.D., Stud.Pal. XX 54.1) Aurelius Hermophilus, son of Horion, ex-cosmete of the great ancient illustrious and most august city of Hermopolis Also in these cases, the participle can appear without the article (19) or with the article (20). On a synchronic level, the participle can be considered as fully lexicalised, as proved by the fact that it can be accompanied by an adjective: (20) Φλ(αουίῳ) Ἰωσῆφ τῷ αἰδεσίµῳ πολιτ(ευοµένῳ) καὶ ῥιπαρίῳ τῆς Ὀξυρ(υγχιτῶν) (V A.D., SB XVIII 13596.1–3) to Flavius Joseph the venerable councilor and riparius of the city of the Oxyrhynchites 63 Rijksbaron, The Syntax and Semantics 133, already pointed out that “some- times a participle can be analysed both as an apposition and a satellite.” He also says that participles that follow a noun usually denote a (semi)permanent feature of the antecedent. ELEONORA CATTAFI 511 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 The choice of a participle instead of a noun indicating the political role (e.g. κοσµητής) could have here been made because of the greater ductility of the participle, with the possibility of being inflected by tense: the mark of tense64 allows the writer to express synthetically whether the political position was exercised in the past, as in (19), in the present (20), or even in the future (21): (21) µετὰ τὴν ὑπατείαν Φλ(αουίου) Βιβιανοῦ τοῦ λαµπρ(οτάτου) καὶ τοῦ δηλωθησοµένου Φαµενὼθ κα (V A.D., SB XVIII 13596.1) the year after the consulship of Flavius Bibianus the most illus- trious and the one to be appointed, 21 Phamenoth The participial forms of the verb ἔχω in the papyri also are a source of ambiguity, as they are rarely accompanied by the article and can cover a range of different values, from the more verbal uses with the meaning of “keeping” (22) to more adjecti- val uses referring to a quality of the head noun (23): (22) ὁ προειρηµένος Παβᾶνος π̣[αρατυ]χὼν ἔχω̣ν̣ µετὰ χῖρας ξοΐδιον ἀνελεῖν τοὺς χοίρους βουλόµ̣[ενος] (IV A.D., P.Oxy. LIV 3771.9– 11) the said Pabanus happened by, having a chisel in his hands, and wanted to kill the pigs (23) ἔστιν δὲ σηµεῖον τῆς ταφῆς· σινδών ἐστιν ἐκτὸς ἔχων χρῆµα ῥόδι- νον (II A.D., Chr.Wilck. 499.9–12) there is an identification mark on the mummy: a linen shroud is on the outside, rose-colored Since it is especially difficult to distinguish between relative and non-relative uses of postnominal participles when these are not accompanied by the article,65 for the purposes of numerical comparison between finite and participial relative clauses I will provide in sections §5 and §6 figures both (i) for participial rela- tive clauses in general and (ii) for those in which the participle is preceded by the definite article. 64 Although in the case of participles, as noted by Pompei, Il participio 117– 120, the value expressed is more aspectual rather than temporal. 65 But as we have seen, the presence of the article is not consistent. 512 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 5. Comparison between participial and finite relative clauses: case marking As pointed out above, Mugler and Hayes have conducted investigations on the competition between participial and finite relative clauses: we can take these studies as a starting point and verify whether the same factors identified by them as affecting the choice between the two options might be relevant also for the analysis of documentary papyri. In order to better understand the relationship between participial and finite relative clauses, Mugler highlights three functional categories in which participles directly compete with finite relative clauses, which he attempts to explain:66 (i) In the function of ornamental or characteristic epithet, e.g. the formula Ὀλύµπια δώµατ ̓ ἔχοντες compared with a similar ex- pression introduced by (τ)οἱ and followed by a finite verb, the variation of case is the relevant feature, as some cases are ex- cluded for metrical reasons. (ii) In the function of defining persons, things, and institutions, e.g. the participle of ἀνάσσω compared with a relative pronoun and a finite form of the same verb, the finite relative clause shows a preference for the genitive singular and never occurs in the masculine nominative or feminine accusative form. (iii) In the function of evocation of past events and anticipation of future events (“rappel et anticipation”), e.g. ὃς ᾔδη τά τ᾽ ἐόντα τά τ᾽ ἐσσόµενα πρό τ᾽ ἐόντα (Il. 1.70, but mostly post-Homeric), a great increase of participles in historiographical texts creates an effect which can be attributed to a conscious decision whereby an apparent “substitution” of relative clauses with participles happens when the subject is relativised. In Herodotus in par- ticular, the choice can be linked to the nominative form of the participles, which is rare in the antecedent of the relative clause. Although metrical reasons are not pertinent to documentary papyri, the case distribution of the head noun, which in the participial strategy also corresponds to the case of the participle, 66 Mugler, R Phil 16 (1942) 146–160. ELEONORA CATTAFI 513 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 indicated by Mugler as a relevant parameter of variation, can be observed in the two constructions: Nominative Genitive Dative Accusative Finite RCs 34 (11%) 105 (33%) 52 (16%) 126 (40%) Partic. + Article 1440 (18%) 3242 (40.5%) 989 (12.5%) 2342 (29%) All partic. RCs 558 (31%) 764 (42%) 146 (8%) 344 (19%) TABLE 2: Case of head noun in finite and participial relative clauses In general terms, both finite and participial relative clauses are attested with head nouns marked by all cases. Although the differences encountered in papyri do not justify a choice between the two strategies based on case-marking, the rarity of the nominative as the head noun of finite relative clauses can be confirmed also for the papyrological examples; conversely, finite relative clauses more often refer back to accusative head nouns, whereas the participial strategy is more frequently marked by the genitive case. 6. Comparison between participial and finite relative clauses: restrictiveness Like finite relative clauses, relative participles can be either restrictive (24) or non-restrictive (25): (24) µετ̣έλαβον πα[ρ]ά τινων ἀπὸ Ἰβιῶ̣ν̣ος σή̣µερον ἐ̣λθόντω[ν] (II A.D., P.Giss.Apoll. 9.3–4) I have learned from men who arrived today from Ibion (25) Ἰ̣σ̣ί̣δ̣ωρος ὁ̣ κ̣[α]ὶ̣ [νῦν ἀπ]ο̣σταλεὶς πρὸς ὑπο̣δ̣ο̣χ̣ὴν τῶν καµη- λαρίων (IV A.D., CPR XXIII 28.4–5) Isidoros, who was once again sent with the task of receiving the camel-drivers According to Hayes,67 restrictiveness is the main factor affect- ing the choice of participial and finite relative clauses in New Testament Greek: in a subject relative clause with a restrictive meaning, the participle will always be preferred. A complemen- tary distribution can be observed overall in the New Testament, with participles primarily being used restrictively, viz. in 84% of the instances, and finite subject relative clauses primarily being used non-restrictively, in 80% of instances. 67 Hayes, An Analysis of the Attributive Participle 213–216. 514 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 As emerges from Table 3, however, participial clauses in papyri show a different situation: on the one hand, as in the New Testament, the pronoun strategy is mainly used to express non- restrictive meanings. On the other hand, in contrast to Hayes’ observations, the participial strategy is also mainly used non- restrictively, although with a more balanced proportion of re- strictiveness compared to finite relative clauses. Multiple reasons may account for a higher frequency of non-restrictive relative clauses (both finite and non-finite) in documentary papyri. For instance, 115 finite relative clauses with a subject function in the corpus are introduced by ὅστις, a pronoun normally used in non- restrictive contexts.68 At the same time, a common use of par- ticipial relative clauses in papyri is in the formula ὁ προκείµενος, “the aforesaid,” which necessarily carries a non-restrictive meaning, as it refers to someone previously identified.69 Restrictive Non-restrictive Finite RCs 70 (18%) 324 (82%) Participial + Article 2484 (32%) 5292 (68%) All participial RCs 655 (37%) 1124 (63%) TABLE 3: Restrictiveness in finite and in participial relative postnominal clauses Moreover, the presence of non-restrictive clauses can be con- nected to the features of the head noun: as pointed out by 68 This observation reflects a broader textual tendency for papyri rather than implying a specific order of choices in production: it highlights a re- curring correlation in usage, without assuming that writers first selected the pronoun and only then determined the (non-)restrictive interpretation. 69 A reviewer suggests that such examples might be pragmatically re- strictive, as they would serve to indicate to the reader that the referent, e.g. Philoxenus in (15), is the one already mentioned in the document, rather than anyone else with the same name. However, this is not the primary function of these expressions in papyrus texts: rather than disambiguating among mul- tiple possible referents, these participial forms (“aforesaid, aforementioned”) refer back to individuals introduced earlier in the same document (Philoxenus was mentioned in line 1 of the papyrus) and thus function as anaphoric markers, serving to maintain referential continuity within highly formulaic contexts. ELEONORA CATTAFI 515 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 Nikolaeva,70 restrictive relative clauses are generally headed by common nouns, non-restrictive relative clauses by personal pro- nouns and proper nouns. In the New Testament, the (few) oc- currences found by Hayes71 of non-restrictive participles were connected to proper names such as divine names, proper names of supernatural beings and humans, and personal pronouns. As documentary papyri are particularly filled with mentions of names (cf. (25)), including family members, wrongdoers, pos- sessor of items, and authoritative figures of various kinds, from which subject relative clauses might depend, their presence could account for the participial strategy often expressing non- restrictive meanings. 7. Contexts of direct comparison The approach of Mugler was focused on individuating specific functional categories in which to conduct a comparison between finite and participial relative clauses. In this section, I also look at some contexts where it is possible to establish a direct com- parison in documentary papyri, namely (i) the attribution of names with the verb καλέω “call,” and (ii) the use of participial and finite relative clauses in death notices with the verb τελευτάω “die,” and I discuss the possible motivating features of the varia- tion between the different constructions. (i) A first context where finite and participial options are found is the attribution of names with καλέω: this can be expressed either by a finite relative clause introduced by ὅς or by the par- ticiple of the verb. The number of attestations, however, is far from balanced, as the finite option is used only in two occur- rences (26, 27), while the participial clause (28) accounts for more than a hundred:72 70 I. Nikolaeva, in K. Brown (ed.), Encyclopedia of Language and Linguistics2 X (Amsterdam 2006) 501–508, at 502. 71 Hayes, An Analysis of the Attributive Participle 216. 72 A third option is also available, namely the construction with the relative pronoun and no verb expressed, which is mainly used for aliases. 516 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 (26) ἔγγιστα τῆς διώρυγος, ὃ καλεῖται Παλαι[ὰ] Διῶρυξ (II A.D., Chr.Mitt. 316.22) near the canal, which is called the Old Canal (27) ἐν ἀριθµῷ ἀνδρῶν οἳ καλοῦνται ναυτοκολυµβηταί (II A.D., P.Mich. III 174.4) a certain class of men who are called sailor-divers (28) [Ἰού]δας ὁ καλούµενος Κίµβε̣ρ̣ (II A.D., P.Yadin I 18.21) Judas, who is called Cimber In terms of text types, (26) is from a contract of will and (27) from a petition addressed to the prefect, while the participle is found in all text types,73 with a prevalence in contracts; moreover, no chronological evolution of a formula is at play, since the two options are attested in the same period. In this case, the finite relative clauses might have been discouraged by the necessity of using a non-active finite form of the verb, while the writers were more familiar with the forms of the participles. Moreover, the case of (27) is quite peculiar: the choice of the finite relative clause could have been influenced by the fact that it was not a proper name to be assigned, as in the prototypical examples with the participle, but the definition of a category. (ii) The second functional context74 involves the verb τελευτάω 73 I list here only some other examples of the participial constructions, illustrating its range of applications, from persons to places: ἀπὸ κώµης [κα- λο]υµέν[ης Ἀ]φρ[οδιτ]ῶ̣ν (VI A.D., P.Cair.Masp. I 67032.12–13), τοῦ γεωργί[ου] Βίκτ̣ο̣ρ̣ος καλου̣µένου Βησ̣έ̣λικος (VI A.D., P.Michael 45.19), τῆς οἰκίας καλου- µένης Πκολᾶ (III A.D., Chr.Mitt. 172.10), το̣ῦ̣ ἁγίου̣ εὐκτηρίου [ἄ]ββα Ἰωάννου βαπτιστο̣ῦ̣ κα̣λ̣[ο]υµένου [Π]εριπάτου (VI A.D., CPR IX 32.6–7). The produc- tivity of the constructions extends also to the verb ἐπικαλέω: δ]ιὰ Διονυσίου τοῦ καὶ Πλάντα ἐπικαλουµένου Λούπ̣ου Διονυσίου (II A.D., P.Ryl. II 157.2). 74 A third context where an alternation is observed, although more in- directly, is the domain of dates: the “present” day can be expressed with a participle as in τῆς ἐνεστώσης ἡµέρας (II A.D., BGU III 710.6), but also with a noun further specified with the indication of the date introduced by the feminine form of the pronoun: σήµερον ἥτις ἐστὶν ιθ̣ (II A.D., P.Oxy. LXVI 4542.5–6). On this note, R. Mascellari, La lingua delle petizioni nell’Egitto romano: evoluzione di lessico, formule e procedure dal 30 a.C. al 300 d.C. (Florence 2021) 333, reports an observation by Bureth, stating that the participle ἐνεστώς remains linked to years, but in the case of the month it is found only before A.D. 138. ELEONORA CATTAFI 517 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 in relative clauses, which usually appears in contracts and ac- companies a personal name or the indication of a family rela- tion, which constitutes the head noun of the clause. The verb can be introduced by the pronoun ὅς in a finite relative clause (29) or can appear in participial relative clauses (30): (29) ὑπὸ τοῦ ἀµφοτέρων πατρὸς Λυσιµάχου τοῦ Διδύµου ὃς τετελεύ- τηκεν (I A.D., P.Mich. V 262.14) by the father of both of them, Lysimachos, son of Didymos, who is dead75 (30) Κατιλλίου Οὐαριανοῦ πατρός µου τετελευτηκότος ἀδιαθέτου (III A.D., P.Oxy. IX 1201.20–21) to my father Catillius Varianus who has died intestate76 The reason motivating this variation is here mainly chrono- logical and situational: the finite forms are earlier, as they are attested only in the first century, and they all belong to two archives, that of Kronion and that of Harthotes. As for the in- stances with the participles, these date from at least two centuries later, with three examples from the participial group appearing in the archive of Ammonius. Interestingly, in a passage from the archive of Kronion, both realisations of the formula are given in a pleonastic construction, signaling that the option of the participle was already present at 75 The other occurrences of the finite construction, all I A.D., are: τοῦ Λυσᾶ ὑπαρχόντων, ὃς τετελεύτηκεν (P.Mich. V 326.61); τοῦ ἑτέρου αὐτοῦ ὑειοῦ Ψενκήβκιος, ὃς τετελεύτηκεν (V 322a.2–3); τοῦ ἑτέρου ἡµῶν ἀδελφοῦ Εὐτύχου δευτέρου ὃς τετελεύ⟨τη⟩κεν (V 305.18), τοῦ ἑτέρου ἡµῶν ἀδελφοῦ Εὐτύχου δευ- τέρου ὣς τετελεύ⟨τη⟩κεν (26), and τοῦ ἑτέρου ἡµῶν ἀδελφοῦ Εὐτύχου δευτέρου ὃς τετελεύτηκεν (21–22); τῷ ἐµοῦ Ἀρείου πατρεὶ Διδύ[µῳ] Εὐηνείου, ὃς ὁµοίως τετελεύτηκεν (P.Mil. I 7.15–16) and σ̣ὺ̣ν̣ τῶι ἑατῆς ἀδελφῶι Ἁρπατ̣ο̣θ̣οῆι, ὃς τετελεύτηκεν (13–15). 76 The other occurrences of the participial construction are, in headed relative clauses: τῷ προ]κ̣ειµένῳ [τετελευτηκ]ότι µου̣ υἱῷ (III A.D., SB XX 15188.13); µητρός µου?] τετελευτηκυίης ἐπʼ ἐµ[οὶ κλ]η̣[ρονόµωι] ἀκολούθως (III A.D., PSI X 1101.8–9); ὁ µειζότερος [υ]ἱ̣[ὸ]ς ἐµοῦ τῆς προκειµένης Τεεῦτος τελευτῶν Παπνούθιο̣ς̣ τ̣ὸ̣ ὄ̣νο̣µ̣α̣ (IV A.D., P.Lips. I 28.8–9); in free relative clauses: τοῦ τετελευτηκ⟨ότ⟩ος (II–III A.D., P.Oxy. LIX 3993.24); τ̣ῶ̣ι̣ τε̣τελευτη- κότι κα[τ’ ἀγχιστείαν διαφέρουσι (IV A.D., P.Ammon. II 38.26); τοῖς τῶι τελευ- τ(ήσαντι) κατ’ ἀ[γ]χ̣ι̣σ̣(τείαν) προσήκ[ο]υ̣[σ]ι̣ (IV A.D., P.Ammon. II 41.40) 518 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 the time when the finite relative clause was mostly used: (31) τῷ τετελευτηκότι ἠµῶν πάντων ἀδελφῶι Λυσᾶτι, ὃς τετελεύτηκεν (I A.D., P.Mich. V 326.47) the deceased brother of us all, Lysas, who is dead 8. Extralinguistic factors In this section we move to more general considerations, focus- ing on the role of two extralinguistic factors in the choice of the participle, namely formality and diachrony.77 As to stylistic characteristics, the use of participles instead of finite relative clauses can give a different appearance to a text.78 This has been observed also in modern texts: e.g., the use of par- ticipial relatives in academic discourse shows a significant differ- ence between the soft and hard sciences,79 with finite relative clauses being frequent in the humanities, and participial clauses increasing as one moves from the soft to the hard sciences. In Ancient Greek, some stylistic uses of participles have been noticed in ancient rhetoric;80 in terms of register, it was shown by Thoma in an analysis of participles in a fourteenth-century text that, if we compare high-register and low-register versions of the same text, the high versions show a high frequency in the use of participles, as opposed to finite verbs, while in the low versions the situation is reversed.81 Finally, Breuer has found 77 Moreover, there might be a role played by language contact in the distribution of participles: Nardi, Journal of Historical Sociolinguistics 14 (2024) 427–471, observes that, in papyri, one third of the participial clauses occur in documents with a Semitic background. 78 According to Lehmann, Der Relativsatz 157, brevity is one advantage of the relative participle compared to the relative clause, while a disadvantage is the lower degree of variability of the non-finite form. 79 B. Gray, Linguistic Variation in Research Articles: When Discipline Tells Only Part of the Story (Amsterdam 2015). 80 B. L. Gildersleeve, “On the Stylistic Effect of the Greek Participle,” AJP 9 (1888) 137–157. 81 C. Thoma, “Grammatical Metaphor and the Function of Participles in High-register Versions of the Life of Aesop,” in A. Georgakopoulou et al. (eds.), ELEONORA CATTAFI 519 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 that, in the Martyrium Pionii, dating from the third century, biblical verses containing attributive participles are quoted by different characters substituting the participial construction with a finite relative clause, possibly for the sake of stylistic variation.82 To evaluate the attestations of the two constructions in papyri on the basis of formality,83 we can look at all types of participial relative clauses and finite relative clauses in subject position: relative participles seem to be more used in formal than informal texts compared to finite relative clauses:84 Formal Informal Finite head-external 346 (88%) 46 (12%) Participial head-external 7524 (95%) 422 (5%) TABLE 4: Distribution of finite and participial relative clauses according to formality The possibility of a diffusion of participial relative construc- tions over time has been pointed out by López Romero in order to explain the reduced number of (finite) subject relative clauses Standard Languages and Language Standards: Greek, Past and Present (Farnham 2009) 137–158. Thoma however does not consider attributive participles, but rather adverbial participles. 82 J. Breuer, “Subordination and Competing Constructions in Greek Acts of Christian Martyrs: A Case Study,” in Subordination and Insubordination 277– 300, at 296–297. 83 The labels “formal” and “informal” are assigned manually on the EVWRIT database on the basis of the social distance perceived in each text. Although these labels do not necessarily coincide with “private” and “public” documents in papyrological contexts, private letters usually belong to the “in- formal” category, while official letters, contracts, and petitions usually belong to the “formal” category. 84 In turn, this might be motivated by the necessity of more conciseness in formal texts, or the possibility of more easily accumulating relative clauses in the case of participles; participial relative clauses could have also played a role of building the structure of the texts. 520 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 from Homer to Herodotus.85 The same tendency seems to con- tinue in the Post-Classical period, as can be observed in Table 5: if we compare the attestations of finite and participial relative clauses according to the chronology of the texts, we see that finite relative clauses with a subject function decrease between the Roman (I–III) and the Late Antique period (IV–VIII), while participial relative clauses increase, showing that diachrony seems to be one of the most important factors in this variation:86 Century I II III IV V VI VII VIII Finite 19 (22%) 79 (36%) 107 (27.5%) 47 (13%) 25 (9%) 101 (14.5%) 17 (10.5%) 0 (0%) Participial 68 (78%) 142 (64%) 282 (72.5%) 310 (87%) 252 (91%) 593 (85.5%) 145 (89.5%) 20 (100%) TABLE 5: Finite and participial relative clauses distributed by century87 Moreover, if we combine the distribution of finite and par- ticipial relative clauses, examining both the period and the text types, we can see that the use of participial relative clauses in- creases in all text types, but especially in contracts. This might be connected to the spread of certain formulae, such as ὁ προ- κείµενος, “the aforesaid,” in this text type: 85 López Romero, Oraciones de relativo 96: “El nominativo es el caso pre- ponderante salvo en Heródoto. Con todo, sufre un retroceso con el paso del tiempo: en Homero representa el 57,2% de los casos, en Heródoto el 24,8% y en Tucídides el 28,8%. Mugler (1939, pp. 32–33) explica el predominio en términos de accesibilidad: es el caso, dice, cuyas funciones son más sencillas y requieren menos esfuerzo. Al retroceso, en cambio, le atribuye una causa sintáctica: la proliferación de las construcciones de participio atributivo.” 86 Also in English academic writing, participles are said to be increasing diachronically at the expense of finite relative clauses: M. Hundt, D. Denison and G. Schneider, “Relative Complexity in Scientific Discourse,” English Language and Linguistics 16 (2012) 209–240. 87 For both finite and participial relative clauses, when it was not possible to assign a precise date to the papyrus but only a general indication of the period (e.g. 200–350 A.D.), I have picked the earlier of the centuries (thus III A.D.). ELEONORA CATTAFI 521 ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 Letters (I–III) Letters (IV–VIII) Petitions (I–III) Petitions (IV–VIII) Contracts (I–III) Contracts (IV–VIII) Finite 39 (31%) 37 (20%) 31 (31.5%) 21 (19%) 124 (29.5%) 128 (11%) Participial 86 (69%) 151 (80%) 67 (68.5%) 91 (81%) 298 (70.5%) 1062 (89%) TABLE 6: Finite and participial relative clauses distributed according to text type in the Roman vs Late Antique periods Finally, it should be pointed out that the rise of participial relative clauses does not correspond to the rise of the entire category of participles in the Post-Classical period: as noted by Manolessou,88 some types of participles, such as the predicative participle and some uses (final, conditional) of the circumstantial participle, are actually reduced both in the New Testament89 and in the papyri, while other uses, such as absolute participial constructions, significantly expand in the Koine.90 We can there- fore now add to the second group relative participles as well. 9. Conclusions This paper has investigated the participial strategy in a corpus of Greek documentary papyri. Starting from typological con- siderations and then moving to Ancient Greek, the degree of equivalence of the participial strategy with the pronominal strategy was determined. The types of relative clauses that can be expressed by the participle were described in syntactic terms, showing also that the pronominal strategy is more widespread in externally-headed constructions in documentary papyri. It was also seen that the presence and position of the article with attributive participles can considerably vary, and that it is applied less systematically than in Classical Greek: this causes ambiguity in relative participles especially when the participial 88 Manolessou, in Advances in Greek Generative Syntax 245. 89 Where, however, according to Boyer, Grace Theological Journal 5 (1984) 176, one word every twenty is still a participle. 90 For the use of the participle with the accusative in complementation in the papyri, see Bentein, JGL 17 (2017) 3–36. 522 FINITE AND NON-FINITE RELATIVE CLAUSES ————— Greek, Roman, and Byzantine Studies 65 (2025) 494–522 forms are not preceded by the article, in expressions such as the illiteracy formula or the stating of political roles. Following the work of Mugler and Hayes, parameters such as the case-marking of the antecedent and the restrictiveness of the relative clause were investigated, showing that an interpretation of the choice between the two constructions only according to these features cannot be supported in the case of documentary papyri. Both participial and finite relative clauses in subject posi- tion are particularly used with a non-restrictive function, which can be linked to the character of the head noun. Some functional contexts where the two constructions directly compete were observed in the papyri, and extralinguistic factors were also taken into account: in stylistic terms, participial rela- tive clauses are more used in formal than in informal texts. Finally, an important element turns out to be the diachrony of the attestations, both in a specific formula and in general terms, with participial relative clauses increasing from the Roman to the Late Antique period, especially in contracts, in contrast to finite relative clauses. More research is needed in order to under- stand whether these findings are relevant only for documentary papyri or can be extended to other stages of the language, and how the situation evolves in the subsequent periods of Greek.91 August, 2025 Ghent University Blandijnberg 2 9000 Ghent, Belgium eleonora.cattafi@ugent.be 91 I wish to thank Morris Callens for his work on the database annotations, Anna Pompei for the discussions about the category of the participle, Klaas Bentein for giving feedback to the preliminary versions of this manuscript, and the anonymous reviewers for their suggestions. Part of this research was carried out thanks to a grant of the Academia Belgica (Stipendium Aca- demiae Belgicae).