Citizens of the Wor(l)d? Metaphor and the Politics of Roman Language

Abstract

In his discussion of Roman wind-names, Seneca the Younger employs a striking metaphor to describe the integration of the name of the south-east wind, Eurus, into Latin. The name Eurus, Seneca says, has been ‘granted citizenship’. This is one of six instances of the metaphor of ‘granting citizenship to words’ in surviving ancient texts. In this article, I use this metaphor as an entry-point to reconsider the importance of citizenship and language to ancient conceptions of Roman identity and status. The metaphor is revelatory of ancient thinking about what citizenship meant, what it depended on, and to whom and on whose authority it should be granted, questions that became urgent as citizenship spread across the Empire. Different versions of the metaphor offer tellingly divergent views of citizenship and of language. These reflect the tensions between origin and culture, inclusion and exclusion, cosmopolitanism and nativism, in contemporary notions of what it meant to be or belong as Roman.

In Book 5 of his Quaestiones naturales, Seneca the Younger discusses the relationship between Greek and Roman names for the winds. He employs a striking metaphor to explain the integration of the name for the south-east wind, Eurus, into Latin. The name Eurus, Seneca says, has ‘now been granted citizenship and does not enter our language as a foreigner’ (‘eurus iam ciuitate donatus est et nostro sermoni non tamquam alienus interuenit’).¹ This metaphor of ‘granting citizenship to words’ is not unique: the same metaphor appears in five other passages by four other authors, including one by Dio in Greek.² Commenting on Aulus Gellius’ use of this metaphor, Simon Swain dismissed it as a ‘cliché’.³ In this article, however, I argue that the metaphor is more productive than Swain allows for, that it is highly revealing of ancient conceptions of both language and citizenship, and that it offers a viewpoint onto the contemporary politics and hierarchies of being Roman.

In describing the entry of new words into Latin in terms of granting them citizenship, the metaphor ties together legal-political and linguistic ways of belonging, and draws a conceptual connection between language and wider markers of Romanness. It sheds light not only on ancient thinking about linguistic change and integration, but also on ancient thinking about citizenship itself. Citizenship is usually regarded as the clearest and most secure marker of Roman identity. I argue, however, that the metaphor provides a way into destabilising this view. It reflects the contested value and place of citizenship as citizenship spread across the Empire. It reveals contemporary uncertainty about how citizenship should relate to and rank against other markers of Romanness, notably language and origin, and to whom and on whose authority citizenship should be granted. This is especially so since the terms of the metaphor are not fixed; its variations reflect the parameters of live contemporary debates about language, citizenship and belonging.

I begin (Section I) by setting out the evidence for the metaphor and my approach to it. I then (II) interrogate the contexts of the metaphor. It appears in passages that, though centred on discussions of language, are politically engaged, and is part of broader conceptual connections between language and politics. The metaphor is the product of a period when the concepts of Roman citizenship and Roman language, and the relationship between the two, were being worked out. Having thus situated the metaphor, I use it to explore three sets of questions and debates about both language and citizenship; as the metaphor itself indicates, these themes are closely entwined. I look first (III) at the process of making citizens and the agency or authority behind decisions about citizenship and language; then (IV) at the value and meaning of citizenship once granted; then finally (V) at the relationship between citizenship and Latin right across the process of enfranchisement.

The metaphor is revealing of ancient debates about citizenship, language and identity. These debates were tied to wider uncertainties about issues including migration, integration and imperial control. To claim to ‘make a word a citizen’ was not an empty turn of phrase, because it was not universally agreed what being a citizen meant or who had the right to decide. Employing this metaphor meant engaging with the real, and contested, contemporary world of the citizenship.

I MAKING METAPHOR WORK

There are six extant examples of the specific metaphor of granting citizenship to words, in five different authors, four writing in Latin (Seneca, Suetonius, Quintilian, Aulus Gellius) and one writing in Greek (Dio).⁴ The two Senecan instances are, as I will discuss, importantly distinct from one another. All six passages are given below:

ab oriente hiberno eurus exit, quem nostri uocavere uulturnum — T. Livius hoc illum nomine appellat in illa pugna Romanis parum prospera in qua Hannibal et contra solem orientem exercitum nostrum et contra uentum constitutum uenti adiutorio ac fulgoris praestringentis oculos hostium uicit; Varro quoque hoc nomen usurpat — sed et eurus iam ciuitate donatus est et nostro sermoni non tamquam alienus interuenit.

The wind Eurus comes from east-south-east.⁵ We used to call it ‘Vulturnus’: Livy calls it this in [his description of] that unfortunate battle for the Romans in which Hannibal, assisted by the wind and the glare that dazzled his enemies’ eyes, defeated our army which was drawn up facing the rising sun and against the wind. Varro also uses that name, but now Eurus too has been granted citizenship and does not enter our speech as if it were a foreigner.

Seneca, Quaestiones naturales 5.16.4–6

nobis uidetur observatio collegisse et rerum saepe factarum inter se conlatio, per analogian nostri intellectum et honestum et bonum iudicant. Hoc uerbum cum Latini grammatici ciuitate donauerint, ego damnandum non puto, puto in ciuitatem suam redigendum. utar ergo illo non tantum tamquam recepto, sed tamquam usitato. We believe that it is inference due to observation, a comparison of events that has occurred frequently; our philosophical school holds that the honourable and the good have been comprehended by analogy. Since Latin grammarians have granted this word [‘analogy’] citizenship, I do not think that it ought to be condemned, I think its citizenship should be restored. I shall, therefore, use the word, not only as received, but as normally used.

Seneca, Epistulae 120.4 (trans. after Gummere)

quare, si fieri potest, et uerba omnia et uox huius alumnum urbis oleant, ut oratio Romana plane uideatur, non ciuitate donata. If possible, then, all our words and our pronunciation should carry the scent of an upbringing in the city, so that our speech seems really Roman, not simply having been given citizenship.

Quintilian, Institutio oratoria 8.1.2–3

‘fuisset autem uerbum hoc [‘nanus’] a te ciuitate donatum aut in Latinam coloniam deductum, si tu eo uti dignatus fores, essetque id inpendio probabilius quam quae a Laberio ignobilia nimis et sordentia in usum linguae Latinae intromissa sunt’. [Apollinaris to Fronto]: ‘But that word [‘nanus’; ‘dwarf’] would have been granted citizenship or led into a Latin colony by you if you had deigned to use it, and it would be more acceptable than the excessively shameful and crude words that were introduced into Latin usage by Laberius’.

Aulus Gellius, Noctes Atticae 19.13.3

hic idem cum ex oratione Tiberi uerbum reprehendisset, adfirmante Ateio Capitone, et esse illud Latinum, et si non esset futurum certe iam inde, ‘mentitur’, inquit, ‘Capito; tu enim, Caesar, ciuitatem dare potes hominibus, uerbis non potes’. When this same [Porcellus] had criticised a word in one of Tiberius’ speeches, and Ateius Capito declared that it was Latin, and if not, that it would certainly be from now on, Porcellus answered: ‘Capito lies; for you, Caesar, can give citizenship to men, but not to words’.

Suetonius, De grammaticis et rhetoribus 22

καί τινος Ἀτεΐου Καπίτωνος εἰπόντος ὅτι εἰ καὶ μηδεὶς πρόσθεν τὸ ὄνομα τοῦτ᾿ ἐφθέγξατο, ἀλλὰ νῦν γε πάντες διὰ σὲ ἐς τὰ ἀρχαῖα αὐτὸ καταριθμήσομεν, Πορκελλός τις ὑπολαβὼν ἔφη σύ, Καῖσαρ, ἀνθρώποις μὲν πολιτείαν Ῥωμαίων δύνασαι δοῦναι, ῥήμασι δὲ οὔ.

And a certain Ateius Capito stated: ‘Even if nobody has deployed this word before, we shall now count it amongst traditional words because of you’. But a certain Porcellus said in reply: ‘Caesar, you can grant Roman citizenship to men, but not to words’.

Cassius Dio 37.17.1–4

Specifically, the metaphor describes the integration of foreign words into Latin; the technical term for these integrated words might be borrowings or loan-words.⁶ It triggers thinking about the relationship between Latin and other languages, and about Latin itself.

There is no decisive evidence of direct intertextuality between the passages, except in the case of Dio, who gives the metaphor within the same anecdote as Suetonius and whose Greek is a near-direct translation of the Latin (πολιτείαν […] δοῦναι for ‘ciuitatem dare’). Therefore, either Dio drew directly on Suetonius or both made use of common material.⁷ There are, however, differences of context even between them: Suetonius gives the metaphor within the biography of the grammarian Porcellus, while Dio gives it as part of the wider narrative of Tiberius’ reign.⁸ Notably, Suetonius, Dio and Gellius are quoting others’ oral uses of the metaphor, hinting at the existence of the metaphor beyond these texts. In general, the metaphor does not seem to have been a fixed expression passed down directly between texts, but seems rather to have been part of a shared language. Variations and inflections in this shared language are highly revealing of linguistic and political attitudes.

In addition to the six examples of citizenship grants to words, I also include in my analysis two closely related metaphors from Columella and Cicero. These describe ‘grants of citizenship’ to agriculture and philosophy respectively:

et ut agricolationem Romana tandem ciuitate donemus (nam adhuc istis auctoribus Graecae gentis fuit) iam nunc M. Catonem Censorium illum memoremus, qui eam Latine loqui primus instituit […] And so that we might at last grant Roman citizenship to agriculture (for up to now it has belonged to Greek authors), let us now recall that eminent Marcus Cato the Censor, who first taught her to speak in Latin [there follows a list of subsequent authors who have written about agriculture in Latin].

Columella, De re rustica 1.pref. 12

‘ne tu’, inquam, ‘Cato, uerbis illustribus et id quod uis declarantibus! itaque mihi uideris Latine docere philosophiam et ei quasi ciuitatem dare; quae quidem adhuc peregrinari Romae uidebatur nec offerre sese nostris sermonibus’. ‘Indeed, Cato’, I said, ‘your language is crystal clear; it conveys your meaning exactly. In fact I feel you are teaching philosophy to speak Latin, and are therefore as if giving her citizenship. Up until now she has seemed a foreigner at Rome, hesitant to speak in our language’.

Cicero, De finibus 3.40

Both these metaphors are also relevant to the relationship between citizenship and language. Columella says that the translation of agricultural treatises into Latin was an important prerequisite (but not the clinching factor) for the citizenship grant. Similarly Cicero states that philosophy was ‘as if’ (‘quasi’) granted citizenship when it was translated into Latin.⁹

My approach in this article borrows from recent work that has used vocabulary and phraseology to illuminate ancient patterns of thought or what might be termed the Roman ‘social imaginary’.¹⁰ Metaphor is an especially fruitful source for this kind of analysis, since it can be understood as evidence of underlying conceptual connections between the different spheres or ‘domains’ that it links (in this case, citizenship and language).¹¹

A key tenet of the concept of the Roman ‘social imaginary’ is that vocabularies and expressions do not merely represent abstract modes of thought but actively constitute social realities by both reflecting and shaping them.¹² Analysis of modes of expression can therefore reveal habits and parameters of social and political practice. Such analysis can be conducted independently of an author or speaker's intentions; modes of expression deliver insights into ancient realities whether or not authors consciously intended them to do so.¹³ Not all instances of the citizenship metaphor were necessarily conscious salvos in a debate about citizenship and language. Nevertheless, fault-lines and questions about these topics can be extracted from the metaphor.¹⁴ I show that different instances of the metaphor expose revealingly divergent views of ‘citizenship’. Specifically, they vary in the value and meaning they ascribe to citizenship within Roman politics and identity, and in where they locate the authority to grant it. There are variations not only within the two domains of the metaphor, but also in the relationship conceived between them.

I concentrate on the specific metaphor of ‘granting citizenship’ because two of my key questions in this article are how important citizenship was to ancient understandings of Romanness and how tightly citizenship and language were bound. I contend, however, that this metaphor is part of wider metaphorical and conceptual linkages between the spheres of language and politics.¹⁵ The metaphor sits alongside a wide range of other metaphorical vocabulary that shows how frequently linguistic phenomena were understood in terms of political and social relationships.¹⁶ Varro compares words to slaves; Fronto states that speech has imperium; the Roman genre of comedy is described as togata (the toga obviously being another key marker of Roman social and political status); several authors draw a connection between political stasis and language change.¹⁷ Some of these other political metaphors occur in close proximity to examples of the citizenship metaphor. Within the Gellius passage, the grammarian Apollinaris contrasts the possibility of granting a word citizenship with the possibility of ‘establishing it in a Latin colony’ (‘in Latinam coloniam deductum’).¹⁸ Like many other Latin authors, Seneca describes a foreign word as alienus and Quintilian describes words as peregrina;¹⁹ in both cases, words are described by the same vocabulary used to denote different groups of people across the Empire, thus conceptually linking linguistic and political status.²⁰ The term barbarus is the prime example here: it shows the deeply embedded connection between foreign people and foreign languages.²¹ Taken together, these metaphorical expressions are evidence of strong linkages between the linguistic and political spheres. This wider backdrop makes the specific citizenship metaphor all the more pointed.

The political charge underlying vocabularies and expressions was also recognised in ancient thought. Fronto articulates it explicitly when he states that the emperor's speech is a material part of his imperium (123.16–18 VdH):

imperium autem non potestatis tantummodo uocabulum sed etiam orationis est: quippe uis imperandi iubendo uetandoque exercetur.

But imperium is a term not only of power but also of speech; for the force of ruling is exercised by ordering and by forbidding.

This passage demonstrates ancient understandings of the power that linguistic expression could hold; we are close here to an ancient version of ‘speech-act’ theory.²² Notably, grants of citizenship are themselves speech acts and/or expressions of imperium; this provides another linkage between citizenship and language. Fronto's observation also explains the consistent ancient interest in emperors’ language, and in the relationship between emperors and language (on which more below). This interest has relevance to the citizenship metaphor, where uncertainties about imperial control over both language and citizenship are visible.

The theory and background outlined in this section demonstrate that metaphors are revelatory of contemporary thinking; they both reflect and open up real-world questions about the domains they cover. Another key takeaway is that the metaphor of granting citizenship to words was one of a broad range of political metaphors used to describe language, and the frequency of these metaphors shows the conceptual linkages between the political and linguistic spheres.

II THE WOR(L)D OF THE CITIZENSHIP

The contexts where the citizenship metaphor appears demonstrate its political and historical bite. I deal first with the immediate context of the metaphor, the passages within which it is located, and then the broader historical context.

All the Latin passages except that from Seneca's Epistulae are, ostensibly, primarily about linguistic issues. The Quintilian passage is part of a technical treatise about the rules of oratory; Gellius reports a conversation about linguistic matters; Suetonius’ anecdote forms part of his De grammaticis, and is used to illustrate the strictness of the grammarian Porcellus, whom Suetonius describes as ‘the harshest critic of the Latin language’ (‘sermonis Latini exactor molestissimus’).²³ The passage of Seneca's Quaestiones naturales is a detailed philological and historical discussion of Latin and Greek wind names. The Columella passage also has the practice of writing and translation as its subject; his citizenship metaphor is embedded in a long passage describing the work of earlier Greek, Roman and Carthaginian agricultural writers. Similarly, Cicero employs his metaphor within discussion of the style and language of philosophy. In Seneca's Epistulae, the metaphor might seem closest to being a cliché, tangential to the main subject of the passage. It appears in an aside justifying Seneca's use of the term analogia within his philosophical discussion of how people acquire knowledge of what is good; Seneca makes the case that they do so by analogy.²⁴ The metaphor of granting citizenship to the term analogia is, however, deployed as part of Seneca's argumentation, not as an empty turn of phrase. As he explains, in using analogia, Seneca is making a choice, both linguistic and philosophical, that some might reject, and he therefore needs to justify his decision.

The concerns of these passages are, however, not narrowly linguistic, but are bound up with wider themes of Roman power, politics and identity. This is clear even from the ostensibly technical and scholarly passage of Seneca's Quaestiones naturales. Seneca's commentary makes clear that wider issues of Roman power and identity are at stake in language choices. Williams sees the naming of the winds as a ‘charged affair’, part of the moralising and ‘Romanising’ vision of empire that Seneca constructs in the Quaestiones naturales.²⁵ The varying versions of the relationship between Latin and Greek across the different wind names help to construct the Roman worldview that Seneca offers. The reference to the Battle of Cannae that the name Vulturnus provokes suggests that traditional names of the winds could recall important moments of collective Roman memory and self-definition. Seneca's broader argument about ‘our’ simultaneous adoption of and superiority over Greek ways of structuring the world, about Greek as both part of and alien to ‘our’ worldview, is made through the different relationships between Greek and Latin words.

The Suetonius and Dio passages, in which emperor and grammarian clash over language policy, bring language choices directly into the political arena. This is one of several places where Suetonius demonstrates that debate about language is both philological and political. He regularly uses language choices as a marker of emperors’ characters: perhaps the most explicit statement of this is in the Gaius (29.1) when Suetonius says that Caligula ‘crimes was increased the magnitude of his crimes by the brutality of his language' (‘immanissima facta augebat atrocitate uerborum’). Emperors’ attitudes to and usage of foreign languages, especially Greek, are central to Suetonius’ characterisation of them, and differing views about the propriety of Greek are important points of contention across Suetonius’ texts.²⁶ It is notable, too, that Dio deploys this anecdote about language and citizenship as part of his broader historical narrative of Tiberius’ reign.

There are also close links between language and imperial politics in the Gellius passage. The conversation between Fronto, Apollinaris and Festus Postumius takes place in imperial space, in the courtyard of the Palatine palace while they await an audience with the emperor.²⁷ Both Fronto and Apollinaris have friends in high places. Gellius tells us twice that Apollinaris was consulted by Erucius Clarus, city prefect and two-time consul.²⁸ Fronto's linguistic authority also developed in the chambers of imperial power. He was tutor to Marcus Aurelius; the debate about what else he was to him has filled many chapters.²⁹ Fronto is regularly exercised by the correct relationship between Greek and Latin, negotiating the correct form of lingua Romana, ‘the Roman language’, in letters to the emperor.³⁰ Language was for him a political matter: we have already seen his statement that language has imperium, and, like Suetonius, he shows interest in emperors’ language, linking decline in their speech with decline in the quality of imperial rule.³¹

So these passages evince the close relationship between language and political issues of power and status even beyond the citizenship metaphor. Importantly, as with the metaphor itself, it is consistently in situations when multiple languages are at play that these issues rise to the surface.

The broader historical context of the metaphor is also significant. In general, we should not imagine Roman citizenship as a commodity with a fixed value and meaning, with the only change being how widely it was spread. In fact the substance and value of citizenship was constantly shifting. Sherwin-White's classic account of Roman citizenship is structured around chronological stages of its development; Wallace-Hadrill stresses the ‘fluidity’ of citizenship even into Late Antiquity.³² Some of this fluidity is visible in the metaphor.

The extant examples of the metaphor date from the long transitional period between the limited extent of citizenship in the late Republic and the enfranchisement of all free men in the Empire in a.d. 212. The metaphor speaks to contemporaneous anxieties and uncertainties about citizenship. But is it possible to be any more historically specific, to relate the metaphor to any more distinct moments or developments? The possible history of the metaphor beyond extant texts means that some caution is prudent. There are, however, indications within the surviving evidence that the metaphor had particular salience to the Principate. All five Latin examples of the specific metaphor of granting citizenship to words date from the first and second centuries a.d. I suggest they reflect the growing complexities surrounding the relationship between citizenship, language and belonging under the Empire.

The use of political metaphor to describe language, outlined in Section I, shows that a link between language and political status dated back into the Republic. We come very close to the metaphor of granting citizenship to words, without quite getting there. Varro and Cicero both describe foreign words as peregrini, but they stop short of admitting these words as citizens. There are, moreover, some telling differences between the republican Ciceronian metaphor and later imperial versions, and between the citizenship-of-words metaphor and Cicero and Columella's metaphors of granting citizenship to philosophy or agriculture. Granting citizenship to philosophy (or to agriculture, as in Columella) — that is, to respected disciplines that the Roman elite aspired to claim for their own, and for which there was no native equivalent — was a different prospect from granting citizenship to foreign words that might compete with native Latin ones. The relationship between citizenship and Latin is more straight-lined in Cicero than in Columella's otherwise similar metaphor. In Cicero, the translation into Latin and the grant of citizenship go hand-in-hand (‘itaque mihi uideris Latine docere philosophiam et ei quasi ciuitatem dare’). In Columella, however, the equation between Latin and citizenship is less direct: he says that although Cato taught philosophy to speak Latin long ago, it is only now ‘at last’ that agriculture is getting citizenship (‘ut agricolationem Romana tandem ciuitate donemus […] iam nunc M. Catonem Censorium illum memoremus, qui eam latine loqui primus instituit’). That is, unlike Cicero, Columella implies that learning Latin was not in itself enough for citizenship. I argue that this difference reflected new configurations of language and citizenship as citizenship spread across the Empire.

The changing relationship between Latin and citizenship is the first of three contemporary developments (the others being concerns over the excessive extension of citizen rights, and the debate over who had authority to grant citizenship) that form a backdrop to the metaphor. Sherwin-White sees the fraying of the link between citizenship and Latin as a particular feature of the post-Augustan principate: ‘the connection of citizenship with Italian birth or origin, and later its connection with Latin culture, is gradually loosened’.³³ During the Republic, citizenship had predominantly been granted to people who spoke Latin, or at least who lived in areas with strong Latin influence.³⁴ This remained true even after the Social War (91–88 b.c.), the conflict between Romans and Italian allies that prompted the biggest republican extension of citizenship. A case has often been made for the use of Italic languages as a political weapon in this conflict and its aftermath, but a more common view now is that the war intensified, rather than initiated, an existing shift to Latin in the Italian peninsula that was completed soon after its conclusion.³⁵ The recipients of Roman citizenship at the war's end had mostly adopted Latin early and willingly.

This late republican moment, however, marked the cusp of the spread of citizenship beyond Latin-speaking regions and its attendant questions. Cicero's Pro Archia, delivered in 62 b.c. to defend the citizenship granted under post-Social-War legislation to a Greek poet from Antioch, provides insights into contemporary thinking about citizenship and language. In making his case for citizenship, Cicero downplays, or argues round, Archias’ foreignness, and especially his linguistic foreignness. He draws a distinction between people based on linguistic characteristics: Archias is not one of those poets who writes in an overbearing and ‘foreign’ (‘peregrinus’) style.³⁶ This is a slippery choice of words, because Cicero has just had to tackle the problem of Archias writing in Greek, which he does by arguing that Greek poetry is equivalent, and complementary to, Latin since it is more widely read, and that Archias uses Greek to promote Roman achievements.³⁷ Cicero's argument about language here is clearly not self-evident; he anticipates that some will use Archias’ writing in Greek as an argument against his enfranchisement.³⁸ The normal expectation here seems to be that citizenship and Latin were linked. Cicero suggests that they can be separated, but only with special pleading.

According to Sherwin-White, Augustus maintained a link between Latin and citizenship; he concentrated on enfranchising areas with strong Latin influence. Areas in Spain and Southern Gaul to which he granted citizenship tended to be areas with high levels of Italian (and therefore by this stage Latin-speaking) immigration; areas with majority local populations received Latin rights rather than full citizenship.³⁹ It was only later that citizenship began to be granted to people who did not necessarily speak Latin, and who lived in areas where languages other than Latin continued to be used even after their inhabitants had been granted citizenship.

The extension of citizenship beyond the borders of Latin-speaking regions created space for people to question what the correct relationship between Latin and citizenship was, and how closely they needed to follow one another. It is precisely this space for questioning the relationship between citizenship and language that the metaphor exploits, given that it is deployed to describe the actual or potential integration of new foreign words into Latin. This question was flirted with in the Republic, but became more urgent under the Principate. All this shows that links between citizenship and language went beyond the metaphorical, and added to the metaphor's force.

The debate about the relationship between citizenship and the Latin language was part of a second key concern, namely how widely citizenship should be granted. Under Claudius, worries about excessive grants of citizenship are especially visible.⁴⁰ Clotho's accusation in the Apocolocyntosis that the emperor wanted to see all Greeks, Gauls, Britons and Spaniards in a toga is comically overstated, but like all the best jokes it caught the contemporary mood.⁴¹ Another more serious passage from Seneca's De beneficiis also suggests that the emperor was toying with a wider extension of citizenship: Seneca floats the possibility of the emperor granting citizenship to all the Gauls (‘quid ergo […] si princeps ciuitatem dederit omnibus Gallis’).⁴² Although the question is asked as part of a rhetorical exercise, it is supposed to be a plausible, not fantastical, scenario. And whilst the text was written in the earliest years of Nero's reign, the princeps imagined here is conceivably Claudius, especially in light of Seneca's criticisms of Claudius’ over-generosity with benefits in Book 1 of the same work.⁴³ Another related change prompting this debate was the increase in grants of citizenship to individuals (or at least their increased visibility and regularisation), usually as a reward for service.⁴⁴ These grants of citizenship provided a model for grants of citizenship to individual words.

Concern about the extension of citizenship could also provoke anxiety about its loss of value. Tacitus, for example, thinks back wistfully to the Republic when citizenship was rare and therefore precious because it was linked to merit (‘eoque Romana ciuitas olim data, cum id rarum nec nisi uirtutis pretium esset’).⁴⁵ Pliny's concerns are different. He states that the financial burdens on those now granted citizenship are harsher than those imposed on citizens of longer standing; he is surprised that new citizens should regard citizenship as so valuable that they are prepared to accept these burdens.⁴⁶ Both authors suggest, in different ways, that citizenship, and citizens, were not always equal or equally valuable.⁴⁷ As I show in Section IV, this is a message that also reverberates in the metaphor.

None of this means that citizenship grants really were out of control in the first and second centuries or under Claudius in particular. The problem was one of perception, not reality. Myles Lavan's recent quantitative modelling of citizen numbers before a.d. 212 has shown that right up until the Constitutio, the growth of citizenship was more modest than generally assumed.⁴⁸ I argue in this article that this new model makes sense of the metaphor and vice versa. Close scrutiny of new citizenship grants and policing of the boundaries of citizenship such as we see in the metaphor makes most sense in a world where citizenship was not yet excessively common and remained to some degree exclusive.

The third relevant contemporary development is the intensification of debate about who had authority to grant citizenship. Under the Republic, grants of citizenship had generally been made collectively by the Senate. There were republican precedents for grants of citizenship by individuals (Marius’ grant to two cohorts in 101 b.c.,⁴⁹ Pompeius Strabo's grant to Spanish horsemen in 90/89 b.c.,⁵⁰ Julius Caesar's mass grants to the inhabitants of Cisalpine Gaul and the Spanish town Gades in 49 b.c.),⁵¹ but the early Principate was the first time that citizenship was routinely granted on one individual's authority. In the period contemporary to the metaphor, there were other more recent changes or challenges to the authority and initiative behind citizenship grants. The first known instance of a community petitioning for citizenship occurred during Claudius’ reign.⁵² This process of community-initiated citizenship grants became normalised from the second century onwards, but in the first century this process was still new and it marked the establishment of an alternative source of initiative behind citizenship grants than the emperor. There is also evidence for differences of opinion between emperor and Senate over citizenship grants. One prominent example is the senatorial opposition that Claudius faced when he proposed making men from Gallia Comata eligible for the senate, in a speech recorded both in Tacitus and on a bronze tablet from Lyon.⁵³ The issue here was of course not whether or not to grant citizenship, since the Gauls in question already held it, but citizenship still played a role in the debate; the question was whether the right to hold senatorial office should follow from citizenship. I argue below that the metaphor of granting citizenship to words responded to this contemporary questioning of authority over citizenship. The emperor's power to grant citizenship is directly challenged in Suetonius and Dio, but the question of authority runs across the other instances too.

The metaphor therefore both reflects and responds to contemporary concerns about what citizenship meant, what it depended on, and to whom and by whom it should be granted. These were concerns right across a broad period from the late Republic to the Constitutio; but they became especially urgent in the early Principate, as citizenship extended, however gradually, beyond Latin borders and the emperor provided a focus for discussions. In the next sections, I scrutinise the different instances of the metaphor more closely and use them to explore three sets of questions surrounding citizenship and language, which are all pertinent to the contemporary context outlined here. The linkage between the domains of citizenship and language powers the metaphor; and so the conditions and characteristics of citizenship shape its commentary on language, and vice versa.⁵⁴ This metaphor was not fossilised or inert; its variations reveal fault-lines in live contemporary debates.

III MAKING CITIZENS

I turn first to the process of making new citizens and in particular the questions about by and to whom citizenship should be granted.

Who has the power to admit words as citizens is a clear point of contention across the passages. Most immediately, this is a concern over the authority and agency behind the expansion of the Latin language, but it also reflects the broader contemporary debate about the authority behind citizenship grants.⁵⁵ In Quintilian and Seneca's Quaestiones naturales, the answer to the question of who grants citizenship to words is left vague: they describe the act of granting citizenship using a passive verb without an agent (‘ut oratio […] non ciuitate donata; eurus iam ciuitate donatus est’). The passages of Columella and Seneca's Epistulae are the only ones where the grant of citizenship is described by an active verb. Columella does not make the subject of the verb explicit, but he may be hiding behind the first person plural to claim authority over agriculture's citizenship for himself. Seneca gives a clearer answer in his Epistulae, stating that Latin grammarians have given analogia citizenship (‘hoc uerbum cum Latini grammatici ciuitate donauerint’). Grammarians’ authority to grant citizenship is not, however, absolute even here. Seneca suggests that some (unspecified) people have ignored grammarians’ advice and found fault with the word analogia, leaving him to make the case for its citizenship to be restored. The technical language of legal status runs through Seneca's commentary here; he describes the proposed restoration of citizenship using the same verb (‘redigere’) that is used, for example, in a passage of Suetonius’ Divus Claudius when Claudius returns a man to peregrine status.⁵⁶ Grammarians’ authority is also indeterminate in Gellius and especially in Suetonius and Dio. In Gellius, Fronto asks the grammarian Apollinaris whether nani is acceptable Latin; Apollinaris replies that Fronto himself has the power to determine whether or not a word becomes a citizen. That is, Fronto's view seems to be that linguistic authority rests with grammarians, whilst the grammarian Apollinaris defers the question back to Fronto. The nature and extent of Fronto's own linguistic authority is not straightforward.⁵⁷ It is difficult to describe him narrowly as a grammarian, and Apollinaris’ statement of Fronto's power over language follows Fronto's own anxieties that his power might be limited. In the anecdote described by Suetonius and Dio, grammarians’ authority competes directly with the emperor's. Whilst Capito argues that Tiberius has power over the boundaries of Latin, Porcellus denies this, stating that ‘you, Caesar, can give citizenship to men but not to words’ (‘tu […] ciuitatem dare potes hominibus, uerbis non potes’; σύ […] ἀνθρώποις μὲν πολιτείαν Ῥωμαίων δύνασαι δοῦναι, ῥήμασι δὲ οὔ).

So the different versions of the metaphor reflect debate within contemporary society about the location of linguistic authority. As their appearance in three of the passages shows, part of this debate was specifically about the role of grammarians. These men had growing power as, in Kaster's phrase, ‘guardians of language’, but there was some unease about their position.⁵⁸ The metaphor also speaks to a more technical Roman linguistic debate about whether language change occurred naturally or was actively imposed.⁵⁹ Citizenship, which could itself be actively granted or naturally acquired, was one framework to structure this debate. In Seneca's Epistulae, the metaphor appears precisely in the context of the contrast between the word as received (‘recepto’) and as normally used (‘usitato’). Seneca proposes multiple sources of linguistic authority in his discussion of the wind names in his Quaestiones naturales. Alongside the impersonal process of granting citizenship, he appeals to general usage (‘nostri’), named individuals (Varro and Livy), and offers his own opinion (‘mihi non uidetur’).⁶⁰ Ultimately, he does not provide a clear answer as to which linguistic authority prevails, giving different answers in the case of different words and as required to suit his own linguistic choices.

This debate about authority over language is not narrowly linguistic; it shows again the relevance of language to wider social and political concerns. Part of the anxiety about grammarians gaining power was specifically an anxiety about their low social status.⁶¹ In the Suetonius and Dio passages, there is an extra, and explicitly political, dimension to the debate. The clash between emperor and grammarian, as Porcellus denies Tiberius’ authority over Latin, is part of a wider debate about the emperor's power over language that runs through and beyond Suetonius’ text. This passage, in which Tiberius’ power to determine the boundaries of Latin is checked, contrasts with a passage of Suetonius’ Tiberius (71). Here, Tiberius is in control and sets the linguistic strategy for the senate to follow. He apologises for his use of one ‘foreign’ (‘peregrino’) word (‘monopolium’) and determines that another Greek word (ἔμβλημα), used in a senatorial decree, must be replaced by a native one. This fits with other evidence of a tightening of language policy under Tiberius, comprising hostility to the use of Greek and a desire to maintain strict Latinity, at least in the public sphere.⁶² The famous passage of Valerius Maximus (2.2.2) recording Roman magistrates’ avoidance of Greek can be read as reflective not of the situation in the Republic when it is set, but of the situation in the Tiberian era when it was written.⁶³

Both the Porcellus anecdote and other evidence, however, suggest that imperial decisions about language did not go unchallenged. It is notable that Suetonius himself has no hesitation in using the word monopolium even in a parallel context; in an earlier passage of Tiberius (30), he describes how Tiberius consulted the senate ‘de uectigalibus ac monopoliis’, using the term monopolium with integrated ablative ending and without comment. The term emblema meanwhile appears in the writings of Pliny the Elder and Quintilian, and would probably have been recognisable to Suetonius’ readership.⁶⁴ Claudius’ attempts to introduce three new letters might also be considered here: Suetonius tells us that he brought them into general use, but Tacitus says that their usage was short-lived.⁶⁵ And the princeps is notably absent from the multiple sources of linguistic authority that Seneca offers. We are left with the impression that the question of who has authority over language does not have a fixed answer, that the emperor's power to control language was not absolute. Though the citizenship metaphor shows that language was a political matter, there was not consensus over how far political authority applied to it.

I think we can go further. The metaphor of granting citizenship to words did not only reflect uncertainty about the extent of imperial power over language; it also reflected anxiety about imperial power per se. Although discussions about language were politicised, they simultaneously provided the distance needed to question the emperor's authority. Dio is clear that Porcellus was sailing close to the wind, but remained just the right side of propriety: ‘Tiberius did no harm to [Porcellus] over this, although he spoke very frankly’ (ἐκεῖνον μὲν οὖν οὐδὲν ἐπὶ τούτῳ κακόν, καίπερ ἀκρατῶς παρρησιασάμενον, ἔδρασε). The metaphor of citizenship brings the political and linguistic spheres together to show the contact between them, but Porcellus pulls back from actually intruding into politics. Challenging the emperor's authority to grant citizenship to a word was not in the end the same as challenging his authority to grant citizenship to a person.

There is other evidence to suggest that language was consistently one domain in which it was more possible to criticise the emperor than elsewhere. As we have seen, both Suetonius and Fronto link language to imperial character; Suetonius offers regular commentary on emperors’ language in his De vita Caesarum. In a famous example from the Historia Augusta, Septimius Severus’ African accent in Latin is remarked upon scornfully: note the disparaging sed in the comment ‘canorus uoce sed Afrum quidam usque ad senectutem sonans’; (‘his voice was melodious but he retained an African accent right up to old age’).⁶⁶ The same text tells us that Severus was ‘much embarrassed’ (‘multum […] erubesceret’) by his sister's Latin, to the point that he had to send her away from Rome.⁶⁷ Her status as a member of the imperial family was not enough to compensate for the deficiency of her language; Severus was forced to capitulate to popular pressure. The setting of the Gellius passage in the courtyard of the imperial palace is also pertinent to this borderline positioning of language and politics; the conversation about language is relevant to imperial authority, but stops just short of encroaching on controversial ground. We could, however, follow Keulen here in reading Gellius as implicitly critical or satirical of Fronto; he does not have a firm grip on language, despite portraying himself as an expert.⁶⁸ On this reading, we might also understand this passage as implicitly critical of the emperor Marcus Aurelius, whose linguistic attitudes and practices were tightly bound to Fronto's own. This criticism perhaps also extends to grants of citizenship themselves. It may not be coincidental that Marcus Aurelius apparently had something of a reputation for excessive liberality with citizenship: the fourth-century historian Aurelius Victor commented that under Marcus, Roman citizenship was given indiscriminately to all (‘data cunctis promiscue ciuitas Romana’).⁶⁹

Contemporary developments and debates surrounding the location of authority behind citizenship grants, outlined in Section II, made the citizenship metaphor especially applicable to contests over linguistic and political authority. This might also reflect a broader shift in the location of authority under the Principate.⁷⁰ Although the emperor had ultimate authority, he could delegate this authority to professionals, including lawyers and grammarians. Policing of this delegated authority was pertinent to the spheres of both language and citizenship.

The metaphor also engages with debates about whom citizenship should be granted to, and whether those of certain origins were more worthy of citizenship than others. Here again the metaphor is indicative of attitudes to both language and politics.

The instances of the metaphor imply a hierarchy of languages. It is notable that all the words granted citizenship that we can identify are of Greek origin, suggesting that Greek words were more promising candidates for citizenship than words from other languages. In Gellius, it is precisely because nanus is of Greek origin, and not barbarous, that it is deemed worthy of citizenship. In his Quaestiones naturales, Seneca is exclusively interested in Greek versus Latin wind names: as Adams points out, the wind names carry regional inflections that Seneca ignores.⁷¹ The (partial) exception to the privileged position of Greek is Quintilian. Here, all those granted citizenship are contrasted unfavourably with those from Rome; Quintilian regards even the regional accent of Padua, whose inhabitants were granted citizenship back in 49 b.c., with disdain. Quintilian here represents the narrowest definition of Roman language and identity, centred exclusively on the city of Rome. In general, the broad distinction between Greek and other languages reflects the fact that the boundary of acceptable Roman language could be drawn either around Latin alone or around Latin-and-Greek in contrast to other languages.⁷²

Yet even the citizenship of Greek words could be precarious and contested. In several examples of the metaphor, a word's citizenship is questioned. One clear difference of opinion is over analogia: Seneca tells us that Latin grammarians granted it citizenship, but that this status has subsequently been lost and he must therefore argue for its restoration. The lack of consensus over words’ right to enfranchisement is indicative of the ongoing uncertainty about the place of Greek in relation to Latin. This relationship was never settled, as some have tried to argue.⁷³ The treatment of Greek words in the citizenship metaphor shows rather that the relationship between Greek and Latin continued to be a live issue right across the two hundred years that the passages cover. Individual Greek words did not maintain a static position in relation to Latin. Linguistic integration was not a one-way process; as the passage of Seneca's Epistulae reminds us, citizenship could be lost as well as gained. All this shows the complex stratigraphy of language contact and the constantly shifting boundaries between Latin and other languages.

There is, however, something of a mismatch between the hierarchy of languages visible in the metaphor and the hierarchy of people when it came to real-life citizenship grants. While Greek words are those most readily granted citizenship, there was reluctance to grant citizenship to Greek-speaking people.⁷⁴ Greeks are one of the groups that Clotho mocks Claudius for attempting to give citizenship to. And Greek-speaking migrants specifically are the target of Umbricius’ rant in Juvenal's third satire, which I explore in more detail in the next section. This could be evidence of the limits of the metaphorical mappings between citizenship and language.⁷⁵ But the apparent mismatch between the people and the languages deemed worthy of citizenship is itself revealing. First, it is evidence of the sharp distinction that Romans could draw between the Greek language and modern Greeks.⁷⁶ Second, the mismatch casts the debate about who was worthy of citizenship into relief. And third, this mismatch was not total. During the early Principate, block grants of citizenship in the Greek East remained rare; here, grants of citizenship to worthy individuals were more common.⁷⁷ The need for close scrutiny of individual candidates for citizenship from the Greek East made a good model for the granting of citizenship to individual Greek words, and reflected the hesitation about the status and worth of both the Greek language and Greek-speaking people.

Usage of the citizenship metaphor reflects the subjective and contested nature of decisions about the extension of both language and citizenship. Different people located the authority over these spheres in different places. Not everybody regarded borrowings from other languages as equally or acceptably ‘Roman’, just as not everybody regarded people from different parts of the empire as equally valid candidates for citizenship. Some saw the boundaries of language as more permeable than others, in the same way that some regarded the boundaries of citizenship as more permeable than others. In the different instances of the metaphor we see a spectrum of these attitudes.

IV CHALLENGING CITIZENSHIP

Having dealt in the previous section with the process of making citizens, in this section I use the metaphor to consider the meaning and value of citizenship once granted. It is here that the metaphor provides a way in to destabilise our view of citizenship as a stable and secure marker of Romanness.

Although the metaphor apparently describes a single linguistic process, crucially the words that have been ‘granted citizenship’ do not all have the same linguistic form or status. The metaphor actually describes different degrees of linguistic integration in different passages. It is especially striking that the metaphor does not necessarily imply total integration to the point that foreign words granted citizenship are indistinguishable from native Latin ones.

Both Quintilian and Seneca draw a distinction between native and nativised language. The passage of Seneca's Quaestiones naturales is especially interesting, because the name Eurus that has been ‘granted citizenship’ can be compared both to words granted citizenship in other passages and to the other wind-names within this passage. On the one hand, Eurus is integrated to the extent that it has a Latin nominal ending and is written in Latin script.⁷⁸ This distinguishes it from less integrated Greek wind names in the passage (θρασκίας, εὐρόνοτος, λευκόνοτος), which appear with Greek morphology and in Greek script. Eurus, however, also contrasts with the native Latin name of the same wind, Vulturnus. Although claiming that Eurus is ‘not a foreigner’ (‘non […] alienus’), Seneca also states that ‘our people’ (‘nostri’) use Vulturnus.⁷⁹ The memories of Cannae that the name Vulturnus provokes underlines the point that the name Eurus is empty of this shared cultural and historical capital. Citizenship, moreover, does not represent the final end-point of integration in this passage. The name Zephyr is more firmly rooted in Latin: while Eurus’ former status as a foreigner is remembered, Seneca tells us that Zephyr is used ‘even by those who do not speak Greek’ (‘zephyrum esse dicent tibi etiam qui Graece nesciunt loqui’). In the passage of Seneca's Epistulae, the word granted citizenship is also potentially marked out as distinct from native Latin words, since it appears with a Greek accusative ending (‘analogian’).⁸⁰

In Gellius, however, the word cleared for citizenship, nanus, is more integrated than in the other passages.⁸¹ Even as Fronto enquires whether he is right to use the word, it appears with Latin accusative ending and in Latin script (‘Fac me […] ut sim certus an recte supersederim nanos dicere’). This marks a transformation from its pre-citizen status, which we are shown when Apollinaris describes how Greeks use the word and it appears with Greek accusative ending and in Greek script (‘νάνους enim Graeci uocaverunt brevi atque humili corpore homines paulum supra terram extantes’).⁸² That is, citizenship here marks a change from visible foreignness to full integration. This view of citizenship as a process of transformation and assimilation is reinforced if we interpret the other colonia metaphor in this passage not as an alternative to citizenship, but as a step towards it; those that were not yet deemed worthy of citizenship could be placed in the halfway house of a colony.

Citizenship therefore does not mean the same thing in all the passages. It ranges from denoting full linguistic integration, so that an enfranchised word is morphologically identical to a native Latin one, to denoting language that remains distinct from native speech. The different degrees of integration within and between the passages, indicated by script and morphology, match the matrix of linguistic integration identified in modern sociolinguistic analysis; as in the modern world, the definition of different linguistic phenomena in antiquity was not fixed.⁸³ The varied script and morphology of words ‘granted citizenship’ is evidence that the same linguistic phenomenon was not always regarded as equally integrated. Linguistic integration operates on a continuum.⁸⁴ All this shows again the unstable boundaries between Latin and other languages. Importantly, the metaphor shows that assessments of a word's place in Latin were not narrow or technical linguistic judgements. Linguistic features like script and morphology provide clues to a word's status but are not determining factors. Judgements about a word's status were always to some degree subjective.

The different degrees of integration implied by ‘citizenship’ in the metaphor are not only significant for our understanding of ancient attitudes to language. They also indicate the fluidity and uncertainty of the meaning of citizenship as a political status. And they should prompt us to think again about the place and value of citizenship in hierarchies of being and belonging as Roman.

The most obvious challenge to the value of citizenship is in Quintilian, where speech that has been granted citizenship is contrasted unfavourably with that which is ‘really Roman’ (‘plane Romana’). Quintilian's statement that citizenship is not a sufficient condition of being ‘really Roman’ is, I think, important to stress in the face of a weight of scholarship that regards citizenship as the ultimate marker and definition of Romanness. Ando states that ‘one becomes Roman by becoming juridically Roman’;⁸⁵ Wallace-Hadrill sees the primacy of citizenship as the key contrast between the categories of Roman and Greek, marking the distinction between a political and a cultural or ethnic identity;⁸⁶ Hall reduces Roman identity to solely a legal-juridical status.⁸⁷

In the Quintilian passage, being ‘really Roman’ means having had an upbringing in Rome. Adams notes the unusualness of this narrow definition of ‘Roman’, centring on the city. He finds that the adjective Romanus, when applied to speech or language as it is in the Quintilian passage, denoted the Latin of the city of Rome specifically only until the Augustan period at the latest.⁸⁸ Adams downplays the significance of the Quintilian passage in breaking this chronology, but I argue that it should be taken more seriously. Once we look beyond the application of the label ‘Roman’ to language, we see that Quintilian is not exceptional in his definition of being Roman here. The two modes of being Roman that Quintilian contrasts in his deployment of the metaphor — one dependent on citizenship and a narrower one, dependent on origin or background — correspond to two poles of a wider contemporary debate about citizenship and its position in definitions of Romanness.

This debate is dramatised in the near-contemporary text of Juvenal's third satire.⁸⁹ The protagonist Umbricius’ rant against foreigners who have migrated to Rome culminates in his statement that ‘there is no space for any Roman’ in the city (‘non est Romano cuiquam locus hic’).⁹⁰ Umbricius’ definition of Romanus here is, like Quintilian's, decidedly narrow and nativist. For him, being ‘Roman’ also means having been brought up in the city; in lines 81–5, he evokes his childhood in Rome in traditional, nostalgic terms and expresses incredulity that someone ‘blown to Rome on foreign winds’ should gain job and housing opportunities ahead of him. There is also a social dimension to Umbricius’ definition of ‘Roman’; the ‘Romans’ he champions are the unskilled poor. Umbricius’ is the rant of these Romans against the citizens-of-the-world that have taken over the city and successfully marketed their skills there. The Roman citizenship is itself a point of slippage or confusion in Umbricius’ speech. Umbricius addresses his words to the Quirites, the traditional term for the Roman citizen body. Umbricius intends the term in its original sense, when citizenship was restricted to Rome and Italy. By the second century a.d., however, the Roman citizen body obviously included the citizens-of-the-world that Umbricius attempts to exclude from his vision of true Romanness. Though Umbricius sets a Protogenes or Diphilus against the ‘Roman’, there is inscriptional evidence from the city of Rome of people with the cognomina Protogenes and Diphilus as Roman citizens.⁹¹ Although this irony is lost on Umbricius, it was presumably not lost on Juvenal, who is playing on the clash between different worldviews.

The comparative evidence of Juvenal shows that Quintilian's deployment of the citizenship metaphor did not represent an empty cliché, but rather reflected a position in a contemporary debate about the place of citizenship. This debate involved a clash between a definition of Romanness centred on citizenship that was inclusive of people from across the empire and a narrower, exclusive definition centred on origin that privileged people from the city of Rome itself.

In the other instances of the metaphor, challenges to the value of citizenship are less direct than in Quintilian. As I have set out, however, there are nonetheless differences in the status implied by citizenship. The matrix of integration within the passage of Seneca's Quaestiones naturales — where, as we have seen, there is a stage of belonging beyond citizenship — also depends upon the notion that citizenship was a necessary but not sufficient condition of being Roman. Beyond his privileging of Greek words, Seneca is less clear than Quintilian or Juvenal that origin is what really matters; here cultural knowledge and acceptance by the wider population are significant in enabling one to belong.

The Gellius passage again offers a contrast. There, citizenship is higher in the hierarchy of belonging. The word nanus, which Apollinaris deems worthy of citizenship, is favourably contrasted with ‘the excessively shameful and crude’ (‘ignobilia nimis et sordentia’) words that Laberius introduced into Latin usage. Citizenship is here a judgement of quality: nanus is an acceptable candidate for citizenship because it is not barbarous and crude. More significant is the nature of the words introduced by Laberius which Apollinaris regards as inferior to the naturalised word nanus.⁹² Gellius has already given details of these words in an earlier passage.⁹³ Although some of the novel words that Laberius uses in his mimes are of foreign origin, some are Latin words that were previously restricted either to technical contexts or to common speech.⁹⁴ Apollinaris’ claim in 19.13 that the enfranchised nanus is ‘more acceptable’ (‘probabilius’) than Laberius’ words can therefore be read as an implicit acknowledgement that naturalised foreign citizens can be more worthy of belonging than native ones. Here, social status is what really matters. This is the reverse both of Quintilian's claim that origin trumps citizenship and especially of Umbricius’ claim that origin matters more than social status; it is in fact precisely the attitude that social status should triumph and foreigners flourish that Umbricius attacks.

These passages therefore reveal different possible hierarchies of Romanness, and show that citizenship competed with origin, culture and social status within these hierarchies. In different instances of the metaphor, citizenship is defined and valued in different ways; it is not a fixed or straightforward marker of being Roman. The metaphor thus reflects the lack of agreement about what being Roman and being a citizen meant in a contemporary world where the Roman population was growing in size and diversity.

V CITIZENS OF THE WORD

In Sections III and IV, I looked at different stages of enfranchisement and showed that usages of the metaphor reveal parameters of contemporary thought about both language and citizenship. I now turn to the relationship between language and citizenship. In particular, the metaphor offers insights into contemporary questions about whether Roman citizenship should be dependent on speaking Latin, and about where language stood in relation to citizenship in hierarchies of belonging as Roman. These questions also spark wider issues surrounding migration, assimilation and the perception and tolerance of difference amongst new citizens.

Traditionally, there was a close connection between speaking Latin and being a Roman citizen.⁹⁵ Cicero asserts this connection strongly: he argues that language was one of the key attributes that bonded Roman citizens together and that speaking good Latin was an essential skill of a Roman citizen.⁹⁶ This connection is also present in the Ciceronian version of the citizenship metaphor, where speaking Latin and gaining citizenship go hand in hand.

As I discussed in Section II, this traditional connection between Latin and citizenship came under pressure as citizenship was extended beyond Latin-speaking regions and peoples. According to Sherwin-White, this effected a ‘gradual loosening’ of the link between citizenship and Latin language and culture, especially in the post-Augustan principate.⁹⁷ It might be tempting to understand the metaphor as a product of this loosening; the concept of granting citizenship to foreign words seems to follow naturally from the granting of citizenship to people who did not necessarily speak Latin. And as we have seen, citizenship is not dependent on full linguistic integration in every instance.

The different versions of the metaphor, however, show that contemporary takes on the relationship between citizenship and language were more varied than a one-way loosening. There were continued attempts to assert the link between Latin and citizenship. In Columella's version of the metaphor, Latin remained a pre-requisite for citizenship, even if it was not enough on its own. In his Quaestiones naturales, Seneca draws a distinction between a word entering Latin as a foreigner and a word becoming a citizen, with some degree of Latinisation required for enfranchisement.

These varied metaphorical views of the relationship between citizenship and language mirror real-world evidence of responses to the spread of citizenship beyond traditional Latin-speaking borders. Here also there were attempts to (re)assert the connection between citizenship and Latin. Suetonius tells an anecdote about Claudius stripping a man of citizenship because he did not speak Latin.⁹⁸ Dio again gives the same anecdote, adding that Claudius said explicitly that ‘it was not fitting that somebody should be Roman who had no knowledge of the Roman language’.⁹⁹ Inscriptional evidence from the city of Rome suggests that a high proportion of migrants in general integrated by using Latin,¹⁰⁰ but that there was an especially pronounced link between the use of Latin and the advertisement of citizenship.¹⁰¹ In other evidence, however, we get nearly the opposite impression. The accusations in Seneca that Claudius wanted to spread citizenship widely implied a perception that he was willing to spread citizenship to people who did not necessarily know Latin.

The agency behind these different configurations of the relationship between Latin and citizenship was itself multi-directional, just as we see varied agency behind citizenship grants in the metaphor. The assertion of the link between Latin and citizenship was made from the bottom up as well as the top down. This is evident in the choice that new migrant citizens made to use Latin in inscriptions at Rome; there is no evidence of enforcement to do so, and the use of other languages was a possible choice.¹⁰² Indeed, Latin was most common precisely among those whose primary claim to status and belonging was through citizenship; the use of other languages was more frequent among higher-status groups who had other attributes to advertise.¹⁰³ That is, it was new citizens themselves who reinforced and exploited the link between Latin and citizenship. I want to suggest that it was in the face of these claims to status through citizenship and through Latin that alternative definitions and hierarchies of Romanness such as we see in the metaphor — based on origin, social status or accent — were made. There were fears that Latin and citizenship were becoming too widespread, and these fears prompted a search for more exclusive markers of Roman status.

The metaphor also offers insights into the closely related question of the perception and tolerance of linguistic difference amongst new migrant citizens. As we have seen, the degree of linguistic assimilation that citizenship implied or required varied within the metaphor; some new citizens stand out linguistically from native words, others are invisible migrants. It is, however, unclear how far even assimilation brought acceptance; there is evidence that some were reluctant to acknowledge the citizenship and/or the right to belong even of those who legally and linguistically integrated. Fronto, Tiberius and Seneca all check up on or draw attention to the status of seamlessly integrated migrant citizens. Varying levels of tolerance can also be seen beyond the metaphor. One example is Claudius’ grant of citizenship to Alpine tribes in a.d. 46.¹⁰⁴ The case revolves around a group of tribesmen who have, apparently mistakenly, appropriated Roman citizenship. Claudius confirms their citizenship on the basis that they have been living and serving as citizens and nobody has noticed the difference for years; these tribesmen are indistinguishable from legitimate citizens. He, however, clearly anticipates pushback from those who are unwilling to grant citizenship even to these assimilated hidden migrants; it is worth remembering that the whole investigation was triggered by an unhappy informer. Juvenal also dramatises these anxieties in his third satire, where Umbricius has a problem with both assimilated and unassimilated migrants and is hazy on their relationship to citizenship. He complains that the foreigners who invade Rome are highly visible and audible, highlighting language as a defining marker of foreignness;¹⁰⁵ but he also complains about those who assimilate too well, to the point that he worries they will become ‘the guts and masters of our great houses’ (‘uiscera magnarum domuum dominique futuri’).¹⁰⁶ It is perhaps this latter complaint that comes closer to reality.¹⁰⁷ Languages other than Latin are strikingly rare at Rome in either inscriptions or graffiti.¹⁰⁸ Just three per cent of inscriptions from the city are in Greek and there is only a handful of inscriptions in other languages.¹⁰⁹ Anxieties about linguistic assimilation, even if misplaced, add another dimension to the unsettled relationship between citizenship, integration and belonging with which the metaphor plays. Different expectations of linguistic assimilation reveal again the clash between different conceptions of what citizenship meant.

Finally, the usages of the metaphor also reflect the debate about where citizenship ranked relative to language in hierarchies of Roman status and belonging. In Columella, citizenship outranks language: speaking Latin is a necessary, but not sufficient, condition of gaining citizenship. In Gellius, the Latinisation of a word is not necessarily a guarantee of citizenship either; Apollinaris holds out the alternative possibility of a word ‘entering a Latin colony’ should it not be granted full citizenship. In Quintilian, however, the ranking of citizenship and language is reversed. Language — specifically and narrowly the Latin of the city of Rome — is primary; here it is citizenship that is an insufficient condition of being really Roman. Strikingly, though he regards even the hint of a non-urban accent as undermining the quality of a person's speech, he also says that an accent can be learned, that by speaking the right kind of Latin one can transcend one's enfranchised status to pass as someone who really counts as Roman. This view further complicates the relationship between language, citizenship and belonging. It tantalisingly raises the question of how far Quintilian's logic might extend, whether someone who uses the right kind of language might ever pass for Roman even without citizenship.

The metaphor therefore shows the instability of the link between language and citizenship, and the instability of their relative positions in conceptions and hierarchies of Romanness. It complicates Sherwin-White's narrative of a gradual loosening of the connection between citizenship and Latin during the Principate. Whilst there is some indication of this loosening in the development of the metaphor, there is also evidence of a reassertion of the connection. In a contemporary world where the link between Latin and citizenship could be claimed and exploited by new and/or lower status citizens, another response was to narrow the definition of Romanness to make other criteria count for more. The relationship between Latin and citizenship did not develop linearly, but was constantly in construction and deconstruction; the metaphor was itself part of these processes.

VI CONCLUSION: A WIDER WOR(L)DVIEW

This article has argued that the metaphor of ‘granting citizenship’ to words was not a cliché, but delivers insights into ancient thinking about both citizenship and language. It reflects and responds to contemporary debates in the context of the extension of citizenship beyond traditional Latin-speaking borders about what Roman citizenship and Roman language should mean; what citizenship should depend upon; to whom and by whom it should be granted. It also raises wider questions including about identity, belonging, migration and imperial power.

The metaphor should make us rethink aspects of Roman citizenship and identity more broadly. It destabilises the notion of citizenship as central to contemporary definitions of Romanness. The different versions of the metaphor reveal different conceptions of the place that citizenship occupied in hierarchies of belonging. Some people questioned the value of citizenship and citizenship had to compete against (and sometimes lost out to) other markers or definitions of being Roman. The challenge to the centrality of citizenship also problematises the common notion of Roman identity as inclusive, a notion that is often dependent on Roman generosity with the franchise.¹¹⁰ Ancient understandings of Romanness that de-prioritise citizenship and especially those that prioritise origin make ‘Roman’ into a much more exclusive category. Indeed, I have shown the clash between inclusive and exclusive definitions of Romanness within and beyond the different versions of the metaphor.

The metaphor also raises the related question about Roman tolerance of difference, the extent to which new migrant citizens were expected to integrate. There were again differences of opinion here. Some resisted integration, while even invisible, naturalised migrants were too much for others. Lying behind these debates is the question of agency, of who gets to decide both specifically about language or citizenship, and more broadly about what defined Romanness.¹¹¹ Multiple different answers are proffered in the evidence I have discussed here; the power even of the emperor over the boundaries of language, citizenship and Romanness is open to challenge. I have suggested that there is a social as well as political dimension to these debates about authority and belonging. They reflect anxieties about social status, and there was a contest between top-down and bottom-up definitions of Romanness.

The metaphor is, of course, also revealing of the relationship between citizenship and language. It was employed in a contemporary context when the link between Latin and citizenship was fraying, or at least looked as if it might fray. It does not, however, straightforwardly support Sherwin-White's view of a gradually loosening connection between Latin and citizenship; it in fact indicates attempts to reinforce this connection. Overall, the question of how far citizenship should depend on language, or of the relative ranking of citizenship and language in hierarchies of Roman status and identity, went unresolved. But it was clear that language was a political matter. The metaphor is part of a wider conceptual mapping between the linguistic and political spheres. Simultaneously, the distance of language from politics made it a sphere where sensitive political commentary could be made.

I want to end by reflecting again on how the metaphor fits with Lavan's recent model of modest growth of citizenship up to a.d. 212.¹¹² Based on this model, he argues against the idea that citizenship had already lost all value before the Constitutio. The metaphor can be made to fit with Lavan's model in two ways. First, I stress that it is the perception of excessive citizenship grants that matters, not the reality of its spread. Accusations like Seneca's in the Apocolocyntosis cannot be proved or disproved by surviving epigraphic or demographic evidence.¹¹³ Instead, they should be taken on their own terms as evidence of the perception of imperial citizenship policies. There is recognition of this problem of perception within ancient evidence. In the Lyon tablet, Claudius bends over backwards to present his actions as cognate with the policies of Augustus and Tiberius.¹¹⁴ Even if there was not in reality a flood of new citizens, there was a perception or fear that there might be. And we should not try to explain this perception away. The metaphor is one reflection of these anxieties. Second, the gap between the perception of an undesirable explosion in citizenship numbers and the reality of modest growth is in itself telling. Concern about the declining value of citizenship and close scrutiny of citizenship grants, such as we see in the metaphor, makes most sense in a world where citizenship was not yet excessively common; while citizenship remained to some degree exclusive, there remained an incentive to police its boundaries. It was only in a world where citizenship had become commonplace that such policing might become redundant. In such a world, the metaphor of granting citizenship to words might also become a cliché. But in the world of the first and second centuries, when the metaphor flourished, citizenship — both real and metaphorical — was not to be taken for granted.