ARCHAEOLOGY VERSUS TACITUS’ AGRICOLA, A FIRST CENTURY WORST CASE SCENARIO
A lecture given to the Theoretical Archaeology Group conference, held in Dublin, 15th December, 2001
Birgitta Hoffmann
In 1425 Poggio, the Pope’s secretary and book-collector in Rome, got a letter from the monastery of Hersfeld in Germany, informing him that after checking his list of desiderata against the books preserved in their library, a volume of hitherto unknown works of Tacitus had been identified. In the ensuing correspondence it became clear that these ‘new’ works were the Dialogue on the Orator, the Germania and the Agricola.
It took another 30 years for this manuscript to make its way from Germany to Italy, where it seems that, from the 1470s onwards, a number of people made hand-written and printed copies, before the original manuscript vanished, only to partly resurface in the early 20th century in Iesi in Northern Italy.
Given the problems that the Germania and Agricola in particular have caused, there must be quite a few Romanists, who wish that the manuscripts might have gone astray in the post en-route to Rome, but instead the Agricola has risen to become one of the most read volumes of Roman history in Britain.
The text itself is, according to most of its editors, a form of funeral eulogy for Tacitus’ father-in-law. It is full of the ingredients of a gripping historical tale: a good story, a believable main character, and a battle scene, indeed – only the happy ending falls far short of expectations. The same ingredients now make ‘Meet the Ancestors’ such a success on television, and we should not be surprised that from at least the 18th century onwards, it became a popular pastime of vicars and military personnel to try to find the location of the reported events, usually by associating them with ‘Roman antiquities’ in the fields of their parishes.
By the nineteenth century Agricola had made it into the children’s book section in a walk-on-part in Puck-of-Pook’s hill and into folk songs, still popular today among Romanists and certain Northumbrians alike. More importantly he had become a stock character and the stuff that school books are made of and it now seems impossible to write a history of Britain without him. Agricola’s musings on invading Ireland even ensured him an honourable mention in textbooks on the Irish Iron Age.
In the period from 1963 to 1981 a series of comprehensive histories of Roman Britain appeared. All gave a general account of the history and archaeology of the province and all have become core text books for Roman Britain courses. The first, Ian Richmond’s Penguin history, treats the Agricola as a historical account and reproduces its story-line more or less unchanged, usually without reference to archaeology.
Sheppard Frere’s “Britannia” is probably the longest account and includes the historical data, rarely critically assessed, but with some added ideas about the chronology and archaeology. The latter are usually hedged with ‘perhaps’, ‘possibly’ and any number of subjunctive clauses, suggesting a possible association but staying clear of straightforward equation.
Peter Salway’s “Roman Britain” is slightly more critical of Tacitus, but still reproduces the main sequence of events. The archaeology is now used to provide clear indicators as to where the action actually took place, but still some doubts are raised.
The peak of this development was reached with M.Todd’s Fontana History of Roman Britain of 1981. In eight pages he gives us a seductively written account of Agricola’s exploits following Tacitus and sometimes exceeding him, as when he provides a description of the weather. The archaeology is seamlessly integrated into the narrative. Most doubts are stilled. The locations are clearly identified and any surviving ‘shoulds’ are easily overlooked. Indeed one wonders, why nobody made a film out of this: ‘General of the Glens’, the lead played by Oliver Reid or even Russel Crowe and accompanied by beautiful long-panned views of the Highlands.
With this background it is hardly surprising, to find that Agricola is probably the second best known Roman in Britain and that lecturers find it increasingly hard, to put the ‘critical’ back into the student essays on the topic. But is this really, how the history and archaeology of the late first century should be handled? What has in effect happened is that historiography has been culled for the dates and events whilst archaeology has been mined for the places and material. Together a beautiful story emerges. Archaeology the handmaiden of history, history the quarry of archaeology and all in one nicely packaged story.
As a result, research in the last 20 years has often been concerned with the ever finer differentiation of the picture. ‘Perdomita Britannia, statim missa’, well of course, this must mean a quick and sudden evacuation of Northern Britain…and the numismatic research obliges us with the date: 86 AD, or perhaps early 87. The large temporary camps leading up to Durno must surely, we are told, be associated with the run-up to Mons Graupius, and hey presto: a series of large marching camps all dated to exactly 83/4 AD. Season-exact dates, who could possibly beat that? Perhaps if we could go back to the text, or more likely the Penguin translation we could reread it again and get the exact time of day it all happened…….
The reader must excuse the sarcastic tone here, but what I have just described is in fact the worst case scenario for our discipline. It is seductive and easy, and unfortunately wrong, and it happened because historians tried to work with archaeological data and archaeologists with historical sources when both were unaware of the limits of the others’ material.
I believe there is a better way of approaching the evidence, but it demands either the co-operation of specialists from different disciplines or the thorough training of students in both subjects. After learning about C14, dendrochronology and other archaeosciences, archaeologists have in the last 30 years come to understand the value of specialist knowledge and the need to secure it for their own work. There is one field, however, where this has not been the case: Historical archaeology.
In the long bygone days of Haverfield and co. Roman archaeologists tended to be trained historians and often also philologists. They learned to deal with texts and the relevant critical apparatus and assumed that most people interested in the subject would read the texts in the original and understand the finer nuances. Archaeology was for getting your hands, or more to the point – somebody else’s hands, dirty in the summer. Archaeological methodology was often in its infancy – context was something that happened to Tacitus, but not on site.
Since then the situation has changed somewhat. Both subjects have developed their own very specialist theories and methodologies to deal with their source material. Some of them require in-depth training and are not easily accessible to the uninitiated, but are nevertheless essential. A proper inter- or rather cross-disciplinary approach is rare, but when it happens it can be extremely fruitful.
So let us see what we can do when we disentangle the different disciplines and ask the specialists to make their own independent contributions to the study of late first century AD Britain.
Let us begin with the archaeology. Traditionally, and under the influence of the texts the first century history of the North of Britain has always been portrayed as follows. Until the late sixties northern England was only partially integrated into the province of Britannia, while most of it was run as a semi-independent client kingdom under Queen Cartimandua of the Brigantes. At the end of Nero’s reign the kingdom became unstable and eventually the Queen had to be rescued by Rome. After a brief violent exchange the client kingdom was incorporated into the province. At this point Agricola became governor of Britain, and it is traditionally him, who is credited with the original foundation of any site north of a line from Chester to York.
Interestingly Tacitus’ Agricola is far less explicit about Agricola being the first governor in this area than a lot of his commentators and fans, think it is, and careful reading should have prepared the archaeological community for what was to come.
In the late 1970’s excavations in Annetwell Street, Carlisle produced large timbers, which when dated by dendrochronology produced felling dates of 72/73 AD. The early date was corroborated by finds of early Flavian Samian pottery and coins from elsewhere in the city, including the fort. Together, this left very little doubt that here was one fort, that was definitely built before Agricola became governor in 77 or 78 AD. Having established such a date for one site and the cultural assemblage to go with it, further studies quickly showed, that other supposedly ‘Agricolan’ bases like Ribchester, Blennerhasset, Castleford and possibly Papcastle produced similarly early material.
David Shotter has since singled out a number of forts even further north that have also produced comparatively large numbers of early coinage, pointing to activity at the sites in the early seventies AD, including Castledykes, Newstead, Camelon, Strageath and Cardean.
Indeed, only Camelon and Strageath have produced equally high levels of Domitianic coinage, suggesting that at these five forts, occupation may have started much earlier than is traditionally assumed.
Nor is the coinage the only finds’ group to produce unusually high levels of early material. The site of Drumquhassle near Loch Lomond has yielded Terra Nigra, a pre-Flavian and very early Flavian type of pottery that was imported from Gaul. At sites like Carlisle, Newstead, Inchtuthil and Castleford glass vessels have been found which are 20 years behind the times for an assemblage of the 80s AD. Traditionally these have been associated with heirlooms or traders dumping old-fashioned material on less discerning customers in the North, but in view of the Carlisle date there are easier ways of explaining this glass.
It might be argued that a lot of this material culture made its way into the North as the possessions of British civilians living on the site of the later forts, and thereby ‘rescue’ the traditional account, but unfortunately, there is yet another strand of conflicting evidence. In the last five years the work of the Roman Gask Project and others north of the Antonine Wall, have discovered multiple structural phases on a number of Flavian military sites. The phenomenon is best known from the series of fortlets and watch towers on the Gask Ridge between Doune and Bertha, but recently two distinct building phases have also been recognised in the fort of Cardean and there are indications that Strageath may also had an earlier phase. All this suggests that the period of occupation in the North was substantially longer than the ‘traditional’ short chronology envisaged in the aftermath of Agricola’s victory at Mons Graupius, instead the associated material suggests that occupation started well before the time suggested by Tacitus.
This archaeological evidence must subtract from what we have traditionally seen as Agricola’s achievement, for it now appears that a number of northern sites, from Manchester to Strathmore, may already have been in occupation during the early 70s. The fact that these dates have been acquired purely by the use of dendrochronology and the analysis of the surviving archaeological record, show that the information in Tacitus’ text presents difficulties. They do not mean that Agricola never existed or that he just sat in London twiddling his thumbs, but that the story told in the Histories of Roman Britain certainly needs re-evaluating.
Having established that simply combining the text and the archaeology in the manner of Frere, Salway and Todd is no longer possible; the next question is, what is wrong with the text? Is Tacitus actually lying?
So far we have dealt with the Agricola as an interesting framework providing useful chronological data and historical snapshots, but this is not what the Agricola was written for. First and foremost the Agricola is a piece of Latin literature; it is not a simple list of important events in Britain from 77 to 84. Are we perhaps trying to do the equivalent of reconstructing French history with the help of The Three Musketeers?
Under the circumstances, it would probably be best to start with the text itself: Earlier I made a cheap jibe about the Penguin translation, which is not fair to the translator who on the whole did a fine job, but it does give us a text that looks nice, easy and so beautifully uncomplicated whereas, unfortunately, the reality is very different. As already stated we only have one source for the text (which now only partly survives) along with a number of complete 15th century copies. There are quite a few instances where the Latin does not make sense and is obviously wrong, or where bits of a sentence have been omitted, and some of the time these omissions are already present in the 10th century manuscript. Things like that happen in manuscript traditions, but they are a nuisance.
The problem is made worse by some of the 15th century copyists who thought that they knew how to improve the Latin of the original and, as is the case with certain computer spell-checkers, they actually made it worse. But that also is the normal fate of manuscripts and because of it, the original Latin text editions have large sections, called the critical apparatus: explaining why the editors reconstructed the text in the way they did and what other options there are. If the manuscript tradition is bad, one can easily find half of the page filled with the alternative readings and suggestions. Unfortunately the first thing that happens in bilingual versions like the Loebs is that most of this apparatus vanishes and, if you use a translation, there is usually no way of knowing that there were problems with the text in the first place.
After this introduction, it is a relief to know that there are experts on the case who are trying to improve the text of Tacitus as much as they can. But they are very busy people and they don’t talk much to archaeologists, and you meet them about as commonly on excavations as you meet archaeologists in the manuscript and rare print rooms of libraries. Going back to the original text or – in the absence of a degree in Latin literature – talking to someone who can explain when the text is corrupt and when the evidence is sound, should be essential to anyone dealing with the Agricola from an archaeological perspective. But strangely, it only rarely happens.
Less straightforward is the question of what the Agricola actually is. Most commentators tell us, that it is an obituary or rather a hagiography on Tacitus’ father-in-law. Well actually this is not what Tacitus says he wants to write. Introductions are probably the most under used sections of any book, but that makes them no less essential: in his introduction to the Agricola, Tacitus bemoans the virtues of the good old times when good men did great things and then wrote about them themselves, and how under Domitian it had become an act of treason to do this. He, as a sign that Domitian’s rule is over, is, therefore, going to provide a great man who did great things in bad times with a biography, an autobiography being out of the question, as the gentleman, Agricola, is dead. One of the best known ‘autobiographies’ of Tacitus’ good old times, are Caesar’s Commentaries on the War in Gaul and interestingly, Tacitus makes numerous direct comparisons between Caesar and Agricola. So it seems, that besides writing a hagiography, he is writing a comparative biography, Agricola and Caesar in comparison, to show how times have changed for the worse. And he makes it clear that this is a moral tract – not a historical essay.
From the 15th century onwards, people have found Tacitus hard-going. It has long been recognised by literary experts that he is full of style and not very interested in facts. In 1969 Wellesley criticised the Agricola as ‘Rhetoric wins an easy victory over fact’ and one of my old tutors summarised the situation as ‘Tacitus never let a fact get in the way of a good phrase’. People might say that this is an exaggeration, but a few years ago Tony Woodman, one of the leading scholars on Tacitus, identified a battle in the Annals, that Tacitus had copied from himself at another point in his historical account, to make the point of history sometimes repeating itself. And this case is not unique:
One of the very few other sources that mention Agricola is the historian Cassius Dio. He reports that when Agricola was governor in Britain, there was some fighting and that he then proved, by sailing around it, that Britain was an island and that for this reason Titus accepted his 15th acclamation as imperator in 79AD. Tacitus also mentions the circumnavigation of Britain in the Agricola, but he conveniently dates its successful conclusion to the end of the seventh season as the crowning achievement, to happen at the same time as the decisive victory at Mons Graupius in 83/84AD. So we don’t get quite the truth, as the circumnavigation really happened four years earlier. Instead by shifting the date Tacitus creates a much more satisfying effect, on a par with some of the best Hollywood blockbusters.
The problem is: if he does this once, how secure is the rest of the account? I have already mentioned the Agricola/Caesar parallels. Apart from the straight comparisons, the text is full of allusions to the Commentaries. Both last seven seasons. Both start off with a double blitzkrieg, Caesar against the Helvetii and Ariovist, Agricola with the Ordovices and Anglesey. Both consider invading an island at roughly the same time in their career and both describe a night attack on a legion, where the general saves the situation at the last moment. And lastly both culminate in a final battle preceded by great speeches about civilisation and freedom. One recurring event may have been coincidence, but the pattern here is too persistent. What seems to be happening is literary imitation at a very sophisticated level. For those familiar with the Commentaries, it would be quickly apparent, but the trick is to make the story believable even without the background. Some readers may be familiar with Terry Pratchett’s “Wyrd Systers” alluding to “Macbeth” and the classical references in the film ‘Oh brother where art thou’. In both cases the imitation is easy enough to see, but not necessary to the enjoyment of the book or film, and both have a twist at the end. Far from claiming that Tacitus invented every detail, therefore, I would like to suggest that he might have rearranged the order of events to create a more satisfying story-line. The seven years of governorship may have given him the idea in the first place, some of the battles may really have happened, and we know that Agricola was credited with showing that Britain was an island, but the events may not necessarily have happened in the order or on the scale portrayed.
Wellesley said in 1969 that it is no exaggeration to say that whole tracts of the narrative have become meaningful only thanks to the activity of the archaeologist. To judge from the record a number of people used this as their cue to try to prove Tacitus right. I believe the history and archaeology of the late first century AD in Britain would be more interesting for everybody concerned, if we ‘let go of Tacitus’ hand and started behaving like grown-ups’ as David Woolliscroft has put it, using whatever means are appropriate to the source material to gain a better understanding of this period, using textual criticism, archaeological analysis and literary studies as equal partners, working independently and comparing the results at frequent intervals. And we should perhaps also recall that the Roman writer Tertullian wrote about 100 years after the completion of the Agricola:
”Cornelius Tacitus, however, – who, to say the truth, is most loquacious in falsehood”.