Terms of Address on Right Periphery in Greek Tragedy

© sandra rodríguez-piedrabuena, 2021 | doi:10.1163/9789004440265_005 This is an open access chapter distributed under the terms of the CC BY-NC-ND 4.0 license. Terms of Address on Right Periphery in Greek Tragedy Sandra Rodríguez-Piedrabuena 1 Introduction The purpose of this paper is to analyse whether the position of a given term of address (TAd) bears any meaning for the understanding of the utterance and the context to which it relates (Dik 2007: 253).1 The idea that the placement of TAds can be meaningful has been considered under the framework of Conversation Analysis (Sidnell 2010: 261–264). As will be seen, CA has provided many cluesastohowthisresearchonGreekdramaistobeconducted.Thereareobvious limitations. For instance, stand-alone items are rarely found due to formal constraints,2 such as the metre (Shalev 2003: 359 n. 12). Accordingly, the focus of this paper will be on TAds at the end of a clause, i.e. on the right periphery (RP). RP-TAds are rare compared to the frequency with which TAds appear at the start of a clause (left periphery [LP]) or in the middle (M). LP-TAds (266) are, in fact, more than twice as frequent as RP-TAds (125) in the sample (see Section 2.1).The question arises as to whether there are differences concerning the distribution of RP-TAds and that of the other positions, and what implications this may have. The distribution of LP-, M- and RP-TAds from a selected sample of Greek drama will be examined in accordance with the same range of parameters. Few studies on TAds have been conducted and research on the placement of TAds in Ancient Greek remains scarce. Bassett (1934: 140 n. 1) expresses an interestinwhetherthevocativeappearsatthestart,inthemiddle, orattheend of a clause. No research on this issue, however, has come to fruition. Fraenkel (1965) focuses on M-TAds. The way in which he describes the position of TAds mainly relies on syntactic categories. Donati (2013: 74) only considers LP-TAds. Dickey (1996: 195) includes a section on the ‘position of addresses within the 1 The following abbreviations are used in this paper:TAd = term of address, LP = left periphery, M = term of address in the middle, RP = right periphery, IC = independent clause-TAd, DP = doubtful position, 2PP = second pair part of an adjacency pair, ECC = extra-clausal constituent, FTA = Face Threatening Act. 2 See Parkinson (1985: 37) and Haverkate (1984: 69) on the use of TAds as stand-alone items. terms of address on right periphery in greek tragedy 71 sentence’ (197). By ‘postpositive vocatives’, she means only those vocatives that do not head a sentence, but which are immediately preceded by another word or phrase.3 I subscribe to the approach by Dik (2007: 253–254), which is opposed to taking only the line and not the clause as a unit when analysing word order in Greek drama, thereby leading to the worn-out metri causa argument. I speak of left and right periphery because only free and not bound TAds are considered in this paper (following Dickey [1996: 5–6, 23], Parkinson [1985: 36], and Kleinknecht and Souza [2017: 260]). TAds are not bound if they are extraclausal constituents4 that relate to the corresponding clause while being syntactically independent. From a formal standpoint, these TAds are prototypically in the vocative, including the usage that is traditionally known as nominativus pro vocativo. Note that vocatives can hardly be considered as cases under the framework of traditional case grammar since they perform no syntactical function within the clause nor within any noun phrase of the proposition.5The following are examples of TAds on LP and RP: [1] LP: Euripides Alcestis 476–477 (Heracles to the chorus) ξένοι,Φεραίας τῆσδε κωμῆται χθονός, Ἄδμητον ἐν δόμοισιν ἆρα κιγχάνω; Strangers, dwellers in this land of Pherae, will I find Admetus at home?6 3 In a later work, Dickey (2010: 330–331) speaks of ‘expressive addresses’, but only from a semantical standpoint. Wendel (1929) is mainly a compilatory work. Scott (1903, 1904, 1905) explainstheuseandomissionof ὦ.Brioso-Sánchez(1971)questionedmanyof Scott’sfindings. Bassett (1934) focuses especially on omission. Donati (2013: 116–136) analyses ὦdiachronicallyandconsidersitacaseof grammaticalisation.McClure(1995:50–56)attemptstoestablish femalespeechinEuripidesbyconsideringTAds,amongotherfeatures.Dickey’s seminal work (1996: 20) generally leaves out tragedy. Rutherford (2012: 101–109) addresses the issue briefly but acutely. See Dickey (2002) on Latin forms of address and Brioso-Sánchez (1971: 35) for prior bibliography on the vocative. Berger (forth.) has recently devoted a study on a similar topic regarding TAds in Latin. According to this work, TAds are important elements of turndesign andtheyfunction onthreedifferentlevels:dialogical (turn-taking),interpersonal (stance-taking and politeness) and textual (discourse organisation). 4 ECCs as coined in Functional Grammar (Dik 1997: 379–407). See also Haverkate (1984: 67): ‘[…] vocatives typically fill the peripheral slots at the beginning or the end of the sentence’. 5 Haverkate (1984: 68). The matter remains controversial: see Donati (2013: 19–74, 95–96). 6 Unless stated otherwise, translations for Eur. Alc. are taken from Conacher (1988), for Eur. Andr. from Lloyd (1994), for Soph. Trach. from Jebb (21908). The remaining translations are my own. 72 rodríguez-piedrabuena [2] RP: Euripides Andromache 234–237 Ερ.τί σεμνομυθεῖς κεἰς ἀγῶν’ἔρχῃ λόγων, ὡς δὴ σὺ σώφρων,τἀμὰ δ’οὐχὶ σώφρονα; Αν.οὔκουν ἐφ’οἷς γε νῦν καθέστηκας λόγοις. Ερ.ὁ νοῦς ὁ σός μοι μὴ ξυνοικοίη,γύναι. He. Why do you preach and enter into a contest of words, alleging that you are virtuous and I am not virtuous? An. You are certainly not on the evidence of your recent words. He. May your ‘sense’ never be mine, Andromache. Combined with TAds in the middle of a clause, a further category should be added, labelled here as ‘Independent-Clause-TAd’ (IC-TAd). A typical example is given by what Lee (1997: 212) calls ‘extended apostrophe’, e.g. Eur. Ion 492– 506, which generally consists of a vocative in combination with a relative clause.7 Lastly, the position of a small number of cases is hard to establish, mainly due to the various possibilities of punctuation or due to asyndeta, i.e. when no particle or postpositive makes a clear split between utterances (Dik 2007: 11).This often happens when aTAd lies between two interrogatives, since it can be difficult to decide whether theTAd should be ascribed to the previous ortothefollowingquestion.Thepositioncan alsobe difficulttoestablishwhen subordinationisloose,asinthecaseof certaincorrelatives,andwheneverthere is no conjunction clearly establishing the ties between clauses.8 These TAds could correspond to ‘pivotal turn extensions’ according to Clayman (2012), where they simultaneously complete the prior syntactic unit and initiate the subsequent unit as a resource for extending turns at talk. Unmarked cases seem to be in LP since they often appear just as part of the mechanics of conversation, in that they are used ‘to get the attention of the addressee or to single him out of a group’.9 From the standpoint of Conversation Analysis (CA), this implies establishing contact or selecting the next addressee, respectively. As will be seen, this is not the case when it comes to RP-TAds. In principle, characters who use TAds are deferential or at least considerate towards their addressee, excluding, obviously, overt insults. Bassett (1934: 142) speaks of the ‘vocative of courtesy’, i.e. the vocative addressed in 7 See Kannicht (1969: ad 1451–1464). 8 Other cases include e.g. Eur. Alc. 510 Ἄδμητε,καὶ σὺ χαῖρε,Θεσσαλῶν ἄναξ. The phrase Θεσσαλῶν ἄναξ seems to be an apposition to Ἅδμητε rather than an independent RP-TAd. 9 Dickey (1996: 198–199); Parkinson (1985: 37). terms of address on right periphery in greek tragedy 73 the first speech of one character to another, which can be omitted in further turns. Other reasons for omission include that of the addressee being greatly inferior to the speaker in status, the urgency of the situation, and the identity of the addressee remaining unknown (Bassett 1934: 144–146; Dickey 1996: 193, 247). Consequently, the omission of TAds could be felt as being rude, abrupt, or unfriendly.10 Eustathius (806.27) describes Hector as rude when compared to Nestor because the latter begins a similar call for volunteers with ὦ φίλοι (Il. 20.204), whereas Hector (Il. 20.303) uses no TAd. I have looked into whether RP-TAds fulfil conversational functions, such as contact establishment, next addressee selection, and turn allocation, and comparedthem with LP-and M-TAds.I havealso examinedwhich types of TAds are more frequently employed in RP, which kinds of characters utter these TAds, and in connection with which speech acts they do so. These and other parameterscould eventuallyshedlighton whetherRP-TAdsbehaveinthe sameway as do LP- and M-TAds. 2 Method 2.1 Corpus I have selected the following plays for my research: Aesch. Supp.; Soph. El., Trach.; Eur. Alc., Heracl., Andr., Ion; Ar. Nub. The base texts are drawn from TLG online, although later editions and commentaries have been checked in every case.11 My preference is for the manuscript reading, unless it is impracticable. The following sample of TAds, classified in terms of position, emerges from this corpus: LP-TAds (266), M-TAds (272), RP-TAds (125), IC-TAds (51), uncertain (51), total (765). 2.2 Parameters Thedistribution of everyTAdhas been analysed in accordancewith thefollowing parameters: CA-functions (i.e. turn allocation and contact establishment), 10 Rutherford (2012: 101, 103–109, esp. 105); Dickey (1996: 193). The same applies to other languages, such as Egyptian Arabic (Parkinson 1985: 39). See also Venegas-Lagüéns (1991: 135). From a broader perspective, this could be analysed under the framework of the impoliteness strategy ‘withhold politeness’, consisting on ‘the absence of politeness work where it would be expected’ (Culpeper 1996: 357, 2016: 425). 11 The main editions and commentaries followed are: Johansen and Whittle (1980); Jebb (31894), Finglass (2007); Jebb (21908), Davies (1991); Dale (1954), Conacher (1988), Parker (2007); Wilkins (1993); Stevens (1971), Lloyd (1994); Lee (1997), Martin (2018); Starkie (1966), Dover (1968), Wilson (2007) respectively. 74 rodríguez-piedrabuena TAd types, speech act types, and non-redressive action. I consider as ‘nonredressive action’ every case where a TAd is uttered in the context of an FTA (Face Threatening Act) that is performed explicitly and as a direct speech act (bald-on-record FTAs). Threemoreparametersdeservetobementioned,namelydistance(Dfactor), i.e. the degree of familiarity between participants (Brown and Levinson 21987: 15–16, 74–82, 249–253), text type, and TAd-variation. Distance has turned out not to be a relevant parameter. Out of 765 TAds, there are 402 examples with a negative D factor (i.e. characters are close to each other), compared to 146 wheretheDfactorispositive(i.e.charactersarenot familiarwith eachother).12 This is, perhaps, due to the fact that characters in tragedy are often acquaintances or relatives.13 Similarly, TAds in any position are more common in dialogue than in long speeches (exceeding 10 lines), odes, and distichs by the chorus. Nonetheless, RP-TAds belong, more significantly (72.8%, 91), in dialogue than the TAds in the other positions do, since this figure exceeds both the percentage of TAds in dialogue of the whole sample (55.69%, 426), as well as the corresponding percentage of LP-TAds (53.01%, 141) and M-TAds (49.26%, 134). Finally, the variation of TAds within a speech or play raises an interesting question (Haverkate 1984: 71–76), but this, unfortunately, lies beyond the scope of this research. The use of ὦalso falls outside the scope, especially since the sample is too limited for the purpose. The matter is, moreover, truly controversial (Dickey 1996: 205). 2.2.1 TAd Types My TAd typology roughly follows that of Dickey (1996),14 although it is simplified in accordance with what was actually found in the sample. Wendel’s labelling is added in brackets and followed by an example: A. Names: a. Anthroponym (Eigenname). b. Patronymic, origin (Patronymika,Völkername). Patronymics convey formality and respect (Dickey 1996: 52–56, 90–98, 248). B. Age (Lebensalter) and kinship terms (e.g. πόσι). When παῖ is addressed to slaves it is probably derived from an age term and not from a kinship 12 The remaining 217 cases are ‘non-applicable’ since the degree of closeness of the participants is difficult to ascertain. 13 Dramatic plots seem to work more effectively when focused on family relationships: see Rutherford (2012: 103) and Aristotle (Poet. 1453b19–22). 14 See also Rutherford (2012: 101). terms of address on right periphery in greek tragedy 75 term (i.e. the address to one’s own child) (Dickey 1996: 258). In any case, παῖ, both when addressed to slaves and to one’s own offspring, is in this paper classified as an ‘age’ term (Rutherford 2012: 105 n. 107) in contrast to τέκνον, which is clearly a kinship term, and possibly more emotional. C. Indefinite addresses and ξένε. The standard TAd was the anthroponym, to which Dickey (1996: 46, 2010: 329) refers as the first name, ‘unless the addressee was a woman, a child, a slave, a foreigner, or a close relative of the speaker’. Deviations thereof were accordingly salient except for γύναι when a male addressed a female who was not his wife. Thus, γύναι was the standard TAd for respectable women in these contexts (Dickey 1996: 243–248; Wendel 1929: 60). D. Doubtful TAds (Beinamen).15 Among others, I include ὦ μέλε,ὦ τᾶν (see McClure 1995: 55) as well as ὦ δαιμόνιε. E. Invocations, e.g. ὦ λέκτρον. Dickey (1996: 24) only includes exclamations when they are actual addresses or prayers to the gods. For a more accuratecomparison,alltheexamplesof freeTAdsareconsideredinthispaper, including TAds to offstage characters, self-addressed TAds, and oaths. F. Titles (rank), e.g. ἄναξ,δέσποινα, and δμῶες. G. Endearment terms, e.g. ὦ φίλταθ’Ἕκτορ. The endearment term in the example is just ὦ φίλταθ’, but ὦ φίλταθ’Ἕκτορ is counted as unit and analysed as an endearment term. The same is true in H. H. Deferential terms, e.g. ὦ σοφώτατοι θεαταί. I. Derogatory terms, e.g. ὦ μῖσος. 2.2.2 Speech Act Types Busse (2006: 112) claims that ‘speech act theory is of particular importance’ in her study on the vocative constructions in Shakespeare. I have implemented a typology of speech acts based on Risselada (1993: 32–49). Speech acts are classified in accordance with two dimensions: orientation towards the speaker or towards the addressee;16 and whether ‘the content […] primarily refers to action, information or emotions’ (Risselada 1993: 35). The following diagram displaysthesetwocriteria,combinedinthe formof intersectinghorizontaland vertical axes respectively, and should be taken as a continuum since no clear- 15 Beiname (‘epithet’) is something different than ‘doubtful TAd’, but the content denoted by these two terms is roughly the same. 16 I have not considered orientation towards a definite or indefinite third person (Risselada 1993: 36), since this can be subsumed by the declarative type and, depending on the context, can be understood as an impersonalisation strategy under the framework of Politeness Theory (Brown and Levinson 21987: 191; House and Kasper 1981: 168). 76 rodríguez-piedrabuena figure 1 Typology of speech acts modified from Risselada 1993: 37 cutcategorisationisintended.Accordingly, anumberof borderlinespeech acts are represented in brackets (Figure 1). I have included several modifications (in italics), especially by further elaborating on the speech acts about emotions.17 Risselada (1993) focused on the subtype of directive speech acts, which are inturnclassified inaccordancewiththescheme below.18Therearetwocriteria: the degree to which the speech act is presented as compulsory,19 and whether the benefit is addressed towards the speaker or the addressee. Again, it is a continuum (Figure 2). I have also elaborated on the commissive speech acts by distinguishing between promises, threats, and warnings in accordance with two criteria, namely whether the speaker is the agent and whether a benefit or detriment is addressed more or less directly towards the recipient, i.e. the eventual perlocutionary effect. Thus, in promises, the speaker is the agent and there can be a benefit directly towards the addressee. In turn, the agent does not explicitly identify with the speaker in warnings and the damage is not explicitly oriented towards the addressee. In contrast, threats are detrimental for the addressee 17 While reproaches and laments relate to the past and the state of affairs is not likely to change (Risselada 1993: 78), speech acts about emotions which belong in a more non-factual dimension, e.g. wishes and complaints, lie between expressive and directive speech acts. 18 Hindelang (1978: 19, 119, 131, 141) also distinguishes between bindende and nicht-bindende Aufforderungen, which, in turn, are subclassified; see also Haverkate (1984: 19). 19 Thus,inrequests‘thedecisionwhetherornottorealizethecontentis left totheaddressee’ (Risselada 1993: 47; see also Berger on bindingness in this volume). terms of address on right periphery in greek tragedy 77 figure 2 Typology of directive speech acts Risselada 1993: 48 and the speaker is the agent. Finally, the concept of ‘metadirectivity’ should be considered for a better account of the speech act types: […] there is […] a metadirective element in every speech act type. This metadirective element results from the systematic perlocutionary effects […] this concept […] is needed for an adequate description of utterances which contain the imperative of a ‘perlocutionary’ verb (e.g. crede mihi ‘believe me’ or dic mihi ‘tell me’). These metadirective utterances […] are often misinterpreted as ordinary directives. They constitute, in fact, the functional counterparts of performative expressions like I tell you that or I ask you whether […]. Just as I promise that I will be here is not an assertive speech act, but an emphatically expressed commissive, believe methatIneversawhimdothatbeforeisnotadirectivebutanemphatically expressed assertion.20 An example of this in Greek is Soph. Trach. 598 (τί χρὴ ποεῖν;σήμαινε. ‘what are thy commands? Give me my charge.’). σήμαινε is not an order but a metadirective relating to the previous question. 2.2.3 CA-Functions (Turn Allocation and Contact Establishment) Here, I consider functions that are normally studied under the framework of Conversation Analysis (see Sacks et al. 1974: 701), to which Parkinson referred (1985: 36) in the following way: TAds ‘play an important role in some speech acts, and also an important part in the “mechanics” of conversation’. From a broader perspective, Geluykens (1987: 121) similarly considered two major functional domains (informational and interactional) when analysing ‘tails’ (right-dislocations) under the framework of Functional Grammar. He does in 20 Risselada (1993: 44–45; her emphases). 78 rodríguez-piedrabuena fact point out that ‘tails are an interactional process rather than a mere word order variation’ and that ‘tails are interactional patterns, and they have to be studied in relation to the turn-taking process’ (Geluykens 1987: 122, 127). In line with this proposal, the distribution of TAds in terms of position is analysed according to whether or not LP-, M- and RP-TAds fulfil a CA-function. A TAd performs a CA-function whenever it is used for the establishment of contact (see Mastronarde 1979), for turn assignment (when the speaker selects the next addressee or allocates the next turn in trilogues), and in second pair parts of an adjacency pair (2PP). On the other hand, TAds do not always perform one of these CA-functions, since they can be used even when the speaker has already established contact and/or has already selected the next addressee. These cases are labelled as interactive. Finally, TAds that do not perform CA- functions can also be non-interactive, whenever there is no direct interaction, e.g. invocations, TAds to offstage characters, and self-references. Thus, TAds performing a non-CA-function are labelled as either interactive or not interactive. 3 Data The following data lead to results in the form of tendencies, since no TAd distribution is absolutely neat. However, there are trends that are well supported by the data which point to differences, particularly between LP- and RP-TAds, and which support the idea that TAd position is not random but can convey a variety of nuances. For several parameters some subtypes are badly attested in the sample. For instance, there are only eight examples of suggestions and warnings; in all, there are four threats (Table 4). The low percentages of these cases probably have less to do with the TAd position than with the fact that the sample is small. For this reason, results in Section 4 are drawn only from those parameters that are sufficiently attested. 3.1 TAd Position and the Mechanics of Conversation There are significant differences in how the TAd position relates to the mechanics of conversation. LP-TAds are employed to establish contact between the characters (30.83%). If they are not uttered to fulfil turn-allocation functions, then they are not used in direct interactions (39.1%) but in invocations to the gods,tooffstagecharacters,orinself-references.M-TAdsaremoreorlessevenly distributedacrossthevariouscategories.Conversely, RP-TAdsarerarelyused to establish contact and are uttered mostly when they are not strictly needed for turn allocation, since contact is already established and/or the next addressee terms of address on right periphery in greek tragedy 85 I have deliberately excluded the distribution of TAd position by gender and status: tests of the type run in this paper can lead to misleading results in these categories, since they fail to account for the differences regarding whether female characters interact with other female characters or, e.g., with male servants, whether the women are older than their interlocutors, etc. Even more difficulties arise regarding status, as this can be a shifting attribute not only in each interaction but also throughout a play (e.g. E. Ion, Ar. Nub.). 4 Results The most significant findings are summarised in Figure 3. There is a hierarchy concerning the analysed parameters as well as correlations between said parameters. CA-functions are the defining parameter, as they constrain the use or omission of a given TAd in the first place.23 Unlike RP-TAds, LP-TAds can be unmarked in that they are often required simply for the establishment of contact or the allocation of the next turn. In turn, RP- TAds seldom play such a part in the mechanics of conversation. In this sense, RP-TAds tend to be marked since they are of a more expletive nature.24 Second, if we rule out the cases when LP- and RP-TAds are ‘mechanical’ and just compare the remaining examples where LP- and RP-TAds are both unconstrained by the mechanics of conversation, then we find that LP-TAds are mainly used for invocations, self-addresses, or addresses to off-stage characters, i.e. they are no longer directly interactive. In contrast, unconstrained RP-TAds are still directly interactive and convey nuances which range from self-initiated repair strategies to reinforcement of the previous clause.25 Self-initiated repair is linked to the choice of respectful or rank TAds, such as ἄναξ (see [5]); it takes place when the urgency of the situation pushes the content into the first position. Nevertheless, these cases remain rare. The other parameters also correlate: for instance, orders are more common with non-constrained RP-TAds and this in turn is related to more RP-TAds in the context of bald-on-record actions. TAd-types also correlate, as not many RP-TAds are deferential. 23 See Haverkate (1984: 69): ‘[…] the attention-getting function of the vocative needs to be set apart from its other functions, since it is not related to the internal structure of the speech act.’ 24 This is regardless of the fact that, from a traditional syntactical and semantic standpoint, any vocative is optional, see Haverkate (1984: 68). 25 See Haverkate (1984: 71–72), precisely with RP-TAds examples. 86 rodríguez-piedrabuena LP RP CA-Functions CONTACT. Omission is salient since abruptness is intended or the situation is urgent. NON-INTERACTIVE: invocations, offstage, and self-addresses INTERACTIVE, although contact is already established and the turn allocated. Consequently, omission is not salient. TAd type INVOCATIONS ENDEARMENTterms KINSHIP terms INDEFINITEADDRESSES *Deferential terms are remarkably rare, whereas derogatory terms are well represented. Speech act type EMOTIONS ACTIONS Directive speech acts REQUESTS(NON-BINDING) ORDERS(BINDING) Emotions LAMENTS,PRAYERS REPROACHES, COMPLAINTS, WISHES figure 3 Distribution patterns of LP- and RP-TAds terms of address on right periphery in greek tragedy 87 5 Discussion As seen, there are distribution patterns that argue for differences between LP- and RP-TAds. However, it is difficult to pin down a single function of RP-TAds as opposed to LP-TAds, especially since the context has a significant bearing on each specific case.26 Thus far, RP-TAds relate to bald-on-record orders, reproaches, and complaints; they are uttered especially by the young and the old. Most importantly, they are employed when there is no longer any need either for the establishment of contact or for the selection of the addressee. In this regard, while the omission of LP-TAds can be salient, that which is salient concerning the use of RP-TAds is precisely the fact that they are used (see Kleinknecht and Souza 2017: 261). Accordingly, there should be an input of some sort, even in the deviant cases where RP-TAds do not follow the abovementioned trends, e.g. the fewer cases of RP-TAds in the context of interrogatives, factual statements, prayers, and generic reflections. The observed distribution patterns (see Figure 3) can shed light on the input of RP-TAds in these non-prototypical cases, as will be presently seen. In view of the majority of the examples, it seems that RP-TAds, probably utteredwith distinctiveintonation,adda noteof insistence,impatience,worry, or unease with a varying degree of im/politeness depending on each case.27 This can be made explicit by the lexical items—a fact that is not very revealing as to the nuance of the RP-TAd itself, since the semantic elements already convey the character’s discomfort, e.g. Ar. Nub. 790, 854–858. More revealing are the cases in which querulousness, insistence, or unease can be hinted at by the context in which the RP-TAd plays a part, regardless of the lexical elements: [3] Aristophanes Clouds 80 Στρ.Φειδιππίδη,Φειδιππίδιον.Φε.τί,ὦ πάτερ; Str. Pheidippides! Pheidippides, my little one! Ph. What, o father? 26 See Dik (1997: 383); Sonnenhauser and Noel Aziz Hanna (2013: 14); Kleinknecht and Souza (2017: 262). 27 NB: in natural conversation there are other situations that can account for RP-TAds which areprobablynotfound in dramatictexts,atleastnotinthissample.Forinstance,ateacher could pose a question in a classroom and then select the next speaker, especially if there are no volunteers for the answer or because the speaker was not actually thinking about any specific addressee when uttering the question. Cf. Parkinson (1985: 37). 88 rodríguez-piedrabuena Strepsiades is insisting on calling his son and waking him up. Pheidippides wantsto continue sleeping, and he already asked his father to let him do so (Ar. Nub. 38). Consequently, he is annoyed and the way he replies, together with a certain intonation in antilabe, could convey his querulous attitude.28 LP- and M-TAds can also convey unease, insistence, querulousness, or reproach but the context needs to be semantically overt in this respect. On the other hand, when a character utters an insult in the form of a TAd, needs to establish contact as well, or to allocate a turn, said TAd moves mechanically intothe firstposition (i.e.LP).Intheabsence of suchconstraints,M-orRP-TAds arepreferred.This is because CA-functionsprevailin the case of LP-TAdsunlike RP-TAds. In contrast, RP-TAds can, as it were, ‘modalise’ the meaning of speech acts that are initially neutral, even in the absence of explicit lexical items or a clearer given context, as in the following examples, which are precisely nonprototypical considering the distribution patterns of RP-TAds: a. Generic reference. The RP-TAd in [4] makes it explicit that what Heracles is expressing is not just a general reflection but a reproach to Admetus for his deceitful behaviour: [4] Euripides Alcestis 1007–1008 φίλον πρὸς ἄνδρα χρὴ λέγειν ἐλευθέρως, Ἄδμητε. A man should speak frankly to his friends, Admetus. The entire clause is not as neutral as it would be without the RP-TAd, which is not needed for the selection of the addressee.29 b. Statements30. There are other instances in which there is no bald-on-record action but where the characters insist upon something. Admetus insists on hosting Heracles: 28 See also Ar. Nub. 86–93, 731–736, 776–784; Eur. Ion 999, 1211; Soph. El. 673–675. 29 See also Aesch. Supp. 99–102. 30 See Eur. Alc. 516–517, where there is a prima facie indifferent declarative sentence with an RP-TAd, which, if uttered with a certain intonation at the end, would probably increase Admetus’ discomfort with the situation later disclosed (Eur. Alc. 629). See also Ar. Nub. 1325; Eur. Alc. 529, 536–539; Andr. 1073–1074; Supp. 403; Ion 309, 645; Soph. El. 924–925 (ὦ τάλαινα), 1179–1182. terms of address on right periphery in greek tragedy 89 [5] Euripides Alcestis 536–539 Ηρ. … εἴθ’ηὕρομέν σ’, Ἄδμητε,μὴ λυπούμενον. Αδ.ὡς δὴ τί δράσων τόνδ’ὑπορράπτεις λόγον; Ηρ.ξένων πρὸς ἄλλων ἑστίαν πορεύσομαι. Αδ.οὐκ ἔστιν,ὦναξ· μὴ τοσόνδ’ἔλθοι κακόν. Her. … I wish, Admetus, that I had not found you in mourning. Ad. What action do you hint at by these words? Her. I’ll travel to the hearth of other hosts. Ad. My lord, that cannot be! May such a dreadful thing not come to pass. In view of examples like these, it is possible to distinguish between two types of contexts that would account for the placement of TAds at RP: (a) Contexts, often of urgency, in which the content pushes itself into first position and an RP-TAd is added as self-initiated repair. This would be a property shared with tail material from the standpoint of Functional Grammar.31 Thus, Admetus shifts from the first name (Ἡράκλεις), which he had been previously using (Eur. Alc. 517, 529), to a rank term (ἄναξ) in [5] in order to balance out and avoid a possible note of disrespect after his hasty reaction (οὐκ ἔστιν). Example [3] could also be interpreted along these lines. (b) Contexts in which the RP-TAd is added as a modifier of tone, generally as reinforcement of the previous utterance. In example [5], (a) would be possible: Admetus is self-repairing a potential FTA. This, however, does not seem to be exactly an FTA but instead a way of insisting that there is no problem if Heracles wants to stay at home. Alternatively, another interpretation is possible, which relates to a usage found in modern Spanish and which would be closer to (b). We can distinguish between two expressions with different punctuation and inflection: [6] No, señor, ‘no sir’ [7] **Noseñor / Noseñora 31 Geluykens (1987: 122); Dik (1997: 403). 90 rodríguez-piedrabuena I purposely misspelled [7] in order to represent the difference in intonation and punctuation. No. [6] corresponds to ‘no sir’; it is respectful and formal, and it normally appears in the second term of an adjacency pair. However, [7] is far from respectful. It appears in contexts where there is familiarity between the speakers and the speaker feels close enough to overtly disagree with the addressee. Otherwise, it is impolite. With noseñor the speaker seems to reinforce his stance by distancing himself from the addressee. In abstract terms, expressions ending in noseñor would result in ‘closeness + reinforcement of one’s own stance by generating distance’.32 This seems to match that which Admetus does in [5]: he had been addressing his friend by his first name (Eur. Alc. 517, 529), and now he insists on his invitation and disagrees with him, increasing the distance by the shift of the TAd type (from Ἡράκλεις to ἄναξ) in RP. Note that noseñor seems to have been pragmaticalised, if not already grammaticalised, in Spanish, as can be found in its invariable form (only masculine33) and in contexts other than direct dialogues, becoming an emphatic stance-taking expression: [8] J. Sámano, “Con lo que tenía, debía haber firmado a Karpin” (El País). Mi meta no es salir en EL PAÍS o en la primera del As o el Marca, no busco eso, no quiero notoriedad. Mucha gente piensa que como técnico de Primera mi ambición es ir al Madrid, al Barça, al Athletic … ¡Pues no señor! No es mi caso, no tengo ninguna necesidad. My goal is not to be in EL PAÍS nor in the front page of the AS or Marca.I’m not looking for that, I don’t seek fame. Many people think that my aspirationas a PremierLeague manageris tobecome the headcoachof Madrid, Barça, Athletic … Notat all,mister! It is not the case for me, I don’t need it. Another example in which an RP-TAd appears in the context of a declarative is Eur. Alc. 821 (Θε.γυνὴ μὲν οὖν ὄλωλεν Ἀδμήτου,ξένε ‘it is Admetus’ wife who is dead, stranger’ [my translation]). However, the utterance is not so much informative as tense. Thus, Dale states that ‘the Servant might intend 821 as a reproachful correction (μὲν οὖν)’.34 The indefinite RP-TAd in the last line of the 32 In this regard, see also Haverkate (1984: 77–78) on the ironic use of señorito and of usted as ‘typical of conflictive forms of verbal interaction’ by generating distance through the code-switching. 33 Cf. Kleinknecht and Souza (2017: 257, 259, 261, 280–281). 34 Dale (1954: 111). See also Parker (2007: 214). terms of address on right periphery in greek tragedy 91 stichomythia can also contribute towards conveying the servant’s uneasy attitude and his insistence on the fact that Heracles should leave now. This is comparable to an example by Haverkate (1984: 75–76), in which an RP-TAd appears ina similarcontext of a declarativefromwhich aninferencecanbe drawn: Está lloviendo a cántaros, Adela (‘It’s pouring with rain, Adela’). Haverkate relates this example to the implication ‘You cannot play tennis now’ and concludes: In the performance of indirect speech acts vocatives […] serve as a signal for the hearer to infer that a multiple speech act is performed by the speaker. c. Prayers. In Aesch. Supp. 811–816, there is an RP-TAd in the context of a prayer (γαιάοχε παγκρατὲς), which is already rare. This prayer is also unusual in that a present imperative is used instead of an expected aorist imperative. Note Johansen and Whittle’s comment (1980: ad 814–815): The shift from an aor. (811 ἔπιδε) to a pres. imperat. […] would indicate an emotional crescendo—an interpretation which is not improbable here but cannot be supported by other indications in the context. It seems that the need of ‘other indications’ is provided by the RP-TAd and that this prayer could be considered especially urgent and emotional not just becauseof the presentimperativebut alsobecause of the positionof theinvocation. Further, note that γαιάοχε παγκρατές in RP is a strong epithet which, added to the prayer, can work as a way of reinforcement. All these examples allow us to conclude that RP-TAdsareless often employed tofulfil CA-functions than to reinforce or change the illocutionary force of (indirect) speech acts (Busse 2006: 114–115). Both Bassett (1934: 146) and Dickey (1996) consider TAds as devices for characterisation. TAds can be a source for characterisation not only through semantic variation35 or omission (Bassett 1934: 145–146) but also through position, especially since RP-TAds are rare. Thus, there are (clusters of) RP-TAds instead of the expected LP- or M-TAds in the first words36 of troubled, hos- 35 Alreadyat the verybeginningof her monograph,Dickey(1996)considersTAd-variationas a device for characterisation, e.g. Hdt. 1.35.3, 41.1, 42.1, 45.2; Pl. Phd. 80b–81c, and Socrates inPl. Ap. (Dickey1996:179). See,likewise,Wendel (1929: 62–63)and Rutherford(2012:102– 109). 36 See Rutherford (2012: 101) on the relevance of the first words addressed to a character and of the use of TAds. 92 rodríguez-piedrabuena tile, and unsympathetic characters: Thanatos (Eur. Alc. 30), Alcmene (Heracl. 646–647[?],648,1050–1051), Peleus (Andr.590, 631),Pheres(Alc.614, 620),Hermione (Andr. 170, 234–237), Menelaus (Andr. 309–313), and Ion (Ion 237–240, 244, 255, 289, 309, 333, 339, 372, 379). These characters are either young or old and/or unsympathetic. In light of these examples, the correlation between RP- TAds and age seems to be of a secondary nature, in that RP-TAds do not depend on age in the way that, for instance, the use of the term παῖ does (which is used predominantly by significantly older characters). RP-TAds seem to occur particularly frequently with eagerness or excitement in behaviour. This in turn appears to be linked to a tendency to portray impetuous characters either as old or young. 6 Conclusion TAds are mostly uttered in upward interactions, i.e. from lower-status characters to higher-status characters. In this regard, their omission can be salient from the standpoint of politeness. Regardless of their position, TAds were generally considered as devices for the establishment of contact and turn allocation.These considerations are still valid with regard to LP- and M-TAds. In contrast, RP-TAds are generally uttered when contact has already been established orwhenthenextaddresseehasalreadybeenselected.Thus,theomission isnot salient,andthisleadstothequestionastowhytheyareusedatall.Ontheother hand, RP-TAds are commonly uttered by impetuous characters, e.g. characters who are hostile and/or young/old. In sum, LP- and RP-TAds are not randomly distributed. This leads us to the conclusion that not only the lexical meaning or the omission allow us to recognisetheinputof aspecificTAdbutalsoitsposition.LP-TAdsrelatetopoliteness, to CA functions, or to both. In either case, omissions are salient, in that they are either rude or urgent. RP-TAds are a way of conveying nuances that range from insistence, impatience, uneasiness, or worry to even greater feelings of hostility that challenge the addressee. A few instances could be interpreted as self-initiatedrepair: generally, when the content pushes itself into the first position and the RP-TAd is a respectful term. Thus, depending on the TAd type and the context, the degree of im/politeness varies.37 From a wider perspective, this is yet another example of how meaningful word order is, especially in 37 Similar conclusions were reached by Hill (1994: 325) on the familiariser dude in American English. terms of address on right periphery in greek tragedy 93 Ancient Greek, where word order is syntactically less constrained than in other languages, such as English. Acknowledgements I am most thankful to the reviewers, to Emilia Ruiz Yamuza, and to Professor Mastronarde for improving my work. I would like to thank Severin Spring and Mevina Puorger. Above all, I feel deeply grateful to Gunther Martin for his invaluable help and comments on previous versions of this draft. I am responsible for every remaining infelicity. References Bassett, S., (1934), ‘The Omission of the Vocative in Homeric Speeches’, American Journal of Philology 55, 140–152. Berger, Ł., (forthcoming), ‘Position and Function of Nominal Address in Roman Comedy’,in 20thInternationalColloquiumonLatinLinguistics(UniversidaddelasPalmas de Gran Canaria, 17th–21st June 2019). Brioso-Sánchez, M., (1971), ‘El vocativo y la interjección ὦ’, Habis 2, 35–48. Brown, P. & Levinson, S.C., (21987 [1978]), Politeness: some Universals in Language Usage, Cambridge. Busse, B., (2006), Vocative Constructions in the Language of Shakespeare, Amsterdam. Clayman, S.E., (2012), ‘Pivotal Address Terms’, Journal of Pragmatics 44, 1853–1867. Conacher, D.J., (1988), Euripides, Alcestis, Warminster. Culpeper, J., (1996), ‘Towards an Anatomy of Impoliteness’, Journal of Pragmatics 25, 349–367. Culpeper, J., (2016), ‘Impoliteness Strategies’, in A. Capone & J.L. Mey (eds.), Interdisciplinary Studies in Pragmatics, Culture and Society, Cham, 421–445. Dale, A.M., (1954), Euripides, Alcestis, Oxford. Davies, M., (1991), Sophocles, Trachiniae, Oxford. Dickey, E., (1996), Greek Forms of Address: from Herodotus to Lucian, Oxford. Dickey, E., (2002), Latin Forms of Address: from Plautus to Apuleius, Oxford/New York. Dickey, E., (2010), ‘Forms of Address and Markers of Status’, in E.J. Bakker (ed.), A Companion to the Ancient Greek Language, Malden/Oxford/Chichester, 327–337. Dik, H., (2007), Word Order in Greek Tragic Dialogue, Oxford. Dik, S.C., (1997), The Theory of Functional Grammar, vol. 2, Berlin/New York. Donati, M., (2013), Il vocativo nel processo identitario dell’interazione linguistica: prospettive dalle lingue classiche, Munich. 94 rodríguez-piedrabuena Dover, K.J., (1968), Aristophanes, Clouds, Oxford. Finglass, P.J., (2007), Sophocles, Electra, Cambridge. Fraenkel, E., (1965), Noch einmal Kolon und Satz, Munich. Geluykens, R., (1987), ‘Tails (Right-dislocations) as a Repair Mechanism in English Conversation’, in J. Nuyts & G. de Schutter (eds.), Getting One’s Words into Line: on Word Order and Functional Grammar, Dordrecht, 119–129. Haverkate, H., (1984), Speech Acts, Speakers and Hearers: Reference and Referential Strategies in Spanish, Amsterdam. Hill, R.A., (1994), ‘You’ve Come a Long Way, Dude: a History’, American Speech 69, 321– 327. Hindelang, G., (1978), Auffordern: Die Untertypen des Aufforderns und ihre sprachlichen Realisierungsformen, Göppingen. House, J. & Kasper, G., (1981), ‘Politeness Markers in English and German’, in F. Coulmas (ed.), Conversational Routine: Explorations in Standardized Communication Situations and Prepatterned Speech, The Hague/Paris, 157–185. Jebb, R.C., (31894 [1867]), Sophocles, The Electra, Cambridge. Jebb, R.C., (21908 [1892]), Sophocles, The Trachiniae, Cambridge. Johansen, H.F. & Whittle, E., (1980), Aeschylus,The Suppliants, 3 vols., Copenhagen. Kannicht, R., (1969), Euripides, Helena, Heidelberg. Kleinknecht,F.&Souza, M.,(2017),‘VocativesasaSourceCategory forPragmaticMarkers’, in C. Fedriani & A. Sansò (eds.), Pragmatic Markers, Discourse Markers and Modal Particles. New perspectives, Amsterdam, 257–287. Lee, K.H., (1997), Euripides, Ion, Warminster. Lloyd, M., (1994), Euripides, Andromache, Warminster. Martin, G., (2018), Euripides, Ion, Berlin/Boston. Mastronarde, D., (1979), Contact and Discontinuity: some Conventions of Speech and Action on the Greek Tragic Stage, Berkeley. McClure,L., (1995),‘Female Speech and CharacterizationinEuripides’,inF.DeMartino & A.H. Sommerstein (eds.), Lo spettacolo delle voci, Part 2, Bari, 35–60. Parker, L.P.E., (2007), Euripides, Alcestis, Oxford. Parkinson, D., (1985), Constructing the Social Context of Communication: Terms of Address in Egyptian Arabic, Berlin/New York. Risselada, R., (1993), Imperatives and Other Directive Expressions in Latin:a Study in the Pragmatics of a Dead Language, Amsterdam. Rutherford, R., (2012), Greek Tragic Style: Form, Language, and Interpretation, Cambridge. Sacks, H., Schegloff, E.A., & Jefferson, G., (1974), ‘A Simplest Systematics for the Organization of Turn-taking for Conversation’, Language 50, 696–735. Sámano, J., ‘Con lo que tenía, debía haber firmado a Karpin’, El País, retrieved from https://elpais.com/diario/2002/12/02/deportes/1038783614_850215.html [14/05/ 2019].