MARSEILLE AND THE PIRENNE THESIS, I:

GREGORY OF TOURS, THE MEROVINGIAN KINGS,

AND “UN GRAND PORT” 1

S.T. Loseby

Introduction

Gregory of Tours, the late sixth-century Frankish historian and hagiog-

rapher, never went to Marseille, as far as we know. Indeed, if, as seems
likely, the manuscript reading which places the house of his mother Ar-
mentaria at Chalon-sur-Saône in Burgundy, and not at Cavaillon in
Provence, is to be preferred, then there is no reason to imagine that
the bishop of Tours even ventured into the Midi, let alone as far as the
shores of the Mediterranean.2 The Merovingian kings likewise seem to
have eschewed first-hand experience of the delights of Provence, that
milieu of sophisticated senators and philosophising judges, whose polite
society is shunned by Abbot Domnolus of Paris in one of Gregory’s
more obvious jokes.3 But if one could ask Gregory, or one of his
Merovingian kings, to think of a port, I suspect that, however unfamil-
iar it was to them personally, Marseille would still have sprung instinc-
tively to their minds. Carolingian rulers and their chroniclers show
scarcely any more interest in Provence than their Merovingian prede-
cessors, but their unfamiliarity with the south, I would suggest, is con-
ceptual as well as literal. While it is not entirely clear which port they

1
Earlier drafts of this paper, or of elements within it, were given in 1994 at confer-
ences and seminars in Edinburgh, Oxford, Leeds, London and Birmingham during my
tenure of a British Academy Postdoctoral Research Fellowship. I am grateful to those
present on these occasions and to Ruth Featherstone for their many helpful comments,
and to the Academy for its support.
2
Gregory of Tours, De virtutibus sancti Martini episcopi, III.60 ed. B. Krusch, Monumenta
Germaniae Historica Scriptores Rerum Merovingicarum I.2, (Hannover, 1885), pp. 584–661. Cf.
Idem., Gloria confessorum, 84 ed. B. Krusch, Monumenta Germaniae Historica Scriptores Rerum
Merovingicarum I.2, (Hanover, 1885), pp. 744–820). For discussion see R. Van Dam, Saints and
their miracles in late Antique Gaul (Princeton, 1993), p. 283, n. 93.
3
Gregory of Tours, Historiae, VI.9 ed. B. Krusch and W. Levison, Monumenta Germaniae
Historica Scriptores Rerum Merovingicarum I.1, (Hannover, 1937–51). As is often the case in
Gregory’s writings, the precise butt of this joke is nevertheless far from clear.
204 S.T. LOSEBY

would have named in response to our hypothetical question, it seems

unlikely that their answer would have been Marseille.
This, of course, is the contrast between the Merovingian and Car-
olingian periods which made early medieval Marseille an important el-
ement in the thinking of Henri Pirenne. In Mahomet et Charlemagne, the
most evolved (if still incomplete) formulation of his celebrated thesis,
Marseille recurs as a choice illustration for Pirenne’s version of the
transformation of the Roman world. In the first place, in support of
his contention of the persistence of the exchange-networks of Antiquity
well beyond the fall of the western Roman Empire, he assiduously as-
sembles documentary (and, in passing, numismatic) evidence, indica-
tive of trade via Marseille. Indeed, he employs a little licence to claim
for Marseille one or two additional anecdotes of an explicitly or im-
plicitly commercial flavour which are not specifically located there in
the primary sources.4 Even into the seventh century, he claims,
‘Marseille nous donne tout a fait l’impression d’un grand port’.5 Sec-
ondly, and in contrast, he asserts that such documentary evidence for
trade via Marseille comes abruptly to a halt after the first decades of
the eighth century, a silence which for him echoes the arrival of the
Arabs off the coast of Provence. By 720 Marseille was doomed, be-
cause the Mediterranean had become a barrier: ‘la mer nourricière
s’est fermée devant lui’.6 The economic geography of Frankia was
henceforth reorientated towards the north, and towards the heartlands
of Carolingian power.
Whilst remaining respectful of the breadth and ambition of
Pirenne’s vision, which has dominated subsequent discussions of early
medieval exchange-networks to a considerable, perhaps even an un-
healthy extent, historians have since been concerned to criticise and to
modify his thesis in various ways. The ammunition at their disposal has in
recent years been greatly reinforced by the availability of ever-in-
creasing quantities of archaeological evidence, a resource scarcely
available to Pirenne himself. One influential work of synthesis in par-
ticular – subtitled Archaeology and the Pirenne thesis - provides a succinct
and cogent statement of what I think still passes for current orthodoxy.7

4
H. Pirenne, Mahomet et Charlemagne, 3rd. ed. (Brussels, 1937). Examples can be
found on p. 80 (Syrians), and p. 98 (slaves). It is a measure of Pirenne’s influence upon
subsequent scholarship that some of his assumptions have since tended to acquire the
status of fact.
5
Pirenne, Mahomet et Charlemagne, p.77.
6
Pirenne, Mahomet et Charlemagne, p.168.
 MARSEILLE AND THE PIRENNE THESIS I 205

Long-distance Mediterranean exchange-networks did indeed survive

the fall of the Roman Empire in the West, but were nevertheless in
progressive decline, a decline which had reached terminal proportions
by the end of the sixth century. The effect of the rapid progress of the
Arabs around the southern shores of the Mediterranean was at most
to deliver the coup de grâce to an economic network which was already
as good as dead. However, both this and other recent works of synthe-
sis have paid no regard to the evidence from Marseille.8 This omission
is odd, firstly because of the abundance and interest of the archaeolog-
ical material recovered from the port in the last quarter of a century,
and secondly because his perception of the fluctuating fortunes of
Marseille is such a significant component of Pirenne’s analysis. For
whatever one chooses to make of Pirenne’s thesis, in one sense he was
right: Marseille should be a touchstone in any evaluation of post-Ro-
man Mediterranean trade. This paper seeks to reconsider Pirenne’s
first, positive contention about Marseille, that it continued to flourish
through the sixth century and into the seventh.9 His account and in-
terpretation of its subsequent demise will be left for reconsideration in
a second paper. Before turning to the new archaeological evidence
from Marseille, and then trying to tie it back into the other categories
of data, some general geographical background will help to explain its
peculiar significance.

The geographical context

Marseille is located some forty kilometres from the present-day mouth

of the Grand Rhône. The historic heart of the city is its sheltered, deep-
water harbour, known today as the Vieux-Port, a superb natural
anchorage which sufficed to meet the needs of commercial shipping
until as recently as the early nineteenth century. However, this mari-
time advantage is offset to landward by the city’s tiny and infertile in-

7
R. Hodges and D. Whitehouse, Mohammed, Charlemagne and the origins of Europe. Ar-
chaeology and the Pirenne thesis (London, 1983).
8
This is a failing common both to critiques of the Pirenne thesis and general discus-
sions of the ancient and medieval Mediterranean economy. A notable exception is K.
Randsborg, “Barbarians, classical antiquity, and the rise of western Europe”, Past and
Present 137 (1992), pp. 8–24.
9
For a more comprehensive presentation of the evidence discussed in this paper,
with extended and additional argumentation, see S.T. Loseby, Marseille in Late Antiquity
and the early Middle Ages (unpublished Oxford D.Phil thesis, 1993), especially chs. 4–5.
206 S.T. LOSEBY

terland. The exiguous basin in which Marseille lies is barricaded in be-

hind a continuous ring of hills, rising up barely ten kilometres or so
from the rear of the Vieux-Port. For this reason, noted the geographer
Strabo, writing in the early first century AD, the inhabitants of
Marseille have ‘been compelled to look to the sea rather than the land’
for their livelihoods.10 Marseille may be in Provence, but it is not really
of Provence: the inadequate integration of the city and the region re-
mains a problem today. Instead, Marseille looks away from Provence
to the far greater commercial hinterland opened up by its proximity to
the Rhône corridor, one of the very few, and perhaps the easiest, of
the natural north-south routes through the Pyrenees-Alps-Carpathians
mountain system of the European continent, and a vital artery of ex-
change and communications between the Mediterranean and northern
Europe. Marseille may be some distance from the nearest of the Rhône
mouths, but the hazards of the delta area, (low-lying, with poor visibil-
ity, sundry sandbanks and constantly shifting channels), have been such
as to prevent the establishment of any permanently viable port within
or closer to the corridor. Rivals to Marseille’s pre-eminence have pe-
riodically emerged, such as Arles in tandem with an outport at Fos in
the Roman period, or, in the eleventh century, St-Gilles.11 But over
the longue durée Marseille, which requires a minimum of infrastructural
support, and is not directly susceptible to the caprices of the river, has
seen them all off to remain the principal seaport for exchange via the
Rhône corridor.
The mediation of interregional traffic was the original idea behind
the foundation of Marseille by colonists from the Greek city-state of
Phocaea in Asia Minor in around 600 BC. The Phocaeans came late
to the vogue for colonisation and they had to venture further; Marseille
was the first Greek colony in southern Gaul, and would remain the
greatest.12 The site of Marseille is also characteristic, in its landward

10
Strabo, Geographica, IV.1.5, ed. and tr. H.L.Jones, 8 vols. (Cambridge, Ma., 1969–82).
11
For some brief comments on the viability of Arles as a seaport, cf. S.T. Loseby,
“Marseille: a late Antique success story?”, Journal of Roman Studies 82 (1992), pp. 165–85,
at 181–2, and, more generally Loseby, Marseille in Late Antiquity and the early Middle Ages;
S.T. Loseby, “Arles in Late Antiquity: Gallula Roma Arelas and urbs Genesii”, Towns in
transition: urban evolution in late Antiquity and the early middle ages, eds. N. Christie and S.T.
Loseby (Aldershot, 1996), pp. 45–70.
12
For Greek Marseille see most recently the many papers in Marseille grecque et la
Gaule (Études massaliètes, 3), eds. M. Bats, G. Bertucchi, G. Congès and H. Tréziny (Lattes/
Aix-en-Provence, 1992), with anterior bibliography.
 MARSEILLE AND THE PIRENNE THESIS I 207

isolation and maritime focus, of Phocaean colonisation, for the Pho-

caeans were, in Aristotle’s words, ‘practisers of emporia’.13 Marseille
duly satisfies the criteria used to define that category of settlement
known variously as an emporium, gateway community, or port of
trade.14 It will, for example, have generally been a provider of guaran-
teed landing, offloading and storage facilities; it has a large population
but a small political area; it is primarily concerned with wholesale rath-
er than retail trade. But most important of all, it is situated at a passage
point between two natural and cultural regions - in this case the Med-
iterranean and northern Europe – ‘separated not only by geography,
but also by their civilisations, such that products commonplace in the
one can assume considerable value in the other’.15 The latter quota-
tion, which is of course the guiding principle of long-distance trade, is
taken from a recent discussion of the emporial function of our Phocae-
an colony in the years after its foundation. I would argue that it applies
equally well to Marseille in the sixth and seventh centuries AD. There
is a tendency in early medieval economic history to see emporia as
phenomena of the northern seas16: Hamwic and Dublin, Hedeby and
Birka, Quentovic and Dorestad are the places which spring most readi-
ly to mind in this context. But down in the Mediterranean Marseille
was also a functioning emporium, at least in the Merovingian period.
Long-distance trade had been the raison d’être of Marseille from its foun-
dation and, devoid as it is of any significant hinterland, the port has
always depended, for its periods of greatest prosperity and significance,
on through traffic rather than the production or distribution of any lo-
cal agricultural surplus or on manufacturing industry. In the nine-
teenth century, for example, Marseille waxed fat on the French colo-
nisation of Africa, only to suffer subsequently during the long process
of disengagement. And in the age of Gregory of Tours, the ability of
Marseille to function as an emporium is an equally valuable pointer to
the vitality of contemporary interregional exchange. For the fortunes

13
Aristotle, quoted in Athenaeus of Naucratis, Deipnosophistae, XIII, 576a, ed. and tr.
C.B. Gulick, (London/New York, 1927–41).
14
K. Polanyi, “Ports of trade in early societies”, Journal of Economic History 23 (1963),
pp. 30–45.
15
J.-P. Morel, “Marseille, ville phocéenne”, Marseille 160 (1991), pp. 10–13, at p. 12
(my translation).
16
e.g. R. Hodges, Dark Age Economics: the origins of towns and trade AD 600–1000 (Lon•
don, 1982), pp. 47–86.
208 S.T. LOSEBY

of Marseille can indeed be used in precisely the way Pirenne used

them, as will become apparent in moving from context and theory to
practice.

The archaeological evidence

It is neither feasible nor appropriate for me to attempt here a compre-

hensive survey of the late antique and early medieval archaeological
data recovered from Marseille in recent years. Instead, I will merely
pick out one or two features of particular interest for summary consid-
eration. The first is a general point about late antique Marseille. The
vast majority of the many archaeological operations carried out across
the city over the last twenty years or so, latterly by the urban archae-
ological unit constituted in 1985, have, as so often in urban environ-
ments, been rescue excavations, limited in time, scope and resources,
and responding to an agenda dictated by developers rather than ar-
chaeologists.17 But the basic evidence of occupation recovered in these
excavations nevertheless reveals one significant feature of fifth- and
sixth-century Marseille. While many cities in Gaul were contracting
sharply in size in Late Antiquity, Marseille was not. There are various
indications that its Hellenistic wall-circuit, enclosing an area of around
50 ha (the precise course of its northern side remains uncertain), re-
mained the basis of the city’s late antique defences. The most obvious is
the fore-wall recovered in excavations at the Bourse site, clearly de-
vised to conform to and protect the Hellenistic defences at the north-
eastern corner of the port, and dated to the fifth century.18 The recent
discoveries of late antique occupation levels on sites right across the
area enclosed by these walls meanwhile suggests no significant reduc-
tion in the scale of intra-mural settlement before the end of the sixth
century.19 Indeed, the Bourse excavations have indicated some subur-

17
The results of these excavations are excellently presented in summary form in
Marseille, itinéraire d’une memoire: cinq années d’archéologie municipale, eds. L.-F. Gantès and
M. Moliner (Marseille, 1990), Musées de Marseille, Le temps des découvertes: Marseille, de Protis à
la reine Jeanne, (Marseille, 1993).
18
M. Bonifay, “Éléments d’évolution des céramiques de l’antiquité tardive à
Marseille d’après les fouilles de la Bourse (1980–1981)”, Revue archéologique de Narbonnaise 16
(1983), pp. 285–346, at pp. 287–9; P.–A. Février, “Aux origines de quelques villes médiévales du
Midi de la Gaule”, Rivista di studi liguri 49 (1983), pp. 316–35, at pp. 327–8.
19
Gantès and Moliner, Marseille, itinéraire d’une mémoire: cinq années d’archéologie munici-
pale, especially pp. 93–6; A. Hesnard, “Une nouvelle fouille du port de Marseille, place Jules-
Verne”, Comptes-rendues de l’academie des inscriptions et belles-lettres (1994), pp. 195–217.
 MARSEILLE AND THE PIRENNE THESIS I 209

ban expansion of fifth-seventh-century date around and ultimately

onto the shallow and silting watercourse outside the east gate of the
city, a suburb which the Ste-Barbe/Puget III excavations, 200 metres
north of the Corne, suggest may have extended outside the walls for
some distance away from the port.20 It would of course be premature
to push this evidence too far. The material recovered in the various
intra-mural excavations in particular has not yet been fully published,
and is likely in many cases to remain difficult to interpret in default of
opportunities for further or more extensive excavation. Nevertheless,
the apparent continuities of occupation and walled area, and the de-
velopment of a new extra-mural suburb of artisanal character contrast
sharply with the evidence from most cities in late antique Gaul of im-
ploding landscapes characterised by reduced enceintes enclosing frac-
tions of rapidly contracting settlements.21 The details may remain open
to debate. But in purely spatial and comparative terms post-Roman
Marseille was indeed the ‘grand port’ which Pirenne envisaged.
Alongside this general, if impressionistic, picture of the relative vital-
ity of Marseille in the fifth and sixth centuries, can be set the specific
evidence afforded by the one extraordinary site where the detailed and
extensive excavation of archaeological deposits of this period has been
possible: the Bourse. A full report of these excavations has yet to ap-
pear, but the sequence and the material found in one area in particu-
lar, the Corne, has been published in a series of excellent and expedi-
tious articles.22 The Corne had been the site of an inner harbour, cre-
ated out of a marshy creek at the north-east corner of the Vieux-Port,
and defined by a stone quay erected in the course of the first century
AD. However, this harbour had always been susceptible to silting be-
cause of the outfall of a nearby spring. By the third century elaborate

20
Bourse suburb: R. Guéry, “Le port antique de Marseille”, Marseille grecque et la
Gaule (Études massaliètes, 3) eds. M. Bats, G. Bertucchi, G. Congès and H. Tréziny (Lattes/
Aix-en-Provence, 1992, pp. 109–21. Puget III: Gantès and Moliner, Marseille, itinéraire
d’une mémoire: cinq années d’archéologie municipale, pp. 53–8.
21
The best overview of late Antique towns in Gaul is P-A. Février, “Vetera et nova: le poids
du passé, les germes de l’avenir”, Histoire de la France urbaine, I, ed. G. Duby (Paris, 1980), pp.
393–493.
22
M. Bonifay, “Éléments d’évolution des céramiques de l’antiquité tardive a
Marseille d’après les fouilles de la Bourse (1980–1981)”; idem, “Observations sur les am-
phores tardives a Marseille d’après les fouilles de la Bourse (1980–1984)”, Revue
archéologique de Narbonnaise 19 (1986), pp. 269–305; D. Foy and M. Bonifay, “Eléments
d’évolution des verreries de l’antiquité tardive a Marseille d’après les fouilles de la Bourse
(1980)”, Revue archéologique de Narbonnaise 17 (1984), pp. 289–308.
 210 S.T.LOSEBY

Fig. 1. Corne, main excavation: proportions of pottery by form and

origin. After Bonifay, “Ceramiques”, p. 304, fig. 13.

Fig. 2. Corne, main excavation: proportions of fine wares only. Proportions derived
from sherd counts. Only significant categories are shown. All periodisation is necessarily
approximate. After Bonifay, “Ceramiques”, p. 304, fig. 14.
 MARSEILLE AND THE PIRENNE THESIS I 211

Fig. 3. Corne, sondages: proportions of pottery by form and origin.

After Cavaillès-Llopis, “Ceramiques”, p. 171, fig. 8.

Fig. 4. Corne, sondages: proportions of fine wares only. Proportions derived from
sherd counts. Only significant categories are shown. All periodisation is necessarily
approximate. After Cavaillès-Llopis, “Ceramiques”, p. 171, fig. 8.
212 S.T. LOSEBY

efforts to contain this problem had broken down.23 The continuing ac-
cumulation of alluvial deposits and the tipping of rubbish thereafter
combined to create what the analysis of fauna and flora finds show to
have been a fetid dump, circled by swarms of flies, and picked over by
at least three species of vulture.24 But in the fifth and sixth centuries
this singularly noxious site went through at least two phases of redevel-
opment. The mud was dredged and new quays erected, all with the
clear intention of restoring it as a functioning watercourse. It was only
in the seventh century that the unequal battle against the silt was once
again abandoned, and the adjacent suburban occupation encroached
onto the alluvium.25 Again there are difficulties in putting this data in
context: the reasons for the revival of this inner harbour cannot be
clear without further knowledge of the state of similar harbour facilities
elsewhere around the port. Even so, it looks like a prima facie indication
of commercial vitality.
This impression is borne out by the ceramic assemblages recovered
from the well-stratified fifth-, sixth- and seventh-century deposits in
sampled areas of the Corne (Figs. 1–4).26 The fine ware pottery falls
into two main categories. One is the ware which goes by a variety of
names including dérivées des sigillées paleochrétiennes (DSP), grey/orange
stamped ware, and even, in older reports, the misleading Visigothic
ware. Variants of this ware were produced between the end of the
fourth and the seventh centuries at various centres in southern Gaul
from Aquitaine to Provence, very probably including Marseille.27 But
23
Guéry, “Le port antique de Marseille”; M. Euzennat, “Fouilles de la Bourse à
Marseille”, Comptes-rendus de l’academie des inscriptions et belles-lettres (1976), pp. 529–52, esp.
pp. 545–50.
24
L. Jourdan, La faune du site gallo-romain et paléochrétien de la Bourse, (Paris, 1976), esp.
summary at pp. 302–4.
25
Description, phasing, and dating in Bonifay, “Éléments d’évolution des
céramiques de l’antiquité tardive à Marseille d’après les fouilles de la Bourse (1980–
1981)”, esp. pp. 290–7, 303–26, with modifications in Bonifay, “Observations sur les am-
phores tardives a Marseille d’après les fouilles de la Bourse (1980–1984)”, pp. 270–1.
26
See previous note and for the smaller pottery sample obtained in a keyhole exca-
vation on the other side of the port, M.-T. Cavaillès-Llopis, “Céramiques de l’antiquité
tardive à Marseille”, Documents d’archéologie méridionale 9 (1986), pp. 167–95. I will make no
serious attempt here to compare these assemblages with other Mediterranean contexts;
for discussion along these lines see Loseby, Marseille in Late Antiquity and the early
Middle Ages.
27
J. Rigoir, “Les sigillées paléochrétiennes grises et orangées”, Gallia 26 (1967), pp. 177–244,
is the basic discussion. Among the many subsequent updates see e.g. J. and Y. Rigoir,
“Les derivées des sigillées dans la moitié sud de la France”, SFECAG, Actes du
Congrès de Reims (1985), pp. 49–56; “La céramique du haut moyen-âge en France
meridional: éléments comparatifs et essai d’interprétation”, La ceramica medievale nel
Mediterraneo occidental Siena/Faenza 1984 (Florence, 1986), pp. 27–54, at pp. 40–2.
 MARSEILLE AND THE PIRENNE THESIS I 213

while it is becoming increasingly apparent that some DSP production

reached export markets in Spain and Italy, it seems primarily to have
enjoyed an intra-regional distribution, unlike the main imported fine
ware found at the Bourse, African Red Slip (ARS), the dominant in-
ternational pottery of late Antiquity. In contrast, the main eastern fine
ware of mis period, Late Roman C, is represented in the Corne de-
posits until the early seventh century, but only by a few statistically in-
significant sherds.28 This dominance by ARS of the imported fine ware
assemblage at Marseille is typical of western Mediterranean sites.29
What is more unusual is its persistence. There is no sign of any decline
in the supply of ARS to Marseille before the period when Gregory of
Tours was writing, in the late sixth century. Indeed, the percentages of
ARS in the sampled levels from the Corne are generally rising, wheth-
er as a proportion of the whole pottery assemblage or of the fine wares
alone. Finally, the range of imported common wares represented in the
fifth-seventh century levels at the Corne is also worthy of notice, not
because they arrive in any significant quantity, but rather because they
come from regions of production scattered all over the Mediterra-
nean.30 These, like the few sherds of Late Roman C, hint at the en-
during economic integration of the Mediterranean and at the potential
diversity of Marseille’s commercial relationships in a way which the
other ceramic evidence does not.
The amphora data recovered from the Corne gives a similarly pos-
itive and internally consistent impression of the vitality of exchange via
Marseille in the post-Roman period (Fig. 5).31 While fine-ware pottery
is a traded commodity in its own right, amphorae were, of course, con-
tainers, generally for foodstuffs, and in particular for oil, wine, and var-
ious nutritious varieties of that favourite ancient condiment, fish-sauce.

28
Bonifay, “Èléments d’évolution des céramiques de l’antiquité tardive à Marseille
d’après les fouilles de la Bourse (1980–1981)”, p. 322, and more generally F. Mayet and
M. Picon, “Une sigillée phocéenne tardive (“Late Roman C ware”) et sa diffusion en Occident”,
Figlina 7 (1986), pp. 129–42.
29
Concise overview in S. Tortorella, “La ceramica fine da mensa africana del IV al
VII secolo d.C”, Società romana e impero tardoantico, III, le merci, gli insediamenti, ed. A.Giar-
dina, (Rome/Bari, 1986), pp. 211–26.
30
“Importations de céramiques communes mediterranéennes dans le midi de la
Gaule”, A cerâmica medieval no Mediterrâneo occidental, Lisbon 1987 (Lisbon, 1991), pp. 27–47.
Eighteen different types of imported common wares are attested at Marseille.
31
Bonifay, “Observations sur les amphores tardives à Marseille d’après les fouilles de la
Bourse (1980–1984)”. For an overview of late Antique amphora production and distribution see C.
Panella, “Le anfore tardoantiche: centri di produzione e mercati preferentiali”, Società romana e
impero tardoantico, III, pp. 251–84.
214 S.T. LOSEBY

Fig. 5. Corne, main excavation: percentages of amphorae by region of production. Derived from
Bonifay, “Amphores”, pp. 297–302–4, with minor modifications (for more detailed observations
on the amphora assemblage cf. Loseby, “Marseille”, p. 185).

Although amphora imports from Spanish and Italian sources are pe-
tering out in the fifth century, these containers were continuing to ar-
rive at Marseille in the sixth century from Africa and from various ar-
eas of the eastern Mediterranean from the Aegean to the Nile, for con-
venience lumped together here under a broad ‘eastern’ classification.32
Nevertheless, African amphorae were increasingly dominant, and by
the time of Gregory of Tours they enjoyed the lion’s share of the mar-
ket. This trend culminated in the virtual disappearance of eastern am-
phora imports from the latest, seventh-century levels at the Bourse, im-
plying some contraction, or at least change, in the pattern of exchange
via Marseille.
Now there are all sorts of difficulties with archaeological evidence of
this type, even with well-stratified, carefully - excavated and meticu-
lously - analysed assemblages such as these. It is not my intention here
to worry about questions of phasing or counting methods, but rather
to highlight the potential pitfalls of ostensibly coherent statistical evi-
dence. The first is obvious. The above percentages are indicative only
of relative and not absolute volumes of trade. African amphorae may, for
example, represent around one-fifth of the mid-fifth-century sample

32
For more detailed discussions of amphora types see M. Bonifay and F. Villedieu,
“Importations d’amphores orientales en Gaule (Ve-VIIe siècle)”, Recherches sur la céramique
byzantine, eds. V. Déroche and J.-M. Speiser (Paris, 1989), pp. 17–46; M. Bonifay and D.
Pieri, “Amphores du Ve au VIIe siècle à Marseille: nouvelles données sur la typologie et
le contenu”', Journal of Roman Archaeology 8 (1995), pp. 94–120.
 MARSEILLE AND THE PIRENNE THESIS I 215

compared with nine-tenths of its seventh-century equivalent, but this

tells us nothing reliable about the real quantities of African imports ar-
riving at Marseille in either period. A little less obvious perhaps, but
slightly more controllable, is the point that these figures measure con-
tainers and not commodities. Since the capacities of African amphorae
were far greater than those of their eastern equivalents, the relative vol-
ume of the imported goods therein will tend to be underestimated.
One approximate calculation has for example suggested that although
eastern containers predominate in the mid-fifth-century sample, the
quantities of African and eastern goods represented thereby, would be
roughly equivalent.33 A further complication is introduced by the like-
lihood that some indefinable proportion of these eastern amphorae are
reaching Marseille not via direct contact with their regions of produc-
tion but via intermediary ports, indeed via African ones. This unpleas-
ant possibility is implicit (in reverse) in Gregory of Tours’ account of a
Frankish embassy to Constantinople venturing via Carthage.34 It has
been given new force in recent years by the discovery of wrecks such
as those at Port-Cros in the Iles d’Hyères or in the Anse St-Gervais off
Fos, dating from the sixth and early seventh centuries respectively.35
The cargoes of these unfortunate vessels included both eastern and Af-
rican amphorae. Knowing roughly where an amphora is manufactured
is one thing. Postulating direct exchange between production centre
and findspot is another.
These problems of interpretation could be expanded upon or mul-
tiplied. But they pale into insignificance beside the more serious limi-
tations of the archaeological evidence. For the fundamental problem is
that most of the exchange which the documentary sources show to
have been carried on via Merovingian Marseille is what one might call
‘archaeologically challenged’, because either the commodities them-
selves, for example papyrus or slaves, or the containers in which they
were transported, such as boxes, skins or barrels, are generally irre-
trievable by excavation. In short, archaeology privileges exchange ei-
ther in, or of, ceramics, a problem which haunts any attempt to recon-
struct late antique exchange-networks on the basis of excavated evi-

33
Bonifay and Villedieu, “Importations d’amphores orientales en Gaule (Ve-VIIe siècle)”, pp.
37–9.
34
Gregory of Tours, Historiae X.2.
35
L. Long and G. Volpe, "Lo scavo del relitto tardoantico della Palud (Isola di Port-
Cros, Francia)”, Vetera Christianorum 31 (1994), pp. 211-33; Gallia Informations,
1987-88, i, 12.
216 S.T. LOSEBY

dence alone. In the case of Marseille this has two important conse-
quences. First, the archaeology only shows trade in one direction, from
the Mediterranean to Frankia. Gaul had stopped exporting goods in
amphorae as early as the fourth century; as previously stated, the var-
ious southern Gallic varieties of DSP fine ware, although produced
throughout late Antiquity, enjoyed a regional rather than an interna-
tional distribution. For knowledge of any reciprocal traffic via Marseille
we therefore have to rely on documentary sources which, here at least,
afford only limited help. Secondly, as the ARS workshops go out of
production in the course of the seventh century, Marseille (and indeed
Provence as a whole) is plunged into an archaeological void punctuated
only by the odd Carolingian chancel fragment or sherd of imported
pottery such as Forum Ware, and by local common ware manufactures
which cannot at present be dated with any precision. The disappear-
ance of good diagnostic pottery evidence in the course of the seventh
century reduces in particular the contribution which the archaeology
of Marseille can make to the testing of the second part of Pirenne’s
thesis. But it also hampers consideration of the latter part of its sup-
posed prosperity under the Merovingians. Our ceramic-led perception
of post-Roman patterns of exchange makes it tempting, because only
one side of its activity is archaeologically visible, to regard the ability
of Marseille to function as an emporium in this period as limited. More
specifically, it makes the disappearance of African ceramic imports
seem emblematic of the failing fortunes of Marseille. It must be admit-
ted that the latter is scarcely a positive sign. But it must also be remem-
bered that the archaeology can tell only part of the story.
Leaving this pessimistic litany of caveats and concerns to one side,
the archaeological evidence does provide some additional support for
Pirenne’s interpretation of Marseille in the Merovingian period as a
‘grand port’. Firstly and specifically, the statistics derived from the ar-
chaeological data can, for all their flaws, be used to illustrate something
more positive about exchange via Marseille in the time of Gregory of
Tours than the mere fact of its continuing existence. The best evidence
here lies not in the amphorae, analysis of which is beset by sundry vari-
ables, but in the pottery (Figs. 1–4). The steady rise in the proportions
of ARS in sampled levels at Marseille is significant because pottery is
a traded commodity in its own right, and in statistical terms this is
therefore a discrete set of data. Although we cannot determine the size
of the Marseille pottery market in any given period, the increasing
share of it captured by ARS imports shows not just the enduring via-
 MARSEILLE AND THE PIRENNE THESIS I 217

bility of this traffic, but also its growing popularity in relation to locally-
produced fine and common wares.36 This would seem to be an index
of the ready availability and affordability of ARS, and of its continuing
importation to Marseille in some quantity down to the period around
600. Secondly, and more generally, the archaeology of Marseille in this
period suggests that the port was insulated from the general trend
which saw most towns shrink considerably in extent, and provides one
potential explanation for this in the endurance of Marseille’s Mediter-
ranean trade-links, and especially those with Africa. For while the his-
tory of eastern amphora imports to Marseille is broadly consistent with
that suggested by the contemporary assemblages excavated elsewhere
around the western Mediterranean, the persistence into the seventh
century of African imports to Marseille seems less typical. The appar-
ent decline in African imports on Spanish and Italian sites after the
Byzantine reconquest of Africa in 533 has encouraged the interpreta-
tion of the latter episode as the catalyst for a reduction in African trade
with other regions of the western Mediterranean, in part because the
reimposition of the annona would have led to the requisitioning by the
state of much of the available agricultural surplus, bringing ‘Africa into
the eastern circuit as a peripheral rather than a central region’.37 But
against a general background of decline, the exceptional persistence of
imports to Marseille suggests that there was still some surplus African
agricultural production available for distribution to western consumers
and indeed that Frankia may have become a preferred market for such
African exports as were continuing in the late sixth and seventh cen-
turies.38 In short, the archaeology of Marseille suggests that Pirenne

36
Indeed, the steady deterioration in the quality of DSP makes it increasingly diffi-
cult in later levels to distinguish it from local common wares: Bonifay, “Éléments d’evo-
lution des céramiques de l’antiquité tardive à Marseille d’après les fouilles de la Bourse
(1980–1981)”, pp. 332–4; Cavaillès-Llopis, “Céramiques de l’antiquité tardive à
Marseille”, pp. 177–80, esp. figs. 15–17. For a general summary, “Importations de
céramiques communes mediterranéennes dans le midi de la Gaule”, p. 42, suggesting
evidence for the continuation of DSP manufacture into the seventh century is confined
to Marseille.
37
C. Wickham, “Marx, Sherlock Holmes, and late Roman commerce”, Journal of
Roman Studies 78 (1988), pp. 183–93, at p. 193. This is an especially sophisticated version of
a widely-held opinion, e.g. Hodges and Whitehouse, Mohammed, Charlemagne and the Or-
igins of Europe, pp.28–30; Tortorella, “La ceramica fine da mensa africana del IV al VII
secolo d.C”, pp.220–2; Panella, “Le anfore tardoantiche: centri di produzione e mercati
preferentiali”, pp. 263–5, and, more generally, C. Panella, “Le merci: produzioni, itiner-
ari, e destini”, in Società romana e impero tardoantico, III, pp. 431–59, especially pp. 456–7.
38
For some brief speculation as to the reasons, Loseby, “Marseille: a late Antique
success story?”, pp. 182–3.
218 S.T. LOSEBY

was right in the importance he attached to the port. But it also suggests
that Marseille may have been rather exceptional. If we now return to
the categories of evidence which Pirenne did have at his disposal, these
impressions can be further strengthened.

The written record

The crucial source for the commercial vitality of Merovingian

Marseille is Gregory of Tours. It must nevertheless be admitted that in
terms of direct evidence for the mediation of exchange via Marseille,
Gregory is not that much use. Indeed, on the only occasion when he
mentions the origins of a commercial vessel putting into the port – a
ship from Spain – he maddeningly neglects to identify what it was car-
rying, except to observe that it was ‘the usual cargo’, preferring to focus
his attention on the plague which the inhabitants of Marseille unwit-
tingly introduced into their homes with their purchases.39 And the spe-
cific (but unprovenanced) commodities which he does name in connec-
tion with Marseille number but three. Two of these, oil and an enig-
matic liquamen, perhaps some sort of fish-sauce, arrive at the port in a
single shipment, seventy jars of which are promptly stolen.40 The im-
port of the third emerges only in a jibe. Stung by an accusatory letter
from his eminent suffragan, Bishop Felix of Nantes, Gregory observes
that if only Felix were bishop of Marseille instead, then oil and other
(sadly unspecified) goods would no longer be shipped there, only pa-
pyrus, so as to give the libellous Felix more scope for his defamations
of the good, that is to say the likes of Gregory and his family.41 And
that is the sum total of Gregory’s specific references. To build up a list
of commodities exchanged via Merovingian Marseille we must have
resource to other documents. Like Gregory, Marculf’s Formulary and
the grants to the great northern Frankish monastic houses of St-Denis
and Corbie of privileges at the Provençal ports of Marseille and Fos
lay particular emphasis on the traffic in oil.42 However, the famous

39
Gregory of Tours, Historiae IX.22.
40
Gregory of Tours, Historiae IV.43.
41
Gregory of Tours, Historiae V.5.
42
Marculf, Formulae, Supplementum, 1, ed. K.Zeumer, Monumenta Germaniae Historica
Formulae I (Hannover, 1886), pp. 107–12. St-Denis: Gesta Dagoberti 18, ed. B. Krusch,
Monumenta Germaniae Historica Scriptores Rerum Merovingicarum II (Hannover, 1888), pp. 401–
25; Diplomata regum francorum e stirpe merowingica, nos. 61, 67, 82, ed. G.H. Pertz, Monumenta
Germaniae Historica Diplomata I (Hannover, 1872); Corbie: Diplomata regum francorum e stirpe
merowingica, no. 86.
 MARSEILLE AND THE PIRENNE THESIS I 219

Corbie diploma of 716 also highlights a cornucopia of other merchan-

dise available at Fos43, a veritable delicatessen of herbs and spices,
fruits and nuts, in quantities large and small, as well as Spanish skins,
papyrus and, of course, fish-sauce. Altogether this remarkable docu-
ment is indicative of the importation into Frankia of an astonishing
range of exotica from around the Mediterranean.44 The written sourc-
es, like the archaeology, are regrettably less informative about ex-
change via Marseille in the opposite direction, from the northern world
to the Mediterranean, although there are recurring indications of a
lively traffic in slaves.45
There exists, in short, a reasonable range of documentary evidence
for the long-distance movement of goods via Provence which, when it
is specific, tends to be centred upon Marseille.46 In matters of detail
the relationship between the written record and the archaeology re-
mains uneasy, although there is one noteworthy case where they dove-
tail with remarkable precision; the Gaza wine mentioned by Gregory
finds its archaeological counterpart in the increasing proportions of the
chestnut-coloured Carthage Late Roman Amphora type 4 present in
the late sixth-century assemblage at the Bourse, a type of amphora

43
For simplicity’s sake I am assuming here that the information pertaining to Fos
can be extrapolated to Marseille without serious risk of error.
44
Pirenne made much of the Corbie diploma in particular: Mahomet et Charlemagne,
pp. 89–91. My use of these privileges within the context of this paper is admittedly anach-
ronistic, since they fall in the period between the mid-seventh and early eighth centuries,
running from the reign of Dagobert I down to that of Chilperic II. But I think it is rea-
sonable to use the information they contain as the best illustration of the range of com-
modities likely to have been imported through the Provencal ports in the preceding cen-
tury. I will be returning in more detail to these privileges and the problems with their
interpretation in a later paper.
45
For Pirenne, Marseille was the great slave-market of the period: Mahomet et Char-
lemagne, p.99. See Vita Boniti, 3, ed. B. Krusch, Monumenta Germaniae Historica Scriptores Re-
rum Merovingicarum VI (Hannover, 1913), pp. 110–39; Vita Eligii, I.10 ed. B. Krusch, Mon-
umenta Germaniae Historica Scriptores Rerum Merovingicarum TV (Hannover, 1902), 634–741;
Gregory I, Registrum epistularum, VI. 10, ed. D. Norberg, 2 vols., Corpus christianorum series
latina, 140 and 140A (Turnhout, 1982). Stricdy speaking, the last two sources do not spec-
ify precisely where the events they describe took place, although Marseille is the most
likely location in both cases. I make no attempt here to piece together all the documen-
tary data for exchange via Marseille. For collections of the textual evidence for trade in
this period, see D. Schwärzel, Handel und Verkehr des Merowingerreiches nach den schriftlichen
Quellen (Marburg, 1983) and more generally D. Claude, Der Handel im westlichen Mittelmeer
während des Friihmittelalters (Untersuchungen zu Handel und Verkehr der vor- und frühgeschichtlichen
Zeit in Mittel- und Nordeuropa, 2) (Göttingen, 1991).
46
In this sense Pirenne’s occasional assumption that any commercial episode documented in
Provence must have taken place at Marseille, though misleading, is understandable.
220 S.T. LOSEBY

thought to have carried wine from the Gaza region. Indeed, its ship-
ment inland is confirmed by the presence of fragments of such ampho-
rae in sixth-century levels on the avenue A. Max site in Lyon, which,
very tidily, is precisely where Gregory records its regular donation to
the church of St.Mary for use in the sacrament.47 But this exceptional
coincidence between archaeological and historical data is offset by
much larger discrepancies. Spanish trade, for example, is mentioned
in the written sources, but leaves no trace in the archaeology; African
imports, on the other hand, dominate the archaeological record, but
are never explicitly documented. All this highlights once again the lim-
itations of both categories of evidence, and the necessity of assessing
the exchange networks of this period from an interdisciplinary stand-
point. This is not the place to discuss such dislocations in detail. But it
should be observed that although the two types of data do not neces-
sarily match, then the cumulative effect of the discrepancy as far as the
commercial vitality of Marseille is concerned is, even so, a positive one.
This effect is heightened if we return to the written sources, and in
particular to Gregory of Tours, in whose works it is not so much the
direct references to the trade of Marseille (the extent of which, as we
have seen, is somewhat disappointing) but the indirect indications of
the peculiar significance of Marseille within late sixth-century Frankia
which are the most telling. According to Raymond Van Dam ‘any dis-
cussion of Merovingian Gaul based primarily [as most are] on the writ-
ings of Gregory of Tours will resolutely reflect his own parochial
world’.48 If so, then Marseille was part of that world. Gregory may nev-
er have gone anywhere near Marseille, but he knew plenty of people
who had; his deacon Agiulf, his priest John, and his maternal uncle
Gundulf to name but three. Agiulf had been relic-collecting at Rome;
John went to Marseille to trade; Gundulf was the agent of a Frankish
king.49 Between them they exemplify the interconnected ways – travel,

47
For Gaza wine at the Bourse and elsewhere in Gaul, Bonifay and Villedieu, “Im-
portations d’amphores orientales en Gaule (Vc-VIIe siècle)”, pp. 20–21, 27–30, and for
the important Adolphe Max site in general, F. Villedieu, Lyon St-Jean: les fouilles de I’avenue
Adolphe Max (Documents d’archéologie en Rhône-Alpes, 3) (Lyon, 1990). Documentary
references: Gregory of Tours, Historiae VII, 29; Gloria Confessorum, 64. It needed a miracle
to stop a member of the Lyonnais clergy from replacing the offering of Gaza wine with
teeth- jarring vinegar, and reserving the former for his own enjoyment.
48
R. Van Dam, Leadership and community in late Antique Gaul (Berkeley, 1985), p. 201.
49
For Agiulf, Gregory of Tours, Historiae X. 1; Gloria Martyrum, 82, ed. B. Krusch,
Monumenta Germaniae Historica Scriptores Rerum Merovingicarum I.2 (Hannover, 1885), 484–
561; John: Gregory of Tours, Vita Patrum, VIII.6, ed. B. Krusch, Monumenta Germaniae
Historica Scriptores Rerum Merovingicarum I.2 (Hannover, 1885), 661–744; Gundulf: Gregory of
Tours, Historiae VI. 11.
 MARSEILLE AND THE PIRENNE THESIS I 221

trade, and high politics — in which Marseille was tied into the Merov-
ingian polity, and through which the port imposed itself upon the con-
sciousness of Gregory up in Tours.
In the writings of the bishop of Tours and in other documentary
sources of the late sixth and early seventh centuries, such as the letters
of his namesake Gregory the Great, the normal route for travellers be-
tween northern Europe and the Mediterranean is via Provence and the
Rhône corridor.50 In almost all specified cases their journey involved
taking ship to or from Marseille. It was the habitual port of passage for
a steady flow of ambassadors, churchmen, missionaries, pilgrims, even
the occasional pretender, whose movements reflect the dominant long-
distance trade axis of the day. And although he had never been on
such a journey himself, the dominance of that route is firmly fixed in
Gregory’s mind, for twice in the Histories, his mental map of the
Merovingian world finds instinctive expression at a level beyond the
mere description of actual events. The first example is in the afore-
mentioned jibe at Felix, when Marseille occurs to Gregory as the nat-
ural commercial context in which to set the bishop of Nantes’ imagined
need for endless reams of papyrus.51 The second occurs when, on its
return from Constantinople, a Frankish embassy achieves a spectacu-
larly unhappy landfall at Agde, and Gregory feels the need to explain
why it had not put in at Marseille.52 For that, it would appear, was the
normal thing to do. Seen from Gregory’s perspective at the heart of
Frankia, Marseille was the obvious Mediterranean port-of-call, wheth-
er for political travellers or commercial traffic. By comparison, the na-
scent trade-networks of the northern seas have not seriously impinged
themselves upon his consciousness.53 For Gregory, Marseille was the
Merovingian port. For him, as for Pirenne, the reorientation of the
commercial geography of Frankia from south to north has not yet tak-
en place.
Gregory’s world view is reflected in the attitudes of the Merovingian
kings. Like him, they hardly ever venture south in person. But they are

50
e.g. Gregory of Tours, Historiae VI.2, VI, 6; X.l; X.2; Gloria martyrum, 82;Gregory
I, Registrum epistularum, VI.52; IX.209; XI.41, etc.
51
Gregory of Tours, Historiae V.5.
52
Gregory of Tours, Historiae VI.2.
53
Fleeting references to “the son of a king of Kent”, Danish sea-raiders, and interest-
free loans to the merchants of Verdun appear to be as far as Gregory”s imagination or
interest stretches in this direction: Historiae IV.26; IX.26; III.3; III.34. But for the very
real activities of the Merovingians in the North Sea area see I.N. Wood, The Merovingian
North Sea (Alingsås, 1983).
222 S.T. LOSEBY

similarly aware of the enduring significance of Marseille. For the par-

ticular nature of their economic interests in the port we again need to
look beyond Gregory’s narrative, where the motivations behind their
actions are largely implicit. It is possible to identify two specific mech-
anisms established by the Merovingian kings in an effort to profit from
their main Mediterranean outport. The first emerges in the relatively
explicit information afforded by seventh-century sources. These show
Marseille and the nearby port of Fos to be at the head of a chain of
royal toll-stations extending up the Rhône corridor.54 Although it is
impossible to determine empirically the value of customs-dues to the
royal fisc, the rarity with which either the right to collect the tolls or
immunity from them was granted away, not to mention the apparent
enthusiasm of the Merovingians for increasing the number of locations
at which they might be levied, suggests that the operation of this system
was potentially lucrative.55 Those in charge of the toll-station at
Marseille, the telonarii, are variously described in grants to St-Denis as
actores regii, iudices publici, or viri inlustri.56 They appear to be men of
some status, directly answerable to the king, and also responsible for
the administration of a cellarium fisci attached to the toll-station. A sim-
ilar institution is attested at Fos in the Corbie diploma of 716.57
The charters show the cellaria at Marseille and Fos to be royal ware-
houses of some description. Their exact workings, nevertheless, remain
rather mysterious. They could be for the holding of goods in bond, for
the storage pending collection of goods on which due toll has already
been paid or remitted, or simply for goods purchased by royal agents
on behalf of the crown, or indeed, and most likely, a combination of
all three.58 But three things about the cellaria seem clear enough. Firstly,
they are attested only at the Provencal ports. Secondly, the quantities

54
Gesta Dagoberti, 18; cf. Marculf, Formulae Supplementum, I.
55
R. Kaiser, “Steuer und Zoll in der Merowingerzeit”, Francia 7 (1979), pp. 1–18.
56
Diplomata regum francorum e stirpe merowingica, nos. 61, 82.
57
Diplomata regum francorum e stirpe merowingica, no. 86. If the cataplus mentioned by
Gregory at Marseille in Historiae. IV.43 is identical with a cellarium, as seems probable,
then this institution can safely be pushed back into the sixth century.
58
H. Pirenne, “Le cellarium fisci: une institution économique des temps mérov-
ingiens”, reprinted in his collected essays, Histoire économique de l’occident médiéval (Paris,
1951), pp. 104–12. More convincing is F.-L. Ganshof, “Les bureaux de tonlieu de
Marseille et de Fos: contribution à I’histoire des institutions de la monarchic franque”,
Études historiques à la mémoire de N. Didier (Paris, 1960), pp. 125–33. The potential affinities
between these cellaria and the Byzantine apotheke are intriguing, though the precise oper-
ations of the latter are equally debatable: see M.F. Hendy, Studies in the Byzantine monetary
economy c.300–1450 (Cambridge, 1985), pp. 626–34, 654–62.
 MARSEILLE AND THE PIRENNE THESIS I 223

of merchandise specified in the charters – or at least the quantities of

oil – are such that even if they are taken with a substantial pinch of
salt, we still need to be thinking of a cellarium more in terms of a ware-
house than a lockable cupboard in the office of the actores regii from
which the occasional precious amphora of imported extra virgin olive
oil might be produced for the very fortunate. Thirdly, whatever their
precise workings, the cellaria seem indicative of a system of some so-
phistication, presided over by officials of rank, and intended to maxi-
mise both the exaction of tolls and the efficient provision of imported
goods to the crown and its chosen beneficiaries. All this implies that
either the tolls, or the commodities, or indeed both, were of sufficient
value to justify royal maintenance of such a mechanism of control.
The numismatic evidence from Marseille, referred to by Pirenne
only in passing, via the works of his friend Maurice Prou, tells a similar
story.59 Here again can be perceived the development by the Merov-
ingians of a particular mechanism of control and profit, centred upon
Marseille. In late sixth-century Frankia, the minting of an increasingly
debased gold coinage was generally left up to the local initiative of in-
dividual moneyers, and struck neither in the name of the Frankish
kings, nor of the eastern emperor.60 This is not the case at Marseille,
where the 570s saw the emergence of the distinctive gold coinage
known as ‘quasi-imperial’. This was struck in the name of the Byzan-
tine emperor, but conformed to the lighter Germanic gold standard.61
The simultaneous and die-linked appearance of this coinage in c.575
at Marseille and other mints on the lower and middle Rhône (princi-
pally Arles, Uzès and Viviers) argues strongly for its direction by a cen-
tral authority, an impression reinforced by its regularity of weight, the
absence of moneyers’ names, and the restriction of its minting to cities,
all exceptional for Frankish coins of this period. The subsequent history
of this coinage is dominated by and in all probability directed from
Marseille. From the reign of Chlothar II (613–29) onwards, this coinage

59
Pirenne, Mahomet et Charlemagne, pp. 108–9.
60
P. Grierson and M. Blackburn, Medieval European Coinage, I: the Early Middle Ages
(5th–10th centuries) (Cambridge, 1986), pp. 111–28.
61
S.E. Rigold, “An imperial coinage in southern Gaul in the sixth and seventh cen-
turies?”, Numismatic Chronicle, 6th ser. 14 (1954), pp. 93–133; J. Lafaurie, “Les monnaies
de Marseille du VIe au VIIIe siècle”, Bulletin de la Société Française de Numismatique 36 (1981),
pp. 68–73; Grierson and Blackburn, Medieval European Coinage, I: the early Middle Ages (5th–
10th centuries), pp. 128–31. But see especially M. Hendy, “From public to private: the
western barbarian coinages as a mirror of the disintegration of late Roman state struc-
tures”, Viator 19 (1988), pp. 29–78, at pp.68–70.
224 S.T. LOSEBY

is minted in the name of the Merovingian kings and not the Byzantine
emperor, but it is in all other respects identical, and, in the numismatic
context of seventh-century Frankia, equally anomalous.
Again, some fundamental things about the quasi-imperial and royal
coinages of late sixth- and seventh-century Marseille seem clear. In the
first place, they are of high value, and can only have been of practical
monetary use in major transactions.62 Secondly, they were intended for
use within northern Europe, as their clearly-marked Germanic weight
standard demonstrates. Confirmation that these solidi Gallicani were not
acceptable around the Mediterranean appears in a letter of Gregory the
Great, who in 595 demanded that rents from the papal estates in
Provence be paid in kind not coin, because ‘in our land Gallic solidi
cannot be spent’.63 The distribution pattern of these coinages corrob-
orates this further. It corresponds closely to the dendritic river-system of
France, and extends up into Frisia and across the Channel into
south-east England.64 Finally, the appearance and persistence, through
much of the seventh century, of the quasi-imperial and royal coinages
suggests that control of the lower Rhône valley money-supply was suf-
ficiently lucrative to make the issuing of a distinct and carefully-man-
aged coinage worthwhile here, when elsewhere in Frankia it was not.
But this last point begs the crucial question of by whom precisely it was
managed. The quasi-imperials have, with good reason, been described
as ‘the most puzzling group of coins struck in Merovingian Gaul’.65
The key to this puzzle lies in the identity of the central minting author-
ity, for on that must hinge the problem of why these coins should be
struck in the name of the emperor. It cannot be a civic authority, since
the coinage was introduced simultaneously at several centres. But nor

62
It is noteworthy in this respect that while the rest of Frankia minted only tremisses
(one-third of a solidus) in this period, the quasi-imperial and royal coinages retained the solidus as well as
the smaller denomination.
63
Gregory I, Registrum epistularum VI. 10. Gregory may have found this out the hard way: cf.
Registrum epistularum III.33, of two years earlier, where he fulsomely accepts payments from
Provence..
64
Lafaurie, “Les monnaies de Marseille du VI e au VIII e siècle”, p. 72, for a distribution
map of the Marseille gold coinages.
65
P.Grierson, “The patrimonium Petri in illis partibus and the pseudo-imperial coinage in
Frankish Gaul”, Revue belge de numismatique 105 (1959), pp. 95-111, at p.95. This is by far
the most serious attempt to put the coins in context. Its conclusions are nevertheless
retracted in Grierson and Blackburn, Medieval European Coinage, I: the Early Middle Ages
(5th-10th centuries), p.130; but cf. Hendy, “From public to private: the western barbarian
coinages as a mirror of the disintegration of late Roman state structures” , p. 70.
 MARSEILLE AND THE PIRENNE THESIS I 225

can it be a single royal authority since, in the divided Merovingian pol-

ity of the 570s, the main issuing centres pertain to different kingdoms.
The political division of the main mint sites has usually been seen
as the problem. Instead, I think it may provide a clue to the solution.
Happily, the first appearance of the quasi-imperials can be put into a
political context, since it occurs in precisely the period to which Gre-
gory of Tours devotes the bulk of his Histories. And here we return to
those indirect indications of Marseille’s significance which underpin
the mechanisms of control outlined above. In the Histories the recurrent
power-struggles in Marseille operate on two levels. The first is that of
local animosities and conflicts among the learned Provençal élite. The
intensity of their rivalries is evident, even if the reasons which underlie
them frequently are not. This is partly because in the Histories these
conflicts are articulated only within the wider context of the pan-
Frankish second level, that of the close and complex political and social
relations between the magnates of Marseille and their lords and mas-
ters at the royal courts.66 In Gregory’s narrative and to a lesser extent
in the poems of his friend and contemporary Venantius Fortunatus, the
patricius Dynamius, Bishop Theodore and the lesser actors on the com-
paratively well-lit stage of late sixth-century Marseille stand out not as
passive bystanders on the periphery of the Merovingian world, bending
this way and that with the winds of political change, but as men inex-
tricably embroiled (on different sides) in the high-court power-struggles
at the centre. Their peripheral location may enable them to scheme
with some success, but their freedom of action is limited, and its exer-
cise is risky: a duke is always likely to be sent in by the vigilant kings,
and neither patrician nor bishop can resist the central power for long.
The reach of the Merovingian kings comfortably extends as far as
Marseille. But Gregory’s narrative reveals more than that, as can be
seen if the high politics is divorced from the local dimension in which it
is embedded in the Histories.
In his peculiarly episodic way Gregory of Tours shows how, in the
internecine and often frankly bewildering game of contemporary
Merovingian power-politics, Marseille was a bargaining counter of sin-
gular importance. Splicing the episodes together, and excluding the lo-
cal dimension, the first element of note is the peculiar treatment of

66
The main episodes are Historiae IV.43; VI.7; VI.11, though minor but significant allusions
to some of the leading characters involved in them and to Marseille lurk throughout the Histories.
For detailed discussion see Loseby, Marseille in Late Antiquity and the early Middle Ages, ch. 5.3.
226 S.T. LOSEBY

Marseille in the divisio regni which followed the death of Chlothar I in

561. Although the bulk of Provence fell to Guntram, Marseille went to
Sigibert.67 Now, if the mechanics of the various partitions of the
Merovingian kingdom are difficult enough, the imperatives behind
them frequently defy explanation. But in the case of Marseille, a city-
territory of negligible size and less agricultural potential, the raison d’etre
of this peculiar arrangement must be sought in the profits to be made
from its trade. This is further brought out in the concession to Gun-
tram of half of Marseille which was made by the magnates of Sigibert’s
heir, the minor Childebert II, ‘on his father’s death’.68 This was clearly
the price of Guntram’s goodwill, sealed by a treaty at Pompierre in
577.69 It is equally clear, I think, that this refers not to any territorial
division of Marseille or of the lands dependent upon it, but to the fi-
nancial division of what makes the port of such capital importance to
the Frankish kings, namely the receipts from tolls on its trade. But in
581 a coup in Childebert’s kingdom sees a new faction of dominant
magnates reverse existing policy by abandoning the arrangement with
Guntram in favour of an alliance with Childebert’s other surviving roy-
al uncle, Chilperic.70 Childebert duly demands the conceded half of
Marseille back, highlighting how central it was to the original deal.71
When Guntram refuses, he strives to assert his control by force, but
with only temporary success. However, in 584, Gregory baldly reports
that Guntram restored Marseille wholly to Childebert, of his own ac-
cord; the explanation is to be found two chapters further back in Gre-
gory’s narrative, where Childebert’s ambassadors cite Guntram’s con-
trol over Marseille and his harbouring of fugitives from Childebert’s
kingdom as the reasons for their enmity.72 Not surprisingly, therefore,
the return of Marseille is soon followed by a renewal of the old alliance
between Childebert and Guntram.73 If Marseille is not among the cit-

67
The best (though I think slightly erroneous) discussion of the impact of this division
on Provence is E. Ewig, “Die fränkischen Teilungen und Teilreiche (511–613)”, reprinted
in E. Ewig, Spätantikes und Fränkisches Gallien (Zurich/Munich, 1976), pp. 114–71, at pp.
135–8. Sigibert’s dominions form the so-called “Austrasian corridor” discussed for exam-
ple in R. Buchner, Die Provence in merowingischer Zeit: Verfassung – Wirtschaft – Kultur (Stut-
tgart, 1933), pp. 10–11. But it is doubtful whether this division is best envisaged in terri-
torial terms.
68
Gregory of Tours, Historiae VI.11.
69
Gregory of Tours, Historiae V.17.
70
Gregory of Tours, Historiae VI.1; VI.3.
71
Gregory of Tours, Historiae VI.11.
72
Gregory of Tours, Historiae VI.33; VI.31.
73
Gregory of Tours, Historiae VI.4l.
 MARSEILLE AND THE PIRENNE THESIS I 227

ies mentioned in the text of the treaty which the two kings finally con-
cluded at Andelot in 58774, then this is because its return was of pri-
mary importance, a necessary precondition for any such agreement.
The close watch which the Merovingian kings exercised over
Marseille, and the central significance which the port held in their con-
flicts in the 580s, is explicable only in terms of the enduring commer-
cial vitality noted by Pirenne from the written sources, and now re-
emphasised by the archaeology. The institution of the cellarium fisci rep-
resents one attempt by the Merovingian kings to maximise their ex-
ploitation of this resource. To my mind, the emergence of the quasi-
imperial coinage in precisely the period described by Gregory of Tours
is another. The problem of responsibility for this coinage would be sim-
plified is we could admit the possibility of economic co-operation be-
tween kingdoms in tapping the traffic plying the lower Rhône valley
trade-route. Gregory’s narrative shows that the necessary political con-
ditions for such an arrangement – in the form of an alliance between
Guntram and Childebert – had indeed existed in the latter part of the
570s, the very period in which the quasi-imperials were introduced.75
The existence of such a compromise would then provide one possible
explanation for the decision to mint the coins in the name of the east-
ern emperor.76 And royal control from the outset would help to ac-
count for the seamless transition under the sole rule of Chlothar II
from the quasi-imperial to the royal coinage. No-one would doubt that
the royal coinage of the seventh century was controlled by the Merov-
ingian kings. The same surely goes for its quasi-imperial antecedent,
otherwise identical in every respect. In any event, the name on the
coins matters less than the destination of the profits, which, as the re-
peated impositions of Merovingian authority indicate, is always likely
to have been the crown. And if such a deal had been struck between
kingdoms, then no wonder Marseille was right at the heart of Merov-
ingian politics in the time of Gregory of Tours.

74
Gregory of Tours, Historiae IX.20.
75
These two kingdoms between them include all the mint-sites at which the quasi- imperial
coinages were struck.
76
This is a rather more specific version of the “compromise” solution suggested by
Lafaurie, “Les monnaies de Marseille du VI e au VIIIe siècle”, p. 71, and followed in
Grierson and Blackburn, Medieval European Coinage, I: the early Middle Ages (5th-10th centu-
ries), p. 130.
228 S.T. LOSEBY

Conclusion: ‘Un grand port’

The new archaeological evidence combines with the reconsideration of

the old documentary and numismatic data to suggest that Pirenne’s in-
stincts about the significance of Marseille in the post-Roman period,
and about the commercial basis of its vitality, were entirely justified.
Marseille flourished in the age of Gregory of Tours and on into the
seventh century because it dominated a still active trade-route between
Frankia and the Mediterranean, and because it could function effec-
tively as an emporium, the role for which it was founded, and which
over the longue durée has marked its periods of greatest importance and
prosperity. But Marseille cannot have been prospering in a commercial
vacuum. Its dynamism therefore suggests that, to misquote Mark
Twain, rumours of the death of Mediterranean trade before 600 have
been greatly exaggerated. And if we prefer to emphasise the complex-
ity of post-Roman Mediterranean exchange-networks, comprehension
of which ‘would require the mind of one of those chess-players who
can play fifty games at once while blindfolded’,77 this need not pre-
clude our recognition of the Merovingian kings as grandmasters, and
of Marseille as their preferred opening. The scale of Merovingian pow-
er and the sophistication of their administration has until recently been
underestimated by historians, their attention distracted by the lethal
combination of Gregory of Tours’ very individual historiographical
agenda and the dexterity of Carolingian propaganda.78 The Merovin-
gians, like all early medieval rulers, needed to maximise their exploita-
tion of available resources, and this dictated the close integration of
Marseille into the Merovingian polity and the development of partic-
ular mechanisms of control and profit centred on the port.79 None of
this makes sense unless there was something worth exploiting, and in
the case of Marseille that something can only have been trade: the ar-

77
C. Wickham, “L’ltalia e l’alto medioevo”, Archeologia Medievale XV (1988), pp. 105- 24,
at p. 111.
78
A much more positive view of the Merovingians is for example offered in I.N.
Wood, The Merovingian Kingdoms, 450–751 (London, 1994).
79
For a stimulating discussion of early medieval kingship and ‘economic policy’, see
J.R. Maddicott, “Trade, industry and the wealth of King Alfred”, Past and Present 123
(1989), pp. 3–51, with the subsequent critiques by R. Balzaretti and J.L. Nelson and re-
sponse by Maddicott in “Debate: trade, industry and the wealth of King Alfred”, Past
and Present, 135 (1992), pp. 142–188. Like the latter, I think that early medieval kings were
keen ‘to don the purple of commerce’ and would take vigorous steps to do so if they
considered the rewards sufficiently worth their while.
 MARSEILLE AND THE PIRENNE THESIS I 229

chaeological evidence now affirms that this was indeed the case. Gre-
gory of Tours may never have gone to Marseille, but he knew a great
deal, indeed perhaps all too much, about the port, its magnates, and
its importance in Merovingian politics. In his day Marseille was indeed
a ‘grand port’. Whether it would remain so for as long as Pirenne
imagined will be the subject of a second paper.