Hearing from the Tabula Imperii Romani is like getting a letter from someone you thought had died. It will certainly surprise some to learn that the project is still alive, and indeed this seems to be only the third fascicle to have been produced in the last decade. So a new sheet inevitably raises the question whether the TIR still serves a purpose. And since this fascicle (hereafter K-32) is edited by the experienced topographer Paolo Sommella, who is also the president of the International Commission for the TIR, it should be a good test case. All the more so, perhaps, because there is in this terrain not only the coverage offered (on a smaller scale) by the Barrington Atlas of the Greek and Roman World; for much of it, there is also (on a larger scale) the Atlante dei siti archeologici della Toscana organized by Mario Torelli (1992), not to mention the ambitious Carta archeologica della provincia di Siena organized by Riccardo Francovich (seven volumes so far, from 1995 onwards) or other resources.
At least the organizers of this fascicle, true to the TIR’s tradition, have made it affordable. Quasar currently lists it at the remarkably low price of 32 euros, which means that while it is inevitably far less handsome then Barrington, penniless academics can afford it.
Some will complain that the publishers did not supply us with a CD, and it is astonishing that there is no indication that anyone considered digitizing the whole project, all the more so because K-32 is a repertory of information much more than it is an attempt to produce a real map of a piece of the Roman Empire. Friends of Italy will recognize this as part of a certain technological negligence within the academic world that has also left the country with, for instance, extremely few of its humanities or social science periodicals online. Barrington was to some extent overtaken by technology while it was being produced, but there was no excuse for bringing out an undigitized compendium of information of this kind in 2006.
What in fact do we have here? (1) A black-and-white map on the scale 1:1,000,000 running from 44 to 40 degrees N and from 6 degrees to 12 degrees E, i.e., a map that includes a part of Provence and most of Toscana (and some of Lazio too, for it reaches as far S as Pyrgi), all of Corsica and the northern half of Sardinia; (2) two black-and-white maps on the scale 1:250,000 covering between them 44 to 42 degrees N and from 10 to 12 degrees E, i.e. the same section of Toscana on a much larger scale (the editor says that these maps are 1:200,000, but that is a typo); (3) a topographical index of the larger area including 1020 entries.
The compiler of such a map is faced with numerous decisions about what to include and how to include it. A great strength of Barrington is that the editor justified the major choices, while each contributor was required to comment on the specific problems his or her map presented. The Siena project mentioned above discusses its own procedures at length. K-32 simply does things, without discussion or justificatory pleas.
The most interesting of the compiler’s decisions was to include a generous number of the names of late-antique massae that can be inferred from toponyms (making use of the work of the ottocento scholars Emanuele Repetti and Silvio Pieri), even when there is no other evidence (Samprugnano = Sempronianum, and more entertainingly San Prugnano in the Arno valley above Florence also = Sempronianum). I considered doing this in Barrington, but preferred to stick to the notion that among habitation sites only sizeable settlements and known upper-class villas should be included. The trouble with K-32’s decision is that it was put into effect in an apparently random way, for anyone who knows the territory or looks at a 1:250,000 map, not to mention a 1:25,000 map, is aware that there are vastly more such names than K-32 recognizes (hundreds of them).
The geographical pattern that emerges nonetheless gives comfort to those who suppose that the low-lying coastal areas were very unhealthy in late antiquity: from the River Cecina southwards there is no such name anywhere near the coast (Magliano is something of an exception). Yet this is only partially confirmed by the fifth-century and later evidence assembled by Annalisa Marzano, Roman Villas in Central Italy (Leiden, 2007). A still more complex implication concerns forest-coverage and deforestation: areas very short on archaeological finds also tend to lack massa names, and the most reasonable explanation of such ‘blank’ areas is that they remained heavily wooded for all or most of antiquity. This applies, for example, to a substantial wedge of territory NW of Volterra. One might think that it also applied to another area, S of Siena, beyond the ‘Masse di Siena’, in the upper valley of the Ombrone: but that is only because K-32 neglects the examples that are to be found there.
K-32 has a number of defects. The underlying maps are unaltered contemporary ones. Now, the region’s altered hydrography is often deeply problematic — what, for example, is the cartographer of the Roman Empire to do with the upper Val di Chiana? But we know for instance that the coastline all the way from Luni to Livorno was very different from what it is now, and we can make a reasonable guess at the ancient shape of the Lacus Prilius near Grosseto. K-32 is not in fact a map of ancient territory at all: it is map of Provence, Toscana, and so on, with archaeological finds approximately marked. Yet it would not have taken a huge effort, with Barrington already in existence, to have registered the known and probable physical changes.
The underlying 1:250,000 map is faint, so faint that the contours and even the coastline are largely invisible. Hence no one who wanted to judge the plausibility of the many Roman roads hypothesized here could possibly do so on the basis of these maps. One of their worst features is in fact the huge network of hypothetical and often unhistorical roads that the compiler has invented, almost entirely on the basis of itineraries that contradict each other or refer to much later conditions. The modest Roman town of Saena seems here to resemble Rome, so many roads lead to it. One such road runs NW across the hills roughly to where Poggibonsi lies now (not the site of an ancient town) and then down the Val d’Elsa to the Arno and thence on to Luca. Now there may well have been a Roman road of some sort running down the Val d’Elsa, but there is no evidence for it, at least in print. Such a route is, however, known from an itinerary of 1191 — when Siena was of course relatively far more important. Furthermore, according to K-32 at least, the Romans were not content with this road, but with a profligacy worthy of a late-industrial government built another road from Siena to the same point on the Arno; this ran across hilly country, was roughly parallel to the other road and never more than ten kilometres away from it. All this raises in acute form how we are to diagnose Roman roads, and about that K-32 has nothing to say.
Centuriation is another area of weakness. When we know of it, we know of it in detail from aerial photographs and large-scale maps: the TIR contents itself with a symbol next to the name of the capoluogo. Survey archaeology presents another problem: a number of areas have been surveyed, and might have hoped that K-32 would have found some way of indicating the fact if not all the results.
Barrington was criticized by some for putting 1400 years of history and archaeology into single maps (a choice that was unavoidable). But at least it attempted to distinguish periods. The TIR lumps Hellenistic, Roman and late-antique together indiscriminately — perhaps an unavoidable consequence of the decision to economize. And K-32 is by no means up-to-date. Its bibliography contains no item more recent than 2001, which means that it knows nothing of such important contributions as Annapaola Mosca’s monograph on the Via Cassia (2002). And it is reluctant to take on board recent discoveries. A glaring example is the whereabouts of the town of Statonia: an inscription published in 1994 makes it highly probable that it was located at Piammiano (= *Plaminianum?!) on the Tiber above Orte, but K-32 continues to locate it, with a question mark it is true, far away to the W at Castro, an old hypothesis for which the evidence was never strong (the compiler cites in favour of this opinion a string of scholars who wrote before the new inscription was published, and fails to cite an essential article published in 1995).
There is in fact a deeper kind of neglect. One will not be surprised to see the River Serchio running directly into the sea, contrary to the clear evidence of ancient writers that it flowed into the Arno at Pisa, because this has become the received doctrine. But to take a case that I know at first hand, the course of the Via Traiana Nova north of Bolsena/Volsinii (which cannot have been the site of Etruscan Volsinii — how often does this have to be repeated?): K-32 simply neglects the published research that it includes in its own bibliography.
K-32 therefore shows that the TIR needs to re-think its mission. Anyone who is at work on a new sheet must do it very differently.
It has been most unpleasant to write this review, for the well-regarded editor was faced with a formidable task. To end in a more positive fashion, I will record the whereabouts of an ancient monument that has often been mentioned in the literature, including K-32, and even in guide-books, but never as far as I know with clear enough directions that anyone could actually find it. When I was compiling the Pisae sheet of Barrington I spent a long futile morning looking for the city’s often-mentioned Roman aqueduct (not to be confused with a much later structure that is a prominent features of the countryside NE of the city). Ten years later, I learned its location — in a conversation at a Pisan dinner table where no archaeologists were present. A few days later a group of us went to see the six splendid arches that survive in a private garden. Take the strada regionale that leads out of San Giuliano Terme to the NNW. About 400 m. after the intersection with Via XX Settembre, if you look carefully on the left, you can see part of the aqueduct quite close to the road. The villa’s proprietress has a robust guard dog but is very hospitable.