Jowett’s Thucydides: A corpus-based analysis of translation as political intervention

Thucydides ’ History of the Peloponnesian War is a key text in the classical Greek canon and an important source of insights into the structures and tensions at the heart of ancient Athenian democracy. Consequently, modern interpretations of his analysis have repeatedly played a major role in shaping debates on the viability and desirability of democratic rule. This paper aims to build on previous discussion of Benjamin Jowett ’ s 1881 translation of Thucydides by applying a comparative corpus-based methodology to explore how this translator ’ s own personal politics shaped his re-presentation of this text. The analysis reveals a striking emphasis on the position and activity of democratic leaders throughout Jowett ’ s version, strongly consistent with the ideology of leadership that he developed during his career as Master of Balliol College, Oxford.


Introduction
First come I. My name is J-W-TT. There's no knowledge but I know it. I am the Master of this College, What I don't know isn't knowledge. (Beeching [1881(Beeching [ ] 1955 Modern interest in the classical Greek historian Thucydides stems in large part from the insights he provides into the successes and failures of ancient Athenian democracy (Mara 2015, 315). Specifically, his master-narrative of the rise and fall of Athens as Greece's leading city-state has repeatedly been used as a "foil" for ideological debates from the mid-seventeenth century onwards, co-opted by proponents of every shade in the political spectrum (Harloe and Morley 2012, 10). His supreme cultural authority has been deployed both in support of greater democracy and as a "devastating critique" of its flaws; as an illustration of the achievements that were possible under a system of equal citizen rights and as a warning against the dangers of "mob rule" (Harloe and Morley 2012, 10;Hoekstra 2012, 26).
This was especially the case in Britain during the nineteenth century when a flurry of new interpretations of the ancient historian's text appeared in direct response to the stirrings of liberal democracy throughout Europe, and to the French and American Revolutions in particular (Turner 1981, 3). In the 1830s, Bloomfield (1830) and Arnold (1835) both published new critical editions with extensive introductions and notes on the Greek; three volumes of Grote's (1846Grote's ( -1856 History of Greece relied heavily on Thucydides as a primary source; Hobbes ([1629] 1843) first Greek-English translation was reprinted five times during this period; and no less than six full or partial re-translations of the work were made between 1829 and 1881 (Lianeri 2002, 12). Many of these re-presentations attempted to impress upon their readers "how short is the distance from the civilized inhabitant of Athens or Corinth to the dweller in London or Vienna" (Crawley [1874(Crawley [ ] 1910, and thus to highlight the relevance of the insights contained in the History for their contemporary society. The present case-study seeks to examine one of the foremost translations from this period, produced by Benjamin Jowett in 1881, and to compare it with those of his immediate contemporaries, Crawley ([1874Crawley ([ ] 1910 and Wilkins (1873). As such, it aims to contribute to a growing body of research which has developed in recent years with the goal of better understanding modern-era interpretations of Thucydides and of picking apart the variations between them. Most notably, this includes Greenwood's (2015) essay "On Translating Thucydides", which provides a valuable methodological framework for the analysis of Thucydidean translations in the form of a multi-dimensional model with which to comprehend the norms shaping the production of each new version of the text. This model encourages researchers to examine the influence of the scholarly context in which a translation was created, the role of earlier readings in guiding later translators' choices, the prevailing "translation culture" and broader expectations of the target context, as well as the individual translator's own aims, interests and working methods. Through a comparative analysis of the multiple renderings of Book 1, Chapters 20-22, Greenwood's (2015, 100-107) article additionally highlights the difficulty and frequent ambiguity of Thucydides' Greek, and therefore, "the difference that translation makes" for the modern reception of this classical author. As she notes, this famous passageand the recurring motif it contains the search for historical truth as physical exertionis "vital for Thucydides' presentation of his credentials as a historian and for the construction of his persona" (2015,104). Yet, Greenwood (2015, 105) argues, due to the complexity of the challenges of rendering this author's dense Greek into more or less accessible English, none of the translations reviewed consistently brings out this metaphor, and a crucial dimension of Thucydides' account of his methods is lost.
A striking omission in Greenwood's otherwise comprehensive discussion is the absence of consideration of the political and ideological factors which might have shaped the production of these translations. Such influences could conceivably be classed within her category of "expectancy norms", reflecting "the prevailing biases in society at large" (2015, 100), but her model does not draw explicit attention to the political uses and abuses of Thucydides (cf. Bagby 1994) and especially the extent to which translations of this author might have been informed by contemporary debates about how the modern world should be governed. The intersection of politics and the translation of Thucydides has been touched upon to a certain degree in earlier work by Beard (2010), Greenwood (2012, Harloe and Morley (2012) and Holmberg (2003), but the ways in which nineteenth-century British translators of Thucydides might have intervened in key ideological debates during this period over the viability and desirability of democratic rule remains largely under-explored. Indeed, to my knowledge, the only study to have engaged on a close textual level with such concerns is Lianeri's (2002) diachronic comparison of six different translations of the History from Hobbes ([1629] 1843) through to Jowett's (1881). Consequently, in the section that follows, I discuss her analysis in detail: this is primarily as a means of introducing the contextual issues at play in the three translations of Thucydides that are the focus of this paper and of explaining my interest in Jowett's version in particular. It also allows me to voice a number of criticisms of Lianeri's approach and, in doing so, to advocate the advantages of a corpus-based methodology over previous approaches to this area of study.

Translation and the establishment of liberal democracy in nineteenthcentury England
Concentrating on each translator's rendering of one of the most frequently cited passages in Thucydides' History (the so-called "funeral oration" given by Pericles in Book 2), Lianeri's central aim is to show how each of these interpretations was shaped by changes in attitudes towards the concept of democracy that occurred during the 250-year period in which these six translations were undertaken. First, she demonstrates how Hobbes ([1629] 1843) version was produced in a socio-political context in which the idea of government by the majority received very negative appraisals: "democracy […] was not an institution that established political liberty and social justice but, rather, a menace to social order and coherence, threatening prosperity and cultural development alike" (Lianeri 2002, 7). Consequently, Lianeri (2002, 5) argues, where the twentieth-century Loeb translator Smith (1919, 323) has Pericles announce (in Book 2, Ch. 37): It is true that our government is called a democracy, because its administration is in the hands, not of the few, but of the many.
Hobbes ([1629] 1843, 191) offers quite a different image of the Athenian polity by rendering the same passage as follows: We have a form of government […] which, because in the administration it hath respect not to a few, but to the multitude, is called a democracy. Lianeri (2002, 5) contends that the democracy of Thucydides' Athens is thus re-framed by Hobbes not as a system of government in which power and authority are bestowed on "the many", but as one in which the administration merely takes into account the views and interests of "the multitude". In her analysis, the possibility that ordinary citizens might have successfully ruled themselves in an ancient Greek city-state is expunged from the text and brought into line with the dominant political ideologies of the receiving culture.
By contrast, Lianeri suggests that Jowett's (1881) translation is the first of the History's interpretations to render "the source-text idea that democracy is a polity by which political power is 'in the hands of the many and not of the few'" (2002,21). This rendering, she argues, was "profoundly influenced" by the thought of cultural critic Matthew Arnold, who openly recognized the potential of widening suffrage as a means of wresting British politics from the grip of the corrupt and politically incapable aristocracy, whilst furthering the interests of the ascendant middle classes (2002,20). Arnold's vision was not, however, based on the principles of social equality and the flattening of political hierarchies; instead, it saw democracy as an institution which would still require the intervention of a higher political authority, a select group of highly educated meritocrats who would be able to guide the actions of the majority and save it from its worst excesses. For this reason, Lianeri (2002, 21) also attributes Jowett's translational choices in another passage of Pericles' speech to Arnold's influence: focusing this time on Chapter 40 of Book 2, she draws our attention to the way in which Jowett's (1881, 119) version introduces a rigid distinction between the active "originators" of policy and the general public, who are involved more simply as passive "judges" on political affairs: An Athenian citizen does not neglect the state because he takes care of his own household; and even those of us who are engaged in business have a very fair idea of politics […] and if few of us are originators, we are all sound judges of a policy.
"This translation", Lianeri (2002, 21) posits, "changes the source text's representation of Athenian citizens as actively engaged in the political institution of the city". As proof of this, we are again invited to refer back to Smith's (1919) "literal" rendering of this same passage. Here, Athenian society is not so clearly divided and indeed the boundaries between the public and private spheres of a citizen's life are deeply blurred: And you will find united in the same persons an interest at once in private and in public affairs, and in others of us who give attention chiefly to business, you will find no lack of insight into political matters […] and we Athenians decide public questions for ourselves or at least endeavour to arrive at a sound understanding of them.
Lianeri's close textual analysis thus reveals important insights into the significance of the changes that have occurred with respect to the interpretation of Thucydides' text and underlines the importance of context-specific political considerations for the study of these translation phenomena.
That said, and with the aim of further developing Lianeri's initial findings, there are a number of criticisms that we might launch at her approach. To begin with, while the scare quotes in her description of Smith's (1919) translation as the most "literal" rendering of the Greek show some acknowledgement of the risks involved in discussing any act of interpretation in this way, Lianeri's repeated use of Smith's version to demonstrate how Hobbes or Jowett have "distorted" the meaning of the source text exaggerates the extent to which Thucydides is at all clear at these points in the text, and therefore, the extent to which Smith's translation can be in any way "literal". In fact, as Hornblower's (1991, 305-306) extensive commentary indicates, the complexity and ambiguity of the ancient historian's language in these passages of Pericles' speech leave open multiple understandings to modern readers: with reference to several other interpretations (including those by Warner 1954;Rhodes 1988;Classen and Steup 1900;and Rusten 1989), Hornblower (1991, 305-306) notes that it is not evident whether Thucydides [is] making Pericles claim that the democracy is actually not fully participatory at all, but run by an elite (Jowett; Rusten)? Or merely that "we Athenians are all of us, by contrast with the sheep-like Spartans, intelligently involved in policy-makingeven when our deliberations result in something less than a definite decision"?
From the source-text evidence of this section of the History, therefore, it is not fully obvious that all Athenian citizens were "actively engaged in the political institution of the city", as Lianeri (2002, 21) suggests. Rather, we must take heed of Greenwood's (2015, 100-107) emphasis on the opacity of Thucydides' Greek, recognize that many, apparently contradictory readings are possible, and argue that what is important here are the reasons why certain translators opted for one rendering while others might have chosen another. 1 Secondly, by framing Jowett's version as the direct and almost inevitable product of Matthew Arnold's ideas, introduced in the late nineteenth century, Lianeri does not acknowledge the fact that Arnold was a former student of Jowett's at Balliol College, Oxford (Gordon and White 2010, 5). Consequently, her analysis overlooks the possibility of the latter's influence on the former from the 1840s onwards and downplays the strength of the translator's own political convictions, deeply held and widely documented throughout his personal letters, biographies and other writings (see Campbell 1897a, 1897b;Quinn and Prest 1987;Tollemache 1904). Indeed, when considering Jowett's translation, it is important to take into account his pre-eminence in Victorian society as Master of Balliol (as the epigraph above satirically illustrates) and his role as one of the progenitors of British idealism during this period (Gordon and White 2010, 5): it was he who first introduced the philosophical study of Hegel and Plato to the Oxford curriculum, having spent the summer of 1844 studying with J. E. Erdmann in Germany, and he who taught one of the first courses in so-called "political economy" (Abbott and Campbell 1897a, 131-132;Allard 2004, 11;Gordon and White 2010, 5). 2 Thus, the belief that democracy can work as a system of government only if controlled by the intervention of a higher political authority should be attributed not solely to Arnold, as Lianeri (2002, 20) indicates, but also to Jowett's own Hegel-and Plato-inspired "ideology of leadership" (Gordon and White 2010, 5;Stray 1998, 122). As Gordon and White (2010, 5) discuss, strong affinities can be drawn between, on the one hand, Hegel's insistence in his Philosophie des Rechts [The Philosophy of Right] on the necessity of establishing "a rigorously educated 'universal class' of civil servants and teachers who were to maintain the spiritual unity of the state against individualistic pressures from below" and, on the other hand, the way Jowett saw it as his "duty" throughout his career "to guide the leaders of the people and to train their teacherscustodire custodes et docere doctores" (Tollemache 1904, 50-51). His pupils were to be "Platonic guardians for Britain and its Empire" (Stray 1998, 122), whose education, common sense and ability would counterbalance what he considered to be the "unfitness of the lower classes to govern themselves" in the emerging democratic society (Jowett 1882;cited in Abbott and Campbell 1897b, 210). Only in this way, Jowett expresses in his letters, might the democracy that he understood as "the politics of the future" be prevented from developing into anarchy (Balliol Archivist 2017). It is one of the aims of this paper, therefore, to investigate Jowett's translation not simply as a reflection of Arnold's thought, but more importantly as part of the "high-profile public discourse" on the continued importance of effective popular leadership that the translator developed throughout his career at Oxford (Stray 1998, 122). In other words, his Thucydides is framed here as part of a series of deliberate interventions within the key political debates of the time by a specific and engaged individual.
The third issue with Lianeri's analysis relates to the fact that, through her focus on broad historical shifts in attitudes to democracy, Lianeri largely fails to acknowledge the continued ideological heterogeneity of the society in which Jowett's Thucydides was produced. To be precise, Lianeri appears to promote the view, critiqued by Brownlie (2006, 156), "that there are different time periods each with a different set of norms/ideologies, which explains the changing characteristics of translations". This impression is heightened by the way in which she mentions only fleetingly Crawley's ([1874] 1910) version of the History and entirely omits all reference to Wilkins (1873) partial translation of the work's key speeches (including that of Pericles). Indeed, by neglecting to compare Crawley's, Wilkins' and Jowett's texts directly, Lianeri downplays the extent to which these constitute near-simultaneous alternative readings, produced in virtually the same cultural context (cf. Pym 1998, 82). By reading Jowett's letters, for instance, we find that although the work was published in 1881, the Master of Balliol began his translation in 1871 and completed much of the work in 1875, that is, just a year after Crawley published his version (Quinn and Prest 1987, xxviii). Moreover, collecting biographical details on Crawley and Wilkins reveals that, like Jowett, these two translators were also based at Oxford colleges: Crawley was a fellow at Worcester College between 1866 and 1880 (Lee and Chubbuck 2004), while Wilkins was a fellow and later librarian at Merton between 1848 and 1887 (Foster 1891(Foster , 1554. Interestingly, however, and especially so given these contextual similarities, major differences can be observed between the three translations, differences which cannot merely be attributed to questions of style (cf. Willett 1999). Therefore, through a comparison with Crawley's and Wilkins' translations, this paper additionally aims to highlight the unbounded possibilities for different readings and renderings of Thucydides' source text during a period of great social, political and economic upheaval in Victorian Britain, as well as the agency of individual translators who do not necessarily subscribe to the dominant ideologies of the day (cf. Brownlie 2006, 156).
The final critique of Lianeri's analysis of Jowett's translation relates to the fact that her comparisons are essentially limited to two short passages of Thucydides' voluminous text, both taken from one speech delivered by Pericles in Book 2 (Ch. [35][36][37][38][39][40][41][42][43][44][45][46][47]. This choice of dataset is entirely justified, given that it is in this funeral oration that Thucydides provides some of the most explicit definitions of the values and beliefs of the ancient Athenian people, especially with respect to democratic forms of government. Nevertheless, it is difficult to support any attempt to make general claims about any of these translations on the basis of just a few phrases extracted from this 220,000-word History. It is for this reason that the third objective of this paper is to make a case in favour of computer-aided corpus-based approaches to the study of Thucydides in translation. As I will explain in more detail below, the use of electronic concordancing software can help us to identify and investigate with significantly greater ease differences between and patterns within translations, taking into account the full length of each work as a complete text. For example, searching for the keyword "democracy" within a digitized copy of Jowett's translation will retrieve not only the passages on which Lianeri concentrates, but fifty six other uses of this word as it appears throughout this whole volume. Moreover, the researcher need not limit him-or herself to the investigation of a single term but can explore the use of lexical items associated with a much wider constellation of interconnected concepts related to democratic political structures and discourse: citizenship, sovereignty, rights, equality, etc. Patterns in each translator's interpretation of the nature of Athenian democracy can thus be discussed not merely with respect to specific passages of specific speeches, but as clear indications of their overall presentation of the source text and culture. These findings may then be used to support, complicate or challenge Lianeri's (2002) initial argument regarding Jowett's text.
Situating the current study: The Genealogies of Knowledge project A corpus is a collection of texts, selected according to a specific set of criteria and held in electronic format (Olohan 2004, 1). The use of corpora within the field of translation studies now constitutes a well-established area of research and has already proved highly fruitful with respect to a number of concerns within the discipline, including not least the study of translator's style (Baker 2000;Saldanha 2011;Winters 2007) and the identification of patterns specific to translated text (Baker 1993;Olohan 2003;Olohan and Baker 2000;Tirkkonnen-Condit 2004). As one of the largest and most ambitious ventures in this direction to date, the AHRC-funded Genealogies of Knowledge (GoK) project is innovative for its shift in focus towards using corpora to investigate the role of translation and other sites of re-mediation in the evolution, transformation and contestation of political and scientific discourse across languages, cultures and centuries. 3 Specifically, a core objective of our research programme is to build a set of large and diverse corpora containing as many different translations, re-translations, commentaries and critical editions of texts relevant to these themes as possible, produced at different times, in different languages and by different writers. We then provide the research community with restricted access to these resources from remote locations via a specially developed suite of online corpus analysis tools, available on our project website. 4 This software includes a Keyword-in-Context (KWIC) concordancer as well as a series of plugins which allow for the graphical visualization of linguistic patterns (Luz 2011;Luz and Sheehan 2014).
In terms of corpus design, the starting point for the GoK project is the (re-)interpretation of Classical Greek thought in three historical lingua francas: Medieval Arabic, Classical and Medieval Latin and Modern English. 5 Only texts produced in these four languages, during these four time periods and of relevance to the development of two constellations of concepts are currently selected for inclusion in our corpora. The first constellation relates to the body politic, includingbut not limited tosuch ideas as polis, polity, democracy, civil society, citizenship, nation, state, natural law and human rights. The second consists of concepts that underpin modern science: e.g. experiment, observation, evidence, proof, episteme, truth, falsehood, aetiology, causation, justification, fact, validity and expertise. By comparing and contrasting the ways in which these concepts are discussed, re-interpreted and contested in the corpus texts, we thus hope to shed new light on key moments of change in the development of these ideas as they have travelled across time and space.
As argued above, modern translations of Thucydides are widely acknowledged as having played an important part in the development of many political concepts such as democracy and citizenship in particular and, for this reason, ten English versions of Thucydides have now been included in the GoK Modern English corpus (see Table 1). These include Jowett's (1881) text and footnotes, as well as those of his two direct contemporaries, Crawley ([1874Crawley ([ ] 1910 and Wilkins (1873). It is additionally worth mentioning that this corpus also contains multiple re-translations of the writings of Plato, Aristotle, Herodotus, Hippocrates and Plutarch, many of which will form the focus of other casestudies (see Jones 2019).

Jowett's ideology of leadership and the functioning of a democratic state
Given that my aim in this paper is to explore the extent to which Jowett's own ideology of leadership might have informed his translation of the History, a clear starting point from which to begin the analysis is to generate a list of keywords relevant to the different participants involved in governing the city-states of Thucydides' Greece. This can be done using the GoK concordance browser's built-in Word Frequency Lists tool to produce a table of the 1,000 most commonly occurring lexical items in each of the three translations. Having merged these three lists in a Microsoft Excel spreadsheet, all of the nouns that could be used to refer to the members of Greek society, whether individually or collectively, can then be selected out of this list, producing the list of keywords shown in Table 2.
Once the file containing Jowett's translation (mod000019) has been selected within the GoK subcorpus selection interface, each of these nouns can be entered into the search field of the browser and the concordances retrieved. The same operation can then be carried out on Crawley's (mod000020) and Wilkins' (mod000148) texts, and a series of both quantitative and qualitative analyses can be made by comparing patterns within each version.
The most straightforward of these comparisons involves examining the extent to which they each deploy certain terms to describe the unnamed protagonists of Thucydides' History in preference to a range of possible alternatives. Such an analysis is greatly facilitated by means of corpus tools because of the way that they are able to retrieve and display  all instances of a keyword at the click of a mouse. In a split second, the software thus delivers information which would have taken the researcher many months to gather manually. These quantitative results should, of course, be regarded with a healthy degree of suspicion, especially given that some of the keywords listed are polysemous and so could be used in ways other than to describe protagonists in the History (e.g. "judges" may feature both as a noun and as a verb). It should also be made clear that direct comparison between the figures retrieved for Jowett and Wilkins' translations would not be valid given that the latter is only a partial translation. Even with these reservations, some intriguing patterns do still emerge through this lens that can be followed more closely through qualitative analysis. Indeed, as Table 3 shows, these statistics would suggest that Jowett uses nouns associated with individual members of a political elite with strikingly higher frequency than Crawley does: the keywords "magistrate(s)" and "leader(s)" in particular are each found twenty one and forty one times respectively in his text, against eleven and twenty five in his rival's. It is also interesting to find that the keyword "council" appears more often in Jowett's version (twenty eight hits against twenty) as does "chief" (forty one hits against thirty six). By contrast, if we look at the bottom half of the table (i.e. at those terms that are relatively more common in Crawley's text), Crawley appears to prefer more abstract terminology to describe the different layers in the political structure of ancient Greek societies, selecting with greater frequency nouns such as the "masses", the "commons", the "government", the "authorities", the "few" and the "many". 6 This preliminary analysis would, therefore, suggest that Jowett's language use presents a more specific and explicit set of roles for individual actors in democratic political systems in comparison with his contemporary.
In order to be able to draw firmer conclusions from this data, however, we must delve deeper into the co-text surrounding each instance of each keyword retrieved. In this more qualitative mode of analysis, comparison with Wilkins' text is valid given that the length of each text is no longer at issue: it is no longer a question of how many times each translator uses a term, but in what manner it is deployed. Indeed, by comparing the ways in which Wilkins and Jowett each use the synonyms "leader" and "chief", we can observe that the former mostly selects these terms to refer to the relations between city-states in a confederacy. As the concordance pasted below shows (Figure 1; lines 1, 2, 4 and 6), the "leaders" to which Wilkins refers are in four out of seven cases either Athens or Sparta, both of which are thus presented as occupying a privileged position of political, economic and military dominance over their respective sets of allies. Out of the exceptions (Figure 1; lines 3, 5 and 7), only line 7 refers to a leader operating within a democracy (in this case Athens), Melos being an oligarchy and Sparta a form of joint monarchy. Similarly, Figure 2 shows that "chief" is mainly used as an adjective synonymous with "principal", and in those few cases in which it refers to an individual or social group, this is overwhelmingly in relation to Athens or Sparta as heads of the two opposing confederacies (e.g. lines 14, 16, 17 and 18).
The concordance in Figure 3 shows that Crawley's text makes more regular reference to specific individuals acting as "chiefs" or "leaders" with a certain power and influence over their fellow citizens. That said, it is interesting to find that this is only ever in connection with "the multitude" or "the commons", i.e. the common people, a specific class of citizens in society (e.g. "Androcles, the chief leader of the commons"line 9). Thus, Crawley's and Wilkins' translations appear to contrast strongly with Jowett's in this respect: through his word choices, the Master of Balliol seems to go out of his way to foreground the presence of "leaders" within the democracies of ancient Greece, and their influence within this political system as a whole. Unlike his contemporaries, we see in Figure 4 that he explicitly writes about "the leaders of the Corcyraean democracy" (line 18), "the leader of the [Athenian] state" (line 28), "the leaders of the [Athenian] state as a whole" (line 29) and "the Athenian leaders" (lines 5 and 39). Searching for the keyword "leadership" adds a further two unique collocations of this kind: "the leadership of the democracy" and "the leadership of the people". Taking into consideration Jowett's ideology of leadership, as discussed above, this textual pattern is unlikely to be a coincidence; rather, it adds further weight to the suggestion that Jowett's interpretation of Thucydides' text was strongly shaped by his own political beliefs regarding the need to maintain strict political hierarchies in the face of shifts in his contemporary Britain towards greater democracy.
The full significance of this general pattern is most convincingly illustrated by focusing on individual concordance lines which stand out as containing particularly unusual collocations and comparing more closely each text as regards their translations of a specific  section of the source. 7 Indeed, contrasting the passage (Book 6, Ch. 89) in which Jowett (1881, 472) uses the phrase "leadership of the people" with the corresponding sections of Crawley's and Wilkins' versions reveals a striking difference in each translator's interpretation of the Greek phrase "ἡ προστασία […] τοῦ πλήθους". On the one hand, we find that Wilkins (1873, 238) translates this section of a speech by Alcibiades in a way that drastically limits this character and his family's influence within Athenian democracy. According to his interpretation of "ἡ προστασία […] τοῦ πλήθους", these protagonists are no more than "persevering champions of popular power", opposed to despotic government like the majority of their fellow citizens. Below is the Greek source, taken from the edition used by Jowett and Wilkins (Poppo 1828, 189), followed by Wilkins (1873, 238) translation: 8 τοῖς γὰρ τυράννοις αἰεί ποτε διάφοροί ἐσμεν (πᾶν δὲ τὸ ἐναντιούμενον τῷ δυναστεύοντι δῆμος ὠνόμασται) καὶ ἀπ᾽ ἐκείνου ξυμπαρέμεινεν ἡ προστασία ἡμῖν τοῦ πλήθους.
in common language, popular government is habitually opposed to despotic government. It was thushatred of the one engendering partiality for the otherthat we continued the persevering champions of popular power.  Jowett's (1881, 472) version, on the other hand, provides a rather different rendering: Any power adverse to despotism is called democracy, and my family have always retained the leadership of the people in their hands because we have been the persistent enemies of tyrants.
Specifically, the way in which Jowett expressly frames Alcibiades and his family as having "retained the leadership of the people in their hands" implies a much more permanent and legitimate role for a political elite in democracies. The "προστασία […] τοῦ πλήθους" is discussed as a position of leadership that pre-exists Alcibiades and his family's occupancy, as a formally established and necessary feature of this system of government. Crawley's ([1874Crawley's ([ ] 1910 Alcibiades, by contrast, holds only a much more temporary status as one who "acts as" or like a leader: We have always been hostile to tyrants, and all who oppose arbitrary power are called commons; hence we continued to act as leaders of the multitude. Certainly, the meaning of ἡ προστασία is ambiguous, given that the noun can be taken both in the sense of "standing before" and "standing up for", in this case, the mass of the people or multitude (τοῦ πλήθουςsee e.g. Liddell and Scott 1889, 697). Indeed, as Ober (1989, 316) notes, the related term προστάτης meant not only one who stood first in the eyes of the people and who physically stood before the people in order to address them, but [also] one who interposed himself between the people and the dangers that threatened them.
Nevertheless, according to Connor (1971, 113), we should not assume that ἡ προστασία constituted any kind of fixed, formal or otherwise uniquely privileged status in Athenian society: an individual referred to as a προστάτης did not enjoy a monopoly of power in the city, nor did their perceived standing in the hierarchy ever remain even temporarily unchallenged (1971,113). Thus, while Jowett's "leadership of the people" is to some extent an acceptable translation of ἡ προστασία, it can be seen to add a level of formal recognition and long-lasting influence to this role in Greek politics which is largely absent from the source-language text, as well as from the alternative renderings provided by Crawley and Wilkins. Looking at a second concordance line which, it turns out, is also extracted from this speech (Book 6, Ch. 89), we find the same pattern of divergent interpretations of the nature of ἡ προστασία repeated. Where in the Greek (Poppo 1828, 190), Thucydides has Alcibiades say "ἡμεῖς δὲ τοῦ ξύμπαντος προέστημεν" (προέστημεν being the first person plural intransitive aorist form of the verb προίστημι with which προστασία shares a common stemsee Liddell and Scott 1889, 677), Wilkins (1873, 238) has Alcibiades say "[m]y family supported the cause of popular government" and Crawley ([1874] 1910, 463) writes "our party was that of the whole people". Clearly, then, these two translators understand "ἡμεῖς δὲ τοῦ ξύμπαντος προέστημεν" to mean "we stood up for the whole" (i.e. as protectors and supporters of everyone altogether) rather than "we stood before the whole" (i.e. as leaders of everyone altogether). Jowett (1881, 472), on the other hand, again consciously or unconsciously exploits the ambiguity of the Greek to foreground the existence of a political hierarchy in Athens and the supreme position of power occupied by Alcibiades' family: he has the Athenian noble announce "we were the leaders of the state as a whole, and not of a part only". His translation stands in marked contrast, therefore, with those of his contemporaries, neither of which suggests these characters were able to exert direct control on policy making, over and above the popular assembly.
Importantly, it is not only Alcibiades and his family who are presented by Jowett in this way. Looking at the other lines in the concordances generated above for the keyword "leader" (Figures 1-4), we find that Pericles too, for example, is introduced as "the leader of the state and the most powerful man of his day" (Book 1, Ch.127 -1881, 79). The centrality and official legitimacy of this individual's agency in the running of the democratic state is thus once more portrayed as unambiguous; he is firmly placed at the top of the political hierarchy, performing a unique "leader" function which is made to appear a permanent and integral feature of the Athenian constitution. Crawley ([1874] 1910, 82), on the other hand, offers a quite different perspective, staying closer to sense of the original Greek ("ὢν γὰρ δυνατώτατος τῶν καθ᾽ ἑαυτὸν καὶ ἄγων τὴν πολιτείαν" -Poppo 1825, 380) by describing Pericles instead as "the most powerful man of his time, and the leading Athenian statesman". Unlike Jowett's, his rendering does not preclude the existence of other powerful Athenians, nor does it suggest Pericles enjoys anything more than a certain degree of temporary influence over his fellow citizens. Crucially, in Crawley's account, the participle ἄγων is not fully nominalized into "the leader"; Pericles is only the "leading" statesman, an individual who is particularly prominent in state affairs at this specific point in time.
As for Wilkins' text, direct comparison is not possible in this case because, as discussed previously, his translation focuses only on the key speeches contained within the History and therefore, he omits this passage entirely. Yet, by searching for Pericles' name in the GoK corpus browser, it is interesting to discover that not once does Wilkins use the term "leader" in connection with this character (see Figure 5). Instead, in his paratextual introduction to Pericles' speech in Book 2, Chapter 60 (line 4), Wilkins (1873, 78) presents the Athenian as someone who is merely involved in the "administration of public affairs": The philo-Laconian party at Athens had thus been encouraged to denounce the policy of Pericles, who, in the following speech, vindicates his administration of public affairs, and endeavours to reassure the failing resolution of his audience.
As was the case with Wilkins' characterization of Alcibiades, this interpretation significantly limits Pericles' influence within Athenian democracy and, in comparison with Jowett's text, provides the reader with an alternative impression of the extent of his political agency. Finally, it is important to note that there are of course exceptions to this general pattern. For example, at the end of Book 8, Chapter 89 of Thucydides' History, we find that it is Crawley, not Jowett, who translates the noun προστάτης by "leader" and that it is Jowett in this case who chooses the term "champion". Specifically, in his rendering of the sentence "ἠγωνίζετο οὖν εἷς ἕκαστος αὐτὸς πρῶτος προστάτης τοῦ δήμου γενέσθαι" (Poppo 1828, 644), Crawley ([1874Crawley ([ ] 1910 writes "it was now a race between them as to which should first become the leader of the commons", while Jowett (1881, 610) suggests instead "[a]ccordingly every one was struggling hard to be the first champion of the people himself". On the face of it, then, this particular comparison of these two translator's texts would appear to contradict the argument constructed in this article, namely, that Jowett's version tends to exaggerate the role of strong leaders in the democracy Thucydides describes. However, the important point here is that, although certainly curious, such examples do not dismantle the validity of the claims presented so far, but rather serve to further emphasize the value of a corpus-based methodology for the study of multiple re-translations. As argued above, concordancing tools greatly facilitate the identification and comparison of different patterns of interpretation between rival versions of the same original work across the full length of each targetlanguage text. They thus allow the researcher to move beyond analyses concentrated on one or two specific passages only, and to develop an understanding of each translator's overall presentation of the source text and the culture of which it is a part. While analysis of this one passage (Book 8, Chapter 89) alone might give the impression that Crawley's text more clearly emphasises positions of leadership, examination of concordances for the keyword "leader" as it is used throughout Jowett's and Crawley's texts reveals the former's word choices here to be an intriguing exception rather than the general rule.

Summary and conclusions
This paper has attempted to build on Lianeri's (2002) initial analysis of Jowett's translation of Thucydides' History by means of a comparative corpus-based approach. By contrasting Jowett's translation with Crawley's and Wilkins' contemporaneous interpretations, it has sought to show how corpus tools can greatly facilitate the identification and investigation of differences between and patterns within multiple re-writings of the same text. The analysis has revealed a striking emphasis on the position and activity of democratic leaders throughout Jowett's text, strongly consistent with the ideology of leadership that he developed during his career at Oxford. My findings can thus be said both to support and complicate Lianeri's discussion of this translation: on the one hand, they show that the way in which Jowett translated Pericles' funeral oration was by no means an exception within his text as a whole and indeed that similar hierarchical divisions within Greek democracy are brought to the fore at many different points in his version of the History. On the other hand, my investigations into Jowett's biography and my comparisons with Crawley's and Wilkins' translations have highlighted the need to consider these re-framing devices not simply as a product of the influence of Matthew Arnold on late Victorian society as a whole, or a product of the general ideological climate of the period, but also as part of a series of interventions by a specific and highly politically engaged individual. online by radical activist groups and counter-hegemonic globalisation movements. Unlike the premodern and modern corpus texts, these are not for the most part translations of content produced elsewhere, so our interest in these discourses is directed instead towards the role of English as lingua franca for the production and contestation of political and scientific knowledge in the internet age. 6. In order to help distinguish between uses of "few" and "many" that do and do not refer to the political elite and the masses respectively, the keywords were in these cases entered into the search field alongside the definite article: "the + few" and "the + many". 7. Unlike some online tools (e.g. Perseus Digital Library), the GoK corpus browser does not allow for the alignment of different translations. That said, parallel passages can still be easily found by searching through the concordance lines generated for similar keywords across the three translations. 8. It is not clear from the introduction to Crawley's translation which edition of the source he relied on. While unfortunate, this is not problematic for our analysis here given that there are no differences at this point in the History between Poppo's (1828) edition and any of the other likely source texts for Crawley's version (Frost 1854, 92;Arnold 1846, 420). The same is true of all the other passages on which the last section of my analysis focuses.